Analog SFF, October 2007

by Dell Magazine Authors

  One thing that made the Texas City explosion worse was that there was additional organic matter present that could be oxidized. The AN was in a mixture that included rosin and paraffin as mechanical binders. Note (Rxn. 1) that there is oxygen left over when AN decomposes completely, and so if there's other material present to oxidize, even more energy is released.

  This is the basis for ammonium nitrate—fuel oil (ANFO) explosive: Mix a little fuel oil with the AN and you get a much bigger bang. ANFO among the general public has a horrible reputation because of its use in the 1995 terrorist bombing in Oklahoma City—but as all Analog readers know, technology is morally neutral. The same explosive that causes a terrorist atrocity can also dig a mine or a canal.

  In fact, ANFO is a very useful, cheap, and safe explosive. It is used by the metric ton in mining these days. Nevada gold mines in particular, which are routinely recovering gold present in the ore at parts-per-million levels, have to bust up lots of rock. You mix the ANFO up in real time, in the blast hole on site just before it's used.

  Another, similar compound is ammonium perchlorate, NH[4]ClO[4] (hereafter “AP"). It's also pretty inert at ordinary temperatures but contains a lot of stored energy, even more than AN:

  (Rxn 2) 4 NH[4]ClO[4] (right arrow) 2 N[2] + 4 HCl + 6 H[2]O + 5 O[2] + 658.4 kJ/mo

  Oxygen and chlorine have even less desire to stay bonded than do oxygen and nitrogen.

  For this reason AP is the main ingredient in the fuel for the Shuttle's SRBs. AP decomposition yields even more extra oxygen than does AN (Rxns 1 vs. 2), so you can get extra thrust if you mix in something for the oxygen to oxidize. The SRBs use aluminum metal and organic binders.

  Also as with AN, AP's endothermicity has caused some disasters. Most spectacular was the 1988 explosion, which killed two people, at an AP manufacturing plant in Henderson, Nevada, just outside Las Vegas. It was triggered by an accidental fire started by welders. Some eight million pounds of AP ultimately went blooie. The plant was making AP for the Shuttle SRBs under NASA contract, but since NASA was still paying for AP production but was not taking delivery(!), they were stockpiling the AP. (Yeah, government contracts and all that ... )

  Compounds like AN and AP obviously are a big improvement over saltpeter, but it's still a nuisance to add additional fuel materials to make a propellant “cocktail.” What about modifying such molecules so that they are “self-contained” with fuel? One way is with “alkylammonium” ions. Basically, you can swap one or more of the hydrogen atoms on the ammonium ion with a short hydrocarbon chain. Even the simplest, though—replacing one of the hydrogens with a methyl (CH3) group—is a bit too much of a good thing. Now there's extra fuel rather than extra oxygen:

  2 NCH[3]H[3]ClO[4] (right arrow) N[2] + 2 HCl + 5 H[2]O + CO + CO[2]

  Still, the principle is clear. Why hasn't it been used? Presumably because alkylammonium salts are expensive, due to their less-than-straightforward syntheses. There are many steps and lots of (unwanted) byproducts that lead to waste and further purification issues.

  So let's talk about synthesis issues for a minute.

  * * * *

  On Shaking and Baking and Synthesis

  A few years back I saw a paper trumpeting the synthesis of some intricate organic compound heretofore known only in an obscure organism—a lichen, as I recall. The synthesis involved about twenty steps, with an utterly abysmal ratio of finished product to input raw materials—the “yield,” as organic chemists call it. And even so, the synthesis still required other organic reagents as raw materials.

  My reaction at the time was, “Aren't we missing something? The lichen started with only CO[2], H[2]O, and sunlight!” Organisms are capable of chemical syntheses that put present technology to shame. And the way they do so is an inspiration to would-be nanotechnologists.

  Present synthesis techniques are often derided as “shake and bake.” Reagents are mixed together and then the statistics of colliding molecules take over. Usually, you get a mixture of possible products, only one of which you want, so then you have to purify the mess to get it out. Then you react it with the next reagent (or set of reagents) and do it all again. Lots of the steps, moreover, usually merely involve putting on and taking off “molecular masking tape"—what an organic synthesist calls “protective groups,” simply clusters of atoms stuck onto the molecular framework to keep parts of it from reacting so that other parts can react. It would be better if we could just react the parts of the molecule we wanted, without having to worry about shielding the other parts.

  If (say) your yield at each step averages 80%, then if you had 20 steps your total yield is (0.8)[20]= 1.15%. There's lots of room for improvement.

  Biology gets much better yields by using highly specific catalysts (enzymes), a catalyst being a substance that speeds up the rate of a chemical reaction without undergoing an overall change itself. If a catalyst is good ("selective") enough, it essentially excludes all the other products except the one you want, so that yields can approach 100%.

  It's more than just highly specific catalysts, though. Biology uses highly specific synthetic assembly systems,in which the product of one reaction is handed off into the next: Consider the sequential molecular assembly of (say) a nucleic acid chain.

  So nanotechnological approaches to synthesis will first of all involve highly selective catalysts. That's already happening—catalysts in fact are one of the best examples of proto-nanotechnology (see “Toward a Not-Just-Diamond Age,” Analog, March 2007). More than this, though: We need assembly systems, structured molecular constructs that receive raw reagents and produce the finished molecular product, just as happens in biosystems. Some work has already been done on attaching reagents to substrates, to control the angles at which the incoming molecules can interact. We could even modulate the substrate by (say) applying an electrical charge to it, either to direct molecular assembly using electrostatic attraction/ repulsion, or to induce electrochemical reactions.

  The goal is to produce finished molecules with as little energy expenditure as possible, with as few byproducts as possible, and using the simplest starting materials possible.

  Just as the lichen does!

  * * * *

  Things That Go Boom

  Well, as I've already implied, there's a close connection between rocket monopropellants and explosives. In both cases we're looking for molecules that have a lot of stored energy. In a rocket, of course, we want to release that energy in a highly controlled manner. But often the very same stuff can be used both as a propellant and as an explosive, as we've already seen with gunpowder. With high explosives, though, “detonation"—a catastrophic shock reaction propagating through the substance at speeds of kilometers per second—must be avoided at all costs. At least in some cases proper engineering can keep the reaction rate from running wild like that. (Of course, it's just what we want if we're instead using the compound as an explosive.)

  Traditional high explosives are nearly all based on nitrogen-oxygen groups in association with an organic backbone. Just as in AN, the nitrogen acts as a “spacer,” keeping the oxygens from reacting with hydrogen (and carbon) elsewhere in the molecule. “Organic nitrates,” or esters, are made by reacting nitric acid with an OH ("hydroxyl") group on the organic molecule. Nitroglycerin(e), for example, is formed by reaction with glycerin (glycerol)(4):

  [FOOTNOTE 4: All right, if you want to be really formal it's 1,2,3-propanetriol.]

  3 HNO[3] (nitric acid) + C[3]H[5](OH)[3] (glycerol) (right arrow) C[3]H[5](NO[3])[3] (nitroglycerin) + 3 H[2]O (water)

  And again, the oxygen in the nitrates would really rather abandon the nitrogen:

  4 C[3]H[5](NO[3])[3] (right arrow) 12 CO[2] + 10 H[2]O + O[2] + 6 N[2] + 4057.6 kJ/mol[2]

  Boom! (Actually, it wouldn't be that clean; you'd get nitrogen oxides as products, too, which would cut down the overall energy.)

  And, it doesn't take much of a nudge to happen. Nitroglycerin is famous for its touchiness. Alfred Nobel, of course
, is famous for “taming” nitroglycerin(5) by absorbing it into a porous silica to make dynamite, which is much more stable to handling. (Even so, be careful. Dynamite that's been lying around for a long time tends to be decorated with little beads of “sweated out” nitroglycerin! Old dynamite's a traditional hazard around abandoned mining areas.)

  [FOOTNOTE 5: Curiously, tiny amounts of nitroglycerin are used as a drug. Yes, the “nitroglycerin” tablets heart patients take are the very same nitroglycerin—obviously packaged in a nonexplosive way! It's a “vasodilator,” meaning that it dilates blood vessels, which is useful for people at risk from strokes or heart attacks.]

  Another nitrate ester is pentaerythritoltetranitrate (PETN), C(CH[2]NO[3])[4]. It's both more powerful and less sensitive than nitroglycerin—though it's still sufficiently sensitive, it's not much used any more.

  Nitrocellulose, made by reacting nitric acid with cellulose, is another ester and is sufficiently well behaved that it's the basis for nearly all modern smokeless powders, used as the propellant in firearms. Even so, though, it's easy to blow up even a modern firearm with the wrong formulation. The powder motes are carefully shaped and sized to control the burning rate for different cartridges.

  Alternatively, nitric acid can react directly with the carbon backbone, leading to a nitro (NO[2]) group directly bonded to a carbon. Trinitrotoluene (TNT, C[6]H[2]CH[3](NO[2])[3]) is the most familiar example. It's not very shock sensitive and can even be melted and poured! Obviously this was a great convenience for loading munitions.

  The “nitroamine” explosives have a nitro group directly bonded to another nitrogen, rather than to a carbon. RDX (cyclotrimethylenetrinitramine, (CH[2])[3](NNO[2])[3]) is the “canonical” example. It's a white crystalline solid that's often billed as the “most powerful conventional high explosive,” and it's remarkably inert. It merely burns at room temperature and doesn't even detonate when hit with small-arms fire. For this reason it's the main ingredient in plastic explosives. Very probably the reason for its relative inertness is the extra nitrogen “spacer” keeping the oxygens away from the methylene (CH[2]) groups.

  If you were going to ask about perchlorate esters or perchlorate derivatives, they make nitroglycerin look like a paragon of stability. They're waytoo touchy for practical uses! The perchlorate ion, ClO[4]-, at ordinary temperatures is unreactive. It consists of a symmetrical tetrahedron of oxygens with the chlorine stuck in the middle, and its very symmetry makes it difficult to react. Attach something to one of the oxygens, though, as in an ester, and you've got a “handle” for breaking up the tetrahedron much more easily. So far as I know you can't even attach a ClO[3] group (what you'd get by replacing one of the oxygens) directly to a carbon—it just falls apart too fast.

  A more fundamental problem with all these sorts of compounds is the dead weight (well, okay, dead “mass") of the nitrogen or chlorine atom. It just acts as a molecular spacer to keep the “fuel” and “oxidizer” parts of the molecule apart. Otherwise, it's a drag on the energy generation and on the specific impulse. The double nitrogen spacers in nitroamine explosives, which probably account for their remarkable inertness, impose even more of an overhead. A further overhead cost is the energy needed to break the bonds so that reaction can even occur.

  This is why endothermic compounds and explosives, as a rule, yield lower specific impulse than the simple combustion of bipropellants such as oxygen and fuel. Of course, the rocket pioneers such as Goddard pointed this out nearly a century ago. (And it's also why “fuel-air explosives"—basically aerosols of fuel—are of such interest to the military. Potentially they're much more energetic than ordinary high explosives.)

  Again, though, bipropellants have other problems. So let's now look at some different ways to make energetic compounds.

  * * * *

  Strains and Bonding

  The very chemical bonds themselves provide a way to pack more energy into a molecule. Acetylene (C[2]H[2]) is a familiar example. Everyone knows it has an exceedingly hot flame, which is why it's used in welding. But why does it have such a hot flame?

  Because it's an endothermic compound as well, and so the extra energy gets added to the combustion. In fact, acetylene is unstable at modestly high pressure. It tends to react with itself ("polymerize") into a bunch of waxy goo that settles to the bottom of the tank. This is notgood news if you're expecting gaseous C[2]H[2] to come out of the tank to run your welder!

  So why is it endothermic? The two carbons are connected by a triple bond, and that bond has higher energy than a single or double bond. (Basically it's holding the carbons too close together, so that their mutual electrostatic repulsion is higher.)

  So if we put carbon-carbon triple bonds in something, they will store energy. Compounds based on acetylene are well known: Calcium “carbide,” CaC[2], is an example. It's actually an “acetylide,” in modern terminology. The two carbon atoms are linked by a triple bond and make up the “acetylide” ion, C[2][-2]. Formally, the two hydrogens in the acetylene molecule have been replaced by a calcium. An acid is something that releases hydrogen ions, and a “salt” of an acid contains a metal ion that replaces the hydrogen. So we can consider CaC[2] as a calcium salt of acetylene—considered as an acid(!). What, then, about acetylides as high-energy fuels?

  One problem is that acetylene's “acidity” is purely formal. In fact, acetylene is so unwilling to give up its hydrogens that when water is added to CaC[2] acetylene bubbles up. The C[2] ion grabs hydrogen off the water molecules instead! (This is why calcium “carbide” is traditionally used in miners’ lamps, by the way. The acetylene from wetting it runs the lamp.) But triple bonds can be put into more complicated compounds—the problem is that conventional syntheses of such things are extremely expensive. (See where the nanotechnology's coming in?)

  Strained compounds provide another path to extra energy. Bonds between atoms that result from blending the orbitals from the isolated atoms—so called “covalent” bonds—have definite preferred angles at which they occur. The four single bonds around a carbon atom, for example, define a regular tetrahedron. Sometimes, though, the position of the other atoms in the molecule will not allow the bonds to occur at this preferred angle. Such a distorted bond is strained, and it has higher energy due to that strain.(6)

  [FOOTNOTE 6: Here's the physical mechanism for the strain energy. A chemical “bond” occurs because, due to the rules of quantum mechanics, the electrons tend to be located between the two atoms that are bonded. The negative electrical charge then tends to “screen” the mutual repulsion of the two atoms so that they stick together. (Since the atoms have given up electrons into the bond, they're left with a residual positive charge.) In a strained molecule this screening is not perfect because the electron density does not lie exactly between the atoms, so the atoms still feel some mutual electrostatic repulsion.]

  An excellent and potentially very useful, example of a strained compound is cubane, C[8]H[8]. In cubane the eight carbon atoms occur at the corners of a cube, with the eight hydrogens attached at each corner to fill out each carbon's four bonds. Obviously, the angle from one carbon atom to its three neighbors is 90 degrees. That's what a cube is! This is a far cry from the tetrahedral angle (approximately 117 degrees) that carbon prefers for single bonds. So the bonds are some 27 degrees from where they “want” to be and are thus seriously strained. In fact, cubane's heat of formation is some +602 kJ/mol (positive due to the fact that it is an endothermic compound and takes energy to make), and the strain energy is +695 kJ/mol. In other words, cubane is endothermic due to the strain alone.

  Furthermore, although the cubane framework is highly endothermic, it's also highly unreactive at ordinary temperatures. As one chemist has put it, it's “thermodynamically a bomb but kinetically a rock.” Even at 230-260 degrees C, some 180 kJ/mol of input energy is required to make it react. Because of this inertness, cubane is even nontoxic(!).

  Cubane has high density, too, the highest of any hydrocarbon: 1.29 gr
ams per cubic centimeter. That's almost a third again denser than water! Remember that high density is another desirable feature of a propellant, to minimize tankage mass and bulk.

  Quite apart from its possible use in explosives and fuels, cubane derivatives are also potentially of great interest as nanotech building blocks. The problem remains their routine synthesis. The synthesis of cubane itself was a tour de force—the molecule had been speculated to be impossible to make simply because of its thermodynamic instability. Even though there's now a lively literature on cubane and its derivatives, they're still extremely complicated and expensive to synthesize.

  Which means that they're hardly practical for most applications, particularly when required in tank-car lots!

  * * * *

  A Question of Liquidity

  We've seen various ways to get endothermic compounds. The problem is that they're not pumpable. Remember? The whole point was to get to a throttle-able monopropellant. So we need to turn our attention to endothermic liquids.

  Or more specifically, endothermic liquids that can stand a fair amount of mechanical shock. After all, nitroglycerin is liquid! It's just a wee bit too delicate to be pumped.

  We really should be able to design a compound that's liquid at room temperature and also has the kinetic stability of (say) RDX, plus much higher stored energy from the incorporation of cubane and/or triple bonds into the hydrocarbon part of the molecule. For example, octa-nitrocubane C[8](NO[2])[8], with each corner hydrogen replaced by a nitro group, is estimated to be about 25% more powerful, mass for mass, than other nitrated explosives. (It's been synthesized, but in too small a quantity to test!) It's also thought to be shock insensitive.