The reaction of the [CpFe(CO)2]- Anion with pentafluorochlorobenzene: Nucleophilic aromatic substitution by halogen - Metal exchange

Article abstract of DOI:10.1002/(SICI)1521-3765(19980710)4:7<1169::AID-CHEM1169>3.0.CO;2-H

In the reaction of the [CpFe(CO)₂]- anion with pentafluorochlorobenzene, [CpFe(CO)₂]₂ and p-chloroperfluoropolyphenyls C₆F₅(C₆F₄)(n)-C₆F₄Cl (n = 0-3) are formed alongside the expected chlorine substitution product [C₆F₅Fe(CO)₂Cp]. The relative yield of these products depends on the ratio of initial reactants, solvent (THF, ether, DMSO, MeCN), counterion (K⁺, K⁺18-crown-6, [(C₆H₅)₃P=N=P(C₆H₅)₃]⁺ [PPN⁺]) and temperature. The highest yield of [C₆F₅Fe(CO)₂Cp] (? 50%) was obtained in THF, with a slight excess of pentafluorochlorobenzene added to [CpFe(CO)₂]K at room temperature. In the presence of proton donors tBuOH or PhCH(Et)CN the reaction does not yield [C₆F₅Fe(CO)₂Cp] and p-chloroperfluoropolyphenyls at all; instead C₆F₅H and [CpFe(CO)₂]₂ are formed. This fact and other data point to the [C₆F₅]^- anion as the key intermediate in all of the observed transformations, among them the formation of [C₆F₅Fe(CO)₂Cp]. The evidence from the use of radical traps (R₂PH and PEt₃) indicates that free radicals are not involved in the major reaction pathway. Halogen-metal exchange is proposed as the first step of the reaction mechanism. The subsequent competition between the possible reactions of the exchange intermediates, [CpFe(CO)₂Cl] and [C₆F₅]^-, is discussed to account for the highly variable ratio of the final products.

Full text of DOI:10.1002/(SICI)1521-3765(19980710)4:7<1169::AID-CHEM1169>3.0.CO;2-H

FULL PAPER
The Reaction of the [CpFe(CO)2] Anion with Pentafluorochlorobenzene:
Nucleophilic Aromatic Substitution by Halogen ± Metal Exchange
G. A. Artamkina, P. K. Sazonov, V. A. Ivushkin, I. P. Beletskaya*
Abstract: In the reaction of the
[CpFe(CO)2] anion with pentafluoro-
chlorobenzene, [CpFe(CO)2]2 and p-
chloroperfluoropolyphenyls C6F5(C6F4)n-
C6F4Cl (n ˆ 0 ± 3) are formed alongside
the expected chlorine substitution prod-
uct [C6F5Fe(CO)2Cp]. The relative yield
of these products depends on the ratio of
initial reactants, solvent (THF, ether,
pentafluorochlorobenzene added to
[CpFe(CO)2]K at room temperature. In
the presence of proton donors tBuOH or
PhCH(Et)CN the reaction does not yield
[C6F5Fe(CO)2Cp] and p-chloroperfluoro-
polyphenyls at all; instead C6F5H and
[CpFe(CO)2]2 are formed. This fact and
other data point to the [C6F5] anion as
the key intermediate in all of the ob-
served transformations, among them the
formation of [C6F5Fe(CO)2Cp]. The evi-
dence from the use of radical traps
(R2PH and PEt3) indicates that free
radicals are not involved in the major
reaction pathway. Halogen ± metal ex-
change is proposed as the first step of
the reaction mechanism. The subsequent
competition between the possible reac-
tions of the exchange intermediates,
[CpFe(CO)2Cl] and [C6F5] , is discussed
to account for the highly variable ratio of
the final products.
‡
DMSO, MeCN), counterion (K ,
Keywords: carbonyl complexes
´
‡
‡
ˆ ˆ
K 18-crown-6, [(C6H5)3P N P(C6H5)3]
halogen ± metal exchange ´ iron ´
nucleophilic aromatic substitutions
´ reaction mechanisms
‡
[PPN ]) and temperature. The highest
yield of [C6F5Fe(CO)2Cp] ( ꢀ 50%) was
obtained in THF, with a slight excess of
that distinguish them from nucleophilic aromatic substitution
reactions with conventional nucleophiles.[5±7] These reactions
do not obey one of the basic mechanistic criteria of the
addition ± elimination mechanism, the element effect (i.e.
higher reactivity of aryl fluorides as compared to other
haloarenes).[8] Reverse leaving group ability order[2, 5±7] (I >
Brꢁ Cl > F) resembles the reactivity observed in the oxida-
tive addition to unsaturated metal complexes.[9] Reduction of
the organic substrate and oxidation of the nucleophile
frequently observed as a side process[2, 5a] is another point
not consistent with the classical SN2Ar mechanism. Up to now
no reaction scheme to account for these differences has been
generally accepted.
We have previously studied the kinetics of fluorine
substitution in perfluorinated arenes by carbonylmetallates,
when reduction does not interfere with nucleophilic substitu-
tion.[5] An unusual ion-pairing effect has been observed:
close-contact ion pairs were more reactive than free ions and
solvent-separated ion pairs.[5] With the primary aim of ration-
alising the nature of the reactivity pattern distinguishing
carbonylmetallates we have now extended our research to the
substrate having both fluorine and chlorine moieties in one
molecule, where competition between several reaction path-
ways is possible. The reaction of pentafluorochlorobenzene
with the dicarbonylcyclopentadienylferrate anion [CpFe-
(CO)2] chosen as a model in the current study has already
been reported by Bruce and Stone;[2a] however, the necessary
study of the entire product composition is lacking and certain
Introduction
Nucleophilic substitution with metal carbonyl anions, carbon-
ylmetallates, may be considered as a model of oxidative
addition of organic halide to transition metal, which is a key
step of various processes catalysed by transition metal
complexes, such as carbonylation or cross-coupling. Only
aliphatic nucleophilic substitution with carbonylmetallates
was studied in detail, the operation of both SN2 and radical
reaction mechanisms being demonstrated.[1] The interaction
of carbonylmetallates with polyfluoroarylhalides, which rep-
resents a well-established synthetic approach to s-polyfluor-
oaryl metal complexes,[2, 3] has not yet received appropriate
mechanistic consideration. However, recent studies have
shown that s-aryl metal carbonyl complexes, especially di-
or poly-metal-substituted aromatics, are of considerable
interest. They can serve as models of intermediates in catalytic
processes, for the study of intramolecular charge transfer and
as a base of new materials possessing unique electronic and
optical properties.[4]
The reactions of carbonylmetallates as well as other metal-
centred anions with aryl halides possess a number of features
[*] Prof. Dr. I. P. Beletskaya, Dr. G. A. Artamkina, P. K. Sazonov,
V. A. Ivushkin
Moscow State University, Department of Chemistry
119899, Moscow, B-234, Vorobꢀyovy Gory (CIS)
E-mail: beletska@org.chem.msu.su
Fax : (‡ 7)095-938-18-44
Chem. Eur. J. 1998, 4, No. 7
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1169 $ 17.50+.25/0
1169

I. P. Beletskaya et al.
FULL PAPER
points, such as the formation of proton-containing complex p-
HC6F4Fp (Fp ˆ [CpFe(CO)2]) in pure THF, were not con-
firmed in our work.[6] Moreover, we observed additional
products that were not reported earlier, such as p-chloroper-
fluoropolyphenyls, which led us to carry out a comprehensive
examination of the reaction mechanism, including the effects
of radical and carbanionic traps. Considerable attention is
given to various factors controlling the course of the reaction,
such as the reactant ratio, the solvent, counterion and
temperature. Important as it is for the optimisation of reaction
conditions of carbonylmetallates with haloarenes, to the best
of our knowledge no similar investigation has ever been
attempted previously.
Editorial Board Member:[*] Irina
Beletskaya studied at Moscow
State University, where she re-
ceived her PhD in 1958 and the
degree of Doctor of Chemistry in
1963. The subject of the latter was
Electrophilic Substitution at Satu-
Results and Discussion
rated Carbon. She became a full
Professor in 1970 and in 1974 a
Corresponding Member of the
The course of the reaction: The starting carbonylferrate is
obtained quantitatively in the form of FpK[10] by reduction of
the dimer Fp2 with sodium/potassium alloy; thus, we were able
to determine not only the relative distribution but also the
absolute yields of the products. Reaction occurred instanta-
neously on the addition of C6F5Cl (15 ± 50% excess) to the
solution of carbonylferrate salt,[a] making direct kinetic
measurements impossible. The basic set of the products
Academy of Sciences (USSR), of which she became a full
member (Academician) in 1992. She is currently Head of the
Laboratory of Organoelement Compounds, Department of
Chemistry at Moscow State University.Professor Beletskaya is
Chief Editor of the Russian Journal of Organic Chemistry and
Chair of the Editorial Advisory Board for the
IUPAC Chemistry in the 21st Century mono-
graphs. From 1989 to 1991 she was President of
the Organic Chemistry Division of IUPAC. She is
the author of over 500 articles and 4 monographs,
and the recipient of the Lomsonov Prize (1979),
the Mendeleev Prize (1982) and the Nesmeyanov
Prize (1991). Her current research interests are
transition metal catalysis in organic synthesis,
organic derivatives of lanthanides, and carbanions
and nucleophilic aromatic substitution.
Abstract in Russian:
[Eq. (1)] was the same in all of the experiments, while the
product distribution changed considerably depending on the
reaction conditions.
In the reactions of carbonylmetallates with activated aryl
halides the higher yields of transition metal s-aryl complexes
are usually obtained at lower temperatures ( 788C).[3d, 5, 11, 12]
In contrast, the yield of the nucleophilic substitution product,
C6F5Fp, in the reaction of FpK with C6F5Cl is higher at room
temperature (50 ± 60%) than at 508C (13 ± 20%), when the
dimer Fp2 and other products of the redox process[b] are
primarily formed (entries 1 ± 3 Table 1). These include chlor-
operfluoropolyphenyls C6F5H and FpCl, though only traces of
the latter were observed. Its presence in the reaction mixtures
was not confirmed by 1H NMR spectra, probably owing to the
overlapping of the Cp-group signals with C6F5Fp. Pentafluoro-
benzene could arise both from proton and hydrogen-atom
abstraction from the solvent or Cp group by the intermediate
pentafluorophenyl anion or radical respectively.
The principal products of the reduction of C6F5Cl were
chloroperfluoropolyphenyls, C6F5(C6F4)nC6F4Cl, never previ-
[a] The rate constant k ˆ 2.79 ´ 104 Lmol 1 s 1 in DMSO has been reported
in the literature.[7]
[b] In the context of this work the term redox process corresponds only to
the net result of the transformation, oxidation of Fp and reduction of
C6F5Cl, but not to its mechanism.
[*] Members of the Editorial Board will be introduced to readers with their
first manuscript.
1170
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1170 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 7

Halogen±Metal Exchange
1169±1178
Table 1. Temperature, solvent polarity and cation effects on the course of the reaction of FpM with C6F5Cl.
Entry
Reactants and reaction conditions[a]
Relative yields [%][b]
‡
M
Solvent
T [8C]
Ratio n(C6F5Cl)/
n(Fp )
C6F5Fp
Fp2
C6F5(C6F4)nC6F4Cl
(n ˆ 1 ± 3)
C6F5H
7.5
1
2
3
4
5
6
7
8
K
K
K
THF
THF
THF
THF/Et2O[d]
THF/DMSO[e]
50
1.44
1.47
1.25
1.36
1.23
1.17
1.39
1.26
22
78
(30)
43 (32)
98
71 (48)
83
15
(8)
11
50
27 (17)
11
16
RT
RT
RT
RT
RT
RT
50
(49)
57 (43)
2
29 (22)
17
2.5
15.5
3
Li[c]
K
K/18-crown-6[f] THF
PPN
K
3
‡
THF
THF[g]
19
19
81
81
13
4
[a] The initial concentrations of reactants were 0.03 ± 0.09m. RTˆ room temperature. [b] Isolated yields are shown in brackets. Product ratio was determined
by means of 1H and 19F NMR spectroscopy with internal standard ((Me3Si)2O or C6H6 in 1H, PhCF3 or C6F6 in 19F NMR). The yields are normalised to 100%
total yield of Fp-containing products, while the corresponding experimental value varied in the range 75 ± 120%. The yield of chloroperfluoropolyphenyls
was calculated per aromatic ring. [c] The Li salt was used since the K salt is insoluble in ether. [d] The mixture contained ꢀ 30% THF. [e] The mixture
contained ꢀ 30% [D6]DMSO. [f] n(18-crown-6)/n(FpK) ˆ 2. [g] CH3OH (4 equiv) was introduced immediately after the mixing of the reactants.
ously detected in reactions of that type. Mass spectra of the
isolated individual homologues of chloroperfluoropolyphen-
yls (n ˆ 0 ± 3) exhibit high-intensity molecular ion peaks, while
the fragmentation is negligible. Small amounts of Fp-substi-
tuted perfluoropolyphenyls, Fp(C6F4)nFp (n ˆ 2 ± 5), were also
detected in the reaction, but the individual components could
not be separated because of their close Rf values. Nevertheless
reactants (C6F5Cl and FpK) were mixed at 508C (entry 8,
Table 1). The yields of the products, including C6F5H,
remained the same as in the blank experiment (entry 1,
Table 1). Thus the intermediate [C6F5] anion must be rapidly
consumed before it could be protonated by methanol.
The stoichiometry of the reaction is also highly suggestive.
Less than one mole of C6F5Cl is consumed for one mole of
Fp . When Fp2 is primarily formed, the reaction balance
approximates to Equation (a), that is, roughly two moles of
FpK per mole of C6F5Cl.
1
the H NMR spectrum of the mixture displayed a character-
istic set of signals in the Cp region (d ˆ 5.1 ± 5.2) with
consecutively decreasing intensity.
Taking into consideration the linear structure of the
phenylene chain in chloroperfluoropolyphenyls, evident from
19F NMR spectra, the most obvious route to their formation
from C6F5Cl appears to consist of a series of nucleophilic
substitution reactions with the [C6F5] anion (Scheme 1,
(2n ‡ 2)FpK ‡ (n ‡ 2)C6F5Cl
!
(a)
(n ‡ 1)Fp2 ‡ C6F5ꢂC6F4†nC6F4Cl ‡ (n ‡ 1)KCl ‡ (n ‡ 1)KF
nˆ0
3
Solvent and counterion effects: The
ion pairing in the carbonylmetallate
strongly influences the course of the
reaction, revealing an unusual bell-
shaped dependence, the highest yield
of C6F5Fp being observed in the reac-
tion with potassium salt in THF at
room temperature (entries 2 ± 3, Ta-
ble 1). Factors increasing the dissocia-
tion of the carbonylmetallate salt, for
example the addition of more polar
solvent (DMSO), crown ether or the
‡
Scheme 1. Possible pathways for p-chloroperfluoropolyphenyl formation.
exchange of the K counterion for the
‡
bulky PPN cation, reduced the ratio
of C6F5Fp to the products of the redox process (entries 5 ± 7,
Table 1). But a similar yet much greater effect is caused by the
use of Li salts with partial exchange of THF for the less polar
diethyl ether, when stronger ion pairing and aggregation of
carbonylmetallate is expected to take place (entry 4, Table 1).
Data is scarce concerning cation and solvent effects on the
product distribution pattern (redox vs. substitution) in the
reactions of carbonylmetallates with haloarenes bearing
heavier halogens (Cl, Br, I). Certain studies were conducted
only for the reactions of carbonylferrate salts with tricarbo-
nylchromium haloarene complexes, when the use of bulky
cations and low temperatures resulted in higher yields of the
substitution product.[12, 16] We observed similar behaviour
route 1). It should be mentioned that the reactions of
perfluoroarenes with nucleophiles belong with the classic
examples of nucleophilic aromatic substitution, and could
hardly be regarded as specific C ± F bond activation as was
suggested in several recent publications.[13] The substitution of
the fluorine para to chlorine on the aromatic ring with a
carbon-centred nucleophile agrees well with the predictions
derived from previous studies.[5a, 6, 14] The possibility of tetra-
fluorodehydrobenzene participation in these reactions may be
excluded, as in that case the aromatic rings could not be
concatenated para[15] (Scheme 1, route 2).
To evaluate the rate of the reaction of the [C6F5] anion
with C6F5Cl, methanol was added immediately after the
Chem. Eur. J. 1998, 4, No. 7
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1171 $ 17.50+.25/0
1171

I. P. Beletskaya et al.
FULL PAPER
Table 2. The effect of anion traps on the course of the reaction of C6F5Cl with Fp at room temperature.
Entry
Reactants and reaction conditions[a]
Relative yields [%][b]
Cation Solvent/Trap
M
Ratio
n(C6F5Cl)/n(Fp )
C6F5Fp Fp2
C6F5(C6F4)nC6F4Cl C6F5H Other products
(n ˆ 0 ± 3)
Yields [%]
‡
1
2
K
K
THF (no trap)
MeCN/MeCN
1.25
1.24
57
26
43
67
11
0
2.5
25.5
p-ClC6F4CH2CN
p-HC6F4Fp
p-ClC6F4CH2CN
p-HC6F4Fp
p-ClC6F4OtBu
o-ClC6F4OtBu
p-HC6H4OtBu
p-HC6F4Fp
4.5
7.5
9.3
9
39
5
3
4
PPN
K
MeCN/MeCN
1.31
27
0
64
91
0
0
28
44
THF/tBuOH (16 equiv) 1.24
3
9
5
6
K
K
PhCH(Et)CN (3.5 equiv) 1.17
0
97
69
0
25
7
p-ClC6F4C(Et)(Ph)CN 10
p-HC6F4C(Et)(Ph)CN
p-HC6F4Fp
8.5
3
THF/C6F5Cl (10 equiv[c]
)
9.72
31
34
[a] The initial concentrations of reactants were 0.03 ± 0.09 M. [b] Concerning the determination and calculation of the product yields see footnote [b] to
Table 1. [c] The solution of FpK was added dropwise to the solution of C6F5Cl.
Scheme 2. Products originating from protonation of the [C6F5] carbanion by alcohol or CH acid.
recently in nucleophilic vinylic substitution reaction between
Fp and chlorodifluorostyrene.[17] Also worth mentioning is
one more work reporting that the use of FpLi in ether/hexane
solution in the reaction with C6F4Hal2 (all isomers, Hal ˆ Cl,
Br) furnishes the monosubstitution products FpC6F4Hal in
moderate yield,[11] yet nothing is known about the course of
the same reactions in THF. This data is hardly compatible with
our current results, and does not help us in answering the
question why ether and large cations caused similar effects.
fluoropolyphenyls, since the latter were not observed in the
reaction carried out in MeCN. The dehalogenation process in
that case yields only proton-containing products, C6F5H and
p-HC6F4Fp. It is noteworthy that the reduction of the C6F5Cl
becomes the major reaction pathway in MeCN, the ratio
C6F5Fp/Fp2 decreasing more than threefold compared with
that in THF (entries 1 ± 3, Table 2), but this effect could be
also caused by the change of medium polarity.
Anion scavengers added to the FpK solution prior to the
addition of C6F5Cl prevent the nucleophilic substitution
pathway of the reaction entirely; not even traces of the
product C6F5Fp are detected (Scheme 2; entries 4,5, Table 2).
We investigated the effect of both O ± H and C ± H acids,
choosing for that purpose tBuOH and PhCH(Et)CN, which
are stronger acids than C6F5H (pKa ꢀ 25[19]), but less acidic
than FpH (pKa ˆ 19.2[21]). (In fact, because of hydrogen
bonding, the acidity of tBuOH is heavily concentration-
dependent, varying from pKa ˆ 20 for the neat liquid to pKa ˆ
30 for the dilute solution in DMSO.[20] The pKa value of
PhCH(Et)CN is possibly slightly higher than of PhCH2CN
(pKa ˆ 21.3 in DMSO[20]). It was shown that Fp is protonated
The effect of anion scavengers: The results obtained indicate
that aryl carbanions play an important role in the reaction of
FpK with C6F5Cl, at least in the reduction of aryl halide.
Carrying out the reaction in the presence of anion scavengers
(proton donors) is likely to provide further evidence of the
[C6F5] anion participating in the process. The solvent MeCN
turned out to be sufficiently acidic[c] to protonate the [C6F5]
carbanions that are involved in the formation of chloroper-
[c] The acidity of MeCN media may be even higher than reported for
MeCN (pKa ˆ 29) in dilute solution in DMSO.[18]
1172
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1172 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 7

Halogen±Metal Exchange
1169±1178
only to a small extent even in
THF/tBuOH (1:1) mixed sol-
vent.[22]) The reaction pattern
observed was the same with
both traps, the yield of the
dimer Fp2 being near to quanti-
tative and only proton-contain-
ing products (mainly C6F5H)
being formed from the reduc-
tion of C6F5Cl (entries 4,5, Ta-
ble 2).
The protonation of penta-
fluorophenyl carbanion by
tBuOH or CH acid gave rise to
another anion derived from the
proton donor. These anions en-
ter into nucleophilic substitu-
tion reactions with C6F5Cl and
C6F5H,
displacing
fluorine
chiefly in the para position,
though minor amounts of ortho
isomers were also detected
Scheme 3. General mechanism of the reaction.
(Scheme 2). This reaction con-
considering the large difference between the redox potentials
sumes [tBuO] anion quantitatively, since the total yields of
the products containing the tBuO group and of the products
derived from the protonation of pentafluorophenyl carbanion
are approximately the same (about 50%).[d] The total yield of
substitution products with [CH2CN] and [PhC(Et)CN]
carbanions is considerably smaller and does not exceed
20%, indicating the operation of yet another reaction path-
way where carbanions are consumed. The same is true for the
transformations of the [C6F5] carbanion, the observed yield
of chloroperfluoropolyphenyls being usually three to seven
times lower than of Fp2 (Table 1), while according to the
reaction balance [Eq. (a)] it should not be less than half[d, e] of
the yield of Fp2.
Several conclusions may be drawn from the results of anion
scavenging experiments. a) The substitution of chlorine for
the Fp group does not involve direct attack of the carbon-
ylmetallate nucleophile on carbon; this rules out the SN2Ar
mechanism. Strictly speaking it should not be referred to as a
nucleophilic substitution reaction. b) Pentafluorophenyl car-
banion is proved to be the essential intermediate in the
formation of the chlorine substitution product, C6F5Fp. c) The
overall reaction scheme must consist of at least two steps, the
first one, producing [C6F5] carbanions, the same for all
products formed in the following step(s). Halogen ± metal
exchange (HME) as the first step of the reaction (reaction A,
Scheme 3) appears to be the only plausible choice, provided
the alternative of a radical reaction pathway is positively
excluded. It should be pointed out that double electron
transfer could also produce [C6F5] carbanions.
2.02 V,[7] EoFp
ˆ
Red
of the reactants (E
ˆ
.
=Fp
C F Cl=C F Cl
.
6
5
6 5
1.16 V[23]). The ET rate constant (k < 10 4 Lmol 1 s )
calculated from Eberson formulae based on Marcus ± Hush
theory[24] is strongly underestimated, as the reaction happens
within the time of mixing even at 1008C.
1
The effect of radical scavengers on the product composition
was carefully examined. Secondary phosphines are good
hydrogen-atom donors and their use is well established in the
detection of the radical steps in various nucleophilic substi-
tution reactions;[25] we utilised iPr2PH and tBu2PH to inter-
.
cept C6F5 radicals. The product distribution data for these
experiments summarised in Table 3, and the high yield of
.
C6F5Fp in particular, imply that C6F5 is not involved in the
reaction (entries 2,3, Table 3). It seems somewhat puzzling
that even a slight increase in the yield of C6F5Fp was observed
in comparison with the standard experiment. An increased
amount of C6F5H is also formed while the proportion of
chloroperfluoropolyphenyls is reduced in the presence of the
secondary phosphine (entries 2,3, Table 3); however, it is not
clear whether this results from radical or anion trapping.
The ability of tertiary phosphines to intercept metal
carbonyl radicals, even when they are formed as short-lived
intermediates, arises from facile ligand substitution typical of
these species.[26] When the reaction was performed in the
presence of PEt3 (entry 4, Table 3) the primary organometal-
lic products remained the same, yet the yield of C6F5Fp was
distinctly lower and a variety of additional, presumably
phosphine-substituted products were formed, though each in
1
trace amount. Judging from H, 19F and 31P NMR spectra of
On the electron-transfer mechanism: Outer-sphere electron
transfer (ET) from FpK to C6F5Cl does not seem feasible
the reaction mixture the majority of the new products are not
dimeric iron carbonyls but belong to the [ArFFe(Lx)Cp] type.
[26a, c]
Taking into account that the starting carbonylmetallate
and the reaction products are inert towards phosphine
substitution, we have to admit that Fp radicals are formed
as intermediates in a certain step of at least one of the reaction
[d] Relative to the amount of FpK entering in the reaction.
[e] For example, if only C6F5C6F4Cl is formed its stoichiometric yield
should be 100%.
.
Chem. Eur. J. 1998, 4, No. 7
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1173 $ 17.50+.25/0
1173

I. P. Beletskaya et al.
FULL PAPER
Table 3. The effect of radical traps on the course of the reaction of C6F5Cl with FpK in THF at room temperature.
Entry
Reactants and reaction conditions[a]
Ratio
Relative yields [%][b]
Trap
C6F5Fp
Fp2
C6F5(C6F4)nC6F4Cl
C6F5H
n(C6F5Cl)/n(Fp )
(n ˆ 0 ± 3)
1
2
3
4
±
1.25
57
71
65
35
43
29
35
48
11
±
3
2.5
11
9
iPr2PH (9 equiv)
tBu2PH (13 equiv)
Et3P[c] (4 equiv)
1.11
1.40
1.11
3
4
[a] The initial concentrations of reactants were 0.03 ± 0.09m. [b] Concerning the determination and calculation of the product yields see footnote [b] to
Table 1. [c] The yield of [{h4-C6F5C5H5}Fe(CO)2PEt3] ꢀ 5% and the yields of other Cp-containing products ꢀ 12%.
pathways. The only isolated and identified product containing
phosphine was [{h4-(C6F5)C5H5}Fe(CO)2PEt3] (yield 5%).
Complexes of this type have recently been obtained with
high yield in the three-component reaction of FpHal, RLi and
R3P,[27] and adopting the mechanism proposed[28] the forma-
tion of the complex in our case is described in Scheme 4. The
parallel with the ease of halogen transfer (I > Brꢁ Cl > F).[31]
The HME mechanism now proposed for the reaction of Fp
with C6F5Cl, can apparently be extended to other reactions of
carbonylmetallates with halopolyfluoroaromatics, giving an
insight into the reverse leaving-group effect that hitherto
lacked adequate explanation.[2, 5±7] It is particularly remark-
able that halopolyfluoroar-
enes are sufficiently activated
to react according to an
SN2Ar
mechanism,[6, 8, 14]
though they can be metal-
lated with certain halophilic
reagents, such as alkyl-
lithiums.[15, 31b]
The question concerning
the origin of specific halophil-
Scheme 4. Mechanism of the formation of [h4-(C6F5C5H5)Fe(CO)2PEt3], based on ref. [28]. The oxidant Ox is other
than C6F5Cl.
ic reactivity of metal-centred
detection of such a complex provides extra evidence for HME
in the reaction.
Thus, experiment not having confirmed the hypothesis of a
radical reaction pathway producing [C6F5] carbanions, the
following discussion will be based on the framework of the
HME mechanism (Scheme 3).
anions is certainly intriguing, but lies beyond the scope of our
current study. At present one could refer to the HSAB
principle[32] predicting the interaction of the soft base with the
soft acid center, in our case the metal-centered anion with the
halogen atom[31a] of the aryl halide. The higher energy of
metal ± halogen as compared to carbon ± metal bonds
may be also taken into consideration.[33] It was shown recently
that even in o-nitroiodobenzene, representative of
a
The halogen ± metal exchange mechanism: Nucleophilic ar-
omatic substitution by an HME mechanism is rather uncom-
mon; the few examples known are limited to the reactions of
metal-centred nucleophiles with nonactivated halobenzenes,
where the SN2Ar substitution mechanism is unfeasible. Thus,
for the reactions of the Me3Sn anion with halobenzenes the
mechanism with halogen transfer in the solvent cage as a key
step was established by Kuivila et al.[29]
typical SN2Ar-reactive substrate, soft nucleophiles such as
thiol anions attack not carbon but iodine atoms, the sub-
stitution product being possibly formed through this path-
way.[34]
Scheme 3 summarises the reactions that can follow the
initial HME step. It is important to note that in the halogen
transfer process along with pentafluorophenyl carbanions an
equivalent amount of FpCl should be generated. However, no
FpCl was detected in the experiments when pentafluorophen-
yl carbanion was trapped with anion scavengers. Instead Fp2
was formed almost quantitatively. We suppose that FpCl is
formed only as an intermediate, and converted to Fp2 in the
rapid reaction with FpK (the reported value of the rate
The effect of MeCN, similar to that observed in our study,
has been reported in the reaction of another metal-centred
ꢃ
nucleophile, [Re(O)(R ± C C ± R)2] (R ˆ Et, Ph) with hal-
obenzenes.[30] Instead of the s-aryl rhenium complex observed
in benzene solution, when the reaction was carried out in
ꢃ
MeCN cyanomethyl complex NCCH2 ± [Re(O)(RC CR)2]
‡
was formed with the equivalent amount of benzene.[30] The
authors ascribed the effect to the interception of phenyl
radicals by MeCN.[30] Probably these results may be reinter-
preted in favour of phenyl anions being trapped by the
solvent, taking into account the clear evidence obtained on
this point in the current study of a similar reaction.
constant for the reaction of Fp PPN with FpCl in MeCN is
103 Lmol 1 s 1[35]) which is assumed to follow the electron
transfer (step C in Scheme 3).[26a, 35] The reported value of the
‡
rate constant for the reaction of Fp PPN with FpCl in MeCN
.
is 103 lmol 1 s 1.[35] In our case Fp radicals should recombine
predominantly in the solvent cage, since they are not trapped
by triethylphosphine (entry 4, Table 3).
ꢃ
In reactions of both Me3Sn and [Re(O)(R ± C C ± R)2]
nucleophiles[29, 30] the leaving group effect is the opposite of
that usually observed in nucleophilic aromatic substitution
(F ꢁ Cl,Br,I), the reactivity of halobenzenes increasing in
1174
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1174 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 7

Halogen±Metal Exchange
1169±1178
When the pentafluorophenyl carbanion is not captured by
anion scavengers, reaction C competes with its coupling with
FpCl [step B and Eq. (2)], thus decreasing the yield of
C6F5Fp.
Conclusion
In the reaction of pentafluorochlorobenzene with the carbon-
ylferrate anion, nucleophilic substitution of chlorine and aryl
halide reduction are not competing pathways. All of the
observed transformations take place in the framework of a
single mechanism. Halogen transfer between aryl halide and
metal-centred anion, generating pentafluorophenyl carban-
ion, is proposed as the key step of the entire process.
Investigation of the reactions of other carbonylmetallates
([Re(CO)5] , [Mn(CO)5] , etc.) with halopolyfluorobenzenes
is currently in progress in an attempt to obtain further
evidence to confirm the halogen ± metal exchange mecha-
nism, and to estimate the scope of this unusual reactivity. The
question concerning the mechanism of the reactions of
carbonylmetallates with perfluoroarenes is of particular
interest in the context of the characteristic ion-pairing effect
in these reactions.
We expected to obtain FpCl as the final product in the
experiment when FpK was added dropwise to the tenfold
excess of C6F5Cl (entry 6, Table 2). The rates of the reactions
of the anions Fp and [C6F5] with C6F5Cl (A and D in
Scheme 3) are increased, but the reactions consuming FpCl (B
and C) are not affected. The fact that not even traces of FpCl
could be detected in this experiment forces us to assume that
the increase in the production rate of FpCl is not sufficient to
counterbalance its disappearance. However, the product
composition did change, the ratio of products C6F5Fp to Fp2
decreasing almost threefold and the yield of chloroperfluor-
opolyphenyls, especially p-ClC6F4C6F5, considerably increas-
ing (entry 6, Table 2). These results point to the role of C6F5Cl
as an anion scavenger. The interception of the pentafluor-
ophenyl anion by C6F5Cl (reaction D, Scheme 3) competes
with the reaction of the anion with FpCl (B). This conclusion
is confirmed by the detection of Fp-substituted perfluoro-
polyphenyls in standard runs because it obviously indicates
that chloroperfluoropolyphenyls are formed in the reaction
when Fp is still present, that is, concurrently with C6F5Fp.
Thus, the first halogen transfer step of the reaction,
common to all products formed (A in Scheme 3), appears to
be rate-limiting, given the absence of FpCl among the
products. The actual product distribution is determined in
the subsequent rapid steps of the reaction (B, C and D ). Each
of the two exchange products, FpCl and [C6F5] , participates
in a side reaction, C or D respectively, giving rise to Fp2 and
the reduction products of C6F5Cl. The fact that there are two
reactions in competition with the reaction yielding C6F5Fp
(reaction B) suggests a possible explanation of the solvent and
counterion effects on the product ratio. We believe that the
temperature effect has its origin in ion pairing and is related to
the increase of the dielectric constant at lower temperatures
(for THF e ˆ 7.4 at 258C and 11.6 at 708C[37]). Let us
consider the following arguments: Ion pairing is known to
decrease the reactivity of carbanions in nucleophilic aromatic
substitution.[5a, 6, 36] On the other hand the reactions of
carbonylmetallates with metal carbonyl halides or corre-
sponding cationic complexes exhibit the reverse reactivity
pattern: tight ion pairs react faster than free ions.[26a] No data
is available concerning the effect of ion pairing in the
reactions of carbanions with metal carbonyl halides (B,
Scheme 3); however, let us speculate that it is less dramatic
than in the competitive reactions C and D. As a result of
strong ion pairing in the low-polarity solvent (Et2O/THF) the
reaction of FpCl with Fp (C) becomes the fastest in the
competition, consuming almost all the FpCl initially formed.
The fastest process in polar solvent or with bulky cations will
be the interception of [C6F5] anion by C6F5Cl (D), where it
prevails over the production of C6F5Fp. On the other hand it
may be that in these conditions the reaction of FpCl with
[C6F5] anion may itself produce a substantial amount of Fp2
instead of C6F5Fp. That the highest yield of C6F5Fp was
obtained in THF indicates that all competitive pathways are
relatively slow in the solvent of medium polarity.
Experimental Section
A vacuum-line technique was used for handling carbonylferrate and
carbanion salts. All other air-sensitive compounds, such as neutral metal
carbonyl derivatives, were handled in a vacuum or under argon with
Schlenk techniques. Argon was purified by passage through oxygen
scavenger and 4 Š molecular sieve columns.
Solvents and reagents: All solvents were distilled from an appropriate
drying agent. MeCN, [D3]MeCN and [D6]DMSO were vacuum-transferred
from the sodium salt of 9-benzylfluorene; THF, [D8]THF and Et2O from
sodium benzophenone ketyl. (Me3Si)2O, C6F5Cl, C6H5CF3, 18-crown-6-
ether were fractionally distilled prior to use. Carbonylferrate dimer (Fp2)
was purified according to the previously described method.[17] Bis(diphe-
‡
nylphosphinoiminium) chloride (PPN Cl ) and LiCl were dried in vacuum
(10 2 Torr) at 120 ± 1408C for 1.5 h.
Break-seal ampoules were used for the transfer of substances. Volatile
reagents and standards were degassed and vacuum-transferred from the
appropriate drying agent: C6F5Cl, C6F6, (Me3Si)2O, C6H6, PEt3 and
C6H5CF3 from sodium mirror, iPr2PH and tBu2PH from P2O5, tBuOH
from tBuOK, Ph(Et)CHCN from NaH. The 19F NMR spectrum of C6F5Cl
agrees with that reported in the literature.[38] Small volumes (<0.5 mL) of
the reagents were placed in thin-walled glass phials that were evacuated
and flame-sealed. For the low-temperature experiments C6F5Cl was dosed
as THF solution.
Spectroscopic methods and reaction mixture analysis: 1H (400 MHz) and
19F (376.3 MHz) NMR spectra were obtained on a Varian VXR-400
spectrometer at room temperature. The chemical shifts d for 1H NMR
spectra were reported relative to TMS and referred to the signals of the
internal reference standards, either (Me3Si)2O (d ˆ 0.05) or C6H6 (d ˆ 7.27).
19F NMR chemical shifts d were reported in ppm downfield from CFCl3
relative to signals of the internal reference standards C6F6 (d ˆ 162.9) or
C6H5CF3 (d ˆ 61.5). The routine analysis of the reaction mixtures was
performed by NMR spectroscopy in undeuterated solvents, in which case
1H NMR spectra were recorded using long pulse delay (PD ꢄ 7 s) and short
impulses (PWˆ 1 ms). We observed that pulse delay variation (1 ± 30 s) has
no impact on the relative integral intensity of the signals in 19F NMR
spectra. The yields of the products were calculated separately from 1H and
19F NMR spectra by comparison of the integral intensity of corresponding
signals, Cp groups in particular, to that of the internal standard (one of the
above listed). If no fluorine standard was added, then the signal of C6F6
impurity in C6F5Cl was used for the chemical shift reference in 19F NMR
spectra and the product yields were calculated relative to that of C6F5Fp,
1
determined from the H NMR spectrum.
IR spectra were recorded on a UR-20 spectrophotometer in 0.2 mm CaF2
cell. Mass spectra were recorded on a MS-890 spectrometer operating with
Chem. Eur. J. 1998, 4, No. 7
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1175 $ 17.50+.25/0
1175

I. P. Beletskaya et al.
FULL PAPER
Table 4. Spectral data for the reaction products.
Solvent 1H NMR d 19F NMR d[a]
IR nÄCꢃO
MS m/z
1
[cm
]
Fp2
THF
4.82(s)
1791(s),
1960(s),
1999(s)
Fp2
FpCl
FpCl
X
MeCN 4.80(s)
THF 5.12(s)
MeCN 5.05(s)
‡
‡
‡
Fp
THF
5.11(s)
105.38(F1), 163.59(F2),
161.18(t,F3)
1999(s),
2042(s)
344 [M] , 316 ([M CO] , 288 [M 2CO] , 268 [M
‡
‡
‡
2CO HF] , 232 [C6F5Cp] , 213 [C6F5Cp F] ,
194 [C6F3Cp] , 193 [C6F3Cp H] , 168 [C6F5H] , 175
‡
‡
‡
Fp
MeCN 5.01(s)
THF
105.73(F1), 163.90(F2),
161.34(t, F3)
‡
‡
‡
[C6F2Cp] , 174 [C6F2Cp H] , 140 [CpFe]
‡
C6F4Cl
(C6F4)2Cl
137.52, 137.91, 139.87,
161.11(F2), 150.69(t, F3)
136.85, 136.93, 137.68,
139.67, 137.16, 160.98(F2),
150.34(t, F3)
136.85, 136.93, 137.16,
137.68, 139.69, 160.98(F2),
150.34(t, F3)
350 [M] (100%)
‡
THF
498 [M] (100%)
‡
(C6F4)3Cl
(C6F4)4Cl
H
THF
THF
THF
646 [M] (100%)
‡
136.58, 136.77, 137.09,
137.64, 139.69, 160.96(F2),
150.26(t, F3)
794 [M] (100%)
138.61(F1), 162.57(F2),
154.71(t, F3)
[{h4-C5H5}Fe(CO)2PEt3][b] THF
145.214 (F1), 164.13(F2),
159.84(t, F3)
1918(s),
1976(s)
462 [M] , 434 ([M CO] , 406 [M 2CO] , 344 [M
‡
‡
‡
‡
‡
PEt3] , 288 [M 2CO PEt3] , 268 [M 2CO PEt3
HF] , 232 [C6F5Cp] , 213 [C6F5Cp F] , 194 [C6F3Cp] ,
193 [C6F3Cp H] , 168 [C6F5H] , 175 [C6F2Cp] , 174
[C6F2Cp H] , 140 [CpFe] , 118 [PEt3]
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
X'
Y
Fp
Fp
H
C6F4Fp
THF
THF
5.10
5.17(s)
107.30 (F1), 140.78 (F2)
106.55 (F1), 141.39 (F2)
‡
‡
‡
1993(s),
2042(s)
650 [M] , 594 [M 2CO] , 538 [M 4CO] , 362
‡
‡
‡
[HC6F4C6F4Cp] , 298 [HC6F4C6F4H] , 186 [FeCp2] , 121
‡
[FeCp]
Fp
Fp
Cl
H
Cl
H
(C6F4)2Fp
(C6F4)3Fp
CH2CN
C(Ph)(Et)CN
C(Ph)(Et)CN
tBuO
THF
THF
MeCN
THF
THF
THF
THF
5.20(s)
5.21(s)
105.41 (F1), 141.33 (F2)
105.22 (F1), 141.15 (F2)
144.06, 147.50
137.00 (F1), 137.70 (F2)
135.89, 140.49
140.37 (F1), 151.43 (F2)
141.08 (F1), 150.18 (F2)
Cl
tBuO
X''
Y'
Cl
tBuO
THF
140.03 (dd, F1), 158.62,
162.14 (t, F2, F3), 148.05 (F4)
[a] All signals are second-order multiplets except where noted otherwise. [b] 31P NMR chemical shift for [{h4-C6F5C5H5}Fe(CO)2PEt3] is d ˆ 63 relative to
H3PO4 (external standard).
electron impact ionisation, 70 eV ionising energy, direct inlet method, with
a source temperature of 1508C.
C6F5Cl (0.15 ± 0.22 mmol) were used. In all of the experiments the substrate
was in excess (15 ± 50%); the concentration of Fp varied in the range of
0.03 ± 0.13 molL 1. Various additives (18-crown-6, radical or anion trap),
C6F5Cl, NMR standard(s) were introduced by breaking thin-walled glass
phials containing these substances (usually in the sequence above). The
solution was rapidly stirred by magnetic stir bar or by manual shaking of the
apparatus at the moment when the substrate was added. The reaction
proceeded instantly; the mixture immediately turned red/black in colour
and turbid. After 10 ± 20 min about 0.6 ± 1 mL of the reaction mixture was
filtered through Celite into the NMR sample tube fused to the apparatus,
and deuterated solvent was added to the filtrate ( ꢀ 0.2 mL of [D8]THF or
CD3CN). The NMR sample tube was then frozen in liquid nitrogen and
General procedure for the reaction of Fp with C6F5Cl: FpK was obtained
by reducing Fp2 with NaK2.8 alloy in THF according to the literature
procedure;[10] the resulting anion solution was then filtered through a fine-
‡
pore glass filter. PPN Fp was prepared beforehand and FpLi was obtained
‡
[39]
in situ by the exchange reactions of FpK with PPN Cl
and LiCl
accordingly. If necessary THF was replaced with another solvent. In the
experiments from which the reaction products were isolated, 0.8 ± 1.3 mmol
of Fp was allowed to react with 0.9 ± 2.0 mmol of C6F5Cl. When only NMR
analysis was performed smaller amounts of Fp (0.11 ± 0.17 mmol) and
1176
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1176 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 7

Halogen±Metal Exchange
1169±1178
sealed off with flame. The apparatus was filled with argon and in some cases
Chem. 1968, 80, 835 ± 841; Angew. Chem. Int. Ed. Engl. 1968, 7, 747 ±
753; d) R. Chukwa, A. D. Hunter, B. D. Santarsiero, Organometallics
1991, 10, 2141 ± 2152.
a
sample was withdrawn for analysis by IR spectroscopy. In the
preparative-scale experiments the remaining solution was transferred
under argon pressure through a teflon hose to the column with a Celite
filter. The solution was filtered into a round-bottomed flask and evaporated
with silica gel at reduced pressure.
[4] W. Beck, B. Niemer, M. Wieser, Angew. Chem. 1993, 105, 969 ± 996;
Angew. Chem. Int. Ed. Engl. 1993, 32, 923 ± 1110.
[5] a) I. P. Beletskaya, G. A. Artamkina, A. Yu. Milꢀchenko, P. K.
Sazonov, M. M. Shtern, J. Phys. Org. Chem. 1996, 9, 319 ± 328;
b) G. A. Artamkina, A. Yu. Milꢀchenko, I. P. Beletskaya, O. A.
Reutov, J. Organomet. Chem. 1986, 311, 199 ± 206; c) G. A. Artamki-
na, A. Yu. Milꢀchenko, I. P. Beletskaya, O. A. Reutov, ibid. 1987, 321,
371 ± 376.
[6] G. A. Artamkina, A. Yu. Milꢀchenko, I. P. Beletskaya, O. A. Reutov,
Dokl. Akad. Nauk. 1989, 304, 616 ± 621.
[7] T. V. Magdesieva, I. I. Kukhareva, G. A. Artamkina, I. P. Beletskaya,
K. P. Butin, Zh. Org. Khim. 1994, 30, 591 ± 597.
[8] a) J. F. Bunnett, R. E. Zahler, Chem. Rev. 1951, 49, 273; b) J. Miller in
Aromatic Nucleophilic Substitution (Eds.: C. Eaborn, N. B. Chapman),
Elsevier, New York, 1968; c) T. J. de Bour, I. P. Direx in The Chemistry
of the Nitro and Nitroso Groups, Vol. 5 (Ed.: H. Feuer), Mir, Moscow,
1972, p. 371 (Russian translation).
[9] J. K. Stille, K. S. Y. Lay, Acc. Chem. Res. 1977, 10, 434 ± 442.
[10] J. E. Ellis, E. A. Flom, J. Organomet. Chem. 1975, 99, 263 ± 268.
[11] S. C. Cohen, J. Chem. Soc. Dalton Trans. 1973, 553 ± 555.
[12] G. B. Richtor-Ado, A. D. Hunter, N. Wichriwska, Can. J. Chem. 1990,
68, 41 ± 48.
Isolation of the reaction products: The dry residue from the reaction of
FpK (prepared from 236 mg, 0.667 mmol Fp2) with C6F5Cl (396 mg,
1.96 mmol) in THF (31.2 mL) at RT was subjected to column chromatog-
raphy on silica gel (40/100, L ˆ 15 cm, d ˆ 1.5 cm). Elution with petroleum
ether produced a colourless band from which a mixture of chloroperfluor-
opolyphenyls (17.9 mg, ꢀ 8%, approximate yield calculated for one
aromatic ring) was obtained. Elution with CH2Cl2/petroleum ether (1:4)
afforded two bright yellow bands. The first band after removal of the
solvent afforded crude C6F5Fp (225 mg, 0.654 mmol, 49%). The second
band consisted primarily of Fp(C6F4)nFp (n ˆ 2 ± 5, 10.0 mg, ꢀ 2.5%,
approximate yield calculated for one aromatic ring) contaminated with
C6F5Fp. Elution with CH2Cl2/petroleum ether (1:1) gave a purple band
from which Fp2 (70.0 mg, 0.396 mmol, 30%) was obtained. Elution with
Et2O gave a brick-red band from which a red-brown solid (10.4 mg) was
obtained, in which FpCl was found by TLC comparison with an authentic
sample. Isolated substances were dried under vacuum. Individual homo-
logues of chloroperfluoropolyphenyls were isolated and Fp(C6F4)nFp was
purified by preparative-plate chromatography on Silpearl. Pure C6F5Fp was
obtained by crystallisation from C6H6/hexane (1:1) at 308C.
[13] a) B. E. Edelbach, W. D. Jones, J. Am. Chem. Soc. 1997, 119, 7734 ±
7742; b) M. Shmulinson, A. Pilz, M. S. Eisen, J. Chem. Soc. Dalton
Trans. 1997, 2483 ± 2486.
[14] R. D. Chambers, D. Close, D. L. H. Williams, J. Chem. Soc. Perkin
Trans. II 1980, 778 ± 781.
[15] a) S. C. Cohen, M. L. Reddy, D. M. Roe, A. J. Tomlinson, A. G.
Massey, J. Organomet. Chem. 1968, 14, 241 ± 251; b) S. C. Cohen, D. E.
Fenton, D. Show, A. G. Massey, ibid. 1967, 8, 1 ± 8.
[16] J. A. Heppert, M. A. Morgenstern, D. M. Scherubel, F. Takusagawa,
M. R. Shaker, Organometallics 1988, 7, 1715 ± 1723.
[17] G. A. Artamkina, M. M. Shtern, P. K. Sazonov, I. P. Beletskaya, Zh.
Org. Khim. 1996, 32, 1319 ± 1328 [Russ. J. Org. Chem. 1996, 32, 1271 ±
1280 (English translation)].
Addition of reagents in reverse order: Solution of FpK (0.216 mmol) in
THF (1.2 mL) was added dropwise (over 5 min) through a capillary to the
rapidly stirred solution of C6F5Cl (2.10 mmol) in THF (3.2 mL). After
complete mixing and addition of the NMR standard, part of the reaction
mixture was transferred to an NMR sample tube. The apparatus was then
filled with argon and a sample for IR spectroscopy was taken.
Product structure assignment: Spectra are presented in Table 4. All
1
isolated compounds were characterised by H and 19F NMR spectroscopy
and mass spectrometry. Metal carbonyl derivatives were also characterised
by IR spectroscopy. The spectral properties of C6F5Fp,[40, 41] p-
FpC6F4C6F4Fp,[40, 41] [{h4-C6F5C5H5}Fe(CO)2PEt3][27, 42] are in good agree-
ment with the literature data for these or similar compounds.
A number of reaction products were not isolated, but identified as follows.
Signals in 19F NMR spectra of the reaction mixture corresponding to C6F5H
and p-HC6F4Fp were identified by comparison with the spectrum of an
authentic sample of C6F5H and the reported[40] spectrum of p-HC6F4Fp. The
products resulting from the interaction of carbanions ([CH2CN] ,
[PhC(Et)CN] ) and [tBuO] with C6F5Cl or C6F5H were identified
similarly. To recognise their signals in the spectrum of the complex product
mixture obtained in the reaction of FpK with C6F5Cl, separate test
experiments were conducted. Carbanion salts were prepared by trans-
metalation of PhCH(Et)CN and MeCN with Ph3CK. An excess of substrate
was added to the carbanion or [tBuO] solution in the appropriate solvent
(THF or MeCN) and the 19F NMR spectrum of the resulting solution was
recorded. The signals of the major substitution products observed in the
test experiments coincided with the signals in question to within 0.1 ppm.
These products were assigned structures in accordance with general
regularities of fluorine chemical shifts and coupling constants.[38, 40]
[18] O. A. Reutov, I. P. Beletskaya, K. P. Butin, CH Acids, Nauka, Moscow,
1980, p. 26 ± 25 (in Russian).
[19] Ref. [18], pp. 19, 21, 40.
[20] Ref. [18], pp. 25, 26.
[21] Ref. [1c], ch. 3.
[22] T. Bauch, A. Sanders, C. V. Magatt, P. Waterman, D. Judelson, W. P.
Giering, J. Organomet. Chem. 1975, 99, 269 ± 279.
[23] M. Tilset, V. D. Parker, J. Am. Chem. Soc. 1989, 111, 6711 ± 6717.
[24] L. Eberson, Acta Chem. Scand. 1984, B38, 439 ± 459.
[25] a) G. F. Smith, H. G. Kuivila, R. Simon, L. Sultan, J. Am. Chem. Soc.
1981, 103, 833 ± 839; b) E. C. Ashby, R. N. De Priest, W.-Y. Su,
Organometallics 1984, 3, 1718 ± 1727.
[26] a) K. Y. Lee, D. J. Kuchynka, J. K. Kochi, Organometallics
1987, 6, 1886; b) A. Fox, J. Molito, A. Poe, J. Chem. Soc.
Chem. Commun. 1981, 1052; c) T. L. Brown, D. R. Kidd, J. Am.
Chem. Soc. 1976, 100, 4095; d) B. H. Byers, T. L. Brown, ibid. 1977, 99,
2527.
Acknowledgements: We gratefully acknowledge the financial support from
the Russian Fundamental Science Foundation (Grants 97-03-33161a and
96-15-97484) and from the program Integration of High School with the
Russian Academy of Sciences (Grant 234). We thank Dr. Yu. A. Veytz
(MSU, Chemistry Department) for the gift of iPr2PH and tBu2PH.
[27] L.-K. Liu, L. S. Luh, Organometallics 1994, 13, 2816 ± 2824.
[28] S. L. Gipson, L.-K. Liu, R. U. Soliz, J. Organomet. Chem. 1996, 526,
393 ± 395.
[29] H. R. Wursthorn, H. G. Kuivila, G. F. Smit, J. Am. Chem. Soc. 1978,
100, 2779 ± 2789.
[30] R. Conry, J. P. Meyer, Organometallics 1991, 10, 3160 ± 3166.
[31] a) N. S. Zefirov, D. I. Makhonꢀkov, Chem. Rev. 1982, 82, 615; b) T. V.
Talalaeva, K. A. Kocheshkov, Methods of Organometallic Chemistry,
Vol. 1, Nauka, Moscow, 1971, pp. 215 ± 251 (in Russian).
[32] a) R. G. Pearson, J.Am. Chem. Soc. 1963, 85, 3533 ± 3539; b) Hard and
Soft Acids and Bases (Ed.: R. G. Pearson), Stroudsbourg, Dowden,
1973, Parts I, II, VI.
[33] a) M. E. OꢀNeill, K. Wade in Comprehensive Organometallic Chem-
istry, Vol. 1 (Ed.: G. Wilkinson), Oxford, Pergamon, 1982, pp. 5 ± 9;
b) ref. [1c], p. 104.
Received: December 4, 1997 [F914]
[1] a) R. E. Dessy, R. L. Pohl, R. B. King, J. Am. Chem. Soc. 1966, 88,
5121; b) M. I. Darensbourg, Progr. Inorg. Chem. 1985, 37, 221 ± 274;
c) J. P. Collman, L. S. Hegedus, J. R. Norton, R. G. Finke, Principles
and Applications of Organotransition Metal Chemistry, University
Science Books, Mill Valley, California, 1987, ch. 5.
[2] a) M. I. Bruce, F. G. A. Stone, J. Chem. Soc. A 1966, 1837 ± 1842;
b) N. J. Ishikawa, S. Ibuki, Bull. Chem. Soc. Jpn. 1974, 47, 2621 ± 2622.
[3] a) R. B. King, Adv. Organomet. Chem. 1964, 2, 157; b) R. B. King,
Acc. Chem. Res. 1970, 3, 417; c) M. I. Bruce, F. G. A. Stone, Angew.
[34] S. Montanari, C. Paradisi, G. Scorrano, J. Org. Chem. 1993, 58, 5628 ±
5631.
Chem. Eur. J. 1998, 4, No. 7
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1177 $ 17.50+.25/0
1177

I. P. Beletskaya et al.
FULL PAPER
[35] W. S. Striejewsky, R. F. See, M. L. Churchill, J. D. Atwood, Organo-
metallics 1993, 12, 4413 ± 4419.
[39] a) Y. Zhen, W. G. Feighery, C. K. Lai, J. D. Atwood, J. Am. Chem.
Soc. 1989, 111, 7832; b) J. K. Ruff, Inorg. Synth. 1974, 15, 84 ± 90.
[40] M. I. Bruce, J. Chem. Soc. A 1968, 1459 ± 1464.
[41] M. I. Bruce, M. A. Thomas, Org. Mass Spectrom. 1968, 1, 835 ± 846.
[42] P. M. Treichel, R. L. Shubkin, Inorg. Chem. 1967, 6, 1328 ± 1334.
[36] C. Germain, Tetrahedron 1976, 32, 2389 ± 2393.
[37] M. Szwarc, Carbanions, Living Polymers and Electron Transfer
Processes, Mir, Moscow, 1971, pp. 166 ± 167 (Russian translation).
[38] P. Bladon, D. W. A. Sharp, J. M. Winfield, Spectrochim. Acta, 1964, 20,
1033 ± 1042.
1178
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0407-1178 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 7