Interplay of weak interactions: An iridium(III) system with an agostic tert-butyl but a nonagostic isopropyl group

Article abstract of DOI:10.1021/om010802i

The solid state structure and NMR data of 2-substituted benzoquinoline Ir(III) complexes show that the pendant group G does not necessarily become agostic despite the presence of an available nearby coordination site at the Ir center. Thus for G = t-Bu the agostic structure is observed, but for G = i-Pr, the potentially agostic i-Pr methyl groups point away from the metal. ONIOM(B3PW91/UFF) calculations reproduce the experimental observations and suggest that several ligand-ligand interactions (parallel aromatic stacking and C-H?H-Ir interactions) successfully compete in energy with the agostic interaction. Fluorobenzene (PhF) was found to be a much better noncoordinating polar solvent for synthesis of the t-Bu complex than dichloromethane.

Full text of DOI:10.1021/om010802i

Organometallics 2002, 21, 575-580
575
In ter p la y of Wea k In ter a ction s: An Ir id iu m (III) System
w ith a n Agostic ter t-Bu tyl bu t a Non a gostic Isop r op yl
Gr ou p
Eric Clot,† Odile Eisenstein,*,† Tiffany Dube´,‡ J ack W. Faller,*,‡ and
Robert H. Crabtree*,‡
Laboratoire de Structure et Dynamique des Syste`mes Mole´culaires et Solides (UMR 5636),
Universite´ Montpellier 2, 34095 Montpellier Cedex 5, France, and Department of Chemistry,
Yale University, 225 Prospect Street, New Haven, Connecticut 06520-8107
Received September 6, 2001
The solid state structure and NMR data of 2-substituted benzoquinoline Ir(III) complexes
show that the pendant group G does not necessarily become agostic despite the presence of
an available nearby coordination site at the Ir center. Thus for G ) t-Bu the agostic structure
is observed, but for G ) i-Pr, the potentially agostic i-Pr methyl groups point away from the
metal. ONIOM(B3PW91/UFF) calculations reproduce the experimental observations and
suggest that several ligand-ligand interactions (parallel aromatic stacking and C-H‚‚‚H-
Ir interactions) successfully compete in energy with the agostic interaction. Fluorobenzene
(PhF) was found to be a much better noncoordinating polar solvent for synthesis of the t-Bu
complex than dichloromethane.
Metalloenzyme active sites have a variety of func-
tional groups close to the metal binding site that
modulate the reactivity of the metal site in interesting
ways. For example, the distal histidine of hemoglobin
is believed to affect the relative affinity of the metal
binding site for CO versus O2.1
We have attempted to apply these ideas to organo-
metallic systems. The ready cyclometalation of 7,8-
benzoquinoline has allowed access to a series of species
of type 1, where a labile site (X in 1a ; L′ in 1b) is located
very close to a pendant group G. The aim has been to
see how the presence of G affects binding, structure, and
reactivity at the labile site. We have described a number
of applications of pendant amino groups (1, G ) NH2).
For example, in 1a (G ) NH2; X ) F), protonation gives
an HF complex (1b, G ) NH2; L′ ) FH.) stabilized by
F-H‚‚‚N hydrogen bonding between the bound HF and
G.2a When G ) NH2 and L′ is H2 in 1b, we see reversible
proton transfer from coordinated H2 to G.2b
low affinity for external ligands, and for the G ) t-Bu
case, the product is an agostic compound with a CH
bond of the t-Bu ligand coordinated to the sixth site.
For G ) i-Pr, in contrast, the potentially agostic i-Pr
methyl groups point away from the metal, and evidence
discussed here shows this species is nonagostic.
Agostic groups3 are normally very common whenever
a vacant coordination site in a complex is located in the
vicinity of a nearby sp3 or sp2 C-H bond, so the failure
of the isopropyl to become agostic in a system where
the corresponding tert-butyl is agostic attracted our
attention. The influence of the carbon skeleton on the
choice of being agostic has already been noted.4 In these
earlier systems the observed agostic interaction was
shown to be the result of a subtle balance between an
attraction between the CH bond and the electron-
deficient metal and repulsions between various groups
within the ligands. Hybrid QM/MM methods have
proved to be very efficient in describing such situations.5
In this work ONIOM (B3PW91/UFF) calculations have
been carried out to find the origin of the different
behavior in the two groups: G ) i-Pr vs t-Bu.
Resu lts a n d Discu ssion
Syn th esis of Liga n d s a n d Com p lexes 2 a n d 3.
Benzoquinoline readily undergoes nucleophilic attack
(1) Collman, J . P.; Fu, L. Acc. Chem. Res. 1999, 32, 455.
(2) (a) Clot, E.; Eisenstein O.; Crabtree, R. H. New J . Chem. 2001,
25, 66. (b) Patel, B.; Lee, D.-H.; Rheingold A. L.; Crabtree, R. H.
Organometallics 1999, 18, 1615. Lee, D.-H.; Patel, B. P.; Clot, E.;
Eisenstein O.; Crabtree, R. H. Chem. Commun. 1999, 297.
(3) Brookhart, M.; Green, M. L. H. J . Organomet. Chem. 1983, 250,
395. Brookhart, M.; Green, M. L. H.; Wong, L. L. Prog Inorg. Chem.
1988, 36, 6: 1. Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32,
789.
Turning to the case of G ) i-Pr and t-Bu, we now
report the effects of these large G groups on the
properties of 1b. The L′ binding site in 1b now has a
(4) Cooper, A. C.; Clot, E.; Huffman, J . C.; Streib, W. E.; Maseras,
F.; Eisenstein, O.; Caulton, K. G. J . Am. Chem. Soc. 1999, 121, 97.
J affart, J .; Etienne, M.; Maseras, F.; McGrady, J . E.; Eisenstein, O. J .
Am. Chem. Soc. 2001, 123, 6000.
* Corresponding authors. E-mail: odile.eisenstein@lsd.univ-montp2.fr
(O.E.); robert.crabtree@yale.edu (R.H.C.); jack.faller@yale.edu (J .W.F.).
† Universite´ Montpellier 2.
‡ Yale University.
(5) Maseras, F. Chem. Commun. 2000, 1821.
10.1021/om010802i CCC: $22.00 © 2002 American Chemical Society
Publication on Web 01/10/2002

576 Organometallics, Vol. 21, No. 3, 2002
Clot et al.
Ta ble 1. Cr ysta llogr a p h ic Da ta for 2 a n d 3
the location of the carbon of the methyl at 2.722(4) Å
from Ir is entirely typical of an agostic system. NMR
evidence, discussed below, confirms the assignment.
2
3
formula
fw
a, Å
C
53H47Cl2F6IrNP3
C53H47Cl2F6IrNP3
1038.93
11.1421(1)
11.4068(3)
17.9943(6)
80.692(2)
81.153(2)
79.941(2)
2203.6(1)
2
1168.00
20.3262(4)
15.3816(3)
15.2066(3)
90
90
90
The isopropyl group of 2 is directed so the secondary
CH, rather than the primary CH’s, is pointing toward
the metal. The observed Ir‚‚‚H distance of 2.77 Å is
considered too long for an agostic interaction, and 2 is
therefore identified as a five-coordinate square pyrami-
dal species. As the hydrogen atoms are not located well
by X-ray crystallography and C-H distances appear
shorter (∼0.95 Å) than the true distance (∼1.10 Å), a
more realistic assessment of distances can be made by
calculating an ideal position of the H based on the
carbon positions. In a later section we discuss a com-
putational study that locates this hydrogen more se-
curely in a very similar position. The X-ray method
tends to underestimate C-H distances, and so in agostic
systems it tends to overestimate Ir‚‚‚H distances. Re-
gardless, none of the computations suggest an agostic
interaction. The only ligand that might have escaped
crystallographic detection would be a coordinated H2,
but the NMR spectroscopic study would certainly have
detected this ligand.
b, Å
c, Å
R (deg)
â (deg)
γ (deg)
V, Å3
4754.3(1)
4
Z
cryst syst
space group
orthorhombic
Pnma (#62)
1.632
30.91
-90
0.71069
0.031
triclinic
P1 bar (No. 2)
1.566
31.66
-90
0.71069
0.033
F
calcd, g cm-3
µ (Mo KR) cm-1
temperature, °C
λ (Å)
R(Fo)
Rw(Fo)
GOF
0.036
1.08
0.029
1.34
at the 2-position as a result of the polarization of the
ring CN multiple bond; the N-oxide shows even greater
reactivity. We have synthesized the new species bq-(i-
Pr) and bq-(t-Bu) from the alkyl Grignard reagents and
the benzoquinoline N-oxide by a known method.6
Cyclometalation of 2-substituted benzoquinolines with
[IrH2L2(H2O)2]PF6 normally affords an aqua complex
(1b , L′ ) H2O), the water coming from the moist
CH2Cl2 solvent. With the bulky substituted benzoquino-
lines used here, in contrast, the products were [IrH-
(bq-R)L2]PF6 (2, R ) i-Pr; 3, R ) t-Bu), lacking an aqua
group or any other external ligand.
In all other respects both structures are entirely con-
ventional with no unusual intra- or intermolecular con-
tacts. A conclusion that 2 is more stable in the non-
agostic form would not be completely justified on the
basis of X-ray data because two tautomers could be in
equilibrium in solution and the least stable might have
crystallized. For this reason, we have examined the
NMR spectroscopic data.
Sp ectr oscop y of 2 a n d 3. The case of 2 is the
simplest. The NMR spectrum shows the terminal hy-
dride as a triplet (2J HP ) 14.5 Hz.) at δ -12.47. There
is no other high-field hydride or dihydrogen resonance
visible in the temperature range used (20 to -80 °C).
The isopropyl group appears as an intense (6H) doublet
at δ 0.64 (3J HH ) 8.78 Hz), assigned to the isopropyl
CH3 group, and a multiplet of unit intensity at δ 2.67
and assigned to the isopropyl secondary CH. No signifi-
cant change is seen in the temperature range men-
tioned. The IR spectrum shows the Ir-H stretch at 2184
cm-1. Both the NMR and IR data are consistent with a
nonagostic structure because the chemical shift of the
i-Pr CH proton and the ν(CH) IR bands are normal; in
particular there is no sign of agostic ν(CH) bands at low
energy. Using a scaling between calculated ONIOM
frequencies and experimental values of 0.9205, we
predict an Ir-H stretching frequency for the agostic
isomer for G ) i-Pr at 2173 cm-1 and an agostic C-H
stretch at 2549 cm-1. While the ν(Ir-H) is found close
to that predicted, the more important ν(C-H) is not seen
in the predicted region, either in the crystal or in
solution, consistent with the absence of the agostic
tautomer in the experimental case.
Unexpectedly, for bq-t-Bu, the cyclometalation gave
very little product using the usual CH2Cl2 solvent.
Moving to PhFsrarely encountered as a reaction
solventsgave a satisfactory yield. PhF is both less
coordinating and less oxidizing than CH2Cl2, while still
being polar enough to dissolve salts such as 3.
Str u ctu r es of Com p lexes 2 a n d 3. X-ray crystal-
lographic study of the two species (Tables 1 and 2;
Figures 1 and 2) showed that the t-Bu is agostic, as one
might expect, whereas the isopropyl is not. The t-Bu
complex (3) has one of its methyl groups pointing
directly toward the metal and one of the hydrogen atoms
of this methyl is clearly bound to the sixth site at
iridium, as is typical in agostic complexes. All hydrogen
atoms in the molecule were found in the difference
Fourier and were refined. Although positions found by
X-ray determinations are not particularly accurate,
refining all hydrogens provides an internal calibration
of the variations and errors in observed C-H distances.
Even without this X-ray evidence for hydrogen positions,
For 3, the NMR spectrum again shows the terminal
hydride as a triplet (2J HP ) 14.7 Hz) but now at δ
-14.34. Again, there is no high-field hydride or dihy-
drogen resonance visible in the temperature range used.
At room temperature, the tert-butyl methyl groups
appear as a singlet at δ 0.63. On cooling there is a
decoalescence into a pair of singlets at δ 0.73 and -0.06
in a 2:1 intensity ratio that clearly correspond to agostic
and terminal methyl groups of the tert-butyl. Line shape
(6) Cappelli, A.; Anzini, M.; Vomero, S.; Mennuni, L.; Makovec, F.;
Doucet, E.; Hamon, M.; Bruni, G.; Romeo, M. R.; Menziani, M. C.; De
Benedetti, P. G.; Langer, T. J . Med. Chem. 1998, 41, 728-741.

Interplay of Weak Interactions
Organometallics, Vol. 21, No. 3, 2002 577
Ta ble 2. Metr ic Da ta for th e Str u ctu r es Discu ssed a
2x
2q
2m
2′q
2′m
3x
3q
3m
Ir‚‚‚H(nonag)
Ir‚‚‚C(nonag)
Ir‚‚‚H(ag)
Ir‚‚‚C(ag)
Ir-H(hydr)
Ir-N(bq)
Ir-C(bq)
Ir-P
2.77(6)
3.454(8)
2.648
3.438
2.731
3.500
2.015
2.741
1.622
1.946
2.040
2.323
2.322
2.912
3.189
2.030
2.853
1.579
2.168
2.015
2.379
2.348
2.100
2.358
3.674
4.903
1.122
120.0
1.98(3)
2.722(4)
1.56(4)
2.124(3)
2.004(3)
2.335(1)
2.334(1)
2.12(5)
2.46(5)
3.77(3)
4.43(3)
1.05(4)
120.0(3)
2.062
2.781
1.605
2.145
2.017
2.319
2.323
2.967
3.057
2.073
2.864
1.578
2.178
2.008
2.110
2.348
2.384
2.353
3.67
1.49(8)
2.154(6)
1.981(7)
2.324(1)
2.324(1)
2.16(7)
2.16(7)
4.02(3)
4.02(3)
1.591
2.186
2.003
2.323
2.323
2.908
2.956
1.578
2.204
2.005
2.361
2.362
2.139
2.142
3.637
3.705
Ir-P′
CH‚‚‚HIr
C′H′‚‚‚HIr
Ar‚‚‚Ar′
Ar‚‚‚Ar′′
C-H(ag)
N-C(R)-C(â)
5.02
1.120
121.4
1.131
118.0
1.131
119.4
118.3(6)
117.6
118.9
a
2, nonagostic i-Pr; 2′, agostic i-Pr; 3, agostic tert-butyl; x, experimental; q, quantum only; m , QM/MM; bq ) benzoquinolinato; ag )
agostic; hydr ) hydride; Ar ) arene centroid bq B ring; Ar′, Ar′′ ) Ph centroids PPh3 B rings.
intermediates in H/D exchange reactions, where the
barriers were predicted to be 11.7 (1,3) and 2.5 (1,1′)
kcal/mol, respectively; in the first case, this is close to
the 1,3 barrier found here.7 We ascribed the low 1,1′
barrier height to the metal remaining bound to the same
carbon throughout in the 1,1′ process but having to shift
from one carbon to another in the 1,3 process and so
losing most of the M‚‚‚alkyl binding energy in the
transition state.
Com p u ta tion a l Stu d y
In what follows, 2x refers to the experimental values,
2q to the pure quantum studies, and 2m to the QM/
MM calculations; metric data are defined in the same
way. The same labeling is used for the tert-butyl
complex, 3.
The fact that the isopropyl complex, 2x, is nonagostic
when it could apparently have become agostic was a
puzzle that led us to carry out computational work to
try to identify the factors involved. Calculations were
carried out using pure quantum mechanical (QM)
methods (DFT, B3PW91) using the model compound 2q
having PH3 ligands. In the i-Pr case, both agostic and
nonagostic situations could be modeled starting from the
i-Pr crystal structure for the nonagostic situation and
starting from the t-Bu crystal structure for the agostic
case, but replacing one terminal methyl of the t-Bu
group by H. Previous calculations have proved capable
of predicting agostic or nonagostic character in such
cases and have identified the factors that control the
outcome.4,5
F igu r e 1. ORTEP diagram of the cation in [IrBq-i-Pr-
(PPh3)2H]PF6‚CH2Cl2, 2, showing 50% probability ellip-
soids.
The optimized nonagostic structure (2q) so obtained
closely fits the experimental one (2x, Figure 1), as
shown by the comparison data of Table 2. The hydrogen
that lies closest to the metal, also present in the experi-
mental structure, is probably better located by the com-
putational work than by the X-ray method. The Ir‚‚‚H
distance is 2.648 Å, and the CH distance is 1.106 Å. This
is very different from the usual agostic situation, and
the CH bond pointing toward the metal is not signifi-
cantly elongated compared to the most appropriate
reference value of 1.102 Å calculated for the correspond-
ing unambiguously unbound secondary CH bond in the
agostic tautomer. This work confirms that the i-Pr
compound is genuinely nonagostic and five-coordinate.
F igu r e 2. ORTEP diagram of the cation in [IrBq-t-Bu-
(PPh3)2H]BF4, 3, showing 50% probability ellipsoids.
analysis gives an exchange barrier of 9.7 kcal/mol. This
exchange processsa 1,3 shift of the metal from one
methyl group to the otherscan be frozen out but the
protons of the agostic Me group remain a single sharp
peak to -90° so the corresponding 1,1′ agostic shifts
exchanging the metal between geminal CH bonds of a
single CH3 groupsis much easier. We have reported this
order of barriers in studies of Ir(I) alkane species,
(7) Ge´rard, H.; Eisenstein, O.; Lee, D-H.; Chen, J .; Crabtree, R. H.
New J . Chem. 2001, 25, 1121.

578 Organometallics, Vol. 21, No. 3, 2002
Clot et al.
Ta ble 3. Com p a r ison of Rela tive Ca lcu la ted
tautomer. This is fully consistent with the NMR and
IR spectroscopic observations suggesting that the non-
agostic tautomer is present in solution, as well as with
the X-ray structure.
En er gies for Agostic 2′q vs Non a gostic 2qa
QM
QM/MM
QM part of
compound
(kcal/mol) (kcal/mol) QM/MM (kcal/mol)
Analysis of the QM and MM contributions to the total
energy in the QM/MM calculations shows that the QM
energies are identical (within 0.2 kcal mol-1) for the two
isomers. Thus the small energetic preference for the
agostic isomer found in the pure QM calculations has
been lost. This means the preference for the nonagostic
isomer has to come from the MM contribution.
2′q (agostic)
2q (nonagostic)
-0.7
0.0
-3.2
0.0
-0.2
0.0
a
Most stable isomer indicated by negative sign in each case.
This information gives some insight into why an aqua
complex is not formed for 2. The i-Pr secondary CH
hydrogen being 2.65 Å (calcd) from the Ir prevents the
water from binding because a typical Ir-OH2 distance
range is 2.2-2.4 Å, which would put the water oxygen
far too close (<1 Å) to this H. This steric effect excludes
water from the site.
The bound primary CH of the -CH(CH3)2 group in
the calculated agostic tautomer, 2′q, has an elongated
C-H distance of 1.131 Å and a short Ir‚‚‚C distance of
2.777 Å, entirely in accord with the normal situation
found for true agostic structures. The shorter Ir-N
distance in the agostic (2.137 Å) versus the nonagostic
(2.186 Å) isomer is consistent with the presence of an
attractive Ir‚‚‚HC interaction in the agostic structure.
The agostic carbon is not in the bq plane and on the
opposite side of the bq plane from the agostic H; this is
probably the result of the C-H bond, rather than the
CH hydrogen, being the true ligand.8 The Ir-N bond is
readily distorted in the sense that its length varies
considerably both in experimental and calculated struc-
tures in this series of compounds.2 This is probably due
to an intrinsically weaker bonding of a pyridine versus,
for example, the bq aryl, but also may be a result of the
high trans influence hydride weakening this bond. Any
attractive Ir‚‚‚HC interaction would be expected to cause
the bq ligand to tilt in such a way as to bring the CH
closer to the metal. Given the relative inflexibility of
the bq Ir-C(aryl) bond, the more flexible Ir-N bond
might be expected to shorten, as indeed is observed. The
bq ligand itself is relatively rigid, as shown by the very
small changes in the bond angles, for example in the
NR-Câ-Cγ angle (see diagram 2q).
P P h 3/bq, CH‚‚‚HIr , a n d CH‚‚‚Ir In ter a ction s in
Com p lexes 2 a n d 3. Close examination of the struc-
tures of 2 and 3 suggests that the agostic interaction is
just one of a number of structural motifs that are
expected to affect the energetics of the system. These
motifs seem to have competitive energy contributions,
and the agostic interaction is obviously not dominant
since the nonagostic structure is preferred for 2. The
Ir‚‚‚HC geometry will not be described further since it
is similar to that in the QM work.
One such contribution is aromatic parallel (||) stacking
between a P-Ph substituent and the benzoquinoline.
Because of the cis arrangement of PPh3 and bq groups
and the pyramidality at P, the stacking cannot be
perfectly parallel; the experimental dihedral angle is
close to 20°, the minimum achievable without strain.
We will call the pyridine ring of the benzoquinoline the
A ring, the central bq ring the B ring, and the cyclom-
etalated benzo ring the C ring. Of the three, only the B
ring is far enough away from the metal to participate
in significant || stacking interactions with a cis PPh3
group. Although such effects have been relatively little
discussed in the literature,9 they may be generally
important in metal complexes. Another potential at-
tractive interaction, perpendicular ( ) stacking, does not
seem to be important in this structure.
The energy difference between the two tautomers, 2q
and 2′q, at this level of calculation is 0.7 kcal/mol in
favor of the agostic tautomer, 2′q, a value that is in
reasonable agreement with experiment considering the
errors possible in such calculations (Table 3).
We decided to undertake a combined quantum me-
chanical/molecular mechanics calculation to take into
account the steric effects of the phenyl groups in the
real PPh3 ligands. This was done via the ONIOM
method (see Computational Details). The agostic and
nonagostic tautomers were once again optimized, re-
sulting in the metric parameters shown in Table 2. The
increased steric bulk of the phosphine on going from
pure QM (2q) to QM/MM (2m , ONIOM(B3PW91/UFF))
has the expected effect of causing the i-Pr group to
recede slightly from the metal. The agostic CH bond
moves slightly away from the metal and shortens
slightly relative to the pure QM case, as can be seen by
comparing the metric data in Table 2. The energy
difference between the two tautomers at this level of
calculation is 3.2 kcal/mol in favor of the nonagostic
For the t-Bu (3) compound, both the experimental (3x)
and the QM/MM structure (3m , Figure 3) show there
is one aromatic stacking interaction between the bq B
ring and one Ph group (ring B′) of one PPh3 ligand. The
centroid-centroid distances from the Ph ring to the bq
B ring are 3.67 Å (B′) and 5.02 Å (B′′) (3m ) versus 3.57
Å (B′) and 5.48 Å (B′′) (3x, Figure 2). In contrast, for 2,
the agostic structure (2′m , Figure 4) also has one such
interaction, but the observed nonagostic structure (2x
(9) Munakata, M.; Ning, G. L.; Suenaga, Y.; Kuroda-Sowa, T.;
Maekawa, M.; Ohta, T. Angew. Chem., Int. Ed. 2000, 39, 4555.
Ezuhara, T.; Endo, K.; Matsuda, K.; Aoyama, Y. New J . Chem. 2000,
24, 609. Scudder, M. L.; Goodwin, H. A.; Dance, I. G. New J . Chem.
1999, 23, 695-705. Cundari, T. Personal communication, 2001.
(8) Crabtree, R. H.; Holt, E. M.; Lavin, M.; Morehouse, S. M. Inorg.
Chem. 1985, 24, 1986.

Interplay of Weak Interactions
Organometallics, Vol. 21, No. 3, 2002 579
F igu r e 3. Diagram of the QM/MM calculated structure
(ONIOM(B3PW91/UFF)) of the agostic tert-butyl, 3m . The
G group is marked with solid spheres. The π-stacked rings
are identified with an open circle around their labels.
CH‚‚‚HM contacts are dotted and marked ‘b’. The CH‚‚‚M
interaction is dotted and marked ‘a’.
F igu r e 5. Diagram of the QM/MM calculated structure
(ONIOM(B3PW91/UFF)) of the nonagostic isopropyl, 2m .
The G group is marked with solid spheres. The π-stacked
rings are identified with an open circle around their labels.
CH‚‚‚HM contacts are dotted and marked ‘b’.
In line with the above argument, the PPh3 ligands
are almost perfectly eclipsed in nonagostic 2x, where
B′‚‚‚B‚‚‚B′′ stacking occurs, but are staggered in agostic
3x and 2′m , where only B′‚‚‚B stacking is present.
The t-Bu (3m ) compound also contains one
C-H‚‚‚H-Ir motif in which an ortho-CH bond of a PPh3
phenyl group points toward the metal hydride with the
result that the H‚‚‚H nonbonding distance is 2.110 Å
(ONIOM); the experimental situation is very similar,
but the H’s are less well located by X-ray and X-H
bonds are subject to systematic shortening. We have
previously drawn attention to the fact that C-H‚‚‚H-M
interactions are quite common in the structures of
transition metal hydride complexes in general.10 In this
prior work we emphasized cases where the H positions
were determined by neutron diffraction to avoid sys-
tematic errors associated with X-ray data.
For 2m , there are two CH‚‚‚HM interactions in the
nonagostic case, with d(H‚‚‚H) of 2.142 and 2.139 Å
(ONIOM); the experimental structure 2x shows a very
similar arrangement, but the H’s are not well located.
On going to the agostic structure, 2′m , these change to
2.099 and 2.358 Å, meaning that only one of the two
original CH‚‚‚HM interactions is still present. Once
again, we have a factor that could destabilize the agostic
form and account for its being less stable.
The P-Ir-P angle increases significantly on going
from the agostic to the nonagostic structures: P-Ir-P
is 159° in the agostic tBu (3m ), 166° in the agostic i-Pr
(2′m ), and 174° in the nonagostic i-Pr (2m ). An increase
in the number of π-π stacking interactionssfavored by
more ideally trans phosphinesscauses the phosphine
ligands to move away from the unhindered hydride and
toward the bq to maximize the π-π attractions in 2m ,
as also shown by the trend in H-Ir-P angles: 80.5° in
agostic 3m , 80° in agostic 2′m , and 87.3° in nonagostic
2m .
F igu r e 4. Diagram of the QM/MM calculated structure
(ONIOM(B3PW91/UFF)) of the agostic isopropyl, 2′m . The
G group is marked with solid spheres. The π-stacked rings
are identified with an open circle around their labels.
CH‚‚‚HM contacts are dotted and marked ‘b’. The CH‚‚‚M
interaction is dotted and marked ‘a’. One aryl H has been
removed to avoid obsuring key parts of the diagram.
and 2m ) has two, one on one side of the bq B ring and
the other on the other side. This is shown from the
centroid-centroid distances from the PPh3 B′ and B′′
ring to the bq B ring of 3.637 and 3.705 Å (2m , Figure
5), versus 3.765 Å (twice) (2x), entirely appropriate for
face-to-face stacking. In the agostic structure 2′m one
Ph ring (B′) remains within stacking distance (3.674 Å)
of the B ring, but the other (B′′) moves away (4.903 Å).
The loss of one favorable || aromatic stacking interaction
on going from the nonagostic to the agostic structure
could therefore be a factor that favors the observed
structure.
(10) Richardson, T.; Koetzle, T. F.; Crabtree, R. H. Inorg. Chim. Acta
1996, 250, 69.

580 Organometallics, Vol. 21, No. 3, 2002
Clot et al.
mL) and extracted with CHCl3 (3 × 40 mL). The combined
extracts were dried over Na2SO4 and treated overnight with
MnO2 (65 g). The filtered chloroform solution was evaporated
to dryness, and the residue was chromatographed on a silica
column using benzene as the eluent to yield 1.4 g (60%) of
yellow oil, which partially crystallized on standing. Anal. Calcd
for C16H15N‚0.2H2O: C, 85.50; H, 6.85; N, 6.23. Found: C,
85.45; H, 6.62; N, 6.24. 1H NMR (C6D6, 500 MHz): 9.77 (d,
J H-H ) 8.5, 1H), 7.30-7.60 (m, 4H), 7.23 (d, J H-H ) 9.6,
1H), 7.15 (s, 1H), 6.93 (d, J H-H ) 9.6, 1H), 3.17 (m, 1H), 1.42
(d, J H-H ) 8.5, 6H).
Thus we see three main types of interactions at work
in these structures: || aromatic stacking, C-H‚‚‚H-Ir
interactions, and an agostic C-H‚‚‚Ir. Each seems to
have broadly similar energy contributions, and thus no
one interaction is dominant. The agostic interaction may
therefore fail to occur when this leads to the loss of other
favorable interactions.
Con clu sion s
Agostic interactions are not always dominant in
determining structure but compete with other weak
interactions such as those between ligands. Here, we
report a case where the agostic tautomer is disfavored.
The resulting vacancy at the metal is stabilized by the
high trans effect bq aryl ligand in the trans position.
Ligand-ligand interactions, in this case π-stacking, can
dominate and prevent the agostic interaction.
2-ter t-Bu tyl-7,8-ben zoqu in olin e was obtained as a pale
yellow solid via the method described above under Ar but using
t-BuMgCl and CH2Cl2 as the eluent. Yield ) 50%, 1.24 g. Anal.
Calcd for C17H17N‚0.1H2O: C, 86.15; H, 7.25; N, 5.90. Found:
1
C, 86.02; H, 6.97; N, 5.33. H NMR (C6D6, 500 MHz): 9.68 (d,
J H-H ) 8.3, 1H), 7.20-7.60 (m, 6H), 7.12 (s, 1H), 1.47 (s, 9H).
2-I s o p r o p y l-7,8-b e n z o q u in o la t o (h y d r id o )b is (t r i-
ph en ylph osph in e)ir idiu m (III) Hexaflu or oph osph ate (2).
[Ir(H)2(PPh3)2(H2O)2][PF6] (500 mg, 0.556 mmol) was dissolved
in degassed, moist CH2Cl2 (20 mL) under Ar and cooled to 0
°C. A solution of 2-isopropyl-7,8-benzoquinoline (123 mg, 0.556
mmol) in 10 mL of CH2Cl2 was added dropwise over 10 min.
The resulting pale orange solution was allowed to stir at room
temperature for 1 h. The solution was concentrated to 20 mL,
and 20 mL of anhydrous diethyl ether was added as an upper
layer. Orange crystals of the product accumulated upon
standing overnight at room temperature. Yield: 57%, 343 mg.
Anal. Calcd for C52H45F6IrNP3‚CH2Cl2: C, 59.32; H, 4.38; N,
1.30. Found: C, 59.12; H, 4.56; N, 1.33. IR (KBr, cm-1): 2184
(νIr-H). 1H NMR (CD2Cl2, 500 MHz): -12.47 (t, 2J H-P )
14.50, 1H, Ir-H), 0.64 (d, J H-H ) 8.78, 6H, CH3), 2.67 (m,
1H, MeCH), 6.7-7.4 (m, 36H, arom.), 8.07 (s, 1H). 31P{1H}
NMR (CD2Cl2, 202.4 MHz): 25.45 (s).
2-t er t -B u t y l-7,8-b e n zo q u in o la t o (h y d r id o )b is (t r i-
p h en ylp h osp h in e)ir id iu m (III) Tetr a flu or obor a te (3). A
suspension of [Ir(H)2(PPh3)2(H2O)2][BF4] (400 mg, 0.486 mmol)
in degassed fluorobenzene (30 mL) was cooled under Ar to 0
°C. A solution of 2-tert-butyl-7,8-benzoquinoline (114 mg, 0.486
mmol) in fluorobenzene (10 mL) was added dropwise over 30
min, and the resulting mixture was stirred at room temper-
ature for 18 h. Addition of moist pentane (20 mL) resulted in
the precipitation of an off-white solid, which was collected and
recrystallized from CH2Cl2/pentane (1:1) to give the product
as colorless crystals. Yield: 30%, 151 mg. Anal. Calcd for
Com p u ta tion a l Deta ils
All calculations were performed with the Gaussian 98 set
of programs11 with the ONIOM method.12 The complexes Ir-
(H)(bq-G)(PPh3)2+ (2m , G ) i-Pr; 3m , G ) tBu) were optimized
at the ONIOM(B3PW91/UFF) level,13 where the QM part was
Ir(H)(bq-G)(PH3)2+. The G group was explicitly included in the
QM part. The iridium atom was represented by the relativistic
effective core potential (RECP) from the Stuttgart group (17
valence electrons) and its associated (8s7p5d)/[6s5p3d] basis
set,14 augmented with an f function (R ) 0.95). The phosphorus
atoms were also treated with Stuttgart’s RECPs and the
associated basis set,15 augmented by a polarization d function
(R ) 0.387). A 6-31G(d,p) basis set16 was used for the hydrides
and for both nitrogen atoms. The remaining atoms were
treated by a 6-31G basis set. The molecular mechanics
calculations were performed with the UFF force field.17 The
model systems Ir(H)(bq-G)(PH3)2+ (2q, G ) i-Pr; 3q, G ) t-Bu)
were also optimized at the B3PW91 level in pure QM calcula-
tions. The nature of the extrema (ONIOM(B3PW91/UFF) and
QM(B3PW91)) was checked through analytical computation
of the vibrational frequencies.
Exp er im en ta l Deta ils
2-Isop r op yl-7,8-ben zoqu in olin e. A solution of 7,8-benzo-
quinoline N-oxide6 (2.06 g, 10.55 mmol) in 40 mL of anhydrous
THF was cooled to 0 °C. Subsequent treatment with i-PrMgCl
(2.0 M solution in diethyl ether, 15.3 mL in 20 mL of THF)
over a 10 min period resulted in a color change to opaque
brown. After stirring at room temperature under Ar for 4 h,
the reaction mixture was poured into 10% aqueous NH4Br (100
C
53H47F4IrNBP2: C, 61.27; H, 4.52; N, 1.34. Found: C, 61.14;
H, 4.49; N, 1.39. IR (KBr, cm-1): 2166 (νIr-H), 2479, 2572
(agostic Ir-CH3). 1H NMR (CD2Cl2, 500 MHz): -14.34 (t,
2J H-P ) 14.67, 1H, Ir-H), 0.63 (s, 9H, CH3), 6.49 (d, J H-H
) 7.90, 1H), 6.62 (t, J H-H ) 8.20), 7.10-7.50 (m, 34H), 8.15
(d, J H-H ) 8.57, 1H). 31P{1H} NMR (CD2Cl2, 202.4 MHz):
20.41 (s). The PF6 salt was obtained in the same way in similar
yield starting from the appropriate bis-aqua PF6 salt.
Cr yst a llogr a p h ic Det a ils. Single crystals of 2 and 3
suitable for X-ray analysis were formed upon standing from a
methylene chloride/diethyl ether or a methylene chloride/
pentane solution, respectively. Crystallographic data are sum-
marized in Table 1. The structures were determined from data
collected with a Nonius KappaCCD diffractometer at -90 °C.
Lorentz and polarization corrections were applied to all data.
The structures were solved by direct methods (SIR92) using
the teXan crystal structure analysis package, and the function
minimized was ∑w(|Fo| - |Fc|)2 in all cases. Hydrogen atoms
were located in the difference maps and were refined. Com-
pound 2 crystallized in the space group Pnma, and the iridium,
as well as many ligand atoms, lie on special positions in a
mirror plane; hence the cation has a plane of symmetry
through the Bq ligand.
(11) 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.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.;
Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, Komaromi, P. I.; Gomperts, G.; 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.; J ohnson, B. G.; Chen, W.; Wong, M.
W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.; J . A. Pople J . A.
Gaussian 98; Gaussian, Inc.: Pittsburgh, PA, 1998.
(12) Svensson, M.; Humbel, S.; Froese, R. D. J .; Matsubara, T.;
Sieber, S.; Morokuma, K. J . Phys. Chem. 1996, 100, 19357.
(13) (a) Becke, A. D. J . Chem. Phys. 1993, 98, 5648. (b) Perdew, J .
P.; Wang, Y. Phys. Rev. B 1992, 82, 284.
(14) Andrae, D.; Ha¨ussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123.
(15) Bergner, A.; Dolg, M.; Ku¨chle, W.; Stoll, H.; Preuss, H. Mol.
Phys. 1990, 30, 1431.
(16) Hariharan, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28, 213.
(17) Rappe, A. K.; Casewitt, C. J .; Colwell, K. S.; Goddard, W.; A.;
Skiff, W. M. J . Am. Chem. Soc. 1992, 114, 10024.
Ack n ow led gm en t. We thank the NSF for funding.
OM010802I