The reaction of aspirin with base

Article abstract of DOI:10.1016/j.tetlet.2011.04.096

Aspirin anion appears to exist only fleetingly, rearranging via acetyl transfer to the ortho carboxylate group, as indicated by IR, UV and NMR. The resulting mixed anhydride cyclises to the more stable bicyclic orthoacetate isomer, a process facilitated b

Full text of DOI:10.1016/j.tetlet.2011.04.096

Tetrahedron Letters 52 (2011) 3561–3564
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
The reaction of aspirin with base
⇑
Sosale Chandrasekhar , Honnaiah Vijay Kumar
Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Aspirin anion appears to exist only fleetingly, rearranging via acetyl transfer to the ortho carboxylate
group, as indicated by IR, UV and NMR. The resulting mixed anhydride cyclises to the more stable bicyclic
orthoacetate isomer, a process facilitated by time and increasing pH. Mechanistic possibilities are dis-
cussed to explain these intriguing observations.
Received 3 February 2011
Revised 20 April 2011
Accepted 24 April 2011
Available online 30 April 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Acetylsalicylic acid (1b, Scheme 1) is an over-the-counter phar-
maceutical that has acquired the status of a household remedy, its
generic term ‘aspirin’ having passed into common parlance.1 Its
well-known analgesic effect has now been supplemented by the
recognition of its anti-coagulant and anti-inflammatory properties.
The simplicity of its molecular structure, however, belies the rather
complex mechanisms mediating its chemical reactivity and phar-
macological action.2,3
A chemical reaction of particular relevance to aspirin’s stability
in biological fluids is hydrolysis. Early work by Edwards,4 later ex-
tended by Garrett,5 indicated that aspirin was unstable in nearly all
pH regions, the pH-rate profile being characterised by a minimum
at pH ꢀ 2.2 and a broad plateau region at pH ꢀ 5–10. An initial
mechanistic proposal based on intramolecular nucleophilic cataly-
sis by carboxylate was later overturned by the extensive work of
Fersht and Kirby,6a in favour of mechanistic general base catalysis.
It is noteworthy, however, that the balance between the two alter-
natives is a subtle one: the former mechanism, which is perhaps
‘deceptively obvious’ in the case of the parent aspirin, is actually
preferred in the 3,5-dinitro derivative.6b
In the neutral pH range, and in the high field region, only a sin-
glet resonance was observed at d 2.36. Interestingly, this was
accompanied by a second, minor singlet at d 1.76, at pH ꢀ 8.0.
The d 1.76 peak also intensified with increasing pH, reaching a
maximum ratio of 1:0.8 at pH ꢀ 12. (Overnight at pH 8.0, the d
2.36 peak disappeared leaving behind only the peak at d 1.76.)
There was also insignificant hydrolysis (if any), as 1b could be
recovered in 80% yield upon acidification of the mixture at pH
8.0 after 12 h. NMR also did not indicate the formation of acetic
acid. These changes in the high field region were accompanied
by changes in the aromatic region.
The pKa of aspirin is 3.5, so it would be practically completely
deprotonated at neutral pH.4–6 This indeed seemed to indicate that
the peak at d 2.36 was due to the O-acetyl methyl group of the as-
prin anion (2). Also, this apparently cyclised at high pH to hemior-
tho ester anion 3a, the peak at d 1.76 being attributed to the
quaternary methyl group in 3a. (The C-Me group in orthoacetates
resonates ꢀd 1.45,9 and the downfield shift in 3a may be attributed
to the electron-withdrawing C@O group.)
The protonation of 3a would lead to its conjugate acid 3b,
although the position of equilibrium is unclear. Analogous hemior-
tho acids are believed to possess pKa ꢀ 11,10a which possibly indi-
cates the predominance of 3b at most pH values employed herein.
In any case, 1H NMR did not indicate an equilibrium mixture, with
no variation in chemical shifts with pH; this, again, indicates the
presence of 3b rather than 3a.
The above—apparently straightforward—explanation, however,
does not account for the intensification of the d 1.76 resonance,
with both time and increasing pH. An alternative explanation
would be that the resonance at d 2.36 is due not to carboxylate an-
ion 2, but to some other species that interconverts with 3b at basic
pH.
Furthermore, aspirin hydrolysis has been studied by UV spec-
troscopy, but it is particularly intriguing that the deprotonation
of aspirin is accompanied by a decrease in absorption intensity.4
Salicylic acid (1a) itself displays a high intensity short wavelength
band (236 nm) and
a low intensity long wavelength band
(302 nm); upon deprotonation, both these suffer marginal hypso-
chromic shifts and lowering in intensity. In the case of aspirin,
however, increasing pH leads to a gradual weakening of the long
wavelength band at 275 nm, with its near disappearance at pH ꢀ 5.
The difference in the behaviour of salicylic acid and aspirin to-
wards base is indeed striking. In order to gain further insight, we
have studied the changes in the 1H NMR spectrum of aspirin in
D2O at various pH values, as described below (Fig. 1).7,8
An intriguing possibility is that aspirin anion (2) initially forms
the mixed anhydride 4a (possibly the source of the d 2.36 reso-
nance). On this basis 2 has only a fleeting existence. Also, the rela-
tively high pKa of the phenolic hydroxyl group in 4a (ꢀ9) ensures
⇑
Corresponding author. Tel.: +91 80 22932689; fax: +91 80 23600529.
E-mail address: sosale@orgchem.iisc.ernet.in (S. Chandrasekhar).
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.04.096

3562
S. Chandrasekhar, H. V. Kumar / Tetrahedron Letters 52 (2011) 3561–3564
H OH
O
O-
O
H
O
O
O
OR
H2O
HO-
(i)
O
O
-
CO2
CO2
H
1a-1c
O
3a
O
3b
2
(i) Base (applies to 1b)
1
a
b
c
R
H
Ac
Me
O-
O
O-
O
pH > 10
O
OR
H
O
4 R
a H
O
O
O
b
Me
HO-
O
OAc
O
O
O
4a,b
7
5
6
Scheme 1. Consequences of deprotonating aspirin 1b, the mixture essentially consisting of the mixed anhydride 4a at moderate pH, with increasing amounts of orthoester 3b
at higher pH; also, 3b apparently predominates overwhelmingly at equilibrium, and a parallel intermolecular route between 2 and 4a appears likely.
Figure 1. Scanned images (relevant parts) of the 400 MHz 1H NMR spectra of aspirin (1b) in D2O under the following conditions (13C: vide infra): (a) 1b; (b) 1b at pH 8.3 (<
10 min.); (c) 1b at pH 8.3 (12 h); (d) 1b at pH 11.8 (< 10 min.). The O-acetyl Me signal (d 2.36, originally from 1b, but from 4a in the other cases) is marked ‘x’, and the evolving
orthoester (3b) Me signal (d 1.76) is marked with an asterisk. (Solvent related peaks and integration markings have generally been elided; the horizontal axes represent d
values relative to sodium dimethylsilapentanesulfonate (DSS); Me singlet peaks are generally truncated.) Base studies. Aspirin was admixed with Na2CO3 in D2O (at 25 °C) in
the molar ratios [1b:base (pH)]: 1:1 (8.32), 1:3 (9.45), 1:6 (10.63), 1:10 (11.78); pH was measured with indicator paper, and cross-checked in H2O with a pH meter. The NMR’s
were run within ꢀ 10 minutes upon basification, and again after 12 h (Fig. 1c); acidification of this mixture led to the recovery of 1b in 80% yield, indicating negligible
hydrolysis (if any). (Fig. 1a) dC (CDCl3) 169.95 (ArC@O), 169.77 (MeC@O), 151.19 (O-CAr), 134.87 (CAr), 132.48 (CAr), 126.15 (CAr), 123.96 (CAr), 122.15 (CAr), 20.99 (CMe). (Fig. 1b;
presumed to be largely 4a): dC (D2O) 173.14 (ArC@O), 157.22 (MeC@O), 133.93 (CAr), 132.31 (CAr), 129.16 (CAr), 125.09 (CAr), 122.08 (CAr), 118.98 (CAr), 23.32 (CMe). (Fig. 1c;
presumed to be essentially 3b): dC (D2O) 176.20 (C@O), 162.24 (O-CAr), 136.71 (CAr), 133.98 (CAr), 133.18 (CAr), 122.10 (CAr), 120.69 (CAr), 119.01 (Corthoester), 26.11 (CMe).
that it is deprotonated substantially only at relatively high pH.10b
The resulting anion (5) would then cyclise via 3a to 3b.
Firstly, the formation of anhydride 4a from anion 2 may not be
mediated by 3a at all: this indicates a preferred intermolecular
pathway. Although intramolecular routes are generally preferred,
it is possible that, because of steric congestion at the ortho-disub-
stituted site in 2, the O-acetyl group adopts an unfavourable con-
formation for its internal transfer (i.e. pointing away from the
carboxylate).
Also, the disappearance of the d 2.36 peak in the NMR overnight
(vide supra) indicates a considerable barrier to the cyclisation of 5
to 3a. But intriguingly, the initial rapid formation of 4a prior to its
slow but almost complete cyclisation to 3b, begs the question
why—or how—4a was formed at all! (Thus, 4a could not have been
formed via the more stable 3b, it would seem.) However, there are
two possible explanations for this apparent anomaly, as follows.
Alternatively, it is possible that protonation of 3a to 3b is slower
than collapse to 5 and 4a. Although proton transfer is generally

S. Chandrasekhar, H. V. Kumar / Tetrahedron Letters 52 (2011) 3561–3564
3563
faster than C–C cleavage, both steric and electronic factors may
facilitate this exception: the pKa (ꢀ11, vide supra) of 3a indicates
only moderate basicity, and the quaternary orthoester centre is
also congested.
The results so far do not allow a distinction between these two
alternatives. However, they do indicate the following order of ther-
modynamic stability: 3b>4a>2. The observations also indicate con-
siderable kinetic control over the proportion of 3b to 4a, a fast
(possibly intermolecular) route leading to 4a, which cyclises slowly
over time to 3b.
In fact, the formation of 4a is not unprecedented. Kinetic studies
on several analogs of 4a have been reported, and indicate (pseudo
first order) t1/2 ꢀ 103 s in dioxane-H2O.11 The equilibrium forma-
tion of 4a had also been suspected earlier, on the basis of the
marked acetylating ability of 1b in pyridine solution.12
The predominance of the mixed anhydride 4a (over 2), however,
remains intriguing, as anhydrides are generally highly reactive and
unstable towards nucleophiles. Clearly, the rearrangement of a car-
boxylic ester to an apparently more stable carboxylic anhydride
calls for explanation. There are two possible reasons for this obser-
vation. (Note that the slow formation of 3a, followed by its rapid
breakdown via 4a, constitutes the purported nucleophilic catalysis
mechanism of aspirin hydrolysis.4–6
)
Firstly, carboxylate 2 is a charged species. Indeed, ionic bonds
are generally more thermodynamically stable than covalent ones,
and carboxylate anions are well-solvated in water. However, these
effects may well be countervailed by repulsions between the neg-
atively-charged carboxylate group and the lone pairs on the oxy-
gen atoms of the ortho acetyl group in 2 (noting, again, the
congested ortho disubstituted environment).
Secondly, anhydride 4a would be stabilised by resonance-as-
sisted hydrogen bonding as indicated in 6. Similar interactions
are known to exist in salicyl aldehyde (and tautomers of b dicar-
bonyl compounds in general).13
Figure 2. Scanned images of C@O IR bands in CHCl3. 1b with pyridine: (a) 1:0; (b)
1:1; (c) 1:2; 4b only: (d).
In fact, the formation of anhydride 4a is also indicated by IR
spectroscopy (Fig. 2). When chloroform solutions of 1b are treated
with increasing amounts of pyridine, the original twin C@O peaks at
1754 cmÀ1 (acetyl) and 1692 cmÀ1 (CO2H) are almost completely
replaced by equally intense peaks at 1767 cmÀ1 and 1713 cmÀ1
[Fig. 2(a)–(c)]. The effect is clearly discernible with two molar
equivalents of pyridine. The twin bands observed are characteristic
of carboxylic anhydrides,14 and hence may be attributed to 4a. For
comparison, the O-methyl ether 4b was prepared from o-anisic acid
1c, and apparently possessed analogous spectral characteristics to
4a [Fig. 2(d)]. (The lower C@O IR values of 4a presumably indicate
chelation, cf. 6; this is also indicated by an upfield shift of
ꢀ6 ppm for the 13C resonance of the acetyl CO group occuring at d
157.22.)14
spectra, and these conclusions also explain certain UV spectral
changes observed by previous workers.
Acknowledgements
We thank University Grants Commission (New Delhi) for a D. S.
Kothari postdoctoral fellowship to H.V.K. We also remain grateful
to a referee for several constructive suggestions which led to a
vastly improved manuscript.
References and notes
The implications of these observations for the mechanism of
aspirin hydrolysis are interesting. Fersht and Kirby established
the mechanism as mechanistic general base catalysis, essentially
on the basis that the mixed anhydride 4a is unreactive (thus ruling
out intramolecular nucleophilic catalysis).6 Therefore, the forma-
tion of 4a would represent a reversible cul de sac, the hydrolysis
occurring via 2 that is present in equilibrium concentrations.
It is noteworthy, however, that the second acceleration of the
hydrolysis reaction beyond pH 10 has apparently not been ex-
plained.4,5 Interestingly, this may now be attributed to attack of
hydroxide ion on the lactone carbonyl group in 3a (cf. 7, which
indicates overall bonding changes, final products being salicylate
and acetate); or, for that matter, the attack of hydroxide ion on 5.
In conclusion, the reaction of aspirin with base appears to lead
to complex changes, essentially involving the rearrangement of the
anion to the mixed anhydride and its hemiortho ester isomer (3b).
The changes are pH- dependent with 3b predominating at high pH.
All these are indicated by appropriate changes in the IR and NMR
1. The Merck Index, 13th ed.; O’Neil, M. J., Ed.; Merck: Whitehouse Station, NJ,
2001; # 856, pp 145–146.
2. Botting, R. M. Pharmacol. Rep. 2010, 62, 518–525.
3. Vane, J. R.; Botting, R. M. Thromb. Res. 2003, 110, 255–258.
4. Edwards, L. J. Trans. Faraday Soc. 1950, 723–735.
5. Garrett, E. R. J. Org. Chem. 1961, 26, 3660–3663. and references cited therein.
6. (a) Fersht, A. R.; Kirby, A. J. J. Am. Chem. Soc. 1967, 89, 4857–4863; (b) Fersht, A.
R.; Kirby, A. J. Ibid 1968, 90, 5826–5832.
7. Experimental procedures. Aspirin (1b) was prepared by acetylation of salicylic
acid,8 purified by recrystallisation and characterised by IR, NMR and mp (132–
136 °C; lit.8 128–135 °C). O-Me ether 4b: A mixture of o-anisic acid (0.152 g,
1.0 mmol) and NaH (0.024 g, 1.00 mmol) in dry THF (10 ml) was stirred under
dry N2 for 0.5 h at 0 °C. (The NaH was calculated as a 60% suspension in mineral
oil, and was pre-washed with dry pentane). The mixture was treated dropwise
with acetyl chloride (0.085 ml, 0.094 g, 1.2 mmol) in dry ether (5.0 ml) and
stirred for 2.0 h. The mixture was diluted with ether (10 ml) and washed with
NaHCO3 solution (5 ml), followed by ice-cool water (5 ml). The organics were
dried briefly (Na2SO4), and volatiles removed in vacuo at <30 °C; the resulting
yellow semi-solid (0.151 g, 0.78 mmol, 78%) was identified spectroscopically as
4b; vmax (neat) 2947 (w), 2842 (w), 1780 (s), 1733 (s), 1601 (s) 1580 (m), 1490
(s), 1466 (m), 1438 (m), 1370 (w), 1018 (s) cmÀ1; dH (CDCl3) 7.88 (m, 1 H, ArH),
7.56 (m, 1 H, ArH), 7.01 (m, 2 H, ArH), 3.92 (s, 3 H, OMe), 2.32 (s, 3 H, –CO–Me);
dC (CDCl3) 166.99 (ArC@O), 161.50 (MeC@O), 159.85 (MeO-CAr), 135.53 (CAr),

3564
S. Chandrasekhar, H. V. Kumar / Tetrahedron Letters 52 (2011) 3561–3564
132.95 (CAr), 120.40 (CAr), 112.09 (CAr), 55.94 (OMe), 21.97 (–CO–Me) (one CAr
could not be clearly identified); HRMS (C.I.): m/z 195.0670 (M+H+, Calcd for
10H11O4 195.0657).
10. (a) Chandrasekhar, S.; Karri, P. Tetrahedron Lett. 2006, 47, 2249–2251; (b)
Smith, M. B.; March, J. March’s Advanced Organic Chemistry: Reactions,
Mechanisms and Structure, 6th ed.; John Wiley: Hoboken, 2007; pp 359–364.
11. Garrett, E. R. J. Pharm. Sci. 1962, 51, 113–116.
12. Davidson, D.; Auerbach, L. J. Am. Chem. Soc. 1953, 75, 5984–5986.
13. Krygowski, T. M.; Zachara-Horeglad, J. E. Tetrahedron 2009, 65, 2010–2014.
14. Kemp, W. Organic Spectroscopy, 3rd ed.; Macmillan: London, 1994; pp 78,
199–200.
C
8. Furniss, B. S.; Hannaford, A. J.; Rogers, V.; Smith, P. W. G.; Tatchell, A. R. Vogel’s
Textbook of Practical Organic Chemistry, 4th ed.; Longman: Harlow, 1986; pp
831–832.
9. Pouchert, C. J. In The Aldrich Library of NMR Spectra, 2nd ed.; Aldrich Chemical:
Milwaukee (WI), 1983; Vol. 1, pp 185–186.