Org Prep Daily

November 19, 2011

Medchem Immigrant in the Polymer Republic

Filed under: industry life — milkshake @ 2:30 pm

Five months on – and there is no looking back. With potassium metal freshly cut, with the glassware, solvent and monomer lines pumped down overnight to 20 mTorr, ready or not, macromolecules, here I come.


  1. THIS. Is porn.

    I remember urban legends from undergraduate times of the need for setups like this one, if one wanted to accomplish cation polymerizations. You use potassium metal, so I suppose this is not you? Any case, shock and awe.

    Comment by DrFreddy — November 20, 2011 @ 4:10 pm

  2. the small dark flask in the middle is potassium organometalics solution that we use for initiation. The addition funnels are double-jacketed (the outer jacket is vacuum, like in Dewar). The heating/cooling lines are not connected yet to the jackets. Apart from the vacuum manifold there is one more in the back for the monomer but I was not allowed to photograph the monomer delivery setup.

    This is a living anionic polymerization but it is equally finicky – a trace of oxygen or moisture gives rise to reactive impurities that also cause initiation over time and you get parasitic polymers with wrong functional groups at the terminus, and since you can’t remove them from the product they ruin the batch.

    Comment by milkshake — November 20, 2011 @ 4:37 pm

  3. Are you doing EO polymerization? This hood looks really, really familiar.

    Comment by khahn4 — November 21, 2011 @ 12:16 pm

    • Yes, you should remember the hood and the glassware well since you run the EO polymerizations for our company… Janni says hi – he also added that everything you synthesized for him worked well in the assays.

      Comment by milkshake — November 21, 2011 @ 1:38 pm

      • Ha, that’s great to hear! Tell everyone I said hello.

        Comment by khahn4 — November 21, 2011 @ 4:22 pm

  4. What’s that funny piece of glassware on a heating mantle connected to the vacuum hose?

    Comment by JH — November 22, 2011 @ 12:42 pm

    • It is air-cooled simple diffusion pump. There is a hot high-MW liquid in it refluxing at high temperature in the vacuum (some kind of silicone oil I am told), and the vacuum line passes through it on its way to oil pump, much like as if you were doing vacuum distillation through a column, and somehow this greatly improves vacuum, from 30-40 mTorr that the oil pump produces on its own (regular 2 stage Welsh belt-driven) down to bellow 2 mTorr on a fully closed manifold. I have still not fully comprehended how this diffusion pump magic works (since there is no outlet where the scrubbed material would go) but somehow it helps to shovel the pumped stream along and into the regular pump. It does noticeably improve the vacuum when you turn it on, once the silicone fluid reflux starts inside. (You must not forget to let it cool and close it under vacuum before you shut down the main oil pump – letting difusion pump baking under air over weekend is a major screwup as the silicon liquid oxidizes and decomposes and may harden into glassy resin that is exceptionally difficult to dig out and clean from the pump). Another factor that helps is to get high vacuum in the vac line is to fill the cold trap(s) with liquid N2 (rather than with dry ice/acetone that you see in organic labs)

      Comment by milkshake — November 22, 2011 @ 3:33 pm

      • Why of course. I know what it is now that you tell me, but didn’t realize that there were one part glass constructions of such pumps, directly connected to a vacuum line. Makes perfect sense though.

        I understand that it works by having streams of the vaporized compound (if you can call them streams at that pressure) be directed so that collisions will make other gas molecules/atoms go more in the direction of outlet than inlet, thus creating a pressure difference. Molecules of the vaporized compound on the other hand will soon hit a cold wall and flow back to the heated reservoir.

        Comment by JH — November 22, 2011 @ 4:11 pm

  5. Totally unrelated question, out of the blue. Have you ever tried to make an amide (2º) from a benzylamine HCl salt and an ester, under so-called Merck conditions (THF, iPrMgCl, -20 ºC)? I decided to go the simple way, hydrolyse to acid, couple with EDCI, but was curious to try it. Should have. I know it works wonders for making Weinreb amides. But Secondary?…

    Comment by HPCC — November 29, 2011 @ 9:32 am

    • This is an old method, using Grignards, but you may need to heat your ester with the formed magnesium amide, and Mg amides are very strong bases so many functional groups are incompatible. I think a better system is from Weinreb that uses Me3Al in toluene from Aldrich, Me2Al-NR2 are less basic and Al can activate the ester more than Mg; I have had good luck with Me3Al-promoted amidation of esters in the past.
      There is another good method for aminolysis of esters that uses a strongy basic cyclic guanidine TBD (from Aldrich) as a catalyst: JOC 2009, 74, 9490-6 and TL 2007, 48, 3863-6

      Comment by milkshake — November 29, 2011 @ 1:31 pm

      • Milkshake, I wish I had read that beforehand. Isolating that acid made me swear all day, it was a pyridinecarboxylic acid with some greasy tail… Something that would have resisted both Grignard and AlMe3 conditions. Thanks!

        Comment by HPCC — November 29, 2011 @ 4:55 pm

        • I am not sure, sodium amide adds to pyridines reversibly in the 2 position under somewhat forcing conditions, Chichibabin amination of pyridines is based on this. But TBD-promoted aminolysis should be safe.
          Hydrolysis of water-soluble substrates like your pyridine-acid: I would do HCl reflux, evaporate, then you can take the hydrochloride of free acid directly to acyl chloride and it does not even have to be that dry for the next step (thionyl chloride excess takes care of the moisture)

          Comment by milkshake — November 29, 2011 @ 5:25 pm

          • Oooooooh, didn’t think of that, especially that in the 2-position, instead of a hydrogen, I had an amine of some sort, thus a decent leaving group, hence a nightmare to come. And that was my final reaction before delivering the API!!! I am glad I went the painful way; EDCI coupling underway. Rotovapping ANYTHING off that b!@tch generated large foamy bubbles, EtOAc, 2-MeTHF, MTBE, MeOH, DCM, water, you name it. And it was sparingly soluble in EtOAc, anything else, forget about it. AND it was a powder so fine that it would clog anything you wanted to use to filter it – my first intuition… Reaction took 15 minutes, workup and isolation took about 8 hours. On 1.5 g scale…

            Now that I know how nasty it is, I’d use KOTMS next time to get the anhydrous K salt. Which, upon suspending in PhMe, or even DMF, can turn into the nicest acid chloride using oxalyl chloride.

            Comment by HPCC — November 29, 2011 @ 8:25 pm

          • Final update. 97% yield when extracted from a pH 3 instead of pH 6 aqueous solution, using the almighty, near-magical 2-MeTHF solvent. Clean phase separation, it seems my greaseball zwitterion “knows” it wants to be surrounded by an organic buddy under those conditions. A little Reaxys helped me fish out conditions, it seems 90+% people hydrolyse with NaOH (aq.), unlike my original LiOH – I suspect the residual LiCl fudges up the separation too… (If 1M LiCl magically washes off DMPU and NMP off EtOAc or DCM organic layers, I wouldn’t be surprised it would want to pull my pyridine-acid-long chain alkyl aryl ether into aqueous oblivion).

            Anyhow. Hope I get enough material this time around. But weighing out that bastard, pure white solid, akin to titanium white pigments, was pure joy at 5:50 PM tonight.

            Comment by HPCC — December 7, 2011 @ 8:59 pm

  6. Hey Milkshake,
    I’m setting up a distill to dry dichloromethane over CaH2. Is there anything I could use as an indicator for dryness, I mean in the sense of the beautiful blue colour from the ketyl radical that you get when you dry THF over Na/benzophenone.
    Hope you can help out…

    Comment by Young Padawan — December 5, 2011 @ 9:15 am

    • Unfortunately there is only Karl Fisher titration as far as I know. Since CaH2 works well for moisture and HCl removal but it is pretty inefficient for removal of alcohols, so make sure that your DCM is stabilized with alkenes (such as amylene or cyclohexene). Many commonly-sold grades of DCM are stabilized with up to 0.5% of methanol (for example the HPLC grade is always stabilized with methanol) and CaH2 cannot remove it, only P2O5 or conc. sulfuric acid will. Make absolutely sure that your DCM is olefin-stabilized – call the tech support and look up MSDS (they have to declare the methanol content there) – I got screwed in a big way, with Fisher common grade (no sign about MeOH content on the can or in the catalog but sure enough 0.45 vol% it was); it turns out Rh2(OAc)4-catalyzed cyclopropanations really do not like the presence of an alcohol since the carbene inserts readily into O-H bond.

      Comment by milkshake — December 5, 2011 @ 10:14 am

      • Thanks Milkshake for the info.
        I’ll check on that solvent that I am using, though I think it’s only amylene stabilized. So karl-Fischer it is…

        Comment by Young Padawan — December 5, 2011 @ 5:56 pm

      • Wouldn’t 4Å mol sieves remove methanol?

        Comment by JH — December 6, 2011 @ 10:30 pm

        • Basic alumina will remove acids, alcohol, and water. Probably your best bet.

          Comment by Kai — January 17, 2012 @ 9:55 pm

        • they might soak up trace quantities of MeOH – but 4.5 mL/L is a bit too much

          Comment by milkshake — December 7, 2011 @ 10:48 am

  7. Totally lab porn heh. I love it!

    Comment by Ryan Mercer — December 5, 2011 @ 1:12 pm

    • and you have not seen the colors: Right now I am quenching another batch and it is bright green, as it should be. (The second – a bit more exploratory – experiment is turning weird pink-brownish). And you will get a large bucket of a very fluffy snow-like polymer at the end if things go right, with a tall narrow erect peak on GPC, with a PDI value below 1.05 and a beautiful proton NMR

      Comment by milkshake — December 5, 2011 @ 1:30 pm

  8. Hi Milkshake

    I see a 3 l and a 5 l flask… and magnetic stirring? Years ago I worked on polyurethanes and polyvyniys and I always used overhead stirring…

    Comment by processchemist — January 16, 2012 @ 12:44 pm

    • I agree, it is not a very efficient stirring (though the new rare-earth mag stirbars are awesome). It does not matter in this case as the polymer stays in solution and the solution viscosity is quite low. The main problem here is that everything has to be strictly air-free and the overhead-stirring is hard to do if you need ultra-high vacuum at the beginning and the ability to seal the system for a mild over-pressure after the monomer addition (the monomer is an easy to-condense gas). I think the best way to do this on large scale would be an autoclave, like the kind you would use for hydrogenations. Maybe with agitation by shaking.

      Comment by milkshake — January 16, 2012 @ 5:23 pm

      • An autoclave with a magnetically coupled stirrer should do the job on kilolab-pilot scale, if the viscosity is low. In my lab I still have an old magnetical coupling with a standard ground joint, but I never managed to cut a stirrer shaft to the non standard lenght required. Honestly I have no nostalgy of my polymer days…

        Comment by processchemist — January 17, 2012 @ 4:32 am

        • I have no nostalgia for my medchem past – traditional medchem derivatomania is a drudgery, it does not pay and it is bound to end in crushing disappointment and unemployment. I don’t have much fondness for big pharma and for big academia either…
          I think you had not experienced the awesome properties of our polymers: They are white like fresh-fallen snow and have a consistency of a lyophilized ice cream. (Few steps later they look more like cornstarch or confectioners sugar but they are still white). By their proton NMR spectra, some of the polymers could be mistaken for a small molecule, so beautiful their signals are – in fact NMR is the way to monitor the purity and conversion. Our polymers are meant to be used for drug formulation so we do not worry about bulk material properties, rheology. Instead we obsess about purity and narrow PDI. (I am telling you, its not just converting to a new religion, it is rather like joining a sect.)

          Comment by milkshake — January 17, 2012 @ 5:29 am

        • Also, I had a look at the product list of your company – very impressive if you are making all this in house. You mentioned doing isocyanate polymer chemistry. Do you have experience with kilo-scale phosgene work, for example for preparing aminoacid N-carboxanhydrides (NCAs)?

          Comment by milkshake — January 17, 2012 @ 5:48 am

          • Thanx, I wish you a long citizenship in the polymer republic… I’ll sure have a deeper look at all the chemistry you posted. No diethyl ether! For the sake of anyone in charge of the scale up, evaluate an alternate solvent! Isopropyl ether (if you don’t have to evaporate it/ distill it) (I used it routinely without any of the fireworks I saw with diethyl ether), MTBE/Heptane mixtures, for sure there’s something else that works.

            Comment by processchemist — January 17, 2012 @ 10:53 am

          • we use Fluka-made phosgene solution in toluene but it is getting expensive and increasingly hard to buy so I might be doing some molten triphosgene cracking on scale soon…
            I wish you best luck and sensible customers. Also if you can use anything from Org Prep Daily for your catalog of building blocks please do so. By the way, can you apply diethyl ether in large volumes in your company – i.e. to precipitate and filter stuff from diethyl ether on a 50L scale – or is it a major regulatory/fire hazard problem?

            Comment by milkshake — January 17, 2012 @ 10:31 am

          • My polimers were glues (to be dissolved, used in coating for some biomedical application), a pain were sun does not shine in downstream processing (from the pilot plant the production was about 3 tons/year). Thanx for the appreciation (yes, all the products are in house made, except for 4), but as you eventually know, to be impressive is barely sufficient to survive, in the last years climate.
            No phosgenations at our shop (regulatory problems, but my partner has a strong experience from the old days, with a caged bird used as detector on the plant…).

            Comment by processchemist — January 17, 2012 @ 9:57 am

  9. This setup is amazing. It’s so clean!!! lol

    Comment by deathbypuppy — April 20, 2012 @ 1:35 am

    • yeah we clean it in a rather caustic base bath that does a good job at removing stuff from glass

      Comment by milkshake — April 20, 2012 @ 11:50 am

  10. Ah, a variant on Zhang’s methodology I take it? ( )

    You might want to also get dispersity values (i.e. what was PDI, IUPAC changed it in 2009) via MALDI-ToF (low MW) or MALS (high MW) as column broadening on the SEC is an issue at such low D’s.

    Comment by Bryn Monnery — July 15, 2012 @ 9:43 am

    • A correct guess you made – I should add that the method works fine for some alcohol initiators but the scope is pretty limited, and K-naphthalenide is far from ideal for scale up so we moved away from this initiation system in the end… MALDI based PDI assay would be great but right now we are paying something like $50/sample to get on a MALDI MS instrument that does not belong to us so we don’t use it on routine basis. GPC triple detector does a reasonable job in this particular case. It gets trickier with hydrophobic and ambiphilic polymers.

      Btw, you have an interesting thesis research project – what are you doing to bring about the endosomal escape, once your polyplex gets internalized? Getting your nucleic acid inside the cell is just one of several things that need to happen for this to work. Are you using plasmid DNA or do you work with siRNA? (Turns out, their polymer-complexing properties are very dissimilar.)

      Comment by milkshake — July 16, 2012 @ 8:42 pm

  11. Writing up now and never got round to touching DNA. We got hung up making uniform l-PEI with no tailing or shouldering (like your PEG, the parent poly(oxazoline)s had tall SEC eluograms with PDI = 1.01 in the LS cell, beautiful 1H-NMR). Sisyphian task – took 2.5 years of my 3 years lab time. However i’m familiar with the issue and I’d caution against making the assumption that polyplexes of any stripe enter via endocytosis at all. If they do the assumption is that they exit by the “proton sponge hypothesis” – the polymer buffers, increases the osmotic potential and the endosome bursts. Appears to be untrue though, with localised defects causing the rupture (the name bandied about but not published is “the sting effect”). My PhD thesis (which I’m writing now) ends at an experiment to determine the chemical underpinning of the “sting”, but it’s nowhere near publishing so can’t really say much about it. Sorry.

    Comment by Bryn Monnery — July 23, 2012 @ 2:26 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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

The Shocking Blue Green Theme. Blog at


Get every new post delivered to your Inbox.

Join 163 other followers

%d bloggers like this: