Mechanism of carbon-fluorine bond activation by (C5Me5)Rh(PMe3)H2

Article abstract of DOI:10.1021/ja970723r

The complex Cp*Rh(PMe₃)H₂ (Cp* = C₅Me₅) reacts with C₆F₆, C₆F₅H, C₁₂F₁₀, or C₁₀F₈ in pyridine or 1:1 pyridine/benzene to give the C-F cleavage products Cp*Rh(PMe₃)(aryl(F))H in high yield. Kinetic studies reveal that the reaction has autocatalytic character, and fluoride ion is shown to be responsible for the catalysis. The anion [Cp*Rh(PMe₃)H]^- reacts rapidly with C₁₂F₁₀ or C₁₀F₈ to give the same C-F cleavage products as Cp*Rh(PMe₃)H₂. A mechanism initiated by deprotonation of Cp*Rh(PMe₃)H₂ followed by nucleophilic attack of the resulting anion on the polyfluoroaromatic with subsequent loss of fluoride is proposed. The fluoride ion continues the cycle by deprotonating Cp*Rh(PMe₃)H₂.

Full text of DOI:10.1021/ja970723r

7734
J. Am. Chem. Soc. 1997, 119, 7734-7742
Mechanism of Carbon-Fluorine Bond Activation by
(C5Me5)Rh(PMe3)H2
Brian L. Edelbach and William D. Jones*
Contribution from the Department of Chemistry, UniVersity of Rochester,
Rochester, New York 14627
ReceiVed March 6, 1997X
Abstract: The complex Cp*Rh(PMe3)H2 (Cp* ) C5Me5) reacts with C6F6, C6F5H, C12F10, or C10F8 in pyridine or
1:1 pyridine/benzene to give the C-F cleavage products Cp*Rh(PMe3)(arylF)H in high yield. Kinetic studies reveal
that the reaction has autocatalytic character, and fluoride ion is shown to be responsible for the catalysis. The anion
[Cp*Rh(PMe3)H]- reacts rapidly with C12F10 or C10F8 to give the same C-F cleavage products as Cp*Rh(PMe3)H2.
A mechanism initiated by deprotonation of Cp*Rh(PMe3)H2 followed by nucleophilic attack of the resulting anion
on the polyfluoroaromatic with subsequent loss of fluoride is proposed. The fluoride ion continues the cycle by
deprotonating Cp*Rh(PMe3)H2.
Introduction
[Ru(dmpe)2(C6F5)H] with loss of HF and formation of some
Ru(dmpe)2(H)(HF2).11
The activation of carbon-fluorine bonds by metal complexes
has proven to be a chemical challenge. The dissociation energy
of the C-F bond has been cited as high as 154 kcal mol-1 for
C6F61a (vs 110 kcal mol-1 for C6H61b). Despite this, numerous
examples of C-F bond activation by metal complexes have
been cited.2 Recently, Aizenberg and Milstein have reported
homogenous catalytic C-F bond hydrogenation by a late
transition metal complex.3 Catalytic formation of perfluoro-
aromatic compounds by reductive defluorination of saturated
perfluorocarbons by an early transition metal complex has also
been achieved.4
Our group, in conjunction with Perutz and co-workers, has
reported photochemically promoted C-F bond cleavage with
(C5Me5)Rh(PMe3)(C2H4)5 and (C5H5)Ir(PMe3)H2.5b Irradiation
of (C5Me5)Rh(PMe3)(C2H4) in C6F6 leads to initial formation
of (C5Me5)Rh(PMe3)(η2-C6F6), and continued photolysis leads
to (C5Me5)Rh(PMe3)(C6F5)F. Photolysis of (C5H5)Ir(PMe3)H2
in C6F6 generated (C5H5)Ir(PMe3)(C6F5)H and (C5H5)Ir(PMe3)-
(η2-C6F6). Neither of these reactions proceeded thermally.
Thermal C-F bond cleavage of C6F6 by electron rich late
transition metal complexes is less common.6-9 Recently, Perutz
and co-workers carried out C-F activation on various poly-
fluoroaromatics at -78 °C with cis-[Ru(dmpe)2H2] (dmpe )
Me2PCH2CH2PMe2).10 The reaction with C6F6 produced trans-
Despite the activity in metal-based C-F bond cleavage, little
is known about the mechanisms of these reactions. In this paper,
we show that (C5Me5)Rh(PMe3)H2 (1) is capable of thermal,
selective C-F activation with a variety of polyfluoroaromatic
compounds (eq 1). The relative reactivity of 1 with different
fluoroaromatics is examined, and the high selectivity observed
is discussed. In addition, evidence is presented that elucidates
the mechanism of C-F cleavage in these polyfluorinated
aromatic substrates.
1 + arylF-F f Cp*Rh(PMe3)(arylF)H + HF
Results and Discussion
(1)
Thermal Reactions of Cp*Rh(PMe3)H2 with Polyfluori-
nated Arenes. The thermolysis of 1 in pyridine-d5 with 12
equiv of C6F6 at 85 °C results in conversion to air-sensitive
(C5Me5)Rh(PMe3)(C6F5)H (2) in good yield (81%, Scheme 1).
The 1H NMR spectrum exhibits a hydride resonance at δ
-12.664 (ddd, JP-H ) 47.6, JRh-H ) 24.4, JF-H ) 12.5 Hz, 1
H). The 19F NMR spectrum shows five fluorine resonances of
equal area, indicating that there is hindered rotation around the
Rh-aryl bond. Two downfield multiplets at δ 95.6 and 90.5
are assigned as belonging to the two ortho fluorines. The 31P-
{1H} NMR spectrum shows a doublet of triplets (δ 5.50, JRh-P
) 142.0, JF-P ) 10.0 Hz), indicating coupling of phosphorus
to both ortho fluorines.12 The value of JRh-P of 142.0 Hz is
typical for a (C5Me5)RhIII complex.13 Chlorination of the
product with CHCl3 results in the formation of (C5Me5)Rh-
(PMe3)(C6F5)Cl (2-Cl), which was confirmed by comparison
with an authentic sample.5b The 31P{1H} NMR spectrum of
this complex shows a doublet of doublets, indicating coupling
of phosphorus to only one ortho fluorine with JP-F ) 22 Hz
due to the hindered rotation around the aryl-Rh bond.
Compound 2 must similarly only couple to one of the two ortho
fluorines at a time, but rotation around the aryl-Rh bond is
X Abstract published in AdVance ACS Abstracts, August 1, 1997.
(1) (a) Smart, B. E. Mol. Struct. Energ. 1986, 3, 141-191. (b) Smart,
B. E. In The Chemistry of Functional Groups, Supplement D; Patai S.,
Rappoport, Z., Eds.; John Wiley & Sons: New York, 1983; Chapter 14.
(2) For a comprehensive review of C-F activation by metal reagents,
see: Kiplinger, J. L.; Richmond, T. G.; Osterberg, C. E. Chem. ReV. 1994,
94, 373-431.
(3) (a) Aizenberg, M.; Milstein, D. Science 1994, 265, 359-361. (b)
Aizenberg, M.; Milstein, D. J. Am. Chem. Soc. 1995, 117, 8674-8675.
(4) Kiplinger, J. L.; Richmond, T. G. J. Am. Chem. Soc. 1996, 118,
1805-1806.
(5) (a) Jones, W. D.; Partridge, M. G.; Perutz, R. N. J. Chem. Soc., Chem.
Commun. 1991, 264-266. (b) Belt, S. T.; Helliwell, M.; Jones, W. D.;
Partridge, M. G.; Perutz, R. N. J. Am. Chem. Soc. 1993, 115, 1429-1440.
(6) Fahey, D. R.; Mahan, J. E. J. Am. Chem. Soc. 1977, 99, 2501-2508.
(7) Blum, O.; Frolow, F.; Milstein, D. J. Chem. Soc., Chem. Commun.
1991, 258-259.
(8) Hofmann, P.; Unfried, G. Chem. Ber. 1992, 125, 659-661.
(9) Bach, I.; Po¨rschke; Goddard, R.; Kopiske, C.; Kru¨ger, C.; Rufinska,
A.; Seevogel, K. Organometallics 1996, 15, 4959-4966.
(10) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. J. Chem. Soc., Chem.
Commun. 1996, 787-788.
(11) Whittlesey, M. K.; Perutz, R. N.; Greener, B.; Moore, M. H. J.
Chem. Soc., Chem. Commun. 1997, 187-188.
(12) Selmeczy, A. D.; Jones, W. D.; Partridge, M. G.; Perutz, R. N.
Organometallics 1994, 13, 385-396.
(13) Klingert, B.; Werner, H. Chem. Ber. 1983, 116, 1450-1462.
S0002-7863(97)00723-3 CCC: $14.00 © 1997 American Chemical Society

Mechanism of C-F Bond ActiVation by (C5Me5)Rh(PMe3)H2
Scheme 1. Reactions of 1 with Polyfluorinated Arenes
J. Am. Chem. Soc., Vol. 119, No. 33, 1997 7735
Figure 2. ORTEP drawing of (C5Me5)Rh(PMe3)(2-perfluorobiphenyl)-
Cl, 4-Cl. Ellipsoids are shown at the 30% probability level.
Figure 1. ORTEP drawing of (C5Me5)Rh(PMe3)(C6F4H)Cl, 3-Cl.
Ellipsoids are shown at the 30% probability level.
(Scheme 1). This product displayed 1H and 31P spectral features
similar to those of 2 and 3. The 19F NMR spectrum showed
nine distinct fluorine multiplets. Conversion to the chloro
derivative permitted determination of the X-ray structure,
indicating that selective C-F activation had occurred at the para
position to give (C5Me5)Rh(PMe3)(4-perfluorobiphenyl)Cl (4-
Cl) (Figure 2). The aryl-aryl dihedral angle is 62.9°. The
19F NMR spectrum again exhibits nine fluorine resonances,
indicating not only hindered rotation about the Rh-aryl bond
but also the aryl-aryl bond.
Thermolysis of 1 with 0.5 equiv of perfluorobiphenyl, or
thermolysis of 1 with 1 equiv of 4, gives 4,4′-bis[(C5Me5)Rh-
(PMe3)H]perfluorobiphenyl (5). While two diastereomers are
possible, the 19F NMR spectrum shows only six fluorine
resonances rather than the expected eight. The two downfield
ortho resonances integrate to two fluorines each while the four
upfield resonances integrate to one fluorine each, suggesting
that the two ortho fluorines of each diastereomer have coincident
chemical shifts. The 1H NMR spectrum shows only one hydride
resonance at δ -12.470 (ddd, JP-H ) 44.0 Hz, JRh-H ) 24.0
Hz, JF-H ) 16.0 Hz, 1 H) as well as resonances corresponding
rapid enough with the smaller hydride ligand to result in a triplet
in the 31P NMR spectrum with an average JP-F of 10 Hz.
The thermal reaction of 1 with C6F5H in pyridine produces
a single C-F cleavage product in high yield, (C5Me5)Rh(PMe3)-
(C6F4H)H (3). The 1H NMR spectrum shows an aromatic
resonance at δ 6.929 (tt, JH-F ) 9.6, 6.9 Hz). The triplet of
triplet pattern is consistent with a hydrogen coupling to two
ortho and two meta fluorines, indicating that C-F activation
occurred exclusively para to the hydrogen in C6F5H. An X-ray
crystal structure of the chloro derivative (C5Me5)Rh(PMe3)-
(C6F4H)Cl (3-Cl) (Figure 1) confirmed the site of C-F
activation. The 19F NMR of 3 displays four types of fluorine,
indicating hindered rotation around the Rh-aryl bond. As in
2, the 31P NMR spectrum shows a doublet of triplets for 3, but
a doublet of doublets for 3-Cl. In contrast to the thermal
reactions, irradiation of 1 in the presence of C6F5H leads only
to the C-H activated product, (C5Me5)Rh(PMe3)(C6F5)H.
Similarly, the thermolysis of 1 in pyridine-d5 with 12 equiv
of perfluorobiphenyl at 85 °C results in near quantitative
conversion to (C5Me5)Rh(PMe3)(4-perfluorobiphenyl)H (4)

7736 J. Am. Chem. Soc., Vol. 119, No. 33, 1997
Edelbach and Jones
Figure 4. Plot of the disappearance of 1 upon reaction of 1 (0.035 M)
with perfluorobiphenyl ([, 0.66 M; 9, 0.52 M; 2, 0.36 M; b, 0.23
M) in a 1:1 pyridine:benzene solvent mixture at 92 °C.
Figure 3. Ball and stick drawing showing both rotomers of (C5Me5)-
Rh(PMe3)(C10F7)Cl, 6-Cl.
where no added base was present. The same organometallic
products, 2-6, were obtained although in lower yields (50-
70%). Also observed in these reactions were larger quantities
of [(C5Me5)Rh(PMe3)2H]+ (5-15%) and a trace of arylF-H.
A small amount of black solid was also formed in these
reactions. The higher extent of decomposition can be attributed
to the lack of a base to complex the HF produced. Reactions
produced slightly higher yields (by about 5%) when run in a
1:1 mixture of benzene/pyridine than in pure pyridine.
to a single C5Me5 ring and a single PMe3 group, disguising the
fact that two diastereomers are present. The 31P{1H} NMR
spectrum of 5 shows the characteristic doublet of triplet pattern,
indicating coupling of both ortho fluorines to phosphorus.
Apparently, the environment around the two halves of the
1
diastereomers is so similar that the H and 31P resonances are
coincident. The meta 19F resonances, however, experience
significantly different environments in the two diastereomers
and consequently are distinct. Treatment of 5 with CHCl3
Rates of Reaction of 1 with Perfluoroarenes. Several
qualitative observations were initially made concerning the
reaction of 1 with the perfluoroarenes. First, different arenes
react at different rates, spanning a range of a factor of about
25. Second, the rate is dependent on the concentration of
perfluoroarene. Third, the reactions are much faster in pyridine
than in benzene. And fourth, a plot of [1] vs time does not
appear to be consistent with either a first- or second-order
reaction. Typical plots are shown in Figure 4 for the reaction
of 1 with different concentrations of perfluorobiphenyl. The
reaction can be seen to increase in rate and then slow down
only as 1 is nearly depleted. This type of kinetic behavior, while
not rare, is not common, and is typical of autocatalytic or
product-catalyzed reactions.15 In this type of reaction, a product
of the reaction is capable of reacting with the starting material,
thereby increasing the observed rate as the reaction continues.
The rate is also seen to increase with substrate concentration.
produced chloro derivatives of the two diastereomers. The 19
F
NMR spectrum shows eight fluorine resonances; in this case
the ortho fluorines for each diastereomer were resolved.
Integration indicates that one diasteriomer dominates with a ratio
of 3:2. The chloro derivative of 5 resisted attempts at forming
single crystals, perhaps as it was a mixture of diastereomers.
The selective nature of C-F activation was probed further
by reacting 1 with perfluoronaphthalene. 1H NMR and 19F
NMR spectroscopy reveals the presence of two products in a
ratio of approximately 1.25:1. The 1H NMR spectrum consists
of one C5Me5 group, as well as two overlapping hydrides each
of which have a doublet of doublet of doublet splitting pattern.
Two distinct trimethylphosphine groups could be assigned. The
31P NMR spectrum shows two doublet of triplet splitting
patterns, indicating that C-F activation occurred in the â
position. The 19F NMR spectrum shows 14 peaks, and
integration allows for the assignment of seven peaks for each
product. The NMR data are consistent with the formation of
two rotomers of (C5Me5)Rh(PMe3)(â-perfluoronaphthyl)H (6).
X-ray crystallography of the chloro derivative 6-Cl shows a
disorder as a result of the presence of the two rotomers. The
solid state structure confirms that C-F bond activation occurs
only at the â position (Figure 3).
An obvious product that could be responsible for autocatalysis
in the present system is the HF produced (eq 1). As HF is a
weak acid in water, it could serve as a proton source in the
reactions under study. Protonation of 1 could lead to loss of
H2 to generate an electrophilic metal cation that might be capable
of cleaving the C-F bond of the polyfluoroaromatic. To test
this possibility, a solution of 1 and perfluorobiphenyl in 1:1
pyridine/benzene was reacted with pyridinium chloride. Reac-
tion occurred rapidly at 92 °C to produce Cp*Rh(PMe3)Cl2 and
H2. No C-F activation product was observed. Use of
pyridinium tetrafluoroborate as the acid was found to give C-F
activation product 4 in ∼50% yield upon heating, but the rate
of formation of product is slower than in the absence of added
acid under comparable conditions. In addition, a substantial
quantity of the decomposition product [Cp*Rh(PMe3)2H]+ is
observed. On the basis of these observations, it seems unlikely
that protons catalyze the reaction.
In all of the above reactions, a very broad resonance is
observed at ∼δ 14.3 ppm, assigned as a mixture of pyridine‚
- 14
HF and HF2
.
Also observed are trace amounts of [(C5Me5)-
Rh(PMe3)2H]+.13 Presumably, the HF produced in the C-F
activation reaction reacts with a small portion of the reactant
and/or product to generate the byproducts and some fluoride
ion, which is trapped as bifluoride. The reactions of 1 with the
above fluoroaromatics were also examined in C6D6 solvent,
(14) An authentic sample of pyridine‚HF (Aldrich) in pyridine-d5 displays
a broad singlet at ∼δ 11.4. Tetrabutylammonium bifluoride, Bu4N+HF2
,
(15) (a) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms,
2nd ed.; McGraw-Hill, Inc.: New York, 1995. (b) Moore, J. W.; Pearson,
R. G. Kinetics and Mechanisms, 3rd ed.; John Wiley & Sons: New York,
1981.
-
in pyridine-d5 displays a sharp triplet (JHF ) 120.5 Hz) at δ 18.44. A mixture
of pyridine‚HF and fluoride ion shows a broad resonance at ∼δ 14, which
moves with the relative amounts of the two components.

Mechanism of C-F Bond ActiVation by (C5Me5)Rh(PMe3)H2
J. Am. Chem. Soc., Vol. 119, No. 33, 1997 7737
Figure 5. Plot of the disappearance of 1 (0.035 M) upon reaction with
perfluorobiphenyl (0.36 M) in a 1:1 pyridine:benzene ratio at 69 °C
with 0.21 M fluoride (9), with 0.0077 M fluoride (2), and without
fluoride (b).
Figure 6. Plot of the disappearance of 1 (0.035 M) upon reaction with
perfluorobiphenyl (0.36 M) at 69 °C in 10% pyridine/benzene (b), in
50% pyridine/benzene (2), and neat pyridine (9).
Scheme 2
Another possibility is that the fluoride ion produced catalyzes
the reaction. Addition of n-Bu4NF to a mixture of 1 and
perfluorobiphenyl in 1:1 pyridine/benzene resulted in rapid
reaction at 69 °C to give 4 in high yield. In the absence of
added fluoride ion at this temperature the reaction is much
slower (Figure 5). Not only does fluoride ion greatly accelerate
this reaction, but the behavior of the concentration vs time plots
now appears approximately exponential. With fluoride present
the reaction now proceeds at moderate rates even at room
temperature. The half-life for reaction in pyridine solvent with
0.21 M F- is ∼14 h, whereas without fluoride the half-life is
96 h. In benzene at room temperature (without added fluoride),
only traces of product are seen over this time period.
Reaction of 1 with perfluorobiphenyl in 1:1 pyridine/benzene
containing ∼1 equiv of pyridine‚HF (but no additional fluoride)
did not result in an increase in the rate of reaction. The yield
of 4 is only about 50%, and about 25% yield of [Cp*Rh-
(PMe3)2H]+ is observed.
Reaction of the Cp compound CpRh(PMe3)H2 with perfluo-
robiphenyl in pyridine was also examined. The reaction
proceeds cleanly at room temperature to give CpRh(PMe3)-
(perfluorobiphenyl)H in good yield. This reaction is about twice
as fast as that of 1 under the same conditions (pyridine solvent,
no added fluoride).
For comparison, the reaction of 1 with 1 equiv of C6F5Cl or
C6F5Br was examined. In both cases, a rapid reaction ensues
at room temperature and (C5Me5)Rh(PMe3)(H)(X) (X ) Cl or
Br) and C6F5H are obtained as the major products. Only trace
amounts of (C5Me5)Rh(PMe3)(C6F5)H are observed. The
formation of these types of products is typical of other metal
hydrides reacting with alkyl and aryl halides (halide * F), which
are believed to react by a free radical chain mechanism.16
Mechanism of C-F Cleavage. Fluoride ion has a clear
influence on the rate of reaction of 1 with polyfluoroaromatics.
Pyridine as solvent is also found to accelerate the rate of these
reactions compared to benzene solvent. Since both pyridine
and fluoride ion can act as bases, it seems reasonable to postulate
that deprotonation of 1 might be responsible for the reaction
with the fluoroaromatic.17 The resulting metal anion could
undergo nucleophilic aromatic substitution to give the products
2-6 and (re)generate a fluoride ion. With pyridine as the initial
base, the fluoride ion combines with the pyridinium ion to give
pyridine‚HF, which is largely undissociated. With fluoride as
the initial base, the fluoride ion reacts with HF to give bifluoride
ion (Scheme 2). Metal carbonyl anions are well-known to
undergo nucleophilic aromatic substitutions on polyfluoroaro-
matics to give metal-aryl complexes, and substitution occurs
with the regiochemistry observed in the above reactions.2,18,19
Since the deprotonated anion of 1 is postulated to be the
nucleophile that displaces fluoride ion, independent examina-
tions of the reaction of the anion [Cp*Rh(PMe3)H]-, (7), with
perfluorobiphenyl and perfluoronaphthylene were examined.
Anion 7 was prepared by deprotonation of 1 with n-BuLi at
-78 °C.20 Upon treatment of a pyridine-d5 solution of
perfluorobiphenyl with 7, an instantaneous reaction occurs at
room temperature to produce 4 (and also some 5). Similarly,
reaction of 7 with a pyridine solution of perfluoronaphthalene
immediately produces 6 (same 1:1.25 ratio of both rotomers).
The effect of pyridine concentration on the rate of reaction
also supports a deprotonation mechanism. Figure 6 shows how
the rate of reaction increases upon changing the solvent from
10% pyridine in benzene to 50% pyridine in benzene to neat
pyridine. The rate increase is dramatic, and is attributable to
the shift in the deprotonation equilibrium (first step in Scheme
2). This shift is due in part to the increase in pyridine
(18) (a) Smart, B. E. In Chemistry of Organic Fluorine Compounds II;
Hudlicky, M., Rappoport, Z., Eds.; ACS Monograph 187; American
Chemical Society: Washington, DC, 1995; pp 996-997. (b) Rodionov, P.
P.; Furin, G. G. J. Fluorine Chem. 1990, 47, 361-434. (c) Rodionov, P.
P.; Osina, O. I.; Platonov, V. E.; Yakobson, G. G. Bull. Soc. Chim. Fr.
1986, 6, 986-992.
(16) (a) Walling, C.; Cooley, J. H.; Ponaras, A. A.; Racah, E. J. J. Am.
Chem. Soc. 1966, 88, 5361-5363. (b) Kuivila, H. G. Acc. Chem. Res. 1968,
1, 299-305. (c) Green, M. L. H.; Knowles, P. J. J. Chem. Soc., Perkin
Trans. 1 1973, 989-991. (d) Kinney, R. J.; Jones, W. D.; Bergman, R. G.
J. Am. Chem. Soc. 1978, 100, 635-637.
(17) Fluoride is known to be an excellent base in nonprotic solvents,
some 12 orders of magnitude better in DMSO than in water. See: Bordwell,
F. G. Acc. Chem. Res. 1988, 21, 456-463.
(19) An alternative mechanism for the fluoride ion displacement from
the fluorocarbon by the metal anion would involve electron transfer from
[Cp*Rh(PMe3)H-] to the arene generating fluoride ion and an arylF radical.
This radical could then recombine with the metal radical to generate the
observed metal product. Any radicals formed, however, would have to
recombine within the radical cage (Vide infra).
(20) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1986, 108, 7332-
7346.

7738 J. Am. Chem. Soc., Vol. 119, No. 33, 1997
Edelbach and Jones
concentration and in part to the increase in solvent polarity, the
latter effect tending to favor the formation of ions.
Scheme 3
The enhanced reactivity of the C5H5 compound compared to
1 can also be accommodated by an initial deprotonation of the
dihydride. One expects the less electron-donating Cp ligand
to make CpRh(PMe3)H2 more acidic, and consequently a higher
steady state concentration of anion should be present to react
with the fluorocarbon. In addition, the Cp compound will
produce a less sterically hindered nucleophile than 1.
Free radical intermediates can be virtually ruled out in these
reactions. When the thermal reaction of 1 with C6F6 was carried
out in the presence of the radical trap 9,10-dihydroanthracene
(10 equiv) in benzene, the rate of reaction and yield of 2 were
not affected. 1H NMR spectroscopy and GC-MS showed only
a trace of anthracene formation. A similar reaction between 1
and C6F6 was performed in neat isopropylbenzene as trap and
the reaction monitored by 31P NMR spectroscopy. Again,
neither the rate of reaction nor yield of 2 was affected. 1H NMR
and GC-MS showed only a trace amount of C6F5H in the
isopropylbenzene reaction. These results strongly indicate free
radicals are not involved.
occurs readily,24 and the fluoride ion then removes a proton
from the acidic25 metal cation. Finally, recombination of the
resulting radical pair generates the product.
This type of mechanism is inconsistent with the observations
made in the reactions of 1 with polyfluoroaromatics. First, free
radical intermediates are not involved. Second, the acceleration
by fluoride ion is unaccounted for. Third, electrochemical
measurements on the oxidation potential of 1 and the reduction
potential of the polyfluoroaromatics indicate that the electron
transfer is uphill by in excess of 2 V. We have redetermined
the polyfluoroarene reduction potentials measured by Pez,26 and
have obtained similar values for the irreversible reductions (-2.4
to -3.0 V vs Fc/Fc+). While the true reduction potentials might
be more positive than this, it is very unlikely that they are off
by more than a few hundreds of millivolt. The oxidation of 1
is also irreversible by CV (EPA ) -0.336 V), but the true
reversible potential is not likely to be far from this value either.
The above redox values result in an uphill electron transfer on
the order of 2.0 to 2.5 V, which would place a minimum barrier
to the reaction of ∼45 kcal/mol. This barrier corresponds to a
rate that is over a billion billion times slower than the observed
rate of reaction of 1 with the slowest substrate, C6F5H.
Consequently, electron transfer from 1 to the perfluoroarene is
extremely unlikely to be involved.
The elucidation of an anionic nucleophilic mechanism in the
reaction of 1 with polyfluoroaromatics suggests that a similar
mechanism might be considered as an alternative to the electron
transfer mechanisms in other similar systems in the literature.
This possibility seems even more reasonable in light of the high
mismatch of redox potentials between the electron transfer
partners (metal complex vs polyfluoroaromatic). In support of
this concept, the system reported by Milstein3b is considerably
more efficient when a base is present (NEt3, K2CO3). Regio-
selectivities are identical with that observed in the present
studies.27 Systems that have postulated electron transfer from
the metal to the polyfluoroaromatic include Rh(PMe3)4H,3b Ru-
(dmpe)2H2,10 MeIr(PEt3)3 (which cyclometallates before electron
transfer),7 and PtH2(PCy3)2.23 We note that Hughes has recently
Earlier C-H and C-C bond cleavage studies involving 1
occur by way of formation of the unsaturated intermediate
[Cp*Rh(PMe3)]. This intermediate can be ruled out in the
present C-F bond activations, since reaction with C6F5H results
in C-F activation rather than C-H activation. Indeed, pho-
toinduced elimination of dihydrogen from 1 or thermolysis of
Cp*Rh(PMe3)PhH in the presence of C6F5H produces exclu-
sively Cp*Rh(PMe3)(C6F5)H.21
As mentioned earlier, these reactions appear to follow
autocatalytic or product-catalyzed behavior. That is, the rate
of reaction is observed to increase as the reaction proceeds. The
stoichiometric product of the reaction is not fluoride ion, but
HF as indicated in eq 1, so that this increase in rate is not
expected. As the reaction proceeds, however, some 1 undergoes
decomposition by reaction with HF, thereby generating surplus
F-. This additional fluoride can act to increase the rate of
reaction by increasing the quantity of strong base, making the
reaction appear autocatalytic. In pure benzene solvent, this
decomposition is more significant. Also, the reaction could be
initiated by adventitious base in the solvent or in the glass wall
of the container.22 Again, once the reaction starts, fluoride ion
would be produced which carries the cycle. With added
fluoride, decomposition produces little additional fluoride so
that the reaction now appears approximately first order in 1. A
detailed kinetic analysis of this system has not proven possible
due to the variable small quantity of decomposition that produces
fluoride ion.
Comparison with Other C-F Cleavage Reactions. Re-
cently proposed mechanisms for polyfluoroaromatic C-F bond
10
activation by Rh(PMe3)4H,3b PtH2(PCy3)2,23 and Ru(dmpe)2H2
all involve electron transfer from the metal to the fluorocarbon
in the first step. This mechanism involves initial outer-sphere
electron transfer to form a 17-electron metal cation radical and
a fluorinated aromatic anion radical (Scheme 3). Fluoride loss
(24) (a) Brown, J. K.; Williams, W. G. Trans. Faraday Soc. 1968, 64,
298-307. (b) Symons, M. C. R.; Selby, R. C.; Smith, I. G.; Bratt, S. W.
Chem. Phys. Lett. 1977, 48, 100-102.
(21) This observation appears to also rule out any other oxidative addition
pathway involving a 16-electron intermediate, which would be expected to
similarly prefer C-H activation. Other possible 16-electron intermediates
could arise from η5 f η3 ring slippage or hydride ligand migration to the
C5Me5 ring to give an η4-C5Me5H ligand. See: Jones, W. D.; Kuykendall,
V. L.; Selmeczy, A. D. Organometallics 1991, 10, 1577-1586.
(22) Hughes, R. P.; Lindler, D. C.; Rheingold, A. L.; Yap, G. P. A.
Organometallics 1996, 15, 5678-5686.
(25) Ryan, O. B.; Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990,
112, 2618-2626.
(26) Marsella, J. A.; Gilicinski, A. G.; Coughlin, A. M.; Pez, G. P. J.
Org. Chem. 1992, 57, 2856-2860.
(27) A referee commented that it would be difficult for fluoride ion to
deprotonate a species such as HRh(PMe3)4 or Ru(dmpe)H2. Fluoride ion is
an excellent base, however, and the acidities of these transition metal
hydrides have not been determined.17
(23) Hintermann, S.; Pregosin, P. S.; Ruegger, H.; Clark, H. C. J.
Organomet. Chem. 1992, 435, 225-234.

Mechanism of C-F Bond ActiVation by (C5Me5)Rh(PMe3)H2
J. Am. Chem. Soc., Vol. 119, No. 33, 1997 7739
pyridinium tetraphenylborate,31 and (C5Me5)Rh(PMe3)(C6F5)Cl5b have
been previously reported. Isopropyl benzene, chloropentafluoroben-
zene, bromopentafluorobenzene, 9,10-dihydroanthracene, pyridinium
poly(hydrogen fluoride) (∼30% pyridine-70% hydrogen fluoride),
tetrabutylammonium fluoride hydrate (TBAF‚3H2O), tetra-n-butylam-
monium hexafluorophosphate, and the fluorinated aromatic compounds
were purchased from Aldrich Chemical Co. The liquids were stirred
over sieves, freeze-pump-thawed three times, and vacuum distilled
prior to use. Tetra-n-butylammonium hexafluorophosphate was dried
at 120 °C under high vacuum for 48 h.
Scheme 4
All 1H NMR and 31P NMR spectra were recorded on a Bruker
AMX400 spectrometer. All 19F NMR spectra were recorded on a
1
Varian 500 spectrometer. All H chemical shifts are reported in ppm
(δ) relative to tetramethylsilane and referenced by using chemical shifts
of residual solvent resonances (C6H6 δ 7.15, pyridine δ 7.55). 19F NMR
spectra were referenced to external C6H5CF3 (δ 0.00 with downfield
chemical shifts taken to be positive). 31P NMR spectra were referenced
to external 30% H3PO4 (δ 0.0). GC-MS was conducted on a 5890
Series II gas chromatograph fitted with an HP 5970 Series mass
selective detector. Photolysis was done with an Oriel Corp. UV Lamp
Model 6137. Analyses were obtained from Desert Analytics. An
Enraf-Nonius CAD4 diffractometer and a Siemens SMART system with
a CCD area detector were used for X-ray structure determination.
Thermolysis of (C5Me5)Rh(PMe3)H2 (1) with Various Fluorinated
Aromatic Compounds. All reactions were carried out by using the
following general procedure. A sample of (C5Me5)Rh(PMe3)H2 (8.5
mg, 0.027 mmol) was dissolved in 0.60 mL of pyridine-d5 or benzene-
d6 in a resealable NMR tube along with 12 equiv of the fluoroaromatic
compound. Dioxane (or dimethoxyethane) was added as an internal
standard. The NMR tube was heated in an oil bath at 85 °C, and the
reaction was followed by 1H NMR spectroscopy until the starting
material was depleted. NMR spectroscopic data, yields (NMR yields
based on an internal standard in pyridine-d5), and approximate half-
lives (t1/2) are summarized below. NMR spectroscopic data in C6D6
are given in the Supporting Information.
suggested an intramolecular nucleophilic attack on a perfluo-
roaryl ligand by a deprotonated Cp* methyl group.22
Catalytic Reactions. Finally, the reaction of (C5Me5)Rh-
(PMe3)(C6F5)H with excess C6F6 was carried out under hydro-
gen to see if the reaction could be made catalytic by cleavage
of the Rh-ArylF bond with H2. A potential catalytic scheme
is outlined above (Scheme 4). A solution of 1 and C6F6 in C6D6
was heated at 135 °C for 25 days under 1 atm of H2. 1H NMR
spectroscopy of the reaction mixture revealed the presence of
1.4 equiv of C6F5H based on the starting metal complex. No
(C5Me5)Rh(PMe3)(C6F5)H or (C5Me5)Rh(PMe3)H2 was seen.
The same reaction performed without H2 resulted in the
formation of 1 equiv of C6F5H. Reactions performed under
higher pressures of H2 or in 1:1 pyridine/benzene also showed
little evidence for catalysis.
NMR Spectroscopic Data and Yields for (C5Me5)Rh(PMe3)-
(arylF)H Complexes. (C5Me5)Rh(PMe3)(C6F5)H. 1H NMR (pyridine-
d5): δ 1.777 (d, JRh-H ) 2.0 Hz, 15 H), 1.101 (d, JP-H ) 10.0 Hz, 9
H), -12.664 (ddd, JP-H ) 47.6 Hz, JRh-H ) 24.4, JF-H ) 12.5 Hz, 1
H). 31P{1H} NMR (pyridine-d5): δ 5.50 (dt, JRh-P ) 142.0 Hz, JF-P
) 10.0 Hz). 19F NMR (C6D6): δ 95.6 (m, 1 Fortho), 90.5 (m, 1 Fortho),
38.6 (t, 1 Fpara), 37.7 (m, 1 Fmeta), 36.6 (m, 1 Fmeta). Yield ) 81%. t1/2
(85 °C) = 12 h.
Conclusions
The organometallic complex (C5Me5)Rh(PMe3)H2 is capable
of thermally cleaving C-F bonds of a variety of fluoroaromatics.
The cleavage is selective and the rate of C-F activation
increases as the number of fluorines increases. The C5H5 analog
was shown to be able to activate the C-F bond of C12F10 more
rapidly than the C5Me5 species. A mechanism involving
nucleophilic aromatic substitution was indicated, with pyridine
or fluoride as the base that produces a metal anion that can
attack the polyfluoroaromatic compound. This system has the
necessary balance between Bro¨nsted acidity of the metal hydride
and nucleophilicity of the conjugate base for the reaction to
succeed. Our studies suggest that catalytic C-F bond cleavage
by a transition metal hydride might be possible in basic solvents
containing fluoride ion as a promoter.
(C5Me5)Rh(PMe3)(4-perfluorobiphenyl)H. 1H NMR (pyridine-
d5): δ 1.811 (d, JRh-H ) 2.4 Hz, 15 H), 1.143 (d, JP-H ) 10.1 Hz, 9
H), -12.483 (ddd, JP-H ) 48.9 Hz, JRh-H ) 24.1 Hz, JF-H ) 10.8 Hz,
1 H). 31P{1H} NMR (pyridine-d5): δ 5.39 (dt, JRh-P ) 141.7 Hz, JF-P
) 9.7 Hz). 19F NMR (C6D6): δ 95.8 (m, 1 Fortho), 90.1 (m, 1 Fortho),
61.5 (m, 1 Fmeta), 61.1 (m, 1 Fmeta), 59.5 (m, 1Fortho′), 57.9 (m, 1 Fortho′),
47.5 (m, 1 Fpara), 38.4-38.2 (overlapping m, 2 Fmeta′). Yield ) 87%.
t1/2 < 30 min.
(C5Me5)Rh(PMe3)(2,3,5,6-C6HF4)H. 1H NMR (pyridine-d5):
δ
6.929 (tt, JH-F ) 9.6, 6.9 Hz), δ 1.788 (d, JRh-H ) 2.0 Hz, 15 H),
1.103 (dd, JP-H ) 10.0 Hz, JRh-H ) 1.2 Hz, 9 H), -12.634 (ddd, JP-H
) 49.0 Hz, JRh-H ) 25.4 Hz, JF-H ) 12.9 Hz, 1 H). 31P{1H} NMR
(pyridine-d5): δ 5.03 (dt, JRh-P ) 144.7 Hz, JF-P ) 10.2 Hz). 19F
NMR (C6D6): δ 93.4 (m, 1 Fortho), 88.2 (m, 1 Fortho), 59.9 (t, 1 Fmeta),
58.4(m, 1 Fmeta). Yield ) 85%. t1/2 = 20 h.
Experimental Section
General Considerations. All manipulations were performed under
an N2 atmosphere, either on a high-vacuum line with modified Schlenk
techniques or in a Vacuum Atmospheres Corporation Glovebox.
Tetrahydrofuran, benzene, and toluene were distilled from dark purple
solutions of benzophenone ketyl. Alkane solvents were made olefin-
free by stirring over H2SO4, washing with aqueous KMnO4 and water,
and distilling from dark purple solutions of tetraglyme/benzophenone
ketyl. Benzene-d6, and pyridine-d5 were purchased from Cambridge
Isotope Lab., and distilled under vacuum from dark purple solutions
of benzophenone ketyl, and stored in ampules with Teflon sealed
vacuum line adapters. The preparations of [PPN]+[BH4],28 (C5-
Me5)Rh(PMe3)H2,29 Li[(C5Me5)Rh(PMe3)H],20 (C5H5)Rh(PMe3)H2,30
(C5Me5)Rh(PMe3)(2-perfluoronaphthalene)H. 19F NMR indicated
that rotomer 1 dominated by a ratio of 1.25:1. Rotomer 1: 1H NMR
(pyridine-d5): δ 1.829 (overlaps with Rotomer 2, 15 H), 1.180 (d, JP-H
) 9.7 Hz, 9 H), -12.627 (ddd, JP-H ) 50.2 Hz, JRh-H ) 26.1 Hz, JF-H
) 12.6 Hz, 1 H). 31P{1H} NMR (pyridine-d5): δ 5.51 (dt, JRh-P
)
141.8 Hz, JF-P ) 10.9 Hz). 19F NMR (C6D6): 110.7 (m, 1 F), 103.6
(29) Isobe, K.; Bailey, P. M.; Maitlis, P. M. J. Chem. Soc., Dalt. Trans.
1981, 2003.
(30) Partridge, M. G. D.Phil. Thesis, University of York, U.K., 1992.
(31) Pyridinium tetraphenylborate was prepared following the general
methodology developed by Wittig and co-workers: (a) Wittig, G.; Raff, P.
Justus Liebigs Ann. Chem. 1951, 573, 195-209. (b) Wittig, G.; Keicher,
G.; Ruckert, A.; Raff, P. Justus Liebigs Ann. Chem. 1949, 563, 110-126.
(c) Wittig, G.; Keicher, G.; Ruckert, A.; Raff, P. Chem. Ber. 1955, 88,
962-976.
(28) Kirtley, S. W.; Andrews, M. A.; Bau, R.; Grynkewich, G. W.; Marks,
T. J.; Tipton, D. L.; Whittlesey, B. R. J. Am. Chem. Soc. 1977, 99, 7154-
7162.

7740 J. Am. Chem. Soc., Vol. 119, No. 33, 1997
Edelbach and Jones
Table 1. Summary of Crystallographic Data for 3-Cl, 4-Cl, and 6-Cl
crystal parameters
chemical formula
formula wt
cryst syst
space group
Z
3-Cl
4-Cl
6-Cl
C19H25ClF4PRh
498.73
monoclinic
C2/c (No. 15)
8
C25H24ClF9PRh
664.78
monoclinic
P21/c (No. 14)
4
C23H24ClF7PRh‚H2O
620.77
monoclinic
P21/c (No. 14)
4
a, Å
b, Å
c, Å
14.153(5)
9.585(8)
30.45(2)
96.18(4)
4107(4)
1.61
4011
3838
235
9.139(7)
8.082(2)
35.903(27)
93.95(6)
2645.6(2.4)
1.67
5374
5038
334
9.7329(10)
8.3407(9)
31.741(3)
98.026(7)
2551.4(5)
1.62
9459
3330
468
â, deg
vol, Å3
F
calc, g cm-3
no. of data collected
no. of unique data
no. of params varied
2
R1(Fo), wR2(Fo ), (I > 2σ(I))
R1(Fo), wR2(Fo ), all data
0.0725, 0.1901
0.0813, 0.1964
2
R1(Fo), wR2(Fo), (I > 3σ(I))
goodness of fit
0.0367, 0.0380
1.69
0.0491, 0.0546
1.53
1.23
(m, 1 F), 53.2 (dt, 1 F), 51.9 (dt, 1 F), 48.2(m, 1 F), 39.9 (m, 1 F),
39.2 (m, overlaps with Rotomer 2, 1F). Rotomer 2: 1H NMR (pyridine-
d5): δ 1.829 (overlaps with Rotomer 1, 15 H), 1.154 (d, JP-H ) 9.8
4,4′-[(C5Me5)Rh(PMe3)Cl]2perfluorobiphenyl. 1H NMR (CDCl3):
δ 1.657 (d, JRh-H ) 3.6 Hz, 15 H), δ 1.527 (d, JP-H ) 10.8 Hz, 9 H).
31P{1H} NMR (CDCl3): δ 5.26 (dd, JRh-P ) 143.1 Hz, JF-P ) 9.2
Hz). 19F NMR indicated that diastereomer 1 dominated by a ratio of
3:2.19F NMR (CDCl3): diasteriomer 1, δ 90.2 (m, 3 Fortho), 86.2 (m, 3
Hz, 9 H), -12.451 (ddd, JP-H ) 48.1 Hz, JRh-H ) 26.1 Hz, JF-H
)
11.0 Hz, 1 H). 31P{1H} NMR (pyridine-d5): δ 4.98 (dt, JRh-P ) 142.2
Hz, JF-P ) 10.7 Hz). 19F NMR (C6D6): δ 116.8 (m, 1 F), 98.2 (m, 1
F), 54.7 (dt, 1 F), 51.2 (dt, 1 F), 46.9 (m, 1 F), 40.1(m, 1 F), 39.2 (m,
overlaps with Rotomer 1, 1 F). Yield (both rotomers combined) )
90%. t1/2 < 30 min.
Fortho), 60.4 (m, 3 Fmeta), 59.0 (m, 3 Fmeta); diasteriomer 2, δ 89.7 (m,
2 Fortho), 86.8 (m, 2 Fortho), 61.5 (m, 2 Fmeta), 58.2 (m, 2 Fmeta). Anal.
Calcd for C38H48Cl2F8P2Rh2: C, 45.85; H, 4.86. Found: C, 45.76; H,
4.79.
4,4′-[(C5Me5)Rh(PMe3)H]2perfluorobiphenyl. 1H NMR (C6D6):
δ 1.684 (s, 15 H), 0.786 (d, JP-H ) 10.0 Hz, 9 H), -12.470 (ddd, JP-H
) 44.0 Hz, JRh-H ) 24.0, JF-H ) 16.0 Hz, 1 H). 31P{1H} NMR
(C6D6): δ 5.01 (dt, JRh-P ) 143.1, JF-P ) 9.2 Hz). 19F NMR (C6D6):
δ 94.0 (m, 2 Fortho), 89.0 (m, 2 Fortho), 59.6 (dd, 1 Fmeta), 59.3 (dd, 1
X-ray Structural Determination of 3-Cl. Orange-red crystals were
obtained by layering a CH2Cl2 solution of 3-Cl with hexanes. A single
crystal having approximate dimensions of 0.30 × 0.30 × 0.45 mm3
was mounted on a glass fiber with epoxy. Lattice constants were
obtained from 25 centered reflections with values of ø between 5° and
70° on an Enraf Nonius CAD4 diffractometer. Cell reduction revealed
a primitive monoclinic crystal system. Data were collected at -20 °C
in accord with the parameters found in Table 1. The intensities of
three representative reflections which were measured after every 60
min of X-ray exposure time remained constant throughout the data
collection indicating crystal and electronic stability. The Molecular
Structure Corporation TEXSAN analysis software package was used
for data reduction, solution, and refinement. The space group was
assigned as C2/c on the basis of systematic absences and a Z value of
8. A Patterson map solution of the structure was used to locate the
rhodium atom. The structure was expanded with the DIRDIF program
to reveal all non-hydrogen atoms. An absorption correction was applied
by using the program DIFABS following isotropic refinement. Aniso-
tropic refinement of all non-hydrogen atoms allowed for the use of a
difference Fourier map for the location of the hydrogen atoms whose
coordinates were subsequently idealized. Full-matrix least-squares
anisotropic refinement of the non-hydrogen atoms (with hydrogen atoms
attached to carbon atoms in idealized positions) was executed until
convergence was achieved. The structure was refined with R1 ) 0.0367
and Rw ) 0.0380.32 Fractional coordinates and thermal parameters are
given in the Supporting Information.
F
meta), 58.2 (dd, 1 Fmeta), 58.0 (dd, 1 Fmeta).
Chlorination of (C5Me5)Rh(PMe3)(arylF)H Compounds. The
general procedure employed is described here for (C5Me5)Rh(PMe3)(4-
perfluorobiphenyl)H. To an ampule containing a hexane solution of
(C5Me5)Rh(PMe3)(4-perfluorobiphenyl)H (50 mg, 0.079 mmol) was
added an excess of CHCl3 (12.7 µL, 0.158 mmol) at room temperature.
The brown solution turned red/orange immediately along with the
formation of a red precipitate. (C5Me5)Rh(PMe3)(C12F10)Cl was isolated
on a thin-layer silica chromatography plate with a 95:5 (v/v) solution
of CH2Cl2-THF. Elemental analysis and NMR data are summarized
below in C6D6 solvent.
NMR Spectroscopic Data and Elemental Analysis for (C5Me5)-
Rh(PMe3)(arylF)Cl Complexes. (C5Me5)Rh(PMe3)(4-perfluorobi-
phenyl)Cl. 1H NMR (C6D6): δ 1.264 (d, JRh-H ) 3.2 Hz, 15 H), 1.128
(d, JP-H ) 10.8 Hz, 9 H). 31P{1H} NMR (C6D6): δ 5.27 (dd, JRh-P
)
140.3 Hz, JF-P ) 22.2 Hz). 19F NMR (C6D6): δ 93.8 (m, 1 Fortho),
88.0 (m, 1 Fortho), 62.0 (m, 1 Fmeta), 61.2 (m, 1 Fmeta), 60.5 (m, 1 Fortho′),
58.0 (m, 1 Fortho′), 48.0 (m, 1 Fpara), 39.0 (dt, 1 Fmeta′), 38.7 (dt, 1 Fmeta′).
Anal. Calcd for C25H24ClF9PRh: C, 45.17; H, 3.64. Found: C, 45.25;
H, 3.52.
(C5Me5)Rh(PMe3)(2,3,5,6-C6HF4)Cl. 1H NMR (C6D6): δ 6.541
(tt, JH-F ) 8.8, 6.8 Hz), δ 1.269 (d, JRh-H ) 3.2 Hz, 15 H), 1.131 (d,
X-ray Structural Determination of (C5Me5)Rh(PMe3)(perfluoro-
biphenyl)Cl (4-Cl). Orange-red crystals were obtained by layering a
CH2Cl2 solution of 4-Cl with hexanes. A single crystal of dimensions
0.60 × 0.37 × 0.11 mm3 was mounted on a glass fiber with epoxy.
Lattice constants were obtained from 25 centered reflections with values
of ø between 5° and 70° on an Enraf Nonius CAD4 diffractometer.
Cell reduction revealed a primitive monoclinic crystal system. Data
were collected at -40 °C in accord with the parameters found in Table
1. The space group was uniquely assigned as P21/c on the basis of
systematic absences. The structure was solved and refined by using
JP-H ) 10.8 Hz, 9 H). 1P{1H} NMR (C6D6): δ 5.41 (dd, JRh-P
)
141.4 Hz, JF-P ) 22.8 Hz). 19F NMR (C6D6): δ 91.6 (m, 1 Fortho),
86.3 (m, 1 Fortho), 61.0 (m, 1 Fmeta), 58.5 (m, 1 Fmeta). Anal. Calcd for
C19H25ClF4PRh: C, 45.76; H, 5.05. Found: C, 45.47; H, 4.98.
(C5Me5)Rh(PMe3)(2-perfluoronaphthalene)Cl. Rotomer 1: 1H
NMR (C6D6): δ 1.297 (d, JRh-H ) 3.2 Hz, 15 H), 1.187(d, JP-H
)
10.8 Hz, 9 H). 31P{1H} NMR (C6D6): δ 5.87 (dt, JRh-P ) 141.0 Hz,
JF-P ) 26.7 Hz). 19F NMR (C6D6): δ 108.2 (m, 1 F), 100.2 (m, 1 F),
53.0-52.5 (m, 2 F), 49.2 (m, 1 F), 40.2 (m, overlaps with Rotomer 2,
2F). Rotomer 2: 1H NMR (C6D6): δ 1.289 (d, JP-H ) 3.2 Hz, 15 H),
1.151 (d, JP-H ) 10.8 Hz, 9 H). 31P{1H} NMR (C6D6): δ 5.13 (dd,
JRh-P ) 141.3 Hz, JF-P ) 25.5 Hz). 19F NMR (C6D6): δ 116.8 (m, 1
F), 95.0 (m, 1 F), 56.2(m, 1 F), 50.9 (m, 1F), 46.9 (m, 1 F), 40.6 (m,
1 F), 40.2 (m, overlaps with Rotomer 1, 1 F). Anal. Calcd for C23H24-
ClF8PRh‚H2O: C, 44.50; H, 4.06. Found: C, 43.98; H, 3.76.
(32) Using the TEXSAN package, R ) (∑||Fo| - |Fc||)/∑|Fo|, Rw
)
[∑w(|Fo| - Fc|)2]1/2/∑w|Fo |, where w ) [σ2(Fo) + (FFo )2]1/2 for the non-
Poisson contribution weighting scheme. The quantity minimized was Σw(|Fo|
- |Fc|)2. Source of scattering factors fo, f′, and f′′: Cromer, D. T.; Waber,
J. T. International Tables for X-ray Crystallography; The Kynoch Press:
Birmingham, England, 1974; Vol IV, Tables 2.2B and 2.3.1.
2
2

Mechanism of C-F Bond ActiVation by (C5Me5)Rh(PMe3)H2
J. Am. Chem. Soc., Vol. 119, No. 33, 1997 7741
the TEXSAN software package as described above for 3-Cl, giving
final parameters of R1 ) 0.0491 and Rw ) 0.0546.
by dissolving the oil in anhydrous THF and stirring over 4 Å molecular
sieves for 24 h. The solvent was removed and the oil stored in dry
benzene at -25 °C. A trace of water is always present. This solution
loses its potency within several weeks.
X-ray Structural Determination of (C5Me5)Rh(PMe3)(perfluoro-
naphthyl)Cl (6-Cl). Orange-red crystals were obtained by layering a
CH2Cl2 solution of 6-Cl with hexanes. A single crystal of dimensions
0.40 × 0.35 × 0.10 mm3 was mounted on a glass fiber with epoxy.
Data were collected at 20 °C on a Siemens SMART CCD area detector
system employing a 3 kW sealed tube X-ray source operating at 1.5
kW; 1.3 hemispheres of data were collected over 7 h, yielding 9459
observed data after integration by using SAINT (see Table 1). Laue
symmetry revealed a monoclinic crystal system, and cell parameters
were determined from 5832 unique reflections.33 An absorption
correction was applied by using SADABS. The space group was
assigned as P21/c on the basis of systematic absences by using XPREP,
and the structure was solved and refined by using the SHELX95
package. For a Z value of 4 there is one independent molecule within
the asymmetric unit. Examination of a difference map following
isotropic refinement of the (C5Me5)Rh(PMe3) fragment showed evidence
for a rotational disorder of the perfluoronaphthyl group. In the
refinement model, the two different orientations were allowed to refine
independently while restraining intraring distances to be similar by using
the SHELX SAME instruction. The occupancies of the two orientations
were fixed at 1:1. In the final model, all non-hydrogen atoms were
refined anisotropically (on F2), with hydrogens included in idealized
locations. A water molecule was included in a crystal void. Final
agreement factors were R1 ) 0.0725 and wR2 ) 0.1901.34 Fractional
coordinates and thermal parameters are given in the Supporting
Information.
Reaction of (C5Me5)Rh(PMe3)H2 with Perfluorobiphenyl with
and without TBAF. Two resealable NMR tubes containing (C5Me5)-
Rh(PMe3)H2 (0.035 M) and perfluorobiphenyl (0.36 M) were prepared
with dioxane as an internal standard in 1:1 pyridine-d5/C6D6 solvent.
TBAF (0.21 M) was added to one of the samples, and the tubes were
heated in an NMR probe at 69 °C. Both reactions were monitored by
1H NMR until the starting material was depleted. The approximate
half-lives were 460 min (90% yield) in the absence of fluoride and 10
min (93% yield) in the presence of fluoride. Another experiment was
performed with 0.0077 M added fluoride ion. Two similar samples
were prepared in neat pyridine-d5 solvent and allowed to react at
ambient temperature (21 °C). The approximate half-lives were 90 h
(87% yield) in the absence of fluoride and 14 h (93% yield) in the
presence of fluoride.
Reaction of (C5H5)Rh(PMe3)H2 with Perfluorobiphenyl.
A
resealable NMR tube containing (C5H5)Rh(PMe3)H2 (0.035 M) and
perfluorobiphenyl (0.36 M) in pyridine-d5 was prepared with dioxane
as an internal standard. The tube was left to stand at room temperature
1
(21.5 °C). The reaction was monitored by H NMR until the starting
material was depleted. The approximate half-live was 48 h (95% yield).
1H NMR (pyridine-d5): δ 5.680 (s, 5 H), 1.561 (dd, JP-H ) 11.0 Hz,
JP-Rh ) 1.0 Hz, 9 H), -12.052 (dd, JP-H ) 40.4 Hz, JRh-H ) 23.6 Hz,
1 H). 31P{1H} NMR (pyridine-d5): δ 14.5 (d, JRh-P ) 141.7 Hz).
Photolysis of 1 in the Presence of C6F5H. A resealable NMR tube
containing a 0.5 mL C6D12 solution of (C5Me5)Rh(PMe3)H2 (0.0173
mmol) and pentafluorobenzene (0.0865 mmol) was irradiated for ca. 3
h with a water-filtered 200-W Hg Applied Photophysics lamp. The
only product formed was (C5Me5)Rh(PMe3)(C6F5)H resulting from
C-H activation of pentafluorobenzene.
Reaction of (C5Me5)Rh(PMe3)H2 with Various Concentrations
of Perfluorobiphenyl. Four NMR tubes were prepared with varying
concentrations of perfluorobiphenyl (0.66, 0.52, 0.36, and 0.23 M) in
1:1 pyridine-d5/C6D6. The concentration of (C5Me5)Rh(PMe3)H2 in
each tube was maintained at 0.035 M. Dioxane (1 µL) was added as
an internal standard. The NMR tubes were heated at 92 °C, and the
reaction was followed by 1H NMR spectroscopy until the starting
material was depleted. The approximate half-lives (t1/2) for each
Thermolysis of (C5Me5)Rh(PMe3)H2 with C6F5Cl. A sample of
(C5Me5)Rh(PMe3)H2 (10.7 mg, 0.029 mmol) was dissolved in ap-
proximately 0.4 mL of C6D6 in a resealable NMR tube along with C6F5-
Cl (3.74 µL, 0.029 mmol) and 1 µL of dimethoxyethane as an internal
standard. The volume was brought to 0.6 mL with C6D6. The reaction
was monitored by 1H NMR and 31P NMR at ambient temperature. After
concentration was determined: 0.66 M, t1/2 = 38 min; 0.52 M, t1/2
44 min; 0.36 M, t1/2 = 64 min; 0.23 M, t1/2 = 88 min.
=
Reaction of (C5Me5)Rh(PMe3)H2 with Perfluorobiphenyl and
Pyridinium Chloride. A resealable NMR tube containing (C5Me5)-
Rh(PMe3)H2 (0.035 M) and perfluorobiphenyl (0.36 M) in 1:1 pyridine-
d5/C6D6 was prepared with dioxane as an internal standard. Hydrogen
chloride gas (12 equiv based on 1) was condensed into the tube. The
tube was heated in the NMR probe at 92 °C. The reaction was
1
380 h H NMR and 31P NMR showed the formation of (C5Me5)Rh-
(PMe3)HCl as the dominant organometallic species along with a small
amount of (C5Me5)Rh(PMe3)Cl2. 1H NMR spectroscopy and GC-MS
showed that 1 equiv of C6F5H had been formed.
Thermolysis of (C5Me5)Rh(PMe3)H2 with C6F5Br. A sample of
(C5Me5)Rh(PMe3)H2 (10.7 mg, 0.029 mmol) was dissolved in ap-
proximately 0.4 mL of C6D6 in a resealable NMR tube along with C6F5-
Br (3.65 µL, 0.029 mmol) and 1 µL of dimethoxyethane as an internal
standard. The volume was brought to 0.6 mL with C6D6. The reaction
was monitored by 1H NMR and 31P NMR at ambient temperature. After
1
monitored by H NMR spectroscopy until the starting material was
depleted. The main product was (C5Me5)Rh(PMe3)Cl2; a singlet at δ
4.37 corresponding to free H2 was also observed.
Reaction of (C5Me5)Rh(PMe3)H2 with Perfluorobiphenyl and
Pyridinium Tetrafluoroborate. A resealable NMR tube containing
(C5Me5)Rh(PMe3)H2 (0.035 M) and perfluorobiphenyl (0.36 M) and
pyridinium tetrafluoroborate (0.14 M) in 1:1 pyridine-d5/C6D6 was
prepared with dioxane as an internal standard. The tube was heated in
1
91 h, H NMR and 31P NMR showed the formation of (C5Me5)Rh-
(PMe3)HBr. 1H NMR spectroscopy and GC-MS showed that 1 equiv
of C6F5H had been formed.
1
the NMR probe at 92 °C. The reaction was monitored by H NMR
Generation of (C5Me5)Rh(PMe3)HCl. (C5Me5)Rh(PMe3)Cl2 (6.7
mg, 0.0175 mmol) and [PPN]+[BH4]- (14.6 mg, 0.026 mmol) were
combined in a Teflon sealed ampule with 2 mL of THF in the drybox.
The suspension was stirred at room temperature for 45 min then filtered
through cotton with hexanes. The solvent was removed under vacuum
and transferred to an NMR tube with C6D6. 1H NMR revealed a
mixture of compounds: (C5Me5)Rh(PMe3)Cl2 (ca. 50%), (C5Me5)Rh-
(PMe3)H2 (30%), and (C5Me5)Rh(PMe3)HCl (20%). 1H NMR
(C6D6): δ 1.623 (d, JRh-H ) 1.6 Hz, 15 H), δ 1.151 (d, JP-H ) 10.8
Hz, 9 H), -11.754 (dd, JP-H ) 52.4, JRh-H ) 20.0, 1 H). 31P{Selective
1H} NMR (C6D6): δ 8.03 (dd, JRh-P ) 143.2 Hz, JH-P ) 47.5 Hz).
Generation of (C5Me5)Rh(PMe3)HBr. (C5Me5)Rh(PMe3)Br2 (8.8
mg, 0.0186 mmol) and [PPN]+[BH4]- (18.5 mg, 0.033 mmol) were
combined in a Teflon sealed ampule with 2 mL of dry THF in the
drybox. The suspension was stirred at room temperature for 1 h and
then filtered through cotton with hexanes. The solvent was removed
under vacuum and transferred to an NMR tube with C6D6. 1H NMR
spectroscopy until the starting material was depleted. The half-life of
1 was =65 min and the yield of 4 was 53%. The complex [Cp*Rh-
(PMe3)2H]+ was formed in 15% yield.
Reaction of (C5Me5)Rh(PMe3)H2 with Perfluorobiphenyl and
Pyridinium Hydrogen Fluoride. A resealable NMR tube (equipped
with a Teflon insert) containing (C5Me5)Rh(PMe3)H2 (0.035 M),
perfluorobiphenyl (0.36 M), and pyridinium poly(hydrogen fluoride)
(0.035 M) in 1:1 pyridine-d5/C6D6 was prepared with dimethoxyethane
as an internal standard. The tube was heated in the NMR probe at 92
°C The reaction was monitored by 1H NMR until the starting material
was depleted. The half-life of 1 was ∼66 min and the yield of 4 was
55%. The complex [Cp*Rh(PMe3)2H]+ was formed in 23% yield.
Anhydrous TBAF. TBAF‚3H2O was heated under high vacuum
with stirring at 40 °C for 48 h.35 The resulting oil was further dried
(33) It has been noted that the integration program SAINT produces cell
constant errors that are unreasonably small, since systematic error is not
included. More reasonable errors might be estimated at 10× the listed values.
(34) Using the SHELX95 package, R1 ) (∑||Fo| - |Fc||)/∑|Fo|, wR2 )
[∑[w(Fo2 - Fc )2]/∑w|Fo )2]}1/2, where w ) [σ2(Fo ) + (aP)2 + bP] and P
(35) Cox, D. P.; Terpinski, J.; Lawrynowicz, W. J. Org. Chem. 1984,
49, 3216-3219.
2
2
2
2
2
) [f‚(maximum of 0 or Fo ) + (1 - f)Fc ].

7742 J. Am. Chem. Soc., Vol. 119, No. 33, 1997
Edelbach and Jones
spectroscopy revealed a mixture of compounds: (C5Me5)Rh(PMe3)Br2
(ca. 20%), (C5Me5)Rh(PMe3)H2 (20%), and (C5Me5)Rh(PMe3)HBr
(60%). 1H NMR (C6D6): δ 1.659 (d, JRh-H ) 1.6 Hz, 15 H), δ 1.187
(dd, JP-H ) 10.4 Hz, JRh-H ) 0.8 Hz, 9 H), -12.047 (dd, JP-H ) 50.8,
JRh-H ) 20.0, 1 H). 31P{Selective 1H} NMR (C6D6): δ 6.02 (dd, JRh-P
) 143.8 Hz, JH-P ) 48.0 Hz).
Table 2. Peak Reduction Potentials (EPC) and Peak Oxidation
Potentials (EPA) for Various Polyfluoroaromatics,
(C5Me5)Rh(PMe3)H2 and (C5H5)Rh(PMe3)H2 (V vs Ferrocene/
Ferrocenium Ion Couple)a
perfluoroaromatics
EPC (V)
metal complex
(C5Me5)Rh(PMe3)H2 -0.336
(C5H5)Rh(PMe3)H2 +0.279
EPA (V)
perfluoronaphthalene
perfluorobenzene
perfluorobiphenyl
pentafluorobenzene
-2.42
-2.45
-2.61
-2.97
[(C5Me5)Rh(PMe3)H]xLix.20 This orange material was prepared as
previously reported. 1H NMR (pyridine-d5): δ 2.292 (d, JRh-H ) 1.0
Hz, 15 H), 1.343 (d, JP-H ) 7.3 Hz, 9 H), -16.525 (overlapping dd,
JP-H ) JRh-H ) 44.6 Hz, 1H). 31P{1H} NMR (pyridine-d5): δ -6.258
(d, JRh-P ) 232.4).
Reaction of [(C5Me5)Rh(PMe3)H]xLix with Perfluorobiphenyl. A
pyridine-d5 solution of [(C5Me5)Rh(PMe3)H]xLix (10.5 mg, 0.035 mmol)
was added to a pyridine-d5 solution of perfluorobiphenyl (85 mg, 0.254
mmol) at room temperature. 1H and 31P NMR spectroscopy indicated
the immediate quantitative formation of 4 and 5 in a ratio of
approximately 3 to 1.
a Experimental conditions: [polyfluoroaromatics] ) 3.0 × 10-3 M;
[metal complex] ) 1.0 × 10-3 M; [supporting electrolyte] ) 0.1 M
tetrabutylammonium hexafluorophosphate; temperature ) 23 ( 2 °C,
scan rate ) 100 mV/s; platinum working electrode, platinum wire
auxiliary electrode, and a silver wire reference electrode were used for
the measurements. Potentials were referenced to a ferrocene internal
standard. Tetrahydrofuran was the solvent used.
Reaction of [(C5Me5)Rh(PMe3)H]xLix with Perfluoronaphthalene.
A pyridine-d5 solution of [(C5Me5)Rh(PMe3)H]xLix (10.5 mg, 0.035
mmol) was added to a pyridine-d5 solution of perfluoronaphthalene
(69 mg, 0.254 mmol) at room temperature. 1H and 31P NMR
spectroscopy indicated the immediate quantitative formation of both
rotomers of 6 in the same 1.25:1 ratio as the thermal reactions with 1.
Electrochemical Studies. The cyclic voltametry measurements of
(C5Me5)Rh(PMe3)H2 and (C5H5)Rh(PMe3)H2 were carried out under
an inert atmosphere with a EG & G PARR Potentiostat/Galvanostat
Model 263A in a nitrogen drybox. The polyfluoroaromatic compounds
were run under a nitrogen purge in a polarographic cell. The working
electrode was a 3 mm diameter platinum disk which was polished prior
to each use. A platinum wire served as the auxiliary electrode, and a
silver wire served as a pseudoreference electrode. The CV experiments
were performed in THF at room temperature with ferrocene as an
internal reference, 0.1 M in supporting electrolyte (tetra-n-butylam-
monium hexafluorophosphate) and 3.0 mM in the polyfluoroaromatics
and 1.0 mM in the metal complexes. The scan rate in all experiments
was 0.1 V/s.
potentials (EPA) for the polyfluoroaromatics and the metal complexes
are reported versus the ferrocene/ferrocenium ion couple (Table 2). The
peak reduction potentials for perfluorobenzene and perfluoronaphthalene
are in close agreement to those reported by Pez et al.26
Acknowledgment is made to the U.S. Department of Energy,
grant FG02-86ER13569, for their support of this work. We also
thank Dr. Witold Paw for his assistance with the electrochemical
measurements, Dr. Charles Campana of Siemens for assistance
with the structure of 6, and Prof. J. P. Dinnocenzo for valuable
discussions.
Supporting Information Available: Tables of bond dis-
tances and angles, atomic positions, and thermal parameters for
3, 4, and 6 as well their NMR spectroscopic data in C6D6 (18
pages). See any current masthead page for ordering and Internet
access instructions.
All of the species tested displayed irreversible behavior with a scan
rate of 0.1 V/s. The peak reduction potentials (EPC) and peak oxidation
JA970723R