Facile and reversible cleavage of C-F bonds. Contrasting thermodynamic selectivity for Ru-CF2H vs F-Os=CFH

Article abstract of DOI:10.1021/ja001646u

In the presence of a catalytic amount of F^-(CsF), Me₃SiR(f) (R(f) = CF₃ and C₆F₅) exchanges R(f) with fluoride of the 16-electron complexes MHF(CO)L₂ (M = Ru, Os; L = P(i)Pr₃, P(t)Bu₂Me) to give Me₃Si-F and the unsaturated pentafluorophenyl complexes, MH(C₆F₅)(CO)L₂, or (when R(f) = CF₃) saturated fluorocarbene complexes, MHF(CF₂)(CO)L₂, via α-fluorine migration. X-ray crystal structure and solution ¹⁹F NMR studies reveal that, in the ground state, the three atoms of the CF₂ group lie in a plane perpendicular to the P-Ru-P axis so that the π-back-donation is maximized and the carbene substituents are inequivalent. Having hydride trans to the CF₂ ligand, MHF(CF₂)(CO)L₂ is a kinetic product, which converts to a thermodynamic isomer. For Ru, the final product is a 16e complex, RuF(CF₂H)(CO)L₂, formed by combination of CF₂ and hydride. For Os, the product is an 18e complex, OsF₂(=CFH)(CO)L₂, resulting from exchange of one carbene fluoride with the hydride. The distinct difference between Os and Ru demonstrates the principle that third-row transition metals show a pronounced tendency toward a higher oxidation state. The isomerization mechanism involves phosphine dissociation as a slow step. Coordinatively saturated RuHF(CF₂)(CO)L₂ reacts with CO within the time of mixing to give the F and CF₂ recombination product, RuH(CF₃)(CO)₂L₂. This unexpectedly fast carbonylation reaction, as well as ¹⁹F spin saturation transfer experiments, reveals the existence of a fast α-F migration equilibrium between RuHF(CF₂)(CO)L₂ and RuH(CF₃)(CO)L₂ in solution. In sharp contrast, the Os analogue does not have such a fast equilibrium, and therefore it does not react with CO at room temperature. At higher temperature, reaction occurs forming the hydride and fluoride exchanged product, Os(CHF₂)(F)-(CO)₂L₂. The contrasting behavior of Ru vs Os regarding stability of fluoroalkyl and fluorocarbene is discussed on the basis of the theoretical calculations, which also provide insight into the isomerization of RuHF(CF₂)-(CO)L₂. Hydrogenolysis of Ru(CF₂H)F(CO)L₂ liberates CH₂F₂, forming RuHF(CO)L₂.

Full text of DOI:10.1021/ja001646u

8916
J. Am. Chem. Soc. 2000, 122, 8916-8931
Facile and Reversible Cleavage of C-F Bonds. Contrasting
Thermodynamic Selectivity for RusCF2H vs FsOsdCFH
Dejian Huang, Patrick R. Koren, Kirsten Folting, Ernest R. Davidson,* and
Kenneth G. Caulton*
Contribution from the Department of Chemistry and Molecular Structure Center, Indiana UniVersity,
Bloomington, Indiana 47405-4001
ReceiVed April 4, 2000
Abstract: In the presence of a catalytic amount of F-(CsF), Me3SiRf (Rf ) CF3 and C6F5) exchanges Rf with
fluoride of the 16-electron complexes MHF(CO)L2 (M ) Ru, Os; L ) PiPr3, PtBu2Me) to give Me3Si-F and
the unsaturated pentafluorophenyl complexes, MH(C6F5)(CO)L2, or (when Rf ) CF3) saturated fluorocarbene
complexes, MHF(CF2)(CO)L2, via R-fluorine migration. X-ray crystal structure and solution 19F NMR studies
reveal that, in the ground state, the three atoms of the CF2 group lie in a plane perpendicular to the P-Ru-P
axis so that the π-back-donation is maximized and the carbene substituents are inequivalent. Having hydride
trans to the CF2 ligand, MHF(CF2)(CO)L2 is a kinetic product, which converts to a thermodynamic isomer.
For Ru, the final product is a 16e complex, RuF(CF2H)(CO)L2, formed by combination of CF2 and hydride.
For Os, the product is an 18e complex, OsF2(dCFH)(CO)L2, resulting from exchange of one carbene fluoride
with the hydride. The distinct difference between Os and Ru demonstrates the principle that third-row transition
metals show a pronounced tendency toward a higher oxidation state. The isomerization mechanism involves
phosphine dissociation as a slow step. Coordinatively saturated RuHF(CF2)(CO)L2 reacts with CO within the
time of mixing to give the F and CF2 recombination product, RuH(CF3)(CO)2L2. This unexpectedly fast
carbonylation reaction, as well as 19F spin saturation transfer experiments, reveals the existence of a fast R-F
migration equilibrium between RuHF(CF2)(CO)L2 and RuH(CF3)(CO)L2 in solution. In sharp contrast, the Os
analogue does not have such a fast equilibrium, and therefore it does not react with CO at room temperature.
At higher temperature, reaction occurs forming the hydride and fluoride exchanged product, Os(CHF2)(F)-
(CO)2L2. The contrasting behavior of Ru vs Os regarding stability of fluoroalkyl and fluorocarbene is discussed
on the basis of the theoretical calculations, which also provide insight into the isomerization of RuHF(CF2)-
(CO)L2. Hydrogenolysis of Ru(CF2H)F(CO)L2 liberates CH2F2, forming RuHF(CO)L2.
Introduction
dimethoxyethane)7 and Hg(CF3)2(DME).8 However, these meth-
ods are usually quite specific or involve highly toxic metals.
Therefore, more general, milder, and safer synthetic routes are
in demand.
Transition metal complexes with polyfluoroalkyl ligands are
receiving significant attention because they are crucial in the
polyfluorocarbon C-F bond activation and functionalization
processes, synthesis of fluorocarbon-containing organic com-
pounds with pharmaceutical interest, and development of
organometallic catalysts soluble in polyfluorocarbon solvents.1,2
The chemistry of transition metal perfluoroalkyl complexes was
studied about 30 years ago, with the focus on new synthetic
methods, mainly by Stone and co-workers.3 Some pioneering
work on metal trifluoromethyl complexes, the simplest perfluo-
roalkyl, was also carried out at that time.4 Since then, several
methods and reagents have been developed to synthesize this
type of complex, such as oxidative addition of CF3I or CF3Br
to low-valent metals,5 decarbonylation of trifluoroacetyl com-
plexes,6 and application of Cd(CF3)2(DME) (DME ) 1,2-
A convenient CF3 transfer reagent, Me3SiCF3, has been
exploited extensively in organic synthesis for several years.9
In sharp contrast, its application in organometallic synthesis
has not been previously reported. Reported here are the first
example of such applications, and, surprisingly, these yield Ru
and Os difluorocarbene complexes. Constructive transformation
of coordinated fluorocarbyl ligands to valuable organic com-
pounds is an important research topic relevant to utilization of
perfluorocarbon wastes.10 It is well documented that the
perfluoroalkyl metal-C bond is stronger than that in hydro-
carbon analogues and is thus relatively inert toward migratory
insertion, an important type of reaction for transition metal alkyl
complexes applied in organic synthesis.11 On the other hand,
in many cases the R-fluorine of a perfluoroalkyl group is
(1) (a) Hughes, R. P. AdV. Organomet. Chem. 1990, 31, 183. (b)
Kiplinger, J. L.; Richmond, T. G.; Osterberg, C. E. Chem. ReV. 1994, 94,
373. (c) Murphy, E. F.; Muragavel, R.; Roesky, H. W. Chem. ReV. 1997,
97, 3425.
(5) King, R. B.; Stafford, S. L.; Treichel, P. M.; Stone, F. G. A. J. Am.
Chem. Soc. 1961, 83, 3604.
(2) (a) Hughes, R. P.; Husebo, T.; Holiday, B. J.; Rheingold, A. L.;
Liable-Sands, L. M. J. Organomet. Chem. 1997, 548, 109. (b) Pozzi, G.;
Cavazzini, Q. S.; Fontana, S. Tetrahedron Lett. 1997, 38, 7605. (c) Koch,
D.; Baumann, W.; Leitner, W. Angew. Chem., Int. Ed. Engl. 1997, 36, 1628.
(3) (a) Treichel, P. M.; Stone, F. G. A. AdV. Organomet. Chem. 1964,
1, 143. (b) Stone, F. G. A. Pure Appl. Chem. 1972, 30, 551. (c) Bruce, M.
I.; Stone, F. G. A. Prepr. Inorg. React. 1968, 4, 117.
(6) Hensley, D. W.; Warster, W. L.; Stewart, R. P. Inorg. Chem. 1981,
20, 645.
(7) Krause, L. J.; Morrison, J. A. J. Am. Chem. Soc. 1981, 103, 2995.
(8) Emeleus, H. J. The Chemistry of Fluorine and its Compounds;
Academic Press: New York, 1969.
(9) Prakash, G. K. S.; Yudin, A. K. Chem. ReV. 1997, 97, 757.
(10) Stiens, D.; Richmond, T. G. Chemtracts: Inorg. Chem. 1998, 11,
900.
(4) McClellan, W. R. J. Am. Chem. Soc. 1961, 83, 1598.
10.1021/ja001646u CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/31/2000

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8917
activated and is subject to electrophilic attack. Therefore, the
majority of reaction pathways of perfluoroalkyl complexes
involve R-F abstraction by Lewis acids to give difluorocarbenes,
which are often highly hygroscopic and further react with water
to give coordinated carbonyl ligands and release HF.12 In some
cases, coordinated water is sufficiently acidic to protonate the
R-fluoride.13,14 Rarely has an R-fluoro migration of an unsatur-
ated metal perfluoroalkyl complex (which functions as an
internal Lewis acid) been reported.15 Presented here is one such
example where a fast reversible R-F migration occurs on an
Ru complex and double R-F migration occurs on an analogous
Os complex.16
Results
Figure 1. ORTEP diagram of RuHF(CF2)(CO)(PtBu2Me)2, 4. Hydro-
gen atoms are omitted except for the one bound to Ru. Selected bond
lengths: Ru1-F4, 2.065(1); Ru1-C5, 1.952(3); Ru1-C2, 1.820(2);
Ru1-P18, 2.4048(4); Ru1-P8, 2.4105(6); C5-F7, 1.323(2); C5-C6,
1.341(3). Angles: P18-Ru1-P8, 159.23(3); C2-Ru1-F4, 178.38-
(10); Ru1-C5-F6, 129.77(13); Ru1-C5-F7, 127.10(8); F6-C5-
F7, 103.09(23).
Reaction of MHF(CO)L2 (M ) Os, Ru; L ) PtBu2Me,
PiPr3) with Me3SiCF3. In the presence of a catalytic amount
(5-10 mol percent) of CsF, Me3SiCF3 reacts with MHF(CO)-
L217 to give MHF(CF2)(CO)L2, 4, 5, and 6 (eq 1).18 The reaction
doublets with large coupling constants (45-50 Hz) to CF2
fluorine atoms and the smaller coupling (12 Hz) to metal-bound
fluoride. The 31P{1H} signal is a doublet of triplets due to two
types of fluorine atoms. The carbene carbon resonance appears
at low field (254 ppm for 4, and 225 ppm for 6) as a triplet of
multiplets with large 1J(CF) values (495 Hz for 4, and 445 Hz
for 6), in agreement with the presence of a metal-bound CF2
unit. At room temperature, two broad 19F resonances are detected
with a 2:1 ratio, at low and high field (for CF2 and M-F,
respectively). Since the CF2 is a single-faced π-acid, its
orientation becomes an issue. Upon cooling of the sample to
-80 °C, the CF2 peak in 4 decoalesces (δ(Fa) - δ(Fb) ) 10
is quite solvent dependent. In a polar solvent such as THF or
fluorobenzene, the reaction is complete in 30 min (M ) Ru),
while in benzene the reaction takes longer (5 h). The solvent
dependence is likely to be related to the better solubility of CsF
in polar solvents. The osmium analogue, 3, reacts more
sluggishly and requires heating in fluorobenzene (80 °C) in order
to complete the reaction in 30 min. While 4 and 5 react instantly
with water, 6 reacts slowly (over 3 h), in all cases to produce
HF19 and MHF(CO)2L2, which has been synthesized indepen-
dently from 1-3 and CO. To completely exclude moisture,
surface silylated glassware must be used. Complexes 4-6 have
been characterized spectroscopically and by an X-ray single-
crystal study (for 4). The most distinctive NMR spectral features
are the hydride peak at -2 to -3 ppm as a triplet of triplets of
2
ppm) to one AMX (X ) H) pattern with a large J(FF) value
(221 Hz). These results reveal that the ground-state geometry
of the CF2 ligand has the p orbital of the carbene carbon parallel
to the P-M-P axis. By adopting this orientation, the CF2 ligand
avoids competing for the same dB electrons with CO. Thus, it
maximizes the π-back-donation from the metal. The CF2 ligand
in crystallographically characterized [CpFe(CO)(PPh3)(CF2)]BF4
also adopts an orientation for maximum back-donation.12d
The structure of 4 was determined by X-ray single-crystal
analysis. The ORTEP diagram is depicted in Figure 1. As in
common six-coordinate d6 complexes, 4 adopts an octahedral
geometry with the two sterically demanding phosphines trans
to each other and the strong π-donor F trans to the strong
π-acceptor CO. Thus, the push-pull stabilization is maxi-
mized.20 The hydride is located trans to the CF2 ligand, which
also, in the solution state, lies in the plane perpendicular to the
P-Ru-P plane, as suggested from the low-temperature 19F
NMR spectrum. The Ru-F distance is 2.065 Å. This value is
comparable to the Ru-F distance (2.011 Å) of RuF2(CO)2-
(PPh3)2, in which the F is also trans to CO.21 The Ru-CF2
distance is 1.952 Å. No comparable value can be found since
this complex is the first structurally characterized six-coordinate
Ru difluorocarbene complex.22 A structurally characterized five-
coordinate RuCF2 complex, Ru(CO)2(CF2)(PPh3)2, has a shorter
Ru-CF2 bond (1.829 Å).23 Apparently, the back-donation in
(11) (a) Brothers, P. J.; Roper, W. R. Chem. ReV. 1988, 88, 1293. (b)
Morrison, J. A. AdV. Inorg. Chem. Radiochem. 1983, 27, 293.
(12) Some examples of hydrolysis of perfluoroalkyl: (a) Reger, D. L.;
Dukes, M. D. J. Organomet. Chem. 1978, 157, 67. (b) Michelin, R. A.;
Ros, R.; Guadalupi, G.; Bombieri, G.; Benetollo, F.; Chapuis, G. Inorg.
Chem. 1989, 28, 840. (c) Clark, G. R.; Hoskins, S. V.; Roper, W. R. J.
Organomet. Chem. 1982, 234, C9. (d) Crespi, A. M.; Shriver, D. F.
Organometallics 1985, 4, 1830. (e) Koola, J. D.; Roddick, D. M.
Organometallics 1991, 10, 591. (f) Richmond, T. G.; Crespi, A. M.; Shriver,
D. F. Organometallics 1984, 3, 314.
(13) Hughes, R. P.; Lindner, D. C.; Rheingold, A. L.; Liable-Sands, L.
M. J. Am. Chem. Soc. 1997, 119, 11544.
(14) Hughes, R. P.; Rose, P. R.; Rheingold, A. L. Organometallics 1993,
12, 3109.
(15) Photolysis of CpMo(CO)3(COCF3) in frozen argon matrix at ca. 12
K has been monitored by IR spectroscopy, and the product was proposed
to be CpMo(CO)2(F)(dCF2). Campen, A. K.; Mahmoud, K. A.; Rest, A. J.
Willis, P. A. J. Chem. Soc., Dalton Trans. 1990, 2817.
(16) Part of this work has appeared as a communication: Huang, D.;
Caulton, K. G. J. Am. Chem. Soc. 1997, 119, 3185. The essential need for
fluoride ion catalysis in the syntheses (vide infra) was not recognized in
this communication.
(17) Poulton, J. A.; Sigalas, M. P.; Folting, K.; Steib, W. E.; Eisenstein,
O.; Caulton, K. G. Inorg. Chem. 1994, 33, 1476.
(20) Caulton, K. G. New J. Chem. 1994, 18, 25.
(21) Brewer, S. A.; Coleman, K. S.; Fawcett, J.; Holloway, J. H.; Hope,
E. G.; Russell, D. R.; Watson, P. G. J. Chem. Soc., Dalton Trans. 1995,
1073.
(22) A spectroscopically characterized Ru(II)-CF2 complex has been
reported; see: Clark, G. R.; Hoskins, S. V.; Roper, W. R. J. Organomet.
Chem. 1982, 234, C9.
(18) For other examples of group (X) transfer to M-F from Me3SiX,
see: (a) Doherty, N. M.; Critchlow, S. C. J. Am. Chem. Soc. 1987, 109,
7906. (b) Hoffman, N. W.; Prokopuk, N.; Robbins, M. J.; Jones, C. M.;
Doherty, N. M. Inorg. Chem. 1991, 30, 4177.
1
(19) HF is demonstrated by NMR (toluene-d8) spectroscopy. H NMR:
1
1
10.75 (d, JHF ) 445 Hz). 19F NMR: -187 (d, JHF ) 445 Hz).

8918 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
the latter complex is larger due to the lower oxidation state of
the metal and the CF2 carbon is not expected to be as
electrophilic as in 4. A quantum computation (below) reveals
this lengthening to also be a result of the carbene being trans
to hydride. The Ru-CF2 distance (1.829 Å) of Ru(CO)2(CF2)-
(PPh3)2 compares to the Ru-CO distance (1.870-1.915 Å) in
the same molecule, showing the multiple bond character of both.
Moreover, bending of the P-Ru-P angle (of 4) away from the
carbene effects rehybridization of the filled dB orbital to enhance
back-bonding to the carbene.24
Isomerization of RuHF(CF2)(CO)L2. 4 has a long lifetime
in a nonpolar solvent such as benzene (1 week) at 20 °C.
However, it rearranges within 5 h in THF to RuF(CF2H)(CO)-
L2, 7 (eq 2), which is isolated as yellow crystals from pentane.
5 also isomerizes to 8 in THF, but much more slowly (over 48
L2, RuH(CF2)(CO)L, or RuH(CF2)(CO)L2+. We tested all three
possibilities individually.
(a) The Effect of Added Co on the Isomerization. To test
for the rate suppression in a CO dissociation mechanism, we
stirred RuHF(CF2)CO)L2 with CO (1 atm) in THF at 20 °C.
Surprisingly, the saturated complexes 4 and 5 react with CO;
within the time of mixing the yellow color of 4 and 5 gave
way to a colorless solution. NMR spectral analysis reveals
formation of RuH(CF3)(CO)2L2, 9 and 10 (eq 3), the product
of the combination of CF2 and F (not hydride) with a CF3 ligand.
9 and 10 were characterized spectroscopically. The most
characteristic feature is the hydride resonance, which is a triplet
of quartets with large coupling to the phosphine and small
coupling to the CF3. The existence of a CF3 group is also
supported by the 13C{1H} NMR spectrum, which gives a quartet
of triplets with a large 1JCF value (358 Hz). This 1J(C(sp3)-F)
coupling constant typically increases with increasing fluorine
content.26 The geometries of 9 and 10 are deduced from the
fact that there are two virtual triplets for the methyl protons of
tBu (9) and two doublets of virtual triplets for the methyl protons
h) and with formation of unidentified byproducts. In the solid
state, however, 4 and 5 are converted to 7 and 8 cleanly in about
6 months. A similar phosphine dependence of the isomerization
rate on the phosphine cone angle, i.e., Pr3P vs Bu2Me in Ru
and Os carbene (CH2) complexes, has been attributed to the
different size of the two phosphine ligands, with PtBu2Me being
larger than PiPr3.25
i
t
i
of Pr (10) accounting for the trans phosphine geometry. The
13C NMR spectrum also reveals two CO signals, one a triplet
of quartets (trans to CF3) and the other a triplet. At room
temperature the NMR signals of 9 are broad due to the hindered
rotation around the Ru-P bond.27 This hindered rotation does
not give rise to two isomers for 10, which has sharp and well-
resolved NMR signals over the temperature range from +60 to
-80 °C. If 13CO is used in the reaction of eq 3, the 13CO is
found exclusively trans to the hydride. The 1H NMR spectrum
of the isotopomer of 10, RuH(CF3)(CO)(13CO)L2, 11, reveals
a hydride doublet of triplets of quartets with large coupling
constant (2J(CH) ) 29 Hz) to trans 13CO; consistent with this,
the proton-coupled 13C NMR spectrum gives a doublet of triplets
The most characteristic NMR feature of 7 is the proton
resonance of the CF2H group, a triplet (2J(HF) ) 59 Hz, 3J(HF)
) 5.6 Hz) of doublets at low field (8.2 ppm). Consistent with
this, the 19F NMR spectrum has a doublet of triplets of doublets
2
with the largest J(HF) coupling (59 Hz) to the CF2H proton.
The (Ru)-F signal is a broad peak at high field (-236 ppm).
The CF2H carbon resonance is a triplet of triplets of doublets
1
at 140 ppm with a J(FC) value (293 Hz) much smaller than
that of CF2 (495 Hz). Unlike the correlation of 1J(CH) with the
s orbital component of the carbon atom, there is no quantitative
relationship between carbon orbital hybridization and the
magnitude of 1J(CF).26 However, electron deficiency at carbon
of quartets at δ 203.8 with the same doublet coupling (2JCH
29 Hz).
)
1
often increases the J(CF) value, as is also true in this case.26
The rapid reaction between six-coordinated 4 (or 5) and CO
led us to suspect that there is a fast equilibrium between six-
coordinated RuHF(CF2)(CO)L2 and the unseen five-coordinated
RuH(CF3)(CO)L2 with CF3 trans to the CO (eq 4). Thus, we
The geometry of 7 is derived from the 13C{1H} NMR spectrum,
which gives a doublet of multiplets at 205 ppm for the carbonyl
ligand with a large coupling constant with Ru-F (75 Hz),
revealing the mutual trans disposition of the F-Ru-CO unit.
Consistent with this, the CO stretching frequency is low (1917
cm-1).
Isomerization Mechanism. 4 is a six-coordinate, saturated
complex, which is normally not fluxional. Hydride being trans
to the CF2 blocks rapid 1,2-H migration; this accounts for the
metastability of 4. To isomerize to 7, ligand dissociation is
essential before H and CF2 can become cis and combine. The
possible dissociating ligands are CO, L, and F-. THF can
stabilize one of these five-coordinate intermediates: RuHF(CF2)-
carried out a 19F spin saturation transfer (SST) study of 4 in
toluene. At 25 °C, saturation of the CF2 resonance does not
cause an observable intensity change of the Ru-F signal.
However, at elevated temperatures (>60 °C), irradiation of the
CF2 group causes a drastic loss of intensity of the fluoride bound
to the Ru. This result corroborates the presence of the fast and
reversible C-F bond cleavage and argues against CO dissocia-
tion being necessary for the 4 f 7 rearrangement.
(23) Clark, G. R.; Hoskins, S. V.; Jones, T. C.; Roper, W. R. J. Chem.
Soc., Chem. Commun. 1983, 719.
(24) (a) Werner, H.; Laubender, M. S.; Lehmann, C.; Herbst-Irmer, R.
Organometallics 1997, 16, 2236. (b) Gusev, D. G.; Kuhlman, R.; Sini, G.;
Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1994, 116, 2685.
(25) Huang, D.; Spivak, G. J.; Caulton, K. G. New J. Chem. 1998, 22,
1023.
(26) Breitmaier, E.; Voelter, W. Carbon-13 NMR Spectroscopy, 3rd ed.;
VCH: New York, 1987; p 160.
(27) For hindered rotation of complexes such as this, see: Notheis, J.
U.; Heyn, R. H.; Caulton, K. G. Inorg. Chim. Acta 1995, 229, 187.

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8919
(b) The Effect of Added Phosphine Ligand. Addition of
equimolar Bu2PMe to a THF-d8 solution of 4 inhibits the
Scheme 1
t
isomerization over 24 h at 20 °C. The only species observed
by 1H and 31P NMR spectra are the free phosphine and 4, ruling
out possibility of adduct formation of 4 with added phosphine.
Therefore, a preliminary step in the isomerization is reversible
phosphine dissociation.
(c) Fluoride Dissociation as Another Possible Mechanism.
We have proved that phosphine dissociation is sufficient for 4
to isomerize. Will F- dissociation also trigger the process?
Although the SST experiment reveals the rapid cleavage and
re-formation of the Ru-F bond, this process does not require
the dissociation of F-; instead, it more likely to be a migratory
insertion process, and the F- does not spontaneously dissociate
from Ru. Addition of a hydrogen-bonding donor, indole
(catalytic amount),28 to help F- dissociation does not accelerate
the isomerization in benzene. However, when fluoride is fully
abstracted by Me3SiOTf, the combination of CF2 with hydride
occurs. Addition of 1 equiv of Me3SiOTf to the benzene solution
of 4 generates Ru(CF2H)(OTf)(CO)L2, 12, cleanly within the
time of mixing at 25 °C (eq 5). The fast combination of CF2
is a quartet (J(PF) ) 6 Hz). The CF3 fluoride exhibits a well-
resolved doublet of triplets at -7.5 ppm. Removal of all the
volatiles regenerates 4 along with a small amount of 7. Similar
to CD3CN, addition of F- (as the anhydrous NMe4+ salt) rapidly
gives an adduct, 15, with CF3 trans to hydride (Scheme 1). The
hydride of 15 is an apparent sextet of doublets with the same
coupling constant with CF3 fluoride and the phosphine (20 Hz).
The smallest coupling constant, that of Ru-F (5.4 Hz), suggests
that F is cis to the hydride. The CF3 fluoride resonance is a
doublet of apparent quartets (3J(FH) ) 21 Hz, 3J(PF) ) 3J(FF)
) 6 Hz) at -10 ppm, and the fluoride peak appears as a
multiplet at high field (-325 ppm). The31P{1H} NMR spectrum
is a doublet of quartets (coupling with Ru-F and CF3). The
CO stretching frequency is lower (1910 cm-1) than that of 4,
in agreement with better π-donation of Ru in 15. The different
geometry of 15 compared to those of 13 and 10 is inconsistent
with F- addition trans to hydride in RuH(CF3)(CO)L2 and
strongly implies that the added fluoride attacks the CF2 carbon
instead. Less nucleophilic CsF does not interact with 4. To probe
the initial reaction site of F-, anhydrous n-Bu4NCl was used as
cheap “isotope” labeling reagent, but it fails to give any chloride
adduct of 4. Instead, it only accelerates the isomerization of 4
to 7; no Cl and F exchange is observed.
DFT Calculations. We carried out these calculations to learn
the relative energies of various isomeric structures, to learn more
about the equilibrium in eq 4 (including structural features of
the CF3 complex), and to compare ruthenium to osmium. Details
of the computational method, especially the need to model
fluorine adequately, are given in the Experimental Section. We
will first summarize aspects of the calculated structures. This
will consist of (a) comparison of the various structures for a
given metal and (b) comparison of Ru and Os analogues. Among
all 10 structures, M-P distances vary insignificantly (2.413-
2.445 Å).
with H in this reaction is apparently caused by dissociation of
the weak ligand triflate instead of phosphine. 12 is fully
characterized spectroscopically. The CF2H proton appears at low
field as a triplet of triplets with large 2J(HF) (58 Hz) and small
3J(HP) values (3 Hz). Two virtual triplets indicate diastereotopic
tBu. The 31P{1H} signal is a triplet, and the CF2H fluoride
appears as a doublet of triplets. In the 13C{1H} NMR spectrum,
the CO resonance is a triplet of triplets with a small CF coupling
constant (7.6 Hz), which supports the cis disposition of CF2H
and CO. The CO stretching frequency is high (1944 cm-1
)
compared to that of RuH(OTf)(CO)L2 (1921 cm-1),29 revealing
that CF2H is a weaker F-donor ligand than hydride. Although
the complex can be isolated as orange crystals from toluene at
-40 °C, it decomposes in 1 day at room temperature to give
HL+ and other, unknown species.
Reaction of 4 with Other Nucleophiles. The CF2 carbon of
4, and the Ru of an unobserved isomer of 4, [RuH(CF3)(CO)-
L2], are two potential positions for a nucleophile to attack. We
have demonstrated that CO attacks Ru of [RuH(CF3)(CO)L2]
irreversibly to give RuH(CF3)(CO)2L2. Similarly, CH3NC binds
4 to give 13 (Scheme 1), which is characterized spectroscopi-
cally. The most informative peak is the hydride resonance, which
is a triplet of quartets (J(FH) ) 6.3 Hz, J(PH) ) 26 Hz) at
-7.9 ppm. The CF3 fluoride appears at -13.7 ppm as a doublet
of broad triplets (J(FH) ) 6 Hz, J(PF) is not well resolved),
and the 31P signal is a broad peak at 64 ppm. These values
compare well with that of 9, suggesting that the two have similar
geometries. The much weaker donor CD3CN also attacks Ru
but reversibly to give RuH(CD3CN)(CF3)(CO)L2, 14. The
hydride of 14 appears at -9.7 ppm as a triplet of quartets (J(PH)
) 22 Hz, J(HF) ) 19 Hz). The significantly large JHF of 14
(vs 13) suggests it has a different geometry. The 31P NMR signal
Both MHF(CO)(PH3)2 species (M1) are square pyramids,
with H apical. Bond lengths for Ru and Os are identical to within
0.02 Å, but the H-M-F is 10° larger for Os.
The species (M2) with CF3 trans to hydride optimize to a
stationary point where one F is bonded to both carbon and
the metal, interacting with the site trans to CO; we will refer
to this as an agostic interaction. There is clearly an alterna-
tive isomer, M2′, but this is found not to have an agostic
interaction. Perhaps the trans effect of hydride prevents this
(28) Wessel, J.; Lee, J. C., Jr.; Peris, E.; Yap, G. P. A.; Fortin, J. B.;
Ricci, J. S.; Sini, G.; Albinati, A.; Koetzle, T.; Eisenstein, O.; Rheingold,
A. L.; Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1995, 34, 2507.
(29) Huang, D.; Huffman, J. C.; Eisenstein, O.; Caulton, K. G. J. Am.
Chem. Soc. 1997, 119, 7398.
agostic donation. It is worth noting that isomer M2′ cannot easily
form the observed next product, M3, while M2 can (the

8920 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
Figure 2. Optimized structures (distances in angstroms, angles in degrees) for MHF(CO)(PH3)2 and isomers of MH(CF3)(CO)(PH3)2, M ) Ru (Os
in parentheses), projected onto their mirror symmetry plane. Hydrogens of PH3 have been removed for simplicity. M2 and M4 each have one
fluorine on carbon eclipsed by the one shown.
emerging CF2 is trans to the hydride in M2), and also that lone
pairs of the agostic F in M2 can to some extent participate in
a push/pull interaction with the carbonyl ligand. Most bond
lengths differ insignificantly (0.01 Å) between Ru2 and Os2;
the exception is the MCF3 group, involving the agostic F
(denoted F*): the interaction of F* is stronger to Os than to
Ru, as judged by a shorter (by 0.19 Å) M-F distance and a
longer (by 0.06 Å) C-F* distance. As if to compensate the
CF3 carbon of M2, the remaining two C-F bonds in Os2 are
shorter (by 0.01 Å) than they are in Ru2. For comparison, in
H3Si-CF3, the calculated C-F bond length is 1.360 Å. The
distance Os-C is shorter by 0.01 Å and M-C-F* is smaller
(by 10°). To accommodate the F*/Os interaction, OC-Os-
CF3 is larger (by 11°). The angle F-C-F here and in the other
structures shown in Figure 2 is always ∼103°, so it is not
diagnostically useful. The structure of a relevant comparison
compound, X, shows the most remarkable change to be the rapid
development of a shorter Ru-CFx bond in Ru2, as RudC
(carbene) character develops concurrently with C-F* bond
stretching. This is also evident by comparing to the Ru-CF3
distance in Y, a conformer of Ru2.
Osmium is thus a stronger π base (i.e., more reducing) than
Ru. This is also evident in the MsCO distances, which are
shorter to Os in M1, M3, M4, and M5. M4 can be profitably
compared to M1 to evaluate the difference between H and CHF2
ligands in a square pyramid. The changes are seen to be mainly
angular: the apical/basal angles. In the M3 f M4 isomerization,
the M-F bond shortens by up to 0.04 Å, indicating more F f
M π donation in unsaturated M4.
The biggest change from M3 to M5 is the shortening of the
M/CFx distance (by about 0.1 Å). This explains why the M/CFx
bond does not lengthen as much as expected on going from
MdC (M3) to MsC (M4): the larger trans influence of hydride
(vs fluoride) makes the MdC bond long in M3. In M5, the
MsF distance trans to CO is shorter than that trans to CHF.
Distances vary by no more than 0.02 Å between Ru5 and
Os5.
The calculated energies of the various products (Figure 3)
also reveal some insights, with the proviso that, by studying
simultaneously two metals from the same periodic group, one
learns by comparison. At the level of theory employed, we seek
trends more than absolute agreement; errors of perhaps 5 kcal/
mol will not obscure the true, underlying periodic trends.
(1) The reaction of MHF(CO)(PH3)2 with H3Si-CF3 is
calculated to be approximately thermoneutral ((4 kcal/mol) for
both M ) Ru and Os. This shows that, despite the very strong
Si-F bond which is formed, the reactants have comparable
stabilities. This suggests that the M-F bond, while synthetically
very useful, is itself quite strong.
(2) The (experimentally unobserved) MH(CF3)(CO)L2 species
show a structure which presages an unexpected feature of the
chemistry reported here: C-F cleavage. The R-agostic F species
is a minimum, and its geometry can lead easily only to a product
with the CF2 carbene ligand trans to H, where subsequent
combination of these ligands is unfavorable.
On isomerizing from M2 to M3, the CF bonds shorten
considerably (0.02-0.05 Å) as the C-F* bond is cleaved, the
resulting CF2 group rotates 90° in order to π-accept from the
dπ orbital which is not donating to CO, and the M/CFx bond
shortens by up to 0.14 Å. The M/CF2 bond is not as short as
the M/CO bond, however. Os3 differs from Ru3 mainly by a
shorter MdC bond and a correspondingly longer MsH bond.
(3a) The full cleavage of the C-F bond, to form M3 from
M2, is much more favorable for M ) Os (-16 kcal/mol) than

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8921
Figure 3. Calculated energies of the speices shown in Figure 2, comparing M ) Ru (solid lines) to M ) Os (dashed lines). For simplicity, PH3
ligands are omitted from the drawings.
for Ru (thermoneutral). This is in agreement with the experi-
mental fact that Ru3 is detectably in equilibrium with a CF3
isomer at 60 °C, while this is not the case for Os3. Osmium
thus shows a greater preference than ruthenium for being
saturated and having more π-acid ligands (i.e., to be more
oxidized).
(3b) Isomer Ru3′, which may be relevant to finally forming
the CHF2 ligand (H has migrated “toward” the carbene), is only
1 kcal/mol higher in energy than Ru3. This shows that H has
an energetically comparable effect whether it is trans to CO or
to CF2.
predicted disparate behavior of these two metals will be revisited
below, after further consideration of the reactivity of 4.
Reaction of 4 with Me3SiCCH. The fast equilibrium of 4
and its five-coordinated isomer RuH(CF3)(CO)L2 indicates that
the two species are energetically similar. Logically, by changing
the ligand environment, the equilibrium might be induced to
favor the unsaturated Ru-CF3 species. One way to alter the
ligands is by addition of Ru-H to a CC multiple bond, which
converts the Ru-H compound to the Ru-C analogue. There-
fore, we tested the reaction of 4 with Me3SiCCH in an attempt
to make the vinyl analogue of 4. Addition of 2 equiv of Me3-
SiCCH to a benzene solution of 4 gives Me3SiCHdCH2
(identified by 1H NMR) and Ru(CF3)(CCSiMe3)(CO)L2, 17, in
5 h at 20 °C. If only 1 equiv of Me3SiCCH is added, partial
conversion to 17 occurs, and no other intermediate is detected.
This observation indicates that Ru(CHdCH(SiMe3))(CF3)(CO)-
L2, which is not observed in the reaction, is more reactive toward
Me3SiCCH and gives 17 with release Me3SiCHdCH2. Reactions
of transition metal vinyl complexes with terminal alkynes have
been reported to give alkenes and alkynyl complexes.30 17 is
highly soluble in common nonpolar solvents and is isolated as
a brown solid from tetramethylsilane. The CF3 fluorine reso-
nance is a triplet (J(PF) ) 12 Hz), and correspondingly, the
phosphine signal is a quartet. The CO stretching frequency is
quite high (1944 cm-1), and the CC triple bond stretching
frequency is normal (2019 cm-1). In contrast to the case for 4,
at room temperature there is no evidence for R-F migration of
17 to form Ru(CCSiMe3)F(CF2)(CO)L2; 17 is not thermally
stable, which excludes a higher temperature search for a possible
CF2 isomer. Apparently, unlike the pure σ-donor hydride, the
alkynyl ligand of 17 can donate its π-electrons to Ru and thus
stabilize the unsaturated metal center in a way unavailable to
the more Lewis acidic [RuH(CF3)(CO)L2]. This result demon-
strates that the equilibrium position for R-F migration is highly
dependent on the Lewis acidity of the metal.
(3c) Isomer Os3′′, where the carbene has migrated “toward”
H, is 6 kcal/mol less stable than Os3. This must be attributed
to competition between the two π-acid ligands CO and CF2 for
back-bonding and the decreased push/pull interaction between
F and CO in Os3′′.
(4) The transformation from M3 to M4 is very favorable for
both metals, although much more so for Ru than for Os. This
metal dependence of ∆E3-4 can be attributed to the same effects
cited above: a greater preference for saturation, and the greater
reducing power of the 5d metal.
(5) Consistent with this logic, the isomerization of Os4 to
Os5 is thermoneutral, while it is unfavorable by 20 kcal/mol
for Ru. Thus, for Ru, any benefit from achieving an 18-valence
electron count must be offset by the diminished reducing power
of this 4d metal: it is less able than osmium to tolerate the
additional π-acid ligand CF2.
In broad strokes, then, the calculations agree with experiment
for ruthenium, and they give a quantitative measure of the
stability of isomers, both observed and unobserved. The
Isomerization of OsH(F)(CF2)(CO)(PtBu2Me)2, 6. In marked
contrast to its Ru counterpart, 4, 6 is persistent in THF at room
temperature for 1 week, even with heat (65 °C, 12 h). To get
an assessment of the difference in M-P bond dissociation
energies for Ru vs Os, we optimized the structure of the product
of eq 6. These reaction energies (after BSSE correction of results
(30) Santos, A.; Lopez, J.; Galan, A.; Gonzalez, J. J.; Tinoco, P.;
Echavarren, A. M. Organometallics 1997, 16, 3482.

8922 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
Scheme 2. (L ) PtBu2Me)
from an MP2 calculation) are large for both metalss39 kcal/
mol for Os and 32 for Rusbut these will certainly be smaller
for bulky PtBu2Me. Given the nearly identical bond lengths for
Ru and Os, due to the lanthanide contraction, this error should
be comparable for the two metals. Therefore, a 7 kcal/mol
increase in M-P bond dissociation energy on going from Ru
to Os is in agreement with the fact that the HsMdCF2 f Ms
CF2H isomerization, which is much slower for Os, involves a
dissociative preequilibrium. The higher reaction energy reflects
a greater reluctance for a 5d metal to be unsaturated compared
to its 4d analogue.
Isomerization can be forced if the solution is heated to higher
temperature (110 °C, in a sealed tube) in either THF or toluene
to give, instead of OsF(CF2H)(CO)L2 (cf. 7), OsF2(dCFH)-
(CO)L2, 18 (Scheme 2). The net change from 6 and 18 is the
exchange of one CF2 fluorine with hydride. 18 is characterized
spectroscopically. The 31P NMR spectrum is a doublet of
doublets with coupling constants with Fb (45 Hz) twice as large
as that with Fa (see Scheme 2 for definitions of Fa and Fb). The
carbene proton exhibits a signal at low field (15.2 ppm) as a
doublet of doublets of doublets due to the coupling with three
gives some unexpected results, indicating that it does not merely
serve as a source of HF. Reaction of RuHF(CF2)(CO)(PtBu2-
Me)2 with NEt3‚3HF (Ru:N molar ratio 10:1) in benzene-d6 for
2 h at room temperature causes partial (30%) conversion of
RuHF(CF2)(CO)(PtBu2Me)2 to free H2 (as evidenced by a singlet
at 4.5 ppm) and Ru(CF3)F(CO)(PtBu2Me)2 (eq 7), along with a
trace amount of Ru(CF2H)F(CO)(PtBu2Me)2. The reaction is
2
different fluorines (large J(FH) (80 Hz), medium J(HFb) (18
Hz), and small J(HFa) (8 Hz)). The proton-decoupled carbene
carbon resonance appears at low field (275 ppm) as a doublet
of doublets of multiplets (not well-resolved coupling with
phosphine and Fa), with large 1J(CF) (374 Hz) and small J(CFb)
(99 Hz). Two equivalent phosphine ligands rule out the static
geometry with the CFH plane parallel to the P-Os-P axis.
The CFH ligand either rotates fast or lies in the plane
perpendicular to the P-Os-P axis (which may give rise to two
rotamers), with fast rotation being more likely. The CO signal
completed by addition of more (1 equiv of HF per Ru in total)
NEt3‚3HF to give Ru(CF3)F(CO)(PtBu2Me)2 as the major
product (90%). The 31P{1H} NMR spectrum of Ru(CF3)F(CO)-
3
(PtBu2Me)2 is a doublet of quartets (2JPF ) 23 Hz, JPF ) 10
Hz), the CF3 fluorides appear as doublet of triplets (JPF ) 10
3
Hz, JFF ) 13 Hz) at 9.5 ppm, and the Ru-bound fluoride
2
is a doublet of triplets of doublets with large J(CFa) (80 Hz).
exhibits an overlapping quartet of triplets at -210 ppm, which
The 19F NMR spectrum shows some interesting features. The
fluorocarbene resonance appears at low field (80.6 ppm) as an
apparent triplet due to the same coupling constants with the
carbene proton and with Fb. The two metal-bound fluorides are
at much higher field (-274 for Fb, -271 for Fa) and strongly
coupled with each other (J(FaFb) ) 121 Hz).
The isomerization of 6 to 18 can also be promoted by water.
Addition of 10% (mole) water to 6 in toluene at room
temperature slowly forms a small amount of OsHF(CO)2L2, 19,
and HF in 3 h. Then 6 is converted to 18 in 24 h, and the HF
also disappears (i.e., reacts with the glass walls). 19 is
independently synthesized by addition of stoichiometric CO to
OsHF(CO)L2. We propose that the formation of the 18 is
catalyzed by acid (HF), which protonates the fluoride of 6 to
form the bifluoride adduct [OsH(FHF)(CF2)(CO)L2], 20 (not
observed). FHF- is a relatively weak ligand31,32 and dissociates
to form a five-coordinated [OsH(CF2)(CO)L2]+, 21. The hydride
and the CF2 now can combine, followed by R-F abstraction.
FHF- coordination and dissociation of HF complete the catalytic
cycle.
is close to the chemical shifts of other five-coordinate ruthenium
3
fluorides, RuR(F)(CO)L2. The small JFF coupling constant
indicates a cis arrangement of CF3 and fluoride. Indeed, in the
13C{1H} NMR spectrum, the CO signal appears as a doublet of
multiplets with a large 2JCF (64 Hz) value, consistent with CO
being trans to fluoride. 5 also reacts with equimolar NEt3‚3HF
to give Ru(CF3)F(CO)(PiPr3)2 as the major product (61%), which
possesses spectra similar to those of its PtBu2Me counterpart.
However, a catalytic (0.05 equiv) amount of NEt3‚3HF does
not cause conversion of 6 to OsF2(CFH)(CO)(PtBu2Me)2, 18;
only 6 is recovered after 24 h at 20 °C.
Ru(CF3)(F)(CO)2(PtBu2Me)2. Addition of excess CO to a
benzene solution of Ru(CF3)F(CO)(PtBu2Me)2 immediately
causes a color change from yellow to colorless. 1H NMR
analysis of the mixture is consistent with a product Ru-
(CF3)F(CO)2(PtBu2Me)2. Two t-Bu proton resonances reveal that
the two CO’s are cis to each other. However, the 31P{1H} NMR
spectrum exhibits an extremely broad peak at 44 ppm, which
sharpens as the temperature is raised to 60 °C. The 19F NMR
spectrum behaves similarly, and no 19F signals were detected
at room temperature, but at 60 °C two broad peaks are located
for CF3 (-16.3 ppm) and Ru-F (-376 ppm). The broadening
of the signals is likely caused by slow rotation of Ru-P bonds
in this sterically crowded molecule. Similar NMR signal
broadening was also noticed in the spectra of RuH(CF3)(CO)2-
(PtBu2Me)2.
Reaction of MHF(CF2)(CO)L2 with NEt3‚3HF. To verify
that HF can, indeed, catalyze the isomerization of MHF(CF2)-
(CO)L2, we tested the reaction of 4-6 in the presence of
catalytic amount of NEt3‚3HF as an HF source. This reagent
(31) Whittlesey, M. K.; Perutz, R. N.; Greener, B.; Moore, M. H. Chem.
Commun. 1997, 187.
(32) Murphy, V. J.; Hascall, T.; Chen, J. Y.; Parkin, G. J. Am. Chem.
Soc. 1996, 118, 7428.
Reaction of 6 with Me3SiOTf. Replacement of the fluoride
by the weaker ligand triflate also causes exchange of the CF2

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8923
Scheme 3
Scheme 4
Table 1. Selected NMR Spectral Data for 24a and 24b
chemical shifts
nuclei
major isomer
minor isomer
OsdCFH
14.2, dd, 2JHF ) 80, 3JHF ) 5.8 15.9, dd, JHF ) 72,
those for 2JPF (21 vs 41 Hz) and 3JFF (<10 vs 82 Hz). Addition
of 1 equivalent more of Me3SiOTf replaces the other fluoride
of 24 to give clean formation of Os(OTf)2(dCFH)(CO)L2, 25.
The proton-decoupled carbene hydrogen peak is a sharp doublet
at room temperature. Even at -70 °C, no rotamers are observed
by NMR spectroscopy (1H, 19F, and 31P), indicating much faster
rotation around this OsdCFH bond.
3JHF ) 15
OsdCFH
117.8, d, JHF ) 79,
3JFF not resolved
102.7, dd, JHF ) 72,
3JFF ) 82
45.3, d, JPF ) 41
178.5, m
Os-PtBu2Me 35.7, d, JPF ) 21
Os-CO
178.7, t, JCP ) 12 Hz
fluoride with hydride. The reaction of 6 with Me3SiOTf is slow
enough to suggest a mechanism for the analogous reaction for
ruthenium (eq 5). Addition of 1 equiv of Me3SiOTf at room
temperature gives, in 10 min, four products (Scheme 3): OsH-
(CF2)(OTf)(CO)L2, 22, Os(CF2H)(OTf)(CO)L2, 23, and two
isomers of OsF(OTf)(CHF)(CO)L2, 24a and 24b. 22 and 23
convert to 24a/b in 3 h. The most characteristic spectral feature
of 22 is the hydride as a virtual triplet of triplets at -2.0 ppm
(JFH ) 85 Hz, JPH ) 27 Hz). The virtual triplet is caused by
coupling between the inequivalent carbene fluorines. The 19F
NMR spectrum of 22 has an ABX (X ) H) pattern for the
carbene fluorides; one peak appears at 107.1 ppm (JFF ) 155
Hz, JFH ) 33 Hz) and the other at 104.5 ppm (JFF ) 155 Hz,
JFH ) 52 Hz). In support of the virtual triplet assignment of
the hydride peak, the sum of JHF (33+52) is the same as JFH
(85 Hz). The geminal F-F coupling constant of 22 is the same
as that of 6 (155 Hz); this strengthens our structural assignment
of 22. The supporting evidence for 23 is the characteristic CF2H
proton at 7.24 ppm (cf. 7.72 ppm of 12) as a triplet of triplets
Reaction of OsHF(dCF2)(CO)L2 with CO. In sharp contrast
to the case for 4, carbon monoxide does not react with 6 at
room temperature, indicative of no fast equilibrium similar to
that between 4 and [RuH(CF3)(CO)L2] (eq 4). With heating to
110 °C over 12 h in toluene-d8, CO converts 6 cleanly to OsF-
(CF2H)(CO)2L2, 26 (Scheme 4). The spectroscopic features of
26 include a triplet of triplets of doublets for the CF2H proton
at low field (7.7 ppm, near that of 7), due to coupling with 19
F
and 31P, with a large (50 Hz) 2J(HF) value. The 13C{1H} NMR
spectrum of the CF2H group is a triplet of doublets of triplets
1
with a large (217 Hz) J(CF). One CO signal appears as a
2
doublet of multiplets with a large J(CF) (70 Hz) splitting,
indicative of a trans F-Os-CO arrangement. Similar to the
case for 8, the 19F resonances of CF2H and the 31P resonance
of the phosphine of 26 are both broad, and no coupling
information can be deduced. The broadening may be caused
by hindered rotation around the P-Os bonds, as is often seen
in similar six-coordinated Ru(II) complexes.27
2
with large JFH (62 Hz). The fluorine resonance of the CF2H
Mechanism of Reaction of 6 with CO. Based on the reaction
of Ru analogue 4, two possible processes may happen on heating
6 in the presence of CO, as shown in Scheme 4. Combination
of CF2 with F could yield five-coordinate OsH(CF3)(CO)L2,
which then binds one CO to give OsH(CF3)(CO)2L2, 27. If the
reaction is monitored after 1 h of heating, a small amount of
27 is formed, along with some 26. The spectroscopic features
of 27 are very much like those of its Ru counterpart 9. Due to
the hindered rotation around the Os-P bond, rotamers inter-
convert slowly at room temperature; the hydride and the fluoride
signals are therefore broad. At 60 °C, the signals sharpen. The
1H NMR spectrum signal of the hydride is a triplet of quartets
at -6.74 ppm, and the 19F NMR spectrum signal is a broad
peak at -2.6 ppm. After longer heating, 27 disappears, and 26
is the only product. 27 can apparently isomerize to form 26.
Alternatively, dissociation of a phosphine from 6, followed by
insertion of CF2 to Os-H and recoordination of lost L, could
provide [Os(CF2H)F(CO)L2], 28 (cf. 7), which would be trapped
by CO to form 26. Or, in the absence of CO, R-F migrates to
generate 18. To test if 18 can be transformed to 26 under CO
group is at -54.1 ppm (JFH ) 61 Hz, JPF ) 16 Hz). These
values are close to that of 12. Consistent with this, the 31P{1H}
NMR spectrum of 23 is a triplet (JPF ) 16 Hz). 22 and 23 are
not long-lived species and convert to 24a/b over 3 h at 20 °C.
The 31P{1H} NMR spectra of 24a and 24b coalesce to one broad
signal at 20 °C, as do the isomeric CFH protons and fluorines.
At low temperature (-60 °C), the signals (31P, 19F of CFH and
1
all H signals) decoalesce to two peaks with a 2:1 ratio. The
NMR data of 24a and 24b are compiled in Table 1. The
geometry in each around Os can be deduced from the lack of
a large JCF coupling between OsF and Os-CO, indicative of
their cis orientation. In addition, two virtual triplets for the tBu
protons suggest two trans phosphines in each. The CO stretching
frequencies of these two isomers are very close (1967 and 1962
cm-1) and relatively high compared to those of 18 (1939 cm-1),
in agreement with CO being trans to a weaker π-donor ligand,
triflate. However, 24a and 24b have significant, large differences
in chemical shifts (10 ppm difference in 31P shifts). The coupling
constants among the nuclei are also very different, especially

8924 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
atmosphere, we heated pure 18 (prepared from thermal isomer-
ization of 6) in toluene under CO (110 °C). To our surprise, a
new product, OsF2[(CHF(PtBu2Me)](CO)2L, 29, was formed
cleanly by a process which does implicate phosphine dissocia-
tion from 18. Since the reaction of 6 with CO did not form 29,
it is unlikely that 18 is generated under these reaction conditions.
Characterization of 29. Multinuclear NMR spectroscopy
provides sufficient information to establish the geometry of 29.
The formation of a P/C bond is an indication of the electrophilic
nature of the CHF ligand of 18. The presence of several
magnetically active nuclei and the lack of symmetry due to a
chiral carbon cause a rather complicated NMR spin system. The
31P{1H} NMR spectrum has two signals, a doublet of doublets
(II) complexes without a π-donor ligand. The CO stretching
frequency (1917 cm-1) of 30 is close to that of RuH(OTf)-
(CO)(PtBu2Me)2 (1921 cm-1), in agreement with C6F5 being a
very strong electron-withdrawing group. Unlike the phenyl
analogue, RuH(Ph)(CO)L2, which reductively eliminates Ph-H
readily upon heating in arene solvent, 30 does not eliminate
C6F5H, even at 100 °C in toluene. The stability of 30 (vs RuH-
(Ph)(CO)L2) can be attributed to a much stronger Ru-C6F5
bond.
Hydrogenolysis of Ru(CF2H)F(CO)L2: Dependence on
Phosphine Ligands. Ru(CF2H)F(CO)(PtBu2Me)2 does not show
any reactivity with H2 at room temperature (or 80 °C) in benzene
over 24 h, as the free H2 peak and the peaks for 7 remain
unchanged. The reaction, however, is promoted by addition of
CsF (in excess, 80 °C, 5 h) to give only a small amount of
t
(Pa, JPF ) 58 and JPP ) 17 Hz) and a multiplet (Pb). Two Bu
groups on each phosphine are no longer equivalent, and
accordingly four doublets are observed. The 19F resonance of
carbon-bound fluorine appears at high field (-209 ppm) as a
doublet of doublets with similar 2JFH (51 Hz) and 2JPF (58 Hz)
splittings. The assignment of the coupling constants is confirmed
by selective 1H or 31P spin decoupling.33 The two metal-bound
fluorides appear at higher field as broad doublets with 2JFF (141
Hz) (close to 121 Hz of 18). Large (83 and 88 Hz) 2JFC values
for the inequivalent carbonyls support the mutually trans
geometry of CO and F. Consistent with this, two CO stretching
bands (1989 and 1911 cm-1) with similar intensity (indicating
a cis arrangement) were found. The resonance of the chiral
carbon appears as a doublet of doublets of multiplets at 91 ppm
with large 1JCF (182 Hz) and 2JPC (82 Hz). This large coupling
clearly reveals the presence of a trans phosphine ligand.
Phosphine attacking a carbene has been observed before.34
Roper and co-workers treated Ru(CH2)(Cl)(NO)(PPh3)2 with the
strong π-acidic ligand CF2dCF2 to form Ru(η2-CF2d
CF2)(Cl)(NO)(CH2PPh3)PPh3, which has a PsC bond.35 π-
Acidic ligands (CO and CF2dCF2) diminish the back-donation
to carbene carbon, which then is susceptible to nucleophilic
attack. Thus, it appears that addition of the π-acid CO to 18
encourages nucleophilic attack by PtBu2Me on the carbene
carbon.
1
CH2F2, which is detected by 19F and H NMR spectra. The
organometallic products are more complicated and include
RuHF(CO)(PtBu2Me)2 (47%), Ru(H)2(CO)2(PtBu2Me)2 (28%),
and RuHF(CO)2(PtBu2Me)2 (8%). We attribute the CO source
(giving dicarbonyl products) to some decomposed Ru com-
plexes. In sharp contrast (eq 9), the PiPr3 analogue 8 reacts
with H2 at room temperature over 12 h to form CH2F2 as a
major organic product, along with a 1:2 ratio of RuHF(CO)-
(PiPr3)2 and RuHF(CO)2(PiPr3)2 as the major organometallic
products. The reaction also generates an acid (HF), which
8
H2, benzene8 CH2F2 + RuHF(CO)(PiPr3)2
(9)
appears (1H NMR) as a broad singlet at 12 ppm and a broad
peak at -188 ppm in 19F NMR spectrum. HF is responsible
for significant broadening of 19F resonances of RuHF(CO)-
(PiPr3)2 and RuHF(CO)2(PiPr3)2 and loss of coupling information
(1H and 31P{1H} NMR) of JHF and JPF. These couplings reappear
after addition of a small amount of CsF, which removes HF
(the peak at 12 ppm disappears in the presence of CsF). The
HF source is apparently the reaction of RuHF(CO)(PiPr3)2 with
excess of H2, together with the “decomposition” reaction, which
liberates CO. Combining this result with eqs 1 and 2, a catalytic
hydrogenation of Me3SiCF3 to CH2F2 can be established in
principle (eq 10).
Synthesis of MH(C6F5)(CO)L2. We attempted to expand the
scope of CsF-catalyzed reactions to the transfer of another Rf
group. While Me3SiC6F5 is not as reactive as Me3SiCF3, heating
(65 °C) of 1 or 3 and Me3SiC6F5 for several hours in THF gives
five-coordinate complexes MH(C6F5)(CO)L2, 30 (M ) Ru) and
31 (M ) Os) (eq 8), in good yield. The hydride signal of 30 is
RuHF(CO)L2
Me3SiCF3 + H2
Discussion
8 Me3SiF + CH2F2 (10)
Transition metal fluoride complexes have been used as
synthetic precursors, in combination with trimethyl silyl reagents
Me3Si-R, for introduction of other functional groups. The
driving force of the reaction is the formation of a strong Si-F
bond.7 Our research group has applied this strategy in synthesis
of formally 16-electron Ru, Os, and Ir complexes.36 The reaction
often proceeds smoothly and quantitatively. To extend this
reaction to a fluorinated R group would open a safe and
convenient route to fluoroalkyl complexes. In our earlier
communication, the reaction of Me3SiCF3 was claimed with
RuHF(CO)L2, 1 (L ) PtBu2Me). We subsequently found that
the sample of 1 used in the reaction was contaminated by trace
CsF from the synthesis of 1 by salt metathesis of RuHCl(CO)-
L2 with CsF in acetone. Indeed, pure 1 does not react with Me3-
SiCF3 under the same conditions (THF, 20 °C). If, however, a
catalytic amount of CsF is added to the pure sample of RuHF-
(CO)L2, the reaction proceeds smoothly. This serendipitous
a triplet of apparent triplets at high field (-27.6 ppm), in
agreement with the hydride being trans to the vacant site. There
are five 19F NMR peaks at room temperature, indicating slow
rotation around the Ru-C6F5 bond, and the C6F5 lies perpen-
1
dicular to the P-Ru-P axis. H and 19F NMR spectra are
unchanged to -60 °C. This observation rules out the possibility
for agostic Ru-F-C bonding from the ortho C-F bond of C6F5.
Therefore, 30 is a rare example of a persistent 16-electron Ru-
(33) Irradiation at the frequency of P (or H) of CHFP causes collapse of
the fluoride signal (-209 ppm) to a doublet.
(34) Burch, R. R.; Calabrese, J. C.; Ittel, S. D. Organometallics 1988,
7, 1642.
(35) Burrell, A. K.; Clark, G. R., Rickard, C. E. F.; Roper, W. R. J.
Chem. Soc., Dalton Trans. 1991, 609.
(36) Cooper, A. C.; Huffman, J. C.; Caulton, K. G. Inorg. Chim. Acta
1998, 270, 261.

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8925
discovery seems to make possible a general route to trifluo-
romethyl complexes (eq 11). We thus screened the reaction with
to steric repulsion of the bulky phosphines and the substituent
of the olefin. However, steric repulsion cannot explain the small
angle of transient species RuHCl(CH2)(CO)(PtBu2Me)2 (2JPP
)
LnM-F + Me3Si-CF3 CsF(cat.)8 LnM-CF3 + Me3Si-F
166 Hz), where CH2 is not large enough to demand that the
phosphines bend away from it.25 4 also has an P-Ru-P which
deviates significantly from linearity. In contrast, the (O)C-
M-X and H-M-L′ angles of all these complexes are close to
180°. On the other hand, when the hydride is replaced by a
π-donating ligand halide, the P-M-P angles are much larger
and closer to 180°. Thus, in OsCl2(dCHCHdCHPh)(CO)(Pi-
Pr3)2, P-M-P is 172°,41 and that of OsCl2(dCHPh)(CO)-
(iPr3)2 is 167.5°.24a Similarly, P-M-P is close to linear in
(11)
other unsaturated Ru/Os fluorides. However, only OsHF(CO)-
L2 shows a positive result. The fluorides Ru(Ph)(F)(CO)L2, RuF-
(NO)L2, and RuF(NO)(CO)L2 are inert to Me3SiCF3 even with
heating (80 °C). Therefore, reaction 11 is valid only for certain
types of fluoride complexes.
Mechanism of Fluoride Ion Catalysis. It has now been
persuasively demonstrated37,38 that catalytic fluoride reacts with
Me3SiCF3 to give trigonal bipyramidal Me3Si(CF3)2- (and Me3-
SiF), and thus this anion is likely to be the CF3- transfer reagent
functioning under our synthetic conditions. A computational
search for a reaction path involving the two unactivated reagents
(M6) was unsuccessful. The lack of reactivity with other metal
42
OsCl2(dCCl2)(CO)(PPh3)2 and RuCl2[dC(F)(OCH2CMe3)]-
(CO)(PPh3)2.43 The distortion of the hydride complex can be
better explained if the electronic rather than steric factors are
considered. The common feature of these complexes is the
presence of a single-faced π-acceptor L′, which orients its
LUMO in-phase with the highest energy metal dπ orbital so
that the back-donation is maximized. When L′ is trans to a strong
trans-directing ligand such as hydride, the M-L′ σ bond is
weakened. This is compensated by bending the P-M-P angle
(away from the L′), thus lowering the energy of the metal dπ
orbital which affects back-donation. The degree of bending is
thus dependent on the π-acidity of the L′. CH2 and CHR are
more π-acidic than CF2; thus, a smaller P-M-P angle is
expected. The P-M-P bond angle of these carbene complexes
is a manifestation of the changing oxidation state of the metal.
When the hydride is replaced by a weaker trans-directing and
π-donating ligand halide, M-L′ σ bonding is strengthened, and
the repulsion between metal dπ and halide p electrons effectively
increases π-donation without bending the P-M-P angle. Six-
coordinate Ru and Os complexes, MH3XL2 and MH2X2L2, all
of which are unsaturated, are known to have nonoctahedral
geometry, and the electronic origin has studied both theoretically
and experimentally.24b
fluoride complexes mentioned earlier suggests that the metal
-
centers are not sufficiently Lewis acidic for CF3 (not a
particularly strong nucleophile) to attack. Indeed, even the much
stronger nucleophile MeLi reacts with RuPhF(CO)L2 only
slowly (over 1 week at room temperature). The 16e Ru(0)
complex Ru(NO)(F)L2 adopts a square planar geometry, with
the LUMO well protected by the four ligands and not readily
accessible for nucleophilic attack. Also, RuF(NO)(CO)L2 is
known to be a weak Lewis acid.39
One question arises from this mechanism: which F- dis-
sociates in eq 1, that from Ru-F, or that from CF3? It seems
more likely that the Ru-bound fluoride is a better leaving group
than the carbon-bound one. If the Ru-F bond breaks, the
primary product is five-coordinated [RuH(CF3)(CO)L2] with CF3
trans to hydride, in which R-F migration gives the observed
isolated product. Dissociation of F- from CF3 would form the
product directly.
The Origin of Distortion of Six-Coordinate Octahedral
Complexes. It is a recurring phenomenon that saturated six-
coordinated “octahedral” Ru and Os complexes with the
geometry shown below have a remarkable distortion of one trans
pair of ligands away from a 180° angle, but the reason is not
well understood. Werner and co-workers reported a 141.38(4)°
r-F Migration: Effects of Metals and Ligands. One well-
known reaction of CF3 complexes is that in which, in the
presence of Lewis acids as F- scavenger, the CF3 ligand is
transformed to a difluorocarbene. In the compounds discussed
here, the unsaturated metal acts as an internal Lewis acid, and
the abstracted fluoride becomes coordinated cis to CF2, thus
providing a pathway for subsequently re-forming the CF3 group.
The thermodynamic preference between MCF3 and FsMdCF2
is highly dependent on the identity of metal and of the ancillary
ligands, which directly affect the Lewis acidity of the metal.
For example, while RuH(CF3)(CO)L2 spontaneously isomerizes
to RuHF(CF2)(CO)L2, the similar compound Ru(CCSiMe3)-
(CF3)(CO)L2 is persistent in solution. The π-donor alkynyl
ligand in the latter apparently decreases the Lewis acidity of
the Ru. Whereas 4 is in fast equilibrium with RuH(CF3)(CO)-
L2 at room temperature, the Os analogue 6 is static (based on
the fact that there is no observable fast fluoride exchange of
Os-F and CF2 by 19F NMR and based on its slow reaction
with CO). Nonetheless, at high temperature (100 °C), the
reaction of 6 with CO forms OsH(CF3)(CO)2L2, which indicates
OsH(CF3)(CO)L2 as a possible intermediate. Both of these
observations are consistent with the DFT calculations (Figure
3). Being a 5d metal, Os is a stronger π-donor than Ru, and
CF2 is consequently less electrophilic.
P-Os-P of complex OsHCl(dCH(SiMe3)(CO)(PiPr3)2 but
provided no explanation.24 Esteruelas and colleagues discovered
a case with P-Os-P of 144° in OsH(OH)(η2-CH2dCHs
OC(O)CH3)(CO)(PiPr3)2.40 The small angle is retained in
solution because 2JPP (165 Hz) is small compared to the normal
250 Hz for a linear P-M-P group. The authors attributed this
(37) Kolomeitsev, A.; Bissky, G.; Lork, E.; Movchun, V.; Rusanov, E.;
Kirsch, P.; Ro¨schenthaler, G.-V. Chem. Commun. 1999, 1017.
(38) Maggiarosa, N.; Tyrra, W.; Naumann, D.; Kirij, N. V.; Yagupolskii,
Y. L. Angew. Chem., Int. Ed. 1999, 38, 2252.
(41) Esteruelas, M. A.; Lahoz, F. J.; On˜ate, E.; Oro, L. A.; Zeier, B.
Organometallics 1994, 13, 1662.
(39) Ogasawara, M.; Huang, D.; Streib, W. E.; Huffman, J. C.; Caulton,
K. G. J. Am. Chem. Soc. 1997, 119, 8642.
(40) Edwards, A. J.; Elipe, S.; Esteruelas, M. A.; Lahoz, F. J.; Oro, L.
A.; Valero, C. Organometallics 1997, 16, 3828.
(42) Clark, G. R.; Marsden, K.; Roper, W. R.; Wright, L. J. J. Am. Chem.
Soc. 1980, 102, 1206.
(43) Hoskins, S. V.; Pauptit, R. A.; Roper, W. R.; Waters, J. A. J.
Organomet. Chem. 1984, 269, C55.

8926 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
Any general statement that C-F bonds are “strong” and
unreactive certainly requires modification for F geminal with a
transition metal, and this need not be an unsaturated metal.
Heterolytic C-F splitting44,45 is facilitated both kinetically and
thermodynamically because of product stabilization by C/M
double bond formation. It follows that this should be promoted
most by LnM moieties which are good π bases (eq 12),
and this is certainly true of the octahedral d6 species studied
here. The fact that our unsaturated target molecules MH(CF3)-
7 is the favored form for Ru, saturated 18 is the more stable
form for Os. An analogous contrasting thermodynamic prefer-
ence has been rationalized47 for the analogous pairs A and B
on the basis of the 5d metal preferring (1) a higher oxidation
number, (2) an 18-valence electron count, and (3) more metal
ligand bonds. Similar reasoning applies to 7 vs 18. The
(CO)L2 (M ) Ru, Os) need not pay the energetic price of full
charge separation in eq 12 (i.e., F- remains coordinated to M)
is only an added benefit of an already attractive rearrangement.
Rotation of the Fluorocarbene. The orientation of fluoro-
carbene, a single-face π-acceptor, is determined by the several
highest occupied metal dπ orbitals. In all fluorocarbene com-
plexes reported here, the ground state has the carbene pπ orbital
perpendicular to the P-M-P axis, with rotation rates highly
dependent on the other ligands. For example, in 6, two carbene
fluoride signals have coalesced at 20 °C. When the Os-F is
replaced by triflate (22, Scheme 3), two relatively sharp carbene
fluoride resonances are observed at room temperature. The
observation of two different fluoride peaks for 22 is not due to
the larger difference in the chemical shifts. In fact, the chemical
shift difference of the two fluorides of 22 is smaller (4 ppm)
than that observed for 6 (10 ppm). There is no significant steric
difference between 6 and 22. Although CF3SO3 is much larger
than fluoride, it lies away from the CF2 group. However, there
is a large difference in the electronic influence of these two
ligands. Fluoride, being a good π-donor ligand, has filled-filled
repulsion between the p electron lone pairs and dπ electrons
(the lower two dπ orbitals) of Os, which increases the energy
level of the dπ. As a consequence, the higher energy geometry
(with the CF2 plane rotated 90° as shown in eq 13) gains more
stabilization from π-back-donation. This leads to a lower rotation
preference for 18 over 6 differs in the substructure C vs D,
which may result from the unfavorable trans influence in C and
the favorable push/pull influence in D. Metal/ligand interactions
must dominate the thermodynamics here since all evidence
suggests C-F bonds are stronger than C-H bonds, which, taken
alone, should favor C, in contrast to our observations. The CHF
carbene should be less of a π-acid than CF2 due to greater F f
C π-donation in the latter.
R-F abstraction occurs not only in unsaturated transition metal
CF3 complexes; it is a rather general phenomenon reported for
several main group fluoromethyl complexes. The difference for
the late transition metal complexes is that the carbene is
stabilized by coordination to the metal. In the case of a metal
that has no π-donation ability but is Lewis acidic, the resulting
carbene fragment subsequently undergoes oligomerization.
Earlier attempts to synthesize CF3Li and CF3MgI produced only
LiF and MgF2 by F abstraction.48 Eujen and Hoge demonstrated
that solvent-free Cd(CF3)2 (more acidic than Cd(CF3)2(DME))
readily decomposes above -50 °C to give CdF2 and (CF2)n (n
) 2, 3).49 Also related is the unsuccessful earlier attempt to
synthesize (Rf)3B (Rf ) perfluoroalkyl). The products are BF3
and uncharacterized polymers.50 FCH2BF2 also decomposes at
room temperature to give BF3.51 While the thermodynamic
driving force for F abstraction for these main group elements
result from the high M-F bond energies, for the late transition
metals the driving force could be that the metal gains more
electrons (from 16 to 18) and the carbene is stabilized by the
π-basic metal.
barrier for the carbene ligand. The much weaker ligand triflate
does not have significant filled-filled repulsion between oxygen
lone pairs and the Os dπ electrons, and so the CF2 rotation barrier
is higher. A similar explanation also accounts for the rotation
of the OsdCFH bonds of 24 and 25. 24 has fluorine trans to
carbene, while the carbene of 25 is trans to triflate. The energy
gap is larger between the high-energy dπ-orbital and the lower
energy ones in 24. An extreme situation for the large rotation
barrier for single-faced π-acid ligand occurs when it is bound
to a d2 metal.46
19F Chemical Shifts of Ru and Os Complexes. 19F NMR
spectra have been the most informative tool in characterizing
the complexes. The one advantage of this tool comes from the
high sensitivity of the 19F chemical shift, which spanned a wide
range from +550 (XeF6) to -500 ppm.52 While the chemical
shift moves downfield as the F content increases (-271.9
(CH3F), -143.6 (CH2F2), -78.6 (CHF3), and -62.3 ppm (CF4)),
Figure 4 depicts the range of ligand chemical shifts, which
shows the following: (1) MdCF2 fluoride appears at low field,
OsdCFH moves to a little higher field, and they are both shifted
Contrasting 4d/5d Behavior. Comparative studies of analo-
gous reactions for ruthenium and osmium species MHF(CO)L2
have revealed results which, we feel, are quite surprising: redox
isomeric forms are preferred for Ru and Os. While unsaturated
(47) Spivak, G. J.; Coalter, J. N.; Oliva´n, M.; Eisenstein, O.; Caulton,
K. G. Organometallics 1998, 17, 999.
(44) Reger, D. L.; Dukes, M. D. J. Organomet. Chem. 1978, 153, 67.
(45) Clark, G. R.; Hoskins, S. V.; Roper, W. R. J. Organomet. Chem.
1982, 234, C9.
(46) Antin˜olo, A.; Carrillo-Hermosilla, F.; Fajardo, M.; Garcia-Yuste,
S.; Otero, A.; Camanyes, S.; Maseras, F.; Moreno, M.; Lledos, A.; Lluch,
J. M. J. Am. Chem. Soc. 1997, 119, 6107.
(48) (a) Emele´us, H. J.; Haszeldine, R. N. J. Chem. Soc. 1949, 2948.
(b) Haszeldine, R. N. J. Chem. Soc. 1954, 1273.
(49) Eujen, R.; Hoge, B. J. Organomet. Chem. 1995, 503, C51.
(50) Parson, T. D.; Baker, E. D.; Burg, A. B.; Juvinall, G. L. J. Am.
Chem. Soc. 1961, 83, 250.
(51) Goubeau, J.; Rohwedder, K. H. Ann. Chem. 1957, 604, 168.

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8927
3), even though these are not analogous products (an unsaturated
F-Ru-CF2H species vs a saturated F2Os(CHF) species).
(2) Replacing H on Ru by a weak π-donor ligand CCSiMe3
speeds the conversion to a unsaturated CF3 complex (eq 6).
(3) The reagents CO, Me3NC, MeCN, and perhaps Me3-
SiCCH supplement spin saturation transfer in proving eq 4: R-F
migration wherein saturated F-RudCF2 is in equilibrium with
an unsaturated, but less stable Ru-CF3 isomer.
(4) Reaction of 6 with CO to give CF2H species suggests
that R-F and/or R-H migration to the carbene does occur for
Os, but more slowly than for Ru. However, in contrast to the
case for Ru, the Os-CO bond energy is necessary to make such
a reaction exergonic.
One perhaps surprising result of these studies is that the
conversion of eq 14 is very favorable. How can this be
rationalized? This can be analyzed as the conversion of M-H
to M-F and of C-F to C-H. While the C-F bond dissociation
energy can be 40 kcal/mol stronger than that of a C-H bond,
this is probably not true of F on a carbon attached to a metal.
Moreover, in general, M-F is stronger than M-H, especially
in an unsaturated molecule, and especially when CO is present
to create a push/pull stabilization. The transformation of eq 15
is perhaps equally surprising, but the above several factors
apply equally well here, and agree with the observed exother-
micity.
Figure 4. 19F chemical shift of M(dRu/Os) fluorocarbyl and fluoride.
The chemical shifts are referenced to CFCl3 (0 ppm) as an external
standard.
downfield in comparison to vinylic CF2 and CHF units. This
trend is comparable to that of the low-field shift of carbene
protons, which move from a normal value of 5-6 ppm to around
20 ppm. (2) Chemical shifts of M-CF3 are in a relatively narrow
range from 18.6 to -13.7 ppm, and those of M-CF2H move
to higher field (-52 to -95 ppm). They are also downfield in
comparison to HCF3 and CH2F2. (3) Five-coordinate Ru/Os
complexes of MRF(CO)L2 have lower chemical shifts (-184.5
to -238.9 ppm) than the six-coordinate ones (-270 to -391.3
ppm), each of which falls in a broad (>100 ppm) range. The
upfield shift for saturated (six-coordinate) M-F relative to that
of five-coordinate ones suggests that the former have a more
electron-rich fluoride. This is further supported by the fact that
(C2H5)2O‚BF3 also has a higher 19F chemical shift (-153 ppm)
than that of BF3 (-131 ppm).52
Experimental Section
All reactions and manipulations were conducted using standard
Schlenk and glovebox techniques. Solvents were dried and distilled
under argon and stored in airtight solvent bulbs with Teflon closures.
Most reagents are commercially available except for NMe4F, which
was obtained by dehydration of NMe4F hydrate.54 RuHF(CO)(PtBu2-
Me)2,55 RuHCl(CO)(PiPr3)2,56 and OsHCl(CO)(PtBu2Me)2 were prepared
according to the literature. All NMR solvents were dried, vacuum-
transferred, and stored in an argon-filled glovebox. H, 31P, 19F, and
1
13C NMR spectra were recorded on a Varian Gem XL300 or a Unity
I400 spectrometer. Chemical shifts are reported in ppm and referenced
to residual solvent peaks (1H, 13C), external H3PO4 (31P), or external
CFCl3 (19F). Infrared spectra were recorded on a Nicolet 510P FT-IR
spectrometer. Elemental analyses was performed on a Perkin-Elmer
2400 CHNS/O elemental analyzer at Indiana University.
Conclusions
We have demonstrated that, in the presence of fluoride as
catalyst, Me3SiCF3 reacts with the unsaturated Ru(II)/Os(II)
fluoride, MHF(CO)L2, to form difluorocarbenes, which are
kinetic products. These isomerize to give different thermody-
namic products, unsaturated RuF(CF2H)(CO)L2 or saturated
OsF2(CFH)(CO)L2. These are the first examples of reversible
R-F and (irreversible) double R-F migration. Addition of a
Brønsted acid has been shown to trigger the conversion of
HMCF3 to MCF2H on Rh and on Ru.53
The work reported here also permits some additional conclu-
sions:
(1) Replacing F on Ru and Os in 4 and 6 speeds the
conversion to the thermodynamic product (eq 5 and Scheme
RuHF(CF2)(CO)(PtBu2Me)2. RuHF(CO)(PtBu2Me)2 (2.0 g, 4.2
mmol) was dissolved in THF (20 mL). Me3SiCF3 (0.63 mL, 4.6 mmol)
and CsF (30 mg, 0.2 mmol) were then added. The color changed from
orange to yellow in 30 min. The volatiles were removed, and the residue
was dissolved in diethyl ether (10 mL) and cooled to -40 °C for 1 h.
Pale yellow crystals formed which were filtered, washed with cold
diethyl ether, and dried. Yield: 1.2 g (54%). Anal. Calcd for C20H43F3-
OP2Ru: C, 46.23; H, 8.34. Found: C, 46.77; H, 8.23. 1H NMR (C6D6,
20 °C): 1.38 (vt, N ) 7 Hz, 6H, PCH3), 1.24 (vt, N ) 13.2 Hz, 18H,
3
PC(CH3)3), 1.17 (vt, N ) 13.5 Hz, 18H, PC(CH3)3), -3.04 (ttd, JHF
) 51 Hz, 2JHP ) 23 Hz, 2JHF ) 7.5 Hz). 31P{1H} NMR (C6D6, 20 °C):
68.9 (dt, 3JPF ) 13 Hz, 2JPF ) 18 Hz). 19F NMR (C6D6, 20 °C): -354
(54) Christe, K. O.; Wilson, W. W.; Wilson, R. D.; Bau, R.; Feng, J. J.
Am. Chem. Soc. 1990, 112, 7619.
(55) Poulton, J. T.; Sigalas, M. P.; Eisenstein, O.; Caulton, K. G. Inorg.
Chem. 1993, 32, 5490.
(52) Harris, R. K.; Mann, B. E. NMR and Periodic Table; Academic
Press: New York, 1978; p 99.
(53) Burrell, A. K.; Clark, G. R.; Jeffrey, J. G.; Rickard, C. E. F.; Roper,
W. R. J. Organomet. Chem. 1990, 388, 391.
(56) Esteruelas, M. A.; Werner, H. J. Organomet. Chem. 1986, 303, 221.

8928 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
Table 2. Crystallographic Parameters of
RuHF(CF2)(CO)(PtBu2Me)2 (4)
35.5 (vt, N ) 17 Hz, PC(CH3)3), 35.3 (vt, N ) 16 Hz, PC(CH3)3), 30.3
(vt, N ) 4.9 Hz, PC(CH3)3), 29.5 (vt, N ) 5 Hz, PC(CH3)3), 6.5 (vtt,
N ) 10.4 Hz, 4JFC ) 3 Hz, PCH3) ppm. IR (C6D6): νCO ) 1917 cm-1
.
formula
color
a
b
c
V
Z
C20H43F3OP2Ru
pale yellow
15.662(2) Å
22.825(3) Å
14.085(1) Å
5035.29 Å3
8
fw
519.57
Pbca
-172 °C
0.71069 Å
1.371 g cm-1
7.7 cm-1
0.039
RuF(CF2H)(CO)(PiPr3)2. Solid RuHF(CF2)(CO)(PiPr3)2 was trans-
space group
T
formed to RuF(CF2H)(CO)(PiPr3)2 on standing at room temperature for
1
6 months. H NMR (THF-d8, 20 °C): 8.05 (td, JFH ) 60 Hz, JHF ) 5
λ
Fcalc
µ
Hz, 1H RuCF2H), 2.52 (m, 6H, PCH(CH3)2), 1.32 (18H, PH(CH3)2),
1.28 (m, 18H, PCH(CH3)2). 31P{1H} NMR (THF-d8, 20 °C): 46.2 (dvt,
N ) 40 Hz). 19F NMR (THF-d8, 20 °C): -52.0 (dvt, JHF ) 60 Hz, N
) 40 Hz, CF2H), -238.9 (br, w1/2 ) 77 Hz, Ru-F). IR (THF): ν(CO)
R (Fo)
Rw (Fo)
0.046
) 1917 cm-1
.
Ru(CF2H)(OTf)(CO)(PtBu2Me)2. To RuHF(CF2)(CO)(PtBu2Me)2
(200 mg, 0.38 mmol) dissolved in toluene (5 mL) and cooled to -78
°C was added Me3SiOTf (74.5 µL, 0.38 mmol). After the solution was
stirred for 15 min and warmed to room temperature, the volatiles were
removed in vacuo, and the residue was recrystallized from toluene
layered with pentane at -40 °C. Yield: 130 mg (55%). Anal. Calcd
for C21H43F5O4P2Ru: C, 40.84; H, 7.02. Found: C, 40.53; H, 6.60. 1H
NMR (C6D6, 20 °C, 300 MHz): 7.72 (tt, 2JFH ) 60 Hz, 3JPH ) 3.3 Hz,
1H, CF2H), 1.56 (vt, N ) 5.7 Hz, 6H, PCH3), 1.05 (vt, N ) 13.5 Hz,
18H, PC(CH3)3), 0.89 (vt, N ) 12.9 Hz, 18H, PC(CH3)3). 19F NMR
(ν1/2 ) 50 Hz, Ru-F), 122 (ν1/2 ) 300 Hz, CF2). The CF2 signal
decoalesces to an AMX (X ) H) pattern at -80 °C. At this temperature,
two rotamers with a ratio of 9:1 can be seen in 19F NMR: for the
2
major rotamer, 128 (3JHF ) 37 Hz, JFF ) 221 Hz), 114.8 (3JHF ) 46
2
2
Hz, JFF ) 221 Hz); for the minor rotamer, 131.0 (3JHF ) 44 Hz, JFF
) 216 Hz), 119.4 (3JHF ) 54 Hz, 2JFF ) 216 Hz). 13C{1H} NMR (C6D6,
20 °C): 254.4 (tm, 1JCF ) 498 Hz, CF2), 204.2 (dt, 2JFC ) 70 Hz, 2JPC
) 12 Hz, CO), 36.5 (vt, N ) 18 Hz, PC(CH3)3), 36.3 (vt, N ) 21.6
Hz, PC(CH3)3), 29.81 (m, PC(CH3)3), 29.06 (m, PC(CH3)3), 5.75 (dvt,
3JFC ) 10 Hz, N ) 22 Hz, PCH3). IR (C6D6): νCO ) 1937 cm-1. X-ray
crystallographic parameters are given in Table 2.
2
3
(C6D6, 20 °C, 282 MHz): -54.9 (dt, JHF ) 60 Hz, JPF ) 17.5 Hz,
CF2H), -79.4 (s, O3SCF3).13C{1H} NMR (toluene-d8, -20 °C, 75
Crystal Structure of RuHF(CF2)(CO)(PtBu2Me)2. One of the
crystals, grown from diethyl ether solution layered with pentane at -40
°C, was attached to a glass fiber with silicone grease and transferred
to the goniostat, where it was cooled to -172 °C for characterization
and data collection. Systematic extinctions uniquely identified the space
group as Pbca (No.61). No correction for absorption was carried out
(µ(Mo KR) ) 7.79 cm-1). The structure was solved by a combination
of direct methods (MULTAN-78) and Fourier techniques. When all of
the non-hydrogen atoms had been refined using anisotropic thermal
parameters, the largest peak in the difference map was 0.90e/A3 in a
position that was suitable for the expected hydride atom, H*. The peak
was introduced as H*, and the coordinates and thermal parameters were
allowed to vary. The final difference map was essentially featureless,
the largest peak being 0.58 e/A3 and the deepest hole 0.57 e/A3. The
crystallographic parameters are available as Supporting Information.
RuHF(CO)(PiPr3)2. RuHCl(CO)(PiPr3)2 (1.0 g, 2.0 mmol) and CsF
(1.0 g, 6.5 mmol) were mixed in acetone (30 mL) and stirred at room
temperature for 12 h. After evaporation of all the volatiles, the residue
was extracted with pentane and filtered through a Celite pad. The filtrate
was concentrated to ca. 10 mL and cooled to -40 °C, for 12 h. The
orange precipitate was filtered, washed with pentane, and dried.
Yield: 0.8 g, 85%. Anal. Calcd for C19H43FOP2Ru: C, 48.58; H, 9.23.
Found: C, 48.96; H, 9.53. 1H NMR (C6D6, 20 °C): 2.30 (m, 6H,
CH(Me)2), 1.25(dvt, 3JHH ) 4.8 Hz, NPH ) 13.8 Hz, 18H, PCH(CH3),
1.20(dvt, JHH ) 4.8 Hz, N ) 13.8 Hz, 18H, PCH(CH3)., -23.7(td, JHF
) 3 Hz, JPH ) 14.1 Hz, 1H Ru-H). 31P{1H} NMR (C6D6, 20 °C):
59.1 (d, JPF ) 16 Hz).
2
3
1
MHz): 204.6 (tt, JPC ) 16 Hz, JCF ) 7.6 Hz, CO), 129 (tt, JCF
)
3
1
309 Hz, JPC ) 6.3 Hz, CF2H). 120.7 (quartet, JPF ) 318 Hz, O3-
SCF3), 36.6 (vt, N ) 19 Hz, PC(CH3)3), 36.0 (vt, N ) 19 Hz, PC(CH3)3),
29.4 (s, PC(CH3)3), 28.9 (s, PC(CH3)3), 4.78 (vt, N ) 18 Hz, PCH3).
31P{1H} NMR (C6D6, 20 °C, 121 MHz): 47.3 (vt, N ) 34 Hz). IR
(C6D6): νCO ) 1944 cm-1
.
RuH(CF3)(CO)2(PtBu2Me)2. RuHF(CF2)(CO)(PtBu2Me)2 (100 mg,
0.19 mmol) was dissolved in diethyl ether (5 mL). The solution was
freeze-pump-thaw degassed and then stirred under 1 atm of CO gas.
The solution color changed immediately to pale yellow. After the
solution was cooled to -40 °C for several hours, white crystals formed.
These were filtered, washed with diethyl ether, and dried in vacuo.
Yield: 85 mg (82%). Anal. Calcd for C21H43F3O2P2Ru: C, 46.06; H,
7.91. Found: C, 46.26; H, 7.87. 1H NMR (toluene-d8, 60 °C, 300
MHz): 1.40 (vt, N ) 5.7 Hz, 6H, PCH3), 1.23 (vt, N ) 12.9 Hz, 18H,
PC(CH3)3), 1.22 (vt, N ) 13.2 Hz, 18H, PC(CH3)3), -6.51 (tq, 2JPH
)
23.4 Hz, 3JFH ) 6.6 Hz, 1H, Ru-H). 19F NMR (toluene-d8, 60 °C, 300
MHz): -1.39 (broad, CF3). 31P{1H} NMR (toluene-d8, 60 °C, 121
MHz): 63.5. 13C{1H} NMR (toluene-d8, 60 °C, 75 MHz): 10.65 (br,
s PCH3), 30.2, 30.4 (s, PtBu), 36.9, 37.1 (vt, N ) 15 Hz, PC(CH3)3),
148.0 (qt, 1JCF ) 358.6 Hz, 3JPC ) 12 Hz, CF3), 203.7 (t, 2JPC ) 6 Hz,
2
3
CO cis to CF3), 205.8 (tq, JPC ) 14 Hz, JCF ) 15 Hz, CO trans to
CF3). IR (C6D6): 2029, 1927 (ν(CO)).
RuH(CF3)(CO)(13CO)(PtBu2Me)2. When 13CO was used in the
above experiment, RuH(CF3)(CO)(13CO)(PtBu2Me)2 was obtained. 1H
NMR(toluene-d8, 60 °C, 300 MHz): 1.39 (vt, N ) 6 Hz, 6H, PCH3),
1.22 (vt, N ) 12.9 Hz, 18H, PtBu), 1.21 (vt, N ) 13.2 Hz, 18H, PtBu),
RuHF(CF2)(CO)(PiPr3)2. The same procedure as for the synthesis
of RuHF(CF2)(CO)(PtBu2Me)2 was used. Yield: 60%. 1H NMR (C6D6,
20 °C, 300 MHz): 2.34 (m, 6H, PCH3), 1.20 (dvt, JHF ) 6.8 Hz, N )
12.4 Hz, 18H, PCH(CH3)), 1.17 (dvt, JHH ) 7.2 Hz, N ) 12.4 Hz,
18H, PCH(CH3), -3.22 (ttd, 3JHF ) 49.2 Hz, JPH ) 20.4 Hz, 2JHF ) 8
2
2
3
-6.51(dtq, JCH ) 29 Hz, JPH ) 25 Hz, JFH ) 6.6 Hz, 1H, Ru-H).
13C NMR (toluene-d8, 60 °C, 75 MHz): 203.8 (doublet of apparent
sextet, JCH ) 29 Hz, JFC ) JPC ) 6 Hz, Ru-13CO). The 13C{1H}
NMR resonance for the CO carbon showed only an apparent sextet.
This proves that the 29 Hz coupling constant is from the trans hydride.
There are two rotamers27 of this product due to the hindered rotation
of the Ru-P bond; at room temperature they interconvert slowly,
causing broadening of the NMR signals and loss of spin coupling. At
higher temperatures (>60 °C), the rotation is faster and the two isomers
coalesce; therefore, the NMR signals become sharper. This problem
can be avoided if PiPr3 is used in place of PtBu2Me. Thus, the NMR
spectrum of RuH(CF3)(CO)(13CO)(PiPr3)2 (prepared by the same method
as the PtBu2Me analogue) at 25 °C in C6D6 shows sharp signals. 13C
2
3
2
2
Hz, Ru-H). 31P{1H} NMR (C6D6, 20 °C): 69.9 (dt, JPF ) 21.9 Hz,
3JPF ) 10.9 Hz). 19F NMR (C6D6, 20 °C): 119.7 (br, V1/2 ) 175.3 Hz,
2F, CF2), -391.7 (br, V1/2 ) 49 Hz, 1F, Ru-F).
Ru(CF2H)(F)(CO)(PtBu2Me)2. RuHF(CF2)(CO)(PtBu2Me)2 (500
mg, 1 mmol) was dissolved in THF (5 mL), and the solution was kept
at room temperature for 6 h. Evaporation of THF left a yellow solid
which was recrystallized from diethyl ether to afford 370 mg of yellow
crystals (74%). Anal. Calcd for C20H43F3OP2Ru: C, 46.23; H, 8.34.
Found: C, 46.56; H, 8.17. 1H NMR (THF-d8, 20 °C, 300 MHz): 8.21
NMR (C6D6, 75 MHz, 20 °C): 204 (dtq, 2JCH ) 28 Hz, 2JPC ) 3JFC
)
2
3
(td, JHF ) 59 Hz, JFH ) 5 Hz, 1H, CF2H), 1.38 (vt, N ) 13.2 Hz,
18H, PC(CH3)3), 1.33 (vt, N ) 12.3 Hz, PC(CH3)3), 1.34 (overlapping
with the signal of PC(CH3)3, PCH3). 31P{1H} NMR (THF-d8, 121
1
2
6.4 Hz, Ru-CO). H NMR (C6D6, 300 MHz, 20 °C): -6.1 (dtq, JCH
) 28 Hz, 2JPH ) 21 Hz, 3JFH ) 7 Hz). Again, the large coupling constant
(28 Hz) of hydride to the 13CO supports the trans disposition of the
these ligands.
2
MHz): 44.5 (overlapping doublet of virtual triplets, N ) 40 Hz, JPF
) 20 Hz). 19F NMR (THF-d8, 20 °C, 282 MHz): -236 (tm, 2JPF ) 20
2
3
3
Hz, Ru-F), -52 (dtd, JFH ) 59 Hz, JPF ) 17 Hz, JFF ) 6.3 Hz,
Ru(CF3)(CCSiMe3)(CO)(PtBu2Me)2. RuH(F)(CF2)(CO)(PtBu2Me)2
(0.30 g, 0.58 mmol) was dissolved in C6H6 (5 mL). To the yellow
solution was added Me3SiCCH (160 µL, 1.2 mmol). The solution color
CF2H). 13C{1H} NMR (THF-d8, 20 °C): 205.1 (dm, JCF ) 75 Hz,
CO), 140.3 (ttd, JCF ) 293 Hz, JPC ) 7.5 Hz, JCF ) 1 Hz, CF2H),
2
1
2
2

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8929
changed immediately to orange. After the solution was stirred for 30
min, the volatiles were removed, and the residue was dissolved in
tetramethylsilane and filtered through a Celite pad. The filtrate was
concentrated to ca. 2 mL and cooled at -40 °C. Orange crystals were
obtained after 2 days. 1H NMR (C6D6, 20 °C, δ): 0.37 (s, 9H, SiMe3),
1.08 (vt, N ) 12.6 Hz, 18H, PtBu), 1.15 (vt, N ) 12.6 Hz, 18H, PtBu),
1.70 (vt, N ) 6.3 Hz, 6H, PMe). 19F NMR (C6D6, 20 °C, δ): 18.6 (t,
changed from bright orange to pale yellow. The volatiles were stripped
off, and the residue was extracted with diethyl ether and filtered. The
filtrate was concentrated to ca. 2 mL and cooled to -40 °C for 4 days.
White crystals formed. These were separated, washed with cold diethyl
ether, and dried (100 mg (60%)). Anal. Calcd for C20H43F3OP2Os: C,
1
39.51; 7.13. Found: C, 39.55; H, 7.56. H NMR (toluene-d8, 20 °C,
300 MHz): 1.46 (vt, N ) 7.5 Hz, 6H, PCH3), 1.23 (vt, N ) 13.5 Hz,
18H, PC(CH3)3), 1.19 (vt, N ) 13.5 Hz, 18H, PC(CH3)3), -2.46 (ttd,
JPF ) 11.8 Hz, CF3). 31P{1H} NMR (C6D6, 20 °C, δ): 45.75 (q, JPF
)
11.8 Hz, PtBu2Me).
3JHF ) 44.7 Hz, JPH ) 26.4 Hz, JHF ) 12 Hz, 1H, Os-H). 31P{1H}
2
2
NMR (toluene-d8, 20 °C): 41.0 (dt, 2JPF ) 26 Hz, 3JPF ) 2.4 Hz). 19
F
RuH(CF3)(CNCH3)(CO)(PtBu2Me)2. RuHF(CF2)(CO)(PtBu2Me)2
(10 mg, 0.019 mmol) was dissolved in C6D6 (0.5 mL), and CNCH3 (1
µL) was added to the solution. The yellow color of the solution faded
immediately. Spectroscopic analysis revealed clean formation of RuH-
(CF3)(CNCH3)(CO)(PtBu2Me)2. 1H NMR (400 MHz, 20 °C): 2.34 (s,
3H, CNCH3), 1.54 (vt, N ) 5.3 Hz, 6H, PCH3), 1.34 (vt, N ) 12.3 Hz,
PC(CH3)3), 1.31 (vt, N ) 12.3 Hz, PC(CH3)3), -7.90 (tq, JPH ) 26
Hz, JFH ) 6.3 Hz, 1H, Ru-H). 31P{1H} NMR (162 MHz, 20 °C): 64.3
(br). 19F NMR (376 MHz, 20 °C): -13.7 (doublet of broad triplets,
J(FH) ) 6H, Ru-CF3).
NMR (toluene-d8, 20 °C, 282 MHz): -361.0 (V1/2 ) 113.8 Hz); CF2
resonances are coalesced to baseline. At 50 °C, a broad peak is seen at
94.7 (V1/2 ) 759 Hz). At -60 °C, the peak decoalesces to two AX
peaks: one, at δ 104.9, is a doublet of doublets (2JFF ) 155 Hz, JHF
3
) 35 Hz), and the other is a doublet of doublets of doublets (2JFF
)
155 Hz, 3JFH ) 51 Hz, 3JFF ) 28 Hz) centered at δ 85.3. Although no
definitive assignment of these two peaks is possible, the latter peak
shows coupling with (Os)-F, while the former does not. It is therefore
likely that the coupling is through space and that (C)-F, which shows
the coupling with (Os)-F, is proximal. 13C{1H} NMR (toluene-d8, 20
RuH(CF3)(NCCD3)(CO)(PtBu2Me)2. RuHF(CF2)(CO)(PtBu2Me)2
was dissolved in CD3CN (0.5 mL). NMR spectroscopic data revealed
clean formation of RuH(CF3)(NCCD3)(CO)(PtBu2Me)2. 1H NMR (300
MHz, 20 °C): 1.35 (N ) 12.4 Hz, PC(CH3)3), 1.32 (vt, N ) 12.5 Hz,
PC(CH3)3), -9.73 (tq, JPH ) 22 Hz, JFH ) 19 Hz, Ru-H). 31P{1H}
(121 MHz, 20 °C): 58.7 (q, JPF ) 6 Hz, Ru-P). 19F NMR (282 MHz,
20 °C): (dq, JFH ) 19 Hz, JPF ) 6 Hz, Ru-CF3).
NMR Monitoring of the Reaction of NMe4F with RuHF(CF2)-
(CO)(PtBu2Me)2. RuHF(CF2)(CO)(PtBu2Me)2 (10 mg, 0.019 mmol)
and anhydrous NMe4F (7.0 mg, 0.086 mmol) were mixed in THF-d8
(0.5 mL) in an NMR tube. Fifteen minutes later, the light yellow
heterogeneous solution was analyzed, revealing ca. 85% formation of
[RuH(CF3)(F)(CO)(PtBu2Me)2]NMe4. 1H NMR (300 MHz, 20 °C): -8.5
(apparent sextet of doublets, 3JPH ) 3JFH ) 20 Hz, 2JFH ) 5.4 Hz, 1H,
Ru-H), 1.33 (vt, N ) 14.4 Hz, PC(CH3)3, 1.38 (vt, N ) 14.4 Hz,
PC(CH3)3) (PCH3 peak is covered by the tBu proton resonances), 3.31
(s, 12H, NMe4). 31P{1H} NMR (121 MHz): 21 (dq, 2JPF ) 16 Hz, 3JPF
) 6H, Ru-P). 19F NMR (282 MHz, 20 °C): -325 (m, Ru-F), -10.5
(doublet of apparent quartets, JHF ) 21 Hz, JPF ) JFF ) 6 Hz, Ru-
CF3). IR (THF-d8): 1910 (ν(CO)).
1
2
2
°C, 100 MHz): 235.3 (tdt, JCF ) 447 Hz, JCF ) 13 Hz, JPC ) 4.5
2
2
Hz. CF2), 182.2 (dt, JFC ) 72 Hz, JPC ) 7.5 Hz, CO), 36.6 (vt, N )
20 Hz, PCMe3) 36.3 (vt, N ) 20 Hz, PCMe3), 29.8 (s, PC(CH3)3, 29.6
s, PC(CH3)3), 5.75 (dvt, JCF ) 8.4 Hz, N ) 29 Hz, PCH3).
OsF2(dCHF)(CO)(PtBu2Me)2. In a sealed NMR tube, OsHF-
(CF2)(CO)(PtBu2Me)2 (10 mg, 0.018 mmol) was dissolved in toluene-
d8 and heated at 100 °C for 8 h to cleanly give OsF2(dCHF)(CO)-
1
(PtBu2Me)2 (see 18 for F labeling). H NMR (toluene-d8, 20 °C, 300
2
3
3
MHz): 15.2 (dd, JFH ) 80 Hz, JFbH ) 18.4 Hz, JFaH ) 8.3 Hz, 1H,
(Os)dCHF), 1.41 (vt, N )7.7 Hz, PCH3), 1.22 (vt, N ) 13.5 Hz, 18H,
PC(CH3)3), 1.19 (vt, N ) 13.5 Hz, 18H, PC(CH3)3). 31P{1H} NMR
(toluene-d8, 20 °C): 31.4 (dd, JFbP ) 43 Hz, J(FaP) ) 20 Hz). 13C{1H}
OsHF(CO)(PtBu2Me)2. OsHCl(CO)(PtBu2Me)2 (300 mg, 0.52 mmol)
and CsF (240 mg, 1.6 mmol) were mixed in acetone (20 mL) and stirred
at room temperature for 2 h. The solution was filtered, and the solid
on the frit was washed with acetone. The combined filtrate and washings
were evaporated in vacuo, and the residue was dissolved in pentane
and filtered to remove a small amount of insoluble white solid. The
filtrate was cooled at -40 °C for 3 h before orange crystals were grown.
Yield: 0.21 g (69%). Anal. Calcd for C19H43FOP2Os: C, 40.86; H,
1
NMR (toluene-d8, 20 °C): 275.5 (ddtd, JFC ) 364 Hz, J(FbC) ) 81
3
2
Hz, JPC ) 17 Hz, JFaC ) 9 Hz, OsdCHF), 181.7 (dtd, J(FaC) ) 80
Hz, J(PC) ) 7.5 Hz, J(FbC) ) 2.8 Hz, Os-CO), 37.0 (vt, N ) 11 Hz,
PCMe3), 36.0 (vt, N ) 11 Hz, PCMe3), 29.8 (s, PC(CH3)3), 29.4 (s,
2
2
PC(CH3)3), 3.54 (vt of dd, N ) 27 Hz, JCF ) 5.8 Hz, JCF ) 3.6 Hz,
PCH3). 19F{1H} NMR (toluene-d8, 20 °C): 80.6 (dm, J(FFb) ) 80 Hz,
(Os)dCHF), -274.2(ddt, J(FbFa) ) 121 Hz, J(PFa) ) 20 Hz), J(FFa)
) 12 Hz,Os-Fb), -270.7(ddt, J(FaFb) ) 121 Hz, J(FFb) ) 79 Hz,
J(PFb) ) 45 Hz, Os-Fb) IR (C6D6): 1937 (ν(CO)). The 31P{1H} NMR
spectrum broadens only slightly at -70 °C, and so does the carbene
proton signal (with loss of spin coupling to Fa and Fb), indicative of
fast rotation around the OsdCFH bond even at low temperature.
Reaction of OsHF(CF2)(CO)(PtBu2Me)2 with Water. OsHF-
(CF2)(CO)(PtBu2Me)2 (30 mg, 0.049 mmol) was dissolved in C6D6 (0.5
mL) in an NMR tube. Water (0.1 µL) was added. Ten minutes after
the addition, there was not much change in the NMR signals of OsHF-
(CF2)(CO)(PtBu2Me)2, and the water appeared at 1.45 ppm as a broad
singlet. Three hours later, the water was consumed, and OsHF(CO)2-
(PtBu2Me)2 was formed along with HF. 1H NMR (300 MHz, 20 °C) of
HF: 10.8 (d, JHF ) 445 Hz). 19F NMR of HF: -187.4 (d, JFH ) 445
Hz).57 At this moment, starting material dominates, although a small
amount of OsF2(CFH)(CO)(PtBu2Me)2 is also formed. After 24 h at
room temperature, OsF2(CFH)(CO)(PtBu2Me)2 and OsHF(CO)2(PtBu2-
Me)2 are the only phosphine-containing products.
1
7.76. Found: C, 40.90; H, 7.75. H NMR (C6D6, 20 °C, 300 MHz):
1.34 (vt, N ) 5 Hz, 6H, PCH3), 1.24 (vt, N ) 12.6 Hz, 18H, PC-
(CH3)3), 1.241 (vt, N ) 14 Hz, 18H, PC(CH3)3), -32.32 (dt, JHF ) 9.3
Hz, JPH ) 14 Hz, 1H, Os-H). 31P{1H} NMR (C6D6, 20 °C, 121
MHz): 44.5 (d, JPF ) 26 Hz). 19F NMR: -184.5 (dt, JHF ) 9 Hz, JPF
) 26 Hz). IR (C6D6, cm-1): 1877 (νCO).
OsHF(CO)2(PtBu2Me)2. OsHF(CO)(PtBu2Me)2 (30 mg, 0.059 mmol)
was dissolved in toluene-d8 (0.5 mL) and degassed. The solution was
charged with CO (1 atm). Within the time of mixing, the color faded.
1
NMR analysis revealed clean formation of OsHF(CO)(PtBu2Me)2. H
NMR (C7D8, 20 °C, 400 MHz): 1.50 (vt, N ) 6.6 Hz, 6H, PCH3),
1.30 (vt, N ) 13.5 Hz, 18H, PC(CH3)3), 1.26 (vt, N ) 12.8 Hz, 18H,
PC(CH3)3), -2.36 (dt, JHF ) 13 Hz, JHP ) 21 Hz, Os-H). 31P{1H}
NMR (C7D8, 162 MHz, 20 °C): 35.7 (d, JPF ) 30 Hz). 19F NMR (C7D8,
376 MHz, 20 °C): -391.3 (JFH ) 12 Hz, JPF ) 30 Hz). 13C{1H} NMR
(C7D8, 75 MHz, 20 °C): 189.7 (dt, JCF ) 11.5 Hz, JPC ) 5.7 Hz, CO,
cis to F), 183.6 (dt, JFC ) 68 Hz, JPC ) 7 Hz, CO, trans to F), 35.6
(vtd, N ) 23 Hz, JCF ) 6.7 Hz, PC(CH3)3), 29.7 (s, PC(CH3)3), 29.4
(s, PC(CH3)3), 5.4 (vtd, JCF ) 7.4 Hz, N ) 27 Hz, PCH3). IR (C7D8):
1969.6, 1896.3 (ν(CO)).
OsF2(CFH(PtBu2Me))(CO)2(PtBu2Me). OsF2(CFH)(CO)(PtBu2Me)2
(30 mg, 0.047 mmol) was dissolved in toluene-d8 (0.5 mL). The solution
was freeze-pump-thawed three times and charged with 1 atm of CO.
The mixture was heated in a 110 °C oil bath for 24 h. NMR
spectroscopic analysis revealed clean formation of OsF2(CFH(PtBu2-
OsHF(CF2)(CO)(PtBu2Me)2. OsHF(CO)(PtBu2Me)2 (150 mg, 0.27
mmol) was dissolved in fluorobenzene (5 mL). To the mixture was
added CsF (2.8 mg, 0.018 mmol), followed by Me3SiCF3 (42 µL, 0.27
mmol). The mixture was stirred at 80 °C for 30 min. The solution color
(57) The scalar coupling constant for HF in the gas phase is measured
as 529 Hz: Muenter, J. S.; Klemperer, W. J. Chem. Phys. 1970, 52, 6033.

8930 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Huang et al.
Me))(CO)2(PtBu2Me). The solvent was removed, and the residue was
washed with pentane and dried to give a white powder. Anal. Calcd
for C21H43F3O2P2Os: C, 39.60; H, 6.81. Found: C, 40.00; H, 6.47. 1H
8.2 Hz, 6H, PCH3), 1.03 (vt, N ) 12 Hz, PC(CH3)3), 0.86 (vt, N ) 13
Hz, PC(CH3)3). 31P{1H} NMR (-60 °C): 47.2 (d, JPF ) 44 Hz).
19F NMR (376 MHz, -60 °C): 86.3 (dd, JFH ) 72 Hz, JFF ) 75 Hz,
OsdCFH), -285.6 (m, overlapping with that of the major isomer, Os-
F).
2
NMR (toluene-d8, 20 °C, 300 MHz): 7.21 (dm, JFH ) 47 Hz, 1H,
CHF(PtBu2Me)), 1.68 (d, J ) 13.8 Hz, 3H, PCH3), 1.48 (d, J ) 8.1
Hz, 3H, PCH3), 1.38 (d, J ) 12.6 Hz, 9H, PC(CH3)3), 1.34 (d, J )
12.6 Hz, 9H, PC(CH3)3), 1.24 (d, J ) 14.4 Hz, 9H, PC(CH3)3), 0.80
(d, J ) 13.2 Hz, 9H, PC(CH3)3). 31P{1H} NMR (toluene-d8, 20 °C,
Os(dCFH)(OTf)2(CO)(PtBu2Me)2. In an NMR tube, OsHF-
(CF2)(CO)(PtBu2Me)2 (20 mg, 0.036 mmol) was dissolved in C6D6 (0.5
mL). To the solution was added Me3SiOTf (12.8 µL, 0.072 mmol),
and the mixture was kept at room temperature for 3 h. NMR
spectroscopic analysis revealed clean formation of Os(dCFH)(OTf)2-
2
121 MHz): 49.9 (dd, JPF ) 58 Hz, JPP ) 17 Hz, CHF(PtBu2Me),
24.3 (m, Os-PtBu2Me). 19F NMR (toluene-d8, 376 MHz, 20 °C): -209.2
(dd, 2JHF ) 48 Hz, 2JPF ) 58 Hz, CHF(PtBu2Me)), -274.0 (dm, 2JFF
)
1
(CO)(PtBu2Me)2. H NMR (20 °C, 300 MHz): 15.6 (d, JFH ) 76 Hz,
2
142 Hz, Os-F), -279 (dm, JFF ) 134 Hz, Os-F). 13C{1H} NMR
1H, OsdCFH), 1.83 (vt, N )7.9 Hz, 6H, PCH3), 1.13 (vt, N ) 14.3
Hz, 18H, PC(CH3)3), 0.97 (vt, N ) 14.3 Hz, 18H, PC(CH3)3). 31P{1H}
NMR (20 °C): 45.4 (s). 19F NMR (20 °C): -79.0 (V1/2 ) 15 Hz, s).
RuH(C6F5)(CO)(PtBu2Me)2. RuHF(CO)(PtBu2Me)2 (0.30 g, 0.64
mmol), Me3SiC6F5 (134 µL, 0.70 mmol), and CsF (5 mg, 0.032 mmol)
were mixed in THF (10 mL) and refluxed for 12 h. The solvent was
removed in vacuo and the residue extracted with toluene and filtered.
The filtrate was concentrated to 3 mL and cooled to -40 °C for 3
days. Bright yellow needles formed, which were filtered, washed with
cooled pentane, and dried. Yield: 230 mg (58%). Anal. Calcd for
C25H43F5OP2Ru: C, 48.61; H, 7.02. Found: C, 48.99; H, 6.90. 1H NMR
(C6D6, 20 °C, 300 MHz): 1.06 (vt, N ) 13 Hz, 36H, PC(CH3)3), 0.538
(vt, N ) 6 Hz, 6H, PCH3), -27.6 (tt, JPH ) 19 Hz, JFH ) 5.4 Hz, 1H,
Ru-H). 31P{1H} NMR: 55.2 (s). 19F NMR: -111.1 (dm, J ) 33 Hz,
ortho F), -120.8 (dm, J ) 35 Hz, ortho F), -163.2 (m, meta F),
-163.4 (t, J ) 20 Hz, para F), -165.3 (m, meta F). IR (C7D8): 1917
(ν(CO)).
OsH(C6F5)(CO)(PtBu2Me)2. The same procedure as above was used,
but the reaction required 24 h for completion. The complex has been
synthesized from OsH(Ph)(CO)L2 and C6F5H, and the spectroscopic
data are published.58
Reaction of RuHF(CF2)(CO)(PtBu2Me)2 with NEt3‚3HF. In an
NMR tube, RuHF(CF2)(CO)L2 (10 mg, 0.020 mmol) was dissolved in
toluene-d8 (0.5 mL). To the solution was added NEt3‚3HF (1.0 µL,
0.02 mmol) via syringe. The color changed gradually from colorless
to yellow over a period of 2 h at room temperature. NMR spectral
analysis reveals consumption of starting material and formation of free
H2 (4.50 ppm, s), Ru(CHF2)F(CO)L2 (10%), and RuF(CF3)(CO)L2
(90%).
2
(toluene-d8, 20 °C, 100.6 MHz): 185.0 (d, JCF ) 83 Hz, Os-CO),
2
1
182.8 (d, JCF ) 88 Hz, Os-CO), 94.0 (ddm, JCF ) 184 Hz, JPC
)
112 Hz, Os-CHF(tBu2Me)), 36.2 (d, JPC ) 17 Hz, PC(CH3)3), 34.7
1
1
(d, JPC ) 32 Hz, PC(CH3)3), 34.3 (d, JPC ) 36 Hz, PC(CH3)3), 4.51
(d, JPC ) 25.8 Hz), 4.51 (m, PCH3). IR (toluene-d8, cm-1): 1990, 1912
(ν(CO)).
OsF(CF2H)(CO)2(PtBu2Me)2. OsHF(CF2)(CO)(PtBu2Me)2 (20 mg,
0.036 mmol) was dissolved in toluene-d8 (0.5 mL), and the solution
was degassed, charged with CO (1 atm), and heated for 12 h. NMR
spectroscopic analysis revealed clean formation of OsF(CF2H)(CO)2-
(PtBu2Me)2. 1H NMR (20 °C): 7.70 (ttd, 2JFH ) 50 Hz, JPH ) 5.3 Hz,
3JFH ) 3.3 Hz, Os-CF2H), 1.47, (vt, N ) 6.3 Hz, PCH3), 1.27, vt, N
) 13.2 Hz, PC(CH3)3), 1.19 (N ) 13 Hz, PC(CH3)3). 19F NMR (20
°C): -95.0 (br, V1/2 ) 336 Hz, 2F, CF2H), -361.5 (t, JPF ) 22 Hz,
Os-F). 13C{1H} NMR (20 °C): 187.7 (m, COb), 183.0 (dm, JCF ) 70
1
2
Hz, COa), 133.3 (tdt, JCF ) 271 Hz, JFC ) 16.3 Hz, JPC ) 7.7 Hz,
Os-CF2H), 37.7 (vt, N ) 22 Hz, PCMe3), 37.6 (vt, N ) 22 Hz,
PCMe3), 30.2 (s, PC(CH3)3), 30.1 (s, PC(CH3)3), 6.09 (vt, N ) 26 Hz,
PCH3). 31P{1H} NMR (20 °C): 16.0 (br).
OsH(CF3)(CO)2(PtBu2Me)2. When the reaction of OsHF(CF2)(CO)-
(PtBu2Me)2 and CO was monitored after 1 h of heating, a small amount
of OsH(CF3)(CO)2(PtBu2Me)2 was observed, along with some OsF-
(CF2H)(CO)2(PtBu2Me)2. After longer heating, the CF3 complex disap-
peared, and only OsF(CF2H)(CO)2(PtBu2Me)2 remained. 1H NMR
(toluene-d8, 300 MHz, 60 °C): -6.73 (tq, JPH ) 24 Hz, JFH ) 7.5 Hz,
Os-H), 1.51 (vt, N ) 6.6 Hz, PCH3). The tert-butyl protons are
overlapping with those of OsHF(CF2)(CO)(PtBu2Me)2 and are not
assigned. 31P{1H} NMR (60 °C, 121 MHz): 26 (br). 19F NMR (60
°C): -2.88 (br). The spectral features are similar to those of the Ru
analogue, RuH(CF3)(CO)2(PtBu2Me)2.
NMR data for RuF(CF3)(CO)L2 follow. 1H NMR (300 MHz): 1.27
(vt, N ) 12.9 Hz, PtBu), 1.19 (vt, N ) 4.8 Hz), 6H PCH3), 1.13 (vt, N
) 12.6 Hz, 18H, PtBu). 31P{1H} NMR (121 MHz): 39.3 (q, JPF ) 9
Hz). 19F NMR (282 MHz): 9.50 (t, JPF ) 10 Hz). However, the metal-
bound fluoride is not observed. To gain information on the missing
fluoride, we added CsF (ca. 10 mg) to the mixture, and the mixture
was stirred at room temperature for 2 h. NMR spectral analysis of the
solution shows the presence of coordinated fluorides. 31P{1H} NMR:
Os(CFH)(F)(OTf)(CO)(PtBu2Me)2. In an NMR tube, OsHF-
(CF2)(CO)(PtBu2Me)2 (20 mg, 0.036 mmol) was dissolved in toluene-
d8 (0.5 mL). To the solution was added Me3SiOTf (6.4 µL 0.036 mmol).
After 30 min, the 31P{1H}NMR spectra revealed three products: OsH-
(OTf)(CF2)(CO)(PtBu2Me)2, Os(CF2H)(OTf)(CO)(PtBu2Me)2, and OsF-
(CHF)(OTf)-(CO)(PtBu2Me)2. After 3 h at room temperature, OsF-
(CHF)(OTf)(CO)(PtBu2Me)2 was the only product.
2
3
3
40.9 (dq, JPF ) 23 Hz, JPF ) 10 Hz), 19F NMR: 9.50 (dt, JFF ) 13
1
Spectroscopic data for OsH(OTf)(CF2)(CO)(PtBu2Me)2 follow. H
3
Hz, JPF ) 10 Hz), -210 (tq, JFP ) 23 Hz, JFF ) 13 Hz, Ru-F). No
NMR: -1.96 (virtual triplet of triplets, N ) 84 Hz, JPH ) 27 Hz, Os-
H). 13P{1H} NMR: 46.1 (s). 19F NMR: 107.1 (dd, JFF ) 155 Hz, JHF
) 33 Hz, OsdCF2), 104.4 (dd, JFF ) 155 Hz, JFH ) 52 Hz, OsdCF2),
-76.8 (s, CF3SO3).
significant change in the 1H NMR spectrum is observed. 13C{1H} NMR
(100 MHz): 205.0 (dm, JCF ) 65 Hz, Ru-CO), 137.5 (qt, 1JCF ) 354
Hz, 2JPC ) 9.5 Hz, CF3), 35.8 (vt, N ) 16 Hz, PC(CH3)3), 35.0 (vt, N
) 17 Hz, PC(CH3)3, 29.8 (vt, N ) 4 Hz, PC(CH3)3), 29.5 (brt, PC-
1
Spectroscopic data for Os(CF2H)(OTf)(CO)(PtBu2Me)2 follow. H
3
(CH3)3), 5.93 (vtd, N ) 21 Hz, JCF ) 1.5 Hz, PCH3).
NMR: 7.26 (tt, JHF ) 62.4 Hz, JPH ) 3 Hz, Os-CF2H). 19F NMR:
-54.1 (dt, JFH ) 61 Hz, JPF ) 16 Hz, Os-CF2H), -76.6 (s, CF3SO3).
31P{1H} NMR: 32.7 (t, JPF ) 16 Hz).
RuF(CF3)(CO)2(PtBu2Me)2. The above solution was charged with
CO (1 atm), and the color changed from yellow to colorless im-
mediately. 1H NMR (400 MHz, 60 °C): 1.47 (vt, N ) 6 Hz, 6H, PCH3),
1
Spectroscopic data for OsF(CHF)(OTf)(CO)(PtBu2Me)2 follow. H
1.32 (vt, N ) 14 Hz, 18H), 1.24 (vt, N ) 12 Hz, 18H, PC(CH3)3). 19
F
NMR (20 °C): 15.2 (brd, JFH ) 87 Hz, 1H, OsdCFH), 1.60 (vt, N )
7.7 Hz, 6H, PCH3), 1.11 (vt, N ) 13 Hz, 18H, PC(CH3)3), 1.04 (vt, N
NMR (376 MHz, 60 °C): -16.3 (br, w1/2 ) 158 Hz, CF3), -376 (br,
w1/2 ) 59 Hz, Ru-F). 31P{1H} NMR (162 MHz, 60 °C): 44.0 (br, w1/2
) 93.5 Hz).
) 13 Hz, 18H, PC(CH3)3). 31P{1H} NMR (20 °C): 41.8 (br, V1/2
)
436 Hz, Os-P).
Reaction of RuHF(CF2)(CO)(PiPr3)2 with NEt3‚3HF. To a benzene
solution of RuHF(CF2)(CO)(PiPr3)2 (10 mg, 0.02 mmol) was added
NEt3‚3HF (1 µL). After 2 h at room temperature, 31P{1H} and 1H NMR
spectra of the solution reveal formation of Ru(CF2H)F(CO)(PiPr3)2
(10%) and Ru(CF3)F(CO)(PiPr3)2 (61%) and a small amount of
unknown products that contain phosphine ligands.
1
NMR data of the major isomer follow. H NMR (-60 °C): 14.2
2
3
(dd, JFH ) 80 Hz, JFH ) 5.8 Hz, OsdCFH), 1.60 (br, 6H, PCH3),
1.15 (br, 18H, PC(CH3)3), 0.72 (br, 18H, PCCH3)3). 31P{1H} NMR (-60
°C): 37.2 (d, JPF ) 19 Hz, Os-PtBu2Me). 19F NMR (376 MHz, -60
°C): 101.4 (d, JFH ) 79 Hz, OsdCFH), -75.8 (s, CF3SO3), -285.6
(br, Os-F).
Spectroscopic data for the minor isomer follow. 1H NMR (-60
°C): 15.9 (dd, 2JFH ) 72 Hz, 3JFH ) 15 Hz, OsdCFH), 1.41 (vt, N )
(58) Renkema, K. B.; Bosque, R.; Streib, W. E.; Maseras, F.; Eisenstein,
O. J. Am. Chem. Soc. 1999, 121, 10895.

Facile and ReVersible CleaVage of C-F Bonds
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 8931
Spectral data for Ru(CF3)F(CO)(PiPr3)2 follow. 31P{1H} NMR (162
set used was a generalized basis consisting of LANL2DZ on Ru and P
and D95* 65 on everything else, plus an f-type polarization function
with an exponent of 0.400 on the Ru. The basis set LANL2DZ is Los
Alamos ECP plus double-ú valence on Ru and P.66 The ECP covers
the 1s, 2s, and 2p electrons, while all others are treated explicitly. To
make the calculations feasible, the PtBu2Me groups were replaced with
PH3, and the methyl groups in the silicon species were replaced with
H. All calculations were performed with Cs symmetry. It was necessary
to change the integration grid from its default size to minimize the
grid noise so that geometry optimizations could converge. The new
grid has 99 radial points and 434 angular Lebedev points.
It was found to be absolutely necessary to include polarization
functions on each center to get the bond lengths and energetics correct.
The most profound effect is on the fluorines. Without polarization
functions, the bond length is overestimated by several hundredths of
an angstrom. In a separate study of diatomic CF, the same trend was
observed. A plot of the difference of the B3LYP density between the
basis set with polarization functions and the basis set without them
clearly shows that inclusion of polarization functions leads to a
significant buildup of electron density between the centers leading to
stronger bonding characteristics and shorter bond length. The inclusion
of polarization functions yields a bond length and Dc much closer to
the experimental value:
MHz): 45.0 (q, JPF ) 10.5 Hz). 19F NMR (376 MHz): 9.0 (t, JPF
)
1
10.5 Hz). H NMR (400 MHz): 2.43 (m, 6H, PCH), 1.14 (m, 36H,
PCH(CH3)2).
Reaction of Ru(CF2H)(F)(CO)(PtBu2Me)2 with H2 in the Presence
of CsF. Ru(CF2H)F(CO)(PtBu2Me)2 (10 mg, 0.02 mmol) was dissolved
in benzene-d6. The solution was degassed, charged with H2 (1 atm),
1
and stirred at room temperature for 24 h. The H NMR spectrum of
the mixture reveals unchanged starting material and a sharp free H2
peak. The solution was then heated for 4 h in 80 °C oil bath. 1H NMR
assay reveals no observable reaction. To the tube was added ca. 10 mg
1
of CsF, and the heterogeneous mixture was heated for 5 h. H, 31P,
and 19F NMR spectra show consumption of Ru(CF2H)F(CO)(PtBu2-
Me)2 and formation of CH2F2, RuHF(CO)L2 (47%), Ru(H)2(CO)2L2
(28%), and RuHF(CO)2(PtBu2Me)2 (8%). 19F NMR (282 MHz): -143
(2JHF ) 50 Hz, CH2F2).
Reaction of Ru(CF2H)F(CO)(PiPr3)2 with H2. In an NMR tube,
Ru(CF2H)F(CO)(PiPr3)2 (10 mg, 0.02 mmol) was dissolved in benzene-
d6 and the solution was freeze-pump-thawed three times before H2
1
(1 atm) was introduced. After 10 min the H NMR spectrum of the
mixture does not show any reaction, as the H2 remains free (4.5 ppm).
The tube was tumbled for 24 h at room temperature; 31P{1H} NMR
spectral analysis shows the disappearance of Ru(CF2H)F(CO)(PiPr3)2
and formation, mainly, of RuHF(CO)2(PiPr3)2 and RuHF(CO)(PiPr3)2.
The 31P{1H} NMR spectrum of the latter, however, shows only a singlet
at 59 ppm in the reaction mixture as there is HF formed in the reaction.
Dc (kcal/mol)
bond length (Å)
without pol. (D95)
with pol. (D95*)
experimental67
110.2
128.8
132.6
1.3597
1.2932
1.2718
1
The H NMR (400 MHz): 12 (br, HF), 4.7 (t, JHF ) 50 Hz, CH2F2),
4.04 (td, JPH ) 19 Hz, JHF ) 6 Hz, RuH of RuHF(CO)2(PiPr3)2), -23.4
(t, JPF ) 18 Hz, Ru-H of RuHF(CO)(PiPr3)2). 19F NMR (376 MHz):
-143.0 (t, JHF ) 50 Hz, CH2F2), -188 (br, HF). 19F peaks for RuHF-
(CO)2(PiPr3)2 and RuHF(CO)(PiPr3)2 are very broad but were sharpened
by addition of CsF (ca. 5 mg) to the reaction mixture.
Acknowledgment. This work was supported by the Petro-
leum Research Fund, administered by the American Chemical
Society. We also thank Johnson Matthey/Aesar for material
support.
Computational Details. All calculations were performed with the
Gaussian 98 package.59 The level of theory used was B3LYP,60 which
is a hybrid functional consisting of Becke’s exchange,61 Slater’s
exchange,62 exact Hartree-Fock exchange, VWN correlation,63 and the
nonlocal (gradient) part of the LYP correlation64 functionals. The basis
Supporting Information Available: Full crystallographic
details, positional and thermal parameters, and bond distances
and angles (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
(59) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.;
Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.6;
Gaussian, Inc.: Pittsburgh, PA, 1998.
JA001646U
(62) Hohenberg, P.; Kohn, W. Phys. ReV. B 1964, 136, 864. Kohn, W.;
Sham, L. J. Phys. ReV. A 1965, 140, 1133. Slater, J. C. In Quantum Theory
of Molecular and Solids. Vol. 4: The Self-Consistent Field for Molecular
and Solids; McGraw-Hill: New York, 1974.
(63) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(64) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. Miehlich,
B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200.
(65) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum: New York, 1976; pp 1-28.
(66) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. Wadt, W.
R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. Hay, P. J.; Wadt, W. R. J.
Chem. Phys. 1985, 82, 299.
(60) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(61) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
(67) Carrington, A.; Howard, B. J. Mol. Phys. 1970, 18, 225.