Ketonization of the unusually acidic elongated enol generated by flash photolytic decarboxylation of p-formylphenylacetic acid in aqueous solution

Article abstract of DOI:10.1139/V07-095

Rates of photolysis of p-formylphenylacetic acid were measured flash photoytically in perchloric acid and sodium, hydroxide solutions, and also in acetic acid, biphosphate ion, and tris-(hydroxym.ethyl)methaneam.monium ion buffers, using H₂O and D₂O as solvents. The results provide rate profiles and solvent isotope effects, which indicate that photolysis occurs through an elongated enol intermediate. This enol is unusually strongly acidic, by some two to three pQ_a units, when compared with simple non-elongated enols.

Full text of DOI:10.1139/V07-095

Pagination not final/Pagination non finale
1
Ketonization of the unusually acidic elongated
enol generated by flash photolytic decarboxylation
of p-formylphenylacetic acid in aqueous solution
Yvonne Chiang, Kirill Kolmakov, and A. Jerry Kresge
Abstract: Rates of photolysis of p-formylphenylacetic acid were measured flash photoytically in perchloric acid and
sodium hydroxide solutions, and also in acetic acid, biphosphate ion, and tris-(hydroxymethyl)methaneammonium ion
buffers, using H O and D O as solvents. The results provide rate profiles and solvent isotope effects, which indicate
2
2
that photolysis occurs through an elongated enol intermediate. This enol is unusually strongly acidic, by some two to
three pQ units, when compared with simple non-elongated enols.
a
Key words: flash photolysis, elongated enols, rate profiles, solvent isotope effects.
Résumé : Faisant appel à la photolyse éclair, on a mesuré les vitesses de photolyse de l’acide p-formylphénylacétique
dans des solutions aqueuses ou dans le l’eau lourde d’acide perchlorique et d’hydroxyde de sodium ainsi que dans
l’acide acétique en présence d’ion biphosphate et de tris(hydroxyméthyl)méthaneammonium comme tampons. Les résul-
tats permettent d’obtenir des profils de vitesse et des effets isotopiques de solvant qui indiquent que la photolyse se
produit par le biais d’un intermédiaire énolique allongé. Cet énol est particulièrement acide, par des valeurs de pQ de
a
deux à trois unités supérieures à celles d’énols simples non allongés.
Mots-clés : photolyse éclair, énols allongés, profils de vitesse, effets isotopiques de solvant.
[
Traduit par la Rédaction] Chiang et al.
4
Introduction
Product formation could therefore occur either by direct
protonation of methylene carbon (eq. [3]) or by protonation
of carbonyl oxygen to give an elongated enol intermediate,
Photodecarboxylation of p-benzoylphenylacetic acid (1)
when carried out in aqueous solution, gives p-methylbenzo-
phenone (2) as its reaction product (eq. [1]) (1).
5
, whose ketonization would afford the observed product
eq. [4]). In previously reported work (2), we carried out this
(
decarboxylation flash photolytically and detected a short-
lived transient species that we identified as elongated enol 5.
This enol has proved to be an unusually strong acid, several
orders of magnitude more acidic than other simple enol
analogs without the elongating cyclohexadienyl spacer.
O
O
hv
CO2
Ph
Ph
-
[
1]
2
CO H
CH
3
1
2
O
O
This substance is formed through a carbanion intermediate,
whose negative charge is delocalized from methylene carbon
H+
[
3]
Ph
Ph
-
CH3
(
3) onto carbonyl oxygen (4) (eq. [2]).
CH2
2
3
O
O
-
O
OH
O
[
2]
Ph
Ph
H+
Ph
Ph
Ph
-
2
CH
CH
2
CH
2
CH2
CH
3
3
4
[
4]
4
5
2
Received 16 May 2007. Accepted 18 July 2007. Published on
the NRC Research Press Web site at canjchem.nrc.ca on
We have now added to that previous study a companion
investigation of the photodecarboxylation of p-formyl-
phenylacetic acid (6) (eq. [5]). We found that this reaction
1
7 October 2007.
This article is dedicated to the memory of Keith Yates.
O
O
1
Y. Chiang, K. Kolmakov and A.J. Kresge. Department of
[
5]
hv
CO2
H
H
Chemistry, University of Toronto, Toronto, ON M5S 3H6,
Canada.
-
2
CO H
CH
3
1Corresponding author (e-mail: akresge@chem.utoronto.ca).
7
6
Can. J. Chem. 86: 1–4 (2008)
doi:10.1139/V07-095
© 2007 NRC Canada

Pagination not final/Pagination non finale
2
Can. J. Chem. Vol. 86, 2008
also occurs through an elongated enol intermediate, which is
also unusually strongly acidic.
sample (80 mg) with no detectable impurities (HPLC, TLC)
and mp 128 to 129 °C was obtained, starting from 150 mg
of the material, by recrystallization (2 times) from CH Cl –
2
2
hexane (1:3 v/v) under an argon atmosphere. (Baker et al.
5), who originally obtained 4-formylphenylacetic acid with
Experimental Section
(
a very poor yield, reported mp 131 °C). The spectroscopic
data (NMR, MS) listed here confirm the structure.
Materials
p-Formylphenylacetic acid was prepared by a Sommelet
reaction (3) on 4-chloromethylphenylacetic acid, itself syn-
thesized by chloromethylation of phenylacetic acid (4).
1
H NMR (CDCl , 300 MHz) δ: 3.75 (s, 2H, CH ), 7.46
3
2
(
1
(
(
d, 2H, J = 8 Hz, p-C H ), 7.86 (d, 2H, J = 8 Hz, p-C H ),
6 4 6 4
13
0.01 (s, 1H, CHO). C NMR (CDCl , 75 MHz) δ: 41.05
3
CH ), 130.09, 130.19, 135.56 and 140.30 (C ), 176.51
2
A
r
4
-Chloromethylphenylacetic acid
A 10 mL capacity vial, equipped with septa and stirring
+
CO H), 191.90 (C=O). MS (EI) m/z: 164 M (85), 119
2
+
+
M
– CO -H (55), 91 M –CO –OH (100). HRMS (EI)
2
2
bar, was purged with argon, and 0.8 mL (1.78 g,
.0068 mol) of SnCl was introduced through a syringe. Dry
+
calcd. for (M ) C H O : 164.0473; found 164.0476.
9
8
3
0
4
All other materials were best available commercial grades.
phenylacetic acid (1.36 g, 0.010 mol) was dissolved in 2 mL
(
(
2.12 g, 0.026 mol) of monochloromethyl ether
ClCH OCH ) at room temperature and the solution was
Kinetics
2
3
Rate measurements were made using either a conventional
flash lamp (microsecond) flash photolysis system (6), or an
slowly added into the vial within 20 min upon vigorous stir-
ring at 0 °C. The reaction mixture was left at room tempera-
ture for 1 h under argon and then quenched with 30 mL of
ice-cold water upon vigorous stirring. An amorphous white
precipitate was separated. This was washed with water and
dissolved in 20 mL of CHCl , and the solution was dried
over MgSO . The solvent was evaporated to dryness, and the
eximer laser (nanosecond) system (7) operating at λ
=
exc
2
48 nm, which have already been described (6, 7). This re-
vealed a short-lived transient species, subsequently identified
as an elongated enol intermediate (vide infra) with broad UV
absorbance centered at λ = 330 nm. Decay of this
absorbance, for the most part, followed the first-order rate
law well, and rate constants were obtained by least-squares
fitting of a single exponential function. In some cases, how-
ever, the data fit a double exponential function better than a
single exponential function, and a double exponential was
therefore used. In this situation, one of the two rate con-
stants obtained changed its value when the reacting solution
was degassed, and the other unchanging rate constant was
then taken to be that for ketonization of the elongated enol
intermediate.
3
4
residue was recrystallized from a mixture of CH Cl
2
2
(
10 mL) and hexane (20 mL) to afford 0.82 g (44%) of a
product whose melting point (149–151 °C) was consistent
1
with the literature (4). Its H NMR (CDCl , 300 MHz) had
the following signals: δ: 3.65 (s, 2H, CH ), 4.57 (s, 2H,
CH Cl), 7.26 (d, 2H, J = 9 Hz, p-C H ), 7.35 (d, 2H, J =
3
2
2
6
4
9
Hz, p-C H ).
6
4
4
-Formylphenylacetic acid
Hexamethylenetetramine (840 mg, 0.0060 mol) and 4-
chloromethylphenylacetic acid (740 mg, 0.0044 mol) were
heated under reflux for 1 h in 20 mL of chloroform. The sol-
vent was then evaporated to one half of its initial volume,
and 990 mg (75%) of a colorless quaternary ammonium salt
separated. This was dried and then heated under reflux in
Product identification
The reaction investigated was shown to be the decar-
boxylation of eq. 5 by HPLC analysis using an authentic
sample of the expected product p-methylbenzaldehyde (7).
1
0 mL of 50 vol% aqueous acetic acid for 1.5 h under an ar-
gon atmosphere. Upon cooling to ambient temperature, the
Results
mixture was quenched with a solution of 10 g of NH Cl in
4
4
1
0 mL of H O and was then extracted with CH Cl (3 ×
Rates of decay of the broad UV absorption band centered
2
2
2
5 mL). The extract was washed with a solution of NH Cl
at λ
= 330 nm, generated by flashing 4-formylphenyl-
4
max
(
1.0 g) and 0.5 mL of concd. HCl in 10 mL of H O and was
acetic acid, were determined in aqueous perchloric acid and
sodium hydroxide solutions, and also in acetic acid, biphos-
phate ion, and tris-(hydroxymethyl)methaneammonium ion
2
dried over MgSO . The solvent was removed under vacuum
4
to afford 460 mg (70%) of crude product, which was
chromatographed on a column containing 8 g of silica gel
with mobile phase MeCN–CH Cl (1:1 v/v). The main frac-
buffer solutions, using H O as solvent; some rate measure-
2
ments were made as well in perchloric acid and sodium hy-
2
2
tion furnished 330 mg (44%) of a relatively pure (>90%,
NMR) compound. The course of the separation was moni-
tored by TLC on SIL-G/UV 254 plates with acetone–
acetonitrile (3:1) as the mobile phase. Further, an analytical
droxide solutions and biphosphate ion buffers, using D O as
2
solvent. The ionic strength of all reacting solutions was kept
constant at 0.10 mol/L with sodium perchlorate as the inert
electrolyte. The results are summarized in Tables S1–S3.2
,3
2
Supplementary data for this article are available on the journal Web site (canjchem.nrc.ca) or may be purchased from the Depository of Un-
published Data, Document Delivery, CISTI, National Research Council Canada, Ottawa, ON K1A 0R6, Canada. DUD 5207. For more in-
formation on obtaining material refer to cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml.
In perchloric acid solutions the absorbance from the transient species of interest here diminished as the acid concentration increased, consis-
tent with decarboxylation occurring through the carboxylate form of the substrate, and HPLC analysis showed the formation of other
3
(
uninvestigated) products. These interferences could be diminished somewhat by using deoxygenated solutions and accumulating data from
several replicate flash photolysis shots.
©
2007 NRC Canada

Pagination not final/Pagination non finale
Chiang et al.
3
The measurements in buffers were made in a series of so-
lutions of constant buffer ratio, and therefore constant
hydronium ion concentration, but varying total buffer con-
centration. Buffer catalysis was strong, and observed rate
constants within a given buffer solution series were accu-
rately proportional to buffer concentration. The data were
therefore analyzed by least-squares fitting of the linear
buffer dilution expression shown in eq. [6]. The zero buffer
Fig. 1. Rate profiles for the ketonization of elongated enol 8 in
H2O (᭺) and D (᭝) solutions.
concentration intercepts, k , obtained in this way, plus the
int
rate constants determined in perchloric acid and sodium hy-
droxide solution, are displayed as the rate profiles shown in
Fig. 1.
[
6]
kobs = kint + kbuff[buffer]
The shape of these rate profiles is characteristic of enol
ketonization reactions, which are known to occur by rate-
determining β-carbon hydronation of either the enol or its
enolate ion by any available acid (8). Because these profiles
refer to solvent-related species, the available acids will be ei-
+
ther the hydronium ion, written here as L , or a solvent wa-
zontal sections of the rate profiles, and the other rate and
equilibrium constants are defined by eq. [7].
ter molecule, L O, as shown in the reaction scheme of
eq. [7]. The acid-catalyzed descending diagonal portions at
2
4
+
E
E
+
[
8]
kobs = k [L ] + k + k ′Q /(Q + [L ])
L
uc
o
a
a
the high acidity ends of these profiles then represent β-car-
bon hydronation of the enol by hydronium ion. These sec-
tions are followed by short “uncatalyzed” horizontal
portions, which could be due either to β-carbon hydronation
of enol by solvent water or to the ionization of enol to
enolate and hydronium ions followed by β-carbon
hydronation of the enolate ion by hydronium ion. This latter
process generates hydronium ions in its rapid pre-
equilibrium ionization step and then uses them up in its rate-
determining step, to give an overall reaction independent of
hydronium ion concentration.
Least-squares fitting using this expression gave for H O
2
3
–1 –1
solutions: kH = (8.11 ± 0.48) × 10 (mol/L) s , k
=
uc
1
–1
E
–9
(
2.62 ± 0.28) × 10 s , Q = (9.20 ± 0.58) × 10 mol/L,
pQa = 8.04, and k ′ = (2.82 ± 0.05) × 10 s , and results for
D O solutions that, when combined with their H O counter-
a
E
5 –1
o
2
2
parts, gave the isotope effects: k /k = 4.49 ± 0.09,
H
D
E
E
5
(
k ′) /(k ′) = 7.74 ± 0.36, and (Q ) /(Q ) = 3.41 ± 0.14.
o H o D a H a D
The lines through the data points shown in Fig. 1 were
drawn using these results. It may be seen that the data fit the
reaction scheme of eq. [7] well and thus provide strong evi-
dence that the process under observation is an enol
ketonization reaction.
OL
O
E
a
Q
H
H
L+
Further support for the assignment of the rate profiles of
+
Fig. 1 to enol ketonization comes from the isotope effects
[
7]
CH
2
CH
2
L+
k
L+
k
L
'
obtained by comparing results determined in H O solution
L
2
with those obtained in D O. The rate constants k and k ′
2
L
o
L
2
O
k
o
2 o
L O k
'
refer to rate-determining substrate hydronation, and isotope
effects on these quantities can therefore be expected to con-
tain primary isotope effect components in the normal direc-
tion (k /k > 1) of appreciable magnitude. These isotope
These short horizontal sections are followed by base-
H
D
catalyzed ascending diagonal portions that can be assigned
to the ionization of enol to enolate ion followed by
hydronation of enolate by solvent water. This process pro-
duces hydronium ions in its rapid pre-equilibrium step but
does not use then up in its rate-determining step, to give an
overall reaction whose rate is inversely proportional to
hydronium ion concentration and has the appearance of hy-
droxide ion catalysis. Finally at sufficiently low acidity, the
position of the enol–enolate equilibrium shifts over to the
side of the enolate ion, and the process reduces to simple
hydronation of enolate β-carbon by solvent water. This gives
the final horizontal rate profile plateau.
effects, however, will also contain secondary components,
with that on k in the inverse direction (k /k < 1) and that
L
H
D
on k ′ in the normal direction (k /k > 1) (9). This will make
o
H
D
the isotope effect on k ′ considerably stronger than that on
o
k ′, which is nicely consistent with our observation, i.e.,
H
(k ′) /(k ′) = 7.7 whereas k /k = 4.5. The isotope effect on
o
H
o
D
H
D
E
Q
is also consistent with this quantity being the acidity
a
constant of an enol ionizing as an oxygen acid. Such equilib-
rium isotope effects lie in the normal direction and generally
fall in the range 3–5 (8); that of course agrees well with the
E
E
observed value (Q ) /(Q ) = 3.4.
a
H
a
D
Thus, both solvent isotope effects and rate profiles agree
that decarboxylation of p-formylphenylacetic acid occurs
through an elongated enol intermediate, 8, as shown in
eq. [9].
The rate law that applies to this reaction scheme is shown
in eq. [8], where kuc is used to designate the process of an
uncertain mechanism represented by the central short hori-
4
The symbol “L” is used to denote either protium or deuterium, “H” is used specifically for protium, and “D” specifically for deuterium.
This is a concentration equilibrium constant applicable at the ionic strength (0.10 mol/L) of the solutions in which it was determined.
5
©
2007 NRC Canada

Pagination not final/Pagination non finale
4
Can. J. Chem. Vol. 86, 2008
[
9]
[10]
OH
O
O
O
OH
O
H
H+
+
H
H
hv
CO2
H
H
H
-
-
CH2
CH
2
CH2
2
CO H
CH
2
CH
3
8
10
11
8
7
6
Discussion
The present results show the elongated enol 8 with pQaE
Acknowledgement
=
8
.04, to have an unusually strongly acidic hydroxyl group;
We are grateful to the Natural Sciences and Engineering
Research Council of Canada for its generous financial sup-
port of our work.
simple enols ionizing as oxygen acids ordinarily have acidity
constants in the range pQaE = 10–11 (8). An especially apt
comparison of the present system can be made with the enol
of acetaldehyde 9, which lacks the elongating cyclo-
OH
References
OH
H
H
CH
2
1. M. Xu and P. Wan. Chem. Commun. (Cambridge), 2147
(2000); L.A. Huck, M. Xu, K. Forest, and P. Wan. Can. J.
Chem. 82, 1760 (2004).
CH
2
9
8
2
. Y. Chiang, A.J. Kresge, I. Onyido, J.P. Richard, P. Wan, and
M. Xu. Chem. Commun. (Cambridge, U.K.), 4231 (2005).
3. S.J. Angyal. In Organic reactions. Vol. 8. Wiley, NY. 1954.
hexadienyl spacer
of enol 8; acetaldehyde enol has pQaE = 10.50 (6), some two
and a half pQ units weaker than the acidity constant of elon-
gated enol 8. This striking difference is a result of the fact
that ionization of enol 8 (eq. [10]) creates an anion whose
negative charge can be delocalized from enolate oxygen 10
onto methylene carbon 11 (eq. [10]); this converts a cyclo-
hexadienyl moiety into a benzene ring and thus benefits
from a gain in benzene resonance energy. A similar shift of
negative charge density is of course possible in the union-
ized enol, but that involves generation of charge and is con-
sequently not as energetically advantageous. The anion is
consequently stabilized more than the unionized enol, and
that reduces the energy gap between the enol and enolate
ion, making the enol a stronger acid than it would otherwise
be.
Chap. 4.
4
5
6
7
8
. A.R. Mitchell, B.W. Erickson, M.N. Ryabtsev, R.S. Hodges
and R.B. Merrifield. J. Am. Chem. Soc. 98, 7357 (1976).
. J.W. Baker, J.A.L. Brieux, and D.G. Saunders. J. Chem. Soc.
4
04 (1956).
. Y. Chiang, M. Hojatti, J.R. Keeffe, A.J. Kresge, N.P. Schepp,
and J. Wirz. J. Am. Chem. Soc. 109, 4000 (1987).
. J. Andraos, Y. Chiang, G.C. Huang, A.J. Kresge, and J.C.
Scaiano. J. Am. Chem. Soc. 115, 10605 (1993).
. J.R. Keeffe and A.J. Kresge. In The chemistry of enols. Edited
by Z. Rappoport. Wiley, NY. 1990. Chap. 7.
9. A.J. Kresge, R.A. More O’Ferrall, and M.F. Powell. In Iso-
topes in organic chemistry. Vol. 7. Edited by E. Buncel and C.
C. Lee. Elsevier, Amsterdam, The Netherlands. 1987. Chap. 4.
©
2007 NRC Canada