Synthesis of piano stool complexes employing the pentafluorophenyl- substituted diphosphine (C6F5)2PCH 2P(C6F5)2 and the effect of phosphine modifiers on hydrogen transfer catalysis

Article abstract of DOI:10.1021/om070089i

Ruthenium, rhodium, and indium piano stool complexes of the pentafluorophenyl-substituted diphosphine (C₆F₅) ₂PCH₂P(C₆F₅)₂ (2) have been prepared and structurally characterized by single-crystal X-ray diffraction. The η⁵,κP-Cp-P tethered complex [{(η⁵,κP-C₅Me₄CH₂C ₆F₄-2-P(C₆F₅)CH₂P(C ₆F₅)₂}-RhCl₂ (9), in which only one phosphorus is coordinated to the rhodium, was prepared by thermolysis of a slurry of [Cp*RhCl(μ-Cl)]₂ and 2 and was structurally characterized by single-crystal X-ray diffraction. The tethering occurs by intramolecular dehydrofluorinative coupling of the η⁵- pentamethylcyclopentadienyl ligand and κP,κP-coordinated 2. The geometric changes that occur on tethering force dissociation of one of the phosphorus atoms. The effects of introducing phosphine ligands to the coordination sphere of piano stool hydrogen transfer catalysts have been studied. The complexes of fluorinated phosphine complexes are found to transfer hydrogen at rates that compare favorably with leading catalysts, particularly when the phosphine and cyclopentadienyl functionalities are tethered. The highly chelating η⁵,κP,κL-Cp-PP complex [(η⁵,κP,κP-C₅Me₄CH ₂-2-C₅F₃N-4-PPhCH₂CH ₂PPh₂)RhCl]BF₄ (1) was found to out-perform all other complexes tested. The mechanism of hydrogen transfer catalyzed by piano stool phosphine complexes is discussed with reference to the trends in activity observed.

Full text of DOI:10.1021/om070089i

Organometallics 2007, 26, 2659-2671
2659
Synthesis of Piano Stool Complexes Employing the
Pentafluorophenyl-Substituted Diphosphine (C6F5)2PCH2P(C6F5)2
and the Effect of Phosphine Modifiers on Hydrogen Transfer
Catalysis
Andrew C. Marr,* Mark Nieuwenhuyzen, Ciara L. Pollock, and Graham C. Saunders
School of Chemistry and Chemical Engineering, Queen’s UniVersity Belfast, DaVid Keir Building,
Belfast, BT9 5AG, United Kingdom
ReceiVed January 30, 2007
Ruthenium, rhodium, and iridium piano stool complexes of the pentafluorophenyl-substituted
diphosphine (C6F5)2PCH2P(C6F5)2 (2) have been prepared and structurally characterized by single-crystal
X-ray diffraction. The η5,κP-Cp-P tethered complex [{(η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}-
RhCl2] (9), in which only one phosphorus is coordinated to the rhodium, was prepared by thermolysis of
a slurry of [Cp*RhCl(µ-Cl)]2 and 2 and was structurally characterized by single-crystal X-ray diffraction.
The tethering occurs by intramolecular dehydrofluorinative coupling of the η5-pentamethylcyclopentadienyl
ligand and κP,κP-coordinated 2. The geometric changes that occur on tethering force dissociation of one
of the phosphorus atoms. The effects of introducing phosphine ligands to the coordination sphere of
piano stool hydrogen transfer catalysts have been studied. The complexes of fluorinated phosphine
complexes are found to transfer hydrogen at rates that compare favorably with leading catalysts, particularly
when the phosphine and cyclopentadienyl functionalities are tethered. The highly chelating η5,κP,κL-
Cp-PP complex [(η5,κP,κP-C5Me4CH2-2-C5F3N-4-PPhCH2CH2PPh2)RhCl]BF4 (1) was found to out-
perform all other complexes tested. The mechanism of hydrogen transfer catalyzed by piano stool phosphine
complexes is discussed with reference to the trends in activity observed.
particularly in the presence of base. An extensive role for
aromatic facially capping ligands, such as η5-pentamethylcy-
clopentadienyl (Cp*) and η6-p-cymene ligands, has resulted from
the ability of these spectator ligands to boost the enantioselec-
tivity of the reaction. The high stereoselectivities achieved for
the hydrogenation of alkyl aryl ketones using arene ruthenium
catalysts have been attributed to the stabilizing interaction of
the capping group with the aryl functionality of the ketone in
the transition state.18 Cp* rhodium(III) complexes have been
commercially exploited for the asymmetric transfer hydrogen
of aryl alkyl ketones.19 More recently Wills and co-workers have
demonstrated that tethering the amine to the aryl group of a
hydrogen transfer catalyst can lead to an increase in catalyst
stability and activity and improved enantioselectivity.20-22
The two main classes of mechanism quoted for transition-
metal-catalyzed hydrogen transfer are the inner sphere hydridic
route and the metal-ligand bifunctional route.23 Examples of
these mechanisms are given in Schemes 1 and 2. Hydridic
Introduction
Transfer hydrogenation is defined as the reduction of an
unsaturated substrate with hydrogen derived from a reductant
other than hydrogen gas. The substrates are most commonly
ketones or imines and the hydrogen donor a secondary alcohol
or formic acid. During the past decade this reaction has become
one of the major transition-metal-mediated transformations in
organic synthesis. This success can be attributed to the
combination of synthetic usefulness and ease of operation.
Transfer hydrogenation has been the subject of recent book
chapters1,2 and reviews.3-9 An iridium hydride complex was
the first reported transition metal catalyst employed for the
transfer of hydrogen.10,11 Soon afterward it was found that simple
phosphine complexes of ruthenium and rhodium, such as [RuCl2-
(PPh3)3]12 and [RhCl(PPh3)3],13-17 are also active catalysts,
* Corresponding author. E-mail: a.marr@qub.ac.uk.
(1) Gladiali, S.; Alberico, E. In Transition Metals for Organic Synthesis,
2nd ed.; Beller, M., Bolm, C., Eds.; Wiley-VCH: New York, 2004, Vol.
2, p 145.
(2) Gladiali, S.; Mestroni, G. In Transition Metals for Organic Synthesis,
1st ed.; Beller, M., Bolm, C., Eds.; Wiley-VCH: New York, 1998; p 97.
(3) Gladiali, S.; Alberico, E. Chem. Soc. ReV. 2006, 35, 226.
(4) Samec, J. S. M.; Ba¨ckvall, J. E.; Andersson, P. G.; Brandt, P. Chem.
Soc. ReV. 2006, 35, 237.
(5) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. ReV. 2004,
248, 2201.
(6) Blaser, H. U.; Malan, C.; Pugin, B.; Spindler, F.; Steiner, H.; Studer,
M. AdV. Synth. Catal. 2003, 345, 103.
(7) Ba¨ckvall, J. E. J. Organometal. Chem. 2002, 652, 105.
(8) Palmer, M. J.; Wills, M. Tetrahedron: Asymmetry 1999, 10, 2045.
(9) Zassinovich, G.; Mestroni, G.; Gladiali, S. Chem. ReV. 1992, 92, 1051.
(10) Trocha, J.; Henbest, H. B. Chem. Commun. 1967, 545.
(11) Henbest, H. B. Proc. Chem. Soc. 1964, 361.
(12) Sasson, Y.; Blum, J. J. Org. Chem. 1975, 40, 1887.
(13) Chowdhury, R. L.; Ba¨ckvall, J. E. Chem. Commun. 1991, 1063.
(14) Mu¨ller, D.; Umbricht, G.; Weber, B.; Pfaltz, A. HelV. Chim. Acta
1991, 74, 232.
(15) Gladiali, S.; Chelucci, G.; Chessa, G.; Delogu, G.; Soccolini, F. J.
Organomet. Chem. 1987, 327, C15.
(16) Kvintovics, P.; James, B. R.; Heil, B. Chem. Commun. 1986, 1810.
(17) Gladiali, S.; Pinna, L.; Delogu, G.; de Martin, S.; Zassinovich, G.;
Mestroni, G. Tetrahedron Asymmetry 1990, 1, 635.
(18) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org. Chem. 2001,
66, 7931.
(19) Blacker, A. J.; Mellor, B. J. (Zeneca Ltd.) WO9842643, 1998.
(20) Matharu, D. S.; Morris, D. J.; Clarkson, G. J.; Wills, M. Chem.
Commun. 2006, 3232.
(21) Cheung, F. K.; Hayes, A. M.; Hannedouche, J.; Yim, A. S. Y.; Wills,
M. J. Org. Chem. 2005, 70, 3188.
(22) Cross, D. J.; Houson, I.; Kawamoto, A. M.; Wills, M. Tetrahedron
Lett. 2004, 45, 843.
(23) Noyori, R; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97.
10.1021/om070089i CCC: $37.00 © 2007 American Chemical Society
Publication on Web 03/31/2007

2660 Organometallics, Vol. 26, No. 10, 2007
Marr et al.
Scheme 1. Monohydride Inner Sphere Mechanism for Hydrogen Transfer from a Secondary Alcohol to a Ketone (M is a
transition metal complex)
Scheme 2. Metal-Ligand Bifunctional Outer-Sphere
Mechanism for Hydrogen Transfer from a Secondary
Alcohol to a Ketone
Scheme 3. Formation of Cp-PL Ligands by Intramolecular
Dehydrofluorinative Coupling
Hydrogen transfer is extremely important in biology, as the
redox processes that fuel life are controlled by passing protons
and electrons down electron transport chains, thus separating
the primary oxidant and reductant and releasing the energy of
reaction slowly. Thus the reduction of cofactors such as
nicotinamide adenine dinucleotide (NAD+) by hydrogen transfer
from organic chemicals is key to life. Hembre and co-workers27
have demonstrated that [Cp*Ru(κP,κP-dppm)(H)] and [Cp*Ru-
(κP,κP-dppm)(η2-H2)]+ (dppm ) bis(diphenylphosphino)-
methane) catalyze the reduction of the pyridinium NAD+ model
N-methylacridinium using hydrogen. [Cp*Rh(κN,κN-bpy)H]+
(bpy ) 2,2′-bipyridine) has been shown to catalyze the transfer
of hydrogen from formic acid to the NAD+ model N-benzyl-
1,4-dihydronicotinamide and thus enable the cofactor-dependent
biocatalytic hydroxylation of a hydroxybiphenyl to the corre-
sponding catechol.28 Similarly [Cp*Rh(κN,κN-bpy)(H2O)]2+ has
been used to transfer hydrogen from formate to flavin adenine
dinucleotide (FAD) in order to assist the operation of monooxy-
genase enzymes in the absence of NAD+.29
mechanisms are expected for catalysts derived from metal
phosphine complexes such as [RhCl(PPh3)3] (Scheme 1).
Addition of a ligand containing a Brønsted base, such as DPEN
(1,2-diphenylethylenediamine), to a piano stool complex such
as cymene ruthenium chloride, [(η6-p-cymene)RuCl(µ-Cl)]2,
allows the activation of the hydrogen donator by an alternative
mechanism in which the metal and ligand act cooperatively
(Scheme 2).
Recently new opportunities for the exploitation of hydrogen
transfer chemistry have emerged. Among these are the use of
racemization catalysts to facilitate dynamic kinetic resolution
(DKR)24,25 and the use of hydrogen transfer reactions to model
and assist enzymes.26-28
In DKR a homogeneous catalyst can be exploited to improve
the yield of a kinetic resolution. Ba¨ckvall has reviewed the
subject.25 When an enantioselective biocatalytic reaction is
carried out on a racemic mixture of a chiral alcohol, the
maximum yield possible is 50%. However, the addition of a
hydrogen transfer catalyst that can racemize the alcohol increases
this potential yield to 100%. Although many catalysts for
racemization are known, the operation of biological and
chemical catalysis together has proven difficult and selectivies
and activities of the resultant systems are often disappointing.
One success in this area was the commercialization of the DKR
of esters from racemic alcohols by the combination of racem-
ization and lipase activity.29
In order to be used in the same pot as a biocatalyst, a chemical
catalyst must be robust and its reactivity must be very specific
so that no side reactions occur with the biopolymers, which
contain enticing soft donor groups such as thiols and thiothers.
It is important to engineer a metal center with a high degree of
reactant specificity and a low tendency to undergo unwanted
substitution reactions, and therefore the comparison of homo-
geneous catalysts for hydrogen transfer promoted by highly
chelating ligands containing good donor atoms is pertinent.
Ideally the active center should be protected inside an activating
pocket similar to that observed at metalloenzyme active sites.
Such a highly stable pocket is afforded by the tethered
cyclopentadienyl ligands η5,κP,κP-Cp-PP and η5,κP-Cp-P
containing a κP,κP- and κP-coordinated diphosphine.
Previously reported [(η5,κP,κP-C5Me4CH2C5F3N-2-PPhCH2-
30
(24) Dinh, P. M.; Howarth, J. A.; Hudnott, A. R.; Williams, J. M. J.;
Harris, W. Tetrahedron Lett. 1996, 37, 762.
(25) (a) Pamies, O.; Ba¨ckvall, J. E. Chem. ReV. 2003, 103, 3247. (b)
Huerta, F. F.; Minidis, A. B. E.; Ba¨ckvall, J. E. Chem. Soc. ReV. 2001, 30,
321.
(26) Hembre, R. T.; McQueen, S. J. Am. Chem. Soc. 1994, 116, 2141.
(27) Lutz, J.; Hollmann, F.; Vinh Ho, T.; Schnyder, A.; Fish, R. H.;
Schmid, A. J. Organomet. Chem. 2004, 689, 4783.
(28) de Gonzalo, G.; Ottolina, G.; Carrea, G.; Fraaije, M. W. Chem.
Commun. 2005, 3724.
CH2PPh2)RhCl]BF4 (1) was selected as the first potential
catalyst. Compound 1 is readily prepared by the thermolysis
reaction between [Cp*RhCl(µ-Cl)]2 and (C5F4N-4)PhPCH2CH2-
PPh2, which involves intramolecular dehydrofluorinative cou-
pling of the two coordinated ligands (Scheme 3). The chirality
of this complex is advantageous, as it imparts the potential to
(30) Bellabarba, R. M.; Nieuwenhuyzen, M.; Saunders, G. C. Organo-
metallics 2003, 22, 1802.
(29) Verzijl, G. K. M.; de Vries, J. G. (DSM) WO0190396, 2001.

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2661
be an asymmetric hydrogen transfer catalyst. Access to chiral
complexes of η5,κP-Cp-P ligands is more problematic. Previous
routes have involved elaborate, multiple-step, low-yielding
syntheses.31 Intramolecular dehydrofluorinative coupling has
been used to prepare achiral complexes of η5,κP-Cp-P ligands
using monodentate phosphines, but the yields and purity are
not as favorable as for reactions using chelating diphosphines.32
However, the convenience of intramolecular dehydrofluorinative
coupling led us to examine other strategies. A chelating ligand
is necessary for rapid and clean dehydrofluorinative coupling,
but the product is a complex of a η5,κP,κL-Cp-PL ligand.
However, it may be possible to convert this to a η5,κP-Cp-P
ligand either by cleaving the PL linkage and removing L or by
using geometric changes that occur on tethering to disfavor
coordination of the L ligating group; if necessary, L can then
be converted to a poorly coordinating group by functionalization.
One group of chelating PL ligands that may partially dissociate
on tethering are those that give small bite (P-M-L) angles. It
has been established that coupling of the phosphine to cyclo-
pentadienyl reduces both the Cp-M-P angle and the M-P
bond length,30,32 which may constrain the geometry of the Cp-
PL ligand sufficiently to force dissociation of the L functionality.
Ethylene-bridged diphosphines favor chelation, giving bite
angles of >80°.33 In contrast, methylene-bridged diphosphines,
R2PCH2PR2, either chelate to give strained four-membered rings
with P-M-P bite angles of ca. 72° 33 or bridge between two
metals, depending on the nature of the substituent R. The strain
of the four-membered ring typically leads to the diphosphines
bridging between metals, but sterically demanding substituents,
such as tert-butyl and cyclohexyl, cause chelation to be
favored.34,35 The dehydrofluorinative coupling reaction requires
that at least one substituent be an aryl group bearing ortho
fluorine atoms, and since fluorinated aryl groups are much more
sterically demanding than the non-fluorinated analogues,36,37 the
fluoro-substituted analogue of dppm, (C6F5)2PCH2P(C6F5)2 (2),
is an ideal candidate for the synthesis of η5,κP-Cp-P complexes
that are chiral by virtue of the coordinated phosphorus bearing
four different substituents. These complexes fulfill the require-
ments of the second type of complex for our study. Comparisons
of their catalytic activity with those of piano stool complexes
of 2 would then provide useful insight into the effect of tethering
on catalytic activity.
Figure 1. (a) Molecular structure of (C6F5)2PCH2P(C6F5)2, 2.
Thermal ellipsoids are at the 30% level. (b) View of the pairing of
molecules of 2. Thermal ellipsoids are at the 30% level. Hydrogen
and fluorine atoms are omitted for clarity.
bis(dichlorophosphino)methane to pentafluorophenylmagnesium
bromide. The structure of 2 (Figure 1) was determined by single-
crystal X-ray diffraction. Selected bond distances and angles
of 2 are given in Table 1. Two of the pentafluorophenyl rings
are approximately coplanar (deviation 8.9°) and twisted relative
to the plane defined by P-C-P by ca. 35°. The planes of the
other pair are approximately parallel (deviation 8.1°), with a
separation of ca. 3.5 Å, and are virtually perpendicular (83° to
90°) to those of the other rings. The latter aryl rings are
positioned on the same side of the plane defined by P-C-P.
Although this arrangement is atypical of uncomplexed Ar2PCH2-
PAr2, it is common for metal complexes of dppm.39 The
structure reveals an unprecedented arrangement comprising pairs
Here we report the syntheses of piano stool complexes of 2,
and a η5,κP-Cp-P complex derived from it, and the catalytic
activities of a wide variety of phosphine-substituted piano stool
complexes including 1 for hydrogen transfer reactions. We
assess the new catalysts relative to important literature catalysts
and demonstrate the excellent potential of η5-Cp,κP,κP-diphos-
phine, η5,κP-Cp-P, and η5,κP,κP-Cp-PP complexes. Part of
the synthetic work has been communicated.38
Results and Discussion
(39) For example: (a) Cowie, M.; Dickson, R. S. Inorg. Chem. 1981,
20, 2682. (b) Mague, J. T. Inorg. Chem. 1983, 22, 45. (c) Delavaux, B.;
Chaudret, B.; Taylor, N. J.; Arabi, S.; Poilblanc, R. J. Chem. Soc., Chem.
Commun. 1985, 805. (d) Balch, A.; Waggoner, K. M.; Olmstead, M. M.
Inorg. Chem. 1988, 27, 4511. (e) Kirss, R. U.; Eisenberg, R. Inorg. Chem.
1989, 28, 3372. (f) Knox, S. A. R.; Lloyd, B. R.; Morton, D. A. V.; Orpen,
A. G.; Turner, M. L.; Hogarth, G. Polyhedron 1995, 14, 2723. (g) Wang,
L.-S.; Cowie, M. Organometallics 1995, 14, 3040. (h) Cauzzi, D.; Graiff,
C.; Lanfranchi, M.; Predieri, G.; Tiripicchio, A. J. Chem. Soc., Dalton Trans.
1995, 2321. (i) Alvarez, M. A.; Bois, C.; Garcia, M. E.; Riera, V.; Ruiz,
M. A. Angew. Chem., Int. Ed. Engl. 1996, 35, 102. (j) Bruce, M. I.; Low,
P. J.; Werth, A.; Skelton, B. W.; White, A. H. J. Chem. Soc., Dalton Trans.
1996, 1551. (k) Cotton, F. A.; Daniels, L. M.; Haefner, S. C.; Kuhn, F. E.
Inorg. Chim. Acta 1999, 287, 159. (l) Qin, Z.; Jennings, M. C.; Puddephat,
R. J. Inorg. Chem. 2001, 40, 6220. (m) Snaith, T. J.; Low, P. J.; Rousseau,
R.; Puschmann, H.; Howard, J. A. K. J. Chem. Soc., Dalton Trans. 2001,
292. (n) Rowsell, B. D.; Trepanier, S. J.; Lam, R.; McDonald, R.; Cowie,
M. Organometallics 2002, 21, 3228.
Synthesis and Structure of (C6F5)2PCH2P(C6F5)2 (2).
Diphosphine 2 was synthesized in 81% yield by addition of
(31) Butenscho¨n, H. Chem. ReV. 2000, 100, 1527.
(32) Bellabarba, R. M.; Nieuwenhuyzen, M.; Saunders, G. C. Organo-
metallics 2002, 21. 5726.
(33) Dierkes, P.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans.
1999, 1519.
(34) Puddephat, R. J. Chem. Soc. ReV. 1983, 12, 99.
(35) Eisentra¨ger, F.; Go¨thlich, A.; Gruber, I.; Heiss, H.; Kiener, C. A.;
Kru¨ger, C.; Notheis, J. U.; Rominger, F.; Scherhag, G.; Schultz, M.; Straub,
B. F.; Volland, M. A. O.; Hofmann, P. New J. Chem. 2003, 27, 540.
(36) Tolman, C. A. Chem. ReV. 1977, 77, 313.
(37) White, D.; Colville, N. J. AdV. Organomet. Chem. 1994, 36, 95.
(38) Marr, A. C.; Nieuwenhuyzen, M.; Pollock, C. L.; Saunders, G. C.
Inorg. Chem. Commun. 2006, 9, 407.

2662 Organometallics, Vol. 26, No. 10, 2007
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 2
Marr et al.
P(1)-C(1)
P(1)-C(11A)
P(2)-C(21A)
C-C
1.845(2)
1.843(2)
1.851(2)
1.369(3)-1.396(3)
P(2)-C(1)
P(1)-C(11B)
P(2)-C(21B)
C-F
1.846(2)
1.838(2)
1.840(2)
1.340(2)-1.353(2)
P(1)-C(1)-P(2)
106.46(10)
101.94(10)
102.91(10)
98.58(9)
-48.1
-151.4
C(1)-P(1)-C(11A)
104.17(9)
C(1)-P(1)-C(11B)
C(11A)-P(1)-C(11B)
C(1)-P(2)-C(21B)
102.06(10)
104.54(9)
75.5
C(1)-P(2)-C(21A)
C(21A)-P(1)-C(21B)
C(11A)-P(1)‚‚‚P(2)-C(21B)
C(11B)-P(1)‚‚‚P(2)-C(21B)
C(11A)-P(1)‚‚‚P(2)-C(21A)
C(11B)-P(1)‚‚‚P(2)-C(21A)
-27.8
Table 2. Selected Bond Lengths (Å) and Angles (deg) for 5, 6, 8, and 9
48
r
5
[Cp*RhCl(dfppe)]BF4
6a
[Cp*IrCl(dfppe)]BF4
8‚CH2Cl2
9
†
Cn -Ma
1.849(16)
2.367(4)
2.386(4)
1.869(7)
2.342(2)
2.362(2)
2.380(2)
1.870(7)
1.886(1)
2.3278(14)
2.311(2)
1.751(7)
1.804(6)
2.2624(10)
M-P
2.328(5)b
2.2705(18)
2.3454(18)
2.3819(17)
M-Cl
2.3998(13)
2.399(7)
2.3891(13)
2.4085(16)
2.3952(16)
1.842(5)
1.849(5)
127.7(2)
P-CH2
1.837(14)
1.854(15)
137.9(5)
138.3(5)
125.1(5)
1.825(6)
1.830(7)
131.0(2)
134.2(2)
121.9(2)
1.845(14)b
1.832(6)
1.833(5)
131.36(3)
133.97(4)
121.10(4)
1.835(7)
1.852(6)
134.6(2)
140.0(2)
125.7(2)
†
Cn -M-P
138.4(2)b
†
Cn -M-Cl
123.2(2)
123.8(2)
121.5(2)
90.73(6)
92.17(6)
90.88(5)
P-M-Cl
80.03(14)
80.38(14)
83.26(7)
85.36(7)
81.19(18)b
85.46(5)
83.96(5)
79.77(6)
82.40(6)
Cl-M-Cl
P-M-P
72.17(4)
98.7(2)
93.9(4)
92.8(4)
84.05(8)
72.4(3)
96.3(10)
94.5(5)b
84.32(5)
72.19(6)
95.1(3)
93.8(2)
96.8(2)
P-C-P
106.9(3)
112.80(18)
M-P-CH2
105.5(2)
109.5(2)
106.4(2)
109.9(2)
†
b
a Cn represents the centroid of the η5-cyclopentadienyl C5 or η6-arene C6 ring. 6 possesses a plane of symmetry defined by Cp†, Ir, Cl, and B leading
to identical pairs of Ir-P and P-CH2 distances and Cp†-Ir-P, P-Ir-Cl, and Ir-P-CH2 angles.
of molecules (Figure 1) with short intermolecular P‚‚‚P distances
of 3.318 Å (∑van der Waals’ radii 3.80 Å40). The P-C distances
and P-C-P angle are identical within 3σ to those of dppm
(1.828(5) to 1.868(5) Å and 106.2(3)°).41
As expected the 31P{1H} NMR spectrum of 2 displays a
second-order pattern consistent with an [A[X]4[Y]4[Z]2]2 spin
system centered at δ -52.1. This value is similar to that of the
ethylene-bridged analogue (C6F5)2PCH2CH2P(C6F5)2 (dfppe) (δ
-44.142). Unfortunately the complicated natures of the spin
system and 19F NMR spectrum did not permit the coupling
constants to be obtained by simulation.
An in situ NMR experiment indicated that monoxide, 3, and
then dioxide, 4, were formed on addition of hydrogen peroxide
to 2. The identity of 4 was confirmed by mass spectrometry.
As expected, the 31P{1H} NMR spectrum of 3 exhibited a
doublet at δ 15.0 (2JPP ) 76 Hz), assigned to (C6F5)2PO, and a
doublet of quintets of quintets at δ -60.1, assigned to (C6F5)2P,
and that of 4 exhibited a singlet at δ 8.9.
resonances and with δP at lower frequency by ca. 30 ppm. The
values of δP for 7 and 8 are at lower frequency to that of [(η6-
p-cymene)RuCl(κP,κP-dppm)]BF4 by ca. 26 ppm. The 19F
46
NMR spectra of salts 5-8 exhibit broadened signals for the
ortho and meta fluorine resonances consistent with hindered
rotation about the P-C6F5 bonds. From variable-temperature
19F NMR spectra values of the free energy of activation, ∆Gq,
for rotation about one pair of P-C6F5 bonds of 56 ( 2 and 54
( 2 kJ mol-1 were calculated47 for 5 and 6, respectively, and
a minimum value of ∆Gq of 61 kJ mol-1 was calculated47 for
both 7 and 8. The spectra did not allow calculation of ∆Gq for
rotation about the other pair of P-C6F5 bonds.
The structures of 5, 6, and 8 were determined by single-crystal
X-ray diffraction (Figures 2, 3, and 4). Selected bond distances
and angles are given in Table 2. For comparison the pertinent
data of the ethylene-bridged rhodium and iridium analogues48
are included. The cation of 5 is symmetric within 3σ about the
plane defined by Cp†, Rh, and Cl, and that of 6 has crystallo-
graphically imposed symmetry about the plane defined by Cp†,
Ir, and Cl. In contrast the cation of 8 is not symmetric. The
cations adopt the expected three-legged piano stool structures
Synthesis and Structures of Piano Stool Complexes of 2.
Treatment of [Cp*MCl(µ-Cl)]2 (M ) Rh or Ir) with 2 equiv of
2 in the presence of tetrafluoroborate gave the yellow solids
[Cp*MCl(κP,κP-2)]BF4, 5 (M ) Rh) and 6 (M ) Ir), in 92%
and 66% yields, respectively. The salts [(η6-arene)RuCl(κP,κP-
2)]BF4 (7, arene ) p-cymene, 4-isopropyltoluene; 8, arene )
mesitylene, C6Me3-1,3,5) were prepared similarly from [(η6-
†
with similar Cn -M-Cl (ca. 125°) and P-M-Cl (79° to 83°)
angles. The Cp†-M-P angles of 5 and 6 are similar and are
significantly larger than those of the ethylene-bridged analogues.
The P-M-P bite angle is identical for the three cations and
1
arene)RuCl(µ-Cl)]2 and 2 in high yield. The H and 31P{1H}
(44) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledo´s, A.; Pons, V.;
Heinekey, D. M. J. Am. Chem. Soc. 2004, 126, 8813.
(45) Valderrama, M.; Contreras, R. J. Organomet. Chem. 1996, 513, 7.
(46) Daguenet, C.; Scopelliti, R.; Dyson, P. J. Organometallics 2004,
23, 4849.
(47) Sandstro¨m, J. Dynamic N.M.R. Spectroscopy; Academic Press:
London, 1982.
NMR spectra of 5 and 6 are similar to those of [Cp*MCl(κP,κP-
dppm)]BF4 (M ) Rh,43 Ir44,45), but without phenyl hydrogen
(40) Bondi, A. J. Phys. Chem. 1964, 68, 441.
(41) Schmidbauer, H.; Reber, G.; Schier, A.; Wagner, F. E.; Mu¨ller, G.
Inorg. Chim. Acta 1988, 147, 143.
(42) Cook, R. L.; Morse, J. G. Inorg. Chem. 1982, 21, 4103.
(43) Valderrama, M.; Contreras, R.; Bascun˜an, M.; Alegria, S.; Boys,
D. Polyhedron 1995, 14, 2239.
(48) 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.

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2663
Figure 2. Structure of the cation of [Cp*RhCl{κP,κP-(C6F5)2-
PCH2P(C6F5)2}]BF4, 5. Thermal ellipsoids are at the 30% level.
Hydrogen atoms are omitted for clarity.
Figure 5. Structure of the S enantiomer of [{η5,κP-C5Me4CH2C6F4-
2-P(C6F5)CH2P(C6F5)2}RhCl2], 9. Thermal ellipsoids are at the 30%
level. Hydrogen atoms are omitted for clarity.
are identical within 3σ, and the P-C-P angles are similar, but
are at least 6° more acute than that of 2. That of 8 is identical
with that of [(η6-p-cymene)RuCl(κP,κP-dppm)]BF4.46
Thermolysis Reactions between [(ηn-CnRn)MCl(µ-Cl)2]2
and 2. The reaction between [Cp*RhCl(µ-Cl)]2 and 2 equiv of
2 in refluxing benzene gave the singly linked neutral complex
[{η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}RhCl2], 9, in
70% yield. Complex 9 was fully characterized by elemental
analysis, mass spectrometry, and NMR spectroscopy, and its
structure determined by single-crystal X-ray diffraction (Figure
5). The 31P{1H} NMR spectrum displays two mutually coupled
resonances at δ 43.1 and -66.6 (2JPP ) 185 Hz). The former
shows further coupling to rhodium (1JRhP ) 170 Hz) and the
2
latter to fluorine (3JPF ) 38 Hz). The magnitude of JPP is
Figure 3. Structure of [Cp*IrCl{κP,κP-(C6F5)2PCH2P(C6F5)2}]BF4,
6. Thermal ellipsoids are at the 30% level. Hydrogen atoms are
omitted for clarity.
significantly larger than that of 28.1 Hz reported for
[Cp*RhCl2(κP-Ph2PCH2PPh2)],43 but is consistent with larger
couplings occurring for more electronegative substituents.49 The
19F NMR spectrum exhibits 13 resonances integrating for 19
fluorine atoms and indicates hindered rotation about one P-C6F5
bond, presumably that of the phosphorus atom coordinated to
rhodium, which is expected to be the most sterically crowded.
∆Gq was calculated47 to be 46 ( 1 kJ mol-1 from variable-
temperature spectra. The 1H NMR spectrum exhibits resonances
consistent with those of rhodium complexes in which a
cyclopentadienyl ring is linked to a phosphine: two mutually
coupled resonances, one also showing coupling to phosphorus,
assigned to the hydrogen atoms of the methylene bridge, and
four resonances, three of which possess coupling to one
phosphorus atom, assigned to the methyl groups.30,32,50
Important bond distances and angles of 9 are given in Table
2. The Rh-P distance and Cp†-Rh-P angle are consistent with
those of rhodium complexes in which a cyclopentadienyl ring
is linked to a phosphine by a tetrafluorobenzyl or similar
group.30,32,50 The Cp†-Rh-P and Cp†-Rh-Cl angles are
smaller than those of 5, and the P-Rh-Cl and Cl-Rh-Cl
angles are close to 90°. Significantly the Rh-P distance is 0.1
Å shorter than that of 5 and the P-C-P angle is identical to
that of 2.
Figure 4. Structure of [(η6-mesitylene)RuCl{κP,κP-(C6F5)2PCH2P-
(C6F5)2}]BF4, 8. Thermal ellipsoids are at the 30% level. Hydrogen
atoms are omitted for clarity.
similar to that of the three-legged piano stool complex of dppm
[(η6-p-cymene)RuCl(κP,κP-dppm)]BF4 (71.29(6)°).46 This angle
is 12° smaller than those of the ethylene-bridged rhodium and
iridium analogues [Cp*MCl{κP,κP-(C6F5)2PCH2CH2P(C6F5)2}]-
BF4 (M ) Rh 84.05(8)°, M ) Ir 84.32(8)°).48 The Ir-P distance
is shorter than the Rh-P distances, which is consistent with
the ethylene-bridged analogues. The asymmetry in the cation
of 8 is reflected in the Ar†-M-P angles and Ru-P distances.
The former differ by ca. 5° and the latter by ca. 0.07 Å. The
mean Ru-P distance (2.308(2) Å) is similar to the Ru-P
distances of [(η6-p-cymene)RuCl(κP,κP-dppm)]BF4 (2.316(2)
and 2.309(2) Å).46 The P-CH2 distances of the three cations
Since [{η5,κP,κP-C5Me3[CH2C6F4-2-P(C6F5)CH2]2-1,3}RhCl]+
is formed by thermolysis of [Cp*RhCl(µ-Cl)]2 and (C6F5)2PCH2-
(49) Finer, E. G.; Harris, R. K. Prog. Nucl. Magn. Reson. Spectrosc.
1971, 6, 61.
(50) (a) Atherton, M. J.; Fawcett, J.; Holloway, J. H.; Hope, E. G.;
Karac¸ar, A.; Russell, D. R.; Saunders, G. C. J. Chem. Soc., Chem. Commun.
1995, 191. (b) Atherton, M. J.; Fawcett, J.; Holloway, J. H.; Hope, E. G.;
Russell, D. R.; Saunders, G. C. J. Organomet. Chem. 1999, 582, 163.

2664 Organometallics, Vol. 26, No. 10, 2007
Marr et al.
Scheme 4. “Clip on, Flip off” Reaction Scheme
Figure 6. Structure of [(η6-mesitylene)Ru(µ-Cl)3RuCl{κP,κP-
(C6F5)2PCH2P(C6F5)2}], 11. Thermal ellipsoids are at the 30% level.
Hydrogen atoms are omitted for clarity.
Table 3. Selected Bond Lengths (Å) and Angles (deg) for 11
Ru(1)-P(1)
Ru(1)-Cl(1)
Ru(1)-Cl(3)
P(1)-C(10)
Ar†-Ru(2)a
Ru(2)-Cl(3)
2.216(4)
2.388(4)
2.417(4)
Ru(1)-P(2)
Ru(1)-Cl(2)
Ru(1)-Cl(4)
2.222(4)
2.470(4)
2.489(4)
1.796(15)
2.404(4)
2.426(4)
Table 4. Transfer Hydrogenationa of Acetophenone
Catalyzed by Ruthenium Complexes
1.849(14) P(2)-C(10)
percentage
1.643(16) Ru(2)-Cl(2)
entry
catalyst
conversion
2.474(4)
Ru(2)-Cl(4)
1
2
3
4
5
6
7
8
9
1/2 [(η6-toluene)RuCl(µ-Cl)]2
24
57
65
90
68
77
21
67
85
1/2 [(η6-mesitylene)RuCl(µ-Cl)]2
1/2 [(η6-p-cymene)RuCl(µ-Cl)]2
1/2 [(η6-mesitylene)RuCl(µ-Cl)]2 + DPEN
1/2 [(η6-p-cymene)RuCl(µ-Cl)]2 + DPEN
[(η6-mesitylene)RuCl(dppe)]BF4
[(η6-mesitylene)RuCl(2)]BF4, 8
P(1)-Ru(1)-P(2)
P(1)-Ru(1)-Cl(2)
P(1)-Ru(1)-Cl(4)
P(2)-Ru(1)-Cl(2)
P(2)-Ru(1)-Cl(4)
74.01(15)
178.21(15)
102.85(15)
105.03(15)
176.32(16)
P(1)-Ru(1)-Cl(1)
P(1)-Ru(1)-Cl(3)
P(2)-Ru(1)-Cl(1)
P(2)-Ru(1)-Cl(3)
Cl(1)-Ru(1)-Cl(2)
Cl(1)-Ru(1)-Cl(4)
Cl(2)-Ru(1)-Cl(4)
P(1)-C(10)-P(2)
Ar†-Ru(2)-Cl(3)
Cl(2)-Ru(2)-Cl(3)
Cl(3)-Ru(2)-Cl(4)
86.94(16)
100.84(15)
86.08(15)
99.41(15)
91.49(14)
95.74(14)
78.15(14)
94.2(7)
Cl(1)-Ru(1)-Cl(3) 171.44(14)
1/2 [(η6-mesitylene)Ru2(2)Cl(µ-Cl)3], 11
[(η6-p-cymene)RuCl(2)]BF4, 7
Cl(2)-Ru(1)-Cl(3)
Cl(3)-Ru(1)-Cl(4)
Ar†-Ru(2)-Cl(2)
Ar†-Ru(2)-Cl(4)
Cl(2)-Ru(2)-Cl(4)
80.78(14)
79.15(14)
129.5(5)
132.9(5)
80.67(14)
133.2(5)
80.96(14)
79.27(14)
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM,
[acetophenone] ) 0.2 M, T ) 75 °C, [KOH] ) 2.5 mM, reaction time 2 h;
percentage conversions are the average of three experiments.
a Ar† represents the centroid of the η6-C6 ring.
identity of 10 was confirmed by high-resolution mass spec-
trometry and its 1H, 19F, and 31P{1H} NMR spectra, which show
features similar to 9. In particular, the 31P{1H} NMR exhibits
a doublet resonance at δP 11.0 and a doublet of quintets
resonance at δP -69.2 with a mutual coupling of 181 Hz, and
CH2P(C6F5)2 via the cation [Cp*RhCl{κP,κP-(C6F5)2PCH2-
CH2P(C6F5)2}]+,48 it is reasonable to presume that the thermol-
ysis reaction between [Cp*RhCl(µ-Cl)]2 and 2 proceeds via
[Cp*RhCl(κP,κP-2)]+, the cation of 5. In order to test this
hypothesis, 5 was treated with bis(1,8-dimethylamino)naphtha-
lene (Proton Sponge). NMR spectroscopy indicated that 2 was
formed in low yield, along with other, unidentified, rhodium-
and phosphine-containing compounds. A maximum yield of
50% is expected from the 1:1 chloride:rhodium ratio of 5.
Attempts to increase the yield by adding chloride were unsuc-
cessful. However the observations give credence to the inter-
mediacy of the cation of 5 in the thermolysis reaction between
[Cp*RhCl(µ-Cl)]2 and 2 and allow us to propose a “clip on,
flip off” reaction scheme (Scheme 4). The first steps involve
coordination of 2 to give [Cp*RhCl(κP,κP-2)]+, presumably as
the chloride salt. This undergoes dehydrofluorinative carbon-
carbon coupling, which tethers the η5-cyclopentadienyl ligand
to one phosphine moiety. The geometry change associated with
this process, in particular the large reductions in Cp†-Rh-P
angle and Rh-P bond distance, so increase the strain within
the cation that the untethered phosphine moiety is forced to
dissociate, releasing the strain in the P-C-P angle. In further
support of this scheme attempts to enforce η5,κP,κP coordination
of the ligand by treatment of 9 with sodium tetrafluoroborate
or sodium hexafluoroantimonate were unsuccessful.
1
the H NMR spectrum displays resonances at δ 3.93 and 3.77
with a mutual coupling of 18.7 Hz characteristic of a methylene
group linking an η5-cyclopentadienyl ligand to C6F4, in addition
to four methyl resonances, three of which show coupling to
one phosphorus atom.
The reaction between [(η6-mesitylene)RuCl(µ-Cl)]2 and 2
equiv of 2 in refluxing methanol/dichloromethane yielded [(η6-
mesitylene)Ru(µ-Cl)3RuCl(κP,κP-2)] (11) in 96% yield with
respect to ruthenium. The identity of 11 was elucidated by
single-crystal X-ray diffraction (Figure 6), with which the
analytical and NMR spectroscopic data are entirely consistent.
Complexes of this type have been reported previously, but their
syntheses involved reaction of an η6-arene ruthenium complex
and a ruthenium diphosphine complex rather than displacement
of an η6-arene ligand.51,52 Selected bond distances and angles
are given in Table 3. The structure possesses approximate
octahedral geometry about Ru(1) and three-legged piano stool
geometry about Ru(2) and is similar to those of [(η6-C6H4RR′-
1,4)Ru(µ-Cl)3RuCl(κP,κP-Ph2P(CH2)4PPh2)] (R ) R′ ) H;51
R ) Me, R′ ) CHMe252). The terminal Ru-Cl distances are
The reaction between [Cp*IrCl(µ-Cl)]2 and 2 equiv of 2 in
refluxing benzene afforded a mixture comprising predominantly
[{η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}IrCl2], 10. Un-
fortunately all attempts to isolate 10 were unsuccessful. The
(51) de Araujo, M. P.; Valle, E. M. A.; Ellena, J.; Castellano, E. E.; dos
Santos, E. N.; Batista, A. A. Polyhedron 2004, 23, 3163.
(52) da Silva, A. C.; Piotrowski, H.; Mayer, P.; Polborn, K.; Severin, K.
Eur. J. Inorg. Chem. 2001, 685.

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2665
Table 5. Liquid Sampling of Transfer Hydrogenationa of Acetophenone Catalyzed by Ruthenium Complexes
percentage conversion
catalyst
10 min
20 min
30 min
1 h
1.5 h
76
2 h
1/2 [(η6-mesitylene)RuCl(µ-Cl)]2 + DPEN
[(η6-p-cymene)RuCl(2)]BF4, 7
23
47
44
50
48
69
89
71
89
78
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM, [acetophenone] ) 0.2 M, T ) 75 °C, [KOH] ) 2.5 mM.
Table 6. Transfer Hydrogenationa of Acetophenone Catalyzed by Rhodium Complexes
percentage
entry
catalyst
conversion
1
2
3
4
5
6
7
8
9
1/2 [(η5-C5Me4H)RhCl(µ-Cl)]2
35
80
90
63
68
37
50
29
49
81
94
1/2 [Cp*RhCl(µ-Cl)]2
1/2 [Cp*RhCl(µ-Cl)]2 + DPEN
1/2 [Cp*RhCl(µ-Cl)]2 + (S)-phenylglycinol
Cp*RhCl2(PPh3)
[(η5-C5Me4H)RhCl(dppe)]BF4
[Cp*RhCl(dppm)]BF4
[Cp*RhCl(dfppe)]BF4
[Cp*RhCl(2)]BF4, 5
10
11
[{η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}RhCl2], 9
[(η5,κP,κP-C5Me4CH2-2-C5F3N-4-PPhCH2CH2PPh2)RhCl]BF4, 1
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM, [acetophenone] ) 0.2 M, T ) 75 °C, [KOH] ) 2.5 mM, reaction time 2 h; percentage
conversions are the average of three experiments.
Table 7. Liquid Sampling of Transfer Hydrogenationa of Acetophenone Catalyzed by Rhodium Complexes
percentage conversion
catalyst
10 min
60
20 min
92
30 min
94
1 h
98
1.5 h
2 h
98
[(η5,κP,κP-C5Me4CH2-2-C5F3N-4-PPhCH2CH2PPh2)RhCl], 1
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM, [acetophenone] ) 0.2 M, T ) 75 °C, [KOH] ) 2.5 mM.
Table 8. Transfer Hydrogenationa of Acetophenone
racemization (as part of a dynamic kinetic resolution) of
secondary alcohols, respectively.19,29 A relatively low substrate
concentration and reaction time were chosen in order to
minimize the back-reaction and thus give comparable results;
higher turnover frequencies are achievable using higher substrate
concentrations. Reproducibility was found to vary with the
temperature of the reaction; higher temperatures gave more
reproducible results.
Catalyzed by Iridium Complexes
percentage
conversion
entry
catalyst
1
2
3
1/2 [Cp*IrCl(µ-Cl)]2
1/2 [Cp*IrCl(2)]BF4, 6
1/2 [Cp*IrCl(dfppe)]BF4
40
5
1
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM,
[acetophenone] ) 0.2 M, T ) 75 °C, [KOH] ) 2.5 mM, reaction time 2 h;
percentage conversions are the average of three experiments; dfppe )
(C6F5)2PCH2CH2P(C6F5)2.
Ruthenium Catalysts. The percentage conversions of ac-
etophenone to rac-phenylethanol for a range of aryl ruthenium
catalysts tested under identical conditions are given in Table 4.
The increased conversion for the mesitylene and cymene rings
(entries 2 and 3) compared with the toluene derivative (entry
1) indicates that under these conditions increased electron
density on the metal increases the rate of hydrogen transfer.
This is consistent with the improvements observed when the
σ-donor DPEN is added (entries 4 and 5). When the amine is
substituted for bis(diphenylphosphino)ethane (dppe, entry 6),
rate enhancement is again observed, but this is lesser, perhaps
due to the ability of the amine to react by the metal-ligand
bifunctional mechanism. The behavior of the dfppm (2)
complexes (8, 11, and 7, entries 7-9) is difficult to predict, as
2 varies from a poison (in complex 8, entry 7) to a good
promoter (in complex 7, entry 9). The reduction in reaction rate
when 2 and mesitylene are coordinated may be due to steric
bulk, as the spatial distribution of substituents around the metal
site is more even on mesitylene than cymene. The activity of 7
is remarkable for a catalyst with no Brønsted basic ligand site
and was deemed worthy of further study.
identical within 3σ for the three complexes (2.384(4) Å), and
the distances of the bridging Ru-Cl bond trans to chloride are
similar (2.407(2) to 2.4343(8) Å), but the distances of the
bridging Ru-Cl bond trans to phosphorus range from 2.472-
(4) to 2.5960(8) Å. The Ru-P distances of 11 are shorter than
those of 8, and consequently the P-Ru-P angle is slightly
larger.
The 19F NMR spectrum indicates hindered rotation about one
pair of P-C6F5 bonds. A minimum value of ∆Gq of 61 kJ mol-1
was calculated47 from variable-temperature 19F NMR spectra.
The spectra did not allow calculation of ∆Gq for rotation about
the other pair of P-C6F5 bonds.
Comparison of Catalytic Activity. Catalyst testing was
conducted to assess the effects of coordinating chelating
phosphine ligands on the ability of piano stool complexes to
catalyze hydrogen transfer. Of particular interest was the
comparison between a phenyl and fluorinated phenyl functional-
ity and the effects of tethering the phosphine and capping group
to make a highly chelating pocket. The reaction chosen for
comparison was the hydrogen transfer from propan-2-ol to
acetophenenone, as this is the substrate most commonly reported
in the literature (Scheme 1). Conditions chosen are consistent
with those reported in industrial patents for the operation of
catalysts for the reduction of alkyl aryl ketones and the
The best two catalysts, 1/2 [(η6-mesitylene)RuCl(µ-Cl)]2 +
DPEN and 7, were sampled after various times in order to better
measure the rate of catalysis. The results are presented in Table
5. Under these conditions we found that phosphine-substituted
catalyst 7 (turnover frequency, TOF, based on measurement after
10 min, 282 h-1) was twice as active as the amine-promoted

2666 Organometallics, Vol. 26, No. 10, 2007
Chart 1. Conversion of Acetophenone as a Function of Time
Marr et al.
catalyst precursor 1/2 [(η6-mesitylene)RuCl(µ-Cl)]2 + DPEN
(TOF based on measurement after 10 min, 138 h-1).
center unless the phosphine was tethered to the cyclopentadienyl,
at which point the activity of the catalyst increased significantly.
The iridium phosphine complexes tested were very poor
catalysts.
Rhodium Catalysts. The results of hydrogenation of ac-
etophenone catalyzed by rhodium complexes are given in Table
6. The comparison of tetramethyl and pentamethyl cyclopen-
tadienyl derivatives (entries 1 and 2) would suggest that
increasing the electron density can increase the rate of transfer
hydrogenation, and consistent with this is the slight increase in
conversion achieved for the tetramethyl cyclopentadienyl com-
plex on coordinating dppe (entry 6). There was also an increased
conversion on the addition of DPEN to pentamethylcyclo-
pentadienyl rhodium (entry 3), but this must be attributed to
the metal-ligand bifuntional mechanism (Scheme 2), as all other
bidentate modifiers added were found to decrease the conversion
of acetophenone in 2 h (entries 4, 7, 8, and 9) relative to the
parent chloride (entry 2). The poisoning effect of the phosphine
ligands tested is reversed by tethering the phosphine and
cyclopentadienyl functionality together (either as a Cp-P or
Cp-PP complex), as illustrated by entries 10 and 11. Complex
9 had hydrogen transfer activity similar to [Cp*RhCl(µ-Cl)]2,
whereas 1 was the most active hydrogen transfer catalyst tested.
The excellent activity of 1 warranted further study. The
hydrogen transfer activity as a function of time is presented in
Table 7. The TOF of the catalyst, based on the 10 min
A graph of the conversion of acetophenone as a function of
time for the excellent catalysts 1/2 [(η6-mesitylene)RuCl(µ-Cl)]2
+ DPEN, 1, and 7 is given (Chart 1). 1 outperforms the other
catalysts and shows a profile consistent with a high turnover
that becomes rate limited at lower substrate concentrations.
Amine-promoted ruthenium performs as expected for a slower
catalyst. The profile of 7 reveals that this reaction slowed
considerably long before full conversion was achieved. This is
a result of either a catalyst that is poor at low substrate
concentrations or one that slowly decomposes under the reaction
conditions. In order to test the stability of piano stool complexes
of chelating phosphines under hydrogen transfer conditions, in
situ NMR studies were conducted.
In Situ NMR Studies. In order to verify that the active
catalysts contained phosphine-bound species and try to deter-
mine the resting state of the catalyst, selected catalysts were
analyzed by 31P NMR during catalytic activity.
In Situ NMR of a Ruthenium Catalyst. [(η6-mesitylene)-
RuCl(κP,κP-dppe)]BF4 dissolved in propan-2-ol at 343 K
exhibited one resonance at δP 71.8 in the 31P{1H} NMR
spectrum. The addition of potassium hydroxide led to the
disappearance of this signal and the appearance of resonances
at δP 82.7 and 35.4 (Figure 7). The addition of acetophenone
did not change the spectrum. During catalytic activity a further
resonance at δP 83.2 grew in. The resting state appeared to
resonate at δP 35.4, as this signal increased in intensity relative
to the other resonances throughout. After 2 h the solution was
cooled, and the room-temperature spectrum consisted of a major
resonance at δP 82.7, a lesser resonance at δP 35.4, and a
resonance at δP 64.2 of low intensity. Removal of the solvent
and addition of chloroform led to the complete regeneration of
the initial spectrum (δP 71.8), revealing the catalyst to be intact.
In the presence of propan-2-ol and deuterium-labeled propan-
2-ol (d8) a strong broad signal was observed at δP 82.1 in
measurement, was 360 h-1
.
Iridium Catalysts. The activity of iridium-based catalysts
was found to be comparatively low under these conditions
(Table 8).
Summary of Results from the Initial Catalyst Tests.
Ruthenium catalysts exhibited a slightly lower inherent rate of
hydrogen transfer and benefited from the coordination of good
donor chelating ligands; the “capping” aryl group and bidentate
ligand must be matched to prevent steric overcrowding and
maximize rates. The rhodium catalysts benefited from a pen-
tamethyl cyclopentadienyl group as a capping ligand, and the
electron donation from this ligand makes the attachment of
bidentate donor ligands unnecessary. A reduction in rate was
observed for chelating phosphine ligands attached to the metal

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2667
Figure 7. 31P spectrum resulting from exposing [(η6-mesitylene)RuCl(κP,κP-dppe)]BF4 to hydrogen transfer conditions.
Scheme 5. Formation of the Ruthenium Hydride
addition to the δP 82.7 resonance, whereas the resonances
observed at δP 35.4 in pure propan-2-ol were absent. We
therefore suggest that [(η6-mesitylene)Ru(H)(κP,κP-dppe)]+ and
its deuterium analogue resonate at δP 82-83, and these species
have a close association with the organic reactants, leading to
differences in chemical shift for [(η6-mesitylene)Ru(H)(κP,κP-
dppe)]OC3H6 (δP 82.7) and [(η6-mesitylene)Ru(H)(κP,κP-dppe)]-
Potassium hydroxide was added to the reaction, and the resultant
OC8H8 (δP 83.2). When the proton-decoupled spectrum was
1
solution gave one doublet at δP -66.1 with JRhP ) 213 Hz.
recorded and rapidly followed by a proton-coupled spectrum,
These results indicate the phosphine remains chelated to rhodium
the resonance at δ 82.7 collapsed in intensity. The resulting
during catalysis. The shift to lower frequency is similar to the
resonance is believed to be a doublet of multiplets, JPH ) 38
shift that occurred in the [(η6-mesitylene)RuCl(dppe)]BF4
Hz, consistent with a hydride complex.53,54
experiment, and by analogy we assign the resonance at δP -66.1
Substitution of propan-2-ol by propan-1-ol led to the forma-
tion of a resonance at δP 49.2 but no resonance at δP 35.4. On
to [Cp*Rh{OCH(CH3)2}(κP,κP-2)]+.
Conclusions from in Situ NMR Experiments. In both cases
(phosphine-modified Ru(II) and Rh(III)) the spectrum resulting
from the addition of potassium hydroxide in propan-2-ol
contained a resonance shifted to lower frequency by δP 33-
36. We assign this resonance to the primary product of the
removal of chloride and association of prop-2-oxide. The
addition of the substrate does not greatly affect the spectrum,
and in the middle of catalytic activity the resonance at lower
frequency dominates. We conclude that the mechanism, as
depicted in Scheme 1, lies to the right-hand side with a rate-
limiting insertion migration reaction as depicted (for Ru) in
Scheme 5. This explains why the addition of a Brønsted basic
site and the initiation of a ligand-assisted pathway as depicted
in Scheme 2 increase the rate of the reaction.
addition of potassium hydroxide, the intensity of this resonance
increased as the reaction proceeded, behaving similarly to the
resonance at δ 35.4 observed in propan-2-ol. We therefore assign
the resonance at δP 49.2 to [(η6-mesitylene)Ru(OCH2CH2CH3)-
(κP,κP-dppe)]+ and that at δP 35.4 to [(η6-mesitylene)Ru{OCH-
(CH3)2}(κP,κP-dppe)]+.
Attempts to isolate and characterize a prop-2-oxide derivative
failed. The presence of a large quantity of a prop-2-oxide
complex in solution during catalysis would suggest that the rate
could be increased by shifting the â-H transfer equilibrium
toward the hydride (Scheme 5). The formation of a crowded
transition state for this reaction may explain why some
combinations of capping group and bidentate ligand lead to low
rates of hydrogen transfer.
The low hydrogen transfer activity of the parent ruthenium
dichloride complexes suggests that the removal of chloride is
limiting the rate of reaction in the absence of phosphine donors.
The addition of phosphine to Ru helps to stabilize the primary
product of chloride dissociation and therefore increase the rate.
This is not the case for pentamethylcyclopentadienyl rhodium
catalysts, and they show a high rate of reaction without
promoters.
In Situ NMR Spectroscopy of a Rhodium Catalyst. The
31P NMR spectrum of salt 5 in propan-2-ol and acetophenone
at 343 K exhibited a doublet at δP -33.2 with 1JRhP ) 130 Hz.
(53) Harding, P. A.; Robinson, S. D.; Henrick, K. J. Chem. Soc., Dalton
Trans. 1988, 415.
(54) (a) Guan, H.; Iimura, M.; Magee, M. P.; Norton, J. R.; Janak, K. E.
Organometallics 2003, 22, 4084. (b) Josh, F. L.; Roundhill, D. M.
Organometallics 1992, 11, 1749.

2668 Organometallics, Vol. 26, No. 10, 2007
Table 9. Transfer Hydrogenationa of Acetophenone at Higher Substrate Concentration
Marr et al.
catalyst
conversion/%
rate/mol dm-3 h-1
TOF/h-1
[(η5,κP,κP-C5Me4CH2-2-C5F3N-4-PPhCH2CH2PPh2)RhCl]BF4, 1
[(η6-p-cymene)RuCl(2)]BF4, 7
65
52
13
0.65
0.52
0.13
325
260
65
[(η6-mesitylene)RuCl(dppe)]BF4
a All reactions were conducted in propan-2-ol, [catalyst] ) 2 mM, [acetophenone] ) 2 M, T ) 75 °C, [KOH] ) 2.5 mM, reaction time 2 h.
Table 10. Oppenauer Oxidationa of rac-Phenyl Alcohol
catalyst
conversion %
TOF
1/2 [Cp*RhCl(µ-Cl)]2
57
48
16
74
29
24
8
[(η5,κP,κP-C5Me4CH2-2-C5F3N-4-PPhCH2CH2PPh2)RhCl]BF4, 1
[(η6-p-cymene)RuCl(2)]BF4, 7
1/2 [Cp*IrCl(µ-Cl)]2
37
a All reactions were conducted in propan-2-one, [catalyst] ) 2 mM, [rac-phenyl alcohol] ) 0.2 mM, T ) 57 °C, [KOH] ) 2.5 mM, reaction time 2 h.
The precise metal environment engineered by our highly
chelating ligands stimulates desired reactivity but also protects
the catalytic center from decomposition. The catalysts 9 and 1
with “tethered” ligands of Cp-P and Cp-PP functionalities
promote hydrogen transfer relative to the parent chloride and
bidentate phosphine complexes. We have successfully created
a pocket for reactivity that stimulates hydrogen transfer.
Higher Substrate Concentrations. The commercial success
of industrial reactions such as hydrogen transfer depends upon
conversion of a certain quantity of substrate per unit time, and
therefore we thought it relevant to test our catalysts at a higher
substrate concentration. The conversions of acetophenone over
2 h using catalysts 1, 7, and [(η6-mesitylene)RuCl(κP,κP-dppe)]-
BF4 for 10 times higher substrate concentration are given in
Table 9. Again the best catalyst was the tethered cyclopenta-
dienyl phosphine rhodium(III) complex 1, TOF 325 h-1, 0.65
with untethered and tethered η5,κP,κP-cyclopentadienyl groups,
and both classes of complex have shown promise as hydrogen
transfer catalysts. In tests to compare piano stool rhodium
complexes containing η5,κP,κP-Cp-PP, η5,κP-Cp-P and un-
coupled Cp and diphosphine functionalities, the activity of the
η5,κP,κP-Cp-PP complex 1 was unexpectedly high, and this
result may enable the design of highly stable and enantioselec-
tive hydrogen transfer catalysts.
Experimental Section
General Considerations. The compounds bis(dichlorophosphi-
no)methane (Strem) and bromopentafluorobenzene (Fluorochem)
were used as supplied. The complexes [Cp*RhCl(µ-Cl)]2,57 [Cp*IrCl-
(µ-Cl)]2,57 [(η6-cymene)RuCl(µ-Cl)]2,58 and 132 were prepared
as described. The complex [(η6-mesitylene)RuCl(µ-Cl)]2 was
prepared by heating [(η6-cymene)RuCl(µ-Cl)]2 in mesitylene as
described for the preparation of tetra- and hexamethylbenzene
complexes.59 The preparation of 2 was performed under dinitrogen
using diethyl ether dried by distillation from sodium/benzophenone
and stored over molecular sieves (4 Å) under dinitrogen. No
precautions to exclude air or moisture were taken for the other
preparations.
1H, 19F, and 31P NMR spectra were recorded at 25 °C using
Bruker DPX300 or DRX500 spectrometers. 1H NMR spectra
(300.01 or 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. EI, LSIMS, and ES mass spectra were recorded
on a VG Autospec X series mass spectrometer. Elemental analyses
were carried out by ASEP, The School of Chemistry and Chemical
Engineering, Queen’s University Belfast.
(C6F5)2PCH2P(C6F5)2 (2). Bis(dichlorophosphino)methane (0.6
cm3, 4.6 mmol) was added by syringe to a solution of C6F5MgBr
in diethyl ether (100 cm3), freshly prepared from C6F5Br (7 g, 0.27
mol) and magnesium (0.56 g, 0.23 mol). Heat was evolved, and a
white precipitate was formed. The mixture was left at ambient
temperature for 18 h, then opened to the atmosphere. Water (30
cm3) and dichloromethane (40 cm3) were added and the organic
layer was separated. The aqueous layer was extracted with
dichloromethane (3 × 20 cm3) and the combined organic solution
dried over magnesium sulfate. The solution was filtered, concen-
trated by rotary evaporation, and filtered through 4 cm of neutral
mol dm-3 h-1
.
Oxidizing Alcohols to Aldehydes. The synthesis of an
aldehyde by catalytic oxidation (dehydrogenation) via hydrogen
transfer to butanone or propan-2-one is termed Oppenauer
oxidation.55 This reaction is of great interest to chemists
attempting to prepare commodity chemicals from the side
products of agriculture, as many of the molecules derived from
natural sources are polyols. Reduction of these alcohols to
carbonyls is useful, as ketones and aldehydes have rich and
useful reaction chemistries. Recently derivatives of group 9
cyclopentadienyl complexes were found to be excellent catalysts
for Oppenauer oxidation, opening up the possibility of simple
oxidation of aliphatic alcohols in propan-2-one or butenone.56
We compared the activity of 1, 7, [Cp*RhCl(µ-Cl)]2, and
[Cp*IrCl(µ-Cl)]2 as catalysts for the oxidation of rac-phenyl
alcohol (Table 10).
Oppenauer oxidation employing Ru and Rh catalysts is slower
in propan-2-one than hydrogenation of the ketone in propan-
2-ol, and this reflects the lower reaction temperature. [Cp*RhCl-
(µ-Cl)]2, 1, and particularly [Cp*IrCl(µ-Cl)]2 perform well, but
the ruthenium catalyst 7 is comparatively poor at catalyzing
this oxidation.
Conclusion
The pentafluorophenyl-substituted diphosphine (C6F5)2PCH2P-
(C6F5)2 (2) is a versatile ligand that enabled the preparation of
a range of piano stool complexes of Ru, Rh, and Ir. Piano stool
complexes have been prepared and structurally characterized
(57) White, C.; Yates, A.; Maitlis, P. M. Inorg. Synth. 1992, 29, 228.
(58) Bennett, M. A.; Smith, A. K. J. Chem. Soc., Dalton Trans. 1983,
1571.
(59) Tocher, D. A.; Gould, R. O.; Stephenson, T. A.; Bennett, M. A.;
Ennett, J. P.; Matheson, T. W.; Sawyer, L.; Shah, V. K. J. Chem. Soc.,
Dalton Trans. 1983, 1571.
(55) Oppenauer, R. V. Recl. TraV. Chem. Pays-Bas 1937, 56, 137.
(56) (a) Suzuki, T.; Morita, K.; Tsuchida, M.; Hiroi, K. J. Org. Chem.
2003, 68, 1601. (b) Hanasaka, F.; Fujita, K. I.; Yamaguchi, R. Organo-
metallics 2006, 25, 826.

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2669
alumina (6% H2O). The alumina was washed with dichloromethane
(3 × 10 cm3). Removal of the solvent from the combined filtrate
and washings by rotary evaporation afforded 2 as a white solid:
2.78 g (81.4%). 1H NMR (CDCl3): δ 3.81 (2H, t, 2JPH ) 6.6 Hz).
19F NMR (CDCl3): δ -130.28 (8F, m, Fortho), -147.81 (4F, t, 2JFF
) 20.9 Hz, Fpara), -159.77 (8F, m, Fmeta). 31P{1H} NMR (CDCl3):
δ -52.1 (second-order pattern of [A[X4][Y4][Z2]2]2 spin system).
LSIMS, m/z (rel int): 744 (23%, M+), 577 (100%, [M - C6F5]+).
HRLSIMS: calcd for C25H2F20P2, 743.9313; found M+, 743.9323.
Anal. Calcd for C25H2F20P2: C, 40.32; H, 0.27. Found: C, 40.05;
H, 0.43.
Fpara), -147.98 (2F, t, 3JFF ) 20.0 Hz, Fpara), -153.31 (0.8F, 10BF4),
-153.36 (3.2F, 11BF4), -159.08 (4F, m, Fmeta), -161.23 (2F, br,
F
meta), -162.01 (2F, br, Fmeta). 31P{1H} NMR (CD3Cl): δ -27.0
(s). ESMS, m/z: 1015 (100%, [M - BF4]+), 881 (9%, [M - BF4
- Cl]+). HRESMS: calcd for C34H1635ClF20P2102Ru, 1014.9140;
found M+, 1014.9141. Anal. Calcd for C35H16BClF24P2Ru: C,
38.13; H, 1.45. Found: C, 37.35; H, 2.77.
[(η6-Mesitylene)RuCl{κP,κP-(C6F5)2PCH2P(C6F5)2}]BF4 (8).
[(η6-Mesitylene)RuCl(µ-Cl)]2 (0.190 g, 0.30 mmol), 2 (0.446 g,
0.60 mmol), and NaBF4 (0.132 g, 1.20 mmol) were treated as for
the synthesis of salt 5. Salt 8 was obtained as a green solid and
was recrystallized from dichloromethane (0.600 g, 91.9%). 1H NMR
(CD3Cl): δ 6.38 (1H, m, CHH′), 5.91 (3H, s, C6H3), 5.34 (2H, s,
CH2Cl2), 4.77 (1H, m, CHH′), 2.30 (9H, s, C6Me3). 19F NMR
(CD3Cl): δ -125.78 (4F, br, Fortho), -129.82 (4F, br, Fortho),
-139.20 (2F, t, 3JFF ) 21.0 Hz, Fpara), -142.69 (2F, t, 2JFF ) 20.5
Hz, Fpara), -152.79 (0.8F, 10BF4), -152.84 (3.2F, 11BF4), -154.41
(4F, m, Fmeta), -156.77 (4F, m, Fmeta). 31P{1H} NMR (CD3Cl): δ
-25.4 (s). LSIMS, m/z: 1001 (100%, [M - BF4]+), 966 (8%, [M
- BF4 - Cl]+). HRLSIMS: calcd for C34H1435ClF20P2102Ru,
1000.8987; found M+, 1000.8966. Anal. Calcd for C35H14BClF24P2-
Ru.CH2Cl2: C, 35.83; H, 1.37. Found: C, 35.73; H, 1.41.
[{(η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}RhCl2] (9). A
slurry of [Cp*RhCl(µ-Cl)]2 (0.141 g, 0.23 mmol) and 2 (0.312 g,
0.42 mmol) in benzene (100 cm3) was heated at reflux for 95 h.
After cooling, the solution was concentrated to ca. 50 cm3 by rotary
evaporation and hexane (50 cm3) added. The resulting red-brown
precipitate was filtered off, washed with hot hexane (3 × 10 cm3),
and dried in Vacuo (0.285 g, 70%). 1H NMR ((CD3)2CO): δ 5.15
(1H, dd, 2JPH ) 14.3 Hz, 2JHH′ ) 14.3 Hz, PCHH′P), 4.35 (1H, dd,
(C6F5)2PCH2P(O)(C6F5)2 (3) and (C6F5)2P(O)CH2P(O)(C6F5)2
(4). Aliquots of aqueous hydrogen peroxide were added to a sample
of 2 in CDCl3. The formation of 3 and 4 was monitored by 19F and
31P{1H} NMR spectroscopy. Removal of the solvents by rotary
evaporation gave a sample of 4, which was characterized by mass
spectrometry. 3: 19F (CDCl3): δ -131.73 (8F, m br, Fortho) -142.68
3
(4F, t, JFF ) 21.0 Hz, Fmeta), -159.46 (8F, m, Fpara). 31P{1H}
3
3
(CDCl3): δ 15.0 (d, JPP ) 78 Hz, PO), -60.1 (dquintquint, JPP
) 78 Hz, JPF ) 34 Hz, JPF ) 10 Hz, P). 4: 1H NMR (CDCl3):
3
5
δ 4.02 (4F, t, JPH ) 14.9 Hz). 19F NMR (CDCl3): δ -132.22
2
2
(8F, m, Fortho), -142.28 (4F, t, JFF ) 20.4 Hz, Fpara), -157.70
(8F, m, Fmeta). 31P{1H} NMR (CDCl3): δ 8.9 (s). LSIMS, m/z (rel
int): 776 (63%, M+), 609 (100%, [M - C6F5]+). HRLSIMS: calcd
for C25H2F20O2P2, 775.9210; found M+, 775.9220.
[Cp*RhCl{κP,κP-(C6F5)2PCH2P(C6F5)2}]BF4 (5). [Cp*RhCl-
(µ-Cl)]2 (0.068 g, 0.11 mmol), 2 (0.149 g, 0.2 mmol), and NaBF4
(0.52 g, 4.7 mmol) were treated as for the synthesis of [Cp*RhCl-
{(C6F5)2PCH2CH2P(C6F5)2}][BF4].48 Salt 5 was obtained as an
1
2JPH′ ) 14.3 Hz, 2JHH′ ) 14.3 Hz, PCHH′P), 4.18 (1H, dd, 2JHH′
)
)
orange solid (0.187 g, 92%). H NMR ((CD3)2CO): δ 5.66 (1H,
4
2
2
m, CHH′), 4.93 (1H, m, CHH′), 1.78 (15H, t, JPH ) 6.1 Hz,
18.6 Hz, JPH ) 6.1 Hz, 1 H, C5CHH′C6F4), 3.96 (1H, d, JHH′
C5Me5). 19F NMR ((CD3)2CO): δ -128.71 (4F, br, Fortho), -130.31
4
18.6 Hz, C5CHH′C6F4), 1.94 (3H, d, JPH ) 7.6 Hz, CH3), 1.86
(3H, d, 4JP,H ) 5.6 Hz, CH3), 1.83 (3H, s, CH3), 1.40 (3H, d, 4JPH
) 0.9 Hz, CH3). 19F NMR ((CD3)2CO): δ -121.25 (1F, m),
-129.61 (2F, m), -130.14 (2F, m), -132.61 (2F, m), -136.99
(1F, m), -149.29 (1F, td, J ) 20.2 Hz, J ) 9.0 Hz), -150.26 (1F,
m), -152.23 (1F, m), -152.43 (1F, m), -158.43 (1F, t, J ) 20.2
Hz), -162.37 (2F, m), -163.02 (2F, m), -163.54 (2F, m). 31P{1H}
3
(4F, br, Fortho), -145.90 (2F, t, JFF ) 20.3 Hz, Fpara), -147.42
(2F, t, 2JFF ) 20.3 Hz, Fpara), -152.54 (0.8F, 10BF4), -152.60 (3.2F,
11BF4), -159.81 (4F, br, Fmeta), -161.16 (4F, br, Fmeta). 31P{1H}
1
NMR ((CD3)2CO): δ -34.0 (dm, JRhP )132 Hz). LSIMS, m/z:
1017 (100%, [M - BF4]+), 982 (23%, [M - BF4 - Cl]+).
HRLSIMS: calcd for C35H1735ClF20P2Rh, 1016.9237; found M+,
1016.9229. Anal. Calcd for C35H17BClF24P2Rh: C, 38.00; H, 1.54.
Found: C, 37.78; H, 1.44.
1
2
NMR ((CD3)2CO): δ 43.1 (ddm, JRhP )170 Hz, JPP ) 185 Hz,
PC6F4CH2), -66.6 (dquint, 2JPP ) 185 Hz, 3JPF ) 38 Hz, P(C6F5)2).
LSIMS: 997 (100%, [M - Cl]+), 961 (29%, [M - 2Cl]+).
HRLSIMS: calcd for C35H1635ClF19P2Rh 996.9167; found M+,
996.9173. Anal. Calcd for C35H16Cl2F19P2Rh‚1/2CH2Cl2: C, 39.63;
H, 1.59. Found: C, 40.65; H, 1.55.
[Cp*IrCl{κP,κP-(C6F5)2PCH2P(C6F5)2}]BF4 (6). [Cp*IrCl(µ-
Cl)]2 (0.100 g, 0.134 mmol), 2 (0.187 g, 0.251 mmol), and NaBF4
(0.119 g, 0.250 mmol) were treated as for the synthesis of salt 5.
1
Salt 6 was obtained as a yellow solid (0.203 g, 65.5%). H NMR
[{(η5,κP-C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}IrCl2] (10). A
slurry of [Cp*IrCl(µ-Cl)]2 (0.106 g, 0.133 mmol) and 2 (0.198 g,
0.266 mmol) in benzene (100 cm3) was heated at reflux for 72 h.
After cooling, the solution was concentrated to ca. 50 cm3 by rotary
evaporation and hexane (50 cm3) added. The resulting yellow
precipitate was filtered off and washed with hot hexane (3 × 10
cm3). The product was extracted into dichloromethane and the
solution filtered through Celite. The solvent was removed by rotary
evaporation and put through a Celite plug eluting with dichloro-
methane. The solvent was removed by rotary evaporation and the
solid dried in Vacuo, yielding 0.12 g of impure 10 as a yellow
powder. All attempts to purify 10 were unsuccessful, and charac-
terization is based on the mass spectral data and comparison of the
((CD3)2CO): δ 5.75 (1H, m, CHH′), 4.05 (1H, m, CHH′), 1.691
(15H, t, 4JPH ) 3.9 Hz, C5Me5). 19F NMR ((CD3)2CO): δ -128.23
3
(4F, br, Fortho), -130.04 (4F, br, Fortho), -140.46 (2F, t, JFF
)
3
20.6 Hz, Fpara), -142.42 (2F, t, JFF ) 20.7 Hz, Fpara), -154.80
(0.8F, 10BF4), -154.86 (3.2F, 11BF4), -155.72 (4F, m, Fmeta),
-157.22 (2F, br, Fmeta), -157.80 (2F, br, Fmeta). 31P{1H}
NMR ((CD3)2CO): δ -73.1 (s). LSIMS, m/z: 1107 (100%, [M -
BF4]+), 1072 (8%, [M - BF4 - Cl]+). HRLSIMS: calcd for
C35H1735ClF20P2193Ir, 1106.9803; found M+, 1106.9846. Anal.
Calcd for C35H17BClF24P2Ir: C, 35.20; H, 1.42. Found: C, 34.53;
H, 1.58.
[(η6-p-Cymene)RuCl{κP,κP-(C6F5)2PCH2P(C6F5)2}]BF4 (7).
[(η6-p-Cymene)RuCl(µ-Cl)]2 (0.100 g, 0.163 mmol), 2 (0.243 g,
0.326 mmol), and NaBF4 (0.06 g, 0.601 mmol) were treated as for
the synthesis of salt 5. Salt 7 was obtained as an orange solid and
was recrystallized from acetone (0.330 g, 91.9%). 1H NMR
1
NMR spectroscopic data with those of 9. H NMR (CD3Cl): δ
2
2
5.21 (1H, ddm, JPH ) 13.1 Hz, JHH′ ) 13.1 Hz, PCHH′P), 4.31
2
(1H, m, PCHH′P), 3.93 (1H, dm, JHH′ ) 18.7 Hz, 1 H,
C5CHH′C6F4), 3.77 (1H, dm, 2JHH′ ) 18.7 Hz, C5CHH′C6F4), 1.91
(3H, d, 4JPH ) 5.4 Hz, CH3), 1.84 (3H, s, CH3), 1.77 (3H, d, 4JP,H
3
2
(CD3Cl): δ 6.27 (2H, d, JHH ) 6.9 Hz, C6H4), 6.09 (2H, d, JHH
) 6.9 Hz, CH), 4.92 (1H, m, CHH′), 4.01 (1H, m, CHH′), 2.58
(1H, sept, 3JHH ) 6.8 Hz, CH(CH3)2), 1.89 (3H, s, CH3), 1.16 (6H,
d, 3JHH ) 6.8 Hz CH(CH3)2). 19F NMR (CD3Cl): δ -128.95 (4F,
br, Fortho), -129.76 (4F, br, Fortho), -145.94 (2F, t, 3JFF ) 20.5 Hz,
) 4.2 Hz, CH3), 1.27 (3H, d, JPH ) 0.9 Hz, CH3). 19F NMR
4
(CD3Cl): δ -121.20 (1F, m), -128.71 (2F, m), -129.67 (2F, m),
-130.56 (2F, m), -135.86 (1F, m), -146.04 (1F, m), -146.58
(1F, t, J ) 20.3 Hz), -148.28 (1F, t, J ) 20.6 Hz), -149.32 (1F,

2670 Organometallics, Vol. 26, No. 10, 2007
Table 11. Crystal Data and Structure Refinement for Compounds 2, 5, 6, 8‚CH2Cl2, 9, and 11
Marr et al.
2
5
6
8‚CH2Cl2
9
11
formula
C25H2F20P2
C35H17BClF24P2Rh
C35H17BClF24-
P2Ir
C34H14BClF24P2Ru‚
CH2Cl2
C35H16Cl2F19-
P2Rh
C34H14Cl4F20-
P2Ru2
fw
744.21
0.62 × 0.38 ×
0.20
1104.60
1193.89
1172.65
1033.23
1208.33
0.31 × 0.12 ×
0.10
cryst dimens, mm
0.40 × 0.22 ×
0.64 × 0.50 ×
0.98 × 0.28 ×
0.56 × 0.28 ×
0.19
0.34
0.26
0.26
T, K
153(2)
triclinic
P1h
298(2)
orthorhombic
Pna2(1)
298(2)
orthorhombic
Pnma
153(2)
monoclinic
P21/c
153(2)
orthorhombic
Pna2(1)
298(2)
triclinic
P1h
cryst syst
space group
unit cell dimens
a, Å
b, Å
c, Å
10.440(3)
11.061(3)
12.053(3)
82.896(4)
67.141(4)
80.910(4)
1263.5(5)
2
17.933(3)
9.7034(16)
21.676(4)
90
90
90
3771.8(11)
4
1.945
17.95(4)
21.60(5
9.71(2)
90
90
90
3765(14)
4
2.106
15.047(2)
12.572(2)
21.976(4)
90
109.372(3)
90
3921.7(11)
4
1.986
14.6421(13)
12.4541(11)
19.7153(17)
90
90
90
3595.2(5)
4
1.909
10.410(4)
11.037(5)
18.457(7)
104.793(6)
92.372(6)
104.433(6)
1973.2(14)
2
R, deg
â, deg
γ, deg
V, Å3
Z
calc density, g cm-3
F(000)
1.956
724
2.034
1168
2160
2288
2288
2024
θ, deg
1.84-28.89
2.56-24.99
1.89-28.29
1.43-28.35
1.93-28.28
1.15-25.00
abs coeff, mm-1
total no. of data
no. of unique
data, Rint
final R indices
[I > 2σ(I)]
0.339
0.758
3.851
0.831
0.841
1.241
14 325
5531, 0.0879
31 277
6608, 0.0353
8755
4020, 0.2562
44 514
9092, 0.1307
40 403
8342, 0.0753
14 571
6927, 0.1009
R1 ) 0.0426
R1 ) 0.0499
R1 ) 0.1139
R1 ) 0.0665
R1 ) 0.0447
R1 ) 0.0815
wR2 ) 0.1034
R1 ) 0.0648
wR2 ) 0.1124
1.034
wR2 ) 0.1254
R1 ) 0.0626
wR2 ) 0.1387
1.029
wR2 ) 0.1633
R1 ) 0.2497
wR2 ) 0.2736
1.082
wR2 ) 0.1452
R1 ) 0.1470
wR2 ) 0.1849
1.009
wR2 ) 0.0856
R1 ) 0.0705
wR2 ) 0.0955
1.017
wR2 ) 0.1883
R1 ) 0.1857
wR2 ) 0.2547
0.919
R indices (all data)
GoF on F2
largest diff peak
and hole, e Å-3
Flack param
0.420, -0.365
1.845, -0.598
5.517, -4.844
1.901, -0.932
1.308, -1.216
1.468, -1.095
0.00
0.61(3)
t, J ) 20.5 Hz), -155.73 (1F, t, J ) 21.7 Hz), -159.36
(2F, m), -160.06 (4F, m). 31P{1H} NMR (CD3Cl): δ 11.0
(dm, JPP ) 181 Hz, PC6F4CH2), -69.2 (dquint, JPP ) 181 Hz,
3JPF ) 35 Hz, P(C6F5)2). LSIMS: 1087 (100%, [M - Cl]+).
HRLSIMS: calcd for C35H1635ClF19P2191Ir 1084.97500; found M+,
1084.971845.
The structure was solved by direct methods and refined with the
program package SHELXTL version 5.62 The non-hydrogen atoms
were refined with anisotropic thermal parameters. Hydrogen atom
positions were added, and idealized positions and a riding model
with fixed thermal parameters (Uij ) 1.2Ueq for the atom to which
they are bonded (1.5 for CH3)) were used for subsequent refine-
2
2
2
2
[(η6-Mesitylene)Ru(µ-Cl)3RuCl{κP,κP-(C6F5)2PCH2P-
(C6F5)2}] (11). A solution of [(η6-mesitylene)RuCl(µ-Cl)]2 (0.100
g, 0.17 mmol) and 2 (0.254 g, 0.34 mmol) in dichloromethane (20
cm3) and methanol (70 cm3) was heated at reflux for 20 h. The
solvent was removed by rotary evaporation and the resulting brown
solid extracted with dichloromethane (3 × 20 cm3). The extract
was filtered and concentrated to 20 cm3 by rotary evaporation.
Addition of diethyl ether (10 cm3) precipitated the product as a
ments. The function minimized was ∑[w(|Fo| - |Fc| )] with
2
reflection weights w-1 ) [σ2|Fo| + (g1P)2 + (g2P)] where P )
2
2
[max|Fo| + 2|Fc| ]/3. CCDC 633807 (2), 633810 (5), 633809 (6),
633808 (8), 294507 (9), and 633811 (11) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
Catalyst Testing. Ketone Reduction, Substrate:Catalyst )
100:1. The catalyst precursor (0.01mmol) was added to acetophe-
none (0.12 g, 1 mmol) in propan-2-ol (4.875 mL), and the mixture
was heated to 75 °C under dinitrogen. Potassium hydroxide (0.125
cm3, 0.1 M in propan-2-ol) was added to the solution, which was
left stirring at that temperature for the reaction time (usually 2 h).
The solution was filtered through a silica plug, eluting with ethyl
acetate (2 × 5 cm3). The solvent was removed by rotary evaporation
to yield an acetophenone/phenyl alcohol mixture. The conversion
was determined via 1H NMR spectroscopy with the ratio between
the methyl resonance exhibiting a singlet at δ 2.47 for acetophenone
and a doublet at δ 1.35 for phenyl alcohol measured and given as
a % conversion.
Ketone Reduction, Substrate:Catalyst ) 1000:1. The catalyst
precursor (0.01 mmol) was added to acetophenone (1.20 g, 10
mmol) in propan-2-ol (4.875 cm3), and the mixture was heated to
75 °C under dinitrogen. Potassium hydroxide (0.125 cm3, 0.1 M in
propan-2-ol) was added to the solution, which was left stirring at
that temperature for 2 h. The solution was filtered through a silica
plug, eluting with ethyl acetate (2 × 5 cm3). The solvent was
1
brown solid (0.342 g, 95.8%). H NMR (CD3Cl): δ 6.26 (1H, m,
CHH′), 5.08 (3H, s, C6H3), 5.04 (1H, m, CHH′), 2.02 (9H, s,
C6Me3). 19F NMR (CD3Cl): δ -128.18 (8F, m, Fortho), -147.46
(4F, m, Fpara), -159.62 (4F, m, Fmeta). 31P{1H} NMR (CD3Cl): δ
-11.3 (s). LSIMS, m/z: 1209 (100%, [M - Cl]+), 1172 (70%, [M
- 2Cl]+). HRLSIMS: calcd for C34H1435Cl3F20P2102Ru2, 1207.7087;
found M+, 1207.7085. Anal. Calcd for C34H14Cl4F20P2Ru2: C,
33.77; H, 1.16. Found: C, 33.93; H, 1.24.
X-ray Crystallography. Crystals of (C6F5)2PCH2P(C6F5)2 (2)
were grown from petroleum ether (bp 40-60 °C), [Cp*RhCl(κP,κP-
2)]BF4 (5) from chloroform, [Cp*IrCl(κP,κP-2)]BF4 (6) and [(η6-
mesitylene)RuCl(κP,κP-2)]BF4 (8) from dichloromethane, [{(η5,κP-
C5Me4CH2C6F4-2-P(C6F5)CH2P(C6F5)2}RhCl2](9)fromdichloromethane/
hexane, and [(η6-mesitylene)Ru(µ-Cl)3RuCl(κP,κP-1)] (11) from
propan-2-one. Crystal and refinement data are given in Table 11.
Diffraction data were collected on a Bruker SMART diffractometer
using the SAINT-NT60 software with graphite-monochromated Mo
KR radiation. Lorentz and polarization corrections were applied.
Empirical absorption corrections were applied using SADABS.61
(60) SAINT-NT; Bruker AXS Inc.: Madison, WI, 1998.
(61) Sheldrick, G. M. SADABS; University of Go¨ttingen: Germany 1996.
(62) Sheldrick, G. M. SHELXTL version 5; Bruker AXS Inc.: Madison,
WI, 1998.

Piano Stool Phosphine Complexes for Hydrogen Transfer
Organometallics, Vol. 26, No. 10, 2007 2671
removed by rotary evaporation to yield an acetophenone/phenyl
alcohol mixture. The conversion was determined via 1H NMR
spectroscopy with the ratio between the methyl resonance exhibiting
a singlet at δ 2.47 for acetophenone and a doublet at δ 1.35 for
phenyl alcohol measured and given as a % conversion.
Alcohol (Oppenauer) Oxidation. The catalyst precursor
(0.01 mmol) was added to rac-phenyl ethanol (0.12 g, 0.1 mmol)
in degassed propan-2-one (4.875 cm3) and stirred at 57 °C
under dinitrogen. Potassium hydroxide (0.125 cm3, 0.1 M in
propan-2-one) was added to the solution, which was left stirring
at that temperature for 2 h under nitrogen. The solution was
filtered through a silica plug, eluting with ethyl acetate (2 × 5 cm3).
The solvent was removed by rotary evaporation to yield an
acetophenone/phenyl alcohol mixture. The conversion was deter-
mined via 1H NMR spectroscopy with the ratio between the methyl
resonance exhibiting a singlet at δ 2.47 for acetophenone and
doublet at δ 1.31-1.48 for phenyl alcohol63 measured and given
as a % conversion.
Acknowledgment. We thank the EU European Social Fund
for funding (to C.L.P.), the EPSRC mass spectrometry service
at Swansea, Prof. Paul Pringle of Bristol University, and Dr.
Pete Licence of Nottingham University.
Supporting Information Available: A listing of atomic
coordinates, anisotropic displacement parameters, bond distances,
and bond angles for 2, 5, 6, 8‚CH2Cl2, 9, and 11. This material is
available free of charge via the Internet at http://pubs.acs.org.
OM070089I
(63) Bianchi, D.; Cesti, P.; Bathstel, E. J. Org. Chem. 1988, 53, 5533.