Synthesis and properties of compressed dihydride complexes of iridium: Theoretical and spectroscopic investigations

Article abstract of DOI:10.1021/ja048775l

Reaction of [Cp*Ir(P-P)CI][B(C₆F₅)₄] (P-P = bisdimethydiphosphinomethane (dmpm), bisdiphenyldiphosphinomethane (dppm)) with [Et₃Si][B(C₆F₅)₄] in methylene chloride under 1 atm of hydrogen gas affords the dicationic compressed dihydride complexes [Cp*Ir(P-P)H₂][B(C₆F ₅)₄]₂. These dicationic complexes are highly acidic and are very readily deprotonated to the corresponding monohydride cations. When the preparative reaction is carried out under HD gas, the hydride resonance exhibits J_HD = 7-9 Hz, depending upon the temperature of observation, with higher values of J_HD observed at higher temperatures. A thermally labile rhodium analogue, [Cp*Rh(dmpm)(H ₂)][B(C₆F₅)₄]₂, was prepared similarly. A sample prepared with HD gas gave J_HD = 31 Hz and J_HRh = 31 Hz, allowing the Rh complex to be identified as a dihydrogen complex. Quantum dynamics calculations on a density functional theory (DFT) potential energy surface have been used to explore the structure of the Ir complexes, with particular emphasis on the nature of the potential energy surface governing the interaction between the two hydride ligands and the Ir center.

Full text of DOI:10.1021/ja048775l

Published on Web 06/25/2004
Synthesis and Properties of Compressed Dihydride
Complexes of Iridium: Theoretical and Spectroscopic
Investigations
Ricard Gelabert, Miquel Moreno, Jose´ M. Lluch,* Agust´ı Lledo´s,* Vincent Pons, and
D. Michael Heinekey*
Contribution from the Departament de Qu´ımica, UniVersitat Auto`noma de Barcelona,
08193 Bellaterra, Barcelona, Spain, and Department of Chemistry, UniVersity of Washington,
Seattle, Washington 98195-1700
Received March 3, 2004; E-mail: lluch@klingon.uab.es
Abstract: Reaction of [Cp*Ir(P-P)Cl][B(C6F5)4] (P-P ) bisdimethydiphosphinomethane (dmpm), bisdiphen-
yldiphosphinomethane (dppm)) with [Et3Si][B(C6F5)4] in methylene chloride under 1 atm of hydrogen gas
affords the dicationic compressed dihydride complexes [Cp*Ir(P-P)H2][B(C6F5)4]2. These dicationic
complexes are highly acidic and are very readily deprotonated to the corresponding monohydride cations.
When the preparative reaction is carried out under HD gas, the hydride resonance exhibits JHD ) 7-9 Hz,
depending upon the temperature of observation, with higher values of JHD observed at higher temperatures.
A thermally labile rhodium analogue, [Cp*Rh(dmpm)(H2)][B(C6F5)4]2, was prepared similarly. A sample
prepared with HD gas gave JHD ) 31 Hz and JHRh ) 31 Hz, allowing the Rh complex to be identified as a
dihydrogen complex. Quantum dynamics calculations on a density functional theory (DFT) potential energy
surface have been used to explore the structure of the Ir complexes, with particular emphasis on the nature
of the potential energy surface governing the interaction between the two hydride ligands and the Ir center.
1. Introduction
substitution, with the distance in the H-H complex predicted
to be approximately 10% longer than in the corresponding T-T
complex.4 These predictions were subsequently verified by
experiment.5 These experiments rely on the well-established
inverse correlation between H-D coupling and H-H distance
Transition metal dihydrogen complexes have emerged as a
fascinating area of research in coordination chemistry. Early
work in this area focused on dihydrogen complexes with a
largely intact hydrogen molecule coordinated to the metal center
(RHH < 1.0 Å).1 Subsequently, a modest number of complexes
have been reported which have longer H-H distances up to
ca. 1.4 Å. Such complexes have been termed elongated
dihydrogen complexes and are of particular interest in that they
may represent an arrested intermediate state in the very
important oxidative addition reaction.2 Fascinating properties
have been reported for some of these molecules. Particularly
interesting are a family of Ru complexes exemplified by
[Cp*Ru(dppm)H2]+ (dppm ) bisdiphenylphosphinomethane),
where RHH ) 1.10 Å has been established by neutron diffrac-
tion.3 A theoretical study of this complex using density
functional theory (DFT) calculations showed that the two
hydrogen atoms are bound in a soft and highly anharmonic
potential energy surface (PES). This study led to the intriguing
prediction that RHH in this complex will increase slightly with
temperature, due to population of low-lying vibrational excited
states and that the bond distance will be sensitive to isotopic
RHH.
6-8 Thus, the observed values of JHD in partially deuterated
samples of the Ru complex decrease with increasing tempera-
ture, signaling that the thermally averaged value of RHH is
increasing.
We now report the preparation of dicationic Rh and Ir
analogues of these ruthenium complexes. The preparation of
these complexes required the development of new synthetic
methodologies for halide abstraction from positively charged
precursors. For Rh, an elongated dihydrogen complex is
observed. In the case of Ir, the resulting dicationic complexes
are best described as compressed dihydride complexes, based
on T1 and JHD data. In contrast to the previous studies of cationic
Ru complexes, the observed value of JHD increases with
increasing observation temperature. An improved procedure has
been developed for the correlation of RHH with JHD for
(4) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledo´s, A. J. Am. Chem. Soc.
1997, 119, 9840.
(5) Law, J. K.; Mellows, H.; Heinekey, D. M. J. Am. Chem. Soc. 2002, 124,
1024.
(6) Heinekey, D. M.; Luther, T. A. Inorg. Chem. 1996, 35, 4396.
(7) Maltby, P. A.; Schlaf, M.; Steinbeck, M.; Lough, A. J.; Morris, R. H.;
Klooster, W. T.; Koetzle, T. F.; Srivastava, R. C. J. Am. Chem. Soc. 1996,
118, 5396.
(8) Gru¨ndemann, S.; Limbach, H. H.; Buntkowsky, G.; Sabo-Etienne, S.;
Chaudret, B. J. Phys. Chem. A 1999, 103, 4752.
(1) General reviews: (a) Kubas, G. J. Metal Dihydrogen and σ-Bond
Complexes: Structure, Theory and ReactiVity. Kluwer: New York, 2001.
(b) Heinekey, D. M.; Oldham, W. J., Jr. Chem. ReV. 1993, 93, 913. (c)
Morris, R. H.; Jessop, P. G. Coord. Chem. ReV. 1992, 121, 155. (d)
McGrady, G. S.; Guilera, G. Chem. Soc. ReV. 2003, 32, 383.
(2) Heinekey, D. M.; Lledo´s, A.; Lluch, J. M. Chem. Soc. ReV. 2004, 33, 175.
(3) Klooster, W. T.; Koetzle, T. F.; Jia, G.; Fong, T. P.; Morris, R. H.; Albinati,
A. J. Am. Chem. Soc. 1994, 116, 7677.
9
10.1021/ja048775l CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 8813-8822
8813

A R T I C L E S
Gelabert et al.
2.28 (t, 3JH-P ) 2 Hz, 15 H, C5(CH3)5 of cis-IrH2), 2.20, 2.01 (m, 6 H,
P(CH3)2 of cis-IrH2), 2.39 (br. s, 15 H, C5(CH3)5 of trans-IrH2), 2.25,
2.17 (br. s, 6 H, P(CH3)2 of trans-IrH2), -10.53 (t, 2JH-P ) 6 Hz, cis-
IrH2), -10.60 (t, 2JH-P ) 20 Hz,. trans-IrH2). 31P{1H} δ (CD2Cl2, 202
MHz, 240 K) -76.2 (s).
complexes of this type, which has eliminated a previously
reported discrepancy between relaxation and coupling data.
Calculations of the electronic structure of complex [Cp*Ir-
(dmpm)H2]2+ (12+) (dmpm ) bisdimethylphosphinomethane)
have been carried out in combination with a nuclear dynamics
study. The structure of 12+ deduced from these studies is consis-
tent with the experimental observation. The calculated potential
energy surface for the interaction of the two hydride ligands
with the Ir center leads to an understanding of the temperature
dependence of the observed HD coupling constant in 1-d1.
A portion of the results described here has been previously
communicated.9
The Ir complex 2 and the Rh complex [Cp*Rh(dmpm)(H2)]-
[B(C6F5)4]2 were prepared similarly.
[Cp*Ir(dppm)H2][B(C6F5)4]2 (2). 1H NMR, δ (CD2Cl2, 750.13
MHz, 300 K) 8-7.4 (m, 20 H, P(C6H5)2, 5.82 (m, 1 H, CH), 4.2 (m,
1 H, CH), 1.75 (s (broad), 15 H, C5(CH3)5, cis-IrH2), 2.17 (s (broad),
15 H, C5(CH3)5, trans-IrH2), -9.31 (s (broad), 2 H, cis-IrH2), -9.50
2
(t, JH-P ) 17 Hz, 2 H, trans-IrH2).31P sel. {1H} δ (CD2Cl2, 303.65
MHz, 300 K) -23.42 (t, trans-IrH2, 2JH-P ) 17 Hz), -54.73 (s (broad),
cis-IrH2).
2. Experimental Details
[Cp*Rh(dmpm)(HD)][B(C6F5)4]2. 1H {31P} NMR, δ (CD2Cl2,
750.13 MHz) 5.5 (m, 1 H, CH), 3.9 (m, 1 H, CH), 2.12 (s, 15 H, C5-
2.1. General Procedures. Unless stated otherwise, all manipulations
were carried out under argon using Schlenk techniques. Phosphine
ligands were obtained from Strem Chemicals and used as received.
Hydrogen gas was purchased from Airgas and passed through a column
of activated molecular sieves prior to use. HD (g) and D2 (g) were
used as received from Cambridge Isotopes. Elemental analyses were
performed by Galbraith. NMR spectra were recorded on Bruker AC-
200, DPX-200, DRX-499, AM-500, and DMX-750 spectrometers.
Proton NMR spectra were referenced to the solvent resonance with
chemical shifts reported relative to TMS. 31P chemical shifts were
referenced to external 85% H3PO4. The NMR studies were carried out
in high quality 5 mm NMR tubes, utilizing deuterated solvent distilled
from standard drying agents. The conventional inversion-recovery
method (180-τ-90) was used to determine the relaxation times T1. In
each experiment the waiting period was longer than 10 times the
expected relaxation rate. Ten variable delays were employed, utilizing
appropriate pulse widths. Workup of spectra used for precise measuring
of coupling constants used zero filling to 128K data points prior to
Fourier transform.
1
(CH3)5), 1.97, 1.94 (s, 3 H, P(CH3)2), -3.61 (d of t, JRh-H ) 31 Hz,
1JH-D ) 31 Hz, 1 H, Rh(HD)).
[Cp*Ir(dppm)H][B(C6F5)4]. 1H NMR, δ (CD2Cl2, 750 MHz) 7.57,
7.52, 7.34 (m, 20 H, P(C6H5)2, 5.88 (d of t, 2JH-H ) 16 Hz, 2JH-P ) 10
Hz, 1 H, CH), 4.23 (d of t, 2JH-H ) 16 Hz, 2JH-P ) 12 Hz, 1 H, CH),
2
1.89 (b s, 15 H, C5(CH3)5), -14.35 (t, JH-P ) 25 Hz, 1 H, IrH). 31P
sel {1H} δ (CD2Cl2, 303.65 MHz, 300 K) -40.08 (d, 2JH-P ) 25 Hz).
[Cp*Ir(dmpm)H][B(C6F5)4]. 1H NMR, δ (CD2Cl2, 200 MHz) 5.01
4
(m, 1H, CH), 3.58 (m, 1H, CH), 2.27 (t, JH-P ) 2.5 Hz, 15 H, C5-
(CH3)5), 2.15, 1.91 (m, 6 H each, P(CH3)3), -15.09 (t, 2JH-P ) 25 Hz,
1 H, IrH). 31P sel. {1H} δ (CD2Cl2, 303.65 MHz, 300 K) -68.6 (d,
2JH-P ) 25 Hz).
[Cp*Ir(dmpm)Cl][B(C6F5)4]. A vial was charged with 170 mg of
(Ph3)CB(C6F5)4 (0.184 mmol), 84 mg of [Cp*Ir(dmpm)Cl]Cl (0.155
mmol), and 5 mL of methylene chloride. After the solution was stirred
for 1 h, the solvent was evaporated, and the remaining solid was washed
with 4 × 5 mL ether. Crystals were obtained by slow diffusion of Et2O
into a solution of [Cp*Ir(dmpm)Cl][B(C6F5)4] in CH2Cl2. The anion
exchange was monitored by 1H NMR spectroscopy, since the resonances
due to the protons of the CH2 and PMe2 groups are very sensitive to a
change of anion. 184 mg of [Cp*Ir(dmpm)Cl][B(C6F5)4] was recovered,
2.2. Synthesis. [Cp*Ir(dmpm)Cl]Cl. The ligand dmpm (55 µL,
0.348 mmol) was added to a solution of 70 mg (0.175 mmol) of
[Cp*IrCl2]2 in 30 mL of degassed methanol. After the solution was
stirred for 2 h under reflux, it was concentrated to approximately 5
mL, and the product precipitated by addition of 5 mL of Et2O.
Recrystallization in CH2Cl2/Et2O (50/50) at -26 °C affords 160 mg
[Cp*IrdmpmCl]Cl‚2H2O (78%). Anal. calcd for C15H29Cl2IrP2‚2H2O:
1
100% yield. H NMR δ (CDCl3, 499.85 MHz) 4.76 (m, 1 H, CH),
4
3.49 (m, 1 H, CH), 1.92 (t, JP-H ) 2.5 Hz, 15 H, C5(CH3)5), 1.80,
2
1.74 (“t”, JP-H ) 6 Hz, 3 H, P(CH3)2). 31P{1H} δ (CDCl3, 202.34
MHz) -58.5 (s).
1
C, 31.58; H, 5.83. Found: C, 31.65; 5.89 H. H NMR δ (CDCl3, 500
2.3. X-ray Structure of [Cp*Ir(dmpm)Cl][B(C6F5)4]. Yellow
crystals suitable for X-ray diffraction were obtained by diffusion of
Et2O into a solution of [Cp*Ir(dmpm)Cl][B(C6F5)4] in CH2Cl2 and
mounted on glass capillaries in oil. Diffraction measurements were made
on a crystal fragment of dimensions 0.24 × 0.24 × 0.19 mm in a
nitrogen stream at 130 K on a Nonius Kappa CCD diffractometer using
graphite-monochromated radiation (λ ) 0.710 73 Å). Crystal-to-detector
distance was 30 mm, and exposure time was 10 s per degree for all
sets. The scan width was 2.0°. Data collection was 81.4% complete to
28.35° in θ. A total of 12 856 reflections were collected covering the
indices h ) -11 to 11, k ) -11 to 15, and l ) -19 to 19. A total of
8265 reflections were symmetry independent, and the Rint ) 0.0761
indicated that the data were of average quality. Indexing and unit cell
refinement indicated a triclinic P lattice. The space group was found
to be P1h (No. 2) with cell parameters a ) 11.5620(4) Å, b ) 12.2440-
(5) Å, c ) 15.1400(8) Å, R ) 74.2690(18)°, â ) 80.1290(15)°, and γ
) 84.017(4)°. The cell volume was 2028.71 (15) Å3, and the calculated
density was 1.928 g/cm3, with Z ) 2. The data were integrated and
scaled using hkl-SCALEPACK. Solution by direct methods produced
a complete heavy atom phasing model closely related to the proposed
structure. All hydrogen atoms were placed using a riding model. All
non-hydrogen atoms were refined anisotropically by full-matrix least
squares. Tables of data collection, solution, and refinement details,
crystal data, atomic coordinates, and anisotropic thermal parameters
are included in the Supporting Information.
MHz) 6.14 (m, 1 H, CH), 3.49 (m, 1 H, CH), 1.95 (t, 4JP-H ) 2.5 Hz,
15 H, C5(CH3)5) 2.19, 1.77 (“t”, 2JP-H ) 6 Hz, 3 H, P(CH3)2). 31P{1H}
δ (CDCl3, 202.34 MHz) -59.4 (s). The previously reported complexes
[Cp*Ir(dppm)Cl]+ and [Cp*Rh(dmpm)Cl]+ were prepared similarly.
1
[Cp*Ir(dppm)Cl]Cl. H NMR, δ, (CDCl3, 500 MHz) 7.57, 7.49,
2
2
7.27 (m, 20 H, P(C6H5)2), 6.57 (d of t, JH-H ) 15.5 Hz, JH-P ) 9.5
Hz, 1 H, CH), 4.67 (d of t, 2JH-H ) 15.5 Hz, 2JH-P ) 13 Hz, 1 H, CH),
1.80 (t, JP-H ) 2.5 Hz, 15 H, C5(CH3)5). 31P{1H} δ (CDCl3, 202.34
4
MHz) -37.1 (s).
1
[Cp*Rh(dmpm)Cl]Cl. H NMR, δ, (CDCl3, 499.85 MHz) 4.8 (d
2
2
2
of t, JH-H ) 14 Hz, JH-P ) 11 Hz, 1 H, CH), 3.39 (d of t, JH-H
)
2
4
14 Hz, JH-P ) 13.5 Hz, 1 H, CH), 1.88 (t, JP-H ) 3.5 Hz, 15 H,
C5(CH3)5), 2.13, 1.77 (“t”, JP-H ) 6 Hz, 3 H, P(CH3)2). 31P{1H} δ
2
1
(CDCl3, 202.34 MHz) -22.09 (d, JRh-P ) 109 Hz).
[Cp*Ir(dmpm)H2][B(C6F5)4]2 (1). In a typical experiment, an NMR
tube fitted with a high vacuum Kontes valve was charged with 40 mg
of Ph3CB(C6F5)4. Et3SiH (1 to 2 mL) was added by vacuum transfer.
After the yellow-orange mixture was stirred overnight, excess Et3SiH
was removed by pumping, and 10 mg of [Cp*Ir(dmpm)Cl]Cl was
added. CD2Cl2 was added by vacuum transfer, and the tube was
1
degassed and flame sealed under 1 atm of H2 (or HD). H NMR δ
(CD2Cl2, 499.85 MHz, 240 K) 5.01 (m, 1 H, CH), 3.70 (m, 1 H, CH),
(9) Pons, V.; Heinekey, D. M. J. Am. Chem. Soc. 2003, 125, 8428.
9
8814 J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004

Compressed Dihydride Complexes of Iridium
A R T I C L E S
3. Calculational Details
3.1. Electronic Structure Calculations. For complex 12+, a global
Cs symmetry has been enforced throughout as the sole geometrical
constraint. The complete set of electronic structure calculations have
been performed with the GAUSSIAN 98 program.10 The DFT formal-
ism has been used with the three-parameter hybrid functional of Becke
and the Lee, Yang, and Parr’s correlation functional, more widely
known as B3LYP.11,12 An effective core operator has been used to
replace the inner electrons of the iridium atom,13 while its outer electrons
were described with the basis set associated with the pseudopotential
of Hay and Wadt with a standard double-ú LANL2DZ contraction.10,13
For the phosphorus atoms, the same pseudopotential and basis set were
used, enlarged with d polarization functions.14 The remainder of the
atoms of the system were described with the standard split-valence
6-31G basis set,15 except for the five carbon atoms in the cyclopenta-
dienyl ring and the one between the phosphorus atoms, for which the
6-31G(d) basis set was used,16 and for the hydrogen atoms directly
bound to the iridium, for which the 6-31G(p) basis set was used
instead.17 Localization of stationary points was achieved through use
of the Schlegel algorithm within a redundant internal coordinate
system.10,18
Figure 1. ORTEP drawing of the cationic portion of [Cp*Ir(dmpm)Cl]B-
(C6F5)4. Hydrogen atoms have been omitted for clarity.
3.2. Nuclear Dynamics Calculations. This work is concerned with
the dynamics of the dihydrogen ligand in complex 12+. Because of the
light masses of the atoms involved, a static study based on locating
and characterizing stationary points is not enough and their quantum
nature has to be taken into account explicitly. Within the adiabatic
separation of nuclear and electronic motions (Born-Oppenheimer
approximation), the stationary states of nuclear motion are obtained
by solving the nuclear Schro¨dinger equation:
Table 1. Selected Bond Distances and Angles for
[Cp*Ir(dmpm)Cl][B(C6F5)4]
distances, Å
angles, deg
P1-Ir1-P2
P1-C13-P2
Cp*-Ir1-P1
Cp*-Ir1-P2
Cp*-Ir1-Cl1
Ir1-Cl1
Ir-P1
2.3917(16)
2.2989(18)
2.2904(18)
1.850
71.81(6)
94.2 (4)
135.93
137.13
123.13
Ir-P2
Ir-Cp* centroid
(Tˆnuc + U(R))Ψ ) EΨ
(1)
reaction of 4 equiv of dmpm and dppm, respectively, with
[Cp*IrCl2]2 in refluxing methanol (dmpm ) bisdimethylphos-
phinomethane). The Rh complex [Cp*Rh(dmpm)Cl]Cl was
prepared similarly. Recrystallization from methylene chloride/
diethyl ether affords these complexes in high purity and adequate
yield.
Metathesis of [Cp*Ir(dmpm)Cl]Cl with (Ph3)CB(C6F5)4 af-
fords [Cp*Ir(dmpm)Cl] B(C6F5)4. Crystals suitable for X-ray
diffraction were grown from methylene chloride/diethyl ether
solutions. The structure of this complex is shown in Figure 1.
Relevant bond distances and angles are listed in Table 1.
Reaction of [Cp*Rh(dmpm)Cl]Cl in methylene chloride under
H2 with 2 equiv of [Et3Si][B(C6F5)4] affords a new thermally
labile dicationic species. In addition to the expected resonances
for the phosphine and Cp* ligands in the 1H NMR spectrum, a
very broad hydride resonance was detected at -3.55 ppm. When
the preparative reaction was carried out under HD gas, the
spectra shown in Figure 2 were obtained. The measured JHD
value is 31 Hz, which is approximately equal to JHRh, leading
to the observed four line pattern.
Here, R is a 3N-6 dimensional vector describing the arrangement of
the nuclei, and U(R) is the potential energy hypersurface, obtained by
solving the electronic Schro¨dinger equation for each nuclear arrange-
ment and adding to the electronic energy obtained in this way the
internuclear repulsion energy for that arrangement of nuclei. Tˆnuc is
the nuclear kinetic energy operator. Solving eq 1 using the discrete
variable representation (DVR) methodology of Colbert and Miller19
yields the stationary states that describe internal molecular motion, or,
in other words, the vibrational states of the molecule, as a set of energies
and nuclear wave functions.
4. Results
4.1. Synthesis. The precursor chloride complexes [Cp*Ir-
(dmpm)Cl]Cl and [Cp*Ir(dppm)Cl]Cl have been prepared by
(10) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Menucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ciolowski, J.; Ortiz, J.
V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaroni, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(11) Lee, C.; Yang, W.; Parr, G. R. Phys. ReV. B 1988, 37, 785.
(12) Becke, A. D. J. Chem. Phys 1993, 98, 5648.
(13) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(14) Ho¨llwarth, A.; Bo¨hme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.; Jonas,
V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys.
Lett. 1993, 208, 237.
Reaction of [Cp*Ir(dmpm)Cl]Cl in methylene chloride with
2 equiv of [Et3Si][B(C6F5)4] under H2 affords [Cp*Ir(dmpm)-
H2][B(C6F5)4]2 (1). [Et3Si][B(C6F5)4] was generated by reaction
of Et3SiH with (Ph3C)B(C6F5)4. A similar reaction with the
dppm analogue gives [Cp*Ir(dppm)H2][B(C6F5)4]2 (2). Both 1
and 2 are characterized as mixtures of two isomeric dihydride
species with both transoid and cisoid variants of a capped square
pyramidal geometry (see Scheme 1).
(15) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257.
(16) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.;
DeFrees, D. J.; Pople, A. A. J. Chem. Phys. 1982, 77, 3654.
(17) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(18) Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. J. Comput. Chem.
1996, 17, 49.
In the case of 1, the cis/trans ratio is 97:3; for complex 2,
the ratio is 65:35. The hydride resonance for the trans isomer
exhibits much larger coupling to the adjacent 31P nuclei in both
(19) Colbert, D.; Miller, W. H. J. Chem. Phys. 1992, 96, 1982.
9
J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004 8815

A R T I C L E S
Gelabert et al.
Scheme 1. Reaction of Ir Chloride Complexes with [Et3Si][B(C6F5)4] under Hydrogen
Figure 4. (a) Partial 1H {31P} NMR spectrum of [Cp*Ir(dmpm)HD]2+
,
1-d1. (b) 1H NMR spectrum (250 K, CD2Cl2, 750 MHz). The asterisk
designates the resonance for the trans isomer.
Figure 2. Partial 1H{31P} NMR spectrum of [Cp*Rh(dmpm)HD]2+ at (a)
500 MHz (CD2Cl2, 265 K) and (b) 750 MHz (CD2Cl2, 250 K).
1
[B(C6F5)4] and [Cp*Ir(dppm)H][B(C6F5)4], as indicated by H
and 31P NMR spectroscopy (see discussion).
4.2. T1 and H-D Coupling in Complexes 1 and 2.
Relaxation time determinations for the hydride resonances
corresponding to the cis isomers of 1 and 2 were carried out
using standard inversion recovery methods. For 2, T1(min) ) 157
ms was measured at 255 K (750 MHz), while 1 gives T1(min)
)
145 ms at 240 K (500 MHz).
Complexes 1 and 2 do not incorporate deuterium in the
hydride ligands when methylene chloride solutions were exposed
to one atmosphere of D2 gas for several weeks at 250 K or for
several hours at room temperature. Partially deuterated samples
were ultimately prepared by carrying out the synthetic procedure
1
under one atmosphere of HD gas. A representative H NMR
spectrum of the hydride region of complex 1-d1 is shown in
Figure 4.
Using H{31P}spectra, measurement of JHD values for 1-d1
1
Figure 3. (Top) Partial (hydride region) 1H NMR spectrum of [Cp*Ir-
(dppm)H2][B(C6F5)4]2 (2). (Bottom) 1H {31P} NMR spectrum (200 K, 500
MHz, CD2Cl2).
and 2-d1was carried out at over a range of temperatures. In
contrast to the previously reported Ru analogue of 2, where
JHD values slightly decreased with temperature, the observed
values of JHD for 1 and 2 increase significantly when the
observation temperature is increased (see Figure 5). For complex
2, the HD coupling declines slightly at higher temperatures due
to a slow exchange process with the trans isomer, which was
confirmed by EXSY NMR spectroscopy.
complexes. The signal due to the cis isomer is broad, due to
rapid dipole-dipole relaxation (see below). The 1H NMR
spectrum of the hydride region for complex 2 is shown in Figure
3. While 1 and 2 are moderately stable thermally at ambient
temperature (10% decomposition after 12 h), attempts to isolate
these complexes were hampered by their very high acidity. Even
in rigorously dried solvents, traces of water lead to deprotonation
to the corresponding monohydride cations [Cp*Ir(dmpm)H]-
These observations also allowed for evaluation of the isotope
shift caused by deuteration, which is upfield for both 1 and 2.
9
8816 J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004

Compressed Dihydride Complexes of Iridium
A R T I C L E S
Figure 7. Optimized structures for the dihydrogen minimum (top left),
dihydride minimum (top right), and librational transition structure (bottom).
All hydrogen atoms except the dihydrogen ligand have been omitted. The
point of view for the structure of the librational transition structure is
different from the other two to highlight the orientation of the dihydrogen
ligand.
Figure 5. Plot of JHD versus temperature for 1-cis-d1 and 2-cis-d1. Estimated
uncertainties are (0.3 Hz at low temperatures and (0.2 Hz at higher
temperatures.
Table 2. Geometries of the Stationary Points for Complex 12+ and
Energies Relative to the Dihydride Minimum
a
RHH
R
Ir-H2
H−Ir−H R
R
P−Ir−P
E
Ir-P
Ir-C(Cp)
structure
dihydrogen
(Å) (Å)
(deg)
(Å)
(Å)
(deg) (kcal/mol)
0.93 1.65 31.7 2.40 2.30
1.63 1.37 61.3 2.40 2.32
71.4
70.6
71.2
71.6
1.4
0.0
1.7
6.3
dihydride
TS dihydrogen-dihydride 1.12 1.55 39.7 2.40 2.31
librational TS 0.84 1.73 26.9 2.41 2.30
a C(Cp) stands for the five carbon atoms forming the cyclopentadienyl
ring.
Localization of the stationary points revealed that both
dihydrogen and dihydride structures exist within Cs symmetry,
all of them cisoid. The most stable structure found is the
dihydride structure, with RHH ) 1.63 Å. Only 1.4 kcal/mol
higher in energy is the dihydrogen structure, with RHH ) 0.93
Å. The transition state that connects both structures (with a
unique imaginary frequency of 619i cm-1) is only 1.7 kcal/mol
higher in energy than the dihydride structure and, consequently,
only 0.3 kcal/mol above the dihydrogen structure, a vanishingly
small barrier. A second transition state structure for the
librational motion of the H2 ligand is 6.3 kcal/mol above the
dihydride minimum, which indicates that a slow rotation regime
of the H2 unit can be assumed. It has been characterized as a
transition structure with a unique imaginary frequency of 546i
cm-1. Representations of the structures of these stationary points
are depicted in Figure 7. The transition state for the dihydrogen
to dihydride isomerization is not shown, due to its similarity to
the dihydrogen structure. Geometrical parameters and energies
for these structures are tabulated in Table 2.
Figure 6. Chemical shifts as a function of temperature for the hydride
resonances of 1 and 1-d1. The trans isomer shows no measurable isotope
shift.
At 300 K, the resonance for 1-d1 is observed 100 ppb upfield
of that for 1. This difference increases slightly with decreasing
temperature, reaching 134 ppb at 223 K (see Figure 6).
4.3. Theoretical Studies. The theoretical study consists of
two parts: First, electronic structure calculations are performed
to obtain an approximation to the potential energy (PE) surface
of the complex. Second, nuclear dynamics calculations are
carried out on this surface to obtain the vibrational energy levels
and the coresponding wave functions.
For complex 12+, DFT calculations with the B3LYP func-
tional have been used to determine electronic energies and wave
functions. The nuclear dynamics calculations allow the deter-
mination of vibrational eigenstates connected to the motion of
the hydrogen atoms. To this end, the matrix representation of
the Hamiltonian operator for such motion was constructed by
using the standard discrete variable representation (DVR)
method, as implemented by Colbert and Miller.19 Diagonaliza-
tion of this matrix yielded vibrational energy levels and wave
functions. The procedure followed to carry out the nuclear
dynamics calculations is essentially a 1D version of the
procedures previously reported.4
Nuclear dynamics calculations were undertaken using a one-
dimensional PE surface or, more properly, a PE profile. A
relaxed PE surface was derived by fixing the H-H distance at
different values and allowing the complex to relax to a restricted
minimum. Because all atoms have changed their positions from
one point to the next, the PE profile is constructed using as“path”
a collective variable s, whose value is determined by finding
9
J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004 8817

A R T I C L E S
Gelabert et al.
Figure 9. Calculated temperature dependence of the HH distance.
Figure 8. Relaxed potential energy profile for the elongation of the H-H
distance, and probability density for the first vibrational states. The abscissa
values correspond to distances in mass-weighted Cartesian coordinates.
mann average of RHH . The mean HH distance decreases with
temperature, as depicted in Figure 9.
Table 3. Vibrational Energy Levels and Geometrical Expectation
Values of the RHH Distance
5. Discussion
E
s i
RHH i
(Å)
5.1. Synthesis. The preparation of the cationic metal halide
precursors follows closely the procedure reported by Valder-
rama20 and co-workers, for the preparation of [Cp*Rh(dppm)-
Cl]Cl and [Cp*Ir(dppm)Cl]Cl. The previous procedure employs
2 equiv (per metal atom) of the appropriate bidentate phosphine
to a methanol solution of [Cp*IrCl2]2 or [Cp*RhCl2]2. We have
found that addition of 4 equiv of phosphine reduces the
formation of byproducts and simplifies workup of the reaction
mixture. The new complex, [Cp*Ir(dmpm)Cl]Cl, was conve-
niently prepared in 78% yield by this procedure.
state
(kcal/mol)
(amu1/2 Å)
0
1
2
3
4
5
6
7
8
9
0.65
1.55
2.18
3.00
4.07
5.27
6.58
7.97
9.42
10.91
1.85
1.51
1.23
1.45
1.48
1.51
1.54
1.58
1.61
1.65
1.54
1.37
1.19
1.34
1.36
1.37
1.39
1.40
1.42
1.44
The X-ray structure of the B(C6F5)4 salt shows the expected
geometry, with no unusual intermolecular contacts between the
anion and the cation. The closest C‚‚‚F distances are 3.048 Å
and 3.219 Å for the Cp* and dmpm ligands, respectively. The
Ir-Cl bond distance is 2.44 Å, comparable to the value 2.38 Å
reported for the ruthenium analogue Cp*Ru(dmpm)Cl.21 Since
previous reports indicated that the Ru bound chloride was very
labile, we were optimistic that halide abstraction reactions from
the Rh and Ir centers would be feasible.
the distance in mass-weighted Cartesian coordinates between
pairs of consecutive points (more on this in the Discussion
section). In this PE profile, the dihydrogen minimum is at s )
0.76 amu1/2 Å, and the dihydride one is at s ) 1.96 amu1/2 Å.
Figure 8 displays the relaxed PE profile as well as the probability
2
density (|Ψi| ) for the lowest vibrational states.
Building the DVR matrix representation of the nuclear motion
Hamiltonian over the PE profile U(s) and diagonalizing give a
set of stationary states corresponding to vibrational energy levels
and nuclear wave functions for a system confined to the
monodimensional potential U(s). Finally, from the vibrational
wave functions, it is straightforward to derive expectation values
for the collective coordinate s and obtain from it the corre-
sponding H-H distances by reversing the fit used when
computing the path variable s in the fitting of the PE profile.
The expected geometry for each vibrational state i is determined
by calculating the expectation value for the collective variable
s using eq 2:
In contrast to the facile chloride displacement observed for
Cp*Ru(dmpm)Cl, no reaction was observed when methylene
chloride solutions of [Cp*Ir(dmpm)Cl]Cl were treated with
NaB(C6F5)4 under hydrogen gas. With AgPF6 or AgBF4 under
hydrogen, reaction is rapid but leads to complex mixtures of
intractable products. Seeking a highly reactive halide abstracting
agent which can be combined with a noncoordinating anion,
we found that the silyl cation [Et3Si][B(C6F5)4] reported by
Lambert and co-workers22 gives efficient chloride abstraction
from Rh and Ir chlorides. Complexes 1 and 2 and the new Rh
complex [Cp*Rh(dmpm)(H2)][B(C6F5)4]2 were conveniently
prepared by this method (see Scheme 1). The Rh complex is
thermally labile, but the Ir complexes 1 and 2 are stable at room
temperature. As expected for dicationic hydride complexes, these
species are extremely acidic, being readily deprotonated by even
traces of water. Typical of this reactivity is deprotonation of
complex 1, which affords [Cp*Ir(dmpm)(H)]+. An attempt to
Ψi|sˆ|Ψi
Ψi|Ψi
s )
(2)
i
In eq 2, Ψi is the wave function for the ith vibrational state.
Expectation values for the HH distance, HH i, for each
R
vibrational state are estimated by interpolation of the s values
i
(20) Valderrama, M.; Contreras, R.; Bascun˜an, M.; Alegria, S. Polyhedron 1994,
14 (15-16), 2239.
used in building the relaxed PE profile of Figure 8. The final
results are gathered in Table 3. The expectation value of the
HH distance determined for the ground state is 1.54 Å. The
effect of temperature was estimated by computing the Boltz-
(21) Smith, D. C., Jr.; Haar, C. M.; Luo, L.; Li, C.; Cucullu, M. E.; Mahler, C.
H.; Nolan, S. P.; Marshall, W. J.; Jones, N. L.; Fagan, P. J. Organometallics
1999, 18, 2357.
(22) Lambert, J. B.; Zhang, S.; Ciro, S. M. Organometallics 1994, 13, 2430.
9
8818 J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004

Compressed Dihydride Complexes of Iridium
A R T I C L E S
regenerate 1 by treatment of [Cp*Ir(dmpm)(H)]+ with triflic
acid gave no reaction, consistent with the very high acidity of
1.
5.4. H-D Coupling in 1 and 2. H-D couplings were
determined over a range of temperatures for samples of 1-d1
and 2-d1. Interestingly, the observed couplings increase sig-
nificantly with increasing temperature, ranging for complex 1
from ca. 7.5 Hz at 220 K up to ca. 9 Hz at 300 K. This
temperature dependence contrasts with the behavior previously
reported for the singly charged Ru analogues, where a modest
decrease in H-D couplings with temperature was observed for
[Cp*Ru(dppm)(H2)]+.
In an earlier communication on the properties of 1,9 it was
noted that the JHD value of 8.1 Hz observed at 240 K is
consistent with a H-H distance of 1.34 Å, using the correlation
between H-H distance and H-D coupling developed by
Limbach and Chaudret.8 This value is not consistent with the
distance 1.49 Å determined from the relaxation data described
above. This discrepancy has led us to a reexamination of the
relationship between JHD and H-H distance.
5.2. Structure of the Rhodium Complex. Given the
important role of Rh complexes in homogeneous hydrogenation,
the relative scarcity of well documented dihydrogen complexes
of rhodium is surprising.23 Based on the observed JHD value of
31 Hz, [Cp*Rh(dmpm)(H2)][B(C6F5)4]2 is formulated as a
dihydrogen complex with a H-H distance of ca. 0.90 Å. The
previously reported singly charged Ru analogue with the same
ligand set has JHD ) 17 Hz, consistent with a H-H distance of
1.2 Å. Presumably, this large structural difference can be
attributed to significantly greater back-donation from the
electron-rich Ru center into the σ* orbital of the bound H2
ligand. The dicationic Rh center lacks sufficient electron density
in metal d orbitals for efficient back-donation to the bound H2.
Ru dihydrogen complexes with comparable structures are
known, but electron-withdrawing coligands are required, e.g.,
[Cp*Ru(CO)2(H2)]+, which has JHD ) 28 Hz.24
5.3. Structure of the Iridium Complexes. Complexes 1 and
2 exhibit two different hydride resonances, with quite different
H-P couplings. Based on previous reports for cationic Ru and
Os analogues, these can initially be assigned to cisoid and
transoid isomers in a capped square pyramidal geometry. For
complex 1, the predominant cisoid isomer has JHP ) 6 Hz, while
the minor transoid isomer exhibits JHP ) 20 Hz. This situation
is similar to that reported by Jia and co-workers for the Os cation
[CpOs(dppm)H2]+, where the cisoid isomer has JHP ) 6.5 Hz,
while the transoid isomer has JHP ) 31.8 Hz.25 The cis/trans
ratio varies significantly, depending upon the identity of the
phosphine ligand, with the more sterically demanding dppm
ligand leading to a ratio of 65:35, while the more compact dmpm
ligand gives a ratio of 97:3. This is readily explicable in terms
of steric interactions between the phosphine ligand phenyl rings
and the methyl groups of the Cp* ligand, which tend to favor
the transoid geometry.
Early analysis of this correlation focused on the H-H distance
range 0.8-1.2 Å, where abundant structural data from neutron
diffraction and solid-state NMR studies are available.6,7 This
led to the development of a simple inverse linear relationship
between JHD and RHH. Subsequently, Limbach and Chaudret
reported a more sophisticated analysis based on application of
the bond-valence bond-length concept to hydride and dihydrogen
complexes.8 To express JHD as a function of RHH or vice versa,
1
they based their model on the relation derived by Hush, JHD
) 43PHH.
27 A two-bond term was added to account for magnetic
couplings in trans polyhydrides. This additional term is pro-
portional to the square of PMH, the metal hydrogen bond order.
The distance R°HH is set to 0.74 Å. The constants bMH and bHH
used by Limbach and Chaudret are decay parameters previously
determined to be 0.404 Å in N-H‚‚‚H hydrogen bonded
systems. The M-H distance R°MH was set to 1.6 Å.8 The
following expression was obtained, where JLL′ is the magnetic
coupling between L and L′ and γ (L) is the gyromagnetic ratio
of L (L, L′ ) H, D, T):
A useful method for characterization of dihydrogen complexes
is the measurement of dipole-dipole relaxation rates in the 1H
NMR resonance for bound hydrogen. At the temperature where
the maximum rate of relaxation is observed (T1(min)), use of the
methodology developed by Halpern and co-workers26 allows
the determination of the H-H distance in a dihydrogen or
dihydride complex. For complexes 1 and 2, the minimum
relaxation times measured for the resonance due to the cisoid
isomers are 219 ms for 1 and 155 ms for 2 (750 MHz).
Assuming the slow rotation regime, the H-H distances derived
by this method are 1.49 Å and 1.45 Å, respectively.
2
JLL′ ) JHHγ(L)γ(L′)/γ(H)2 ) 1JLL′°PHH + 2JLL′°(1 - PHH
)
or
1JHD ) (43 - 22 JHD)exp
0.74 - RHH
+
[
]
b
2
0.74 - RHH
[
2JHD exp
+ 2JHD (3)
{
}
]
b
2
Assuming b ) 0.404 and JHD ) -2.8 Hz, eq 3 can be
rearranged and simplified to become
(23) Several Rh polyhydride complexes which include a dihydrogen ligand are
known. For example, [Cp*Rh(PMe3)(H2)H]+ (Taw, F. L.; Mellows, H.;
White, P. S.; Hollander, F. J.; Bergman, R. G.; Brookhart, M.; Heinekey,
D. M. J. Am. Chem. Soc. 2002, 124, 5100) and Tp′Rh(H2)H2 (Bucher, U.
E.; Lengweiler, T.; Nanz, D.; von Philipsborn, W.; Venanzi, L. M. Angew.
Chem., Int. Ed. Engl. 1990, 29, 548. Gelabert, R.; Moreno, M.; Lluch, J.
M.; Lledo´s, A. Organometallics 1997, 16, 3805). The only well documented
Rh complex reported to date with a single dihydrogen ligand is (PCP)Rh-
(H2), for which JHD ) 33 Hz has been reported: Xu, W.-W.; Rosini, G.
P.; Gupta, M.; Jensen, C. M.; Kaska, W. C.; Krogh-Jespersen, K.; Goldman,
A. S. J. Chem. Soc., Chem. Commun. 1997, 2273.
RHH ) 0.74 - 0.404 ln{0.5(17.3571 - [301.27 - 4(1 +
0.351 43 JHD)]0.5)}(Å) (4)
In the model of Limbach and Chaudret, the decay parameter
2
b and the two-bond coupling JHD are fixed to 0.404 Å and
-2.8 Hz, respectively. These values are based on previous work
and experimental data available. To verify this model, Limbach
and Chaudret tabulated H-H and M-H distance data for a
variety of dihydrogen and polyhydride complexes determined
by neutron and X-ray diffraction and plotted them against the
(24) Chinn, M. S.; Heinekey, D. M.; Payne, N. G.; Sofield, C. D. Organome-
tallics 1989, 8, 1824.
(25) Jia, G.; Ng, W. S.; Yao, J.; Lau, C.-P.; Chen, Y. Organometallics 1996,
15, 5039.
(26) Desrosiers, P. J.; Cai, L.; Lin, Z.; Richards, R.; Halpern, J. J. Am. Chem.
Soc. 1991, 113, 4173. Some confusion about the units employed in the
equations describing dipole-dipole relaxation has been recently clarified:
Bayse, C. A.; Luck, R. L.; Schelter, E. J. Inorg. Chem. 2001, 40, 3463.
(27) Hush, N. S. J. Am. Chem. Soc. 1997, 119, 1717.
9
J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004 8819

A R T I C L E S
Gelabert et al.
the most stable structure. This conclusion would be in agreement
with the relative oxidative power of iridium and ruthenium:
because ruthenium is more difficult to oxidize, [Cp*Ru(dppm)-
(H2)]+ has a more stable dihydrogen minimum whereas complex
1 has a more stable dihydride minimum. One has to be cautious
with this interpretation, however, because it is only accurate as
far as the location of the minima on the potential energy
hypersurface itself is concerned. The structure of a substance
as determined from experiment is not necessarily related on a
one-to-one basis with a given minimum in the potential energy
hypersurface, because nuclear motion (which occurs even at 0
K) is missing from the latter. As far as nuclei are concerned, it
is only a static picture. This situation is similar to what was
encountered in the other elongated dihydrogen complexes
studied before and even in apparently very different systems
such as the so-called “low-barrier hydrogen bonds”:29 the fact
that the “relevant” portion of the potential energy surface is
very flat and anharmonic and that inclusion of vibrational motion
(through consideration of the zero-point energy) is a must to
account for the geometry of the complex. Complex 12+ differs
from the other elongated dihydrogen complexes for which
nuclear dynamics calculations have been done in that there is
an actual potential energy barrier involved.
The motion of the light nuclei can be approximately ac-
counted for by computing the full potential energy hypersurface
U(R), where R is a 3N-6 dimensional vector of internal
coordinates describing a given arrangement of the nuclei. This
is difficult because of the high dimensionality, and some
modeling is needed to reduce the dimensionality of the system.
The simplest approach is to reduce the dimensionality of U(R)
to just one (monodimensional treatment). Generation of an
acceptable monodimensional approximation to the full hyper-
surface that retains the key features of the problem is crucial
here. This is in essence the process of building up a reduced
potential energy surface depending on just one dynamical
variable, the latter corresponding to some geometrical parameter
of the complex (one of the components of R or a function of
these).
Figure 10. RHH as a function of JHD. The data used in the fittings are
those published in ref 8. The blue line represents the relationship reported
by Limbach and Chaudret (ref 8). The red line is derived using parameters
discussed in the text.
values of JHD reported for these complexes. As shown in Figure
10 (blue line), a reasonable fit of these data to eq 3 results. A
significantly better correlation results from fitting the data with
2
b and JHD defined as variable parameters. A least-squares fit
against the experimental data allowed us to obtain the values
of 0.494 Å for b and -3.04 Hz for 2JHD. The plot of the resulting
relation is shown as the red line in Figure 10.
Using this modified relationship between JHD and RHH, the
observed value of JHD for complex 1 of 8.1 Hz at 240 K
corresponds to RHH ) 1.45 Å, which is in good agreement with
the value determined from T1(min) at this temperature of 1.49 Å.
5.5. Temperature Dependence of H-H Bond Distances
in Complexes 1 and 2. The observed values of JHD increase
significantly with increasing temperature, suggesting that the
H-H bond distances in these complexes are decreasing with
increasing temperature, in contrast to the situation with the
cationic Ru analogue of 2, where a small increase in HH distance
was observed at higher temperatures. The observed couplings
are independent of the magnetic field employed, which rules
out the possibility that residual dipolar coupling could be
responsible for the temperature dependence of the coupling.28
In an earlier report, some of us had suggested that the
observations for the iridium complex were qualitatively con-
sistent with a dihydrogen/cis-dihydride equilibrium, with two
distinct structures of similar energy in rapid equilibrium.9
Attempts to quantitatively model this equilibrium using two
limiting structures and reasonable estimates for the limiting
couplings were not successful.9 To resolve this problem, we
have turned to combined quantum dynamics + density func-
tional theory theoretical calculations.
A natural choice seems to be RHH itself, for our interest lies
in the dynamics of the H-H bond. This would mean construct-
ing a potential energy profile by picking different geometries
of the complex corresponding to different values of RHH and
computing the potential energy of these structures. A problem
arises in this process when selecting the values of the rest of
the geometric parameters of the complex, in particular, the Ir-
H2 distance. As can be seen in Table 2, the value of this
parameter changes noticeably in the set of stationary points
dynamically relevant (1.37 Å in the dihydride minimum, 1.65
Å in the dihydrogen minimum), so any “arbitrary” choice will
be incorrect. Besides, in the previous elongated dihydrogen
complexes studied, it was discovered that the dynamics of the
H-H and M-H2 bonds are inextricably intertwined.30
5.6. Theoretical Studies. In contrast to the singly charged
Ru analogues where a dihydrogen complex was the sole
minimum, the calculated structures for stationary points in
complex 12+ suggest that a dihydride structure with RHH ) 1.63
Å is slightly lower in energy than the dihydrogen structure with
RHH ) 0.93 Å (see Table 2 and Figure 7). The very small
calculated value of the energy barrier connecting the dihydrogen
and dihydride minima makes it unlikely that either minimum
corresponds to a separate chemical entity with real existence,
for including vibrational zero-point energy could wash out the
barrier (see Figure 8). Neither minimum correctly describes the
structure found experimentally. Nonetheless, based on electronic
structure results alone, one could conclude that complex 1 from
a theoretical point of view is a dihydride complex, this being
A solution exists that manages to combine both motions in a
single dynamical variable: constructing a kind of “minimum
energy path” coordinate. In this simplified representation, the
potential energy profile is built with the energy of those points
obtained by constraining RHH to a set of values to span the
(29) Garcia-Viloca, M.; Gelabert, R.; Gonza´lez-Lafont, A.; Moreno, M.; Lluch,
J. M. J. Am. Chem. Soc. 1998, 120, 10203.
(30) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledo´s, A. Chem. Phys. 1999,
241, 155.
(28) Luther, T. A.; Heinekey, D. M. J. Am. Chem. Soc. 1997, 119, 6688.
9
8820 J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004

Compressed Dihydride Complexes of Iridium
A R T I C L E S
complete range of structures from dihydride to dihydrogen and
then to optimize the resulting structures subject to this constraint
on RHH. The potential energy profile would, in this way, be
relaxed with respect to all variables except RHH, visit all relevant
stationary points found, and represent in an appropriate way
those areas of the potential energy surface in which the Ir-H2
distance follows RHH adiabatically.
through areas with small curvature (or, in other words, of low
frequencies). This has a bearing on the zero-point energy
estimated, which mainly comes from the lowest frequency
modes, so it is underestimated. If other relevant degrees of
freeedom were included, zero-point energy would be increased,
and if the shape of the potential energy profile is taken into
account, the expectation values RHH i would likely change
slightly.
One cannot use this relaxed potential energy profile as a
function of RHH directly in the dynamical calculations: the
reason is that, by optimizing the complete complex at each point,
any two points in this profile differ not only by the increment
in H-H distance but also in the positions of the rest of nuclei.
Therefore, the value of the mass in the kinetic operator in the
nuclear Schro¨dinger equation is unclear. The solution to this
problem is to define a new dynamical variable s, which
corresponds to the path length along this potential energy profile
in mass-weighted Cartesian coordinates. In this way, the mass
is implicitly included in the kinetic energy operator, and the
dynamical equations can be solved without any ambiguity in
terms of s. To translate the relaxed potential energy profile from
RHH to s, one must orient all the optimized structures in the
same way, order them from smallest to largest RHH value, and
then compute the length of the linear vector in mass-weighted
coordinates that links any two consecutive structures in the
complete configurational space. We think that this proposal is
a reasonable option to capture, within the intrinsic limitations
of a monodimensional treatment, the dynamical features arising
from the motion of the light dihydrogen ligand.
The unusual shape of the PE profile, with a lower minimum
in the dihydride region and a secondary shallow minimum in
the dihydrogen region, has a bearing on the expectation values
of the RHH distance in vibrational excited states. In effect, the
first few states display decreasing RHH as the energy of the
state increases, leading to a decrease in the mean H-H distance
as temperature increases, as seen experimentally. The mean
H-H distance measured at a certain temperature can be
estimated by computing the Boltzmann average of RHH . The
results are depicted in Figure 9, where the mean H-H distance
decreases with temperature, in good agreement with experiment.
The RHH of 1.49 Å derived from T1(min) at 240 K can only be
truly compared with mean values coming from a Boltzmann
average at this temperature. Figure 9 shows that RHH at T ≈
250 K is approximately 1.50 Å, in good agreement with
experimental data.
The very small calculated value of the energy barrier
connecting the dihydrogen and dihydride minima makes it
unlikely that either minima corresponds to a separate chemical
entity with real existence, for including vibrational zero-point
energy could wash out the barrier. Neither minimum correctly
describes the structure found experimentally. Contrary to the
cases of the ruthenium complexes, in this case the most stable
minimum is that of the dihydride structure, and the delocaliza-
tion of the ground vibrational state pushes the expectation value
toward shorter values of the H-H distance. Accordingly, this
is a rather extreme case of an elongated dihydrogen complex,
and a term like “compressed dihydride” may be a more
appropriate descriptor for complexes of this type. Being aware
of the limitations of theoretical calculations, we think that the
description we propose here is reasonable and explains the
available experimental results.
The potential energy profile built in this way, because of the
reasons explained above, can be appropriately called a “relaxed
potential energy profile”. Figure 8 displays the relaxed PE
profile. The path length s is measured from an arbitrarily chosen
point, picked to be a clear dihydrogen structure sufficiently high
in energy (in particular, RHH ) 0.50 Å, which has RIr-H2
)
1.41 Å, and whose energy is 60.6 kcal/mol above the dihydride
minimum). In this profile, the dihydrogen minimum is at s )
0.76 amu1/2 Å, and the dihydride minimum appears at s )1.96
amu1/2 Å.
As described in the Results section, diagonalization of the
DVR matrix representation of the nuclear motion Hamiltonian
over the PE profile U(s) gives a set of stationary states
corresponding to vibrational energy levels and nuclear wave
functions for a system confined to the monodimensional
potential U(s). Expectation values for the collective coordinate
s were derived from these, and the corresponding H-H distances
were obtained by reversing the fit used when computing the
path variable s in the fitting of the PE profile. Hence, we expect
that a low-temperature neutron diffraction experiment will show
a single structure with RHH ) 1.54 Å (see Table 3), even though
with a certain width or uncertainty, as described by the
probability density for the vibrational state V ) 0 in Figure 8.
Some comments are warranted on the effects brought about
by this monodimensional model. As explained, the path s is
actually a kind of “minimum energy path” approach to the full
3N-6 dimensional potential energy hypersurface that captures
the relevant features for the dynamics of the dihydrogen ligand.
However, one might wonder whether there is any flaw in this
model that might affect the results and conclusions. Actually,
the path being a monodimensional relaxed path, it crosses the
dynamically relevant area of the potential energy hypersurface
6. Summary and Conclusions
The previously reported discrepancy between RHH values for
1 determined by relaxation and JHD methods has been clarified
by the development of an improved correlation between RHH
and JHD. Good agreement is found between these experimental
results and the results derived from quantum nuclear dynamics
on a DFT potential energy profile, which also provide an
explanation for the novel temperature dependence of the HH
distance. Our results suggest that complexes such as 1 and 2
cannot be described adequately by a single structure, in the sense
of the word usually used by chemists. Further, terms such as
“elongated dihydrogen” or “compressed dihydride” should be
used with caution. We favor the latter for 1 and 2 because the
lowest minimum is actually a dihydride, with the caveat that
the PE surface is extremely flat and the hydrogen atoms are
extensively delocalized.
The complexes reported here allow us to evaluate the
interactions of H2 with a family of metal fragments of the form
[CpM(PP)]n+. As previously reported for Ru and Os (n ) 1),
9
J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004 8821

A R T I C L E S
Gelabert et al.
dihydrogen complexes (Ru) or cis dihydride complexes (Os)
are observed. This is consistent with the general observation
that the second row transition element prefers to form dihy-
drogen complexes and maintain a lower formal oxidation state
than that of the corresponding third row element. In both cases,
variable amounts of a trans dihydride isomer are observed,
depending upon the steric and electronic contributions of the
chelating phosphine ligand PP. In the Ru family, a wide range
of HH distances are observed, depending upon the coligands.
As previously reported, certain ligand combinations leading to
HH distances of ca. 1.1 Å give novel temperature-dependent
values of RHH, due to the very soft PES corresponding to motion
of the H atoms.
equilibrium with ca. 20% of a WII dihydride complex.31 Iridium
complexes 1 and 2 are best described as compressed dihydride
complexes with HH distances of about 1.5 Å. In the case of Ir,
variable amounts of a trans dihydride isomer are also observed.
Acknowledgment. RG gratefully acknowledges the Minis-
terio de Ciencia y Tecnolog´ıa of Spain for a Ramo´n y Cajal
research contract. This research was supported by the Spanish
Ministerio de Ciencia y Tecnolog´ıa under Projects BQU2002-
00301 and BQU-2002-04110-C02-02. D.M.H. is grateful to the
U.S. National Science Foundation for support of this research.
Acquisition of the 750 MHz NMR spectrometer was supported
by the National Science Foundation and the Murdoch Charitable
Trust.
Supporting Information Available: Details of the X-ray
structure determination for [Cp*Ir(dmpm)Cl][B(C6F5)4]. This
material is available free of charge via the Internet at
http://pubs.acs.org.
For the dicationic species (n ) 2), we find that Rh gives a
conventional dihydrogen complex with a short HH distance of
ca. 0.9 Å. In contrast to Ru, no trans dihydride isomer is
observed. This effect of charge in this case is analogous to the
early observations on Re cations versus neutral W complexes,
where [(PCy3)2Re(CO)3(H2)]+ is entirely a dihydrogen complex,
while the neutral tungsten analogue (PCy3)2W(CO)3(H2) is in
JA048775L
(31) Heinekey, D. M.; Radzewich, C. E.; Voges, M. H.; Schomber, B. H. J.
Am. Chem. Soc. 1997, 119, 4172.
9
8822 J. AM. CHEM. SOC. VOL. 126, NO. 28, 2004