Combined computational and experimental study of substituent effects on the thermodynamics of H2, CO, arene, and alkane addition to iridium

Article abstract of DOI:10.1021/ja010547t

The thermodynamics of small-molecule (H₂, arene, alkane, and CO) addition to pincer-ligated iridium complexes of several different configurations (three-coordinate d⁸, four-coordinate d⁸, and five-coordinate d⁶) have been investigated by computational and experimental means. The substituent para to the iridium (Y) has been varied in complexes containing the (Y-PCP)Ir unit (Y-PCP = ν3-1,3,5-C₆H₂[CH₂ PR₂]₂Y; R = methyl for computations; R = tert-butyl for experiments); substituent effects have been studied for the addition of H₂, C-H, and CO to the complexes (Y-PCP)Ir, (Y-PCP)Ir(CO), and (Y-PCP)Ir(H)₂. Para substituents on arenes undergoing C-H bond addition to (PCP)Ir or to (PCP)Ir(CO) have also been varied computationally and experimentally. In general, increasing electron donation by the substituent Y in the 16-electron complexes, (Y-PCP)Ir(CO) or (Y-PCP)Ir(H)₂, disfavors addition of H-H or C-H bonds, in contradiction to the idea of such additions being oxidative. Addition of CO to the same 16-electron complexes is also disfavored by increased electron donation from Y. By contrast, addition of H-H and C-H bonds or CO to the three-coordinate parent species (Y-PCP)Ir is favored by increased electron donation. In general, the effects of varying Y are markedly similar for H₂, C-H, and CO addition. The trends can be fully rationalized in terms of simple molecular orbital interactions but not in terms of concepts related to oxidation, such as charge-transfer or electronegativity differences.

Full text of DOI:10.1021/ja010547t

Published on Web 08/17/2002
Combined Computational and Experimental Study of
Substituent Effects on the Thermodynamics of H2, CO, Arene,
and Alkane Addition to Iridium
Karsten Krogh-Jespersen,* Margaret Czerw, Keming Zhu, Bharat Singh,
Mira Kanzelberger, Nitesh Darji, Patrick D. Achord, Kenton B. Renkema, and
Alan S. Goldman*
Contribution from the Department of Chemistry and Chemical Biology, Rutgers,
The State UniVersity of New Jersey, New Brunswick, New Jersey 08903
Received February 28, 2001. Revised Manuscript Received June 7, 2002
Abstract: The thermodynamics of small-molecule (H2, arene, alkane, and CO) addition to pincer-ligated
iridium complexes of several different configurations (three-coordinate d8, four-coordinate d8, and five-
coordinate d6) have been investigated by computational and experimental means. The substituent para to
the iridium (Y) has been varied in complexes containing the (Y-PCP)Ir unit (Y-PCP ) η3-1,3,5-C6H2[CH2-
PR2]2Y; R ) methyl for computations; R ) tert-butyl for experiments); substituent effects have been studied
for the addition of H2, C-H, and CO to the complexes (Y-PCP)Ir, (Y-PCP)Ir(CO), and (Y-PCP)Ir(H)2. Para
substituents on arenes undergoing C-H bond addition to (PCP)Ir or to (PCP)Ir(CO) have also been varied
computationally and experimentally. In general, increasing electron donation by the substituent Y in the
16-electron complexes, (Y-PCP)Ir(CO) or (Y-PCP)Ir(H)2, disfavors addition of H-H or C-H bonds, in
contradiction to the idea of such additions being oxidative. Addition of CO to the same 16-electron complexes
is also disfavored by increased electron donation from Y. By contrast, addition of H-H and C-H bonds or
CO to the three-coordinate parent species (Y-PCP)Ir is favored by increased electron donation. In general,
the effects of varying Y are markedly similar for H2, C-H, and CO addition. The trends can be fully
rationalized in terms of simple molecular orbital interactions but not in terms of concepts related to oxidation,
such as charge-transfer or electronegativity differences.
or R-H.5,6 We present evidence that this principle may be
Introduction
seriously flawed.
The addition of CO and the “oxidative” addition of molecules
belonging to the general class R-H (where R ) H or a group
of comparable electronegativity such as silyl or hydrocarbyl)
are requisite steps in most organometallic catalytic cycles.1 Even
catalyses that do not directly involve these reactions frequently
contain at least one such reaction in reverse (e.g., C-H
elimination in the case of polymerization or oligomerization of
olefins). A thorough understanding of the factors (electronic,
steric, etc.) that govern the thermodynamics of these reactions
is therefore critical to any attempts at rational design or even
fine-tuning of organometallic catalysts.2 It is our belief, however,
that the present level of understanding of such factors is
remarkably limited. For example, currently a primary guiding
principle concerning the thermodynamics of these reactions is
that increased electron richness favors addition of either CO3,4
The functionalization of alkanes is an area where the
development of efficient catalysts is of great interest.7 In the
past few years, complexes containing “pincer”-ligated iridium
fragments, (PCP)Ir (where PCP ) η3-1,3-C6H3[CH2PR2]2; in
this paper, R ) methyl for computations, R ) tert-butyl for
experiments), have been found to be among the most efficient8,9
and certainly the most selective10 catalysts for the dehydrogen-
ation of alkanes. The thermodynamics of H2 addition to the
catalytic species (eq 1) is of particular interest in the context of
dehydrogenation, since the catalytic cycle presumably involves
the transfer of hydrogen from an alkane to the metal center (eq
3). Equation 3 is less exothermic than simple addition of H2 by
an amount equal to the enthalpy of alkane dehydrogenation (ca.
(5) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.; John
Wiley & Sons: New York, 1988; pp 1189-1194.
(6) Collman, J. P. Acc. Chem. Res. 1968, 1, 136-143.
(7) For some recent reviews of homogeneously catalyzed hydrocarbon func-
tionalization, see (a) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.;
Peterson, T. H. Acc. Chem. Res. 1995, 28, 154-162. (b) Shilov, A. E.;
Shul’pin, G. B. Chem. ReV. 1997, 97, 2879-2932. (c) Sen, A. Acc. Chem.
Res. 1998, 31, 550-557. (d) Guari, Y.; Sabo-Etiennne, S.; Chaudret, B.
Eur. J. Inorg. Chem. 1999, 1047-1055.
(8) (a) Gupta, M.; Hagen, C.; Flesher, R. J.; Kaska, W. C.; Jensen, C. M. Chem.
Commun. 1996, 2083-2084. (b) Gupta, M.; Hagen, C.; Kaska, W. C.;
Cramer, R. E.; Jensen, C. M. J. Am. Chem. Soc. 1997, 119, 840-841.
(9) Xu, W.; Rosini, G. P.; Gupta, M.; Jensen, C. M.; Kaska, W. C.; Krogh-
Jespersen, K.; Goldman, A. S. Chem. Commun. 1997, 2273-2274.
(10) Liu, F.; Pak, E. B.; Singh, B.; Jensen, C. M.; Goldman, A. S. J. Am. Chem.
Soc. 1999, 121, 4086-4087.
* Authors to whom correspondence should be addressed. E-mail:
krogh@rutchem.rutgers.edu (K.K.-J.); agoldman@rutchem.rutgers.edu
(A.S.G.).
(1) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and
Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, CA, 1987; pp 523-668.
(2) Martinho Simoes, J. A.; Beauchamp, J. L. Chem. ReV. 1990, 90, 629-
688.
(3) Werner, H. Angew. Chem., Int. Ed. Engl. 1968, 7, 931-941.
(4) (a) Sulfab, Y.; Basolo, F.; Rheingold, A. L. Organometallics 1989, 8, 2139-
43. (b) Basolo, F. Polyhedron 1990, 9, 1503-35.
9
10.1021/ja010547t CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 10797-10809
10797

A R T I C L E S
Krogh-Jespersen et al.
23-30 kcal/mol),11 shown in eq 2.
We report both theoretical and experimental approaches
toward the study of substituent effects for addition of small
molecules to complexes (Y-PCP)Ir and several other complexes
that contain the (Y-PCP)Ir unit.
[M] + H2 h [M](H)2
∆H1
(1)
We have synthesized the p-methoxy (Y ) OCH3) and
p-methoxycarbonyl [Y ) C(O)OCH3] substituted PCP ligands,
which represent electron-rich and electron-poor derivatives of
the parent ligand, respectively. The effect of these substituents
on the thermodynamics of H2 addition to (Y-PCP)Ir(CO) are
measured directly. We recently reported the products of arene
C-H addition to (PCP)Ir;22 equilibrium measurements reveal
the effects of derivatizing the arene undergoing C-H addition.
First-principles electronic structure calculations render the study
of a much wider range of para substituents feasible, and more
importantly, a fuller range of reactions may be examined. The
experimentally observed substituent effects are in excellent
agreement with the computed values; therefore, for the closely
related reactions where experimental values are not accessible,
we feel that we have excellent reason to be confident in the
accuracy of the calculated substituent effects.
alkane h H2 + alkene
∆H2 ≈ 23-30 kcal/mol (2)
[M] + alkane h [M](H)2 + alkene
∆H3 ) ∆H1 + ∆H2
(3)
On a more fundamental level, the PCP ligand affords an
excellent opportunity to vary ligand X in a fragment of the
general form trans-ML2X and to observe the resulting effect
on the thermodynamics of small-molecule additions. trans-
ML2X fragments (M ) Co, Rh, Ir; L typically a phosphine; X
typically an anionic group, especially halide) are found in a
wide range of important catalysts.1 Varying halides and related
ligands results in changes at the metal center that are substantial
and fairly complex; this has been particularly well established
with respect to energies of H2 addition.12-19 We have found
that the lighter, more electronegative halogens result (paradoxi-
cally) in increased electron density at the metal center of trans-
IrL2X(CO)20 and a decreased tendency toward oxidative addition
of H2;21 however, partly because bond lengths and other factors
are affected, elucidating the relative importance of steric,
electrostatic, and σ/π effects is a complex task. In the case of
the PCP ligand, the phenyl group is the equivalent of X in trans-
ML2X; systematic changes of the substituent at the position para
to the metal allow us to leave the ligand qualitatively unchanged
while varying electronic factors (exclusively) in an incremental
and controlled fashion.
Computational and Experimental Methods
Density Functional Calculations. Standard computational methods
based on density functional theory23 and implemented in the GAUSS-
IAN98 series of computer programs24 have been employed. Specifically,
we have made use of the three-parameter exchange functional of
Becke25 and the correlation functional of Lee, Yang, and Parr26
(B3LYP). The Hay-Wadt relativistic, small-core effective core po-
tential and corresponding basis set (split valence double-ú) were used
for the Ir atom (LANL2DZ model),27 and the second- and third-row
elements carried all-electron, full double-ú plus polarization function
basis sets (Dunning-Huzinaga D95(d)28 for Li, B, C, N, O, F;
McLean-Chandler29 for P). Hydrogen atoms in H2 or a hydrocarbon,
which formally become hydrides in the product complexes, were
described by the triple-ú plus polarization 311G(p) basis set;30 regular
hydrogen atoms present in the PCP ligand or in alkyl, aryl, boryl, and
amino groups carried a double-ú 21G basis set.31
(11) NIST Standard Reference Database Number 69, February 2000 release:
http://webbook.nist.gov/chemistry/.
(20) The inverse halide order for electron richness has been reported in other
systems; see, for example, Zietlow, T. C.; Hopkins, M. D.; Gray, H. B. J.
Am. Chem. Soc. 1986, 108, 8266-8267.
(21) (a) Abu-Hasanayn, F.; Krogh-Jespersen, K.; Goldman, A. S. Inorg. Chem.
1993, 32, 495-496. (b) Abu-Hasanayn, F.; Goldman, A. S.; Krogh-
Jespersen, K. J. Phys. Chem. 1993, 97, 5890-5896. (c) Abu-Hasanayn,
F.; Krogh-Jespersen, K.; Goldman, A. S. Inorg. Chem. 1994, 33, 5122-
5130.
(22) Kanzelberger, M.; Singh, B.; Czerw, M.; Krogh-Jespersen, K.; Goldman,
A. S. J. Am. Chem. Soc. 2000, 122, 11017-11018.
(23) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and Molecules;
University Press: Oxford, U.K., 1989.
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; 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.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Mailick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98 (Revision A.5); Gaussian
Inc.: Pittsburgh, PA, 1998.
(12) The M(PMe3)4X2 (M ) Mo, W) systems have been especially well studied
in this respect. Complexes of the lighter halides were found to react more
favorably with H2, providing an important contrast with H2 addition to
IrL2(CO)X, which is closely related to the (Y-PCP)Ir(CO) system examined
in the present work. More generally, however, conclusions drawn from
the Mo/W work are in good agreement with those of the present study
concerning the dominance of specific orbital interactions (filled-filled or
filled-empty), and in particular π donation from the ligand being varied,
in determining substituent effects. (a) Rabinovich, D.; Parkin, G. J. Am.
Chem. Soc. 1993, 115, 353-354. (b) Murphy, V. J.; Rabinovich, D.;
Hascall, T.; Klooster, W. T.; Koetzle, T. F.; Parkin, G. J. Am. Chem. Soc.
1998, 120, 4372-4387. (c) Hascall, T.; Rabinovich, D.; Murphy, V. J.;
Beachy, M. D.; Friesner, R. A.; Parkin, G. J. Am. Chem. Soc. 1999, 121,
11402-11417.
(13) (a) Rachidi, I. E.-I.; Eisenstein, O.; Jean, Y. New J. Chem. 1990, 14, 671-
677. (b) Riehl, J.-F.; Jean, Y.; Eisenstein, O.; Pe´lissier, M. Organometallics
1992, 11, 729.
(14) Clot, E.; Eisenstein, O. J. Phys. Chem. A 1998, 102, 3592-3598.
(15) (a) Lunder, D. M.; Lobkovsky, E. B.; Streib, W. E.; Caulton, K. G. J. Am.
Chem. Soc. 1991, 113, 1837-1838. (b) Poulton, J. T.; Folting, K.; Streib,
W. E.; Caulton, K. G. Inorg. Chem. 1992, 31, 3190-3191. (c) Caulton, K.
G. New J. Chem. 1994, 18, 25-41.
(16) Hauger, B. E.; Gusev, D.; Caulton, K. G. J. Am. Chem. Soc. 1994, 116,
208-214.
(17) Albinati, A.; Bakhmutov, V. I.; Caulton, K. G.; Clot, E.; Eckert, J.;
Eisenstein, O.; Gusev, D. G.; Grushin, V. V.; Hauger, B. E.; Klooster, W.
T.; Koetzle, T. F.; McMullan, R. K.; O’Loughlin, T. J.; Pe´lissier, M.; Ricci,
J. S.; Sigalas, M. P.; Vymenits, A. B. J. Am. Chem. Soc. 1993, 115, 7300-
7312.
(18) (a) Le-Husebo, T.; Jensen, C. M. Inorg. Chem. 1993, 32, 3797-3798. (b)
Lee, D. W.; Jensen, C. M. J. Am. Chem. Soc. 1996, 118, 8749-8750. (c)
Lee, D. W.; Jensen, C. M. Inorg. Chim. Acta 1997, 259, 359-362.
(19) Li, S. H.; Hall, M. B.; Eckert, J.; Jensen, C. M.; Albinati, A. J. Am. Chem.
Soc. 2000, 122, 2903-2910 and references therein.
(25) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(26) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(27) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(28) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry; Schaefer, H.
F., Ed.; Plenum: New York, 1976; pp 1-28.
(29) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639.
(30) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980,
72, 650.
(31) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J. Am. Chem. Soc. 1980, 102,
939.
9
10798 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
Table 1. Substituent Parameters Used for LFER Fits
Reactant and product geometries were fully optimized by gradient
methods.32 The (PCP)Ir fragment has the aryl group canted relative to
the P-Ir-P axis (C2 point group), so most of the complexes studied
possess no molecular symmetry (C1). In selected cases, higher symmetry
(Cs, C2V) was imposed to illustrate the importance of σ vs π effects.
Although a few calculations used PR2 ) PH2, our general computational
model for (PCP)Ir has methyl groups attached to the phosphorus atoms
[i.e., PR2 ) P(CH3)2], a compromise between the use of hydrogen atoms
and the alkyl groups actually employed in the catalytic systems (e.g.,
iPr or tBu). The computed reaction energies represent purely electronic
energies and are not corrected for thermal or vibrational effects.
Experimental. General experimental methods are included in the
Supporting Information. Also in the Supporting Information are
spectroscopic data, elemental analyses, and synthetic procedures for
the following compounds: 1,3-bis[di(tert-butyl)phosphinomethyl]-5-
methoxybenzene (MeO-PCP-H), (MeO-PCP)IrHCl, (MeO-PCP)IrH4,
(MeO-PCP)IrH2, methyl 3,5-bis[di(tert-butyl)phosphinomethyl]benzoate
(CH3OC(O)-PCP-H), [CH3OC(O)-PCP]IrHCl, and [CH3OC(O)-PCP]-
IrH4. Syntheses of the ligands, like those for the corresponding iridium
hydrido chlorides, were based on syntheses reported by Moulton and
Shaw33 for the parent ligand. Reduction of the hydrido chlorides to the
hydrides followed the methods of Kaska and Jensen and co-workers.34,35
Spectroscopic data and methods for in situ generation of the (PCP)Ir
aryl hydrides are also included in the Supporting Information.
Linear Free-Energy Relationship (LFER) Analysis. Since all
substituent effects examined involved phenyl para substituents, the
standard Hammett substituent parameter, σp, was initially used for a
linear free energy analysis of all data sets. Reasonably good correlations
of calculated ∆E values with σp were found for all reactions with
multiple data points. The resulting reaction single parameters are
expressed as Fsp, to distinguish them from dual-parameter F values (see
below), and likewise, standard Hammett substituent parameters are
written as σsp. In an attempt to dissect the substituent effects further
into σ and π effects, we analyzed these six reactions using dual-
parameter models wherein calculated ∆E values are fit to an equation
of the following form (σ is the substituent parameter and F is the
reaction parameter; R and I denote resonance and inductive, respec-
tively):
a
b
Y
σsp
σR(BA)
σIc
NH2
-0.66
-0.27
0.06
-0.82
-0.61
-0.45
0.00
0.12
0.27
0.50
0.00
0.30
0.65
OCH3
F
H
0.00
C(O)OCH3
NO2
0.52
0.78
0.14
0.15
a Reference 79. b Based on benzoic acid pH; ref 37. c Inductive effect;
ref 37.
Table 2. Computed LFER Parameters for Addition Reactions of
Eqs 4, 6, and 9-16, Varying Y (Y-PCP) or Z (Z-C6H5 or
Z-C6HMe2)a
eq
reaction
Fsp
FR
FI
4 [Y-M] + H2 f [Y-M]H2
4.0(5) 4.4(7)
2.3(3) 2.5(1)
3.2(5) 3.7(2)
2.7(12)
1.9(1)
2.0(3)
6 [Y-M] + Ph-H f [Y-M](Ph)(H)
6 [Y-M] + n-Bu-H f [Y-M](n-Bu)(H)
[M] + Z-Ar-H f [M](Z-Ar)H
-6.0(9) -4.5(10) 8.2(17)
9 [M] + Z-Ar-H f [M](Z-Ar-horizontal)H
10 [M] + Z-Ar-H f [M](Z-Ar-vertical )H
-8.2
0.7
-6.3
-7.2
11 [Y-M](CO) + H2 f trans-[Y-M](CO) H2 -1.3(2) -1.6(1) -0.4(1)
11 [Y-M](CO) + H2 f cis-[Y-M](CO)H2
-1.3(2) -1.6(1) -0.4(1)
12 [Y-M](CO) + Ph-H f [Y-M](CO)H(Ph) -0.9(4) -1.4(2) -0.04(15)
12 [M](CO) + Z-Ar-H f [M](CO)(Z-Ar)H
13 [Y-M]H2 + H2 f [Y-M]H4
-4.7
-2.5(5) -3.0(2) -1.2(3)
2.7(4) 2.9(5) 2.1(10)
-2.6(4) -3.1(2) -1.1(3)
-2.5(3) -2.9(2) -1.3(3)
-1.5(3) -1.9(2) -0.6(2)
-7.0
14 [Y-M] + CO f [Y-M](CO)
15 [Y-M]H2 + CO f trans-[Y-M]H2(C O)
15 [Y-M]H2 + CO f cis-[Y-M]H2(CO)
16 [Y-M](CO) + CO f [Y-M](CO)2
a Positive values of F indicate that addition is favored by increasing
electron donation by Y or Z. M ) (PCP)Ir; [Y-M] ) (Y-PCP)Ir.
dual-parameter models. For all reactions investigated, the signs (positive
or negative) of the reaction parameters were the same for all six scales
used. Thus, the answers to the questions of major concern here are
independent of the choice of dual-parameter model. Overall, good fits
were obtained with all the dual-parameter scales (for each model, the
r2 values averaged over all reactions ranged from 0.955 to 0.978). For
the single-parameter Hammett para-substituent model, the average r2
value was 0.935. The standard Taft-Lewis dual-parameter model gave
correlations as good or better than the others (average r2 ) 0.978), so
we invoke only this dual-parameter model and the accompanying σR(BA)
scale in our discussions of individual reactions (Table 1).
∆E ) FRσR + FIσI
For all reactions and for all six scales, much better correlations were
found for the π/resonance than for the σ/inductive parameters. On the
basis of both the calculated magnitudes of the reaction parameters and
the quality of the correlations, it appears that π/resonance effects are
generally considerably more important than σ/inductive effects. Cal-
culated reaction parameters are summarized in Table 2. A more detailed
description of the LFER analysis is in the Supporting Information.
“Resonance” effects are assumed to be largely attributable to π
interactions, while inductive effects are assumed to correlate with σ
donation. It should be noted, however, that these parameters are derived
empirically on the basis of para- and meta-substituent effects, and the
π/resonance and σ/inductive correlations are in no way rigorous.37,38
Thus, the parameters only crudely reflect the π/σ properties of the para
carbon. Numerous dual-parameter models have been proposed, all of
which prove valuable but with limited generality.37 In particular, several
resonance parameter scales (σR) have been proposed by Taft et al.37
and by others.38 These scales complement several scales applicable for
inductive effects. Of these perhaps the most widely used is the benzoic
acid scale [σR(BA)] based upon the reaction used by Hammett as the
standard for the original σP treatment.37 We investigated six different
Results and Discussion
Addition of H2 to (Y-PCP)Ir (Standard Linear Free-
Energy Relationships). Perhaps the two most effective systems
for catalytic alkane dehydrogenation are those that contain the
fragments (PCP)Ir8-10 or Rh(PMe3)2Cl.39,40 We have previously
calculated that H2 addition to either fragment is exoergic by
ca. 25-30 kcal/mol, approximately the enthalpy of alkane
dehydrogenation (eq 2).9 Thus, eq 3 is approximately thermo-
neutral for these fragments, and we believe this is a critical factor
contributing to the catalytic effectiveness of these species
(particularly for transfer-dehydrogenation). By contrast, (PCP)-
Rh adds H2 only weakly, and consequently alkane dehydroge-
(32) Schlegel, H. B. Modern Electronic Structure Theory; Yarkony, D. R., Ed.;
World Scientific Publishing: Singapore, 1994; pp 459-500.
(33) Moulton, C. J.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1976, 1020-
1024.
(34) Nemeh, S.; Jensen, C.; Binamira-Soriage, E.; Kaska, W. C. Organometallics
1983, 2, 1442-1447.
(35) (a) Gupta, M.; Hagen, C.; Flesher, R. J.; Kaska, W. C.; Jensen, C. M. Chem.
Commun. 1996, 2083-2084. (b) Gupta, M.; Hagen, C.; Kaska, W. C.;
Cramer, R. E.; Jensen, C. M. J. Am. Chem. Soc. 1997, 119, 840-841.
(36) Morales-Morales, D.; Redon, R.; Wang, Z.; Lee, D. W.; Yung, C.;
Magnuson, K.; Jensen, C. M. Can. J. Chem. 2001, 79, 823-829.
(37) Ehrenson, S.; Brownlee, R. T. C.; Taft, R. W. In Progress in Physical
Organic Chemistry; Streitwieser, A. S., Taft, R. W., Eds.; John Wiley &
Sons: 1973; Vol. 10, pp 1-80.
(39) (a) Maguire, J. A.; Goldman, A. S. J. Am. Chem. Soc. 1991, 113, 6706-
6708. (b) Maguire, J. A.; Petrillo, A.; Goldman, A. S. J. Am. Chem. Soc.
1992, 114, 9492-9498.
(40) Rosini, G. P.; Soubra, S.; Wang, S.; Vixamar, M.; Goldman, A. S. J.
Organomet. Chem. 1998, 554, 41-47.
(38) Wells, P. R. In Linear Free Energy Relationships; Academic Press: London,
1968; pp 1-45.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10799

A R T I C L E S
Krogh-Jespersen et al.
Table 3. Computed Absolute and Relative Reaction Energies for
resonance effects span a range of 4.3 kcal/mol, whereas induc-
tive effects span only 1.8 kcal/mol. Independent of the greater
value of FR, in this as in all other cases the statistical certainty
of the correlation with resonance effects is also greater (the P
value associated with FR is PR ) 0.7%; PI ) 10% for FI).
Addition of H2 to (Y-PCP)Ir: Supporting Evidence for
the Importance of π Effects. Further evidence for the
importance of resonance or π effects is obtained from calcula-
tions conducted with C2V symmetry imposed on the (Y-PCP)Ir
fragment. The presence of the mirror planes establishes a
rigorous σ/π separation, and rotation of the locally planar NH2,
NO2, or BH2 groups by 90° from coplanarity with the aryl group
will virtually eliminate their π-accepting/donating ability.
Addition of H2 to (Y-PCP)Ira
symmetry imposed (C )b
2v
Y
∆E
∆∆E
∆E
∆∆E
NH2
-26.37
-2.36
-27.68
-24.87
-2.97 (NH2 coplanar)
NH2(90)
OCH3
F
-0.17 (NH2 orthogonal)
-25.80
-24.94
-24.82
-24.01
-22.11
-20.89
-1.78
-0.93
-0.81
0.00
1.90
3.13
Li
H
-24.71
0.00
C(O)OCH3
NO2
NO2(90)
BH2
BH2(90)
-21.34
-23.95
-20.51
-24.92
3.36 (NO2 coplanar)
0.76 (NO2 orthogonal)
3.63 (BH2 coplanar)
-20.51
3.51
-0.21 (BH2 orthogonal)
a Reaction energies, in kilocalories per mole, were calculated from eqs
4 and 5. b NH2(90), BH2(90), and NO2(90) refer to calculations in which
the respective group is held orthogonal to the PCP aryl ring. In all other
calculations the group was either coplanar (constrained symmetry) or
approximately coplanar with the aryl ring (unconstrained).
nation by this fragment is strongly endoergic.9 Conversely, H2
addition to Ir(PMe3)2Cl is too highly exoergic (ca. 60 kcal/mol;
eq 1) for this species to act as a dehydrogenation catalyst.
Obviously, the thermodynamics of the alkane dehydrogenation
step and, more generally, the catalytic effectiveness of fragments
trans-ML2X are strongly dependent on the electronic properties
of X. Thus, both for reasons specific to the (PCP)Ir system and
for reasons much more general, it is of interest to explore the
energetic effects resulting from systematic variations in the
electronic properties of the aryl group in (PCP)Ir.
As seen in Table 3, the computed energy of H2 addition to
(Y-PCP)Ir fragments (∆E4, eq 4) generally becomes more
negative with increasing electron-donating ability of Y. This is
most conveniently described in terms of ∆∆E, the energy for
the homodesmotic (isodesmic) reaction 5:41
rotation of the Y-group in (Y-PCP)Ir prior to substrate addition
The BH2 group (for which there are no published substituent
values), when oriented in the PCP aryl plane, is the most
unfavorable substituent examined for eq 5: ∆E4 ) 3.6 kcal/
mol, which may be compared with ∆E4 ) -3.0 kcal/mol for
the most favorable substituent, the NH2 group. Rotating NO2,
NH2, or BH2 perpendicular to the aryl ring plane, so that
conjugation cannot occur with the aryl π system, results in very
small absolute values of ∆E4 for all three substituents. Upon
90° rotation of the group, the small ∆E4 values are almost
identical in the case of NH2 (-0.17 kcal/mol) and BH2 (-0.21
kcal/mol), the strongest π donor and π acceptor, respectively.
Calculations on p-Li-substituted (Y-PCP)Ir afford ∆E4 )
-0.85 kcal/mol, which is surprisingly close to the value found
when Y ) F: ∆E4 ) -0.93 kcal/mol. Li is obviously not an
intrinsic π donor; however, as a strong σ donor it effects a
significant build-up of negative charge in the σ-type orbitals of
the carbon to which it is bound. This excess negative charge in
the σ system results in a polarization of the π cloud toward the
metal and thus Li-PCP may be significantly π-donating relative
to unsubstituted H-PCP. To further gauge the “π-donating”
ability of Li we have calculated the barrier to rotation around
the B-C bond of several derivatives (p-Y)C6H4-BH2 (i.e., the
difference in energy between the ground-state conformation,
which has coplanar BH2 and phenyl units, and the conformer
in which the two units are perpendicular).
(Y-PCP)Ir + H2 h (Y-PCP)Ir(H)2
∆E4
(4)
(H-PCP)Ir(H)2 + (Y-PCP)Ir h
(H-PCP)Ir + (Y-PCP)Ir(H)2
∆∆E4 (5)
The magnitudes of these substituent effects are on the order of
several kilocalories per mole (Table 3), which implies that they
have the potential to affect reaction rates by several orders of
magnitude, if they are manifest in any way in a rate-determining
transition state (directly or in a preequilibrium). For the single-
parameter LFER analysis, the value of Fsp (the single-parameter
reaction coefficient, see Methods section and Table 2) is 4.0(5)
(P ) 0.15%; errors in parentheses represent one standard
deviation in all cases); the one exception to the correlation of
∆∆E with σsp (the standard Hammett substituent parameter) is
fluorine (σsp ) 0.06; ∆∆E4 ) -0.93 kcal/mol). Fluorine,
however, is π-donating as well as σ-withdrawing. If only
resonance effects are considered, no anomalies are found among
the substituents examined. The energy of H2 addition is found
to increase with increasing value of the resonance parameter
(see Methods section) as follows: NH2 < OCH3 < F < H <
C(O)OCH3 < NO2. By use of the σR(BA)/σI dual-parameter
model, reaction exoergicity is found to correlate with both
inductive and resonance substituent values with the reaction
parameters FR ) 4.4(7) and FI ) 2.7(12). The importance of
resonance effects is therefore greater, not only because FR > FI
but also because the range of σR values is greater; the implied
rotation about the H2B-C bond as a measure of the
π-donating ability of Y
The magnitude of this barrier is expected to primarily reflect
the double-bond character in the B-C bond and hence provide
a measure of the π-donating/accepting ability of Y as transmitted
to the carbon adjacent to B. The calculated B-C rotational
barriers are 9.2, 10.5, 11.5, and 13.3 kcal/mol for Y ) C(O)-
OCH3, H, F, and OCH3, respectively, and they correlate
reasonably well with the σR(BA) parameters: F ) -4.6(10); r2
) 0.92. The calculated B-C rotational barrier for Y ) Li (12.7
kcal/mol) lies between those for F and OCH3, indicating that
(41) Sung, K. J. Org. Chem. 1999, 64, 8984-8989 and references therein.
9
10800 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
42
Table 4. Computed Absolute and Relative Reaction Energies for
Benzene and n-Butane C-H Addition to (Y-PCP)Ira
p-LiC6H4 is a better π donor than is p-FC6H4 and thus
implying that Li-PCP is more π-donating than F-PCP. If
increased Y-PCP f Ir σ donation were to significantly favor
the addition, then Li-PCP, at least comparable to F-PCP as a π
donor and undoubtedly a better σ donor than F-PCP, would be
expected to engender a significantly greater exoergicity for H2
addition, in contrast with the calculated result (Table 3). This
suggests that, despite the positive value of FI for reaction 4,
increased σ donation has a negligible or even unfavorable effect
on the thermodynamics of H2 addition to (PCP)Ir.
R ) Ph
R ) n-butyl
∆∆E
Y
∆E
∆∆E
∆E
OCH3
-7.73
-6.66
-5.84
-5.08
-1.07
0.00
0.82
1.58
2.57
4.19
5.46
5.28
-1.62
0.00
1.27
1.89
H
C(O)OCH3
NO2
a Reaction energies, in kilocalories per mole, were calculated from
eq 6.
In summary, statistical analysis of the energetic effects
produced by substituents with known σ values, in conjunction
with calculations on less conventional substituents (e.g., Li,
BH2), strongly supports the conclusion that the exoergicity of
H2 addition to (Y-PCP)Ir is highly sensitive to the π-donating/
withdrawing ability of Y (favored by π donation). This is
consistent with our previous study on addition to L2IrX
complexes in which it was also found that H2 addition energies
are determined largely by the degree of π-donating/withdrawing
ability of X,43 and with earlier work by Eisenstein in which the
importance of X f M π donation in L2IrX(H)2 complexes was
elucidated.13,14 Dual-parameter analysis suggests that σ donation
also favors H2 addition, albeit to a significantly lesser extent
than π donation; however, the statistical support for this
conclusion is weak, and additional calculations (Y ) Li, rotated
NH2, and BH2) seem to indicate that any such effect is negligible
or even inverse.
H-Ir-Cbutyl ∼ 71°) are best described as distorted trigonal-
bipyramidal.47 This complicates any attempt at direct compari-
sons of H-H and aryl-H additions to (PCP)Ir.
We have restricted the calculations to examinations of
benzene and n-butane adding to the parent complex (Y ) H),
the two derivatives we have actually synthesized [Y ) OCH3
and C(O)OCH3], and the nitro complex (Y ) NO2). It can be
seen from Table 4 that increased electron donation favors
addition of phenyl-H bonds to (Y-PCP)Ir. However, the effects
appear to be slightly weaker than found for addition of H2
[single-parameter model, Fsp ) 2.3(3) for arene versus 4.0(5)
for addition of H2; dual-parameter model, FR ) 2.5(1) and FI
) 1.9(1) for arene versus FR ) 4.4(7) and FI ) 2.7(12) for H2].
Thus, the equilibrium for an arene-H2 exchange reaction (eq
7) is calculated to lie slightly to the right when Y′ is more
electron-donating than Y:
(Y′-PCP)Ir(R)(H) + (Y-PCP)Ir(H)2 h
(Y-PCP)Ir(R)(H) + (Y′-PCP)Ir(H)2 (7)
Addition of Arene C-H bonds to (PCP)Ir: Para-
Substituted PCP Ligands. The thermodynamics of C-H
addition to (PCP)Ir and the related substituent effects are,
obviously, of interest in the context of (PCP)Ir-catalyzed
hydrocarbon functionalization:
Note that any mechanism for (Y-PCP)Ir-catalyzed alkane
dehydrogenation, olefin hydrogenation, or several other catalytic
reactions presumably involves species of the type (Y-PCP)Ir-
(R)(H) and (Y-PCP)Ir(H)2, though not necessarily the species
(Y-PCP)Ir. Thus, in that context, the relatiVe substituent effects
for R-H versus H2 addition (as expressed in eq 7) are arguably
as important as are the substituent effects for each reaction
individually (i.e., eqs 4 and 6).
We have experimentally measured the equilibrium constant
for eq 7 for Y′/Y ) OCH3/H and R ) Ph. Starting with either
(CH3O-PCP)Ir(Ph)(H) and (H-PCP)Ir(H)2 or the converse pair,
the same value is obtained, K ) 3.9 ( 0.2 (25 °C), correspond-
ing to ∆G7 ) -0.8 kcal/mol. Assuming that ∆S ≈ 0 (as seems
reasonable in view of the closely related nature of the species
involved) and that ∆H7 ≈ ∆E7, we obtain an experimental ∆E7
≈ -0.8 kcal/mol, in excellent agreement with the calculated
value of -0.7 kcal/mol [Tables 3 and 4; -1.78 - (-1.07) )
-0.71 kcal/mol].
The above experimental results are consistent with the results
from the full range of arenes used to calculate ∆E6: increased
electron donation by Y-PCP favors Ph-H addition but to a
lesser extent than it favors H2 addition. This is probably due to
increased π donation stabilizing the distorted trigonal-bipyra-
midal structure (calculated for the dihydrides) more than the
square-pyramidal structures (calculated for the aryl hydrides);
this differential stabilization has been reported previously for
isomeric ML2XH2 complexes.13 Consistent with this conclusion,
(Y-PCP)Ir + R-H h (Y-PCP)Ir(R)(H)
(6)
More generally, substituent effects on the thermodynamics of
C-H addition to transition metal complexes are not well
elucidated. It seems reasonable to presume that substituent
effects for C-H addition would correlate with those for H2
addition, but the quality of this correlation (assuming it does
exist) is not clearly established.44,45
Unlike the alkyl hydrides (presumed intermediates in (PCP)-
Ir-catalyzed alkane dehydrogenation), the aryl hydrides have
recently been observed and isolated.22 Thus, in the interest of
keeping the calculations as closely related as possible to
experimental efforts, we focus more on aryls than alkyls in this
paper. It is important to note, however, that the geometry of
the aryl hydrides is calculated to be pseudo-square-pyramidal
with apical hydride (CPCP-Ir-Cphenyl ∼ 174°; H-Ir-Cphenyl
∼
90°).46 In contrast, the dihydrides (CPCP-Ir-H ∼ 151°;
H-Ir-H ∼ 58°) and alkyl hydrides (CPCP-Ir-Cbutyl ∼ 159°;
(42) Calculated electron populations in the “unoccupied” p orbital of the BH2
group of p-Y-C6H4BH2 are in accord with the rotational barrier being
comparable for Y ) MeO and Y ) Li. These populations for several
substituents Y are as follows: C(O)OMe, 0.120; H, 0.130; F, 0.136; MeO,
0.157; Li, 0.160. Direct evidence for polarization by Li is revealed by
calculations on Y-C6H5. Populations of the pπ orbitals at the C1 and C4
positions, respectively, are calculated as follows for the respective sub-
stituents Y: H (1.000, 1.000); Li (0.891, 1.028); F (0.965, 1.031).
(43) Krogh-Jespersen, K.; Goldman, A. S. In Transition State Modeling for
Catalysis; ACS Symposium Series 721; American Chemical Society:
Washington, DC, 1998; pp 151-162.
(46) Krogh-Jespersen, K.; Czerw, M.; Kanzelberger, M.; Goldman, A. S. J.
Chem. Inf. Comput. Sci. 2001, 41, 56-63.
(47) See Supporting Information for selected structural parameters (bond lengths
and angles) for all calculated parent complexes and derivatives.
(44) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw, J. E. J.
Am. Chem. Soc. 1987, 109, 1444-1456.
(45) Labinger, J. A.; Bercaw, J. E. Organometallics 1988, 7, 926-928.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10801

A R T I C L E S
Krogh-Jespersen et al.
Table 5. Computed Absolute and Relative Reaction Energies for
Addition of Arene ([p-Z-R2C6H2]-H) C-H Bond to (PCP)Ira
the addition of the alkyl-H bonds is more favored by increased
electron donation [Fsp ) 3.2(5)] than is addition of the less
electron-rich aryl-H bonds [Fsp ) 2.3(3); see Table 4 and the
discussion of alkyl-H bond addition below]. This result is
obviously inconsistent with charge-transfer effects but can be
explained in terms of the alkyl hydrides possessing a structure
that is much closer to the trigonal-bipyramidal geometry of
(PCP)IrH2.
Addition of Arene C-H Bonds to (PCP)Ir: Para-
Substituted Arenes. The direct observation of C-H addition
products that undergo arene exchange affords the opportunity
to directly measure the thermodynamic effects of substituents
located on the arene addendum. We have recently reported that
1,3-xylene undergoes addition selectively at the 5-position
(giving a 3,5-dimethylphenyl hydride).22 The equilibrium con-
stant for eq 8a (Ar′ ) xylyl; Ar ) phenyl) is ca. 40 at -38 °C,
which corresponds to ca. 7 on a per-C-H bond basis and ∆G
) -0.9 kcal/mol. This is in perfect agreement with the
calculations, which predict ∆E8 ) -0.9 kcal/mol for benzene/
xylene [-6.66- (-5.74) ) -0.92 kcal/mol].
product
unconstrained (C )
aryl horizontal (C )
aryl vertical (C )
s
1
s
Z
∆E
∆∆E
∆E
∆∆E
∆E
∆∆E
NO2
C(O)OCH3
Cl
H
OCH3
NH2
-12.83
-8.67
-8.40
-6.66
-5.54
-4.41
-6.18
-2.01
-1.74
0.00
1.12
2.25
-8.65
-5.30
-6.74 -4.56
-3.35
0.00
-2.18
0.00
2.63
+5.98
-3.61 -1.43
a Reaction energies are given in kilocalories per mole.
methoxy: 6.2 vs 1.1 kcal/mol. In contrast, the effect exerted
by a nitro group on the PCP ring of the iridium complex is
only slightly greater than that of the methoxy group: 1.6 vs
1.1 kcal/mol (Table 5). The nitro group has both large electron-
withdrawing inductive and resonance effects, whereas methoxy
has an electron-withdrawing inductive effect but an electron-
donating resonance effect. Thus, these results suggest that both
π and σ electron-withdrawing groups (Z) on the benzene ring
very strongly favor addition, in contrast to the effects on the
PCP ring (Y), where π donation favors addition but the effect
of σ donation is much less significant or even unfavorable. The
relatively greater importance of π versus σ effects, for Y versus
Z, is reflected in the reaction parameters: varying Y in (Y-
PCP)Ir gives FR ) 2.5(1) and FI ) 1.9(1) (Table 2), while
varying Z for addition of the C-H bond at the 4-position in
C6H5Z gives FR ) -4.5(10) and FI ) -8.2(17).
(PCP)Ir(Ar′)(H) + ArH h (PCP)Ir(Ar)(H) + Ar′H (8a)
Monosubstituted benzenes can add to (PCP)Ir to yield several
phenyl hydride isomers and rotamers. To simplify the deter-
mination of equilibrium constants for eq 8, which involves two
arenes and therefore potentially twice as many isomers, we
therefore chose to compare ortho-substituted m-xylenes, i.e., 2-Z-
1,3-dimethylbenzenes (eq 8b). For each of these species, only
one aryl C-H bond undergoes addition.
Despite the negligible or even unfavorable effect of increasing
σ donation from Y-PCP, in all cases discussed above the
addition is favored by electron-donating substituents, Y, on PCP
as well as by electron-withdrawing substituents, Z, on the arene
undergoing addition. This generalization would appear to
support, at least on a qualitative level, the concept of C-H
addition being an “oxidative” process, i.e., a process driven in
large part by the transfer of charge from metal to addendum.
We believe, however, that this is not the case. The aryl hydride
product has two empty d orbitals, only one of which is of π
symmetry with respect to either the axis of the metal-carbon
bond formed upon C-H addition or the axis of the metal-
carbon(PCP) bond. That empty d(π) orbital is of appropriate
symmetry to overlap with the PCP π orbitals in accord with
the favorable effect of increased π donation from PCP. By
contrast, the π orbitals of the added arene overlap with a filled
d orbital, since the resulting aryl ligand is approximately
perpendicular to the PCP plane as in the structure labeled (Y-
PCP)Ir(horizontal-Ar-Z)(H).
We were successful for two xylyl substituents: nitro and
chloro. The relative equilibrium constants at -38 °C are KNO
:
2
KCl:KH ) 4600:43:1. The corresponding experimental values
of ∆∆G8b ≈ ∆∆E8b are -3.9 and -1.8 kcal/mol, in good
agreement with calculated values of -5.1 and -1.6 kcal/mol,
respectively. It bears noting that the effect of the nitro group is
calculated to be less when on xylene (-5.1 kcal/mol) than on
benzene (-6.2 kcal/mol). This is presumably due to the
neighboring methyl groups forcing the NO2 group out of
coplanarity with the ring and thus decreasing its ability to act
as a π acceptor (the calculated dihedral angle between the NO2
and phenyl planes is ∼50°). Likewise, the computed substituent
effect of MeO is negligible in the xylene case (0.01 vs 1.12
kcal/mol on benzene) because the MeO methyl group rotates
to form a dihedral angle of 90° with the aryl plane, effectively
eliminating π conjugation by the oxygen lone pair.
For all substituents examined in arene (p-ZC6H4-H) addition
to (Y-PCP)Ir, the effect on the arene ring undergoing addition
is opposite in sign to the effect exerted when the same
substituent is positioned on the PCP aryl ring: electron-rich
substituents favor addition when on the PCP ring [Fsp ) 2.3(3)]
and disfavor it when on the addendum ring [Fsp ) -6.0(9)].
The magnitudes of the relative effects, however, are highly
dependent on the location of the substituent. The nitro group,
when substituted on an adding benzene ring, yields a strikingly
high value for |∆∆E|, a value more than 5 times that of
9
10802 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
Due at least in part to steric factors, the horizontal structure
represents the most favorable geometry, even with methyl
groups on the P atoms. However, if we computationally replace
the methyl groups with H atoms, it becomes possible for the
phenyl group to rotate 90° without encountering excessive steric
stress, and its π orbitals can now interact with the same empty
d(π) orbital that overlaps with the PCP π orbitals. The energies
calculated for formation of the two isomers, with mirror
symmetry (Cs) imposed in both cases, are shown in Table 5.
For (Y-PCP)Ir(horizontal-Ar-H)(H), the geometry and the
reaction thermodynamics are similar to the system with me-
thylated P atoms and no imposed symmetry. Formation of the
vertical isomer (eq 10) is less favorable than that of the
horizontal isomer (eq 9) for Y ) Z ) H, but only by 1.2 kcal/
mol (3.35- 2.18). Much more importantly, however, the effect
of Varying substituent Z is Very different for eqs 9 and 10. In
particular, horizontal addition (eq 9) is disfavored by the strongly
π-donating NH2 in the Z position (∆∆E ) 6.0 kcal/mol versus
Z ) H; Table 5). In contrast, when the arene is added vertically
(eq 10), Z ) NH2 exerts a favorable effect on the reaction energy
(∆∆E ) -1.4 kcal/mol versus Z ) H). Evidently, π donation
from the para substituent on the arene can faVor addition to
(PCP)Ir, depending upon the isomer formed, in contradiction
with the idea of the reaction being driven or promoted largely
by transfer of charge from metal to ligand.
Addition of Alkyl C-H Bonds to (PCP)Ir. We have
recently reported calculations46 on the thermodynamics of
addition for different hydrocarbon C-H bonds to (H-PCP)Ir.
A large energy difference between alkane and arene addition is
calculated; Table 4 shows a difference of 11 kcal/mol favoring
benzene vs the primary C-H bond of n-butane.49 This result is
experimentally well precedented for late-metal systems; for
example, Wick and Jones50 have found a difference of 9.35 kcal/
mol between addition of benzene and n-pentane C-H bonds.
Early metal systems show much smaller differences (e.g., 1.2
kcal/mol for benzene vs n-butane addition to [tBu3SiO]2Tid
NSitBu3).51,52 In terms of the present system, the difference may
be attributable to a combination of σ effects (cf. the large value
of FI ) -8.2 for addition of H-C6H4Z) and π effects, with
alkyl being more π donating as it lacks the capacity of aryl to
act as a π acceptor. Although π donation into the empty metal
d(π) orbital should be favorable, this may be outweighed by
unfavorable interactions between filled d(π) orbitals present in
late metals and filled orbitals in alkyl ligands (Pauli steric
repulsion).53 Note in this context that the unfavorable effect of
the NH2 group for the horizontal addition of aminobenzene
(∆∆E9 ) +6.0 kcal/mol; Table 5) is of much greater magnitude
than the favorable effect for the vertical addition (∆∆E10 ) -1.4
kcal/mol).
Concerning ancillary ligand substituent effects, electron-
donating substituents on the PCP ligand favor addition of alkane
but less so than addition of H2. Thus the values of ∆E7 are
positive for alkyl-H addition also but less so than for phenyl-
H; i.e., the favorable effect of increased electron donation by
Y-PCP increases as phenyl-H < alkyl-H < H-H. The alkyl
hydride geometries are quite similar to those of the dihydrides
(CPCP-Ir-Cbutyl ∼ 160° and Cbutyl-Ir-H ∼ 70° vs CPCP-Ir-H
) 150° and H-Ir-H ∼ 60°); thus, alkyl-H addition probably
provides a better comparison with H2 addition. It is difficult to
distinguish π effects from σ effects but the major conclusions
concerning ancillary ligand effects for either arene or alkane
addition seem clear: C-H addition to (Y-PCP)Ir is favored by
increased electron donation from Y but to a lesser degree than
is H2 addition.
When located in the Y position, NH2 favors either horizontal
(eq 9) or vertical addition of arene (eq 10) (∆∆E ) -2.7 and
-1.9 kcal/mol, respectively; Z ) H). For vertical addition (eq
10), the favorable effect of Y ) NH2 (-1.9 kcal/mol) is quite
similar to the effect exerted by NH2 at the Z position (-1.4
kcal/mol). The fact that a substituent can exert effects of the
same direction and magnitude when located on either addendum
or ancillary ligand is obviously inconsistent with the idea that
substituent effects are dominated by the transfer of charge from
the metal to the addendum (or from addendum to metal). Instead,
this result is fully consistent with the proposal that the major
effect of the NH2 substituent derives from π donation into the
empty d orbital, which overlaps equally with both the PCP and
the addendum aryl groups of (Y-PCP)Ir(vertical-Ar-Z)(H).
H2 and C-H Addition to (Y-PCP)Ir(CO): Para-Substi-
tuted PCP Ligands. The calculated and experimental results
discussed above demonstrate an increased tendency toward
H-H and C-H addition for the more electron-rich (Y-PCP)Ir
derivatives and an increased tendency for addition of more
electron-deficient arenes. The single exception to this gener-
alization is the effect of the π-donating NH2 group on benzene,
when the aryl group is added perpendicular to the equatorial
plane. It may be tempting to consider this one exception to be
a mere anomaly and to rationalize the general effects in terms
of the definition of “oxidative addition”, i.e., a reaction that
In contrast to NH2, which exerts a substituent effect that is
predominantly π (σR ) -0.82; σI ) 0.12), NO2 exerts
predominantly a σ effect (σR ) 0.15; σI ) 0.65). Whereas Z )
NH2 affects eqs 9 and 10 in opposite directions, the effect of Z
) NO2 is found to be favorable for both eqs 9 and 10. The
value of ∆∆E is less favorable for vertical (-4.6 kcal/mol) than
for horizontal addition (-5.3 kcal/mol), presumably due to the
energetically unfavorable effect of π withdrawal for eq 10, but
apparently the strongly σ-withdrawing effect of NO2 is much
more important than the relatively weak π withdrawal (which,
by itself, should make a positive contribution to ∆∆E of eq
10). In addition to the much greater value of σI versus σR for
NO2, this observation is in accord with the very large inductive
parameter (FI ) -8.2) determined in the symmetry-uncon-
strained reactions; FI is expected to be, and apparently is,
approximately independent of the aryl group orientation.48
(49) n-Butane was chosen, rather than a smaller alkane, to allow for future studies
on primary vs secondary C-H addition and internal vs terminal â-elimina-
tion reactions.
(50) Wick, D. D.; Jones, W. D. Organometallics 1999, 18, 495-505.
(51) Bennett, J. L.; Wolczanski, P. T. J. Am. Chem. Soc. 1997, 119, 10696-
10719.
(52) For excellent discussions of the kinetic and thermodynamic selectivity
exhibited by transition metal complexes toward C-H bonds, particularly
in terms of factors other than steric, see the following and references
therein: (a) Bennett, J. L.; Vaid, T. P.; Wolczanski, P. T. Inorg. Chim.
Acta 1998, 270 (1-2), 414-423. (b) Wick, D. D.; Jones, W. D.
Organometallics 1999, 18, 495-505.
(53) (a) Ziegler, T.; Tschinke, V.; Becke, A. J. Am. Chem. Soc. 1987, 109, 1351-
1358. (b) Ziegler, T.; Tschinke, V.; Versluis, L.; Baerends, E. J.; Ravenek,
W. Polyhedron 1988, 7, 1625-1637.
(48) Accordingly, dual-parameter analysis for the vertical addition gives FR
)
0.7 and FI ) -7.2; since there are only three data points, there is no
overdetermination and no error analysis is possible.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10803

A R T I C L E S
Krogh-Jespersen et al.
Table 6. Computed Absolute and Relative Reaction Energies for
Addition of H2, Ph-H, and CO to (Y-PCP)Ir(CO)a
example, the heavier, more electropositive halogens were found
to favor H2 addition to IrL2(CO)X (although it was not realized
that such complexes were actually less electron-rich). With
reaction energies as the substituent effect criterion, the calculated
values in Table 6 lead squarely to the conclusion that the
prototypical “oxidative addition” reaction represented by eq 11
is not really oxidative at all. The Taft-Lewis reaction param-
eters are FR ) -1.64(6) and FI ) -0.45(11); thus the
unfavorable effect of increased electron donation is not large,
but the statistical significance is high (PR ) 0.01%; PI ) 2.7%)
and clearly in the direction opposite that of an “oxidative”
addition. On the basis of these values, the range of resonance
effects (1.6 kcal/mol) is significantly greater than the range of
inductive effects (0.3 kcal/mol).
product
(Y-PCP)Ir(CO)H
(Y-PCP)Ir(CO)HPh
(Y-PCP)Ir(CO)2
2
Y
∆E
∆∆E
∆E
∆∆E
∆E
∆∆E
BH2
NO2
-10.89 -0.76
-10.70 -0.57
-13.09 -1.08
-12.76 -0.75
-12.54 -0.53
12.93
12.84
13.15
-0.22
-0.32
0.00
C(O)OCH3 -10.51 -0.39
H
Li
F
OCH3
NH2
-10.13
-9.70
-9.61
-9.20
-8.91
0.00
0.42
0.51
0.93
1.21
-12.01
0.00
-12.03 -0.02
-11.41
-11.15
-10.81
0.60
0.86
1.20
13.94
14.20
0.78
1.04
a Reaction energies, in kilocalories per mole, were calculated from eqs
11, 12, and 16.
Experimental results unambiguously corroborate the calcu-
lated trend: addition of H2 to ([(CH3O)C(O)]-PCP)Ir(CO) (νCO
) 1930.0 cm-1) is more favorable than to the unsubstituted
complex (νCO ) 1927.7 cm-1), while H2 addition to the electron-
rich MeO-PCP complex (νCO ) 1925.5 cm-1) (2 atm of H2)
was not observable (K < 0.01 atm-1). Measured values of K11
involves removal of charge from the metal and therefore should
be favored by increased electron density on the metal center.
However, in the case of four-coordinate d8 complexes IrL2-
(CO)X, we have previously argued that increased electron
density (more specifically, increased π donation) disfaVors
addition of H2.21 Unfortunately, the comparison between
complexes with different ligands X (mostly halides) is not
straightforward: complexes of the less electronegative halogens
add H2 more readily (seemingly consistent with the oxidative
addition concept) but paradoxically, their complexes are less
electron-rich.20 Close examination reveals that significant
variables other than electron-donating ability are involved,
including the extent of orbital overlap and the metal-ligand
bond length, which results in greater electrostatic effects for
the smaller halides.21 The present (PCP)Ir systems, involving
different substituents at the para position, are in some respects
much more amenable to analysis. We therefore examined
addition of H2 to (PCP)Ir(CO) derivatives for the sake of
comparison with both the corresponding three-coordinate PCP
complexes and the four-coordinate carbonyl halide (Vaska-type)
analogues.
(22 °C) are as follows: C(O)OCH3-PCP, K11 ) 0.056 atm-1
;
H-PCP, K11 ) 0.036 atm-1 57-59
.
Addition of Arene C-H Bonds to (PCP)Ir(CO): Para-
Substituted Arenes. Like addition of H2, the addition of C-H
bonds to (Y-PCP)Ir(CO) (eq 12) is disfavored (computationally)
by increased electron-donating ability of Y-PCP (Table 6).
Although we have recently reported the isolation of the
benzene CΗ addition product shown in eq 12 (Z ) H), (PCP)-
Ir(CO)(Ph)(H),22 the substituent effects for addition are difficult
to test experimentally for two reasons: (i) presumably the
equilibrium lies far toward benzene elimination, precluding
equilibrium studies, and (ii) the kinetics of elimination are
extremely slow even at 110 °C, precluding calorimetric studies.
Nevertheless, in view of the experimentally confirmed substitu-
ent effects for H2 addition, there appears to be no reason to
doubt the validity of these calculated results.
Addition of H2 to either (PCP)Ir(CO) or (C(O)OCH3-PCP)-
Ir(CO) results in the formation of both cis and trans isomeric
dihydrides, in accord with a report by Milstein for addition to
the di(isopropyl)phosphino analogue:54
(Y-PCP)Ir(CO) + H2 h (Y-PCP)Ir(CO)(H)2 (11)
Does the correlation between ∆E and the electron-donating
ability of Y suggest that addition of H2 or C-H bonds to (Y-
PCP)Ir(CO) is best viewed as a “reductive addition”?60 Such a
concept would imply a transfer of electron density from the
The substituent effects are very similar for both isomers but
we have focused on the more symmetrical trans isomer, which
was found by Milstein and co-workers54 to be thermodynami-
cally favored in accord with our calculations. The reaction
energies, shown in Table 6, clearly illustrate that H2 addition
is disfaVored by increasing electron donation by the para
substituent on PCP [Fsp ) -1.3(2); Psp ) 0.3%]. Note that
varying the PCP para substituent raises none of the complicating
factors introduced when the halide is varied in IrL2(CO)X; for
example, upon substituting H with Li (or, experimentally,
methoxycarbonyl with methoxy), the primary result can only
be an increase of electron density at the para carbon. The concept
of “oxidative addition” as proposed by Vaska and co-work-
ers55,56 was in large part based on substituent effects; for
(57) Although the ∆∆E values (i.e., the substituent effects) are the focus of
this work, it should be noted that the calculated and absolute values for
the thermodynamics of addition are reasonably consistent (especially
considering the different alkyl groups on the phosphorus atoms in
experiments versus calculations). For example, for unsubstituted (PCP)Ir-
(CO), the experimental values for H2 addition are Kcis ) 0.068 atm-1 and
Ktrans ) 0.036 atm-1. Converting to reciprocal molar units gives ∆G )
-1.66 kcal/mol and -1.29 kcal/mol (ref 55). By use of the value for the
entropy of addition of H2 to Vaska’s complex, -29 eu (based on H2 in
solution; refs 55 and 58) gives ∆H ) -10.22 and -9.85 kcal/mol,
respectively. This compares with the respective calculated values of ∆E )
-6.50 and -10.13 kcal/mol.
(58) Based upon the solubility of H2 in 1-octene at 298 K, 0.0040 M/atm-1
:
Purwanto; Deshpande, R. M.; Chaudhari, R. V.; Delmas, H. J. Chem. Eng.
Data 1996, 41, 1414-1417.
(59) For formation of the cis isomer, the same trend is observed: [(CH3O)C-
(O)]-PCP, K11 ) 0.19; H-PCP, K11) 0.068. It also bears note that addition
of H2 to the di(tert-butyl)phosphino analogue (the actual ligand used in all
experiments described in this work) is significantly less favorable than to
the isopropyl analogue.54
(54) Rybtchinski, B.; Vigalok, A.; Bendavid, Y.; Milstein, D. Organometallics
1997, 16, 3786-3793.
(55) Vaska, L. Acc. Chem. Res. 1968, 1, 335-344.
(56) Vaska, L.; Werneke, M. F. Ann. N.Y. Acad. Sci. 1971, 172, 546-562.
(60) Crabtree, R. H.; Quirk, J. M. J. Organomet. Chem. 1980, 199, 99-106.
9
10804 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
Table 7. Computed Absolute and Relative Reaction Energies for
Addition of H2 and CO to (Y-PCP)Ir(H)2
addendum to the metal center; this should be faVored by
increased electron richness on the para position of a benzene
derivative (p-ZC6H4-H) undergoing C-H addition. To the
contrary, there appears to be a strong inVerse correlation based
on the two substituents that we have calculated: methoxyben-
zene adds less favorably to (PCP)Ir(CO) than benzene by 1.0
kcal/mol, and nitrobenzene adds more favorably by 5.3 kcal/
mol.
a
product
[M]H b
trans-[M](CO)H
cis-[M](CO)H
2
4
2
Y
∆E
∆∆E
∆E
∆∆E
∆E
∆∆E
BH2
NO2
-15.79 -1.62 -47.09 -1.79 -43.56 -1.87
-15.46 -1.30 -46.64 -1.34 -43.14 -1.45
C(O)OCH3 -15.18 -1.01 -46.29 -0.99 -42.79 -1.10
H
F
Li
OCH3
NH2
-14.17
-13.47
-13.43
-12.81
-12.15
0.00 -45.30
0.70 -44.44
0.74 -44.79
1.36 -43.77
2.02 -43.18
0.00 -41.69
0.86 -41.05
0.51 -41.06
1.54 -40.36
2.13 -39.68
0.00
0.64
0.63
1.33
2.01
Thus, somewhat counterintuitively, C-H addition to (Y-
PCP)Ir(CO) is favored by electron-withdrawing para substituents
on either the PCP ring (Y) or the adding arene (Z). [This is
analogous to the case of addition to (Y-PCP)Ir being favored
by electron-donating p-NH2 located on either the PCP ring or
the adding arene, for arene adding vertically to (Y-PCP)Ir (eq
10).] The arene addendum effect for eq 12 shows a greater
magnitude as might be expected, since the bond being formed
is closer to the substituent. However, less predictably, the
relative importance of inductive and resonance effects is
apparently quite different for the Y and Z substituents. The
limited data yield reaction parameters FR ) -4.7 and FI ) -7.0,
implying comparable resonance and inductive contributions if
the full range of substituents were varied at the Z position. In
contrast, for substitution on the Y-PCP ring reaction parameters
are FR ) -1.4(2) and FI ) -0.04(15). This is consistent with
the reaction parameters obtained for addition of H2 to (Y-PCP)-
Ir(CO) [FR ) -1.64(6) and FI ) -0.45(11)], where the
resonance effect is much greater than the inductive effect (by a
factor of >5 over the full range of substituents).
In our study of H2 addition to Vaska’s complex, we
considered several factors to explain the observation that the
lighter halides disfavored H2 addition.21 There appeared to be
an initial-state effect, wherein the better π donors such as X )
F stabilized the square-planar complex via donation into the
empty Ir(pz) orbital. An apparently more important effect was
manifest in the product: repulsive interactions (electrostatic and/
or covalent) between the lone pairs on X and the additional
electrons resulting from H2 addition. The present observations
support the explanation of repulsive interactions between the
ancillary ligand (X or PCP) and the electron density contributed
by the addendum (hydride and hydrocarbyl ligands): increased
π electron density from either PCP or the adding aryl group is
clearly indicated to result in a less favorable addition reaction.
This cannot be explained solely in terms of either initial-state
effects or charge transfer.
a Reaction energies, in kilocalories per mole, were calculated from eqs
13 and 15. b [M] ) (Y-PCP)Ir.
H2 Addition to (Y-PCP)Ir(H)2. Although addition to Ir(I)
is far more common, many examples are known of H2 addition
to Ir(III). Addition of H2 to complexes of the type IrL2X(H)2,
where X is a halide, has received the most attention;14,16-19 the
products have been shown to be dihydrogen complexes, viz.,
IrL2X(H)2(H2). The addition has been studied theoretically and
experimentally. The lighter halides have been shown to disfavor
coordination of H2, and this has been proposed to be due to
increased X f M π bonding14,16 although, as discussed above
in the context of varying X in species IrL2X(CO), it is difficult
to conclusively rule out other factors when the halide ligand is
varied.
H2 can also add to Ir(III) to give Ir(V) complexes. (PiPr3)2-
IrH5 was shown by neutron diffraction to be a classical
pentahydride61 and has been shown to undergo reversible
elimination of H2.62 (PCP)IrH4 was reported by Kaska and
Jensen and co-workers35 and, on the basis of T1min values, was
proposed to be a classical tetrahydride.63,64 Our calculations on
this species indicate that the tetrahydride is indeed the most
stable isomer, though the dihydrogen complexes cis- and trans-
(PCP)Ir(H)2(H2) are calculated to be only slightly higher in
energy. Further investigation of the dihydrogen/dihydride
equilibria is beyond the scope of this paper; herein we focus
on the full oxidative addition:
(Y-PCP)Ir(H)2 + H2 f (Y-PCP)Ir(H)4
(13)
The results of calculations of eq 13 are shown in Table 7. It is
found that increased electron donation by Y strongly disfaVors
the addition. Dual-parameter analysis indicates that π donation
is the dominant factor [FR ) -3.0(2); FI ) -1.2(3)]. Y ) Li
and Y ) F show similar reaction energies, an indication that
increased σ donation probably does not play a significantly
disfavorable role. The π donation effect is presumably due in
part to stabilization of the coordinatively unsaturated (Y-PCP)-
Ir(H)2 complex [as proposed for addition to IrL2X(H)213]. There
is also structural evidence for destabilizing effects in the product;
this is probably best described in terms of repulsive interactions
between the PCP π orbitals and the additional electron density
contributed by H2 to the conjugating (formally empty) orbital
Minimizing electron-electron repulsion is apparently the
dominant driving force within the π system. Sigma withdrawal
by substituents on the adding arene (Z) is strongly favorable;
however, the relative importance of σ effects is much less
pronounced for the PCP substituents (Y). While electron-
electron repulsion probably also plays a role in the σ system,
detailed analysis of the results suggests that there is an additional
component to the effects, one that is indicative of a true
oxidative addition. Such a charge-transfer component would
partially offset the unfavorable effect of increased electron
donation (resulting in increased Pauli repulsion) in the case of
the PCP substituent, whereas it would enhance the already
favorable effect of an electron-withdrawing group on the arene
undergoing addition. Hence the reaction parameters for eq 12
(especially FI) are much greater for Z than for Y.
(61) Garlaschelli, L.; Khan, S. I.; Bau, R.; Longoni, G.; Koetzle, T. F. J. Am.
Chem. Soc. 1985, 107, 7212-13.
(62) Goldman, A. S.; Halpern, J. J. Am. Chem. Soc. 1987, 109, 7537-7539.
(63) Gupta, M. Ph.D. Thesis, University of Hawaii, 1997.
(64) McLoughlin, M. A.; Flesher, R. J.; Kaska, W. C.; Mayer, H. A.
Organometallics 1994, 13, 3816-3822.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10805

A R T I C L E S
Krogh-Jespersen et al.
in the plane perpendicular to the P-Ir-P axis.13 The evidence
for this is presented in more detail in the Geometries section
below.
CO Addition to (Y-PCP)Ir. Addition of CO to the parent
three-coordinate fragment (H-PCP)Ir is calculated to be exoergic
by 59.2 kcal/mol (eq 14, Y ) H):
(calculations shown in Table 6). The reaction parameters are
markedly similar to those found for addition of H2 to the same
complex: FSP ) -1.5(3); FR ) -1.9(1); FI ) -0.6(2) (cf. -1.3,
-1.6, and -0.4, respectively, for H2 addition66). Presumably
these results are closely related; most obviously in terms of
initial-state stabilization and likely also in terms of filled-filled
orbital repulsions in the products. One difference of note is the
effect when Y ) Li: for H2 addition, Li and F have very similar
effects, but for CO addition, Li is significantly more favorable.
This indicates that increased σ donation favors addition of CO
relative to H2; perhaps this is due to the absence of a ligand
trans to PCP in the CO adduct.
(Y-PCP)Ir + CO f (Y-PCP)Ir(CO)
(14)
This is 25 kcal/mol less than the calculated value for addition
of CO to Ir(PMe3)2Cl (84 kcal/mol); this large difference is
similar to that for H2, which is computed to add to Ir(PMe3)2Cl
33 kcal/mol more favorably than to (H-PCP)Ir.65 As might be
expected, increasing electron richness on the PCP ligand favors
the thermodynamics of CO addition, presumably due to
improved Ir(d)-CO(π*) back-bonding. The magnitude of the
effect (see Supporting Information) is less than that for H2
addition to (Y-PCP)Ir, but the ordering of relative energies is
essentially the same as observed for H2 addition (Table 3). This
trend is also reflected by reaction parameters [FR ) 2.9(5); FI
) 2.1(10)] that are both ca. 70% of the corresponding values
for H2 addition (4.4 and 2.7). As with H2, this indicates a greater
overall energy range for resonance effects than for inductive
effects (2.9 vs 1.7 kcal/mol). Also, as with H2, consideration of
the nonparametrized substituents such as (rotated) NH2, BH2,
NO2, or Li (see Supporting Information) suggests that σ donation
is even less significant than indicated by the dual-parameter
LFER analysis.
The addition of CO to (Y-PCP)Ir(Ph)(H), eq 17, is formally
analogous to CO addition to the dihydride, eq 15:
(Y-PCP)Ir(Ph)(H) + CO f cis-(Y-PCP)Ir(Ph)(H)(CO)
(17)
However, the phenyl hydrides possess a square-pyramidal
geometry (H apical) rather than the distorted trigonal-bipyra-
midal (tbp) geometry of (PCP)Ir(H)2. Since the distorted tbp
structure has been shown to be particularly well stabilized by
π donation,13 the inverse correlation of Ir-CO bond dissociation
energy and electron richness might not be expected in the case
of eq 17. Indeed, there is no clear correlation with the electron-
donating ability of Y-PCP. The reaction parameters are FR )
-0.7(1) and FI ) +1.0(1). These values are very small and the
only examples in this work where FR and FI are of opposite
sign. Thus, we give them little weight; however, the absence
of a significant correlation with electron-donating ability does
at least offer one additional example contradicting the notion
that increasing electron richness should necessarily result in
increased M-CO bond strengths.67-69
CO Addition to (Y-PCP)Ir(H)2. Like the oxidative addition
of H2 or C-H bonds, the addition of CO to late transition metal
centers is generally assumed to be favored by increased electron
richness at the metal center.3,4 The above example of CO
addition to (Y-PCP)Ir (eq 14) conforms to this expectation. In
contrast to addition to (PCP)Ir, however, addition of CO to
(PCP)Ir(H)2
Overview of R-H and CO Addition to (PCP)Ir: Geom-
etries and Structural Trends. Experimentally, X-ray crystal
structures have been determined previously for two of the
(Y-PCP)Ir(H)2 + CO f (Y-PCP)Ir(CO)(H)2 (15)
8b
complexes calculated in this work: (PCP)IrH2 and (PCP)Ir-
is found to be disfaVored by increased electron-donating ability
(CO)36 [both with bis(tert-butyl)phosphino groups]. Calculated
iridium-ligand bond distances and angles are in excellent
agreement with the crystallographic values (within 0.01 Å and
1°, respectively).70
of the Y-PCP ligand (Table 7); the reaction parameters [Fsp
)
-2.6(4); FR ) -3.1(2); FI ) -1.1(3)]66 are of the same
magnitudesbut opposite directionsas those for CO addition
to (PCP)Ir. This result is closely related to the inverse
dependence on electron-donating ability of Y-PCP found for
the “converse” reaction: addition of H2 to (PCP)Ir(CO) (eq 11)
to give the same product as in eq 15. (Indeed, these are not
fully independent results: from the calculations of eqs 4, 11,
and 14, a thermodynamic cycle can be completed: eq 15 )
eq 11 + eq 14 - eq 4.)
The calculated lengths of the Ir-H and Ir-C bonds that are
formed upon R-H and CO addition to (Y-PCP)Ir show no
significant variations as functions of the substituent Y. However,
the (Y-PCP)C-Ir bond lengths show distinct variations that are
small but clearly systematic (Table 8).47,71 The correlation
between bond lengths and σsp can be parametrized according
to eq 18, where δsp is the correlation factor (in units of
CO Addition to (Y-PCP)Ir(CO). As discussed above, an
inVerse correlation between electron donation and M-CO bond
dissociation energy is found for (Y-PCP)Ir(CO)(H)2. Is this an
anomalous result? It would appear not to be: addition of CO
to (Y-PCP)Ir(CO) is also calculated to show such an inverse
correlation:
(67) Results concerning these six-coordinate d6 carbonyls are no doubt closely
related to reports, both computational (ref 68) and experimental (ref 69),
demonstrating that M-CO bond dissociation energies in complexes of the
form [M(CO)5X-] (M ) Cr, Mo, W) vary inversely with the π-donating
ability of X.
(68) Macgregor, S. A.; MacQueen, D. Inorg. Chem. 1999, 38, 4868-4876.
(69) (a) Graham, J. R.; Angelici, R. J. Inorg. Chem. 1967, 6, 2082. (b)
Darensbourg, D. J. AdV. Organomet. Chem. 1982, 21, 113.
(70) Crystallographic and calculated values, respectively, are as follows for
(PCP)IrH2 (ref 8b): C-Ir, 2.124(13) vs 2.120; P-Ir, 2.308(2) vs 2.304;
C-Ir-P, 82.41(6) vs 81.8. For (PCP)Ir(CO) (ref 36) (molecule 1 of the
two molecules in the unit cell): (PCP)C-Ir, 2.102(8) vs 2.106; (CO)C-
Ir, 1.873(10) vs 1.878; P-Ir, 2.298(2) vs 2.314; (PCP)C-Ir-P, 81.800(2)
vs 80.7.
(Y-PCP)Ir(CO) + CO f (Y-PCP)Ir(CO)2
(16)
(65) Rosini, G. P.; Liu, F.; Krogh-Jespersen, K.; Goldman, A. S.; Li, C.; Nolan,
S. P. J. Am. Chem. Soc. 1998, 120, 9256-9266.
(66) Values given are for formation of the trans dihydride; values for the cis
isomer are equal within the statistical margin (see Table 7).
9
10806 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
Table 8. Computed CPCP-Ir Bond Lengthsa and Correlation
with Hammett Parameter (σsp) for (Y-PCP)Ir and Its H2 and CO
Adducts71
b
Y
M
MH
2
M(CO)
MH
4
M(CO)2
M(CO)H
2
H
F
1.979 2.120
1.980 2.117
1.983 2.112
1.967 2.121
1.960 2.120
1.985 2.108
1.985 2.116
1.962 2.123
2.106
2.104
2.104
2.103
2.098
2.102
2.106
2.099
2.142
2.142
2.141
2.137
2.131
2.145
2.147
2.133
2.172
2.173
2.176
2.160
2.149
2.178
2.180
2.151
2.124
2.123
2.125
2.115
2.108
2.126
2.129
2.111
OCH3
CO2CH3
NO2
NH2
Li
Figure 1. Schematic illustration of the interaction between the CPCP p(π)
orbital and the filled dxy orbital of (PCP)Ir, indicating that increased π
donation by the PCP ring (or decreased π withdrawal) is energetically
unfaVorable. Thus the most π-electron-rich species, (H2N-PCP)Ir, has the
longest CPCP-Ir bond. Orbital shading indicates occupancy. Orientation and
axes are indicated in the box.
BH2
-1.8(3) 0.9(2) -0.2(2) -0.9(2) -2.0(4) -1.3(3)
c
δsp
a Bond lengths are given in angstroms. b [M] ) (Y-PCP)Ir. c Correlations
are given in picometers.
picometers):
d(CPCP-Ir) ) D + σspδsp
(18)
Figure 2. Schematic illustration of the interaction between the CPCP p(π)
orbital and the empty (PCP)Ir dxy orbital of (PCP)IrH2, indicating that
increased π donation by the PCP ring (or decreased π withdrawal) is
energetically faVorable. Thus the most π-electron-rich derivative, (H2N-
PCP)IrH2, has the shortest CPCP-Ir bond.
For the parent 14-electron complex, CPCP-Ir bond lengths
increase with increasing electron richness [δsp ) -1.8(3)]. For
example, the CPCP-Ir bond length of electron-rich (NH2-PCP)-
Ir is 0.025 Å greater than that of (NO2-PCP)Ir (Table 8).
Addition of CO to give (Y-PCP)Ir(CO) results in a much smaller
range of CPCP-Ir bond lengths among the various complexes
[with low statistical significance; δsp ) -0.2(2)]; presumably,
CPCP f Ir π donation becomes more favorable (or less
unfavorable) due to decreased electron density in the dxy orbital
(see Figure 3 below for definition of axes), which results from
Ir f CO back-bonding. Addition of H2 to give (Y-PCP)Ir(H)2
π* orbitals; it is difficult to choose between these two
possibilities, but the implications for the addition reactions are
the same and the difference does not have a bearing on our
major conclusions. For the sake of simplicity we frame the
following discussion, and Figures 1-3, in terms of interactions
with the filled PCP π orbitals.)
The unfavorable interaction indicated in Figure 1 is mitigated
by the addition of CO, owing to Ir-CO π donation; accordingly,
the magnitude of δsp decreases substantially [from -1.8(3) to
-0.2(2)]. Even more pronounced is the effect of H2 addition,
which results in a formally unoccupied dxy orbital. (Note that
this does not necessarily imply that the metal is “oxidized”;
donation by the “hydrides” into other Ir orbitals must also be
accounted for, but the occupancy of the dxy orbital is reduced.)
Increased π donation by CPCP is now favorable, and the
correlation with electron-donating ability of Y is inverse: δsp
) 0.9(2). This interaction (see Figure 2) is strictly analogous
to that proposed in seminal work by Eisenstein and co-workers13
on ML2XH2 systems and is related in a more general sense to
work by Caulton and others on the stabilization of unsaturated
complexes by X f M π bonding.15,72
Addition of CO to (PCP)IrH2 effects another reversal of the
sign of δsp as the d-orbital occupancy is rearranged so that the
PCP p(π) orbitals again interact with a (formally) filled dxy
orbital. In addition to the occupied dxy orbital, the CPCP p(π)
orbital can interact, unfavorably, with the hydride orbitals and/
or with the Ir(py) orbital (partially occupied due to donation by
the hydrides). This is indicated in Figure 3.
results in an inVerse correlation with electron richness: δsp
)
0.9(2), presumably reflecting CPCP f Ir π donation into the
“empty” dxy orbital in the equatorial plane. If either CO or H2
is added to (Y-PCP)Ir(H)2, the correlation reverses back to give
negative values of δsp: -1.3(3) and -0.9(4) for (Y-PCP)Ir(H)4
and trans-(Y-PCP)Ir(CO)(H)2, respectively. Likewise, addition
of CO to (Y-PCP)Ir(CO), to give (Y-PCP)Ir(CO)2, also results
in longer (PCP)C-Ir bonds for electron-richer derivatives: δsp
) -2.0(2).
The directions and magnitudes of these correlations with
CPCP-Ir bond lengths, and particularly the differences between
the various complexes [e.g., the value of δsp for (Y-PCP)IrH2
as compared with that for (Y-PCP)Ir(CO)H2], may be easily
rationalized in terms of simple molecular orbital (MO) interac-
tions. In the case of the 14-electron parent complex, (Y-PCP)-
Ir, there is a d8 configuration and a filled d orbital of the correct
symmetry to interact with the PCP p(π) orbitals. Accordingly,
CPCP-Ir distances are longer for the more electron-rich species
such as (H2N-PCP)Ir, reflecting a greater filled-filled repulsion.
This is illustrated in Figure 1. (Alternatively the trend may
reflect a diminished favorable interaction with the empty PCP
(71) The P values for the regression analysis of the correlation of the CPCP-Ir
bond length with the single-parameter σ, σsp, range from 0.3% to 2.1% as
follows (percent values): (Y-PCP)Ir, 0.3; (Y-PCP)IrH2, 2.1; (Y-PCP)IrH4,
1.0; (Y-PCP)IrH2(CO), 0.9; (Y-PCP)Ir(CO)2, 0.6. Further support for the
significance of these correlations is obtained with substituents Y ) Li and
Y ) BH2, for which σsp values are unavailable. For example, for all
complexes that give an inverse correlation of σsp with CPCP-Ir bond length
(negative δsp), the BH2 derivative gives the second shortest CPCP-Ir bond
length (after Y ) NO2). In addition the correlations found between the
different complexes give P values that are essentially zero; for example,
for the 18-electron species (Y-PCP)IrH4, (Y-PCP)IrH2(CO), and (Y-PCP)-
Ir(CO)2, the respective P values are 1.6 × 10-5, 3.5 × 10-5, and 5.8 ×
10-5. Note that the geometries for each set of derivatives are obtained
independently.
Thus, CPCP-Ir bond lengths are calculated to be consistently
greater for the more electron-rich derivatives of all the 18-
electron complexes. This suggests that the additions to the
corresponding 16-electron complexes are not less favorable
merely because of stabilization of the 16-electron reactants.
Apparently, increased π donation from the PCP ligand also
results in destabilization of all the calculated 18-electron
(72) Bryndza, H. E.; Domaille, P. J.; Paciello, R. A.; Bercaw, J. E. Organo-
metallics 1989, 8, 379-385.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10807

A R T I C L E S
Krogh-Jespersen et al.
Figure 3. Schematic illustration of the interaction between the CPCP p(π)
orbital and the filled (PCP)Ir dxy orbital of (PCP)Ir(CO)H2, indicating that
increased π donation by the PCP ring (or decreased π withdrawal) is
energetically unfaVorable. Thus the most π-electron-rich derivative, (H2N-
PCP)Ir(CO)H2, has the longest CPCP-Ir bond.
Figure 4. Effect of increased π donation by Y-PCP (or decreased π
withdrawal) on the thermodynamic relationships between the H2/CO adducts
of (Y-PCP)Ir.
the addition reaction; this is completely inconsistent with the
substituent effects being determined primarily by charge-transfer
components. Instead, these results can all be explained in terms
of addition of R-H or CO to (PCP)Ir resulting in decreased
electron density in one of the d orbitals in the xy plane, a process
that may be favored by increased π donation from either the
PCP carbon or the vertically added phenyl.
Computationally (and experimentally for H2 addition), in-
creased electron-donating ability of the para substituent Y is
found to disfavor R-H or CO addition to (Y-PCP)Ir(CO) (Table
6). The H2 result in particular is somewhat ironic since (PCP)-
Ir(CO) is a derivative of Vaska’s complex, the system that was
the basis for the original characterization of H2 addition as an
“oxidative” process.55,56 However, the observed substituent effect
does not imply the converse, i.e., that the process is a “reductive
addition”; increased electron donation from the substituent Z
disfavors addition of the p-C-H bond of ZC6H5 with an even
stronger substituent effect than found for Y in Y-PCP. The trend
is best explained in terms of additional Pauli (filled-filled)
repulsions operative in the 18-electron products (point iv, above)
as compared with the 16-electron reactants (points ii and iii);
thus increased electron donation from either ancillary ligand
(PCP) or addendum (arene) is unfavorable.
Increased electron-donating ability of the para substituent Y
is found to disfaVor H-H or CO addition to (Y-PCP)Ir(H)2
(Figure 4; Table 7). In the case of CO addition, there is
widespread agreement that metal-π* back-bonding plays a
more important role than CO-to-metal σ donation in determining
bond dissociation energies (at least in the case of anionic and
neutral metal carbonyls);73 increased electron richness is,
accordingly, widely believed to generally favor increased
M-CO bond dissociation energies.3,4 However, as with H2
addition to (Y-PCP)Ir(CO), the interactions described above
(points ii and iv) are apparently dominant.
The fact that the direction of the observed trends for C-H
and H-H addition can be successfully explained without
invoking oxidative/reductive components (i.e., charge transfer
or polarity) does not eliminate the possibility that such factors
do contribute to the reaction energies. Indeed, close examination
of the complete body of reaction parameters seems to suggest
that the C-H additions (and therefore, presumably, H2 additions
as well) are in fact somewhat oxidative. In particular, the very
negative values of FI derived for all three additions of
p-ZC6H4-H in which the Z is varied [vertical and horizontal
addition to (PCP)Ir and addition to (PCP)Ir(CO)] would seem
products [including (PCP)IrH4, (PCP)Ir(CO)(R)(H) and (PCP)-
Ir(CO)2] due to the effect of increased repulsive interactions
between filled orbitals.
Overview of R-H and CO Addition to (PCP)Ir: Ther-
modynamics. Using (PCP)Ir as the core metal fragment, we
have examined substituent effects for the addition of small
molecules (hydrocarbons, H2, and CO) to late-metal complexes
with several prototypical configurations including d8 three-
coordinate, d8 four-coordinate, and d6 five-coordinate. Five
classes of complexes resulting from the addition of R-H (R )
H or hydrocarbyl) and/or CO to (PCP)Ir have been investi-
gated: (PCP)Ir(R)(H), (PCP)Ir(CO), (PCP)Ir(CO)2, (PCP)Ir-
(CO)(R)(H), and (PCP)Ir(H)4.
The substituent effects on the thermodynamic relationships
between these species can be fully rationalized in terms of
fundamental MO interactionssbut not in terms of electron
richness or oxidation/reduction. Key considerations are as
follows:
(i) The parent complex, (PCP)Ir, is destabilized by increased
π donation from the PCP π system (Figure 1).
(ii) The 16-electron distorted tbp complex, (PCP)Ir(H)2, is
strongly stabilized by π bonding from the PCP π system (Figure
2).13
(iii) The 16-electron (PCP)Ir(CO) complex and quasi-square-
pyramidal (PCP)Ir(aryl)(H) are also stabilized by π donation
from PCP, though probably to a lesser extent than (PCP)Ir-
(H)2.13
(iv) The 18-electron species (PCP)Ir(CO)(H)2, (PCP)Ir(CO)-
(R)(H), and (PCP)Ir(CO)2 are destabilized by increasing π
donation from PCP; in all cases this likely results from filled-
filled interaction of the PCP π orbital with increased electron
density in the Ir dxy and/or py orbitals.
The thermodynamic relationships indicated in Figure 4 all
readily follow from the simple considerations noted in points
i-iv.
As a result of the (relative) stabilizing/destabilizing effects
noted in points i-iii, addition of R-H or CO to 14-electron
three-coordinate (Y-PCP)Ir is significantly faVored by increased
electron-donating ability of Y; this appears to suggest that such
additions are “oxidative” (Tables 2-4). Furthermore, the
addition of p-ZC6H4-H to (PCP)Ir is favored by electron-
withdrawing Z groups, when the arene is added perpendicularly
to the PCP plane (Table 5); this is also consistent with C-H
addition being an “oxidative” process. However, if the phenyl
group is added so that it is coplanar with the PCP ligand, a
strongly π-donating group (Z) on the arene (p-amino) favors
(73) Davidson, E. R.; Kunze, K. L.; Machado, F. B. C.; Chakravorty, S. J. Acc.
Chem. Res. 1993, 26, 628-635.
9
10808 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Thermodynamics of Small-Molecule Addition to Iridium
A R T I C L E S
to suggest as much. However, any such oxidative component
is apparently not dominant; for example, in the case of addition
to (Y-PCP)Ir(CO) it is not sufficiently strong to offset the
increased electron-electron repulsion resulting from C-H
addition, and thus increasing electron donation from Y disfavors
addition of C6H6 to (Y-PCP)Ir(CO).
of either CO or RH to the 14-electron species (Y-PCP)Ir is
favored by increased electron donation from the para substituent
Y. However, further addition of either CO or RH to the resulting
16-electron complexes, (Y-PCP)IrRH or (Y-PCP)Ir(CO), is
generally disfaVored by increased electron donation by Y. The
trends for the energies of CO addition, H-H addition, and C-H
bond addition to the (Y-PCP)Ir complexes are all found to be
very similar. All the trends elucidated, encompassing the effects
of varying both the PCP ligand as well as added arene, may be
simply explained in terms of filled-filled and filled-empty MO
interactions operative in the respective reactants and products,
rather than by any net transfer of charge to or from the addenda.
Thus, despite widespread assumptions, electron-rich ancillary
ligands do not appear to generally favor the thermodynamics
of addition of H2, hydrocarbons, or CO. Undoubtedly there are
true charge-transfer components influencing substituent effects
in the present systems and others; however, this work indicates
that such effects are not generally dominant.77,78
In varying aryl ring substituents (as in this work) and in the
case of varying halides,12-21 it seems that π effects and the
effects of specific orbital interactions are dominant. In addressing
the question of a genuinely oxidative component, it may prove
helpful to focus on other types of ligands, though complications
presented by other ligands may prove equally problematic.
Considering other systems, it is known that the tris(p-tolyl)-
phosphine and tris(p-methoxyphosphine) analogues of Vaska’s
complex [Ir(PPh3)2(CO)Cl] add H2 more favorably than the
parent, though the difference is slight (∆∆G ) 0.2 and 0.3 kcal/
mol, respectively);56 this suggests a small favorable effect of
increased σ donation. More generally, the fairly limited ex-
amples of metal-hydrogen BDEs that have been determined
(enthalpies of single H-atom addition to odd-electron complexes)
indicate dependencies on the electron-donating ability of the
ancillary ligands that range from small and positive to small
and inverse. Tilset found negligible effects in comparing several
neutral-complex couples including CpMo(CO)3H (Cp ) C5H5,
C5Me5), Co(CO)3LH [L) PPh3, P(OPh)3], and (C5H5)W(CO)2-
LH (L ) PMe3, CO);74 a comparison of LM(CO)3 complexes
(L ) Tp, Tp′, Cp; M ) Mo, W) indicated an inverse relationship
between M-H BDE and electron richness.75 The most extensive
study of M-H BDEs was conducted by Wang and Angelici76
on cationic complexes; a slight dependence on ancillary ligand
electron-donating ability was found for complexes of Fe, Mo,
Os, and W, while zero dependence was found for [CpIr(CO)-
(PR3)H]+ (varying R) or [(C5MenH5-n)Ir(cyclooctadiene)H]+.
Conclusion. Substituent effects have been investigated,
experimentally and computationally, for manifold small-
molecule addition reactions with (PCP)Ir as the core metal
fragment. The experimentally observed substituent effects are
all in excellent agreement with the computed values. Addition
Acknowledgment. Financial support by the National Science
Foundation (Grant CHE-9704304) and a computer equipment
grant (DBI-9601851-ARI) are gratefully acknowledged.
Supporting Information Available: Experimental details and
complete data for computed absolute and relative reaction
energies of all reactions investigated (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA010547T
(77) With respect to our conclusion we feel that it is important to note an
important caveat: relatiVe energies of addition are often more important
than absolute energies. Even assuming that addition of CO or R-H is not
oxidizing (or even if such additions are assumed to be reducing), increased
electron donation by ancillary ligands such as Y-PCP probably does favor
the addition of CO or R-H relatiVe to the addition of “classical” two-
electron donor ligands (e.g., amines or trialkylphosphines) that engage in
a significant net transfer of charge to the metal center (ref 78). In other
words, equilibria such as the following (e.g., L ) NR3 or PR3) will
presumably be pushed to the right by increased electron donation by
ancillary ligands: M-L + RH ) M(R)(H) + L; M-L + CO ) M(CO)
+ L.
(78) For example, Bryndza and Bercaw reported that π-donating ancillary ligands
(X) substantially accelerate the rates of dissociation of PMe3 from Cp*-
(PMe3)2RuX; presumably these kinetics correlate closely with the Ru-P
BDEs: Bryndza, H. E.; Domaille, P. J.; Paciello, R. A.; Bercaw, J. E.
Organometallics 1989, 8, 379-385.
(79) Exner, O. In AdVances in Linear Free Energy Relationships; Chapman, N.
B., Shorter, J., Eds.; Plenum Publishing Company: New York, 1972; pp
28-29.
(74) Tilset, M. P.; Vernon D. J. Am. Chem. Soc. 1989, 111, 6711-17.
(75) Skagestad, V.; Tilset, M. J. Am. Chem. Soc. 1993, 115, 5077-83.
(76) Wang, D.; Angelici, R. J. J. Am. Chem. Soc. 1996, 118, 935.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10809