Org Prep Daily

July 12, 2010

Look ma, no bromine!

Filed under: mechanisms — milkshake @ 3:57 pm

After heating up some simple 2-chloro-4-amino-5-bromopyrimidines with an excess of p-MeS-aniline, I got the corresponding 2-anilino-substituted products  – but with the bromine clipped off. The aniline was apparently the reducing agent in this case, producing lots of deep-blue colored oxidized aniline stuff in the process.

I suppose it has to do with the chemical similarity of this aniline with the leuco form of methylene blue, a known reducing agent.


  1. Maybe try the aluminum amide?

    Comment by Al — July 13, 2010 @ 2:33 am

  2. I observed the same debromination of pyrimidine when tried to deoxybrominate susbtituted 5-bromo-4-hydroxypyrimidine. I just wonder how POBr3 can act as a reductant.

    Comment by petr — July 13, 2010 @ 8:02 am

  3. We’ve seen something similar when we used a steel pressure vessel for the reaction chamber. In that case the amine was ammonia instead of an aniline.

    Comment by Larry — July 13, 2010 @ 8:10 am

  4. Aaaah, kinase inhibitors…

    Comment by kinase — July 14, 2010 @ 5:02 pm

    • nothin’ to be ashamed of: we do what we must because we can.

      Comment by milkshake — July 14, 2010 @ 6:17 pm

  5. Is it possible to perform a Sonogashira coupling with a substrate (the aryl iodide part) that has a free secondary amine further off? Or do primary and secondary amines have to be protected?



    Comment by Javaslinger — July 19, 2010 @ 10:21 pm

    • I was doing Sonogashira with these diaminobromopyrimidines with an unprotected amino and the coordinating properties of the substrate were definitely a problem – but I found a nice system with PdCl2.Dppf and NEt3 (1.5 mol% CuI) in THF at 60C. This system is supposed to work with simple aryl bromides at RT but in this case heating for 1 day was necessary, and it went very cleanly. Though when I tried the corresponding iodo-diaminopyrimidines, the reaction was initially faster (with the Dppf.PdCl2 catalyst system) but it stopped at partial conversion – I think iodide coordination shut down the already-heavily coordinated Pd. Also NEt3.HI seems quite soluble in THF whereas NEt3.HBr precipitates out. So in this case (with unprotected aminos) I would recommend using the bromo starting material:

      by the way, if you want to cyclize the Sonohashira product to azaindoles, a procedure is here:

      Comment by milkshake — July 20, 2010 @ 7:36 am

    • I have done sonogashira couplings of p-iodoanilines with TMS-acetylene without any protecting group, in good yield. Toluene, room temperature, 1.5-3 mol % Pd(PPh3)2Cl2, same of CuI, distilled triethylamine as base. 100% conversion.

      Comment by opsomath — July 20, 2010 @ 2:04 pm

  6. Pardon the thread hijack,

    Anybody have experience using nHexyllithium?

    It sounds appealing in that its essentially non-pyrophoric and supposedly has reactivity equivalent to nBuLi. However, nobody seems to use it much. I was able to dig up one reference using it to make LDA (which I would like to do) but its relative reactivity was not discussed.


    Comment by MrOrange — July 26, 2010 @ 11:54 am

    • I remember that some pharma process people like to use Hex-Li on scale in place of BuLi, both these reagents are supposed to be practically equivalent for deprotonations/halogen exchange but Hex-Li does not produce butane gas on quench/warm-up. I only saw it in a paper though – I never used HexLi myself

      Comment by milkshake — July 26, 2010 @ 3:22 pm

  7. No so much problems with BuLi on scale as long as you have a flare or the equivalent to burn the butane (that is stripped usually during distillations, venting out from the reactor or from the vacuum system). Once upon a time cyclohexyl lithium was the safest choice, but the problem is always been the price.

    Comment by processchemist — August 5, 2010 @ 10:04 am

  8. Hi Milkshake,

    I was wondering if you knew any literature resources that describe general strategies for optimization of a new reaction. All the stuff I have found detail statistical methods that deal with continuous variables and not discrete variables. I feel like I am throwing shi# at the wall to see what sticks and there must be a better way. Thanks!

    Comment by Blue — September 2, 2010 @ 5:30 pm

    • hi Blue, I do not have a lit reference but I did some methodology optimization myself. One rather important thing was to change one parameter at time (if you do several changes at once you do not know what did what). It helps to be reasonably organized, try things in a methodical fashion – go through the list of catalysts, solvents, reaction temperature etc. When you are optimizing the conditions a really useful thing is to develop a good way to analyze your reaction without having to work it up and isolate the products – that saves lots of time and also allows you to run a small-scale experiments that are just enough to weight out the starting materials and set up conveniently. When you are done with the analysis you throw the reactions away. For example one time my products were quite volatile but the product enantiomers separated nicely on chiral GC, so only thing I needed to measure the yield and ee was to find an internal standard and do a calibration on the peak integral – that way I could run and analyze several experiments a day (that helped to fill up the tables of reaction parameters vs Y and ee rather quickly.)

      Also if your reaction has problems (side reactions, poor conversion, poor ee or dr) it is helpful to try to understand it mechanistically. You can make a hypothesis about the effect of polarity/coordination ability of the solvent on the presumed transition state, the effect of temperature, kinetically vs thermodynamicaly controlled processes – and then see if changing the reaction conditions has an effect that goes along with your hypothesis.

      I should mention that some reactions are not easy to optimize – especially when there are too many variables and more-than-one rate-controlling steps happen in the catalyst cycle that are hidden from your view – so you may want to try to find as close lit precedent as possible.

      Comment by milkshake — September 3, 2010 @ 12:57 pm

      • Thanks Milkshake.

        Comment by Blue — September 6, 2010 @ 11:10 am

  9. Somebody knows if it is possible to purify by silica gel flash chromatogrphy 3-stannyl-pyridines? Or they occours severe degradation?

    Comment by madforit — September 8, 2010 @ 12:35 pm

    • I did not work with ArSnBu3 compounds but a colleague next to me did – and he columned them on silica. I remember though there is some strange stability issue with boronic acids derived from pyridine, apparently they de-borylate easily – I hope its not something you are running into.

      Maybe you can try to use de-activated silica (slurry the silica for the column in your mobile phase, add dropwise some triethylamine with stirring, in a quantity something like 1 mL NEt3/100g of silica, stir for few min and pack the column with the slurry).

      Comment by milkshake — September 8, 2010 @ 2:07 pm

  10. Thank you so much, milkshake.
    Anybody have experience with alkilation of indole type nitrogen with bromoethylamine derivatives?

    Comment by madforit — September 10, 2010 @ 9:42 am

    • I have not done it but I have seen aminoalkylation of phenols with chloroethyl amines. Indole NH is quite acidic so even a mild base like NaOMe will deprotonate it. I would use DMF, 0C, add 3 equivs of NaH in mineral oil, stir for 30 min, then cool to -20 and add 1.1 equiv of chloroethylamine hydrochloride as a solid and stir it on the melting ice bath to RT overnight.

      Comment by milkshake — September 10, 2010 @ 10:45 am

    • KOH in DMSO has always worked very well for me: see Perkin 1, 1973, 499. This won’t work so well if your amine is unprotected though.

      Comment by Martyn — September 10, 2010 @ 11:11 am

  11. Thank you so much guys.

    Comment by madforit — September 10, 2010 @ 11:36 am

  12. Hi Milkshake,
    Off topic, but you mentioned using anhydrous DMSO in a comment in “In the Pipeline”. Could you let me know how to make DMSO anhydrous?

    Comment by chacko — September 13, 2010 @ 9:36 pm

  13. Hey Chacko, you should check out an awesome set of papers by Smithers on drying solvents. He worked in Malaysia, so he should know. Here’s the one on DMSO – JOC 1978, v. 43. p 3966

    Comment by James — September 14, 2010 @ 2:09 pm

    • dry DMSO: I would add a spoon of CaH2 (you can also use NaH in this case), stir the mix on a 60C oil bath for an hour or so, then do vacuum distillation from a 60C bath . You don’t want to heat the DMSO much or it will become smelly from Me2S.

      Comment by milkshake — September 15, 2010 @ 12:21 pm

      • Thanks a lot. I’ll try it out

        Comment by chacko — September 21, 2010 @ 4:40 am

  14. Hi MS,
    Trying to couple functionalized pyrrolidine with 2-chloro-5-fluoro-4,6-dimethoxypyrimidine using different Buchwald amination conditions (Ba(OH)2 (different bases tried)/ Toluene/water / 2(2-ditertbutylphosphine)-biphenyl-palladium(II)acetate but gave only 20% conversion. Need to try with X-Phos/BINAP/S-Phos and also need to vary the solvents and bases. Do you have any suggestions for this transformation? I am unable to synthesize corresponding Iodo derivatives as substrates.

    Alternatively, I have been trying to do SNAr reactions. What could be the better LG (chloro) for SNAr reactions with 2-chloro-5-fluoro-4,6-dimethoxypyrimidine. F at 2-postion would be better for SNAr reaction but difficult to prepare for this system. Do you have any suggestions other than F?



    Comment by marto — May 11, 2011 @ 4:49 am

    • I think your di-methoxy-substituted chloropyrimidine has strongly coordinating properties. For such substrates Buchwald biaryl-derived class of hindered phosphines with Pd does not work well. (With Buchwald ligands the rate acceleration is achieved by the bulky phosphine destabilizing bis-phosphine-Pd complex; mono-phosphine-Pd complexes even with a very hindered phosphine are more active catalysts than di-phosphine complexes: they insert faster into Ar-X because they are less coordinationally saturated. This acceleration effect is ruined if your substrate and product bind strongly and coordinationally saturate Pd – using more bulky phosphines then actually makes things worse).
      A good way to deal with such strongly coordinating substrates is to use rigid diphosphine ligands like Xanphos or Dppf and heat the mix up.

      In your case I would use Xanphos, a very robust diphosphine, as a (R3P)2Pd complex. For example, try Xanphos (6 mol%) with Pd2(dba)3 (3 mol%) (it has two Pd equivalents) and excess of Cs2CO3 solid as a slurry in peroxide-free THF or dioxane. You may need to run your reaction in a pressure vial above 100C because the reflux temperature may not be high enough.

      But there is a simple alternative that I would suggest to try right away: A direct SN-Ar without any transition metal. Pyrrolidines are superb nucleophiles. Even though your pyrimidine is likely a weak electrophile, chloropyrimidines can be activated by a catalytic amount of acid especially in protic media. I suggest that you mix your chloropyrimidine with 1.1 equivalent of your pyrrolidine in a moderately small volume of 2-propanol, add 2 drops of methanesulfonic acid and heat the mix in a pressure vial to 100-140C. Then dilute with water and precipitate with bicarbonate (there will be foaming). As a solvent in this reaction, 2-propanol can be replaced with 1-methoxy-2-propanol which is higher boiling. (The reaction is easier to sample and monitor by HPLC if you can avoid using a pressure flask, by going to a higher-boiling solvent)

      Comment by milkshake — May 11, 2011 @ 12:18 pm

  15. Hi MS,

    Thanks very much for the suggestions.



    Comment by marto — May 11, 2011 @ 1:19 pm

  16. Hi MS,

    I am having the problem of the cleavage of sulfamide group (2- position) in a pyrimidine ring during the Aldol condensation of substituted pyrimidine aldehyde (5-position) and carbanion generated from aliphatic keto compound (NaH used as base) under basic conditions. I am unable to reduce this side product (5-10%) and please suggest me how it can be minimized.



    Comment by Marto — April 22, 2012 @ 7:57 am

    • how is the sulfamide connected (is it 2-pyrimidinyl-SO2NR2 or is it 2-pyrimidinyl-NR-SO2R) ? Does the sulfonamide group have free NH ?- What is the structure of the cleaved product (if you know it)? These are important details that are unclear from your description.

      By the way, you can do aldol reaction under non-basic conditions, for example Mukaiyama aldol, with silyl enol ether and a Lewis acid catalyst like Cu(Triflate)2

      Comment by milkshake — April 22, 2012 @ 1:46 pm

  17. Sorry for the poor description.

    It is 2-Pyrimidinyl-NR-SO2R and the NH is protected as alkyl group. The cleaved product is without sulfonamide group and aldehyde group being converted to acid.

    I have summarized the problems as follows

    (a ) I beleive that aldehyde is converted to acid by a Cannizaro type reaction (I am looking for alcohol as other side product as well). I need to avoid this impurity by some means to improve the condensed product. One important change what I can think of is to reduce the contact time between aldehyde and base. Please suggest me if you have any experience.

    The reaction rate is maximum in the initial 1 h. After wards the degradation of aldehyde is happening instead of condensation.

    (b) My aldehyde got very poor solubility in many solvents. Please suggest me some polar aprotic solvents for the aldol condensation and if you have any suggestions for coordination solvents as well . Carbanion also have very less solubility in many solvents.

    Presently, I am using > 20 V for the reaction, this is not ideal in terms of process perspective

    (c) Mukiyama aldol reaction is unsuccessful for unknown reasons. The route change is not allowed at this stage as enough efforts have been done.

    Please help me.



    Comment by marto — April 23, 2012 @ 10:51 am

    • Unfortunately this kind of sulfonamide cleavage is going to be very facile, tertiary sulfonamides are cleaved even with a mild base like K2CO3+ methanol, and with other nucleophiles like amines and alkoxides easily and in case of pyrimidine substiution you have this problem exacerbated with strong electron withdrawing effect (pyrimidine electron deficiency is comparable to dinitrophenyl so the corresponding pyridiminylamino anion will be a great leaving group) and the coordination effect of ring nitrogens that can direct metal alkoxides is also bad news. I believe the desired aldol product that ends up as alkoxide anion is in fact the agent that is desulfonating your product. So your problem is of a systemic character and a minor change like different solvent or base etc is not going to help, you need to either change the reaction mechanism or your aldehyde or enolate substrate structure.

      I think your only chance if you dont want to change the reactant structure is to go less basic/more Lewis acidic when generating the enolate. I would try TiCl4 1 or 2 equivalents, with 2 to 4 equivalents of iPr2NEt at low temperature maybe -78C in dichloromethane, to enolize your ketone, and then add your aldehyde as a solid and warm up to room temperature. Another (much milder and easier to work up system) would be several equivalents of ZnCl2 or ZnBr2 with triethylamine, at 0 oC. Another (less friendly workup) agent would be Bu2BOTf and iPr2NEt, a low temperature. Then there are things like silicon tetrachloride enolates – see old work from Denmark and Evans group on Lewis-acid + tertiary amine promoted aldol reactions. There was also a version promoted by BF3.Et2O if I remember correctly

      Of course if you use Lewis acids to promote aldol condensation you cannot use dipolar aprotic solvents, or protic solvents so DCM or THF would be your best bet. Since you have several electron-pair donors in your aldehyde substrate you may end up using several equivalents of Lewis acid.

      Sorry I can’t help you more because I have not done any of this by myself/

      Comment by milkshake — April 23, 2012 @ 7:52 pm

      • I think probably the best process-friendly modification would be to switch your ketone to an alpha bromoketone derivative and do a simple Reformatsky with zinc dust

        Comment by milkshake — April 23, 2012 @ 8:48 pm

  18. Hi MS,

    Thanks very much for the suggestions.


    Comment by marto — April 23, 2012 @ 11:19 pm

RSS feed for comments on this post. TrackBack URI

Leave a Reply to opsomath Cancel 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: