Steric effects and mechanism in the formation of hemi-acetals from aliphatic aldehydes

Article abstract of DOI:10.1002/poc.3138

Some physical properties (pK_a, log P_OW, boiling points) of hexanoic acid 1 (X = COOH) and its seven isomers 2, 3, 4, 5, 6, 7, 8 (X = COOH) are reported. Hexanal 1 (X = CHO) and its seven isomeric aldehydes 2, 3, 4, 5, 6, 7, 8 (X = CHO) are shown to equilibrate, in methanol solution, with their hemi-acetals. Logarithms of equilibrium constants correlate with values of E_s for the isomeric C₅H₁₁ substituents, and with logs of relative rates for saponification of the corresponding methyl esters with ρ = 0.52, reflecting the reduced steric demand of hydrogen compared to oxygen in the quaternization of ester and aldehydic carbonyl groups. Rates of equilibration have also been measured in buffered methanol. For hexanal, with a 2:1 Et₃N:AcOH buffer, the buffer-independent contribution is dominated by the methoxide catalysed pathway. Rates in this medium have been determined for isomers 1, 2, 3, 4, 5, 6, 7, 8 (X = CHO), and their logarithms do not correlate with logarithms of equilibrium constants for hemi-acetal formation or with substituent steric parameters derived from ester formation or saponification, indicating that the steric changes associated with full quaternization of the carbonyl group are not mirrored in the transition structures for hemi-acetal formation. It is suggested that transition states for hemi-acetal formation are relatively early so that steric interactions are effectively those between the nucleophile and ground state conformations of the aldehydes. A comparison of the entropies of hemi-acetal formation with entropies of activation has provided a basis for a suggested transition structure. Comparisons with acid chloride hydrolyses are made. Copyright 2013 John Wiley & Sons, Ltd. Logarithms of equilibrium constants for formation hemi-acetals of hexanal and its seven isomeric aldehydes correlate well with values of E_s for the isomeric C₅H₁₁ substituents, and with logs of relative rates for saponification of the corresponding methyl esters. Logarithms of rate constants for hemi-acetal formation do not, indicating that the steric changes associated with full quaternization of the carbonyl group are not mirrored in the transition structures for hemi-acetal formation. The reasons for this are discussed. Copyright

Full text of DOI:10.1002/poc.3138

Special Issue Article
Received: 23 January 2013,
Revised: 11 April 2013,
Accepted: 17 April 2013,
Published online in Wiley Online Library: 10 June 2013
(wileyonlinelibrary.com) DOI: 10.1002/poc.3138
Steric effects and mechanism in the formation
of hemi-acetals from aliphatic aldehydes†
Graham Dawa, Andrew C. Regana, C. Ian F. Watta* and Evan Wooda
Some physical properties (pKa, log POW, boiling points) of hexanoic acid 1 (X = COOH) and its seven isomers
2–8 (X = COOH) are reported.
Hexanal 1 (X = CHO) and its seven isomeric aldehydes 2–8 (X = CHO) are shown to equilibrate, in methanol solution,
with their hemi-acetals. Logarithms of equilibrium constants correlate with values of Es for the isomeric C5H11
substituents, and with logs of relative rates for saponification of the corresponding methyl esters with r = 0.52,
reflecting the reduced steric demand of hydrogen compared to oxygen in the quaternization of ester and aldehydic
carbonyl groups. Rates of equilibration have also been measured in buffered methanol. For hexanal, with a 2:1 Et3N:
AcOH buffer, the buffer-independent contribution is dominated by the methoxide catalysed pathway. Rates in this
medium have been determined for isomers 1 – 8 (X = CHO), and their logarithms do not correlate with logarithms of
equilibrium constants for hemi-acetal formation or with substituent steric parameters derived from ester formation
or saponification, indicating that the steric changes associated with full quaternization of the carbonyl group are not
mirrored in the transition structures for hemi-acetal formation. It is suggested that transition states for hemi-acetal
formation are relatively early so that steric interactions are effectively those between the nucleophile and ground
state conformations of the aldehydes. A comparison of the entropies of hemi-acetal formation with entropies of
activation has provided a basis for a suggested transition structure. Comparisons with acid chloride hydrolyses
are made. © 2013 The Authors. Journal of Physical Organic Chemistry published by John Wiley & Sons Ltd.
Keywords: aliphatic aldehydes; methyl hemi-acetals; rates and equilibria; steric effects; mechanism
be assigned to a small difference in steric demand in the transi-
tion states for acid- and base-catalysed ester reactions.
In an earlier paper,[1] we noted that the volumes and surface
Log kHOꢀ ¼ 0:963ðꢁ0:044Þ log kHþ þ 0:022ðꢁ0:060Þ (1)
areas of the saturated hydrocarbon groupings of molecular for-
INTRODUCTION
mula C5H11 were remarkably independent of both connectivity
Logarithms of relative rates for the acid-catalysed esterifications
provide the traditional steric substituent parameters, Es, and we
and conformation. Following a suggestion by Charton,[2] we
have pursued the possibility that these residues then present a
have tabulated those for the C5H11 isomers in our earlier paper.[1]
set of eight readily available substituents (see Fig. 1) whose
It is clear that relative rates for the ester saponifications might
polarity, polarizability and solvation properties show little variation
equally be used to compare steric demand within this series.
and might be used to probe substituent effects on functional
As a preliminary to the study of reactivity presented in this
group reactivity which are purely steric in origin, i.e. arising only
paper, we report some physical measurements on the set of
from interactions between filled orbitals on atoms of the substitu-
carboxylic acids. Many of these are already available in the liter-
ature, and we have found no significant disagreement with
earlier measurements, but the collection presented here (Table 1)
offers, at least, an internal consistency in terms of operator,
ent and the reacting set in the functional group and reagent.
We have already compared relative reactivities in acid-
catalysed methyl esterification of hexanoic acid and its isomers,
1 – 8 (X = COOH) with those for saponification of the same set
reagents and apparatus. Within the set, refractive indices show
of esters. In devising his set of group steric parameters, Taft,[3,4]
suggested these reactions might have similar sensitivities to
steric interaction with the substituent and pointed out that
similar experiments applied to m- or p-substituted benzoic acids
or benzoates demonstrated the very different sensitivities of
these reactions to polar effects, with r = ꢀ0.23 for the acid-
catalysed esterification[3] and r = +2.26 for saponification[5] in
Hammett treatments. The plot of logarithms of relative rates,
kHOꢀ, for the saponification reactions against those, kH+, for
the acid-catalysed esterifications for isomers, 1 – 8 (X = COOMe
or COOH), Fig. 2, yielded a good linear correlation (see Eqn 1,
R2 = 0.988) with a slope close to one, with reactivities decreasing
130-fold from 1 (X = COOR) to 8 (X = COOR). The correlation
excludes any major variation of electronic properties within the
series; the minor deviation from unit slope might reasonably
* Correspondence to: C. Ian F. Watt, The School of Chemistry, University of
Manchester, Manchester M13 9PL, UK.
E-mail: Ian.Watt@manchester.ac.uk
†
This article is published in Journal of Physical Organic Chemistry as a Special
issue: In Memoriam edited by Michael Page (University of Huddersfield,
Chemistry, Queensgate, Huddersfield, HD1 3DH, United Kingdom).
a
G. Daw, A. C. Regan, C. I. F. Watt, E. Wood
The School of Chemistry, University of Manchester, Manchester M13 9PL, UK
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
The copyright line for this article was changed on July 16, 2014.
J. Phys. Org. Chem. 2013, 26 1048–1057
© 2013 The Authors. Journal of Physical Organic Chemistry published by John Wiley & Sons Ltd.

STERIC EFFECTS IN THE FORMATION OF HEMI-ACETALS
electron donation from methyl substituents at the a-position,
but that suggestion is difficult to reconcile with the finding that
3,3-dimethylbutanoic acid, 5 (X = COOH), with the methyl substi-
tution at the b-position, has almost equal acidity with pKa = 4.97.
We suggest that the more plausible explanation for the slightly
reduced acidities of 8 and 5 (X = COOH) is steric hindrance to sol-
vation by hydrogen bonding, tighter in the negatively charged
carboxylate anion than in the neutral carboxylic acid. If the effect
of these substituents on reactivity was purely electronic in origin,
and the saponification of methyl esters had r = 2.23 (against
Hammett constants), then the pKa spread could account for a
factor of only 6.9 in the reactivity across the eight isomers, rather
than the observed 130-fold.
The major portion of this work is a study of the addition of
methanol to the corresponding aldehydes, 1 – 8 (X = CHO) in
dilute methanol solution, yielding the hemi-acetals (see Fig. 3),
selected because it offered the possibility of measurement of
both rates and equilibria, a situation remarkably rare in organic
chemistry. This then permits the informative testing of the
relationship between ΔGreaction and ΔG{ for the same reaction
(a possible Type 1 linear free energy relationship[6]) as well as
of correlation of those measurements with substituent parame-
ters defined through other reactions.
Figure 1. Structures of the eight isomers of C5H11
X
little variation, 1.413 (ꢁ0.002) at 21 ꢂC, with no indication of
trend with structure, at least supportive of constant polarizabil-
ity. Other properties (octanol/water distribution coefficient, boil-
ing point) show some variation but with no obvious relationship
to structure. The variation, however, is considerably less than
that associated with addition or removal of a single CH2 group.
For example, 1.55 < log Pow <1.98 for the set, while for the ho-
mologues, pentanoic, hexanoic and heptanoic acid values are
1.39, 1.98 and 2.49, respectively. For the boiling points, 186 < BP
204 ꢂC within the isomer set, while those for pentanoic, hexanoic
and heptanoic acid are 184, 204 and 223 ꢂC respectively.
RESULTS
Preparative chemistry
Hexanal 1 (X = CHO), 2-methylpentanal 4 (X = CHO) and 2-
ethylbutanal 7 (X = CHO) were commercially available. All other
aldehydes were prepared by oxidation of the corresponding
primary alcohols, themselves available by lithium aluminium
hydride reduction of methyl esters of the corresponding acids.
A number of oxidation methods were investigated, but, in our
hands, a minor modification of the method of Anelli et al.[7]
(oxidation by bleach, catalysed by TEMPO and potassium
bromide) was most effective, routinely giving “GC yields” of over
75%. Isolated yields of analytically pure material, however, were
usually ca 30%, with the losses arising mainly in ensuring that
the material produced was free from contamination by solvent
or hydrate. A representative procedure is given in the experi-
mental section, but we note that there were particular difficulties
in preparing aldehyde 8 (X = CHO) which, on standing in impure
form, deposited a white solid, yet to be identified, but which was
not the oxidation product. Distilled material was much more
stable, but was stored under an inert atmosphere and distilled
immediately before use.
Over the set, 4.67 < pKa < 5.00, with 2,2-dimethylbutanoic
acid, 8 (X = COOH) being the weakest acid with pKa = 5.00. This
reduced acidity might be associated with a small inductive
Figure 2. Comparison of relative reactivities in acid-catalysed esterifica-
tion and in ester saponification
Table 1. Some physical properties of the isomeric acids
aLog P(OW) bpKa
ꢂ
Acid 1–8 (X = COOH)
BP C
Initial observations
Hexanoic 1
204.2–204.8
198.6–198.9
198.4–198.5
195.9–196.1
186.0–186.4
191.4–191.8
190.8–191.0
192.9–193.3
1.98
1.75
1.68
1.74
1.64
1.55
1.74
1.63
4.81
4.78
4.68
4.75
4.97
4.75
4.67
5.00
The behaviour of solutions of the aldehydes in d4-methanol was
first examined by 1H-NMR spectroscopy. Figure 4 shows the
4-methylpentanoic 2
3-methylpentanoic 3
2-methylpentanoic 4
3,3-dimethylbutanoic 5
2,3-dimethylbutanoic 6
2-ethylbutanoic 7
2,2-dimethylbutanoic 8
aPOW against dil. HCl to suppress any effect of dissociation of
the carboxylic acid group
bMeans of three separate determinations agreeing to ꢁ .005
Figure 3. Hemi-acetal formation by addition of methanol to aldehydes
1–8 (X = CHO)
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
J. Phys. Org. Chem. 2013, 26 1048–1057
wileyonlinelibrary.com/journal/poc

G. DAW ET AL.
each contain a stereogenic centre, and the more complex methine
signals, shown in Fig. 5, are believed to reflect the introduction of
the new stereogenic centre in formation of hemi-acetals, but with
no indication of any strong diastereoselectivity.
Measurements of rates and equilibria
The NMR observations described above allow measurements
of equilibrium constants for hemi-acetal formation by signal
integration and, in principle at least, also rates of approach to
equilibrium. In practice, the timescales for equilibration in the
buffered solutions were inconveniently short for this method,
and we reverted to use of UV–vis spectroscopy. Solutions of all
these aldehydes in methanol show an n ! p* absorption band
with lmax = 287(ꢁ5) nm and emax ꢃ 55. Dimethyl acetals of satu-
rated aldehydes are transparent at these wavelengths, and we
assume that the hemi-acetals show the same behaviour in their
UV–vis spectra as the full acetals. Absorbances at lmax, after
rapid addition of the aldehydes to buffered methanol mixtures
and mixing, showed first-order decay to new equilibrium values,
yielding a rate constant, kobs, for the approach to equilibrium,
which is the sum of forward and reverse rate constants, (kf + kr),
where the constant methanol concentration is subsumed into
kf. Initial and final values of the absorbances, Ao and Aeq yield
values for the equilibrium constant with Ke = (Ao ꢀ Aeq)/Aeq
and the relationships between rates and equilibrium then allow
extraction of separate values for both kf and kr.
In view of the sensitivity of rates to solution acidity, we first
examined the behaviour of hexanal itself in a series of buffer mixtures
to identify conditions under which the mechanism of acetal
formation was likely to be the same for all isomers. Table 2 collects
rate constants for equilibrations of hexanal in methanol containing
four different triethylamine:acetic acid buffers and one
triethylamine:citric acid buffer. Plots of rate constant against total
buffer concentration are linear (Fig. 6A), allowing estimates of
buffer-independent contributions, ko, which are included in the table.
The buffer-independent rate constants may contain contribu-
tions from spontaneous reaction, ks, and from lyate and lyonium
catalysed reactions, i.e. Eqn 2.
Figure 4. The equilibration of hexanal (initially ca 0.06 M) and its hemi-
acetal in d4-methanol, observed by 1H-NMR spectroscopy over approxi-
mately 6 h
behaviour of hexanal itself, 1 (X = CHO) with the most obvious
indicators of reaction being the diminution of the signals from
the aldehydic hydrogen (at d9.70) and adjacent CH2 group
(at d2.47) and the matching appearance of a new triplet at higher
field. The times over which these changes occurred were variable
from about an hour to several days, with strong sensitivity to traces
of acid or base, even in carefully cleaned NMR tubes. This irrepro-
ducibility of time scale could be eliminated by buffering with
0.05 M Et3N:HOAc in 1:2 molar ratio, with times for equilibration
at 30 ꢂC then conveniently reduced, reproducibly, to a matter of
minutes. The changes in spectra associated with the aldehyde
chemistry were unaffected by presence of the buffer.
The appearance of the new triplet might be associated either
with hemi-acetal formation as shown in the figure, or with
formation of the full dimethyl acetal. This latter possibility
was tested by comparison with an authentic sample of 1,1-
1
dimethoxyhexane, whose H-NMR spectrum, separately and in
addition to the equilibrated mixture, presents a one-hydrogen
triplet well resolved, 0.2 ppm upfield from that observed in the
equilibration mixture, showing clearly that the product of equil-
ibration observed under these conditions is not the full acetal.
Experiments with the other isomers similarly indicated that the
products of equilibration with methanol under these very mild
conditions were not full acetals.
ko ¼ ks þ kMeO-½MeOꢀꢄ þ kMeOH2 ½MeOH2þꢄ
(2)
þ
Methanol is a well-characterised solvent,[8,9] with a reliable mea-
surement of its autoprotolysis constant (1.995 ꢅ 10ꢀ17 at 25 ꢂC).
Constants are also available for dissociation of acetic acid
(2.69 ꢅ 10ꢀ10 M), citric acid (7.41 ꢅ 10ꢀ9 M) and triethylammonium
ion (2.19 ꢅ 10ꢀ11 M) in methanol, so that concentrations of MeOꢀ
and MeOH+2 can be calculated for buffers used in this study, and
Table 2 includes calculated values of pH (=ꢀ log[MeOH+2]).
Figure 6B plots values of log ko against pH and shows increas-
ing rate as the buffers become more basic with a reasonable fit
to limiting lines of slope zero and one (dashed lines). The solid
curve in this plot is the best fit of Eqn 3 to the data (R = 0.978)
yielding values for ks = 1.84(ꢁ0.51)ꢅ10ꢀ3 sꢀ1 and kMeO- = 4.15
(ꢁ0.62)ꢅ104 Mꢀ1 sꢀ1 for the rate constants.
For equilibrations of aldehydes 1, 2, 5, 7 and 8 (X= CHO), the
new methine signals appeared at d4.373 (t, J = 5.5Hz), 4.351 (t,
J = 5.5Hz), 4.470 (t, 5.0Hz), 4.291 (d, J = 5.3 Hz) and 4.048 (s), respec-
tively. The hydrocarbon portions of aldehydes 3, 4 and 6 (X= CHO)
log ko ¼ log ðks þ kMeO-½MeOꢀꢄÞ
(3)
Values here are comparable with those obtained by Pocker
et al.[10] for additions of water to propanal, iso-butyraldehyde
and pivaldehyde (values for kHO
ꢀ
were, respectively,
Figure 5. Methine signals (400 MHz H-NMR spectra) from the hemi-
acetals formed from aldehydes 3, 4 and 6 (X = CHO)
2.35 ꢅ 103, 1.77 ꢅ 103 and 6.33 ꢅ 103 Mꢀ1 sꢀ1 at 0 ꢂC). Even with
© 2013 The Authors. Journal of Physical Organic Chemistry
wileyonlinelibrary.com/journal/poc
J. Phys. Org. Chem. 2013, 26 1048–1057
published by John Wiley & Sons Ltd.

STERIC EFFECTS IN THE FORMATION OF HEMI-ACETALS
Table 2. Rate constants for equilibrations of hexanal in methanolic buffers at 25 ꢂC
Buffer
Et3N: AcOH
Et3N: AcOH
Et3N: AcOH
Et3N: AcOH
Et3N:Citric
cratio
1:2
9.51
2:3
9.74
3:2
10.49
2:1
10.72
3.5:4.0
8.92
-log[MeOH+2]
a[Buffer]tot /M
bkobs /sꢀ1
0.075
0.050
0.025
d0.000
19.2 ꢅ 10ꢀ3
14.1 ꢅ 10ꢀ3
9.34 ꢅ 10ꢀ3
4.35 ꢅ 10ꢀ3
2.20 ꢅ 10ꢀ2
1.68 ꢅ 10ꢀ2
1.10 ꢅ 10ꢀ2
0.56 ꢅ 10ꢀ2
8.95 ꢅ 10ꢀ2
6.91 ꢅ 10ꢀ2
4.45 ꢅ 10ꢀ2
2.27 ꢅ 10ꢀ2
16.2 ꢅ 10ꢀ2
12.9 ꢅ 10ꢀ2
9.37 ꢅ 10ꢀ2
5.99 ꢅ 10ꢀ2
4.22 ꢅ 10ꢀ2
2.92 ꢅ 10ꢀ2
1.58 ꢅ 10ꢀ2
0.27 ꢅ 10ꢀ2
aConcentrations are sums of the separate base and acid concentrations.
bReactions were run in duplicate, and individual values of kobs were reproducible to ꢁ2%.
cMolar ratios of triethylamine and either acetic acid or citric acid.
dExtrapolated values (ko)
demonstrated absence of large electronic factors in the reactivity
with the set, we assume a similar pattern.
For each aldehyde, rates have been determined by UV–vis
spectroscopy at three different buffer concentrations. In all cases,
rates were linearly dependent on buffer concentration yielding
values for the rate constant for the buffer catalysed reaction,
kbuff, and extrapolation to zero buffer concentration to yielding
ko. Since equilibrium constants are also available for each
aldehyde, we have extracted values of the rate constants for
formation of hemi-acetal, kbuff(f) and ko(f) in this medium.
Table 3 collects those rate constants and the equilibrium
constants, Keq, for hemi-acetal formation, which were, as
expected, independent of buffer concentration. The span in
reactivities is 47 for the buffer-independent reaction (ko)
and 29 for the buffer-catalysed process (kbuff), both much
reduced from the ca 130-fold spread for the ester reactions.
The span of equilibrium constants is notably less again
(0.69 < Keq < 10.6), so that the sensitivity to substituents effects is
larger in the progress to transition than to product state.
Figure 6. Dependences of rates of hemi-acetal formation for hexanal in
methanol at 25 ꢂC with Et3N:HOAc buffers in 2:1 (□), 1.5:1(⋄) and 1:2(○)
molar ratios (plot A), and of log ko on pH (plot B)
the uncertainties, because of the balance of concentrations, it is
clear that in all the Et3N:AcOH buffers used, ko is dominated by
the methoxide-catalysed contribution (kMeO-[MeOꢀ]). In the case
of the 2:1 molar ratio Et3N:HOAc buffer (calcd. pH = 10.72), the
fraction is greater than 99.9%. We have not been able to carry
out similar full examinations of dependences on buffer composi-
tion on rates for the remaining aldehydes, but in view of the
Tests for internal free energy relationships are shown in Fig. 7A
and Fig. 7B, with the dashed line providing a reference of unit
slope. Figure 7A shows that there is a reasonable linear correlation
(R2 = 0.984, Eqn 4) between logarithms of rates, relative to hexanal,
Table 3. Rates and equilibrium constants for hemi-acetal formation from aliphatic aldehydes 1–8 (X = CHO) in methanol at 25 ꢂC
with 2:1 Et3N:HOAc buffer
Aldehyde 1 – 8 (X = CHO)
103 ko(f)a
101 kbuff (f)a
bKeq
/sꢀ1
/Mꢀ1 sꢀ1
Hexanal 1
57.4 (ꢁ1.7)
69.2 (ꢁ3.1)
48.5 (ꢁ2.9)
9.92 (ꢁ0.34)
56.3 (ꢁ4.8)
12.2.(ꢁ0.4)
5.77 (ꢁ3.6)
1.45 (ꢁ0.8)
20.6 (ꢁ0.5)
20.0 (ꢁ0.1)
13.3 (ꢁ0.1)
5.90 (ꢁ0.09)
16.3 (ꢁ1.3)
3.06 (ꢁ1.3)
2.11 (ꢁ.0.10)
0.69 (ꢁ0.02)
10.5
10.6
5.62
4.03
2.49
1.15
1.17
0.95
4-methylpentanal 2
3-methylpentanal 3
2-methylpentanal 4
3,3-dimethylbutanal 5
2,3-dimethylbutanal 6
2-ethylbutanal 7
2,2-dimethylbutanal 8
aUncertainties in rate constants are standard errors based on the linear regression analysis for three buffer concentrations.
Reactions were run in duplicate, and individual values of kobs were reproducible to ꢁ2%.
bUncertainties in the values of the equilibrium constants arise almost wholly from that in the initial absorbance (see experimental
section) and are estimated to be less than ꢁ10%.
© 2013 The Authors. Journal of Physical Organic Chemistry
J. Phys. Org. Chem. 2013, 26 1048–1057
wileyonlinelibrary.com/journal/poc
published by John Wiley & Sons Ltd.

G. DAW ET AL.
product states, and the temperature dependences of rates and
equilibria were examined for aldehydes 1, 5, 7 and 8 (X = CHO).
Because reactions of 1 and 5 in the 2:1 Et3N:HOAc buffer were
too fast for observation at T ≥ 30 ꢂC with available apparatus,
reactions were compared in the less basic 1:2 Et3N:HOAc buffer,
but one in which buffer-independent reactivity remains domi-
nated by the methoxide-catalysed pathway. Table 4 collects
the equilibrium constants, Ke, and first-order rate constants ko(f)
for the buffer-independent formation of the hemi-acetals over
20 < T < 40 ꢂC, and the values for the enthalpy, entropy and free
energy changes derived from the temperature dependences.
Figure 7. Internal correlations between rates and equilibria for formation of
hemi-acetals (H-As) from aldehydes 1–8 (X = CHO). Graph A plots logarithms
of relative rates for buffer-catalysed reaction against those for buffer-
independent reaction. Graph B plots logs of relative rates for the buffer-
independent reaction against the logarithms of the equilibrium constants
For the variation of equilibrium constants ([Hemi-acetal]eq
/
[Aldehyde]eq), Van’t Hoff treatment yields enthalpy changes
between ꢀ33.5 < ΔH < ꢀ31.4kJmolꢀ1, rather less than the value
calculated for the reaction (ꢀ41.4 kJ molꢀ1) in the gas phase using
Benson’s[11] group increments for conversion of the double bond
of a general aliphatic aldehydic carbonyl group to a hemi-acetal.
The entropy changes are large and negative (ꢀ92.6 > ΔS >
of the buffer-catalysed (kbuff rel) and buffer-independent (ko rel
)
reactions, with the buffer-catalysed reaction being less sensitive
to the effects of substituent variation.
110.0 J.Kꢀ1 molꢀ1) and comparable with that (ꢀ76 J.Kꢀ1 molꢀ1
)
found by Pocker[10] for hydration of propanal. Entropic contribu-
tions (27.6 > ꢀTΔS > 32.8 kJ molꢀ1) to the free energy change
almost cancel the enthalpic change. If the concentration of
methanol (for liquid methanol, [MeOH] = 24.59 M at 25 ꢂC)
is taken into account in the equilibrium expression (i.e. Ke =
[Hemi-acetal]eq/[MeOH][Aldehyde]eq), all entropies become
more negative by 23.4J.Kꢀ1 molꢀ1 so that the range is (ꢀ126.0>
logkbuff rel ¼ 0:726 ðꢁ0:040Þ log korel ꢀ 0:037ðꢁ0:036Þ (4)
Figure 7B, in contrast, shows that the correlation between
logarithms of rate and equilibrium constants is poor.
For external correlation with steric parameters, we have used
logarithms of relative rates of saponification of the correspond-
ing methyl esters as our reference points. Figure 8A shows a
reasonable linear relationship between logarithms of equilibrium
constants for formation of hemi-acetals and logarithms of rates
constants for ester saponification (see Eqn 5, R2 = 0.986)
ΔS > ꢀ134.4 J.Kꢀ1 molꢀ1). Since liquid methanol[12] has
S
(298)= 128 J.K. molꢀ1 of which the major contribution is transla-
tional,[13] the entropy changes found for the equilibrium processes
correspond closely with those expected from the reaction
stoichiometry.
Eyring treatment of the rate constants (pseudo-first order, ko
(f)) yields entropies of activation which are significantly more
negative (ꢀ160.8 > ΔS{ > ꢀ211.0 J K molꢀ1) than those for the
equilibration by factors between 1.62 and 1.92, so that entropic
contributions (47.9 > ꢀTΔS > 62.9 kJ molꢀ1) dominate the free
energies of activation.
logKrel ¼ 0:520 ðꢁ0:025Þ logkrelðsapnÞ þ 0:010 ðꢁ0:033Þ (5)
Figure 8B plots logarithms of relative rate constants for the
buffer-independent formation of the hemi-acetals against the
same reference points, and it is clear that no simple relationship
exists between reactivities in the reactions. Use of the more
traditional Es-values (based on relative rates of acid-catalysed
esterification) does not alter the picture.
DISCUSSION
For these reactions, also, the availability of both rates and
equilibria offers the possibility of an informative comparison of
enthalpy and entropy changes in progress to transition and
The esterification and hemi-acetal reactions are compared in
Fig. 9. Saponification of the methyl esters is first order in both
ester and hydroxide, reflecting their union in an activated
complex often modelled by the tetrahedral intermediate arising
from full addition of hydroxide to the carbonyl. Hemi-acetal
formation also involves quaternization of a carbonyl group, but
differs in replacement of the ester alkoxy residue by a smaller
hydrogen atom (Van der Waals radii for O and for H are 1.72 Å
and 1.44 Å, respectively[14]), with corresponding reduced steric
demand. The proportionality and the reduced slope of the corre-
lation line, seen in Fig. 8A and Eqn 5, seem entirely consistent
with the original Taft proposition that the steric effects in the
transition states of the ester reactions are well modelled by those
in the familiar tetrahedral intermediates.
The absence of correlation between rates and equilibria in
hemi-acetal formation (Fig. 7B) and the similar absence of
correlation between rates of hemi-acetal formation and rates of
ester saponification (Fig. 8B) signals that the changes in steric
interaction in progress to transition structures for hemi-acetal
formation are not modelled by those associated with full
quaternization of the carbonyl group. This work has shown already
Figure 8. Plot of log Krel for hemi-acetal formation against log krel for
saponification of methyl esters (graph A) and plot of logs of relative rate
constants of hemi-acetal formation (log krel) against logs of relative rates
of methyl ester saponification (graph B). Saponification rates are taken
from reference 1
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
wileyonlinelibrary.com/journal/poc
J. Phys. Org. Chem. 2013, 26 1048–1057

STERIC EFFECTS IN THE FORMATION OF HEMI-ACETALS
Table 4. Temperature dependences of rate and equilibrium constants for hemi-acetal formation for aldehydes 1, 5, 7 and
8 (C = CHO)
T /ꢂC
ako(f) /sꢀ1
aKeq
Equilibrium and activation parameters
Hexanal 1 (X = CHO)
20.0
25.0
30.0
40.0
2.44 ꢅ 10ꢀ3
3.94 ꢅ 10ꢀ3
4.39 ꢅ 10ꢀ3
7.72 ꢅ 10ꢀ3
13.6
10.4
8.60
5.59
ΔHr = ꢀ33.5 (ꢁ1.1)kJ molꢀ1
ΔSr = ꢀ92.6 (ꢁ5.5) J Kꢀ1 molꢀ1
ΔGr(298) = ꢀ5.9 (ꢁ2.1) kJ molꢀ1
ΔH{ = 39.1 (ꢁ4.5) kJ molꢀ1
ΔS{ =ꢀ160.8 (ꢁ18.1) J Kꢀ1 molꢀ1
ΔG{(298) = 87.0 (ꢁ10.9) kJ molꢀ1
3,3-dimethylbutanal 5(X = CHO)
20.0
25.0
30.0
40.0
1.57 ꢅ 10ꢀ3
2.15 ꢅ 10ꢀ3
2.82 ꢅ 10ꢀ3
4.49 ꢅ 10ꢀ3
3.23
2.49
2.08
1.39
ΔHr = ꢀ32.0 (ꢁ0.4) kJ molꢀ1
ΔGr(298) = ꢀ2.4 (ꢁ 0.9) kJ molꢀ1
ΔH{ = 37.2 (ꢁ 1.5) kJ molꢀ1
ΔSr = ꢀ99.5 (ꢁ1.4) J Kꢀ1 molꢀ1
ΔS{ = ꢀ171.3 (ꢁ 4.9)J Kꢀ1 molꢀ1
ΔG{(298) = 88.3 (ꢁ 2.9) kJ molꢀ1
2-ethylbutanal 7(X = CHO)
20.0
25.0
30.0
40.0
1.62 ꢅ 10ꢀ4
1.46
1.17
0.96
0.64
ΔHr = ꢀ31.4 (ꢁ0.3) kJ molꢀ1
ΔGr(298) = ꢀ0.4 (ꢁ0.6) kJ molꢀ1
ΔH{ =43.7 (ꢁ2.4) kJ molꢀ1
ΔSr = ꢀ103.9 (ꢁ1.0) J K molꢀ1
2.42 ꢅ 10ꢀ4
3.04 ꢅ 10ꢀ4
5.88 ꢅ 10ꢀ4
ΔS{ = ꢀ168.2 (ꢁ7.9) J Kꢀ1 molꢀ1
ΔG{(298) = 93.8 (ꢁ4.8) kJ molꢀ1
2,2-dimethylbutanal 8(X = CHO)
20.0
25.0
30.0
40.0
5.37 ꢅ 10ꢀ5
5.85 ꢅ 10ꢀ5
8.77 ꢅ 10ꢀ5
1.33 ꢅ 10ꢀ4
1.16
0.95
0.75
0.50
ΔHr = ꢀ32.6 (ꢁ1.7) kJ molꢀ1
ΔSr = ꢀ110.0 (ꢁ5.5) J Kꢀ1 molꢀ1
ΔGr(298) = 0.2 (ꢁ3.3) kJ molꢀ1
ΔH{ = 34.0 (ꢁ4.5) kJ molꢀ1
ΔS{ = ꢀ211.0 (ꢁ14.0) J Kꢀ1 molꢀ1
ΔG{(298) = 96.6 (ꢁ8.9) kJ molꢀ1
aFor uncertainties, see footnotes to Table 3.
Figure 9. Relationships between rates of saponification of esters and
equilibria in hemi-acetal formation
that energy changes in hemi-acetal formations for this set of alde-
hydes are small (based on Ke = [Hemi-acetal]eq/[MeOH][Aldehyde]
eq, 4.1 < ΔG298 < 8.7 kJ molꢀ1) compared to those for tetrahedral
intermediate formation for esters, which should be close to activation
energies for saponification of methyl esters of aliphatic esters,[15–17]
typically with ΔG{ > 85 kJ molꢀ1. Bema–Hapothle[18] considerations
would then suggest that the transition structures for the addition to
aldehyde should be more reactant-like with relatively small develop-
ment of bonding between the incoming nucleophile and the
carbonyl carbon, and correspondingly small displacements of the
central carbon from the plane of its original three ligands. Our mea-
surements (graphed in Fig. 7B and Fig. 8B) show that aldehydes, 1,
2, 3 and 5 react at closely similar rates. These isomers are all are
mono-substituted at the a-carbon, i.e. C4H9–CH2–CHO. The remaining
four isomers, all branched at the a-position, are significantly less
reactive, and we suggest that this pattern is consistent with reactivity
reflecting steric interaction of the incoming nucleophile with the
ground state conformations accessible to the aldehydes.
Figure 10. Conformations of simple aliphatic aldehydes and vectors for
approach of nucleophile to carbonyl carbon
for rotation about the bond between carbonyl carbon and
a-carbon are less than 10 kJ molꢀ1 and energy differences
are less than 5 kJ molꢀ1, so that all conformations should be
populated at room temperature.
For aldehydes with a single a-alkyl substituent (other than t-
butyl), the alkyl group eclipses the oxygen in lowest energy
conformation (A in Fig. 10) with the other rotamers (B and C in
Fig. 10) being slightly higher in energy. The substituent eclipsing
the > C = O bond seems unlikely to have major steric interaction
with a nucleophile approaching the carbonyl group along the
Burgi–Dunitz angle,[22,23] but the others certainly do, with the
steric demand of an alkyl group being the larger. When both
are hydrogen as in conformation A, the interaction is minimal
so that approach may occur equally from either face of the
carbonyl group, while in conformations B and C an alkyl group
closes one face. All conformations thus possess at least one open
face. For aldehydes with two alkyl substituents at the a-position,
steric interactions are unfavourable for the conformation in
NMR observations,[19,20] supported by computational studies,[21]
on simple aliphatic aldehydes suggest that conformations
predominate in which the carbonyl oxygen is eclipsed either by
hydrogen or alkyl at the a-position as shown in Fig. 10. Barriers
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
J. Phys. Org. Chem. 2013, 26 1048–1057
wileyonlinelibrary.com/journal/poc

G. DAW ET AL.
which the remaining a-hydrogen eclipses the > C = O bond
(conformation F) and one face only is open in the remaining
two (conformations D and E). If the relative reactivities of open
and closed faces, as defined above, are similar in all isomers,
then while total reactivities must reflect conformational popula-
tion distributions,[24] aldehydes with branching at the a-position
can never be more reactive than isomers with a single a-alkyl
group. Isomer 8 (X = CHO), with full substitution at the a-
carbon, may be taken to give an estimate of the reactivity of a
closed face and comparing this with the reactivities of 1, 2, 3
and 5 (X = CHO) provides a very crude estimate that open faces
are between 30- and 90-fold more reactive than closed faces.
Our interpretation thus requires that the transition states in
these additions are ‘early’, with relatively little perturbation of
geometry and bonding in the transition structure. We believe
the picture presented is internally consistent but must draw
attention to the measurements of Hill and Milosevich[25] of
secondary a-deuterium isotope effects on hydrate and ethyl
hemi-acetal formation from pentanal, C4H9CO.L (L = H or D).
The equilibrium a-deuterium isotope effects reported for both
hydrate and ethyl hemi-acetal formation are large and inverse,
KH/KD = 0.72 and 0.73, respectively, and consistent with the
change of the attachment of the isotopically labelled hydron
from trigonal to tetrahedral carbon. The secondary kinetic
a-deuterium isotope effect for formation of the ethyl
hemi-acetal, for the buffer-catalysed reaction using a acidic
buffer (1:1 HOAc;AcO–) is similar, kH/kD = 0.74, suggesting
that force constants changes in product and transition state
formations are alike, i.e. that the transition state is late.[26] The
discrepancy between the indications from variation of alkyl substit-
uents in our work and the isotopic substitution reported by Hill and
Milosevic is challenging, and at this stage, we can only note the
different reaction conditions and suggest that there are important
differences in transition structures for acid- and base-catalysed
reactions of oxygen nucleophiles to aldehydes.
tion and so is difficult to reconcile with the very large negative
entropies of activation we have found in our kinetic measurements
for hemi-acetal formation. We suggest that these require, not only
that the aldehyde carbonyl groups are sufficiently reactive to
engage a fully solvated methoxide, but that achievement of
transition state calls for involvement of an additional molecule of
methanol. The exact organisation of such an assembly must be
speculative, but it seems reasonable that the additional molecule
of methanol hydrogen bonds to the developing negative charge
on the carbonyl oxygen. A possible arrangement is shown below
(Fig. 11), in which the methoxide functions as a general base,
catalysing addition of one of its solvating methanol molecules,
with the additional molecule of methanol placed to minimise
nuclear motion in the eventual collapse to neutral product hemi-
acetal (pKa ꢃ 13.3)[36] with regeneration of a fully solvated
methoxide anion. Similar cyclic arrays have already been consid-
ered for hydrations of aliphatic aldehydes.[10,37,38]
To conclude this discussion, we make comparison between the
kinetics of formation of the hemi-acetals from aldehydes and of hy-
drolyses of the corresponding set of acid chlorides 1 – 8 (X= COCl).
In our earlier work,[1] we found that the logarithms of rates of
hydrolyses of the acid chlorides did not correlate with logs of rates
of methyl ester formation, or of ester saponification. The behaviour
is shown in Fig. 12A, and we took that as evidence against mecha-
nisms for acid chloride involving rate-limiting formation of a tetra-
hedral intermediate by addition of water to the acid chloride
carbonyl, but did suggest that the reactivity of aldehydic carbonyls
might offer a better model for the behaviour of acid chlorides than
that of esters. In the event, Fig. 12A, bears a remarkable qualitative
resemblance to Fig. 8B, with both showing the close grouping in
reactivity of isomers with no a-branching.
Figure 12B now plots logs of rates of acid chloride hydrolyses
against those for hemi-acetal formation and, despite the relative
small span of reactivities in both cases (27-fold for the acid
We have not yet considered the detailed nature of the nucleo-
phile for the methoxide-catalysed additions, but this must be
consistent with the very large negative entropies of activation
found. Isotopic fraction factors for methoxide in methanol have
been interpreted,[27–29] in terms of the existence of a tri-solvate,
with the anion strongly hydrogen bonded to three molecules of
methanol. For reactions in which methoxide behaves as a base
towards carbon acids,[30–32] this fully solvated anion is apparently
not the active species. Rather, desolvation, releasing solvating
methanol molecules into bulk solvent to provide more reactive
forms, precedes proton abstraction. Solvent isotope effects for
methanolysis of phenyl carbonates[33] and of phenyl esters[34,35]
have similarly been interpreted in terms of release of at least one
of the solvating methanols prior to the nucleophilic attack at
carbonyl carbon. We have not been able to measure solvent
isotope effects in this work, but suggest that release of solvating
methanol should increase rather than decrease entropy of activa-
Figure 12. Plot A, logs of relative rates for hydrolysis of acid chlorides
against logs of relative rates for ester saponification. Plot B, of plot of logs
of relative rates for hydrolysis of acid chlorides against logs of relative
rates hemi-acetal formation
Figure 11. A possible mechanism and transition structure to account for the large negative entropies of activation
© 2013 The Authors. Journal of Physical Organic Chemistry
wileyonlinelibrary.com/journal/poc
J. Phys. Org. Chem. 2013, 26 1048–1057
published by John Wiley & Sons Ltd.

STERIC EFFECTS IN THE FORMATION OF HEMI-ACETALS
6, 7 (X = COOH) and m = 0.029 M for 8 (X = COOH). Tabulated values are
the means of three titrations agreeing to ꢁ0.005.
chlorides and 47-fold for the aldehydes), a reasonable correlation
exists (Eqn 6, R2 = 0.951).
Measurements of partition coefficients (Log POW
)
logkrelðRCOClÞ ¼ 0:670ðꢁ0:063Þ logkrelðH-AÞ
þ0:054ðꢁ0:057Þ
(6)
A sample of the carboxylic acid (ca 0.025 g) was weighed out (ꢁ0.0001 g)
into a screw-capped polypropylene centrifuge tube. Octanol (1 ml) and
0.001 M aqueous HCl solution (5 ml) were then added and the tube
shaken vigorously for 10 min forming a cloudy suspension. The tube
was then centrifuged at 3600 rpm for 30 minutes at 20 ꢂC resulting in
complete separation of the two layers, and samples were taken from
the water layer using a syringe. Air was blown out from the air-filled
syringe while the needle passing through the organic phase and only
when in the water layer was a sample of the lower aqueous layer taken
in to the syringe. Some of this was expelled as the needle was withdrawn
through the organic layer, and the needle was wiped dry thoroughly
before transferring a sample of the aqueous layer to a GC vial. This
complete aqueous layer was analysed by GLC on a 30 m ꢅ 0.25 mm poly-
dimethylsiloxane capillary column (Restek RTX-1). The method used held
the oven at 50 ꢂC for 5 min before a ramp of 3 ꢂC a minute until 95 ꢂC and
then a second ramp of 20 ꢂC per minute, until 320 ꢂC where it was held
for 2 min. Peaks were identified for the carboxylic acid, and octanol,
and the ratio of areas of the octanol (solubility in water at 20 ꢂC,
0.042 g in 100 g),[39] and carboxylic acid peaks used to compare relative
amounts carboxylic acid in the aqueous layer, assuming that the
response factors of the flame ionisation detector for octanol and the acid
were in the ratio of 8 to 6 (i.e. the carbon numbers in the molecular
formulae[40]). Each measurement was run in triplicate. This procedure
yielded log Pow = 1.98 for hexanoic acid, slightly larger than the
published value (log Pow = 1.92).[41]
On this basis it is clear that, provided transition states are early,
mechanisms involving rate-limiting formation of tetrahedral
intermediates, or activated complexes, are not excluded by the
behaviour of these aliphatic acid chlorides.
CONCLUSION
We have found, like others before us, that structure–reactivity
correlations using substituent parameters intended to quantify
their steric interaction with the reacting site can be very poor,
even when the functionality and the bonding changes are
closely related to those in the reference reaction. Such failures
have often been ascribed to variable and unquantified electronic
contributions to the substituent parameters, but this work has
confirmed that poor correlations may persist even when the
substituent set is chosen very carefully to ensure electronic homo-
geneity. We have suggested here that poor correlations then signal
markedly different extents of bonding change at the transition
states of the reference reaction and the reaction under examina-
tion. The corollary is, that while substituent steric parameters
should be used with caution as predictors of reactivity, those for
a substituent set such as the isomeric saturated C5H11 hydrocarbon
residues have value as a probe for variation of extent of bonding
changes in the transition state in similar reactions.
Esterification of the acids and reduction of methyl esters
(representative procedure)
A round-bottom flask was charged with 2,2-dimethylbutyric acid (60.0 g,
520 mmol) which was warmed to 40 ꢂC. Thionyl chloride (75.0 ml,
1040 mmol) was added dropwise, and the resultant solution was heated
under reflux for 30 min. The acid chloride product was distilled from the
reaction mixture using a Vigreux column, and the residue added
dropwise to an ice-cooled solution of methanol (30 ml, 780 mmol) in
CH2Cl2 (50 ml). This mixture was then left to stir overnight at room
temperature, before being washed with saturated sodium bicarbonate
solution (50 ml) and distilled water (50 ml). The organic phase was
separated, then dried over magnesium sulfate before distillation at atmo-
spheric pressure yielded methyl 2,2-dimethylbutanoate as a colourless
liquid (40.0 g., 60 %, bp 124–128 ꢂC).
Methyl 2,2-dimethylbutanoate (10.0 g, 76.9mmol) was added dropwise
down the condenser of a reflux apparatus to a stirred suspension of lithium
aluminium hydride (1.75 g, 46.0mmol) in diethyl ether (40ml). The resultant
mixture was then stirred at room temperature for 1 h before careful addi-
tion of a saturated solution of aqueous sodium sulphate to form a white
granular precipitate from which the organic layer could be decanted. The
organic layer, and further ethereal washings were combined, and the ether
removed by distillation at atmospheric pressure through a Vigreux column.
The residue was then finally subjected to bulb-to-bulb distillation to afford
2,2-dimethylbutan-1-ol (6.0g, 76 %, bp 132–135ꢂC) as a colourless oil.
EXPERIMENTAL SECTION
The compounds
The carboxylic acids 1 – 8 (X= COOH) were available from our earlier stud-
ies.1 Hexanal, 2-ethylbutanal, 2-methylpentanal and 3,3-dimethylbutanal
were purchased from Sigma-Aldrich. All other aldehydes were prepared
by oxidation of the corresponding primary alcohol, using an adaptation
of the method of Anelli et al.,[7] and one example is detailed below.
Of the precursor primary alcohols, hexanol, 4-methylpentanol, 2-
methylpentanol and 3,3-dimethylbutanol were commercially available. All
others were prepared by reduction of the methyl esters of the correspond-
ing carboxylic acids with lithium aluminium hydride. Exemplary methods
are described below. All aldehydes were distilled shortly before use in
kinetic or equilibrium measurements and shown to be >99.95% by GLC
analysis on a 30 m ꢅ 0.25 mm polydimethylsiloxane capillary column (Restek
RTX-1). 1H- and 13C-NMR spectra showed no extraneous peaks (see below)
and were consistent with structures and literature data where available.
Measurements of acid dissociation constants (pKa)
A 0.05 M aqueous sample solution of each carboxylic acid was prepared,
except in the case of 2,2-dimethylbutanoic acid 8 (X = COOH), which was
only 0.03 M, being restricted by its lower aqueous solubility. Each carbox-
ylic acid (15 ml of solution was then titrated, at 20 ꢂC against aqueous
NaOH solution (0.750 M), using a Radiometer Copenhagen TitraLab VIT
90 video titrator (ABU91 auto burette and a SAM90 sample station)
adding 0.3 ml aliquots and allowing each 20 s between each addition
before recording solution pH, using a glass electrode calibrated against
standard buffers at pH 4 and pH 10 (Hydrion buffer capsules). Non-
linear regression of calculated pH to the experimental values using pKa
as variable yielded values for pKas at m = 0.047 M for acids 1, 2, 3, 4, 5,
Oxidation of primary alcohols (representative procedure)
A round-bottom flask was charged with 2,2,6,6-tetramethylpiperidin-1-
oxyl (TEMPO) (0.08 g 0.50 mmol), 4-methylpentanol (5.11 g, 50 mmol)
and CH2Cl2 (20 ml). Potassium bromide (1.20 g, 10.1 mmol in 2.5 ml water)
was added to this solution, and the resulting mixture was stirred
vigorously while being cooled to ꢀ10 ꢂC in a NaCl-ice bath. A mixture
of aqueous sodium hypochlorite solution (66 ml of 1.5 M, 0.10 mmol)
and sodium bicarbonate (0.91 g, 10.8 mmol) was added over a period
of 20–30 min, keeping the temperature below 15 ꢂC. The mixture was left
to stir for a further 5 min at room temperature before the organic phase
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
J. Phys. Org. Chem. 2013, 26 1048–1057
wileyonlinelibrary.com/journal/poc

G. DAW ET AL.
was separated and diluted in pentane (20 ml). This solution was then
washed with 1 M HCl (10 ml) in KI (0.16 g, 0.96 mmol), 10 % aqueous
sodium thiosulphate solution (20 ml), sodium bicarbonate solution
(10 ml) and water (10 ml). The organic phase was dried over sodium
sulfate and distilled at atmospheric pressure first through a Vigreux col-
umn, to remove the majority of the solvent, and finally by bulb-to-bulb
distillation to afford 4-methylpentanal (1.50 g, 30%) as a colourless liquid.
diluted in methanol to produce two new acetic acid solutions of 0.0333M
and 0.0166 M giving a total of three different concentrations of buffer.
In a typical kinetic experiment, aliquots of methanolic buffer (3 ml)
were pipetted into stoppered quartz 1 cm UV cells, which were placed
in the thermostatted block of the UV spectrometer and allowed to equil-
ibrate to the set temperature for at least 20 min. Spectra were recorded
on a Varian Cary Bio-50 single beam spectrometer using a multiblock
whose temperature was controlled by a Varian Cary PCB 150 water
Peltier system. Initially, the spectrum of the appropriate hexanal isomer
was recorded by scanning between 400 and 200 nm to identify the
maximum of the n-p* carbonyl excitation (lmax = 287 nm). Reaction
kinetics were then monitored by adding 10 ml of the appropriate
hexanal to a fresh buffer solution, with agitation by a Teflon paddle
(to make them ca 0.03 M in aldehyde) while monitoring absorbance
at 287 nm. Mixing was complete in less than 2 s, and initial absorbances,
ca 1.3, were obtained by extrapolating curves back to initial addition.
The changes in absorbance were first order and monitored for at least
five half-lives. Rate constants were extracted by non-linear least squares
fitting of Acalc = Ainf – ΔA e –kt to the data. A plot of the rate constant at
different concentrations of buffer against concentration of acetic acid
produced a straight line, which could be extrapolated back to deter-
mine the rate constant at zero buffer concentration (see main text)
Experiments were repeated with a second buffer mixture containing
initially 0.045 M Et3N and 0.03 M in HOAc.
Hexanal[42] dH (300 MHz, CDCl3) 9.78 (1H, t, J = 1.8 Hz), 2.43 (2H, dt,
J = 7.2 and 1.8 Hz), 1.65 (2H, q, J = 7.2 Hz), 1.36–1.29 (4H, m) and 0.91
(3H, t, J = 6.9 Hz); dC (300 MHz, CDCl3) 202.9, 43.9, 31.3, 22.4, 21.8, 13.8.
4-Methylpentanal[43] dH (300 MHz, CDCl3) 9.79 (1H, t, J = 2.1 Hz),
2.48–2.42 (2H, m), 1.68–1.51 (3H, m), 0.93 (6H, d, J = 6.3 Hz); dC
(300 MHz, CDCl3) 203.0, 42.0, 30.8, 27.7, 22.3.
3-Methylpentanal[44] dH (500 MHz, CDCl3) 9.70 (1H, t, J = 2.5 Hz), 2.34
(1H, ddd, J = 16.0, 5.5 and 2.0 Hz), 2.16 (1H, ddd, J = 16.0, 8 and 2.5 Hz),
1.92 (1H, oct, J = 7.0 Hz), 1.35–1.27 (1H, m), 1.23–1.17 (1H, m), 0.89 (3H,
d, J = 6.5 Hz) and 0.84 (3H, t, J = 7.5 Hz); dC (300 MHz, CDCl3) 203.2, 50.7,
29.8, 29.5, 19.5, 11.3.
2-Methylpentanal[45] dH (300 MHz, CDCl3) 9.64 (1H, d, J = 2.1), 2.37
(1H, dhx, J = 6.9 and 2.1 Hz), 1.76–1.63 (1H, m), 1.47–1.27 (3H, m), 1.11
(3H, d, J = 6.9 Hz), 0.95 (3H, t, J = 6.9 Hz); dC (400 MHz, CDCl3) 205.5, 46.1,
31.6, 20.1, 13.9, 13.3.
Acknowledgements
This work was supported by a grant from EPSRC (UK). We thank
Dr Paul Christian (University of Manchester) for the use of his
centrifuge. This work arose from enjoyable and useful conversa-
tions with Professor Peter Guthrie (University of Western Ontario),
and we thank him and Dr T.W. Bentley (University of Swansea)
for prepublication comment on the MS.
3,3-Dimethylbutanal[46] dH (300 MHz, CDCl3) 9.86 (1H, d, J = 3.0 Hz), 2.28
(2H, d, J = 3.0 Hz), 1.09 (9H, s); dC (300 MHz, CDCl3) 203.6, 56.6, 30.9, 29.8.
2,3-Dimethylbutanal[47] dH (200 MHz, CDCl3) 9.67 (1H, d, J = 3.0 Hz),
2.30–2.00 (2H, m) 1.04 (3H, d, J = 10.2 Hz), 1.00 (3H, d, J = 10.2 Hz), 0.92
(3H, d, J = 10.2 Hz); dC (400 MHz, CDCl3) 205.8, 52.4, 28.4, 20.6, 18.7, 9.5.
2-Ethylbutanal[48] dH (300 MHz, CDCl3) 9.59 (1H, d, J = 3 Hz), 2.18–2.08
(1H, m), 1.75–1.46 (4H, m), 0.93 (6H, t, J = 7.5 Hz); dC (300 MHz, CDCl3)
205.6, 54.9, 21.4, 11.4.
REFERENCES
[1] A. C. Regan, C. I. F. Watt, J Phys Org Chem 2007, 20(3), 195–205.
[2] M. Charton, B. Charton, J Org Chem 1979, 44, 2284.
[3] R. W. Taft Jr., J Am Chem Soc 1952, 74, 2729, and 1953; 75, 4231.
[4] R. W. Taft Jr., J Am Chem Soc 1953, 75, 4231.
[5] C. K. Hancock, P. C Falls, J Am Chem Soc 1961, 83, 4214–6.
[6] A. Williams, Free Energy Relationships in Organic and Bio-Organic
Chemistry, Royal Society of Chemistry, Cambridge UK, 2003.
[7] P. L. Anelli, F. Montanari, F. S. Quici, Org Syn CV 8, 1993, 367.
[8] F. Rived, I. Canals, E. Bosch, M. Roses, Analytical Chemistry Acta 2001,
439, 315–333.
[9] In: Tables of Rate and Equilibrium Constant for Heterolytic Organic
Reactions (Ed: V. Palm), VINITI, Moscow, 1975–85.
[10] Y. G. Pocker, D. G. Dickerson, J Phys Chem 1969, 73, 4005.
[11] S. W. Benson, F. R. Criuckshank, D. M. Golden, G. R. Hagen, H. E.
O’Neal, R. Shaw, R. Walsh, Chem Rev 1969, 69, 279.
2,2-Dimethylbutanal[49] dH (300 MHz, CDCl3) 9.47 (1H, s), 1.54 (2H, q,
J = 7.5 Hz), 1.06 (6H, s), 0.86 (3H, t, J = 7.5 Hz); dC (300 MHz, CDCl3) 206.7,
46.0, 29.8, 20.8, 8.6.
Behaviour of solutions observed by NMR
A stock buffer solution of deuterated methanol containing 0.050 M deuter-
ated acetic acid and 0.025 M triethylamine was initially prepared. Eight vials
were prepared each containing a hexanal (10 ml) isomer and 1 ml of the stock
solution. The samples were left overnight before being transferred to a
Bruker 400 MHz NMR spectrometer and allowed to equilibrate at 25 ꢂC for
at least 20 min. A 1H NMR spectrum of each isomer was then acquired,
recording accurate integration values for the aldehyde (ꢃ 9.5 ppm) and
hemi-acetal (4.5–4 ppm) resonances, and the equilibrium constant,
K= [Hemi-acetal]eq/[Aldehyde]eq was evaluated as the ratio of integrals.
[12] J. R. S Machado, W. B. Street, J Chem Eng Data 1983, 28, 218.
[13] If the liquid is treated as a dense gas, the Sackur –Tetrode equation
gives Strans = 98.7 J.Kꢀ1 molꢀ1
.
[14] From W. D. Cornall P. Cieplak, C. I. Bayley, I. R. Gould, K. M. Merz,
D. M. Ferguson, D. C. Spelmayer, T. Fox, J. W. Caldwell, P. A. Kolman,
J Am Chem Soc 1995, 117, 5170.
[15] For collected kinetic data on ester saponifications see, A.J. Kirby, in;
Compr. Chem. Kinet (Eds: C. H. Bamford, C. F. H. Tipper), Elsevier,
Amsterdam, 1972, 10, 57–207, Ch. 2 Hydrolysis and formation of
esters of organic acids.
Kinetics of reactions of aldehydes with methanol
[16] H. A. Smith, C. H. Reichardt, J Am Chem Soc 1941, 63, 605.
[17] H. A. Smith, J Am Chem Soc 1940, 62, 1176.
[18] W. P. Jencks, Chem Revs 1985, 85, 511, and references therein.
[19] R. J. Abrahams, J. A. Pople, Mol Phys 1960, 3, 609.
[20] G. J. Karabatosos, N. His, J Am Chem Soc 1956, 87, 2864–70.
[21] C. H. Langley, J.-H. Lii, N. L. Allinger, J Comput Chem 2001, 22,
1396–1425.
Methanol was taken from a 2.5 l bottle of anhydrous methanol (Aldrich
0.005% water), which was freshly opened and then stored under nitro-
gen atmosphere. Triethylamine was distilled from KOH pellets through
a 20 cm Vigreux column and stored under nitrogen. Acetic acid was
redistilled through a 20 cm Vigreux column before taking a centre cut
which was stored under nitrogen.
A buffer solution of methanol containing 0.050 M acetic acid and 0.025 M
triethylamine was initially prepared. This stock solution was then further
[22] H.-B. Burgi, J. Dunitz, E. Shefter, J Am Chem Soc 1973, 95, 5065–7.
[23] H.-B. Burgi, J. D. Dunitz, Struct Correl 1994, 1, 163–204.
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
wileyonlinelibrary.com/journal/poc
J. Phys. Org. Chem. 2013, 26 1048–1057

STERIC EFFECTS IN THE FORMATION OF HEMI-ACETALS
[24] Effectiively Winstein-Holness kinetics, see J. I. Seeman, W. A. Farone,
J Org Chem 1978, 43, 1854.
[38] M. Eigen, Discussion Faraday Soc 1965, 39, 7.
[39] M. Janado, Y. Yano, Y. Doi, H. Sakamoto, J Soln Chem 1983, 12, 741–54.
[40] M. Morvai, I. Palyka, I. Molnar-Perl, J Chromatographic Science 1992,
30(11), 448–52.
[41] C. J. Smith, T. A. Perfette, R. Garg, C. Hansch, Food Chem Toxicology
2004, 42, 9–15.
[25] E. A. Hill, S. A. Milosevich Tetrahedron Lett 1976, 50, 4553–5.
[26] A similar resuly is reported for the hydration of p-nitrobenzaldehyde.
J. R. Pires, A. S. Stachissini, L. do Amaral, J Phys Org Chem 1994, 7, 192–5.
[27] R. A. More O’Ferrall, J Chem Soc, Chem Commun 1969, 114–5.
[28] J. M. A. Al-Rawi, J. P. Bloxsidege, J. A. Elvidge, J. R. Jones, R. A. More
O’Ferrall, J Chem Soc, Perkin Trans 2 1979, 1593–9.
[42] T. Tsuda,T. Hayashi, H. Satomi,T. Kawamoto , T. Saegusa, J Org Chem
1986, 51, 537–540.
[29] V. Gold, S. Grist, J Chem Soc, B 1971, 11, 1665–70.
[30] V. Gold, S. Grist, J Chem Soc, B 1971, 11, 2282–5.
[31] L. Baltzer, N. A. Bergmann, J Chem Soc, Perkin Trans 2 1982,
313–19.
[32] L. Baltzer, N. A. Bergmann, Acta Chem Scand 1982, A36, 39–45.
[33] J. Stratton, B. Gatlin, K. S. Venkatsubban, J Org Chem 1992, 57, 3237–40.
[34] W. P. Huskey, R. L. Schowen, Gazz Chem Ital 1987, 117, 409–16.
[35] C. D. Bryan, K. B. Schowen, R. Schowen, Can J Chem 1996, 74, 931.
[36] R. P. Bell, D. P. Onswood, Trans Faraday Soc 1962, 58, 1557.
[37] R. P. Bell, P. G. Evans, Proc Roy Soc 1966, A291, 297.
[43] R. C. Ebersole, W. O. Godtrfredsen, S. Vangedal, E. Caspi, J Am Chem
Soc 1974, 96, 6499–6507.
[44] H. C. Brown, G. W. Kabalka, J Am Chem Soc 1970, 92, 712–714.
[45] M. Boelens, P. J. Valois, H. J. Wobben, A. Gen, J Agr Food Chem
1971, 5, 984–991.
[46] M. Ceruti, I. Degani, R. Fochi, Synthesis 1987, 79–82.
[47] K. D. Raner, J. Lusztyk, K. U. Ingold, J Am Chem Soc 1988,
110, 3519–3524.
[48] B. W. Gung, M. M. Yanik, J Org Chem 1996, 61, 947–951.
[49] Y. Odic, M. Pereyre, J Organomet Chem 1973, 55, 273–294.
© 2013 The Authors. Journal of Physical Organic Chemistry
published by John Wiley & Sons Ltd.
J. Phys. Org. Chem. 2013, 26 1048–1057
wileyonlinelibrary.com/journal/poc