Oxidation-state-dependent reactivity and catalytic properties of a Rh(I) complex formed from a redox-switchable hemilabile ligand

Article abstract of DOI:10.1021/ja9723601

The synthesis and characterization of a phosphinoalkylarene, redox-switchable hemilabile ligand (RHL), (η⁵-C₅H₅)Fe(η⁵-C₅H₄C₆H₄OCH₂CH₂PPh₂) (1), are reported. This ligand, which incorporates a redox-active ferrocenyl group, exhibits oxidation-state-dependent bonding properties and, hence, affords electrochemical control over the electronic and steric environments of bound transition metal centers. Two equivalents of 1 complex to Rh(I) to yield a bis(phosphine), η⁶-arene complex [(η¹:η⁶-(η⁵-C₅H₅)Fe(η⁵-C₅H₄C₆H₄OCH₂CH₂PPh₂))(η¹-(η⁵ -C₅H₅)Fe(η⁵- C₅H₄C₆H₄OCH₂CH₂PPh₂))Rh]⁺BF₄^- (2). Single-crystal X-ray diffraction studies of 2·1.25CH₂Cl₂, as well as solution spectroscopic data of 2, are consistent with this formulated piano-stool geometry. Foremost, the properties of 2 as a function of bound RHL state-of-charge are extensively investigated. Interestingly, 2D ¹H NMR exchange spectroscopy (EXSY) studies demonstrate significantly faster intramolecular η⁶-arene, free arene exchange rates only upon oxidation of the ligand chelated to the Rh(I) center, as found in 2²⁺. This faster exchange rate was used as a qualitative measure of the increased lability of the η⁶-aryl group upon RHL oxidation. Moreover, activation parameters measured for the arene-arene exchange reaction of 2²⁺ also support a decrease in the Rh(I)-arene interaction only upon oxidation of the ferrocenyl group on the bound η⁶-arene moiety. In addition, changes in the stoichiometric and catalytic reactivity of 2 upon RHL oxidation are consistent with the observed charge-dependent arene-arene exchange behavior. Significantly, labilization of a weakly bound η⁶-arene moiety in 2²⁺ results in substantial increases in both the acetonitrile bonding affinity and allyl ethyl ether isomerization activity of the Rh(I)-RHL complex. In contrast, no significant changes in the reactivity of 2 were observed upon oxidation of the ligand which contained the ferrocenylarene not bound to the Rh(I) center, as found in 2⁺. This dependence of complex reactivity on RHL oxidation state demonstrates the utility of such novel metal complexes for the reversible, electrochemical control of transition metal center small molecule uptake or catalytic activity via the selective labilization of weakly coordinating groups upon ligand oxidation.

Full text of DOI:10.1021/ja9723601

J. Am. Chem. Soc. 1997, 119, 10743-10753
10743
Oxidation-State-Dependent Reactivity and Catalytic Properties
of a Rh(I) Complex Formed from a Redox-Switchable
Hemilabile Ligand
Caroline S. Slone,† Chad A. Mirkin,*,† Glenn P. A. Yap,‡ Ilia A. Guzei,‡ and
Arnold L. Rheingold‡
Contribution from the Department of Chemistry, 2145 Sheridan Road, Northwestern UniVersity,
EVanston, Illinois 60208, and Department of Chemistry and Biochemistry,
UniVersity of Delaware, Newark, Delaware 19716
ReceiVed July 14, 1997X
Abstract: The synthesis and characterization of a phosphinoalkylarene, redox-switchable hemilabile ligand (RHL),
(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2) (1), are reported. This ligand, which incorporates a redox-active ferrocenyl
group, exhibits oxidation-state-dependent bonding properties and, hence, affords electrochemical control over the
electronic and steric environments of bound transition metal centers. Two equivalents of 1 complex to Rh(I) to
yield a bis(phosphine), η6-arene complex [(η1:η6-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))(η1-(η5-C5H5)Fe(η5-
C5H4C6H4OCH2CH2PPh2))Rh]+BF4- (2). Single-crystal X-ray diffraction studies of 2‚1.25CH2Cl2, as well as solution
spectroscopic data of 2, are consistent with this formulated piano-stool geometry. Foremost, the properties of 2 as
a function of bound RHL state-of-charge are extensively investigated. Interestingly, 2D 1H NMR exchange
spectroscopy (EXSY) studies demonstrate significantly faster intramolecular η6-arene, free arene exchange rates
only upon oxidation of the ligand chelated to the Rh(I) center, as found in 22+. This faster exchange rate was used
as a qualitative measure of the increased lability of the η6-aryl group upon RHL oxidation. Moreover, activation
parameters measured for the arene-arene exchange reaction of 22+ also support a decrease in the Rh(I)-arene
interaction only upon oxidation of the ferrocenyl group on the bound η6-arene moiety. In addition, changes in the
stoichiometric and catalytic reactivity of 2 upon RHL oxidation are consistent with the observed charge-dependent
arene-arene exchange behavior. Significantly, labilization of a weakly bound η6-arene moiety in 22+ results in
substantial increases in both the acetonitrile bonding affinity and allyl ethyl ether isomerization activity of the
Rh(I)-RHL complex. In contrast, no significant changes in the reactivity of 2 were observed upon oxidation of the
ligand which contained the ferrocenylarene not bound to the Rh(I) center, as found in 2+. This dependence of
complex reactivity on RHL oxidation state demonstrates the utility of such novel metal complexes for the reversible,
electrochemical control of transition metal center small molecule uptake or catalytic activity via the selective labilization
of weakly coordinating groups upon ligand oxidation.
Introduction
centers have demonstrated that a metal steric environment can
greatly affect small molecule bonding properties1 and catalytic
activity.2 Furthermore, tuning of the electron-donating or the
electron-withdrawing properties of ligands also has been shown
to impact the electronic nature of a transition metal complex
and, therefore, its reactivity.3 Traditionally, such modulation
of transition metal complex reactivity has focused on changes
in the ligand’s properties before being bound to a metal center.
Recently, however, electrochemical methods have been used
to control the properties of metal complexes through the use of
redox switchable hemilabile ligands (RHLs).4-8
Strategies for controlling transition metal center reactivity
have relied on the synthetic manipulation of either the steric or
the electronic properties of bound ligands.1-3 Studies of ligands
which impose steric constraints on reactive transition metal
* Author to whom correspondence should be addressed.
† Northwestern University.
‡ University of Delaware.
X Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) (a) Gladysz, J. A.; Boone, B. J. Angew. Chem., Int. Ed. Engl. 1997,
36, 550. (b) Boone, B. J.; Klein, D. P.; Seyler, J. W.; Me´ndez, N. Q.; Arif,
A. M.; Gladysz, J. A. J. Am. Chem. Soc. 1996, 118, 2411. (c) Fernandez,
A. L.; Prock, A.; Giering, W. P. Organometallics 1996, 15, 2784. (d)
Beckett, R. P.; Burgess, V. A.; Davies, S. G.; Krippner, G. Y.; Sutton, K.
H.; Whittaker, M. Inorg. Chim. Acta 1996, 251, 265. (e) Mizutani, T.; Ema,
T.; Tomita, T.; Kuroda, Y.; Ogoshi, H. J. Am. Chem. Soc. 1994, 116, 4240.
(f) Collman, J. P.; Zhang, X.; Wong, K.; Brauman, J. I. J. Am. Chem. Soc.
1994, 116, 6245.
The ligand reported herein, (η5-C5H5)Fe(η5-C5H4C6H4OCH2-
CH2PPh2) (1), is one of an extensive series of RHLs studied in
our group which afford electrochemical control over both the
steric and electronic environments of bound transition metal
(3) (a) Schnyder, A.; Hintermann, L.; Togni, A. Angew. Chem., Int. Ed.
Engl. 1995, 34, 931. (b) RajanBabu, T. V.; Ayers, T. A.; Casalnuovo, A.
L J. Am. Chem. Soc. 1994, 116, 4101. (c) Casalnuovo, A. L.; RajanBabu,
T. V.; Ayers, T. A.; Warren, T. H. J. Am. Chem. Soc. 1994, 116, 9869. (d)
RajanBabu, T. V.; Ayers, T. A. Tetrahedron Lett. 1994, 35, 4295. (e) Chang,
S.; Heid, R. M.; Jacobsen, E. N. Tetrahedron Lett. 1994, 35, 669. (f)
Hawkins, J. M.; Loren, S.; Nambu, M. J. Am. Chem. Soc. 1994, 116, 1657.
(g) Nishiyama, H.; Yamaguchi, S.; Kondo, M.; Itoh, K. J. Org. Chem. 1992,
57, 4306. (h) Haltermann, R. L.; McEvoy, M. A. J. Am. Chem. Soc. 1992,
114, 980. (i) Jacobsen, E. N.; Zhang, W.; Gu¨ler, M. L. J. Am. Chem. Soc.
1991, 113, 6703. (j) Baldy, C. J.; Morrison, D. L.; Elliott, C. M. Langmuir
1991, 7, 2376. (k) Schrock, R. R.; DePue, R. T.; Feldman, J.; Schaverien,
C. J.; Dewan, J. C.; Liu, A. H. J. Am. Chem. Soc. 1988, 110, 1423.
(2) (a) Abbenhuis, H. C. L.; Burckhardt, U.; Gramlich, V.; Martelletti,
A.; Spencer, J.; Steiner, I.; Togni, A. Organometallics 1996, 15, 1614. (b)
Togni, A.; Burckhardt, U.; Gramlich, V.; Pregosin, P. S.; Salzmann, R. J.
Am. Chem. Soc. 1996, 118, 1031. (c) Lindner, E.; Fisahn, P.; Fawzi, R.;
Steimann, M. Chem. Ber. 1996, 129, 191. (d) van Rooy, A.; Kamer, P. C.
J.; van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J.; Veldman, N.; Spek,
A. L. Organometallics 1996, 15, 835. (e) Myers, A. W.; Jones, W. D.
Organometallics 1996, 15, 2905. (f) Kotha, S.; Tetrahedron 1994, 50, 3639.
(g) Burk, M. J.; Feaster, J. E.; Nugent, W. A.; Harlow, R. L. J. Am. Chem.
Soc. 1993, 115, 10125. (h) Noyori, R. Science 1990, 248, 1194. (i) Whitesell,
J. K. Chem. ReV. 1989, 89, 1581. (j) Chan, A. S. C.; Pluth, J. J.; Halpern,
J. Inorg. Chim. Acta 1979, 37, L477.
S0002-7863(97)02360-3 CCC: $14.00 © 1997 American Chemical Society

10744 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Slone et al.
centers.4 These ligands have a substitutionally inert group and
a substitutionally labile one. When bound to a transition metal
center, such as Rh(I), the bonding affinity of the substitutionally
labile moiety of the RHL can be controlled by adjusting the
oxidation state of a redox group that is covalently attached to
it, eq 1. Furthermore, it has been shown that the magnitude of
this effect can be tailored by altering the RHL phosphine
substituents,4a varying the weak bonding group (e.g. ether,4a,c-d
thioether,4a η6-arene,4b etc.), or controlling the distance between
the redox-active group and the transition metal center.4a
strategies which focus on changes in the electronic properties
of the metal center itself.7,8 For example, many have shown
that oxidation or reduction of a metal center can change the
bonding strength or hapticity of a bound ligand.7 Kochi and
co-workers, for instance, have shown that reduction of (η5-C5H4-
Me)Mn(CO)2(NCMe) labilizes a bound acetonitrile ligand
toward substitution reactions.7a In contrast to this metal-based
strategy, others have used substitutionally inert redox-active
ligands to affect the electronic properties of bound transition
metals without oxidation of the metal center itself.8 Wrighton
and co-workers, for example, have demonstrated that oxidation
of the cobaltocene group in [(1,1′-bis(diphenylphosphino)-
cobaltocene)Re(CO)4]PF6 increases the rate of nucleophilic
-
attack of N3 at a CO ligand by a factor of 54008a and that
cobaltocene oxidation affects the rates of catalytic reactions
mediated by a 1,1′-bis(diphenylphosphino)cobaltocene Rh(I)
complex.8b Since a substitutionally inert redox-active ligand
was used, however, this strategy only yielded electrochemical
modulation of transition metal electronic properties and not a
selective opening and closing of coordination sites at the metal
center as with the RHL approach.
Both metal-based and ligand-based strategies have been used
to alter the properties of metal complexes electrochemically.
For instance, systems which incorporate redox-active groups
into macrocyclic ligands, such as crown ethers, thiocrowns, and
cryptands, have been used to electrochemically control macro-
cycle bonding constants for metal ions.5,6 Although the majority
of such redox-active sequestering agents have been designed
to complex alkali and alkaline earth cations,5 a number of
macrocycles which target late transition metal ions also have
been investigated.6 These systems, however, are not desirable
for the reversible modulation of transition metal center reactivity
since oxidation of the redox-active ligand can result in expulsion
of the metal center of interest. Complementary to the modula-
tion of ligand bonding via ligand oxidation are metal-based
Past work in our group has illustrated the utility of RHL
complexes for the reversible electrochemical control over both
the electronic and steric properties of bound transition metal
centers.4 In this report, we demonstrate how changes in the
properties of a phosphinoalkylarene RHL complex, [(η1:η6-(η5-
C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))(η1-(η5-C5H5)Fe(η5-
C5H4C6H4OCH2CH2PPh2))Rh]+BF4- (2), upon ligand oxidation
result in substantial differences in complex reactivity. Specif-
ically, we address how the thermodynamic perturbation in
complex free energy that occurs upon RHL oxidation directly
translates into changes in complex stoichiometric and catalytic
reactivity. Therefore, by selectively labilizing weakly bound
η6-arene moieties in the Rh(I)-RHL complex 2, we have been
able to demonstrate for the first time that the small molecule
bonding properties and the catalytic activity of that complex
can be substantially affected simply upon ligand oxidation.
(4) (a) Allgeier, A. M.; Slone, C. S.; Mirkin, C. A.; Liable-Sands, L.
M.; Yap, G. P. A.; Rheingold, A. L. J. Am. Chem. Soc. 1997, 119, 550. (b)
Sassano, C. A.; Mirkin, C. A. J. Am. Chem. Soc. 1995, 117, 11379. (c)
Singewald, E. T.; Mirkin, C. A.; Stern, C. L. Angew. Chem., Int. Ed. Engl.
1995, 34, 1624. (d) Allgeier, A. M.; Singewald, E. T.; Mirkin, C. A.; Stern,
C. L. Organometallics 1994, 13, 2928.
(5) (a) Beer, P. D.; Chen, Z.; Grieve, A.; Haggitt, J. J. Chem. Soc., Chem.
Commun. 1994, 2413. (b) Beer, P. D. EndeaVour 1992, 16, 182. (c) Beer,
P. D.; Tite, E. L.; Ibbotson, A. J. Chem. Soc., Dalton Trans. 1991, 1691.
(d) Medina, J. C.; Goodnow, T. T.; Rojas, M. T.; Atwood, J. L.; Lynn, B.
C.; Kaifer, A. E.; Gokel, G. W. J. Am. Chem. Soc. 1992, 114, 10583. (e)
Chen, Z.; Gokel, G. W.; Echegoyen, L. J. Org. Chem. 1991, 56, 3369. (f)
Beer, P. D.; Keefe, A. D.; Sikanyika, H.; Blackburn, C.; McAleer, J. F. J.
Chem. Soc., Dalton Trans. 1990, 3289. (g) Andrews, M. P.; Blackburn, C.;
McAleer, J. F.; Patel, V. D. J. Chem. Soc., Chem. Commun. 1987, 1122.
(6) (a) Beer, P. D.; Nation, J. E.; Harman, M. E.; Hursthouse, M. B. J.
Organomet. Chem. 1992, 441, 465. (b) Tendero, M. J. L.; Benito, A.;
Mart´ınez-Ma´n˜ez, R.; Soto, J.; Paya`, J.; Edwards, A. J.; Raithby, P. R. J.
Chem. Soc., Dalton Trans. 1996, 343. (c) Sato, M.; Suzuki, K.; Asano, H.;
Sekino, M.; Kawata, Y.; Habata, Y.; Akabori, S. J. Organomet. Chem. 1994,
470, 263. (d) Sato, M.; Shigeta, H.; Sekino, M.; Akabori, S. J. Organomet.
Chem. 1993, 458, 199. (e) Sato, M.; Tanaka, S.; Akabori, S.; Habata, Y.
Bull. Chem. Soc. Jpn. 1986, 59, 1515. (f) De Santis, G.; Fabbrizzi, L.;
Licchelli, M.; Pallavicini, P.; Perotti, A. J. Chem. Soc., Dalton Trans. 1992,
3283.
Results
Synthesis, Characterization, and Solid-State Structure of
2. The synthesis of 1 and its complexation to Rh(I) to form
the 18-electron, bis(phosphine), η6-arene piano-stool complex
(2) was reported in a preliminary communication.4b RHL 1 was
synthesized by the reaction of ferrocenylphenol9 and ClCH2-
CH2Cl in distilled H2O at reflux for 15 h in the presence of
n
excess KOH and a catalytic amount of Bu4NI to form (η5-
C5H5)Fe(η5-C5H4C6H4OCH2CH2Cl) in 46% yield, eq 2. The
reaction of (η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2Cl) with KPPh2
in THF at 0 °C gave ligand 1 in 80% yield, eq 2. Ligand 1 and
(7) For transition metal oxidation-state-dependent ligand bonding strength
see the following: (a) Hershberger, J. W.; Kochi, J. K. J. Chem. Soc., Chem.
Commun. 1982, 212. (b) Kuchynka, D. J.; Amatore, C.; Kochi, J. K. Inorg.
Chem. 1986, 25, 4087. (c) Hecht, M.; Schultz, F. A.; Speiser, B. Inorg.
Chem. 1996, 35, 5555. (d) Yeung, L. K., Kim, J. E.; Chung, Y. K.; Rieger,
P. H.; Sweigart, D. A. Organometallics 1996, 15, 3891. For transition metal
oxidation-state-dependent ligand hapticity, see the following: (e) Connelly,
N. G.; Emslie, D. J. H.; Metz, B.; Orpen, A. G.; Quayle, M. J. J. Chem.
Soc., Chem. Commun. 1996, 2289. (f) Bowyer, W. J.; Geiger, W. E. J. Am.
Chem. Soc. 1985, 107, 5657. (g) Harman, W. D.; Sekine, M.; Taube, H. J.
Am. Chem. Soc. 1988, 110, 2439. (h) Kuchynka, D. J.; Kochi, J. K. Inorg.
Chem. 1988, 27, 2574.
(8) (a) Lorkovic, I. M.; Wrighton, M. S.; Davis, W. M. J. Am. Chem.
Soc. 1994, 116, 6220. (b) Lorkovic, I. M.; Duff, R. R., Jr.; Wrighton, M.
S. J. Am. Chem. Soc. 1995, 117, 3617. (c) Kotz, J. C.; Nivert, C. L. J.
Organomet. Chem. 1973, 52, 387. (d) Kotz, J. C.; Nivert, C. L.; Lieber, J.
M.; Reed, R. C. J. Organomet. Chem. 1975, 91, 87. (e) Elschenbroich, C.;
Stohler, F. Angew. Chem., Int. Ed. Engl. 1975, 14, 174.
its precursor were fully characterized by 1H NMR spectroscopy,
31P NMR spectroscopy, electron ionization (EI) mass spectrom-
etry, and elemental analysis.
(9) Weinmayr, V. J. Am. Chem. Soc. 1955, 77, 3012.

Redox-Switchable Hemilabile Ligand Synthesis
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10745
Table 1. Crystallographic Data for 2‚1.25CH2Cl2
formula
C61.25H56.50BCl2.5F4Fe2O2P2Rh
1276.55
triclinic
formula weight
crystal system
space group
a, Å
b, Å
c, Å
P1h
15.9857(2)
19.2497(2)
19.77310(10)
70.9728(4)
88.8718(7)
77.3017(1)
5602.53(10)
4
R, deg
â, deg
γ, deg
V, Å3
Z
crystal color, habit
D(calc), g cm-3
µ(Mo KR), cm-1
temp, K
orange, block
1.513
10.35
173(2)
size, mm3
diffractometer
radiation
R(F), %a
R(wF2) %a
0.30 × 0.30 × 0.20
Siemens P4/CCD
Mo KR (λ ) 0.710 73 Å)
4.10
Figure 1. ORTEP diagram of cation 2‚1.25CH2Cl2. Thermal ellipsoids
are drawn a 30% probability.
12.60
2
2
a Quantity minimized: R(wF2) ) ∑[w(Fo - Fc2)2/∑[w(Fo )2]1/2; R
The Rh(I) piano-stool complex 2 was synthesized by the
reaction of [RhCl(COT)2]210 (COT ) cyclooctene) with 2 equiv
of AgBF4 in CH2Cl2, followed by dilution and reaction of the
resulting solution with 4 equiv of 1 at -78 °C, eq 3. Pure
samples of 2 were fully characterized by 1H NMR spectroscopy,
31P NMR spectroscopy, high-resolution fast atom bombardment
(FAB+) mass spectrometry, and a single-crystal X-ray diffraction
study.
) ∑∆/∑∆(Fo), ∆ ) |(Fo - Fc)|.
Table 2. Selected Distances (Å) and Angles (deg) of
2‚1.25CH2Cl2
Rh(1)-P(1)
Rh(1)-P(2)
Rh(1)-C(45)
Rh(1)-C(46)
Rh(1)-C(47)
Rh(1)-C(48)
Rh(1)-C(49)
Rh(1)-C(50)
2.2827(8) Rh(2)-P(1′)
2.2457(8) Rh(2)-P(2′)
2.302(3) Rh(2)-C(45′)
2.345(3) Rh(2)-C(46′)
2.383(3) Rh(2)-C(47′)
2.358(3) Rh(2)-C(48′)
2.360(3) Rh(2)-C(49′)
2.366(3) Rh(2)-C(50′)
2.2771(9)
2.2393(9)
2.303(3)
2.339(3)
2.375(3)
2.371(3)
2.372(3)
2.357(3)
Rh(1)-arene centroid 1.877(3) Rh(2)-arene centroid 1.877(3)
P(1)-Rh(1)-P(2) 96.38(3) P(1′)-Rh(2)-P(2′) 96.07(3)
C(45)-C(46)-C(47) 120.0(3) C(45′)-C(46′)-C(47′) 119.7(3)
C(46)-C(47)-C(48) 119.9(3) C(46′)-C(47′)-C(48′) 120.2(3)
C(47)-C(48)-C(49) 119.2(3) C(47′)-C(48′)-C(49′) 119.6(3)
C(48)-C(49)-C(50) 121.2(3) C(48′)-C(49′)-C(50′) 120.2(3)
C(49)-C(50)-C(45) 118.9(3) C(49′)-C(50′)-C(45′) 118.9(3)
C(50)-C(45)-C(46) 120.8(3) C(50′)-C(45′)-C(46′) 120.8(3)
Single-crystals of 2‚1.25CH2Cl2 suitable for X-ray diffraction
were grown from the the slow diffusion of pentane into a CH2-
Cl2 solution of 2. Two crystallographically independent, but
chemically equivalent, ion pairs were located in the asymmetric
unit. An ORTEP diagram of one of the two cations is shown
in Figure 1, and crystallographic data are presented in Table 1.
Selected bond distances and angles are given for both crystal-
lographically independent cations in Table 2. The observed
Rh-C bond distances are in the expected range for monomeric
Rh(I) compounds with a bis(phosphine), η6-arene piano-stool
geometry (Rh-C ) 2.217-2.516 Å), whereas the monodentate
Rh-P bond distances are slightly longer than literature values
(Rh-P ) 2.217-2.251 Å).11 The structure of 2‚1.25CH2Cl2
also compares well with that of an analogous compound, [(η1:
η6-C6H5OCH2CH2PPh2)(η1-C6H5OCH2CH2PPh2)Rh]+BF4- (3),
previously reported by our group, Chart 1.12 In particular, both
independent cations are similar to 3 in that they possess a Rh-P
bond distance for the chelating phosphine ligand which is shorter
(2.2457(8) and 2.2393(8) Å) than the Rh-P bond distance for
the monodentate phosphine ligand (2.2827(9) and 2.2771(9) Å).
Chart 1
Similar to 3, the bound arene ligand in 2‚1.25CH2Cl2 shows
small deviations from planarity; the average deviation for the
two independent cations is 0.02 (Å). This is 4 times the
deviations from planarity for the phosphorus-bound phenyl rings.
Through extensive theoretical calculations done by other
researchers, it has been shown that arenes bound to transition
metal centers, such as Rh(I), adopt boat conformations as a result
of the interaction between a metal d-orbital and an η6-arene
π-orbital.13 Indeed, several complexes which are isoelectronic
with 2, such as 3, have been reported to contain arene ligands
with such deformations from planarity.11-13 However, unlike
the boat conformation observed for 3, only the bond between
the Rh center and the C atom connected to the chelate arm is
significantly shorter than the other Rh-C bonds in 2‚1.25CH2-
Cl2. Perhaps, the attachment of the ferrocenyl moiety to the
aryl group creates unfavorable steric interactions which prevent
the formation of a boat conformation as observed for 3. This
(10) Porri, L.; Lionetti, A.; Allegra, G.; Immirzi, A. Chem. Commun.
1965, 336.
(11) (a) Singewald, E. T. ; Slone, C. S.; Stern, C. L.; Mirkin, C. A.;
Yap, G. P. A.; Liable-Sands, L. M.; Rheingold, A. L. J. Am. Chem. Soc.
1997, 119, 3048. (b) Alvarez, M.; Lugan, N.; Donnadieu, B.; Mathieu, R.
Organometallics 1995, 14, 365. (c) Westcott, S. A.; Taylor, N. J.; Marder,
T. B.; Baker, R. T.; Jones, N. J.; Calabrese, J. C. J. Chem. Soc., Chem.
Commun. 1991, 304. (d) Townsend, J. M.; Blount, J. F. Inorg. Chem. 1981,
20, 269. (e) Albano, P.; Aresta, M.; Manassero, M. Inorg Chem. 1980, 19,
1069.
(12) Singewald, E. T.; Shi, X.; Mirkin, C. A.; Shofer, S. J.; Stern, C. L.
Organometallics 1996, 15, 3062.
(13) Muetterties, E. L.; Bleeke, J. R.; Wucherer, E. J.; Albright, T. A.
Chem. ReV. 1982, 82, 499.

10746 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Slone et al.
steric interaction also is reflected in the longer Rh-arene
centroid distances for 2‚1.25CH2Cl2 (1.877(3) Å for both
independent cations) as compared with the Rh-arene centroid
distance in 3 (1.840(3) Å). Also, the structure of the ferrocenyl
moieties in both independent cations of 2‚1.25CH2Cl2 are similar
to analogous aromatic substituted ferrocenes studied by other
workers in that the angles between the planes of the substituted
cyclopentadienyl and aryl rings are relatively small (angles
between the planes range from 13.3° to 19.3°).14
Compound 2 also has been spectroscopically characterized,
and all solution data are consistent with its solid-state structure.
The 31P NMR spectrum of 2 exhibits two resonances at δ 34.9
and δ 31.3 which are assigned to magnetically and chemically
inequivalent P atoms due to the coupling to each other and the
103Rh nucleus. Each resonance appears as a doublet of doublets
(JRh-P ) 204.4 and 209.7 Hz; JP-P ) 39.3 Hz) with coupling
constants that are in the expected region for Rh(I) complexes
with this type of geometry.11,12 Also, all resonances in the 1H
NMR spectrum of 2 have been assigned either by comparison
with the spectrum of the free ligand 1 or by a 1H COSY NMR
experiment.
Synthesis and Characterization of 2+ and 22+. To explore
the effects of oxidation of the redox-active ferrocenyl groups
in 2 on its reactivity, the oxidized forms (2+ and 22+) were
synthesized and isolated. The mixed-valence complex 2+ and
the doubly-oxidized complex 22+ were formed by the reaction
of 2 with either 1 or 2 equiv of AgBF4, respectively, in CH2Cl2
at room temperature, eqs 4 and 5. Complexes 2+ and 22+ were
isolated from the resulting Ag0 mirrors in extremely high yields
Figure 2. The 31P NMR spectra in CD2Cl2 of all oxidation states (a)
before the addition of oxidant (2), (b) after the addition 1 equiv of
AgBF4 (2+), and (c) after the addition of 2 equiv of AgBF4 (22+).
observation indicates that the mixed-valence form (2+) contains
a ferrocenium group in the monodentate, and not the chelated
phosphine ligand, respectively. Note, the cyclic voltammetry
of 2 in CH2Cl2 shows the ferrocenyl group on the unbound arene
ligand to be 143 mV easier to oxidize than the ferrocenyl group
on the arene bound to the cationic Rh(I) center (vida infra).
Upon the addition of a second equivalent of AgBF4, the 31P
NMR resonances for the chelated and the monodentate phos-
phine ligands shifted even more upfield to δ 33.4 and 29.0,
respectively. Therefore, the observed changes in the 31P NMR
spectra of these complexes easily identify ligand oxidation state
and Rh(I)-RHL complex purity.
1
and characterized by H NMR and 31P NMR spectroscopy.
Unlike the 31P NMR spectra of 2+ and 22+, the assignment
1
of the resonances in the corresponding H NMR spectra was
not straightforward. Many resonances were broad and expe-
rienced large shifts upon ligand oxidation and changes in
complex concentration. However, a comparison of the 1D
Although the oxidation reactions resulted in complexes with
paramagnetic ferrocenium centers, which are known to cause
large contact and dipolar shifts and line broadening,15 NMR
spectroscopy was still a useful characterization tool. The
resonances in the 31P NMR spectra, in particular, were very
well resolved since the 31P nuclei in 2+ and 22+ are far from
the ferrocenium centers and, therefore, have little interaction
with the unpaired electron spin density. The changes in the
31P NMR spectrum of 2 upon sequential oxidation were small,
but highly reproducible, Figure 2. Upon the addition of 1 equiv
of AgBF4 to 2, the 31P NMR resonances at δ 34.9 and 31.3
shifted upfield to δ 34.3 and 30.3, respectively. From a 31P-
{1H}-1H HETCOR experiment of compound 3, we conclude
that those resonances in the 31P NMR spectrum which are the
most upfield correspond to the P atom in the monodentate ligand
on Rh(I).12 Thus, upon oxidation of 2 to 2+, the chemical shift
of the 31P NMR resonance of the monodentate phosphine ligand
changed the most (∆δmono ) 1.0 and ∆δchel ) 0.6). This
1
spectra and an examination of the H COSY spectra of 2, 2+,
and 22+ at the same concentration and temperature (0.024 M in
CD2Cl2 at 20 °C) afforded assignment of all resonances in these
complexes. In particular, the methylene resonances R to the
phosphine groups (CH2PPh2) were characteristic of ligand
oxidation state. For instance, upon the formation of 2+, the
multiplet corresponding to the R protons of the monodentate
phosphine ligand shift from δ 2.06 to 1.79, while the position
of the quartet corresponding to the R protons in the chelated
ligand shift from δ 1.71 to 0.95. Furthermore, these methylene
resonances in 22+ are observed at δ 1.53 and 0.27 for the
monodentate and chelated phosphine ligands, respectively.
NMR Kinetic Studies. Previously, it was determined that
complex 2 undergoes a degenerate intramolecular, free arene,
η6-arene exchange reaction.4b In this study, the rate of this
unique exchange reaction was utilized as a measure of the
relative lability of the η6-arene ligand on the Rh(I) center in all
states-of-charge of 2 (Scheme 1). The rates of the exchange
process for the reduced (2), mixed-valence (2+), and doubly-
oxidized (22+) forms were determined by 1H NMR 2D exchange
spectroscopy (EXSY) experiments. In general, EXSY is an
extremely useful technique for detecting dynamic processes with
(14) (a) Beer, P. D.; Sikanyika, H.; Blackburn, C.; McAleer, J. F.; Drew,
M. G. B. J. Chem. Soc., Dalton Trans. 1990, 3295. (b) Loubser, C.; Imrie,
C.; van Rooyen, P. H. AdV. Mater. 1993, 5, 45. (c) Allen, F. H.; Trotter, J.;
Williston, C. S. J. Chem. Soc. A 1970, 907.
(15) (a) Sharp, R. R. In Nuclear Magnetic Resonance: Webb, G. A.,
Ed.; The Chemical Society: London, 1994; Vol. 24, p 469. (b) Materikova,
R. B.; Babin, V. N.; Solodovnikov, S. P.; Lyatifov, I. R.; Petrovsky, P. V.;
Fedin, E. I. Z. Naturforsch. 1980, 35b, 1415.
rates of 102-10-2 s-1 16
.

Redox-Switchable Hemilabile Ligand Synthesis
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10747
Table 3. Rates and Activation Parameters for the Arene-Arene
Exchange Reaction
∆Gq
∆Hq
∆Sq
(293)
a
complex
k (s-1
)
(kcal/mol)b
(kcal/mol)b (cal/mol K)b
2
0.36 ( 0.02 +17.50 ( 0.05
0.33 ( 0.03 +17.55 ( 0.05
0.51 ( 0.04 +17.29 ( 0.06
+22 ( 2
+20 ( 2
+13 ( 2
+15 ( 5
+9 ( 5
-15 ( 7
2+
22+
a Arene-arene exchange rate at 20 °C. b Activation parameters from
Eyring plots.
Figure 3. The CH2PPh2 region of the 1H NMR EXSY spectrum of 2
in CD2Cl2 at 30 °C with tm ) 0.5 s.
Scheme 1
Figure 4. Eyring plots for all oxidation states: (open circle) reduced
complex, 2; (crossed circle) mixed-valance complex, 2+; (filled circle)
double oxidized complex, 22+
.
the Eyring equation.18 The Eyring plots obtained for 2, 2+, and
22+ are shown in Figure 4, and the activation parameters are
given in Table 3. From these data, it can be seen that ∆Gq
(293)
for this process does not vary considerably; however, the
activation parameters, ∆Hq and ∆Sq, for 2, 2+, and 22+ exhibit
experimentally significant variations. For example, the ∆Hq
of exchange is less for the doubly oxidized complex (22+) than
for both the reduced (2) and the mixed-valence (2+) complexes;
these values for 2 and 2+ are the same within experimental error.
Moreover, the ∆Sq of exchange for 22+ is of a different sign
than the ∆Sq values for both 2 and 2+.
Quantification of the Thermodynamic Perturbation of 2
upon RHL Oxidation. The cyclic voltammogram of 2 in CH2-
Cl2/0.1 M nBuNPF6 exhibits three reversible waves (Figure 5a).
The first wave is assigned to the FeII/FeIII couple of the
ferrocenyl group on the unbound arene since the E1/2 value of
-30 mV vs FcH/FcH+ is identical, within experimental error,
to that of the free ligand (E1/2 ) -27 mV) under similar
conditions. The second wave at E1/2 ) +113 mV vs FcH/FcH+
corresponds to the oxidation of the ferrocenyl group attached
to the bound arene ligand and the wave at E1/2 ) +630 mV vs
FcH/FcH+ is assigned to a RhI/RhII couple based on previous
work done in our group.12 In fact, the latter oxidation is an
internal diagnostic marker that implies no unusual rearrangement
by the complex upon ferrocenyl oxidation. The piano-stool
geometry of 2 is one of the few geometries which stabilize
Rh(II) with these types of ligands.12
1
Since the H NMR CH2PPh2 resonances of the RHL are
diagnostic of ligand oxidation state, these resonances were
monitored in the 2D EXSY experiments. A 1H EXSY spectrum
of 2 is shown in Figure 3. In this plot, the one-dimensional
NMR spectrum corresponding to the resonances of the R-me-
thylene protons in the chelated and monodentate phosphine
ligands is plotted along the diagonal, and the observed cross
peaks are indicative of those protons undergoing chemical
exchange. The rates of the exchange reaction were determined
by obtaining the volumes of the diagonal and cross peaks for
each complex and relating them to the mixing time (see the
Experimental Section).16 The measured rates of the intramo-
lecular arene-arene exchange reaction (Scheme 1) at 20 °C are
given in Table 3. These exchange rates demonstrate an
increased lability of the η6-arene ligand in the doubly oxidized
complex 22+ (k ) 0.51 ( 0.04 s-1) as compared to the labilities
of the bound arenes in the reduced complex 2 and the mixed-
valence complex 2+ (k ) 0.36 ( 0.02 s-1 and k ) 0.33 ( 0.03
s-1, respectively).17
From the cyclic voltammogram of 2, it was possible to
quantitate the effects of ligand oxidation. For systems that
involve more than two oxidation states, ladder diagrams relating
reactants and products through chemical and electron transfer
processes can be used in conjunction with E1/2 values to obtain
important thermodynamic information. The ladder diagram
(17) For the arene-arene exchange reaction for 2+, a subsequent fast
electron transfer step most likely occurs after the rate determining step.
Since this fast step is thermodynamically favorable, the values for k, ∆Gq,
∆Hq, and ∆Sq are a measure for the step in which the ferrocenyl arene is
displaced from the Rh(I) center.
In addition, rate measurements over a range of temperatures
gave the activation parameters for the exchange process from
(18) Sandstro¨m, J. Dynamic NMR Spectroscopy; Academic Press:
London, 1982.
(16) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935.

10748 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Slone et al.
also is assumed to be negligible (e.g. the electronic communica-
tion between the two ferrocenyl groups in 1,2-diferrocenyltet-
ramethylethane can not be measured by electrochemical meth-
ods; the ethoxy-Rh-arene spacer in 2 is much longer than a
tetramethylethyl spacer).20
The two ferrocenyl-based waves in the cyclic voltammogram
in Figure 5a were better resolved using differential pulse
voltammetry. The measured ∆E1/2 of 143 mV corresponded
well to the cyclic voltammetry data, and was used to calculate
1
a Kox of 3.8 × 10-3 and a ∆G of 3.3 kcal/mol at 20 °C,
Schemes 1 and 2. The change in free energy (∆G) is the
thermodynamic perturbation of the complex that results from
RHL ligand oxidation and could be termed the thermodynamic
“RHL effect”. Therefore, by simply changing the oxidation state
of 2, we have been able to destabilize the metal complex by
3.3 kcal/mol. Presumably, the majority of this effect is
associated with a decrease in Rh(I)-arene bond strength upon
RHL oxidation. It is also important to note that the change in
free energy of the metal complex upon replacing the ferro-
cenylarene ligand with an arene moiety containing a ferrocenium
group can only be measured from a single cyclic voltammogram
when intramolecular degenerate exchange reactions are possible.
In contrast, the absolute thermodynamic perturbation that results
from ligand oxidation cannot be obtained if only nondegenerate
processes are possible. In the latter case, cyclic voltammetry
merely yields a ratio of equilibrium constants for the reactions
involving the oxidized and reduced species.
Figure 5. Cyclic voltammetry of 2 measured at 20 mV/s at a glassy
n
carbon electrode (8.0 × 10-5 cm2): (a) in CH2Cl2/0.1 M Bu4NPF6
n
and (b) in 1% CH3CN in CH2Cl2/0.1 M Bu4NPF6.
Scheme 2
Ligand Oxidation-State-Dependent Acetonitrile Bonding.
In the reduced state, complex 2 reacts with acetonitrile at room
temperature to form a mixture of the square-planar, cis- and
trans-bis(acetonitrile) adducts, 4 and 5, in a ratio of 3:1, eq 6.
The 31P NMR spectrum of 4 and 5 exhibits resonances at δ
35.9 and 22.2, respectively. Each resonance appears as a
doublet with JRh-P values of 174.8 Hz for the cis adduct and
131.7 Hz for the trans adduct. The reactivity of 2 toward
acetonitrile was similar to the reactivity of the analogous
complex 3.12 Interestingly, when 2 was reacted with acetonitrile
shown in Scheme 2 relates arene-arene exchange equilibria to
the electrochemical reactions involving ligand oxidation. Using
the Nernst equation, a relationship between the arene-arene
exchange equilibria of 2 and one of its oxidized forms (2+) was
derived: ∆E1/2 ) -(RT/nF ) ln(Kox1/Kred).19 Significantly, since
Kred is the equilibrium constant for a degenerate reaction, and
at low temperature (-78 °C), the cis adduct 4 was formed
exclusively. As the temperature was then raised to 25 °C, the
cis adduct slowly converted to the trans adduct 5 until a 3:1
cis to trans ratio was obtained. At 25 °C, the observed cis to
trans ratio of 3:1 corresponds to a ∆G(298) of -0.65 kcal/mol
for the reaction going from 5 to 4. Furthermore, upon removal
of acetonitrile (48 h exposure to vacuum), both 4 and 5 were
completely converted back to complex 2.
Another way to quantify the “RHL effect” is to compare the
equilibria between 2 and its acetonitrile adducts 4 and 5 before
and after oxidation of the ligand chelated to the metal center. If
the Rh(I)-arene bond is indeed weakened upon oxidation of a
pendant redox-active ferrocenyl group, then the Rh(I) center’s
1
therefore equal to 1, it is possible to directly calculate Kox if
the E1/2 values for the electrochemical reactions going from 2
to 2+ and 2′ to 2+′ are measurable. Even though the E1/2 for
the latter process is not experimentally obtainable, it can be
assumed to be the same as the E1/2 for the process going from
2+ to 22+. This is a reasonable assumption since the electronic
communication between the ferrocenyl group attached to the
free arene and the formally charged Rh(I) center is negligible.
Indeed, the E1/2 of ferrocenyl group attached to the free arene
in complex 2 is virtually identical to that of the uncomplexed
ligand 1. Furthermore, because of the distance between the two
ferrocenyl groups in 2, electronic communication between them
(19) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; John Wiley
& Sons: New York, 1980.
(20) Morrison, W. H., Jr.; Krogsrud, S.; Hendrickson, D. N. Inorg. Chem.
1973, 12, 1998.

Redox-Switchable Hemilabile Ligand Synthesis
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10749
Table 4. Catalytic Data for 2, 2+, and 22+ at 20 °C
Scheme 3
catalyst
substrate:catalyst ratioa
turnover frequencyb
2
10:1
10:1
100:1
100:1
100:1
0.20 ( 0.07
0.66 ( 0.15
0.41 ( 0.06
0.39 ( 0.03
0.90 ( 0.16
22+
2
2+
22+
a [Substrate] ) 1.76 M in CD2Cl2 and [catalyst] ) 0.176 and 0.0176
M for subtrate to catalyst ratios of 10:1 and 100:1, respectively. b Values
are an average of two experiments.
bonding affinity for small molecules such as CH3CN would be
expected to increase upon oxidation. To determine if this was
the case, the electrochemical behavior of 2 in the presence of
CH3CN was studied, Figure 5b. The cyclic voltammetry of 2
in CH2Cl2/0.1 M nBuNPF6 is markedly different from 2 in 1%
CH3CN in CH2Cl2/0.1 M nBuNPF6. Upon the addition of CH3-
CN to the electrochemical cell, the three waves that were
Figure 6. Isomerization studies using a 100:1 substrate to catalyst ratio
([Rh] ) 0.017 M and [ethyl allyl ether] ) 1.7 M in CD2Cl2 at 20 °C)
of each oxidation state: (open circle) reduced complex (2); (crossed
circle) mixed-valance complex (2+); and (filled circle) doubly oxidized
complex (22+).
n
observed for 2 in CH2Cl2/0.1 M BuNPF6 disappear. Instead,
a single wave with an E1/2 value of -20 mV vs FcH/FcH+ is
observed. This oxidation potential is consistent with the
displacement of the η6-arene ligand in 2 by CH3CN, as was
observed via NMR spectroscopy, and the simultaneous oxidation
of both free ferrocenyl groups. In addition, the well-defined,
reversible RhI/RhII couple at +630 mV vs FcH/FcH+ was
replaced by a broad, irreversible wave with an Epa value of +992
mV vs FcH/FcH+. This electrochemical irreversibility is
common for 16-electron, square-planar Rh(I) complexes with
coordination environments similar to those found in 4 and 5.4a
The difference in E1/2 values associated with the oxidation
of the ligand chelated to the Rh(I) center and the oxidation of
the same ligand after CH3CN displacement, ∆E1/2, can be used
to estimate the ratio of equilibrium constants for the displace-
ment of a ferrocenium, η6-arene ligand compared to the
displacement of the more electron rich ferrocenyl, η6-arene
ligand by CH3CN (Kox2/Kox1), Scheme 3. Using the Nernst
equation, which relates the equilibria involving 2+, 4+, and 5+
to those involving 22+, 42+, and 52+, a relationship between
complexes isomerize such allyl substrates.21a Herein, catalyst
concentrations were kept constant and substrate-to-Rh catalyst
ratios of either 10:1 or 100:1 were used. In order to quantify
the extent to which oxidation of the ligand bound to the Rh(I)
center affected its catalytic activity, the isomerization activties
for 2, 2+, and 22+ were investigated. For the runs using the
mixed-valence (2+) and doubly oxidized complexes (22+), the
catalyst was generated prior to each experiment and 1H and 31
P
NMR spectroscopy were utilized to confirm complete and clean
oxidation reactions. The concentrations of the products (cis-
1
and trans-ethyl 1-propenyl ether) were followed by H NMR
1
2
∆E1/2 and Kox2/Kox was derived; ∆E1/2 ) -(RT/nF ) ln(Kox
/
spectroscopy, and the transition metal catalyst was monitored
by 31P NMR spectroscopy. The initial isomerization rates were
obtained from linear fits of the initial slopes of plots of the
concentration of the products (cis and trans combined, in mol/
L) versus time (min) and normalized to the moles of catalyst
used to yield initial turnover frequencies.
The data for each set of catalytic conditions are given in Table
4, and representative examples of the catalytic data for 2, 2+,
and 22+ are shown in Figure 6. Significantly, it was observed
that the rate of isomerization was consistently faster for the
doubly oxidized form (22+) than the reduced form (2), under
identical conditions. Over numerous experiments, it was found
that this difference, on average, was 2.2 and 3.3 times faster
for the 100:1 and the 10:1 substrate-to-catalyst ratios, respec-
tively. Moreover, the mixed-valence complex (2+) exhibited
an isomerization activity which was identical to the reduced
complex (2), within experimental error.22 Therefore, only upon
oxidation of the ferrocenyl group on the η6-arene ligand was a
substantial difference in the isomerization activity observed.
Although the plots given in Figure 6 are not linear, the data is
Kox1).19 For the reactions depicted in the square-wave diagram
shown in Scheme 3, the measured ∆E1/2 value of 133 mV
corresponds to a Kox2/Kox ratio of 1.8 × 102. Since the E1/2
1
values for 2 and its CH3CN adducts were measured in CH2-
Cl2/0.1 M nBuNPF6 and 1% CH3CN in CH2Cl2/0.1 M nBuNPF6,
respectively, any solvent effects on half-wave potentials were
assumed to be negligable.
Ligand Oxidation-State-Dependent Isomerization Studies.
Another way to assess the change in the properties of the
transition metal center upon ligand oxidation is to monitor the
catalytic activity of complex 2 as a function of the ligand’s state-
of-charge. To demonstrate that RHL oxidation can significantly
affect the reactivity of an RHL complex, we chose a well-
understood catalytic transformation for Rh(I) transition metal
centers, the isomerization of allyl substrates.21 The reaction in
eq 7 was chosen because it is known that bis(phosphine)Rh(I)
(21) (a) Bergens, S. H.; Bosnich, B. J. Am. Chem. Soc. 1991, 113, 958.
(b) Inoue, S.; Takaya, H.; Tani, K.; Otsuka, S.; Sato, T.; Noyori, R. J. Am.
Chem. Soc. 1990, 112, 4897.

10750 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Slone et al.
very reproducible. It appears that the rate acceleration for the
doubly oxidized species occurs at the beginning of the catalytic
reaction and then levels off from there on. Perhaps, the product
olefin inhibits the reaction. Furthermore, as the reaction
proceeded, the activity of all catalysts decreased. This decrease
in activity was coincident with the appearance of a doublet in
Scheme 4
the 31P NMR spectra during the reaction: δ 41.3 (d, JRh-P
)
181.1 Hz). Presumably, this new compound inhibits the
isomerization reaction. However, since only the initial rates
were studied, where there is no evidence of this doublet or any
other additional species, we assume that the differences in the
initial rates were not affected by species which appeared at much
later reaction times.
benzene. Indeed, there are several examples in which the
relative stabilization of an η4-intermediate is utilized, in part,
to explain the increased lability of aromatic ligands with strongly
electron-withdrawing or electron-donating substituents when
compared to unsubstituted analogs.24a,f,h,26b,i In the case of 22+
,
Discussion
the strongly electron-withdrawing ferrocenium group would be
expected to facilitate the formation of an η4-intermediate, such
as complex 6 in Scheme 4. This initial ring slippage would be
expected to affect the rate of the arene-arene exchange reaction.
In related work done in our group, a cis-phosphine, cis-ether
intermediate, such as 7 in Scheme 4, was experimentally
determined to be important in the arene-arene exchange
process.12
Compound 2 is an ideal reaction center for studying the
effects of RHL oxidation state on complex stoichiometric and
catalytic reactivity. In particular, the rates of intramolecular
arene-arene exchange can be used as a measure of the relative
lability of the weakly coordinating aryl groups in 2, 2+, and
2
2+. Significantly, it was observed that the exchange rate for 2
substantially increased upon oxidation to 22+, Table 3. Such
arene labilization may be termed the kinetic “RHL effect”, and
this observed increased arene lability can be understood in two
ways. First, it is well-known that decreasing the amount of
electron density on a π-ligand by decreasing the number of
electron-donating groups decreases the metal-ligand bond
strength and, as a result, often increases a π-ligand’s lability.11a,23
Thus, the effect demonstrated herein is similar to previously
observed substituent effects; however, with 2 one can electro-
chemically interconvert an electron-donating ferrocenyl group
to an electron-withdrawing ferrocenium group without the use
of conventional synthetic steps. Second, it has been shown that
arene ligands substituted with strongly electron-donating24a-c,h,i
or electron-withdrawing substituents24a-g can facilitate arene ring
slippage through the stabilization of an η4-intermediate relative
to the unsubstituted case. Therefore, since arene-arene ex-
change reactions commonly involve arene haptotropic shifts,25,26
those arenes which ease initial ring slippage to an η4-arene
would be expected to display increased lability as compared to
Interestingly, the arene-arene exchange rates for 2 and 2+
do not differ within experimental error, Table 3 and Figure 4.
Since Kox1 was determined to be 3.8 × 10-3, the vast majority
of complex 2+ contains a ferrocenylarene bound to the Rh(I)
center. Therefore, the similarity in exchange rates for 2 and
2+ indicate that the displacement of the bound aryl group is
important in the intramolecular exchange reaction for all RHL
oxidation states; a subsequent fast electron transfer most likely
occurs for 2+ after the rate-determining step.17 Furthermore,
the exchange pathways for 2 and 2+ likely involve similar
mechanisms with a large extent of Rh(I)-arene bond breakage
in the transition state as evidenced by their similar large, positive
∆Hq and ∆Sq values, Table 3. Such interpretations are
reasonable for coordinatively saturated 18-electron complexes,
many of which are known to undergo ligand exchange reactions
that are dissociative in nature.25 Moreover, large positive ∆Hq
and ∆Sq values are suggestive of dissociative pathways, whereas
smaller positive ∆Hq and negative ∆Sq values reflect associative
mechanisms.25a
(22) Only a 100:1 substrate-to-catalyst ratio was used for 2+. At a 10:1
substrate-to-catalyst ratio, the resonances of the 1H NMR spectra of 2+
interfere with the accurate integration of both the substrate and product
resonances.
(23) (a) Bowyer, W. J.; Merkert, J. W.; Geiger, W. E.; Rheingold, A. L.
Organometallics 1989, 8, 191. (b) Koelle, U.; Fuss, B.; Rajasekharan, M.
V.; Ramakrishna, B. L.; Ammeter, J. H.; Bo¨hm, M. C. J. Am. Chem. Soc.
1984, 106, 4152. (c) Yur’eva, L. P.; Peregudova, S. M.; Nekrasov, L. N.;
Korotkov, A. P.; Zaitseva, N. N., Zakurin, N. V.; Vasil’kov, A. Y. J.
Organomet. Chem. 1981, 219, 43. (d) Hamon, J.-R.; Astruc, D.; Michaud,
P. J. Am. Chem. Soc. 1981, 103, 758.
(24) (a) Watts, W. E. in ComprehensiVe Organometallic Chemistry:
Wilkinson, G., Ed.; Pergamon: New York, 1982; Vol. 8, p 1050. (b) Hunter,
A. D.; Mozol, V.; Tsai, S. D. Organometallics 1992, 11, 2251. (c) Hunter,
A. D.; Shilliday, L.; Furey, W. S.; Zaworotko, M. J. Organometallics 1992,
11, 1550. (d) Clack, D. W.; Kane-Maguire, L. A. P. J. Organomet. Chem.
1978, 145, 201. (e) Traylor, T. G.; Hanstein, W.; Berwin, H. J.; Clinton,
N. A.; Brown, R. S. J. Am. Chem. Soc. 1971, 93, 5715. (f) Trahanovsky,
W. S.; Wells, D. K. J. Am. Chem. Soc. 1969, 91, 5870. (g) Wells, D. K.;
Trahanovsky, W. S. J. Am. Chem. Soc. 1969, 91, 5872. (h) Le Maux, P.;
Saillard, J. Y.; Granjean, D.; Jaouen, G. J. Org. Chem. 1980, 45, 4524. (i)
Pidcock, A.; Smith, J. D.; Taylor, B. W. J. Chem. Soc. A 1969, 1604.
(25) (a) Howell, J. A. S.; Burkinshaw, P. M. Chem. ReV. 1983, 83, 557.
(b) Ku¨ndig, E. P.; Desobry, V.; Grivet, C.; Rudolph, B.; Spichiger, S.
Organometallics 1987, 6, 1173. (c) Howell, J. A. S.; Dixon, D. T.; Kola,
J. C.; Ashford, N. F. J. Organomet. Chem. 1985, 294, C1. (d) Muetterties,
E. L.; Bleeke, J. R.; Sievert, A. C. J. Organomet. Chem. 1979, 178, 197.
(26) (a) Traylor, T. G.; Goldberg, M. J. J. Am. Chem. Soc. 1987, 109,
3968. (b) Traylor, T. G.; Stewart, K. J.; Goldberg, M. J. J. Am. Chem. Soc.
1984, 106, 4445. (c) Pidcock, A.; Smith, J. D.; Taylor, B. W. J. Chem.
Soc. A 1967, 872. (d) Zingales, F.; Chiesa, A.; Basolo, F. J. Am. Chem.
Soc. 1966, 88, 2707.
The ∆Hq value for 22+ is small as compared with those for
2 and 2+. This observation is consistent with a decrease in the
Rh(I)-arene bond strength accompanying RHL oxidation or
alternatively, a significant amount of bond formation ac-
companying bond breakage in the rate determining step. Also,
note the substantial differences in the ∆Sq values for the doubly
oxidized complex (22+) as compared to those measured for 2
and 2+, Table 4. Considering the proposed steps for the arene-
arene exchange given in Scheme 4, the differences in the
mechanisms for 2 (or 2+) and 22+, implied by the two very
different ∆Sq values, can be easily understood. For both 2 and
2+, the haptotropic shift to yield the η4-arene ligand (6) is not
favored, and thus, it is the η6 to η4 haptotropic shift which is
the slow step in the arene-arene exchange reaction. This
breakage of the Rh(I)-arene bond in the transition state is
consistent with the large positive ∆Sq values observed for 2
and 2+. In contrast, since 22+ favors arene slippage more than
2 and 2+, the coordination of the internal ether of the free
ferrocenyl-aryl ether is the important step in the reaction. The
negative ∆Sq and the small positive ∆Hq for the arene-arene
exchange reaction involving 22+ are consistent with Rh(I)-ether
bond formation accompanying arene ring slippage in the
transition state. Finally, if the differences in the activation
parameters for 2, 2+, and 22+ were merely due to different

Redox-Switchable Hemilabile Ligand Synthesis
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10751
solvation of the charged transition states, then the measured ∆Sq
for the mixed-valence form (2+) should differ significantly from
that for 2, but it does not.
complex charge or solvation. In that case, one would expect
significant differences in the activities for 2 and 2+.
Conclusions
The first reactivity consequence of the increased lability of
the Rh(I)-arene bond upon oxidation of the chelated ligand
was observed by studying complex 2 and its charge-dependent
interactions with CH3CN. From a comparison of the cyclic
voltammetry data of 2 in the absence and presence of 1% CH3-
CN in CH2Cl2, a 100-fold increase in its CH3CN bonding
affinity was observed upon moving from its totally reduced to
totally oxidized state (22+). These oxidation-state-dependent
equilibria can be viewed as a measure of the thermodynamic
“RHL effect”. Note that this reactivity pattern is consistent with
the 3.3 kcal/mol destabilization of 2 that accompanies double
Redox-switchable hemilabile ligands (RHLs), such as 1, offer
tremendous control over both the steric and electronic environ-
ments of bound transition metal centers and, thus, metal complex
reactivity. The Rh(I)-RHL complex (2) has provided a novel
system for the facile investigation of both the kinetic and
thermodynamic effects associated with complex oxidation. With
2, we have shown that ligand oxidation not only accelerates an
intramolecular exchange reaction, but also changes the mech-
anism through which such a reaction occurs. Furthermore, we
have demonstrated that RHLs can be utilized to modulate
transition metal complex reactivity via the selective and
electrochemically induced opening and closing of coordination
sites at a reactive transition metal center. When optimized, such
systems may be useful for catalyst storage and activation and,
therefore, the easy use and recovery of expensive homogeneous
transition metal catalysts. Presently, we are improving the RHL
design in an effort to understand fundamental factors that will
yield the maximum oxidation-state-dependent effects on com-
plex reactivity.
oxidation to form 22+
.
Numerous η6-arene transition metal complexes are catalyti-
cally active.27,28 In some cases, the arene ligand stays bound
to the metal center throughout the reaction cycle.27b,c Of more
interest to us, however, are cases in which the arene ligand is
substitutionally labile on the metal center and acts to stabilize
the complex in the absence of substrate or coordinating solvent.28
For several such complexes, it was shown that initial arene
displacement28a-c or ring slippage28d-f was essential for the
occurrence of reactions at the bound transition metal center.
Furthermore, it has been observed that too strong of a metal-
arene bond can result in complete deactivation of a reactive
metal catalyst.29 Since the EXSY data given in Table 3 are
consistent with arene labilization in 22+, the lability of the η6-
aryl group in 2, 2+, and 22+ was expected to have an effect on
complex catalytic activity. To test this, a catalytic reaction was
chosen where the initial displacement of the arene moiety and
an opening of coordination sites at the metal center were
important steps in the reaction cycle.
Experimental Section
General. All reactions were carried out under nitrogen using
standard Schlenk techniques or in an inert atmosphere glovebox unless
noted otherwise. All solvents were treated as previously described.4a,12
10
Ferrocenylphenol9 and [RhCl(COT)2]2 were prepared according to
literature procedures. RhCl3‚xH2O was used on loan from Johnson-
Matthey Chemical Co. Ethyl allyl ether was purchased from Aldrich
Chemical Co. and distilled over K2CO3 and degassed prior to use. All
other chemicals were purchased from Aldrich Chemical Co. and used
as received.
The isomerization activity studies of 2, 2+, and 22+ performed
herein demonstrate for the first time that the catalytic activity
of an RHL complex can be significantly altered upon ligand
oxidation. Note that the doubly oxidized complex (22+) exhibits
a significant increase in its catalytic activity upon ligand
oxidation when compared to the reduced (2) or the mixed-
valence (2+) forms, Table 4. This observation is consistent with
the proposed decrease in the Rh(I)-arene-bonding interaction
that accompanies oxidation of 2+ to 22+. Furthermore, this
observation suggests that η6-arene displacement or ring slippage
in the starting 18-electron complex is an important step in the
catalytic cycle. In addition, the observation that the isomer-
ization activity of 2+ does not differ from that of 2 is extremely
important. This confirms that the lability of the bound η6-arene
moiety in 2, 2+, and 22+ is the major factor in this system that
controls isomerization rates. It is unlikely that these oxidation-
state-dependent activities are merely due to differences in
Physical Measurements. One-dimensional 1H NMR spectra were
recorded on either a Varian Gemini 300 MHz or a Varian Unity 400
MHz FT-NMR spectrometer. 31P{1H} NMR spectra were recorded
on either a Varian Gemini 300 MHz or a Varian Unity 400 MHz FT-
NMR spectrometer at 121 and 162 Hz, respectively, and referenced
versus the external standard of 85% H3PO4. 2D 1H NMR COSY
experiments were performed on a Varian Unity 400 MHz FT-NMR
spectrometer. 1H NMR exchange spectroscopy (EXSY) spectra were
recorded on a Varian Unity 400 MHz FT-NMR spectrometer using a
standard NOESY pulse sequence with tm varying with temperature.
Electrochemical measurements were carried out on a PINE AFRDE4
or AFRDE5 bipotentiostat using a glassy carbon working electrode (8.0
× 10-5 cm2), a Pt mesh counter electrode, and a Ag wire reference
n
electrode. In all cases a 0.1 M solution of Bu4NPF6 was used as the
supporting electrolyte. All electrochemical data are referenced versus
the FcH/FcH+ [Fc ) (η5-C5H5)Fe(η5-C5H4)] redox couple. Electron
ionization (EI) and fast atom bombardment (FAB) mass spectra were
recorded on a Fisions VG 70-250 SE mass spectrometer. Elemental
analyses were done by either Desert Analytics, Tucson, AZ, or
Schwarzkopf Microanalytical Laboratory Inc., Woodside, NY.
(27) (a) Muetterties, E. L.; Bleeke, J. R. Acc. Chem. Res. 1979, 12, 324.
(b) Amaratunga, S.; Alper, H. J. Organomet. Chem. 1995, 488, 25. (c) Amer,
I.; Alper, H. J. Am. Chem. Soc. 1990, 112, 3674. (d) Davies, S. G.; Green,
M. L. H.; Mingos, D. M. P. Tetrahedron 1978, 34, 3047. (e) Tsonis, C. P.;
Farona, M. F. J. Organomet. Chem. 1976, 114, 293. (f) White, J. F.; Farona,
M. F. J. Organomet. Chem. 1973, 63, 329.
(28) (a) Bo¨nnemann, H.; Goddard, R.; Grub, J.; Mynott, R.; Raabe, E.;
Wendel, S. Organometallics 1989, 8, 1941. (b) Fairlie, D. P.; Bosnich, B.
Organometallics 1988, 7, 936. (c) Klabunde, K. J.; Anderson, B. B.; Bader,
M.; Radonovich, L. J. J. Am. Chem. Soc. 1978, 100, 1313. (d) Choe, S.-B.;
Kanai, H.; Klabunde, K. J. J. Am. Chem. Soc. 1989, 111, 2875. (e) Johnson,
J. W.; Muetterties, E. L. J. Am. Chem. Soc. 1977, 99, 7395. (f) Shmidt, F.
K.; Levkovskii, Y. S.; Saraev, V. V.; Ryutina, N. M.; Kosinskii, O. L.;
Bukunina, T. I. React. Kinet. Catal. Lett. 1977, 7, 445.
Synthesis of (η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2Cl). Ferrocen-
ylphenol9 (2.92 g, 0.011 mol), KOH (1.20 g, 0.023 mol), dichloroethane
(41.4 mL, 0.525 mol), and a catalytic amount of the phase transfer
catalyst n-Bu4NI were refluxed with 150 mL of distilled water in air.
After 15 h, 150 mL of distilled water and 150 mL of CH2Cl2 were
added to the crude reaction mixture. The organic phase was separated,
washed once with distilled water, and dried over MgSO4. Column
chromatography of the crude product on silica gel was performed.
Starting material was eluted with 100% pentane and a second band
was eluted by increasing the percentage of CH2Cl2 until a 100% CH2-
Cl2 eluent was obtained. This band was concentrated by vacuum
evaporation of the solvent to yield an analytically pure orange
microcrystalline sample of (η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2Cl),
(yield ) 46%, 1.73 g, 0.051 mol). 1H NMR (CDCl3): δ 3.81 (t, JH-H
) 5.93 Hz, 2H, CH2Cl), 4.01 (s, 5H, η5-C5H5), 4.22 (t, JH-H ) 5.94
(29) (a) Nickel, T.; Goddard, R.; Kru¨ger, C.; Po¨rschke, K.-R. Angew.
Chem., Int. Ed. Engl. 1994, 33, 879. (b) Fish, R. H.; Kim, H.-S.; Babin, J.
E.; Adams, R. D. Organometallics 1988, 7, 2250. (c) Crabtree, R. H.;
Mellea, M. F.; Quirk, J. M. J. Am. Chem. Soc. 1984, 106, 2913. (d)
Lucherini, A.; Porri, L. J. Organomet. Chem. 1978, 155, C45.

10752 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Slone et al.
Hz, 2H, OCH2), 4.25 (t, JH-H ) 1.80 Hz, 2H, η5-C5H4), 4.55 (t, JH-H
) 1.71 Hz, 2H, η5-C5H4), 6.84 (m, 2H, C6H4), 7.39 (m, 2H, C6H4).
MS (EI): M+ ) 340 m/z. Anal. (C18H17ClFeO) C, calcd 63.47, found
63.47; H, calcd 5.03, found 4.83.
7.13 (m, 20H, PPh2), 14.85 (b, ferrocenium), 17.18 (b, ferrocenium).
31P NMR (CH2Cl2): δ 33.4 (dd, JRh-P ) 203.0 Hz, JP-P ) 39.4 Hz),
29.0 (dd, JRh-P ) 210.3 Hz, JP-P ) 39.4 Hz).
Reaction of 2 with Acetonitrile. In an NMR tube, complex 2 was
reacted with neat CD3CN to form a mixture of cis- and trans-bis-
(acetonitrile) adducts, [cis-(η1-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2-
PPh2))2(CH3CN)2Rh]+BF4- (4) and [cis-(η1-(η5-C5H5)Fe(η5-C5H4C6H4-
Synthesis of (η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2) (1).
A
THF solution of KPPh2 (0.50 M, 1.11 mmol) was added slowly to a
solution of (η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2Cl) (0.361 g, 1.05
mmol) in 70 mL of THF at 0 °C. The reaction mixture was allowed
to warm to room temperature over 12 h, and the solution was
concentrated to dryness in vacuo. A CH2Cl2 solution of the crude
reaction mixture was added to a flask equipped with 20 g of silica gel,
leaving the KCl behind, and the solvent was removed in vacuo. Column
chromatography of the crude product was performed in a glovebox
under nitrogen. Three bands were eluted with a 1:1 pentane to CH2-
Cl2 solution. The first band contained mostly HPPh2 and was discarded.
The second band contained mostly starting material and was saved.
The third and major band contained 1. Vacuum evaporation of the
solvent afforded the desired ligand 1 as a yellow microcrystalline solid
(yield ) 80%, 414 mg, 0.85 mmol). 1H NMR (CDCl3): δ 2.62 (t,
-
OCH2CH2PPh2))2(CH3CN)2Rh]+BF4 (5), in a 3:1 ratio, respectively.
Upon removal of the CD3CN by exposure to vacuum for 48 h, only
the arene-coordinated product (2) was observed. In CD2Cl2 with 2
equiv of CH3CN a similar mixture of 4 and 5 is formed. If this reaction
is performed at -78 °C, only the cis adduct (4) forms. As the
temperature is raised to 25 °C, the cis adduct slowly converts to the
trans adduct until a ratio of 3:1 is reached. Data for 4. 1H NMR (CH2-
Cl2, -78 °C): δ 1.62 (s, 6H, CH3CN), 2.07 (m, 4H, CH2PPh2), 3.98
(s, 10H, η5-C5H5), 4.43 (m, 4H, CH2O), 4.32 (m, 4H, η5-C5H4), 4.60
(m, 4H, η5-C5H4), 6.69 (d, JH-H ) 6.3 Hz, 4H, C6H4), 7.23-7.38 (m,
24H, C6H4 and PPh2). 31P NMR (CH2Cl2): δ 35.9 (d, JRh-P ) 174.8
Hz). Data for 5. 1H NMR (CH2Cl2, 25 °C): δ 1.94 (s, 6H, CH3CN),
2.23 (m, 4H, CH2PPh2), 4.01 (s, 10H, η5-C5H5), 4.27 (m, 4H, η5-C5H4),
4.41 (m, 4H, CH2O), 4.57 (m, 4H, η5-C5H4), 6.64 (d, JH-H ) 8.0 Hz,
4H, C6H4), 7.29-7.55 (m, 24H, C6H4 and PPh2). 31P NMR (CH2Cl2):
δ 22.2 (d, JRh-P ) 131.7 Hz).
JH-H ) 7.55 Hz, 2H, CH2PPh2), 4.00 (s, 5H η5-C5H5), 4.09 (q, JH-H
)
7.62 Hz, 2H, CH2O), 4.23 (t, JH-H ) 1.83 Hz, 2H, η5-C5H4), 4.53 (t,
JH-H ) 1.86 Hz, 2H, η5-C5H4), 6.71 (m, 2H, C6H4), 7.34 (m, 4H, C6H4
and PPh2), 7.48 (m, 8H, PPh2). 31P NMR (CDCl3): δ -22.2 (s);
FAB-HRMS: M+ calcd 490.1149, found 490.1123 m/z. Anal.
(C30H27FeOP): C, calcd 73.48, found 73.16; H, calcd 5.55, found 5.18.
NMR Kinetic Studies. 2D exchange spectroscopy (EXSY) was
used to determine the rates and activation parameters of the arene-
arene exchange reactions of 2, 2+, and 22+. A standard 2D NOESY
pulse sequence (d1 (relaxation) - π/2 - t1 - π/2 - tm (mixing) - π/2
- t2 (aquisition)) was used on a Varian 400 MHz FT NMR
n
E1/2 ) -27 mV vs FcH/FcH+ (CH2Cl2/0.1 M Bu4NPF6).
Synthesis of [(η1:η6-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))-
(η1-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))Rh]+BF4- (2). [RhCl-
(COT)2]210 (40.2 mg, 0.112 mmol) and AgBF4 (22.7 mg, 0.116 mmol)
were reacted in 4 mL of CH2Cl2 for 50 min. The resulting reaction
mixture was filtered over Celite to remove a light gray precipitate and
diluted with 300 mL of CH2Cl2. A solution of 1 (110.0 mg, 0.224
mmol) in 150 mL of CH2Cl2 was added dropwise at -78 °C. After
3.5 h, the solvent was removed in vacuo to yield a red-orange powder.
Pure samples of 2 were obtained by layering a CH2Cl2 solution of the
product with diethyl ether three to six times (yield ) 45%, 120 mg,
0.050 mmol). 1H NMR (CD2Cl2): δ 1.71 (q, 2H, CH2PPh2 chelated),
2.06 (m, 2H, CH2PPh2 free), 3.80-4.02 (m, 7H, CH2O chelated and
η5-C5H5 free), 4.06 (s, 5H, η5-C5H5 chelated), 4.14 (m, 2H, CH2O free),
4.20 (m, 2H, η5-C5H4), 4.23 (m, 2H, η5-C5H4), 4.34 (m, 2H, η5-C5H4),
4.65 (m, 2H, η5-C5H4), 6.66 (m, 2H, C6H4 free), 6.86 (m, 4H, η6-C6H4),
7.12 (m, 8H, PPh2), 7.25-7.48 (m, 14H, C6H4 free and PPh2). 31P
NMR (CD2Cl2): δ 34.9 (dd, JRh-P ) 204.4 Hz, JP-P ) 39.3 Hz), 31.3
(dd, JRh-P ) 209.7 Hz, JP-P ) 39.3 Hz). FAB-HRMS: M+ calcd
1083.1353, found 1083.1313 m/z.
1
spectrometer. The region of the H NMR spectrum which contained
the CH2PPh2 resonances was used to measure rate data since those
resonances were well resolved for all oxidation states of 2. Mixing
times (tm) were varied with temperature. The rate of the exchange
reaction was determined by comparing the volumes of the cross peaks
(VAB and VBA) and the diagonal peaks (VAA and VBB) and evaluating
the mixing time in the equation k ) 1/tm ln[(r + 1)/(r - 1)], where r
) (VAA + VBB)/(VAB + VBA).16 From a plot of ln(k/T) versus 1/T
modeled according to the Eyring equation, ∆Hq and ∆Sq values for
the exchange reaction were calculated.18 Using the Eyring equation
in the form ∆Gq ) -RT ln(hk/kbTκ), where κ is 0.5, ∆Gq
values
(293)
were calculated. These values agreed well with those obtained using
∆Hq and ∆Sq in the equation ∆Gq ) ∆Hq - T∆Sq. To confirm these
data, each point on the Eyring plot was measured in two to four separate
experiments. From these data, errors in the exchange rates (∆k) were
1
estimated for 2, 2+, and 22+. Using a CD3OD H NMR standard, the
average fluctuation in the probe temperature from the desired values
(∆T) was determined to be (0.5 °C. Errors in ∆Gq were obtained by
applying ∆k and ∆T to the Eyring equation.18 The analysis of errors
in ∆Hq and ∆Sq was performed by fitting all plots of ln(k/T) vs 1/T to
give the minimum and maximum slopes. These slopes, and the
resulting y-intercept values, were used to calculate the minimum and
maximum values of ∆Hq and ∆Sq, respectively. The largest differences
from the best fit of the experimental data were taken as generous
estimates of the errors in ∆Hq and ∆Sq.
Synthesis of [(η1:η6-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))-
(η1-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))Rh]2+(BF4-)2 (2+). 2
(18 mg, 0.0154 mmol) and 1 equiv of AgBF4 (3 mg, 0.0154 mmol)
were reacted in 1.0 mL of CD2Cl2. After 1 min, an Ag0 mirror was
observed and a dark burgundy red solution was collected upon filtration
over Celite. Once the oxidation reaction was shown to be complete
by NMR (spectroscopic yield >98%), the sample was used for kinetic
NMR and reactivity studies. 1H NMR (CH2Cl2): δ 0.95 (q, JH-H
)
7.65 Hz, 2H, CH2PPh2 chelated), 1.79 (m, 2H, CH2PPh2 free), 3.79
(m, 2H, CH2O chelated), 3.81 (m, 2H, η5-C5H4), 3.89 (m, 2H, η5-C5H4),
3.95 (dt, JP-H ) 21.0 Hz, JH-H ) 4.40 Hz, CH2O free), 4.10 (s, 5H,
η5-C5H5), 4.56 (b, ferrocenium), 6.38 (m, 4H, C6H4), 6.69 (m, 4H, η6-
C6H4), 6.81-7.31 (m, 20H, PPh2), 8.62 (b, ferrocenium), 11.0 (b,
ferrocenium). 31P NMR (CH2Cl2): δ 34.3 (dd, JRh-P ) 203.9 Hz, JP-P
) 39.5 Hz), 30.3 (dd, JRh-P ) 210.7 Hz, JP-P ) 39.5 Hz).
Isomerization Studies of 2, 2+, and 22+. In a 5-mm diameter NMR
tube equipped with an air-free screw cap top and Teflon septa was
placed either 2 (0.010 mmol), 2+ (0.010 mmol), or 22+ (0.010 mmol)
in 0.48 mL of CD2Cl2. To the catalyst solution was added 0.12 mL of
ethyl allyl ether via an air-free syringe. The tube was then vigorously
shaken for 30 s and then placed into the probe at 20 °C. The progress
1
of the reaction was monitored by H NMR, and the Rh catalyst was
monitored by 31P NMR spectroscopy. Initial rates of isomerization
were obtained by plotting product concentration (cis and trans
combined) versus time. The best fit of the initial slope of these data
was taken as the initial rate, and all rates were normalized for the moles
of catalyst present to give the initial turnover frequencies. Each catalytic
run was performed at least two times to confirm the results.
Synthesis of [(η1:η6-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))-
(η1-(η5-C5H5)Fe(η5-C5H4C6H4OCH2CH2PPh2))Rh]3+(BF4-)3 (22+). 2
(18 mg, 0.0154 mmol) and 2 equiv of AgBF4 (6 mg, 0.0308 mmol)
were reacted in 1.0 mL of CD2Cl2. After 1 min, a Ag0 mirror was
observed and a very dark burgundy red solution was collected upon
filtration over Celite. Once the oxidation reaction was shown to be
complete by NMR (spectroscopic yield >98%), the sample was used
for kinetic NMR and reactivity studies. 1H NMR (CH2Cl2): δ -6.28
(b, ferrocenium), -3.61 (b, ferrocenium), 0.27 (q, JH-H ) 7.52 Hz,
2H, CH2PPh2 chelated), 1.53 (m, 2H, CH2PPh2 free), 2.58 (b, ferro-
cenium), 3.67 (m, 2H, CH2O chelated), 3.52 (b, ferrocenium), 3.79
(m, 2H, CH2O free), 4.32 (m, 4H, C6H4), 5.85 (m, 4H, η6-C6H4), 6.30-
Crystallographic Structure Determination. Single crystals of
2‚1.25CH2Cl2 suitable for the X-ray diffraction study were grown by
the slow diffusion of pentane into a CH2Cl2 solution of 2. Crystal,
data collection, and refinement parameters are given in Table 1. The
systematic absences in the diffraction data were consistent for the space
groups P1 and P1h. The E-statistics strongly suggested the centrosym-

Redox-Switchable Hemilabile Ligand Synthesis
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10753
metric space group P1h that yielded chemically reasonable and com-
putionally stable results of refinement. The structure was solved by
direct methods, completed by a subsequent difference Fourier syntheses,
and refined by a full-matrix least-squares procedures. Two indepen-
dently but chemically equivalent ion pairs were located in the
asymmetric unit. Also, there were 2.5 solvate molecules of methylene
chloride in the asymmetric unit. All non-hydrogen atoms were refined
with anisotropic displacement coefficients. One of the solvate mol-
ecules was disordered over an inversion center. One chlorine atom in
the disordered solvent molecule was furthered disordered between two
positions, 70:30, and was refined isotropically, and the hydrogen atoms
in the molecule were ignored due to disorder. Hydrogen atoms were
treated as idealized contributions.30 No attempt to locate the hydrogen
atom on the disordered solvent molecule was made. Supporting
Information contains further details of the structure determination.
Acknowledgment. We acknowledge the NSF (CHE-
9625391 and CHE-9357099) and the Petroleum Research Fund
(No. 30759-AC3) for generously funding this research. C.A.M.
also acknowledges an A. P. Sloan Foundation Fellowship and
a Camille Dreyfus Teacher-Scholar Award. C.S.S. acknowl-
edges an ARCS Foundation Fellowship.
Supporting Information Available: A detailed description
of the X-ray diffraction study of 2‚1.25CH2Cl , including tables
2
of experimental details, atom positional parameters, B(eq)
values, bond lengths, bond angles, and anisotropic displacement
parameters (17 pages). See any current masthead page for
ordering and Internet access instructions.
(30) All software and sources of scattering factors are contained in the
SHELXTL (version 5.03) program library (G. Sheldrick, Siemens XRD,
Madision, WI).
JA9723601