Mechanism of the lithium-iodine exchange in an iodothiophene

Article abstract of DOI:10.1139/v05-173

Solutions of 2-lithio-5-methylthiophene (4) were characterized using DNMR techniques and shown to be a mixture of monomer and dimer in THF-Et₂O (3:2). The hypervalent iodine ate complex 5 (Ar₂I^-Li ⁺), a presumed intermediate in the Li-I exchange with 2-iodo-5-methylthiophene, was observed by ¹³C and ⁷Li NMR spectroscopy at low temperatures (-130°C). At higher temperatures, the ate complex coalesced with 2-lithio-5-methylthiophene. A kinetic scheme was developed, which accounts for the exchange of the monomer 4M, dimer 4D, and 2-iodo-5-methylthiophene (6) with the ate complex 5. The rates of the various exchanges were obtained through a DNMR analysis of the variable temperature ¹³C and ⁷Li NMR spectra, and the thermodynamic and activation parameters were calculated. The monomer 4M and the ate complex 5 have similar reactivity as aryl donors in the Li-I exchange reaction, but 4M is at least 1000 times as reactive as the dimer 4D towards the iodide.

Full text of DOI:10.1139/v05-173

1577
Mechanism of the lithium–iodine exchange in an
iodothiophene1
Hans J. Reich and Wesley L. Whipple
Abstract: Solutions of 2-lithio-5-methylthiophene (4) were characterized using DNMR techniques and shown to be a
mixture of monomer and dimer in THF–Et2O (3:2). The hypervalent iodine ate complex 5 (Ar2I–Li+), a presumed inter-
7
mediate in the Li–I exchange with 2-iodo-5-methylthiophene, was observed by 13C and Li NMR spectroscopy at low
temperatures (–130 °C). At higher temperatures, the ate complex coalesced with 2-lithio-5-methylthiophene. A kinetic
scheme was developed, which accounts for the exchange of the monomer 4M, dimer 4D, and 2-iodo-5-methylthiophene
(6) with the ate complex 5. The rates of the various exchanges were obtained through a DNMR analysis of the variable
7
temperature 13C and Li NMR spectra, and the thermodynamic and activation parameters were calculated. The mono-
mer 4M and the ate complex 5 have similar reactivity as aryl donors in the Li–I exchange reaction, but 4M is at least
1000 times as reactive as the dimer 4D towards the iodide.
Key words: halogen–metal exchange, lithium iodinate, iodine ate complex, lithium reagent, aggregate reactivity.
Résumé : On a caractérisé des solutions de 2-lithio-5-méthylthiophène (4) en faisant appel à des techniques de RMN
dynamique (RMND) et on a pu démontrer que, dans le THF–Et2O (3 : 2), elles comprennent un monomère et un dimère.
7
Faisant appel à la RMN du 13C et du Li à basse température (–130 °C), on a pu observer le complexe 5 de l’iodate
hypervalent (Ar2I–Li+), un intermédiaire supposé dans l’échange Li–I avec le 2-iodo-5-méthylthiophène. À des tempéra-
tures supérieures, il y a coalescence du complexe are avec le 2-lithio-5-méthylthiophène. On a développé un schéma ci-
nétique qui permet de tenir compte de l’échange du monomère 4M, du dimère 4D et du 2-iodo-5-méthylthiophène (6)
avec le complexe ate 5. On a obtenu les vitesses des divers échanges par une analyse RMN des spectres RMN du 13C
7
et du Li à des températures variables et on en a calculé les paramètres thermodynamiques et d’activation. Vis-à-vis
des donneurs aryles, dans une réaction d’échange Li–I, les réactivités du monomère 4M et du complexe ate 5 sont
semblables; toutefois, vis-à-vis de l’iode, le monomère 4M est au moins mille fois plus réactif que le dimère 4D.
Mots clés : échange halogène–métal, iodate de lithium, complexe ate de l’iode, réactif de lithium, réactivité d’un agrégat.
[Traduit par la Rédaction] Reich and Whipple 1587
Introduction
also been detected as intermediates in the Li–Mg exchange
(14). Both experimental (7e, 15) and computational (3d, 8)
studies have addressed the ease of formation of various ate
complexes. The relative rates of the lithium–metalloid ex-
changes follow the trend: I > Te >> Sn >> Br > Se >> S.
In previous reports of the lithium–iodine exchange of
phenyllithium (1) with iodobenzene (2) we have kinetically
and spectroscopically identified the stable hypervalent ate
complex 3 as present in solution (7b, 7c, 7e). Direct involve-
ment of 3 in the exchange process could not be proven, since
analysis of the kinetic processes was complicated by the un-
favorable equilibrium between phenyllithium monomer and
dimer, and the degenerate exchange between 2 and 3.
The lithium–metalloid exchange, independently discov-
ered by Gilman et al. (1) and Wittig et al. (2) is an effective
route for the formation of many organolithium reagents inac-
cessible by other methods. Understanding the mechanism of
this process provides greater insight and predictability into
the factors that influence reactivity. The reaction has com-
manded substantial research interest over the past 60 years
(3a), where four-centered (4), SET (3a–3c, 5), SN2 (6), and
ate complex mechanisms (7a, 8) have been suggested. Inves-
tigation of the lithium–metalloid exchange identified ate
complexes as permissible intermediates, including those of I
(7b, 7c, 9), Te (7c–7e, 10a, 10b), Se (7d, 7f, 10b), Sn (7g,
11, 12), Si (13), and Hg (7b, 7c). Iodine ate complexes have
Li
-
Li
+
I
I
1
2
3
Received 25 April 2005. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 1 December 2005.
We report here our results on the solution structure of 5-
methyl-2-thienyllithium (4) and its behavior in the lithium–
iodine exchange using low-temperature, multinuclear NMR
techniques. The use of 4, with its favorable NMR properties,
ease of handling, and considerably higher monomer to dimer
ratio compared to PhLi, has allowed measurement of the
rates and equilibria of exchange between monomer 4M and
H.J. Reich2 and W.L. Whipple. Department of Chemistry,
University of Wisconsin, 1101 University Avenue, Madison,
Wisconsin 53706, USA.
1This article is part of a Special Issue dedicated to organic
reaction mechanisms.
2Corresponding author (e-mail: reich@chem.wisc.edu).
Can. J. Chem. 83: 1577–1587 (2005)
doi: 10.1139/V05-173
© 2005 NRC Canada

1578
Can. J. Chem. Vol. 83, 2005
Fig. 1. 13C NMR spectra in THF–ether (3:2). (A) 0.10 mol/L of
4 at –133 °C (the inset showing the C-2 signals is from a
0.2 mol/L Li-labeled sample, where the dimer C-2 signal was
processed by multiplying the FID with a Gaussian function be-
fore Fourier transformation (LB = –6.1, GB = 0.1). The coupling
constants are believed accurate to 0.2 Hz (D = 4D, M = 4M).
(B) A 1:1 ratio of 4 and 6 with 9 equiv. of HMPA (T = 2-
methylthiophene, protonation product). (C) The iodide 6.
dimer 4D (eq. [1]), monomer and ate complex 5 (eq. [2]),
and ate complex and 2-iodo-5-methylthiophene (6) (eq. [3]).
It has also allowed an estimate of the maximum reactivity of
the dimer with the iodide in the Li–I exchange (eq. [4]).
6
k1
Li
[1]
2
Li
Li
S
S
S
k-1
4M
4D
128.2
137.4
13C NMR Spectra
THF–ether (3:2)
∆H1 = 6.4 0.3 kcal/mol
∆H = 5.3 0.3 kcal/mol
∆S1 = -2.3 2 eu
∆S 13.6 2 eu
[4D]
[4M]2
3
K1
=
73.2
146.3
4
=
C
B
I
1
1
S
5
∆H1 = 0.92 0.06 kcal/mol
∆S1 = 11.2 0.3 eu
6
C-2 δ 73.2
5
4
3
-120 °C
128.7
Li+
Li
4 (0.13 mol/L)
6 (0.13 mol/L)
126.1
k2
-
-
[2]
I
+
I
138.5
150.5
S
Li
I
S
S
S
k-2
THF–ether (3:2)
9 equiv. HMPA
S
S
T
T T
T
5
4M
6
5
2
∆H2 = 1.50 0.4 kcal/mol
∆H = 10.4 0.4 kcal/mol
∆S2 = -21.9 2 eu
∆S = 21.8 2 eu
[5]
1
6
1
6Li-C
J
= 15.3 Hz
J
Li-C = 7.8 Hz
K2
=
174.5
2
2
[4M] [6]
δLi 1.59
δ
Li 1.27
M-2
S
S
Li
D-2
Li
182.3
143.0
126.6
∆H2 = -10.2 0.3 kcal/mol
∆S2 = -48.6 1.5 eu
139.7
Li
S
132.6
126.4
136.8
4M
4D
Li+
182 180 178 176 174
-
I
A
I
4 (0.10 mol/L)
-133 °C
S
k3
D-5
M-5
D-3
M-3
D-4
M-4
S
S
5
+
[3]
6
+
k
3
Li+
-
I
152 150 148 146 144 142 140 138 136 134 132 130 128 126 124
ppm
I
S
S
S
5
6
∆H3 = 3.70 0.2 kcal/mol
∆H-3 = 3.70 0.2 kcal/mol
∆S3 = -7.8 1 eu
∆S-3 = -7.8 1 eu
2. Solutions with a 1:1 ratio of lithium reagent 4 and io-
dide 6 with excess HMPA present (Fig. 1B). In such so-
lutions K2 becomes very large, leaving 5 as the only
detectable species (7a, 7b) and allowing its NMR identi-
fication and characterization.
Li
3. Solutions with an excess of lithium reagent over iodide
Li
Li
S
I
S
S
4M
+
7
(Fig. 5). Line-shape fitting of the Li NMR spectra of
k4
4D
+
[4]
these samples allow estimation of k2 at each tempera-
ture, using as input the values of K1, K2, k1, and k3 ob-
tained from samples 1 and 4.
Li+
k
4
-
I
S
6
S
S
5
4. Solutions with an excess of iodide 6 over lithium re-
agent 4 (Fig. 6). In these solutions 4M, 5, 6, and, at
temperatures above –85 °C, 4D, are in rapid exchange
and it is possible to measure the temperature depend-
ence of the ate complex association constant K2 from
analysis of the averaged peak positions in the 13C NMR
spectra, using K1 obtained from samples 1 and the
chemical shifts of the individual species. Line-shape
analysis of these samples also allowed measurement of
k3 and an estimate of k4.
Results and discussion
The NMR study we set out to perform used four kinds of
samples, all in THF–ether (3:2) solution (this mixed solvent
system is close in donor strength to that of pure THF and al-
lows the measurement of NMR spectra down to –135 °C). In
each case variable temperature Li, Li, and 13C NMR spec-
6
7
tra were measured. They are:
1. Solutions of the lithium reagent 4 to define the aggre-
gates (4M and 4D) present (Figs. 1A and 2). The
concentration and temperature dependence of the equi-
librium constant (K1) and the temperature dependence
of the exchange rates k1 and k–1 were determined from
these samples (Figs. 3 and 4).
The solution structure of 2-lithio-5-methylthiophene
Previous studies of 2-thienyllithium reagents have in-
cluded X-ray structures of the ether solvated tetramer of 4
(7h), several substituted lithiothiophenes,3 and 2-lithiobenzo-
thiophene (16a), as well as a report of the solution 13C NMR
spectrum of 2-lithiothiophene itself (17). We have recently
3 Several single crystal X-ray structures have been reported in the Cambridge Structural Database as private communications (1999). 3-
Dimethylaminomethyl-2-lithiothiophene dimer (LEDSES): A.L. Spek, M.T. Lakin, and R. den Besten. 3-Methoxy-2-lithiothiophene
(HIGPOC): A.L. Spek and N. Veldman. 3-Bromo-2-lithiothiophene TMEDA dimer (HIGXUQ): A.L. Spek. 2-Lithiothiophene dimer
(JUJXER): A.L. Spek and W.J.J. Smeets.
© 2005 NRC Canada

Reich and Whipple
1579
Fig. 4. Temperature dependence of the free energy of activation
Fig. 2. Selected 13C and Li NMR spectra of a variable tempera-
ture study of 0.101 mol/L 5-methyl-2-thienyllithium (4) in THF–
ether (3:2). These spectra were simulated to obtain k1, k–1, and
K1 (eq. [1]). Monomer (M), dimer (D).
7
(calculated from a form of the Eyring equation, ∆G‡ = RT[23.76 +
ln(T/kr)] = ∆H‡ – T∆S‡) for the monomer–dimer exchange (k1 (᭹)
from line-shape analysis of the spectra in Fig. 2, eq. [1]), ate
complex formation (k2 (᭡) from Fig. 6, eq. [2]), and ate com-
plex – iodide exchange (k3 (ٗ) from Fig. 7, eq. [3]).
7Li NMR
13C NMR
T (°C)
-31
2 4M
4D
9.0
8.0
7.0
6.0
5.0
4.0
9.0
8.0
7.0
6.0
5.0
4.0
k-1
-52
-62
k1
k-2
k2
G
G
k-3
-80
k3
-91
130 150 170 190 210 230 250
T (K)
130 150 170 190 210 230 250
T (K)
-102
-120
7
Fig. 5. Variable temperature (A) 13C and (B) Li NMR spectra
of a solution of 4 (0.117 mol/L) and 6 (0.046 mol/L) in THF–
ether (3:2).
M-2 δ 182.3
D-2 δ 174.5
D-5
D-3
M-3
D-4 M-4
M-5
-133
13C NMR
4 (0.117 mol/L)
6 (0.046 mol/L)
2 4M
4M
4D
7Li NMR
A
T (°C)
B
+
6
5
146 144 142 140 138 136 134 132 130 128 126 124
ppm
2
1
ppm
-70.4
-80.5
5
3
4
Fig. 3. (A) Concentration dependence of the 4M and 4D ratio in
THF–ether (3:2) at –121 °C. (B) Temperature dependence of free
energy (∆G = RT ln(Keq)) for the monomer dimer equilibrium
(K1, eq. [1]) and the ate complex equilibrium (K2, eq. [2]).
-85.9
A
1.0
0.5
B
-1.2
-1.4
-1.6
-1.8
-2.0
-2.2
-2.4
Excess Lithium Reagent
-91.7
-97.3
Slope: 1.9 0.13
K1 = 12.6 L/mol
-121 °C
0.0
K1
-0.5
-1.0
-1.5
-2.0
-2.5
K2
G
K2 (Li)
-102.7
-1.8 -1.7 -1.6 -1.5 -1.4 -1.3 -1.2 -1.1
log [M]
120 140 160 180 200 220 240
T (K)
-108.3
-110.9
-113.7
4M
5
reported detailed NMR studies of a series of lithiothio-
phenes, including some with ortho chelating groups (7i).
Clean solutions of 4 were readily available since the pre-
cursor 2-methylthiophene is easily metalated with n-BuLi
(17, 18), and the pure reagent can be crystallized from THF
or ether (7h). Figure 1A reports the 13C NMR spectrum of
4M
5
4D
4D
5
3
4
-118.6
-124.7
5
2
5
3
4
6
95.5% Li-labeled 4 (19) in THF–ether (3:2) at –121 °C.
There are two sets of carbon signals, corresponding to two
distinct lithium species (19a). Each showed fully resolved
13C–6Li scalar coupling of the ipso carbon signals at
182.3 ppm (1:1:1 triplet, 1J = 15.3 Hz) and 174.5 ppm
4D 4M
5
4M
5
3
4
-130.9
3
2
1
0
-1
150
145
140
135
130
125
1
(1:2:3:2:1 pentet, J = 7.8 Hz), respectively (19b). The as-
ppm
ppm
signment of the two species as monomer 4M and dimer 4D
derives from the lithium–carbon coupling, as well as from a
variable concentration study, which showed that the two spe-
cies differ in aggregation state by a factor of two (slope of
experiments in various mixtures of THF and ether showed
that the dimer signals increasingly predominated as the frac-
tion of ether increased. The carbon assignments in the 13C
NMR spectra were based on identification of C-2 from the
the log[M] vs. log[D] plot was 1.9
0.13, K1 =
12.6 (mol/L)–1 at –128 °C, Fig. 3a) (7j, 7k, 16b). In addition,
© 2005 NRC Canada

1580
Can. J. Chem. Vol. 83, 2005
Fig. 6. Variable temperature NMR spectra of solutions of 4 and
6 in THF–ether (3:2) with excess iodide. The two panels are
from experiments with slightly different concentrations. (A) 13C
NMR spectra. The dotted spectra are simulations of the C-3 res-
3 signals), and hence the thermodynamic parameters for the
dimer–monomer equilibrium (Fig. 3B).
Lithium–iodine exchange of 2-lithio 5-methyl-thiophene
Solutions containing both 4 and 5 show complex DNMR
behavior because of the presence of four species (4M, 4D, 5,
and 6) whose relative concentrations are strongly tempera-
ture dependent and which are all undergoing mutual ex-
change involving at least three different rate constants (k1,
k2, and k3). Figure 1 shows spectra of 4 and 6 and a 1:1 mix-
ture of the two, the latter with 9 equiv. of hexamethylphos-
phoric triamide (HMPA) present, which has the effect of
greatly increasing K2 (7a, 7b). The presence of a new spe-
cies in Fig 1B, the ate complex 5, can be readily seen, since
the 13C signals for C-3, C-4, and C-5 are upfield of all three
constituents, and so cannot simply be a dynamic average of
them.
7
onance, as described in the text. (B) Li NMR spectra. The in-
sets are vertical expansions showing the build-up and eventual
coalescence of 4D with the average of the other signals as tem-
perature increased.
13C NMR
4 (0.128 mol/L)
6 (0.166 mol/L)
2 4M
4D
B
T (°C)
7Li NMR
4 (0.123 mol/L)
6 (0.156 mol/L)
A
4M
5 + 6
+
6
5
T (°C)
-50
5* + 6*
-59
Excess Iodide
5
3
4
-64
-73
-61
-69
-76
-81
-87
-79
-90
-99
Examination of samples in THF–ether (3:2) with varying
ratios of iodide and lithium reagent revealed that there were
two limiting behaviors. With excess lithium reagent the sig-
nals of three of the components present (4D, 4M, and 5)
could be decoalesced. There is not a detectable amount of 6
at the lower temperature limit of the solvent system (Fig. 5).
On the other hand, when iodide was in excess there was
(with a minor exception) a single, strongly temperature de-
-92
-98
7
pendent set of 13C and Li signals throughout the accessible
64x
-102
-107
temperature range (Fig. 6).
The spectra of Fig. 5 can be interpreted as follows. At the
lowest temperatures the signals of the ate complex 5 are
decoalesced from those of the monomer 4M and dimer 4D.
The formation constant (K2) of 5 is high, and since there is
excess lithium reagent, there is no detectable iodide 6. Be-
tween –130 and –115 °C the signals of 5 and 4M coalesce
(eq. [2]). The signals of 4D are not appreciably affected by
this process. Between –110 and –90 °C the 4D signals co-
alesce with the averaged signals of 5 and 4M. Starting at ca.
–110 °C, the shift of the equilibrium of eq. [2] (K2) away
from ate complex toward the iodide and lithium reagent be-
comes apparent as indicated by the downfield shift of the
three visible signals (the C-3, C-4, and C-5 signals of 5 are
all upfield of those of 4M, 4D, and 6).
Figure 6 cannot be interpreted with only the eq. [1] and
[2] processes, since the lowest temperature spectrum should
show excess iodide 6 and the ate complex 5 in slow ex-
change, just as was seen for excess 4 in Fig. 5 (the process
k2 is slow on the NMR timescale at this temperature). We
propose that here, as seen for the PhLi–PhI system (7b, 7c),
there is another exchange process that averages the signals
of 5 and 6. The ate complex is behaving as a carbanion do-
nor, and transferring an aryl group to the iodide to form a
new ate complex (k3, eq. [3]). When there is a stoichiometric
excess of lithium reagent, there is no corresponding process
that can average the lithium reagent and ate complex signals,
and so the signals can be decoalesced (Fig. 5).
4D
C-5
C-3
C-4
-117
6
4D
4M
5
4D 4M
5
144 142 140 138 136 134 132 130 128 126
ppm
2
0
ppm
Li–C coupling, C-5 as a quaternary carbon by an INEPT ex-
periment, on the literature assignments for 2-lithiothiophene
(17), and on the correlation between equivalent carbons of
4M, 4D, 5, and 6 provided by the coalescence experiments
(Figs. 2, 5, and 6). Thus, the carbons of 6 could be readily
assigned from shift parameters (20), and this allowed carbon
signal assignments of all of the other compounds.
A variable temperature experiment of 4 is shown in Fig. 2.
7
Each pair of 13C and Li NMR signals for 4M and 4D co-
alesced between –120 and –91 °C. Complete line-shape sim-
ulations of these spectra were used to determine the rates of
interconversion for monomer and dimer. The relationship be-
tween the pseudo-first-order NMR rate constants (kMD and
kDM) obtained from the simulation and the mechanistic rate
constants k1 and k–1 of eq. [1] is straightforward: k–1 is also a
first-order rate constant (it is best to think of a first-order
rate as the fraction of substrate converted per unit time), so
k–1 = kDM. The monomer to dimer conversion k1 is a second-
order rate constant (fractional conversion per unit time per
molar concentration of reactant), and thus kMD = k1[4M].
The activation parameters in eq. [1] were calculated from the
temperature dependence of k1 and k–1 (see Fig. 4). The simu-
lation also gave the concentration of monomer and dimer as
a function of temperature (for the four low temperature
points from the areas of the two C-3 peaks, for the five high
temperature points from the peak position of the averaged C-
We will perform a quantitative analysis of the thermody-
namics and kinetics of the Li–I exchange as revealed by the
spectra of Figs. 5 and 6. First, we will determine the ate
complex association constant K2 from an analysis of chemi-
cal shifts and peak positions in the 13C NMR spectra of
Fig. 6 using the information on k1 and K1 gleaned from
Fig. 2 and similar experiments. This allows us to extract the
© 2005 NRC Canada

Reich and Whipple
1581
7
7
Fig. 7. Exchange matrix used for the simulation of the Li NMR
spectra of Fig. 5B.
Fig. 8. Line-shape simulation of the variable temperature Li
NMR spectra of Fig. 5B (4 (0.117 mol/L) and 6 (0.046 mol/L)
in THF–ether (3:2)). Lower lines are experimental data, upper
dotted lines are simulations using the rate constants shown and
the exchange matrix of Fig. 7.
A
D
M
kAD = k-4 [4M]
k4 [6]
(5)
(4D)
(4M)
kDA
=
i2π(vA)
-1/T-ΣkAX
A
(5)
kAM
kAD
2
7Li NMR Spectra
THF–ether (3:2)
kAM = k-2
i2π(vD)
-1/T-ΣkDX
D
(4D)
=
kDM
E
kDA
kMA = k2 [6]
kDM = 124
T (K)
k4 [6]
k
kDM
=
+
1
kAM = 1690
-110.9
i2π(vM)
-1/T-ΣkMX
2
M
(4M)
kMD
kMA
kMD = k1 [4M] + k 4 [5]
T (K)
-97.3
kDM = 356
kDM = 72
kAM = 27300
-113.7
Li–I exchange rate constant k2 from the spectra of Fig. 5,
and sets the stage for determining k3 as well as estimating
the maximum rate of exchange of the dimer 4D with the io-
dide 6 (k4).
kAM = 819
kDM = 346
kDM = 43
-102.7
-118.6
kAM = 9700
kAM = 389
Determination of the ate complex association constant (K2)
We have used the chemical shift of the C-3 signal in the
13C NMR spectra of an experiment like the one in Fig. 6A to
determine the composition of the sample. At each tempera-
ture above coalescence (–87 to –59 °C) an iterative computer
program was used to calculate the concentration of all spe-
cies for a series of trial ate complex concentrations until the
calculated averaged 13C NMR chemical shift matched the
experimental one. These calculations used the previously de-
termined values of K1 (eq. [1]), the concentrations of the
component species, and the chemical shifts of each com-
pound as input. The shifts of 4M, 4D, and 5 were found to
be temperature independent and those of 6 were extrapolated
from spectra of pure 6 taken at a series of temperatures.
Two additional data points were derived from the 7Li
NMR spectra. At –117, –107, –102, and –98 °C it was possi-
ble to detect the signal of 4D (see insets, Fig. 6B). The area
of this signal was determined at the two temperatures at
which it is best defined (–102 and –107 °C), and these val-
ues were then used to calculate the concentrations of the
other species using the known value of K1.
From the concentrations, K2 can be calculated at each
temperature and the thermodynamic parameters for the ate
complex equilibrium determined (see eq. [2], Fig. 3B). The
large negative entropy of formation of the ate complexes
(–48.6 eu) is similar to that found for the formation of lith-
ium diphenyliodinate 3 (–44 eu) (7c) and other systems (7d,
7f, 13). This loss of entropy derives from loss of trans-
lational and rotational motion for the component iodide and
lithium reagent, as well as the extra THF molecules frozen
by the very electrophilic free lithium cation of the ate com-
plex (7b, 7c), compared to the strongly C-bound cation of
the aryllithium reagent (7l).
kDM = 202
kDM = 33
kAM = 2800
-108.3
5
-124.7
5
kAM = 118
4D 4M
5
4D 4M
5
4
3
2
1
4
3
ppm
2
1
ppm
ence from dynamic effects. Since the lines were fifty to sev-
eral hundred Hz wide in the crucial part of the simulation,
we anticipated that errors in line width would not greatly af-
fect the accuracy of the rate constants determined. We as-
signed inherent line widths of 4 Hz for 5, 8 Hz for 4D, and
10 Hz for 4M based on the low temperature spectra and the
behavior of other related aryllithium species. An iterative
process was used to establish optimum populations and rates
(see the Experimental section) using the exchange matrix
shown in Fig. 7. The exchange matrix includes contributions
from the reaction of dimer with iodide (k4 process, eq. [4]).
The values of k–1 in solutions containing 6 were actually
smaller than those determined for solutions of just the lith-
ium reagent. Since no detectable contribution from k4 is
present, it was set to zero, leaving only kDM (k–1) and kAM (k–2)
to be adjusted (7m).
The simulations are presented in Fig. 8. An excellent
match between experimental and calculated spectra was
achieved, with k–1 within a factor of two of those values de-
termined from solutions of 4 (Fig. 2) and K2 values within a
factor of four of those determined from analysis of the previ-
ous 13C NMR shifts of Fig. 6 (see Fig. 4). There are a num-
ber of potential reasons for these discrepancies: the chemical
shifts may have a slight temperature dependence, the as-
sumption of T1 values for the three lithium species may have
been too crude, and there may have been small variations in
solvent composition between experiments. Both K1 and K2
are very solvent sensitive. For example, K1 decreased by a
factor of 5.7 between 60% and 80% THF in ether at
–117 °C. The values for k2 and k–2 as a function of tempera-
ture are shown in an Eyring plot in Fig. 4, leading to the ac-
tivation parameters reported in eq. [2].
Determination of the rate of Li–I exchange (k2)
We next simulated the NMR spectra of Fig. 5 to deter-
mine values for k2 and k–2. The 13C signals were rather
noisy, and simulations were not very well defined. However,
7
the Li NMR signals had superb signal-to-noise and had the
decisive advantage that the degenerate exchange of iodide
7
and ate complex (eq. [3]) does not affect the Li line shape,
since this process leaves the lithium signal unchanged.
7
A significant problem for simulations of the Li spectra is
Rate of exchange of ate complex with iodide (k3)
We are now in a position to attempt a simulation of the
13C NMR line shifts of Fig. 6, which involves, in addition to
the quite variable quadrupolar T1 relaxation rates (21), which
could not be independently measured because of interfer-
© 2005 NRC Canada

1582
Can. J. Chem. Vol. 83, 2005
Fig. 9. (A) Spectrum of a solution of 4 and 6 at –131 °C.
(B) Optimal simulation. (C) Simulation with no contribution
from the process of eq. [3] (k3) (7m). See Fig. 10 for the defini-
Fig. 10. Exchange matrix for the interactions between 4M, 4D,
5, and 6. See eqs. [1]–[4] for definitions of the rate constants.
k-4 [4M]
k4 [6]
kAD
=
kDA =
tions of kAI and kAM
.
A
D
M
I
2
2
(5)
(4D)
(4M)
(6)
kAM = k-2
k-2
kAI =
kMA = k2 [6]
13C NMR
6 (0.166 mol/L)
4 (0.128 mol/L)
i2π(vA)
-1/T-ΣkAX
A
D
k3 [6]
k-4 [4M]
kAD
kAM
kDM
kAI
kDI
+
+
2
2
i2π(vD)
-1/T-ΣkDX
kDA
kIA = k2 [4M] + k3 [5]
=
E
-131 °C
i2π(vM)
-1/T-ΣkMX
kMA
kMD
kID
kMI
M
I
kDM = k-1 kMD =
k1 [4M] + k 4 [5]
kDI = 0
kMI = 0
kID = 0
i2π(vI)
-1/T-ΣkIX
kIA
kIM
k-4 [4M]
kIM
=
–1
–1
2
kAM = 1.77 s
kAI = 1.77 s
–1
C
5
6 4D
4M
kAM = 1.77 s
kAI = 264 s
–1
C-3
B
A
(7b), where first-order behavior in iodide was demonstrated
by a DNMR study.
138 137 136 135 134 133 132 131 130 129 128 127
ppm
The data obtained show the astonishing result that the
monomeric lithium reagent and the ate complex have com-
parable reactivity as donors in the Li–I exchange. In another
test of ate complex reactivity, the reaction of PhLi–PhI with
hexamethyldisiloxane in THF–HMPA, PhLi is at least a fac-
tor of 105 more reactive than Ph2I–Li+ (7a). This estimate is
based on the reduction in kobs for the silylation reaction
when excess PhI was added to a solution of PhLi in THF–
HMPA. Of course, in this case the PhLi has been greatly ac-
tivated by complexation to HMPA (by a factor of 103 or
more) (7j), so a direct comparison is not very meaningful.
We considered that the k2 process might be approaching dif-
fusion control, and thus suffer from rate compression, but
diffusion-controlled processes at –102 °C in THF–ether
should have rates of ca 1 × 109 s–1 (mol/L)–1, about 103
times faster than the measured bimolecular rate (k2). A more
likely explanation is that the rate-limiting (or partial-
limiting) step for the Li–I exchange (eq. [2]) is C—Li bond
dissociation (CIP to SIP) or solvent association with or dis-
sociation from the lithium cation. Barriers to dissociation of
THF from lithium species can be substantially higher than
the activation energies of the Li–I exchange measured here
(7n, 22, 23). This is in principle testable (the Li–I exchange
should have fractional or zero order in iodide) by repeating
the experiment of Fig. 5 at various concentrations, to estab-
lish whether consistent rates would be obtained. Unfortu-
nately, the concentration of iodide is not easily manipulated
in solutions with excess lithium reagent, the only conditions
where k–2 can be measured.
It is the interference by the process of eq. [3] that prevents
us from demonstrating by DNMR techniques whether the ate
complex is an obligatory intermediate in the Li–I exchange.
This would require showing that the exchange rates of io-
dide and lithium reagent were the same and that the ate com-
plex lifetime was correctly predicted from the lifetimes of
the reaction partners. To achieve this, signals for all three
components (4M, 5, and 6) must be visible in a sample so
that lifetimes of all three can be simultaneously measured.
As can be seen from Figs. 5 and 6, at no point can 13C sig-
nals of all three species be seen. This is in part because at
the low temperatures where rates are slow enough to observe
individual signals, K2 is so large that the concentrations of
4M or 6 (whichever is stoichiometrically limiting) is too
small to detect, and in part because k3 is too large.
the eq. [1] and [2] processes, the bimolecular exchange be-
tween ate complex and iodide (k3, eq. [3]). We used the val-
ues of k1, k2, K1, and K2 calculated from the thermodynamic
and activation parameters in eqs. [1] and [2] to determine
the NMR rates and populations of 4M, 4D, 5, and 6 at each
temperature. Neither the populations nor the k1 and k2 rate
constants were allowed to vary in the simulation. Any ex-
cess broadening of the iodide 6 and ate complex 5 signals
over that predicted by k2 can then be assigned to k3, the io-
dide – ate complex exchange of eq. [3]. This is illustrated in
Fig. 9, which shows simulations of the –131 °C spectrum.
At this temperature the iodide signal would be decoalesced
if kAI = kAM (Fig. 9C), i.e., if k2 was the only mechanism for
exchange. This can be compared with the simulation in
Fig. 9B, which has the rate kAI optimally adjusted to fit the
actual spectrum.
The exchange matrix and the relationship between first-
order NMR rate constants (kAD, kAM, kAI, etc.) and their
mechanistic equivalents (k1, k2, k3, k4, etc.) is shown in
Fig. 10. This simulation is complicated by the fact that tem-
peratures low enough for decoalescence cannot be achieved.
In addition, values for k3 and k–3 are calculated from k2 and
k–2 (k3 = (kIA – k2[4M])/[5], k–3 = (2kAI – k–2)/[6]), and so
the limited accuracy of k2 because of the small temperature
range and scatter of the ∆G‡ vs. T plot (Fig. 4) directly af-
fects the accuracy of the k3 values.
The simulation worked reasonably well (insets in
Fig. 6A), although there were small errors in the chemical
shift of the averaged signal, which is not surprising consid-
ering the accumulation of calculated values that comprise
the input. The most likely source is small errors in the value
of K2 used to calculate the concentrations. Only at the four
lowest temperatures in Fig. 6 were the spectra significantly
affected by the value of k3, so we report data only at these
temperatures. Although there is significant broadening at in-
termediate temperatures (–87 and –81 °C), this is largely due
to the monomer–dimer exchange (k1) and is not very sensi-
tive to k3. Activation parameters are shown in eq. [3]. We
did not perform experiments to verify that the k3 process was
first order in 6 by repeating this experiment with variable io-
dide concentrations. However, such an experiment was pre-
viously performed for a PhLi–PhI solution in THF–HMPA
© 2005 NRC Canada

Reich and Whipple
1583
obligatory intermediate), and as a reactive partner in its own
right in donating a thienyl anion to iodothiophene (eq. [3]).
The latter observation suggests that ate complexes may be
reactive partners in reactions when lithium reagents are pre-
pared by Li–I exchange. The fact that solutions of iodide
and lithium reagent are stable in samples where upwards of
109 Li–I exchanges have occurred during an experiment pro-
vides a strong argument that single electron transfer reac-
tions involving aryl radical intermediates play no role in the
reaction, since cage effects of this magnitude would seem
impossible.
Rate of Li–I exchange for the dimer (k4)
7
The observation that the Li NMR signal of the dimer 4D
could be decoalesced from the averaged signal comprising
the lithiums of 4M and 5 in Fig. 6B shows that the dimer is
not undergoing Li–I exchange (eq. [4]) under conditions
7
where exchange of monomer is rapid. The Li NMR signal
of 4D at –102 °C is 60 Hz wide. At the same temperature
the 13C NMR signal of 4D in a sample containing only 4 is
80 Hz (Fig. 2), so there is no detectable broadening assign-
able to the Li–I exchange of dimer, the k4 process. If we es-
timate that no more than 10 Hz of the 60 Hz line width of
the lithium signal could be due to the k4 process, then the
NMR rate constant kDA ≤ 31 s–1 (π∆ν), and the physical rate
constant k4 = kDA/[6] ≤ 670 s–1 (mol/L)–1. Since the rate of
exchange of the iodide with monomer (k2) is 6.7 × 106 s–1
(mol/L)–1 at –102 °C (see eq. [2]), reaction of dimer with the
iodide is at least a factor of 1000 slower than reaction of the
monomer. Although many reactivity effects of organolithium
reagents have been ascribed to aggregation phenomena, de-
termination of quantitative and qualitative differences be-
tween reactivity of different aggregates of an organolithium
species under the same conditions has rarely been reported
(24, 25). This result shows that the common assumption that
higher aggregates are less reactive than lower ones holds
well in this monomer–dimer case, and that the differences
can be quite large.
Experimental section
General
All glassware was dried in an oven at 110 °C for at least
2 h and flushed with N2 for at least 5 min. All reactions
were performed under an atmosphere of dry nitrogen.
Instrumentation
GC analyses were performed using a 12 m × 0.32 mm
SE-30 capillary column, with He pressure of 6.0 psi (1 psi =
6.894 757 kPa), column flow rate (split ratio 300:1) of
3 mL/min, and column temperature of 90 °C. Retention times
and response factors with respect to n-undecane are as fol-
lows: 2-iodo-5-methylthiophene (2.77 min, 2.39), 2-(methylthio)-
5-methylthiophene (2.87 min, 1.99), n-undecane (2.67 min,
1.00). For a 1:1 molar solution of compound: C11H24, the re-
sponse factor = (peak area C11H24)/(peak area compound).
Summary
A DNMR investigation of 2-lithio-5-methylthiophene (4)
showed this lithium reagent to be a mixture of monomer and
4-center dimer in THF–ether (3:2). Solutions of 4 and the
corresponding iodide 6 undergo Li–I exchange on the NMR
timescale. The ate complex 5, a probable intermediate in the
exchange, is formed with a high association constant at low
temperature (–130 °C), but is largely dissociated at tempera-
tures above –70 °C. A variable-temperature DNMR kinetic
analysis showed that three exchange processes occurred on
the NMR timescale, the monomer dimer exchange (eq. [1]),
the Li–I exchange of the monomeric lithium reagent
(eq. [2]), and direct exchange between the ate complex and
the iodide (eq. [3]). The ate complex 5 is comparable in re-
activity to the monomeric lithium reagent 4M in the Li–I ex-
change process. The data obtained also allowed an estimate
of the upper limit of a fourth process, the Li–I exchange of
the dimer (eq. [4]). The bimolecular rate constants at –102 °C
are shown in the following:
7
6
13C, Li, and Li NMR spectra were obtained on a Bruker
AM-360 spectrometer operating at 90.56, 139.96, and
53.00 MHz, respectively. The NMR spectra were acquired
unlocked and in non-deuterated solvents unless otherwise
stated. For 13C NMR, a 30° pulse of 5 µs repeated every
2.9 s was used. An exponential line broadening of 2–5 Hz
was used, and a 64 K transform was carried out over a spec-
tral width of 20 800 Hz. 13C NMR chemical shifts are re-
ported in ppm relative to TMS (δ 0.0), using THF as an
7
internal reference. Li NMR spectra were recorded over a
spectral width of 1400 Hz (8K data points) using a 10 µs
pulse and 6Li NMR spectra over a spectral width of 2907 Hz
(8K data points) with a pulse width of 10 µs and a recycle
7
6
delay of 4 s. Both Li and Li NMR chemical shifts were
referenced to external 0.30 mol/L LiCl in MeOH (δ 0.0 ≈
–100 °C) (19c). n-Bu-6Li (95.5% enriched) was prepared by
literature procedures (7l).
Probe temperatures were measured (decoupler turned off)
by replacing the sample with an open tube containing the
same solvent and the sensor of a RTD (platinum resistance)
thermometer, which was designed to measure temperatures
from –150 °C to 50 °C, and was calibrated using ice water
(0.01 °C) and CO2 (–78.5 °C). The accuracy of the resis-
tance thermometer was checked with the melting points of
toluene (–95.00 °C) and methylcyclohexane (–126.6 °C) and
were found to be within 0.4 °C of the reported value. Sam-
ples were allowed to equilibrate for at least 8 min in the
probe prior to tuning the spectrometer and acquiring the
spectrum. Probe temperatures were confirmed following the
acquisition of each spectrum.
Rate constant
k ((mol/L)–1 s–1)
krel
1/100
1.0
1.9
<1/1000
6.8×103
6.7×105
1.3×106
<6.6×102
k1 (eq. [1])
k2 (eq. [2])
k3 (eq. [3])
k4 (eq. [4])
The lithium iodine exchange reaction has been mechanis-
tically controversial since its discovery in 1938. The present
DNMR investigation, together with other related papers (7a–
7c), places aryl iodine ate complexes (5) firmly into the
mechanistic picture in two roles: as a permissible intermedi-
ate in the exchange itself (our results do not make it an
DNMR simulations were performed with the Binsch
DNMR program (26) and with WINDNMR (7m).
© 2005 NRC Canada

1584
Can. J. Chem. Vol. 83, 2005
Table 1. Data for the plot of ln[monomer] vs. ln[dimer] at –121 °C.
Solvents
THF and Et2O were freshly distilled from sodium benzo-
[4]
(mol/L)
Peak ratio
(4D/4M)
[4D]
(mol/L)
[4M]
(mol/L)
K1
phenone ketyl prior to use. The THF still was recharged ev-
ery 3 weeks to insure that no benzene (formed by cleavage
of benzophenone) accumulated in the THF. Solutions of
organolithium reagents were titrated by use of 1-propanol
with 1,10-phenanthroline as the indicator (27). 2-
Methylthiophene was commercially available.
((mol/L)–1)
0.21
0.12
0.094
0.062
0.028
1.845
1.332
1.190
0.739
0.510
0.0681
0.0343
0.0255
0.0132
0.0047
0.0783
0.0515
0.0429
0.0357
0.0185
12.5
12.9
13.9
10.4
13.8
Preparation of stock solutions of 2-lithio-5-
methylthiophene (4)
To a solution of 2-methylthiophene (2.4 mL, 25 mmol) in
5 mL of THF at –78 °C was added 13.3 mL of a 1.88 mol/L
solution of n-BuLi or n-Bu6Li (25 mmol). The reaction mix-
ture was warmed to 0 °C for 10 min, then cooled to –78 °C.
Crystals formed in the centrifuge tube within 10 min. The
supernatant was removed from the crystals, which were
washed two times with 6 mL of THF, each time removing
the excess solvent by cannula. The crystals were dissolved in
ca. 15 mL of THF to give a solution of 4 in THF, which was
determined to be 0.95 mol/L. The solution of 4 was stable
for several months at –20 °C.
10 mL of pentane and 1 to 2 drops of 30% NaOH and the
mixture dried with Na2SO4. The organic solution was ana-
lyzed by GC to determine the relative concentrations of the
components vs. n-undecane.
Concentration dependence of 4
Seven samples of 4 in THF–ether (3:2) were prepared by
appropriate dilution of a 0.78 mol/L sample of 4 in THF.
Spectra (13C, 7Li) of each sample were taken at –121 °C. Af-
ter the experiment, the concentration of 4 in each sample
was determined by quenching with Me2S2 and GC analysis.
The spectra from the five samples with the highest concen-
trations were integrated (C-3) by line-shape simulation. The
data are shown in Table 1 and Fig. 3A.
2-Iodo-5-methylthiophene (6)
To a solution of 2-methylthiophene (96.8 µL/mmol,
1.94 mL, 20 mmol) in 20 mL of THF at –78 °C was added a
solution of n-BuLi (1.88 mol/L, 10.6 mL, 20 mmol). The so-
lution of 4 was warmed to 0 °C and added dropwise to a
suspension of 1,2-diiodoethane in THF at –78 °C. The mix-
ture was stirred at –78 °C for 1 h, warmed to 0 °C for 2 h,
and partitioned between 200 mL of Et2O–pentane (1:1) and
150 mL of NaHCO3. The organic layer (purple) was washed
with 50 mL of 0.5 mol/L Na2S2O3, 200 mL of water,
150 mL of brine, and passed through a cone of Na2SO4.
Evaporation of the solvent gave a light brown liquid (4.34 g,
97%), which was purified by Kugelrohr distillation (bp 40–
60 °C (1 mm Hg)) giving a colorless liquid. 1H NMR
(CDCl3, 200 MHz) δ: 2.46 (d, J = 0.94 Hz, 3H), 6.45 (dq,
J = 3.7, 0.9 Hz, 1H), 7.01 (d, J = 3.67 Hz, 1H). 13C NMR
(THF–Et2O (60:40), –116 °C, 90.556 MHz) δ: 15.4, 73.2,
128.34, 137.4, 146.3 (28).
Temperature dependence of 4
A 0.101 mol/L sample of 4 in THF–ether (3:2) was pre-
pared and a series of 13C and 7Li NMR spectra were taken at
accurately determined temperatures. Sample spectra are
shown in Fig. 2. The spectra of the C-3 carbons were simu-
lated. Line widths in the absence of exchange were deter-
mined from the solvent peaks. Below coalescence the line
positions and populations of 4M and 4D were allowed to
vary in the simulation. For determination of the thermody-
namic parameters above coalescence, the populations were
calculated from the chemical shift of the averaged peaks.
Several experiments were performed to determine ∆H1 and
∆S1 at various concentrations: the ∆H1 values varied from
0.8 to 1.4 kcal/mol and the ∆S1 values from 10.5 to 14.5 eu,
giving some measure of the reproducibility of these experi-
ments. One significant variable may be accurately reproduc-
ing solvent composition for the NMR samples, since the K1
values are very sensitive to the ratio of THF to ether.
For determination of the kinetic parameters, the popula-
tions during and above coalescence were calculated from the
∆H1 and ∆S1 values given in eq. [1]. Data used and derived
from the simulations are shown in Table 2. Activation pa-
rameters were determined from the values of k1 and k–1 and
are shown in eq. [1].
The preparation of NMR samples of 2-lithio-5-methyl-
thiophene (4)
To a septum-capped 10 mm NMR tube under N2 was
added 0.05 mL of a 1.42 mol/L solution of 2-lithio-5-
methylthiophene (4) and 1 mL of THF. The NMR tube con-
taining the dilute solution of 4 was rolled so that all of the
surface of the NMR tube and septum came in contact with
the organolithium solution (to dry the NMR tube). The solu-
tion of 4 was removed from the NMR tube with a syringe. A
solution of 4 in THF (1.42 mol/L, 0.20 mL, 0.288 mmol)
was added to the NMR containing a static N2 needle. THF
(1.8 mL) and Et2O (1.2 mL) was added to the tube and the
solution was mixed with a vortex mixer. The contents of the
tube were cooled to –78 °C and the septum was sealed with
parafilm.
The preparation of NMR samples of 4 containing 2-
iodo-5-methylthiophene (6)
The sample preparation is the same as described in the
previous section, except after the NMR tube was dried, Et2O
was added (22 °C) followed by 6. The sample was mixed
with a vortex mixer, cooled to –78 °C, then a solution of 4 in
THF was added. The sides of the NMR tube were washed
with the THF, which was added last.
The determination of concentration for NMR samples
To the NMR solution was added Me2S2 and 25 µL
(0.118 mmol) of n-undecane. The mixture was added to
© 2005 NRC Canada

Reich and Whipple
Table 2. Data for DNMR simulation of the spectra in Fig. 2.
1585
[4D]
(mol/L)
[4M]
(mol/L)
kMD
(s–1)a
k1
kDM
(s–1)c
K1
T (K)
((mol/L)–1 s–1)b
((mol/L)–1)d
20
57
116
293
1 654
3 859
386
1 246
2 931
7 400
43 600
22
47
74
189
995
2 134
3 350
9 380
8.7
13.3
19.8
19.6
21.9
25.1
27.7
31.0
35.8
37.6
41.0
139.9
153.1
163.0
170.4
182.0
192.5
200.7
210.9
220.6
231.0
241.7
0.0242
0.0278
0.0308
0.0307
0.0315e
0.0325e
0.0332e
0.0340e
0.0346e
0.0352e
0.0357e
0.0526
0.0457
0.0394
0.0396
0.0379e
0.0360e
0.0346e
0.0331e
0.0311e
0.0306e
0.0295e
107 000
175 000
582 000
1 180 000
3 300 000
8 410 000
6 405
19 260
36 820
101 000
248 300
16 900
43 950
102 600
aDNMR rate constant for conversion of monomer to dimer nuclei.
bk1 = kMD/[4M] (s1 (mol/L)–1).
cDNMR rate constant for the conversion of dimer to monomer nuclei k–1 = kDM (s–1).
dK1 = [4D]/[4M]2 ((mol/L)–1).
eThese concentrations were calculated from ∆H1 and ∆S1 (eq. [1]).
Table 3. Input and output data for determination of K2 in THF–ether (3:2).
K1
[4D]
(mol/L)
[4M]
(mol/L)
[5]
(mol/L)
[6]
(mol/L)
K2
T (K)
δ for 6
δ obs
((mol/L)–1)
((mol/L)–1)
0.67
1.02
1.91
2.95
4.66
7.15
12.3
27.4
55.9
98.8
229
214.4
209.3
203.6
200.1
197.0
194.1
191.8
186.2
180.8
175.2
171.2
166.2
137.33
137.32
137.30
137.29
137.28
137.28
137.27
137.25
137.24
137.22
136.34
136.20
135.91
135.61
135.19
134.69
133.94
132.67
131.60
32.14
30.50
28.94
27.53
26.57
25.66
24.93
23.17
21.54
19.82
18.64
17.11
0.041 7
0.040 5
0.038 2
0.035 9
0.032 9
0.029 4
0.024 3
0.015 9
0.009 3
0.005 3
0.001 9
0.000 78
0.0360
0.0394
0.0365
0.0361
0.0352
0.0338
0.0312
0.0262
0.0207
0.0163
0.0101
0.0067
0.0037
0.0057
0.0144
0.0150
0.0219
0.0304
0.0433
0.0651
0.0838
0.0962
0.109
0.152
0.150
0.145
0.141
0.134
0.126
0.113
0.091
0.072
0.060
0.047
0.042
130.90
a
a
0.115
397
7
aConcentrations calculated from the fraction of dimer in the Li NMR spectra.
the concentrations of the other species could be determined
by a similar iterative process. The area of this signal was de-
termined at the two temperatures at which it is best defined
(–102 and –107 °C), where 4D corresponds to 1.4% and
0.6% of total peak area. These values were then used to cal-
culate the concentrations of the other species using the
known value of K1.
Determination of the ate complex association constant
(K2)
A sample 0.156 mol/L in 6 and 0.123 mol/L in 4 was pre-
7
pared using the previous procedure. 13C and Li NMR spec-
tra were taken at 13 temperatures between –117 and –59 °C
(Fig. 6). The 13C NMR chemical shift of the averaged C-3
carbon in the spectra from –87 to –59 °C was used to deter-
mine the concentrations of 4M, 4D, 5, and 6. Input parame-
ters were the values of the chemical shifts of the individual
compounds (δ 132.6 for 4M, δ 136.8 for 4D, δ 128.7 for 5,
the shifts of 6 were temperature dependent and calculated
using the formula δ = 0.002 815(T °C) + 137.500), the ob-
served chemical shift in the averaged spectrum, the
stoichiometric concentrations of 4 and 6, and the value of K1
calculated from ∆H1 and ∆S1. The input data were then used
to iteratively calculate values for all concentrations using
test values of K2 until the average δ for all species matched
observed δ for C-3. Data are presented in Table 3.
Determination of ate complex association and
dissociation rate constants (k2 and k–2)
A sample 0.117 mol/L in 4 and 0.046 mol/L in 6 in THF–
7
ether (3:2) was prepared by the general procedure, and Li
and 13C NMR spectra were measured at a series of accu-
7
rately measured temperatures (Fig. 5). The Li signals were
simulated as follows. The population of the exchanging spe-
cies (4M, 4D, 5, and 6) were calculated from the known
concentrations and the values of K1 and K2 at each tempera-
ture, and values for k1 and k–1 were calculated from the acti-
vation parameters of eq. [1]. These were used to start the
simulation. The line width in the absence of exchange for
the simulations was problematic, since it is likely to be both
Two additional K2 values were determined at –107 and
–102 °C, where it is possible to detect the signal of 4D (see
Fig. 6B and insets). From the relative area of the dimer peak
© 2005 NRC Canada

1586
Can. J. Chem. Vol. 83, 2005
7
Table 4. Data for the DNMR simulation of the Li NMR spectra of Fig. 5B (determination of k2 and k–2).
[4D]
[4M]
[5]
[6]
kDM
kAM
(s–1)b
127
369
849
1 750
3 100
12 200
27 300
k2
T (K)
(mol/L)a
(mol/L)a
(mol/L)a
(mol/L)a
(s–1)b
((mol/L)–1 s–1)c
12
29
52
109
181
336
356
148.5
154.6
159.5
162.2
164.9
170.5
175.9
0.0171
0.0182
0.0194
0.0202
0.0213
0.0244
0.0281
0.0370
0.0358
0.0354
0.0353
0.0353
0.0361
0.0372
0.0456
0.0446
0.0427
0.0410
0.0388
0.0319
0.0235
0.0004
0.0014
0.0032
0.0049
0.0073
0.0142
0.0225
3.70×105
3.32×105
3.16×105
4.12×105
4.68×105
7.61×105
7.67×105
aCalculated from K1 (eq. [1]) and K2(Li) at each temperature and fixed during the simulation (pop(dimer) = 2[dimer]).
bDetermined by simulation NMR.
ck2 = kMA/[6].
Table 5. Input and output data for simulation of the 13C NMR spectra in Fig. 6 (determination of k3 and k–3).
[4D]
[4M]
[5]
[6]
kAI
(s–1)b
k–2
(mol/L)c
k–3
(mol/L)d
kIA
(s–1)e
k2
k3
T (K)
(mol/L)a
(mol/L)a
(mol/L)a
(mol/L)a
(mol/L)c
(mol/L)f
2.65×103
7.31×103
1.98×104
5.62×104
4.80×105
1.34×105
3.38×105
7.56×105
1.57×106
1.07×107
1.70×104
4.24×104
9.45×104
1.96×105
9.85×105
2.25×105
3.50×105
5.27×105
7.31×105
9.58×105
1.28×105
3.32×105
7.69×105
1.76×106
1.15×107
16.5
230
2 581
17 570
84 100
141.9
152.5
163.7
173.8
183.0
0.0018
0.0015
0.0015
0.0041
0.0109
0.0035
0.0041
0.0086
0.0146
0.0222
0.127
0.124
0.117
0.105
0.084
0.0394
0.0426
0.0490
0.0606
0.0820
aCalculated from K1 (eq. [1]) and K2 (eq. [2]) at each temperature and fixed during the simulation (pop(dimer) = 2[dimer]).
bNMR rate for conversion of ate complex nuclei to iodide nuclei determined by line-shape simulation.
cCalculated from the activation parameters in eq. [2].
dCalculated: k–3 = (2kAI – k–2)/[6].
eFrom line-shape fitting: kIA = kAI(Pop A)/(Pop I).
fCalculated: k3 = (kIA – k2[4M])/[5].
compound specific (different rates of quadrupolar relaxation
eqs. [1] and [2]. The exchange matrix is shown in Fig. 10,
the numeric data in Table 5, and the simulated spectra in
Fig. 6A.
for different lithium environments) and temperature sensi-
tive. In the absence of actual information, the line widths
were set at 4, 8, and 10 Hz for 5, 4D, and 4M, respectively,
based on values for several related species. The simulations
were performed with the program WINDNMR (7m) using
the exchange matrix of Fig. 7. The relative values of 4M and
4D fit well, but small changes in the relative populations of
5 were required for a good fit for all nuclei. Chemical shifts
were assumed to be temperature independent (no actual
measurements could be performed because dynamic broad-
ening occurred at too low a temperature) and were set by fit-
ting the spectrum at 142 K. After an optimal fit was
achieved, we redetermined K2 at each temperature, shown as
K2(Li) in Fig. 3B (∆H2 = –8.8 1 kcal/mol, ∆S2 = –43.3
1 eu). New concentrations of all species were calculated
with these parameters, and a final fitting was performed with
the calculated concentrations frozen, modifying only kAI and
kAM. Only the spectra from –124 to –97 °C provided good
definition for the individual rates, and data for these temper-
atures are shown in Table 4 and Fig. 4.
Acknowledgement
We thank the National Science Foundation (NSF) for fi-
nancial support.
References
1. H. Gilman, W. Langham, and A.L. Jacoby. J. Am. Chem. Soc.
61, 106 (1939).
2. G. Wittig, U. Pockels, and H. Dröge, Chem. Ber. 71, 1903
(1938).
3. (a) W.F Bailey and J.J. Patricia. J. Organomet. Chem. 352, 1
(1988); (b) W.F. Bailey, T.T. Nurmi, J.J. Patricia, and W.
Wang. J. Am. Chem. Soc. 109, 2442 (1987); (c) E.C. Ashby
and T.N. Pham. J. Org. Chem. 52, 1291 (1987); (d) K.B.
Wiberg, S. Sklenak, and W.F. Bailey. Organometallics, 20, 771
(2001).
4. A.P Batalov. Zh. Obshch. Chim. 48, 1607 (1978).
5. H.R. Ward. Acc. Chem. Res. 5, 18 (1972).
Determination of rate constants for ate complex
exchange with iodide (k3, k–3)
6. H.R. Rogers and J. Houk. J. Am. Chem. Soc. 104, 522 (1982).
7. (a) H.J. Reich, N.H. Phillips, and I.L. Reich. J. Am. Chem.
Soc. 107, 4101 (1985); (b) H.J. Reich, D.P. Green, and N.H.
Phillips. J. Am. Chem. Soc. 111, 3444 (1989); (c) H.J. Reich,
D.P. Green, and N.H. Phillips. J. Am. Chem. Soc. 113, 1414
(1991); (d) H.J Reich, B.Ö. Gudmundsson, D.P. Green, M.J.
Bevan, and I.L. Reich. Helv. Chim. Acta, 85, 3748 (2002);
(e) H.J. Reich, D.P. Green, N.H. Phillips, J.P. Borst, and I.L.
Reich. Phosphorus Sulfur Silicon Relat. Elem. 67, 83 (1992);
The 13C NMR spectra in Fig. 6A were simulated to ex-
tract values for the rate of 5/6 exchange. A sample was pre-
pared 0.166 mol/L in 6 and 0.128 mol/L in 4 in THF–ether
(3:2). In the simulations the population of all of the species
was calculated from the values of K1 and K2 determined
from ∆H1 and ∆S1, and values of k1, k–1, k2, and k–2 were
calculated from the values for the activation parameters in
© 2005 NRC Canada

Reich and Whipple
1587
( f ) H.J. Reich, B.Ö. Gudmundsson, and R.R. Dykstra. J. Am.
Chem. Soc. 114, 7937 (1992); (g) H.J. Reich and N.H. Phil-
lips. J. Am. Chem. Soc. 108, 2102 (1986); (h) D.R. Powell,
W.L. Whipple, and H.J. Reich. Acta Crystallogr. Sect. C, C52,
1346 (1996); (i) K.L. Jantzi, C.L. Puckett, I.A. Guzei, and H.J.
Reich. J. Org. Chem. 70, 7520 (2005); ( j) H.J. Reich, D.P.
Green, M.A. Medina, W.S. Goldenberg, B.Ö. Gudmundsson,
R.R. Dykstra, and N.H. Phillips. J. Am. Chem. Soc. 120, 7201
(1998); (k) H.J. Reich, W.S. Goldenberg, A.W. Sanders, K.L.
Jantzi, and C.C. Tzschucke. J. Am. Chem. Soc. 125, 3509
(2003); (l) H.J. Reich, J.P. Borst, R.R. Dykstra, and D.P.
Green. J. Am. Chem. Soc. 115, 8728 (1993); (m) DNMR sim-
ulations were performed with a version of the computer pro-
gram WINDNMR: H.J. Reich. J. Chem. Ed. Software, 3D, 2
(1996). Available from http://www.chem.wisc.edu/areas/reich/
plt/windnmr.htm); (n) H.J. Reich and K.J. Kulicke. J. Am.
Chem. Soc. 118, 273 (1996).
15. T. Kanda, S. Kato, N. Kambe, Y. Kohara, and N. Sonoda. J.
Phys. Org. Chem. 9, 29 (1996).
16. (a) S. Harder, J. Boersma, L. Brandsma, J.A. Kanters, W.
Bauer, R. Pi, P.v.R. Schleyer, H. Schöllhorn, and U. Thewalt.
Organometallics, 8, 1688 (1989); (b) W. Bauer, W.R. Winches-
ter, and P.v.R. Schleyer. Organometallics, 6, 2371 (1987).
17. R.R. Fraser, T.S. Mansour, and S. Savard. Can. J. Chem. 63,
3505 (1985).
18. W.E. Truce and E. Wellish. J. Am. Chem. Soc. 74, 5177 (1952).
19. For some organolithium aggregates, quadrupolar broadening is
7
severe enough with Li (92.6% natural abundance) that C–Li
6
coupling cannot be resolved. The Li analogs show little or no
quadrupolar broadening and thus couplings are more easily
seen. (a) G. Fraenkel, M. Henrichs, J.M. Hewitt, B.M. Su, and
M.J. Geckle. J. Am. Chem. Soc. 102, 3345 (1980); G.
Fraenkel, A. Chow, and W.R. Winchester. J. Am. Chem. Soc.
112, 6190 (1990); G. Fraenkel, S. Subramanian, and A. Chow.
J. Am. Chem. Soc. 117, 6300 (1995); (b) A.J. Jones, D.M.
Grant, J.G. Russell, and G. Fraenkel. J. Phys. Chem. 73, 1624
(1969); (c) G. Fraenkel and M.P. Hallden-Abberton. J. Am.
Chem. Soc. 103, 5657 (1981).
8. G. Boche, M. Schimeczek, J. Cioslowski, and P. Piskorz. Eur.
J. Org. Chem. 1851 (1998); J. Cioslowski, P. Piskorz, M.
Schimeczek, and G. Boche. J. Am. Chem. Soc. 120, 2612
(1998).
9. W.B. Farnham and J.C. Calabrese. J. Am. Chem. Soc. 108,
2449 (1986).
20. S. Gronowitz, I. Johnson, and A.B. Hörnfeldt. Chem. Scr. 7,
76 (1975).
10. (a) S. Ogawa, Y. Masutomi, N. Furukawa, and T. Erata.
Heteroat. Chem. 3, 423 (1992); (b) Y. Masutomi, N.
Furukawa, and T. Erata. Heteroat. Chem. 6, 19 (1995).
11. A.J. Ashe III, L.L. Lohr, and S.M. Al-Taweel. Organo-
metallics, 10, 2424 (1991).
21. L.M. Jackman and L.M. Scarmoutzos. J. Am. Chem. Soc. 106,
4627 (1984).
22. B.L. Lucht and D.B. Collum. J. Am. Chem. Soc. 116, 6009
(1994).
23. G. Hilmersson and Ö. Davidsson. J. Org. Chem. 60, 7660
(1995).
12. A. Maercker, H. Bodenstedt, and L. Brandsma. Angew. Chem.
Int. Ed. Engl. 31, 1339 (1992).
24. J.F. McGarrity, C.A. Ogle, Z. Brich, and H.-R. Loosli. J. Am.
13. A.H.J.F. de Keijzer, F.J.J. de Kanter, M. Schakel, R.F. Schmitz,
and G.W. Klumpp. Angew. Chem. Int. Ed. Engl. 35, 1127
(1996); A.H.J.F. de Keijzer, F.J.J. de Kanter, M. Schakel,
V.P. Osinga, and G.W. Klumpp. J. Organomet. Chem. 548, 29,
(1997).
Chem. Soc. 107, 1810 (1985).
25. (a) A. Streitwieser, S.S.-W. Leung, and Y.-J. Kim. Org. Lett. 1,
145 (1999); (b) S.S.W. Leung and A. Streitwieser. J. Am.
Chem. Soc. 120, 10557 (1998).
26. D.S. Stephenson and G. Binsch. Quantum Chem. Prog. Exch.
14. (a) V. Schulze, M. Brönstrup, V.P.W. Böhm, P. Schwerdtfeger,
M. Schimeczek, and R.W. Hoffmann. Angew. Chem. Int. Ed.
37, 824 (1998); (b) M. Müller, H.-C. Stiasny, M. Brönstrup, A.
Burton, and R.W. Hoffmann. J. Chem. Soc. Perkin Trans. 2,
731 (1999).
10, 365 (1978).
27. S.C. Watson and J.F. Eastham. J. Organomet. Chem. 9, 165
(1967).
28. S. Hotta. J. Heterocycl. Chem. 38, 923 (2001).
© 2005 NRC Canada