New modes of reactivity in the threshold of the reduction potential in solution. alkylation of lithium PAH (Polycyclic Aromatic Hydrocarbon) dianions by primary fluoroalkanes: A reaction pathway complementing the classical birch reductive alkylation

Article abstract of DOI:10.1002/chem.200700187

Some of the most highly reduced organic species in solution, such as the dianions of PAHs (polycyclic aromatic hydrocarbons) display unexpected reactivity patterns when they react with an appropriate counterpart. As seen before in their reaction with prop

Full text of DOI:10.1002/chem.200700187

DOI: 10.1002/chem.200700187
New Modes of Reactivity in the Threshold of the Reduction Potential in
Solution. Alkylation of Lithium PAH (Polycyclic Aromatic Hydrocarbon)
Dianions by Primary Fluoroalkanes: A Reaction Pathway Complementing
the Classical Birch Reductive Alkylation
Cristóbal Melero, Raquel P. Herrera, Albert Guijarro,* and Miguel Yus*[a]
Dedicated to Professor Jan-Erling Bäckvall on occasion of his 60th birthday
Abstract: Some of the most highly reduced organic species in solution, such as the
dianions of PAHs (polycyclic aromatic hydrocarbons) display unexpected reactivi-
ty patterns when they react with an appropriate counterpart. As seen before in
Keywords: alkylation
alkanes lithium
hydrocarbons (PAHs)
empirical calculations
·
fluoro-
polyaromatic
semi-
their reaction with propene and other alkenes, PAHsÀ2 apparently react with fluoro-
alkanes in a nucleophilic fashion in spite of being generally regarded as powerful
electron-transfer reagents in their reactions with haloalkanes. This methodology
complements the current methodologies on reductive alkylation of polycyclic
arenes by allowing access to a new set of regioisomers, the regiochemistry of
which can be easily predicted by simple MO calculations.
·
·
·
Introduction
both very high-lying and highly delocalized electrons. This
eventually dictates much of their reactivity, which in many
cases is reminiscent of the alkaline metal that they originat-
ed from.[5] For instance, the lithium salts of anionic naphtha-
lene or biphenyls are mostly regarded as ET (electron-trans-
fer) reagents and have seen much success in preparative or-
ganic chemistry as lithiating reagents.[5] In general, arene
radical anions and dianions react with conventional electron
acceptors such as alkyl halides (X=Cl, Br, or I) or carbonyl
compounds through an ET process to give rise to radicals,
which evolve towards more or less complex mixtures of
products. These reaction crudes can consist of: 1) mainly the
carbanionic reagent, that is, the organolithium reagent for
M=Li (e.g., with primary chloroalkanes and haloarenes);
2) mixtures of radical-coupling products and organolithium
reagent (e.g., with secondary and tertiary alkyl chlorides
and alkyl bromides and iodides); or 3) mainly coupling
products (e.g., from the pinacol synthesis and the Wurtz
coupling, particularly with allylic and benzylic halides, alkyl
iodides as well as other organic halides). A mechanistic
study on competitive kinetics[6] reported recently, suggests
that the naphthalene dianion (Li2C10H8) reacts through a
The investigation of new patterns of reactivity in species
that have very high-energy occupied orbitals affords exciting
results, both from the fundamental point of view of reactivi-
ty,[1,2] as well as from a more practical point of view regard-
ing synthesis.[3,4] Highly reduced polycyclic aromatic hydro-
carbons (PAHs) occupy an advantageous position in this
scenario, because of their rich structural diversity and the
easy availability of many of them. In these anionic species,
the p-extended vacant orbitals of the hydrocarbon
ACHTREUNG(p-LUMOs) have been occupied by a number of extra elec-
trons, which often come from the direct reaction with an al-
kaline metal. The resulting polyanionic species thus have
[a] C. Melero, Dr. R. P. Herrera, Dr. A. Guijarro, Prof. Dr. M. Yus
Instituto de Síntesis Orgµnica and
Departamento de Química Orgµnica
Facultad de Ciencias, Universidad de Alicante
Apdo. 99, 03080-Alicante (Spain)
Fax : (+34)965-90-35-49
AHCTREnUNG ucleophilic substitution reaction pathway with primary
E-mail: aguijarro@ua.es
alkyl fluorides.[1] Indeed, highly reduced anionic-PAHs offer
a valuable opportunity to explore the dichotomy between
nucleophilic substitution (SN2) and electron transfer (ET)
yus@ua.es
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
10096
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

FULL PAPER
reactivity. The naphthalene dianion result was extended and
applied to a variety of PAHs, which after reduction to the
Results and Discussion
di
A
Naphthalene (1) has the most negative second reduction po-
tential among all the PAHs except for benzene itself (see
below).[8] We have observed previously that in the presence
of an excess of Li(s), naphthalene is doubly reduced in THF
or THP to its dianion, Li2-1.[9,2] Based on the reduction po-
tential criteria,[8] the remaining PAHs (1–12) are also ex-
pected to be reduced to the corresponding dianions (Li2-1–
Li2-12) under the same conditions, according to Equa-
tion (1). This is certainly so for those PAHs for which spec-
troscopic evidence of the corresponding dianions is avail-
able. However, the actual extent to which these reductions
take place has not been studied. It might vary for different
PAHs and should be understood in the context of different
heterogeneous equilibria depending on a number of varia-
bles [Eq. (1)]. These include the second reduction potential
of the PAH (Eo2), the ion-pairing equilibria of the dissolved
species, the complexation ability of the medium towards the
different species in solution, the concentration, and the tem-
perature, all of which can be further complicated by the ap-
pearance of new crystalline phases. Disproportionation equi-
libra are profoundly affected by the medium (e.g., by small
changes in the solvent),[10] and the complex kinetics inherent
to heterogeneous ET process on metal surfaces and the in-
trinsic high reactivity of the involved species makes the de-
termination of the actual composition of the reaction media
difficult. In addition to spectroscopic claims, we have includ-
ed electrochemical data from the literature on the genera-
tion of PAH dianions so that the arguments that concern
the generation and stability of some of these highly reduced
species do not rest exclusively on spectroscopic grounds.
When available, references for the crystal structure are
given. Further, the experimental regioselectivity can be re-
produced by means of facile semiempirical calculations on
the dianion (Figure 1; see the Supporting Information for
details). In all cases but one (i.e., the biphenyl dianion) the
alkylation takes place at the carbon atom of the PAHÀ2 that
supports the highest coefficient in the HOMO of the di-
tyldihydroarenes by reaction with n-octyl fluoride.[3] Inter-
estingly, we found that this type of reactivity appeared to be
distinctive of PAH dianions (at least for those of the small-
to-medium size that were tested), the corresponding radical
anions remained unreactive under identical conditions,
except for a few PAHs of very high reduction potential, and
therefore enhanced ET ability. This distinctive behavior of
the radical anions and dianions of PAHs has also been mani-
fested before in other reagents, including low-strained alicy-
clic ethers,[2] and especially in terminal alkenes,[4] which dis-
play a distinctly different reactivity with respect to these
two kind of highly reduced reagents. From our preliminary
studies, the electrophilic role of primary alkyl fluorides is
also remarkable. In general, conventional nucleophiles dis-
play low reactivity in their reactions with fluororalkanes.[7]
This is in part due to the high LUMO energies of fluororal-
kanes that prevent easy access to these orbitals by making
certain transition states such as the Sn2 difficult. This seems
to be, however, an advantage for the fluororalkanes when
they react with PAHs dianions. The reactivity pattern that is
displayed suggests a reaction pathway that is increasingly
dominated by the overlap of orbitals (i.e., a nucleophilic
substitution); this corresponds to a decrease in the ET char-
acter of the interaction. In the present article, we expanded
the number of PAH dianions that are suitable for alkylation
with primary fluororalkanes. Both alternant and non-alter-
nant PAHs are considered: naphthalene, biphenyl, phenan-
threne, anthracene, fluoranthene, pyrene, chrysene, tetra-
phene, o-terphenyl, p-terphenyl, acenaphthylene, and 1,1’-bi-
naphthyl. In addition, we provide further data that supports
the hypothesis of a SN mechanism for the alkylation process.
Finally, the factors that control the regioselectivity of the
process are discussed in light of simple MO calculations.
AHCTREaUNG nion. Apparently, reactions at other positions that support
lower HOMO coefficients are not competitive. The corre-
sponding calculated Mulliken charges, which were obtained
from the Mulliken population analysis, are also given. For
the purpose of pinpointing the reactive sites, we have not
seen any qualitative difference in the results that were ob-
tained when performing the density functional theory
(DFT) calculations on the same dianionic substrates, there-
fore the simplest method is reported (see biphenyl dianion
for an example; see also the Supporting Information). Dif-
ferent types of geometries (planar and nonplanar) have
been tested for each molecule, but only those results that
correspond to the lowest true minima that were found in the
potential-energy hypersurface, and were confirmed by the
corresponding vibrational analysis have been reported. Be-
sides 1,1’-binaphtyl and o-terphenyl dianions (C2), all PAHs
adopt a flat conformation in the ground state as dianions, in-
cluding the nonplanar conformationally dependent polyary-
Abstract in Spanish: Algunas de las especies orgµnicas mµs
reducidas en disolución, como los dianiones de HAP (hidro-
carburo aromµtico policíclico), muestran patrones de reactivi-
dad imprevistos cuando reaccionan con el sustrato apropia-
do. Tal y como se vio anteriormente frente a propeno y otros
alquenos, los HAPÀ2 reaccionan aparentemente como nucleó-
filos con fluoruros de alquilo, a pesar de ser considerados ge-
neralmente como poderosos agentes de transferencia electró-
nica frente a halogenuros de alquilo. Esta metodología com-
plementa las actuales metodologías de alquilación reductora
de arenos policíclicos, permitiendo el acceso a un nuevo
grupo de regioisómeros, cuya regioquímica puede ser fµcil-
mente predicha por medio de simples cµlculos de orbitales
moleculares.
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10097

A. Guijarro et al.
Figure 1. PM3 HOMO coefficients and Mulliken charges (in parenthesis) as well as the observed molecular point group of symmetry for the PAH dia-
nions 1À2–12À2
.
lenes such as biphenyl and p-terphenyl (see also phenan-
threne dianion for more specifications).
thalene with BuLi, crystallized (coordinated with N,N,N’,N’-
tetramethyl-1,2-ethanediamine (TMEDA)) and its structure
was determined by X-ray diffraction.[11] The 1H and
13C NMR spectra of Li2-1 have also been reported.[12] Its re-
activity is mainly unexplored, probably due to its prevailing
ET reactivity, although unique reactivity profiles that are
different from ET are beginning to be identified. We found
recently that Li2-1 reacts nucleophilically with terminal al-
kenes to afford carbolithiation products.[4] In its reaction
with primary fluororalkanes, it gives high yields of alkylation
Naphthalene dianion: The naphthalene dianion (1À2,
C10H8À2) is an extremely reduced organic species. In fact, at
present it is over the limit of the reduction potential that is
workable in solution by electrochemical means. By cyclic
voltammetry, naphthalene displayed a first reduction wave
(Eo1 =À2.53 V vs. Ag/AgCl) in DMA/TBAB (DMA=dime-
products in
a regioselective way (Table 1, entries 1–3;
Scheme 1). The reaction of Li2-1 with 1-fluorooctane afford-
ed a mixture of 1-octyl-1,4- and 1-octyl-1,2-dihydronaphtha-
lene in 84% overall yield (1a and 1a’, respectively; 1a/1a’=
1.4:1). This reaction is notorious for several reasons. It
cleanly affords alkylation products in high yields (1a+1a’:
84%), in the absence of any detectable amount of Wurtz
coupling products (0% of n-hexadecane by GLC), while the
analogous reaction with n-chlorooctane, which proceeds
through a well-established ET mechanism,[1,2,13] gives high
yields of n-octane after hydrolysis (89% of n-octane). Al-
though the alkylation reaction of Li2-1 with primary fluoror-
alkanes is believed to proceed through an SN2 transition
state, the potential role of radicals in the reaction could not
be fully discarded. As a first approach to study the radical
reaction pathway, the behavior of 5-hexenyl radical probes
thylamine, TBAB=tetrabutylammonium bromide) that cor-
À
CÀ1
responds to the reduction to the radical anion (1 , C10H8 ).
In contrast to the rest of arenes that were studied, no
second reduction wave could be measured for naphthalene
before solvent discharge, in spite of the careful choice of the
electrolytic media to shift the cathodic limits to very nega-
tive electrochemical potentials.[8] However, as a lithium salt,
the naphthalene dianion (Li2-1, Li2C10H8) is relatively stable
in THP at room temperature and in THF at low tempera-
tures, and can be prepared by direct exposure of the hydro-
carbon to Li(s) in these solvents.[2,9] Compound 1-Li2 was
first prepared by double deprotonation of 1,4-dihydronaph-
10098
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

Alkylation of Lithium PAH Dianions
FULL PAPER
Table 1. Reactions of compounds 1–12 with 1-fluorooctane (A), 6-fluoro-
hex-1-ene (B), and fluoromethylcyclopentane (C). For the structures of
the starting materials and the products see the schemes in the text.
PHA RCH2F
Products (yields [%])[a]
1
1
A
1a (49)
1a (77)[b]
1b (42)
1b (0)
1a’ (35)
1a’ (0)[b]
1b’ (34)
1b’ (0)
2
3
4
1
1
2
B
C
A
1c (0)
1c’ (0)
1c (35)
1c’ (30)
2a (48)
2a’ (33)
([D]2a, 4 7) ([D]2a’, 31)[c]
[c]
2a (87)[b]
2b (48)
2b (70)[b]
2b (0)[b]
3a (22)[d]
3a (78)[e]
4a (36)[f]
4a (75)[g]
5a (53)
2a’ (1.6)[b]
2b’ (32)
2b’ (0)[b]
2b’ (0)[b]
3b (45)[d]
3b (0)[e]
4b (57)[f]
4b (0)[g]
5
2
B
2c (0)
2c (0)
2c’ (0)[b]
2c’ (0)[b]
6
7
2
3
C
A
2c (72)[b] 2c’ (0)[b]
8
4
A
9
5
6
6
6
7
7
7
8
9
A
A
B
C
A
B
10
11
12
13
14
15
16
17
H2-6a[h]
H2-6b[h]
H2-6b (0)
7a (73)
6a (72)
6b (63)
6b (0)
Scheme 1.
H2-6c (0) 6c (0)
H2-6c[h]
6c (46)
reaction of Li2-1 with 6-fluorohex-1-ene was also carried out
under different dilution conditions (formal [Li2-1]=0.1, 0.05,
and 0.01m). In all cases, rearranged products (1c and 1c’), if
any, were below the detection limits.
As a precedent, the reaction of 6-fluorohex-1-ene with the
radical anion sodium naphthalene (NaC10H8) in 1,2-di-
7b (73)
7b (0)
8a (54)
9a (76)
cis-10a
7c (0)
C
7c (54)
A
A
A
A
A
8b (22)[i]
18 10
19 11
20 12
trans-10a (84)[j] 10b
11a (63)
12a (58)
AHCTREmUNG ethoxyethane (DME) at 258C is documented in the litera-
[a] Yields were determined by quantitative GLC and by using decane/
dodecane as an internal standard. [b] CH3CN was used instead of H2O
for hydrolysis. [c] 4-Deuterio-1-octyl-1,4-dihydrobiphenyl ([D]2a, 4 7%,
53:47 dr, >98.5% deuterium incorporation) and 2-deuterio-1-octyl-1,2-
dihydrobiphenyl ([D]2a’, 31%, 62:38 dr, >99.8% deuterium incorpora-
tion) were obtained by using D2O for deuterolysis (diastereomeric ratio
by 1H and 13C NMR spectroscopy, deuterium incorporation by MS and
1H NMR spectroscopy, the natural isotopic distribution was corrected).
[d] Phenanthrene/n-fluorooctane=1:1.2. Yields were determined by
quantitative GLC by using a calibration curve and Ph2CH2 as internal
standard. [e] Ratio phenathrene/n-fluorooctane=5:1, in THP. Yields by
quantitative GLC as in footnote [d]. [f] Anthracene/n-fluorooctane=
1:1.5. Yields were determined by quantitative GLC by using a calibration
curve and diphenylformamide as internal standard. [g] Anthracene/n-flu-
orooctane=5:1. Yields were determined by quantitative GLC as in foot-
note [f]. [h] Identified by 1H NMR spectroscopy in the reaction crude,
along with a second regioisomer, presumably the 1-alkyl-1,5-dihydropyr-
ene (H2-6a’–c’), but isolated as the corresponding 1-alkylpyrene. [i] Ob-
tained after the DDQ rearomatization of an inseparable mixture of 7-
ture.[14] In those studies, only low-boiling products were fully
identified, with emphasis given to the 1-hexene/methylcyclo-
pentane ratios, even though the combined yield was only
48% (Æ10%).[14a] Those studies were consistent with ET as
the key step in the formation of the light hydrocarbons
(C6H12), although major products of the type 1b, 1b’, 1c,
and 1c’ were overlooked.
For secondary and tertiary fluoroalkanes (i.e., 2-fluoro-
AHCTREoUNG ctane and 2-methyl-2-fluoroheptane), the reaction with Li2-
1 is interpreted in terms of an ET process that is followed
by a competition between radical coupling with the radical
anion, and further reduction with either the radical anion or
the dianion. The result is a mixture of coupling and reduc-
tion products that consist of isomeric octyldihydronaphtha-
lenes, plus the corresponding octanes (n-octane or iso-
AHCTREUNG
octane).[1]
octyldihydrobenzo[a]ACHTREUNGanthracenes. [j] cis-10a/trans-10a/10b 1:0.44:0.48.
Biphenyldianion : Biphenyl (2, C12H10) has the highest first
reduction potential (E1o =À2.68 V in DMA/TBAB vs. Ag/
AgCl) among the PAHs that are considered here, and in
general among all the PAHs, except for benzene. Reduction
CÀ1
was studied. 6-Fluorohex-1-ene was treated with an excess
of Li2-1 in THP at 08C (C6H11F/Li2-1=1:10, formal [Li2-1]=
0.2m). After the reaction was complete, hydrolysis afforded
the non-rearranged products 1b and 1b’ in 76% overall
yield, but the rearranged (cyclized) products 1c and 1c’
could not be detected by GLC (Table 1, entry 2, Scheme 1).
To confirm this finding beyond doubt, 1c and 1c’ were syn-
thesized directly from fluoromethylcyclopentane by reaction
with Li2-1 in 65% overall yield (Table 1, entry 3; Scheme 1),
and were used as control substances in GLC analyses. The
of biphenyl affords the corresponding radical anion (2
,
C12H10À), which can be further reduced electrochemically to
biphenyl dianion (2À2, C12H10À2). The second reduction po-
tential lies at very negative cathodic potential (Eo2 =À3.18 V
vs. Ag/AgCl), although, in contrast to naphthalene, it is still
within the measurable range.[8] The reason for this apparent
misplacement of the second reduction potential values (Eo2)
could be interpreted in terms of p-aromaticity. Indeed, at-
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10099

A. Guijarro et al.
tainment of Hückel aromaticity in 2À2 (14 p eÀ) could favor
the second electron uptake, the opposite being true for
naphthalene dianion 1À2, which attains to a certain degree
an antiaromatic character with the second reduction. De-
and 2c’ from fluoromethylcyclopentane by reaction with
Li2-2 (72% overall yield), and by using the pure product as
a control in GLC analyses.
Apparently, the regioselectivity that is displayed by the bi-
phenyl dianion (i.e., alkylation at the C1 position as shown
in Table 1) is not well predicted by means of MO calcula-
tions (Figure 1). PM3 HOMO coefficients predicts that for
2À2 the C4 (or para position) should be the reacting site in
the isolated dianion instead of the C1 (ipso position), which
supports a somewhat lower HOMO coefficient: C1 0.36
(À0.07); C4 0.40 (À0.60) (Mulliken charges in parenthesis).
We have also carried out DFT calculations,[19] by using the
B3LYP exchange–correlation functional,[20] on 2À2 with simi-
lar results. The geometry of the isolated dianion was opti-
1
scriptive work of the biphenyl dianion includes the H and
13C NMR spectra of Li2-2 (Li2C12H10) at À808C in
[D8]THF,[15] and the UV and IR spectra of different alkali
metal salts of biphenyl in sublimed layers.[16] In addition to
that, Li2-2 has been suggested to be a component of the
Li–biphenyl solutions of 2:1 stoichiometry in THF used for
reductive cleavage applications akin to Li(s).[17]
The reaction of a solution of Li2-2 in THF at 08C (see Ex-
perimental Section for details) with n-fluorooctane affords a
mixture of two regioisomers 2a and 2a’ after hydrolysis with
water, in 81% overall yield (Scheme 2, Table 1, entry 4;
(d,p) basis set.[21]
mized in the gas phase by using the 6–311GACHTREUNG
A stationary point of C2h symmetry was found (and con-
firmed by vibrational analysis), which has the following
atomic gross populations for the HOMO and Mulliken
charges (in parenthesis): C1 0.12 (À0.05); C4 0.17 (À0.24).
As in the previous calculations, this indicates that the attack
is expected to take place at the C4 position of 2À2, which is
the carbon atom that supports the highest contribution to
the HOMO of the dianion. The reason for this misbehavior
is not clear yet. We have observed attack at the C4 position
in other types of reactions that involve 2À2, which is in
agreement with the MO calculations.[4] On the other hand,
biphenyl is the only PAH for which this apparent miscon-
duct is observed, the remaining eleven PAHs that were used
in this study are all well behaved in regard to the calculated
and the observed reacting site.
Reductive alkylation of biphenyl with Li(s)/NH3(l) followed
by treatment with bromomethane affords 1-methyl-1,4-dihy-
drobiphenyl (2-HR, R=Me, Scheme 2).[22] Protonation of
2À2 by ammonia is proposed as an intermediate step that
leads to HLi-2 in this synthesis.
Scheme 2. Reductive alkylation of biphenyl under Birch conditions and
alkylation of dilithium biphenyl (Li2-2, Li2C12H10) with n-fluorooctane.
Scheme 1). We have developed a different hydrolysis proto-
col that avoids mixtures of products in the hydrolysis step
(e.g., 2a+2a’) and simplifies the isolation of products. The
best results were obtained by using acetonitrile (CH3CN,
pKa ꢀ25),[18] which is a weaker proton source. Selective pro-
tonation of the delocalized anionic intermediate with aceto-
nitrile affords the 1,4-dihydro derivatives as major products.
We can also take advantage of CH3CN in the protonation
step of different alkylated PAHs (Table 1, footnote [b]).
Deuterolysis with D2O affords [D]2a and [D]2a’ in 78%
overall yield, with excellent deuterium incorporation (see
Table 1, footnote [c]); this supports the presence of a living
dianionic species in the reaction medium that is alkylated by
n-fluorooctane, and is in agreement with Scheme 2.
As in the case of the naphthalene dianion, a potential rad-
icalary reaction pathway was also examined by using 5-hex-
enyl radical probes. The results are summarized in Table 1,
entries 5 and 6 (Scheme 1). When 6-fluorohex-1-ene was
treated with an equivalent of Li2-2 in THF at 08C (formal
[Li2-2]=0.2m), the reaction crude afforded the non-rear-
ranged products 2b and 2b’ in 80% overall yield. Again, the
rearranged (cyclized) products 2c and 2c’ could not be de-
tected by GLC. This was confirmed by direct synthesis of 2c
The analogous reaction that was carried out by using the
C
radical anion of biphenyl (Li-2 , LiC12H10) has a very differ-
ent outcome. By using a substoichiometric amount of Li(s)
with respect to biphenyl in THF at 08C, n-fluorooctane is
mainly reduced to n-octane (95%); only trace amounts of
2a and 2a’ are observed in this reaction by quantitative
GLC (Scheme 3). This is a recurrent behavior that is ob-
served for many of the PAH dianions investigated here, and
it illustrates the clear differences in reactivity between radi-
cal anions and dianions of PAHs.
Phenanthrene dianion: Phenanthrene (3, C14H10) has the
first and second reduction potential at a substantially nega-
tive cathodic potential (Eo1 =À2.49 V, Eo2 =À3.13 V in DMA/
Scheme 3. Reaction of lithium biphenyl (LiC12H10) with n-fluorooctane.
10100
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

Alkylation of Lithium PAH Dianions
FULL PAPER
TBAB vs. Ag/AgCl);[8] it ranks third among the non-substi-
tuted PAHs. As a lithium salt, the phenanthrene dianion
(Li2C14H10, Li2-3) can be prepared by direct exposure of
phenanthrene to Li(s).Compound Li2-3,[23] has been studied
at À708C in THF by NMR spectroscopy.[24] At room tem-
perature, the NMR spectra were not resolved, which is most
likely due to a thermally accessible triplet state of the spe-
cies. The results were interpreted to mean that 3À2 has a cer-
tain degree of intrinsic structural twisting, which prevents its
16 p-electron cloud from developing an excessive antiaro-
matic character.[25] Although PM3 and other semiempirical
calculations (MINDO/3, MNDO, AM1) afford a planar ge-
ometry for 3À2 (C14H10À2, C2v), the implementation of an ex-
tended basis set (3–21G to 6–311G**) to ab initio methods
affords a different non-planar geometry for 3À2 (C2), which
might originate from the systemꢁs driving force to reduce
antiaromaticity.[26]
Anthracene dianion: Anthracene (4, C14H10, Eo1 =À2.04V,
Eo2 =À2.64V in DMA/TBAB vs. Ag/AgCl) [8] is reduced
much more easily than its isomer phenanthrene. The lithium
salt of the anthracene dianion (Li2C14H10, Li2-4) can be pre-
pared either by double deprotonation of the 9,10-dihydroan-
thracene,[28] or by direct reaction of anthracene with Li(s) in
THF;[29] a reaction in which the radical anion (LiC14H10,
[29]
C
Li-4 ) is formed in an initial step as expected. In contrast
1
to Li2-3, the H NMR spectrum of Li2-4 is well resolved at
room temperature.[25,29] The Li NMR spectrum of Li2-4 has
7
also been studied.[30]
As in the case for Li2-3, the reaction of Li2-4 with n-
fluoroACHTREUoNG ctane in THP at 08C affords a separable mixture of
both the mono- and the dialkylated species 4a and 4b in
93% overall yield (Table 1, entry 8, footnote [f]). The 4a/4b
ratio was improved by performing the reaction with an
excess of Li2-4 (Table 1, entry 8, footnote [g], Scheme 4).
Assignment of the cis stereochemistry of the dialkylated
product 4b was done on the basis that a cis geometry is ob-
tained by alkylation of the 9-alkyl-9,10-dihydroanthracen-
10-yl lithium,[31] and it was confirmed by comparison of the
NMR chemical shifts with the cis and trans-9,10-diethyl-
9,10-dihydroanthracene.[32]
The PM3 HOMO coefficients of 3À2 predict that C9
should be the reacting site in the isolated dianion (Figure 1).
It is worth noting that the alkylation takes place at the
carbon that bears the largest HOMO coefficient rather than
at the site of the largest calculated density of charge (C3).
This indicates that the transition state of the alkylation step
is mainly driven by the overlap of orbitals rather than by
Coulombic interactions, in spite of the charged nature of the
Fluoranthene dianion: Fluoranthene (C16H10, Eo1 =À1.78 V,
dianions and the polarization of the C F bond. The reaction
Eo2 =À2.37 V in
methylamine/tributylmethylammonium
À
of Li2-3 with n-fluorooctane in THP at 08C affords 3a and
3b in 67% overall yield when the reagents are present in
stoichiometric amounts (Table 1, entry 7, footnote [d];
Scheme 4). The double alkylation that affords 3b is a com-
iodide (MA/TBMAI please define) vs. Ag/AgCl)[8] is re-
[29]
C
duced with Li(s) to the intermediate radical anion (Li-5 ),
and then to the dianion (Li2-5); the structure of the ion pair
1
was determined by H NMR spectroscopy for the sodium,[29]
and lithium salts,[33] and was studied by 7Li NMR spectrosco-
py[30] in [D8]THF.
The reaction of Li2-5 with n-fluorooctane in THP at 08C
affords 3-octyl-2,3-dihydrofluoranthene (5a) in 53% yield
(Scheme 5, Table 1, entry 9). Interestingly, a classical Birch
approach for the reductive alkylation with Li(s) in THF/
NH3(l) affords the protonated intermediate HLi-5, which
subsequently gives a different set of alkylation products
(Scheme 5). This is consistent with the quenching of the di-
Scheme 4.
plication that can be minimized by employing an excess of
Li2-3 with respect to the n-fluorooctane. By using a 5:1
excess (Table 1, entry 7, footnote [e]), 3a is obtained in 78%
yield in the absence of the dialkylated product. Assignment
of the trans stereochemistry of the dialkylated product 3b
was done on the basis that the trans geometry is obtained by
alkylation of the 9-alkyl-9,10-dihydrophenanthen-10-yl lithi-
um, and it was confirmed by comparison of the NMR chem-
ical shifts with the cis and trans-9,10-diethyl-9,10-dihydro-
phenanthrene.[27] The reaction of the radical anion of phen-
C
anthrene (Li-3 , LiC14H10) has a very different outcome. Re-
C
action of Li-3 with n-fluorooctane in THP at 08C affords
Scheme 5. Reduction of fluoranthene under Birch conditions to afford
HLi-5 (LiC16H11), and alkylation of dilithium fluoranthene (Li2-5,
Li2C16H10) with n-fluorooctane.
only minor amounts of alkylation products (3a <1%, 3b
8%); n-octane is the major reaction product.
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10101

A. Guijarro et al.
ACHTREUNGanion Li2-5 by the ammonia in the reaction media to give
1
HLi-5, as determined by H and 13C NMR spectroscopy.[34]
Pyrene dianion: Pyrene (6, C16H10, Eo1 =À2.13 V, E2o =
À2.86 V in DMA/TBAB vs. Ag/AgCl)[8] is reduced with Li(s)
[35]
C
in THF to the intermediate radical anion (Li-6 ), and with
an excess of Li(s) to the dianion Li2-6 as determined by
7
1H,[36] 13C,[37] and Li NMR spectroscopy.[30] The reaction of a
solution of Li2-6 in THP at 08C with n-fluorooctane affords
two major components after hydrolysis with water, which
were tentatively assigned as 1-octyl-1,9-dihydropyrene (H2-
6a) and a second component, presumably the isomeric 1,5-
dihydropyrene derivative (1-octyl-1,5-dihydropyrene, H2-
6a’-H2) in similar amounts. 1-Alkyldihydropyrenes can
appear as eight possible regioisomers, out of which the 1,2,
1,5 and 1,9-dihydropyrene isomers display only three vinylic
signals. After ruling out the 1,2-dihydro isomer, which has
Scheme 7.
1
an easily recognizable CH2CH aliphatic pattern by H NMR
spectroscopy, the remaining structures are H2-6a and H2-6a’.
The actual structure assignment of regioisomeric dihydro-
pyrenes, in particular those of 1,5 and 1,9-dihydropyrene has
been proposed several times,[38,39] and contested.[40] In our
hands, these two isomers are unstable towards oxidative
rearomatization, and no further efforts to secure their pro-
posed structure have been undertaken. After workup and
chromatographic isolation, the only product that was ob-
tained was 1-octylpyrene (6a), which is derived from H2-6a
and H2-6a’ by oxidation during handling (Scheme 6 and
Table 1, entry 10). As in previous cases, 5-hexenyl radical
probes were tested with similar results (Table 1, entries 11
and 12). When 6-fluorohex-1-ene was treated with an equiv-
alent of Li2-6 in THP at 08C (formal [ Li2-6]=0.2m, see Ex-
perimental Section), the reaction afforded the non-rear-
ranged product H2-6b, which was isolated as 6b in 63%
yield. The rearranged (cyclized) products H2-6c and 6c, if
any, were below the detection limits by GLC. Further confir-
mation of that came by the direct synthesis of H2-6c and 6c
from fluoromethylcyclopentane by reaction with Li2-6 (46%
yield, Table 1, entry 12, Scheme 7).
pyrene,[34] which allows access to a different set of re-
gioisomers. Pyrene reacts with Li/NH3(l), followed by alkyla-
tion
to
afford
9-alkyl-1,9-dihydropyrenes
(HR-6)
(Scheme 6).[38,41] The structure of the monoanionic inter-
mediate HLi-6, which is present in solutions that contain
1
NH3(l) has been established by H and 13C NMR spectrosco-
py,[41a,34] and is consistent with a monoprotonation of the
pyrene dianion by NH3(l). Interestingly, previous attempts to
directly alkylate the dianion of pyrene, which was prepared
in ethereal solvents with conventional electrophiles, was met
with only very limited success. Thus, the dianion of pyrene
(Li2-6) reacts with one equivalent of iodomethane in THF at
À788C to afford only a 9.5% of 1-methylpyrene, along with
a 41% of pyrene after hydrolysis and rearomatization with
2,3-dichloro-5,6-dicyano-p-quinone (DDQ).[34] This exposes
the strong ET character of Li2-6 with respect to convention-
al “electrophiles”, which rather than undergoing SN reac-
tions, behave as ET acceptors when they react with the
pyrene dianion.
Chrysene dianion: Chrysene (7, C18H12) can be doubly re-
duced in a reversible way by cyclic voltammetry at À108C;
it displays E1/2(1) =À0.61 V, E1/2(2) =À1.12 V as first and
second half-wave reduction po-
The regiochemical outcome of this alkylation is again
complementary to the classical Birch reductive alkylation of
tentials in DMF/TMAB vs.
perylene/perylene.[42] For the
sake of comparison with other
PAHs in this paper, a semi-
quantitative value can be ob-
tained by adding the redox po-
tential of perylene vs. Ag/AgCl
to the E1/2 values: Eo1 (pery-
lene)=À1.70 V
in
DMA/
TBAB vs. Ag/AgCl.[8] This is
only an approximate value be-
cause of the different experi-
mental conditions and expres-
sion of the potentials from dif-
ferent sources, but it is still a
Scheme 6. Reductive alkylation of pyrene under Birch conditions and alkylation of dilithium pyrene
(Li2C16H10) with primary fluoroalkanes.
10102
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

Alkylation of Lithium PAH Dianions
FULL PAPER
good approach for our purposes. Chemical reduction of
chrysene with Li(s) was performed in THF (among other sol-
vents) to yield the dianion (C18H12Li2, Li2-7), which has been
342, C26H30, in about 22% GLC yield, assuming that the
response factor is identical to 8a). Treatment of the mixture
with DDQ (CHCl3, RT, 12 h) afforded an unique rearomat-
ized product, which was identified as 7-octylbenzo[a]anthra-
cene (8b, Table 1, entry 16; Scheme 9).The Birch reduction
1
7
identified by H,[43] 13C,[44] and Li NMR spectroscopy.[30] The
reaction of a solution of Li2-7 in THP at 08C with n-fluoro-
ACHTREUNGoctane affords 6-octyl-5,6-dihydrochrysene (7a) after hydrol-
ysis with water and chromatographic isolation (Schemes 7
and 8, Table 1, entry 13). When 6-fluorohex-1-ene was treat-
Scheme 9.
(Li(s)/NH3(l),THF) of benzo[a]anthracene followed by EtOH
protonation yields 7,12-dihydrobenzo[a]anthracene.[47] The
same type of 7,12-dihydro derivatives are observed in the re-
duction of substituted benzo[a]anthracenes, such as 7-
methylACHTREbUGN enzo[a]anthracene which affords 7-methyl-7,12-dihy-
drobenzo[a]anthracene.[48] We have found no data on at-
tempts to obtain alkylated dihidrobenzo[a]anthracenes by
reductive alkylation of benzo[a]anthracene under Birch con-
dition, as seen above for other PAHs. However, given the
information that was collected on reduction potentials, an
estimate for the second reduction potential of benzo[a]an-
Scheme 8. Reductive alkylation of chrysene under Birch conditions and
alkylation of dilithium chrysene (Li2C18H12) with primary fluoroalkanes.
ed with one equivalent of Li2-7 in THP at 08C (formal [Li2-7]
=0.2m, see Experimental Section), the reaction afforded
only the non-rearranged product 7b in 73% yield, without
any detectable amount of 7c. The rearranged product, 7c
was prepared by direct synthesis from fluoromethylcyclo-
pentane by reaction with Li2-7 in 54% yield (Table 1,
entry 15).
The regiochemical outcome of the reductive alkylation of
chrysene under Birch conditions is shown in Scheme 8. By
analogy with the reductive alkylation of pyrene, a parallel
sequence of steps can be anticipated. In the presence of
NH3(l) the crysene dianion (Li2-7), is protonated to afford
HLi-7, which subsequently undergoes alkylation at the C5
position to yield 5-alkyl-5,6-dihydrochrysenes (HR-7).[45]
thracene can be made: Eo2 ꢀÀ1.70 V+
(À1.00); V=À2.70 V
A
in DMA/TBAB vs. Ag/AgCl. This value is more negative
than those of fluoranthene (Eo2 =À2.37 V), anthracene
(Eo2 =À2.64V), and 9-phenylanthracene ( Eo2 =À2.55 V),[8]
among others; all of these vales are reported vs. Ag/AgCl.
For last named compounds it is known that the correspond-
ing dianions are strong bases, which are protonated in NH3(l)
to afford stable monoprotonated monoanionic intermediates
(e.g., HLi-5, Scheme 5); these intermediates can be alkylat-
ed in an additional step. This is the expected behavior for
benzo[a]anthracene under Birch reductive alkylation condi-
tions. As in the cases that were considered before, direct al-
kylation of the dianion with fluoroalkanes represents a com-
plementary synthetic route for the reductive alkylation of
Benzo[a]anthracene dianion: Benzo[a]anthracene (or tetra-
phene, 8, C18H12) can be doubly reduced in two reversible
steps by cyclic voltammetry at À108C in DMF/TMAB; it
displays E1/2(1) =À0.34V, E1/2(2) =À1.00 V as first and second
half-wave reduction potentials vs. perylene–perylene.[42]
Again, for the sake of comparison, a semiquantitative value
benzo[a]
ACHTREaUGN nthracene.
o-Terphenyland p-terphenyldianions
: o-Terphenyl and
p-terphenyl (9 and 10, respectively, C18H14) are electrochem-
ically reduced in two reversible steps (o-terphenyl: Eo1 =
À2.62 V, Eo2 =À2.72 V; p-terphenyl: E1o =À2.40 V, Eo2 =
À2.70 V in DMA/TBAB vs. Ag/AgCl).[49,8] o-Terphenyl
reacts with Li(s) in [D8]THF to afford a dianion (Li2-9),
can be obtained by adding the redox potential of perylene
o
1
vs. Ag/AgCl: E (perylene)=À1.70 V in DMA/TBAB vs.
Ag/AgCl to the E1/2 values.[8] Benzo[a]anthracene is also
chemically reduced with Li(s) in THF to afford the corre-
sponding dianion, Li2-8, which has been characterized by
1H,[46] and 7Li NMR spectroscopy.[30] Dianion Li2-8 reacts
with 1-fluorooctane in THF at 08C to afford 7-octyl-7,12-di-
hydrobenzo[a]anthracene (8a) as the major reaction product
in 54% yield (Table 1, entry 16). An inseparable mixture of
minor isomeric byproducts was also observed in the crude
which has been studied by H and 13C NMR spectroscopy.[15]
1
NMR spectroscopy studies of the lithium salts of the o-ter-
phenyl (Li-10 and Li2-10) are also available.[50] In our hands,
Li2-9 and Li2-10, which were prepared in THP at 08C, react
with 1-fluorooctane to give alkylation at the inner ring, as
predicted by calculations (Figure 1). Hydrolysis affords 3-
octyl-2,3-diphenyl-1,4-cyclohexadiene (9a) in 76% yield
(Table 1, entry 17; Scheme 10), and 3-octyl-3,6-diphenyl-1,4-
cyclohexadiene (cis and trans-10a) and 2,5-diphenyl-5-octyl-
1
mixture and chromatographic fractions by H and 13C NMR
spectroscopy; it gave a unique peak by GLC or GC–MS (m/z:
1,3-cyclohexadiene (10b) in 84% overall yield, as a
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10103

A. Guijarro et al.
at 08C, which affords 5-octyl-1,5-dihydroacenapthylene
(11a) in 63% yield after hydrolysis with water. Compound
11a undergoes a [1,5]-H rearrangement to the more stable
isomeric acenaphthene 11b slowly (within hours), although
both 11a and 11b could be isolated as pure substances
(Scheme 11).
In contrast, the reduction of acenaphthylene with Li(s) in
THF/NH3(l) mixtures affords the stable intermediate HLi-11
(likely by protonation of Li2-11), which has been identified
1
by H and 13C NMR spectroscopy in the reaction media.[34]
Compound HLi-11 can be further alkylated to afford the re-
Scheme 10.
gioisomeric
1-alkyl-1,5-dihydroacenaphthylenes
HR-11
1:0.44:0.48 mixture of the cis
and trans-1,4-dihydro and 1,2-
dihydro isomers, respectively
(Table 1, entry 18; Scheme 10).
Reductive methylation of p-
terphenyl with Li(s)/NH3(l)/THF
and bromomethane or chloro-
methane at À338C under a va-
riety of reaction conditions
gives complex mixtures that
contain both the dimethylated
compounds in the inner ring
(3,6-dimethyl-3,6-diphenyl-
Scheme 11. Reduction of acenaphthylene under Birch conditions to afford HLi-11 (LiC16H11), which can be al-
kylated in a subsequent step, and the alkylation of dilithium acenaphthylene (Li2- 11, Li2C12H8) with n-fluo-
rooctane.
A
G
along
with large amounts of recov-
ered p-terphenyl (10) as their
major components, and other products.[51] A number of ar-
guments are given to explain these facts, which include the
slow kinetics of protonation of the dianion in ammonia, or
(Scheme 11). An equivalent strategy has been described by
quenching Na2-11 in THF with one equivalent of MeOH fol-
lowed by conventional alkylation, which affords 1-alkyl-1,5-
dihydroacenaphthylenes (HR-11) along with its regioisomer-
ic 2a-alkyl-2a,5-dihydroacenaphthylenes.[57]
À
the involvement of NH2 as a base for multiple alkylations.
In any case, it is clear that ET pathways are competitively
involved. A complex mixture is obtained for o-terphenyl
under analogous conditions; monomethylated compounds
prevail at the inner ring (up to four isomers), as well as di-
methylated compounds at the inner ring, along with some
recovered o-terphenyl (9) for CH3Cl at À788C.[51] Calcula-
tions on 9À2 and 10À2 predict the alkylation at the inner ring
well (see Figure 1), but as in the earlier case, it must be re-
marked upon that the alkylation does not take place at the
carbon atom that bears the largest density of charge in o- or
p-terphenyl dianions.
1,1’-Binaphthyldianion : Little information is found in the
literature on the reduced forms of 1,1’-binaphthyl (12,
C20H14), which was attained either electrolytically or by
means of alkali metals. The reduction of 12 with potassium
in DME is reported to initially afford EPR-active solutions
that are further reduced to a diamagnetic species (presuma-
bly a dianion or higher polyanions), along with some conver-
sion to perylene.[58] Under our conditions, binaphthyl is re-
duced with an excess of Li(s) in THP at 08C, the resulting
mixture could be alkylated with 1-fluorooctane as in the
previous instances. After hydrolysis, 4-octyl-3,4-dihydro-1,1’-
binaphthyl (12a) was isolated from the crude (Table 1,
entry 20; Scheme 12). Compound 12a displays hindered ro-
Acenaphthylene dianion: Although no electrochemical data
has been found on the second reduction potential of ace-
naphthylene (11, C12H8), there is enough evidence of the oc-
currence of its dianion. Compound 11 is reduced with Li(s)
[29,52,33]
C
in THF to give the radical anion (Li-11 ),
and the di-
anion,[29] which has been studied as the lithium salt (Li2-11)
by 1H,[53,33] 13C,[53] and 7Li NMR spectroscopy.[30,53] It has
been also obtained by double deprotonation of acenaph-
thene with BuLi·TMEDA,[54] and characterized by single-
crystal X-ray crystallography.[55,56] We have carried out the
reaction of a solution of Li2-11 in THF with 1-fluorooctane
Scheme 12.
10104
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

Alkylation of Lithium PAH Dianions
FULL PAPER
1
Table 2. Carcinogenicity of some PAHs.
tation in the 300 MHz H NMR spectrum at 258C in CDCl3,
but also displays a single set of resonances at 808C. Alkyla-
tion took place principally at the 4position, as expected
from calculations. We have chosen in Figure 1 the transoid-
12À2 instead of the cisoid-12À2 (both of C2 symmetry) as the
more stable conformation of 12À2 (DE(trans–cis) =1.37 kcal
molÀ1). This is not a critical point here, because the same
conclusions are drawn for both symmetries regarding regio-
selectivity.[3]
PAH
IARC Group
PEF
benzo[a]pyrene
benzo[a]anthracene
chrysene
1
1
0.1
0.01
0.001
2A
2B
2B
naphthalene
Group 2B (possibly carcinogenic to humans).[59] A potency equivalency
factor (PEF) with respect to the reference benzo[a]pyrene is also given
(Table 2).[60]Appropriate safety measures must be enforced when manipu-
lating these PAHs, and in general, all PAH derivatives are of unknown
toxicity.
Conclusion
General: All reactions were carried out under an atmosphere of dry
argon in oven-dried glassware. THF was distilled from sodium benzophe-
none ketyl. THP was distilled from Na/K alloy. IR spectra were mea-
sured (neat) with a Nicolet Impact 400 D-FT Spectrometer. NMR spectra
were recorded on Bruker AC-300 or a Bruker Avance-500 spectrometers
with CDCl3 as a solvent at 258C, unless otherwise stated. Chemical shifts
(d) are in ppm relative to internal TMS and coupling constants (J) are in
Hz. LRMS and HRMS were measured with a Shimadzu GC/HS QP-5000
and Finingan MAT95S spectrometers, respectively. Gas chromatographic
analyses (GLC) were determined with a Hewlett–Packard HP-5890 in-
strument equipped with a flame ionization detector (FID) and a 12 m ca-
pillary column (0.2 mm diameter, 0.33 mm film thickness), by using nitro-
In conclusion, we describe new synthetic applications that
are derived from PAH dianions, which apparently react as
nucleophiles with fluoroalkanes in spite of being generally
regarded as powerful and rather intractable electron-trans-
fer reagents. So far, all of the PAHs that were tested, except
for biphenyl, afforded the expected alkylation products. The
reaction products are regiochemically controlled, alkylated
dihydroarenes. These are interesting molecules, in which
one ring is dearomatized; this allows further functionaliza-
tion in the polycyclic framework. In most cases it is possible
to evaluate the reacting site within the polycyclic structure
by simple MO calculations. This methodology therefore
complements the current methodologies on the reductive al-
kylation of polycyclic arenes, and allows access to a new set
of regioisomers, the regiochemistry of which is dictated by
the HOMO coefficients on the dianion. Classical metal/-
gen (2 mLminÀ1
column =608C (3 min) and 60–2708C (158CminÀ1), P=40 kPa as routine
) as a carrier gas, Tinjector =2758C, Tdetector =3008C,
T
working conditions. Dianionic PAHs (Li2-1 to Li2-12) were prepared by
using standard methods of manipulation under an argon atmosphere
from the corresponding PAHs and an excess of mechanically activated
lithium powder at 08C. The best grade PAHs that were commercially
available (Acros, Aldrich), as well as 1-fluorooctane were used. When
given, the concentrations of Li2-1 to Li2-12 are formal, the actual concen-
tration of the dianionic species was not determined. The lithium granules
(Aldrich) were mechanically activated by milling under mineral oil by
using a rotary mill, the resulting lithium powder was washed repeatedly
with dry hexane.
ACHTREUNGammonia reductive alkylations afford, in general, a different
set of alkylated regioisomers, which correspond to the alky-
lation of the monoprotonated monoanionic PAH intermedi-
ate. The transition state for the alkylation seems to be
driven by the overlap of orbitals rather than by polar inter-
actions, despite the charged nature of the dianion and the
General procedure for the reductive alkylation of PAHs 1–12 with fluo-
roalkanes—preparation of compounds 1a–c to 12a: A suspension of lithi-
um (70 mg, ca. 10 mmol) and the corresponding PAH (1–12, 2 mmol) was
prepared by stirring in dry THP (10 mL) at 08C for 1 h under an argon
atmosphere. A solution of the corresponding fluoroalkane (1-fluorooc-
tane, 6-fluorohex-1-ene, or fluoromethylcyclopentane, 2 mmol) in THP
(1 mL) was added dropwise to this suspension. After 30 min, the mixture
was hydrolyzed with water (5 mL) and neutralized with 3m HCl; hydro-
quinone (10 mg) was added, and the organic phase was analyzed by
quantitative GLC, after the addition of a carefully weighed amount of
decane, dodecane, or diphenylmethane (ca. 1 mmol) as an internal stan-
dard. The reaction products were isolated from the reaction crudes with-
out internal standard by extracting with Et2O (3”20 mL), drying over
Na2SO4, removing the solvents in vacuo, and by purifying the resulting
residue by column chromatography (silica gel doped with 5% hydroqui-
none, hexane). Pure isolated products were used in the calibration
curves. For the non-isolable minor isomers, an identical response factor
to the major isomer was assumed. If necessary, a 1% solution of hydro-
quinone was added to the purified products to prevent decay during stor-
age. Alternatively, the yield was also determined by the addition of a
carefully weighed amount of an internal standard (diphenymethane or di-
phenylformamide) to the extracted reaction crudes and by submission to
NMR spectroscopy analysis. As representative examples, a full descrip-
tion of the isolated dihydrochrysenes 7a–c is given in this section:
À
highly polarized character of the C F bond. This can be
seen in the phenanthrene and o- and p-terphenyl examples,
for which the alkylation does not take place at the carbon
atom that bears the highest charge, but at the carbon atom
with the highest HOMO coefficient. In the remaining cases,
the carbon atom that bears the highest coefficient also con-
centrates the charge. This is currently the subject of deeper
studies regarding structural features and their correlations
with reactivity. It is also significant that, in most cases, the
arene radical anion was unable to afford substantial
amounts of alkylated products. In these cases ET is the pre-
vailing reactivity, and octane was the main reaction product.
Finally, an improved hydrolytic protocol (by using acetoni-
trile as a source of protons) that affords an improved ratio
of regioisomers in the last protonation step is also reported.
6-Octyl-5,6-dihydrochrysene (7a): Rf =0.50 (hexane); 1H NMR
(300 MHz, CDCl3): d=0.84(apparent t, J=6.6 Hz, 3H; CH3), 1.07–1.50
(m, 14H; 7”CH2), 2.92–3.06 (m, 1H; CH2CHCH2), 3.20 (dd, J=15.8,
ExperimentalSection
5.8 Hz, 1H;
aromCHHCHCH2), 7.22–7.29 (m, 2H; 2”CHarom), 7.29–7.40 (m, 1H;
CHarom), 7.40–7.58 (m, 2H; 2”CHarom), 7.75–7.88 (m, 3H; 3”CHarom),
CaromCHHCHCH2), 3.50 (dd, J=15.8, 3.6 Hz, 1H;
Caution: It should be noted that the International Agency for Research
on Cancer (IARC) classifies the following PAHs as Group 1 (carcino-
genic to humans), Group 2 A (probably carcinogenic to humans), or
C
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10105

A. Guijarro et al.
7.93 (apparent d, J=8.6 Hz, 1H; CHarom), 8.16 ppm (apparent d, J=
8.5 Hz, 1H; CHarom); 13C NMR (75 MHz, CDCl3): d=14.07 (CH3), 22.61
(CH3CH2), 25.57, 28.62, 29.24, 29.51, 29.61 (5”CH2), 31.82 (CaromCH2),
33.56 (CH2CH2CH), 38.21 (CH), 122.15, 123.68, 124.49, 125.35, 126.12,
126.82, 126.89, 127.22, 127.90, 128.59 (10”CHarom), 130.98, 131.05, 132.33,
133.23, 134.02, 141.03 ppm (6”Carom); IR (film): n=3059, 2924, 2853,
1462, 1377, 816, 757 cmÀ1; MS (70 eV, EI): m/z (%): 344 (1.16) [M+2H]+,
343 (6.40) [M+1H]+, 342 (22.85) [M]+, 230 (19), 229 (100), 228 (36), 43
(12), 41 (11); HR-MS: m/z calcd for C26H30: 342.2348; found 342.2362.
[4] C. Melero, A. Guijarro, M. Yus, Tetrahedron Lett. 2006, 47, 6267–
6271.
[5] M. Yus in The Chemistry of Organolithium Compounds, Part 2, Vol.
1 (Eds.: Z. Rappoport, I. Marek), Wiley, Chichester, 2004, pp. 657–
747.
[6] a) A. Guijarro, R. D. Rieke, Angew. Chem. 1998, 110, 1789–1791;
Angew. Chem. Int. Ed. 1998, 37, 1679–1681; b) A. Guijarro, D. M.
Rosenberg, R. D. Rieke, J. Am. Chem. Soc. 1999, 121, 4155–4167.
[7] R. D. Chambers, S. R. James in Comprehensive Organic Chemistry,
Vol 1 (Eds.: D. Barton, W. D. Ollis), Pergamon Press, New York,
1979, p. 527.
[8] K. Meerholz, J. Heinze, J. Am. Chem. Soc. 1989, 111, 2325–2326.
[9] R. P. Herrera, A. Guijarro, M. Yus, Tetrahedron Lett. 2001, 42,
3455–3458.
[10] G. Levin, B. E. Holloway, M. Szwarc, J. Am. Chem. Soc. 1976, 98,
5706–5709.
[11] J. J. Brooks, W. Rhine, G. D. Stucky, J. Am. Chem. Soc. 1972, 94,
7346–7351.
[12] R. Benken, H. Günther, Helv. Chim. Acta 1988, 71, 694–702.
[13] J. F. Garst, Acc. Chem. Res. 1971, 4, 400–406.
[14] a) J. F. Garst, F. E. Barton II, Tetrahedron Lett. 1969, 10, 587–590;
b) J. F. Garst, F. E. Barton II, J. Am. Chem. Soc. 1974, 96, 523–529.
[15] W. Huber, A. May, K. Müllen, Chem. Ber. 1981, 114, 1318–1336.
[16] a) A. N. Sidorov, Opt. Spektrosk. 1979, 47, 678–683 (Chem. Abstr.
1979, 92:110107; b) J. P. Devlin, J. S. McKennis, C. Thornton, J. C.
Moore, J. Phys. Chem. 1982, 86, 2613–2616.
6-(5-Hexenyl)-5,6-dihydrochrysene (7b): Rf =0.28 (hexane); 1H NMR
(300 MHz, CDCl3): d=1.14–1.56 (m, 6H; 3”CH2), 1.86–2.04(m, 2H;
CHCH2CH2), 2.92–3.05 (m, 1H; CH2CHCH2), 3.20 (dd, J=15.8, 5.6 Hz,
1H; CaromCHHCH), 3.50 (dd, J=15.8, 3.6 Hz, CaromCHHCH), 4.80–4.97
(m, 2H; CH2=CH), 5.72 (ddt, J=17.0, 10.1, 6.7 Hz, 1H; CH=CH2), 7.21–
7.29 (m, 2H; 2”CHarom), 7.30–7.39 (m, 1H; CHarom), 7.42- 7.56 (m, 2H;
CHarom) 7.75–7.89 (m, 3H; CHarom), 7.92 (apparent d, J=8.6 Hz, 1H;
CHarom), 8.15 ppm (apparent d, J=8.3 Hz, 1H; CHarom); 13C NMR
(75 MHz, CDCl3): d=27.01 (CH2), 28.60 (CaromCH2CH), 28.83, 33.37 (2”
CH2), 33.63 (CHCH2CH2), 38.18 (CH2CHCH2), 114.23 (CH2=CH),
122.16, 123.67, 124.51, 125.37, 126.15, 126.86, 126.95, 127.24, 127.90,
128.61 (10”CHarom), 130.95, 130.98, 132.33, 133,23, 134.01 (5”Carom),
138.88 (CH=CH2), 140.91 ppm (Carom); IR (film): n=3062, 2974, 2927,
2853, 1639, 1597, 1486, 1461, 1449, 1430, 1378, 1257, 1030, 993, 909, 860,
815, 762, 736 cmÀ1
;
MS: m/z (%): 316 (0.01) [M+4H]+, 315 (0.13)
[M+3H]+, 314(1.55) [ M+2H]+, 313 (12.36) [M+1H]+, 312 (48.01) [M]+,
230 (19), 229 (100), 228 (44), 226 (12); HR-MS: m/z calcd for C24H24
:
312.1878; found 312.1883.
[17] J. J. Eisch, J. Org. Chem. 1963, 28, 707–710.
[18] R. G. Pearson, R. L. Dillon, J. Am. Chem. Soc. 1953, 75, 2439–2443.
[19] a) P. Hohenberg, W. Kohn, Phys. Rev. B 1964, 136, 864–871; b) W.
Kohn, L. Sham, J. Phys. Rev. A 1965, 140, 1133–1138.
[20] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) P. J. Ste-
phens, F. J. Devlin, C. F. Chablowski, M. J. Frisch, J. Phys. Chem.
1994, 98, 11623–11627.
[21] R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem. Phys.
1980, 72, 650–654.
[22] D. F. Lindow, C. N. Cortez, R. G. Harvey, J. Am. Chem. Soc. 1972,
94, 5406–5412.
[23] K. Müllen, Helv. Chim. Acta 1978, 61, 1296–1304.
[24] a) A. Minsky, A. Y. Meyer, M. Rabinovitz, Angew. Chem. 1983, 95,
45–46;Angew. Chem. Int. Ed. Engl. 1983, 22, 45–46; b) A. Minsky,
A. Y. Meyer, M. Rabinovitz, Tetrahedron Lett. 1982, 23, 5351–5354.
[25] R. Frim, A. Mannschreck, M. Rabinovitz, Angew. Chem. 1990, 102,
919–920; Angew. Chem. Int. Ed. Engl. 1990, 29, 919–921.
[26] A. Ioffe, A. Ayalon, M. Rabinovitz, J. Chem. Soc. Perkin Trans. 2
1994, 1115–1116.
6-(5-Cyclopentylmethyl)-5,6-dihydrochrysene (7c): Rf =0.30 (hexane);
1H NMR (400 MHz, CDCl3): d=0.94–1.13 (m, 2H; 2”CHH of cyclopen-
tyl), 1.32–1.61 (m, 6H; 4”CHH of cyclopentyl, CHCH2CH
1.73 (m, 1H; CHH of cyclopentyl), 1.75–1.86 (m, 2H; CHH of cyclopen-
tyl, CH2CH(CH2)2), 3.02–3.11 (m, 1H; CH2CH(Carom)CH2), 3.21 (dd, J=
ACHTRE(UNG CH2)2), 1.62–
A
ACHTREUNG
15.8, 5.7 Hz, 1H; CaromCHHCH), 3.52 (dd, J=15.9, 3.0 Hz, 1H;
CaromCHHCH), 7.23–7.31 (m, 2H; CHarom), 7.30–7.38 (m, 1H; CHarom),
7.42–7.49 (m, 1H; CHarom), 7.49–7.58 (m, 1H; CHarom), 7.77–7.88 (m, 3H;
CHarom), 7.93 (apparent d, J=8.6 Hz, 1H; CHarom), 8.15 ppm (app. d, J=
8.6 Hz, 1H; CHarom); 13C NMR (100 MHz, CDCl3): d= 25.16, 25.18 (2”
CH2 of cyclopentyl), 28.83 (CaromCH2CH), 32.69, 32.86 (2”CH2 of cyclo-
pentyl), 37.09 (CH2CHACHTREUNG(Carom)CH2), 37.40 (CH2CHACHTRE(UGN CH2)2), 40.21 (CH2 of
cyclopentyl), 122.14, 123.70, 124.55, 125,36, 126.12, 126.82, 126.90, 127.23,
127.85, 128.61 (10”CHarom), 130.95, 131.05, 132.37, 133.22, 133.98,
141.12 ppm (6”Carom); IR (film): n=3061, 3035, 3021, 2948, 2866, 1485,
1469, 1449, 1429, 1378, 908, 817, 762, 734 cmÀ1; MS: m/z (%): 315 (0.10)
[M+3H]+, 314(1.21) [ M+2H]+, 313 (9.58) [M+1H]+, 312 (36.83) [M]+,
230 (19), 229 (100), 228 (43), 226 (11); HR-MS: m/z calcd for C24H24
:
[27] a) P. W. Rabideau, R. G. Harvey, J. Org. Chem. 1970, 35, 25–30;
b) P. W. Rabideau, R. G. Harvey, J. B. Stothers, Chem. Commun.
1969, 1005–1006.
[28] R. G. Harvey, C. C. Davis, J. Org. Chem. 1969, 34, 3007–3009.
[29] R. G. Lawer, C. V. Ristagno, J. Am. Chem. Soc. 1969, 91, 1534–
1535.
312.1878; found 312.1882.
The synthesis of fluorinated starting materials and full description of the
remaining compounds 1a–c to 12a can be found in the Supporting Infor-
mation.
[30] R. H. Cox, H. W. Terry, Jr., L. W. Harrison, Tetrahedron Lett. 1971,
12, 4815–4818.
Acknowledgements
[31] J. L. Mooney, Z. Marcinow, P. W. Rabideau, J. Org. Chem. 1986, 51,
527–532.
This work was generously supported by the Spanish Ministerio de Educa-
ción y Ciencia [grant no. CTQ2004–01261 and Consolider Ingenio 2010
(CSD2007—00006)] and the Generalitat Valenciana (grant no.
GRUPOS05/052 and GV07/36). C.M. and R.P.H. thank the University of
Alicante for fellowships. We also thank Medalchemy S.L. and Chemetall
GmbH for gifts of chemicals.
[32] N. Ahmad, C. Cloke, I. K. Hatton, N. J. Lewis, J. MacMillan, J.
Chem. Soc. Perkin Trans. 1 1985, 1849–1858.
[33] B. C. Becker, W. Huber, C. Schnieders, K. Müllen, Chem. Ber. 1983,
116, 1573–1594.
[34] K. Müllen, W. Huber, G. Neumann, C. Schnieders, H. Unterberg, J.
Am. Chem. Soc. 1985, 107, 801–807.
[35] O. W. Howarth, G. K. Frꢂnkel, J. Chem. Phys. 1970, 52, 6258–6267.
[36] K. Müllen, Helv. Chim. Acta 1978, 61, 2307–2317.
[37] C. Tintel, J. Cornelisse, J. Lugtenburg, Recl. Trav. Chim. Pays-Bas
1983, 102, 231–235.
[1] R. P. Herrera, A. Guijarro, M. Yus, Tetrahedron Lett. 2003, 44,
1309–1312.
[38] C. Tintel, J. Cornelisse, J. Lugtenburg, Recl. Trav. Chim. Pays-Bas
1983, 102, 14–20.
[2] M. Yus; R. P. Herrera, A. Guijarro, Chem. Eur. J. 2002, 8, 2574–
2584.
[3] R. P. Herrera, A. Guijarro, M. Yus, Tetrahedron Lett. 2003, 44,
[39] R. G. Harvey, P. W. Rabideau, Tetrahedron Lett. 1970, 11, 3695–
3698.
1313–1316.
10106
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 10096 – 10107

Alkylation of Lithium PAH Dianions
FULL PAPER
[40] M. A. Hempenius, P. P. J. Mulder, C. Erkelens, H. Zuilhof, W.
Heinen, J. Lugtenburg, J. Cornelisse, J. Org. Chem. 1993, 58, 3076–
3084.
[41] a) C. Schnieders, K. Müllen, W. Huber, Tetrahedron 1984, 40, 1701–
1711; b) C. Tintel, J. Lugtenburg, G. A. J. Van Amsterdam, C. Erke-
lens, J. Cornelisse, Recueil: Recl. Trav. Chim. Pays-Bas 1983, 102,
228–231.
[42] T. Saji, S. Aoyagui, J. Electroanal. Chem. 1983, 144, 143–152.
[43] A. Minsky, A. Y. Meyer, M. Rabinovitz, Tetrahedron Lett. 1982, 23,
5351–5354.
[44] R. E. Hoffman, N. Treitel, E. Shabtai, R. Benshafrut, M. Rabinovitz,
Perkin 2 2000, 1007–1011.
[52] N. J. Flint, B. J. Tabner, J. Chem. Soc. Perkin Trans. 2 1986, 1815–
1820.
[53] B. Eliasson, U. Edlund, J. Chem. Soc. Perkin Trans. 2 1983, 1837–
1842.
[54] L. D. Kershner, J. M. Gaidis, H. H. Freedman, J. Am. Chem. Soc.
1972, 94, 985–986.
[55] W. E. Rhine, J. H. Davis, G. Stucky, J. Organomet. Chem. 1977, 134,
139–149.
[56] I. L. Fedushkin, G. V. Khoroshenꢁkov, M. N. Bochkarev, S. Mühle,
H. Schumann, Russ. Chem. Bull. 2003, 52, 1358–1362.
[57] M. E. Van Loo, J. Lugtenburg, J. Cornelisse, Eur. J. Org. Chem.
2000, 713–721.
[45] R. G. Harvey, J. Org. Chem. 1971, 36, 3306–3311.
[46] A. Minsky, A. Y. Meyer, R. Poupko, M. Rabinovitz, J. Am. Chem.
Soc. 1983, 105, 2164–2172.
[58] S. P. Solodovnikov, S. T. Ioffe, Y. B. Zaks, M. I. Kabachnik, Izv.
Akad. Nauk SSSR Ser. Khim. 1968, 442–444 (Chem. Abstr. 1968,
69:76168.
[47] R. G. Harvey, K. Urberg, J. Org. Chem. 1968, 33, 2206–2211.
[48] H. M. Lee, R. G. Harvey, J. Org. Chem. 1979, 44, 4948–4953.
[49] K. Meerholz, J. Heinze, Electrochim. Acta 1996, 41, 1839–1854.
[50] a) A. Bohnen, W. Heitz, K. Müllen, H. J. Rꢂder, R. Schenk, Makro-
mol. Chem. 1991, 192, 1679–1993; b) C. G. Screttas, M. Micha-Scret-
tas, J. Org. Chem. 1983, 48, 153–158.
[59] http://monographs.iarc.fr/ENG/Classification/index.php
[60] J. F. Collins, J. P. Brown, G. V. Alexeeff, A. G. Salmon, Regul. Toxi-
col. Pharmacol. 1998, 28, 45–54.
Received: February 2, 2007
Revised: August 3, 2007
Published online: November 6, 2007
[51] R. G. Harvey, D. F. Lindow, P. W. Rabideau, J. Am. Chem. Soc.
1972, 94, 5412–5420.
Chem. Eur. J. 2007, 13, 10096 – 10107
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10107