[Warning, long post!]

Dear all,

I am confronted with a seemingly easy transformation. I want to remove a fluorine atom at a fluor phosphonic acid salt. The usual method to achieve this is the basic hydrolysis or the hydrolysis with acid. These methods are not applicable here.

See also: The Hydrolysis of Phosphinates and Phosphonates: A Review https://www.mdpi.com/1420-3049/26/10/2840

  1. Base will not hydrolyze the compound, because it is shielded from nucleophilic attack.
  2. Acid will deteriorate the compound and is also not applicable.
  3. I recently tried to achieve hydrolysis by TMSBr cleavage. I see the intended mass in LCMS, but the compound is likely to be contaminated by other polar species. Also, during the workup traces of acids may emerge. I did the reaction with TMSBr at 60 degrees. Maybe other conditions would lead to a different outcome, but this is not explored yet.


Because these reactions are neither robust nor is the outcome easily quantified in terms of yield and purity, I tried the attachment of a cleavage group like benzyl, that does not require acidic or basic conditions.


This route works nicely in theory, but in reality, I see almost no product. The concept is elegant, but performing 3 reactions to get the final product will lead to a major loss of the final product.

Therefore I still have not a way to get the structure I want yet in a robust fashion. Waylander from this board recommended the use of Siylating agents as a temporary cleaving group and the concept I find to be interesting.


I already observed with other esters, that in the moment the negative charge is neutralized, the fluorine will leave the molecule eagerly. The advantage here is definitely, that a TMS-Ester may also be applicable for easy cleavage, even after the fluorine left the molecule. I did not have the opportunity to test this strategy yet, but I am a little bit concerned, that the -OTMS will be cleaved before the fluorine leaves, leaving me right where I began. However, I did further research on the topic, and I found some interesting reactions where fluorine itself will act as a leaving group.


If I receive the double TMS ester, I would have an easy time, because this would go easily in the organic phase in aq/DCM workup and could be easily purified from any excess fluoride and inorganic byproducts and then could be hydrolyzed with water to the salt in the next step under basic conditions with ammonia.


Do you think double silylating is applicable? Would it be feasible to use for example a mixture of diethylaminotrimethylsilane and Hexamethyldisiloxan (for the second silylation step)? I don't care what the intermediate is, as long as everything can be cleanly hydrolyzed to the free phosphonate salt. The reagents themselves must not introduce acid into the system when in contact with water and all byproducts must be volatile. I found a lot of different silayting agents here, https://technical.gelest.com/brochures/silicon-based-blocking-agents/general-silylation-procedures, but I am not experienced with the usage of such chemicals with individual pros and cons.

  • $\begingroup$ I tried today to silyate my compound in excess with HMDS unter reflux conditions. In LCMS I did not find the TMS-Ester or the free phosphonic acid. I guess it might be possible that the TMS-ester, if it had been formed, just immediately hydrolyses during the analytic run at pH 2.8 for 11 min. I will try Si NMR, I would assume that from my starting material I will only observe one other SI signal and will also do an NMR of HMDS for reference. $\endgroup$
    – raptorlane
    Mar 27 at 18:59
  • $\begingroup$ I fully believe it @raptorlane $\endgroup$
    – user144814
    Mar 29 at 4:57
  • $\begingroup$ What do you believe? 🙂 $\endgroup$
    – raptorlane
    Mar 29 at 16:28
  • $\begingroup$ I think the silylation strategy is interesting, but HMDSO is not a strong silylating agent, so I don't think you'd get your desired fluoride elimination in the last scheme. Have you considered using potassium trimethylsilanolate (KOTMS)? I'm not fully convinced when you say SN2 can't be done due to shielding, because phosphorus is big. You could just be deactivating it from nucleophilic attack after deprotonation. If you silylate the OH/O- and then heat with KOTMS in, say, THF or dioxane or MeCN, you might displace the fluoride and crash out KF. $\endgroup$ Mar 30 at 1:17
  • 1
    $\begingroup$ Let us continue this discussion in chat. $\endgroup$ Mar 30 at 2:52


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.