See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/235899929

Hydrolysis of 1,8-and 2,3-naphthalic anhydrides

and the mechanism of cyclization of 1,8-
naphthalic acid in...

Article in Journal of the Chemical Society Perkin Transactions 2 · December 2001

DOI: 10.1039/b104148g

CITATIONS READS

6 38

8 authors, including:

Santiago F Yunes Hernan Chaimovich

Federal University of Santa Catarina University of São Paulo
20 PUBLICATIONS 86 CITATIONS 205 PUBLICATIONS 3,778 CITATIONS

SEE PROFILE SEE PROFILE

Luis Gustavo Dias Iolanda M Cuccovia

University of São Paulo Instituto de Química
45 PUBLICATIONS 391 CITATIONS 115 PUBLICATIONS 1,891 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Lipopeptide simulation View project

All content following this page was uploaded by Luis Gustavo Dias on 27 November 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue
are linked to publications on ResearchGate, letting you access and read them immediately.
 2

PERKIN
Hydrolysis of 1,8- and 2,3-naphthalic † anhydrides and the
mechanism of cyclization of 1,8-naphthalic acid in aqueous
solutions ‡

Teresa C. Barros,§a Santiago Yunes,b Guilherme Menegon,a Faruk Nome,b

Hernan Chaimovich,a Mario J. Politi,a Luis G. Dias a and Iolanda M. Cuccovia *a
a
Departamento de Bioquímica, Instituto de Química, Universidade de São Paulo, Brazil
b
Departamento de Química, Universidade Federal de Santa Catarina, Florianópolis, Brazil

Received (in Cambridge, UK) 4th May 2001, Accepted 21st September 2001
First published as an Advance Article on the web 24th October 2001

Naphthalene-1,8-dicarboxylic acid, 1,8-Acid, cyclizes spontaneously in acidic aqueous solution to naphthalene-

1,8-dicarboxylic anhydride, 1,8-An, and here we present an ab initio study of the reaction pathway. The effect of
pH on the hydrolysis of 1,8-An was analysed and compared with the hydrolysis of naphthalene-2, 3-dicarboxylic
anhydride, 2,3-An, to naphthalene-2,3-dicarboxylic acid, 2,3-Acid. The values of the pKa’s of 1,8-Acid and 2,3-Acid
were ca. 3.5 and 3.0, for monoanion formation, pKa1, and 5.5 and 5.0 for dianion formation, pKa2, respectively.
Fluorimetric titration demonstrated that the diprotonated 2,3-Acid, AH2, was further protonated to yield AH3⫹. The
pH–rate constant profile for 2,3-An hydrolysis showed a water reaction between pH’s 1.0 and 6.0 and a base catalysed
hydrolysis above pH 7.0. Under no condition was 2,3-An formed from 2,3-Acid. The pH dependent decomposition
kinetics of 1,8-An is complex and, below pH 6.0, the pH–rate constant profile was fitted by assuming that both
AH2 and AH3⫹ are in equilibrium with 1,8-An. The values of the equilibrium constants for 1,8-An formation
from AH2 and AH3⫹ were ca. 4 and 100 in dilute and concentrated acid, respectively. Ab initio calculations
for a possible reaction pathway connecting the undissociated 1,8-Acid to 1,8-An show a transition state
where an intramolecular proton transfer is concerted with oxygen alignment towards the carbonyl
centre. The planar intermediate is then dehydrated yielding a complex between water and 1,8-An.

Introduction thalic acid derivatives.3 Due the importance of modelling

peptide cleavage by using simple organic compounds we studied
Carboxylic anhydrides hydrolyse in water and the reaction can the hydrolysis of 2,3- and N-butyl-1,8-naphthalimides.4 Alk-
be acid or base catalysed.1–6 In aqueous solution even mono- aline hydrolysis of N-butyl-1,8-naphthalimide includes ring
carboxylic acids readily equilibrate with the corresponding opening yielding naphthalamic acid which, at low pH, cyclizes
anhydride, although the equilibrium strongly favours the acid.7,8 to the naphthalene-1,8-dicarboxylic anhydride, 1,8-An.4 The
The equilibrium constants for anhydride formation, KE = alkaline hydrolysis of N-butyl-2,3-naphthalimide, however,
k1/k⫺1 = [anhydride]/[dicarboxylic acid] [eqn. (1)], have been yields naphthalamic acid, which does not cyclize in acid.4 The
determined in several cases.9–12 analysis of imide hydrolysis in this system would be simplified
by understanding the mechanism of 1,8-An formation from
(1) naphthalene-1,8-dicarboxylic acid, 1,8-Acid.4,5
Here we present the study of the equilibrium reaction
The values of KE vary from 10⫺7 to 25 depending on the between naphthalene-1,8-dicarboxylic acid and its anhydride in
structure of the diacid. The unionized acid is the kinetically aqueous acid, the alkaline hydrolysis of 1,8-An, and compare
active species in anhydride formation.12 these data with the hydrolysis of naphthalene-2,3-dicarboxylic
The hydrolysis of cyclic anhydrides has been studied in anhydride (Scheme 1).
aqueous media and in water–solvent mixtures. There are several
correlations between rates and nature of the acid, size and con-
figuration of the anhydride ring for both mono- and bi-cyclic
anhydrides.13
The model system most extensively used to analyse particular
aspects of enzymatic catalysis covalently incorporate reactants
into the same molecule, modelling bimolecular reactions in an
intramolecular system.2 Relative accelerations, obtained by
comparisons of the second and first order rate constants can
reach values as large as 1013 in model systems based upon naph-

† The IUPAC name for naphthalic acid is napthalenedicarboxylic acid.

‡ Electronic supplementary information (ESI) available: tables contain-
ing the values of the rate constants. See http://www.rsc.org/suppdata/
p2/b1/b104148g/
§ Present address: Instituto de Ciências da Saúde, Universidade
Paulista, UNIP, Bauru, São Paulo, Brazil. Scheme 1

2342 J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350 DOI: 10.1039/b104148g

This journal is © The Royal Society of Chemistry 2001
Experimental 1,8-Acid is 15%. Hence in this pH region the anhydride–acid
equilibrium needs to be considered.
Materials
Spectrometric titrations. The pKa1 and pKa2 of 2,3-Acid
Naphthalene-1,8-dicarboxylic anhydride, 1,8-An (Aldrich), (Scheme 2) were determined from the absorbance (255 nm)
was purified by sublimation. Naphthalene-2,3-dicarboxylic
anhydride, 2,3-An, was synthesised and purified using standard
procedures.14 Naphthalene-2,3-dicarboxylic acid, 2,3-Acid
(Aldrich), was used as received. Naphthalene-1,8-dicarboxylic
acid, 1,8-Acid, was prepared by adding 0.03 M NaOH to a
water suspension of 1,8-An (0.01 M) and maintaining the
sample at 50 ⬚C for 5 h. At the end of the reaction absorbance
measurements (at 343 nm) assured total absence of the
anhydride (see below).
All reagents were PA grade or better and organic solvents
were distilled. All aqueous solutions and buffers were prepared
in freshly glass-bi-distilled water.

Methods
Absorbance spectra (UV–visible) and kinetic measurements
were obtained on a Hewlett-Packard HP8452A spectrophoto-
meter equipped with a thermostatted water-jacketed cell
holder (Microquímica MQBTZ99-20), or a Beckman DU-7
spectrophotometer equipped with a Peltier block. Fluorescence
measurements were made on a Hitachi F-2000 spectrofluoro-
meter. pH was measured with a Beckman model Φ71 pH meter.
The electrode was calibrated against standard buffers, in a
thermostatted stirred vessel, at the appropriate temperature Scheme 2
(±0.1 ⬚C). The pH of the buffers was adjusted with NaOH.
Titration of the 2,3-Acid and 1,8-Acid were made on a Radio- measured at several pH’s, by adding 0.025 mL of 1 × 10⫺3 M
meter PHM 82 Standard pH Meter equipped with a glass methanolic solution of the 2,3-Acid to 2.5 mL of 0.10 M buffer.
electrode calibrated with standard buffers at 30 ⬚C. The buffers used were borate, phosphate and citrate from pH 9
to 2.5 and HCl at lower pH’s.
The values of pKa0, pKa1 and pKa2 of 2,3-Acid were deter-
Kinetics
mined from the variation of the uncorrected fluorescence emis-
The kinetics of hydrolysis and cyclization of the substrates were sion intensities at 378 nm using the excitation wavelength fixed
followed at 343 nm for 1,8-An (ε = 12900 M⫺1 cm⫺1) and 365 nm at 300 nm. pKa1 of 1,8-Acid was determined using emission
for 2,3-An (ε = 3600 M⫺1 cm⫺1) in a quartz cell of 1 cm optical intensities at 420 nm and excitation wavelength at 300 nm.
path. The reaction was initiated by adding 20 µL of a stock Excitation and emission slits were fixed to a 10 nm bandpass
solution of the substrates prepared in acetonitrile to 2.0 ml of width. Fluorescence rectangular quartz cells of 1 cm optical
buffer. The final concentration of substrate was 1.0 × 10⫺5 M pathlength were used with a 90⬚ excitation–emission geometry.
for the reactions performed at 30 ⬚C and 1.7 × 10⫺5 M at 50 ⬚C. A methanolic solution, 0.025 mL, of 2,3-Acid 1 × 10⫺3 M was
The buffers used were borate (pH 9.0 to 10.0), phosphate added to 2.5 mL of buffer giving a final concentration 1 × 10⫺5
(pH 6.4 to 8.0), citrate (pH 2.2 to 6.0) and HCl for pH’s lower M. For 1,8-Acid, a 1 × 10⫺3 M solution in NaOH pH 11 was
than 2.0, where the H0 scale was used.15 The buffer concen- used.
trations are detailed in the tables and figure legends. The The values of pKa1 and pKa2 of 1,8-Acid (Scheme 2) were
absorbance values were stored directly on a microcomputer and obtained by absorbance measurements (320 nm) at several pH’s
analysed using the HP 8452 kinetics software. Data analysis in 1.0 cm optical pathlength quartz cells. 1,8-Acid (0.05 mL,
demonstrated first order behaviour for, at least, 4–5 half- 0.01 M) was added to 2.5 mL of 0.10 M buffer yielding a final
lives. Reported rate constants are averages of, at least, three concentration of 2 × 10⫺4 M of 1,8-Acid.
independent determinations differing by no more than 5%. Since absorbance or fluorescence values were obtained within
ca. 20 seconds of 1,8-Acid addition to the buffered solution
pKa determinations no correction on the pKa values regarding the anhydride–acid
equilibrium was necessary.
All determinations were made at 30 ⬚C.
Computational procedures
Potentiometric titrations. The 2,3-Acid was titrated with HCl Ab initio calculations. The potential energy surface for the
using 50 mL of a 0.001 M solution of 2,3-Acid made alkaline formation of 1,8-An was studied by ab initio quantum chemical
with NaOH to pH 11.0, with or without NaCl, under a N2 methods. The gas-phase geometries of reactant, products,
atmosphere. Titrating HCl (0.0667 M) was added to the solu- intermediate and transition states were fully optimised at the
tion with a Hamilton microsyringe. 1,8-Acid was titrated poten- Hartree–Fock (HF) level of theory using the standard 6-31G*
tiometrically with HCl (0.0667 M) under N2 using an aliquot of basis set. Electron correlation energy was included by single-
5.0 mL of a 0.01 M solution of 1,8-Acid (pH 12.0) diluted to point calculations at second-order Møller–Plesset perturbation
50 mL. The initial pH in this solution is near 11.0. Since the theory (MP2), using the same basis set. The characteristics of
pKa2 of 1,8-Acid is near 5.5, and at this pH the 1,8-Acid does all stationary points were evaluated by calculations of vibration
not cyclize (see below), it is not necessary to consider the harmonic frequencies at the HF/6-31G* level. Vibrational har-
anhydride–acid equilibrium. monic frequencies were obtained from second-order analytical
The potentiometric titration of 1,8-Acid at pH’s lower than energy derivatives. Assuming ideal behaviour, partition func-
4.0 were completed in ca. 15 min. On this time scale, the max- tions (translational, rotational, vibrational and electronic) and
imum calculated amount of 1,8-An formed upon cyclization of thermodynamic properties (referring to the standard state of

J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350 2343

Table 1 pKa’s of naphthalic acids and phthalic acids

Compound Method Salt or buffer pKa0 pKa1 pKa2

1,8-Acid Potentiometric 0 3.63 5.53

Potentiometric NaCl 0.02 M 3.52 5.38
Potentiometric NaCl 0.10 M 3.54 5.44
Spectrophotometric Buffer 0.1 M 3.50 5.40
Fluorescence Buffer 0.02 M 3.56
Fluorescence Buffer 0.05 M 3.49
2,3-Acid Potentiometric 0 3.27 5.28
Potentiometric NaCl 0.02 M 3.33 5.09
Potentiometric NaCl 0.10 M 3.05 5.00
Spectrophotometric Buffer 0.1 M 2.80 5.00
Fluorescence Buffer 0.1 M ⫺0.30 2.80 4.90
o-Phthalic acid a 2.89 5.51
m-Phthalic acid a 3.54 4.60
p-Phthalic acid a 3.51 4.82
a
Ref. 28.

1 mol L⫺1 and 298 K) were calculated by employing standard

statistical mechanical theory.16 Low frequency modes were (3)
subtracted from enthalpic and entropic contributions.
Reaction coordinate points were determined using the
intrinsic reaction coordinate (IRC) approach, also at the HF/6-
(4)
31G* level, using the method of Gonzales and Schlegel.17,18
Solvation free energies (related to the transfer of the solute
molecule from the 1 mol L⫺1 vapour to 1 mol L⫺1 aqueous The values of the molar extinction coefficients εAH , εAH and
2
solution at 298 K) were calculated at the MP2/6-31G* level by εA, corresponding to diprotonated, AH2, monoprotonated,
using the polarized-continuum model (PCM) and standard AH, and dianionic species, A, respectively, are presented in the
atomic radii, at the gas-phase determined geometries.19,20 legend of Fig. 1.
Addition of the energetic and entropic contributions to the The values of pKa1 and pKa2 of 1,8-Acid and 2,3-Acid,
MP2/6-31G*//HF/6-31G* electronic energies gave the gas- estimated by spectral measurements (Table 1) are, within the
phase free energies. Aqueous-phase free energies were obtained experimental error, similar to those determined by potentio-
from further addition of solvation (PCM) contributions. metric titration and in the same range of pKa’s of derivatives of
phthalic acid, Table 1.22
Results Photophysical properties of 1,8-Acid and 2,3-Acid have been
described.23 pKa values of both 1,8-Acid and 2,3-Acid were
The values of pKa1 and pKa2, of 1,8-Acid and 2,3-Acid (Scheme determined by changes in fluorescence emission yields with pH
2), calculated from potentiometric titration curves,21 with and are included in Table 1.
and without NaCl are presented in Table 1. NaCl produced a Since 2,3-Acid does not cyclize even at very low pH, the
moderate decrease in the pKa’s of both diacids (Table 1). fluorescence emission of 2,3-Acid was determined at high acid-
The values of pKa1 and pKa2 of 2,3-Acid and 1,8-Acid were ities and another pKa was apparent (Fig. 1C). The pKa in the H0
also determined by UV spectra (Fig. 1). The data in Figs. 1A region, pKa0, was attributed to the protonation of 2,3-AcidH2,
and 1B were fitted using eqn. (2). AbsT and CT are the absorb- AH2, leading to 2,3-AcidH3⫹, AH3⫹ (Scheme 3), as described in
eqn. (5).
(2)
(5)

ance and the total concentration of the diacid, and pKa1 and The fluorescence decrease at high HCl concentration was not
pKa2 refer to the dissociation constants of the diacid, Ka1 and due to quenching of the AH2 fluorescence by chloride ion, but
Ka2 (Scheme 3), as defined by eqns. (3) and (4). rather a significant decrease in quantum yield due to the form-
ation of AH3⫹ species.23 The fluorescence of the 2,3-Acid in 5.0
M HCl decreases 95% compared with that in pH 2.0, while the
addition of 5.0 M NaCl to 2,3-Acid at pH 2.0 leads to only a
30% decrease in the fluorescence intensity.
The values of the relative fluorescence intensity, F, versus pH
were adjusted using eqn. (6) where CT is the total 2,3-Acid
concentration, AH3, AH2, AH, and A are adjustable
parameters (presented in the legend of Fig. 1C) proportional to
the molar absorption coefficient at the excitation wavelength
and the emission yield for the 2,3-Acid species.
The calculated values of the three pKa’s of 2,3-Acid are
presented in Table 1.
The hydrolysis of 2,3-An was studied over a wide pH range.
Scheme 3 The observed rate constant, kψ, for 2,3-An hydrolysis does not

(6)

2344 J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350

Fig. 1 pH effect on the absorbance or fluorescence of: (A) 2,3-Acid: the concentration of 2,3-Acid was 1 × 10⫺4 M and the solid line was calculated
using εA = 1190 M⫺1 cm⫺1; εAH = 1950 M⫺1 cm⫺1; εAH = 1360 M⫺1 cm⫺1 in eqn. (2). (B) 1,8-Acid: the concentration of 1,8-Acid was 2 × 10⫺4 M and the
2

solid line was calculated using eqn. (2) and the following parameters: εA = 1000 M⫺1 cm⫺1; εAH = 1850 M⫺1 cm⫺1; εAH = 2500 M⫺1 cm⫺1. (C) 2,3-Acid:
2

the sample was excited at 300 nm and the fluorescence measured at 378 nm. The solid line was calculated using eqn. (6) and AH3 = 100; AH2 =
3400; AH = 2600 and A = 475. The concentration of 2,3-Acid was 1 × 10⫺5 M, buffers 0.1 M and the temperature 30 ⬚C.

Table 2 Rate and equilibrium constants for the reactions of hydrolysis

of 2,3-An and 1,8-An a

1,8-An 2,3-An

Parameters 30 ⬚C 50 ⬚C 30 ⬚C

kH/M⫺1 s⫺1 1.2 × 10⫺4 1.13 × 10⫺3

kOH/M⫺1 s⫺1 1.3 × 103 3.20 × 103 9.9 × 103
k⫺1/s⫺1; k⫺2/s⫺1 5.0 × 10⫺5 3.30 × 10⫺4 2.8 × 10⫺3
k2/s⫺1 8.0 × 10⫺4 3.90 × 10⫺3
k1/s⫺1 1.5 × 10⫺4 1.35 × 10⫺3
Ka0 (pK0) 2.45 (⫺0.39) 2.45 (⫺0.39)
Ka1 (pKa1) 2.51 × 10⫺4 (3.6) 3.55 × 10⫺4 (3.45)
Ka2 (pKa2) 3.16 × 10⫺6 (5.5)
KE0 115.5
KE1 3.0 4.25
a
Values obtained from adjusting experimental data to kinetic and
Fig. 2 pH effect on observed rate constant of hydrolysis of 2,3-An at equilibrium equations (see text).
30 ⬚C. The line was obtained using eqn. (7) and the parameters used are
in Table 2.

vary from pH 0.0 to 7.0 (Fig. 2), increasing sharply with pH in for OH⫺ attack. The values of the rate constants are presented
alkali. The slope of the linear log kψ–pH function was 1.0 above in Table 2.
pH 7.0 (not shown) and the kψ–pH pattern can be attributed to The spectra of 2,3-Acid, product of the hydrolysis of 2,3-An,
the water attack on 2,3-An up to pH 7.0 and OH⫺ catalysed remained unchanged, even for 5 days at 50 ⬚C in HCl 5.0 M
hydrolysis in alkali.24,25 The data of Fig. 2 were adjusted using (not shown), implying that if any 2,3-An is formed under these
eqn. (7), where k⫺1 is the first order rate constant for water conditions, the equilibrium strongly favours the 2,3-Acid.
attack on anhydride and kOH is the second order rate constant The kψ–pH dependence for 1,8-An hydrolysis is complex,
consisting of an acid catalysed region at pH’s lower than 1.0, a
kψ = k⫺1 ⫹ kOH[OH] (7) well defined plateau between pH 1.2 and 2.4, a pH dependent

J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350 2345

Fig. 3 Observed rate constant for the hydrolysis of 1,8-An and
attainment of equilibrium between 1,8-An and 1,8-Acid as a function
of pH–H0. At 30 ⬚C (0.02 M buffer) the rate constants were obtained
starting with: (a) 1,8-An for hydrolysis (above pH 6.0) and equilibrium
(below pH 6.0) (O) and (b) 1,8-Acid (䊉). At 50 ⬚C (0.1 M buffer) rate
constants for both hydrolysis (above pH 6.0) and equilibrium
attainment were obtained starting the reaction with 1,8-An (∆). The pH
scale at [H⫹] > 1.0 M is given in H0. The lines were obtained using eqns.
(18), (23) and (25) and the parameters used are in Table 2.

region between pH 2.5 and 4.5, a water catalysed region and a

OH⫺ catalysed reaction above pH 7.0 (Fig. 3).
Below pH 4.5 the values of kψ increase to a plateau with an
apparent pK, pKap, near 4.0 at 30 and 50 ⬚C. The values for kψ,
below pH 4.5, were the same whether the reaction was started
by addition of 1,8-Acid or 1,8-An (Fig. 3), demonstrating that
kψ, in this pH region, is related to the rate of diacid–anhydride
equilibrium attainment. Fig. 4 (A) Hydrolysis of 1,8-An followed by UV spectra at pH 7.5,
The quantitative analysis of the kinetic data presented in Fig. 50 ⬚C. Spectra were obtained after intervals of 45 seconds, (——)
3 required the analysis of the diacid–anhydride equilibrium. initial spectrum of 1,8-An and (—ⴢ—ⴢ—) final spectra of 1,8-Acid.
The marked spectral difference between 1,8-Acid and 1,8-An, (B) Absorbances at 340 nm, after attainment of equilibrium between
as observed upon 1,8-An hydrolysis at pH 7.5 (Fig. 4A), allowed 1,8-An and 1,8-Acid at 50 ⬚C, HCl pH 2.5. Scale of pH at values lower
the determination of the anhydride concentration at equi- than zero is in H0. The solid line was obtained using eqn. (16).
librium as a function of pH at 340 nm where only the anhydride
Here, as previously shown in related systems, the mono-
absorbs. The absorbance at equilibrium, AET, was determined at
protonated form of 1,8-Acid, AH, does not cyclize to the
50 ⬚C (Fig. 4B).
anhydride.9,12
The final absorbance obtained using 1,8-An as substrate at
Considering, as described repeatedly,2 that [H2O] is constant,
pH’s higher than 6.0 does not change with pH, consistent with
KE0 and KE1 can be defined as shown in eqns. (8) and (9).
near quantitative hydrolysis of 1,8-An to 1,8-Acid in this pH
range. Furthermore no absorbance change is detected at pH
higher than 7.0 when 1,8-Acid is used as substrate, strongly (8)
suggesting that the dissociated ionic forms of the diacid do not
cyclize.
At pH’s lower than 4.5 the equilibrium absorbance increases
(9)
becoming constant between pH 2.0 and 0. As the acid concen-
tration increases the absorbance increases again reaching a
value consistent with near quantitative anhydride formation The absorbance at the equilibrium at each pH, AET, (Fig. 4B)
(Fig. 4B). As expected, the values of AET at several pH’s were was related to eqns. (8) and (9) and Scheme 3 as follows. AT,
identical using either 1,8-Acid or 1,8-An to initiate the reaction. is taken as the absorbance corresponding to quantitative
From data of Fig. 4B it can be supposed that there are anhydride formation. 1,8-Acid does not absorb at 340 nm;
two equilibrium phenomena, involving differently protonated hence, at any pH absorbance is related only to anhydride con-
species, one between pH 4.5 and 0.0 and another below pH centration. The absorbance contribution from the anhydride
0.0. One of the simplest forms to rationalise these results produced from AH3⫹, in the absence of any other ionic species,
follows. Between pH 0 and 4.5 equilibrium is established, as is defined as AE1, and that from AH2 as AE2. Since the ratio of
described for several related systems,12 between the anhydride 1,8-AcidH2 and 1,8-AcidH3⫹ changes with acidity, the value
and AH2. At [H⫹] > 1.0 M equilibrium is established between of AET will be the sum of the molar fraction of 1,8-AcidH3⫹
the anhydride and the protonated form of 1,8-AcidH2, (α) and 1,8-AcidH2 (β) multiplied by AE1 and AE2, respectively
1,8-AcidH3⫹ (Scheme 3). [eqn. (10)].
In fact protonation of both cyclic anhydrides and dicarboxy-
lic acids were proposed decades ago.26,27 The variation of AET AET = αAE1 ⫹ βAE2 (10)
with pH was analysed by assuming that formation of 1,8-An
can occur from 1,8-AcidH3⫹ and from the neutral 1,8-AcidH2 Eqn. (13) can be derived, given the definitions in eqns.
and can be described by the equilibrium constants KE0 and KE1, (11) and (12), where the subscript E refers to equilibrium
respectively (Scheme 3). concentrations.

2346 J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350

Table 3 Thermodynamic parameters of the equilibrium between 1,8-Acid a and 1,8-An and comparison with the same reaction using DMPAcid
and DMPAn

Reagent Product ∆H/kJ mol⫺1 ∆S/J mol⫺1 K⫺1 ∆G b/kJ mol⫺1

1,8-Acid 1,8-An 10.1 ± 1.9 42.8 ± 5.8 ⫺3.72 × 0.02

DMPAcid DMPAn 16.51 c 37.8 c 4.3 b, c
a
Thermodynamic parameters at pH 2.5 for the reaction of cyclization of 1,8-Acid to 1,8-An were calculated from measured equilibrium constants at
several temperatures using: ∆G = ⫺RTln Keq; ln Keq = ∆S/R ⫺ (∆H/R) × 1/T ; R = 8.317 J K⫺1 mol⫺1. b ∆G was calculated for Keq at 50 ⬚C. c The ∆G
value for the cyclization reaction of DMPAcid to anhydride at 50 ⬚C was calculated from literature data11 for comparison.

shown in Table 2. The pKw values used for the calculation of

(11) [OH⫺] at 30 and 50 ⬚C were 13.833 and 13.2617 respectively.28
The value of k⫺1 was determined at the pH independent region,
between pH 5.5 and 6.0, Table 2.
Between pH 1.0 and 5.0, where the concentration of
anhydride at equilibrium is measurable, the observed rate con-
(12) stant reflects the kinetic process for attainment of equilibrium
between the diacid and the anhydride. Assuming that AH2 is the
kinetically active species in this range of pH the reaction rate
is given by eqn. (19) where k1 is the rate constant for AH2
[AH2]E = [AH3⫹]T ⫺ [AH3⫹]E (13) cyclization.
Using eqns. (8), (9) and (13), eqns. (14)–(16) can be derived.
(19)

(14)
In this pH range the total acid concentration, [1,8-Acid]T, is
given by eqn. (20).

(15) [Acid]T = [AH] ⫹ [AH2] ⫹ [AH3⫹] (20)

Using eqns. (3), (5) and (20), eqn. (21) can be derived.
The experimental AET–pH dependence (Fig. 4B) was ana-
lysed using eqn. (16) and the best fit values, at 50 ⬚C, were Ka0 = [Acid]T = [AH2](1 ⫹ Ka1/[H⫹] ⫹ [H⫹]/Ka0) (21)
2.45 (pKa0 = ⫺0.39), Ka1 = 3.55 × 10⫺4, (pKa1 = 3.45), KE0 = 115.5
and KE1 = 4.25 (Table 2). The value of pKa1 calculated from the Eqns. (19) and (21) give rise to eqn. (22) and kψ for equi-
fit of the equilibrium data with eqn. (16) is very similar to that librium cyclization from AH2, the predominant active species
determined by titration a 30 ⬚C (Table 1). The value of pKa0 between pH 5 and 2, is given by eqn. (23).9
(Table 2) obtained by fitting the equilibrium data between 1,8-
An and 1,8-Acid using eqn. (16) is identical to that determined
for 2,3-Acid by fluorescence titration (Table 1). (22)

(16)

(23)

At pH’s lower than 2.0 the ionic form AH3⫹ contributes

to the equilibrium anhydride formation. The reaction in this
The effect of temperature on KE1 was determined at pH 2.5,
acidity range can be characterised by the equilibrium between
where essentially all the 1,8-Acid is in the AH2 form. The values
AH3⫹ and 1,8-An and an acid catalysed reaction, kH, for the
of ∆S and ∆H, calculated from the linear plots of ln KE1 vs.
cyclization of AH3⫹ (see Fig. 3). The reaction rate is given by
1/T , are shown in Table 3.
eqn. (24), where k2 and k⫺2 are the rate constants for formation
After the determination of the equilibrium constants for
and decomposition of 1,8-An from AH3⫹, respectively.
anhydride formation and pKa’s of 1,8-Acid (Fig. 3) the pH
effect on the rate constants of 1,8-An decomposition can be v = kH [H⫹] [AH3⫹] ⫹ k2[AH3⫹] ⫺ k⫺2[An] (24)
analysed quantitatively.
At pH’s higher than 7, the predominant reaction is the alk- The plot of kψ vs. aH , at values of H0 lower than ⫺1, is linear
⫹

aline hydrolysis of 1,8-An. The reaction rate, v, in this region and from the slope of the plot the value of kH was obtained.
and the observed rate constant, kψ, are given by eqns. (17) and From the pertinent acid–base dissociation equilibria it can be
(18), where kOH is the second order rate constant for OH⫺ shown that eqn. (25) is true.
v = kψ[1,8-An] (17)
(25)
kψ = kOH[OH⫺] ⫹ k⫺1 (18)

attack on the anhydride and k⫺1 is the rate constant of water The kψ vs. pH profiles of Fig. 3, at 30 and 50 ⬚C, were
attack on the anhydride. adjusted using eqns. (18), (23) and (25). Between pH 1.0 and 5.0
The plot of log kψ vs. pH, from pH 7.0 to 9.0, was linear with the only adjustable parameter was k1. Below pH 1.0 both k2 and
a slope of 1.0 at 50 and 30 ⬚C and the calculated value of kOH is k⫺2 were varied to fit the experimental data to eqn. (25). As an

J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350 2347

 0.2 M HCl, additional acid catalysis was not observed.9 The kψ
vs. pH profile for the dimethylmaleic acid reaction is similar to
that obtained here with 1,8-An from pH 1.0 to 6.0. Maleic acid
derivatives cyclize only upon protonation of both carboxylic
groups.9
Spontaneous formation of 1,8-An from 1,8-Acid has been
suggested previously.29,30 We have shown that in the pH 2.0–
5.0 region the equilibrium formation of 1,8-An can be
adequately described by cyclization of AH2. Below pH 1.0 the
values of 1,8-Acid cyclization rate constants increased sharply
with aH⫹.
The pH dependence of kψ for the anhydride–acid equilibrium
was adequately fitted with the equilibria presented in Scheme 3.
It should be noted that the equilibrium constant for anhydride
formation determined experimentally is identical, within
experimental error, to that calculated from the rate constants
presented in Table 2 in the pH 1–5 range. This calculation was
made possible since the rate constant for water attack on
anhydride, k⫺1, was obtained experimentally (Table 2). Such a
comparison was not possible at the higher acidity range since
the rate constant for the water attack on the anhydride, k⫺2, may
contain a term in protonated water. Furthermore the error
Fig. 5 Structure of the possible reactant A (1,8-Acid), transition states in the estimation of the equilibrium constant at high acidity,
B and D, intermediate C and products E (1,8-An ⫹ H2O) in the KE0, is much larger, since there is essentially total conversion of
cyclization reaction. The naphthalic ring atoms are not shown for a
better structure visualisation.
1,8-Acid to 1,8-An.
There is extensive literature concerning hydrolysis of
anhydrides in aqueous acid.27,30–33 The effect of high acid con-
centration (0.5–4 M) on reaction rate is dependent on the
structure of the anhydride.31 Perchloric acid in dioxan–water
(60 : 40) increases the hydrolysis rate of glutaric, homophthalic
and camphoric anhydrides but inhibits the hydrolysis of maleic
and phthalic anhydrides. Univalent salts retard the acid
hydrolysis of these cyclic anhydrides. Therefore, for both mono-
and bi-cyclic anhydrides there is a relationship between sensi-
tivity to acid, size and configuration of the anhydride ring.
However, flexibility of the anhydride ring systems does not
affect the sensitivity to acid, e.g., camphoric anhydride has a
rigid structure, but nonetheless, its hydrolysis is acid-catalysed.
Bunton et al. suggested that the acid hydrolysis occurs by an
A-2 mechanism involving protonation of an oxygen atom and
nucleophilic attack upon the carbonyl carbon atom.31
Fig. 6 Relative MP2/6-31G*//HF/6-31G* activation energies for the
gas phase (䊏) and solution (䊉) for the cyclization reaction of 1,8-Acid.
Leisten, in a cryoscopic study of the equilibrium between
cyclic anhydrides and diacid in 100% sulfuric acid, suggests that
initial estimate we assumed that k⫺1 = k⫺2. The best fit values for there is an equilibrium between the anhydride and the pro-
the rate and equilibrium constants are presented in Table 2. tonated form of the diacid.26 This suggestion is in agreement
The study of the potential energy surface for the formation with the protonation proposed in this work for 1,8-Acid below
of 1,8-An from 1,8-Acid, AH2 form, led us to the following pH 1.0
reaction scheme: 1,8-Acid TS1 Intermediate TS2 Anhydride formation from carboxylic acids is a general
1,8-An ⫹ H2O. The fully optimised structures are shown in Fig. phenomenon. The equilibrium constant between mono-
5 and the free energy profiles for the reaction in the gas phase carboxylic acids and its anhydrides lies strongly toward the
and aqueous solution are presented in Fig. 6. The reaction has acid. Acetic acid, for example, at the boiling point, contains
two distinguishable steps. only ca. 0.03% of the anhydride.7,8,12 For dicarboxylic acids the
(1) An intramolecular proton transfer concerted with oxygen equilibrium constant for anhydride formation, KE, varies from
alignment (O2, Fig. 5) towards the carbonyl centre form- 25 to 1 × 10⫺5 depending on structure.12 Apparently, high KE
ing transition state 1 (TS1) and then a neutral and planar values for anhydride formation are related mainly to strong
intermediate (minima in the potential energy surface), with a non-bonded interactions in the diacid that are relieved in the
tetragonal sp3 carbon (C1). cyclic anhydride.12
(2) The second step is dehydration, through transition state Dicarboxylic anhydrides also form in solution upon mono-
2 (TS2), yielding a complex between water and 1,8-An. Both ester hydrolysis. Monoesters of dicarboxylic acids hydrolyse by
transition states were characterised by a single imaginary fre- intramolecular nucleophilic attack by the carboxylate ion upon
quency and IRC calculations pointed to the expected previous the ester carbonyl yielding anhydrides.34–37 Linear correlations
and forward minima structures in the reaction pathway. between the rate constant for diacid cyclization and ∆pK
(pKa1 ⫺ pKa2), have been described. The ∆pK of a diacid is
largely determined by the existence of an intramolecular hydro-
Discussion gen bonded species in the monoanion. In a study of pKa of 1,2
The pH effect on the observed rate constants for anhydride dicarboxylic acids attached to rigid skeletons intramolecular
solvolysis, kψ, has been studied in few cases. hydrogen bonding increases with the distance between the car-
The value of kψ for decomposition of maleic anhydride boxy groups. The stronger intramolecular hydrogen bonding
derivatives in aqueous solution is pH dependent between occurs in the monoanion in solution when the carboxy group is
pH 3.0 and 5.0 and, since the reactions were followed only up to spaced so that an essentially linear bond of about 2.45 Å lying

2348 J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350

in the intersection of the two planes of the carboxy groups is reactants in the gas phase and obtained good agreement with
formed. The carboxy groups should be attached to a carbon relative rate constants measured in solution. No such series of
skeleton such that no change in conformation of the carbon rate constants is available for the present system. Nonetheless,
skeleton is allowed, but the carboxy group may rotate.22 the calculated mechanism (Figs. 5 and 6) is a possible pathway
The pKa1/pKa2 ratios of 1,8- and 2,3-Acid were ca. 1.6, imply- for the 1,8-Acid cyclization reaction and shows the same
ing an intramolecular bonding of the same effectiveness. How- proximity and strain features present in the reaction studies of
ever, 1,8-Acid cyclizes at pH’s below 4.0 while the 2,3-Acid does Bruice and Lightstone 40 which account for reaction rate
not cyclize, to a measurable extent, even at 50 ⬚C and 6.0 M enhancements.
HCl. The cyclization of 2,3-Acid cannot be observed using our Hawkins’ 11 studies of the hydrolysis of phthalic, PTAn,
methods, probably because the equilibrium strongly favours the and 3,6-dimethylphthalic, DMPAn, anhydrides show that in
acid. Hence, the correlation between ∆pK and anhydride for- aqueous acid DMPAn is in equilibrium with 3,6-
mation does not hold for the comparison between 2,3-Acid and dimethylphthalic acid, DMPAcid, while PTAn is quantitatively
1,8-Acid. In fact, although correlations between ∆pK and rates hydrolysed to phthalic acid, PTAcid, and the equilibrium con-
of anhydride formation do exist in a limited series of diacids, stant cannot be measured. In water PTAn hydrolyses ca. 8 times
they cannot hold for all diacids.38 In order for a diacid to exhibit faster than DMPAn and the cyclization step is ca. 6 times faster
a high ∆pK its geometry should allow the formation of an for DMPAcid. The slower hydrolysis rate of DMPAn is prob-
internal hydrogen bond of about 2.40 Å with little change in ably not due to steric hindrance towards water attack, but to the
molecular strain on going from the open chain to the cyclic relief of steric strain by the methyl groups of DMPAcid, when
form. If, on the other hand, the hydrogen bond in the mono- compared with the corresponding anhydride. Comparison of
anion is formed at the expense of a large increase in strain, the the thermodynamic parameters for the cyclization reaction for
∆pK of the diacid may be small. Thus, an important require- 1,8-Acid and DMPAcid (Table 3) shows that the difference in
ment for fast cyclization, apart from the existence of steric equilibrium constants reflects differences in the enthalpy term
strain in the diacid, appears to be that the carboxy groups are as for the equilibrium and not an entropic difference as has been
closely situated as possible.12 suggested in related systems.40
It has been suggested that strain is released in the intra-
molecular diacid to anhydride conversion.34 As seen from the
calculated structures (see Fig. 5), 1,8-Acid is a rigid and highly Conclusions
strained system, where each carboxy group lies in an opposite Reversible 1,8-An formation from 1,8-Acid was observed in
side of the aromatic ring plane. On the other hand, the 1,8-An is acidified aqueous solution below pH 5.0 while the decom-
a rigid but planar structure, without any carboxy interaction position of 2,3-An leads to quantitative formation of 2,3-Acid.
resulting in strain.10,12,13 It was suggested previously that the The undissociated 1,8-Acid, AH2, as well as the protonated
carboxy group proximity is also determinant in anhydride form, AH3⫹, were the kinetically reactive species. The values of
formation.12,35–38 Here, and in other bridged cyclic anhydrides the equilibrium constant for 1,8-An formation of ca. 4 and 100,
where the equilibrium constant for anhydride formation is rela- respectively, for the AH2 and AH3⫹ forms, indicated that
tively high, 3 carbon atoms and the anhydride bonds, forming a protonation stabilizes the anhydride relative to the dicarboxylic
6-membered ring, separate both carboxy groups.13,39 The dis- acid. Ab initio calculations demonstrated that the pathway
tance between the attacking oxygen (O2) and electrophilic car- for anhydride formation includes a rate determining intra-
bon (C1) is just 2.76 Å in the 1,8-Acid, reaching 2.27 Å in TS1 molecular proton transfer concerted with oxygen alignment
(Fig. 5). As in similar dicarboxylic acids, strain release and car- towards the carbonyl centre.
boxy group proximity are the major factors governing the rate
of anhydride formation from 1,8-Acid. The structures 1,8-Acid
(see Fig. 5, A) and TS1 (B) are quite different and the pathway Acknowledgements
could include a previous proton transfer from one oxygen to the The authors are grateful to the Brazilian Agencies FAPESP,
other in the same carboxy group. Additional calculations would CNPq, and CAPES for financial support. This work is part of
be necessary to fully investigate this putative step. We can the MSc dissertation of T.C.B, IQUSP (1991) and of the PhD
assume, however, that this step may be unimportant kinetically. thesis of S.Y., UFSC (1996). T.C.B. was a CAPES graduate
If such a transfer is not a barrierless process, it would have a fellow and S.Y. was a CNPq and CAPES fellow. G.M. and
smaller energy barrier than a hydrogen transfer between differ- L.G.D. are graduate and post-doctoral fellows from FAPESP.
ent carboxy groups and, because of short transfer distance, The authors thank Dr Jorge Masini, IQUSP, for his helpful
hydrogen tunnelling effects would be present. suggestions and for his computational advice in the calculation
The mechanism here proposed for the cyclization of the of potentiometric pKa’s of the diacids and Professor C. A.
diprotonated 1,8-Acid is similar to those described where the Bunton, UCSB, USA, for advice and helpful suggestions with
reaction of a neutral diacid proceeds towards an intramolecular the manuscript.
carboxylate attack upon the other carboxy function with final
anhydride formation.35–38
The KE1 value obtained with 1,8-An at 50 ⬚C (see results) References
agrees reasonably well with the calculated free energy difference 1 A. J. Kirby, Adv. Phys. Org. Chem., 1980, 17, 183.
between reactant and products (see Fig. 6). Although both the 2 W. P. Jencks, Catalysis in Chemistry and Enzymology, Dover,
calculated and experimental results are in the range obtained New York, 1987.
for other similar systems, our purpose with the calculations 3 S. Y. Yunes, J. C. Gesser, H. Chaimovich and F. Nome, J. Phys. Org.
was not to theoretically obtain the value of the experimental Chem., 1997, 10, 461.
4 T. C. Barros, Formação e Decomposição de Naftalimidas em Solução
rate constants. Using a simplified model (PCM) to describe Aquosa: Dependência Estrutural e Efeito de Micelas, Masters Thesis,
solvation effects does not account for specific solvent–solute Chemistry Institute, Biochemical Department, University of São
interactions (e.g. hydrogen bonds) that are certainly important Paulo, USP, São Paulo, Brazil, 1991.
in this particular reaction in aqueous solution. Theoretical 5 S. F. Yunes, Hidrólise de Monoalquil Ésteres do Ácido 1,8-Naftálico,
calculations for cyclization involving a number of explicit water PhD Thesis, Department of Chemistry, Centro de Ciências F´sicase
molecules, although possible, was not the subject of our present Matemáticas, Universidade Federal de Santa Catarina, UFSC,
Florianópolis, Santa Catarina, Brazil, 1996.
interest. 6 M. Bender, Chem. Rev., 1960, 60, 53.
Bruice and Lightstone 40 have recently described theoretically 7 J. A. Knopp, W. S. Linnell and W. C. Child, Jr., J. Phys. Chem., 1966, 66,
a one step intramolecular cyclization reaction with anionic 1513.

J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350 2349

 8 W. P. Jencks, F. Barley, R. Barnett and M. Gilchrist, J. Am. Chem. 22 L. L. McCoy, J. Am. Chem. Soc., 1967, 89, 1673.
Soc., 1966, 88, 4464. 23 T. C. Barros, G. R. Molinari, P. Berci-Filho, V. G. Toscano and
9 L. Eberson, Acta Chem. Scand., 1964, 18, 1276. M. J. Politi, J. Photochem. Photobiol., 1993, 76, 55.
10 T. Higuchi, L. Eberson and J. D. McRae, J. Am. Chem. Soc., 1967, 24 R. P. Bell, The Proton Chemistry, 2nd edn., 1973, Chapman and
89, 3001. Hall, London, UK.
11 M. D. Hawkins, J. Chem. Soc., Perkin Trans. 2, 1975, 282. 25 L. P. Hammett, Physical Organic Chemistry, 1970, 2nd edn.,
12 L. Eberson and H. Welinder, J. Am. Chem. Soc., 1971, 93, 5821. McGraw-Hill Kogakusha, Japan.
13 L. Eberson and L. Landstrom, Acta Chem. Scand., 1972, 239. 26 J. A. Leisten, J. Chem Soc., 1961, 2191.
14 A. D. Campbell and M. R. Grimett, Aust. J. Chem., 1963, 16, 854. 27 C. A. Bunton, J. H. Fendler, N. A. Fuller, S. Perry and J. Rocek,
15 R. H. Boyd in Solute Solvent Interactions, Ed. J. F. Coetzee and J. Chem. Soc., 1965, 6174.
C. D. Ritchie, 1969, Marcel Dekker, New York. 28 CRC Handbook of Chemistry and Physics, ed. R. C. Weast, 52nd
16 D. A. MacQuarrie, Statistical Mechanics, Harper & Row, New York, edn., 1971, The Chemical Rubber Co., USA.
1976. 29 L. J. Fitzgerald, R. E. Gerkin and G. D. Renkes, Vib. Spectrosc.,
17 K. Fukui, Acc. Chem. Res., 1981, 14, 363. 1991, 2, 269.
18 C. Gonzales and B. H. Schlegel, J. Phys. Chem., 1990, 94, 5523. 30 R. Aroca, S. Corrent and J. R. Menendez, Vib. Spectrosc., 1997, 13,
19 J. Tomasi and J. Pérsico, Chem. Rev., 1994, 94, 2027. 221.
20 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. 31 C. A. Bunton, J. H. Fendler, N. A. Fuller, S. Perry and J. Rocek,
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., J. Chem. Soc., 1963, 5361.
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. 32 C. A. Bunton, N. A. Fuller, S. Perry and V. J. Shiner, J. Chem. Soc.,
Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, 1963, 2918.
M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, 33 C. A. Bunton, N. A. Fuller, S. Perry and I. H. Pitman, J. Chem. Soc.,
S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, 1962, 4478.
K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, 34 E. Gaetjens and H. Morawetz, J. Am. Chem. Soc., 1960, 82, 5328.
J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, 35 T. C. Bruice and U. K. Pandit, J. Am. Chem. Soc., 1960, 82, 5858.
A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, 36 T. C. Bruice and U. K. Pandit, Proc. Natl. Acad. Sci. USA, 1960, 46,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, 402.
C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, 37 T. C. Bruice and W. C. Bradbury, J. Am. Chem. Soc., 1965, 87,
M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. 4846.
Replogl and J. A. Pople, Gaussian 98, Revision A.6, Gaussian, Inc., 38 T. C. Bruice and W. C. Bradbury, J. Am. Chem. Soc., 1965, 87, 4851.
Pittsburgh PA, 1998. 39 H. K. Hall, Jr., J. Org. Chem., 1963, 28, 2027.
21 J. C. Masini, Talanta, 1994, 41, 1383. 40 T. C. Bruice and F. C. Lightstone, Acc. Chem. Res., 1999, 32, 127.

2350 J. Chem. Soc., Perkin Trans. 2, 2001, 2342–2350