Org Prep Daily

November 3, 2007

High temperatures

Filed under: procedures — milkshake @ 10:58 am

I was recently attempting cyclizations that are best done thermally but they did not work at 250C (unlike the literature precedent) so I investigated them after the Russian manner; in diphenyl ether and under pressure at 290 – 340C they confessed everything.

The temperatures needed were outside the range of a Biotage microwave reactor and oil bath. I did not have an alloy bath and I was looking for some cheap high-temperature bath liquid (I needed temperature reproducibility and the sand bath/solid metal block temperature readings are notoriously unreliable). I found that molten sodium nitrite worked well for the purpose: A 60mL beaker was filled with solid NaNO2 and buried in sand bath obtained by filling a heating mantle with sand (the mantle was controlled via Variac transformer). This setup was placed on top of a stirplate. When NaNO2 melted a magnetic stirbar rod was added (a common stirbar from which the Teflon coating was shaved off) and bath behaved nicely in range 280-340C. I also found out that I could extend the range downwards a little, approximately to 260C-340, by adding few spoons of KNO3. The bath starts bubbling a bit at 360C because NaNO2 begins to decompose slowly above 320C but a 20 min heating to 340 was without problems. My main worry was what might have happened if my reaction vial had ruptured and the organics got into the bath: molten NaNO2 is a pretty strong oxidant and tiny bits of paper burst into flames when fed into the bath…

If you have run high-temperature reactions like this – please what kind of bath did you use?

Addendum: A 2:1 (by weight) fused mixture of KNO3 and LiNO3 melts around 140-160C and solidifies at 125C. It should be far more heat resistant than the NaNO2-based mixtures and thus it should be applicable even for very high temperatures (>300C). LiNO3 is not expensive – but it is somewhat hygroscopic. Of course spilling a reaction mixture into the bath would still backfire 🙂 Thanks Kai for the reference.


  1. You’re brave… I would have never considered that… Nice though!


    Comment by Javaslinger — November 3, 2007 @ 12:34 pm

  2. I love the link to the Syriana picture. When I read your post, that sums up quite nicely what first came to mind. I’m not brave enough to do something like this, but maybe molten sodium nitrite isn’t as dangerous as I assume.

    Comment by Chemgeek — November 3, 2007 @ 12:56 pm

  3. The Chemist’s Companion, “Fused Salts”/Eutectics

    CuCl/88 mole-% ZnCl_2 mp 242 C

    Add a little Lewis acid (powdered samarium triflate). Use a heated major hydraulic press in propionitrile/buytronitrile eutectic solvent (mp rises with pressure). Organic synthesis at red heat is a little off the wall (unless you have oxidatively cyclized hexaphenylbenzene with FeCl3/MeNO2. The product C_42H_18 melts higher than Pyrex mp caps.)

    Comment by Uncle Al — November 3, 2007 @ 1:12 pm

  4. Jeez – just use a heating mantle and stick a temperature probe or thermometer into the reaction!

    Comment by Steve — November 3, 2007 @ 1:34 pm

  5. I can’t stick a probe into these react mixtures because the reaction is done in a closed vial under pressure…

    As for the high-temperature reaction being unfashionable: Stoltz and Wood started their impressive Staurosporine tot synthesis by cyclization of oxaldiamide of o-toluidine, onto the methyl groups, to produce 2,2′-bis-indolyl. A closed ampoule, 350C, neat t-BuOK excess as a solvent and base. It was the second step in a long and tricky scheme – so Brian Stoltz had to put hundreds grams of amide through the melt.

    (Two steps later he was doing rhodium chemistry in t-BuCOMe. We asked how ho got this weird solvent – it turned out he went to stockroom in desperation and loaded on cart all liquid chemicals with boiling point below 250C and then proceeded to test one after another as a solvent in his reaction…)

    Comment by milkshake — November 3, 2007 @ 1:41 pm

  6. I’m doing a lot of chemistry in boiling diphenyl ether (Fischer indole synthesis) and it works very well. For this purpose I use only heating mantle and it’s possible to reach 330C without any problem.

    Comment by Edvinas — November 3, 2007 @ 2:34 pm

  7. I just know how hot can be my heating mantles. Nice trick with molten salt, should use it some day. So far I’ve been just putting an opened vial (phenyl ether, Claisen rearrangement) right in the mantle, lifting it with tweezers to cool every once in a while. Takes 5-10 minutes.

    I also tried heatgun rearrangent, but overheated the vial – it was above 300! Still, recovered some product from the black glass left.

    Comment by liquidcarbon — November 4, 2007 @ 7:19 am

  8. liq.c: heat guns are wicked like that! they locally superheat the glass, and not very evenly at all. never use one for a recrystallization, no matter how convenient it is (yeah, learned that the hard way)

    Comment by Ψ*Ψ — November 4, 2007 @ 2:43 pm

  9. I think you can learn how to handle a heatgun for this purpose after some practice 🙂

    Comment by liquidcarbon — November 4, 2007 @ 8:43 pm

  10. I like to use heat gun for re-crystallizations but I was cursing myself when I was re-crystallizing a precious and heat-sensitive carbamoyl chloride, on large scale, and I caramelized the whole batch and had to start over again. The heat-gun re-crystallization worked well enough on small scale but with a big flask to heat and with great self-confidence I ruined the thing.

    Comment by milkshake — November 4, 2007 @ 11:15 pm

  11. I am doing re-crystallizations with the heating gun routinely and they usually work fine, you just need to practice a little bit; everyonce in a while something like what Milkshake just mentioned can happen, but it’s up to you to decide when it’s worth it and when you should be very careful instead!

    Comment by ZAL — November 5, 2007 @ 12:17 pm

  12. I attempted to use a sand bath once. The temp gradients are staggering- 1-2 cm apart and there would be a jump from say, 225 deg C to >350 deg, even after hours of equilibration!

    Comment by Jose — November 6, 2007 @ 11:51 am

  13. My former lab had both a metal bath (Woods metal – pretty messy, oxidizes at high temperatures such as those you use) and a salt bath.

    This is what I found with some searching (translated from

    A mixture of 60% 60% NaNO3+40% KNO3 melts at 222°C and is used as a heat transfer medium in solar-thermal powerplants. The melt is chemically stable up to 590 °C, has a density of 1,79 g/cm³ and the viscosity similar to that of water (2,1 mPa s).

    Addition of sodium nitrite reduces the melting temperature further: A salt mixture called HiTec (53% KNO3 + 40% NaNO2 + 7% NaNO3) melts at 140°C.

    There’s a site that has a lot of mixtures searchable (not the ones above, though; no idea why):
    Seems like most low-melting salt mixtures are either oxidizers or Lewis acids, didn’t see anything benign there. On the other hand, the nitrate mixture is used to vulkanize rubber by injecting the precursor into it, so maybe one doesn’t have to worry that much about combustion.

    Another interesting thing is that some salt melts are used as solvents for e.g. cellulose (“Using the melts of LiClO4·3H2O, NaSCN/KSCN/LiSCN·2H2O and LiCl/ZnCl2/H2O as cellulose solvents,…”) or polyacrylonitrile. Those salts/salt mixtures might not work for your temperature range, though.
    See e.g. these papers:;2-E;2-F

    BTW, lead melts at a little over 300 C, so using that plus a little say Bismuth may also work for your purposes without too much of an expense.

    also, i am a big fan of heatgun recryst, why drop the convenience for the 1% of cases (which might even be anticipated) where it messes things up…

    Comment by kai — November 6, 2007 @ 2:56 pm

  14. some more info regarding the fear of combustion: check out

    Click to access HITEC%20Heat%20Transfer%20Salt.pdf

    to see what kind of tests they did.

    I guess the low volatility of the diphenylether could be a problem, but the stuff seems less dangerous than what I would have thought.

    Comment by kai — November 6, 2007 @ 6:56 pm

  15. Kai,

    this is most useful information – thank you! I am quite amazed about the fire-hazzard test results they did with their HiTec: even solid aluminum is tolerated, at lower temperatures, so is steel. It does not flare with gasoline, grease. But burns with wood, charcoal. Also the temperature range seems much higher than expected, I thought the thermal instability of nitrite could be limiting but they were able to heat it up quite high. (I think anything that goes to 400C is more than enough for organic chemistry).

    I already ordered half kilo of Wood’s alloy but I will make the HiTec too, and compare the two. Thank you!

    Comment by milkshake — November 6, 2007 @ 9:21 pm

  16. Potassium formate is low-melting, at 165C, but it is probably super-hygroscopic and decomposes at higher temperatures. Also KSCN m.p. is 173C

    Comment by milkshake — November 6, 2007 @ 10:25 pm

  17. Metal baths surface oxidize and their fumes are toxic, plus metal fume fever. Their high densities require a steel container and stress your submerged Pyrex. Sprinkle the molten surface with powdered graphite. Bath components are not carbide formers. Floating oxidized dross is reduced back to metal (and the graphite to carbon monoxide). A half-kilo of Woods metal is about 50 ml cold.

    Benzophenone plus biphenyl or diphenyl ether forms eutectics used as high temp heat exchange fluids – in closed systems. Higher glymes go very high, but they peroxidize. Mobil 1 goes mighty high as do vacuum pump oils. You are looking at a terrible mess plus fumes in open air. Have fun!

    Comment by Uncle Al — November 7, 2007 @ 1:41 pm

  18. Yea, I just opened a box from Aldrich – and was unpleasantly surprised how few tiny little Woods alloy sticks I got for my hundred bucks. I actually thought about putting some surfactant-like additive to cover the alloy molten metal surface – we have a huge jar of low-melting zinc undecylenate from some past project – but I think the graphite sprinkle trick looks better. Thank you, Uncle!

    Comment by milkshake — November 7, 2007 @ 2:14 pm

  19. For an unstirred mixture, I have used boiling sulfuric acid in a custom glass vapor bath to get very precise temperature control (300.0 °C) (see US Patent 6,407,089), and have contemplated the use of mixed diethyl/dibutyl phthalate esters for similar precision control in the range 300-340 °C. Phthalates are pretty inert to thermal decomposition/oxidation, but do the reflux part in N2 to be on the safe side. You’d need a good glasssblower to provide you with the stirring capability for use with boiling vapor baths.

    I am a regular reader of your blog. You’re the man!

    Comment by mrj — November 7, 2007 @ 6:29 pm

  20. Uncle Al mentions high temp exchange fluids (DowTherms) and I have taken to about 300C (mantle-sand bath small beaker or immersed coil 2L set up). Cavet is I worked in a plant and had assess as the Engineers got free samples (5 gal) as typically Dow would only sell in drum qunatity. As noted they are designed for closed system as this is above flash points so risky to use in open vessels. I did find most seemed to contain lower boiling components as first heat up saw evolution and occasional bumping that did not occur next use.

    Comment by CMC guy — November 7, 2007 @ 6:30 pm

  21. further thought: for stirring HiTec, you might try those kid’s magnets that come in pairs. They are shaped much like a stir egg and are covered with what looks like some very hard glassy coating. They make funny sounds when you handle two of them together (that’s how they are marketed). One by itself might be perfect for stirring!

    Comment by mrj — November 7, 2007 @ 6:40 pm

  22. kai, thank you for the tables of eutectics: CsBr+LiBr (48.5:51.5 mol ratio, mp 230C) looks innocuous enough and should be very stable – only that CsBr is so damn expensive and hygroscopic…

    KNO3+LiNO3 (59:41 mol ratio, mp 124C) looks non-toxic and stable, (KNO3-LiNO3-NaNO3 53-30-17 is even mp 120C). But this does not solve the oxidizer hazzard.

    Comment by milkshake — November 7, 2007 @ 10:34 pm

  23. I am surprised that Wood’s metal is still available. I would use solder for baths in these ranges. Standard 60/40 leaded solder melts at ~185 C and Sn Ag 3.0 Cu 0.5 at ~220. The lead free stuff (RoHS compliant) is used in wave flow soldering machines, and should be reasonably resistant to oxidation.

    Comment by Eric M — November 10, 2007 @ 12:10 pm

  24. For normal high-temp reactions (220 – 330C) I use Primol, a type of mineral oil common in chemical plants. I think it is made by exxon-mobil. For really high, like 500-650C I use a quartz hot tube.

    Comment by Milo — November 11, 2007 @ 5:07 pm

  25. I wonder if anybody used this:
    “Flaked graphite has advantage over both oil and Wood’s metal. It is clean and inexpensive. It is a good conductor of heat and permits magnetic stirring. In addition, it is thermally stable and not flammmable under normal laboratory conditions.”

    Quoted from Aldrichimica Acta (1988) Vol 21, No 1, p. 2

    Click to access al_acta_21_02.pdf

    NB Flaked graphite is available from e.g. Aldrich (ord nr. 332461)

    in MSDS it is written “Autoignition Temperature 1346 deg F ( 730.00 deg C)”

    Comment by Alexey — November 17, 2007 @ 9:30 am

  26. I did not hear about a graphite bath but I would expect it can have the same problem with uneven heat distribution as sand bath since there is no convection. (I could not find the reference in the Aldrichimica Acta link you provided – but there was a nice sonochemistry review there)

    Comment by milkshake — November 18, 2007 @ 7:14 pm

  27. That link is wrong, the correct link is this one:

    Click to access al_acta_21_01.pdf

    But it doesn’t contain much more than I quoted. On one hand, most probably you are right that it will overheat since there is no convection. On the other hand, it was suggested by prof. A. G. Davies (see this link) and it says that graphite is a good heat conductor (vs. sand). But it doesn’t say for which purposes it fits.

    So, what are you using at the moment?

    Comment by Alexey — November 21, 2007 @ 1:45 pm

  28. P.S. I wonder what was the reason for Davies to use graphite instead of sand?

    Comment by Alexey — November 21, 2007 @ 2:01 pm

  29. Right now I am not doing anything high-temp but I expect that soon we shall melt again.

    Comment by milkshake — November 21, 2007 @ 5:20 pm

  30. We’ve been doing a lot of very high temp rearrangements under pressure in stainless autoclaves and the best safest method of heating to 350+ is the fluidised sand bath, it takes about 2 hours to heat and is good up to 400+ with good heat transfer uniform temperature and no fire risk. previously I have used phosphoric acid and dowtherm and other nasties. The unpleasant persistant flower smell of diphenyl ether is something I won’t forget for a while.

    Comment by Vecktor — May 26, 2008 @ 4:33 pm

  31. I usually use Wood bath. It has low heat capacity and high heat conductivity. As for your situation, you can try sodium acetate. AcONa itself melts 324 C, but its trihydrate, usual commercial form, melts 58. By the slow heating you can reach wide range of temperatures. I dehydrated AcONa*3H2O by melting dozens of times – it’s rather safe and stable.

    Comment by kolpak — September 14, 2008 @ 5:14 pm

  32. by volume, Woods alloy is ridiculously expensive from Aldrich – I suppose making your own would be a far cheaper alternative. I may try the fused sodium acetate bath, this looks very reasonable. The KNO3+LiNO3 bath worked well but I was frightened what would happen if the flask cracked and the stuff spilled in…

    Comment by milkshake — September 14, 2008 @ 5:21 pm

  33. Hi MS,

    Bit off topic..

    I am having work up difficulties after the cyclization reaction using PPA (Polyphosphoric acid) .. I am forming quinoline compound from maleate substituted amino compound (50- 100 g scale reaction).Used NaHCO3 for the neutralization but gave an emulsion and difficult to extract compound from that. Whats the best work up procedure involving PPA?
    Please suggest me if you have any experience.



    Comment by marto — March 11, 2011 @ 8:40 am

    • I would use KOH solution for neutralization, potassium phosphates are far more soluble than sodium phosphates. Polyphosphates have surfactant properties, maybe this is the source of your difficulties – would your product molecule survive reflux with diluted KOH? This could be a good way to work up the reaction before the extraction. Also, using solvents like ether or tBuOMe in extraction supresses the emulsion.

      PPA reaction workup is almost always unpleasant, my experience is limited but I remember that there is Eaton’s reagent, about 7% solution of P2O5 in methanesulfonic acid. Eaton’s reagent was developed as less awful alternative to PPA, so maybe you should give it a try.

      Comment by milkshake — March 11, 2011 @ 11:59 pm

  34. Thanks very much for the help.

    Comment by marto — March 13, 2011 @ 11:20 pm

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

%d bloggers like this: