Recently a discussion came up about the possibility of transesterification of salicylic acid with ethyl acetate.
Nicodem has argued that this is strongly thermodynamically unfavoured because phenol is a special case being an aromatic 'alcohol'
and so this simply won't go. I am paraphrasing here off course.
I think that thermodynamic arguments are all about numbers and one should use calculations as a basis because sometimes the results may be surprising
and intuitive understanding can only go so far.
Now as one might expect, I did not manage to find all the thermodynamic data for the specific case of the salicylic acid transesterification. So I
begun hunting for other model cases.
Sadly, neither was data complete for the simple case of phenyl acetate, in fact, I could not find complete data for ANY esters of phenol.
So I decided to search more broadly and consider functional group metathesis reactions in general where a phenolic swaps out for an alkyl group and an
alkyl alcohol becomes the leaving group. In principle, this thermodynamic suppression effect should be extendible to other such cases.
So since I had not luck with ester data, I begun looking at ether reactions. In this case I was fortunate. Complete data was available for the
following hypothetical reaction:
2PhOH(l) + Et2O(l) <=> Ph2O(l) + 2EtOH(l)
all in the liquid phase.
So for example, one would reflux phenol dissolved in excess diethyl ether for days with a strong acid added as catalyst and let the ether metathesis
run it's course. I do not have data for the mixed ether: phenyl ethyl ether, so I will only consider the extreme case of both groups metastasising to
form diphenyl ether. Surely if this proceeds satisfactorily on thermodynamic arguments, then the half case would do even more so.
Now the analogy to salicylic acid transesterification should be self evident, BUT, this case should be thermodynamically double trouble as we are in
fact employing two phenols and making 2 ethanols... it should be an even harder thermodynamic sell by nicodem's arguments.
Ok, so I have done the calculation and attached the spreadsheet. I would appreciate checking because it's been a long long time that I have done such
calculations and could very well have made mistakes.
Anyhow, the data was all taken from the NIST chemistry webbook.
The results for the liquid phase reaction is:
Keq = 0.015 @298K
As nicodem argued, clearly thermodynamically not favoured, however, also not so low as to prevent such a reaction give some yields
when using excess of the diethyl ether.
So I soldier on...
I further calculated that if one employs a tenfold molar excess of diethyl ether to phenol, not a particularly large excess if one
employs the ether as solvent, then the equilibrium conversion of phenol was calculated to be 22%.
You might say that this is quiet poor and it is, but remember that this is a double phenolic substitution and so doubly suppressed, therefor not so
bad if you consider this to be a worst case scenario.
I therefor argue for now that there is still evidence to suggest that the transesterification of salicylic acid with excess (tenfold plus) ethyl
acetate may be possible and should be experimented on for confirmation.
While it would yield a mixed product, this could still be good enough for instructional purposes, especially if students go on to do a purity test. It
could also simultaneously be used to demonstrate thermodynamic limitations.
Note to moderators: I would prefer this merged with the closed thread on the alternative aspirin synthesis.
It would be only fair to re-open that and give nicodem the opportunity to critic this which is off course also welcome to do so here
in this thread if not merged!
A friend of mine, let me call him 'deltaG' is volunteering some molecular
modelling.
Since the only data missing for the phenyl acetate transesterification is the entropies of formation of phenyl acetate, it would be nice if these
could be determined by ab initio calculation.
Preliminary result of a high level DFT calculation is S = 297J/mol.K. A solvation model was applied with ethyl acetate's dielectric constant, however
this S value seems too high for my liking (it would be great if it were true because it would mean that the transesterification would proceed nicely),
so I remain skeptical of it. It looks like it could be a gas phase value and I need a liquid phase one?
Aside 2:
I was somewhat jubilant to find that diphenyl carbonate was fully thermodynamically characterised, fortuitously, so too propylene carbonate, so I was
excited that in my study on these various phenol-alkyl metatheses about some oxygenated functional group, I might consider the following hypothetical
case study equilibrium as well:
After all, carbonate groups are second cousins of esters
Sadly though, I hit a wall in the most unlikely place in this regard... no dSf data for propylene glycol... can you believe that?! So if anyone has a
value for me, I would GREATLY appreciate it!
[Edited on 25-10-2013 by deltaH]Nicodem - 25-10-2013 at 07:45
While I do not underestimate the usefulness of computational chemistry to answer certain questions that nature makes hard to answer, I still believe
in the first rule of scientific research: "When in doubts, ask the nature". Being an old fashioned I still trust experimental results more than any
computational chemistry software.
See table 3, entry 7 of DOI: 10.1016/j.poly.2008.08.010 (attached). By doing some literature work, you can find other articles of phenyl esters
transesterifications with aliphatic alcohols, if you are still interested about the direction the reaction takes. But in short, quoting the article
(and summarizing all the other articles): "All esters were completely converted to the corresponding products". And this demonstrates the second rule
of scientific research: "One day in the library can save you months of work in the laboratory". (and sometimes one minute of literature work can save
you a year of laboratory work)
Thank you Nicodem for this information. It is very useful (EDIT 1 Potentially not so useful see below)! Sadly I have no access to
much resources but I completely agree with you, experimentation is king. When I was still a student, my supervisor used to love saying "nature doesn't
understand the theory" in such cases I try my best to look into things with what
I have.
At the moment, I am mostly confined to hypothetical investigations, sometimes they are simple to try out at home in the context of my means, in such a
case I will try to do so. I like posting my ideas here because others help me refine them and someday when I have the means, I can implement them.
I was hoping that somebody could volunteer trying this reaction in reverse to see its feasibility, i.e. reflux aspirin with equimolar amounts of
alcohol in a large excess of acidified ethyl acetate and see how much salicylic acid forms.
Thank you again for the time you have invested in this, apologies for upsetting you!
EDIT 1:
Ok just read the paper, they reacted phenyl benzoate with excess methanol (the reverse reaction to what I want) and it went to completion, but the
alcohol was in excess here, so how does that help? I must be missing something, please advise.
So far the results of my ether case study suggests that K~10-2, so if you use an alcohol in excess you would of course fully generate phenol?
But I also showed that if you use a tenfold excess of the ethyl acetate in that case, you can generate 22% conversion of the phenol.
EDIT 2:
I just wanted to stress that my calculations weren't based on the molecular modelling result, that result was too good to believe They were based on actual NIST dHf and dSf data for the ether case (only good data I
could find).
[Edited on 25-10-2013 by deltaH]watson.fawkes - 25-10-2013 at 10:06
As I see it, you haven't shown anything. You've made a prediction. Go verify it or find someone else's verification if you want to think that you've
shown anything.deltaH - 25-10-2013 at 10:30
watson.fawkes, I did a thermodynamic calculation based on experimentally measured results. Now, there is always a chance that I made
a mistake in the calculation (hence why I include the calculation in the attachment), or the experimental data is flawed. Sadly, thermodynamic data
of compounds often are flawed, but in this case, there are multiple measurements and they are very close to each other from many authors (as reported
on the NIST database). This means that I have a high confidence in it until proven wrong.
kmno4 - 25-10-2013 at 10:49
I have found an article (JCED), slightly off-topic, but interesting:
"Phenyl Acetate Preparation from Phenol and Acetic Acid"
I will try to repeat this procedure if I have time.
Hm....
Attachment: 1dd.pdf (140kB) This file has been downloaded 6168 times
deltaH - 25-10-2013 at 10:53
Thank you for that kmno4. I'm having a read through now.
...the title is funny
Ok read it. Yes this looks very promising, there yield of 55% with ten molar excess acid. Again, not the exact case of transesterification, but
certainly relevant in the bigger context of this thermodynamic study with phenolics.
Any chance I persuade you to try the reverse experiment I suggested?
Reflux equimolar aspirin and ethanol in ten fold excess of ethyl acetate with small addition of H2SO4? Analyse to see how much aspirin is left after a
few hours.
Please
[Edited on 25-10-2013 by deltaH]Mildronate - 25-10-2013 at 13:06
if it doesnt work termodynamical it doesnt work. i will calcuate gibs energy next week.deltaH - 25-10-2013 at 13:27
Agreed Mildronate and that would be appreciated... please include calculations when you do for instructional purposes and thanks.bfesser - 27-10-2013 at 09:28
I commend deltaH and others (like blogfast) who predict and explain chemistry through thermodynamics. Yes, experiment decides but what is wrong with
attempting to put some of our predictions and results on a sound mathematical basis?
Professionals take this approach frequently to save time and money: