Boron carbide

Boron carbide is a dark gray ceramic which possesses exceptionally good mechanical properties. Boron carbide is, in fact, the third-hardest industrial material, behind only diamond and cubic boron nitride – and whereas diamond and cBN are produced only as abrasive powders or small bits, large boron carbide parts with complex geometries are commercially available.

It must be noted that boron carbide exists in a broad range of potential molecular arrangements.  It can be anywhere from B4.3C on the carbon-rich extreme, to B14C on the boron-rich extreme.  “Ideal” boron carbide is B13C2, though the vast majority of commercial grades cluster around B4.3C to B4C, where the latter is B4.3C with excess carbon and carbonaceous inclusions.

The mechanical properties of boron carbide, across its various composition ranges and commercial variant types when necessary, are as follows:  

Density: 
2.52 gm/cc, pure B13C2 single-crystal
2.49 – 2.54 gm/cc, sintered or hot-pressed boron carbide ceramic.
2.47 – 2.49 gm/cc, boron-enriched boron carbide ceramic.
2.6 – 2.75 gm/cc, reaction-bonded B4C-Si cermet.

Hardness (Vickers):  

3400-3900 HK0.1 (microhardness) pure B4.3C single-crystal.
2400-3100 HV1, sintered or hot-pressed boron carbide ceramic.
2200-2900 HV1, boron-enriched boron carbide ceramic.
1450-1850 HV1, reaction-bonded B4C-Si cermet.

Fracture toughness:  2.5 – 5 MPa*m1/2

Compressive strength:  ~3000 MPa

Flexural strength:  315-340 MPa (three point bend, average), sintered or hot-pressed boron carbide ceramic.

Melting point: 2400°C ± 100°C

Sonic velocity:  14.035 km/s (L), ~8.6 km/s (T)

Note – Radiation Shielding:  Boron isotope 10B has a very broad neutron capture cross-section and the reaction between a 10B atom and a free neutron does not produce radioactive byproducts, so natural boron carbide, which contains about 20% 10B, and specially-produced 10B-enriched boron carbide, are widely used in nuclear shielding applications.

Note – Amorphization:  At very high pressures, boron carbide’s molecular structure collapses into a low-strength, brittle amorphous phase.  When articles of armor derived from boron carbide are tested in ballistic experiments against very high-velocity threats, or threats with very high sectional densities and compressive strengths, amorphous bands can be observed around impact sites.  This amorphization can result in anomalously poor performance.  For this reason, boron carbide is not utilized in armor rated for heavy AP or very high-velocity threats. 

History 

Boron carbide’s molecular composition has proven difficult to pin down.  It forms so readily when boron is reacted with any carbon source that its discovery was accidental – and it exists in such a wide range of possible stoichiometries, from the carbon-rich extreme of B4.3C to the boron-rich extreme of B14C, that the details of its structure took decades of investigation to ascertain. 

Boron carbide was first discovered in the mid-19th century by a loosely-affiliated network of European researchers – A. Joly, W. Hampe, F. Wohler, K. A. Kuhne, and H. St. Claire-Deville – who were attempting to purify and study elemental boron.  They reported that the reduction of boric anhydride by aluminum in the presence of carbon results in the synthesis of “(i) golden-yellow hexagonal plates of aluminium boride, AlB2; (ii) black plates of aluminium boride, AlB2; (iii) yellow quadratic crystals of adamantine lustre containing carbon and aluminium; and (iv) hard black crystals of a boron carbide, or more probably of several carbides, formed by the alteration of the preceding compounds at a high temp., in the presence of carbon and an excess of boric anhydride.”  

The yellow adamantine boron-carbon-aluminum compound of (iii) was assigned the ternary formula C2Al3B48.  (Boron and carbon regularly form ternary compounds with other elements.)  This was at one point considered an error, and it was assumed that it was regular boron carbide with certain aluminum impurities, but further 20th century investigations made it apparent that the initial assignment was probably correct, so the discovery of the first synthetic ternary boride dates to around 1870.  Likewise, the discovery of boron carbide – and the discovery that its composition varies – can be attributed to those mid-19th century researchers, though, to them, it was just a footnote on the difficult path to pure boron.  

The 19th century was a time when chemists had realized that the world was made up of just a few dozen building blocks, which they then set about trying to purify and characterize.  Fluorine, a well-known constituent of hydrofluoric acid, proved especially difficult to isolate.  The reactive, corrosive, and highly toxic nature of hydrogen fluoride had, by the latter half of the 19th century, killed or injured many famous scientists – including Sir Humphry Davy.  The first to isolate fluorine, and prepare compounds from fluorine gas, was French chemist Henri Moissan.  In 1886, after three years of heroic labor and repeated bouts of fluorine poisoning, Moissan successfully isolated pure elemental fluorine as a corrosive greenish-yellow gas.  He became quite famous on account of this work, won a respectable academic position (and later a Nobel prize,) and then set about preparing for his next great work: The synthesis of diamond.  

Moissan began by analyzing a number of samples of natural diamond and related stones.  (As an aside, in so doing he became the first person to identify natural gemlike silicon carbide, which today is named moissanite in his honor.)  It became apparent that elemental iron was diamond’s constant companion, and other elemental impurities – such as boron and silicon – were also commonly found around diamond deposits.  Moissan assumed that those metals might somehow catalyze diamond formation, and his first line of experimentation involved heating pure metals and metalloids alongside a carbon source, often sugar, in an electric arc furnace.  This never resulted in the production of synthetic diamonds, where high pressures turned out to be indispensable, but did result in the synthesis and characterization of various metal carbides.  Boron carbide was among them.  Synthesized in pure form by Moissan in 1899, it was assigned the formula B6C, which is not far from the “ideal” boron carbide composition of B13C2 or B6.5C.  

Boron carbide subsequently gained minor popularity as a high-performance material for abrasive and metallurgical applications, though it never approached silicon carbide’s industrial importance. 

In 1934, the Norton Company, a supplier of boron carbide ceramic powders and bonded parts, published the results of a research program into the mechanical and chemical properties of the boron carbide materials they were offering on the market.  They found that their grades of material all clustered around a composition of B4.15C, and “B4C” became their preferred nomenclature.  Though they were unaware of it at the time, B4.15C indicates that their material consisted of a solid solution of carbon in carbon-rich boron carbide.    

Not much has changed since 1934 in that regard.  Virtually all commercially-available boron carbide is at or near the carbon-rich extreme, and many grades are solutions of carbon in boron carbide.  This is not necessarily a bad thing, as will later be explained.  In any case, the assignment of “B4C” applied well back in 1934 and is still appropriate today.

The synthesis of boron carbide hasn’t changed much since 1934, either.  The Norton Company utilized the reduction of boric acid with carbon.  Boric acid is dehydrated to B2O3 in the first step of the process, followed by:

2 B2O3 + 7 C → B4C + 6 CO 

The end product is a large ingot of boron carbide, with very coarse grains and some residual porosity.  These ingots are typically milled or crushed in order to produce boron carbide powders, which results in irregular particle shapes, but generally a tight particle size range.  

Though simple, it should be readily apparent that enormous quantities of carbon monoxide gas are released as a byproduct of that synthesis from boric acid.  (2.3 cubic meters CO per kilogram of boron carbide.)  Further, this highly endothermic reaction typically occurs in an arc furnace at a temperature of roughly 2750°C, over many hours, and energy efficiency is low.   The process requires 16,800 kJ/mol boron carbide.  For these reasons, the European and American production of boron carbide is now all but finished.  

There are well-known alternative production methods, for e.g. via magnesiothermic reduction:

2 B2O3 + 6 Mg + C →  B4C + 6 MgO 

Via synthesis directly from the elements:

13 B + 2 C →  B13C2

Or via the reduction of boron trichloride in the presence of hydrogen.  

4 BCl3 + 6 H2 + C →  B4C + 12 HCl

These alternative methods are each attractive.  Magnesiothermic reduction is clean, its magnesium oxide byproduct is economically and industrially useful, and the reaction is highly exothermic and quite energy-efficient.  Synthesis from the elements results in a product with finely-tuned stoichiometry, so that the production of ideal B13C2 or various boron-enriched grades is possible.  The reduction of the trichloride is also fairly simple and attractive.  And yet: All of these methods have always been at least an order of magnitude more expensive than the reduction of boric acid or boron oxide with carbon, so they’re only employed on small scales, sometimes even in single gram quantities, for laboratory use.  As of this writing, none of them are apparently employed on industrial scales. 

Regardless of its method of production, boron carbide, unlike silicon carbide, is easily amenable to hot-pressing with the addition of a slight excess of carbon.  There were also “sintered” grades of material even as far back as the 1930s – they were reported in the Norton Company’s 1934 paper – though these were actually byproducts of reducing B2O3 with carbon above the melting point of B4C, and can be characterized by a coarse grain structure, a heterogeneous chemical composition, and considerable residual porosity, making these early “sintered” grades suitable for certain industrial applications that require temperature or abrasive wear resistance, but unsuitable for any application that might require impact resistance.  Boron carbide powders proved resistant to all other forms of pressureless sintering, so effectively all 20th century applications relied upon hot-pressed material.

Hot-pressing is associated with severe limitations.  As mentioned in our SiC monograph, it is a slow, expensive, intrinsically limited, and low-throughput procedure.  Nevertheless, to the present day, most high-performance boron carbide is made via hot-pressing in stacked dies.  

In 1967, Norton manufactured the first articles of boron carbide armor: Breastplates for helicopter crews.  (Alumina breastplates were already in limited use.)  In 1969, they supplied the army with boron carbide armor for helicopter seats.  Needless to say, both the breastplates and the helicopter seat armor were made of hot-pressed boron carbide.   

At around that time, from the late 1960s onwards, there have been efforts directed at “reaction bonding” boron carbide, in a process inspired by the reaction bonding of silicon carbide.  Results have been mixed.  Attempting to reaction bond boron carbide invariably results in a product with considerably lower hardness and higher density than boron carbide, and with a heterogenous and often complex chemistry, for there are ternary Si-B-C compounds and a wide variety of Si-B compounds that can form in situ, alongside SiC, B4C, and residual silicon.  For these reasons, combined with a cost that’s markedly higher than that of reaction bonded silicon carbide or sintered alumina, reaction bonded boron carbide has always been a niche material. 

Boron carbide’s “amorphization problem,” which will be covered at length later in this monograph, was discovered in the very late 1990s, and became a matter of considerable research interest in the years that followed.  Although the problem appears to be intrinsic to boron carbide and similar icosahedral boron compounds, research is ongoing into amorphization resistant grades that have been doped with other elements, or strengthened via secondary phases. 

The past several years have seen considerable progress in the pressureless sintering of boron carbide, particularly in Adept’s Ti-B4C ceramic. – an emerging material that combines the performance characteristics of hot-pressed grades with a price similar to that of reaction bonded B4C.

Boron Carbide Structure and Basic Mechanical Properties

Boron carbide is the prototypical icosahedral boron compound.  There are numerous other compounds in the same class – such as B6S, B6O, B6P, B6As, AlMgB14, B13N2, and B4Si, among others — which exhibit a very similar structure.  These materials all consist of a 12-atom icosahedra which occupy a rhombohedron’s points, and a short atom chain that connects those icosahedra which runs down the c axis of the rhombohedron.  

 

Compounds like B6P are highly regular and predictable – B6P always describes B12 icosahedra linked by two phosphorous atoms – and it was once assumed that boron carbide might be similar; in other words, it was assumed that “B4C” describes B12 icosahedra linked by C-C-C atomic chains.  As it turns out, this is not the case, B12(CCC) seems very rare, and boron carbide’s chemistry and structure are highly variable. This is because boron can easily substitute for carbon on the chain linking icosahedra, and, to a limited extent, carbon can substitute for boron in those icosahedra themselves.  So B12, B11C, and B10C2 icosahedra are theoretically possible, and these can be linked by CCC, CBC, CCB, CBB, BCB, or BBB chains.

 

To complicate matters further, the 11B isotope of boron and the 12C isotope of carbon have very similar electronic and nuclear properties, and analytical techniques cannot unambiguously distinguish between them.  So a complete determination of boron carbide’s preferred structures has not yet been performed.   

What has been determined, however tentatively, is that the carbon-rich stoichiometry corresponds to a configuration where the icosahedra are B11C and the chains are predominantly (~80%) C-B-C with a fraction (~20%) of C-B-B.  The “ideal” B4C formula of B12(CCC) – pure boron icosahedra linked with three-atom carbon chains – does not appear to arise, or arises only very rarely, for B11C variants are thermodynamically preferred.  

Commercial grades of boron carbide always cluster around the carbon-rich stoichiometry, so B11C(CBC) describes, to a fair approximation, “boron carbide.”  It appears to be the closest approximation of boron carbide’s molecular structure.  It is the generically most likely polytype. 

Less is known about boron-rich grades of boron carbide.  Although “ideal” B13C2 was thought to consist of B12 icosahedra and C-B-C chains, it is now generally supposed that B13C2 consists almost solely of B11C(CBB) units.  It is also supposed that the most boron-rich stoichiometry, at just 6.5% carbon, consists of B12(CBB) units, thus B14C.  This maximally B-rich boron carbide variant has a slightly lower density, at approximately 2.48 gm/cc.  It also appears to exhibit somewhat reduced mechanical properties.  

At this point a few things must be noted:  

First, though B11C(CBC) is apparently the most common polytype in most commercial grades of boron carbide, all boron carbide ceramics contain a wide variety of polytypes.  B11C(CBC) just happens to be predominant.  B10C2(CBC), B12(CCC), and other variants inevitably arise alongside it, to some extent.  There is no such thing as polytypically pure polycrystalline boron carbide.

Second, all commercial grades of boron carbide contain carbon as an impurity – either as a dispersed amorphous phase, or in the form of graphitic particles between boron carbide grains.  This has implications for the molecular and chemical analysis of boron carbide, which is how you can end up with “B4.15C.”  It also influences boron carbide’s dynamic material properties; glassy amorphous carbon and brittle graphitic inclusions can impair boron carbide’s impact strength and function as fracture nucleation sites.  This is, in small part, why boron carbide’s ballistic performance trails its mechanical properties.  

Third, those icosahedra-linking chains are frequently imperfect.  There are often vacancies, e.g. C-[ ]-C, where “[ ]” denotes a missing boron atom.  Two-atom C-C chains are also possible.  In ballistic applications, it is likely that a high concentration of such defects can also have performance implications.

Lastly, it remains difficult to pin down boron carbide’s molecular structure across its very wide stoichiometry range.  It must be emphasized that the above assertions constitute best guesses which, though current, may need to be updated as future investigations bring new information to light. 

Hardness

Boron carbide is an intrinsically hard material – and the hardest of the carbides by a significant margin – on account of its tightly-packed clusters of strongly-bonded atoms.  As a bulk ceramic material, however, boron carbide’s hardness is strongly dependent on its manner of production.  

Boron carbide single-crystals, and absolutely pure samples that have been produced via high-pressure high-temperature (HPHT) techniques, can attain hardness values in the 3000-3800 HV1 range.  This value approaches the hardness values of “superhard” materials like cubic boron nitride and polycrystalline diamond, and for this reason boron carbide is often said to be “the third hardest material after diamond and cubic boron nitride.”  (It is the third hardest common material, but there are a number of exotic compounds, like c-BC2N, that are harder.  Diamond aside, all known superhard compounds contain boron.)   

Conventionally hot-pressed and high-end sintered grades of boron carbide are markedly softer than samples produced via HPHT.  These conventional grades usually attain around 2600-2900 HV1 – and are thus comparable in hardness to conventional ceramic materials such as silicon carbide and tungsten carbide.   

Reaction-bonded boron carbide’s hardness is dominated by its residual silicon filler and is typically 1400-1600 HV1.  

Unlike silicon carbide, boron carbide does not have a variety of polytypes with varying mechanical properties – but it does have a wide range of potential compositions.  Generally, hardness exhibits little variance, but boron-rich compositions below about 13% carbon by weight are apparently about 10% softer than boron carbide with a more carbon-rich stoichiometry.  Whether or not this is a real effect, or merely due to issues in sintering and processing, remains to be seen.  It is known that the boron-rich grade is less dense and has a larger spacing between atoms, so it is very plausible that it has an intrinsically lower hardness. 

Impurity Content and Type

As previously mentioned, all boron carbide grades and powders contain free carbon as an impurity. Aside from carbon, boron carbide’s impurity load is generally low.  Typical commercial grades of boron carbide powder contain up to 0.3% N and O, and up to 0.2% in total metal content – the majority of which are usually iron, silicon, aluminum, and titanium, in descending order of likelihood.  

Hot-pressed boron carbide always contains an excess of C, which is used as a sintering aid.  Until about 1990, carbon was added to boron carbide powder mixes as Novolac resin; since 1990, carbon black has become the more popular free carbon source.  Regardless, the “B4C” designation might be absolutely accurate in the description of hot-pressed boron carbide.

Most pressureless sintered grades, what few there are, use light metals to assist in boron carbide densification.  Aluminum, titanium, and silicon are all popular choices.  These metals liquify during sintering and react to form borides, carbides, and ternary metal-B-C compounds in situ, which can help to cement boron carbide grains in place.

Unlike silicon carbide, which in absolutely pure form is colorless and transparent, pure boron carbide is black, opaque, and reflective.  Its powders, moreover, are always dark gray or black.  As a general rule, it is impossible to determine anything about boron carbide from its color. 

Thermal Properties

Boron carbide is highly refractory and can be used in parts that operate at very high temperatures, so long as those parts operate in inert atmospheres or vacuum.  Under such conditions, boron carbide will retain its hardness even at 1500°C – it is highly resistant to thermal softening.  

Boron carbide’s melting point depends on its stoichiometry.  Pure B13C2 boron carbide single crystals melt at 2447 °C ± 50°C.  Carbon-rich boron carbide and commercial grades of boron carbide have a slightly lower melting point, at roughly 2350°C.  Boron-rich boron carbide also melts at 2350°C.  If boron-rich boron carbide is produced with an excess of B (e.g., B14C+B,) it has an even lower melting point, at approximately 2250°C.  

Though boron carbide has excellent hot hardness and is highly refractory with a melting point much higher than that of alumina, silica, and many other heat-resistant ceramics in common use, it is generally not considered an industrially-useful refractory material.  This is not because of its relatively high cost, it is because boron carbide oxidizes readily in air.   

Though minor oxidation can occur at temperatures as low as 250°C, considerable oxidation starts at approximately 700°C and forms a liquid B2O3 surface layer.  The reaction governing this phenomenon is:

  B4C + 4 O2 → 2 B2O3 + CO2

B2O3 is considerably more volatile and more permeable to oxygen than silica, and has a lower viscosity by approximately an order of magnitude.  B2O3 is also unstable in the presence of water.  If exposed to moisture or atmospheric humidity at an elevated temperature, the B2O3 layer will swell, crack, and spall due to hydrolysis.  For these reasons, boron carbide parts are not operated at elevated temperatures in air.

Polycrystalline boron carbide’s thermal conductivity is approximately λ = 36 W/(m∙К), which is considerably lower than that of polycrystalline SiC.  Its rate of thermal expansion is almost identical to that of SiC.  Because boron carbide is not in use as an electronic material, nor as a heatsink, these properties are of little importance. 

Electronic and Nuclear Properties

In contrast with silicon carbide, the electronic properties of boron carbide are of minor importance.  Boron carbide’s electrical conductivity varies with its boron to carbon ratio and with its free carbon impurity profile.  Though boron carbide is a p-type semiconductor with some potentially attractive electronic properties – including generally low electrical resistance and high thermoelectric powder – industrially-produced boron carbide is presently unsuitable for use as a semiconductor, for not only is very high purity all but impossible to attain, it is also all but impossible to ensure a homogeneous molecular structure and stoichiometry.

Boron carbide does, however, see widespread use in nuclear shielding.  Boron has two stable isotopes, 10B and 11B, and the 10B isotope has a very high thermal neutron absorption cross-section.  (10B: 3837 barn. 11B: 0.005 barn.)  What this means is that 10B atoms efficiently  capture and react with free neutrons.  This reaction between a free neutron and a 10B atom produces non-radioactive helium and lithium.  Only cadmium and some of the rare-earth metals, such as 149Sm and 157Gd, have higher thermal neutron absorption cross-sections than 10B – and when these heavier elements react with free thermal neutrons, they produce unstable radioisotopes.  Moreover, 10B is capable of absorbing neutrons across a broad energy range, in contrast with cadmium isotopes, which aren’t typically used in very high-energy shielding applications. 

Boron carbide produced from natural boron, containing about 20% 10B, has a neutron capture cross-section of approximately 500 barn.  This is adequate for use in a wide variety of shielding applications, particularly in thermal reactors.  When improved performance is required, 10B-enriched boron carbide can be prepared.  10B4C has a density of around 2.25 gm/cc and similar or superior mechanical properties to industrial-grade boron carbide – which, in principle, makes it an extremely attractive armor material, but for its extreme cost.  

Boron carbide in neutron shielding applications is either sintered into a dense ceramic part, crudely sintered into a porous sponge, or used in densely-packed powder form.  Sometimes it is added to plastic or aluminum matrixes; polyethylene, in particular, contains a lot of hydrogen which is useful as a primary radiation shield, and the addition of boron helps to contain secondaries.  

Though pure boron might be a slightly more effective radiation shielding material on a weight basis than boron carbide, it is more reactive and roughly an order of magnitude more expensive than boron carbide, so it is not frequently utilized. Hexagonal (graphitic) boron nitride and boron oxides are occasionally used in place of boron carbide, but are generally much inferior on a weight basis, so they are only used in applications where their chemical or surface properties render them strongly preferable to boron carbide.

Chemical Stability

Boron carbide is highly stable to attack from acids and alkali.  It does not react to any substantial extent with acids such as HCl, HNO3, or H2SO4, and is not dissolved by aqueous alkali.  It may decompose in the presence of hydrofluoric acid, though generally at a low rate.  

Boron carbide is, however, broadly reactive at elevated temperatures.  It begins to oxidize in air at temperatures as low as 250°C, and degradation is considerable past 800°C.  It also reacts with metals and metal oxides at elevated temperatures; for example, if boron carbide and iron are in contact at 1000°C, they will react, releasing carbon monoxide and producing iron boride (Fe2B/FeB).  Boron carbide is also highly reactive with halogens at temperatures of roughly 700°C.  Reactivity with nitrogen begins at 1800°C. 

Flexural Strength

Sintered or hot-pressed polycrystalline boron carbide has an average flexural strength (three-point bend, ASTM C 1181) of approximately 315-340 MPa.  There is considerable variance here – all boron carbide materials will exhibit “long tails” in these tests, with some samples over-performing and exhibiting bending strengths of 550 MPa or more, and some samples under-performing the average by a considerable margin.  This data scatter is largely on account of the fact that flexural strength in polycrystalline boron carbide is dominated by the presence of residual porosity and carbon inclusions.   

Reaction-bonded boron carbide generally has a flexural strength about 20% lower – roughly 250 MPa on average.  

As of this writing, the flexural strength of single-crystal boron carbide has yet to be determined.  This is apparently because high-quality boron carbide single-crystals are extremely difficult to prepare.

Sonic Velocity and Acoustic Impedance

Boron carbide has an extremely high sonic velocity, with hot-pressed samples measuring at 14.035 km/s in the longitudinal direction and 8.9 km/s in the shear or transverse direction.  Thus boron carbide edges out all other technical ceramics.  SiC, TiB2, Al2O3, AlN, and ZrO2 all have significantly lower acoustic velocities than boron carbide – though, with that said, SiC and TiB2 have greater acoustic impedance on account of their higher densities.   

It is, at present, unknown how stoichiometry, carbon inclusions, and sintering method affect this property.  

Surface Chemistry

Boron carbide ceramic parts and powders usually have a surface layer of B2O3.  This would make it much like SiC, which itself always has a SiO2 surface layer, but for certain key differences: The B2O3 layer in boron carbide is slower to form, weaker, and intrinsically less stable.  It’s possible to produce boron carbide powders and parts that do not have an oxide surface, and it’s trivial to remove an oxide surface on boron carbide via mechanical or chemical means.  Thus if a surface oxide layer is desired, it can be produced; if a surface oxide layer is unwanted, it can easily be removed, e.g. by grit blasting at a low temperature.  The generically most probable surface on any given sample of boron carbide is a B2O3 layer a few micrometers deep, but that oxide surface layer can’t be taken for granted.

Boron Carbide Sintering 

The production of dense boron carbide parts has been briefly reviewed in prior sections. Here we will review it in some more detail. 

Boron carbide production via carbothermic reduction – 2 B2O3 + 7 C → B4C + 6 CO – doesn’t usually result in powder.  It most frequently results in the production of dense ingots, and it’s possible to synthesize boron carbide in a guided way so that ingots form in complex shapes such as domes and cones.  These are of poor quality; they universally feature large carbon inclusions, large grains, and internal porosity.  Nevertheless, parts made from as-synthesized boron carbide ingots were used in the early part of the 20th century, especially in parts where abrasion resistance is of especial importance, e.g. in nozzles for grit blasting.  

Hot-pressing is also an old method of boron carbide production, and one that has hardly changed in 100 years.  This involves mixing boron carbide powders with a carbonaceous resin or pure carbon source like graphite or carbon black.  That mixed powder, or a preform derived from that mixed powder, is then inserted into a graphite die, where pressures of 10-50 MPa at temperatures of >2000°C are subsequently applied.  This is always done in an inert atmosphere, or vacuum, to prevent the oxidation of the B4C and graphite die.  Boron-rich boron carbides cannot be produced in a graphite die due to carburization, so they can only be hot-pressed in dies made of boron nitride or graphite dies lined with expensive BN foils.

In the late 1970s, shortly after pressureless sintering was established as a production method for silicon carbide ceramic parts, it was discovered that boron carbide can also be pressureless sintered with the addition of a large amount of excess carbon.  The process, which was described in the 1977 US Patent 4195066, involves mixing carbon-rich boron carbide powders with free carbon or a carbon-based resin, pressing the mixture into a preform at roughly 70% theoretical density, and sintering that preform in an inert atmosphere at 2100-2200°C – very close to the melting point of boron carbide.  The end result is roughly 95% boron carbide and 5% free carbon.  

The mechanical strength, thermal properties, and abrasion resistance of these early pressureless sintered grades are good, but their impact characteristics are suboptimal.  For this reason, boron carbide intended for armor applications has typically been hot-pressed.  

New grades of pressureless sintered boron carbide are challenging that old paradigm.  Novel metallic, metal boride, and metal carbide sintering aids can now – since about 2010 – produce boron carbide ceramics that can be pressureless sintered without much excess carbon, and these new pressureless sintered grades exhibit ballistic performance comparable to hot-pressed grades.  They are gradually seeing increased use in high-end body armor systems.  Different sintering aids aside, the production of these new-type pressureless sintered boron carbide ceramics is much along the lines of the 1977 patent. 

Boron carbide is also amenable to sintering via SPS, flash sintering, and other new techniques.  Production of boron carbide via these methods does not take place on large industrial scales, but rather on lab scales for the production of highly pure polycrystalline samples. 

Boron Carbide Amorphization

That boron carbide performs anomalously poorly – that is, far more poorly than its mechanical properties would otherwise suggest – against high-velocity or high-strength (e.g. cemented carbide cored projectiles,) has been known for decades. But, just about two decades ago, it was shown that this is on account of the fact that boron carbide’s molecular structure simply breaks down at extremely high pressures.  In this, it is unique among ceramic armor materials. 

When boron carbide is subjected to a high-pressure event, such as a very high-velocity ballistic impact, amorphous bands of a boron-carbon glass will form around the site of the impact.  As this amorphous material is far weaker than boron carbide, it will fail prematurely and abruptly, in a brittle, glass-like manner.  

The pressures required to initiate this effect are generally beyond what steel-core small arms AP threats can generate upon impact, so B4C retains its ordered structure (which is very strong) and performs well against steel.  Projectiles with WC-Co cores usually take boron carbide past the breakdown pressure threshold.  Similarly, virtually any impact at 5500-6000+ fps will shock boron carbide into some degree of structural breakdown.  (Which may sound irrelevant but is very important to designers of vehicular armor.) 

Various strategies towards an amorphization-resistant grade, such as silicon doping and boron enrichment, have been proposed in recent years, but don’t appear to work very well.  They may raise the pressure threshold (this has yet to be meaningfully proven in live-fire experiments,) but B4C appears to remain susceptible to structural breakdown.

Furthermore, just about every compound that’s structurally similar to boron carbide — B6S, B6O, B6P, AlMgB14, and others — seems to suffer from the same problem.  Perhaps not to the same extent; the threshold pressure may be higher (B6P) or seemingly lower (AlMgB14).  It does seem, however, that this problem is intrinsic to the structure of icosahedral boron compounds, and that there’s no easy fix.   

For this reason, boron carbide is recommended for armor rated to stop steel-cored projectiles, but not for armor built to withstand faster threats or those with WC-Co cores.