Co-condensation reactions of substituted aromatic compounds with lithium atoms at 77 K

Article abstract of DOI:10.1002/ejic.200390065

Lithium atoms were co-condensed with +I-substituted benzene derivatives like trimethyl(phenyl)silane and tert-butylbenzene in the presence of THF at 77 K, which resulted in C-H bond activation, to produce the aryllithium compound as well as the coupling o

Full text of DOI:10.1002/ejic.200390065

FULL PAPER
Co-Condensation Reactions of Substituted Aromatic Compounds with Lithium
Atoms at 77 K
John P. Dunne,[a] Matthias Bockmeyer,[a] and Matthias Tacke*[a]
Keywords: Lithium / CϪH bond activation / Co-condensation / Density functional calculations
Lithium atoms were co-condensed with +I-substituted ben-
zene derivatives like trimethyl(phenyl)silane and tert-butyl-
benzene in the presence of THF at 77 K, which resulted in
C−H bond activation, to produce the aryllithium compound
as well as the coupling of the substituted phenyl radicals.
Other +I-substituted benzene derivatives like xylene and
mesitylene failed to react with lithium atoms in the presence
of THF. Oxygen and nitrogen donor substituted benzene de-
rivatives (anisole, N,N-dimethylaniline, and N,N-dimethyl-
benzylamine) showed activation of the ortho C−H bond when
co-condensed with lithium atoms in the presence of THF.
However, when thioanisole was co-condensed with lithium
atoms and THF, a C−S bond cleavage occurred instead.
These results were interpreted with the aid of DFT calcula-
tions, which show π- and σ-complexes between the lithium
clusters and benzene derivatives as reactive intermediates.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
Many metals are known to activate methane in a matrix
under photochemical conditions,[3,7] but these types of reac-
tions are of no use for preparative scale chemistry. For ex-
ample, iron atoms when photolysed react with methane to
form CH3FeH at 10 K in an argon matrix.[8]
The activation of carbonϪhydrogen bonds has been a
major topic of research for the last 30 years.[1Ϫ3] Numerous
methods have been developed to perform these reactions.
Most of these have centred on the oxidative addition of
the hydrocarbon moiety to a highly sub-coordinated metal
complex as shown in Figure 1,[4,5] however, CϪH bond ac-
tivation is not limited to transition metal chemistry. An-
other commonly used method is σ-bond metathesis using
extremely electrophilic lanthanide centres as outlined be-
low.[6]
Work performed by this group in the 1990s showed that
lithium atoms could selectively activate the CϪH bond of
benzene and toluene in the presence of THF without the
need for further photochemical excitation.[9Ϫ11] This re-
sulted in the direct synthesis of aryllithium compounds on
a synthetic scale.
One of the key intermediates in this reaction is an
Li2Ϫbenzene π-complex. This complex is capable of under-
going two single electron transfers from the Li2 cluster to
the LUMO of the aromatic ring, resulting in the activation
of the aromatic CϪH bond.
Figure 1. Selected examples of CϪH bond activation using metal
complexes
To date there has been very little preparative-scale work
using metal atoms to perform oxidative addition reactions.
The aim of this paper was to (a) expand the number of
aromatic substrates for this CϪH bond activation, (b) to
study how different substituents alter the selectivity of the
reaction, and (c) to gain an insight into the electronic limits
[a]
Department of Chemistry, University College Dublin,
Belfield, Dublin 4, Ireland
E-mail: matthias.tacke@ucd.ie
458
 2003 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ1948/03/0203Ϫ0458 $ 20.00ϩ.50/0 Eur. J. Inorg. Chem. 2003, 458Ϫ466

Co-Condensation Reactions of with Lithium Atoms at 77 K
FULL PAPER
of this reaction. For this purpose the following ϩI-substi- THF at 77 K, led to the observed lithiated compound 2-
tuted [C(CH3)3, SiMe3, p- and m-xylene] benzene derivat- methoxyphenyllithium, shown by NMR spectroscopy (as
ives were chosen, as well as a series of donor-substituted shown in Scheme 3). The ipso-carbon signal at δ ϭ
7
(OMe, NMe2, CH2NMe2, SMe) benzene derivatives.
169.6 ppm and a singlet at δ ϭ 1.56 ppm in the Li NMR
1
spectrum, as well as the integration in the H NMR spec-
trum clearly indicate a tetrameric structure in C6D6.[16] The
yield of this reaction increased to 30% compared to 12%
when toluene was used.[9,10] As in previous co-condensa-
tions, a further confirmation of the regiochemistry was
gained by the reaction of 2-methoxyphenyllithium with di-
methyl disulfide. The observed product of this reaction was
Results
Experimental Study
The ϩI-substituted benzene derivatives mesitylene and
(meta- or para-)xylene failed to react with lithium atoms at
77 K in the presence of THF as shown in Scheme 1. How-
ever, when an SiMe3 substituent was placed on the aromatic
ring [as for trimethyl(phenyl)silane] a reaction with lithium
atoms was observed at 77 K. The yield for this reaction was
32%, based on the amount of lithium vapourised. A
broadened set of signals corresponding to the species 4,4Ј-
bis(trimethylsilyl)biphenyl was observed in the 13C NMR
spectrum. However, the 7Li NMR spectrum showed a
broad peak at δ ϭ Ϫ2.99 ppm, which indicates the presence
of an Liϩ(THF)4 solvent-separated ion-pair species. These
NMR spectroscopic data in conjunction with the dark red/
black colour of the solution indicates that the broadened
13C NMR signals are actually of the 4,4Ј-bis(trimethylsilyl)-
biphenyl radical anion. It is formed by the reduction of
4,4Ј-bis(trimethylsilyl)biphenyl by excess lithium metal, as
shown in Scheme 2. When the co-condensation product was
redissolved in THF and treated with dimethyl disulfide a
full characterisation was achieved (Scheme 2). This showed
that the major product is 4,4Ј-bis(trimethylsilyl)biphenyl
(80%) and the minor product trimethyl[4-(methylsulfanyl)-
phenyl]silane (20%), identified by NMR spectroscopy and
GC MS.[12,13]
1-methoxy-2-(methylsulfanyl)benzene in
a quantitative
yield, which was identified from NMR spectroscopic data
and GC MS data[17,18] as shown in Scheme 3.
When the sulfur donor thioanisole was used, a different
reaction occurred, resulting in a cleavage of the CϪS bond
instead of CϪH bond activation. This time the major prod-
uct was diphenyl disulfide as shown in Scheme 4. When the
co-condensation product was washed with pentane, it
yielded a white air-sensitive product. This was shown to be
MeLi by treating it with benzaldehyde to produce 1-pheny-
lethanol upon an aqueous work up (as outlined in
Scheme 4). When the pentane solution was treated with
chlorotrimethylsilane (Scheme 4), the lithiated product
identified by NMR spectroscopy and GC MS measure-
ments was shown to be trimethyl(phenylsulfanyl)silane.[19]
Changing the donor substituent to a nitrogen donor like
N,N-dimethylaniline resulted in the formation of 2-(di-
methylamino)phenyllithium as identified by NMR spectro-
scopy in a low yield of 1% as shown in Scheme 3. In the
13C NMR spectrum in C6D6 an ipso carbon signal δ ϭ
1
171.6 ppm, in conjunction with the integration of the H
NMR peaks indicates a tetrameric structure with four co-
ordinated THF molecules resulting in a species of the com-
position Ar4Li4·THF4. It was not possible to determine if
the dimethylaniline group was intramolecularly coordinated
to Li from the NMR spectroscopic data. The reaction of 2-
(dimethylamino)phenyllithium with S2Me2 resulted in the
formation of dimethyl[(2-methylsulfanyl)phenyl]amine, thus
confirming that the regiochemistry is ortho to the NMe2.[20,21]
When N,N-benzylamine was co-condensed with lithium
atoms and THF, it yielded a pyrophoric beige solid in 36%
yield that was only sparingly soluble in aromatic solvents.
NMR spectroscopy in C6D6 confirmed the ortho activation
and the formation of 2-[(dimethylamino)methyl]phenylli-
thium.[21,22] Examination of the NMR spectroscopic data
Scheme 1. Failed reaction of lithium atoms with ϩI-substituted
benzene derivatives
A similar reaction series is seen when tert-butylbenzene indicates a tetrameric structure of the composition Ar4Li4
was co-condensed with lithium atoms and THF at 77 K as in benzene, with no coordinated THF being present as out-
outlined in Scheme 2. The yield for this reaction was lower lined in Scheme 3. The reaction of 2-[(dimethylamino)me-
(1% based on the amount of lithium metal vapourised) than thyl]phenyllithium with Me2S2 confirmed the regiochemis-
the trimethyl(phenyl)silane co-condensation. When the co- try as being ortho to the benzyl group. The isolated product
condensation product was treated in situ with Me2S2, NMR was identified as dimethyl[2-(methylsulfanyl)benzyl]amine
spectroscopic and GC-MS data identified the major prod- by GC-MS and NMR spectroscopic techniques as shown
uct (70%) as 4,4Ј-di-tert-butylbiphenyl[14] and a minor prod- in Scheme 3.[23]
uct (30%) as 1-tert-butyl-4-(methylsulfanyl)benzene.[15]
When a donor-substituted aromatic compound was used
Theoretical Study
the observed direction of lithiation changed from para to
The previously reported mechanism for the reaction of
ortho. Co-condensation of anisole with lithium atoms and lithium atoms with benzene was performed at the
Eur. J. Inorg. Chem. 2003, 458Ϫ466
459

J. P. Dunne, M. Bockmeyer, M. Tacke
FULL PAPER
Scheme 2. The reductive coupling of ϩI-substituted benzene derivatives by lithium atoms
Scheme 3. Reaction of donor-substituted aromatic compounds with lithium atoms in THF
Scheme 4. Co-condensation of lithium atoms with thioanisole in the presence of THF
460
Eur. J. Inorg. Chem. 2003, 458Ϫ466

Co-Condensation Reactions of with Lithium Atoms at 77 K
FULL PAPER
HartreeϪFock level of theory.[9,10] We report here the recal- than it is for toluene or tert-butylbenzene. The second step,
culation of the key intermediates for this benzene co-con- which is the formation of the insertion product 1f, has a
densation, using a DFT method, B3LYP, and the expansion reaction enthalpy of Ϫ15.4 kcal/mol. This is larger than the
of this mechanism to include other ϩI- and donor-substi- reaction enthalpies for the benzene and toluene/tert-butyl-
tuted benzene derivatives. In the benzene system a similar benzene systems.
trend is seen using the B3LYP/6-31G** method as was re-
When p-xylene is used, the energies for each step in the
ported earlier.[9,10] The first step of the reaction, the forma- mechanism also differ slightly compared to the previous
tion of a π-complex (1a), resulted in a calculated reaction cases. For the first step, the formation of 1i, the reaction
enthalpy of Ϫ8.2 kcal/mol. This is larger than the previ- enthalpy (Ϫ9.8 kcal/mol) is the most exothermic of the ϩI-
ously reported value (HF/6-31G**) of Ϫ5.9 kcal/mol.[10] substituted benzene derivatives. The second step, the forma-
The next step is a double electron transfer via the aromatic tion of 1j is the lowest of the reaction enthalpies (Ϫ14.6
system, resulting in the insertion of the Li2 cluster into the kcal/mol) of all the ϩI-substituted benzene derivatives.
CϪH bond (the formation of 1b). This was calculated to
When the substituent on the aromatic ring is a Lewis base
be Ϫ15.8 kcal/mol. This is also larger than the previously (e.g. O, N, P, S donors) the reaction pathway is more com-
calculated value of Ϫ8.0 kcal/mol.[10] These reaction en- plicated. Now there are two possibilities for coordination
thalpies are outlined in Figure 2.
of the lithium cluster as shown in Figure 3. The first pos-
The calculated reaction enthalpies shown in Figure 2 for sibility is the formation of a π-complex (2a), as in the previ-
the intermediate species, from the reaction involving tolu- ous compounds (1aϪ1j). The second possibility is coor-
ene with lithium atoms, agreed with the proposed mechan- dination of the Li2 cluster to the donor moiety in an η1-
ism reported earlier.[9,10] From these calculations it can be fashion (2b). When an oxygen donor is involved, e.g. ani-
seen that the Li2 cluster forms a π-complex (1c) with tolu- sole, the preferred pathway is via the η1 complex. The en-
ene resulting in a reaction enthalpy of Ϫ9.1 kcal/mol. This ergy difference between complex 2a and 2b is Ϫ2.6 kcal/
is larger than in the benzene experiment (Ϫ8.2 kcal/mol). mol, showing that the η1 complex 2b is more stable than
The intermediate 1c then undergoes a double electron the π-complex. As in the previous mechanism the next step
transfer via the aromatic system in an exothermic step of is a double single-electron transfer step from the Li2 cluster
Ϫ15.0 kcal/mol. However, this enthalpy is slightly lower in to the LUMO of the aromatic ring. However, for the anisole
energy than that of the benzene system (Ϫ15.8 kcal/mol). co-condensation this step has a larger reaction enthalpy of
In the resultant species 1d the Li2 cluster has inserted into Ϫ24.7 kcal/mol, compared with the ϩI-substituted derivat-
the para CϪH bond. When the CH3 group of toluene was ives shown in Figure 2. In the η1 species 2b there is an agos-
replaced by a tert-butyl group as in tert-butylbenzene, an tic interaction between the ortho CϪH bond and the coord-
˚
identical series of reaction enthalpies was found.
inated Li2 cluster, with an LiϪH distance of 2.586 A as
When the aromatic compound is trimethyl(phenyl)silane, detailed in Table 1. This explains the experimentally ob-
the energies for each step in the mechanism change slightly served ortho activation of the CϪH bond.
as described in Figure 2. For the first step, the coordination
If the donor atom is changed from an oxygen to a nitro-
of the lithium cluster to the aromatic system to form the π- gen donor, for example N,N-dimethylbenzylamine as out-
complex 1e, the reaction enthalpy is Ϫ8.6 kcal/mol. This lined in Figure 4, the formation energy for 3b, the η1-coord-
value is larger than it is in the benzene system but lower inated Li2 species, is Ϫ15.5 kcal/mol, while for the π-com-
Figure 2. Calculated reaction enthalpies for ϩI-substituted benzene derivatives and Li2
Eur. J. Inorg. Chem. 2003, 458Ϫ466
461

J. P. Dunne, M. Bockmeyer, M. Tacke
FULL PAPER
Figure 4. Calculated reaction mechanism for the co-condensation
of lithium atoms with benzylamine
Figure 3. Calculated reaction mechanism for the co-condensation
of lithium atoms with anisole
Table 1. Agostic HorthoϪLi interaction in η1 complexes
˚
LiaϪHortho [A]
Anisole
2.586
2.777
2.669
3.001
3.109
N,N-Dimethylaniline
N,N-Dimethylbenzylamine
Thioanisole
Dimethylphenylphosphane
plex 3a the reaction enthalpy (Ϫ8.8 kcal/mol) is less exo-
thermic. It is clear from this difference of Ϫ6.7 kcal/mol
that the preferred pathway is by coordination to the donor
moiety. As in the anisole case the energy for the insertion
of the Li2 cluster into the CϪH bond (3c) is higher (at
Ϫ23.1 kcal/mol) than via the π-complex types listed in Fig-
ure 2. In the calculated structure 3b there is also an agostic-
type interaction between the ortho CϪH bond and the co-
ordinated Li2 cluster as detailed in Table 1. However, in this
˚
case the LiϪH distance is longer (2.669 A) than it is in the
˚
anisole species 2b (2.586 A).
Figure 5. Calculated reaction mechanism for the co-condensation
of lithium atoms with N,N-dimethylaniline
A different preference is seen when the N-donor is N,N-
dimethylaniline. In this reaction there is no clear preference
between the η1-coordination mode (4b) and the formation
of a π-complex (4a) as shown in Figure 5. The difference in
energy between these two species is small (Ϫ0.8 kcal/mol) lined in Figure 6. When a similar theoretical study is per-
and is in favour of an η1 complex (4b). As in the previous formed on a phosphorus donor like dimethylphenylphos-
two systems, there is an interaction between the ortho CϪH phane, the η1-coordination mode (5b) is preferred over a π-
bond and the coordinated Li2 cluster 4b with a long LiϪH complex (5a) with a calculated energy difference of Ϫ1.6
˚
˚
interaction of 2.777 A compared to 2.586 A in the anisole kcal/mol (shown in Figure 6). Unlike previous examples
example 2b (outlined in Table 1). The energy for the inser- outlined above, there is no interaction between the ortho
tion into the ortho CϪH bond of the aromatic ring is Ϫ21.9 CϪH bond and the coordinated Li2 cluster as shown in
kcal/mol, which is lower than it is in the other donor species Table 1. This can be seen from an extremely long LiϪH
˚
(anisole and N,N-dimethylbenzylamine).
distance of 3.109 A. The reaction enthalpy for the insertion
For soft donor atoms like phosphorus and sulfur, as de- of the Li2 into the CϪH bond is Ϫ20.6 kcal/mol. This is
fined by the Pearson concept, we see a different trend, out- smaller (compared to Ϫ21.9 kcal/mol and Ϫ24.7 kcal/mol
462
Eur. J. Inorg. Chem. 2003, 458Ϫ466

Co-Condensation Reactions of with Lithium Atoms at 77 K
FULL PAPER
for the anisole system) than the other reaction enthalpies to the unfavourable hard/soft donor interaction. When a
for this step.
single electron is placed onto the free thioanisole there is a
substantial elongation of the SϪMe bond from 1.821 to
˚
˚
1.993 A and the SϪPh bond length decreases from 1.785 to
1.739 A. It was possible to predict the experimental ob-
served CϪS bond activation from this calculation.
Discussion
Previous research in this area showed that toluene re-
acted with lithium atoms in the presence of THF under co-
condensation conditions. The thermodynamic product
benzyllithium was not formed; however, the isolated com-
pound was identified as p-tolyllithium.[9,10] This interested
us in the examination of the effect that other ϩI substitu-
ents have on the reaction. A delicate balance between ϩI
enhancing complex formation ability and LUMO energy,
allowing reductive cleavage of the aromatic CϪH bond, is
needed for successful CϪH activation.
The electron-deficient Li2 cluster 1 as shown in Scheme 5,
will accept electron density from the aromatic system to
form an unusual main group element π-complex (2). When
the substituent on the aromatic ring donates electron den-
sity to the π-system (the ϩI effect), the interaction with the
lithium cluster will strengthen. This can be observed by the
decrease in the calculated LiϪC and LiϪLi bond lengths
shown in Tables 2 and 3. For example in the case of toluene
Figure 6. Calculated reaction mechanism for the co-condensation
of lithium atoms with dimethylphenylphosphane
However, when the same calculation is performed on a
sulfur donor like thioanisole, the situation changes. This
time there is no favoured pathway, with the π-complex 6a
being just 0.1 kcal/mol lower in energy than the η1-species
6b as shown in Figure 7. Again there is no interaction be-
tween the ortho CϪH bond and the η1-coordinated Li2
cluster. In this species the LiϪHagostic distance is longer
˚
the LiϪC distance decreases to 2.610 A, while for mesi-
˚
tylene the LiϪC distance is 2.591 A compared with the
˚
longer LiϪC of 2.626 A for the benzene system. This means
that there will be a stronger interaction between the aro-
matic rings of the ϩI-substituted benzene derivatives [for
example xylene, tert-butylbenzene, mesitylene, and trime-
thyl(phenyl)silane] than benzene. However, the next mech-
anistic step is crucial. This involves the double single-elec-
1
˚
(3.001 A) than it is in the corresponding η complexes with
oxygen and nitrogen donor substituents as outlined in
˚
Table 1. However the LiϪS distance of 2.452 A is signific-
antly larger than in the anisole species (LiϪO distance of
˚
1.948 A). This is a clear indication of weaker bonding due
Scheme 5. Key mechanistic steps in aromatic CϪH bond activation
using lithium atoms
Table 2. Selected LiϪC bond lengths for the ϩI-substituted ben-
zene derivatives
˚
LiϪC distance [A]
Benzene
Toluene
p-Xylene
Mesitylene
tert-Butylbenzene
Trimethyl(phenyl)silane
2.625
2.591
2.570
2.547
2.595
2.601
Figure 7. Calculated reaction mechanism for the co-condensation
of lithium atoms with thioanisole
Eur. J. Inorg. Chem. 2003, 458Ϫ466
463

J. P. Dunne, M. Bockmeyer, M. Tacke
FULL PAPER
Table 3. Selected LiϪLi bond lengths for all calculated intermediates
D*-η1 complex
CϪH activated species
Free Li2
˚
LiϪLi distance [A] in:
π complex
˚
Benzene
Toluene
Trimethyl(phenyl)silane
Anisole
Thioanisole
N,N-Dimethylaniline
N,N-Dimethylbenzylamine
Dimethylphenylphosphane
2.769
2.776
2.781
2.772
2.774
2.795
2.781
2.782
N/A
N/A
N/A
2.746
2.762
2.767
2.785
2.769
N/A
2.723 A
2.281
2.283
2.299
2.299
2.295
2.296
2.293
tron transfer from the π-bonded Li2 cluster 3 to the LUMO then couples to form the substituted biphenyl derivative as
of the aromatic ring. The outcome of the reaction depends outlined in Scheme 2. In this case the SiMe3 substituent sta-
on how the substituents affect the LUMO, and not its ϩI bilises the radical long enough to allow it to form the
ability. If the LUMO energy is increased in relation to tolu- coupled product 4,4Ј-bis(trimethylsilanyl)biphenyl. On
ene and benzene, as it is for mesitylene and xylene (as out- warming in the MVS reactor, excess solid lithium reduces
lined in Table 4), it is impossible to place electron density 4,4Ј-bis(trimethylsilanyl)biphenyl to its radical anion, in an
into it, thereby preventing the reaction from occurring. This analogous reaction to the reduction in solution of 4,4Ј-di-
can be seen from the lack of reaction between lithium atoms tert-butylbiphenyl by lithium metal.[14] From the calculated
and xylene or mesitylene (as described in Scheme 1). If the LUMO energies of these two biphenyl species (Table 4) a
substituent lowers the LUMO energy, as for comparison can be made in relation to the ease of reduction
trimethyl(phenyl)silane and tert-butylbenzene (as shown in of 4,4Ј-bis(trimethylsilanyl)biphenyl. The calculated LUMO
Table 4) the reaction will proceed quite readily. If the energies for 4,4Ј-di-tert-butylbiphenyl is Ϫ0.579 eV, while
LUMO energy of the aromatic ring is lowered too much, a 4,4Ј-bis(trimethylsilanyl)biphenyl has a lower LUMO en-
single electron reduction of the aromatic compound to its ergy of Ϫ0.898 eV. This shows that the reduction of 4,4Ј-
stable radical anion will occur instead of CϪH bond activa- bis(trimethylsilanyl)biphenyl by lithium metal can readily
tion.
occur in THF solution.
However, a secondary CϪH activated product was also
identified. This was the para-lithiated aromatic 4-(trimethyl-
silyl)phenyllithium. It could not be identified in solution
but treating this species with S2Me2 to form trimethyl[4-
(methylsulfanyl)phenyl]silane, allowed a further full charac-
terisation of the substitution pattern on the ring as outlined
in Scheme 2.
The co-condensation of tert-butylbenzene and lithium
atoms in the presence of THF was similar to the reaction
involving trimethyl(phenyl)silane (as detailed in Scheme 2).
The major product was found to be 4,4Ј-di-tert-butylbi-
phenyl. This is formed from the coupling of the tert-bu-
tylphenyl radical. On warming in the MVS machine it is
believed that 4,4Ј-di-tert-butylbiphenyl is reduced by solid
lithium metal to its radical anion. Because of this no direct
NMR spectroscopy was attempted, instead the co-con-
densation product was treated with S2Me2. A minor prod-
uct was observed to be tert-butyl-4-(methylsulfanyl)ben-
zene, this corresponds to a secondary reaction occurring at
77 K, the activation of the para CϪH bond of tert-butyl-
benzene.
Table 4. Calculated LUMO energies of selected reactants using
B3LYP/6-31G**
LUMO [eV]
N,N-Dimethylaniline
Mesitylene
m-Xylene
0.387
0.261
0.197
o-Xylene
p-Xylene
0.194
0.164
Toluene
tert-Butylbenzene
Anisole
0.123
0.102
0.085
Benzene
N,N-Dimethylbenzylamine
Thioanisole
Trimethyl(phenyl)silane
4,4Ј-Di-tert-butylbiphenyl
4,4Ј-Bis(trimethylsilanyl)biphenyl
0.072
0.057
Ϫ0.041
Ϫ0.180
Ϫ0.579
Ϫ0.898
When a Lewis base function is used, the ϩI inductive
effect versus LUMO effect is no longer important, because
the favoured donation mode to the Li2 cluster is now via
the η1-coordination mode through the donor moiety as
shown by a series of DFT calculations described in
Figures 3Ϫ7. This mode of bonding is energetically fa-
For the reaction of trimethyl(phenyl)silane with lithium
atoms in the presence of THF, a reaction sequence different
from the toluene co-condensation was observed. This is
governed by the ability of the SiMe3 group to stabilise an
aromatic radical. The major product identified was 4,4Ј-bis-
(trimethylsilanyl)biphenyl. This is the coupled product of
the free phenyltrimethylsilyl radical. This indicates that a voured over the formation of a π-complex with the Li2 clus-
reduction of trimethyl(phenyl)silane occurs at 77 K, which ter. Due to this change in the coordination mode, a change
464
Eur. J. Inorg. Chem. 2003, 458Ϫ466

Co-Condensation Reactions of with Lithium Atoms at 77 K
FULL PAPER
in regiochemistry occurs, resulting in ortho CϪH bond ac- Conclusion
tivation.
Lithium atoms are able to activate ϩI- and donor-substi-
tuted benzene derivatives. In the case of ϩI-aromatic com-
pounds like trimethyl(phenyl)silane and tert-butylbenzene,
two products were observed, the major product was the
coupled biphenyl and the minor product the CϪH-activ-
ated para-lithiated aromatic compound. However, other ϩI
benzene derivatives with higher LUMO energies than tolu-
ene, like xylenes and mesitylene, failed to react. It appears
that this reaction type has a narrow range based on the
LUMO energy of the aromatic compound. When donor
substituted benzene derivatives like anisole, N,N-dimethyl-
aniline, and N,N-benzylamine were co-condensed with li-
thium atoms in the presence of THF, ortho CϪH bond ac-
tivation was exclusively observed. This results from a mech-
anistic change from a π-bonded Li2 cluster via the phenyl
ring to a η1-bonded Li2 via the donor atom. For this reac-
tion type, the LUMO energy is not as important as in the
series of reactions using ϩI-substituted benzene derivatives.
For the co-condensation of thioanisole with lithium atoms
in THF, an activation of the CϪS bond as well as a reduct-
ive coupling of thioanisole was observed.
When an oxygen donor was used as in the case of anisole,
the NMR spectroscopic data, which is in good agreement
with the literature data,[16] showed an ortho-lithiated prod-
uct. NMR spectroscopic data indicate a tetrameric struc-
ture of the composition Ar4Li4. In contrast to the benzene
and toluene experiments reported previously,[9,10] the yield
was increased to 30% from 12%, which is extremely high
for an MVS experiment. As in the trimethyl(phenyl)silane
co-condensation, treating it with dimethyl disulfide as out-
lined in Scheme 3, allowed a further characterisation of the
lithiated product. This reaction confirms the ortho lithiation
by the formation of 1-methoxy-2-(methylsulfanyl)benzene,
which was identified from NMR spectroscopic data and
GC MS data.[17,18]
When the oxygen donor substituent of anisole is changed
to the softer sulfur atom as in thioanisole, a similar result
was expected. However, no CϪH bond activation was ob-
served, instead NMR spectroscopic data indicated a cleav-
age of the SϪMe bond. This resulted in the formation of
the PhϪS· radical, which dimerises to form diphenyl disulf-
ide. Secondary products were shown to be (MeLi·THF)4
and PhSLi·THF by derivatisation reactions and this is de-
scribed in Scheme 4. This result is not unexpected since the
SϪMe bond strength of 69.4 kcal/mol is weaker than that
of the CϪH bond of 100 kcal/mol. The selectivity for the
SϪC sp3 bond is due to its lower bond strength compared
with the SϪPh sp2 bond of 86.5 kcal/mol.[24]
When the donor substituent was a nitrogen donor as in
N,N-dimethylaniline, the observed product was 2-(dimethyl-
amino)phenyllithium (as outlined in Scheme 3). This reac-
tion type is similar to the anisole case, but with a much
lower yield of 1%. This is possibly due to the high LUMO
energy (0.387 eV) of the N,N-dimethylaniline, as described
in Table 4. This reaction is surprising since some ϩI-substi-
tuted aromatic compounds like mesitylene and xylenes,
which have lower LUMO energies, do not react with lithium
atoms. Presumably, the change in coordination mode to η1
facilitates this reaction.
Experimental Section
General: All experimental procedures were performed using stand-
ard Schlenk techniques under dry argon. THF, toluene, m-, p-xy-
lene, anisole, and tert-butylbenzene were dried and distilled from
sodium benzophenone ketyl under argon. Trimethyl(phenyl)silane,
thioanisole, N,N-dimethylaniline, N,N-dimethylbenzylamine were
dried and distilled from calcium dihydride under argon. All solv-
ents were then degassed by three freeze-pump-thaw cycles. Celite
was stored overnight in an oven at 120 °C and was dried under
vacuum for 3 h before use. For each filtration layers of approxim-
ately 6 cm of Celite were used. 1H and 13C NMR spectra were
obtained using either a Varian 300 MHz or a Varian 500 MHz
NMR machine. All chemical shifts are reported in ppm and are
referenced to TMS. The 7Li NMR spectra in this study were ob-
tained with a Jeol 300 MHz machine. The chemical shifts for the
7Li NMR spectra are reported in ppm and are referenced to an
internal standard of LiBr in D2O. All GC MS samples were run as
dichloromethane solutions using a Finnigan Trace GC MS, with
an RTX-5MS 15-m column.
In an ether solution N,N-dimethylbenzylamine is known
to react with BuLi to form 2-[(dimethylamino)methyl]phen-
yllithium. The factors controlling the selectivity are not well
investigated.[20,21] It is reported as being tetrameric (Ar4Li4)
in the solid state with the NMe2 group intramoleculary co-
ordinated to Li, while in THF solution it is dimeric with
the composition Ar2Li2·4THF. When benzylamine is co-
condensed with lithium atoms and THF as outlined in
Scheme 3, a pyrophoric beige solid is obtained in a larger
yield (36%) than in the N,N-dimethylaniline experiment
Typical Co-condensation Experiment: Lithium metal (0.53 g, 0.076
mol) was vapourised from an aluminium crucible protected by a
stainless-steel inlet at around 800 °C over 90 min and co-condensed
with a mixture of the aromatic compound (20 mL) and THF (80
mL, 1.12 mol). After removal of the solvent mixture, the residue
was dissolved in dry toluene and the solution was filtered through
Celite from the solid lithium hydride by-product and unreacted li-
thium metal. After removal of the toluene in vacuo a dark red solid
was isolated and washed with dry pentane (20 mL). The lithiated
1
(1%). H and 13C NMR spectroscopic data show the pres-
1
aromatic compound was identified by H and 13C NMR spectro-
ence of 2-[(dimethylamino)methyl]phenyllithium with no
coordinated THF. This composition is in agreement with
the published solid-state structure.[21] No activation was de-
tected in the benzyl position. Confirmation of the regioch-
emistry was obtained as in previous co-condensations by
the reaction with S2Me2.
scopy.
Typical Derivatisation of the Lithiated Aromatic Compound with
Me2S2: Dimethyl disulfide (0.3 mol/L) was added dropwise with
vigorous stirring to a solution of of the lithiated aromatic com-
pound (0.25 g) in THF (20 mL) at room temperature. The reaction
Eur. J. Inorg. Chem. 2003, 458Ϫ466
465

J. P. Dunne, M. Bockmeyer, M. Tacke
FULL PAPER
Metal Atoms and Small Clusters in Matrices, in: Chemistry and
Physics of Matrix-Isolated Species (Eds.: L. Andrews, M. Mos-
kowits), North Holland, Amsterdam, 1989, pp. 277Ϫ302.
W. E. Billups, M. M. Konars, R. H. Hauge, J. L. Margrave, J.
Am. Chem. Soc. 1980, 102, 7393.
was thought to be complete when a colour change from red to
yellow occurred. Water (10 mL) was added to this solution and the
organic phase separated. The aqueous phase was washed three
times with dichloromethane (10 mL). The combined organic ex-
tracts were dried with MgSO4 and the solvent was then removed
under reduced pressure, affording the derivatised compound as a
[8]
[9]
M. Tacke, Chem. Ber. 1995, 128, 91.
[10]
[11]
[12]
[13]
M. Tacke, Chem. Ber. 1996, 129, 1369.
M. Tacke, Eur. J. Inorg. Chem. 1998, 537.
1
brown oil. GC MS and H and 13C NMR spectroscopy were used
M. D. Curtis, A. L. Allred, J. Am. Chem. Soc. 1965, 12, 2554.
C. A. Walree, X. Y. van Lauteslager, A. M. Wageningen, J.
W. van Zwikker, L. A. Jenneskens, J. Organomet. Chem. 1995,
496, 117.
T. Cohen, I. H. Jeong, B. Murdryk, M. Bhupathy, M. M. A.
Awad, J. Org. Chem. 1990, 55, 1528.
R. S. Cooke, G. S. Hammond, J. Am. Chem. Soc. 1970, 92,
2739.
S. Harder, J. Boersma, L. Brandsma, G. P. M. v. Mier, J. A.
Kanters, J. Organomet. Chem. 1989, 364, 1.
A. Perez-Encabo, S. Perrio, A. M. Z. Slawin, S. E. Thomas, A.
T. Wierzchleyski, D. J. Williams, J. Chem. Soc., Perkin Trans.
1 1994, 629.
J. H. Bowie, S. O. Lawesson, J. Ø. Madson, G. Schroll, D. H.
Williams, J. Chem. Soc. B 1966, 951.
M. R. Detty, G. P. Wood, J. Org. Chem. 1980, 45, 80.
C. H. Yoder, K. F. Sheffey, R. Howell, R. E. Hess, C. Pascala,
C. D. Schaeffer, J. J. Zuckerman, J. Org. Chem. 1976, 1511.
F. T. Oakes, J. F. Sebastian, J. Organomet. Chem. 1978, 159,
363.
J. T. B. H. Jastrzebski, G. van Koten, M. Konijn, C. H. Stam,
J. Am. Chem. Soc. 1982, 104, 5490.
to identify the product.
Theoretical Methods Used in this Study: With a view to determining
the structures and reaction enthalpies of intermediates in the reac-
tion mechanism, the application of theoretical methods has proven
advantageous. For this purpose the program package used was
GAUSSIAN 98[25] implemented on a DEC Alpha work station
(500 MHz CPU/256 MB RAM) or an Origin 200 eight-processor
cluster (SGI 180 MHz CPU/2 GB RAM). Ab-initio calculations
were performed at the DFT level of theory (B3LYP) using the 6-
31G** basis set for C, H, O, N, S, P, and Li. Harmonic vibrational
frequencies, calculated at the same level, characterised stationary
points and gave the zero-point energy. The difference in the sum of
the electronic and the zero-point energies were interpreted as reac-
tion enthalpies at 0 K. LUMO energies for each aromatic com-
pound were obtained from geometry-optimised structures using
B3LYP/6-31G**.
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
Acknowledgments
K. Ohkata, K. Takee, K. Akiba, Bull. Chem. Soc. Jpn. 1985,
7, 1946.
The authors want to thank Prof. Dr. Dietmar Stalke for the access
to his NMR spectroscopy facilities in Würzburg. This work was
supported by the Centre for High-Performance Computing Ap-
plications at University College Dublin, Ireland.
[24]
[25]
D. F. McMillen, D. M. Golden, J. Phys. Chem. 1984, 88, 3175.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
ery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Mil-
lam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Po-
melli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P.
Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A.
G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Re-
plogle, J. A. Pople, Gaussian 98, Revision A.7, Gaussian, Inc.,
Pittsburgh PA, 1998.
[1]
A. E. Shilov, G. B. Shul’pin, Chem. Rev. 1997, 97, 8; A. E.
Shilov, G. B. Shul’pin, Chem. Rev. 1997, 97, 2905.
[2]
B. A. Arndtsen, R. G. Bergmann, T. A. Mobley, T. H. Petersen,
Acc. Chem. Res. 1995, 28, 154.
[3]
R. H. Crabtree, Chem. Rev. 1995, 95, 987.
[4]
A. H. Janowicz, R. G. Bergman, J. Am. Chem. Soc. 1982,
104, 352.
A. H. Janowicz, R. G. Bergman, J. Am. Chem. Soc. 1983,
154, 3929.
[5]
[6]
N. S. Radu, S. L. Buchwald, B. Scott, C. J. Burns, Organometal-
lics 1996, 15, 19; N. S. Radu, S. L. Buchwald, B. Scott, C. J.
Burns, Organometallics 1996, 15, 3913.
R. H. Hauge, J. L. Margrave, Reactions of First-Row Transition
Received May 6, 2002
[7]
[I02230]
466
Eur. J. Inorg. Chem. 2003, 458Ϫ466