Pure Boron as a Ballistic Material – The LLNL Boron Mystery
I’ve mentioned the LLNL “Light Armor Program” several times here before. There are many interesting things about it, and it was hugely influential in pioneering the development of ceramic body armor. Yet, in keeping with their purely experimental approach and in light of the fact that they didn’t always have access to validated materials or today’s best analytical instruments, the program exhibited many oddities.
There was their share of unusual results – for instance, the very poor performance they attained with silicon carbide, which I might write about later.
They also tested materials that are highly unusual and even quite mysterious. Take a look at the table below.
The beryllium compounds are strange by definition. With their extreme toxicity and exorbitant cost, it should have been clear even then that they have no future. The titanium-beryllium intermetallic TiBe12 is especially strange, as it seems as though it should be a soft compound – it should be qualitatively different from ceramics in a number of important respects – yet it apparently performed very well.
I’m not sure what to make of SiN3 – a very nitrogen-rich silicon nitride – as I don’t believe there is any report of it existing as a bulk material. Si3N4 is common silicon nitride, but Si3N4 has a density of 3.17 gm/cc, which is quite far from the 2.815 gm/cc of the “SiN3” material apparently tested. (Which actually performed much worse than quartzite. This too is an unusual material in a military context, as it’s a natural stone – a form of polycrystalline SiO2 – that’s most commonly used in kitchen countertops.)
But no material raises more questions than the one in the box above; the only non-beryllium material to outperform boron carbide: “B” – pure boron.
Let’s jump straight to the conclusion: What they tested wasn’t pure boron.
Boron is very difficult to make
Near-pure boron is extraordinarily difficult to produce, so difficult that it took over 100 years of effort to isolate. Boron was recognized as a distinct chemical element by 1808, but near-pure samples, at roughly 99% purity, weren’t isolated until 1909. This is because its preparation is a multi-step process that’s very easy to botch, very difficult to get right.
Boron carbide, which isn’t a cheap material, is prepared in a pretty straightforward way:
2 B2O3 + 7 C → B4C + 6 CO
The ingredients – B2O3 and petroleum coke – are inexpensive, and most of the material’s high cost has to do with energy inputs and the work associated with milling ingots and sieving powder.
The most economical synthesis of pure elemental boron also starts with B2O3, which is reduced with magnesium at very high temperatures:
B2O3 + 3 Mg → 2 B + 3 MgO
Unlike the production of B4C, which releases CO as a gas, the reduction of B2O3 with magnesium leaves behind large volumes of solid MgO that need to be carefully removed.
And even after the MgO is taken care of, the magnesiothermic reduction of B2O3 invariably results in a low-quality amorphous boron powder with a very high impurity load. Almost every impurity you can imagine – leftover magnesium and oxygen impurities, as well as silicon, iron, and carbon impurities from the starting B2O3 powder and reactor vessel. All of these are present in various boron compounds, like MgB2, MgB12Si2, MgB2C2, Mg2B6, and many others.
Put it like this: Pure boron is very highly reactive and loves to pick up impurities wherever they can be found.
It’s not practically feasible to refine boron produced in this magnesiothermic manner past 97% purity, and today’s commercial-grade amorphous boron powders are typically only around 93-95% pure. Mg-B and B-C compounds can’t be separated from boron easily, and in many cases they can’t be separated at all.
There are various reports of an aluminothermic route to boron – this one, for instance – but it would doubtless produce great quantities of AlB12 and AlB12C2 as impurities in the final product, and they’d be effectively impossible to filter or wash out. The end result would be very far from pure boron. In fact, aluminothermic reduction might produce as much AlB12 as B.
Sintering Pure Boron
So absolutely pure boron powder is basically unobtainable, but today we can get pretty close. Even material at greater than 99% purity is available, if you’re willing to pay about $10,000 per pound for it. So, with this availability, we’ve learned quite a lot about the characteristics of boron – and there’s one in particular that casts tremendous doubt on the notion that the LLNL experiments utilized pure boron: High-purity boron is effectively unsinterable.
If you’re after a large elemental boron tile for a 6×6” hard armor side plate, there are no good options. You can try conventional sintering, but you’ll be left with a handful of boron dust. You can hot-press it however you like, or even stick it in an SPS furnace, but in most cases the result will be highly porous – with density no higher than 60-65% of theoretical. There are a few experiments that reported better results – including this quite successful one – but all of them were working with very small specimen sizes, very impure powders, HPHT, or all of the above. The one at the link was both small – 4mm thick and 20mm in diameter – and highly impure, and the authors themselves suggest that the good results they obtained can be attributed to a particular impurity. And, besides, their methods would not have been available to the LLNL’s researchers in the 1960s.
Something else worth mentioning is that if boron powders are hot-pressed in a graphite die, a lot of that boron is going to transform into boron carbide in situ. On the one hand, it’d be a shame to turn expensive near-pure boron into B4C by accident during sintering – on the other hand, it might be the only way to get a dense ceramic part.
What was the LLNL’s “B”?
Because we know that it’s effectively impossible today to produce large, fully-dense boron tiles at better than 99% purity – and certainly impossible in any cost-effective engineering context – we can absolutely rule out the notion that the LLNL researchers had tested a pure or near-pure form of polycrystalline boron, made via sintering or hot-pressing. And yet, at the time, there would have been no other way to produce high-quality bulk boron.
So what was the LLNL’s “B”? Giving them the benefit of the doubt, there are two possibilities:
- They tested highly impure B made via magnesiothermic reduction, which was a mixture of MgB2 + B + other magnesium-boron, oxygen-boron, and ternary compounds + boron carbide. This was likely hot-pressed in a graphite die and contained a relatively large volume fraction of boron carbide.
- They tested highly impure “B” made via aluminothermic reduction, which was a mixture of AlB12, AlB12C2, B4C, B6O, and a number of other aluminum-boron and aluminum-boron-carbon phases, with very small amounts of elemental B.
These complex multi-phase ceramics should exhibit low densities and good mechanical properties such as hardness, so it’s not too surprising that they managed to out-perform boron carbide. But they were not pure boron – in fact they were surely very far from pure boron – and I don’t believe that the ballistic characteristics of a pure or near-pure boron have ever been tested.
This all sort of implies that complex Mg-B-C-O and Al-B-C-O mixtures might result in high-performance armor ceramics, but quality control and batch-to-batch variance would present serious difficulties.
