1,2-bis{(pentafluorophenyl)phenylphosphino}ethane: A probe for configurational stability in three-legged piano stool complexes

Article abstract of DOI:10.1021/om050893+

The pentafluorophenyl-substituted diphosphine (C₆F ₅)PhPCH₂CH₂PPh(C₆F₅) has been prepared, as a 1:1.7 mixture of rac (1a) and meso (1b) isomers, in four steps from dppe. The reaction between [Cp*RhCl(μ-Cl)]₂ and 1 in the presence of tetrafluoroborate yielded a mixture of racemic diastereoisomers of [Cp*RhCl(κP,κP-1a)][BF₄] (4a·BF₄) and trans and cis isomers of [Cp*RhCl(κP, κP-1b)][BF₄] (4b·BF₄ and 4c·BF ₄, respectively). On addition of Proton Sponge, 4a and 4c, in which at least one pentafluorophenyl group is close to the pentamethylcyclopentadienyl ligand, underwent rapid dehydrofluorinative carbon-carbon coupling giving trans- and cis-[{η⁵,κP,κP-C₅Me ₄CH₂C₆F₄-2-PPhCH₂CH ₂PPh(C₆F₅)}RhCl]⁺ (5 · and 6), respectively. The latter underwent further dehydrofluorinative carbon-carbon coupling to give two isomers of [{η⁵,κP,κP-C ₅Me₃[CH₂C₆F₄-2-PPhCH ₂]₂}RhCl]⁺ (7). Isomerization of 4b to 4c was observed in chloroform and dimethlysulfoxide. Neither isomerization of 4a to 4b or 4c nor isomerization of 5 to 6 was observed at ambient or elevated temperature in dimethyl sulfoxide. The results provide the first evidence that complexes of η⁵,κP,κL-cyclopentadienyl-phosphine- donor ligands are configurationally stable at high temperature.

Full text of DOI:10.1021/om050893+

996
Organometallics 2006, 25, 996-1003
1,2-Bis{(pentafluorophenyl)phenylphosphino}ethane: A Probe for
Configurational Stability in Three-Legged Piano Stool Complexes
Mark Nieuwenhuyzen, Graham C. Saunders,* and E. C. M. Sarah Smyth
School of Chemistry, Queen’s UniVersity Belfast, DaVid Keir Building, Belfast, BT9 5AG, U.K.
ReceiVed October 17, 2005
The pentafluorophenyl-substituted diphosphine (C6F5)PhPCH2CH2PPh(C6F5) has been prepared, as a
1:1.7 mixture of rac (1a) and meso (1b) isomers, in four steps from dppe. The reaction between [Cp*RhCl-
(µ-Cl)]2 and 1 in the presence of tetrafluoroborate yielded a mixture of racemic diastereoisomers of
[Cp*RhCl(κP,κP-1a)][BF4] (4a‚BF4) and trans and cis isomers of [Cp*RhCl(κP,κP-1b)][BF4] (4b‚BF4
and 4c‚BF4, respectively). On addition of Proton Sponge, 4a and 4c, in which at least one pentafluorophenyl
group is close to the pentamethylcyclopentadienyl ligand, underwent rapid dehydrofluorinative carbon-
carbon coupling giving trans- and cis-[{η5,κP,κP-C5Me4CH2C6F4-2-PPhCH2CH2PPh(C6F5)}RhCl]+ (5
and 6), respectively. The latter underwent further dehydrofluorinative carbon-carbon coupling to give
two isomers of [{η5,κP,κP-C5Me3[CH2C6F4-2-PPhCH2]2}RhCl]+ (7). Isomerization of 4b to 4c was
observed in chloroform and dimethlysulfoxide. Neither isomerization of 4a to 4b or 4c nor isomerization
of 5 to 6 was observed at ambient or elevated temperature in dimethyl sulfoxide. The results provide the
first evidence that complexes of η5,κP,κL-cyclopentadienyl-phosphine-donor ligands are configurationally
stable at high temperature.
both moieties remain bound with inversion occurring via a two-
or four-legged piano stool intermediate (Scheme 1, mechanism
B) and involving dissociation of ligand X.10,13 Either or both
of these may operate depending on the complex and the
conditions.
Introduction
The study of the epimerization of three-legged piano stool
complexes containing stereogenic metal centers, [(ηn-CnRn)-
MX(LL′)]m+ (n ) 5 or 6),1-6 is important to the development
of their use as stereoselective catalysts.7,8 These complexes have
been shown to undergo inversion at the metal with widely
different rates, and two types of mechanism have been
proposed.2,9-13 One type involves hemidissociation, dissociation
of one of the ligating moietes of LL′ (Scheme 1, mechanism
A),8,12 or complete dissociation of LL′, and in the other type
Complexes of chelating ligands in which one moiety is labile
often undergo inversion at the metal more rapidly than those
of ligands that are nonlabile, and it is inferred that in such cases
mechanism A dominates.9 For complexes in which the chelating
ligand is nonlabile, inversion at the metal still occurs by
mechanism B. For example, inversion at the metal has been
observed in complexes of chelating diphosphines, which are
typically considered as nonlabile, such as [Cp*RuCl(Ph2-
PCHMeCH2PPh2)],14 [Cp*RhCl(Ph2PCHMeCH2PPh2)]+,10 and
[Cp*RhCl{Ph2PCH2CH2PPh(C5F4N-4)}]+,15 and this is pre-
sumed to occur by mechanism B. Calculations and mechanistic
studies reveal that the barrier to inversion of [(ηn-CnRn)M-
* Corresponding author. E-mail: g.saunders@qub.ac.uk.
(1) (a) Brunner, H. Eur. J. Inorg. Chem. 2001, 905. (b) Brunner, H.
Angew. Chem., Int. Ed. 1999, 38, 1194. (c) Brunner, H.; Ko¨llnberger, A.;
Burgemeister, T.; Zabel, M. Polyhedron 2000, 19, 1519. (d) Brunner, H.;
Grau, I.; Zabel, M. Organometallics 2004, 23, 3788.
(2) Brunner, H.; Ko¨llnberger, A.; Zabel, M. Polyhedron 2003, 2639.
(3) Ganter, C. Chem. Soc. ReV. 2003, 32, 130.
(4) Consiglio, G.; Morandini, F. Chem. ReV. 1987, 87, 761.
(5) Arena, C. G.; Calamia, S.; Faraone, F.; Graiff, C.; Tiripicchio, A. J.
Chem. Soc., Dalton Trans. 2000, 3149.
(LL′)]n+ is less than 15 kcal mol-1 16
However, to our
.
knowledge a mechanism involving phosphine dissociation has
not been unequivocally ruled out.
(6) Davenport, A. J.; Davies, D. L.; Fawcett, J.; Garratt, S. A.; Russell,
D. R. J. Chem. Soc., Chem. Commun. 1999, 2331.
In complexes in which the carbocylic ligand is joined to the
chelating ligand by a short, rigid linkage, giving an ηn,κL1,κL2
bound ligand, mechanism B is prevented. Brunner has postulated
that complexes of this type of ligand are configurationally stable
at the metal.17 Indeed, room-temperature configurational stability
has been confirmed for [{η6,κP,κN-C6H4(CH2CH2PPh2)(pyr)-
3}Ru(OH2)](OTf)2 (pyr ) bis(trifluoromethyl)pyrazole or cam-
phorpyrazole), but decomposition occurred on heating.18 No
other studies to confirm configurational stability have been
(7) (a) Carmona, D.; Cativiela, C.; Elipe, S.; Lahoz, F. J.; Lamata, L.
P.; Lo´pez-V´ıu, M. P.; Oro, L. A.; Vega, C.; Viguri, F. J. Chem. Soc., Chem.
Commun. 1997, 2351. (b) Faller, J. W.; Grimmond, B. J.; D’Allesi, D. G.
J. Am. Chem. Soc. 2001, 123, 2525. (c) Davenport, A. J.; Davies, D. L.;
Fawcett, J.; Garratt, S. A.; Russell, D. R. J. Chem. Soc., Dalton Trans.
2000, 4432.
(8) Standfest-Hauser, C.; Slugovc, C.; Mereiter, K.; Schmid, R.; Kirchner,
K.; Xiao, L.; Weissensteiner, W. J. Chem. Soc., Dalton Trans. 2001, 2989.
(9) Bellabarba, R. M.; Nieuwenhuyzen, M.; Saunders: G. C. Organo-
metallics 2002, 21, 5726.
(10) Carmona, D.; Lahoz, F. J.; Oro, L. A.; Lamata, L. P.; Viguri, F.;
San Jose´, E. Organometallics 1996, 15, 2961.
(11) Davies, D. L.; Fawcett, J.; Garratt, S. A.; Russell, D. R. J.
Organometallics 2001, 20, 2039.
(12) (a) de Klerk-Engels, B.; Groen, J. H.; Vrieze, K.; A. Mo¨ckel, A.;
Lindner, E.; Goubitz K. Inorg. Chim. Acta 1992, 195, 237. (b) Demerseman,
B.; Renouard, C.; Le Lagadec, R.; Gonzalez, M.; Crochet, P.; Dixneuf, P.
H. J. Organomet. Chem. 1994, 471, 229.
(14) Morandini, F.; Consiglio, G.; Straub, B.; Ciani, G.; Sironi, A. J.
Chem. Soc., Dalton Trans. 1983, 2293.
(15) Bellabarba, R. M.; Nieuwenhuyzen, M.; Saunders, G. C. Organo-
metallics 2003, 22, 1802.
(16) Ward, T. R.; Schafer, O.; Daul, C.; Hofmann, P. Organometallics
1997, 16, 3207.
(13) Alezra, V.; Bernardinelli, G.; Corminboeuf, C.; Frey, U.; Ku¨ndig,
E. P.; Merbach, A. E.; Saudan, C. M.; Viton, F.; Weber, J. J. Am. Chem.
Soc. 2004, 126, 4843.
(17) (a) Brunner, H.; Ko¨llnberger, A.; Mehmood, A.; Tsuno, T.; Zabel,
M. Organometallics 2004, 23, 4006. (b) Brunner, H.; Vale´rio, C.; Zabel,
M. New J. Chem. 2004, 24, 275.
10.1021/om050893+ CCC: $33.50 © 2006 American Chemical Society
Publication on Web 01/07/2006

A Probe for Configurational Stability
Organometallics, Vol. 25, No. 4, 2006 997
Scheme 1
Scheme 2
carried out at elevated temperature. However, if L1 is the
bridgehead and both L1 and L2 can dissociate, then inversion at
the metal is still possible. Thus, despite the three-point attach-
ment of the ligand, a mechanism for inversion at the metal can
be proposed. For ligands in which L1 is a stereogenic center,
inversion at the metal requires inversion at L1 (Scheme 2).
We are interested in η5,κP,κL-cyclopentadienyl-phosphine-
donor (Cp-PL) ligands as a means of providing chiral,
potentially configurationally stable metal complexes as precata-
lysts for organic transformations. This ligand type includes L
groups that are considered as nonlabile, such as phos-
phines,9,15,19,20 and that are labile, such as thioether.9 The latter
are of particular interest since configurational stability would
be combined with ligand hemilability,21 which provides a latent
coordination site at the metal, allowing a greater range of
Scheme 3
reactivity. For inversion at the metal to occur, the phosphine
must dissociate and inversion at phosphorus occur. Thus, if the
phosphine is insufficiently labile or the barrier to inversion at
the phosphorus atom of the dissociated phosphine is sufficiently
high, then configurational stability will result.
Although phosphine moieties of chelating Cp-P ligands are
expected to be typically nonlabile, their dissociation is not
without precedent.22 Furthermore, the synthetic strategy we have
developed for complexes of these ligands (Scheme 3) neces-
sitates a strongly electron-withdrawing polyfluoroaryl phosphine
substituent, which reduces the basicity and increases the cone
angle of the phosphine compared to the nonfluorinated analogue.
(18) (a) Therrien, B.; Ward, T. R. Angew. Chem., Int. Ed. 1999, 38, 405.
(b) Therrien, B.; Ko¨nig, A.; Ward, T. R. Organometallics 1999, 18, 1565.
(19) Atherton, M. J.; Fawcett, J.; Holloway, J. H.; Hope, E. G.; Karac¸ar,
A.; Russell, D. R.; Saunders: G. C. J. Chem. Soc., Dalton Trans. 1996,
3215.
(20) Fawcett, J.; Friedrichs, S.; Holloway, J. H.; Hope, E. G.; McKee,
V.; Nieuwenhuyzen, M.; Russell, D. R.; Saunders, G. C. J. Chem. Soc.,
Dalton Trans. 1998, 1477.
(22) (a) Kettenbach, R. T.; Butenscho¨n, H. New J. Chem. 1990, 14, 599.
(b) Foerstner, J.; Olbrich, F.; Butenscho¨n, H. Angew. Chem., Int. Ed. Engl.
1996, 35, 1234. (c) Yong, L.; Hofer, E.; Wartchow, R.; Butenscho¨n, H.
Organometallics 2003, 22, 5463.
(21) (a) Slone, C. S.; Weinberger, D. A.; Mirkin, C. A. Prog. Inorg.
Chem. 1999, 48, 233. (b) Braunstein, P.; Naud, F. Angew. Chem., Int. Ed.
2001, 40, 680.

998 Organometallics, Vol. 25, No. 4, 2006
Nieuwenhuyzen et al.
Scheme 4
Scheme 5
which is expected to possess spectroscopic properties different
from those of both the cis and trans isomers.9,19 The coupling
reaction is rapid in the presence of a suitable base or at high
temperature. Observation of a decrease in the concentration of
the complex of the Cp-L1P ligand and a concomitant increase
in the concentration of the complex of the CpdL1P ligand would
establish unequivocally that phosphine dissociation and inversion
at phosphorus have occurred. Ultimately this establishes whether
complexes of Cp-PL2 ligands are configurationally stable under
the conditions of investigation. It is noteworthy that this method
does not require enantiomeric enrichment. The latter method
was chosen for our study.
We wished to use a rhodium complex of a Cp-P1P2 ligand.
This is prepared by an intramolecular coupling of the η5-C5R5
and a polyfluoroaryl substituent on P1 disphosphine. Since P2
is also to bear a polyfluoroaryl substitutent, the diphosphine
selected for this study was (C6F5)PhPCH2CH2PPh(C6F5), 1. It
was expected that the rac isomer of 1 would provide a complex
suitable for the study, and that the meso isomer would produce
a complex of the double-linked CpdPP ligand and a complex
in which the cyclopentadienyl ring and diphosphine are not
linked. Here we report the synthesis of 1 and its rhodium piano
stool complexes and a determination of the configurational
stability of a Cp-P1P2 rhodium complex.
On the basis of past observations23 both of these effects are
expected to increase the lability of the phosphine moiety.
Although the barrier to inversion at phosphorus in phosphines
is typically high, 25 to 36 kcal mol-1 24
this is lowered by
,
electron-withdrawing substituents. For example, ∆Gq is 20 kcal
mol-1 for MeOP(Ph)CHMe2 and 18.9 kcal mol-1 for Me3-
25
SiP(Ph)CHMe2.26 The presence of a polyfluoroaryl substituent
is also expected to reduce the barrier to inversion at the
phosphorus.
We wished to determine experimentally whether rhodium
complexes of Cp-PL ligands can undergo inversion. Two
methods were considered. One involves enantiomeric enrich-
ment of a complex and measurement of its optical rotation or
CD spectrum over time. The other involves the investigation
of phosphine dissociation and inversion by employing a
phosphine with one polyfluoroaryl substituent as the L2 moiety
(Scheme 4). Two pairs of enantiomers exist for complexes of
these ligands. The cyclopentadienyl ring and polyfluoroaryl
substituent are either on the same side (cis) or on opposite sides
(trans) of the RhL1P plane. If phosphine dissociation and
inversion do occur, then an equilibrium of the cis and trans
pair of enantiomers would be established. Although the cis and
trans isomers are expected to have different spectroscopic
properties and thus be readily observed by 31P NMR spectros-
copy, it would not be possible to observe the isomerization if
the equilibrium has been established under the conditions of
the synthesis. However, in the case of the cis stereoisomers
intramolecular dehydrofluorinative carbon-carbon coupling can
occur to give a complex of a double-linked CpdL1P ligand,
Results and Discussion
The four-step synthesis of diphosphine 1 is outlined in
Scheme 5. Treatment of Ph2PCH2CH2PPh2 (dppe) with a 10-
fold excess of lithium in THF at 0 °C produced a yellow solution
of (Li+)2[PhPCH2CH2PPh]2- 27
, which on subsequent treatment
with chlorotrimethylsilane, then hexachloroethane, gave PhCl-
PCH2CH2PPhCl.28 Addition of an excess of C6F5MgBr in
diethyl ether gave 1, as a 1:1.7 mixture of rac (1a) and meso
(1b) isomers, which was isolated as a white solid in 54.5% yield.
(The identity of the major isomer was determined from the
spectroscopic data of [Cp*RhCl(κP,κP-1)][BF4] (vide infra).)
It is presumed that 1a is formed as a racemic mixture of the
RR and SS enantiomers. The ratio of isomers is in contrast to
that of 1.7:1 rac:meso reported for 1,2-bis[(2-quinolinemethyl)-
(23) (a) Kemmitt, R. D. W.; Nichols, D. I.; Peacock, R. D. J. Chem.
Soc., Chem. Commun. 1967, 599. (b) Kemmitt, R. D. W.; Nichols, D. I.;
Peacock, R. D. J. Chem. Soc., A 1968, 1898. (c) Kemmitt, R. D. W.; Nichols,
D. I.; Peacock, R. D. J. Chem. Soc., A 1968, 2149.
(27) Dogan, J.; Schulte, J. B.; Swiegers, G. F.; Wild, S. B. J. Org. Chem.
(24) Baechler, R. D.; Mislow, K. J. Am. Chem. Soc. 1970, 92, 3090.
(25) Pabel, M.; Willis, A. C.; Wild, S. B. Inorg. Chem. 1996, 35, 1244.
(26) Baechler, R. D.; Mislow, K. J. Am. Chem. Soc. 1971, 93, 773.
2000, 65, 951.
(28) Mu¨ller, G.; Klinga, M.; Leskela¨, M.; Rieger, B. Eur. J. Inorg. Chem.
2002, 2625.

A Probe for Configurational Stability
Organometallics, Vol. 25, No. 4, 2006 999
9
6JPF, and JPF coupling constants of 43.0, 33.0, 2.0, and 0 Hz,
respectively, for resonances centered at δ -26.56 for 1a and at
δ -26.43 for 1b. These coupling constants are similar to those
of (C6H3F2-2,6)2PCH2CH2P(C6H3F2-2,6)2 (3JPP 47.2, 3JPF 30.1,
6JPF 1.2, and 9JPF -0.4 Hz),20 and the chemical shifts are similar
to that of PPh2(C6F5) (δ -24.7).29 The 19F NMR spectrum shows
only four resonances. Those of the ortho fluorine atoms of 1a
and 1b are coincident at δ -129.68, as are those of the meta
fluorine atoms at δ -160.70. Two resonances for the para
fluorine atoms (δ -150.46 and -150.70) are consistent with
the two isomers of 1. Unfortunately to date attempts to separate
1a and 1b have been unsuccessful.
A small amount of a 1:1 mixture of isomers of the phosphine
monoxide 2, characterized by mass spectrometry and NMR
spectroscopy, was also obtained from the preparation of 1. The
31P{1H} NMR spectrum exhibits two sets of doublets of triplets
at ca. δ 30 and doublet of doublets at ca. δ -25, consistent
with two pairs of enantiomers. The former resonances are
assigned to the phosphine by comparison with those of 1 and
the latter to the phosphine oxide by comparison to those of the
phosphine dioxide, 3, formed by oxidation of an NMR sample
of 1 with hydrogen peroxide. The identity of 3 was confirmed
by mass spectrometry. Two resonances are observed in the 31P-
{1H} NMR spectrum consistent with meso and rac isomers.
Figure 1. (a) Experimental and (b) simulated 31P{1H} NMR
spectrum of 1.
phenylphosphino]ethane, which is synthesized by a similar
route.28 The identity of 1 was confirmed by elemental analysis,
high-resolution mass spectrometry, and NMR spectrosocopy.
The 31P{1H} NMR spectrum of 1 contains a complicated
asymmetric pattern, consistent with two overlapping patterns
arising from A parts of AA′X2X2′ spin systems (Figure 1). The
3
3
spectrum was simulated using relative values of the JPP, JPF,
Scheme 6

1000 Organometallics, Vol. 25, No. 4, 2006
Table 1. Selected NMR and Mass Spectral Data for 4‚BF4, 5‚BF4, 6‚BF4, and 7‚BF4
Nieuwenhuyzen et al.
a
δP (1JRhP/Hz)
δF
δH
LSIMS
b
4a‚BF4 64.6 (138)c
-125.10 (2F, d, 3JFF ) 21.7 Hz, Fortho),
-126.40 (2F, d, 3JFF ) 25.3 Hz, Fortho),
-145.80 (2F, t, 3JFF ) 20.7 Hz, Fpara), -143.24
(2F, t, 3JFF ) 20.2 Hz, Fpara), -157.15 (4F, m,
1.61 (t, 4JPH ) 3.9 Hz, CH3)
851 (M+), 816 ([M - Cl]+)
calcd for C36H29ClF10P2Rh,
851.032841; found M+,
851.032267
54.4 (146)d
F
meta), -159.12 (4F, m, Fmeta
-126.5 (4F, br s, Fortho), -144.85 (2F, t, 3JFF
21.8 Hz, Fpara), -158.55 (4F, br s, Fmeta
)
b
4b‚BF4 60.1 (142)
)
1.46 (t, 4JPH ) 3.8 Hz, CH3)
1.51 (t, 4JPH ) 4.2 Hz, CH3)
)
b
4c‚BF4 51.1 (141 Hz) -127.46 (2F, d, 3JFF ) 18.1 Hz, Fortho),
-143.60 (2F, t, 3JFF ) 20.5 Hz, Fpara), -157.59
(4F, m, Fmeta
)
5‚BF4
74.5 (140)e
53.8 (140) d
-123.80 (m), -134.93 (m), -144.13 (td, J )
22.4, 9.0 Hz), -145.17 (t, J ) 22.4 Hz),
-151.16 (t, J ) 22.5 Hz)
4.35 (dd, 2JHH ) 17.8 Hz, 4JPH ) 17.8 Hz,
C5CHHC6F4), 3.29 (d, 2JHH ) 17.8 Hz,
831 (M+), 895 ([M - Cl - H]+)
calcd for C36H28ClF9P2Rh,
C5CHHC6F4), 1.80 (d, 4JPH ) 4.8 Hz, CH3), 831.026613; found M+,
1.63 (d, 4JPH ) 8.8 Hz, CH3), 0.76 (d, 4JPH
)
831.022610
1.7 Hz, CH3)
f
6‚BF4
7‚BF4
85.1 (142)
63.5 (142)
72.0 (134)g
67.4 (137)h
-121.77 (m), -134.02 (m)g, -134.75 (m)h,
-144.71 (td, J ) 22.3, 9.0 Hz),g -145.55 (td,
J ) 22.4, 9.0 Hz)h, -151.97 (t, J ) 22.4 Hz)h,
-153.21 (t, J ) 22.4 Hz)g
4.60 (m), 3.59 (m),2.13 (d, 4JPH ) 6.5 Hz,
CH3), 0.71 (s, CH3)
811 (M+), 875 ([M - Cl - H]+)
calcd for C36H27ClF8P2Rh,
811.020385; found M+,
811.018631
1
a NMR spectra recorded in CDCl3 except where stated. Many of the H and 19F resonances of 5‚BF4, 6‚BF4, and 7‚BF4 are obscured, and only those
resonances that can be positively assigned to the complexes are given. b Data common to 4a‚BF4, 4b‚BF4, and 4c‚BF4. 1H: δ 7.4-8.0 (m, C6H5), 2.5-3.5
(m, CH2). 19F: δ -153.88 (0.8F, 10BF4-), -153.93 (3.2F, 11BF4-). c P(cis-C6F5), JPP ) 22 Hz. d P(trans-C6F5). e P(C6F4). f NMR spectrum recorded in
2
(CD3)2SO. g Major isomer. h Minor isomer.
Treatment of [Cp*RhCl(µ-Cl)]2 with 2 equiv of 1 in the
presence of ammonium tetrafluoroborate yielded, after recrys-
tallization, the salt [Cp*RhCl(κP,κP-1)][BF4], 4‚BF4, as an
orange solid in moderate yield (Scheme 6). The analytical and
spectral data of 4‚BF4 (Table 1) are entirely consistent with
the formulation. The 1H and 19F NMR spectra, although
consistent with the formulation, are far from simple. In contrast
the 31P{1H} NMR spectrum is simple and indicates the presence
of the three expected isomers. A pair of mutually coupled
resonances at δ 64.6 and 54.4 are consistent with [Cp*RhCl-
(κP,κP-1a)][BF4], 4a‚BF4, and are assigned on the basis of the
reaction with Proton Sponge (vide infra) to the phosphorus
Figure 2. Structure of the SS enantiomer of the cation of
atoms with the pentafluorophenyl substituents respectively cis
[Cp*RhCl{κP,κP-rac-(C6F5)PhPCH2CH2PPh(C6F5)}][BF4], 4a‚BF4.
and trans to the pentamethylcyclopentadienyl ring. Two doublets
Thermal ellipsoids are at the 30% level. Hydrogen atoms are omitted
for clarity. Cp†-Rh 1.867(10) Å, Rh-Cl 2.395(2) Å; Rh-P(1)
2.339(3) Å; Rh-P(2) 2.322(3) Å; Cp†-Rh-Cl 120.7(3)°; Cp†-
Rh-P(1) 134.2(3)°; Cp†-Rh-P(2) 131.4(3)°; Cl-Rh-P(1) 84.99-
(9)°; Cl-Rh-P(2) 84.63(9)°; P(1)-Rh-P(2) 84.08(9)°.
of multiplet resonances at δ 60.1 and 51.1 are assigned to trans-
[Cp*RhCl(κP,κP-1b)][BF4], 4b‚BF4, and cis-[Cp*RhCl(κP,κP-
1b)][BF4], 4c‚BF4, respectively, on the basis of the reaction with
Proton Sponge (vide infra). The ratio of 4a:4b:4c is ca. 12:17:
1. An in situ NMR study revealed that the reaction between
[Cp*RhCl(µ-Cl)]2 and 1 in the presence of tetrafluoroborate
forms 4‚BF4 in virtually quantitative yield with a ratio of 4a:
4b:4c of 10:16:1, confirming the assignments of 1a and 1b and
consistent with the steric congestion in 4c due to the greater
bulk of pentafluorophenyl compared to phenyl substituents.
Attempts to separate the isomers by fractional crystallization
were unsuccessful, but a crystal of 4a‚BF4 suitable for a single-
crystal X-ray diffraction study was obtained (Figure 2). The
geometry, Rh-P and Rh-Cl distances, and P-Rh-P angles
of 4a‚BF4 are consistent with those of [Cp*RhCl(P2)]BF4, in
which the diphosphine bears at least one fluoroaryl substitu-
ent.15,19,20
possible and high-resolution mass spectrometry (Table 1). In
both solvents 4a‚BF4 and 4c‚BF4, but not 4b‚BF4, underwent
rapid reactions on addition of Proton Sponge. The rapid reactions
of 4a and 4c and lack of reaction of 4b are consistent with
observations of the treatment of cis- and trans-[Cp*RhCl{Ph2-
PCH2CH2PPh(C5F4N-4)}][BF4] with Proton Sponge.15 The data
indicated that, as expected, the single-linked product 5 was
formed from 4a. Over time isomerization of 4b to 4c, and
subsequent reaction, was observed (Figures 3 and 4). Ultimately
two products with similar spectroscopic properties were formed
in a ratio of ca. 3:2. The spectroscopic data strongly suggest
that one of these is the expected 1,3-double-linked product 7a,
which is similar to that formed exclusively on coupling dfppe
and Cp*.19 On the basis of the similarity of the data, it is
proposed that the other is the 1,2-linked product, 7b, which is
similar to that formed exclusively on coupling dfppe and η5-
C5Me4H.30 In CDCl3 the reaction occurred with less than the
stoichiometric quantity of Proton Sponge. In (CD3)2SO the
The dehydrofluorinative coupling reactions of 4‚BF4 were
studied by in situ NMR experiments in CDCl3 and (CD3)2SO
1
at room temperature. Although the H and 19F NMR spectra
were complicated, the 31P{1H} NMR spectra were simple and
gave data from which assignments and the course of the reaction
could be determined. The identity of the products of the reaction
1
in CDCl3 were supported by H and 19F NMR data where
(30) Nieuwenhuyzen, M.; Saunders: G. C. J. Organomet. Chem. 2000,
595, 292.
(29) Nichols, D. I. J. Chem. Soc., A 1969, 1471.

A Probe for Configurational Stability
Organometallics, Vol. 25, No. 4, 2006 1001
PCHMeCH2PPh2)]+ in dimethyl sulfoxide at 85 °C (k ) 6.18
× 10-5 s-1).10
The difference in the rates of isomerization of 4b is consistent
with the difference in solvation properties. Chloride dissociation
and ion separation is favored in dimethyl sulfoxide, which is
polar with a large relative permittivity and is both a strong donor
and moderate acceptor (µ ) 3.96 D, ꢀr ) 46.6 at 20 °C,31 donor
number ) 29.8, acceptor number ) 19.332), but not in
chloroform, which is much less polar, is only a moderate
acceptor, and is not a donor (µ ) 1.01 D, ꢀr ) 4.8 at 20 °C,31
acceptor number ) 23.132). Since it might be expected that if
mechanism A predominates in both solvents, the rates would
be comparable, we suggest that mechanism B, which involves
chloride dissociation, predominates in dimethyl sulfoxide. The
slower rate of isomerization in chloroform is consistent with
either mechanism operating in that solvent.
The reaction between [Cp*RhCl(µ-Cl) 2 and 2 equiv of 1 in
refluxing benzene was carried out as an attempt to prepare pure
5 or 7. It was hoped that either product would be precipitated.
Unfortunately, although after 50 h a precipitate was apparent,
this was not of a single compound. The mass spectrum of the
solid showed the presence of the cations of the single- and
double-linked products 5 and 7, and NMR spectroscopy
confirmed the mixture and the presence of minor amounts of
other rhodium phosphine complexes.
To assess the configurational stability of 5, the NMR sample
resulting from the reaction between 4a‚BF4 and an excess of
Proton Sponge was heated at 100 °C and the 31P{1H} NMR
spectrum recorded periodically. No change in the ratio of 5 to
7 was observed, even after 2 months. The result indicates that
neither both nor one of phosphine dissociation and inversion at
phosphorus occurs under these conditions.
Figure 3. (a) 31P{1H} NMR spectrum of 4‚BF4 in CDCl3 at
25 °C: (b) 3 h, (c) 66 h, and (d) 238 h after addition of Proton
Sponge.
Conclusion
In summary, the pentafluorophenyl-substituted diphosphine
(C6F5)PhPCH2CH2PPh(C6F5) (1) was prepared as a 1:1.7
mixture of the rac and meso isomers from dppe. In situ NMR
studies of intramolecular dehydrofluorinative coupling reac-
tions of the salt [Cp*RhCl(κP,κP-1)][BF4] reveal that inver-
sion at the metal occurs slowly in chloroform and more rapidly
in dimethyl sulfoxide. Subsequent NMR studies establish
that phosphine dissociation and inversion at phosphorus do not
occur in [{η5,κP,κP-C5Me4CH2C6F4-2-PhPCH2CH2PPh(C6F5)}-
RhCl][BF4] in dimethyl sulfoxide at 100 °C. This observation
provides evidence that complexes of Cp-PL ligands are
configurationally stable at elevated temperature.
Experimental Section
Figure 4. (a) 31P{1H} NMR spectrum of 4‚BF4 in (CD3)2SO at
25 °C: after addition of (b) less than stoichiometric amount and
(c) excess of Proton Sponge.
General Considerations. The compounds dppe, lithium (Ald-
rich), bromopentafluorobenzene (Apollo), and hexachloroethane
(Lancaster) were used as supplied. Chlorotrimethylsilane (Aldrich)
was purified by vaccum distillation and stored under dinitrogen.
[Cp*RhCl(µ-Cl)]2 was prepared as described.33 The preparation of
1 was performed under argon and dintrogen using diethyl ether,
THF, and dichloromethane dried by distillation under dinitrogen
from sodium/benzophenone, potassium/benzophenone, and calcium
hydride, respectively, and stored over molecular sieves (4 Å). No
precautions to exclude air or moisture were taken for the other
preparations.
reaction required more Proton Sponge, which allowed the
observation of the intermediate 6 as a pair of doublet resonances
of equal integration in the 31P{1H} NMR spectrum (Figure 4).
The increase in intensity of these doublets as 4b was consumed
and their decrease in intensity as 7a and 7b were formed
confirms the intermediacy of 6 and supports the identities of
7a and 7b. In the presence of an excess of Proton Sponge the
reaction reached completion within 24 h, indicating rapid
isomerization of 4b to 4c. In contrast the isomerization in CDCl3
was slow, allowing analysis of the kinetics. As expected, a first-
order rate law was obeyed with k ≈ 2.5 × 10-6 s-1. This is
comparable to the rate of epimerization of [Cp*RhCl(Ph2-
(31) SolVent RecoVery Handbook, 2nd ed.; Smallwood, I. M., Ed.;
Blackwell Science: Oxford, 2002.
(32) The Donor-Acceptor Approach to Molecular Interactions; Gutmann,
V., Ed.; Plenum Press: New York, 1978.
(33) White, C.; Yates, A.; Maitlis, P. M. Inorg. Synth. 1992, 29, 228.

1002 Organometallics, Vol. 25, No. 4, 2006
Nieuwenhuyzen et al.
1H, 19F, and 31P NMR spectra were recorded using Bruker
DPX300 or DRX500 spectrometers. H NMR spectra (300.01 or
Table 2. Crystal Data and Structure Refinement for 4a‚BF4
1
formula
fw
C36H29BClF14P2Rh‚1/2CH2Cl2
981.17
500.13 MHz) were referenced internally using the residual protio
solvent resonance relative to SiMe4 (δ 0), 19F (282.26 MHz)
externally to CFCl3 (δ 0), and 31P (121.45 or 202.46 MHz)
externally to 85% H3PO4 (δ 0). All chemical shifts are quoted in
δ (ppm), using the high-frequency positive convention, and coupling
constants in Hz. Simulations were carried out using the gNMR
simulation package.34 EI and LSIMS mass spectra were recorded
on a VG Autospec X series mass spectrometer. Elemental analyses
were carried out by ASEP, The School of Chemistry, Queen’s
University Belfast.
(C6F5)PhPCH2CH2PPh(C6F5) (1). A solution of dppe (2.9 g,
7.3 mmol) in THF (50 cm3) at 0 °C was added to an excess of
lithium shot (2.9 g, 0.42 mmol) under argon. The mixture was
stirred vigorously at 0 °C for 1 h, allowed to warm to ambient
temperature, and stirred for a further 72 h. The resulting yellow-
brown solution was decanted from the lithium, and chlorotrimeth-
ylsilane was added dropwise by syringe until the color disappeared.
The volatiles were removed under reduced pressure, affording an
off-white solid, which was extracted with dichloromethane (150
cm3). Hexachloroethane (3.5 g, 15.0 mmol) in dichloromethane (50
cm3) was added slowly to the extract and the solution left at ambient
temperature for 72 h. The volatiles were removed under reduced
pressure, affording ClPhPCH2CH2PPhCl as a colorless oil, which
solidified on standing in vacuo. The identity of the chlorophosphine
was confirmed by mass spectrometry and 31P{1H} NMR spectros-
copy. 31P{1H} NMR (CDCl3): δ 93.1. HRLSIMS: calcd for
C14H1435Cl2P2, 313.99431; found for M+, 313.99478.
cryst dimens, mm
T, K
0.42 × 0.18 × 0.08
293(2)
cryst syst
monoclinic
space group
unit cell dimens
a, Å
b, Å
c, Å
P2(1)/n
17.220(4)
12.682(3)
19.14655
110.082(5)
3927.0(17)
4
â, deg
U, Å3
Z
calc density, g cm-3
F(000)
1.660
1956
θ, deg
2.54-23.23
abs coeff, mm-1
total no. of data
no. of unique data, Rint
final R indices [I > 2σ(I)]
0.747
33 406
6913, 0.1497
R1 ) 0.0832
wR2 ) 0.2035
R1 ) 0.1430
wR2 ) 0.2338
0.956
R indices (all data)
GoF on F2
largest diff peak and hole, e Å-3
1.375, -0.826
[M - C6H5(C6F5)PO]+). HRLSIMS: calcd for C26H15F10OP2,
595.043850; found [M + H]+, 595.044586.
(C6F5)PhP(O)CH2CH2P(O)Ph(C6F5) (3). Dioxide 3 was formed
on addition of hydrogen peroxide to an NMR sample of 1 in CDCl3.
1H NMR (CDCl3): δ 7.74 (4H, m), 7.55 (6H, m), 2.82 (2H, m),
2.58 (2H, m). 19F NMR (CDCl3): δ -130.25 (4F, m, Fortho),
-145.31 (2F, m, Fpara), -158.64 (4F, m, Fmeta). 31P{1H} NMR
(CDCl3): δ 30.1 (2s). LSIMS, m/z: 611 (100%, M+), 443 (17%,
[M - C6F5 - H]+). HRLSIMS: calcd for C26H14F10O2P2,
611.038764; found M+, 611.036133.
[Cp*RhCl{KP,KP-rac-(C6F5)PhPCH2CH2PPh(C6F5)}][BF4]
(4a‚BF4), trans-[Cp*RhCl{KP,KP-meso-(C6F5)PhPCH2CH2PPh-
(C6F5)}][BF4] (4b‚BF4), and cis-[Cp*RhCl{KP,KP-meso-(C6F5)-
PhPCH2CH2PPh(C6F5)}][BF4] (4c‚BF4). A slurry of [Cp*RhCl(µ-
Cl)]2 (0.075 g, 0.12 mmol), 1 (0.14 g, 0.24 mmol), and NH4BF4
(0.5 g, 4.8 mmol) was treated as for the synthesis of [Cp*RhCl-
(dfppe)][BF4].19 A mixture of 4a‚BF4, 4b‚BF4, and 4c‚BF4 (ca.
12:17:1) was obtained as an orange solid. Yield: 0.07 g (31%).
Anal. Calcd for C36H29BClF14P2Rh‚CH2Cl2: C, 43.41; H, 3.05.
Found: C, 43.58; H, 3.54. NMR and mass spectral data are given
in Table 1.
Pentafluorophenylmagnesium bromide in diethyl ether (100 cm3),
freshly prepared from C6F5Br (3.5 cm3 g, 0.03 mol) and magnesium
(1.0 g, 0.04 mol), was added to a slurry of the ClPhPCH2CH2-
PPhCl in diethyl ether (50 cm3) with vigorous stirring. A color
change from black to dark brown was observed. The mixture was
stirred for 16 h at ambient temperature. The solution was opened
to the atmosphere, water (30 cm3) and dichloromethane (30 cm3)
were added, and the organic layer was separated. The aqueous layer
was extracted with dichloromethane (2 × 30 cm3). The combined
extracts and organic layer were washed with water (3 × 30 cm3)
and dried over magnesium sulfate. The solution was filtered and
the solvent removed by rotary evaporation to afford a brown solid.
Product 1, as a 1:1.7 mixture of 1a and 1b, was obtained as a white
crystalline solid following chromatography on neutral alumina (6%
H2O) with hexane/dichloromethane (9:1) as eluant. Yield: 2.3 g
1
(54.5%). H NMR (CDCl3): δ 7.46 (4H, m), 7.34 (6H, m), 2.41
X-ray Crystallography. A crystal of 4a‚BF4‚0.5CH2Cl2 (0.42
× 0.18 × 0.08 mm) was obtained by slow evaporation of solvent
from a solution of 4a‚BF4 in dichloromethane. Crystal data are listed
in Table 2. Diffraction data were collected on a Bruker SMART
diffractometer using the SAINT-NT35 software with graphite-
monochromated Mo KR radiation. Lorentz and polarization cor-
rections were applied. Empirical absorption corrections were applied
using SADABS.36 The structure was solved by direct methods and
refined with the program package SHELXTL version 5.37 The non-
hydrogen atoms, except those of the anion (B(1), B(1′), F(11),
F(11′), F(12), F(12′), F(13), F(13′), F(14), and F(14′)) and solvent
(C(1S), Cl(1S), and Cl(2S)), were refined with anisotropic thermal
parameters. Hydrogen atom positions were added and idealized,
and a riding model with fixed thermal parameters (Uij ) 1.2Ueq
for the atom to which they are bonded (1.5 for CH3)) was used for
(4H, m). 19F NMR (CDCl3): δ -129.68 (4F, m, Fortho), -150.46
3
3
(1.2F, t, JFF ) 19.8 Hz, Fpara), -150.70 (0.8F, t, JFF ) 19.8 Hz,
F
para), -160.70 (4F, m, Fmeta). 31P{1H} NMR (CDCl3): δ -26.56
3
(0.63, 2nd order pattern from AA′X2X′2 spin system: JPP ) 43.0
Hz, 3JPF ) 33.0 Hz, 6JPF ) 2.0 Hz), -26.43 (0.37, 2nd order pattern
from AA′X2X′2 spin system: JPP ) 43.0 Hz, 3JPF ) 33.0 Hz, 6JPF
3
) 2.0 Hz). LSIMS, m/z: 578 (100%, M+), 411 (58%, [M -
C6F5]+). HRLSIMS: calcd for C26H14F10P2, 578.041110; found M+,
578.041806. Anal. Calcd for C26H14F10P2: C, 54.03; H, 2.44.
Found: C, 53.69; H, 2.61.
A small amount (<0.05 g) of monoxide 2, as a 1:1 mixture of
meso and rac isomers, was obtained by elution with dichlo-
romethane. 1H NMR (CDCl3): δ 7.4-7.8 (10H, m), 2.54 (3H, m),
2.35 (1H, m). 19F NMR (CDCl3): δ -129.40 (4F, m, Fortho),
-130.47 (2F, m, Fortho), -130.73 (2F, m, Fortho), -145.84 (2F, m,
2
subsequent refinements. The function minimized was ∑[w(|Fo| -
F
para), -150.13 (2F, m, Fpara), -158.78 (4F, m, Fmeta), -160.51
2
2
|Fc| )] with reflection weights w-1 ) [σ2|Fo| + (g1P)2 + (g2P)]
(4F, m, Fmeta). 31P{1H} NMR (CDCl3): δ 30.1 (dm, 3JPP ) 59 Hz,
2
2
where P ) [max|Fo| + 2|Fc| ]/3. CCDC 286537 contains the
3
3
3
PO), 29.6 (dm, JPP ) 59 Hz, PO), -25.3 (dt, JPP ) 59 Hz, JPF
) 33 Hz, P), -25.8 (dt, 3JPP ) 59 Hz, 3JPF ) 33 Hz, P). LSIMS,
m/z: 595 (100%, [M + H]+), 427 (25%, [M - C6F5]+), 303 (72%,
supplementary crystallographic data for this paper. These data can
(35) SAINT-NT; Bruker AXS Inc.: Madison, WI, 1998.
(36) Sheldrick, G. M. SADABS; University of Go¨ttingen: Germany 1996.
(37) Sheldrick, G. M. SHELXTL version 5; Bruker AXS Inc.: Madison,
WI, 1998.
(34) gNMR, version 4.0; Cherwell Scientific Publishing Ltd.: Oxford,
1995.

A Probe for Configurational Stability
Organometallics, Vol. 25, No. 4, 2006 1003
be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
and bond angles for 4a‚BF4‚0.5CH2Cl2. This material is available
free of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: A listing of atomic
coordinates, anisotropic displacement parameters, bond distances,
OM050893+