Solvent Nucleophilicity

Article abstract of DOI:10.1021/ja031828z

The rates of the reactions of benzhydrylium ions (diarylcarbenium ions) with solvent mixtures of variable composition (water/acetonitrile, methanol/acetonitrile, ethanol/acetonitrile, ethanol/water, and trifluoroethanol/water) have been determined photometrically by conventional UV-vis spectroscopy, stopped-flow methods, and laser flash techniques. It has been shown that the first-order rate constants follow the previously published relationship log k(20°C) = s(N+E), where E is an empirical electrophilicity parameter, N is an empirical nucleophilicity parameter, and s is a nucleophile-specific slope parameter. From plots of log k versus E of the benzhydrylium ions are derived the solvent nucleophilicity parameters s and N, the latter of which are designated as N₁ to emphasize that their use in the quoted correlation equation gives rise to first-order rate constants. A linear correlation between N₁ and Kevill's solvent nucleophilicity N_T based on S-methyldibenzothiophenium ions is reported, which allows one to interconvert the two sets of data. Because the N₁ values are directly comparable to the previously reported nucleophilicity parameters N for π-systems (www.cup.uni-muenchen.de/oc/mayr/), the systematic design of Friedel-Crafts reactions with solvolytically generated carbocations becomes possible.

Full text of DOI:10.1021/ja031828z

Published on Web 03/30/2004
Solvent Nucleophilicity
Shinya Minegishi,† Shinjiro Kobayashi,‡ and Herbert Mayr*,†
Contribution from the Department Chemie der Ludwig-Maximilians-UniVersita¨t Mu¨nchen,
Butenandtstrasse 5-13 (Haus F), D-81377 Mu¨nchen, Germany, and IFOC, Kyushu UniVersity,
Higashi-ku, Fukuoka 812-8581, Japan
Received December 19, 2003; E-mail: herbert.mayr@cup.uni-muenchen.de
Abstract: The rates of the reactions of benzhydrylium ions (diarylcarbenium ions) with solvent mixtures of
variable composition (water/acetonitrile, methanol/acetonitrile, ethanol/acetonitrile, ethanol/water, and
trifluoroethanol/water) have been determined photometrically by conventional UV-vis spectroscopy, stopped-
flow methods, and laser flash techniques. It has been shown that the first-order rate constants follow the
previously published relationship log k(20 °C) ) s(N + E), where E is an empirical electrophilicity parameter,
N is an empirical nucleophilicity parameter, and s is a nucleophile-specific slope parameter. From plots of
log k versus E of the benzhydrylium ions are derived the solvent nucleophilicity parameters s and N, the
latter of which are designated as N1 to emphasize that their use in the quoted correlation equation gives
rise to first-order rate constants. A linear correlation between N1 and Kevill’s solvent nucleophilicity NT
based on S-methyldibenzothiophenium ions is reported, which allows one to interconvert the two sets of
data. Because the N1 values are directly comparable to the previously reported nucleophilicity parameters
N for π-systems (www.cup.uni-muenchen.de/oc/mayr/), the systematic design of Friedel-Crafts reactions
with solvolytically generated carbocations becomes possible.
Scheme 1
Introduction
The development of quantitative scales of solvent nucleo-
philicity has intrigued chemists for several decades.1,2 Such
scales are important for the theory of nucleophilic substitutions
and for designing syntheses where the solvent may compete
with other nucleophiles for the electrophile under consideration.
In 1948, Grunwald and Winstein presented the relationship
(1), which expresses the rates of SN1 solvolyses of substrates
RX by a substrate-specific parameter m (m ) 1 for tert-butyl
chloride) and a solvent-specific parameter, the solvent-ionizing
power Y (Y ) 0 for 80% aqueous ethanol).3
nucleophilicities of alcohols.5 In the extended Grunwald-
Winstein eq 2,6 the electrophilic term mY of eq 1 is accompanied
by the nucleophilic term lN, where l is the sensitivity toward
changes in solvent nucleophilicity N.
log(k/k0)RX ) mY + lN
(2)
While the solvent-ionizing power Y can easily be determined
by investigating substrates which solvolyze without nucleophilic
solvent participation (limiting SN1, l ) 0), it is difficult to find
reactions which are entirely controlled by the nucleophilic term,
that is, substitutions with m ) 0.
log(k/k0)25°C ) mY
(1)
where k and k0 are the first-order rate constants for the solvolysis
of RX in a given solvent and the standard solvent 80% aqueous
ethanol (v/v), respectively, at 25°C. Equation 1 holds for SN1
reactions with a rate-determining ionization step where nucleo-
philic solvent participation is absent.4
Peterson and Waller7 derived a scale of nucleophilicities of
solvent molecules from the rates of reactions of solvent mole-
cules with tetramethylenehalonium ions in liquid sulfur dioxide
(Scheme 1).
Three years later, Winstein, Grunwald, and Jones reported
that SN2 reactions proceed considerably faster in aqueous
alcohols than in mixtures of acetic acid and formic acid of equal
ionizing power Y and assigned these differences to the higher
More widely employed is the NOTs scale. Schadt, Bentley,
and Schleyer defined the sensitivity of solvolysis rates of methyl
tosylate on solvent nucleophilicity as l ) 1.8 With the assump-
tion of equal nucleophilicities of acetic and formic acid, as
derived from Peterson’s and Waller’s work, Bentley and
Schleyer concluded that the sensitivity of methyl tosylate
† Universita¨t Mu¨nchen.
‡ Kyushu University.
(1) Bentley, T. W.; Llewellyn, G. Prog. Phys. Org. Chem. 1990, 17, 121-
159.
(5) Winstein, S.; Grunwald, E.; Jones, H. W. J. Am. Chem. Soc. 1951, 73,
2700-2707.
(2) Kevill, D. N. AdVances in QuantitatiVe Structure-Property Relationships;
Charton, M., Ed.; JAI Press: Greenwich, CT, 1996; Vol. 1, pp 81-115.
(3) Grunwald, E.; Winstein, S. J. Am. Chem. Soc. 1948, 70, 846-854.
(4) More exactly, eq 1 would hold for any situation where nucleophilic solvent
participation is equal in both the substrate under consideration and the
standard, tert-butyl chloride.
(6) Winstein, S.; Fainberg, A. H.; Grunwald, E. J. Am. Chem. Soc. 1957, 79,
4146-4155.
(7) Peterson, P. E.; Waller, F. J. J. Am. Chem. Soc. 1972, 94, 991-992.
(8) Schadt, F. L.; Bentley, T. W.; Schleyer, P. v. R. J. Am. Chem. Soc. 1976,
98, 7667-7675.
9
5174
J. AM. CHEM. SOC. 2004, 126, 5174-5181
10.1021/ja031828z CCC: $27.50 © 2004 American Chemical Society

Solvent Nucleophilicity
A R T I C L E S
Scheme 2. Abbreviations and Electrophilicity Parameters E of
solvolyses on solvent-ionizing power is m ) 0.3. Substitution
of this value into eq 2 yields eq 3 which has widely been used
for the determination of solvent nucleophilicities.8
Benzhydrylium Ions
NOTs ) log (k/k0)MeOTs - 0.3YOTs
(3)
where k is the rate constant for the solvolysis of methyl tosylate
in the solvent under consideration, and k0 is the rate constant
for the solvolysis of methyl tosylate in 80% aqueous ethanol
(v/v).
While the choice of m ) 0.3 for methyl tosylate solvolyses
was still being debated and other factors, for example, m )
0.55, had been proposed,9 Kevill recognized that substrates with
neutral leaving groups provide a superior tool for determining
solvent nucleophilicity. In analogy to earlier studies by Swain,10
the solvolysis rates of 1-adamantyldimethylsulfonium triflate
were found to be affected by solvent variation by less than a
factor of 7, while the solvolysis rates of 1-adamantyl tosylate
varied by 7 powers of 10 within the same group of solvents.
Consequently, it was concluded that in solvolyses of oxonium
and sulfonium ions, that is, substrates with neutral leaving
groups, the contribution of solvent-ionizing power can be
neglected, and Kevill employed the solvolysis rates of the
S-methyldibenzothiophenium ion (SN2 reactions) as the basis
of the solvent nucleophilicity scale NT (eq 4).11
NT ) log(k/k0)MeDBTh
(4)
+
where k is the rate constant for the solvolysis of the S-
methyldibenzothiophenium ion (MeDBTh+) in the solvent under
consideration, and k0 is the rate constant for the solvolysis of
the S-methyldibenzothiophenium ion (MeDBTh+) in 80%
aqueous ethanol (v/v).
Although the solvent nucleophilicity parameters NOTs and NT
defined by eqs 3 and 4 show fairly good correlations with each
other as well as with solvent nucleophilicities derived from
solvolyses of other methylsulfonium or triethyloxonium ions,2
they are relative parameters, which are not linked to other scales
of nucleophilicity. The ongoing controversy on the role of
nucleophilic solvent participation is highlighted by the title of
a recent J. Am. Chem. Soc. article: “Is the tert-Butyl Chloride
Solvolysis the Most Misunderstood Reaction in Organic Chem-
istry?”.12
The development of fast kinetic methods provides a possibil-
ity for the direct measurement of solvent nucleophilicity. Thus,
carbocations have been generated by laser flash induced
heterolysis of suitable precursors, and the rate of decay of these
carbocations in various solvents reflects solvent nucleophilicity,
decoupled from solvent-ionizing power.13 Similar information
has also been obtained with Jencks’ and Richard’s azide clock
method which derives solvent nucleophilicity from the ability
of solvents to compete with azide ions for solvolytically
generated carbocations.14
We have recently suggested the employment of benz-
hydrylium ions as reference electrophiles (Scheme 2) for the
quantitative comparison of nucleophiles of widely differing
structures and reactivities.15
As described in detail in refs 15 and 16, nucleophilicity
parameters N and s, as defined by eq 5, can be obtained by
plotting log k(20 °C) for the reactions of a certain nucleophile
(9) Kevill, D. N.; Rissmann, T. J. J. Org. Chem. 1985, 50, 3062-3064.
(10) Swain, C. G.; Kaiser, L. E.; Knee, T. E. C. J. Am. Chem. Soc. 1958, 80,
4092-4094.
(11) Kevill, D. N.; Anderson, S. W. J. Org. Chem. 1991, 56, 1845-1850.
(12) Gajewski, J. J. J. Am. Chem. Soc. 2001, 123, 10877-10883.
(13) (a) McClelland, R. A.; Banait, N.; Steenken, S. J. Am. Chem. Soc. 1986,
108, 7023-7027. (b) McClelland, R. A.; Kanagasabapathy, V. M.;
Steenken, S. J. Am. Chem. Soc. 1988, 110, 6913-6914. (c) McClelland,
R. A.; Kanagasabapathy, V. M.; Banait, N. S.; Steenken, S. J. Am. Chem.
Soc. 1989, 111, 3966-3972. (d) Chateauneuf, J. E. J. Chem. Soc., Chem.
Commun. 1991, 1437-1438. (e) Das, P. K. Chem. ReV. 1993, 93, 119-
144. (f) Kirmse, W.; Krzossa, B.; Steenken, S. J. Am. Chem. Soc. 1996,
118, 7473-7477. (g) Kirmse, W.; Guth, M.; Steenken, S. J. Am. Chem.
Soc. 1996, 118, 10838-10849. (h) McClelland, R. A. Tetrahedron 1996,
52, 6823-6858.
(14) (a) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 4689-
4691. (b) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 4691-
4692. (c) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 1373-
1383. (d) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 1383-
1396. (e) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 6,
1396-1401. (f) Ta-Shma, R.; Jencks, W. P. J. Am. Chem. Soc. 1986, 108,
8041-8050. (g) Amyes, T. L.; Richard, J. P.; Novak, M. J. Am. Chem.
Soc. 1992, 114, 8032-8041. (h) Richard, J. P.; Amyes, T. L.; Lin, S.-S.;
O’Donoghue, A. C.; Toteva, M. M.; Tsuji, Y.; Williams, K. B. AdV. Phys.
Org. Chem. 2000, 35, 67-116.
9
J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004 5175

A R T I C L E S
Minegishi et al.
Scheme 3
Table 1. Independence of the First-Order Decay of (thq)2CH+ in
50W50AN of the Concentration of DABCO/DABCO-H+ BF4 (1/1)
-
Buffer Additive
+
+
[(thq)2CH ],
[DABCO], [DABCO−H ],
kobs
,
mol L-1
mol L-1
s-1
6.84 × 10-6
8.57 × 10-6
8.06 × 10-6
7.84 × 10-6
8.39 × 10-6
9.44 × 10-4
2.37 × 10-3
3.33 × 10-3
4.33 × 10-3
5.79 × 10-3
1.24 × 10-3
1.22 × 10-3
1.21 × 10-3
1.22 × 10-3
1.22 × 10-3
average k ) 1.22 × 10-3 s-1
(fur)2CHOH18 (0.87 g, 3.2 mmol) was dissolved in a mixture of 50
mL of dry ether and 2.2 mL (17 mmol) of propionic anhydride. A
54% etheral solution of HBF4‚OEt2 (1.34 g, 8.24 mmol) was added.
The mixture was stirred at room temperature for 1 h and cooled with
an ice bath for 20 min. The precipitate was filtered off under nitrogen
and washed successively with cold, dry ether and cold, dry pentane.
The residue was dried in vacuo to yield 0.79 g of a deep-violet powder
(2.3 mmol, 72%). H NMR (300 MHz, CD2Cl2): δ 3.49 (t, J ) 8.6
Hz, 4H), 5.01 (t, J ) 8.6 Hz, 4H), 7.18 (d, J ) 8.7 Hz, 2H), 8.12 (d,
J ) 8.7 Hz, 2H), 8.27 (s, 2H), 8.72 (s, 1H, Ar2CH).
Kinetics. The benzhydrylium salts used in this study are colored
substances whose absorption maxima have been reported previ-
ously.15,19,20 For the investigation of the reactions of the benzhydrylium
ions with solvents, stock solutions of benzhydrylium tetrafluoroborates
in acetonitrile were prepared, then combined with the nucleophilic
solvent, and the disappearance of the benzhydrylium absorbances was
followed photometrically. For investigations of the nucleophilicities
of trifluoroethanol/water mixtures, stock solutions of benzhydrylium
tetrafluoroborates (E < -3) in trifluoroethanol were employed.
Reactions with τ1/2 > 10 s were monitored by conventional UV-
vis measurements using fiber optics and an immersion probe as
described previously.20-22
For the stopped-flow measurements, the stock solutions of the
benzhydrylium salts in acetonitrile or TFE were mixed with the
nucleophilic solvents in the stopped-flow instrument (Hi-Tech
SF-61DX2 controlled by Hi-Tech KinetAsyst2 software) in ratios of
1/1 or 1/10, and the decay of the benzhydrylium absorbance was
followed as described previously.15,20,23
with a series of electrophiles of Scheme 2 versus the corre-
sponding electrophilicity parameters, E.
1
log k(20 °C) ) s (N + E)
(5)
We have now used this method for characterizing the
nucleophilic reactivities of a series of common solvents and
solvent mixtures, and we will compare the solvent nucleophi-
licities thus obtained to the nucleophilicities of other n-, π-,
and σ-nucleophiles.
For the sake of clarity, nucleophilicity parameters of solvents
and solvent mixtures which refer to first-order rate constants
(k1) are designated as N1. For solvent mixtures where only one
component of the solvent is acting as the nucleophile, they may
be converted into ordinary N values (referring to second-order
rate constants) by subtraction of the logarithm of the molarity
(log [Nuc]) of the nucleophilic component. For the solvent
mixtures investigated in this work, [Nuc] ) 2-55 mol L-1
;
that is, the N values referring to second-order rate constants
are 0.3-1.7 units smaller than N1.
The rates of rapid reactions (τ1/2 < 1 ms) were determined with
laser flash photolytically generated benzhydrylium ions which were
obtained from diarylmethyl 4-cyanophenolates or diarylmethyl acetates
(Scheme 3).13b Irradiation by laser flash in a quartz cell was carried
out with a Continuum PL9010 Nd:YAG laser flash apparatus (λ )
266 nm; power/puls ca. 50 mJ). An Osram XBO 150W xenon lamp
was used as a light source for the detection in the UV-vis region.
All rates were measured at 20 °C ((0.2 °C for stopped-flow and
conventional experiments, (1 °C for laser flash experiments). First-
order rate constants kobs (s-1) were obtained by least-squares fitting
of the absorbance data to the single-exponential curve At ) A0
exp(-kobst) + C. As a consequence of the poor solubility of the
benzhydrylium tetrafluoroborates, benzhydryl 4-cyanophenolates, and
benzhydryl acetates in some of the solvent mixtures, up to 1% of a
cosolvent (CH3CN) was employed.
Most solvent nucleophilicity parameters N1 and s presented
in this work are based on the UV-vis photometric detection of
the decay of the benzhydrylium ion concentrations with three
independent kinetic methods (Scheme 3): conventional UV-
vis spectrometry (τ1/2 > 10 s), stopped-flow (10 s > τ1/2 > 10-3
s), and laser flash techniques (10-3 s > τ1/2 > 10-7 s). The
mutual agreement of the rate constants determined by using
these three methods corroborates the reliability of the present
data.
Experimental Section
Materials. Benzhydrylium tetrafluoroborates,15 benzhydryl 4-cyano-
phenolates,13c,17 and benzhydryl acetates13c,17 were prepared as previ-
ously described.
Water was distilled and passed through a Milli-Q water purification
system. Alternatively, HPLC grade water was distilled before use.
Methanol and ethanol were distilled over CaH2. 2,2,2-Trifluoroethanol
was stored over molecular sieves (3 Å) and distilled over CaSO4.
Acetonitrile was dried with molecular sieves (3 Å) or distilled over
diphenylketene.
Results
When the benzhydrylium ions were exposed to the aqueous
or alcoholic solvent mixtures specified in Tables 1 and 2 as
(18) Kelly, D. P.; Jenkins, M. J. J. Org. Chem. 1984, 49, 409-413.
(19) Bartl, J.; Steenken, S.; Mayr, H.; McClelland, R. A. J. Am. Chem. Soc.
1990, 112, 6918-6928.
Synthesis of Bis(2,3-dihydrobenzofuran-5-yl)methyl Tetrafluo-
roborate (fur)2CH+BF4-. Bis(2,3-dihydrobenzofuran-5-yl)methanol
(20) Minegishi, S.; Mayr, H. J. Am. Chem. Soc. 2003, 125, 286-295.
(21) Mayr, H.; Schneider, R.; Schade, C.; Bartl, J.; Bederke, R. J. Am. Chem.
Soc. 1990, 112, 4446-4454.
(15) Mayr, H.; Bug, T.; Gotta, M. F.; Hering, N.; Irrgang, B.; Janker, B.; Kempf,
B.; Loos, R.; Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am. Chem.
Soc. 2001, 123, 9500-9512.
(16) Mayr, H.; Kempf, B.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66-77.
(17) Pham, T. V.; McClelland, R. A. Can. J. Chem. 2001, 79, 1887-1897.
(22) Dilman, A. D.; Ioffe, S. L.; Mayr, H. J. Org. Chem. 2001, 66, 3196-
3200.
(23) Mayr, H.; Ofial, A. R. Einsichten - Forschung an der LMU Mu¨nchen
2001, 20, 30-33.
9
5176 J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004

Solvent Nucleophilicity
A R T I C L E S
Table 2. First-Order Rate Constants (s-1) for the Reactions of Water/Acetonitrile Mixtures with Benzhydrylium Ions at 20 °Ca
cation
E
W
91W9AN
80W20AN
67W33AN
50W50AN
33W67AN
20W80AN
10W90AN
Ph2CH+
5.90
3.63
2.11
1.48
0.00
-1.36
-3.14
-3.85
-4.72
-5.53
-5.89
-7.02
-7.69
-8.22
1.33 × 109 b
3.28 × 107
1.84 × 106
8.55 × 105
1.04 × 105
(tol)2CH+
(ani)PhCH+
(ani)(tol)CH+
(ani)2CH+
(fur)2CH+
(pfa)2CH+
(mfa)2CH+
(dpa)2CH+
(mor)2CH+
(mpa)2CH+
(dma)2CH+
(pyr)2CH+
(thq)2CH+
3.2 × 107 c
2.0 × 106 c
8.2 × 105 c
1.0 × 105 c
3.2 × 107 c
2.1 × 106 c
9.1 × 105 c
1.3 × 105 c
3.47 × 107
1.91 × 106
8.95 × 105
1.01 × 105
3.06 × 107
1.87 × 106
8.29 × 105
9.82 × 104
2.37 × 107
1.83 × 106
6.81 × 105
9.87 × 104
7.11 × 102
1.42 × 101
1.57
1.9 × 106 d
7.8 × 105 d
9.44 × 104
7.99 × 105
9.55 × 104
9.36 × 101
4.93 × 101
4.96 × 101
3.62 × 101
3.39 × 101
3.32 × 101
4.39
3.98
4.47
3.78e
3.58
3.62
4.44
4.12
3.97
1.56
6.73 × 10-1
4.93 × 10-1
4.13 × 10-1
3.31 × 10-1 e
2.85 × 10-1
2.51 × 10-1
8.03 × 10-2
8.08 × 10-2
3.31 × 10-1
2.84 × 10-1
2.83 × 10-1
2.52 × 10-1
2.24 × 10-1
2.17 × 10-1
2.06 × 10-2
5.57 × 10-3 e
2.20 × 10-3 e
4.29 × 10-3
1.66 × 10-3
3.77 × 10-3
1.23 × 10-3
4.05 × 10-3
1.17 × 10-3
4.33 × 10-3
1.22 × 10-3
1.28 × 10-3
a Compositions of solvents are given as (v/v); W ) water, AN ) acetonitrile. b From ref 13d; because of the proximity of the diffusion limit, this value
was not used for the correlation. c From ref 13c. d From ref 17 at 25 °C. e Reference 20.
Scheme 4
well as in Tables S1-S5 in the Supporting Information, an
exponential decay of the benzhydrylium absorbances was
observed. Probably because of solubility problems, the reactions
of (dpa)2CH+ and (pfa)2CH+ with trifluoroethanol/water ) 20/
80 or 10/90 could not be described by single-exponential
functions, and rate constants for these reactions are not given
violet (E ) -11.3)20 with H2O and D2O does not show a kinetic
isotope effect.25
If more electrophilic benzhydrylium ions are employed, the
rates of deprotonation of the oxonium ions in Scheme 4 can be
assumed to remain almost unaffected, while the reverse reac-
tions, that is, the regeneration of the benzhydrylium ions, must
become slower. Consequently, we can conclude that all reactions
monitored by stopped-flow and laser flash techniques in this
work proceed via rate-determining attack of the solvent nu-
cleophiles at the benzhydrylium ions.
The reactions of benzhydrylium ions of electrophilicity -6
< E < -1 with acetonitrile/water mixtures were determined
with stopped-flow techniques by combining solutions of ben-
zhydrylium tetrafluoroborates in acetonitrile with water or
acetonitrile/water mixtures to yield the solvent mixtures listed
in Table 2. Because the stopped-flow instrument used in this
work does not allow us to employ mixing ratios >10:1, we were
unable to study solvent mixtures with water or alcohol contents
>91%.
Benzhydrylium ions of E g 0 have been generated by 20-ns
laser pulses in the corresponding solvent mixtures, and their
decay has been followed by UV-vis spectroscopy. Some of
these data were already available in the literature as indicated
in Table 2.
Figure 1 shows that plots of rate constants (log k1) determined
by the three different methods versus the empirical electrophi-
licity parameters E give linear correlations. This proves the
consistency of the data obtained by the independent methods
as well as the applicability of eq 5 for describing the reactions
of carbocations with these solvents. The fact that some of the
data depicted in Figure 1 as well as in 35 analogous correlations
(shown in the Supporting Information) have been taken from
the literature is an additional confirmation for the reliability of
our data, and vice versa.
in Table S3. For the same reason, Table 2 (and Table S1) does
not give rate constants for the reactions of (dpa)2CH+ with
aqueous acetonitrile containing more than 50% water.
Laser flash photolysis of the benzhydryl 4-cyanophenolates
or acetates24 results in heterolytic or homolytic cleavage of the
Ar2CH-O bond as shown in Scheme 3. Because the absorption
maxima of the benzhydryl radicals are at considerably lower
wavelengths than those of the benzhydryl cations,19 the con-
comitant formation of both species does not affect the observa-
tion of the exponential decay of the benzhydryl cation absor-
bances during the reactions with the nucleophilic solvent.
As described for acetonitrile/water mixtures in Tables 1 and
2, and for many other solvents and solvent mixtures in Tables
S1-S5 of the Supporting Information, the solvent nucleophi-
licities were examined with benzhydrylium ions of widely
differing reactivity. Generally, rate constants covering a range
of 7-10 powers of 10 were employed for characterizing each
of the solvent systems.
Rate constants for the reactions of acetonitrile/water mixtures
with the weak electrophiles (thq)2CH+ (E ) -8.22) and
(pyr)2CH+ (E ) -7.69) were obtained with a conventional
UV-vis spectrometer by injecting concentrated solutions of the
corresponding benzhydrylium tetrafluoroborates in acetonitrile
into the solvent mixtures under consideration. Table 1 shows
that the rate constants for the first-order decay of (thq)2CH+ in
50W50AN mixtures are independent of the concentrations of
DABCO/DABCO-H+ buffer additives.
Because an analogous behavior has been observed for the
consumption of (pyr)2CH+ in acetonitrile/water and of (ind)2CH+
and (thq)2CH+ in ethanol/water mixtures (see Supporting
Information), it is concluded that the reaction of the benzhy-
drylium ion with the solvent molecule and not the successive
proton transfer is rate-determining (Scheme 4). This conclusion
is in accord with Ritchie’s report that the reaction of crystal
Slopes and intercepts of the correlations in Figure 1 (and the
analogous correlations shown in the Supporting Information)
yield the nucleophilicity parameters N1 and s given in Table 3.
In some cases, carbocations of different electrophilicity were
(24) Most experiments were performed with 4-cyanophenolates. Benzhydrylium
acetates were used in highly aqueous solutions, where the solubility of the
benzhydryl 4-cyanophenolates is low, or in solvents of low nucleophilicity,
where the recombination of the benzhydrylium ions with 4-cyanophenolate
anion is fast as compared to the reaction of Ar2CH+ with the solvent.
(25) Ritchie, C. D.; Skinner, G. A.; Badding, V. G. J. Am. Chem. Soc. 1967,
89, 2063-2071.
9
J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004 5177

A R T I C L E S
Minegishi et al.
Figure 1. Plots of log k1 for the decay of benzhydrylium ions in several solvents versus the electrophilicity parameters E of the benzhydrylium ions (20 °C).
Mixtures of solvents are given as (v/v), solvents: M ) methanol, E ) ethanol, W ) water, T ) trifluoroethanol, AN ) acetonitrile.
Table 3. Nucleophilicity (N1) and Slope (s) Parameters for
Solvents
tion employed, these rate constants were not used for the
calculation of the correlation equations.
solventa
N
1
s
The nucleophilicity parameter for methanol listed in Table 3
can be compared to that derived from Ritchie’s rate constants
for the reactions of methanol with tritylium ions and tropylium
ions, N1 ) 6.02 and s ) 1.01.20 These values, which have
explicitly been labeled as “Approximate Parameters” in Table
7 of ref 20, can now be revised. It should be noted, however,
that our previous guess20 was able to reproduce the rate constants
for the reactions of benzhydrylium ions with methanol reported
here, with an accuracy of a factor of 50. Table S6 in the
Supporting Information shows that the rate constants given in
Table 2 and in Tables S1-S5 can be reproduced by the
reactivity parameters in Table 3 with a standard deviation of a
factor of 1.82.
W
5.20
5.16
5.04
5.05
5.05
5.02
5.02
4.56
1.23
2.93
3.20
3.42
3.57
3.77
4.78
5.04
7.44
7.03
6.68
6.28
5.96
5.81
5.54
5.38
7.10
6.94
6.74
6.37
6.06
5.77
5.19
7.54
7.45
7.20
7.01
6.67
6.38
6.04
5.55
0.89
0.91
0.89
0.90
0.89
0.90
0.89
0.94
0.92
0.88
0.88
0.90
0.89
0.88
0.83
0.90
0.90
0.86
0.85
0.87
0.89
0.90
0.94
0.91
0.90
0.90
0.89
0.90
0.90
0.92
0.96
0.92
0.87
0.89
0.91
0.90
0.92
0.94
0.97
91W9AN
80W20AN
67W33AN
50W50AN
33W67AN
20W80AN
10W90AN
T
90T10W
80T20W
60T40W
50T50W
40T60W
20T80W
10T90W
E
90E10W
80E20W
60E40W
50E50W
40E60W
20E80W
10E90W
91E9AN
80E20AN
67E33AN
50E50AN
33E67AN
20E80AN
10E90AN
M
91M9AN
80M20AN
67M33AN
50M50AN
33M67AN
20M80AN
10M90AN
Discussion
The close similarity of all s parameters in Table 3 implies
that the relative nucleophilicities of these solvents and solvent
mixtures are fairly independent of the carbocation electrophi-
licities. As a consequence, the data set reported in this work
might also be described by constant selectivity relationships of
the Ritchie type.26 However, because the s parameters in Table
3 are considerably larger than those of most other nucleophiles
investigated by Ritchie (s ≈ 0.6 for amines, alkoxides, etc.),20
it is not possible to treat the whole set of nucleophilicity
parameters by Ritchie’s equation.26 We, therefore, base the
following discussion on the nucleophilicity parameters N1 and
s as derived from eq 5.
In agreement with a previous report by McClelland,13c the
nucleophilicities of acetonitrile/water mixtures remain almost
constant when the extent of water exceeds 20% (v/v, Figure
2).
In contrast, the nucleophilicities of methanol/acetonitrile and
of ethanol/acetonitrile mixtures increase steadily with the amount
a Mixtures of solvents are given as (v/v), solvents: M ) methanol, E )
ethanol, W ) water, T ) trifluoroethanol, AN ) acetonitrile.
(26) (a) Ritchie, C. D. Acc. Chem. Res. 1972, 5, 348-354. (b) Ritchie, C. D. J.
Am. Chem. Soc. 1975, 97, 1170-1179. (c) Ritchie, C. D. Can. J. Chem.
1986, 64, 2239-2250.
found to react with equal rates, corresponding to k ) 107 s-1
Because this value appeared to be the limit of the instrumenta-
.
9
5178 J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004

Solvent Nucleophilicity
A R T I C L E S
Figure 2. Dependence of the nucleophilicity N1 on the composition of
mixtures of acetonitrile with water, methanol, or ethanol.
Figure 4. Relationship between nucleophilicity parameters N1 and NT from
reactions with benzhydrylium ions and the S-methyldibenzothiophenium
ion, respectively. Mixtures of solvents are given as (v/v), solvents: M )
methanol, E ) ethanol, W ) water, T ) trifluoroethanol, A ) acetone.
For NT, see ref 32. For N1 of 90A10W and 80A20W, see ref 33. (a) The
nucleophilicities N1 were obtained by interpolating data for aqueous
trifluoroethanol from Table 3.
it has been shown, for example, that both individual rate
constants, kEtOH and kH O, grow by a factor of approximately
2
102 when 5% ethanol/95% water was gradually replaced by 95%
ethanol/5% water.27r The rate constants do not increase uni-
formly, however: While the ratio kEtOH/kH O was close to 1 in
2
95% ethanol as well as in 95% water, kEtOH/kH O reached a broad
2
maximum of 4.6 in 15-50% aqueous ethanol. Because of the
Figure 3. Dependence of the nucleophilicity N1 on the composition of
water/alcohol mixtures.
reported dependence of the selectivities kR′OH/kH O on the
2
medium,27r we have not calculated individual rate constants in
this work.
Scheme 5. Selectivities of Carbocations in Alcohol/Water Mixtures
How do the solvent nucleophilicities N1 derived from the
reactions with benzhydrylium ions (Table 3) compare to those
solvent nucleophilicities previously derived from SN2 reactions
of methyl sulfonium ions and methyl tosylate? The good
correlation shown in Figure 4 demonstrates that nucleophilicities
toward carbocations and toward methyl sulfonium ions (SN2
reactions)2 are closely related, which justifies Kevill’s choice
of methyl sulfonium ions as substrates for determining nucleo-
philic solvent participation in SN1 reactions. Correlation eq 6
indicates that nucleophile variation has a considerably stronger
influence on the reactivities toward carbocations than toward
methyl sulfonium ions.
of alcohol. While pure methanol and pure ethanol possess
approximately the same nucleophilicity, methanol/acetonitrile
mixtures are more nucleophilic (by ca. 0.3 logarithmic units)
than the corresponding ethanol/acetonitrile mixtures. Figure 2
shows a remarkable increase of nucleophilicity from 90%
ethanol/10% acetonitrile to pure ethanol, resulting in almost
equal reactivities of pure methanol and pure ethanol. Because
variation of solvent composition over such wide ranges causes
significant medium effects, we will not discuss the formal
second-order rate constants.
N1 ) 1.51NT + 6.79, n ) 17, R2 ) 0.948
(6)
An analogous conclusion has previously been drawn by Bunting,
Figure 3 shows that the nucleophilicity of ethanol/water
mixtures grows with increasing content of alcohol while the
nucleophilicity of trifluoroethanol/water mixtures decreases with
increasing content of alcohol. The steep decrease from 90%
trifluoroethanol to pure trifluoroethanol is particularly obvious.
Product ratios obtained from solvolytically generated car-
bocations in alcohol/water mixtures (Scheme 5) have been the
topic of extensive investigations27 and have been reviewed by
Ta-Shma and Rappoport.27r
(27) (a) Harris, J. M.; Becker, A.; Clark, D. C.; Fagen, J. F.; Kennan, S. L.
Tetrahedron Lett. 1973, 14, 3813-3816. (b) Pross, A.; Koren, R.
Tetrahedron Lett. 1975, 16, 3613-3616. (c) Ando, T.; Tsukamoto, S.
Tetrahedron Lett. 1977, 18, 2775-2778. (d) Bentley, T. W.; Harris, H. C.
J. Org. Chem. 1988, 53, 724-728. (e) Bentley, T. W.; Koo, I. S. J. Chem.
Soc., Perkin Trans. 2 1989, 1385-1392. (f) Kevill, D. N.; Kyong, J. B.;
Weitl, F. L. J. Org. Chem. 1990, 55, 4304-4311. (g) Bentley, T. W.; Koo,
I. S.; Norman, S. J. J. Org. Chem. 1991, 56, 1604-1609. (h) Shimizu, N.;
Osajima, E.; Tsuno, Y. Bull. Chem. Soc. Jpn. 1991, 64, 1145-1152. (i)
Bentley, T. W.; Shim, C. S. J. Chem. Soc., Perkin Trans. 2 1993, 1659-
1663. (j) Bentley, T. W.; Christl, M.; Kemmer, R.; Llewellyn, G.; Oakley,
J. E. J. Chem. Soc., Perkin Trans. 2 1994, 2531-2518. (k) D’Souza, M.
J.; Kevill, D. N.; Bentley, T. W.; Devaney, A. C. J. Org. Chem. 1995, 60,
1632-1637. (l) Kevill, D. N.; Casamassa, A. J.; D’Souza, M. J. J. Chem.
Res., Synop. 1996, 472-473. (m) Kevill, D. N.; Anderson, S. W.; Ismail,
N. H. J. J. Org. Chem. 1996, 61, 7256-7262. (n) Bentley, T. W.; Llewellyn,
G.; McAlister, J. A. J. Org. Chem. 1996, 61, 7927-7932. (o) Bentley, T.
W.; Ebdon, D.; Llewellyn, G.; Abduljaber, M. H.; Miller, B.; Kevill, D.
N. J. Chem. Soc., Dalton Trans. 1997, 3819-3825. (p) Koo, I. S.; Yang,
K.; Kang, K.; Lee, I.; Bentley, T. W. J. Chem. Soc., Perkin Trans. 2 1998,
1179-1184. (q) Bentley, T. W.; Llewellyn, G.; Zoon, H. R. J. Org. Chem.
1998, 63, 4654-4659. (r) Ta-Shma, R.; Rappoport, Z. AdV. Phys. Org.
Chem. 1992, 27, 239-291 and references therein.
It has been found that the ratio kR′OH/kH O derived from
2
product analysis depends on solvent polarity and generally
decreases slightly from 50% aqueous alcohols to pure alcohols.27r
Combination of the absolute rate constants for the decay of laser
flash solvolytically generated benzhydrylium ions (ani)2CH+
with the product ratios yields the individual rate constants kR′OH
and kH O as defined in Scheme 5. For ethanol/water mixtures,
2
9
J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004 5179

A R T I C L E S
Minegishi et al.
Table 4. Approximate Solvent Nucleophilicity Parameters (N1) for
Solvent Mixtures Calculated by Eq 6
who found a linear correlation between the aminolysis rates of
methyl 4-nitrobenzenesulfonate (SN2 reaction) and the corre-
sponding rates of amine additions to the 1-methyl-4-vinylpy-
ridinium ion with a slope of 2.27 ()1/0.44).28 Bunting’s and
Richard’s reports of linear correlations between Ritchie’s N+-
parameters (nucleophilicities toward carbocations)26 and Swain’s
and Scott’s n-parameters (nucleophilicities toward CH3Br)29 with
slopes of 2 also indicate that variations of nucleophiles affect
b
c
solventa
N
T
N
1
70E30W
30E70W
95A5W
-0.20
-0.93
-0.49
-0.42
-0.52
-0.70
-0.83
-0.96
-1.11
-1.23
-0.46
-0.37
-0.54
-0.66
-0.84
-1.12
-3.30
-2.19
-1.76
-0.94
-0.64
-0.34
0.08
6.5d
5.7d
6.1
6.2
6.0
5.7
5.5
5.3
5.1
4.9
6.1
6.2
6.0
5.8
5.5
5.1
1.8
3.5
4.1
5.4
5.8
6.3
6.9
-1.2
1.0
2.4
3.0
3.1
4.1
70A30W
60A40W
50A50W
40A60W
30A70W
20A80W
10A90W
80D20W
70D30W
60D40W
50D50W
40D60W
20D80W
97T3We
80T20We
80T20E
2
reactivities toward electrophilic Csp centers to a larger extent
than toward electrophilic Csp centers.28,30 In accordance with
3
these findings, l-values (according to eq 2) greater than 1 have
been found for solvolyses of chloroformate when nucleophilic
addition to the carbonyl group was rate-determining.31
Because linear correlations between NT, N′T, NOTs, and N′OTs
have previously been reported,2 it is not surprising that the
nucleophilicity parameters N1 derived from reactions with
benzhydrylium ions in this work also correlate linearly with
N′T, NOTs, and N′OTs (Figures S1-S3, Supporting Information).
Because of the good correlation between N1 and NT shown
in Figure 4, we suggest the employment of eq 6 for estimating
N1 of further solvent mixtures from reported NT values (Table
4). Although the data obtained in this way have to be considered
as approximate, they are most useful for designing syntheses
in these solvents.
60T40E
50T50E
40T60E
20T80E
97H3We
90H10We
70H30We
50H50We
HCO2H
-5.26
-3.84
-2.94
-2.49
-2.44
-1.78
The linear correlation shown in Figure 4 implies that the
solvent nucleophilicities N1 toward carbocations reported in this
work are controlled by the same factors as the solvent nucleo-
philicities NT toward methylsulfonium ions. A major advantage
of the new parameters N1 is that they can be combined with
the electrophilicity parameters E of carbocations, using eq 5,
so as to estimate absolute lifetimes of carbocations which are
produced solvolytically in aqueous or alcoholic solutions.33
CH3CO2H
a Unless otherwise stated, mixtures of solvents are given as (v/v),
solvents: M ) methanol, E ) ethanol, W ) water, T ) trifluoroethanol,
A ) acetone, D ) dioxane, H ) hexafluoro-2-propanol. b From ref 2. c A
slope parameter of s ) 0.9 is recommended for these solvents (compare s
parameters in Table 3). d Not by using eq 6, but by interpolating data for
aqueous ethanol from Table 3. e Mixtures of solvents are given as (w/w).
The flattening of the log k versus E correlations for k > 108
L mol-1 s-1 also prevents the calculation of the exact point,
where the enforced change from SN1 to SN2 mechanisms is
taking place. According to Jencks, this point is related to the
lifetimes of intermediates, which cannot be shorter than the
duration of a bond vibration (ca. 10-13 s).35 Having in mind
the flattening of the correlation lines for s(E + N) > 8,34 one
can only derive a lower limit and conclude that solvolysis with
s(E + N) < 13 will not proceed via enforced SN2 type
mechanisms.
Because s ≈ 0.9 for all solvents investigated (Table 3), eq 5
predicts that carbocations have a half-life τ1/2 > 10-10 s-1 if E
+ N1 < 11. As 10-10 s-1 is the time needed for solvent
reorganization, one can conclude that in 80% aqueous ethanol
(N1 ) 6.68), carbocations with E < 4.5 will be thermally
equilibrated, while in trifluoroethanol (N1 ) 1.23), thermal
equilibration will already be reached for carbocations of E <
10. Typical E values for carbocations are ca. 8.5 (for (CH3)3C+),34
5.9 (for Ph2CH+),15 and 0.5 (for Ph3C+).20 It should be noted,
however, that eq 5 has been reported to be limited to second-
order rate constants <108 L mol-1 s-1. For faster reactions, the
magnitude of the rate constants will be overestimated because
of the flattening of the correlation curves.34 As a consequence,
thermal equilibration of the intermediate carbocations may also
occur if the sum (E + N1) is slightly higher than 11.
A further advantage of the N1 parameters listed in Table 3 is
their direct comparability to the previously published N param-
eters of π-systems15,16,36-39 and hydride donors.15,40-43 Thus, it
becomes possible to predict nucleophiles which can intercept
(35) (a) Jencks, W. P. Acc. Chem. Res. 1980, 13, 161-169. (b) Jencks, W. P.
Chem. Soc. ReV. 1981, 10, 345-375 (c) Jencks, W. P. Chem. ReV. 1985,
85, 511-527.
(28) Bunting, J. W.; Mason, J. M.; Heo, C. K. M. J. Chem. Soc., Perkin Trans.
2 1994, 2291-2230.
(29) Swain, C. G.; Scott, C. B. J. Am. Chem. Soc. 1953, 75, 141-147.
(30) Richard, J. P.; Toteva, M. M.; Crugeiras, J. J. Am. Chem. Soc. 2000, 122,
1664-1674.
(36) Mayr, H.; Patz, M. Angew. Chem. 1994, 106, 990-1010; Angew. Chem.,
Int. Ed. Engl. 1994, 33, 938-957.
(37) Mayr, H.; Kuhn, O.; Gotta, M. F.; Patz, M. J. Phys. Org. Chem. 1998, 11,
642-654.
(31) (a) Kevill, D. N.; D’Souza, M. J. J. Chem. Soc., Perkin Trans. 2 1997,
1721-1724. (b) Kevill, D. N.; D’Souza, M. J. Can. J. Chem. 1999, 77,
1118-1122.
(38) Mayr, H.; Patz, M.; Gotta, M. F.; Ofial, A. R. Pure Appl. Chem. 1998, 70,
1993-2000.
(32) Nucleophilicities NT of solvent mixtures (v/v) from ref 2: 0.37 (E), 0.17
(M), 0.16 (90E10W), 0 (80E20W), -0.35 (90A10W), -0.37 (80A20W),
-0.39 (60E40W), -0.58 (50E50W), -0.74 (40E60W), -1.16 (20E80W),
-1.31 (10E90W), -1.38 (W), -1.71 (42T58W ) 50T50W, w/w), -1.85
(52T48W ) 60T40W, w/w), -1.98 (63T37W ) 70T30W, w/w), -2.25
(87T13W ) 90T10W, w/w), -3.93 (T).
(39) (a) Kempf, B.; Hampel, N.; Ofial, A. R.; Mayr, H. Chem.-Eur. J. 2003, 9,
2209-2218. (b) Bug, T.; Hartnagel, M.; Schlierf, C.; Mayr, H. Chem.-
Eur. J. 2003, 9, 4068-4076.
(40) Mayr, H.; Basso, N.; Hagen, G. J. Am. Chem. Soc. 1992, 114, 3060-
3066.
(41) Mayr, H.; Basso, N. Angew. Chem. 1992, 104, 1103-1105; Angew. Chem.,
(33) Denegri, B.; Minegishi, S.; Kronja, O.; Mayr, H. Angew. Chem. 2004, 116,
im Druck; Angew. Chem., Int. Ed. 2004, 43, in print.
Int. Ed. Engl. 1992, 31, 1046-1048.
(42) Funke, M.-A.; Mayr, H. Chem.-Eur. J. 1997, 3, 1214-1222.
(43) Mayr, H.; Lang, G.; Ofial, A. R. J. Am. Chem. Soc. 2002, 124, 4076-
4083.
(34) Roth, M.; Mayr, H. Angew. Chem. 1995, 107, 2428-2430; Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2250-2252.
9
5180 J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004

Solvent Nucleophilicity
A R T I C L E S
solvent polarity (chloroform, dichloromethane, acetonitrile,
nitromethane),21,36,40,45 there is evidence that π-nucleophilicity
increases somewhat in protic solvents.44 As a consequence, it
is not only possible to trap carbocations by π-nucleophiles and
hydride donors which are located above the corresponding
reaction media in Figure 5 but also by those located slightly
below the corresponding solvents. In agreement with this
conclusion, Richard had previously reported that carbocations
which are solvolytically generated in 50% aqueous acetonitrile
(N1 ) 5.05) can be trapped by π-nucleophiles which possess N
parameters greater than 6 even when they are used in lower
concentrations (0.01-0.1 mol L-1).44 Kitagawa’s trapping of
fullerenyl cations (E ca. 7-8)46 by anisole (N ) -1.18, s )
1.20)16 in 9/1 (v/v) anisole/trifluoroethanol must be due to the
high s parameter of this π-nucleophile which becomes important
in fast reactions.
As in previous papers dealing with the application of eq 5, it
should be reminded that the rate constants predicted by eq 5
are usually accurate within a factor of 10-100 if systems with
strong steric shielding (e.g., tritylium ions) or systems which
are affected by anomeric effects (e.g., reactions of alkoxycar-
benium ions with alcohols) are excluded. With these exceptions
in mind, Figure 5 is a useful guide for designing syntheses via
solvolytically generated carbocations. Synthetic investigations
in the Mu¨nchen group have shown that it is possible to base
acid free Friedel-Crafts chemistry on Figure 5.47
Figure 5. Comparison of the nucleophilicity parameters N1 of solvents to
the N parameters of typical π-systems and hydride donors. Mixtures of
solvents are given as (v/v), solvents: M ) methanol, E ) ethanol, W )
water, T ) trifluoroethanol, AN ) acetonitrile.
solvolytically generated carbocations in alcoholic or aqueous
solutions as previously determined for one example by Rich-
ard.44 If the nucleophiles on the left side of Figure 5 are
employed in concentrations of [Nuc] ) 1 mol L-1, the calculated
second-order rate constants become numerically identical to the
pseudo-first-order rate constants (k1Ψ ) k2[Nuc]) with the
consequence that the N and N1 parameters on both sides of
Figure 5 become directly comparable.
Acknowledgment. This paper is dedicated to Professor Reiner
Sustmann on the occasion of his 65th birthday. Financial support
by the Deutsche Forschungsgemeinschaft (Ma 673/20-1) and
the Fonds der Chemischen Industrie is gratefully acknowledged.
We thank Professsor Dennis Kevill and Dr. Armin Ofial for
helpful discussions and Dr. Masahi Kotani and Mr. Masanori
Sakamoto (both Fukuoka) for the instruction on how to use laser
flash photolysis instruments.
If the s parameters are neglected in a first approximation, 1
M solutions of the nucleophiles on the left of Figure 5 may be
expected to react equally fast with carbocations as the solvents
on the right of Figure 5. As a consequence, carbocations may
be trapped by 1 M solutions of nucleophiles of N > 1.2 if
trifluoroethanol is used as the solvent and by nucleophiles of N
> 7.5 if they are generated in methanol. This analysis is not
perfectly correct, however, because it neglects the solvent
dependence of the N parameters. Although we have shown that
the rates of the reactions of carbocations with neutral π-nu-
cleophiles and neutral hydride donors only slightly depend on
Supporting Information Available: Details of the kinetic
experiments, determination of the nucleophilicity parameters,
as well as correlations of N1 with N′T, N′OTs, and NOTs (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA031828Z
(45) (a) Kane-Maguire, L. A. P.; Mansfield, C. A. J. Chem. Soc., Dalton Trans.
1976, 2192-2196. (b) John, G. R.; Mansfield, C. A.; Kane-Maguire, L.
A. P. J. Chem. Soc., Dalton Trans. 1977, 574-578. (c) John, G. R.; Kane-
Maguire, L. A. P. J. Chem. Soc., Dalton Trans. 1979, 873-878. (d) John,
G. R.; Kane-Maguire, L. A. P. Inorg. Chim. Acta 1981, 48, 179-183.
(46) Kitagawa, T.; Lee, Y.; Hanamura, M.; Sakamoto, H.; Konno, H.; Takeuchi,
K.; Komatsu, K. Chem. Commun. 2002, 3062-3063.
(44) Richard, J. P.; Szymanski, P.; Williams, K. B. J. Am. Chem. Soc. 1998,
120, 10372-10378.
(47) Mayr, H.; Hofmann, M.; Minegishi, S.; Hampel, N., unpublished results.
9
J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004 5181