Synthesis and characterization of carbazolide-based iridium PNP pincer complexes. Mechanistic and computational investigation of alkene hydrogenation: Evidence for an Ir(III)/Ir(V)/Ir(III) catalytic cycle

Article abstract of DOI:10.1021/ja501572g

New carbazolide-based iridium pincer complexes (^carbPNP) Ir(C₂H₄), 3a, and (^carbPNP)Ir(H)₂, 3b, have been prepared and characterized. The dihydride, 3b, reacts with ethylene to yield the cis-dihydride ethylene complex cis-(^carbPNP) Ir(C₂H₄)(H)₂. Under ethylene this complex reacts slowly at 70 C to yield ethane and the ethylene complex, 3a. Kinetic analysis establishes that the reaction rate is dependent on ethylene concentration and labeling studies show reversible migratory insertion to form an ethyl hydride complex prior to formation of 3a. Exposure of cis-( ^carbPNP)Ir(C₂H₄)(H)₂ to hydrogen results in very rapid formation of ethane and dihydride, 3b. DFT analysis suggests that ethane elimination from the ethyl hydride complex is assisted by ethylene through formation of (^carbPNP)Ir(H)(Et)(C₂H ₄) and by H₂ through formation of (^carbPNP) Ir(H)(Et)(H₂). Elimination of ethane from Ir(III) complex ( ^carbPNP)Ir(H)(Et)(H₂) is calculated to proceed through an Ir(V) complex (^carbPNP)Ir(H)₃(Et) which reductively eliminates ethane with a very low barrier to return to the Ir(III) dihydride, 3b. Under catalytic hydrogenation conditions (C₂H₄/H ₂), cis-(^carbPNP)Ir(C₂H₄)(H) ₂ is the catalyst resting state, and the catalysis proceeds via an Ir(III)/Ir(V)/Ir(III) cycle. This is in sharp contrast to isoelectronic (PCP)Ir systems in which hydrogenation proceeds through an Ir(III)/Ir(I)/Ir(III) cycle. The basis for this remarkable difference is discussed.

Full text of DOI:10.1021/ja501572g

Article
pubs.acs.org/JACS
Synthesis and Characterization of Carbazolide-Based Iridium PNP
Pincer Complexes. Mechanistic and Computational Investigation
of Alkene Hydrogenation: Evidence for an Ir(III)/Ir(V)/Ir(III) Catalytic
Cycle
Chen Cheng,† Bong Gon Kim,†,§ Damien Guironnet,† Maurice Brookhart,*,† Changjian Guan,‡
David Y. Wang,‡ Karsten Krogh-Jespersen,*,‡ and Alan S. Goldman*,‡
†Department of Chemistry, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, United States
‡Department of Chemistry and Chemical Biology, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08903,
United States
S
* Supporting Information
ABSTRACT: New carbazolide-based iridium pincer com-
plexes (carbPNP)Ir(C2H4), 3a, and (carbPNP)Ir(H)2, 3b, have
been prepared and characterized. The dihydride, 3b, reacts with
ethylene to yield the cis-dihydride ethylene complex cis-
(
carbPNP)Ir(C2H4)(H)2. Under ethylene this complex reacts
slowly at 70 °C to yield ethane and the ethylene complex, 3a.
Kinetic analysis establishes that the reaction rate is dependent
on ethylene concentration and labeling studies show reversible
migratory insertion to form an ethyl hydride complex prior to
formation of 3a. Exposure of cis-(carbPNP)Ir(C2H4)(H)2 to hy-
drogen results in very rapid formation of ethane and dihydride,
3b. DFT analysis suggests that ethane elimination from the ethyl hydride complex is assisted by ethylene through formation of
(
carbPNP)Ir(H)(Et)(C2H4) and by H2 through formation of (carbPNP)Ir(H)(Et)(H2). Elimination of ethane from Ir(III)
complex (carbPNP)Ir(H)(Et)(H2) is calculated to proceed through an Ir(V) complex (carbPNP)Ir(H)3(Et) which reductively
eliminates ethane with a very low barrier to return to the Ir(III) dihydride, 3b. Under catalytic hydrogenation conditions (C2H4/
H2), cis-(carbPNP)Ir(C2H4)(H)2 is the catalyst resting state, and the catalysis proceeds via an Ir(III)/Ir(V)/Ir(III) cycle. This is in
sharp contrast to isoelectronic (PCP)Ir systems in which hydrogenation proceeds through an Ir(III)/Ir(I)/Ir(III) cycle. The
basis for this remarkable difference is discussed.
form cyclooctene (COE) and t-butylethane (TBA) with high
turnover numbers at 200 °C (eq 1);11 this reaction is often
regarded as a benchmark for screening catalysts for transfer
dehydrogenation.
1. INTRODUCTION
Dehydrogenation of alkanes to alkenes using homogeneous
catalysts has received intense interest in recent years.1−4 Dehy-
drogenation reactions can be run in an “acceptorless” mode,5
but most commonly the reaction is run as a transfer dehydro-
genation where a sacrificial alkene is used as a hydrogen accep-
tor. Transfer dehydrogenation is one of the two key reactions in
the dual catalytic system used to achieve alkane metathesis, a
potentially important reaction for converting low molecular
weight hydrocarbons to higher molecular weight hydrocarbons
useful as fuels.6−8 Conversion of linear alkanes to aromatics
under homogeneous conditions has also been achieved via mul-
tiple transfer dehydrogenations coupled with electrocyclic ring
closure of intermediate trienes.9
Following this initial report, extensive investigations of
transfer dehydrogenations using the (tBuPCP)IrH2 pincer
complex and many other PCP-type derivatives have been
reported.12−17 The three most thoroughly examined frame-
works, “PCP”, “POCOP”,18,19 and “PCOP”,7 are shown in
Figure 1. Extensive screening reactions, mechanistic studies,
and DFT investigations of these systems have been
reported.14,19−24
While a number of early reported systems based on late
transition-metal complexes showed promise,10 a major break-
through was the discovery by Kaska and Jensen that the
iridium pincer complex [C6H3-2,6-(CH2P(t-Bu)2)2]IrH2,
(
tBuPCP)IrH2, catalyzes the transfer dehydrogenation reaction
Received: February 14, 2014
Published: April 18, 2014
between cyclooctane (COA) and t-butylethylene (TBE) to
© 2014 American Chemical Society
6672
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
(ΔH = −9.3 kcal/mol). In view of these results we felt it would
be of interest, in the context of transfer dehydrogenation,
to examine analogues of (PCP)Ir and (POCOP)Ir in which
the central iridium-coordinated sp2 carbon is replaced by a less
σ-donating sp2 nitrogen group. Toward this end, ligand 1,
bearing a carbazolide backbone, was targeted for synthesis
along with the iridium complexes 3a,b (Figure 2). These
ligands bear a close relationship to the Ozerov PNP ligand;27
however, based on the pKa values of the neutral ligands,28 the
central nitrogen atom in 3a,b is expected to be a weaker σ
donor than in the Ozerov complexes 2a,b.
This manuscript reports the synthesis and characterization of
the carbazolide ligand 1 and iridium complexes 3a,b. In pre-
liminary screening using the COA/TBE system, the iridium
carbazolide system exhibited quite low transfer dehydrogen-
ation activity at 200 °C without apparent decomposition. Since
alkene hydrogenation is a critical part of the transfer cycle, we
then probed these systems for simple alkene hydrogenation.
These studies have revealed unusual features of these reactions;
especially noteworthy is the observation that H2 can play a
critical role by facilitating reductive elimination of alkane from
the metal center.
Figure 1. Common PCP-type pincer complexes.
The basic mechanism established for transfer dehydrogen-
ation using the COA/TBE system is shown in Scheme 1. In the
Scheme 1. Hydrogen Transfer Between COA and TBE using
PCP- and POCOP-Iridium Pincer Complexes
2. RESULTS
2.1. Synthesis of Carbazole-Based Ligands and
Iridium Complexes. The carbazolide ligand 1′ was prepared
as described in Scheme 2. The 1,8-dibromide precursor 6 was
prepared as previously described.29 Since the bromination is
selective for the 3,6-positions on the carbazole ring, these posi-
tions were protected with methyl groups; subsequent bro-
mination occurs at the 1,8-positions to yield 6. Carboxylation
of 6 gave diacid 7, and reduction of 7 with LiAlH4 to the diol 8
and bromination using PBr3 afforded the dibromide 9. Nucleo-
philic substitution using HP(i-Pr)2 gave the carbPNP ligand 1.
Deprotonation of 1 with LiN(TMS)2 yields a lithium complex,
1′, which is not stable in solution and must be used imme-
diately. Reaction of the in situ generated lithium salt 1′ with
[(C2H4)2IrCl]2 followed by filtration and removal of solvent
and HN(TMS)2 under vacuum gave the Ir(I) ethylene complex
3a as a brown solid in 78% yield.
2.2. Screening 3a for Transfer Dehydrogenation. The
newly prepared complex 3a was screened for the benchmark
reaction of hydrogen transfer between COA and TBE. Using
the same conditions as previously reported for the (tBuPOCOP)
Ir system (0.03 mol % of 3a and a 1:1 molar ratio of COA:
TBE)14 at 200 °C gave traces of cyclooctene but overall poor
activity. Monitoring by NMR spectroscopy shows no catalyst
decomposition or catalyst deactivation (e.g., formation of a
binuclear dimer).16 Rather, surprisingly, under these conditions
of high TBE concentration the resting state of the carbazole-
based iridium complex 3 was shown to be the dihydride species
3b, suggesting that hydrogenation is turnover limiting for
case of the (tBuPCP)Ir system, at low TBE concentration the
catalyst resting state is the dihydride and the turnover-limiting
step is hydrogenation of TBE, while at high TBE concentration
the resting state is the vinyl hydride and dehydrogenation is
turnover limiting.22 For the (tBuPOCOP)Ir system, the resting
state is the alkene complex, and dehydrogenation is turnover
limiting.19
Convenient rates for transfer dehydrogenation for the PCP
and POCOP systems occur at temperatures in the range of ca.
125−200 °C. Thus, temperatures employed for alkane me-
tathesis using these catalysts must also be in this range despite
the fact that the molecular olefin metathesis cocatalysts often
decompose rapidly at these temperatures and operate more
effectively at lower temperatures.25 In part for this reason it is
desirable to develop new transfer dehydrogenation catalysts
that function at lower temperatures.
DFT calculations have shown that the thermodynamic favor-
ability of oxidative addition of nonpolar substrates, including H2
and RH, to the fragment XML2 (M = Ir, Rh) increases as the
σ-donating ability of X decreases.26 The direction of this effect
is opposite that which is generally accepted for oxidative
addition, and moreover, the magnitude is surprisingly large.
For example, addition of CH3−H to Ir(PH3)2F is calculated to
be far more exothermic (ΔH = −34.9 kcal/mol) than addition
to the much more electron-rich metal center of Ir(PH3)2CH3
Figure 2. Comparison of carbazolide and Ozerov-type anionic PNP-type pincer ligands and their iridium complexes.
6673
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
Scheme 2
transfer dehydrogenation catalyzed by complex 3. In light of
these observations, we initiated a detailed study of the olefin
hydrogenation reactions with the carbazolide system using
the simplest olefin, ethylene. These investigations are described
below.
2.3. Ethylene hydrogenation studies employing
carbazolide complexes 3a and 3b. Purging hydrogen
through a benzene solution of ethylene complex 3a at room
temperature (rt) resulted in the formation of the corresponding
iridium(III) dihydride species 3b within 10 min accompanied
by the formation of 1 equiv C2H6 (eq 2). However, when C2H4
the appearance of two nonequivalent hydride resonances, with
the hydride trans to the anionic nitrogen at a lower chemical
shift (−11.83 ppm, dt, JH,H = 5.2 Hz, JH,P = 18.2 Hz) and the
hydride trans to C2H4 at a higher chemical shift (−23.99 ppm,
dt, JH,H = 5.8 Hz, JH,P = 13.6 Hz).31
Binding of C2H4 to iridium dihydride 3b to form iridium
ethylene dihydride 3c is reversible and thermodynamically
favored (K1 ≫ 1, eq 5). The free energy barrier for exchange
of free C2H4 with 3c at −20 °C was estimated to be ∼13 kcal/
mol based on NMR line broadening of both the bound
C2H4 and Ir−H signals. A large positive ΔS⧧ value (40
12 cal·K−1·mol−1) was determined for the exchange, and
the rate of exchange was found to be independent of the
C2H4 concentration, which is consistent with a dissociative
mechanism for C2H4 exchange. At rt, the rate of C2H4
dissociation is fast on the NMR time scale, and the hydride
resonances are not observable as they are broadened into the
baseline.
was purged through a solution of 3b at rt in the absence of H2,
no hydrogenation was observed over 24 h (eq 3). Stoichio-
metric hydrogenation of C2H4 with 3b under a C2H4 atmo-
sphere (eq 4) proceeded at 70 °C with a half-life of ca. 40 min
(ΔG⧧ ∼ 26 kcal/mol)
Although ethylene dihydride complex 3c is the resting state
under C2H4 at rt, it remained to be determined whether
ethylene insertion into an Ir−H bond or reductive elimination
of ethane from the resulting iridium ethyl hydride is rate-
determining for C2H4 hydrogenation. When C2H4 was purged
through a solution of deuterium-labeled 3c-d2 at rt, protio-3c
formed (eq 6). The exchange of hydrogen for deuterium
This behavior is in dramatic contrast to that of the
tBuPOCOP iridium(III) dihydride complex, which hydrogenates
a much bulkier alkene, COE, at −70 °C.19 tBuPCP)IrH2 also
(
reacts with the bulky TBE to give TBA under relatively mild
conditions (ca. 55 °C) and with 1-alkenes rapidly even at sub-
ambient temperatures.22,30
A closer look at the reaction of iridium dihydride 3b with
C2H4 at −50 °C revealed the immediate and complete
conversion to an Ir(III) ethylene dihydride species, 3c, under
excess C2H4 (eq 5). The cis-dihydride geometry is indicated by
indicates that 3c is in equilibrium with the insertion product,
the ethyl hydride complex 3d, with K2 ≪ 1 since 3d was not
observed experimentally (eq 7). This reversible insertion of
C2H4 at rt strongly suggests that the rate-limiting step for
6674
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
a
Scheme 3. Proposed Pathways (Unimolecular and Bimolecular) For the Reaction of 3c with C2H4 To Yield 3a and C2H6
a
Rate constants determined at 75 °C.
hydrogenation of C2H4 with 3b under excess C2H4 is reductive
elimination of C2H6. The rate of H/D scrambling obtained
using C2D4 indicates a barrier to insertion of ΔH⧧ = 21.5
0.8 kcal/mol at −10 °C, with a small ΔS⧧ value of 3.3 2.8 eu,
consistent with an intramolecular rearrangement reaction.
Details of the reductive elimination of ethane from iridium
ethyl hydride 3d were investigated. Kinetic experiments re-
vealed that the rate of stoichiometric hydrogenation of C2H4
with dihydride 3b under excess C2H4 (eq 4) is dependent on
[C2H4] in the range 0.05−0.9 M at 75 °C according to eq 8.
These observations suggest that the ethyl hydride complex
3d, formed by migratory insertion of 3c, is trapped by H2 to
form a six-coordinate dihydrogen complex, 3d-H2, and that this
complex undergoes elimination of ethane at a rate much greater
than ethane elimination from the five-coordinate 3d (eq 9).
kobs = 1.3 0.13 × 10−4 s−1 + 7.2 0.28 × 10−4 M−1 s−1[C2H4]
(8)
The nonzero value of kobs in the limit as [C2H4] approaches
zero is consistent with a unimolecular pathway that proceeds
via elimination of ethane from 3d. The [C2H4]-dependent term
suggests that the major pathway under high C2H4 pressure
involves, in addition to the ethylene molecule that rapidly co-
ordinates to give 3c, a second molecule of ethylene in the tran-
sition state (TS) for reductive elimination of ethane from 3d;
this would be consistent with formation of the six-coordinate
species (carbPNP)Ir(H)(C2H5)(C2H4) (3d-C2H4) prior to ethane
elimination (Scheme 3).
In the reaction of ethylene complex 3a with H2 to form
dihydride 3b and ethane (eq 2), low-temperature NMR experi-
ments revealed the quantitative formation of ethylene dihydride
3c prior to production of any 3b and ethane. The half-life for
conversion of 3c to ethane and dihydride 3b under 1 atm H2 is
ca. 5 min at 25 °C, which is much faster than the conversion of
3c to ethane and ethylene complex 3a under ethylene. (In fact,
this rate of reaction of 3c under H2, to give 3b and ethane,
is only marginally slower than the rate of migratory inser-
tion of 3c indicated by the H/D exchange experiments.)
The dihydrogen complex 3d-H2 is not spectroscopically
detectable under 1 atm H2 and, since the rates of migratory
insertion of 3c and overall hydrogenation under 1 atm H2 are
similar, this indicates that the barrier for release of ethane
from 3d-H2 is significantly less than 21 kcal/mol (DFT studies
discussed below support this view and provide additional
mechanistic details).
Acceleration by H2 of propene and TBE hydrogenation by
the carbazole-based dihydride 3b was also observed. In the
absence of excess H2, no reaction of propene or TBE with 3b
was observed even at 100 °C (eq 10). However, when a mix-
ture of propene/H2 or TBE/H2 was introduced to 3b, rapid
catalytic hydrogenation of olefin was observed at rt (eq 11).
The catalyst resting state in each of these hydrogenations is the
dihydride complex 3b. These observations further support the
contention that alkane elimination from the iridium center in
6675
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
equilibrium lies fully to the right indicating that ΔG° < ca.
−3 kcal/mol. The calculated barrier to dissociation, ΔG⧧
=
7.9 kcal/mol, is somewhat lower than the barrier of ca. 13 kcal/mol
indicated by the experimentally determined rate of exchange
with free ethylene. These results may suggest that ligand
binding energies are understated by the computational method.
It is worth noting, as we compare associative vs dissociative
reaction pathways below, that this propensity would only lead
to a computational bias in favor of dissociative pathways.
In contrast to the kinetically facile addition of ethylene to
dihydride 3b to give the cis-dihydride ethylene complex 3c
(ΔG⧧ = 5.0 kcal/mol), ethylene addition along a symmetrical
pathway to give the trans-dihydride is calculated to have an
extremely high kinetic barrier (ΔG⧧ > 40 kcal/mol) as well as
somewhat unfavorable thermodynamics (ΔG°calc = +9.1 kcal/
mol).33 In agreement with these calculated values, the trans-
dihydride was never observed experimentally in the course of
this work.
the carbazolide system is facilitated by H2. The overall catalytic
cycle is shown in Figure 3. Although we could not observe
Locating a plausible TS for the seemingly simple insertion
reaction of the coordinated ethylene of 3c into the Ir−H bond
turned out to be computationally intricate. The product of
olefin insertion, ethyl hydride 3d, is a pentacoordinate metal d6
species and hence susceptible to pseudo-second-order Jahn−
Teller effects; accordingly, several square-pyramidal (SQP) or
trigonal-bipyramidal (TBP) structures may exist for 3d.34 The
lowest energy conformer of 3d is SQP with ethyl apical and
the carbazolide nitrogen and hydride ligands oriented trans
to each other (>N−Ir−H = 177.7°, see Scheme 4); this is a TC
Figure 3. Proposed catalytic cycle for hydrogenation of propene and
TBE with 3b.
stable alkene dihydride complexes (analogous to the ethylene
dihydride complex 3c) in the reactions of 3b with excess pro-
pene or TBE, deuterium incorporation into the hydride posi-
tions was observed when propene-d6 was purged through a
solution of dihydride 3b at rt, showing that reversible insertion
of propene takes place even at rt (eq 12).
a
Scheme 4. Species 3c, TS-3c-3d, and 3d
a
Metrics are shown (bond lengths in Å) for the coordination sphere
excluding the coordinating P atoms (i.e., for the ligands in the
approximate plane bisecting the P−Ir−P axis).
structure in the notation of Eisenstein and Pelissier.35 The TC
conformer of 3d is 6.1 kcal/mol above 3c in free energy; other
SQP structures located, namely TH and TN, are found to be 9.7
and 35.5 kcal/mol above 3c, respectively. A TBP structure,
denoted YN (>H−Ir−C = 75.9°), is 8.1 kcal/mol above 3c.
Some geometrical parameters pertaining to the lowest energy
conformers of 3c, 3d, and the TS of lowest energy connecting
them, TS-3c/3d, are shown in Scheme 4. In TS-3c/3d the C−
H bond is nearly fully formed (d(C−H) = 1.137 Å), and the
Ir−H distance (d(Ir−H) = 2.28 Å) remains well below the sum
of the van der Waals radii of Ir and H (3.1 Å),36 indicating the
presence of an agostic interaction.37 Examination of the reac-
tion coordinate evaluated at TS-3c/3d shows predominantly
a swinging motion of the hydride ligand, i.e., a substantial in-
crease of the N−Ir−H angle when progressing toward 3d, with
a smaller component best described as loss of the apparent
agostic interaction.
2.4. DFT Studies of the Hydrogenation of Alkenes
with Iridium Carbazole-Based Catalysts. The results of
DFT calculations on the reactions of carbazolide iridium
dihydride 3b with ethylene and dihydrogen are consistent with
and quite helpful in explaining the experimental observations.
The calculations employed the recently developed M11 func-
tionals32 and valence basis sets of triple-ζ plus polarization quality;
complete computational details are provided in the Experimental
Section and in Supporting Information.
2.4.1. Reaction of Dihydride 3b with Ethylene to Afford
Ethyl Hydride 3d. In accord with experiment, the reaction of
3b with ethylene is calculated to yield the cis-dihydride ethylene
complex 3c. The addition is calculated to be exergonic, and the
kinetic barrier is calculated to be quite low, ΔG⧧ = 5.0 kcal/mol,
consistent with the immediate formation of 3c observed upon
addition of ethylene to a solution of 3b. The calculated free
energy of olefin addition, ΔG°calc = −2.9 kcal/mol, is about at
the upper limit consistent with experiment since the observed
Disturbingly, the free energy of TS-3c/3d is calculated to be
32.8 kcal/mol above that of 3c, while the experimental value,
implied by the rate of H/D exchange observed for 3c-d2, is
only 21 kcal/mol. Single point calculations (M11 optimized
6676
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
Figure 4. Calculated free energies (kcal/mol) for the hydrogenation of ethylene by 3b under 1 atm ethylene at 25 °C (favored path in blue).
Structures in brackets ({ }) are not calculated geometries; they are shown to illustrate the barriers estimated (see text) for steps that do not appear to
proceed via a TS. Computational estimates for the energy of TS-3c/3d vary greatly (see text); hence, the barrier for ethylene insertion is not shown.
geometries) at the MP2 level predict an even higher barrier
(44.0 kcal/mol), whereas CCSD calculations (with a smaller
valence basis set) yield 30.8 kcal/mol. Other commonly
used “Minnesota functionals” such as M11-L, M06-L, or M06
produced optimized ethylene insertion TS structures very
similar to TS-3c/3d, but their free energies were only 23.6,
27.7, and 25.9 kcal/mol, respectively, above 3c.38 The over-
estimation of the barrier by single-determinant based methods
may reflect that proper descriptions of the potential energy
surfaces and electronic states of d6 ML5-type species require
minimally the use of two-configuration wave functions.39
TS-3c/3d has a narrow N−Ir−C angle of 86.1° and hence
shows resemblance to a TBP YH-type structure. Whether a
multiconfiguration approach would produce a TS structure and
energy for olefin insertion significantly different from TS-3c/3d
is, however, well outside the scope of the present work.40
2.4.2. Reaction of Ethyl Hydride 3d under an Ethylene
Atmosphere. 2.4.2.1. The Unassisted (Dissociative) Pathway.
For the PCP-type (phenyl-based) pincer analogues of complex
3d, calculations indicate that reductive elimination with loss of
ethane is exergonic. For example, using the computational meth-
ods applied to the present (carbPNP)Ir system we calculate that
ΔG° for loss of ethane from (iPrPCP)Ir(ethyl)(H) is −2.5 kcal/mol
(ΔH° = +10.1 kcal/mol). In contrast, for loss of ethane from
carbPNP iridium ethyl hydride complex 3d ΔG° is predicted to
be +15.6 kcal/mol (ΔH° = 28.9 kcal/mol), representing a very
pronounced difference of 18 kcal/mol for C−H elimination/
addition. The barrier to reductive C−H coupling in 3d is calcu-
lated as ΔG⧧ = 16.6 kcal/mol (Figure 4); this coupling leads to
a C−H σ-bond complex, 3e, with free energy 22.1 kcal/mol
above ethylene dihydride complex 3c. The ethane molecule in
3e is bound quite strongly; ΔH° for dissociation to yield 3f is
11.9 kcal/mol.
complex 3e, but we can provide an estimate for the effective
free energy barrier. We assume that ΔH⧧ ≈ 11.9 kcal/mol
(equal to ΔH°, and thus presumably a lower limit since this
assumes that ΔH⧧ = 0 for the back reaction) and that ΔS⧧ for
this unimolecular, dissociative, process is in the range of 10−
20 eu. These assumptions yield a range of values for ΔG⧧ for
ethane loss from 3e at 25 °C of 5.9 to 8.9 kcal/mol or, in effect,
a TS with free energy from 28 to 31 kcal/mol above that of
resting state 3c.
At 75 °C, the temperature at which the [C2H4] dependence
was experimentally determined, the resulting rate equation
(eq 8) indicates a first-order rate constant of 1.3 × 10−4 s−1 for
3c undergoing overall ethylene insertion and ethane loss; this
rate corresponds to ΔG⧧ = 26.7 kcal/mol. At this tem-
exp
perature, the calculated free energy of ethane complex 3e is
21.7 kcal/mol above 3c (the value is 22.1 kcal/mol at 25 °C).
Again assuming that ΔH⧧ ≈ 11.9 kcal/mol for ethane loss from
3e and ΔS⧧ = 10−20 eu, the estimated free energy (relative
to 3c) of the effective TS for ethane loss is 33.6 kcal/mol
(21.7 kcal/mol +11.9 kcal/mol) minus 3.5−7.0 kcal/mol
(10−20 eu at 75 °C); this is equal to 26.6−30.1 kcal/mol, in
full agreement with the experimental value of 26.7 kcal/mol.
2.4.2.2. The Ethylene-Assisted (Associative) Pathway.
Addition of ethylene to ethyl hydride 3d, can give two isomers,
3d-cis-C2H4 and 3d-trans-C2H4, where the ethylene and ethyl
groups are mutually cis or trans, respectively, with free energies
2.2 and 4.3 kcal/mol above that of 3d plus free ethylene. A
proper TS for the addition was not located, but incremental
scans of the potential energy surfaces reveal that both additions
proceed with virtually no energy barrier. The free energy of the
TS (or the effective TS) may be estimated using the same
procedure as outlined above for dissociation of ethane from 3e
(in this case proceeding in the reverse direction). The en-
thalpy of ethylene loss from 3d-trans-C2H4 is calculated to be
11.9 kcal/mol (coincidentally equal to the value for ethane loss
We have not been able to locate a proper TS on the potential
energy surface for dissociation of ethane from the C−H σ-bond
6677
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
from 3e); if ΔS⧧ is again estimated to be in the range of 10−
20 eu, then ΔG⧧ lies in the range 5.9−8.9 kcal/mol, which
places the TS (for ethylene loss from 3d-trans-C2H4 or ethyl-
ene addition to 3d) at ∼16−19 kcal/mol above the resting state
3c plus ethylene at 25 °C (Figure 4). Elimination from 3d-cis-
C2H4 is calculated to be slightly less favorable, with a TS about
5 kcal/mol higher than that for elimination from 3d-trans-
C2H4.
H2-assisted pathway for the release of ethane from ethylene
dihydride complex 3c is even more favorable than the ethylene-
assisted pathway. The DFT calculations strongly support this
conclusion. As discussed above, the unassisted pathway (involv-
ing dissociative loss of ethane from 3d) appears to have a free
energy barrier (25 °C) of ca. 28 kcal/mol relative to resting
state 3c or 22 kcal/mol above ethyl hydride 3d. In the presence
of an H2 atmosphere, the five-coordinate unsaturated ethyl
hydride complex 3d is calculated to add H2 with no barrier on
the energy surface (just as it readily adds ethylene), to afford a
dihydrogen complex with the dihydrogen ligand trans or cis to
the ethyl group (3d-trans-(H2) or 3d-cis-(H2), respectively);
coincidentally, either species has a free energy 4.9 kcal/mol
above that of 3c (Figure 5). If ΔH⧧ ∼ ΔH° ∼ 11 kcal/mol for
the reverse reaction (loss of dihydrogen from 3d-trans-(H2) or
3d-cis-(H2)) and if ΔS⧧ is estimated to be in the range of 10−
20 eu, then the effective TS for the H2 elimination/addition has
a free energy of 10−13 kcal above 3c. This corresponds to a
free energy of 4−7 kcal/mol above that of 3d, which is fully
consistent with an approximately diffusion-controlled reaction
(e.g., a diffusion-controlled rate of ca. 1010 M−1 s−1;42 and a
concentration of 0.041 M H2, equivalent to 1 atm, corresponds
to a pseudo-first-order rate constant of 4 × 108 s−1 and ΔG⧧ =
5.7 kcal/mol at 25 °C).
Reductive elimination from six-coordinate d6 complexes is
generally not a facile process,41 and indeed ΔG⧧ for elimination
from 3d-trans-C2H4 is substantial, 16.0 kcal/mol at 25 °C, with
a TS (TS-3d/3a) that is 26.4 kcal/mol above the free energy
of 3c plus ethylene (1 atm) (Figure 4). At 75 °C, the value is
28.9 kcal/mol at 1 atm ethylene or 26.5 kcal/mol at 1 mol/L
(28 atm), in excellent agreement with the second-order rate
constant of 7.2 × 10−4 M−1 s−1 obtained at this temperature
(eq 8) which corresponds to ΔG⧧ = 25.5 kcal/mol. Metric
exp
parameters for the isomeric TS’s for ethylene-assisted ethane
elimination are shown in Scheme 5.
Scheme 5. TS for Elimination of Ethane from 3d-trans-C2H4
a
and from 3d-trans-C2H4 (TS-3d/3a)
3d-trans-(H2) readily converts to an Ir(V) species 3g (with
ΔG° = 1.0 kcal/mol) in a nearly barrierless transition. The TS
for this conversion, TS-A, is a mere 0.7 kcal/mol higher in
electronic energy, E, than 3g. Species 3g, although a minimum
on the electronic energy surface, is actually slightly higher than
TS-A in free energy and even enthalpy. It is thus questionable
whether 3g can even be characterized as a true intermediate.
Importantly, however, no matter whether it is a minimum or
a nonstationary point on the reaction coordinate, the Ir(V)
species 3g connects to Ir(III) species 3b via TS-B to lose
ethane; the barrier to this transformation is only 2.0 kcal/mol
(Figure 5). Scheme 6 illustrates the geometry of these species
along the reaction coordinate. The full Ir(V) character of 3g
is clear from the absence of any close interhydride or carbon-
hydride contacts. It can be seen that the reaction coordinate
(3d-trans-(H2) → TS-A → 3g → TS-B → 3b + C2H6) is com-
prised almost exclusively of Hb moving away from Ha combined
with movement of Hc toward the α-carbon of the ethyl group.
These motions occur concomitantly although the motion of Hb
is somewhat more pronounced at earlier points on the reaction
coordinate (cf. TS-A and 3g), while the motion of Hc is some-
what more pronounced later (cf. 3g and TS-B).
In the pathway calculated to be most favorable for the
H2-assisted C−H elimination, the hydrogen undergoing C−H
elimination (Hc in Scheme 6) is derived from the hydride
ligand of 3d, rather than from the incoming H2 molecule. A
pathway in which the ethane hydrogen is derived from the
incoming H2 was calculated, but its TS for elimination (TS-C)
has a free energy of 17.5 kcal/mol which, though not prohi-
bitively high, is significantly above that of TS-B (7.9 kcal/mol).
TS-C might be viewed as being a point on a σ-CAM (σ-complex
assisted metathesis) pathway.43 Note that although TS-C is
higher in energy than TS-B, the σ-CAM-type pathway is appar-
ently not intrinsically unfavorable. In fact, TS-C is calculated
to be slightly lower in energy than TS-D, a C−H elimination
TS which, like TS-C, connects to 3d-cis-(H2) but in TS-D the
eliminating H atom is derived from a hydride ligand of 3b.
(Thus, of the two pathways examined for H2-assisted C−H
a
Metrics are shown (bond lengths in Å) for the coordination sphere
excluding the coordinating P atoms (i.e., for the ligands in the
approximate plane bisecting the P−Ir−P axis).
Reductive elimination of ethane from 3d-trans-C2H4 is stron-
gly exergonic (ΔG° = −21.6 kcal/mol) in striking contrast with
the value obtained for elimination from five-coordinate ethyl
hydride 3d, ΔG° = +15.6 kcal/mol. The TS for the uphill re-
ductive coupling by 3d is, however, calculated to be somewhat
lower than reductive coupling by 3d-trans-C2H4. Reductive
elimination (i.e., loss of ethane), however, is significantly less
favorable from 3d. This can be explained in terms of the free
energy of reductive coupling (ΔG° = 15.0 kcal/mol) to give 3e,
combined with the thermodynamic barrier to loss of ethane
from ethane complex 3e (ΔH° = 11.9 kcal/mol). Alternatively,
and perhaps most simply, the barrier may be explained by
considering the high endothermicity of ethane loss from 3d;
ΔH° = +28.9 kcal/mol (or +36.7 kcal/mol relative to resting
state 3c). Even a total entropic contribution as great as ΔS =
+30 eu in the rate-determining step only lowers ΔG⧧ to ca.
28 kcal/mol above 3c. Consequently, it is ultimately the un-
favorable thermodynamics (specifically the high positive enthal-
py) that results in a very high kinetic barrier for the unassisted
elimination. In the case of the ethylene assisted pathway, the
overall thermodynamics of elimination (driven by ethylene
coordination) are quite favorable. Thus, while there is a fairly
high kinetic barrier to elimination from six-coordinate 3d-trans-
C2H4, in the presence of ethylene (ca. 1 atm or greater) the
associative pathway is still the more favorable one with ΔG⧧ =
26.4 kcal/mol relative to 3c, in excellent agreement with the
experimentally determined dependence of rate on the con-
centration of ethylene (eq 8).
2.4.3. Reaction of Ethyl Hydride 3d under a H2
Atmosphere. The experimental kinetics indicate that an
6678
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
Figure 5. Calculated free energies (kcal/mol) for hydrogenation of ethylene by 3b under 1 atm H2 (favored path in blue). Structures in brackets
({ }) are not calculated geometries; they are shown to illustrate the barriers estimated (see explanation in text) for steps which do not appear to
proceed via a TS.
a
Scheme 6. Species on the Reaction Pathway From 3d-trans-(H2) to Dihydride 3b
a
Metrics are shown (bond lengths in Å) for the coordination sphere excluding the coordinating P atoms (i.e., for the ligands in the approximate
plane bisecting the P−Ir−P axis).
elimination from the intermediate 3d-cis-(H2), the σ-CAM
pathway is the slightly more favorable).
thermodynamic barrier. Both of the resulting adducts are co-
ordinatively saturated 18-valence-electron complexes. Since the
ethane product is derived from the ethyl and hydride ligands
of 3d, not from the incoming dihydrogen molecule, the
H2-assisted pathway can be viewed as a displacement reaction
in analogy with the reaction with ethylene. Elimination from
the six-coordinate d6 ethylene adducts, however, has a sub-
stantial kinetic barrier, ca. 16 kcal/mol in the most favorable
case, reflecting the general and well-explained45 behavior of
such species. In contrast, upon addition of H2, the barrier to
ethane loss is quite low. This may be attributed to the ability of
the added H2 (in contrast with ethylene) to undergo oxidative
cleavage to give an Ir(V) species.
In either the reaction with ethylene or with H2, the
elimination of ethane following the addition is very exothermic.
In the case of the ethylene-assisted reaction, inspection of the
TS (TS-3d/3a shown in Scheme 7) as well as the IRC reveals
that as the ethane dissociates from the complex the geometry of
the remaining coordination sphere is severely distorted from
the square-planar geometry of the final product in which
It is interesting to note that (PONOP)Ir(CH3)(H)+ (PONOP
=2,6-bis(di-tert-butylphosphinito)pyridine), isoelectronic with
3d, adds H2 to form dihydrogen adduct (PONOP)Ir(CH3)(H)-
(η2-H2)+, which can be observed by low-temperature NMR spec-
troscopy.44 The H2 ligand is trans to the terminal hydride. Elimi-
nation of methane from (PONOP)Ir(CH3)(H)+ occurs by a
σ-CAM mechanism in which the hydrogen that is eliminated
originates from the η2-H2 ligand in analogy to the conversion of
3d-cis(H2) to 3b plus C2H6 via TS-C. The barrier to elimination
from the dihydrogen complex (PONOP)Ir(CH3)(H)(η2-H2)+ is
significantly lower (17.9 kcal/mol) than the direct elimination
of methane from methyl hydride (PONOP)Ir(CH3)(H)+, con-
sistent with DFT computations for the carbazolide analogue in
Figure 5.
The comparison of the H2-assisted path for ethane elimi-
nation with the ethylene-assisted path is informative (Figures 4
and 5; summary in Figure 6). Addition of either H2 or ethylene
to ethyl hydride 3d proceeds without any significant kinetic or
6679
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
Figure 6. Comparison of ethane elimination from 3d promoted by addition of H2 (blue) and promoted by addition of ethylene (red).
(i.e., insertion of ethylene into the Ir−H bond of 3c), and it
thus proceeds at a rate comparable to H/D exchange by 3c−d2.
Scheme 7
3. DISCUSSION AND SUMMARY
The much poorer σ-donating ability of the central nitrogen
atom (relative to the central carbon atom of iridium PCP and
iridium POCOP systems) in the iridium carbazolide system
studied here gives rise to a stronger preference for the Ir(III)
oxidation state, relative to Ir(I), for the carbazolide system.
For example, the enthalpy of ethane loss from five-coordinate
ethyl hydride 3d is calculated to be 28.9 kcal/mol vs only
10.1 kcal/mol from (iPrPCP)Ir(H)Et. The free energy of 3d is
calculated as 6.1 kcal/mol above that of the six-coordinate
ethylene cis-dihydride resting state, 3c, and thus the overall
barrier to “unassisted” release of ethane from 3c is quite high.
The high barrier to ethane elimination from 3d results in un-
usual features with respect to hydrogenation of olefins by these
complexes. The (carbPNP)Ir dihydride, 3b, readily adds ethylene
to give 3c. This is rapidly followed by olefin insertion to give
the ethyl hydride, 3d, as shown by rapid H/D exchange
between hydride ligands and ethylene hydrogens. Although
direct elimination of ethane from 3d has a very high barrier, 3d
can easily trap ethylene. The resulting six-coordinate adduct,
specifically the isomer 3d-trans-C2H4, has a significant kinetic
barrier to elimination (ΔG⧧ = 16 kcal/mol). Nevertheless the
overall calculated barrier of 26.4 kcal/mol is less than that
predicted for a direct ethane elimination pathway, and indeed
the absolute value of this barrier and the predicted dependence
on ethylene pressure is in excellent agreement with the experi-
mental rate and the kinetic dependence on ethylene pressure.
A similar “ethylene-promoted” reductive elimination from
the facially substituted cationic iridium complex L3IrH2(C2H4)+
(L = PPhMe2) has been noted by Caulton.46
ethylene is trans to nitrogen. (This reflects Hoffmann’s explana-
tion for the barrier to reductive elimination from such species.)
In contrast, in the case of the H2 adduct (see TS-B, Scheme 7),
C−H bond formation smoothly leads to formation of a d6
octahedral ethane complex with Hb trans to N and Ha trans to
ethane. This is followed by a presumably barrierless dissociation
of ethane to afford a low-energy square-pyramidal species (with
Ha apical, and Hb trans to N) which is analogous to the
microscopic reverse of the barrierless addition of H2 to five-
coordinate ethyl hydride 3d.
While 3g is calculated to be an Ir(V) intermediate complex
preceding TS-B on the reaction coordinate, one might envision
a pathway without any such energy minimum along the reac-
tion coordinate, in which ethane elimination proceeds con-
certedly with H−H oxidative cleavage. But regardless, the
barrier for the H2-assisted pathway is extremely facile because
of the low energy of species (intermediates and/or TS’s) that
are clearly Ir(V) in character; it is these species that allow a
nearly barrierless transition from the dihydrogen adduct of the
ethyl hydride (3d-trans-H2) reactant, to the ethane adduct of
the dihydride product.
Thus, the unassisted elimination of ethane has a high barrier
attributable to the thermodynamic unfavorability of three-
coordinate Ir(I) relative to Ir(III) in the case of the carbazolide−
PNP complex. The ethylene-assisted pathway has a somewhat
lower barrier (ca. 26 kcal/mol) attributable to the kinetics of
interconversion between octahedral Ir(III) and square-planar
Ir(I). The H2-assisted pathway faces neither of these difficulties.
Accordingly, the kinetics of the H2-assisted pathway are
apparently limited only by the kinetics of formation of 3d
The effect of hydrogen on lowering the overall barrier to
hydrogenation of ethylene by 3c is much more pronounced
than the effect of ethylene. DFT calculations suggest that the
ethyl hydride adds H2 to form the η2-dihydrogen complex,
3d-trans-(H2); this barrierless step is followed by rapid reduc-
tive elimination of ethane proceeding via an Ir(V) intermediate,
with an extremely low overall barrier, ΔG⧧ = 3 kcal/mol. This
6680
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
(Chemalog) and 4 Å molecular sieves. All chemicals were purchased
from commercial sources and used without purification unless other-
wise stated. [Ir(C2H4)2Cl]2 and 6 were prepared as described.29,52
NMR spectra were recorded on Bruker NMR spectrometers (AV-400,
lower barrier for reductive elimination relative to the ethylene-
promoted pathway may be explained in terms of the carba-
zoleide pincer engendering a strong preference for the Ir(III)
oxidation state vs Ir(I), the lack of any species with Ir(I) char-
acter along the reaction coordinate, and the fact that there is no
net change in the oxidation state upon release of ethane from
the dihydrogen complex, 3d-H2.
1
DRX-500, and AV-600). H and 13C{1H} NMR spectra were refer-
enced to residual solvent peaks. 31P{1H} NMR spectra were referenced
to 85% aqueous H3PO4 except for spectra obtained below 0 °C. For
species in which two trans phosphorus atoms exhibited strong
coupling, virtual triplets were observed.
Experimental and computational studies suggest that cationic
LL′Ir+ Crabtree-type hydrogenation catalysts47 can function via
Ir(III)/Ir(V)/Ir(III) cycles. Brandt et al.48 carried out compu-
tational and experimental studies on a cationic phosphanoox-
azoline iridium catalyst49 that supported an olefin hydroge-
nation cycle in which an unsaturated Ir(III) olefin dihydride
species is the catalyst resting state; hydrogenation proceeds via
addition of H2 to form an Ir(III) olefin η2-dihydrogen dihydride
complex, followed by migratory insertion and oxidative cleavage
of the dihydrogen ligand to yield an Ir(V) alkyl trihydride.
Facile elimination of alkane from this Ir(V) species followed by
coordination of olefin closes the catalytic cycle. Similar conclu-
sions were reached by Cui et al.50 in their study of cationic
Ir(III) hydrogenation catalysts employing bidentate N-hetero-
cyclic carbene-oxazoline ligands. In a study of the hydroge-
4.2. Synthesis of 3,6-Dimethyl-9H-carbazole-1,8-dicarbox-
ylic Acid (7). To a solution of 6 (2.48 g, 7.02 mmol) in diethyl ether
n
(100 mL) was added BuLi (3.0 mL, 2.5 M in hexane, 7.5 mmol) at
0 °C. After stirring for 1 h at 0 °C, Me3SiCl (0.98 mL, 7.7 mmol) was
added, and the reaction mixture was allowed to warm to 23 °C. After
stirring at this temperature for 1 h, the suspension was cooled to
t
−78 °C, and BuLi (19 mL, 1.6 M in pentane, 30 mmol) was added
slowly. The reaction mixture was stirred at 0 °C for 3 h. At −78 °C,
the reaction mixture was exposed to a CO2 atmosphere and allowed to
warm to 23 °C overnight. Hydrolysis was performed at 0 °C by the
addition of aqueous HCl (50 mL, 1 N). The organic product was
extracted with ethyl acetate, washed with aqueous NaCl (2 × 100 mL,
3 M), dried over anhydrous MgSO4, and filtered, and the solvent was
evaporated. The residue solid was washed with diethyl ether (2 ×
5 mL) to afford the product as a light-yellow solid (1.7 g, 6.0 mmol,
86% yield). 1H NMR (400 MHz, DMSO): δ 10.91 (s, 1H, NH), 8.23
(s, 2H, Ar−H), 7.85 (s, 2H, Ar−H), 2.51 (s, 6H, Ar−CH3). 13C NMR
(150 MHz, DMSO): δ 167.98, 137.88, 128.97, 128.25, 125.93, 123.41,
112.55, 20.86. MS (EI) m/z: 283.08 (calcd: 283.08).
+
nation of ethylene by (triphos)Ir(C2H4)2 Bianchini has sug-
gested that both ethylene and H2 can accelerate the reductive
elimination of ethylene based on the half-order pressure depen-
dence of both hydrogen and ethylene on the rate of hydroge-
nation.51
4.3. Synthesis of (3,6-Dimethyl-9H-carbazole-1,8-diyl)-
dimethanol (8). To a solution of 7 (1.0 g, 3.5 mmol) in dry THF
(30 mL) at −78 °C was added LiAlH4 (500 mg, 13.2 mmol). The
reaction mixture was allowed to warm to 23 °C and then heated at
70 °C for 2 h. The excess LiAlH4 was quenched with aqueous NaOH
(15%, 3 mL) under vigorous stirring, and the reaction mixture was
filtered over a layer of Celite. The solvent was evaporated, and the
residue was dissolved in diethyl ether, washed with H2O, dried over
MgSO4, and filtered, and the solvent was evaporated to afford the
product as a light yellow solid (0.77 g, 3.0 mmol, 85% yield). 1H NMR
(400 MHz, DMSO): δ 10.29 (s, 1H, NH), 7.73 (s, 2H, Ar−H), 7.17
(s, 2H, Ar−H), 5.25 (t, J = 5.65 Hz, OH), 4.85 (d, J = 5.65 Hz, CH2),
2.45 (s, 6H, Ar−CH3). 13C NMR (150 MHz, DMSO): δ 136.40,
127.32, 125.41, 125.25, 123.03, 118.84, 60.79 (CH2), 21.62 (Ar−
CH3). Anal. calcd for C16H17NO2: C, 75.27; H, 6.71; N, 5.49. Found:
C, 73.53; H, 6.46; N, 4.95.
The hydrogen-induced acceleration of olefin hydrogenation
by dihydride 3b is even more dramatic in the case of more
hindered olefins. Neither propylene nor TBE form stable
adducts with 3b and neither react with 3b to give alkane even
at temperatures up to 100 °C. Deuterium labeling shows that
olefin binding and insertion occurs, and alkyl hydrides are
readily formed reversibly at rt, but apparently even at 100 °C
the barrier to reductive elimination is too high to complete the
hydrogenation reaction. These olefins, however, are rapidly
hydrogenated at rt under hydrogen, presumably via addition of
hydrogen to the corresponding alkyl hydride complex.
Our studies of the hydrogenations explain why the carba-
zoleide complex is ineffective for catalytic transfer dehydrogen-
ation. Under transfer dehydrogenation conditions there is no
free hydrogen present, and thus the required hydrogenation
step in the cycle has an exceptionally high barrier. In compari-
son with the much more effective PCP-type pincer complexes,
the difference is ultimately attributable to the very unfavorable
thermodynamics of the Ir(III) → Ir(I) transformations, particu-
larly elimination of alkane from the five-coordinate alkyl
hydride 3d (the high barrier may also be expressed in terms of
the very unfavorable thermodynamics of transfer of hydrogen
from the Ir(III) dihydride to olefin). Notably, preference for
the higher oxidation state is in turn attributable to the much
poorer σ-donating ability of the carbazolide nitrogen relative to
that of the coordinating carbon of the PCP-type pincers. As the
carbazolide ligand biases the system strongly in favor of the
+3 vs +1 oxidation state, future work with this ligand will focus
on rhodium, since, in the case of PCP-type pincer ligands, the
Rh(III) state is not sufficiently accessible to allow an effective
catalytic cycle based on the Rh(I)/Rh(III) couple.
4.4. Synthesis of 1,8-Bis(bromomethyl)-3,6-dimethyl-9H-
carbazole (9). To a solution of 8 (0.50 g, 2.0 mmol) in dry THF
(10 mL) was added PBr3 (1.0 mL, 2.9 g, 10 mmol) at 0 °C. The
reaction mixture was stirred at 0 °C for 3 h and then poured into H2O
(100 mL) at 0 °C. The organic product was extract with ether, washed
with H2O, dried over MgSO4, and filtered, and the solvent was
evaporated to afford the product as a yellow solid (0.65 g, 1.7 mmol,
1
87% yield). H NMR (500 MHz, C6D6) δ 8.36 (s, 1H, NH), 7.62 (s,
2H, Ar−H), 6.78 (s, 2H, Ar−H), 4.26 (s, 4H, CH2), 2.29 (s, 6H, Ar−
CH3). 13C NMR (151 MHz, C6D6) δ 137.59, 129.10, 127.88, 124.94,
121.64, 119.91 (s), 31.22 (s, CH2), 21.20 (s, Ar−CH3). Anal. calcd
for C16H15Br2N: C, 50.42; H, 3.97; N, 3.68. Found: C, 51.77; H, 4.28;
N, 3.61.
4.5. Synthesis of 1,8-Bis((diisopropylphosphino)methyl)-3,6-
dimethyl-9H-carbazole (1). To a solution of 9 (0.38 g, 1.0 mmol) in
degassed CH2Cl2 (20 mL) was added HP(iPr)2 (0.26 g, 2.2 mmol) at
23 °C. The reaction mixture was stirred at 23 °C for 4 h, and the
solvent was evaporated. The residue was washed with pentane (3 ×
10 mL) and redissolved in CH2Cl2 (20 mL). To this solution was
added Et3N (1.0 mL, 0.73 g, 7.2 mmol). After stirring at 23 °C for
15 min, the volatiles were evaporated, the residue extracted with
diethyl ether, filtered, and diethyl ether was evaporated to afford the
4. EXPERIMENTAL SECTION
1
product as a light yellow solid (0.41 g, 0.90 mmol, 90% yield). H
NMR (400 MHz CDCl3): 9.04 (s, 1H, NH), 7.65 (s, 2H, Ar−H), 7.07
(s, 2H, Ar−H), 3.09 (s, 4H, CH2), 2.49 (s, 6H, Ar−CH3), 1.85−1.79
(m, 4H, CH(CH3)2), 1.15−1.07 (m, 24H, CH(CH3)2). 13C NMR
4.1. General Comments. All manipulations were carried out
using standard Schlenk, high-vacuum, and glovebox techniques. Argon
was purified by passing through columns of BASF R3-11 catalyst
6681
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
(100 MHz CDCl3): 137.70, 128.57, 127.77, 124.04, 121.82, 117.92,
26.58 (d, JCP = 20.2 Hz, CH2), 23.71 (d, JCP = 13.5 Hz, CH(CH3)2),
21.56 (s, Ar−CH3), 19.77 (s, CH(CH3)2), 19.36 (s, CH(CH3)2). 31P
NMR (162 MHz, CDCl3) δ 0.43 (s). Anal. calcd for C28H43NP2: C,
73.82; H, 9.51; N, 3.07. Found: C, 73.10; H, 8.98; N, 3.13.
and transition states. Additional computational details are also
available in Supporting Information.
ASSOCIATED CONTENT
* Supporting Information
■
S
4.6. Synthesis of 1,8-Bis((diisopropylphosphino)methyl)-3,6-
dimethylcarbazolide Iridium(I) Ethylene (3a). To a solution of 9
(0.16 g, 0.36 mmol) in benzene (10 mL) was added slowly
LiN(TMS)2 (61 mg, 0.36 mmol) at 23 °C, and the solution turned
dark red instantly. After stirring for 5 min, a solution of [(C2H4)2IrCl]2
(0.10 g, 0.36 mmol) in benzene (5 mL) was added to the reaction
mixture. After stirring for 5 min, the reaction mixture was filtered, and
the volatiles were evaporated to afford 3a as a dark-brown solid
Kinetic data, computational details, tables of thermodynamic
quantities (E, H, G, S) and optimized geometries, absolute
energies, structural depictions and .mol files of species relevant
to Figures 4 and 5. This material is available free of charge via
the Internet at http://pubs.acs.org.
1
AUTHOR INFORMATION
(0.19 g, 0.28 mmol, 78%). H NMR (600 MHz, Tol) δ 7.90 (s, 2H,
■
Ar−H), 6.83 (s, 2H, Ar−H), 2.94 (vt, JPH = 3.3 Hz, 4H, CH2), 2.58 (s,
6H, Ar−CH3), 1.87−1.79 (m, 4H, CH(CH3)2), 1.76 (vt, JPH = 5.0 Hz,
4H, C2H4), 1.04 (dd, J = 14.1, 7.1 Hz, 12H, CH(CH3)2), 0.94 (dd, J =
12.8, 6.4 Hz, 12H, CH(CH3)2). 13C NMR (151 MHz, Tol) δ 149.67
(vt, JCP = 4.2 Hz), 127.46 (vt, JCP = 3.5 Hz), 126.27 (s), 126.17 (s),
121.48 (s), 119.32 (s), 23.59 (vt, JCP = 12.9 Hz, CH(CH3)2), 21.84 (s,
Ar−CH3), 19.80 (s, CH2), 19.04 (s, CH(CH3)2), 18.09 (s,
CH(CH3)2), 16.32 (s, C2H4). 31P NMR (243 MHz, Tol) δ 30.80
(s). Anal. calcd for C30H46IrNP2: C, 53.39; H, 6.87; N, 2.08. Found: C,
53.89; H, 7.01; N, 1.81.
Corresponding Authors
mbrookhart@unc.edu
kroghjes@rutgers.edu
alan.goldman@rutgers.edu
Present Address
§Department of Chemistry Education and Research Institute of
Natural Science, Gyeongsang National University, 660−701
Jinju, Republic of Korea.
Notes
4.7. Observation of 1,8-Bis((diisopropylphosphino)methyl)-
3,6-dimethylcarbazolide Iridium(III) Dihydride (3b). Hydrogen
was purged through a solution of 3a (5 mg, 0.07 mmol) in toluene-d8
The authors declare no competing financial interest.
1
ACKNOWLEDGMENTS
(0.5 mL) for 15 min. The product formed quantitatively. H NMR
■
(600 MHz, Tol) δ 7.90 (s, 2H, Ar−H), 6.91 (s, 2H, Ar−H), 3.08 (t,
J = 3.1 Hz, 4H, CH2), 2.57 (s, 6H, Ar−CH3), 1.80−1.69 (m, 4H,
CH(CH3)2), 0.99 (dd, J = 13.9, 6.9 Hz, 12H, CH(CH3)2), 0.91 (dd,
J = 14.6, 7.2 Hz, 12H, CH(CH3)2), −28.14 (t, J = 12.7 Hz, 2H, IrH).
13C NMR (151 MHz, Tol) δ 149.25 (s), 128.44 (s), 127.08 (s), 126.59
(s), 122.75 (s), 119.76 (s), 26.30 (t, J = 16.0 Hz, CH(CH3)2), 22.57
(t, J = 11.2 Hz, CH2), 21.95 (s, Ar−CH3), 19.43 (s, CH(CH3)2), 19.10
(s, CH(CH3)2). 31P (243 MHz, Tol): 57.03 (s).
This work was supported by the NSF through the CCI Center
for Enabling New Technologies through Catalysis (CENTC)
Phase II Renewal, CHE-1205189.
REFERENCES
■
(1) Findlater, M.; Choi, J.; Goldman, A. S.; Brookhart, M. Catal. Met.
Complexes 2012, 38, 113−141.
(2) Choi, J.; MacArthur, A. H. R.; Brookhart, M.; Goldman, A. S.
Chem. Rev. 2011, 111, 1761−1779.
4.8. Observation of 1,8-Bis((diisopropylphosphino)methyl)-
3,6-dimethylcarbazolide Iridium(III) Ethylene cis-Dihydride (3c).
Ethylene was purged through a solution of 3b (6.7 mg, 0.010 mmol) in
toluene-d8 (0.5 mL) at −50 °C for 1 min, and the product formed
(3) Huang, Z.; Brookhart, M.; Goldman, A. S.; Kundu, S.; Ray, A.;
Scott, S. L.; Vicente, B. C. Adv. Synth. Catal. 2009, 351, 188−206.
(4) Morales-Morales, D. Iridium Complexes Org. Synth. 2009, 325−
344.
(5) Xu, W.; Rosini, G. P.; Gupta, M.; Jensen, C. M.; Kaska, W. C.;
Krogh-Jespersen, K.; Goldman, A. S. Chem. Commun. 1997, 2273−
2274.
1
quantitatively. H NMR (500 MHz, Tol, −50 °C) δ 7.98 (s, 2H,
Ar−H), 6.97 (s, 2H, Ar−H), 3.15−3.00 (m, 4H, CH2), 2.73 (s, 6H,
Ar−CH3), 1.88 (s, 4H, C2H4), 1.86−1.80 (m, 2H, CH(CH3)2), 1.77−
1.71 (m, 2H, CH(CH3)2), 1.00−0.88 (m, 18H, CH(CH3)2), 0.77 (dd,
J = 14.1, 7.1 Hz, 6H, CH(CH3)2), −11.83 (td, J = 18.2, 5.2 Hz, 1H,
IrH), −23.99 (td, J = 13.6, 5.8 Hz, 1H). 13C NMR (126 MHz, Tol,
−50 °C) δ 150.31 (s), 127.29 (s), 126.76 (s), 123.92 (s), 120.47 (s),
119.73 (s), 50.83 (s, C2H4), 26.81 (t, J = 12.9 Hz, CH2), 25.58 (t, J =
15.7 Hz, CH(CH3)2), 23.96 (t, J = 18.2 Hz, CH(CH3)2), 23.01
(s, CH(CH3)2), 22.55 (s, Ar−CH3), 20.64 (s, CH(CH3)2), 18.86
(s, CH(CH3)2), 16.97 (s, CH(CH3)2). 31P NMR (202 MHz, Tol,
−50 °C) δ 43.37 (bs).
(6) Goldman, A. S.; Roy, A. H.; Huang, Z.; Ahuja, R.; Schinski, W.;
Brookhart, M. Science 2006, 312, 257−261.
(7) Nawara-Hultzsch, A. J.; Hackenberg, J. D.; Punji, B.; Supplee, C.;
Emge, T. J.; Bailey, B. C.; Schrock, R. R.; Brookhart, M.; Goldman, A.
S. ACS Catal. 2013, 3, 2505−2514.
(8) Haibach, M. C.; Kundu, S.; Brookhart, M.; Goldman, A. S. Acc.
Chem. Res. 2012, 45, 947−958.
(9) Ahuja, R.; Punji, B.; Findlater, M.; Supplee, C.; Schinski, W.;
Brookhart, M.; Goldman, A. S. Nat. Chem. 2011, 3, 167−171.
(10) (a) Crabtree, R. H.; Mihelcic, J. M.; Quirk, J. M. J. Am. Chem.
Soc. 1979, 101, 7738−7739. (b) Burk, M. J.; Crabtree, R. H. J. Am.
Chem. Soc. 1987, 109, 8025−32. (c) Maguire, J. A.; Boese, W. T.;
Goldman, A. S. J. Am. Chem. Soc. 1989, 111, 7088−7093. (d) Maguire,
J. A.; Petrillo, A.; Goldman, A. S. J. Am. Chem. Soc. 1992, 114, 9492−
9498.
5. COMPUTATIONAL DETAILS
Most computational data quoted in the main text result from
DFT calculations employing the M11 exchange−correlation
functionals;32 the Ir SDD relativistic effective core potential and
associated (6s5p3d) valence basis set;53 and 6-311G(d,p) basis
sets on all other atoms (P, N, C, and H).54 The bulky iPr groups
on P were retained in the computations. Enthalpies (Ho) and
Gibbs’ free energies (Go; T = 298.15 K, P = 1 atm) were ob-
tained from the electronic potential energies (E) using standard
statistical mechanical expressions applicable to the harmonic
oscillator/rigid rotor approximations.
In the Supporting Information, we present energetic results
in the form of three tables. Tables S1 and S2 contain relative
energies pertinent to Figures 4 and 5 in the main text, respectively;
and Table S3 provides absolute energies for relevant minima
(11) Gupta, M.; Hagen, C.; Flesher, R. J.; Kaska, W. C.; Jensen, C. M.
Chem. Commun. 1996, 2083−2084.
(12) Liu, F.; Goldman, A. S. Chem. Commun. 1999, 655−656.
(13) Haenel, M. W.; Oevers, S.; Angermund, K.; Kaska, W. C.; Fan,
H.-J.; Hall, M. B. Angew. Chem., Int. Ed. 2001, 40, 3596−3600.
(14) Gottker-Schnetmann, I.; White, P.; Brookhart, M. J. Am. Chem.
̈
Soc. 2004, 126, 1804−1811.
(15) Kuklin, S. A.; Sheloumov, A. M.; Dolgushin, F. M.; Ezernitskaya,
M. G.; Peregudov, A. S.; Petrovskii, P. V.; Koridze, A. A.
Organometallics 2006, 25, 5466−5476.
6682
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683

Journal of the American Chemical Society
Article
(16) Kundu, S.; Choliy, Y.; Zhuo, G.; Ahuja, R.; Emge, T. J.;
Warmuth, R.; Brookhart, M.; Krogh-Jespersen, K.; Goldman, A. S.
Organometallics 2009, 28, 5432−5444.
ACS: Washington, DC, 2013, INOR-681. The origin of the kinetic
barrier will be discussed in detail in a forthcoming publication.
(34) Jean, Y.; Eisenstein, O. Polyhedron 1988, 7, 405−407.
(35) Riehl, J. F.; Jean, Y.; Eisenstein, O.; Pelissier, M. Organometallics
1992, 11, 729−737.
(17) Punji, B.; Emge, T. J.; Goldman, A. S. Organometallics 2010, 29,
2702−2709.
(36) The van der Waals radii of Ir and H are reported as 202 and 110
pm, respectively. http://periodic.lanl.gov/index.shtml.
(18) Morales-Morales, D.; Redon, R.; Yung, C.; Jensen, C. M. Inorg.
Chim. Acta 2004, 357, 2953−2956.
(37) (a) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983,
250, 395−408. (b) Brookhart, M.; Green, M. L. H.; Parkin, G. Proc.
Natl. Acad. Sci. U. S. A. 2007, 104, 6908−6914.
(19) Gottker-Schnetmann, I.; Brookhart, M. J. Am. Chem. Soc. 2004,
̈
126, 9330−9338.
(20) Krogh-Jespersen, K.; Czerw, M.; Kanzelberger, M.; Goldman, A.
S. J. Chem. Inf. Comput. Sci. 2001, 41, 56−63.
(21) Krogh-Jespersen, K.; Czerw, M.; Summa, N.; Renkema, K. B.;
Achord, P. D.; Goldman, A. S. J. Am. Chem. Soc. 2002, 124, 11404−
11416.
(38) The present authors are not qualified to comment with
authority on the probable causes for this significant disagreement
among the functionals. We note, however, that M11 is a range-
separated hybrid functional containing a contribution of 42.8%
Hartree−Fock (HF) exchange to the short-range interelectronic
interactions but 100% HF exchange to the long-range interactions,
whereas M11-L is a local functional with a dual-range local exchange
functional (i.e., 0% HF exchange). M06-L is also a local functional,
whereas M06 is a hybrid functional with 27% HF exchange; neither
M06-L nor M06 features range-separation. It is tempting to associate
the exceedingly large barrier produced by the M11 functionals, relative
to the barriers predicted by M11-L, M06-L, and M06 functionals and
experiment, with differences in the treatment of electron exchange and,
in particular, the large admixture of HF exchange present in M11.
(39) See, e.g. Bersuker, I. B. The Jahn-Teller Effect: Implications in
Electronic Structure Calculations. In Progress in Theoretical Chemistry
and Physics; Piecuch, P., Maruani, J., Delgado-Barrio, G., Wilson, S.,
Eds.; Springer: New York, 2009; Vol. 19; pp 343−362.
(40) The barrier predicted by M11-L (23.6 kcal/mol) is in fact very
close to the experimental value (21 kcal/mol). We note that M11-L is
intended to ‘... provide broad accuracy for both single-configurational
and multiconfigurational molecules and for solid-state lattice
constants.’ Peverati, R.; Truhlar, D. C. J. Phys. Chem. Lett. 2012, 3,
117−124.
(22) Renkema, K. B.; Kissin, Y. V.; Goldman, A. S. J. Am. Chem. Soc.
2003, 125, 7770−7771.
(23) Gottker-Schnetmann, I.; White, P. S.; Brookhart, M. Organo-
̈
metallics 2004, 23, 1766−1776.
(24) Goettker-Schnetmann, I.; Heinekey, D. M.; Brookhart, M. J. Am.
Chem. Soc. 2006, 128, 17114−17119.
(25) (a) Malcolmson, S. J.; Meek, S. J.; Sattely, E. S.; Schrock, R. R.;
Hoveyda, A. H. Nature 2008, 456, 933−937. (b) Ibrahem, I.; Yu, M.;
Schrock, R. R.; Hoveyda, A. H. J. Am. Chem. Soc. 2009, 131, 3844−
3845. (c) Flook, M. M.; Jiang, A. J.; Schrock, R. R.; Muller, P.;
Hoveyda, A. H. J. Am. Chem. Soc. 2009, 131, 7962−7963.
(26) (a) Wang, D. Y.; Choliy, Y.; Krogh-Jespersen, K.; Hartwig, J. F.;
Goldman, A. S. Abstracts of Papers. In Proceedings of the 240th ACS
National Meeting, Boston, MA, August 22−26, 2010; ACS:
Washington, DC, 2010, INOR-5. (b) Wang, D. Y.; Krogh-Jespersen,
K.; Goldman, A. S. Abstracts of Papers. In Proceedings of the 244th
ACS National Meeting and Exposition, Philadelphia, PA, August 19−
23, 2012; ACS: Washington, DC, 2012, INOR-587. (c) Wang, D. Y.
Ph. D. Thesis, Rutgers University: New Brunswick, NJ, 2012. The
origin of this effect will be discussed in detail in a forthcoming
publication.
(41) (a) Wick, D. D.; Goldberg, K. I. J. Am. Chem. Soc. 1997, 119,
10235−10236. (b) Bartlett, K. L.; Goldberg, K. I.; Borden, W. T. J.
Am. Chem. Soc. 2000, 122, 1456−1465. (c) Bartlett, K. L.; Goldberg,
K. I.; Borden, W. T. Organometallics 2001, 20, 2669−2678. (d) Fekl,
U.; Goldberg, K. I. J. Am. Chem. Soc. 2002, 124, 6804−6805.
(42) Elliot, A. J.; McCracken, D. R.; Buxton, G. V.; Wood, N. D. J.
Chem. Soc. Faraday Trans. 1990, 86, 1539−1547.
(43) Perutz, R. N.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2007, 46,
2578−2592.
(44) Campos, J.; Kundu, S.; Pahls, D. R.; Brookhart, M.; Carmona,
E.; Cundari, T. R. J. Am. Chem. Soc. 2013, 135, 1217−1220.
(45) Saillard, J.; Hoffmann, R. J. Am. Chem. Soc. 1984, 106, 2006−
2026.
(46) Lundquist, E. G.; Huffman, J. C.; Folting, K.; Caulton, K. G.
Angew. Chem., Int. Ed. Engl. 1988, 27, 1165−1167.
(47) Crabtree, R. Acc. Chem. Res. 1979, 12, 331−337.
(48) Brandt, P.; Hedberg, C.; Andersson, P. G. Chem. - Eur. J. 2003,
9, 339−347.
(49) Helmchen, G. n.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336−345.
(50) Cui, X.; Fan, Y.; Hall, M. B.; Burgess, K. Chem. - Eur. J. 2005, 11,
6859−6868.
(51) Bianchini, C.; Farnetti, E.; Graziani, M.; Kaspar, J.; Vizza, F. J.
Am. Chem. Soc. 1993, 115, 1753−1759. The authors note that other
mechanistic interpretations are possible.
(27) (a) Ozerov, O. V.; Guo, C.; Papkov, V. A.; Foxman, B. M. J. Am.
Chem. Soc. 2004, 126, 4792−4793. (b) Weng, W.; Guo, C.; Moura, C.;
Yang, L.; Foxman, B. M.; Ozerov, O. V. Organometallics 2005, 24,
3487−3499. (c) Gatard, S.; Celenligil-Cetin, R.; Guo, C.; Foxman, B.
M.; Ozerov, O. V. J. Am. Chem. Soc. 2006, 128, 2808−2809. (d) Weng,
W.; Guo, C.; Celenligil-Cetin, R.; Foxman, B. M.; Ozerov, O. V. Chem.
Commun. 2006, 197−199.
(28) The pKa values are in DMSO. Source: http://www.chem.wisc.
edu/areas/reich/pkatable/index.htm.
(29) Britovsek, G. J. P.; Gibson, V. C.; Hoarau, O. D.; Spitzmesser, S.
K.; White, A. J. P.; Williams, D. J. Inorg. Chem. 2003, 42, 3454−3465.
(30) Lee, D. W.; Kaska, W. C.; Jensen, C. M. Organometallics 1998,
17, 1−3.
(31) While the stable isomer of (carbPNP)IrH2(C2H4) is the cis-
isomer, 3c, treatment of (tBuPCP)IrH2 with ethylene at −90 °C results
in formation of the trans-dihydride (tBuPCP)IrH2(C2H4). The
increased stability of the cis-dihydride in the case of the carbazole
system is likely in part due to the decreased trans effect of the
carbazide nitrogen in 3a relative to the sp2 carbon the (PCP)Ir
complex thereby favoring the cis-dihydride structure which places
hydride trans to nitrogen, although the greater steric demand of t-Bu vs
i-Pr groups likely also plays an important role. Other iridium ethylene
dihydride complexes of which we are aware are the facially substituted
cationic L3IrH2(C2H4)+ where the hydrides are forced to be cis. See:
(a) Barbaro, P.; Bianchini, C.; Meli, A.; Peruzzini, M.; Vacca, A.; Vizza,
F. Organometallics 1991, 10, 2227−2238. (b) Garcia-Camprubi, A.;
Martin, M.; Sola, E. Inorg. Chem. 2010, 49, 10649−10657.
(32) Peverati, R.; Truhlar, D. G. J. Phys. Chem. Lett. 2011, 2, 2810−
2817.
(52) Onderdelinden, A. L.; van der Ent, A. Inorg. Chim. Acta 1972,
420−426.
(53) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123−41.
(54) (a) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem.
Phys. 1980, 72, 650. (b) McLean, A. D.; Chandler, G. S. J. Chem. Phys.
1980, 72, 5639.
(33) Biswas, S.; Zhou, T.; Wang, D. Y.; Hackenberg, J.; Nawara-
Hultzsch, A.; Schrock, R. R.; Brookhart, M.; Krogh-Jespersen, K.;
Goldman, A. S. Abstracts of Papers. In Proceedings of the 245th ACS
National Meeting and Exposition, New Orleans, LA, April 7−11, 2013;
6683
dx.doi.org/10.1021/ja501572g | J. Am. Chem. Soc. 2014, 136, 6672−6683