Catalytic fluorination by halide exchange with 16-electron ruthenium(II) complexes. X-ray structure of [Tl(μ-F)2Ru(dppe)2]PF6

Article abstract of DOI:10.1021/om010288g

The 16-electron ruthenium(II) complexes [RuCl(dppe)₂]PF₆ (2; dppe = 1,2-bis(diphenylphosphino)ethane), [RuCl(chiraphos)₂]PF₆ (3; chiraphos = (S,S)-3,4-bis(diphenylphosphino)butane), and [RuCl(PNNP)]PF₆ (4; PNNP = (1S,2S)-N,N′-bis[2-(diphenylphosphino)-benzylidene]diaminocyclohexane) catalyze the nucleophilic fluorination of activated alkyl halides with a catalyst loading as low as 1 mol %. The alkyl halides (CH₃)₃CX (X = Br, 5c; X = I, 5d), Ph₂CHBr (6c), and PhCH(Me)Br (7c) are converted to the fluoro analogues in the presence of TlF as the fluoride source. Yields are between 31 and 83%. The chiral complex 4 converts 7c to PhCH(Me)F (7a) with 49% yield after 24 h. At 1% conversion, 7a is nonracemic (16% ee), which indicates that kinetic resolution occurs, albeit at a low level. The fluorination of 1,2-dibromo-1,2,3,4-tetrahydronaphthalene (8c) is highly regioselective and gives 1-fluoro-2-bromo-1,2,3,4-tetrahydronaphthalene (8a) in 68% yield. The difluoro-bridged thallium adduct [Tl(μ-F)₂Ru(dppe)₂]PF₆ (9) was observed by ³¹P NMR during catalysis with 2 and independently prepared by reaction of 2 with TlF (2 equiv). Complex 9 was characterized by ¹H, ³¹P, ¹⁹F, and ²⁰⁵Tl NMR spectroscopy, as well as by X-ray diffraction.

Full text of DOI:10.1021/om010288g

3472
Organometallics 2001, 20, 3472-3477
Ca ta lytic F lu or in a tion by Ha lid e Exch a n ge w ith
16-Electr on Ru th en iu m (II) Com p lexes. X-r a y Str u ctu r e of
[Tl(µ-F )2Ru (d p p e)2]P F 6
Peter Barthazy, Antonio Togni,* and Antonio Mezzetti*
Laboratory of Inorganic Chemistry, Swiss Federal Institute of Technology, ETH Zentrum,
CH-8092 Zu¨rich, Switzerland
Received April 6, 2001
The 16-electron ruthenium(II) complexes [RuCl(dppe)2]PF6 (2; dppe ) 1,2-bis(diphe-
nylphosphino)ethane), [RuCl(chiraphos)2]PF6 (3; chiraphos ) (S,S)-3,4-bis(diphenylphos-
phino)butane), and [RuCl(PNNP)]PF6 (4; PNNP ) (1S,2S)-N,N′-bis[2-(diphenylphosphino)-
benzylidene]diaminocyclohexane) catalyze the nucleophilic fluorination of activated alkyl
halides with a catalyst loading as low as 1 mol %. The alkyl halides (CH3)3CX (X ) Br, 5c;
X ) I, 5d ), Ph2CHBr (6c), and PhCH(Me)Br (7c) are converted to the fluoro analogues in
the presence of TlF as the fluoride source. Yields are between 31 and 83%. The chiral complex
4 converts 7c to PhCH(Me)F (7a ) with 49% yield after 24 h. At 1% conversion, 7a is
nonracemic (16% ee), which indicates that kinetic resolution occurs, albeit at a low level.
The fluorination of 1,2-dibromo-1,2,3,4-tetrahydronaphthalene (8c) is highly regioselective
and gives 1-fluoro-2-bromo-1,2,3,4-tetrahydronaphthalene (8a ) in 68% yield. The difluoro-
bridged thallium adduct [Tl(µ-F)2Ru(dppe)2]PF6 (9) was observed by 31P NMR during catalysis
with 2 and independently prepared by reaction of 2 with TlF (2 equiv). Complex 9 was
1
characterized by H, 31P, 19F, and 205Tl NMR spectroscopy, as well as by X-ray diffraction.
In tr od u ction
halides. A recent example is the stoichiometric reaction
of [PdF(Ph)(PPh3)2] with CH2Cl2 to give CH2ClF and
CH2F2.6 A further example of stoichiometric fluorination
has recently been reported from our laboratories and
involves the 16-electron ruthenium(II) complex [RuF-
(dppp)2]+ (1c; dppp ) 1,3-bis(diphenylphosphino)et-
hane), obtained by reaction of the well-known chloro
analogue7 with TlF (Scheme 1a).8a Complex 1c reacts,
in fact, with tertiary alkyl halides to give the corre-
sponding R-F derivatives (Scheme 1b).8
We now find that the X/F exchange (X ) Br, I) in the
organic molecule can be combined with the halide
metathesis of [RuX(dppp)2]+. The resulting catalytic
reaction exploits TlF as the fluoride source and five-
coordinate ruthenium complexes as catalysts. Among
these complexes, there are 16-electron species of general
formula [RuClP2L2]+ (L ) P or N donor) that we have
recently used as catalysts for the epoxidation and
cyclopropanation of olefins (Chart 1).9
The use of transition-metal complexes as catalysts for
reaction where a new carbon-fluorine bond is formed
is still exceedingly rare, despite the importance of, for
example, organofluorine compounds in medicinal chem-
istry.1 One of us recently reported the first enantiose-
lective electrophilic fluorination of â-keto esters by the
N-F reagent F-TEDA catalyzed by Ti(TADDOLato)
complexes.2 While this reaction appears to proceed via
the formation of a coordinated substrate enolate that
is subsequently attacked by the N-F reagent, a direct
interaction between the metal and the fluorine atom to
be transferred is not required. Similarly, in the recently
reported first asymmetric ring opening of epoxides by
hydrofluorinating agents catalyzed by J acobsen’s chiral
[CrCl(salen)] complexes, the involvement of fluorochro-
mium species was considered as very improbable.3
In contrast to this, a variety of well-defined metal
fluoro complexes, low-valent fluoride salts,4 or high-
valent metal fluorides and oxofluorides5 have been
shown to act as fluoride transfer agents in stoichiometric
reactions, in particular with alkyl, acyl, and silyl
(5) (a) Dukat, W. W.; Holloway, J . H.; Hope, E. G.; Rieland, M. R.;
Townson, P. J .; Powell, R. L. J . Chem. Soc., Chem. Commun. 1993,
1429. (b) Holloway, J . H.; Hope, E. G.; Townson, P. J .; Powell, R. L. J .
Fluorine Chem. 1996, 76, 105.
(6) Grushin, V. V. Angew. Chem., Int. Ed. Engl. 1998, 37, 994. See
also: Fraser, S. L.; Antipin, M. Y.; Khroustslyov, V. N.; Grushin, V.
V. J . Am. Chem. Soc. 1997, 119, 4769. Pilon, M. C.; Grushin, V. V.
Organometallics 1998, 17, 1774.
* To whom correspondence should be addressed. E-mail: A.T.,
togni@inorg.chem.ethz.ch; A.M., mezzetti@inorg.chem.ethz.ch.
(1) See, for instance: (a) Davis, F. A.; Qi, H.; Sundarababu, G. In
Enantiocontrolled Synthesis of Fluoro-Organic Compounds; Solos-
honok, V. A., Ed.; Wiley: New York, 1999; p 1. (b) Hoffman, R. V.;
Tao, J . In Asymmetric Fluoroorganic Chemistry; Ramachandran, P.
V., Ed.; ACS Symposium Series 746; American Chemical Society:
Washington, DC, 2000; p 52. (c) Hiyama, T. In Organofluorine
Compounds; Yamamoto, H., Ed.; Springer: Berlin, 2000; p 137.
(2) Hintermann, L.; Togni, A. Angew. Chem., Int. Ed. 2000, 39, 4359.
(3) Bruns, S.; Haufe, G. J . Fluorine Chem. 2000, 104, 247.
(4) Taverner, S. J .; Heath, P. A.; Clark, J . H. New J . Chem. 1998,
655 and references therein.
(7) Bressan, M.; Rigo, P. Inorg. Chem. 1975, 14, 2286.
(8) (a) Barthazy, P.; Hintermann, L.; Stoop, R. M.; Wo¨rle, M.;
Mezzetti, A.; Togni, A. Helv. Chim. Acta 1999, 82, 2448. (b) Barthazy,
P.; Stoop, R. M.; Wo¨rle, M.; Togni, A.; Mezzetti, A. Organometallics
2000, 19, 2844.
(9) (a) Stoop, R. M.; Bauer, C.; Setz, P.; Wo¨rle, M.; Wong, T. Y. H.;
Mezzetti, A. Organometallics 1999, 18, 5691. (b) Stoop, R. M.; Mezzetti,
A. Green Chem. 1999, 39. (c) Stoop, R. M.; Bachmann, S.; Valentini,
M.; Mezzetti, A. Organometallics 2000, 19, 4117. (d) Bachmann, S.;
Furler, M.; Mezzetti, A. Organometallics 2001, 20, 2102.
10.1021/om010288g CCC: $20.00 © 2001 American Chemical Society
Publication on Web 07/10/2001

[Tl(µ-F)2Ru(dppe)2]PF6
Organometallics, Vol. 20, No. 16, 2001 3473
fluoride (5a ) is as low as 2%. The formation of (But)2O
(10% of starting 5c) suggests that adventitious water
cannot be removed completely although precautions are
taken (see Experimental Part). A control reaction, run
without adding complex 1b, showed that tert-butyl
bromide does not react with TlF under these conditions.
Sch em e 1
Next, we investigated the related complex [RuCl-
(dppe)2]+ (2; dppe ) 1,2-bis(diphenylphosphino)ethane),
which was recently reported by Morris.10 In the presence
of 2 as catalyst (10 mol %), tert-butyl bromide (5c)
reacted at room temperature with TlF (1.1 equiv) to
yield tert-butyl fluoride (5a , 60% yield). Some isobutene
(ca. 30%) was also formed (Table 1, run 2). The use of
tert-butyl iodide (5d ) substantially inhibits the elimina-
tion reaction, speeds up the reaction, and improves the
yield of tert-butyl fluoride (run 3). The chiral analogue
[RuCl(chiraphos)2]PF6 (3; chiraphos ) (S,S)-3,4-bis-
(diphenylphosphino)butane)9a gives similar results with
a catalyst loading of only 2 mol % (run 3). In the
presence of the dppe derivative 2, the benzylic bromide
1-bromo-1,1-diphenylmethane (6c) gave the fluoro de-
rivative Ph2CHF (6a ) in high yield after 24 h of reaction
time (run 5).
Ch a r t 1
Having chiral catalysts at our disposal, we tested
racemic 1-bromo-1-phenylethane (7c) as substrate. With
the (achiral) dppe derivative 2, 7c is fluorinated with
moderate efficiency (runs 6 and 7). In particular, the
catalyst loading can be reduced to 1 mol % without a
major loss in the yield and selectivity, but at the cost of
longer reaction times. The chiraphos derivative 3 is
more active and selective than 2 (run 8). However, GC
analysis on a chiral column indicated that racemic
1-phenyl-1-fluoroethane (7a ) is formed.
Sch em e 2
Thus, we tested the recently reported five-coordinate
complex [RuCl(PNNP)]PF6 (4; PNNP ) (1S,2S)-N,N′-
bis[2-diphenylphosphino)benzylidene]diaminocyclo-
hexane).9b-d Complex 4 gave 7a in moderate yield (49%,
run 9). More importantly, monitoring of the enantio-
meric excess of 1-phenyl-1-fluoroethane by GC during
the reaction indicated that the initially formed product
is nonracemic (Table 2). At 1% conversion, the enan-
tiomeric excess of 7a is 16% ee, as measured by GC on
a chiral column. The enantiomeric excess drops with
longer reaction times and reaches 3% at 100% conver-
sion after 24 h of reaction time. As several measure-
ments of the enantiomeric excess of the products
obtained with achiral 2 (as well as with chiral 3) gave
0 ( 0.1% ee, it is reasonable to assume a 3% ee as
significant.
The above data suggest that kinetic resolution of rac-
7c occurs to some extent. Possible mechanisms encom-
pass classic kinetic resolution (if an SN2 mechanism is
operative) or dynamic kinetic resolution.11 In the latter
case, fast racemization of the substrate would occur,
most probably due to an SN1 mechanism. Thus, the
enantiomeric excess of the unreacted substrate should
give an indication as to which mechanism occurs.
Unfortunately, we were not able to determine the
enantiomeric excess of 7c by means of GC on a chiral
column.
Resu lts a n d Discu ssion
The catalytic fluorination is depicted in Scheme 2,
together with the substrates used. Conditio sine qua non
for the success of the fluorination is that the reaction
solutions be protected from adventitious water and from
the contact with glass surfaces. Two methods were
developed. Small-scale reactions were run in an NMR
tube fitted with a Young valve and a Teflon insert,
which was shaken mechanically during the reaction
time (method A). Solution reactions were prepared in a
glovebox under purified nitrogen (see Experimental
Section). On a larger scale, the reaction solutions were
prepared and kept in Teflon vessels in a glovebox and
stirred magnetically (method B).
We started our investigation with tert-butyl bromide
(5c) as substrate and [RuCl(dppp)2]PF6 (1b)7 (10 mol
%) as the catalyst precursor. Owing to the low boiling
point of 5a (13 °C), the reaction was carried out in an
NMR tube fitted with a Young valve. The reaction of
tert-butyl bromide (5c) with TlF in the presence of 1b
is very slow at room temperature. Raising the temper-
ature to 50 °C accelerates the reaction, and 42%
conversion of 5c is obtained after 12 h (Table 1, run 1).
However, the main product is isobutene (15%), deriving
from the elimination of HBr, and the yield of tert-butyl
(10) Chin, B.; Lough, A. J .; Morris, R. H.; Schweitzer, C. T.;
D’Agostino, C. Inorg. Chem. 1994, 33, 6278.
(11) See, for instance: Keith, J . M.; Larrow, J . F.; J acobsen, E. N.
Adv. Synth. Catal. 2001, 343, 5 and ref 9 therein.

3474 Organometallics, Vol. 20, No. 16, 2001
Ta ble 1. Ca ta lytic F lu or in a tion a
Barthazy et al.
yield (%)
run
cat.
substrate
mol %
method
t (days)
conversn (%)
R-F
ether
olefin
1b
2
3
4
5
6
7
8
9
1b
2
2
3
2
2
2
3
4
5c
5c
5d
5c
6c
7c
7c
7c
7c
8c
10
10
10
2
10
10
1
1
1
10
a
a
a
a
a
a
b
b
b
a
0.5
5
0.2
5
1
4
8
2
1
3
42
100
100
94
93
51
53
86
100
94
2
63
84
57
83
31
34
76
49
68
10
9
8
7
2
15
28
8
31
3
nd
nd
nd
nd
traces
nd
nd
nd
c
10
4
a
Control reactions under the same conditions but without the catalyst showed no reaction for 5c and 8c, 6% conversion (6% yield) for
5d , 2% (1%) for 6c, and 13% (7%) for 7c. The abbreviation nd means not determined. At 50 °C. c Unknown byproducts.
b
Ta ble 2. Con ver sion , Yield , a n d En a n tiom er ic
Excess of 7a a s a F u n ction of Rea ction Tim e
d u r in g th e F lu or in a tion Ca ta lyzed by 4a
indication that a ruthenium complex is involved in the
key step of the X/F exchange (probably halide abstrac-
tion from the substrate) and not only as phase-transfer
agent and bromide (or iodide) scavenger.
t (h)
conversn (%)
yield (%)
ee (%)
[Tl(µ-F )2Ru (d p p e)2]P F 6. Monitoring of the (yellow)
catalytic solutions by means of 31P NMR spectroscopy
showed that the doubly fluoro-bridged TlF adduct [Tl-
(µ-F)2Ru(dppe)2]PF6 (9) is the only detectable metal-
containing species in the reaction catalyzed by [RuCl-
(dppe)2]PF6 (2). We independently prepared 9 by reaction
of 2 with TlF (2 equiv). The yellow solid was isolated
from the CH2Cl2 solutions by precipitation with 2-pro-
panol. The presence of the P4RuF2Tl unit is strongly
suggested by the AA′MM′QQ′X spin system observed
in the 19F, 31P, and 205Tl NMR spectra. The spectra are
not well-resolved (either in CDCl3 or in CH2Cl2), prob-
ably due to a dynamic process that is fast at room
temperature. Lowering the temperature (in CD2Cl2
solution) improves the resolution only in the case of the
31P NMR spectrum. However, the pattern observed is
unambiguous. The 19F spectrum (20 °C) consists of a
1
2
4
6
24
1
28
31
43
quant
1
4
10
19
62
16
13
7
2
3
a
Portions of the reaction solution was periodically sampled and
analyzed by GC (see Experimental Section).
Finally, trans-1,2-dibromo-1,2,3,4-tetrahydronaphtha-
lene (8c) was fluorinated regio- and stereoselectively at
the C(1) position in the presence of 2 (10 mol %) (run
10). The reaction gave trans-1-fluoro-2-bromo-1,2,3,4-
tetrahydronaphthalene as the major product (68% yield)
after 3 days. The elimination product (1,2-dihydronaph-
thalene) is present in traces, along with yet unidentified
products. The retention of configuration at C(1) in the
reaction of 8c can be explained by a neighboring-group
mechanism, that is, involving the rate-determining
formation of a bromonium intermediate. This would
exclude, at least in this case, an SN2 mechanism.
Other substrates did not react (or reacted very slug-
gishly) under the conditions described above. In par-
ticular, primary alkyl halides, such as 2-iodopropane
and cyclohexyl bromide, are completely unreactive. The
fact that the reactivity is higher with typical SN1
substrates suggests that substantial charge separation
is involved in the transition state of the halogen transfer
reaction. A further limitation is the poor tolerance for
functional groups. Protic substrates, such as alcohols
and carboxylic acids, do not react owing to the formation
of HF. Also, R-halogeno ketones are unreactive.
The catalytic system described herein requires shorter
reaction times and milder conditions than those of
traditional methods employing fluoride salts, such as
biphasic fluorination with spray-dried KF or phase-
transfer agents.12 Thus, nearly quantitative conversions
and selectivities up to 88% for the fluorinated products
are obtained. The present system also reacts with less
activated haloalkanes than required by silver and
copper florides.13 Moreover, kinetic resolution takes
place, albeit with a low efficiency, as racemic 6a gives
slightly enantiomerically enriched Ph(Me)CHF in the
reaction with the chiral catalyst 2. This is an additional
1
broad doublet of doublets at δ -296 with a J Tl,F
coupling constant of ca. 800 Hz. The signals are broad,
owing to the nonresolved coupling to the 203Tl and 205Tl
isotopomers (29.5 and 70.5% abundance, respectively).
1
The same value of the J Tl,F coupling constant is
observed for the broad triplet at δ 1055 in the 205Tl NMR
spectrum (20 °C), which arises from the coupling to the
two (chemically equivalent) F nuclei. The room-temper-
ature 31P NMR spectrum consists of the AA′MM′ part
of the spin system, which features overlapping signals
for the two pairs of chemically equivalent P atoms.
However, the chemical shifts of the PA, PA′ and PM, PM′
atoms shift differently upon lowering the temperature.
Thus, resolved signals are observed at -90 °C, that is,
a pseudo-quintet at δ 55.7 for the axial P atoms (J P,P′
=
19 Hz, J P,F = 19 Hz), and a doublet of triplets at δ 54.4
for the equatorial ones. The attribution of the latter
signal to the equatorial phosphines is based on the large
apparent J P,F coupling constant of 146 Hz, which is the
sum J P,F + J P,F′, the trans and cis P-Ru-F coupling
constants. The J P,F + J P,F′ coupling constant is detected
also in the 31P-coupled 19F NMR spectrum. The 31P-
decoupled 19F NMR spectrum shows only the doublet
due to the coupling to thallium.
An X-ray study supports the formulation of 9 (Figure
1, Table 3). The complex cation is approximately C2
symmetric with a pseudo-binary axis through the Ru
and Tl atoms. As a consequence, the Ru(µ-F)2Tl ring is
symmetrical and planar within (0.026 Å. The Ru-F
(12) Rock, M. H. Methoden Organische Chemie (Houben-Weyl), 4th
ed.; Thieme: Stuttgart, Germany, 1986; Vol. E10b/1, pp 50-65.
(13) Yoneda, N.; Fukukara, T.; Yamagishi, K.; Suzuki, A. Chem. Lett.
1987, 1675. See also ref 1b, p 192.

[Tl(µ-F)2Ru(dppe)2]PF6
Organometallics, Vol. 20, No. 16, 2001 3475
Sch em e 3
ages (X ) halide) are very rare,16 and to the best of our
knowledge, 9 is the first µ-fluoro-bridged Tl adduct of a
transition metal reported so far.
It should be noted that the reaction of the dppp
analogue [RuCl(dppp)2]+ (1b) with TlF gives quantita-
tively the dark red, five-coordinate species [RuF(d-
ppp)2]+ (1a ) instead of a thallium adduct.8 We suggest
that this is because the dppp ligand is bulkier than
dppe, which stabilizes a coordinatively unsaturated
species.17 In contrast, the hypothetical [RuF(dppe)]+ is
less stable, and its TlF adduct 9 is formed instead. In
support of this interpretation, the reaction of [RuCl-
(dppe)2]+ (2) with 1 equiv of TlF yields 9 in 50% yield
and unreacted 2 (50% of starting), which supports the
stepwise reaction shown in Scheme 3.
F igu r e 1. ORTEP view of [Ru(dppe)2(µ-F)2Tl]+ (9) (30%
probability ellipsoids).
Ta ble 3. Selected Bon d Dista n ces (Å) a n d An gles
The reactivities of complexes 9 and 1a with alkyl
bromides or iodides are completely analogous. Indeed,
the reaction of the thallium adduct 9 with Ph3C-Cl (2
equiv) gives Ph3C-F and [RuCl(dppe)2]PF6 in nearly
quantitative yield within 12 h. Also, 9 reacts with 1,3-
diphenylallyl bromide or tert-butyl iodide to give the
fluoro alkyl and [RuX(dppe)2]+ quantitatively after 24
h of reaction time.18 As stated above, complex 9 is the
only phosphine-containing species observed in solution
by 31P NMR spectroscopy during the catalytic fluorina-
tion. Similarly, the reaction solutions containing [RuCl-
(chiraphos)2]PF6 (3) as catalyst are yellow, and their 31P
NMR spectra suggest that [Tl(µ-F)2Ru(chiraphos)2]PF6
is present. In contrast, the reaction solutions of dppp
derivative 1b are dark red during catalysis, which
suggests that [RuF(dppp)2]+ is the main species in
solution.
(d eg) for [Tl(µ-F )2Ru (d p p e)2]P F 6 (9)
Tl(1)-F(1)
F(1)-Ru(1)
Ru(1)-P(1)
Ru(1)-P(3)
F(1)‚‚‚H(18)
2.419(7)
2.112(7)
2.351(3)
2.363(3)
2.18
Tl(1)-F(2)
F(2)-Ru(1)
Ru(1)-P(2)
Ru(1)-P(4)
F(2)‚‚‚H(42)
2.419(8)
2.119(7)
2.306(3)
2.299(3)
2.29
F(1)-Tl(1)-F(2)
Ru(1)-F(1)-Tl(1)
F(1)-Ru(1)-P(1)
F(1)-Ru(1)-P(2)
F(1)-Ru(1)-P(3)
F(1)-Ru(1)-P(4)
P(1)-Ru(1)-P(2)
P(1)-Ru(1)-P(4)
P(2)-Ru(1)-P(4)
66.7(2)
107.8(3)
85.9(2)
95.5(2)
84.9(2)
168.0(2)
83.80(9) P(1)-Ru(1)-P(3)
104.6(1)
91.5(1)
F(1)-Ru(1)-F(2)
Ru(1)-F(2)-Tl(1)
F(2)-Ru(1)-P(1)
F(2)-Ru(1)-P(2)
F(2)-Ru(1)-P(3)
F(2)-Ru(1)-P(4)
77.9(3)
107.5(3)
84.2(2)
166.7(2)
86.2(2)
97.0(2)
167.9(1)
104.9(1)
83.9(1)
P(2)-Ru(1)-P(3)
P(3)-Ru(1)-P(4)
F(2)‚‚‚H(42)-C(42) 133.4
F(1)‚‚‚H(18)-C(18) 138.4
bond lengths are significantly longer in 9 (average 2.12
Å) than in the closely related cis-[RuF2(dppp)2] (average
2.06 Å),8b as expected on going from terminal to bridged
fluorides. The Tl-F distances (both 2.42 Å) can only be
compared to that in a Tl(III) fluoro porphyrin (2.441(6)
Å),14 as no other TlI-F distances are known. Despite
the presence of the (µ-F)2 bridge, the F(1)-Ru-F(2)
angle in 9 (77.9(3)°) is similar to that in cis-[RuF2-
(dppp)2] (78.2(1)°). The closest nonbonded contacts of the
Tl atom are in the range 3.30(2)-3.79(2) Å and involve
the carbon atoms of the axial phenyl rings (C(5)-C(10)
and C(29)-C(34)), which form a pocket around the RuF2
fragment. The closest F atom of the [PF6]- anion, F(4),
is 3.96(2) Å away from Tl.
Further features of 9 are the short H‚‚‚F nonbonded
contacts between the fluoro ligands and the ortho H
atoms of both equatorial phenyl rings (Figure 1). The
shortest F‚‚‚H distances (based on idealized H positions
with C-H ) 1.00 Å) are 2.18 and 2.29 Å, which is
significantly shorter than the sum of the van der Waals
radii of F and H (ca. 2.7 Å) (Table 3). Similar contacts
have been already observed in fluoro complexes of
ruthenium and have been attributed to C-H‚‚‚F hy-
drogen bonds.8b,15 Complexes containing Ru-X-Tl link-
Con clu sion
The observation of enantioselection, albeit at a low
level, in the reaction of PhCHBr in the presence of the
chiral complex 4 strongly supports the involvement of
the metal complex in the reaction. The stoichiometric
reactions indicate that both [RuF(P-P)2]+ and [Tl(µ-
F)2Ru(P-P)2]+ react with alkyl halides. At this stage,
we cannot conclude which of these two species is the
(15) (a) Coleman, K. S.; Fawcett, J .; Holloway, J . H.; Hope, E. G.;
Russell, D. R. J . Chem. Soc., Dalton Trans. 1997, 3557. (b) Cockman,
R. W.; Ebsworth, E. A. V.; Holloway, J . H.; Murdoch, H.; Robertson,
N.; Watson, P. G.; In Inorganic Fluorine Chemistry; Thrasher, J . S.,
Strauss, S. H., Eds.; ACS Symposium Series 555; American Chemical
Society: Washington, DC, 1994.
(16) (a) Bianchini, C.; Masi, D.; Linn, K.; Mealli, C.; Peruzzini, M.;
Zanobini, F. Inorg. Chem. 1992, 31, 4036. (b) Blake, A. J .; Christie, R.
M.; Roberts, Y. V.; Sullivan, M. J .; Schro¨der, M.; Yellowlees, L. J . J .
Chem. Soc., Chem. Commun. 1992, 848. (c) Kahwa, I. A.; Miller, D.;
Mitchel, M.; Fronczek, F. R.; Goodrich, R. G.; Williams, D. J .;
O’Mahoney, C. A.; Slawin, A. M. Z.; Ley, S. V.; Groombridge, C. J .
Inorg. Chem. 1992, 31, 3963.
(17) For a discussion of the effects that stabilize the five-coordinate
species [RuX(dppe)2]+ (X ) Cl, Br, I) and [RuX(dppp)2]+ (X ) F, Cl,
Br, I), see: Barthazy, P.; Broggini, D.; Mezzetti, A. Can. J . Chem. 2001,
79, 904.
(18) This protocol yields bromo and iodo complexes of the type [RuX-
(P-P)2]+ (X ) Br, I), avoiding anion metathesis (often an untidy
reaction). For an application, see ref 17.
(14) Coutsolelos, A. G.; Orfanopoulos, M.; Ward, D. L. Polyhedron
1991, 10, 885.

3476 Organometallics, Vol. 20, No. 16, 2001
Barthazy et al.
[RuCl2(dppe)2]21 with TlPF6. The NMR spectra were recorded
on Bruker Avance 250 (for 1H (250 MHz), 31P (101 MHz), and
Tl (144 MHz)) and 300 (19F, 282 MHz) spectrometers. Chemical
shifts δ are in ppm relative to internal SiMe4 (1H), to external
85% H3PO4 (31P), to external [Tl(OH2)6]+ (Tl), and to external
CFCl3 (19F). FAB mass spectra were measured by the analyti-
cal service of Laboratorium fu¨r Organische Chemie of the ETH
Zu¨rich on a ZAB VSEQ instrument. The stoichiometric reac-
tions of complex 9 with Ph3C-Cl, 1,3-diphenylallyl bromide,
and tert-butyl iodide were carried out and quantified as
described in ref 8b.
Ca ta lytic F lu or in a tion (Meth od A). TlF (33 mg, 150
µmol) was suspended in a solution of the organic substrate
(100 µmol) and the catalyst (10 µmol) in CDCl3 (0.4 mL) in a
Young NMR tube fitted with a Teflon liner. The NMR tube
was shaken mechanically at room temperature. Products were
identified by comparison of 1H and 19F NMR data with
literature values and quantified by integration of the 1H NMR
spectra of the reaction solutions. Complex 4 was prepared in
situ from [RuCl2(PNNP)] and TlPF6 as previously reported.9c
Ch a r t 2
active catalyst. If step b in Scheme 3 is reversible, both
1a and 9 could react with R-X via the common
intermediate [RuFL4]+ by elimination of TlF from 9. We
speculate that the 16-electron species [RuFL4]+ acts as
a Lewis acid and interacts with the alkyl halide R-F.
Indeed, electron-deficient (14- or 16-electron) complexes
are known to form adducts with dichloromethane.19
Coordination to the metal center could increase the
polarization of the C-X bond and promote the σ-bond
metathesis as sketched in A (Chart 2). Alternatively,
the five-coordinate ruthenium complex could act merely
as a shuttle of TlF in the organic solvent. The (soluble)
thallium adduct [Tl(µ-F)2Ru(P-P)2]+ could then act as
chloride scavenger and fluoride source as shown in B
(Chart 2). However, the X-ray structure of 9 suggests
that the Lewis acidity of the F-coordinated Tl+ ion is
low, as there are no close contacts to any other atom
except to the fluoro ligands (and two phenyl protons,
as discussed above). The latter observation somewhat
disfavors the hypothetical intermediacy of B in the
fluoride transfer as compared to A (Chart 2).20 Finally,
we explicitly stress that the pictorial descriptions in
Chart 2 are merely working hypotheses and do not
imply yet a mechanistic interpretation.
We have demonstrated the catalytic halide exchange
for selective nucleophilic fluorination of moderately
activated organic bromides and iodides by five-coordi-
nate ruthenium(II) complexes. The present reaction
opens the way to enantioselective nucleophilic fluorina-
tion. TlBr (or TlI) can be recycled. However, to avoid
the use of the toxic and expensive TlF, we are investi-
gating the use of different inorganic fluoride sources,
as well as studying the mechanistic aspects of the X/F
exchange, including the mechanism of enantioselection.
Reaction with ter t-Bu tyl Br om ide (5c). 1H NMR (CDCl3):
δ 1.83 (s, 9H, C(CH3)3, 5c); 1.38 (lit.22 1.30, d, 9H, J F,H ) 21.1
Hz, lit. 20,22 F-C(CH3)3, 5a ); 4.66 (lit.23 4.55, m, 2H, (CH3)Cd
CH2)); 1.75 (t, 6H, J H,H ) 1 Hz, (CH3)CdCH2)); 1.28 (lit.24 1.27,
s, (CH3)3COC(CH3)3). 19F NMR (CDCl3): δ -131.0 (lit.25 -132,
10 lines, 1F, J F,H ) 21.1 Hz, lit.25 21, C-F, 5a ).
1
Rea ction w ith ter t-Bu tyl Iod id e (5d ). H NMR (CDCl3):
δ 1.95 (s, 9H, C(CH3)3, 5d ), data for 5a are as reported above
for 5c.
1
Rea ction w ith 6c. H NMR (CDCl3): δ 6.31 (s, 1H, CBr-
H, 6c); 6.48 (d, 1H, J F,H ) 47.8 Hz, Ph2C(F)-H, 6a ); 5.40 (lit.26
5.38, s, 1H, Ph2CHOC(H)Ph2). 19F NMR (CDCl3): δ -166.9
(lit.25 -169, d, 1F, J F,H ) 46.6 Hz, lit.25 48, Ph2C(H)-F, 6a ).
Rea ction w ith 7c. 1H NMR (CDCl3): 7c: δ 7.26-7.47 (lit.
7.35, m, 5H, arom), 5.23 (lit. 5.22, q, 1H, J H,H ) 6.9 Hz, CBr-
H), 2.06 (lit. 2.0, d, 3H, J H,H ) 6.9 Hz, CH3).27 7a : 1H NMR
(CDCl3): δ 7.18-7.40 (lit. 7.45, m, 5H, arom), 5.64 (lit. 5.55,
dq, 1H, J H,H ) 6.4 Hz, J F,H ) 47.6 Hz, CF-H), 1.66 (lit. 1.60,
3H, dd, J H,H ) 6.4 Hz, J F,H ) 23.9 Hz, CH3).28 19F NMR
3
(CDCl3): δ -167.4 (lit. -167.5, sextet, 1F, J F,H ) 24 Hz, lit.
2
24 Hz, J F,H ) 48 Hz, lit. 48 Hz, C-F, 7a ).28
1
Rea ction w ith 8c. H NMR (CDCl3): 8c: δ 7.12-7.35 (m,
4H, arom), 5.66 (s, 1H, C(1)-H), 4.95 (s, 1H, C(2)-H), 3.21-
3.36 (m, 1H, CH2), 2.76-3.08 (m, 2H, CH2), 2.15-2.23 (m, 1H,
CH2). 8a : 1H NMR (CDCl3): δ 7.14-7.45 (lit. 7.1-7.4, m, 4H,
arom), 5.62 (lit. 5.58, dd, 1H, J H,H ) 5 Hz, J F,H ) 51 Hz, CF-
H), 4.61 (lit. 4.55, m, 1H, CBr-H), 2.65-3.35 (lit. 2.6-3.2, 2H,
m, CH2), 2.12-2.62 (lit. 2.0-2.6, 2H, m, CH2).29 19F NMR
Exp er im en ta l Section
2
(CDCl3): δ -145.2 (lit.29 -146.0, doublet, 1F, J F,H ) 51 Hz,
Gen er a l Con sid er a tion s. All operations were carried out
under argon using standard Schlenk techniques or in a
glovebox (Braun AG) under purified nitrogen. All reagents and
solvents were Fluka purissimum grade or had comparable
purity. Solvents (including deuterated ones for NMR spectros-
copy) and liquid reagents were distilled before use and/or dried
over molecular sieves. As a variation of the published proce-
dure,10 [RuCl(dppe)2]PF6 was prepared by reacting trans-
C-F, 8a ).
Ca ta lytic F lu or in a tion of 6a (Meth od B). All manipula-
tions were performed in a glovebox to exclude rigorously traces
of moisture. TlF (1.1 mmol) was added to a solution of 7c (1
mmol), decane (as internal standard), and the catalyst (10
µmol) in CH2Cl2 (5 mL) in a 15 mL Teflon vessel fitted with a
Teflon-coated magnetic bar. The suspension was stirred at
(21) Chatt, J .; Hayter, R. G. J . Chem. Soc. 1961, 896.
(22) Olah, G. A.; Baker, E. B.; Evans, J . C.; Tolgyesi, W. S.; McIntyre,
J . S.; Bastien, I. J . J . Am. Chem. Soc. 1964, 86, 1360.
(23) White, E. H.; Reefer, J .; Erickson, R. H.; Dzadzic, P. M. J . Org.
Chem. 1984, 49, 4872.
(24) King, J . F.; Lam, J . Y. L.; Dave, V. J . Org. Chem. 1995, 60,
2831.
(19) For examples of complexes containing coordinated chloro al-
kanes that are not stabilized by the chelate effects, see: (a) Fang, X.
G.; Scott, B. L.; J ohn, K. D.; Kubas, G. J . Organometallics 2000, 19,
4141. (b) Huhmann-Vincent, J .; Scott, B. L.; Kubas, G. J . Inorg. Chem.
1999, 38, 115. (c) Huhmann-Vincent, J .; Scott, B. L.; Kubas, G. J . J .
Am. Chem. Soc. 1998, 120, 6808. (d) Huang, D.; Huffman, J . C.;
Bollinger, J . C.; Eisenstein, O.; Caulton, K. G. J . Am. Chem. Soc. 1997,
(25) Lai, C.; Kim, Y. I.; Wang, C. M.; Mallouk, T. E.; J . Org. Chem.
1993, 58, 1393.
119, 7398. For
a recent example of Ag-coordinated CH2Cl2, see:
Fornie´s, J .; Mart´ınez, F.; Navarro, R.; Urriolabeitia, E. P. Organome-
tallics 1996, 15, 1813.
(20) We thank one of the reviewers for this suggestion and for the
observation that the fluoro ligand in 9 could deprotonate But-X in
view of its high basicity. However, we expect that the coordination to
thallium strongly decreases the basicity of fluoride, which also explains
the stability of the thallium adduct itself.
(26) Mizuno, H.; Matsuda, M.; Iino, M. J . Org. Chem. 1981, 46, 520.
(27) Venkatachalapathy, C.; Pitchumani, K. Tetrahedron 1997, 53,
2581.
(28) York, C.; Surya Prakash, G. K.; Olah, G. A.; Tetrahedron 1996,
52, 9.
(29) Shimizu, M.; Nakahara, Y.; Yoshioka, H. J . Chem. Soc., Chem.
Commun. 1989, 1881.

[Tl(µ-F)2Ru(dppe)2]PF6
Organometallics, Vol. 20, No. 16, 2001 3477
room temperature. Reaction control was performed by 1H, 31P,
and 19F NMR and gas chromatography. Spectroscopic data for
7a were identical with those in the literature. Relevant
([M - TlF]+, 16%), 899 ([RuH(dppe)2]+, 100%), 499 ([RuH-
(dppe)]+, 5%). Anal. Calcd for C52H48F8P5RuTl: C, 48.59; H,
3.76. Found: C, 48.71; H, 3.78.
1
features of H and 19F NMR spectra of the reaction solutions
X-r a y Str u ctu r e Deter m in a tion of [Tl(µ-F )2Ru (d p p e)2]
(9). Crystals were obtained by diffusion of hexane into a CH2-
Cl2 solution of 9 at room temperature. Crystal data: C52H48F8P5-
RuTl, yellow prism, Mr ) 1285.19, T ) 293 K, triclinic, P1, a
) 10.7584(3) Å, b ) 11.1670(3) Å, c ) 11.7810(2) Å, R ) 74.980-
(1)°, â ) 63.60°, γ ) 87.030(1)°, V ) 1221.09(5) Å3, F(000) )
632, Z ) 1, Dc ) 1.748 Mg m-3, µ(Mo KR) ) 3.835 mm-1, crystal
size 0.40 × 0.38 × 0.22 mm3, Siemens SMART platform with
CCD detector, normal focus molybdenum-target X-ray tube,
graphite monochromator, ω-scans, h ) -13 to +14, k ) -14
to +14, l ) -11 to +15; 8998 reflections for 1.89° < θ < 30.01°
(7135 unique). Unit cell dimension determination and data
reduction were performed by standard procedures, and an
empirical absorption correction (SADABS) was applied. The
structure was solved with SHELXS-96 using direct methods
and refined by full-matrix least squares on F2 with anisotropic
displacement parameters for all non-H atoms. The crystal is
a 50:50 racemic twin, and the absolute structure parameter
was 0.465(8). Hydrogen atoms were introduced at calculated
positions and refined with the riding model and individual
isotropic thermal parameters. R1 ) 0.0522 and wR2 ) 0.1256
(6430 unique reflections with I > 2σ(I)), R1 ) 0.0634, and wR2
) 0.1401 (all data), GOF ) 1.172. The maximum and mini-
mum difference peaks were 1.651 and -1.300 e Å-3, and the
largest and mean values of ∆/σ were 0.063 and 0.02. Selected
bond lengths and angles are given in Table 3. Atomic coordi-
nates, anisotropic displacement coefficients, and an extended
list of interatomic distances and angles are available as
Supporting Information.
and reference to the literature are given above. The reactions
were quantified by gas chromatography on a Fisons Instru-
ments GC 8000 gas chromatograph equipped with an Optima
δ-3 capillary column (30 m × 0.25 mm, 0.25 µm film) and on
a ThermoQuest Trace GC 2000 Series gas chromatograph
equipped with a Supelco Beta Dex capillary column (30 m ×
0.25 mm, 0.25 µm film). FID detectors were used for signal
detection on both chromatographs. Quantitative analyses were
performed using n-decane as internal standard. Response
factor Rf (1.75) of 1-bromo-1-phenylethane (7c) was determined
by injection of a solution of 7c and decane of known concentra-
tion. The response factor of 7a was determined by comparison
of the area of the product with the integrated intensities from
a one-pulse 1H NMR spectrum of the same solution (δ 1.24
for 1-fluoro-1-phenylethane, 7a ). The temperature program
used for achiral GC analysis was a 5 min isotherm 50 °C, then
5 °C/min until 200 °C. Retention times were 10.8 min for
1-fluoro-1-phenylethane (7a ), 14.3 min for decane, and 19.7
min for 1-bromo-1-phenylethane (7c). The temperature pro-
gram used for chiral GC was isotherm (50 °C). The retention
time for 7a was 43.3 min for the first enantiomer and 46.9
min for the second one.
[Tl(µ-F )2Ru (d p p e)2]P F 6 (9). A CH2Cl2 suspension (30 mL)
of [RuCl(dppe)2]PF6 (1.134 g, 1.05 mmol) and TlF (530 mg, 2.38
mmol) was stirred for 3 h at room temperature. The thallium
salts were filtered off, and the solution was transferred
dropwise into hexane (100 mL) with vigorous stirring. The
light yellow precipitate was filtered off and dried under
1
vacuum: 1.095 g (81%). H NMR (CDCl3): δ 7.9-8.0 (m, br,
4H, Ph H), 7.6-7.7 (m, 8H, Ph H), 7.4-7.5 (m, 8H, PhH), 7.38
(t, 6H, Ph H, J H,H ) 7.3 Hz), 7.28 (t, 2H, Ph H, J H,H ) 7.4 Hz),
7.07 (t, 4H, Ph H, J H,H ) 7.6 Hz), 6.86 (t, 4H, Ph H, J H,H ) 7.6
Hz), 5.86 (t, 4H, Ph H, J H,H ) 8.1 Hz), 3.0-3.2 (m, 2H, PCH2),
2.2-2.4 (m, 4H, PCH2), 1.6-1.8 (m, 2H, PCH2). 31P NMR (CD2-
Ack n ow led gm en t. This work was supported by the
ETH Zu¨rich. We thank Dr. H. Ru¨egger for assistance
with the 19F and 205Tl NMR spectra and D. Broggini and
Dr. M. Wo¨rle for the X-ray crystallographic study.
Cl2, 183 K): δ 55.7 (pseudo-quintet, 2P, J P,P′ = 19 Hz, J P,F′
=
19 Hz, Pax), 54.4 (d × t, 2P, J P,P′ = 19 Hz, J P,F′ = 145 Hz, Peq),
Su p p or tin g In for m a tion Ava ila ble: Details of the X-ray
study of 9. This material is available free of charge via the
Internet at http://pubs.acs.org.
-143.1 (septet, 1P, PF6, J P,F ) 710 Hz). 19F NMR (CDCl3, 293
K): δ 74.5 (d, 6F, J P,F ) 710 Hz, PF6), -296 (br d, 2F, J Tl,F
)
800 Hz, RuF2). 205Tl NMR (CDCl3, 293 K): δ 1055 (br t, 1Tl,
J Tl,F ) 800 Hz, F2Tl). MS (FAB+) m/z: 1141 ([M]+, 12%), 917
OM010288G