Carboxylic acid ionization constants by 19F NMR spectroscopy

Article abstract of DOI:10.1002/(SICI)1099-1395(199910)12:10<751::AID-POC190>3.0.CO;2-P

The ¹⁹F NMR spectra of 26 simple fluorinated carboxylic acids were measured in aqueous solutions of pH 03-10.0. Analysis of the fluorine chemical shift dependence on pH allowed the determination of ionization constants from the titration curves

Full text of DOI:10.1002/(SICI)1099-1395(199910)12:10<751::AID-POC190>3.0.CO;2-P

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 12, 751–757 (1999)
Carboxylic acid ionization constants by 19F NMR spectro-
scopy
Stefan E. Boiadjiev† and David A. Lightner*
Department of Chemistry, University of Nevada, Reno, Nevada 89557-0020, USA
Received 7 April 1999; revised 21 June 1999; accepted 22 June 1999
ABSTRACT: The 19F NMR spectra of 26 simple fluorinated carboxylic acids were measured in aqueous solutions of
pH 0.3–10.0. Analysis of the fluorine chemical shift dependence on pH allowed the determination of ionization
constants from the titration curves; the values agreed with known pKa values. The acidity of a fluorinated bilirubin
precursor was established using this method. Copyright 1999 John Wiley & Sons, Ltd.
KEYWORDS: 19F NMR; fluorinated carboxylic acids; ionization constants; pKa
INTRODUCTION
19F is 0.83 of that of protons) offer an opportunity to use
19F NMR spectroscopy for the examination of ionization
equilibria. The chemical shift (ꢀF) range of a 19F NMR
signal is intrinsically very wide, and therefore the fluorine
nucleus is an excellent, highly sensitive probe of its
environment. This fact was used more than 35 years ago
for pKa determinations of p-fluoroacetophenone (C–H
acidity) and p-fluorobenzamide (N–H acidity),7 and as
recently as 1998 for subtle changes of ꢀF arising from
solvent-induced isotope shifts due to enrichment of water
with H2 18O.8
The literature provides a number of examples in which
19F NMR spectroscopy was used to study the ionization
of protonated amines containing fluorine. Such amines
were designed to have pKas in the physiologically
important range (pH 6.5–8.0) and were used in the
development and application of NMR indicators for non-
invasive and accurate intracellular pH measurement.9–12
To the best of our knowledge, there have been no
systematic reports on titrations of simple carboxylic acids
and determination of their pKas using 19F NMR spectro-
scopy. In this paper we report on the changes of 19F NMR
chemical shifts associated with the ionization of
carboxylic acids in the (strongly) acidic pH range, show
how such changes can be used to measure their pKas and
apply the method to the pKa determination of a
fluorinated bilirubin precursor (1).
The natural yellow–orange pigment bilirubin is the end
product of heme catabolism in mammals.1 Bilirubin
contains two propionic acid groups, and their ionization
behavior is implicated in the pigment’s hepatic transport,
neurotoxicity, formation of gallstones and protein and
lipid membrane binding.2 In a recent investigation of the
inter-relationship between altered acidity of synthetic
bilirubin analogs and their solution properties, such as
stereochemistry, polarity, solubility and excretion, we
introduced substituents with a strong electron-withdraw-
ing effect in the vicinity of the carboxylic acid groups.3,4
Methoxy and methylthio substitution at the a-carbon of
each propionic acid of bilirubin was expected to decrease
the acid pKa by 1 unit, but this did not significantly
change the pigments’ overall properties.3 In contrast, a-
fluoro substitution, which was expected to decrease the
pKa by more than 2 units (CH3CO2H, pKa = 4.76;
FCH2CO2H, pKa = 2.585) led to drastically altered
properties.4 Synthetic a,a'-difluoromesobilirubin-XIIIa
is polar and water soluble, whereas natural bilirubin is
relatively non-polar, lipophilic and completely insoluble
in water. Water solubility was attributed to complete
ionization of the acid groups at pH ꢀ 7. Therefore, we
sought to obtain an independent quantitative estimate of
acidity of a,a'-difluoromesobilirubin-XIIIa and some of
its precursors.
The presence of a fluorine atom, with a nuclear spin
I = 1/2, the natural isotopic abundance of 100% and high
receptivity (a measure of the ease of detecting a nucleus;6
RESULTS AND DISCUSSION
A series of carboxylic acids bearing a single fluorine
reporter atom on an sp3-or sp2-hybridized carbon atom or
trifluoromethyl group attached to an aliphatic or aromatic
carbon atom were studied. The structures of the acids
investigated are shown in Scheme 1.
*Correspondence to: D. A. Lightner, Department of Chemistry,
University of Nevada, Reno, Nevada 89557-0020, USA.
†On leave from the Institute of Organic Chemistry, Bulgarian
Academy of Sciences.
Contract/grant sponsor: National Institute of Health; Contract/grant
number: HD 17779.
Copyright 1999 John Wiley & Sons, Ltd.
CCC 0894–3230/99/100751–07 $17.50

752
S. E. BOIADJIEV AND D. A. LIGHTNER
Titration curves of all the fluorinated carboxylic acids
1–26 were constructed by measuring their 19F NMR
spectra at 17 different pH values in the range 0.3–10.0 in
aqueous solution containing oxalic acid–potassium
oxalate. A small amount (5%, v/v) of DMSO-d6 was
added to preserve the solubility over the range of acids
selected in Table 1. The limiting fluorine chemical shift
values for the free acid and carboxylate anion, their
difference DꢀF and graphically estimated ionization
constants are listed in Table 1. The sensitivity values
shown in Table 1 are approximated by the ratio DꢀF/D
pH, where DpH is a range of Æ0.25 units from the pKa.
This range is the steepest part of the titration curve. The
maximum 19F NMR chemical shift sensitivity to pH
(highest slope) is found, as expected, near the pKa of the
fluorinated carboxylic acid under study. In this pH region
there are comparable concentrations of both the acid and
its conjugate base present in solution. Here, the ratio of
the two species, which interconvert rapidly on the NMR
time-scale, and hence the average observed 19F NMR
chemical shift are altered dramatically by small pH
changes.
In order to explore the feasibility of using 19F NMR to
determine carboxylic acid pKa values, we focused first on
fluorinated carboxylic acids of known pKa, where the pKa
Scheme 1
Table 1. 19F NMR data for carboxylic acids 1±26
b
d
Compound
ꢀF(acid)a
ꢀF(carboxylate)a
DꢀF
Sensitivityc
pKa
1
2
3
4
5
6
7
8
188.56
228.94
175.13
112.77
113.92
106.60
118.91
114.68
117.03
120.03
114.95
118.44
116.54
114.34
110.84
105.80
127.29
o-F 107.35
p-F 102.68
59.53
180.91
217.71
164.35
116.82
114.76
111.03
119.08
115.16
118.42
120.17
115.19
118.99
117.78
114.68
112.86
106.31
118.91
111.84
108.65
59.88
7.65
11.23
10.78
4.05
0.84
4.43
0.17
0.48
1.39
0.14
0.24
0.55
1.24
0.34
2.02
0.51
8.38
4.49
5.97
0.35
0.21
0.46
0.03
0.14
0.25
0.26
0.15
3.82
6.60
5.74
1.98
0.41
2.10
0.10
0.22
0.58
0.07
0.11
0.24
0.73
0.18
1.14
0.26
4.81
2.53
3.35
0.19
0.10
0.22
—
2.67
2.50
2.33
3.31
3.68
4.01
3.96
4.02
4.14
4.62
4.59
4.55
4.08
4.12
4.21
4.44
2.61
3.29
3.29
2.73
3.90
3.77
—
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
62.86
63.20
60.32
62.63
62.48
64.05
73.78
62.64
62.74
60.29
62.49
62.23
63.79
73.63
0.07
0.10
0.15
0.08
4.19
3.99
3.17
4.05
a
Values reported in ꢀ(ppm) from the chemical shift in Hz and the exact spectrometer frequency of the individual measurement.
b
c
d
DꢀF = ꢀF (acid) ꢀF (carboxylate).
DꢀF/DpH.
Determined from the graph of ꢀF vs pH. The spectra of 1–18 were referenced to C6F6 at 162.90 ppm and the spectra of 19–26 to CFCl3 at
0.00 ppm.
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)

CARBOXYLIC ACID IONIZATION CONSTANTS
753
pKas determined in its absence, the pKas of water-soluble
acids were re-determined in the absence of DMSO-d6.
The sodium salt of 2 gave pKa = 2.54 in aqueous oxalate
buffer vs 2.53 in the same buffer with 5% (v/v) DMSO-d6
present.
For additional reference standards to use in calibrating
the 19F NMR method to determine carboxylic acid pKa
values, we turned to the three isomeric fluorobenzoic
acids. The 19F NMR spectra of 3- and 4-fluorobenzoic
acids have previously been measured in acid and base but
without a description of their full titration curves.15 All
three isomeric fluorobenzoic acids also showed very
good coincidence between 19F NMR-determined and
literature pKas: for ortho-isomer 4, 3.31 vs 3.2714; for
meta-isomer 5, 3.68 vs 3.86 Æ 0.04;5,14,16,17 and for
para-isomer 6, 4.01 vs 4.15 Æ 0.0055,14,17 (see Fig. 1 for
titration curves). Sigmoidal curve fitting gave essentially
the same pKa values: 4, 3.31; 5, 3.69; and 6, 3.99. Further
comparisons were made between 19F NMR-determined
pKas and literature values, e.g. for 20 (although with
lower confidence; see below) with an NMR-determined
pKa of 3.90 and potentiometrically measured pKa =
3.75,17 and for 26 with an NMR-determined pKa =
4.05 vs pKa = 4.15 found for the very similar 4,4,4-
trifluorobutanoic acid.18
Figure 1. Variation of 19F NMR chemical shift with pH for
aqueous solutions of ¯uorobenzoic acids: *,4; !,5; ^,6
was determined by classical means. Fluoroacetic acid (2,
as its sodium salt) was examined first for this purpose.
Among all the acids studied, it showed greatest downfield
chemical shift change (DꢀF = 11.23 ppm) upon ioniza-
tion. Plotting ꢀF as function of pH showed a typical
titration curve, from which one could determine the pKa
by approximate graphical means.
The equilibrium (acidity) constant Ka of an acid HA
ionization [Eqn. (1)] is related to the observed chemical
shift by Eqn. (2), where XHA is the mole fraction of the
non-ionized acid:13
Our conclusion from the comparisons with literature
pKas is that 19F NMR spectroscopy is very well suited for
pKa determination in the (strongly) acidic pH range,
much as was shown to be for protonated amines in the
physiological pH region.10–12
‡
As shown in Figure 1, and found in Table 1, when the
fluorine is on a phenyl ring, carboxylic acid ionization
causes an upfield shift of the 19F resonance. This is
consistent with that found in protonated aniline deriva-
tives10 or imidazoles,12 where in all cases the fluorine
was on a phenyl ring. Hybridization at carbon is not the
sole reason for such behavior, because a-fluorocinnamic
acid (17) exhibited a downfield shift, as did 1, 2, 3 and
all trifluoromethyl-containing compounds, except 19
(Table 1). Comparison of DꢀF and sensitivity
(Table 1) confirmed earlier observations that a phenyl
p-system transmits the polar effects to the fluorine well.19
Hence the total change in ꢀF upon ionization of 4, 5 and 6
is 4.05, 0.84 and 4.43 ppm, respectively (Fig. 1).
In order to examine the ability of fluorine to act as a
remote sensor for carboxylic acid ionization, we studied
fluorophenylacetic (7–9) and 3-(fluorophenyl)propionic
acids (10–12) (Figs 2 and 3, respectively). When the
fluorine reporter is insulated from the ionizable COOH by
one (as in 7–9) or two methylene groups (as in 10–12),
the sensitivity is greatly diminished (Table 1). The trend
is toward smaller DꢀF values as the distance between
fluorine and carboxyl increases for a given series, e.g. in
the para-isomers. The limit of detection was reached for
o-fluorophenylacetic acid (7) and o- (10) and m-3-
(fluorophenyl)propionic acid (11), showing DꢀF = 0.17,
0.14 and 0.24 ppm, respectively. Although the titration
HA ‡ H2O „ A ‡ H3O
ꢁ1†
‡
‰A Š‰H3O Š
Ka ˆ
‰HAŠ
ꢀFꢁobserved† ˆ XHAꢀFꢁacid† ‡ ꢁ1 XHA
ꢀFꢁcarboxylate†
†
ꢁ2†
‡
‰H3O ŠꢀFꢁacid† ‡ KaꢀFꢁcarboxylate†
ˆ
‡
‰H3O Š ‡ Ka
‡
When [H3O ] = Ka, the observed ꢀF is the average of the
limiting ꢀF(acid) and ꢀF(carboxylate) shifts. The mean ꢀF
value corresponds to a solution containing equal
concentrations of acid and its conjugate base. Conse-
quently, from the graph point coordinate ꢀF = 1/2
[ꢀF(acid) ‡ ꢀF(carboxylate)], the value of pH = pKa can
be determined. For acid 2, pKa = 2.50 was found by this
approach, in excellent agreement with that in the
literature, pKa = 2.58 Æ 0.04.5,6,14 Alternatively, a sig-
moidal curve-fitting program gave pKa = 2.53, in ex-
cellent agreement.
In order to ensure that the presence of the small amount
of DMSO-d6 (5% v/v) used to maintain solution
homogeneity did not cause a significant deviation from
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)

754
S. E. BOIADJIEV AND D. A. LIGHTNER
Figure 2. Variation of 19F NMR chemical shift with pH for
aqueous solutions of ¯uorophenylacetic acids: *,7; !,8;
^,9
Figure 3. Variation of 19F NMR chemical shift with pH for
aqueous solutions of ¯uorophenylpropionic acids: *,10;
!,11; ^,12
curve can be barely seen in Fig. 3, such small DꢀF values
fall within instrumental and experimental error. Hence
DꢀF values smaller than 0.3 ppm should be considered
unreliable for pKa measurements.
of the para-fluorine did not decrease the pKa of 18 below
that of o-fluorobenzoic acid (4), pKa = 3.31.
In an attempt to lower the NMR detection threshold by
increasing the number of identical fluorines per molecule,
the spectra of trifluoromethyl-containing acids 19–26
were measured. However, the CF3 substituent proved less
effective than a directly attached fluorine, cf. monofluoro
p-fluorobenzoic acid 6 (DꢀF = 4.43 ppm) or even 9 (DꢀF =
1.39 ppm) with the corresponding p-trifluoromethylben-
zoic acid 21 (DꢀF = 0.46 ppm). The reason for such
diminished sensitivity of the trifluoromethyl probe is not
clear but might be rooted in the inherently narrower 19F
chemical shift dispersion range ( 55 to 65 ppm) in
RCF3 compounds over a wide range of R groups, i.e. even
if R = OR' the range is the same.20,21 In contrast, one
fluorine atom on a secondary aliphatic carbon has
chemical shift dispersion range of 160 to 230 ppm,
and in ArF it is 100 to 200 ppm. Thus, the last case
with a wide dispersion might be expected to be the most
responsive to subtle changes in the nature of Ar, while
CF3 would be intrinsically insensitive to changes.
Curiously, the drop in DꢀF with increasing distance
between sensor and ionization center is greater with the
ortho- and meta- isomers than with para-isomers and
greatest with the ortho-isomers. Apparently the number
of intervening bonds is not the only factor. In the o-fluoro
acids 7 and 10, DꢀF drops faster than in the para-isomers;
e.g. in 12, although the fluorine is eight bonds removed
from the ionization center, it still senses it well (DꢀF
= 0.55 ppm), whereas in 10, with a six-bond separation,
DꢀF is only 0.14 ppm (Fig. 2). The influence on pKa,
however, is opposite: o-fluoro acids are stronger than p-
fluoro acids (compare 4 with 6 and 7 with 9). Fluorine
substitution does not exert a significant effect on the
acidity of 10–12. From comparison of the data from 4–
12, it follows that a fluorine at the para-position has the
best sensitivity.
Fluorine detection of ionization is much more effective
when the phenyl ring is conjugated with rather than
insulated from the carboxyl group, as in cinnamic acids
13–15 vs the corresponding hydrocinnamic (phenylpro-
pionic) acids 10–12 (Table 1). For example, DꢀF =
2.03 ppm for 15 is much larger than DꢀF = 0.55 ppm for
the saturated acid 12. Fluorine substitution on the phenyl
ring of 13–15 increases their acidity vs that of the parent
cinnamic acid (pKa = 4.44).14
The 19F NMR spectra of pyrrolepropionic acid 1
measured at various pH are presented in Fig. 4. This a-
fluorinated acid showed a favorable downfield chemical
shift change DꢀF = 7.65 ppm upon ionization, and a
sensitivity of 3.82 ppm per pH unit near the pKa. From
Fig. 4 it can be seen that not only does the chemical shift
change with pH, but so does the splitting pattern of the
fluorine multiplet. The titration curve of 1, displayed in
Fig. 5, had an inflection point at pH 2.67 corresponding to
its pKa. This 19F NMR-determined pKa of 1 is very close
to those of fluoroacetic acid (2) (DꢀF = 11.23 ppm,
pKa = 2.50) and a-fluorocinnamic acid (17) (DꢀF =
8.38 ppm, pKa = 2.61). Only a-fluorophenylacetic acid
(3) is a stronger acid (DꢀF = 10.78 ppm, pKa = 2.33).
Hence the decrease in pKa due to a single fluorine alpha
to the carboxylic acid group is ꢂ2 pK units, an estimate in
2,4-Difluorobenzoic acid (18) offered the opportunity
to follow simultaneously two different fluorine NMR
signals. The para-fluorine of 18 exhibited a larger
chemical shift change (DꢀF = 5.97 ppm) and greater
sensitivity than those of the ortho-fluorine (DꢀF =
4.48 ppm). It was comforting to note that analysis of
the titration curves from either fluorine signal provided
an identical pKa value (3.29). Interestingly, the presence
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)

CARBOXYLIC ACID IONIZATION CONSTANTS
755
Figure 5. Variation of 19F NMR chemical shift with pH for
aqueous solutions of a-¯uoro(pyrrole)propionic acid (1)
Varian Unity Plus spectrometer at 470.254 MHz in
5 mm tubes, and were referenced against external
standards: hexafluorobenzene (5 Æ 0.02 mM in CHCl3)
at ꢀ 162.90 ppm or CFCl3 (10 Æ 0.05 mM in CHCl3)
at ꢀ 0.00 ppm, with shifts upfield from CFCl3 being
negative.20 Typical experimental parameters were flip
angle 55°, interpulse delay 3 s, collecting 128
transients, and spectral width 35 kHz using 70K data
points. Each FID was zero filled to 262K and
multiplied with an exponential function (line broad-
ening 0.1 Hz) prior to Fourier transformation to give a
0.27 Hz digital resolution. Proton decoupling was not
applied and consistently the chemical shift of a selected
prominent line was followed. pH measurements were
made using an Orion Model 811 pH meter with an
Orion Model 91-02 combination electrode calibrated
twice for the range 0.3–10.0 at pH 4.00, 7.41 and
10.00. Potassium tetraoxalate buffers (50 mM)14 were
used throughout, the pH being adjusted with HCl or
KOH. The pH values reported in Table 1 refer to those
of freshly prepared aqueous oxalic acid–oxalate con-
taining solutions before dissolving the fluorinated acid.
A stock solution of the each acid 1–26 (concentration
40 Æ 1 mM) was prepared in DMSO-d6. A 100 ml
aliquot was diluted to a volume of 2 ml at each pH
of the oxalate buffer, giving a total fluorinated acid
concentration of 2.00 Æ 0.05 mM and 5% (v/v) DMSO-
d6 in the NMR samples. The pH values of the NMR
solutions were then rechecked and showed only small
changes of 0.03–0.25 pH units, with larger variations
occurring with the more acidic compounds. These
findings are consistent with previous studies that
showed that very low concentrations of DMSO exert
only a very small effect on buffer pH.22
Figure 4. Changes of 19F NMR chemical shift and 19F1H
coupling pattern of a-¯uoro(pyrrole)propionic acid (1) with
pH
good agreement with literature data on the effect of one
strongly electronegative fluorine substituent. The pKa of
1 provides suggestive evidence that the acidity of the
bilirubin synthesized from 14 would also show a
corresponding increase (ꢂ100-fold), and this doubtless
contributes to its peculiar properties, such as its aqueous
solubility.
Further work is in progress (P. B. Karadakov,
University of Surrey, UK) to understand the observed
19F NMR chemical shifts and DꢀF by applying ab initio
calculations.
CONCLUSIONS
The fluorine nucleus appears to be an excellent probe for
monitoring by NMR ionization equilibria in acidic
aqueous medium. Its wide chemical shift dispersion
range allows accurate pKa determinations for a variety of
carboxylic acids. The method defines the expected high
acidity of a fluorinated bilirubin precursor to be
pKa = 2.67.
The small amount of added DMSO-d6 served as an
internal NMR deuterium lock and was used to maintain
solution homogeneity. Whereas a number of the acids
(e.g. 2, 4, 5 and 25) used in this work are soluble in the
EXPERIMENTAL
The 19F NMR spectra were acquired at 25 Æ 1°C on a
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)

756
S. E. BOIADJIEV AND D. A. LIGHTNER
3
oxalate buffer over the entire pH range, others were not.
When the measurements were repeated on soluble acids
in the absence of DMSO-d6, the same pKas were
measured as in the presence of DMSO-d6.
2JFC = 15.1 Hz), 128.48 (d, JFC = 8.2 Hz), 130.97 (d,
3JFC = 4.3 Hz), 160.88 (d, 1JFC = 246.4 Hz), 168.65 ppm.
Methyl 3-(3-Fluorophenyl)-2-methoxycarbonylpropio-
nate. Obtained in 65% yield, b.p. 126–128°C (1 mmHg).
1H-NMR, ꢀ 3.21 (2H, d, J = 7.8 Hz), 3.66 (1H, t,
1H NMR spectra (at 500.6 MHz) and 13C NMR spectra
(at 125.9 MHz) were measured in CDCl3 and referenced
to the residual CHCl3 1H signal at 7.26 ppm and the
CDCl3 13C signal at 77.00 ppm.
3
J = 7.8 Hz), 3.71 (6H, s), 6.83–6.94 (3H, m, JFH = 9.4,
7.8 Hz), 7.13–7.22 (1H, m) ppm; 13C NMR, ꢀ 34.11,
2
Commercial fluorinated acids 2 (as sodium salt), 3, 4,
5, 6, 8, 19, 20, 21 and 23 were obtained from Aldrich, 7,
9, 17, 22 and 24 from Acros, 13, 15, 18, 25 and 26 from
Oakwood Products and 14 and 16 from Lancaster. The
syntheses of 1, 10, 11 and 12 are described below.
52.32, 53.00, 113.50 (d, JFC = 21.1 Hz), 115.45 (d,
4
2JFC = 21.3 Hz), 124.24 (d, JFC = 2.3 Hz), 129.83 (d,
3
3JFC = 8.4 Hz), 140.06 (d, JFC = 7.4 Hz), 162.53 (d,
1JFC = 245.7 Hz), 168.71 ppm.
Methyl 3-(4-Fluorophenyl)-2-methoxycarbonylpropio-
nate. Obtained in 70% yield, b.p. 125–127°C (1 mmHg).
1H NMR, ꢀ 3.19 (2H, d, J = 7.8 Hz), 3.63 (1H, t,
3-(2,4-Dimethyl-5-methoxycarbonyl-1H-pyrrol-3-yl)-
2-¯uoropropionic acid (1). A mixture of the corre-
sponding dimethyl ester4 (257 mg, 1 mmol), 6 ml of
methanol and 1.25 ml (1.25 mmol) of 1 M aqueous
sodium hydroxide was stirred at room temperature for
24 h. The methanol was evaporated under vacuum. The
residue was dissolved in 10 ml of 0.2 M sodium
hydroxide, extracted with diethyl ether (10 ml) and the
aqueous layer was acidified with 10% hydrochloric acid.
The solution was concentrated under vacuum to ꢂ3 ml
and the precipitated semi-solid product was recrystallized
from methanol–water to afford 104 mg (43%) of acid 1,
m.p. 186–187°C. 1H NMR [CDCl3 ‡ 20% (v/v) DMSO-
d6], ꢀ 1.75 (3H, s), 1.79 (3H, s), 2.44 (1H, m), 2.54 (1H,
3
J = 7.8 Hz), 3.70 (6H, s), 6.96 (2H, dd, JHH = 8.7 Hz,
3JFH = 8.7 Hz),
7.15
(2H,
dd,
3JHH = 8.7 Hz,
4JFH = 5.4 Hz) ppm; 13C NMR, ꢀ 33.72, 52.36, 53.43,
2
3
115.17 (d, JFC = 21.4 Hz), 130.16 (d, JFC = 7.9 Hz),
4
1
133.23 (d, JFC = 2.9 Hz), 161.56 (d, JFC = 244.7 Hz),
168.86 ppm.
Fluorohydrocinnamic acids. General procedure. A
solution of 2.40 g (10 mmol) of the foregoing dimethyl
ester in 20 ml of methanol was mixed with a solution of
2.00 g (50 mmol) of sodium hydroxide in 10 ml of water
and the mixture was refluxed for 3 h. The methanol was
evaporated under vacuum and the residue was acidified
with concentrated hydrochloric acid. Sodium chloride
(ꢂ 5 g) was added and the diacid was extracted with di-
ethyl ether (5–6 Â 25 ml). The residue, after evaporation
of the diethyl ether, was heated on an oil-bath at 145–
150°C for 1 h. After cooling, the crude monoacid was
purified by radical chromatography on silica gel (eluent
1.5–2% methanol in dichloromethane) followed by
recrystallization from dichloromethane–hexane.
2
3
m), 3.34 (3H, s), 4.37 (1H, ddd, JFH = 49.6 Hz, JHH
=
7.9, 4.0 Hz), 10.14 (1H, s), 12.35 (1H, br s) ppm; 13C
NMR [CDCl3 ‡ 20% (v/v) DMSO- d6], ꢀ 9.57, 10.22,
2
1
26.39 (d, JFC = 21.9 Hz), 49.46, 88.07 (d, JFC
=
3
184.0 Hz), 113.81 (d, JFC2 = 2.1 Hz), 115.40, 126.26,
131.14, 160.69, 170.24 (d, JFC = 23.8 Hz) ppm.
Alkylation of dimethyl malonate with ¯uorobenzyl
chlorides. General procedure. To a solution of freshly
prepared from sodium under N2 solution of sodium
methoxide (0.1 mol) in 50 ml of anhydrous methanol was
added 0.1 mol of dimethyl malonate in 10 ml of methanol
over 10 min through a septum. After 10 min of stirring
the corresponding fluorobenzylchloride (0.05 mol) was
added neat over 15 min and the mixture was refluxed for
5 h. After cooling, half of the solvent was evaporated
under vacuum and the residue was diluted with 50 ml of
water and acidified with 5% hydrochloric acid. The
product was extracted with diethyl ether (3 Â 50 ml). The
combined extracts were washed with water (3 Â 30 ml),
dried (anhydrous MgSO4) and filtered and, after remov-
ing the solvent, the product was distilled under vacuum.
3-(2-Fluorophenyl)propionic acid (10). Obtained in
90% yield, m.p. 76–77°C. 1H NMR, ꢀ 2.70 (2H, t,
J = 7.7 Hz), 2.99 (2H, t, J = 7.7 Hz), 7.00–7.11 (2H, m),
7.18–7.27 (2H, m), 11.98 (1H, br s) ppm; 13C NMR, ꢀ
24.21 (d, 3JFC = 2.2 Hz), 34.12, 115.30 (d,
4
2JFC = 21.8 Hz), 124.07 (d, JFC = 3.4 Hz), 126.90 (d,
3
2JFC = 15.6 Hz), 128.20 (d, JFC = 8.1 Hz), 130.54 (d,
1
3JFC = 4.6 Hz), 161.10 (d, JFC = 245.3 Hz), 179.38 (d,
5JFC = 3.0 Hz) ppm.
3-(3-Fluorophenyl)propionic acid (11). Obtained in
85% yield, m.p. 44–45°C. 1H NMR, ꢀ 2.69 (2H, t,
J = 7.6 Hz), 2.96 (2H, t, J = 7.6 Hz), 6.88–7.03 (3H, m,
3JFH = 9.0, 7.7 Hz), 7.21–7.30 (1H, m), 11.96 (1H, br
Methyl 3-(2-Fluorophenyl)-2-methoxycarbonylpropio-
nate. Obtained in 67% yield, b.p. 127–128°C (1 mmHg).
1H NMR, ꢀ 3.25 (2H, d, J = 7.8 Hz), 3.70 (6H, s), 3.75
(1H, t, J = 7.8 Hz), 6.98–7.07 (2H, m), 7.17–7.23
(2H, m) ppm; 13C NMR, ꢀ 28.21, 51.43, 52.13, 114.96
(d, 2JFC = 21.8 Hz), 123.80 (d, 4JFC = 3.4 Hz), 124.25 (d,
s) ppm; 13C NMR,
ꢀ 30.15, 35.25, 113.29 (d,
2
2JFC = 21.0 Hz), 115.18 (d, JFC = 21.2 Hz), 123.89 (d,
4JFC = 1.7 Hz), 129.97 (d, JFC = 8.3 Hz), 142.56 (d,
3
3JFC = 7.3 Hz), 162.85 (d, 1JFC = 245.7 Hz), 179.28 ppm.
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)

CARBOXYLIC ACID IONIZATION CONSTANTS
757
Methods, Part One, edited by K. W. Bently, Chapt. VI, pp. 319–
412. Wiley, New York (1963).
6. H. Gu¨nther, NMR Spectroscopy, 2nd ed., p. 463. Wiley, New York
(1994).
7. R. W. Taft and P. L. Levins, Anal. Chem. 34, 436–437 (1962).
8. J. R. P. Arnold and J. Fischer, J. Chem. Soc., Chem. Commun.
1859–1860 (1998).
9. P. W. Kuchel, in Analytical NMR, edited by L. D. Field and S.
Sternhell, pp. 184–189. Wiley, New York, (1989).
10. C. J. Deutsch and J. S. Taylor, Ann. N. Y. Acad. Sci. 508, 33–47
(1987).
3-(4-Fluorophenyl)propionic acid (12). Obtained in
86% yield, m.p. 86–87°C. 1H NMR, ꢀ 2.67 (2H, t,
J = 7.6 Hz), 2.93 (2H, t, J = 7.6 Hz), 6.98 (2H, dd,
3
3
3JHH = 8.7 Hz, JFH = 8.7 Hz), 7.16 (2H, dd JHH
=
4
8.7 Hz, JFH = 5.5 Hz), 11.69 (1H, br s) ppm; 13C NMR,
2
ꢀ 29.66, 35.70, 115.28 (d, JFC = 21.4 Hz), 129.67 (d,
3JFC = 7.9 Hz), 135.68 (d, JFC = 3.2 Hz), 161.47 (d,
4
1JFC = 244.3 Hz), 179.45 ppm.
11. C. J. Deutsch and J. S. Taylor, Biophys. J. 55, 799–804 (1989).
12. J. L. Harper, R. A. J. Smith, J. J. Bedford and J. P. Leader,
Tetrahedron 53, 8211–8224 (1997).
13. D. L. Rabenstein and T. L. Sayer, Anal. Chem. 48, 1141–1146
(1976).
14. A. I. Vogel, Vogel’s Textbook of Quantitative Chemical Analysis,
5th ed. Longman, Harlow (1989).
15. I. R. Fox, P. L. Levins and R. W. Taft Jr, Tetrahedron Lett. 249–
252 (1961).
Acknowledgements
We thank the US National Institutes of Health (HD
17779) for generous support of this work.
16. A. J. Hoefnagel, M. A. Hoefnagel and B. M. Wepster, J. Org.
Chem. 43, 4720–4745 (1978).
REFERENCES
17. M. Ludwig, V. Baron, K. Kalfus, O. Pytela and M. Vecˇerˇ, Collect
Czech. Chem. Commun. 51, 2135–2142 (1986).
18. E. T. McBee, O. R. Pierce and D. D. Smith, J. Am. Chem. Soc. 76,
3722–3725 (1954).
19. W. Adcock, M. J. S. Dewar, R. Golden and M. A. Zeb, J. Am.
Chem. Soc. 97, 2198–2205 (1975).
1. A. F. McDonagh, in The Porphyrins, edited by D. Dolphin, Vol.
VI, Chapt. 6, pp. 293–491. Academic Press, New York (1979).
2. C. Tribelli and J. D. Ostrow, Hepatology 24, 1296–1311 (1996).
3. S. E. Boiadjiev and D. A. Lightner, J. Org. Chem. 63, 6220–6228
(1998).
20. V. Wray, Annu. Rep. NMR Spectrosc. 10B, 1–507 (1980).
21. C. H. Dungan and J. R. Van Wazer, Compilation of Reported 19F
NMR Chemical Shifts. Wiley, New York (1970).
22. D. L. Holmes and D. A. Lightner, Tetrahedron 51, 1607–1622
(1995).
4. S. E. Boiadjiev and D. A. Lightner, J. Org. Chem. 62, 399–404
(1997).
5. (a) E. P. Serjeant and B. Dempsey, Ionization Constants of Organic
Acids in Aqueous Solution. Pergamon Press, Oxford (1979); (b) J.
F. King, in Elucidation of Structures by Physical and Chemical
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 751–757 (1999)