Skip to main content

Chemical News

All Is Forgiven At 1000 Degrees

I can’t say that I’ve ever done any flash vacuum pyrolysis, and I’ll bet that most chemists reading this haven’t, either. I’ve seen papers on it for many years, including some pretty unusual reactions, but it’s never been a tool that I’ve reached for. Here’s a new paper in the Journal of Organic Chemistry on the technique, a Perspective by Lawrence Scott (now retired from Boston College), and it’s almost enough to make me want to run some of the reactions.

The idea is pretty simple. You take your starting material, in the vapor phase, through a short very-high-temperature zone (up to, say, 1000 degrees C). At those temperatures, an awful lot of reactions become possible (that T-delta-S term in the thermodynamics!), which is one reason why you need to keep things rather dilute, even in the vapor phase. (Needless the say, at 1000C, the vapor phase part is pretty much mandatory – that’s well above the boiling point of, say, liquid sodium metal). As Scott’s synthesis of C60 fullerenes under these conditions shows, you really have to think about conformational energies and strain differently at these temperatures, because the molecules themselves are going to be populating a lot of conformations that you’d otherwise never think about. If there are energetically feasible reactions within reach of these conformers, you’ll get some pretty unusual products.

The Perspective is mostly about the preparation of numerous bowl-shaped hydrocarbons, and (eventually) C60 itself, but you can imagine a lot of other things happening as well. Scott mentions that they did try to make a di-aza compound in one series, but with no success whatsoever. Later calculations showed that the nitriles in the starting material were much less likely to participate in the necessary cyclization reactions than the plain acetylenes that they’d been using, and the reduced aromaticity of the pyridine products helped make the whole process thermodynamically unfavorable.

But it’s not like heterocycles in general can’t be prepared by FVP techniques, of course – here’s a recent review. And here are some nice slides from one of Phil Baran’s group meetings a few years ago, with a good collection of FVP transformations (both heterocyclic and carbocyclic). Just looking over the literature on this sort of thing briefly in this way, you can see that there’s a lot of chemical diversity that’s relatively unexplored. The upswing of interest in flow chemistry over the last few years makes me wonder if this isn’t another possible market for the flow equipment makers – after all, the FVP folks have been doing their work under flow conditions from the very beginning, just from a slightly different (and slightly warmer) perspective.

With such equipment, the whole area looks like a good place for the “Reaction Discovery” techniques that people have been using to find new catalytic reactions. You could imagine taking a plate of model systems through some FVP conditions and just analyzing the products by LC/MS to see what happened to them, looking for new ring-forming reactions, rearrangements, and so on. The only thing I would have to fear, working in the pharma labs, is the wrath of the scale-up chemists if we used one of these scaffolds in a drug candidate. But that could turn into a problem worth solving, under the right conditions. . .

33 comments on “All Is Forgiven At 1000 Degrees”

  1. Chad Irby says:

    …and with suitable modifications, you could make some of your favorite compounds that way!

    ClF3, anyone?

    1. aairfccha says:

      Don’t know about ClF3 but FOOF is produced more or less like that.

      “…mixture of oxygen and fluorine through a 700-degree-heating block…”

      Although I wonder how quick you have to quench the reaction mix since
      “FOOF is only stable at low temperatures; you’ll never get close to RT with the stuff without it tearing itself to pieces. “

      1. Chad Irby says:

        So at 1000 degrees instead of 700, it should be even MORE fun!


      2. fajensen says:

        Quickly, indeed: “liquid oxygen bath” OK? Then the curious observation: “Audible “pings” and pressure excursions occurred when liquid O2F2 dripped onto uncooled portions of the apparatus.”.

        Err, Sneakers, Please! I’ll just go and watch from over there.

    2. A Nonny Mouse says:

      Bromofluoromethane is made in this manner as well.

  2. Nick K says:

    FVP was quite popular some 30 years ago at Imperial College London, where the late Professor Charles Rees did a fair bit. I remember seeing an FVP kit made by Aldrich. I don’t know if it is still available.

    1. Li Zhi says:

      That’s what e-bay’s for, dude :p

  3. anonymous says:

    I mean one can do structure elucidation, mechanism, bond migration with ease I suppose but only on aromatic system. I haven’t seen many papers on aliphatic systems (unless I am wrong)! We have seen papers by Hudlicky and others on synthesis of cyclopropanes at high temperatures but has been superseded by metal based catalysts at lower temperatures. I mean my compounds (peptides, heterocycles) would be toast at that temperature! But, I still think it is cool! I am also thinking Dutch Chemist Prof. Wentrup who also popularized FVT.

  4. Some idiot says:

    As a scale-up chemist, it sounds more like an opportunity than a threat… the (good) scale-up chemists are usually very open for a (good) continuous process…!

  5. Barry says:

    all the discussion of FVP seems to presume that the chemistry is vapor-phase. But heat transfer at very low pressures is poor at best. Isn’t it more likely that much of the chemistry happens on the tube walls? Any thoughts?

    1. Mark Thorson says:

      There would be radiant heat transfer. Don’t most organics absorb IR fairly well? This may resemble photochemistry.

      1. Mark Thorson says:

        Now that I think about it, I wonder if you could make C60 or CVD diamond using LEDs. Even if it’s slower, the lower cost of equipment would let you run large gangs of reactors in parallel. You’d probably want to choose a vapor-phase precursor that’s very sensitive to your LED wavelength. Aromatics absorb strongly in the UV, so maybe something like napthalene illuminated by UV LEDs.

    2. Paul D. says:

      Thermal diffusivity of a gas is inversely proportional to its density. So heat transfer into the gas is very good at low pressure.

  6. Anonymous says:

    No literature access to look things up or even see if they’re mentioned in the review. Both of these are below 1000 C, but: (1) Schleyer’s spectacular synthesis of adamantane by thermal rearrangement of tetrahydro-cyclopentadiene, C10H16 —> C10H16. (2) Prinzbach’s rearrangement of pagodane to dodecahedrane. Then there’s the popular conversion of C(diamond) to C(graphite) at high temperature.

  7. AQR says:

    When I was in grad school in the 70’s, one of the cume questions asked for us to propose a mechanism for a FVP transformation. One of my colleagues answered, “At 1500 degrees, you don’t need a mechanism.”

    1. John Wayne says:

      I hope they passed

  8. AndrewD says:

    We already do very large scale vapour phase reactions at 1000C eg Steam reforming of Hydrocarbons
    Much petrochemical refinery chemistry is vapour phase at high temperature so we process chemist have an installed base to start from if you research people come up with something useful.

  9. Chemperor says:

    A lot of textbook cluster chemistry was done by vacuum pyrolysis in the labs of Prof Brian Johnson and Lord Jack Lewis. The main challenge with this kind of chemistry is isolating and characterizing all of the species found in the “primordial soup” that comes out the end of the reactor. A lot of vacuum pyrolysis work gets criticized because it is challenging to scale up and selectivity is poor. However, these kinds of experiments teach us a lot about what’s possible in a system and provides opportunities to develop specific synthetic techniques to target the most interesting species. All things old become new again!

  10. Barry says:

    Bill vonE. Doering cautioned long ago that chemistry reported with e.g. LAH would more properly be attributed to LAH with an uncharacterized number of THFs in coordination. He spent most of his career working in the vapor phase.

  11. Fazal Majid says:

    Wouldn’t sonochemistry (the use of extreme temperatures from ultrasound-induced cavitation) be a much more convenient mechanism for this, not to mention it can reach local temperatures of 5000K?

    1. Mildweasel says:

      At 5000K you’re doing plasma physics I would think

      1. zero says:

        Xenon compounds? Things even more exotic?

    2. Chemperor says:

      This is probably outside the realm of this thread, but has anyone had great experiences with sonochemistry? I spent a good deal of my advisor’s money several years ago outfitting our lab with a sonic generator with this exact idea in mind. I thought that perhaps I could access unusual inorganic motifs using the technique. In the end, I had nothing to show but several broken flasks and a lot of starting material. I’d like to revisit the idea with my own group, but it’s quite an investment and I’ve been burned–figuratively–once before.

      1. Nick K says:

        My experience with sonochemistry, FWIW, is that it can have a remarkable effect speeding up heterogeneous reactions, especially those involving metals. Several times I was able to make organozincs by immersing the flask in an ultrasonic bath for a minute or two. Without sonication nothing would happen. I also tried the same trick with homogeneous reactions, but observed no effect.

        1. Derek Lowe says:

          That’s been my experience as well. The sonication-induced cleaning of the metal surface can be a very potent effect, and one could easily see how it could accelerate a reaction whose rate-limiting step is a reaction at a fresh surface. Homogeneous reactions, though, are another matter. Conditions inside sonochemical bubbles are very weird indeed, but I don’t think I’ve ever seen an effect in solution.

          1. Me says:

            Nobody else used it to kickstart an annoying Grignard formation/

  12. Dr. Lloyd T J Evans says:

    I actually have seen FVP being done, in the lab I worked in during my doctorate research studies. I never used it myself, but my colleagues on the bench next to mine did. It was a rather specialized setup, optimized to take in cyclooctatetraene (COT) and transform it into pentalene. This involves making an additional bond across the ring between the 1 and 5 positions, kicking out one equivalent of dihydrogen.

    You end up with two fused cyclopentadiene rings in a rather unstable configuration, so the product has to be immediately condensed into a dry ice cooled container. You can’t keep it like this for long, so the usual procedure was to make it in smallish batches of no more than 50 grams, then treat it with butyllithium while still cold to make the dilithium pentalene salt. This is then aromatic and therefore thermally stable, much like how cyclopentadienyl lithium (LiCp) is also aromatic and thermally stable.

    There are other ways to make pentalene of course, but the FVP method from COT seems to be the most efficient and least time-consuming one. The pentalene dianion is of course a ligand in organometallic chemistry, in which it finds almost as widespread and varied use as its single ringed Cp relative. In fact, one of the first reactions which demonstrated that the pentalene ring system could exist was a weirdo rearrangement of a tantalum COT complex.

    COT can be made into an aromatic dianion as well, which also finds use as a sort of Cp equivalent for large metals. In fact, a pair of COT rings sandwiching uranium(IV) is generally known as uranocene. Direct analogues of uranocene can be made for some of the other actinides and lanthanides. But COT can also be put on transition metals just the once, if there are other supporting ligands on the metal.

    The strange transformation happened with a (COT)TaMe3 complex. Someone in the same lab I worked in had tried to convert this to the trichloride complex by treating it with triethylamine hydrochloride. Which did succeed in replacing the methyl groups on the tantalum with chlorides, but had an unexpected effect on the COT ring. It rearranged itself to the pentalene dianion (again by forming a new bond and kicking out dihydrogen) while remaining attached to the tantalum. Any organic chemists want to try and propose a mechanism for that?

  13. John Campbell says:

    I am surprised no one has mentioned the work of the late Prof Hamish McNab on Flash Vacuum Pyrolysis. There is a review of his work here:

  14. LabM says:

    ” At those temperatures, an awful lot of reactions become possible (that T-delta-S term in the thermodynamics!)…”

    Derek, it’s a misconception. As van’t Hoff and Le Chatelier tell us — it’s delta H that determine how equilibrium constants will change with temperature (assuming delta S and delta H do not or weakly depend on T). The quantity you need to look at is not delta G, it’s delta G/RT.

    1. Hap says:

      Except dG = dH-TdS; Keq = K exp(-dG/RT) (as you noted, dG/RT is important here) = K exp (-dH/RT + dS) = Kexp(-dH/RT) * exp (dS). At high T, the effect of the dH term on Keq (and dG) decreases, and so the effect of the dS term on Keq (or dG) increases.

      If dH increases as T, then it would keep up, but a lot of thermo would be useless then. As long as it doesn’t increase faster than T, then eventually at high T, enthalpic contributions lose to entropic ones (assuming that everything isn’t a plasma at that point).

      1. LabM says:

        My point was that “an awful lot of reactions”, perhaps, should have been “an awful lot of endothermic reactions”. No increase in T will get you more products in an exothermic reaction. Since Derek’s statement sounded pretty general, I simply pointed out that one should not ignore delta H in determining what happens at high T.

        Of course, specific conditions (K > 1 at high T) put constraints on the sign and magnitude of delta H and delta S that make your analysis valid, but they are specific not general.

  15. Hap says:

    Yes, the analysis assumes that dH > 0 and that dS 0, then higher temperatures inhibit the reaction, and if dS is small but < 0, then the entropic term won't win out until plasma, and won't help you do a reaction. It would depend how big dH and dS are for heat to help or hurt a reaction (if dH is small, the sign of dS matters a lot, but that's probably not the kind of reaction anyone's discussing, and if dH is large, then dS won't have any positive effect with temperature, because K is large enough that there won't be much increase in product concentration with temperature, or dS < 0 and temperature lowers product concentrations).

    1. tangent says:

      Digging back to my P-Chem elective, here’s what I don’t get: aren’t chemists usually more interested in bond-forming reactions, and don’t these temperatures drive the other direction?

      High T favors the equilibrium for the reaction with dS positive. Doesn’t dS positive tend to mean you’re breaking a molecule into parts, or de-cyclizing, i.e. negative net bonding, so stuff can all thrash around a bigger state space with the increased temperature?

      When people get interesting reactions out of these high-T conditions, do they tend to be like condensation reactions, i.e. they form a bond of interest, but they kick some small molecule loose so they stay near zero dS instead of negative? Or kick multiple junk pieces loose to be positive-dS?

Comments are closed.