Solvent polarity and organic reactivity in mixed solvents: Evidence using a reactive molecular probe to assess the role of preferential solvation in aqueous alcohols

Article abstract of DOI:10.1021/jo048163j

(Chemical Equation Presented) Product selectivities [S = ([ester product]/[acid product]) x ([water]/[alcohol solvent])] are reported for solvolyses of p-methoxybenzoyl chloride (2) in aqueous methanol, ethanol, 2,2,2-trifluoroethanol, n-propyl alcohol, isopropyl alcohol, and tert-butyl alcohol at 25, 35, and 45°C. S values are small and depend significantly on the alcohol cosolvent, varying from 1.3 in methanol to 0.1 in tert-butyl alcohol, but S depends only slightly on the solvent composition, and on the temperature. As S adjusts the product ratios for changes in bulk solvent compositions, it is suggested that preferential solvation by either alcohol or water at the reaction site is not a major factor influencing rates or products. Logarithms of rates of solvolyses of 2 correlate well with Kosower Z values (based on solvatochromism). In contrast, another solvatochromic polarity index, E_T(30), shows "dispersion" in correlations with the solvent ionizing power parameter, Y_OTs, probably due to aromatic ring and other solvation effects.

Full text of DOI:10.1021/jo048163j

Solvent Polarity and Organic Reactivity in Mixed Solvents:
Evidence Using a Reactive Molecular Probe To Assess the Role of
Preferential Solvation in Aqueous Alcohols
,†
†
‡
‡
T. William Bentley,* David N. Ebdon, Eun-Ju Kim, and In Sun Koo
Department of Chemistry, University of Wales, Swansea, Singleton Park, Swansea,
SA2 8PP, Wales, United Kingdom, and Department of Chemical Education and the Research Institute of
Natural Science, Gyeongsang National University, Jinju, 660-701, Korea
cmsolvol@swansea.ac.uk
Received October 18, 2004
Product selectivities [S ) ([ester product]/[acid product]) × ([water]/[alcohol solvent])] are reported
for solvolyses of p-methoxybenzoyl chloride (2) in aqueous methanol, ethanol, 2,2,2-trifluoroethanol,
n-propyl alcohol, isopropyl alcohol, and tert-butyl alcohol at 25, 35, and 45 °C. S values are small
and depend significantly on the alcohol cosolvent, varying from 1.3 in methanol to 0.1 in tert-butyl
alcohol, but S depends only slightly on the solvent composition, and on the temperature. As S
adjusts the product ratios for changes in bulk solvent compositions, it is suggested that preferential
solvation by either alcohol or water at the reaction site is not a major factor influencing rates or
products. Logarithms of rates of solvolyses of 2 correlate well with Kosower Z values (based on
T
solvatochromism). In contrast, another solvatochromic polarity index, E (30), shows “dispersion”
in correlations with the solvent ionizing power parameter, YOTs, probably due to aromatic ring and
other solvation effects.
Introduction
stitutions in aqueous alcohols is that two products are
formed, and quantitative product studies provide ad-
ditional data helping to unravel the complexity.
Organic reactions are frequently carried out in aque-
ous/organic solvent mixtures for practical reasons: e.g.,
to increase solubilities for reactions in aqueous media or
to increase rates of reactions in organic media. For kinetic
and mechanistic studies, mixed solvents have an advan-
tage in providing the opportunity to vary continuously
the reaction conditions, in contrast to the more abrupt
changes arising from studies of substituent effects or the
very subtle changes arising from secondary kinetic
isotope effects. The main disadvantage of mixed solvents
is that interpretations of the data are more complex, but
an important advantage of studies of nucleophilic sub-
1
We now report an investigation of the possibility that
rates and products of organic reactions in mixed alcohol/
water mixtures could be influenced by preferential sol-
1
vation, alternatively referred to as solvent sorting, in
which one or more solvation shells differ significantly in
molar composition from the bulk solvent mixture. As
might be expected, preferential solvation is particularly
important for salts in mixtures of water and organic
cosolvents,2a but it also occurs for the polar probe dye
(
1) Ta-Shma, R.; Rappoport, Z. Adv. Phys. Org. Chem. 1992, 27,
2
76-280 (a), 255-256 (b), 239-291 (c), 244 (d).
†
University of Wales.
Gyeongsang National University.
(2) Marcus, Y. Ion Solvation; Wiley: Chichester, UK, 1985; Chapter
7 (a), p 185 (b), p 184 (c), and p 202 (d).
‡
1
0.1021/jo048163j CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/28/2005
J. Org. Chem. 2005, 70, 1647-1653
1647

Bentley et al.
indicators used to determine solvation indices:2b e.g., for
the solvent polarity index E (30), based on the solvent
dependence of electronic transitions (solvatochromism)
TABLE 1. Product Selectivities (S, Eq 1) for Solvolyses
of p-Methoxybenzoyl Chloride (2) in Methanol/Water (%
v/v)
T
of the pyridinium N-phenolate betaine dye (1). There is
25 °C
set Bb,c
3
5 °C
45 °C
a relatively large increase in E
T
(30) when small amounts
% v/v
set Aa
set Cb
set Cb
set Cb
3a
of water are added to aprotic solvents (e.g., acetonitrile
9
8
0
0
1.36
1.31
1.35
1.24
1.26
1.31
1.32
1.37
1.39
1.37
1.40
1.39
1.19
1.31
1.32
1.25
1.26
1.27
1.30
1.28
1.39
1.36
1.32
1.31
1.31
1.34
1.40
1.32
1.47
1.25
1.34
1.27
1.19
1.25
1.18
1.52
3a
and dioxane ), providing strong evidence for preferential
solvation by water. Aprotic organic cosolvents would be
expected to be more prone to preferential solvation than
70
6
50
0
1.23
1.35
1.43
1
13
4
alcohol cosolvents, but H and C NMR of 1 and recent
4
3
2
0
0
0
1
intermolecular H-NOESY NMR of small neutral organic
5
solutes provided evidence for preferential solvation for
various binary solvents including aqueous alcohols.
There is much current interest in preferential solvation
from observations of solvatochromism of binary aqueous
10
a
Data from from ref 11. b From at least duplicate HPLC
analyses of at least two independent samples; typical errors (3%.
Reference 14.
3b
c
mixtures, but it is very difficult to separate preferential
6
solvation from other solvent effects. The possible role of
preferential solvation in solvolytic reactivity is relevant
to innumerable mechanistic studies of organic substrates
TABLE 2. Product Selectivities (S, Eq 1) for Solvolyses
of p-Methoxybenzoyl Chloride (2) in Ethanol/Water (%
v/v)
1,7-9
in mixed solvents, and has been considered previously,
1
but was not thought to be a “major factor”. Recent
25 °C
set Bb
3
5 °C
45 °C
5
,6
developments have led us to investigate further.
Inert chemical probes are often used for the charac-
terization of solvent mixtures.10 A completely indepen-
dent and direct probe of the solvent composition around
the reaction site is the product ratio, when alcohols or
water react competitively as nucleophiles with a reactive
molecular probe via an extremely short-lived intermedi-
ate. We now report applications of this alternative
approach to studies of solvent effects in binary mixtures
of water and various alcohols, and we discuss solvent
effects other than preferential solvation, which could
influence solvatochromism of 1.
set Aa
set Cc
c
set Cc
%
v/v
set C
9
0
0.71
0.64
0.71
0.70
0.72
0.73
0.75
0.77
0.71
0.66
0.70
0.75
0.74
0.76
0.78
0.78
0.78
0.55
0.57
0.60
0.64
0.68
0.75
0.74
0.77
0.80
0.64
0.51
0.67
0.74
0.76
0.82
0.85
0.85
0.82
0.66
0.66
0.62
0.69
0.85
0.90
0.84
8
0
70
60
5
4
3
2
0
0
0
0
10
a,b,c
As for Table 1.
cation is not a major product-determining step.1
1-13
When
product selectivities (S) are defined by using eq 1, where
S ) ([ester product]/[acid product]) ×
([water]/[alcohol solvent]) (1)
the square brackets refer to concentrations in moles/liter,
product ratios are adjusted for changes in bulk solvent
compositions (i.e., ignoring preferential solvation). S
values were found to be approximately independent of
solvent composition for solvolyses of 2 in methanol/water
and ethanol/water at 25 °C.11 We now extend our previous
work, and discuss the implications for the effect of
preferential solvation on reactivity in alcohol/water mix-
tures.
The chosen probe substrate was p-methoxybenzoyl
chloride (2), which reacts in alcohol/water mixtures to
form a mixture of an ester and an acid. Product studies
and salt effects are consistent with such rapid reactions
of solvent-separated ion pair intermediates with protic
solvents that reaction via a more dissociated (“free”)
Results
Product selectivities (S, eq 1) are shown for solvolyses
in aqueous methanol (Table 1) and ethanol (Table 2) at
three temperatures, and for methanol/ethanol mixtures
at 25 °C in Table 3; there are also data for aqueous
mixtures of tert-butyl alcohol (Table 4), and 2,2,2-tri-
fluoroethanol (TFE), n-propyl alcohol, and isopropyl al-
(
3) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry,
3
rd ed.; Wiley-VCH: Weinheim, Germany, 2003; pp 38-42 (a) and
26-430 (b).
4
(
4) Dawber, J. G.; Williams, R. A. J. Chem. Soc., Faraday Trans. 1
1
986, 82, 3097-3112.
(
5) Bagno, A. J. Phys. Org. Chem. 2002, 15, 790-795.
(
6) Bentley, T. W.; Koo, I. S. Org. Biomol. Chem. 2004, 2, 2376-
(11) Bentley, T. W.; Harris, H. C.; Koo, I. S. J. Chem. Soc., Perkin
Trans. 2 1988, 783-789.
2
380.
(
7) (a) Tommila, E. Acta Chem. Scand. 1955, 9, 975-988. (b) Hyne,
(12) Song, B. D.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 8470-
8479.
J. B.; Robertson, R. E. Can. J. Chem. 1956, 33, 933-941.
8) Harris, J. M.; Clark, D. C.; Becker, A.; Fagan, J. F. J. Am. Chem.
(
(13) Several other substrates (e.g. adamantyl, classified1b as the first
group) behave similarly, but are more hydrophobic and do not possess
a UV chromophore, and so cannot be investigated over a wide range
of aqueous mixtures.
Soc. 1974, 96, 4478-4484.
(
9) Bentley, T. W.; Harris, H. C. J. Org. Chem. 1988, 53, 724-728.
10) Marcus, Y. J. Chem. Soc., Perkin Trans. 2 1994, 1751-1758.
(
1648 J. Org. Chem., Vol. 70, No. 5, 2005

Solvent Polarity and Organic Reactivity in Mixed Solvents
TABLE 3. Product Selectivities (S) for Solvolyses of
p-Methoxybenzoyl Chloride (2) in Methanol/Ethanol (%
v/v) at 25 °C
nine solvent compositions and all three temperatures;
deviations from the average value of 1.32 (23 values at
2
5 °C) show no consistent trends. Even at 35 and 45 °C,
%
av/v
20
1.54
40
1.51
60
1.50
80
1.51
there are only two slightly high S values for the most
rapid reactions in the more aqueous media, possibly due
to insufficiently rapid mixing. A similar lack of variation
in S is observed for TFE, and the average of 24 values is
S
a
Refers to methanol/ethanol product ratios in eq 1; data from
ref 14.
0
.135 (Table 5); possibly S decreases in more aqueous
TABLE 4. Product Selectivities (S, Eq 1) for Solvolyses
of p-Methoxybenzoyl Chloride (2) in tert-Butyl Alcohol/
Water (% v/v)
media at 25 °C.
New results for ethanol (B and C, Table 2) show a small
but consistent increase in S for more aqueous solvents
at all three temperatures, a trend less convincingly
discernible in previous data (A, Table 2). For data set C,
the extreme values at 25 °C (0.55 and 0.80) differ by
<20% from the average value of 0.68 (other averages:
0.74 at 35 °C and 0.75 at 45 °C). All results for n-PrOH
25 °C
3
5 °C
45 °C
set Ba
set Cb
set C
b
set C
b
%
v/v
9
0
0
0
0
0
0
0
0
0
0.12
0.12
0.12
0.12
0.12
0.12
0.13
0.11
0.12
0.11
0.12
0.12
0.12
0.14
0.17
0.14
0.13
0.11
0.11
0.12
0.14
0.15
0.17
0.15
0.14
0.13
0.21
0.28
0.31
0.33
0.22
8
7
6
5
4
3
2
1
(
Table 5) show similar but more marked trends. Results
for i-PrOH (Table 5) and t-BuOH (Table 4) are ap-
proximately constant (except for some values at 45 °C),
although the low values of S automatically compress the
range of values.
There are few previous studies of S for solvolyses in a
wide range of alcohol/water mixtures, but values of 1/S
for solvolyses of 2-bromoadamantane at 120-150 °C
defined the dimensiosolvatic parameter (D, Table 6), for
which steric effects had a major role.1 Steric effects are
apparent from the following order of S for solvolyses of
a
Data from ref 14.; the mobile phase for HPLC contained 0.1%
triethylamine instead of acetic acid. b As for Table 1.
cohol (Table 5) at 25, 35, and 45 °C. All data were
obtained by response calibrated HPLC, and were then
converted to selectivities (S, eq 1).
Product ratios were constant after 8-10 half-lives (up
to several hours) and after longer times (up to 2 days),
showing that products did not slowly interconvert during
the reaction. Also, the ester products were shown to be
stable in acidified 50% alcohol/water, and the acid
product (p-methoxybenzoic acid) was stable in acidified
5a
2
in aqueous alcohols at 25 °CsMeOH (1.32, Table 1) >
EtOH (0.71, Table 2) > i-PrOH (0.36, Table 5) > t-BuOH
0.12, Table 4)sand from selectivities for solvolyses of
(
1
1
-bromoadamantane in 70 and 80% aqueous alcohols at
00 °C (Table 6).1
5b,c
By definition S ) 1 for water, so
water is not in the correct sequence for two of the entries
in Table 6, and other factors such as the number of
hydrogen bonds and/or electrophilicities must contrib-
ute.15
9
0% alcohol/water mixtures, showing that rapid inter-
conversion of products did not occur. Two different HPLC
analyses of the acid were employed to resolve the acid
signal from the relatively large solvent signals; addition
of 0.1% acetic acid to the eluent (ionization suppression),
Other minor solvent effects are also needed to explain
why the ratio of S values at 25 °C for methanol/water
and ethanol/water (1.32/0.71 ) 1.86) is not the same as
the value of 1.52 obtained in methanol/ethanol mixtures
or alternatively 0.1% triethylamine (ion pairing), gave
very similar S values.14
(
Table 3); taking only set B data from Tables 1 and 2,
Three independent measurements in our laboratories
are shown in Table 1 (for methanol/water) and in Table
the ratio of S values is 1.33/0.74 ) 1.80, so the discrep-
ancy is about 20%. Also, for 1-bromoadamantane, the
ratio of the selectivities for water/TFE and water/ethanol
2
(for ethanol/water). Agreement is within (5%. Two
independent measurements of S at 25 °C for t-BuOH/
water (Table 4) and for i-PrOH (Table 5) agree satisfac-
torily, with two exceptions (30% and 40% i-PrOH). A
possible source of error, leading to anomalously high
values of S, occurs for very rapid reactions in highly
aqueous media due to insufficiently rapid dispersal after
(
1.04/1.26 ) 0.83) is not the same as the average
15b
selectivity of 0.68 for ethanol/TFE. Selectivities for TFE
could be low because the CF group is oriented toward
3
the positive charge, so that trapping cannot occur unless
there is time for a rotation.16
11
mixing. Independent results for n-PrOH (Table 5) agree
well, except for 90% alcohol, which also shows an
anomalous temperature dependence.
Implications. Solvent effects on solvolyses of 2 are
1
1
similar to those for 2,6-dimethylbenzoyl chloride and
p-methoxybenzyl chloride,1 and solvolyses of p-meth-
7a
1
7b
oxybenzoyl bromide correlate well with YBnBr
.
These
Discussion
(
15) (a) Oki, M.; Ikeda, H.; Toyota, S. Bull. Chem. Soc., Jpn. 1998,
Trends in S Values. First, we consider whether S
values for individual alcohol/water mixtures are constant
within experimental uncertainties or show consistent
variations with composition or temperature. Results for
methanol (Table 1) appear to be close to constant, for all
71, 749-754. (b) McManus, S. P.; Zutaut, S. E. Isr. J. Chem. 1985, 26,
4
2
2
00-403. (c) McManus, S. P.; Zutaut, S. E. Tetrahedron Lett. 1984,
5, 2859-2862. (d) Ando, T.; Tsukamoto, S. Tetrahedron Lett. 1977,
775-2778.
(16) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106,
373-1383.
1
(
17) (a) Bentley, T. W.; Koo, I. S.; Norman, S. J. J. Org. Chem. 1991,
5
6, 1604-1609. (b) Liu, K.-T.; Hou, I.-J. Tetrahedon, 2001, 57, 3343-
(
14) (a) Ebdon, D. N. Ph.D. Thesis, University of Wales, 1999,
3347. (c) Kevill, D. N.; D’Souza, M. J. J. Phys. Org. Chem. 2002, 15,
881-888. (d) Bentley, T. W.; Koo, I. S. J. Chem. Soc., Perkin Trans. 2
1989, 1385-1392. (e) Liu, K.-T.; Chen, H.-I. J. Chem. Soc., Perkin
Trans. 2 2000, 893-898.
Chapter 6. (b) Bentley, T. W.; Ebdon, D. N. Atualidades de F ı´ sico-
Qu ı´ mica Org aˆ nica; Humeres, E., Ed.; Florian o´ polis, SC, Brazil, ISSN
1
414-0314, 1999; pp 1-14.
J. Org. Chem, Vol. 70, No. 5, 2005 1649

Bentley et al.
TABLE 5. Product Selectivities (S, Eq 1) for Solvolyses of p-Methoxybenzoyl Chloride (2) in n-Propyl Alcohol,
Isopropyl Alcohol, and 2,2,2-Trifluoroethanol/Water (% v/v)
n-PrOH
i-PrOH
TFE
35°C
25°C
25°C
3
5°C
45°C
35°C
45°C
25°C
45°C
set Ba
set Cb
set C
b
set C
b
set Ba
set Cb
set C
b
set C
b
set C
b
set C
b
set C
b
%
v/v
9
0
0
0
0
0
0
0
0
0
0.61
0.56
0.55
0.63
0.60
0.66
0.70
0.73
0.78
0.32
0.42
0.47
0.53
0.60
0.60
0.67
0.75
0.82
0.81
0.67
0.68
0.82
0.75
0.85
0.90
0.37
0.51
0.64
0.64
0.79
0.98
0.91
0.33
0.31
0.32
0.33
0.35
0.36
0.35
0.34
0.32
0.36
0.35
0.34
0.36
0.39
0.43
0.44
0.28
0.29
0.31
0.32
0.31
0.29
0.30
0.32
0.28
0.22
0.30
0.35
0.47
0.53
0.56
0.74
0.49
0.49
0.15
0.15
0.15
0.14
0.15
0.13
0.15
0.11
0.11
0.17
0.16
0.14
0.12
0.13
0.15
0.15
0.16
0.12
0.11
0.10
0.11
0.12
0.14
0.12
8
7
6
5
4
3
2
1
a
Data from ref 14. b As for Table 1.
TABLE 6. Comparison of 1/S with D Values from
N
SCHEME 1. Proposed S 1 Mechanism for
Dimensiosolvatic Effects
Solvolysis of p-Methoxybenzoyl Chloride (2) in
Alcohol/Water, with Product Formation via
Front-Side Collapse of Water- or
solvent
MeOH
EtOH
n-PrOH
i-PrOH
t-BuOH
Alcohol-Separated Ion Pairs
a
1
/S
0.76
1.5
1.0
1.4
1.7
1.3
1.6
2.5
2.8
2.6
4.9
8.3
10
8.7
b
D
1
-AdBrc
a
Relative to water (S ) 1), for solvolyses of p-methoxybenzoyl
b
chloride (2) at 25 °C. Defined as 1/S for solvolyses of 2-bromoada-
c
mantane at 150 °C (from ref 15a). Average of selectivities for
1
-bromoadamantane in 70% and 80% v/v alcohol/water at 100 °C,
quoted in ref 15b as k(water)/k(alcohol), so the values are
equivalent to 1/S in eq 1.
N
results are consistent with S 1 reactions for solvolyses
in ethanol and in more polar solvents. As there is a
tendency for aromatic acid chlorides to undergo a mecha-
nistic change to addition-elimination, depending on the
substituent and solvent,1 it is possible (but unproven)
that the mechanism of solvolyses of 2 could change in
solvents of lower polarity (e.g. tert-butyl alcohol); also,
the transition state may be weakly solvated nucleo-
7c,d
is negligible in mixtures of similar alcohols. This assump-
tion is intuitively reasonable, and is supported experi-
mentally by E (30) data for methanol and ethanol
T
1
7c,e
19a
philically.
showing close to ideal behavior. Assumption 3: Sol-
From the lack of trapping by fluoride ion, the rate of
reaction of the p-methoxyphenyl acylium ion with water
was estimated at k > 1011 s , which corresponds to a
lifetime of the “cation” even shorter than typical lifetimes
volyses of (2) will show a constant value of S (eq 1) if
preferential solvation is absent. This assumption is sup-
ported by the constant S values for solvolyses of 2 in
methanol/ethanol (Table 3), when preferential solvation
is assumed to be absent (assumption 2).
-1 12
(
ns) of solvents in coordination spheres of for example
2d
lithium or sodium cations. Reactions occurring via “free”
carbocations give S > 1 for alcohol/water mixtures.1
As S values for solvolyses of 2 are generally low, and S
If the above three assumptions are accepted, based on
the arguments and experimental evidence discussed
above, it follows that the relatively constant S values for
solvolyses of 2 in alcohol/water mixtures (Tables 1, 2, 4,
and 5) can be explained by similar solvent compositions
at the reaction site and in the bulk solvent (except for
n-PrOH). Although there are not large variations in
selectivities,1 our experimental probe (S values) should
be sensitive to solvent composition at the reaction site
because the mass law effect of concentration on rates
should apply.
c,18
<
1 (except for MeOH), it is assumed that all or most of
the product is formed when a molecule of protic solvent
acts as a nucleophile and attacks a solvent-separated ion
pair intermediate (Scheme 1).1b,11,12,15 This leads us to
propose the following. Assumption 1: Product ratios for
solvolyses of 2, in any particular binary mixture of protic
solvents, will depend on the solvent composition adjacent
to the reaction site at the instant when the product-
forming reaction occurs (as discussed above, product
ratios in a range of binary mixtures will also depend on
properties of the two competing solvents such as elec-
trophilicity, nucleophilicity, steric bulk, and number of
hydrogen bonds). Assumption 2: Preferential solvation
c
Our results are not in clear agreement with indepen-
dent studies of preferential solvation.5 S for each binary
mixture varies within a factor of 2 in all cases except two
(90% n-PrOH at 25 and 45 °C, Table 5). In contrast, for
,19
1
equimolar mixtures of water and n-PrOH, H NOESY
(
18) (a) Richard, J. P.; Rothenberg, M. E.; Jencks, W. P. J. Am.
(19) (a) Bosch, E.; Ros e´ s, M. J. Chem. Soc., Faraday Trans. 1992,
88, 3541-3546. (b) Skwierczynski, R. D.; Connors, K. A. J. Chem. Soc.,
Perkin Trans. 2 1994, 467-472.
Chem. Soc. 1984, 106, 1361-1372. (b) Bentley, T. W.; Ryu, Z. H. J.
Chem. Soc., Perkin Trans. 2 1994, 761-767.
1650 J. Org. Chem., Vol. 70, No. 5, 2005

Solvent Polarity and Organic Reactivity in Mixed Solvents
NMR data show a 5-fold preference for n-PrOH solvating
phenol, and data analyzed by the 1-step (2 state)
preferential solvation by ethanol would also lead to
greater S values for ethanol.
5
exchange model imply a 12-fold preference for n-PrOH
Detailed selectivity data for solvolyses of derivatives
of 2-deoxyglucosoyl isoquinolinium tetrafluoroborates at
65 °C indicate that the 2-deoxyglucosoyl oxacarbenium
ion is so reactive that it is not solvent-equilibrated in
1
9b
solvating 1;
perhaps even more remarkable is the
predicted 4.4-fold preference for MeOH solvating 1 in
aqueous methanol.1
9b
typical alcohol/water mixtures.2 As in our work,
2b
18b
A possible source of discrepancies is that S values only
measure changes (they are not absolute). If, as sug-
SMeOH/water > SEtOH/water, but S varies up to 2-fold with
1
9
gested, the mole ratio of solvent in the solvation shell
is proportional to that in the bulk solvent, constant S
values would be observed because the solvent ratio (eq
solvent composition; S increases as water is added to
ethanol or methanol (possibly due to preferential solva-
tion at the positively charged, relatively hydrophobic
2
2b
1
) is unchanged. Another possibility is that the “true”
reaction site), but is constant or decreases for TFE.
mole ratio of water/alcohol at the reaction site could be
say 4-fold less than the ratio in the bulk solvent, and then
the true S (eq 1) would also be 4-fold lower. Our
counterarguments are the following: (i) S values (based
on bulk solvent compositions) are already surprisingly
low, because alcohols are more nucleophilic than water;20
Solvolysis of R-D-glucopyranosyl fluoride in 55/45 v/v
water/methanol gives <1% yield of methyl â-D-glucopy-
ranoside, possibly because the hydrophilic reaction site
is preferentially solvated by water.2
Solvolyses of 1-adamantyl substrates also proceed via
highly unstable intermediates, but data are usually
available only for more highly alcoholic v/v compositions;
2c,d
(ii) contrary to the expected reduction in preferential
solvation caused by decreasing hydrogen bonding as
temperatures are increased,21 our S values tend to
increase at increased temperatures and our 1/S values
at 25 °C parallel D values at 150 °C within a factor of 2
S values (shown as k
dimethylsulfonium salts vary little (between 1.11 and
W
/k
E
rather than k
E
/k
W
) S) for
1.36) for solvolyses over a wide solvent range (20-90%
v/v ethanol/water) at 70.6 °C;2 chloroformates vary
3a
23b
(Table 6).
between 1.11 and 1.66 for 96-40% v/v ethanol/water at
2
5 °C. Other S values (k
E W
/k ) for 1-bromoadamantane in
A more plausible explanation of the above discrepan-
1d
ethanol/water are also close to constant.
cies is that organic groups distant from the reaction site
may be preferentially solvated. As the reaction site is only
a small part of a typical organic substrate, preferential
solvation of more distant hydrophobic groups within the
substrate will not be reflected in S values: e.g., as
preferential solvation at the molecular level was detected
Comparisons of Solvent Effects on Rate Con-
stants and on Solvatochromism in Alcohol/Water
Mixtures. Preferential solvation has been investigated
3b,19
in detail through solvent effects on solvatochromism,
and it could also influence solvent effects on reactivity.
5
For applications of the Grunwald-Winstein equation for
for phenol in alcohol/water, so perhaps only the phenyl
24
solvent-ionizing power (Y, a measure of solvent polar-
ring is solvated preferentially by the more hydrophobic
alcohol. Although “resolution” at the atomic level would
25
ity ), plots of logarithms of rate constants (k) for various
solvolyses in aqueous alcohols versus various solvent
functions were examined, and mole fraction was selected
5
be desireable, product ratios provide experimental evi-
dence at an intermediate level of “resolution”, but cru-
cially solvent effects at the reaction site are probed (i.e.,
solvent effects on reactivity).
2
4
empirically as the most suitable for interpolations.
However, there is no theoretical basis for the expecta-
tion26 that there should be a linear relationship between
logarithms of rate constants and mole fractions for the
Investigations of other reactive probe substrates are
desirable, especially if additional, positive rather than
negative evidence could be obtained in some cases.
Choices are limited by mechanistic changes, involving for
2
c
full composition range of a binary aqueous mixtures s
volume fraction has recently been advocated for aqueous
mixtures,6 and preferential solvation is not the sole
cause of deviations from a linear relationship (as pro-
,27
1
b
example either more stable, dissociated carbocations,
or prior nucleophilic attack.12,17d Suitable substrates
26
6
posed ) because many other solvent effects are possible.
When Y (based on logarithms of rates of solvolyses of
tert-butyl chloride) is plotted against YCl (based on
solvolyses of the much more hydrophobic 1-chloroada-
mantane), there is very little “dispersion” into separate
correlation lines for aqueous acetone, ethanol, and metha-
nol,28 consistent with little or no effect of preferential
solvation on solvolytic reactivity. Dispersion depends on
the number of electrons adjacent to the sites of positive
should also dissolve rapidly in highly aqueous media.
Cationic substrates provide possible water-soluble alter-
native reactive probes, although they may be more
susceptible to effects of preferential solvation. Solvolyses
of 2-, 3-, and 4-methylphenyldiazonium ions were shown
to have constant selectivities (with S < 1) in both
methanol/water
and
ethanol/water
mixtures.
1
8b
Unexpectedly
SEtOH/water > SMeOH/water, explained by
immediate reaction of a highly destabilized cation with
a solvent nucleophile in the first solvation shell;2
2a
(
23) (a) Kevill, D. N.; Anderson, S. W. J. Am. Chem. Soc. 1986, 108,
1
579-1585. (b) Kevill, D. N.; Kyong, J. B.; Weitl, F. L. J. Org. Chem.
(
20) Kevill, D. N. Advances in Quantitative Structure-Property
Relationships; Charton. M., Ed.; JAI Press: Greenwich, CT, 1996; Vol.
, pp 82-115.
1990, 55, 4304-4311.
(24) Grunwald, E.; Winstein, S. J. Am. Chem. Soc. 1948, 70, 846-
854.
1
(
21) (a) Antonious, M. S.; Tada, E. B.; El Seoud, O. A. J. Phys. Org.
(25) Katritzky, A. R.; Fara, D. C.; Yang, H.; T a¨ mm, K.; Tamm, T.;
Karelson, M. Chem. Rev. 2004, 104, 175-198.
(26) Dawber, J. G.; Ward, J.; Williams, R. A. J. Chem. Soc., Faraday
Trans. 1 1988, 84, 713-727.
(27) Asaad N.; den Otter, M. J.; Engberts, J. B. F. N. Org. Biomol.
Chem. 2004, 2, 1404-1412.
(28) Bentley, T. W.; Carter, G. E. J. Am. Chem. Soc. 1982, 104,
5471-5747.
Chem. 2002, 15, 403-412. (b) Tada, E. B.; Silva, P. L.; El Seoud, O. A.
J. Phys. Org. Chem. 2003, 16, 691-699.
(22) (a) Pazo-Llorente, R.; Bravo-D ´ı az, C.; Gonz a´ lez Romero, E. Eur.
J. Org. Chem. 2003, 17, 3421-3428. (b) Zhu, J.; Bennet, A. J. J. Org.
Chem. 2000, 65, 4423-4430. (c) Banait, N. S.; Jencks, W. P. J. Am.
Chem. Soc. 1991, 113, 7958-7963. (d) We thank a reviewer for drawing
our attention to this experiment.
J. Org. Chem, Vol. 70, No. 5, 2005 1651

Bentley et al.
FIGURE 1. Correlation of Z values (kcal/mol) with logarithms
of rate constants for solvolyses of p-methoxybenzoyl chloride
FIGURE 2. Correlation of E
T
(30) with YOTs at 25 °C; values
of E (30) are calculated from the published transition energies
T
(
2), relative to 80% v/v ethanol/water (k
compositions refer to % v/v organic solvent; slope ) 3.44 (
.12, intercept ) 84.23 ( 0.13, r ) 0.991, n ) 16; Z values are
0
) at 25 °C; all solvent
(
ref 38; see also refs 19b and 39); YOTs values are from ref 40;
all solvent compositions are % v/v; slope ) 2.0 ( 0.2, intercept
53.9 ( 0.4, r ) 0.942 for the 19 solvents in Figure 2 for which
I values are available (ref 30); an improved correlation is E
30) ) (2.38 ( 0.14)YOTs + (6.0 ( 1.3)I + (54.5 ( 0.3), r ) 0.975.
0
)
from ref 31, with additional data for 70% and 80% methanol/
water and 80 and 90% acetonitrile/water (open circles) from a
correlation between Z and the transition energy of the charge-
transfer absorption band of N-ethyl-4-cyanopyridinium iodide
T
-
(
(
ref 32); kinetic data are from ref 17a.
water of 95.4 ( 0.7 kcal/mol, higher and with a more
realistic uncertainty than the original value of 94.6 (
charge,29 and this can be accounted for quantitatively by
31
0.1.
using the aromatic ring solvation parameter (I).30
The Z value for pure water has attracted much
1
0
We now consider the extent to which a special solvation
effect of aromatic rings (and other π-electron systems)
could influence solvatochromic measures of solvent polar-
ity. For Z values, based on the transition energy of the
charge-transfer absorption band of the solvatochromic
attention, and significantly lower Z values of 92.5,
91.8,2 and 91.4 have also been proposed from various
5,35
32
correlations. An extrapolation of the data in Figure 1,
36
based on a very recent direct measurement giving k )
-
1
58 s for solvolyses of 2 in ca. 4% acetonitrile/water at
25 °C, gives a value of Z ) 96.0 ( 0.6 kcal/mol; because
of the presence of acetonitrile, and from our data for 2
in less aqueous media,1 it is likely that k in water at
3
1
dye (3), a plot (Figure 1) against logarithms of rate
constants for solvolyses of 2 shows very little dispersion
into separate correlation lines for aqueous acetone,
methanol, and ethanol. Both substrates (2 and 3) contain
only one aromatic ring, and the difference in anion does
not significantly affect the correlation because the three
binary solvent mixtures have very similar electrophilici-
7d
-1
25 °C is greater than 58 s , thus giving a slightly higher
Z value. A value of Z ) 95.3 ( 0.7 is obtained from a
correlation (r ) 0.99 for 11 solvent compositions of
dioxane/water, ethanol/water, and methanol/water) of Z
vs solvolyses of R-phenylethyl chloride (kinetic data from
ref 37sdata for water required a small extrapolation).
All of the quoted errors are standard errors, so the upper
I
ties (e.g., see the good correlation of YCl with Y based on
33
solvolyses of 1-iodoadamantane ).
and lower limits (at 95% confidence) are wider.
As the lower values1
0,25,32,35
of ca. 92 kcal/mol were
obtained from correlations in which dispersion was
observed, additional solvent effects were present, and so
19b
extrapolations are more uncertain. From published data,
extrapolations of E (30) vs mole fraction (x) of the
approximately linear regions from pure methanol or
(30) value of pure
water, too low by ca. 5 kcal/mol. Consequently, the
discrepancy of ca. 4 kcal/mol between our estimate of Z
T
ethanol up to x ) 0.5 would give an E
T
Earlier plots of Z vs Y showed dispersion and linearity
over a limited solvent range,31 but plots of E
T
(30) vs Y
3
4
not only showed dispersion but were also nonlinear. We
previously reported a plot (showing very little dispersion,
but for a less extensive range of solvents) of Z vs
logarithms of rate constants for solvolyses of p-methoxy-
)
ca. 96 kcal/mol and the lower values is not surprising.
In contrast to the lack of dispersion in Figure 1, a plot
of Z vs Y
difference between acetone/water and methanol/water).
Also, a plot (Figure 2) of E (30) vs YOTs (based on
I
gave substantial dispersion (>4 kcal/mol
33
1
7a
benzyl chloride; extrapolation gave a Z value for pure
T
40
solvolyses of 1- or 2-adamantyl tosylate) shows similar
dispersion. Plots such as Figures 1 and 2 converge
automatically to the data point for pure water, but
(
29) Bentley, T. W.; Dau-Schmidt, J.-P.; Llewellyn, G.; Mayr, H. J.
Org. Chem. 1992, 57, 2387-2392.
(
30) Kevill, D. N.; Ismail, N. H. J.; D’Souza, M. J. J. Org. Chem.
1
994, 59, 6303-6312.
(
(
31) Kosower, E. M. J. Am. Chem. Soc. 1958, 80, 3253-3260.
32) Medda, K.; Pal, M.; Bagchi, S. J. Chem. Soc., Faraday Trans.
(35) Griffiths, T. R.; Pugh, D. C. J. Solution Chem. 1979, 8, 247-
1
4
1988, 84, 1501-1507.
258.
(
33) Bentley, T. W.; Carter, G. E.; Roberts, K. J. Org. Chem. 1984,
9, 5183-5189.
34) Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Liebigs
Ann. Chem. 1963, 661, 1-37.
(36) B a´ scuas, J.; Garc ´ı a-Rio, L.; Leis, J. R. Org. Biomol. Chem. 2004,
2, 1186-1193.
(
(37) Fainberg, A. H.; Winstein, S. J. Am. Chem. Soc. 1957, 79, 1597-
1602.
1652 J. Org. Chem., Vol. 70, No. 5, 2005

Solvent Polarity and Organic Reactivity in Mixed Solvents
dispersion cannot be due solely to preferential solvation
in the aqueous organic mixtures, because for each alcohol/
water mixture the data for the two pure organic solvents
of differences in aromatic ring solvation, one of the
possible solvent effects other than preferential solvation
6,19
T
also contributing to the observed nonlinear plots of E -
(30) vs mole fraction19 or volume fraction for alcohol/
34
(methanol and ethanol) show the largest dispersion. The
pattern of dispersion (MeOH > EtOH > acetone-water)
water mixtures; when these and other possible solvent
effects are ignored, the role of preferential solvation is
overemphasized.6
19
can be explained by using the aromatic ring solvation
30
parameter (I). Water has the lowest I value (-0.45), and
others are in the order 90% acetone/water (-0.17), 90%
dioxane/water (-0.12), 100% ethanol (0.20), and 100%
Experimental Section
3
0
methanol (0.41); details of correlations (given in the
legend to Figure 2) indicate a substantial dependence of
Materials. Commercial samples (Aldrich) of p-methoxy-
benzoyl chloride (2) were purified by reduced pressure distil-
lation, and were checked for purity by HPLC analysis of
methanolysis products; p-methoxybenzoic acid (Aldrich) was
recrystallized from ethanol. Alcohols for selectivity measure-
ments were dried by reaction with magnesium, except for tert-
butyl alcohol, which was distilled from sodium. Solvents for
HPLC were AR grade. HPLC response factors were obtained
from standard solutions of esters, prepared by injecting 10.0
µL of a 5% solution of 2 in dry acetonitrile into 5.00 mL of
alcohol or 40% acetonitrile/water.
T
E (30) on I.
Conclusions
Low and approximately constant product selectivities
for solvolyses of 2 (S, eq 1, Tables 1-5) for a wide range
of aqueous alcohol solvent compositions and at several
temperatures, typically employed in kinetic and mecha-
nistic studies, are consistent with previous conclusions1
that the solvent compositions at the reaction site do not
vary greatly from those of the bulk solvents. However,
small increases in S in more aqueous solvents, particu-
larly for ethanol (Table 2) and n-propyl alcohol (Table 5)
indicate that preferential solvation may occur. Overall,
these results provide further evidence that preferential
solvation/solvent sorting is not a major factor influencing
reactivity for nucleophilic substitutions of neutral organic
substrates in aqueous alcohols (i.e. at the reaction site),
even though substantial preferential solvation may be
a,8,9
Selectivities. A few microliters (5-50) of dilute solutions
(
0.2-5%) of 2 in dry acetonitrile was mixed rapidly with
thermostated solvolysis solvent (5-8 mL) in an air-driven,
41
turbostirred apparatus. Swansea procedures and equipment
14,41
were used to obtain the results coded B in Tables 1-5;
Jinju
results, coded C, were obtained by using the equipment
42
described earlier. The chromatography columns were Spher-
isorb ODS2, and the mobile phase was typically 70% methanol/
water, containing 0.1% acetic acid.
Many experiments were carried out to check that the
products were those formed under kinetic control. An illustra-
-3
tion is a study of a 4 × 10 M solution of p-methoxybenzoic
5
22a,b
observed by NMR. Cationic substrates
may be more
acid in 90% methanol/water, containing the same amount or
a 2-, 4-, or 20-fold excess of HCl; after 1 day only the latter
solution contained any ester (0.6%).
prone to preferential solvation at the reaction site.
Logarithms of rate constants for solvolyses of 2 cor-
relate well with Z values (based on electronic transition
energies for 3) with little or no “dispersion” into separate
correlation lines for each binary mixture (Figure 1);
extrapolation leads to an upward revision of the original
Z value for water (from 94.6 to ca. 96 kcal/mol), contrary
Correlations. Statistical analyses were performed with
Microsoft Excel, and graphs were plotted with KaleidaGraph
(
Synergy Software, Reading, PA).
Acknowledgment. This work was supported by
EPSRC through a research studentship (to D.N.E.). We
are grateful to N. Asaad and A. Bagno for helpful
comments.
10,25,32,35
to literature reports of lower estimates (ca. 92).
Neither solvatochromism of N-alkyl pyridinium iodides
32
(
such as 3) in alcohol/water mixtures (up to C
3
)
nor
solvolyses of 2 in alcohol/water mixtures (Tables 1-5) are
considered to involve substantial preferential solvation.
Also, structural similarities between 2 and 3 lead to
similar effects of aromatic ring solvation, so the good
Note added in proof: Professor A. Bagno has kindly
informed us of a paper (Halle, B. J. Chem. Phys. 2003,
119, 12372-12389) showing that intermolecular NOEs
may contain substantial contributions by spins from the
bulk solution, which may complicate the evaluation of
the results (e.g., ref 5).
correlation (Figure 1) is explained. In contrast E
T
(30) does
not correlate well with YOTs (Figure 2), probably because
JO048163J
(
38) Dimroth, K.; Reichardt, C. Fresenius’ Z. Anal. Chem. 1966, 215,
44-350.
39) Krygowski, T. M.; Wrona, P. K.; Zielkowska, U.; Reichardt. C.
Tetrahedron 1985, 41, 4519-4527.
40) Bentley, T. W.; Llewellyn, G. Prog. Phys. Org. Chem. 1990, 17,
21-158.
3
(
(41) Bentley, T. W.; Jones, R. O. J. Chem. Soc., Perkin Trans. 2 1993,
2351-2357.
(42) Koo, I. S.; Yang, K.; Kang, K.; Lee, I.; Bentley, T. W. J. Chem.
Soc., Perkin Trans. 2 1998, 1179-1183.
(
1
J. Org. Chem, Vol. 70, No. 5, 2005 1653