Carbon-fluorine bond activation coupled with carbon-hydrogen bond formation α to iridium: Kinetics, mechanism, and diastereoselectivity

Article abstract of DOI:10.1021/ja0545012

Reactions of iridium(fluoroalkyl)hydride complexes Cp*Ir(PMe ₃)(CF₂R_F)Y (R_F = F, CF₃; Y = H, D) with LutHX (Lut = 2,6-dimethylpyridine; X = Cl, I) results in C-F activation coupled with hydride migration to give Cp*Ir(PMe ₃)(CYFR_F)X as variable mixtures of diastereomers. Solution conformations and relative diastereomer configurations of the products have been determined by ¹⁹F{¹H}HOESY NMR to be (S_C, S_Ir)(R_C, R_Ir) for the kinetic diastereomer and (R_C, S_Ir)(S_C, R_Ir) for its thermodynamic counterpart. Isotope labeling experiments using LutDCl/Cp*Ir(PMe₃)(CF₂R_F)H and Cp*Ir(PMe₃)(CF₂R_F)D/LutHCl) showed that, unlike a previously studied system, H/D exchange is faster than protonation of the α-CF bond, giving an identical mixture of product isotopologues from both reaction mixtures. The kinetic rate law shows a first-order dependence on the concentration of iridium substrate, but a half-order dependence on that of LutHCl; this is interpreted to mean that LutHCl dissociates to give HCl as the active protic source for C-F bond activation. Detailed kinetic studies are reported, which demonstrate that lack of complete diastereoselectivity is not a function of the C-F bond activation/H migration steps but that a cationic intermediate plays a double role in loss of diastereoselectivity; the intermediate can undergo epimerization at iridium before being trapped by halide and can also catalyze the epimerization of kinetic diastereomer product to thermodynamic product. A detailed mechanism is proposed and simulations performed to fit the kinetic data.

Full text of DOI:10.1021/ja0545012

Published on Web 10/06/2005
Carbon-Fluorine Bond Activation Coupled with
Carbon-Hydrogen Bond Formation r to Iridium: Kinetics,
Mechanism, and Diastereoselectivity
Shaun A. Garratt, Russell P. Hughes,* Ivan Kovacik, Antony J. Ward,
Stefan Willemsen, and Donghui Zhang
Contribution from the Department of Chemistry, 6128 Burke Laboratory, Dartmouth College,
HanoVer, New Hampshire 03755
Received July 7, 2005; E-mail: rph@dartmouth.edu
Abstract: Reactions of iridium(fluoroalkyl)hydride complexes Cp*Ir(PMe3)(CF2RF)Y (RF ) F, CF3; Y ) H,
D) with LutHX (Lut ) 2,6-dimethylpyridine; X ) Cl, I) results in C-F activation coupled with hydride migration
to give Cp*Ir(PMe3)(CYFRF)X as variable mixtures of diastereomers. Solution conformations and relative
diastereomer configurations of the products have been determined by 19F{1H}HOESY NMR to be (SC,
SIr)(RC, RIr) for the kinetic diastereomer and (RC, SIr)(SC, RIr) for its thermodynamic counterpart. Isotope
labeling experiments using LutDCl/Cp*Ir(PMe3)(CF2RF)H and Cp*Ir(PMe3)(CF2RF)D/LutHCl) showed that,
unlike a previously studied system, H/D exchange is faster than protonation of the R-CF bond, giving an
identical mixture of product isotopologues from both reaction mixtures. The kinetic rate law shows a first-
order dependence on the concentration of iridium substrate, but a half-order dependence on that of LutHCl;
this is interpreted to mean that LutHCl dissociates to give HCl as the active protic source for C-F bond
activation. Detailed kinetic studies are reported, which demonstrate that lack of complete diastereoselectivity
is not a function of the C-F bond activation/H migration steps but that a cationic intermediate plays a
double role in loss of diastereoselectivity; the intermediate can undergo epimerization at iridium before
being trapped by halide and can also catalyze the epimerization of kinetic diastereomer product to
thermodynamic product. A detailed mechanism is proposed and simulations performed to fit the kinetic
data.
fluorinated stereocenters can also be challenging.28 While some
naturally occurring molecules contain such chiral centers,29 the
traditional method of synthesis has been to form a new C-F
Introduction
There is a significant effort in the organometallic community
to find ways of functionalizing normally unreactive and strong
bonds such as those between carbon and hydrogen, or carbon
and fluorine. In some ways the latter task is especially
challenging due to the strength of the C-F bond, the strongest
single bond to carbon,1 and has been the subject of much study,
both experimental and computational.2-27 The synthesis of
bond selectively using a fluorinating reagent, typically contain-
(14) Edelbach, B. L.; Jones, W. D. J. Am. Chem. Soc. 1997, 119, 7734-7742.
(15) Cronin, L.; Higgitt, C. L.; Karch, R.; Perutz, R. N. Organometallics 1997,
16, 4920-4928.
(16) Burdeniuc, J.; Jedlicka, B.; Crabtree, R. H. Chem. Ber./Recl. 1997, 130,
145-154.
(1) Smart, B. E. Fluorocarbons. In Chemistry of Functional Groups, Supplement
D; Patai, S., Rappoport, Z., Eds.; Wiley: New York, 1983; Chapter 14, pp
603-655.
(17) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. Chem. Commun. 1996,
787-788.
(18) Kiplinger, J. L.; Richmond, T. G. J. Am. Chem. Soc. 1996, 118, 1805-
1806.
(2) Hughes, R. P.; Laritchev, R. B.; Zakharov, L. N.; Rheingold, A. L. J. Am.
Chem. Soc. 2004, 126, 2308-2309.
(19) Kiplinger, J. L.; Richmond, T. G. Chem. Commun. 1996, 1115-1116.
(20) Kiplinger, J. L.; Richmond, T. G.; Osterberg, C. E. Chem. ReV. 1994, 94,
373-431.
(3) Reinhold, M.; McGrady, J. E.; Perutz, R. N. J. Am. Chem. Soc. 2004, 126,
5268-5276.
(4) Clot, E.; Me´gret, C.; Kraft, B. M.; Eisenstein, O.; Jones, W. D. J. Am.
Chem. Soc. 2004, 126, 5647-5653.
(21) Aizenberg, M.; Milstein, D. Science 1994, 265, 359-361.
(22) Weydert, M.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1993,
115, 8837-8838.
(5) Noveski, D.; Braun, T.; Schulte, M.; Neumann, B.; Stammler, H.-G. J.
Chem. Soc., Dalton Trans. 2003, 4075-4083.
(23) Belt, S. T.; Helliwell, M.; Jones, W. D.; Partridge, M. G.; Perutz, R. N. J.
Am. Chem. Soc. 1993, 115, 1429-1440.
(6) Jones, W. D. J. Chem. Soc., Dalton Trans. 2003, 3991-3995.
(7) Braun, T.; Perutz, R. N. Chem. Commun. 2002, 2749-2757.
(8) Braun, T.; Noveski, D.; Neumann, B.; Stammler, H.-G. Angew. Chem.,
Int. Ed. 2002, 41, 2745-2748.
(24) Klahn, A. H.; Moore, M. H.; Perutz, R. N. J. Chem. Soc., Chem. Commun.
1992, 1699-1701.
(25) Jones, W. D.; Partridge, M. G.; Perutz, R. N. J. Chem. Soc., Chem. Commun.
1991, 264-266.
(9) Braun, T.; Perutz, R. N.; Sladek, M. I. Chem. Commun. 2001, 2254-2255.
(10) Braun, T.; Cronin, L.; Higgitt, C. L.; McGrady, J. E.; Perutz, R. N.;
Reinhold, M. New. J. Chem. 2001, 25, 19-21.
(26) Jones, M. T.; McDonald, R. N. Organometallics 1988, 5, 1221-1223.
(27) Mazurek, U.; Schwarz, H. Chem. Commun. 2003, 1321-1326.
(28) Ramachandran, P. V., Ed.; Asymmetric Fluoroorganic Chemistry; Synthesis,
Applications, and Future Directions; ACS Symposium Series 746; Ameri-
can Chemical Society: Washington, DC, 2000.
(11) McAlexander, L. H.; Beck, C. M.; Burdeniuc, J. J.; Crabtree, R. H. J.
Fluorine Chem. 1999, 99, 67-72.
(12) Braun, T.; Foxon, S. P.; Perutz, R. N.; Walton, P. H. Angew. Chem., Int.
Ed. 1999, 38, 3326-3329.
(29) Filler, R. In Fluorine-Containing Chiral Compounds of Biomedical Interest;
Ramachandran, P. V., Ed.; ACS Symposium Series 746; American
Chemical Society: Washington, DC, 2001; pp 1-20.
(13) Whittlesey, M. K.; Perutz, R. N.; Greener, B.; Moore, M. H. Chem.
Commun. 1997, 187-188.
9
10.1021/ja0545012 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 15585-15594
15585

A R T I C L E S
Scheme 1
Garratt et al.
ing a metal.30-39 A complementary approach to fluorinated
stereocenters would be to break a C-F bond in a selective
manner to give the stereocenter, clearly a more difficult
thermodynamic process due to the strength of the C-F bond.1
However, it is now apparent that aliphatic C-F bonds R to
certain transition metal centers are activated toward attack by
exogenous protic40-43 or Lewis acids,44-48 with a strong bond
to H or a main group element providing significant thermody-
namic compensation for breaking the C-F bond. In some cases,
this process occurs in combination with hydride migration from
metal to the R-C,49-51 providing some encouragement that
selective C-F bond activation could be realized.
Following our initial discovery that cationic iridium-fluo-
roalkyl compounds react with elemental dihydrogen to give free
hydrofluorocarbons (HFCs),52 we demonstrated that this process
involved a heterolytic activation of H2 to give an iridium hydride
and an exogenous proton. Subsequent C-F bond activation by
the exogenous proton afforded defluorination and accompanying
hydride migration to carbon. Specifically, treatment of the
independently prepared hydride complex Cp*Ir(PMe3)(CF2-
CF3)H (1a)53 with CH3CO2H in CD2Cl2 afforded two diaster-
eomers of the acetate product in a 2:1 ratio as shown in Scheme
1.54 When CH3CO2D was used, around 10% of Cp*Ir(PMe3)-
(CDFCF3)OAc (2b) was observed along with the protio
analogue 2a. This result was consistent with two competing
processes; exogenous deuteronation at fluorine with intramo-
lecular H migration to give the principal product and competitive
scrambling of H+ and D+ (and thereby 1a and 1b) via a
presumed η2-HD intermediate. However, the fact that only a
small amount of deuterium was found on the R-carbon in the
product indicated that the dominant process was the former.
This was confirmed by using the corresponding iridium-
deuteride (1b) and CH3CO2H, in which 2b was the dominant
product, with some formation of 2a occurring by H/D exchange.
For each product isotopologue, the diastereoselectivity was about
2:1, and no significant kinetic isotope effects were observed.54
Although H(D)-migration showed some diastereoselectivity,
we subsequently demonstrated that the reaction of Cp*Ir-
(PMe3)(CF2CF2CF3)Me (3) with the weakly acidic 2,6-luti-
dinium chloride (LutHCl) proceeds quantitatively to give
Cp*Ir(PMe3)[CF(Me)CF2CF3]Cl (4) as a single diastereomer.
The relative configurations of the Ir and C stereocenters in 4
were defined crystallographically as (RIr, RC)(SIr, SC).55 No free
methane formation was observed, and use of 2,6-lutidinium-
(D)-chloride (LutDCl) gave no D-incorporation into the CF-
(Me)C2F5 group. This is consistent with external protonation
occurring exclusively at the R-CF bond to the complete
exclusion of protonation at the Ir-CH3 bond to give the kind
of η-methane intermediate found in many other systems.56-66
Furthermore, in a recently published study of C-F bond
activation coupled with vinyl migration, in which the intermedi-
ate is trapped by the pendant vinyl group before anion
coordination, we have provided evidence that the S-enantiomer
of the starting material affords the S-configuration at carbon in
the product and that C-F bond activation is completely
diastereoselective.67 Observation of complete diastereoselectivity
in these other migration reactions prompted a return look at
the original hydride migration system in an attempt to address
the reasons for the lower diastereoselectivity observed and to
further clarify the overall mechanistic features of this reaction.
(30) Gouverneur, V.; Greedy, B. Chem. Eur. J. 2002, 8, 766-771.
(31) Mohar, B.; Baudoux, J.; Plaquevent, J.-C.; Cahard, D. Angew. Chem., Int.
Ed. 2001, 40, 4214-4216.
(32) Piana, S.; Devillers, I.; Togni, A.; Rothlisberger, U. Angew. Chem., Int.
Ed. 2002, 41, 979-982.
(33) Prakesch, M.; Gre´e, D.; Gre´e, R. Acc. Chem. Res. 2002, 35, 175-181.
(34) Muniz, K. Angew. Chem., Int. Ed. 2001, 40, 1653-1656.
(35) Shibata, N.; Liu, Z.; Takeuchi, Y. Chem. Pharm. Bull. 2000, 48, 1954-
1958.
(36) Togni, A.; Mezzetti, A.; Barthazy, P.; Becker, C.; Devillers, I.; Frantz, R.;
Hintermann, L.; Perseghini, M.; Sanna, M. Chimia 2001, 55, 801-805.
(37) Liu, Z.; Shibata, N.; Takeuchi, Y. J. Org. Chem. 2000, 65, 7583-7587.
(38) Differding, E.; Lang, R. W. Tetrahedron Lett. 1988, 29, 6087-6090.
(39) Iseki, K. Tetrahedron 1998, 54, 13887-13914.
(40) Appleton, T. G.; Berry, R. D.; Hall, J. R.; Neale, D. W. J. Organomet.
Chem. 1989, 364, 249-273.
(41) Clark, G. R.; Hoskins, S. V.; Roper, W. R. J. Organomet. Chem. 1982,
234, C9-C12.
(55) Hughes, R. P.; Zhang, D.; Zakharov, L. N.; Rheingold, A. L. Organome-
tallics 2002, 21, 4902-4904.
(42) Burrell, A. K.; Clark, G. R.; Rickard, C. E. F.; Roper, W. R. J. Organomet.
Chem. 1994, 482, 261-269.
(56) Bullock, R. M.; Headford, C. E. L.; Hennessy, K. M.; Kegley, S. E.; Norton,
J. R. J. Am. Chem. Soc. 1989, 111, 3897-3908.
(43) Michelin, R. A.; Ros, R.; Guadalupi, G.; Bombieri, G.; Benetollo, F.;
Chapuis, G. Inorg. Chem. 1989, 28, 840-846.
(57) Johansson, L.; Tilset, M. J. Am. Chem. Soc. 2001, 123, 739-740.
(58) Wik, B. J.; Lersch, M.; Tilset, M. J. Am. Chem. Soc. 2002, 124, 12116-
12117.
(44) Koola, J. D.; Roddick, D. M. Organometallics 1991, 10, 591-597.
(45) Crespi, A. M.; Shriver, D. F. Organometallics 1985, 4, 1830-1835.
(46) Richmond, T. G.; Shriver, D. F. Organometallics 1984, 3, 305-314.
(47) Richmond, T. G.; Crespi, A. M.; Shriver, D. F. Organometallics 1984, 3,
314-319.
(59) Jones, W. D. Acc. Chem. Res. 2003, 36, 140-146.
(60) Hall, C.; Perutz, R. N. Chem. ReV. 1996, 96, 3125-3146.
(61) Gross, C. L.; Girolami, G. S. J. Am. Chem. Soc. 1998, 120, 6605-6606.
(62) Northcutt, T. O.; Wick, D. D.; Vetter, A. J.; Jones, W. D. J. Am. Chem.
Soc. 2001, 123, 7257-7270.
(48) Reger, D. L.; Dukes, M. D. J. Organomet. Chem. 1978, 153, 67-72.
(49) Huang, D.; Caulton, K. G. J. Am. Chem. Soc. 1997, 119, 3185-3186.
(50) Greene, T. R.; Roper, W. R. J. Organomet. Chem. 1986, 299, 245-250.
(51) Burrell, A. K.; Clark, G. R.; Jeffrey, J. G.; Rickard, C. E. F.; Roper, W. R.
J. Organomet. Chem. 1990, 388, 391-408.
(52) Hughes, R. P.; Smith, J. M. J. Am. Chem. Soc. 1999, 121, 6084-6085.
(53) Hughes, R. P.; Kovacik, I.; Lindner, D. C.; Smith, J. M.; Willemsen, S.;
Zhang, D.; Guzei, I. A.; Rheingold, A. L. Organometallics 2001, 20, 3190-
3197.
(63) Stahl, S.; Labinger, J. A.; Bercaw, J. E. Angew. Chem., Int. Ed. 1998, 37,
2181-2192.
(64) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1995, 117,
9371-9372.
(65) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1996, 118,
5961-5976.
(66) Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507-514.
(67) Hughes, R. P.; Laritchev, R. B.; Zakharov, L. N.; Rheingold, A. L. J. Am.
Chem. Soc. 2005, 127, 6325-6334.
(54) Hughes, R. P.; Willemsen, S.; Williamson, A.; Zhang, D. Organometallics
2002, 21, 3085-3087.
9
15586 J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005

Carbon−Fluorine Bond Activation
A R T I C L E S
Scheme 2
Scheme 3
discussion, we will henceforth refer to the major diastereomer
observed in the hydride migrations (in CD2Cl2 solution) as the
kinetic isomer and the minor one the thermodynamic isomer.
The diastereomers of complexes 5, 6, 8, and 9 have all been
fully characterized by NMR spectroscopy. For example, the 1H
NMR spectrum of complex 5a exhibits characteristic Cp*, PMe3,
and R-CHF resonances for each diastereomer, with the R-CHF
of the kinetic diastereomer resonating at δ 6.43 ppm and that
of the thermodynamic diastereomer at δ 6.59 ppm; the 19F
spectrum shows the corresponding fluorine resonances at δ
-191.5 and -192.6 ppm, respectively; the 31P{1H}NMR
spectrum shows the PMe3 of the kinetic diastereomer at δ -30.2
ppm as a doublet (3JPF ) 21.3 Hz) from coupling with the
R-fluorine, while that of the thermodynamic diastereomer
appears as a broad doublet of quartets (3JPF ) 17, 4JPF ) 3 Hz)
at -33.2 ppm from coupling to the R-fluorine as well as the
CF3. Similarly, complex 5b exhibits characteristic Cp* and PMe3
Results and Discussion
Characterization of the Diastereomeric Hydride Migration
Products. Reaction of the perfluoroethyl complexes 1a with
LutDCl or 1b with LutHCl gave compounds Cp*Ir(PMe3)-
(CHFCF3)Cl (5a) and Cp*Ir(PMe3)(CDFCF3)Cl (5b) in an
identical ratio of ∼2:1 (Scheme 2); each compound is formed
as a mixture of diastereomers, only one of which is shown, and
the composition of which varies with reaction conditions (see
later). Similarly a ∼2:1 mixture of complexes 6a and 6b was
obtained from the corresponding perfluoropropyl complexes 7a
(with LutDCl) or 7b (with LutHCl). Once formed, these
diastereomeric products are configurationally stable at room
temperature, and mixtures of varying ratios show no sign of
equilibration.
1
2
resonances in the H NMR spectrum; the H NMR spectrum
shows the R-CDF of the kinetic and thermodynamic diastere-
omers at δ 6.43 and 6.59 ppm, respectively, each appearing as
a broad doublet (2JFD) 8 Hz) from coupling to the R-CDF
fluorine; the 19F NMR spectrum shows the R-CDF fluorine of
the kinetic diastereomer at δ -192.0 ppm as a broad doublet
of quartets of triplets (3JPF ) 22, 3JFF ) 15, 2JDF ) 8 Hz) from
coupling with 31P, CF3, and deuterium, while the corresponding
resonance for the thermodynamic diastereomer resonates at
-193.6 ppm, also as a broad doublet of quartets of triplets (3JPF
Treatment of complexes 1a and 7a with >1 equiv of LutHI
in CD2Cl2 at room temperature afforded, respectively, Cp*Ir-
(PMe3)(CHFCF3)I (8) and Cp*Ir(PMe3)(CHFCF2CF3)I (9) as
mixtures of diastereomers (ratio 1.5:1-2:1), as shown in Scheme
3. Notably, at room temperature the iodide reactions are
complete before NMR spectra can be obtained and are much
faster than the analogous processes with LutHCl, for which
detailed kinetic measurements are possible as described below.
Compound 8 has been made previously by direct oxidative
addition of CF3CFHI to Cp*Ir(CO)2, followed by replacement
of CO in the resultant Cp*Ir(CO)(CFHCF3)I by PMe3 in
refluxing toluene.54 Notably, the diastereomer ratio obtained in
this high-temperature synthesis (1:6) is inverted from the 2:1
ratio obtained by the methodology of Scheme 3. This provides
some evidence that in the hydride migration reactions the
principal observed diastereomer of 8, and presumably also of 5
and 6, is the product of kinetic control. Further kinetic evidence
to confirm this is presented later. For consistency in subsequent
3
2
) 17, JFF ) 15, JDF ) 8 Hz) from coupling with 31P, CF3,
and deuterium; the 31P{1H}NMR spectrum shows the resonance
of the kinetic diastereomer at δ -30.2 ppm as a broad doublet
(3JPF ) 22 Hz) from coupling to the R-CDF fluorine, whereas
that of the thermodynamic diastereomer appears at -33.2 ppm
4
as a broad doublet of quartets (3JPF ) 17, JPF ) 3 Hz) from
coupling to the R-CDF fluorine as well as the CF3. Complexes
6a,b, not unexpectedly, give similar spectra. For diastereomers
of the iodo complex 9 the overall coupling patterns are similar
to those of the chloride analogue 6a, but in the 1H and 19F NMR
spectra, most signals are shifted to higher frequency. The spectra
are very similar to those of 8, which had been previously
prepared by an alternative route.54
The 1D NMR data for these compounds clearly define the
basic connectivities but provide no definitive stereochemical
9
J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005 15587

A R T I C L E S
Garratt et al.
Figure 1. Newman projections down the C-Ir bond showing the major
NOEs observed in the 19F{1H} HOESY experiment on 5a. Only one
enantiomer of each diastereomer is shown.
Figure 3. Newman projections down the C-Ir bond showing the major
NOE interactions observed in the 19F{1H} HOESY and 1H{1H} NOESY
experiments on 9. Only one enantiomer of each diastereomer is shown.
Figure 2. Newman projections down the C-Ir bond showing the major
NOEs observed in the 19F{1H} HOESY experiment on 6a. Only one
enantiomer of each diastereomer is shown.
Figure 4. Newman projection down the C-Ir bond showing the major
NOEs found in the 19F{1H} HOESY experiment on 8. Only one enantiomer
of this diastereomer is shown.
information. While we have not been able to obtain X-ray
quality crystals of a single diastereomer of any of the migration
products to determine the relative stereochemistries of the two
stereocenters, a series of 19F{1H} HOESY studies enabled
assignment of the relative configurations and preferred confor-
mations in solution, as we have previously demonstrated for
similar compounds.68 For complexes 5a and 6a, the 19F{1H}
HOESY experiments were performed on a mixture of diaster-
eomers in CD2Cl2 solution. Newman projections of the con-
formations and relative stereocenter configurations derived from
the observed strong NOE cross-peaks are shown in Figures 1
and 2. For 5a NOEs were found for each isomer between the
â-CF3 and both the Cp* and PMe3, indicating that the CF3 is
oriented between these groups. This is expected, based on
numerous X-ray structures of similar compounds.69 For the
kinetic diastereomer, there is a strong NOE between the CHF
and the PMe3 but a much weaker NOE from CHF to the Cp*;
the R-CF in this conformation is just within the 5 Å NOE
range.70,71 For the thermodynamic diastereomer, there is a strong
NOE between the CHF and the Cp*, indicating they are close
in space, with no correlation found from CHF to the PMe3.
Likewise for 6a (Figure 2) NOEs are found from the â-CF2
groups to both Cp* and PMe3, indicating a close spatial
relationship in each case. The CF3 group of the thermodynamic
isomer also shows a correlation to both Cp* and PMe3; for the
kinetic isomer, the CF3 is only close to the Cp*. For the kinetic
isomer, there is a strong NOE between the CHF and the PMe3,
demonstrating they are again close in space, while for the
thermodynamic isomer there is strong correlation between the
CHF and the Cp*, confirming that the solution conformations
for 6a are analogous to those of 5a.
These NOE correlations define not only the preferred ground
state conformations shown, but also the relative configurations
at C and Ir for the kinetic diastereomer as (SC, SIr)(RC, RIr) and
for the thermodynamic counterpart as (RC, SIr)(SC, RIr). In
Figures 1 and 2 only the (SC, SIr) and (RC, SIr) isomers are
depicted. It is clear that the kinetic diastereomer formed in the
hydride migration reaction has the same relative stereochemistry
as the single diastereomer formed in the analogous methyl
migration and that this is indeed the product of kinetic control.55
For complex 9, 19F{1H} HOESY and 1H{1H} NOESY
experiments were carried out on the 2:1 mixture of diastereomers
formed by reaction of 7a with excess LutHI in CD2Cl2. Newman
projections of the derived structures are given in Figure 3. In
the 19F{1H} HOESY spectrum, NOEs were again found for each
isomer between the â-CF2 and both the Cp* and PMe3 groups,
indicating that the C2F5 fragment is oriented between these
groups. For the thermodynamic isomer, there is a strong NOE
between the CHF and the Cp*, with no correlation found from
CHF to the PMe3. For the kinetic isomer, there is a strong NOE
between the CHF and the PMe3, but as found for 5a, there is
also a weak NOE from CHF to the Cp*. In the 1H{1H} NOESY
(Figure 3), the only significant NOE observed is between the
CHF and the PMe3 group of the thermodynamic isomer,
consistent with the structure shown.
The 19F{1H} HOESY spectrum of the mixture of diastereo-
mers of 8 was also obtained on the 1:6 diastereomeric mixture
of 8, formed by treatment of Cp*Ir(CO)(CHFCF3)I with PMe3
in refluxing toluene.54 A Newman projection of the derived
structure showing the main NOEs for the thermodynamic
1
diastereomer is shown in Figure 4. In the H-19F HOESY
spectrum, NOEs were seen only for the thermodynamic dias-
tereomer but were sufficient to characterize it, and therefore by
a process of elimination, the corresponding kinetic diastereomer.
Correlations were found between the â-CF3 and both the Cp*
and PMe3, indicating that the CF3 fragment is oriented between
these groups. A NOE is also observed between the CHF and
(68) Hughes, R. P.; Zhang, D.; Ward, A. J.; Zakharov, L. N.; Rheingold, A. L.
J. Am. Chem. Soc. 2004, 126, 6169-6178.
(69) Hughes, R. P.; Smith, J. M.; Liable-Sands, L. M.; Concolino, T. E.; Lam,
K.-C.; Incarvito, C.; Rheingold, A. L. J. Chem. Soc., Dalton Trans. 2000,
873-879.
(70) Neuhaus, D.; Williamson, M. P. The Nuclear OVerhauser Effect in
Structural and Conformational Analysis; Wiley-VCH: New York, 2000.
(71) Macchioni, A. Eur. J. Inorg. Chem. 2003, 195-205.
9
15588 J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005

Carbon−Fluorine Bond Activation
A R T I C L E S
Table 1. Rate Constants kobs (L1/2/mol1/2 s) for Reaction of 1a,b
and 7a,b with 2,6-Lutidinium(H/D) Chloride in CD2Cl2 at 0 °C
/
/
compounda
acid
104 kobs(L1 2/mol1 2 s)
kH/kD
[Ir](H)CF2CF3 (1a)
[Ir](D)CF2CF3 (1b)
[Ir](D)CF2CF3 (1b)
[Ir](H)CF2CF3 (1a)
[Ir](H)CF2CF2CF3 (7a)
[Ir](D)CF2CF2CF3 (7b)
[Ir](D)CF2CF2CF3 (7b)
[Ir](H)CF2CF2CF3 (7a)
LutHCl
LutDCl
LutHCl
LutDCl
LutHCl
LutDCl
LutHCl
LutDCl
7.5 ( 0.2
4.2 ( 0.2
5.7 ( 0.2
5.3 ( 0.2
5.8 ( 0.2
4.2 ( 0.2
5.6 ( 0.2
4.5 ( 0.2
1.8 ( 0.2
1.1 ( 0.2
1.4 ( 0.2
1.2 ( 0.2
Figure 5. Interpretation of the half-order rate dependence on [LutHCl].
a [Ir] ) Cp*Ir(PMe3).
Scheme 4
the Cp* ring. Thus, this isomer is analogous in structure to the
thermodynamic isomers of 5a, 6a, and 9, which is also apparent
from their very similar NMR spectra.
With the structures of the product diastereomers firmly
defined we can now move to a detailed discussion of the kinetics
and mechanism of their formation and interconversion.
Kinetic Observations. We have carried out detailed kinetic
studies on the reactions of 1 and 7 with LutH(D)Cl salts in three
solvents: CD2Cl2, DMF-d7, and CD3NO2. Two remarkable
general observations concerning the kinetic data are worthy of
initial comment. First, in all solvents the experimentally
observed rate law was established to be -d[Ir]/dt ) k[Ir]-
[LutH(D)Cl]1/2, with an expected first-order dependence on the
concentration of iridium starting complex but a surprising one-
half-order dependence on the concentration of lutidinium salt.
There is no literature evidence that lutidinium salts are dimeric
in these solvents, so a (dimer) a 2(monomer) preequilibrium
is not a convenient way to explain this half-order dependence.
Our interpretation is summarized in Figure 5: LutHCl is a weak
acid, with a pKa(H2O) of 6.7;72-74 dissociation generates HCl
and lutidine, the small and equal concentration of each of which
has a half-order dependence on [LutHCl]. In agreement with
this interpretation, addition of 2,6-lutidine results in a strong
inhibition of C-F activation. The most significant conclusion
is that the actiVe source of the exogenous proton for the C-F
actiVation reactions is not lutidinium chloride, but rather small
amounts of HCl generated by its dissociation!
activation step is preceded by a dissociative equilibrium which
will also have an unknown isotope effect associated with it.
An Eyring plot of the kinetic data measured over the
temperature range 0 to 15 °C affords a free energy of activation
∆G q(298K) ) 84 ( 4 kJ mol-1 for the reaction of 1a and LutHCl.
An identical value of ∆G q(298K) ) 85 ( 7 kJ mol-1 was
obtained for the reaction of 7a and LutHCl. These are
significantly larger than the value of ∆G q(298K) ) 64 ( 2 kJ
mol-1 previously measured for reaction of 1a with acetic acid,
consistent with the much faster reaction (by a factor of
approximately 50 at 0 °C) observed in the previously reported
system.54
The second general observation is that plots of ln[Ir] versus
time are linear for at least three reaction half-lives, even using
one molar equiv of LutHCl. Clearly, the proton acts as a catalyst
whose concentration remains effectively constant throughout
most of the reaction. (As discussed later, the effect of the halide
counterion must be taken into account under some circum-
stances.) Consequently, the raw kinetic data behave as for a
pseudo first-order reaction; the slope of the linear plot gives
k
obs[LutHCl]1/2, which in turn affords values of kobs
Initial kinetic experiments were performed for the reaction
.
As shown in Scheme 4, the competitive proton exchange
equilibrium between 1a (1b) and Lut(H/D)Cl must be faster
than the protonation of the R-C-F bond (k3, k4 . k1, k2), which
results in complete scrambling of H and D before C-F
activation, and therefore an identical mixture of 5a and 5b from
both reactions. Similar results are observed with the corre-
sponding perfluoropropyl analogues 7a,b. Clearly, when k1 is
similar to k2, the degree of isotopic scrambling in the final
products (5a and 5b) is determined by the equilibrium isotope
effect in the proton exchange equilibrium between 1a (1b) and
Lut(H/D)Cl. In other words, the isotopomer ratio [5a]/[5b]
reflects the equilibrium constant (Keq) for the H/D exchange
process. This is different from the previously reported system
of 1a or 7a with LutDCl or LutHCl and 1b53 or 7b with LutDCl
or LutHCl in CD2Cl2 at 0 °C. Reactions were monitored by 1H
NMR, and concentrations of reactants and products were
obtained by integration of appropriate resonances, using 1,3,5-
trimethoxybenzene as an internal integration standard. The rate
constants kobs (L1/2/mol1/2 s) at 0 °C are shown in Table 1.
Kinetic isotope effects are not large, and any interpretation
related to the C-F activation step is clearly inappropriate due
to their small magnitude and the complication that the C-F
(72) Gero, A.; Markham, J. J. J. Org. Chem. 1951, 16, 1835-1838.
(73) Brown, H. C.; Mihm, X. R. J. Am. Chem. Soc. 1955, 77, 1723-1726.
(74) Brown, H. C.; Kanner, B. J. Am. Chem. Soc. 1966, 88, 986-992.
9
J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005 15589

A R T I C L E S
Garratt et al.
Figure 6. Eyring plots for the reaction of 7a with LutHCl in different solvents.
Table 2. Solvent Effects on the Activation Parameters for the
Reaction of Cp*Ir(PMe3)(C3F7)H (7a) + LutHCl
component rather than the ∆Hq values. Likewise, the signifi-
cantly greater value for ∆Gq when DMF-d7 is used as the
solvent is also a result of the large negative value of ∆Sq in
this solvent. Comparison of rates using CD3NO2 and DMF-d7
makes it clear that solvent donor properties and dielectric
constants are both important in affecting the reaction rate. It
seems likely that strong donor solvent stabilization of the protic
acid component of the starting materials will be a significant
cause of slower reaction rates in DMF. As we will see shortly,
the intermediate formed by C-F activation must be ionic and
therefore the transition state leading to this intermediate must
be a polar one. In DMF the large negative ∆Sq and consequently
slower rate could result from the solvation requirements for this
transition state, while in nitromethane the high solvent dielectric
helps to stabilize this relative to the ground state.
donor no.
dielectric constant
∆
H
q
∆
S
q
∆
G
q(298K)
1
1
a
1
1
solvent
(kJ mol-
)
ꢀra
(kJ mol-
)
(J mol-1 K-
)
(kJ mol-
)
CD2Cl2
CD3NO2
DMF-d7
0
11.3
111.3
8.9
35.9
36.7
73(5)
86(4)
74(2)
-38(18)
14(15)
-73(7)
85(7)
82(6)
96(3)
a Ref 77.
using acetic acid as the protonating agent, in which the
corresponding k1 and k2 values must be larger than k3 or k4 and
scrambling competes very inefficiently with C-F bond activa-
tion.54 The exchange is shown as occurring via a putative
cationic η2-HD intermediate, which is not directly observed in
this system. However similar cationic species, obtained by
protonation of iridium hydrides, have been reported else-
where.75,76
Plots of concentration versus time for the reaction of 7a with
varying amounts of LutHCl in CD2Cl2, two examples of which
are shown in Figure 7, provide considerable insight into the
nature of the intermediates from which the kinetic (RIrRC, SIrSC)
and thermodynamic (RIrSC, SIrRC) diastereomeric products are
formed. When an excess of LutHCl (4.6 equiv) is used, the plot
shows the expected features, with a pseudo first-order decay of
starting material 7a and formation of the kinetic and thermo-
dynamic diastereomers of product 6a. When a smaller excess
of LutHCl (2.0 equiv) is used, the ratio of diastereomers of 6a
varies over the reaction course, as shown in Figure 8, starting
at ∼4:1 (by extrapolation) and giving a final ratio of ∼2.5:1.
When a larger excess of LutHCl (10 equiv) is used, the ratio of
diastereomers remains constant at ∼4:1 over the entire reaction
course. Recall that we have shown that, once isolated from the
reaction mixture, the product diastereomers do not interconvert
or equilibrate at room temperature. However, as shown in Figure
7, when using less than one equiv of LutHCl it is clear that in
the very late stages of the reaction the kinetic product diaste-
reomer is actually consumed to produce its thermodynamic
counterpart; the sudden diVergence in the concentrations
appears to coincide with the time at which the concentration
of aVailable chloride drops to approximately zero.
Solvent effects on the overall reaction rates were significant;
compared to reactions in CD2Cl2 those in DMF-d7 were
considerably slower and those in CD3NO2 considerably faster.
As already mentioned, the experimental rate law is identical in
all three solvent systems. Eyring plots are shown in Figure 6
for the reaction of 7a with LutHCl in the three solvents used.
Values of ∆Hq and ∆Sq obtained from these plots were used
to calculate ∆Gq(298K) for each reaction; the results, along with
values of the donor number and dielectric constants for each
solvent,77 are compared in Table 2. Clearly the overall activation
free energies are strongly influenced by the sign of the ∆Sq
term. While the values of ∆Gq in CD2Cl2 and in CD3NO2 are
statistically indistinguishable, reaction rates are observed to be
significantly faster in the latter solvent, and while the uncertainty
in the calculated ∆Gq values do not reflect this in a statistically
different way, it is clear that the relative values of ∆Gq are
CD3NO2 < CD2Cl2. It is also clear that the faster reaction in
CD3NO2 compared to CD2Cl2 is a result of the sign of the ∆Sq
(75) Tellers, D. M.; Bergman, R. G. Organometallics 2001, 20, 4819-4832.
(76) Oldham, W. J., Jr.; Hinkle, A. S.; Heinekey, D. M. J. Am. Chem. Soc.
1997, 119, 11028-11036.
(77) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry; Wiley-
VCH: Weinheim, Germany, 2003.
9
15590 J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005

Carbon−Fluorine Bond Activation
A R T I C L E S
Scheme 5
Figure 7. Plots of concentration vs time for starting materials and
diastereomeric products in the reactions of 7a with two different equiv of
LutHCl in CD2Cl2.
trapping of 10SS by chloride would afford the observed kinetic
diastereomer of 5a (6a). However, if inversion at iridium occurs
before chloride trapping, 10SS forms 10RS, which can then be
trapped by chloride to give the thermodynamic product dias-
tereomer. In previous work we have shown that inversion at
the metal in such 16-electron cations can be fast on the NMR
time scale.78 In Scheme 5 the cationic intermediates 10 are
shown as being solvated, even in CD2Cl2, as analogous weakly
bound methylene chloride and other chlorocarbon complexes
have been experimentally observed in alkyl analogues of
10.75,79-82
Figure 8. Variation of the ratio of diastereomers of product 6a in the
reaction of 7a with 2.0 equiv of LutHCl in CD2Cl2 at 15 °C.
To ensure that changes in the diastereomeric ratio are not
due to any configurational lability of the R-carbon under the
reaction conditions, isolated 6a was treated with LutDCl to test
whether deuterium incorporation into the R-carbon occurs
through a proton exchange pathway. No deuterium was observed
in 6a after 24 h. This is consistent with the idea that
configurational inversion at iridium is responsible for the change
in diastereomer ratio during the reaction.
The Reaction Mechanism. A reaction mechanism consistent
with the observed kinetic results is shown in Scheme 5 using
the S-enantiomer of racemic starting complex 1a or 7a;
obviously a mirror image scheme must exist for the correspond-
ing R-enantiomer. In a previously published study of C-F bond
activation reactions of 7a coupled with vinyl migration, we have
provided evidence that the S-enantiomer of the starting material
affords the S-configuration at carbon in the product and that
C-F bond activation is completely diastereoselective.67 We have
no reason to suppose that hydride migration should proceed by
a fundamentally different pathway, and so Scheme 5 depicts
an analogous, selective loss of Fa and migration of H to give
cation 10SS as the kinetically formed intermediate. Irreversible
(78) Hughes, R. P.; Lindner, D. C.; Rheingold, A. L.; Liable-Sands, L. M. J.
Am. Chem. Soc. 1997, 119, 11544-11545.
(79) Arndtsen, B. A.; Bergman, R. G. Science 1995, 270, 1970-1973.
(80) Tellers, D. M.; Bergman, R. G. Can. J. Chem. 2001, 79, 525-528.
(81) Tellers, D. M.; Bergman, R. G. J. Am. Chem. Soc. 2001, 123, 11508-
11509.
(82) Tellers, D. M.; Yung, C. M.; Arndtsen, B. A.; Adamson, D. R.; Bergman,
R. G. J. Am. Chem. Soc. 2002, 124, 1400-1410.
9
J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005 15591

A R T I C L E S
Garratt et al.
Scheme 6
Therefore, formation of two diastereomers throughout the
reaction appears to result from competition between intramo-
lecular inversion at iridium in initially formed cationic inter-
mediate 10SS and bimolecular trapping of that intermediate by
chloride. This is consistent with the observation that using larger
excesses of LutHCl results in an approximately constant ratio
of product diastereomers throughout the reaction but that
reduction of the concentration of LutHCl, and hence the
concentration of chloride available for trapping, results in
reduced rates of bimolecular trapping relative to inversion at
iridium and consequently a decrease in the kinetic/thermody-
namic diastereomer ratio over time (Figure 8). The rate of
inversion (k2[10SS]) must be similar to that for trapping of the
cationic intermediate 10SS by chloride (k4[10SS][Cl-]) (see
Scheme 5). If the rate of inversion were much faster than that
for trapping, the diastereomer ratio would reflect the equilibrium
constant (Keq ) k3/k4) for inversion and should not change with
time, assuming that the rates of trapping of 10SS and 10RR were
essentially the same. If the rate constants of inversion (k2, k3)
and trapping (k4, k5) are similar, then as [Cl-] decreases, one
would expect to see a change in product diastereomer ratio.
This change will be greater if intermediate 10RS is favored
thermodynamically over 10SS; i.e., if k2 > k3.
If potentially equilibrating cationic intermediates 10 are
indeed formed, significant solvent effects are expected, and they
are observed. Reactions are at least twice as fast in the higher
dielectric constant solvent CD3NO2 (vide supra) consistent with
a polar intermediate, and significantly, the diastereomer ratio
is inverted compared to that observed in CD2Cl2 under similar
conditions. The use of 2 equiv of LutHCl in CD2Cl2 gives an
initial kinetic/thermodynamic diastereomer ratio of 6a of ∼4:1
decreasing to ∼2:1; in CD3NO2 the ratio is initially 2:3,
changing over the reaction course to 1:2. This is consistent with
the polar intermediate 10 (see Scheme 5) formed after C-F
activation/migration being stabilized in nitromethane, allowing
intramolecular epimerization at iridium to occur more competi-
tively with chloride trapping. Support for this mechanistic
premise was provided when a 3:2 mixture of kinetic/thermo-
dynamic diastereomers of 6a, generated in CD2Cl2 was observed
to be unchanged after 1 day. On dissolution of this same mixture
in CD3NO2, a rapid change in diastereomer ratio to 1:2 was
found; this changed steadily over 3 days until it reached an
apparent equilibrium value of ∼1:15. This is clearly indicative
of irreversible coordination of chloride and consequent con-
figurational stability at iridium in CD2Cl2, but reversible
dissociation of chloride in nitromethane, allowing thermody-
namic control of the diastereomer ratio to prevail via epimer-
ization at iridium.
thermodynamic product once the concentration of chloride drops
to approximately zero (Figure 7). This can occur if untrapped
cation 10SS (or 10RS) can catalyze this transformation, presum-
ably via a halide-bridged intermediate 11, as shown in Scheme
6. By halide exchange via the bridged species 11, a cation such
as 10RS can convert the configurationally stable kinetic diaste-
reomer shown into a configurationally labile cation 10SS; if this
reaction is reasonably fast it must result in conversion of kinetic
product diastereomer to the thermodynamic diastereomer. To
test this premise, a 2:1 diastereomer mixture of 6a was treated
with 0.1 equiv of Cp*Ir(PMe3)(C3F7)(O3SCF3)55 in CD2Cl2 at
21 °C in an NMR tube. The triflate ligand is weakly bound,
and this complex should be an appropriate substitute for the
cationic intermediate 10 in the hydride migration reactions. A
1H NMR spectrum obtained within 5 min of mixing showed a
dramatic change in the diastereomer ratio of 6a from 2:1 to
<1:10, consistent with our prediction. The above results support
our proposal that halide exchange with a cationic intermediate
is responsible for the epimerization observed when [Ir] > [Cl-].
Attempts were made to isolate the cationic intermediate 10
by treatment of 7a with 1 equiv of LutH[B{3,5-C6H3(CF3)2}4],83
containing a very weakly coordinating anion, in CD2Cl2 at room
temperature. The reaction was complete within minutes, appar-
ently giving [Cp*Ir(PMe3)(CHFC2F5)]+[B{3,5-C6H3(CF3)2}4]-.
Like other analogues75,79-82 (vide supra), it is probable that this
In DMF-d7, reactions of 7a with LutHCl were, as mentioned
earlier, much slower than in CD2Cl2, but diastereomer ratios of
2:1 were typically found not to change over the course of the
reaction. Unlike reactions in methylene chloride or nitromethane,
epimerization at iridium does not appear to occur over time,
perhaps because any intermediate cation is configurationally
stabilized by stronger coordination of DMF than either CD2Cl2
or CD3NO2.
Finally, any mechanism must account for the observation that,
while the individual product diastereomers are configurationally
stable in CD2Cl2, reactions using less than one equiv of LutHCl
result in sudden, rapid conversion of the kinetic product to the
1
species contains a coordinated molecule of CD2Cl2. The H
(83) Yandulov, D. V.; Schrock, R. R. J. Am. Chem. Soc. 2002, 124, 6252-
6253.
9
15592 J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005

Carbon−Fluorine Bond Activation
A R T I C L E S
rate constant of ∼103 s-1 85
.
The use of these as initial estimates
Scheme 7. Mechanism and Rate Constants (CD2Cl2; 294 K) from
Simulation Using the IBM CKS Programa,b
for the rate constants (k1, k2, k3) allowed appropriate values for
other rate constants to be derived to give the best fit. Interest-
ingly, to get a satisfactory fit, a value of k8 (<k1) for protonation
of starting material by putative product LutHF was required.
This is relatively insignificant when excess LutHCl is used. Rate
constants for inversion at iridium (k2, k3) are somewhat larger
than those for trapping by chloride (k4, k5). The ratio of k6/k7
also reflects the approximate equilibrium constant (∼15) for
the thermodynamic/kinetic diastereomer products, albeit mea-
sured in CD3NO2 instead of CD2Cl2.
Conclusions
a Ref 84. bk6 and k7 are rate constants for interconversion of diastereomers
of product by untrapped cations 10.
Major features of the mechanism of exogenous proton-
promoted C-F bond activation coupled with hydride migration
at iridium have been clarified. The half-order dependence on
LutHCl concentration strongly suggests that LutHCl is not the
active source of protons but rather dissociates to provide small
concentrations of HCl as the active reagent. Kinetic and
thermodynamic product stereochemistries have been defined
clearly, and in particular the reasons for less than complete
diastereoselectivity have been elucidated. These results beg the
question of why diastereoselectivity is so clean in other
migrating group systems, and experiments designed to probe
these other reactions are underway. These results are also not
able to distinguish whether C-F bond activation and H-
migration occur in a synchronous or sequential fashion. Further
experiments to examine this question are underway.
NMR spectrum showed a single set of resonances for an iridium
complex: doublets at δ 1.72 and 1.58 for the Cp* and PMe3,
respectively, and a doublet of doublets at δ 6.81 for the R-CHF.
In the 19F NMR spectrum, a sharp singlet was found at -83.3
ppm for the CF3 group of the iridium complex along with broad
resonances at -112.6 and -116.5 ppm due to the â-CF2 group
and a broad multiplet at -193.0 ppm due to the R-CHF. The
31P NMR spectrum showed a single peak, at -24.37 ppm,
consistent with a cationic iridium complex. On leaving the
product from reaction of 7a and LutH+[B{3,5-C6H3(CF3)2}4]-
in CD2Cl2 solution for 30 min, some chloride abstraction from
the solvent was observed to give complex 6a. Addition of 1
equiv of NaI in CD3OD at this point produced 9 as a 1:15
mixture of diastereomers, i.e., the thermodynamic isomer
favored.
Experimental Section
General Considerations. All reactions were performed in oven-
dried glassware, using standard Schlenk techniques, under an atmo-
sphere of nitrogen which has been deoxygenated over BASF catalyst
and dried over Aquasorb, or in a Braun drybox. Methylene chloride,
hexane, diethyl ether, tetrahydrofuran, and toluene were dried over an
alumina column under nitrogen.86 NMR spectra were recorded on a
Varian Unity Plus 300 or 500 MHz FT spectrometer. 1H NMR spectra
were referenced to the protio impurity in the solvent; C6D6 (7.16 ppm),
CDCl3 (7.27 ppm), CD2Cl2 (5.32 ppm). 19F NMR spectra were
referenced to external CFCl3 (0.00 ppm). 31P{1H} NMR spectra were
referenced to external 85% H3PO4 (0.00 ppm). The complexes Cp*Ir-
(PMe3)(CF2CF3)H (1a), Cp*Ir(PMe3)(CF2CF3)D (1b), Cp*Ir(PMe3)(CF2-
CF2CF3)H (7a), and Cp*Ir(PMe3)(CF2CF2CF3)D (7b)53 were prepared
by literature procedures. Cp*Ir(PMe3)(CHFCF3)I (8) had been previ-
ously synthesized, by an alternative route.54 LutHCl, LutDCl,87 and Lut-
[B{3,5-C6H3(CF3)2}4]83 were prepared according to literature proce-
dures.
2,6-Lutidinium Iodide. To a solution of 2,6-lutidine (1.2 mL, 10.2
mmol) in THF (25 mL) was added methanol (0.62 mL, 15.25 mmol)
followed by trimethylsilyliodide (3.05 g, 15.25 mmol), each via syringe.
Immediate evolution of HI and precipitation of a white powder was
observed. The mixture was stirred for 30 min and then filtered via
cannula. The solid was washed 4 times with THF (10 mL) and then
dried in vacuo. The solid was recrystallized from CH2Cl2/hexane to
yield a pale yellow microcrystalline solid, yield 1.45 g (60%). 1H NMR
(CDCl3): δ 14.25 (br s, 1H, LutH+), 8.29 (t, J ) 7.8 Hz, 1H, p-PyH),
7.59 (d, J ) 7.8 Hz, 2H, m-PyH), 3.05 (s, 6H, PyMe).
In summary, we conclude that iridium cations 10 are crucial
intermediates during these C-F activation/hydride migration
reactions and play two key roles in the epimerization at iridium
and consequent loss of reaction diastereoselectivity; intramo-
lecular inversion of the cation competes with chloride trapping,
and untrapped cation also catalyzes the interconversion of
product diastereomers. This loss can be minimized by using
high concentrations of LutHCl but not eliminated altogether.
The proposed mechanism was subjected to testing using the
IBM chemical kinetics simulator.84 The mechanism shown in
Scheme 7 with the rate constants shown give satisfactory
simulations of the concentration versus time data for all reagents
and products in CD2Cl2 at 294 K, such as those shown in Figure
8; in particular it is successful in the challenging simulation of
the data using less than 1 equiv of LutHCl, in which dramatic
concentration changes of product diastereomers occur late in
the reaction. The mechanism requires selective C-F bond
activation with hydride migration to give cation 10SS to be rate
limiting, with the subsequent inversion at iridium and trapping
by halide being significantly faster processes. The simulations
are based on reasonable estimates for some key rate constants;
the overall rate constant kobs from reactions at 273 K was
extrapolated to give a reasonable estimate for the room-
temperature value of k1, and the rate of inversion at iridium in
the known cationic complex [Cp*Ir(PMe3)(CF2CF3)-
(H2O)]+(BF4)- is fast on the NMR time scale with an estimated
(85) Hughes, R. P.; Lindner, D. C.; Smith, J. M.; Zhang, D.; Incarvito, C. D.;
Lam, K.-C.; Liable-Sands, L. M.; Sommer, R. D.; Rheingold, A. L. J. Chem.
Soc., Dalton Trans. 2001, 2270-2278.
(86) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518-1520.
(84) Hinsberg, W. D.; Houle, F. A. IBM Almaden Research Center Chemical
Kinetics Simulator. https://www.almaden.ibm.com/st/computational_science/
ck/msim/, September 2005.
(87) Gronberg, K. L. C.; Henderson, R. A.; Oglieve, K. E. J. Chem. Soc., Dalton
Trans. 1998, 3093-3104.
9
J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005 15593

A R T I C L E S
Garratt et al.
4
2
4
3
Cp*Ir(PMe3)(CHFCF3)Cl (5a). A J-Young NMR tube was charged
with Cp*Ir(PMe3)(CF2CF3)H (9.1 mg, 0.0174 mmol) and LutHCl (2.5
mg, 0.0174 mmol). CD2Cl2 (0.5 mL) was vacuum transferred into the
NMR tube. After about 1 h at room temperature, the NMR spectra
showed the complete conversion of Cp*Ir(PMe3)(CF2CF3)H into Cp*Ir-
(PMe3)(CHFCF3)Cl as a 2.7:1 mixture of two diastereomers. The ratio
of the two diastereomers changed over the reaction period, from 4.3:1
to 2.7:1. The solvent was pumped down, and the yellow residue was
extracted into hexane. Evaporation of hexane afforded a yellow solid
(9.3 mg, 99%). Anal. Calcd for C15H25ClF4IrP (540.00): C, 33.36; H,
4.67. Found: C, 33.49; H, 4.64.
(d, JFF ) 13, 3F, CF3), -112.45 (ddd, JFF ) 274, JFP ) 21, JFF
)
2
3
3
21, 1F, â-CF), -116.70 (ddd, JFF ) 274, JFH ) 40, JFF ) 17, 1F,
â-CF), -194.43 (br s, 1F, CHF); 31P{1H}NMR (CD2Cl2) δ -34.0 (dd,
3JPF ) 21, JPF ) 21, PMe3).
4
Cp*Ir(PMe3)(CHFCF2CF3)I (9). A J-Young NMR tube was
charged with Cp*Ir(PMe3)(CF2CF2CF3)H (10 mg, 0.0174 mmol) and
LutHI (8 mg, 0.0348 mmol). CD2Cl2 (0.5 mL) was transferred into the
NMR tube by syringe. After 10 min, the NMR spectra showed the
complete conversion of Cp*Ir(PMe3)(CF2CF2CF3)H into Cp*Ir-
(PMe3)(CHFCF2CF3)I as a 2:1 mixture of two diastereomers. The
solvent was pumped down, and the yellow residue was extracted into
hexane. Evaporation of hexane afforded a yellow solid (12 mg, 99%).
Anal. Calcd for C16H25ClF6IIrP (681.46): C, 28.20; H, 3.70. Found:
C, 28.08; H, 3.62.
The kinetic diastereomer (SC, SIr)(RC, RIr): 1H NMR (CD2Cl2) δ 6.43
2
3
3
4
(dqd, JFH ) 47, JFH ) 12, JPH ) 6, 1H, CHF), 1.68 (d, JPH ) 2,
15H, Cp*), 1.57 (d, 2JPH ) 11, 9H, PMe3); 19F NMR (CD2Cl2) δ -71.0
3
3
2
3
(dd, JFF ) 13, JHF ) 12, 3F, CF3), -191.5 (ddq, JHF ) 47, JPF
)
)
The kinetic diastereomer: 1H NMR (CD2Cl2) δ 6.94 (ddd, JFH
)
2
21, JFF ) 15, 1F, CHF); 31P{1H}NMR (CD2Cl2) δ -30.2 (d, JPF
3
3
47.5 Hz, 3JFH ) 33.8 Hz, 3JPH ) 2.5, 1H, CHF), 1.90 (d, 4JPH ) 2 Hz,
21, PMe3).
15H, C5Me5), 1.78 (d, JPH ) 11 Hz, 9H, PMe3); 19F NMR (CD2Cl2)
2
The thermodynamic diastereomer (RC, SIr) (SC, RIr): 1H NMR (CD2-
4
2
δ -83.43 (d, JFF ) 17 Hz, 3F, CF3), -109.91 (dd, JFF ) 273 Hz,
2
3
3
Cl2) δ 6.59 (dqd, JFH ) 47, JFH ) 11, JPH ) 3, 1H, CHF), 1.72 (d,
3JFF ) 17 Hz, 1F, â-CF), -118.09 (ddd, JFF ) 273 Hz, JFF ) 33.8
2
3
4JPH ) 2, 15H, Cp*), 1.47 (d, JPH ) 11, 9H, PMe3); 19F NMR (CD2-
2
Hz, 3JFF ) 17 Hz, 1F, â-CF), -183.30 (dd, 3JFH ) 33.8 Hz, 1F, CHF);
3
3
4
Cl2) δ -71.1 (ddd, JFF ) 15, JHF ) 11, JPF ) 4, 3F, CF3), -192.9
31P{1H}NMR (CD2Cl2) δ -37.49 (d, JPF ) 17.8 Hz, PMe3).
3
(ddq, 2JHF ) 47, 3JPF ) 17, 3JFF ) 15, 1F, CHF); 31P{1H}NMR (CD2-
The thermodynamic diastereomer: 1H NMR (CD2Cl2) δ 6.75 (ddd,
3
4
Cl2) δ -33.2 (dq, JPF ) 17, JPF ) 4, PMe3).
2JFH ) 47.0 Hz, JFH ) 38.5 Hz, JPH ) 3.0 Hz, 1H, CHF), 1.95 (dd,
3
3
Cp*Ir(PMe3)(CDFCF3)Cl (5b). A J-Young NMR tube was charged
with Cp*Ir(PMe3)(CF2CF3)D (9.1 mg, 0.0174 mmol) and LutDCl (2.5
mg, 0.0174 mmol). CD2Cl2 (0.5 mL) was vacuum transferred into the
NMR tube. After about 1 h at room temperature, the NMR spectra
showed the complete conversion of Cp*Ir(PMe3)(CF2CF3)D into Cp*Ir-
(PMe3)(CDFCF3)Cl as a 2.5:1 mixture of a two diastereomers. The
solvent was pumped down, and the yellow residue was extracted into
hexane. Evaporation of hexane afforded a yellow solid (9.3 mg, 99%).
Anal. Calcd for C15H24DClF4IrP (541.03): C, 33.30; H, 4.66. Found:
4JPH ) 2.0, J ) 1.0 Hz, 15H, C5Me5), 1.66 (dd JPH ) 10.5, J ) 1.0
2
Hz, 9H, PMe3); 19F NMR (CD2Cl2) δ -82.97 (d, JFF ) 12.9 Hz, 3F,
4
CF3), -111.79 (ddd, 2JFF ) 273.3 Hz, 4JFP ) 24.5 Hz, 3JFF ) 17.4 Hz,
2
3
3
1F, â-CF), -116.24 (ddd, JFF ) 273.3 Hz, JFH ) 38.5 Hz, JFF
)
2
3
17.4 Hz, 1F, â-CF), -187.68 (ddq, JFH) 47.0 Hz, JFF ) 17.4 Hz,
3JPF ) 16.1 Hz, 1F, CHF); 31P{1H}NMR (CD2Cl2) δ -41.95 (dd, J )
16.7, 16.1 Hz, PMe3).
Reaction of 7a with 2,6-Lutidinium Tetrakis{3,5-bis(trifluoro-
methyl)phenyl}borate. The solid reagents were weighed into a
J-Young NMR tube in the drybox and CD2Cl2 (0.5 mL) added by
C, 33.46; H, 4.69.
2
The kinetic diastereomer: 2H NMR (CH2Cl2) δ 6.43 (br d, JFD
)
1
1
4
syringe. An immediate color change to yellow was observed. The H
8, CDF); H NMR (CD2Cl2) δ 1.68 (d, JPH ) 2, 15H, Cp*), 1.57 (d,
2JPH ) 11, 9H, PMe3); 19F NMR (CD2Cl2) δ -71.0 (br d, JFF ) 15,
3
NMR spectrum (after 10 min) indicated the reaction was complete,
with apparent formation of [Cp*Ir(PMe3)(CHFC2F5)]+[B{3,5-C6H3-
3
3
2
3F, CF3), -192.0 (br dqt, JPF ) 22, JFF ) 15, JDF ) 8, 1F, CDF);
(CF3)2}4]-. H NMR (CD2Cl2): δ 8.06 (t, J ) 8.1 Hz, 1H, p-py-H),
1
31P{1H}NMR (CD2Cl2) δ -30.2 (br d, JPF ) 22, PMe3).
3
The thermodynamic diastereomer: 2H NMR (CH2Cl2) δ 6.59 (br d,
7.78 (br s, 8H, B-(o-Ar-H), 7.62 (br s, 4H, B-(p-Ar-H), 7.42 (d, J )
8.1 Hz, 1H, o-py-H), 6.82 (dd, J ) 44.4, 40.8 Hz, 1H, CHF), 2.70 (s,
6H, pyMe), 1.72 (d, J ) 2.1 Hz, 15H, C5Me5), 1.58 (d, J ) 10.8 Hz,
9H, PMe3). 19F NMR (CD2Cl2): δ -63.26 (s, B-Ar-CF3), -83.25
(s, CF3), -112.64 (br d, J ) 289.3 Hz, â-CF2), -116.53 (br d, J )
289.3 Hz, â-CF2), -193.05 (br m, CHF). 31P NMR (CD2Cl2): δ -24.37
(br m, PMe3). After 30 min, some 6a was observed by NMR, indicating
chloride abstraction from solvent.
2JFD ) 8, 1D, CDF); H NMR (CD2Cl2) δ 1.72 (d, JPH ) 2, 15H,
1
4
Cp*), 1.47 (d, 2JPH ) 11, 9H, PMe3); 19F NMR (CD2Cl2) δ -71.2 (br
dd, 3JFF ) 15, 4JPF ) 3, 3F, CF3), -193.6 (br dqt, 3JPF ) 17, 3JFF ) 15,
2JDF ) 8, 1F, CDF); 31P{1H}NMR (CD2Cl2) δ -33.2 (br dq, JPF
)
3
4
17, JPF ) 3, PMe3).
Cp*Ir(PMe3)(CHFCF2CF3)Cl (6a). A J-Young NMR tube was
charged with Cp*Ir(PMe3)(CF2CF2CF3)H (10 mg, 0.0174 mmol) and
2,6-lutidinium(H) chloride (5 mg, 0.0348 mmol). CD2Cl2 (0.4 mL) was
vacuum transferred into the NMR tube. After about 1 h at room
temperature, the NMR spectra showed the complete conversion of
Cp*Ir(PMe3)(CF2CF2CF3)H into Cp*Ir(PMe3)(CHFCF2CF3)Cl as a
2.4:1 mixture of two diastereomers. The solvent was pumped down,
and the yellow residue was extracted into hexane. Evaporation of hexane
afforded a yellow solid (10 mg, 99%). The diastereomeric ratio changed
over the reaction course, from 3.7:1 to 2.4:1 Anal. Calcd for C16H25-
General Procedure for NMR Kinetic Experiments. All solid
reagents were weighed into a J-Young NMR tube in the drybox. For
low-temperature reactions, the sealed tube was then taken out of the
box and opened in a liquid nitrogen-cooled Schlenk flask under inert
atmosphere. Precooled NMR solvent (0.5-0.6 mL) was then added
by syringe and the mixture maintained at low temperature until injection
into the NMR probe. For reactions at room temperature or above, the
NMR solvent was added in the drybox. An internal integration standard
of 1,3,5-trimethoxybenzene was used. Concentration versus time plots
were simulated using the IBM chemical kinetics simulator.84
ClF6IrP (590.01): C, 32.57; H, 4.27. Found: C, 32.68; H, 4.09.
2
The kinetic diastereomer: 1H NMR (CD2Cl2) δ 6.57 (ddd, JFH
)
3
3
4
47, JFH ) 38, JPH ) 5, 1H, CHF), 1.66 (d, JPH ) 2, 15H, C5Me5),
1.56 (d, JPH ) 11, 9H, PMe3); 19F NMR (CD2Cl2) δ -83.13 (d, JFF
) 12, 3F, CF3), -115.26 (dd, 2JFF ) 273, 3JFF ) 20, 1F, â-CF), -118.09
(ddd, 2JFF ) 273, 3JFH ) 38, 3JFF ) 18, 1F, â-CF), -195.95 (br s, 1F,
Acknowledgment. R.P.H is grateful to the National Science
Foundation for generous financial support and to Professors Seth
Brown (University of Notre Dame), Chuck Casey (University
of Wisconsin, Madison), and Robert Ditchfield (Dartmouth
College) for extremely helpful discussions.
2
4
CHF); 31P{1H}NMR (CD2Cl2) δ -30.56 (d, JPF ) 20, PMe3).
3
The thermodynamic diastereomer: 1H NMR (CD2Cl2) δ 6.75 (ddd,
2JFH ) 48, JFH ) 40, JPH ) 3, 1H, CHF), 1.72 (d, JPH ) 2, 15H,
3
3
4
C5Me5), 1.45 (d, 2JPH ) 10, 9H, PMe3); 19F NMR (CD2Cl2) δ -82.94
JA0545012
9
15594 J. AM. CHEM. SOC. VOL. 127, NO. 44, 2005