The role of ate complexes in the lithium-sulfur, lithium-selenium and lithium-tellurium exchange reactions

Article abstract of DOI:10.1002/1522-2675(200211)85:11<3748::AID-HLCA3748>3.0.CO;2-0

Hypervalent ate complexes are presumptive intermediates in the metal-halogen, metal-tellurium, and related exchange reactions. The effect of o,o′-biphenyldiyl vs. diphenyl substitution on formation of tellurium ate complexes was studied by a kinetic technique and by NMR spectroscopy. Only a modest increase in the association constant (K_ate) was measured. When Li/M exchanges of o,o′-biphenyldiyl sulfides and selenides were made intramolecular by means of a m-terphenyl framework (12-S, 12-Se, 21), enormous increases (> 10⁹) in the rate of Li/S and Li/Se exchange were observed compared to acyclic models. Apparently, these systems are ideally preorganized to favor the T-shaped geometry of the hypervalent intermediates. For the selenium systems, ate complex intermediates (20-Se, 26) were detected spectroscopically in THF- or THF/HMPA-containing solutions. A DNMR study showed that Li/Se exchange was substantially faster than exchange of the lithium reagents with the ate complex. Therefore, these ate complexes are not on the actual Li/Se exchange pathway.

Full text of DOI:10.1002/1522-2675(200211)85:11<3748::AID-HLCA3748>3.0.CO;2-0

3748
Helvetica Chimica Acta Vol. 85 (2002)
The Role of Ate Complexes in the Lithium-Sulfur, Lithium-Selenium and
Lithium-Tellurium Exchange Reactions
by Hans J. Reich*, Birgir ÷. Gudmundsson, D. Patrick Green, Martin J. Bevan, and Ieva L. Reich
Department of Chemistry, University of Wisconsin-Madison, Madison, WI 53706
This paper is dedicated to Professor Dieter Seebach in honor of his 65th birthday
Hypervalent ate complexes are presumptive intermediates in the metal-halogen, metal-tellurium, and
related exchange reactions. The effect of o,o'-biphenyldiyl vs. diphenyl substitution on formation of tellurium
ate complexes was studied by a kinetic technique and by NMR spectroscopy. Only a modest increase in the
association constant (Kate) was measured. When Li/M exchanges of o,o'-biphenyldiyl sulfides and selenides were
made intramolecular by means of a m-terphenyl framework (12-S, 12-Se, 21), enormous increases (>109) in the
rate of Li/S and Li/Se exchange were observed compared to acyclic models. Apparently, these systems are
ideally preorganized to favor the T-shaped geometry of the hypervalent intermediates. For the selenium systems,
ate complex intermediates (20-Se, 26) were detected spectroscopically in THF- or THF/HMPA-containing
solutions. A DNMR study showed that Li/Se exchange was substantially faster than exchange of the lithium
reagents with the ate complex. Therefore, these ate complexes are not on the actual Li/Se exchange pathway.
Introduction. The lithium-metalloid exchange reaction is the mildest and most
general procedure for the preparation of organolithium reagents. The lithium-bromine
[1][2a] and lithium-iodine exchanges have been the most popular, but the reaction
applies to many of the main-group third-, fourth-, and fifth-row elements. Tin [3a][4],
selenium (first studied by Seebach and Peleties [3b], and tellurium [3c] have been used
extensively. The reaction fails with second-row CÀM bonds like those of chlorides,
sulfides, and phosphines, except in exceptional circumstances, e.g., when there are no
protons that can be metallated, when an unusually stable carbanion is being prepared,
or when a strained ring is being cleaved [5][6].
Hypervalent ate complexes such as 1 3 have been spectroscopically characterized
and are likely intermediates in the degenerate phenyl-phenyl Li/I [7a][7b][8][9], Li/Te
[7b][7c][10a], and Li/Sn [7d][11][12] exchange reactions. Ate complexes of third- and
even second-row metalloids Se, P, and Si can be detected in favorable structures, e.g.,
when the aryl groups are heavily substituted with electronegative halogen atoms (4, see
[10b]) or when o,o'-biphenyldiyl ligands are present (5, see [13] and 6, see [14a]).

Helvetica Chimica Acta Vol. 85 (2002)
3749
Although the identification of ate complexes such as 1 4 suggests strongly that
they are intermediates in the various Li/M exchanges, this has not been unambiguously
demonstrated. There could, for example, be a direct exchange process kd (Eqn. 1) faster
than the path through the ate complex (kate). We address this question here in the
context of studies of the Li/Te, Li/Se, and Li/S exchanges.
The Li/S exchange is rarely observed, and the Li/Se exchange is relatively slow. To
remedy this, and to stabilize potential ate-complex intermediates in the hope that they
could be directly observed, we undertook to prepare sulfur- and selenium-containing
aryllithium reagents [7e] incorporating the o,o'-biphenyldiyl ligand, identified by
Hellwinkel and co-workers [14] and others as providing increased kinetic and
thermodynamic stability to hypervalent structures. This motif has allowed the study
of mono-, bis-, and even tris-biphenyldiyl substituted hypervalent compounds of I, Te,
Se, P, Si, and S, among them5 [13], 6 [14a], 7 [7f][15], 8 [14b][14c][10c], and their S
[10d] and Se [14d][10c] analogs. All of these showed dramatically higher stability than
analogous polyphenyl compounds Ph3I [1b], Ph4Te [1c][10f], Ph4Se [10e][10f], and
Ph4S [10g][10f]. For example, Ph4Te decomposes rapidly at 808 [10f], whereas 8 is
stable to 2088 [14c].
In the course of studying these systems, we identified structures that allowed us to
address the subtle mechanistic question outlined in Eqn. 1. Specifically, a system that
fulfills the following requirements is needed: 1) the three species in the triangular
equilibrium of Eqn. 1 must all be present in detectable concentrations under conditions

3750
Helvetica Chimica Acta Vol. 85 (2002)
where the rates kd, kate, and k'ate can each be measured (this can be problematic because
the rates and the ate complex equilibrium constant are very temperature- and solvent-
dependent, and both must be in the measurable range), 2) the lithium reagent should
be monomeric to simplify the rate measurements (this plagues the PhLi/PhI and PhLi/
Ph2Te, and related thiophene systems, since significant concentrations of kinetically
inactive PhLi dimer are present), and 3) there must be minimal interference from the
direct exchange between the metalloid and the ate complex, in which the ate complex
acts as an anion donor (this process complicates kinetic analysis of Li/I exchanges [7a]).
Results and Discussion. Effect of Biphenyldiyl Ligand on Tellurium Ate Complex
Stability. We began our study by investigating a system where the effect of biphenyldiyl
vs. diphenyl substitution on chalcogen ate-complex stability could be measured. There
was some reason to be concerned, since, in contrast to most selenides, which undergo
Li/Se exchange on treatment with alkyllithium reagents, selenophene [16] and
benzoselenophene [17] are instead metallated by BuLi. Even dibenzoselenophene
[18a] and benzotellurophene [12] can be metallated cleanly under appropriate
conditions. Thus, these systems, rather than being activated toward Li/M exchange,
show an unusual resistance to nucleophilic attack at the heteroatom, perhaps as a result
of aromatic stabilization of the heterocyclic ring. To test the −5-ring effect× on ate-
complex formation, we compared the propensities of Ph2Te and the dibenzotellur-
ophene 10 [14c] to form ate complexes with PhLi.
The presence of the ate complexes 1 and 2 was first detected kinetically by
measuring the rate of reaction of PhLi with electrophiles in the presence and absence of
PhI or Ph2Te [7g]. Rate reductions were observed when the metalloids were added.
This was attributed to the formation of ate complexes that are less reactive than the
PhLi in equilibrium with them. When the rate of metallation of 2-(methylthio)thio-
phene by PhLi was measured with increasing amounts of Ph2Te, PhI, or 10 present, it
decreased progressively, indicating the formation of the unreactive ate complexes
(Fig. 1). The larger decrease in rate for 10 vs. P hTe can be interpreted in terms of a 15-
2
fold larger association constant Kate for 10.
We also performed a variable-temperature 13C-NMR study on the PhLi/10 system
(Fig. 2). At room temperature, the spectrum is the superposition of the spectrum of 10
and PhLi. On lowering the temperature, the chemical shifts of all signals change as 10 is
converted to the ate complex 11. At À708 (Fig. 2,f), ate-complex formation is
essentially complete (Kate ˆ 2300 mÀ1, > 94% 11). Exchange between PhLi, 10, and 11
is still fast on the NMR timescale, although several of the peaks, notably the Co(Ph)
signal and C(2), are already quite broad due to beginning of decoalescence. Between
À708 and À908, the PhLi monomer-dimer exchange goes through decoalescence, and
the peaks for (PhLi)2 appear at their final shift positions. PhLi monomer is still in fast
exchange with 11 at À908 (note the absence of the PhLi monomer signal at d 143).
Between À908 and À1128, the ate-complex exchange of (PhLi)1, 10, and 11 becomes
slow on the NMR time scale, and at À1128, all but one of the expected 24 aromatic C-
signals are at least partially resolved: 4 for PhLi dimer, 4 for PhLi monomer, and 16 for
11 (the ipso signal for PhLi monomer at d 196.5 was not resolved). We were able to
assign all C-atoms except for distinguishing C(x), C(x') for C(2) to C(6). The C-atom of
the apical dibenzotellurophene ring was assigned to the upfield signal in each case

Helvetica Chimica Acta Vol. 85 (2002)
3751
Fig. 1. Inhibition of the rate of reaction of PhLi with 2-(methylthio)thiophene by the addition of PhI, Ph2Te, and 10
Fig. 2. a) 13C-NMR of 10 in THF at À 1108. b h) variable-temperature 13C-NMR studyof PhLi (0.16m) and 10
(0.08m) in THF (dotted lines mark the chemical-shift changes due to formation of 11, decoalescences of
monomer and dimer PhLi, and decoalescence of ate complex and PhLi; the graphics gives the 13C-NMR
chemical shifts of 11); i) temperature dependence and thermodynamic parameters for formation of 11

3752
Helvetica Chimica Acta Vol. 85 (2002)
without direct evidence, but consistent with the behavior of other tellurium and iodine
ate complexes.
The strong downfield shift of the two apical signals of C(1') and Cipso(Ph), at d 176
and 167 provide support for the structural assignment of 11. A variety of hypervalent
metalloid ate complexes with apical aryl groups have such downfield 13C-NMR signals
for the CÀM C-atom: Ph3TeLi (d 174.1) [7b][10a], Ph5TeLi (175.4) [7c], Ph2ILi (166.5)
[7a], Ph4ILi (172.4) [7c], Ph2SnMe3Li (180.0) [7d]. Neutral hypervalent molecules also
show this effect (e.g., in P hI the −apical× C-I signal is at d 161.9, the −equatorial× at d
3
125.2 [7c]).
Experiments with excess telluride (0.08m 10, 0.04m PhLi) showed that 10 and 11
were still in fast exchange at À1108, a process that can be assigned to bimolecular
transfer of phenyl between 10 and 11 (ate complex as phenyl-anion donor), a process
also identified for the Li/I exchange of PhLi/PhI [7a]. With excess PhLi present, the
concentration of 10 is very low, and bimolecular exchange becomes very slow, allowing
the static NMR spectrum of Fig. 2,h to be observed.
Analysis of the chemical shifts of the averaged 13C-NMR signals C(2) to C(6) allows
determination of the fraction of 11 at each temperature from 0 to À508. The amount of
uncomplexed phenyllithium remaining could then be determined at each temperature,
which, in turn, allowed calculation of the concentration of (PhLi)1 and (PhLi)2, by
means of the thermodynamic parameters for the monomer-dimer association constant
reported (DH ˆ À0.38 kcal/mol, DS8 ˆ À9.3 eu) [7h]. From this data, association
constants and thermodynamic parameters for the ate-complex equilibrium of (PhLi)1
and 10 to form 11 could be calculated. The negative entropy seen for this equilibrium is
precedented. Substantial negative entropies are generally seen for ArLi/ate complex
equilibria (ate complexes are favored at low temperature)1), as well as other contact
ion pair/separated ion pair equilibria2), a consequence of the additional solvent
molecules frozen by the free lithium cation. For similar reasons, Li/M exchanges also
show negative entropies of activation3).
The NMR study, like the kinetic study of Fig. 1, showed that 10 formed an ate
complex in THF, but not with the expected much larger association constant (e.g., 10 is
half-associated with PhLi in THF at À308, P hTe is half-associated at ca À708). Since
2
the increase in Kate for incorporation of the biphenyldiyl moiety in 10 was relatively
modest, we did not examine the Se and S analogs. Instead, we chose a m-terphenyl
framework to provide further pre-organization to maximize the stability of the Se and S
ate complexes. This mimics the expected T-shaped geometry and provides additional
entropic advantages since the anion addition to S or Se is intramolecular. Although the
m-terphenyl framework had not been previously used to stabilize hypervalent
metalloids, similar structures with −apical× O- and N-atoms, for example the (2,6-
dicarboxyphenyl)sulfuranide(1À) anion 9 [19a], have been used to stabilize ate
1
)
For equilibration of PhLi/PhI with Ph2ILi, DH8 ˆ À9.9 kcal/mol and DS8 ˆ À44 eu [1c]; for equilibration
of biphenyldiyldimethylsilane/MeLi with biphenyldiyltrimethylsilicate, DH8 ˆ À13.1 kcal/mol and DS8 ˆ
À53.8 eu [13].
2
3
)
For the equilibration of lithium fluorenide contact and separated ion pairs in THF, DH8 ˆ À7.5 kcal/mol
and DS8 ˆ À22 eu [25].
)
Negative entropies of activation have been found for the rates of intermolecular PhLi/ArBr exchange in
Et2O (DH≥ ˆ 14 kcal/mol, DS≥ ˆ À24 eu) [26a] and THF (223 K, DH≥ ˆ 9.5 kcal/mol, DS≥ ˆ À30 eu).

Helvetica Chimica Acta Vol. 85 (2002)
3753
complexes and neutral hypervalent sulfur [19b][20], phosphorus [14e][21], and tin [22]
compounds, among others.
Several precursors to the desired lithium reagents 12-S and 12-Se were prepared as
shown in Scheme 1. Ortho metallation of dibenzothiophene [2b] and dibenzoseleno-
phene [18a] allowed introduction of a D-atom (if needed) and a boronic acid moiety.
The H/D isotope effect was large enough that >90% of the second metallation of (4-
2H)dibenzothiophene occurred at the ortho' position. Boronic acids 13-S and 13-S-d
were coupled with o-iodobromobenzene (o-dibromobenzene gave substantial amounts
of bis-coupling product). Although this synthesis is short and provided a route to the D-
substituted compound 14-d, it was plagued by low yields and difficult purifications. A
more effective (albeit longer) route involved Suzuki coupling [23] of 13-S with o-
nitrochlorobenzene, reduction of the nitro group, and iododediazoniation to form the
iodide 17-S. To permit generation of halide-free lithium reagent 12-S, the iodide was
exchanged with BuLi, and coupled with bromotrimethylstannane to give 18-S. This
synthesis was not useful for preparing isotopically substituted deuterated analogs, since
D-scrambling occurred during the reaction of the diazonium compound with iodide.
Presumably, radical intermediates in this reaction [24] undergo rapid scrambling
through a sulfuranyl radical [19c].
Scheme 1. Synthesis of 12-S and 12-Se
Lithium-Sulfur Exchange. A first estimate of the rate of Li/S exchange was obtained
by preparing the deuterated lithium reagent 12-S-d by treatment of 14-d with tBuLi in
THF at À788 and quenching the reaction with dimethyl disulfide. The deuterium was
completely scrambled within 30 s, as shown by the 0.5 integration of the proton at d 7.8
in the produced methylthio derivative 19 (ˆ12-S with MeS instead of Li, assigned to
2
HÀC(2) (HÀC(2') was obscured by other signals) and by the H-NMR spectrum,

3754
Helvetica Chimica Acta Vol. 85 (2002)
which showed two signals at d 7.5 and 7.8 in a 1:1 ratio. A similar reaction in Et2O as
solvent showed less than 3% scrambling at À788 for 10 min (kTHF > 1500 ¥ kEt2O).
More definitive evidence for a rapid degenerate Li/S exchange was provided by a
13C-NMR spectroscopic study of 12-S prepared in THF/Et2O 80 :20 (Fig. 3). To avoid
interference by lithium halides, solutions of lithium reagents for spectroscopic studies
were prepared by Li/Sn exchange with low halide methyllithium. At À788 and lower,
the structure was static on the NMR timescale as evidenced by 18 signals in the
13C-NMR spectrum corresponding to the 18 C-atoms of 12-S, including the C-Li signal
at 197.4 ppm, a chemical shift strongly suggestive of a monomeric structure [27][7h].
All of the C-signals were assigned. Atoms C(2) and C(2') were directly identified by the
characteristic effect of deuteration (in 12-S-d), and the C(1), C(1'), C(3), and C(3')
were distinguished by the observation of a D-isotope shift. The quaternary atoms C(1),
C(6), C(6') C(7), C(7'), and C(10) were identified in a fully coupled spectrum; of these,
C(6) could be distinguished from C(7) since C(6) was a t due to two 3J(C,H) couplings
(2J(C,H) is < 2 Hz) , whereas C(7) was a d (similarly for C(6') and C(7')). C(10) was
the only quaternary C-atom not broadened in the spectrum at higher temperature
(Fig. 3). The signals of C(1) to C(6) were assignable and distinguished from those of
C(2') to C(6') based on their close similarity to model compounds such as the
trimethylstannyl derivative 18-S. A 13C,1H COSY experiment together with homo-
nuclear 1H-NMR decoupling experiments removed any remaining ambiguities in
assignment of all 1H- and 13C-signals in the NMR spectra. These assignments were also
consistent with those made for the Se analog 12-Se.
Fig. 3. 13C-NMR Spectra of a variable-temperature studyof a THF/Et O 80 :20 solution of 12-S (0.12m)
2
When the temperature was raised, spectral changes characteristic of a dynamic
NMR process occurred. These changes are uniquely consistent with a degenerate Li/S

Helvetica Chimica Acta Vol. 85 (2002)
3755
exchange of 12-S and 12-S'. Two of the signals in the 13C-NMR spectrum remained
sharp throughout, and these can be assigned to C(9) and C(10), whose environment is
unaffected by the process. At room temperature, three pairs of the more closely spaced
signals (C(3) and C(3'), C(4) and C(4'), and C(5) and C(5')) have largely coalesced
(see Fig. 4 for an analysis of the chemical shifts). Of these, the sharpest one (d 123.5) is
at the temperature extrapolated average of signals assigned to C(4) and C(4') which
have the smallest Dd (1.6 ppm) of all exchanging pairs (the excess broadening of signals
above coalescence is roughly proportional to the square of the chemical shift). The
other two (d 123.5, 125.0) have the next smallest Dd values (C(5) and C(5'), 2.8 ppm;
C(3) and C(3'), 3.0 ppm). All other signals have Dd > 7.5 ppm and are still too broad
to be seen at 258. The signal averaging is illustrated in Fig. 4,b, which shows the
calculated positions of the fully averaged spectrum.
Fig. 4. a) 13C-NMR Chemical shifts of 12-S in THF/Et2O 3 :97; b) calculated positions of the 13C-NMR chemical
shifts of 12-S under conditions of rapid exchange (12-S > 12-S'); c) 13C-NMR chemical shifts of fullycoalesced
12-Se in THF/Et2O 3 :97; d) 13C-NMR chemical shifts of 20-Se in THF/Et2O 80 :20, 5 equiv. of HMPA
(hexamethylphosphoric triamide) at À 788
The signals for C(1), C(6), C(7), C(7'), and C(8) in the spectra of Fig. 3 were
subjected to line-shape analysis4) to determine the rate of Li/S exchange. The
unaffected signals for C(9) and C(10) were used to establish the line width in the
absence of exchange. The rate data and an Eyring plot are shown in Fig. 5.
Attempts to force conversion of 12-S to the ate complex by addition of hexamethyl-
phosphoric triamide (HMPA), a technique that has been effective with other metalloids
[7a][7d][7g][13], were unsuccessful. Only decomposition products were detected even
when the sample temperature never exceeded À788. Similar solutions of 12-Se formed
the ate complex 20-Se quantitatively (see below) and were stable for weeks at À788.
Lithium-Selenium Exchange. The large stabilization of the transition state for the
exchange of 12-S compared to acyclic models encouraged us to synthesize the
analogous selenium compound 12-Se (Scheme 1) and study its behavior by NMR
spectroscopy. Fig. 6 shows variable-temperature 13C-NMR spectra of 12-Se in THF/
Et2O 80 :20. At temperatures between 68 and À1268 in this solvent mixture, as well as
4
)
Simulations were performed with a version of the computer program WINDNMR [7i].

3756
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 5. Rates and activation parameters for equilibration of 12-S/12-S' in THF/Et2O 80 :20. Also shown is the
effect on the rate of adding 3 equiv. of PMDTA (pentamethyldiethylenetriamine).
Fig. 6. a) 13C-NMR Spectra of a variable-temperature studyof a THF/Et 2O 80 :20 solution (0.12m) of 12-Se;
b) same sample with 5 equiv. of HMPA added at À 788; c) expansion of the C(1) signal of 20-Se showing the 77Se
satellites
between À798 and À1418 in THF/Me2O/Et2O 8 :67:25 and at À788 in THF/Et2O 3 :97,
at most ten signals, two of them with half-intensity, were visible in the 13C-NMR
spectrum. This is consistent only with the presence of a rapidly equilibrating mixture of
isomeric lithium reagents 12-Se and 12-Se' and/or the static selenium ate(1À) complex
20-Se. Several lines of evidence show that the isomers are in rapid equilibrium, and the
ate complex 20-Se is also present in significant concentrations in solutions containing
> 10% THF. Thus, in contrast to the behavior of 12-S, the chemical shifts of 12-Se in
THF/Et2O 80 :20 are significantly temperature-dependent (especially C(1), which goes
from d 171 at 68 to d 186 at À1268). The signals also show increased broadening at low
temperature as if a decoalescence was occurring. Although the central atoms C(9) and
C(10) are sharper than the others, they also broaden at the lowest temperature; so

Helvetica Chimica Acta Vol. 85 (2002)
3757
slowing of the exchange between 12-Se and 12-Se' cannot be the only process
responsible for the broadening observed.
A 77Se-NMR (7.5% natural abundance, spin ´) study clarified the situation (Fig. 7).
Between 68 and À1138, the Se-signal progressively moved upfield, and broadened
substantially below À788. Between À1138 and À1288, the signal decoalesced into two
peaks, a small one at 450 ppm very close to the original signal at 68, assignable to 12-Se,
and a second one at 405 ppm, which we have assigned to the ate complex 20-Se. We
were unable to achieve decoalescence in the 13C-NMR spectra because of a
combination of insufficient solubility, the resultant problems with sensitivity, and the
smaller chemical-shift changes (in Hz) for most of the 13C- vs. the 77Se-NMR signals.
Fig. 7. a) Variable-temperature 68.68-MHz 77Se-NMR spectra of 12-Se in THF/Et2O 80 :20; b) thermodynamics
of the 12-Se > 20-Se equilibrium, determined bymeasuring the ratio of the two components from the averaged
chemical shifts of the 77S-NMR signal and the 13C-NMR signals for C(1), C(2), and C(8) (DH8 ˆ À3.1 Æ 0.3 kcal/
mol, DS8 ˆ À16.7 Æ 1 eu)
Several experiments with different solvents and cosolvents supported the assign-
ments made above. In solvents containing only a few percent of THF (THF/Me2O/Et2O
8 :67:25 and THF/Et2O 3 :97), the 13C-NMR spectra showed ten C-signals, and the 13C-
and 77Se-NMR chemical shifts were essentially temperature-independent, consistent
with the absence of significant concentrations of 20-Se. The 12-Se and 12-Se' exchange
should have no effect on the 77Se-NMR signal, and none was detected. The 13C-NMR
signals broadened below À1008 as a result of incipient decoalescence between 12-Se
and 12-Se' although the spectrum was still above coalescence at À1418, the lowest
temperature that could be reached.
One consequence of the strong C-Li coordination in ArLi and the weak interaction
between lithium and the counter-ion in the ate-complex intermediates is that ate-
complex formation for the PhI/PhLi [7a], Ph2Te/PhLi [7c], and R3PhSn/PhLi [7d]
systems is strongly solvent dependent, with no detectable complex in Et2O solution but

3758
Helvetica Chimica Acta Vol. 85 (2002)
a significant amount in THF 5). HMPA is especially effective at stabilizing these ate-
complex structures, and that is the case for 12-Se as well. When 5 equiv. of HMPA was
added to a solution of 12-Se/20-Se in THF/Et2O 80 :20, sharp signals were observed in
both the 13C- and 77Se-NMR spectra. The NMR spectra are inconsistent with an
assignment of these signals to rapidly equilibrating 12-Se/12-Se'. This solution contains
only 20-Se, based on the following considerations.
A comparative analysis of the 13C-NMR chemical shifts of 12-S, 12-Se, and 20-Se is
shown in Fig. 4. Comparison of the calculated average signals of 12-S in THF/Et2O
3 :97 (Fig. 4,a and b) with the coalesced signals of 12-Se (Fig. 4,c) showed a close
similarity, with no signal differing by more than 1.3 ppm. On the other hand, the
comparison of the signals of 12-Se with those of a solution containing HMPA (20-Se,
Fig. 4,d) reveals large changes in chemical shifts (e.g., 17 ppm for C(1), 5 ppm for C(6)),
confirming that a major change in structure had occurred. The C(1) signal in 20-Se at d
187 is similar to that for the tellurium ate(1À) complex 11, and far from the d 167
predicted for a rapidly equilibrating structure (average of C(1), ca. 140 ppm, and C(1'),
ca. 195 ppm). In addition, a C,Se coupling (1J(C,Se) ˆ 85 Hz) was resolved for C(1). An
1
equilibrating structure should have a coupling of ca. 50 Hz, the average of J(C,Se) for
C(1) (ca. 100 Hz6) and C(1') (0 Hz). The 77Se-NMR chemical shift of d 411 is close to
the upfield signal at d 406 observed in the decoalesced spectrum Fig. 7 (À1358)7).
The 7Li-NMR is also fully consistent with conversion to 20-Se. At À1158, the broad
‡
s at À0.6 ppm for 12-Se is converted to the characteristic quint. of Li(HMPA)4 on
addition of HMPA, as expected for a separated ion 20-Se with lithium coupled to four
P-nuclei [7j].
Thermodynamics of the Li-Se Exchange Process. The 77Se-MR spectra in Fig. 7 and
13C-NMR spectra in Fig. 6 were used to determine the temperature dependance of the
equilibrium of 12-Se and 20-Se. The shifts for the averaged signal of the exchanging
species were used to calculate the equilibrium composition of each component and thus
DG8. Only the five spectra, from ‡ 68 to À998, could be used because below À998,
decoalescence took place. A 77Se-NMR chemical shift of 406 ppm was used for 20-Se
and 447 ppm for 12-Se. The 13C-NMR signals for C(1), C(2), and C(8) were chosen for
this analysis because they showed the largest chemical-shift change. The signals from
Fig. 4 for pure 12-Se (C(1) 169.4, C(2) 133.5, C(8) 121.1) and Fig. 6 for pure 20-Se (C(1)
186.4, C(2) 128.6, C(8) 117.5) were used as reference signals for the calculations, and
were assumed to be temperature-independent. A plot of DG8 vs. T for Kate is presented
in Fig. 7.
Effect of PMDTA (pentamethyldiethylenetriamine ˆ N-[2-(dimethylamino)ethyl]-
N,N',N'-trimethylethane-1,2-diamine). An intriguing and unexpected effect was
observed when PMDTA was added to a solution of either 12-S or 12-Se in THF/
Et2O 80 :20. For both systems, the presence of 3 equiv. of PMDTA slowed the exchange
5
)
This effect seems also to be present in the transition state leading to the lithium-metalloid exchange since
large solvent effects on rates are typically seen (e.g., from the rate data in Et2O for the PhLi/TolBr
exchange reported by Winkler and Winkler [26] in Et2O and our data in THF [7c], we calculate kTHF
/
kEt O ˆ 13600 at 08).
2
6
7
)
The 1J(C,Se) in dibenzoselenophene is 99.8 Hz, for compound 18-Se it is 100.2 Hz for C(2) and 98.2 Hz for
C(11) (numbering as shown for 12).
)
Small changes in d(Te) for the tellurium ate(1À) complex 2 in THF and THF/HMPA have been reported [10a].

Helvetica Chimica Acta Vol. 85 (2002)
3759
process by at least a factor of 200. For 12-Se, the rate reduction was large enough that
decoalescence could be achieved below À708, so that, at À1118, 16 of the 18 possible
signals were resolved. This rate reduction must have its origins in a relative
destabilization of the ate complex, since the 13C-NMR signals were essentially
temperature-independent down to À1418, and no signal for 20-Se was detectable in the
77Se-NMR spectra down to À1238, conditions under which > 80% of the material was
ate complex in the absence of PMDTA (Fig. 7). Evidently, complexation of PMDTA
stabilizes the lithium reagent 12-Se more than it does the ate complex 20-Se and the
transition state leading to it.
Kinetics of the Li-Se Exchange Process. The rate of the 12-Se/20-Se interchange
could be estimated from the 77Se-NMR spectra of Fig. 7 (À 78 to À1288) by means of
line-shape analysis by two-site exchange. Activation parameters in THF/Me2O 80:20
were DH≥ ˆ 5.7 Æ 0.3 kcal/mol, DS≥ ˆ À5 Æ 2 eu for the 12-Se/20-Se equilibrium.
Unfortunately, 77Se-NMR cannot be used to measure the key 12-Se/12-Se' intercon-
version since the degenerate exchange does not affect the selenium signals.
In principle, the rate of the 12-Se/20-Se exchange as well as the Li/Se exchange (the
12-Se/12-Se' interconversion), kd and kate in Eqn. 1, can be simultaneously determined
from analysis of the low-temperature 13C-NMR spectra, e.g., by the coalescence of the
C(2) and C(2') 13C-NMR signals. In spite of substantial efforts, this was unsuccessful,
partly because of signal-to-noise problems, but more importantly because that rate is
probably faster than can be detected by the dynamic NMR (DNMR) method. This will
be clearer when we discuss the unsymmetrical system 21 below. Thus, we were not able
to establish unambiguously that the ate complex 20-Se was an obligatory intermediate
in the exchange.
Since the 77Se-NMR studies were more effective than 13C-NMR work, we
synthesized the unsymmetrical derivative 21. In this compound, the two isomeric
lithium reagents 21a and 21b are no longer identical, and both the Li/Se exchange and
the exchange with the ate complex are measurable by 77Se DNMR studies. The
precursor trimethylstannyl derivatives 25a and 25b were prepared by a procedure
similar to that used for 12-Se starting with a Suzuki coupling [23] of 2-bromo-4-
isopropylacetanilide with the boronic acid 13-Se (Scheme 2). As was observed in the
preparation of 2H-labeled 17-S, iododediazoniation of the diazonium salt prepared
from 23 gave a 1:1 mixture of the two iodides 24a and 24b. Interestingly, one of the
isomers has a diastereotopic isopropyl group in the 1H- and 13C-NMR spectra,
Scheme 2. Synthesis of 21a/21b

3760
Helvetica Chimica Acta Vol. 85 (2002)
presumably the result of restricted rotation around the ArÀAr bond. Although the
presence of two isomers in the precursor iodide 24 and stannane 25 was a nuisance
synthetically, it was not of concern for the proposed study since the results with 12-Se
show that the lithium reagents 21a and 21b would scramble immediately on formation.
The NMR behavior of 21a/21b was analogous to that of 12-Se. In solvent mixtures
rich in THF at low temperature, the ate complex 26 formed in significant
concentrations. The defining variable-temperature NMR experiments for this system
are displayed in Fig. 8. In Me2O/Et2O 54 :46, no ate complex forms (signal expected at
404 ppm), but the two isomeric lithium reagents 21a and 21b can be decoalesced, with
77Se-NMR signals at d 447 and 453 ppm. They are present in a 26 :74 ratio (p-
tolyllithium and phenyllithium at equilibrium with their bromides are present in a
38 :62 ratio [26b]). A line-shape analysis4) between À1258 and À1468 (Fig. 8,a) gave
DGÀ≥ 136ˆ 6.22 kcal/mol, DH≥ ˆ 6.3 Æ 0.6 kcal/mol, and DS≥ ˆ À2.5 Æ 4.5 eu for the 21a to
21b equilibrium.
Fig. 8. Variable-temperature 68.68-MHz 77Se-NMR spectra of 21a/21b in a) Me2O/Et2O 54 :46 and b) THF/
Me2O/Et2O 12 :48 :40. Signal A and B correspond to 21a and 21b, signal C to the ate complex 26
A similar variable-temperature experiment for the sample in Fig. 8,a to which 12%
THF had been added is shown in Fig. 8,b. Below À1308, a new signal at d 404
decoalesced, which we have assigned to the ate complex 26. As for 12-Se (Fig. 7),
addition of more THF increased the fraction of the signal at d 404, and addition of
HMPA gave complete conversion to the ate complex (signal at d 411 ). Like the other
ate complexes, 26 showed characteristic downfield 13C-NMR shifts for the apical C-
atoms at d 181.9 and 187.2. Quenching of this solution with bromotrimethylstannane
gave 82% yield recovery of the stannyl precursors 25a and 25b (3 :2 ratio), showing
that the structure remained intact despite complete disappearance of the signals for
21a/21b in the 77Se-, 13C-, and 7Li-NMR spectra.

Helvetica Chimica Acta Vol. 85 (2002)
3761
What is remarkable is that, even though the ate-complex 77Se-NMR signal is
relatively sharp below À1458, the lithium reagents 21a and 21b are still in rapid
exchange. Although the lithium-reagent signals did show substantial broadening below
À14988), no sign of decoalescence was detected9). Thus, qualitatively, the Li/Se
exchange (interconversion of 21a and 21b, kd in Eqn. 1) is faster than the exchange of the
ate complex with the lithium reagents (kate).
We have attempted to quantitate this observation by performing a series of three-
site line-shape simulations on the spectrum at À1458. We make the assumption that
neither the 77Se-NMR chemical shifts of 21a and 21b nor their ratio at equilibrium was
changed when 12% of THF was added to the Me2O/Et2O mixture. To simplify
discussion, we assign symbols A, B, and C to 21a, 21b, and 26, respectively. Selecting the
inherent line width for the line-shape simulations is problematic since there is clearly
some broadening process at very low temperatures (see À1558 spectrum) that
selectively broadens the A/B signal compared with the C signal. We have used the line
width of the small amounts of protonated material for this purpose.
Fig. 9 shows simulations which demonstrate that the spectrum at À1458 cannot be
explained if exchange through the ate complex C is the only path (i.e., if kAB and kBA are
zero). If we fit the ate complex peak C, then A and B are well below coalescence
(simulation 2). If we increase kAC and kBC, then the ate complex peak C broadens
Fig. 9. 77Se-NMR Spectra and computer simulations of 21. The top spectrum is in a solution containing no THF,
the bottom one has 12% THF in the Me2O/Et2O mixture. Simulation 1 is a 2-spin simulation of the 21a/21b
exchange at À1468 in Me2O/Et2O (no ate complex present). Simulation 4 is the optimal 3-spin simulation of the
exchange 21a/21b/26, including a large rate for the A/B exchange. In simulations 2 and 3, only processes through
the ate complex C are allowed (kAB ˆ kBA ˆ 0)
8
)
The broadening of the peak at À1558 is possibly a viscosity effect at these very low temperatures. It cannot
be the result of the 21a/21b decoalescence since that requires an unsymmetrical peak shape (see the À1368
spectrum in Fig. 8,a).
9
)
The small shoulder on the side of the peak at 445 ppm in the spectra at À1308 to À1498 is the protonated
compound.

3762
Helvetica Chimica Acta Vol. 85 (2002)
(simulation 3). If kAC and kBC are increased sufficiently to coalesce A and B as in the
actual spectrum, then C is also coalesced with A and B (not shown). Only if we include
both a kAB term and the kAC and kBC terms, we can simulate the observed spectra
(simulation 4). Our best estimate is that kAB is at least 50 times as fast as kAC . To a first
approximation, 26 is not an intermediate in the exchange reaction, which must
principally occur by some sort of direct exchange (kd in Eqn. 1).
We can consider several possibilities for an exchange process that does not involve
the ate complex. Single-electron transfer (SET) reactions have been detected for the
metal-halogen exchange of alkyl bromides [28], but not for alkyl iodides or aryl halides
[8][29]. We can rule out SET mechanisms in the present system because the solutions
of 12-Se and 21 we have studied have undergone well in excess of 1010 Li/Se exchanges
in the course of the NMR experiment. There would have to be a cage effect of
unprecedented magnitude to avoid quenching if even a minuscule fraction of the
exchanges produced radical intermediates. Even the much slower Li/S exchange
occurred > 24000000 times without mishap during just the approximately 1 h it took to
measure the single 13C-NMR spectrum at 258 in Fig. 3 (k ˆ 6800sÀ1).
A better alternative would be a direct SN2 substitution, or an associative
substitution that avoids the spectroscopically detected form of 26. This behavior can
be formulated as follows. Aryllithium reagents like 21a and 21b are strong contact ion
pairs (CIP). The ate complex 26 is a separated ion pair (SIP). At some point in any
exchange process that involves 26, ion separation must occur. We can identify two
‡
limiting processes. Either CÀLi first forms the SIPC À//Li and then CÀ adds to the
‡
CÀSe s* orbital to give CÀSeÀCÀ//Li , or CÀLi adds to CÀSe bond to form the CIP
ate complex CÀSeÀCLi, which can then undergo ion-pair separation to the more
‡
stable SIPC ÀSeÀCÀ//Li (Scheme 3). A very satisfactory explanation for our NMR
experiment is that the rate of dissociation of the CIP-Ate to CIP-Li is faster than its rate
of ion-pair separation to form SIP-Ate, so the reaction effectively bypasses the
observed SIP-Ate complex. Another way of looking at this is in terms of the detailed
mechanism of the conversion of ate complex to lithium reagents. This is an electrophilic
‡
‡
attack of Li on the weak hypervalent CÀSe bond of the ate complex. Since the Li of
the ate complex will be coordinated to four solvent molecules, it is not a very effective
electrophile. If Li loses a solvent molecule and becomes CIP-Ate, the electrophilic
‡
substitution then becomes facile. The SIP-Ate complex is thus a parasitic equilibrium
on the optimum exchange mechanism which proceeds almost entirely through CIP
intermediates (top row in Scheme 3).
It is interesting to note that the free energy of activation (DG=À151 ˆ 5.3 kcal/mol) we
reported for the interconversion of the CIPand SIPof the lithium reagent (ArS) CHLi
2
(Arˆ 3,5-bis(trifluoromethyl)phenyl) [7k] is very close to those measured here for the
A/B to C interconversion at À1468 (DG=AB < 5.3 kcal/mol, DGA=C ꢀ 6.2 kcal/mol)
suggesting that the principal barrier for the Li/Se exchange in the very favorable
geometry of compounds 12-Se and 21 might be association and dissociation of
coordinated THF solvent.
Intermolecular Li/S, Li/Se, and Other Li/M Exchanges. Fig. 10 shows a comparison
of the intramolecular Li/S and Li/Se exchange rates (either kate or kd) reported in this
paper with intermolecular analogs. The Li/I and Li/Te rates were determined from
published spectra [7a][7b], the others were measured by means of the reaction of p-

Helvetica Chimica Acta Vol. 85 (2002)
3763
Scheme 3. Possible pathways for Li/Se exchange. CIP ˆ contact ion pair, SIP ˆ separated ion pair.
TolLi with PhCl, PhBr, Ph2S, and Ph2Se in THF (Eqn. 2) [7c][6a]. The temperature
dependence was measured only for the Li/Br exchange (DGÀ≥ 150ˆ 16.1 kcal/mol, DH≥ ˆ
9.5 kcal/mol, DS≥ ˆ À29.6 eu). The rate for the very slow exchanges Li/S and Li/Cl
could be an upper limit, since there was substantial decomposition of the p-TolLi even
at very low conversions. The comparisons can only be semiquantitative because some of
the rates are for exchanges of RLi with the ate complex that, as we have seen, can be
slower processes than the Li/M exchange itself. In addition, different solvent mixtures
were employed. The intramolecular exchanges had to be run in weaker solvents such as
mixtures of THF, Me2O, and Et2O, which would lower the rates significantly compared
to the intermolecular exchanges, which were run in pure THF.
k2
PhÀSÀPh ‡ p-TolLi
„
PhÀSÀp-Tol ‡ PhLi
(2)
It can be seen that there are very large periodic effects on the rates, with the PhCl/
PhBr/PhI series ca. 1 :108 :1013 in k1. Similarly, the Ph2S/Ph2Se/Ph2Te series is 1:104 :1014
(the PhI and Ph2Te rates are rather inexact extrapolations from DNMR work at
temperatures below À1008). There has been substantial computational effort at
determining energies and structures of ate complexes of the main-group elements [30],
and the calculated energies are consistent with our experimental data.
One measure of the acceleration of an intramolecular process is the −effective
molarity×, defined as the concentration of a bimolecular reactant needed to give the
same rate as an intramolecular process. The extrapolated rate constant for the
degenerate exchange of 12-S at 08 is 1100 sÀ1. The second-order rate constant for
approach to equilibrium of Ph2S and p-TolLi at 08 (Eqn. 3, see below) is ꢁ 2 ¥ 10À7
l mÀ1 sÀ1 (DG≥ ˆ 24.3 kcal/mol). Thus the effective molarity is >5 ¥ 109 for 12-S. It is
qualitatively similar for the p-TolLi/Ph2Se-21 comparison.
The origin of these enormous rate effects is not clear. Five-membered rings in
tetracoordinate Si- and P-compounds, with natural bond angles close to tetrahedral, are
strained, whereas in pentacoordinate hypervalent structures, the apical-equatorial (or
basal) angles are close to 908 and small rings are less strained, resulting, for example, in
enhanced rates of cyclic phosphate ester hydrolysis. However, this −small-ring effect× is

3764
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 10. Summaryof ArLi-halogen and ArLi-chalcogen exchange rates, measured byseveral different techniques .
The DG≥ values are not strictly comparable since they were measured in different ethereal solvents, as indicated.
Two of them (open and filled triangles) correspond to kate (exchange of ArLi with the ate complex), others to an
*
authentic Li/M exchange (kd). The points were measured by means of the PhnM/TolLi exchange in THF
(Eqn. 2). Second-order rates were converted to pseudo-first-order rates by dividing k2 by the concentration of
ArLi. The lines are least-squares fit.
likely to be only a small fraction of the rate acceleration seen here, since 10 showed only
about one order of magnitude increase in Kate (Fig. 1) compared to acyclic analogs. A
rationale is that the natural bond angle in the heavy chalcogens is close to 908 (H2S
92.18, H2Se 90.68 H2Te 90.38). The C-Y-C bond angles in selenides and tellurides are
thus below the tetrahedral angle (e.g., in P hSe, the angle is 1068, in P hTe it is 1018
2
2
[31]), so there is less release of strain in cyclic hypervalent structures. The major
portion of the increase in Li/S and Li/Se exchange rates in 12-S, 12-Se, and 21 must be a
consequence of almost ideal preorganization for 3-center nucleophilic attack of the
CÀLi bond on the CÀY antibonding orbital [7c].
Conclusions.
The m-terphenyl framework in structures 12-S, 12-Se, and 21
provides an ideal geometry for the Li/M exchange, to the extent that the Li/S exchange,
which is barely detectable at room temperature in acyclic systems, becomes fast on the
NMR timescale, a rate acceleration in excess of 109. The Li/Se exchange is similarly
accelerated, and the selenium ate complexes 20-Se and 26 become spectroscopically
detectable at low temperature in THF-containing mixed solvents. A DNMR experi-
ment showed that a direct exchange not involving the ate complexes was ca. 50 times as
fast as exchange of the lithium reagent with the ate complexes, an observation that
could be rationalized in terms of the contact-ion-pair and separated-ion-pair
intermediates. These effects are unlikely to have their principal origins in ring-strain
effects, since the ate complex association of PhLi with the dibenzotellurophene 10 was
only slightly increased as compared to Ph2Te.

Helvetica Chimica Acta Vol. 85 (2002)
3765
Experimental Part
General. All reactions involving lithium reagents were run in oven- or flame-dried flasks under dry N2.
Tetrahydrofuran (THF) and Et2O were freshly distilled from benzophenone ketyl (ˆoxidodiphenylmethyl).
Me2O was obtained from Aldrich and dried with MeLi prior to use. Hexamethylphosphoric triamide (HMPA)
was obtained from Aldrich, distilled from CaH2 at reduced pressure (0.5 mm), and stored under N2 over
molecular sieves. Starting materials were commercially available or a literature citation for their preparation is
given. FC ˆ flash column chromatography. GC: SE-30 capillary columns, flame-ionization detector; mea-
surements corrected for response factors. Multinuclear NMR spectra: Bruker AM-360 spectrometer operating
at 90.56 MHz (13C), 139.962 MHz (7Li), and 68.68 MHz (77Se); 13C-NMR, ca. 800 transients and exponen-
tial multiplication of
2
5 Hz; d(C) in ppm rel. to internal C6H6 (d 129.18), THF (C(2), d 67.96),
20 Hz
Me2O (d 60.08), or Et2O (C(2), d 60.51), J in Hz; 77Se-NMR, exponential multiplication of
2
depending on broadening of signals, spectra referenced externally to Me2Se in CDCl3 at r.t.; temps.
of the experiments measured with an external carbon-13 thermometer (CCl4/D6 acetone 1:1 [32]), a resis-
tance temp. device (RTD), or the tris(trimethylsilyl)methane internal d(C) thermometer [7l]; 20 min
were allowed between acquisitions for the temp. to equilibrate; temps. were generally measured
regularly during the experiment and after each measurement when accurate variable temp. studies were being
performed.
Caution: HMPA is carcinogenic, and organotin and organoselenium compounds are toxic and should be
handled in a hood with due care.
Note. The systematic numbering used for the compound names below is not the conceptually convenient ad
hoc numbering used in Figs. 2 4, 6 and 7.
2,2'-Diiodo-4,4'-dimethyl-1,1'-biphenyl. Into a 2-l beaker was added 4,4'-dimethyl[1,1'-biphenyl]-2,2'-
diamine [33] (38.7 g, 0.182 mol) dissolved in conc. HCl soln. (184 ml) and ice (184 g). The soln. was stirred and
kept below 28 while NaNO2 (27.6 g, 0.40 mol, 2.2 equiv.) in H2O (90 ml) was added dropwise over 30 min. After
15 min, the diazonium-salt soln. was filtered through glass wool/ice into a 4-l beaker containing a cold soln. of KI
(90.6 in 175 ml H2O). During the addition, the KI soln. was kept below 58 and vigorously stirred. Stirring was
continued for 2 h at r.t., excess sodium hydrogensulfite was added, and the mixture was left overnight. The
mixture was filtered, and the brown precipitate was washed with H2O and placed in a 1-l beaker containing
MeOH (800 ml). After boiling for 10 min, the soln. was filtered leaving crude 2-iodonio-4,4'-dimethyl-1,1'-
biphenyl iodide, which was recrystallized from DMF to yield 7.1 g (9%). M.p. 241 2438.
The MeOH filtrate was evaporated and the brown precipitate dissolved in Et2O (500 ml) and extracted
with 3 Â 200 ml portions of 6n HCl. The org. layer was washed with 7% NaHCO3 and NaCl soln. and dried
(Na2SO4). Evaporation gave 14 g of crude 2,2'-diiodo-4,4'-dimethyl-1,1'-biphenyl, which was recrystallized from
Et2O to yield 10.5 g (13.3%). M.p. 108 1138 [14f]. 1H-NMR (CDCl3, 200 MHz): 7.8 (s, 2 H); 7.25 (d, J ˆ 8, 2 H);
7.1 (d, J ˆ 8, 2 H); 2.3 (s, 6 H). 13C-NMR (CDCl3, 50.1 MHz): 146; 139; 130; 128; 99; 20.6.
The acid washings were made basic with 20% NaOH soln., yielding 26 g of crude 3,8-dimethylbenzo[c]-
cinnoline. Recrystallization from AcOEt gave 22.6 g (59.6%). M.p. 186 1888. 1H-NMR (CDCl3, 200 MHz):
8.95 (s, 2 H); 8.38 (d, J ˆ 9, 2 H); 7.68 (d, J ˆ 9, 2 H); 2.6 (s, 6 H).
3,7-Dimethyldibenzotellurophene (10). Into a dried round-bottom flask (500 ml, 3-necked) was placed 2,2'-
diiodo-4,4'-dimethyl-1,1'-biphenyl [14f] (15,75 g, 36.3 mmol). The apparatus was purged with N2 for 2 h, THF
t
(100 ml) added, the soln. cooled to À788, and 1.85m BuLi (86.5 ml, 160 mmol, 4.4 equiv.) added dropwise. The
soln. was warmed to 08 for 20 min, and 2,5-dihydro-3-methyltellurophene 1,1-dichloride (prepared from isoprene
and TeCl4 [34]) (40 mmol, 10.64 g, 1.1 equiv.) in THF (100 ml) was added. After 3 h, solvents were evaporated,
and the solid was dissolved in H2O (200 ml) and CH2Cl2 (200 ml). The org. layer was washed with sat. brine,
filtered through Na2SO4, and evaporated to give 11 g of crude 10. Recrystallization from hexane (decolorized
with charcoal) yielded 9.4 g (84%) of 10. M.p. 156 1578 [14c]. 1H-NMR (CDCl3, 200 MHz): 7.9 (d, J ˆ 8); 7.7 (s,
2 H); 7.24 (d, J ˆ 8, 2 H); 2.33 (s, 6 H). 13C-NMR (CDCl3, 50 MHz): 141.6; 136.4; 132.7; 123.8; 127; 21.2.
Typical Preparation and Standardization of PhLi Solution for Kinetic Studies. An Erlenmeyer flask (50 ml,
24/40) was dried, fitted with septum, and purged with N2. Stock 1.5m PhLi/THF (prepared from halide-free PhLi
crystallized at least twice from Et2O [7h]; 20 ml, 30 mmol) was added and diluted with THF (10 ml). To the soln.
was added undecane (0.634 ml, 3.0 mmol) as a GC standard. A 1.0-ml aliquot of the soln. was syringed into each
of three dry, purged 5-ml round-bottom flasks equipped with septa and stirring bars. Each was quenched with
100 ml (1.1 mmol) of MeSSMe; sat. NH4Cl soln. (ca. 0.10 ml) was added (white precipitate), and the solns. were
dried (Na2SO4). PhI was accurately weighed into each soln. (ca. 40 ml) as GC standard. Analysis of the solns. by
GC gave the concentrations of PhLi and undecane to be 0.90m and 0.11m, resp.

3766
Helvetica Chimica Acta Vol. 85 (2002)
Effect of 10 on the Reaction of PhLi with 2-(Methylthio)thiophene. Six round-bottom flasks (5 ml, long-
necked) were dried, and 10 was weighed into each in the following amounts: 0 g, 0.0723 g (0.2349 mmol,
0.5 equiv.), 0.0279 g (0.09063 mmol, 1 equiv.), 0.1353 g (0.4395 mmol, 1.5 equiv.), 0.0825 g (0.268 mmol,
2.0 equiv.), and 0.2089 g (0.6786 mmol, 2.5 equiv.). The flasks were equipped with stirring bars and septa,
and purged with N2 for 25 min. THF was added to each in the following amounts: 1.25 ml, 1.95 ml, 0.4 ml,
1.22 ml, 0.56 ml, and 1.13 ml. The solns. were cooled to À788, and 1 equiv. of PhLi (1.20m) was added in the
following amounts: 0.25 ml (0.30 mmol), 0.40 ml (0.47 mmol), 0.076 ml (0.0906 mmol), 0.22 ml (0.293 mmol),
0.112 ml (0.134 mmol), and 0.244 ml (0.2714 mmol). After 10 min, 1 equiv. of 2-(methylthio)thiophene was
added in the following amounts: 32 ml (0.30 mmol), 50 ml (0.47 mmol), 9.7 ml (0.0906 mmol), 31.3 ml
(0.293 mmol), 14.3 ml (0.134 mmol), and 29 ml (0.2714 mmol). After 60 min, the reactions were quenched with
MeSSMe (75 ml) and the mixtures allowed to warm to r.t. NH4Cl (50 ml) was added, and the solns. were dried
(Na2SO4). GC Analysis gave the results shown in the Table.
Table 1. Reaction of PhLi with 2-(Methylthio)thiophene in the Presence of 10
PhLi [mmol/equiv.]
10 [equiv.]
PhSMe [mmol]
Product [mmol]a)
% Reaction
% Recovery
0.30
0.47
0.0906
0.293
0.134
0.271
0
0.087
0.2021
0.0462
0.190
0.105
0.220
0.189
0.249
0.039
0.0703
0.0255
0.0352
63
53
43
24
19
13
92
96
94
89
97
94
0.5
1.0
1.5
2.0
2.5
a) Product ˆ 2,5-bis(methylthio)thiophene.
Similar experiments were performed with Ph2Te [35] and PhI. Results are summarized in Fig. 1.
NMR Spectroscopyof PhLi and 10 in THF. Into a dry 10-mm NMR tube was accurately weighed 0.098 g
(0.317 mmol) of 10, and the tube was fitted with a septum and purged with N2. THF (3.5 ml) was added and the
sample cooled to À788, keeping positive N2 pressure in the tube. Stock 1.2m PhLi soln. (0.516 ml, 0.64 mmol)
was added, and the 13C-NMR spectra were obtained at a series of temps. Spectra and d(C) data are reported in
Fig. 2. Similar experiments at several other PhLi/10 ratios confirmed the assignments made.
(4-2H)Dibenzothiophene. To a dry 50-ml flask fitted with a septum was added dibenzothiophene (0.5 g,
2.71 mmol). The flask was cooled to À788 under N2 and 1.06m s-BuLi (3.8 ml, 4.03 mmol) was added over a few
minutes. The yellow-orange soln. was stirred at À788 for 2 h, quenched with D2O (100 ml), and allowed to warm
to r.t. The mixture was partitioned between Et2O/hexane and H2O and the org. layer dried (Na2SO4) and
evaporated: 0.483 g (95%) of a white solid. The product was used without further purification. A small amount
was recrystallized from hexane for analysis. White needles. M.p. 97.0 98.58. 1H-NMR (200 MHz, CDCl3): 8.12
(m, 2 H); 7.82 (m, 1 H); 7.42 (m, 4 H). 13C-NMR (75.403 MHz, CDCl3): 139.35 (s); 139.23 (s, D-isotope shift);
135.6 (s); 126.61 (d); 126.49 (d, D-isotope shift); 124.3 (d); 122.7 (d); 122.5 (1:1:1 t, CD); 121.5 (d). HR-MS:
‡
‡
185.0398 (M , C12H7DS ; calc. 185.0409).
Dibenzothiophene-4-boronic Acid (13-S). Dibenzothiophene (0.4 g, 2.17 mmol) was metallated as
described above. Trimethyl borate (0.62 ml, 5.4 mmol) was added at À788, the soln. warmed to r.t., and then
cooled in an ice bath, and 1n HCl (10 ml) added. The mixture was diluted with Et2O and extracted twice with 1n
NaOH. The alkaline extract was washed with Et2O and acidified with 3 n HCl. The white precipitate was filtered
and washed with H2O and Et2O to afford 0.403 g (81%) of 13-S. White solid. The crude product, which did not
melt or decompose under 3008, was used in the next step. 1H-NMR (200 MHz, (D6)acetone/(D6)DMSO): 8.37
(dd, J ˆ 7.8, 1.3, 1 H); 8.27 (m, 1 H); 8.03 (dd, J ˆ 7.0, 1.3, 1 H); 7.93 (m, 1 H); 7.65 (br. s, 2 H); 7.50 (m, 3 H).
13C-NMR (75.403 MHz, CDCl3): 146.1 (s); 141.4 (s); 135.7 (s); 135.7 (s); 134.4 (d); 127.1 (d); 124.1 (d); 124.3
‡
‡
(d); 124.2 (d); 123.0 (d); 122.0 (d). HR-MS: 228.0419 (M , C12H9BO2S ; calc. 228.0416).
4-(2-Bromophenyl)dibenzothiophene (14-S). A mixture of benzene (15 ml), 13-S (0.296 g, 1.3 mmol), 1-
bromo-2-iodobenzene (0.44 g, 1.1 mmol), [Pd(PPh3)4] (0.50 g, 3 mol-%), 2m Na2CO3 (1.3 ml), and MeOH
(5 ml) was heated to 50 608 under N2 (TLC monitoring, 1% AcOEt/hexane). After 5 h, the product spot was
predominant. The mixture was allowed to stir overnight at r.t. and worked up by partitioning between Et2O/
hexane and H2O. The org. layer was extracted with H2O, 10% NH4OH soln., H2O, and brine, dried (Na2SO4),

Helvetica Chimica Acta Vol. 85 (2002)
3767
and evaporated. The product was passed through a plug of silica gel (2% AcOEt/hexane), excess 1-bromo-2-
iodobenzene was removed by distillation, and the residue was submitted to FC. The eluate (0.25 g) was
recrystallized from CH2Cl2/MeOH: 0.19 g (43%) of 14-S. White crystals. M.p. 85 958 (small amount of Ph3P
contamination). The reaction was not easily reproducible, yields varied from 20 50%. 1H-NMR (270 MHz,
CDCl3): 8.17 (m, 2 H); 7.77 (m, 1 H); 7.74 (dd, J ˆ 7.4, 1.5, 1 H); 7.55 (t, J ˆ 7.5, 1 H); 7.40 (m, 4 H); 7.35 (dd, J ˆ
7.0, 1.4, 1 H); 7.30 (ddd, J ˆ 8.2, 7.0, 2.1, 1 H).
1
The (6-2H) compound 14-S-d was prepared similarly, the H-NMR signal at d 7.77 was missing.
4-[2-(Methylthio)phenyl]dibenzothiophene (19). To an oven-dried and N2-purged 10-ml conical flask, 4-(2-
iodophenyl)dibenzothiophene (17-S; 20.4 mg, 0.0528 mmol) was added. The flask was repurged with N2 for
10 min. Freshly distilled Et2O (3 ml) was added, the soln. cooled to À788, and 2.57m BuLi (0.08 ml,
0.2056 mmol) added. The reaction was quenched with Me2S2 (0.034 ml, 0.3774 mmol) and the mixture allowed
to warm to r.t., diluted with hexanes/Et2O 1:1, washed with H2O (3Â) and brine, and dried (MgSO4). TLC
purification gave 9.8 mg (61% of 19). 1H-NMR (499.89 MHz, CDCl3): 8.20 (dt, 2 H); 7.77 (m, 1 H); 7.56 (t, J ˆ
7.5, 1 H); 7.46 (m, 4 H); 7.40 (m, 1 H); 7.29 (td, J ˆ 7, 1, 1 H); 2.35 (s, 3 H). 13C-NMR (125.70 MHz, CDCl3 ):
140.0 (s); 139.7 (s); 138.9 (s); 137.9 (s); 135.9 (s, 2C); 135.3 (s); 129.6; 128.9; 127.7; 126.7; 125.7; 124.9; 124.5;
‡
‡
124.3; 122.7; 121.7; 120.9; 16.0. HR-EI-MS: 306.0528 (M , C19H14S2 ; calc. 306.0537).
The deuterated compound 19-d was prepared similarly, the 1H-NMR signal at d 7.77 was missing.
Experiments performed in THF with 19-d showed a 0.5-H integration for the signal at d 7.77. 2H-NMR showed
two peaks at d 7.8 and 7.5 in a 1:1 ratio.
4-(2-Nitrophenyl)dibenzothiophene (15-S).
A soln. of dibenzothiophene-4-boronic acid (2.334 g,
10.2 mmol), 1-chloro-2-nitrobenzene (1.50 g, 9.52 mmol), benzene (100 ml), MeOH (10 ml), and 2m Na2CO3
(10 ml, 2 equiv) was sparged with N2 for 20 min before adding [Pd(PPh3)4] (0.37 g, 0.32 mmol). The mixture was
stirred at 708 for 7 h, allowed to cool to r.t., and partitioned between Et2O/hexane (200 ml) and H2O. The org.
layer was washed with H2O, 10% NH4OH soln., H2O, and brine, dried (Na2SO4), and evaporated. Purification
by a short silica-gel column (30% CH2Cl2/hexane), collecting the yellow band, gave 3.13 g (100%) of 15-S.
Yellow solid which was used in the next step. A small amount was recrystallized from CH2Cl2/MeOH for
analysis. M.p. 137 1398. 1H-NMR (200 MHz, CDCl3 ): 8.15 (m, 2 H); 8.07 (m, 1 H); 7.60 (m, 7 H); 7.30 (dd, J ˆ
7.0, 1.3, 1 H). 13C-NMR (90.56 MHz, CDCl3 ): 148.7 (s); 139.2 (s); 139.0 (s); 136.0 (s); 135.6 (s); 134.8 (s); 133.0
(d); 132.6 (s); 131.9 (d); 129.2 (d); 126.9 (d); 126.2 (d); 124.9 (d); 124.7 (d); 124.6 (d); 122.7 (d); 121.9 (d); 121.3
‡
‡
(d). HR-EI-MS: 305.0495 (M , C18H11NO2S ; calc. 305.0510).
2-(Dibenzothiophen-4-yl)benzenamine (16-S). A mixture of 4-(2-nitrophenyl)dibenzothiophene (2.83 g,
9.26 mmol), AcOH (26 ml), EtOH (60 ml) and Fe-powder (2.13 g, 38 mmol) was refluxed for 5 h (TLC
monitoring, 10% AcOEt/hexane), cooled to r.t., and added to H2O (100 ml). Solid Na2CO3 was added carefully
to the mixture until CO2 bubbling had stopped. The greenish slurry was diluted with H2O and extracted with
Et2O (3Â). The org. layer was washed with H2O, dried (Na2SO4) and concentrated to 50 100 ml. The Et2O
soln. was cooled in an ice bath, and conc. HCl soln. was added dropwise until no more of the solid amine
hydrochloride formed. The solid was filtered, washed with cold H2O and cold Et2O and allowed to air dry to give
2.263 g (78%) of the HCl salt as a slightly off-white solid, which was used without further purification. A small
amount of amine 16-S was liberated from the HCl salt for analysis. 16-S: 1H-NMR (200 MHz, CDCl3): 8.16 (m,
2 H); 7.78 (m, 1 H); 7.55 (t, J ˆ 7.5, 1 H); 7.46 (m, 3 H); 7.25 (m, 2 H); 6.86 (m, 2 H); 3.46 (br. s, 2 H). 13C-NMR
(90.56 MHz, CDCl3 ): 143.5 (s); 139.9 (s); 139.5 (s); 135.9 (s); 135.6 (s); 133.9 (s); 129.9 (d); 129.2 (d); 127.5 (d);
126.6 (d); 125.4 (s); 124.9 (d); 124.2 (d); 122.6 (d); 121.6 (d); 120.5 (d); 118.3 (d); 115.7 (d). HR-EI-MS:
‡
‡
275.0765 (M , C18H13NS ; calc. 275.0768).
4-(2-Iodophenyl)dibenzothiophene (17-S). A mixture of 16-S ¥ HCl (1.955 g, 6.27 mmol), conc. HCl soln.
(2.5 ml), ice (100 g), and H2O (10 ml) was cooled in an ice bath, and an aq. NaNO2 soln. (0.455 g, 6.60 mmol)
was carefully added with stirring. After 2 h, an aq. KI soln. (5.2 g, 31 mmol) was added to the bright yellow soln.
with vigorous stirring (brownish color). After stirring for 2 h at 08 and 2 h at r.t., the mixture was extracted with
Et2O/hexane (3Â) and the org. layer washed with H2O and dried (Na2SO4). The dark viscous liquid was passed
through a short silica-gel plug (5% AcOEt/hexane) to remove most of the dark impurities, and the eluate was
purified by FC (15% CH2Cl2/hexane). Recrystallization from CH2Cl2/MeOH gave 1.573 g (65%) of 17-S. White
crystals. M.p. 89.0 91.08. 1H-NMR (200 MHz, CDCl3): 8.20 (m, 2 H); 8.03 (m, 1 H); 7.79 (m, 1 H); 7.55 (t, J ˆ
7.5, 1 H); 7.45 (m, 4 H); 7.33 (dd, J ˆ 7.5, 1.2, 1 H); 7.15 (ddd, J ˆ 8.0, 6.0, 3.0, 1 H). 13C-NMR (90.56 MHz,
CDCl3 ): 145.0 (s); 139.7 (s); 139.5 (s); 139.5 (d); 139.2 (s); 135.7 (s); 135.7 (s); 129.9 (d); 129.7 (d); 128.3 (d);
‡
127.4 (d); 126.8 (d); 124.5 (d); 124.4 (d); 122.7 (d); 121.8 (d); 120.9 (d); 98.9 (s). HR-EI-MS: 385.9611 (M ,
‡
C18H11IS ; calc. 385.9622).

3768
Helvetica Chimica Acta Vol. 85 (2002)
4-[2-(Trimethylstannyl)phenyl]dibenzothiophene (18-S). To dry Et2O (20 ml) at À788 under N2 was added
2.07m BuLi (0.80 ml, 1.66 mmol) and, by cannula, an Et2O/THF (3.0 ml/0.5 ml) soln. of 17-S (0.496 g,
1.28 mmol). After 5 min, an Et2O (3.0 ml) soln. of Me3SnBr (300 ml, 1.5 equiv.) was added and the soln. stirred
for 5 min, allowed to warm to r.t., and quenched with H2O. The mixture was partitioned between Et2O/hexane
and H2O and the org. layer washed 3 Â with H2O, dried (Na2SO4), and evaporated. The product was passed
through a silica-gel plug (5% AcOEt/hexane), and then purified by prep. HPLC (reversed phase MeCN, flow
rate 10 ml/min; tR (protonated compound) 13 s, tR (product) 18.5 s): 0.40 g (74%) of 18-S, which crystallized
upon standing. M.p. 75.0 76.08. 1H-NMR (500 MHz, CDCl3 ): 8.19 (m, 1 H); 8.15 (dd, J ˆ 7.5, 1.0, 1 H); 7.79 (m,
1 H); 7.66 (m, 1 H); 7.53 (t, J ˆ 7.5, 1 H); 7.53 (m, 1 H); 7.44 (m, 4 H); 7.30 (dd, J ˆ 7.2, 1.1, 1 H); À0.14 (s,
J(Sn,H) ˆ 54.9, 52.6, 9 H). 13C-NMR (90.56 MHz, THF): 148.9 (s, J(Sn,C) ˆ 29.3); 143.2 (s); 140.8 (s); 140.6 (s,
J(Sn,C) ˆ 19.0); 140.5 (s); 137.3 (d, J(Sn,C) ˆ 35); 136.6 (s); 136.5 (s); 129.0 (d); 128.9 (d, J(Sn,C) ˆ 9.5); 127.9
(d, J(Sn,C) ˆ 45); 127.9 (d); 127.4 (d); 125.3 (d); 125.0 (d); 123.2 (d); 122.5 (d); 121.2 (d); À8.6 (q, 1J(Sn,C) ˆ
‡
‡
352.8, 337.5, 3C). HR-EI-MS: 409.0078 ([M À CH3] , C20H17SSn ; calc. 409.0073).
Dibenzoselenophene-4-boronic Acid (13-Se). A soln. of dibenzoselenophene [18b] (2.00 g, 8.65 mmol) in
t
THF (50 ml) under N2 was cooled to À788, and 1.5m BuLi (11 ml, 16.5 mmol) was added. After 5 h at À788,
trimethyl borate (4 ml, 35 mmol) was added. The soln., which turned colorless immediately upon the addition,
was allowed to warm to r.t., and worked up as described for 13-S: 1.93 g (81%) of 13-Se. White solid. The crude
product, which did not melt or decompose below 3008, was used without further purification: 1H-NMR
(200 MHz, (D6)acetone/(D6)DMSO): 8.34 (dd, J ˆ 7.9, 1.3, 1 H); 8.26 (m, 1 H); 8.19 (br. s, 2 H); 8.01(m, 2 H);
7.50 (t, J ˆ 7.6, 1 H); 7.41 (m, 2 H). 13C-NMR (75.403 MHz, CDCl3): 146.8 (s); 141.5 (s); 138.5 (s); 138.4 (s);
134.2 (d); 127.1 (d); 126.2 (d); 125.4 (d); 125.0 (d); 124.8 (d); 123.1 (d); 109.6 (s).
4-(2-Nitrophenyl)dibenzoselenophene (15-Se). As described for 15-S, with 13-Se (2.12 g, 7.73 mmol), 1-
chloro-2-nitrobenzene (1.16 g, 7.34 mmol), [Pd(PPh3)4] (0.31 g, 3.5 mol-%), MeOH (10 ml), 2m Na2CO3 (8 ml,
2 equiv), and benzene (100 ml). The product was purified by passing it through a short silica-gel plug: 2.76 g
(100%) of 15-Se. Yellow solid that was pure enough for use in the next step. 1H-NMR (200 MHz, CDCl3 ): 8.15
(m, 2 H); 8.07 (m, 1 H); 7.78 (m, 1 H); 7.51 (m, 6 H); 7.25 (dd, J ˆ 7.4, 1.1, 1 H).
2-(Dibenzoselenophene-4-yl)benzenamine (16-Se). As described for 16-S, with 15-Se (2.76 g, 7.83 mmol),
AcOH (22 ml), Fe-powder (1.8 g, 32 mmol) and EtOH (50 ml): 2.17 g (82%) of 16-Se ¥ HCC. A small amount of
16-Se was liberated from the HCl salt for analysis. 1H-NMR (200 MHz, CDCl3): 8.12 (m, 2 H); 7.80 (m, 1 H);
7.57 (t, J ˆ 7.4, 1 H); 7.36 (m, 5 H); 6.86 (m, 2 H); 3.65 (br. s, 2 H). 13C-NMR (90.56 MHz, CDCl3 ): 143.3 (s);
141.3 (s); 139.6 (s); 138.9 (s); 138.7 (s); 136.8 (s); 129.6 (d); 129.4 (d); 127.5 (d); 127.1 (s); 126.9 (d); 125.9 (d);
125.7 (d); 124.8 (d); 123.0 (d); 121.8 (d); 118.6 (d); 115.9 (d).
4-(2-Iodophenyl)dibenzoselenophene (17-Se). As described for 17-S, with 16-Se (2.17 g, 6.1 mmol), NaNO2
(0.42 g, 6.1 mmol), and KI (5.03 g, 5 equiv.). After FC and recrystallization from CH2Cl2/MeOH, 17-Se (1.61 g,
62%) was obtained. White crystals. M.p. 118 1198. 1H-NMR (200 MHz, CDCl3 ): 8.14 (m, 2 H); 8.00 (m, 1 H);
7.81 (m, 1 H); 7.55 (t, J ˆ 7.6, 1 H); 7.42 (m, 4 H); 7.30 (dd, J ˆ 7.4, 1.1, 1 H); 7.11 (m, 1 H). 13C-NMR
(90.56 MHz, CDCl3 ): 146.2 (s); 141,9 (s); 140.6 (s); 139.7 (s); 139.5 (d); 138.5 (s); 138.3 (s); 129.6 (d); 129.4 (d);
128.4 (d); 127.3 (d); 126.8 (d); 125.9 (d); 125.1 (d); 124.8 (d); 123.1 (d); 121.9 (d); 98.6 (s). HR-EI-MS: 433.9088
‡
‡
(M , C18H11ISe ; calc. 433.9069).
4-[2-(Trimethylstannyl)phenyl]dibenzoselenophene (18-Se). As described for 18-S, with 17-Se (0.37 g,
0.85 mmol), 2.07m BuLi (0.47 ml, 1.15 equiv.), and Me3SnBr (180 ml, 1.8 equiv.). Purification with prep. HPLC
gave 0.57 g (86%) of 18-Se. Colorless thick liquid, which crystallized upon standing. M.p. 92.0 93.08. 1H-NMR
(500 MHz, CDCl3 ): 8.16 (m, 1 H); 8.11(dd, J ˆ 7.9, 1.1, 1 H); 7.83 (m, 1 H); 7.65 (m, 1 H); 7.53 (t, J ˆ 7.6, 1 H);
7.53 (m, 1 H); 7.42 (m, 4 H); 7.25 (dd, J ˆ 7.3, 1.1, 1 H); À0.11 (s, J(Sn,H) ˆ 54.9, 52.6, 9 H). 13C-NMR
(90.56 MHz, THF): 150.4 (s, J(Sn,C) ˆ 29.9); 143.3 (s, J(Sn,C) ˆ 19.2); 142.9 (s, 1J(Sn,C) ˆ 469.7, 448.5); 142.1
(s, J(Se,C) ˆ 98.2); 140.6 (s, J(Se,C) ˆ 100.2); 139.4 (s); 139.1 (s); 137.3 (d, J(Sn,C) ˆ 35.8); 129.0 (d, J(Sn,C) ˆ
9.9); 128.5 (d, J(Sn,C) ˆ 35.2); 127.9 (d); 127.7 (d); 127.4 (d, J(Se,C) ˆ 7.3); 126.5 (d, J(Se,C) ˆ 16.6); 125.8 (d);
125.4 (d); 123.8 (d); 122.3 (d); À8.3 (q, 1J(Sn,C) ˆ 352.3, 336.8, 3C). EI-MS: isotopic pattern matched, no exact
mass found.
N-[4-Isopropyl-2-(dibenzoselenophene-4-yl)phenyl]acetamide (22). As described for 15-S, with dibenzo-
selenophene-4-boronic acid (13-Se; 1.48 g, 5.4 mmol), N-(2-bromo-4-isopropylphenyl)acetamide [36] (1.26 g,
4.9 mmol), [Pd(PPh3)4] (0.17 g, 3 mol-%), MeOH (10 ml), benzene (80 ml), and 2m Na2CO3 (2 equiv.). TLC
monitoring (Et2O/hexane 60 :40) indicated formation of dibenzoselenophene as a by-product. Therefore, more
13-Se (0.2 g, 0.7 mmol) and [Pd(PPh3)4] (0.1 g, 2 mol-%) were added. After a total reaction time of 4.5 h at 60
708, the crude product was prepurified by filtration through a short silica-gel plug followed by FC (3% Et2O/
CH2Cl2): 1.35 g (68%) of 22. Colorless solid. M.p. 82 858; 1H-NMR (300 MHz, CDCl3): 8.20 (m, 3 H); 7.85 (m,

Helvetica Chimica Acta Vol. 85 (2002)
3769
1 H); 7.65 7.30 (m, 6 H); 6.96 (br. s, 1 H); 2.94 (sept., J ˆ 6.9, 1 H); 1.89 (s, 3 H); 1.27 (d, J ˆ 6.9, 6 H).
13C-NMR (75.403 MHz, CDCl3): 168.3 (s); 145.2 (s); 141.4 (s); 139.7 (s); 139.1 (s); 138.5 (s); 135.9 (s); 132.4 (s);
131.7 (s); 127.5 (d); 127.4 (d); 127.2 (d); 127.2 (d); 126.0 (d); 125.7 (d); 125.0 (d); 123.2 (d); 122.3 (d); 122.2 (d);
‡
‡
24.5 (d); 24.0 (q); 23.9 (q). HR-EI-MS: 407.0779 (M , C23H21NOSe ; calc. 407.0791).
4-Isopropyl-2-(dibenzoselenophene-4-yl)benzenamine (23). A mixture of 22 (1.30 g, 3.2 mmol), conc. HCl
soln. (20 ml), and H2O (10 ml) was refluxed for 20 h. The mixture was cooled in ice and made basic with aq.
KOH soln. The mixture was extracted with CH2Cl2 , the org. layer washed with H2O, dried (Na2SO4), and
evaporated: 1.10 g (95%) of 23. Yellow oil that was used without further purification. 1H-NMR (300 MHz,
CDCl3): 8.15 (m, 2 H); 7.85 (m, 1 H); 7.60 7.10 (m, 6 H); 6.80 (d, J ˆ 7.9, 1 H); 2.87 (sept., J ˆ 6.9, 1 H); 1.25 (d,
J ˆ 6.9, 6 H). 13C-NMR (75.403 MHz, CDCl3): 141.5 (s); 141.1 (s); 139.8 (s); 139.3 (s); 139.0 (s); 138.8 (s); 137.2
(s); 127.4 (d); 127.4 (d); 127.4 (d); 127.2 (s); 126.9 (d); 125.9 (d); 125.6 (d); 124.8 (d); 123.0 (d); 121.7 (d); 116.1
‡
‡
(d); 33.3 (d); 24.2 (q). HR-EI-MS 365.0704 (M , C21H19NSe ; calc. 365.0685).
Isomer Mixture of 4-(2-Iodophenyl)-8-isopropyldibenzoselenophene and 4-(2-Iodo-5-isopropylphenyl)di-
benzoselenophene (24a/24b). As described for 17-S, with 23 (1.10 g, 3.01 mmol), NaNO2 (0.22 g, 3.2 mmol),
conc. HCl soln. (1.5 ml), and KI (2.5 g, 5 equiv.). The crude product was purified by FC: 0.82 g (57%) of 24a/
24b. Colorless material. 1H-NMR (300 MHz, CDCl3): 8.20 7.00 (m, 20 H); 3.12 (sept., J ˆ 7.0, 1 H); 2.96 (sept.,
J ˆ 7.0, 1 H); 1.40 (d, J ˆ 7.0, 6 H); 1.31 (d, J ˆ 7.0, 3 H); 1.30 (d, J ˆ 7.0, 3 H). 13C-NMR (75.403 MHz, CDCl3):
149.5 (s); 146.5 (s); 146.1 (s); 145.9 (s); 142.1 (s); 142.1 (s); 141.0 (s); 140.9 (s); 139.8 (s); 139.6 (d); 139.5 (d);
138.7 (s); 138.5 (s); 138.4 (s); 136.9 (s); 129.7 (d); 129.5 (d); 128.4 (d); 128.2 (d); 127.8 (d); 127.5 (d); 127.2 (d);
126.9 (d); 126.0 (d); 125.9 (d); 125.7 (d); 125.1 (d); 125.0 (d); 124.8 (d); 123.1 (d); 121.9 (d); 121.9 (d); 120.8 (d);
‡
‡
98.7 (s); 94.7 (s); 34.2 (d); 33.7 (d); 24.3 (q); 23.9 (q); 23.7 (q). HR-EI-MS: 475.9553 (M , C21H17ISe ; calc.
475.9538).
Isomer Mixture of 8-Isopropyl-4-[2-(trimethylstannyl)phenyl]dibenzoselenophene and 4-[5-Isopropyl-2-
(trimethylstannyl)phenyl]dibenzoselenophene (25a/25b). As described for 18-S, with 25a/25b (0.80 g,
1.68 mmol), BuLi (0.81 ml, 2.5 mmol), and Me3SnBr (400 ml, 3.4 mmol). The product was purified by FC
(hexane): 0.70 g (82%) of 24a/24b. Colorless material. Spectral data for the major isomer after FC purification
of a Me3SnBr-quenched NMR sample: 1H-NMR (75.403 MHz, CDCl3): 8.25 7.30 (m, 10 H); 3.13 (sept., J ˆ 7.0,
1 H); 1.41 (d, J ˆ 7.0, 6 H); À0.05 (s, J(H,Sn) ˆ 55.1, 52.6, 9 H). 13C-NMR (75.403 MHz, CDCl3): 149.7 (s); 145.7
(s); 142.6 (s); 142.5 (s); 141.8 (s); 138.8 (s); 138.3 (s); 137.1 (s); 136.8 (d); 128.4 (d); 128.0 (d); 127.4 (d); 127.0
(d); 125.9 (d); 125.7 (d); 125.1 (d); 121.5 (d); 120.8 (d); 34.2 (d); 24.3 (q); À8.2 (q, J(C,Sn) ˆ 352.9, 336.9).
Preparation of NMR Samples for the Intramolecular Li/S and Li/Se Exchange Studies. The same procedure
was used for both sulfur and selenium lithium compounds and will be referred to as the General Procedure for
such sample preparations. Samples in THF/Et2O 80 :20 were prepared as follows: all equipment, a 10-ml NMR
tube, two 5-ml round-bottom flasks, two 5-ml conical flasks, syringes, needles, and cannulas were oven-dried and
either allowed to cool in a desiccator (syringes, needles, cannulas) or purged with N2. The flasks and NMR tube
were fitted with septa, and N2 pressure was kept on the whole system throughout the sample preparation. For
making 0.10m solns., 0.40 mmol of the starting material (trimethylstannane compounds) was dissolved in 0.5 ml
of dry Et2O in one of the 5-ml conical flasks. The cannula was flushed with ca. 1 ml of MeLi (in Et2O) between
the two 5-ml round-bottom flasks. MeLi (0.5 equiv.) was added into the NMR tube, which was shaken and
cooled in a dry-ice bath (À 788). The Et2O soln. containing the stannane was transferred via cannula into the
NMR tube which was shaken and kept at À788. No color change should be seen at this point. Into the remaining
5-ml conical flask was added dry THF (3.2 ml) and MeLi (0.6 equiv.). The THF/MeLi soln. was slowly
transferred via cannula into the NMR tube. The color of the soln. turned yellow immediately as THF was added
and then darker in the case of selenium compounds (ate-complex formation). Solns. of 3% THF/Et2O were
made in the same way.
Solns. with Me2O as the main solvent were made similarly, with the following changes. Slightly more than
the desired amount of Me2O was condensed into a graduated vial cooled to À788 and fitted with a septum. MeLi
(ca. 0.5 ml) was added to the vial, and the dried Me2O was condensed via cannula into the NMR tube containing
the starting material/MeLi soln. (see above). The remaining solvent (THF and/or Et2O) to be used in an
experiment was dried with enough MeLi to give a total of at least 1 equiv. used in the reaction (generally 10
20% excess MeLi was used). The dried solvent was added via cannula into the NMR tube cooled to À788.
The above methods gave solns. that were almost free of any protonated product. Significant protonation
occurred if MeLi was not added to the THF before adding the THF to the Et2O soln. containing the stannane.
The Li/Sn exchange in Et2O is very slow and MeLi completely dries the Et2O before it starts to react with the
stannane. The lithium reagents were quenched with Me3SnBr and combined with quenched fractions from other
experiments to be purified later. A typical recovery of stannanes was 80%.

3770
Helvetica Chimica Acta Vol. 85 (2002)
Variable-Temperature 7Li- and 13C-NMR Spectroscopyof 4-(2-Lithiophenyl)dibenzothiophene ( 12-S) in
THF/Et2O 80 :20. A sample of 12-S in THF/Et2O 80 :20 was prepared according to the General Procedure with
18-S (0.22 g, 0.51 mmol), THF (3.2 ml), Et2O (0.4 ml), and 1.43m MeLi (0.45 ml). The sample turned yellow
7
upon the addition of MeLi to the trimethylstannane and stayed that way during the experiment. 13C- and Li-
NMR spectra were taken at temps. between r.t. and À788. The sample was stable at À788 for a week, after which
time very little protonation had taken place. Addition of HMPA to the sample resulted in decomposition. NMR
data before addition of HMPA:1H-NMR (360.14 MHz, À788): 8.19 (m, 1 H); 8.02 (d, J ˆ 7.8, 1 H); 7.94 (d, J ˆ
7.2, 1 H); 7.76 (m, 1 H); 7.47 (d, J ˆ 7.8, 1 H); 7.40 (t, J ˆ 7.8, 1 H); 7.33 (m, 2 H); 7.16 (d, J ˆ 7.2, 1 H); 6.77 (m,
7
2 H). 13C-NMR (90.56 MHz, À788): spectra and data in Fig. 3. Li-NMR (139.96 MHz, À1158): 0.66 (s).
In THF/Et2O 3 :97: 13C-NMR (90.56 MHz, 170 K): 196.3 (br. s); 151.6 (s); 147.3 (s); 143.0 (s); 142.3 (d);
139.5 (s); 136.5 (s); 135.9 (s); 126.5 (d); 125.6 (d); 125.0 (d); 124.8 (d); 123.9 (d); 123.5 (d); 122.4 (d); 122.4 (d);
121.9 (d); 117.9 (d). A spectrum is shown in Fig. 4.
Variable Temperature 7Li-, 77Se-, and 13C-NMR Spectroscopyof 4-(2-Lithiophenyl)dibenzoselenophene ( 12-
Se) 20-Se in THF/Et2O 80 :20. The General Procedure was followed, with 18-Se (0.16 g, 0.33 mmol), THF
7
(3.2 ml), Et2O (0.6 ml), and 1.56m MeLi (0.25 ml, 1.2 equiv.). 77Se-, 13C-, and Li-NMR spectra were taken at
temps. between ‡ 68 and À1358. After the variable temp. study (spectra in Figs. 6 and 7), the sample was treated
with 5.0 equiv. of HMPA, and13C- and 7Li-NMR spectra were taken. The sample was quenched with Me3SnBr to
give 0.13 g (80%) of 18-Se with less than 10% protonated product present: 1H-NMR (200 MHz, CDCl3): 8.2 7.3
(m, 11 H); À0.1 (s, J(H,Sn) ˆ 54, 52, 9 H).
The following is selected NMR data for the lithium reagent and the ate complex: 12-Se. 13C-NMR
(90.56 MHz, ‡ 68): 170.6 (s, 2 C); 145.1 (s, 2 C); 142.9 (s, 2 C); 140.1(s); 133.5 (d, 2 C); 125.3 (d); 124.8 (d, 2 C);
123.5 (d, 2 C); 123.2 (d, 2 C); 121.0 (d, 2 C). 77Se-NMR (68.68 MHz, ‡ 68): 444.8 (s). 13C-NMR (90.56 MHz,
À788): 176.7 (s, 2 C); 142.3 (s, 2 C); 140.6 (s, 2 C); 140.2 (s); 131.5 (d, 2 C); 124.4 (d); 124.4 (d, 2 C); 123.0 (d,
2 C); 122.0 (d, 2 C); 119.6 (d, 2 C). 77Se-NMR (68.68 MHz, À788): 432.6 (br. s). 7Li-NMR (139.96 MHz): at
À1138, À0.87 (br. s). at À1258, À1.0 (br. s).
Solution of 12-Se in 3% THF. 13C-NMR (90.56 MHz, À788): 169.4 (br. s, 2 C); 145.1 (s, 2 C); 142.9 (s, 2 C);
140.1 (s); 133.5 (d, 2 C); 125.6 (d); 125,0 (d, 2 C); 123.7 (d, 2 C); 123.5 (d, 2 C); 121.1 (d, 2 C). 7Li-NMR
(139.96 MHz): at À788, 0.4 (br. s); at 160 K, 0.2 (br. s). 77Se-NMR (68.68 MHz, À788): 447.0 (s). A spectrum is
shown in Fig. 4.
Solution of 20 in THF/Et2O 20 :80 with 5 equiv. of HMPA. 13C-NMR (90.56 MHz, À788): data in Fig. 6.
1H-NMR (360.14 MHz, À788): 7.88 (d, J ˆ 7.4, 2 H); 7.84 (d, J ˆ 7.4, 2 H); 7.75 (d, J ˆ 6.4, 2 H); 7.25 (t, J ˆ 7.3,
1 H); 6.96 (t, J ˆ 7.0, 2 H); 6.84 (t, J ˆ 6.9, 2 H). 77Se-NMR (68.68 MHz, À788): 414.0 (s); at À1288. 411.0 (s).
7Li-NMR (139.96 MHz, À1158): À0.5 (quint.). See Fig. 6,b.
In THF/Me2O/Et2O 8 :67:25. A sample of 12-Se was made following the General Procedure (Me2O instead
of THF), with 18-Se (0.18 g, 0.38 mmol), Me2O (3.0 ml), Et2O (0.8 ml), THF (0.3 ml), and MeLi (0.25 ml,
7
0.38 mmol). 13C-, 77Se-, and Li-NMR spectra were taken between À1418 and À798.
Variable-Temperature NMR Spectroscopy of an Isomer Mixture of 8-Isopropyl-4-(2-lithiophenyl)dibenzo-
selenophene and 4-(5-Isopropyl-2-lithiophenyl)dibenzoselenophene (21a/21b) in Me2O/Et2O 54 :46 and in
Me2O/Et2O/THF 48 :40 :12. A sample of 21-Se in Me2O/Et2O 54 :46 was prepared according to the General
Procedure (with Me2O instead of THF), with an isomer mixture of 25a/25b (0.27 g, 0.52 mmol), Me2O (1.5 ml),
Et2O (0.9 ml), and MeLi (0.4 ml, 0.52 mmol). 77Se-NMR spectra were taken at temps. between À1468 and À418
(see Fig. 8,a). Dry THF (400 ml) was added to a 5-ml conical flask fitted with a septum. To the THF was added a
few drops of MeLi, and the dried THF was transferred via cannula into the NMR tube cooled at À788. 77Se-
NMR spectra were taken at temps. between À1578 and À498 (see Fig. 8,b). The sample was quenched with
Me3SnBr (200 ml). After aq. workup, 0.22 g (82%) of the isomeric stannanes (ca. 3 :2) were obtained with less
than 5% contamination of the protonated species: 1H-NMR (300 MHz, CDCl3): 8.25 7.25 (m, 20 H); 3.11
(sept., J ˆ 7.0, 1 H); 2.98 (sept., J ˆ 7, 1 H); 1.39 (d, J ˆ 7.0, 6 H); 1.33 (d, J ˆ 7.0, 6 H); À0.08 (s, J(H,Sn) ˆ 55.0,
52.7, 9 H); À0.09 (s, J(H,Sn) ˆ 55.0, 52.5, 9 H).
In THF/Et2O/Me2O (36 :36 :28. A sample of 21a/21b in THF/Et2O/Me2O 36 :36 :28 was prepared
according to the General Procedure except of a 5-mm NMR tube, with a ca. 1:6 isomer mixture 25a/25b (0.031 g,
0.0605 mmol), THF (0.25 ml), Et2O (0.25 ml), Me2O (0.20 ml), and BuLi (0.07 ml, 0.069 mmol). 13C-, 77Se-, and
7Li-NMR spectra were taken at temps. between À608 and À1438. HMPA (0.075 ml, 0.431 mmol, 7 equiv.) was
added at À788. 13C-, 77Se-, and 7Li-NMR spectra were taken at À1388. The sample was quenched with Me3SnBr
to give a 2 :1 isomer mixture 25a/25b.
Preparation of p-Tolyllithium. A flask containing 15 ml of THF was cooled to À788 under N2, 2.0m tert-
butyllithium (5.4 ml, 10.8 mmol) was added followed by dropwise addition of a soln. of p-bromotoluene

Helvetica Chimica Acta Vol. 85 (2002)
3771
(0.924 g, 5.4 mmol) in THF (3.5 ml). The flask was warmed to 08 for 10 min and stored in the freezer at À208.
Titrations of the sample diluted in propan-1-ol with 1,10-phenanthroline as an indicator [37] showed the soln. to
be 0.194m in p-tolyllithium. This soln. contains 1 equiv. of LiBr.
Determination of the Exchange Rate between Diphenyl Selenide and p-Tolyllithium (TolLi) [6a][26]. A
stirred soln. of 0.194m p-tolyllithium (0.80 ml, 0.155 mmol) in THF was cooled to 08 under N2 in a long-necked,
round-bottom flask. Ph2Se (27 ml, 0.155 mmol) was added. After 5 min, Me3SiCl (20 ml, 0.158 mmol) was added
followed by pentylbenzene (13.3 mg, 0.0897 mmol) as a GC standard and sat. NH4Cl soln. (24 ml). The org. soln.
was stirred vigorously, dried (Na2SO4), and analyzed by GC. The amounts of PhSiMe3 (20.9% reaction) and p-
TolSiMe3 were determined (corrected for response factor), total recovery was 91%. Similar experiments were
run for 10 (34.3% reaction), 20 (52.6%), and 40 min (66.4%), and the infinity point was taken after 180 min
(79.4% of PhSiMe3). The rate constant of exchange (second-order approach to equilibrium), rate of appearance
of PhLi/PhSiMe3) was calculated with Eqn. 3; where t ˆ time in s, K ˆ equilibrium constant ˆ 14.9 (R ˆ P h or
Tol) , C0 ˆ initial concentration of TolLi, and x ˆ molarity of product (PhSiMe3). The exchange rate constant k2
was calculated to be 0.00490 Æ 0.00017 mÀ1 sÀ1 at 08.
ꢀ
ꢁ
ꢀ
ꢁ
K0:5
2C0
x À x=K À C0 À C0=K0:5
C0 ‡ C0=K0:5
k2t ˆ
Á ln
À ln
ꢂ3†
x À x=K À C0 ‡ C0=K0:5
C0 À C0=K0:5
2
‰RSeTolŠ‰PhLiŠ
‰RSePhŠ‰TolLiŠ
‰PhSiMe3Š
‰TolSiMe3Š
K ˆ
ˆ
2
Determination of Exchange Rate between Bromobenzene and p-Tolyllithium. A similar series of
experiments with PhBr instead of Ph2Se were performed at À208 (K ˆ 3.22, k2 ˆ 0.0109 Æ 0.0002 mÀ1 sÀ1),
À308 (K ˆ 3.19, k2 ˆ 0.00487 Æ 0.00011 mÀ1 sÀ1), and À508 (K ˆ 3.32, k2 ˆ 0.000768 Æ 0.000039 mÀ1 sÀ1). An
Eyring plot is shown in Fig. 10.
Determination of Exchange Rate between Diphenyl Sulfide and p-Tolyllithium. Done as described for Ph2Se.
After 17 h at 08, a small peak amounting to 0.27% was seen on GC analysis with the correct retention time for
PhSiMe3. At this point, the material recovery was 60%. On this basis, the rate was calculated with the
equilibrium constant determined for the Ph2Se exchange: k2 ˆ 2.0 ¥ 10À7 mÀ1 sÀ1 at 08. This rate constant can be
considered to be an upper limit.
REFERENCES
[1] a) G. Wittig, U. Pockels, H. Drˆge, Chem. Ber. 1938, 71 1903; b) G. Wittig, K. Clauss, Liebigs Ann. Chem.
1952, 578, 136; c) G. Wittig, H. Fritz, Liebigs Ann. Chem. 1952, 577, 39.
[2] a) H. Gilman, W. Langham, A. L. Jacoby, J. Am. Chem. Soc. 1939, 61, 106; b) H. Gilman, A. L. Jacoby, J.
Org. Chem. 1938, 3, 108.
[3] a) N. Meyer, D. Seebach, Chem. Ber. 1980, 113, 1290; b) D. Seebach, N. Peleties, Chem. Ber. 1972, 105, 511;
D. Seebach, N. Peleties, Angew. Chem., Int. Ed. 1969, 8, 450; D. Seebach, A. K. Beck, Angew. Chem., Int.
Ed. 1974, 13, 806; D. Seebach, N. Meyer, A. K. Beck, Liebigs Ann. Chem. 1977, 846; c) D. Seebach, A. K.
Beck, Chem. Ber. 1975, 108, 314.
[4] D. Seyferth, M. A. Weiner, L. G. Vaughan, G. Raab, D. E. Welch, H. M. Cohen, D. L. Alleston, Bull. Soc.
Chim. Fr. 1963, 1364.
[5] J. Meinwald, S. Knapp, S. K. Obendorf, R. E. Hughes, J. Am. Chem. Soc. 1976, 98, 6643; T. Cohen, L.-C. Yu,
J. Org. Chem. 1985, 50, 3266.
[6] a) M. Schlosser, T. Kadibelban, G. Steinhoff, Liebigs Ann. Chem. 1971, 743, 25; b) J. F. Biellmann, J. B.
Ducep, J. L. Schmitt, J. J. Vicens, Tetrahedron 1976, 32, 1061.
[7] a) H. J. Reich, D. P. Green, N. H. Phillips,J. Am. Chem. Soc. 1989, 111, 3444; b) H. J. Reich, D. P. Green,
N. H. Phillips, J. Am. Chem. Soc. 1991, 113, 1414; c) H. J. Reich, D. P. Green, N. H. Phillips, J. P. Borst, I. L.
Reich, Phosphorus Sulfur 1992, 67, 83; d) H. J. Reich, N. H. Phillips, J. Am. Chem. Soc. 1986, 108, 2102;
H. J. Reich, N. H. Phillips, Pure Appl. Chem. 1987, 59, 1021; e) H. J. Reich, B. ÷. Gudmundsson, R. R.
Dykstra, J. Am. Chem. Soc. 1992, 114, 7937; H. J. Reich, M. J. Bevan, B. ÷. Gudmundsson, C. L. Puckett,
Angew. Chem., Int. Ed. 2002, 41, 3436; f) H. J. Reich, C. S. Cooperman, J. Am. Chem. Soc. 1973, 95, 5077;
g) H. J. Reich, N. H. Phillips, I. L. Reich, J. Am. Chem. Soc. 1985, 107, 4101; h) H. J. Reich, D. P. Green,
M. A. Medina, W. S. Goldenberg, B. ÷. Gudmundsson, R. R. Dykstra, N. H. Phillips, J. Am. Chem. Soc.
1998, 120, 7201; i) H. J. Reich, J. Chem. Ed. Software, 1996, 3D, 2; http://www.chem.wisc.edu/areas/reich/
plt/windnmr.htm; j) H. J. Reich, J. P. Borst, R. R. Dykstra, D. P. Green,J. Am. Chem. Soc. 1993, 115, 8728;

3772
Helvetica Chimica Acta Vol. 85 (2002)
k) H. J. Reich, W. H. Sikorski, unpublished results W. H. Sikorski, Ph. D. Thesis, University of Wisconsin,
Madison, 1997; l) W. H. Sikorski, A. W. Sanders, H. J. Reich, Magn. Reson. Chem. 1998, 36, S118.
[8] W. F. Bailey, J. J. Patricia, J. Organomet. Chem. 1988, 352, 1.
[9] W. B. Farnham, J. C. Calabrese, J. Am. Chem. Soc. 1986, 108, 2449.
[10] a) S. Ogawa, Y. Masutomi, N. Furukawa, T. Erata, Heteroatom Chem. 1992, 3, 423; b) Y. Masutomi, N.
Furukawa, T. Erata, Heteroatom Chem. 1995, 6, 19; c) S. Ogawa, S. Sato, T. Erata, N. Furukawa,
Tetrahedron Lett., 1992, 33, 1915; d) S. Ogawa, Y. Matsunaga, S. Sato, I. Iida, N. Furukawa, J. Chem. Soc.,
Chem. Commun. 1992, 1141; e) S. Ogawa, S. Sato, T. Erata, N. Furukawa, Tetrahedron Lett. 1991, 32, 3179;
f) S. Ogawa, S. Sato, N. Furukawa, Tetrahedron Lett. 1992, 33, 7925; g) S. Ogawa, Y. Matsunaga, S. Sato, T.
Erata, N. Furukawa, Tetrahedron Lett. 1992, 33, 93.
[11] A. J. Ashe III, L. L. Lohr, S. M. Al-Taweel, Organometallics 1991, 10, 242.
[12] A. Maercker, H. Bodenstedt, L. Brandsma, Angew. Chem., Int. Ed. 1992, 31, 1339.
[13] A. H. J. F. de Keijzer, F. J. J. de Kanter, M. Schakel, R. F. Schmitz, G. W. Klumpp, Angew. Chem., Int. Ed.
1996, 35, 1127; A. H. J. F. de Keijzer, F. J. J. de Kanter, M. Schakel, V. P. Osinga, G. W. Klumpp, J.
Organomet. Chem. 1997, 548, 29.
[14] a) D. Hellwinkel, Chem. Ber. 1966, 99, 3628; b) D. Hellwinkel, G. Fahrbach, Chem. Ber. 1968, 101, 574;
c) D. Hellwinkel, G. Fahrbach, Liebigs Ann. Chem. 1968, 712, 1; d) D. Hellwinkel, G. Fahrbach, Liebigs
Ann. Chem. 1968, 715, 68; e) D. Hellwinkel, W. Krapp, Chem. Ber. 1978, 111, 13; f) D. Hellwinkel, Chem.
Ber. 1966, 99, 3660.
[15] K. Clauss, Chem. Ber. 1955, 88, 268; F. M. Beringer, L. L. Chang, J. Org. Chem. 1972, 37, 1516.
[16] Y. K. Yur×ev, N. K. Sadovaya, J. Gen. Chem. USSR 1964, 34, 1814.
[17] S. Harder, J. Boersma, L. Brandsma, J. A. Kanters, W. Bauer, R. Pi, P. v. R. Schleyer, H. Schˆllhorn, U.
Thewalt, Organometallics 1989, 8, 1688.
[18] a) W. J. Burlant, E. S. Gould, J. Am. Chem. Soc. 1954, 76, 5775; b) J. D. McCullough, T. W. Campbell, E. S.
Gould, J. Am. Chem. Soc. 1950, 72, 5753.
[19] a) P. H. W. Lau, J. C. Martin,J. Am. Chem. Soc. 1978, 100, 7077; b) C. W. Perkins, S. R. Wilson, J. C. Martin,
J. Am. Chem. Soc. 1985, 107, 3209; c) C. W. Perkins, R. B. Clarkson, J. C. Martin, J. Am. Chem. Soc. 1986,
108, 3206.
[20] W. Walter, B. Krische, J. Voss, J. Chem Res. 1978, 4101.
[21] S. A. Culley, A. J. Arduengo, J. Am. Chem. Soc. 1984, 106, 1164.
[22] G. van Koten, J. T. B. H. Jastrzebski, J. G. Noltes, A. L. Spek, J. C. Schoone, J. Organomet. Chem. 1978, 148,
233.
[23] N. Miyaura, T. Yanagi, A. Suzuki, Synth. Commun. 1981, 11, 513.
[24] G.-J. Meyer, K. Rˆssler, G. Stˆcklin, J. Am. Chem. Soc. 1979, 101, 3121.
[25] T. E. Hogen-Esch, Adv. Phys. Org. Chem. 1977, 15, 153; D. H. O×Brien, C. R. Russell, A. J. Hart, J. Am.
Chem. Soc. 1979, 101, 633.
[26] a) H. J. S. Winkler, H. Winkler, J. Am. Chem. Soc. 1966, 88, 969; b) H. J. S. Winkler, H. Winkler, J. Am.
Chem. Soc. 1966, 88, 964.
[27] A. J. Jones, D. M. Grant, J. G. Russell, G. Fraenkel, J. Phys. Chem. 1969, 73, 1624; W. Bauer, W. R.
Winchester, P. v. R. Schleyer,Organometallics 1987, 6, 2371; S. H. Bertz, G. Dabbagh, X. He, P. P. Power,J.
Am. Chem. Soc. 1993, 115, 11640.
[28] E. C. Ashby, T. N. Pham, J. Org. Chem. 1987, 52, 1291.
[29] G. A. Ross, M. D. Koppang, D. E. Bartak, N. F. Woolsey, J. Am. Chem. Soc. 1985, 107, 6742.
[30] G. Boche, M. Schimeczek, J. Cioslowski, P. Piskorz,Eur. J. Org. Chem. 1998, 9 1851; J. Cioslowski, P.
Piskorz, M. Schimeczek, G. Boche, J. Am. Chem. Soc. 1998, 120, 2612; K. B. Wiberg, S. Sklenak, W. F.
Bailey, Organometallics 2001, 20, 771.
[31] W. R. Blackmore, S. C. Abrahams, Acta Crystallogr. 1955, 8, 317, 323.
[32] J. J. Led, S. B. Petersen, J. Magn. Reson. 1978, 32, 1.
[33] S. V. Niementowski, Chem. Ber. 1901, 34, 3325.
[34] J. Bergman, L. Engman, J. Am. Chem. Soc. 1981, 103, 2715.
[35] W. R. McWhinnie, M. G. Patel, J. Chem. Soc. Dalton 1972, 199.
[36] D. Valentine Jr., J. W. Tilley, R. A. LeMahien, J. Org. Chem. 1981, 46, 4614.
[37] S. C. Watson, J. F. Eastham, J. Organomet. Chem. 1967, 9, 165.
Received June 18, 2002