3554
V. P. Fitzjarrald, R. Pongdee / Tetrahedron Letters 48 (2007) 3553–3557
pling reactions with other functional groups, there were
no accounts of the reactivity or overall effectiveness of
this process utilizing benzoic acids as starting materials.
To our knowledge, this methodology has not been ex-
plored and we now wish to report our findings regarding
this new application for the Mitsunobu reaction.
Next, we examined the effects of differentially-substi-
tuted benzoic acids as outlined in Table 3. Yields were
generally good overall with ortho- (entry 1) and meta-
(entry 2) substituted benzoic acids working well within
the context of the coupling reaction. Additionally, the
use of electron-rich benzoic acids such as 4-ethoxyben-
zoic acid (26, entry 3) did not hinder the C–O bond
forming event leading to phenyl ester product 29 in
excellent yield. A noteworthy aspect of this reaction is
that ortho-substituted benzoic acids (entry 1) demon-
strated a negligible effect on the efficiency of the
coupling process contrary to that observed when
ortho-substituted phenols were employed (Table 2, en-
tries 4–6 and 10).
Our initial efforts in the development of this methodol-
ogy began with screening various reaction solvents as
well as investigating the effects of temperature as noted
in Table 1. For the coupling of o-toluic acid (3) with
p-cresol (4), taken as our standard reaction, we found
that most of the commonly utilized solvents for the
Mitsunobu reaction fared poorly in terms of reaction
yields. For example, use of methylene chloride (CH2Cl2,
entry 1) or benzene (Ph–H, entry 2) resulted in very low
conversions to the desired coupled product. Further-
more, use of N,N-dimethylformamide (DMF, entry 3)
gave only a 29% yield of ester 5. However, upon employ-
ing tetrahydrofuran (THF, entry 4) as the reaction sol-
vent, the desired phenyl ester 5 was produced in a
synthetically viable 60% yield. Moreover, we were extre-
mely pleased to find that further investigations involving
refluxing THF (entry 5) resulted in a dramatically
increased yield, 99%, of 5.23
The Mitsunobu reaction is widely accepted to involve a
triphenylphosphine–azodicarboxylate adduct such as 30
illustrated in Figure 2a. Following nucleophilic addition
of alcohol 31 to produce phosphonium ion 32 and
hydrazine byproduct 33, an SN2 reaction occurs involv-
ing carboxylate anion 34 resulting in inversion of config-
uration for the isolated ester product 35. In terms of the
reaction reported in this Letter, this pathway would
require a SN2 displacement occurring at a sp2-center.
We believe this to be an unfavorable and therefore unli-
kely process. As such, we believe that the formation of
phenyl esters from benzoic acids and phenols under
Mitsunobu conditions proceeds via a non-classical
mechanism.
After securing the optimal reaction conditions, we
turned our attention to a thorough exploration of the
substrate scope. As outlined in Table 2, we sought to
determine the effects of incorporating various functional
groups as well as differing substitution patterns on the
phenol component. As expected, the reaction works best
with unhindered phenolic nucleophiles (entries 1–3 and
7–9) which generally provided yields in excess of 70%.
As the steric hindrance surrounding the reacting nucleo-
philic oxygen atom on the phenol increases, we observed
a decrease in the overall efficiency of the process with
ortho-substituted phenols (entries 4–6 and 10) resulting
in lower yields than the corresponding meta-substituted
phenols (entries 3 and 8). It is worth noting that the ste-
rically encumbered 2,6-dimethyl phenol (9) provided the
coupled product in a synthetically useful 50% yield, a
demonstration of the overall robustness of this process.
Furthermore, the reaction is tolerant of both strong
electron-donating (entries 7 and 8) as well as electron-
withdrawing (entries 9 and 10) groups.
With this in mind, we invoke the existence of an acyl-
oxyphosphonium ion such as 37 (Fig. 2b) which
originates from adduct 30 by displacement with
carboxylate anion 34. Next, reaction with phenoxide
anion 38 via an addition–elimination manifold produces
the corresponding phenyl ester 39. The intermediacy of
acyloxyphosphonium ions was previously postulated
by Hughes et al.24,25 in his studies on delineating the
mechanism of the Mitsunobu reaction. Later, DeShong
and co-workers26,27 also suggested the existence of an
acyloxyphosphonium ion when he observed that lacton-
ization of bicyclic lactones under Mitsunobu conditions
resulted in either inversion or retention of configuration
depending upon the steric environment surrounding the
nucleophilic alcohol functional group. DeShong and
co-workers went on to demonstrate that treatment of
Table 1. Solvent and temperature effects on the coupling of benzoic acids with phenols via the Mitsunobu reaction
Entry
Solvent
Temperature (°C)
Yielda,b (%)
1
2
3
4
5
CH2Cl2
Ph–H
DMF
THF
25
25
25
25
65
20
17
29
60
99
THF
a Products obtained were >95% pure by 1H and 13C NMR.
b Reported yield for 3.6 mmol scale reaction.