Why does the tetrakis(trimethylphosphine)iridium(III) hydridochloride cation adopt the sterically and electronically unfavorable cis geometry?

Article abstract of DOI:10.1021/om0004848

Although the tetrakis(trimethylphosphine)iridium(III) hydridochloride cation, [HIrCl-(PMe₃)₄]⁺, and similar complexes would intuitively be expected to adopt a trans geometry on electronic and sterical grounds, experimentally the cis geometry is found to prevail. Quantum chemical calculations suggest that the trans and cis structures are nearly isoenergetic, such that the cis:trans equilibrium is dominated by the higher entropy of the lower symmetry cis structure.

Full text of DOI:10.1021/om0004848

4608
Organometallics 2000, 19, 4608-4612
Wh y Does th e Tetr a k is(tr im eth ylp h osp h in e)ir id iu m (III)
Hyd r id och lor id e Ca tion Ad op t th e Ster ica lly a n d
Electr on ica lly Un fa vor a ble Cis Geom etr y?
Ofer Blum,† Ra’anan Carmielli, J an M. L. Martin,* and David Milstein*
Department of Organic Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel
Received J une 9, 2000
Although the tetrakis(trimethylphosphine)iridium(III) hydridochloride cation, [HIrCl-
(PMe3)4]+, and similar complexes would intuitively be expected to adopt a trans geometry
on electronic and sterical grounds, experimentally the cis geometry is found to prevail.
Quantum chemical calculations suggest that the trans and cis structures are nearly
isoenergetic, such that the cis:trans equilibrium is dominated by the higher entropy of the
lower symmetry cis structure.
In tr od u ction
quantum mechanical calculations suggest that entropy
contributions originating in different symmetries domi-
nate the equilibrium between the isomers.
The relative stability of geometrical isomers plays a
major role in determining the course and outcome of
many organometallic reactions, because the intermedi-
ates are often coordinatively unsaturated and capable
of facile isomerization. Hence, controlling these isomer-
izations is at the heart of the design of stereospecific or
enantioselective catalysis. It is currently believed that
ligand effects, both steric and electronic, are the domi-
nant nonenvironmental factors that determine the
relative stabilities of coordination compound isomers.1
Here we demonstrate that neither steric nor electronic
arguments can predict correctly the stable isomer of the
octahedral RMXL4 complexes (R ) strong σ-donor such
as hydride, methyl, and phenyl; M ) Ir(III), Ru(II), and
Os(II); X ) weak σ-donor such as halide, OR, SH, NHR,
PHR, H2O, NH3, CO, CH3CN, and η2-H2; L ) PMe3 and
PMe2(C6H5)). Of the two possible isomers of RMXL4 (1,
2), the trans configuration is predicted to be sterically
more stable, because each of the largest ligands (L) has
only two large neighbors at 90° positions, whereas in
the cis isomer, two of the L groups have three large
neighbors cis disposed. Electronic considerations also
suggest that the trans R-M-X geometry should be the
most stable, since the R group is the best σ-donor (thus,
having the largest trans influence),2 and since X is the
weakest σ-donor. Nevertheless, we observed the trans
to cis isomerization of [HIrCl(PMe3)4]+, indicating that
contrary to electronic and steric arguments, the cis
isomer is thermodynamically more stable. We observed
a similar trans to cis isomerization during the oxidative
addition of water, methanol, and H2S to [Ir(PMe3)4]PF6,
but were unable to isolate the trans products.3,4 Our
Exp er im en ta l a n d Com p u ta tion a l Deta ils
All syntheses and chemical manipulations were carried out
under nitrogen in a Vacuum Atmospheres DC-882 glovebox.
All solvents were reagent grade or better. THF was prepurified
through a column of basic alumina. Both THF and pyridine
were refluxed over Na/benzophenone ketyl, distilled under
argon and stored over activated 4 Å molecular sieves. All
deuterated solvents were purchased from Aldrich, degassed
and dried over a large amount of 3 Å molecular sieves for at
least one week before use. 1H and 31P NMR spectra were
recorded at 400.19 and 161.9 MHz, respectively, using a
Bruker AMX 400 spectrometer. Chemical shifts are reported
in ppm downfield from Me4Si and referenced to the residual
solvent-h1 or downfield from external H3PO4 85% in D2O.
P r ep a r a tion of tr a n s-[HIr Cl(P Me3)4][OSO2-p-tolyl] (1).
p-Toluenesulfonic acid monohydrate (14 mg, 7.36 × 10-2 mmol)
dissolved in dry THF (2 mL) was added to a suspension of the
previously reported5 [Ir(PMe3)4]Cl (41 mg, 7.71 × 10-2 mmol)
in THF (4 mL) at room temperature. The red solid turned
white during 5 min, and the solution became colorless. After
30 min the solid was isolated by filtration, yielding 1 as the
single product.1H NMR (pyridine-d5): 8.51 (d, 3J d
) 8.0 Hz,
H-H
2H, ortho to IrOSO2), 7.12 (d, 3J d
) 7.8 Hz, 2H, ortho to
H-H
CH3), 1.69 (t, virtualJ t
) 3.1 Hz, 36H, 4P(CH3)3), -21.74
H-P
(quintet, 2J quintet
) 14.6 Hz, 1H, Ir-H). 31P{1H} NMR
H-P
(pyridine-d5): -45.6 (s). IR (Nujol): 2163 (strong, νIr-H trans
to a weak σ-donor). Anal. Calcd: C 33.90, H 6.59. Found: C
33.77, H 6.42.
P r ep a r a tion of cis-[HIr Cl(P Me3)4]P F 6 (2). Bubbling dry
gaseous HCl (99.999% purity) through a red THF (10 mL)
6
solution of [Ir(PMe3)4]PF6 (51 mg) for 2 h caused bleaching.
The solvents were stripped off under high vacuum, yielding
complex 2 as a white solid. 1H NMR (pyridine-d5): 1.80 (d,
virtual
2J H-P ) 10.3 Hz, 9H, P(CH3)3 trans to Cl), 1.74 (t,
J
)
H-P
* Corresponding authors. E-mail: comartin@wicc.weizmann.ac.il,
david.milstein@weizmann.ac.il.
2
3.8 Hz, 18H, 2P(CH3)3 mutually trans), 1.54 (dd, J H-P ) 8.2
† Present address: Department of Chemical Engineering, University
of Michigan, Ann Arbor, MI 48109.
4
Hz, J H-H,trans ) 1.0 Hz, 9H, P(CH3)3 trans to H), -11.65 (dtd
2
(apparent dq), J d
) 148.3 Hz, the weighed average of
H-P,trans
(1) (a) Fernandez, A. L.; Prock, A.; Giering, W. P. Organometallics
1996, 15, 2784. (b) Wilson, M. R.; Liu, H.; Prock, A.; Giering, W. P.
Organometallics 1993, 12, 2044. (c) Atwood, J . D. Inorganic and
Organometallic Reaction Mechanisms, 2nd ed.; VCH: New York, 1997.
(2) Appleton, T. G.; Clark, H. C.; Manzer, L. E. Coord. Chem. Rev.
1973, 10, 335.
(4) Milstein, D.; Calabrese, J . C.; Williams, I. D. J . Am. Chem. Soc.
1986, 108, 6387.
(5) Herskovitz, T.; Guggenberger, L. J . Am. Chem. Soc. 1976, 98,
1615.
(3) Blum, O.; Milstein, D. Unpublished results.
(6) Thorn, D. L. Organometallics 1982, 1, 197.
10.1021/om0004848 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 09/28/2000

Geometry of the [HIrCl(PMe3)4]+ Cation
2J t and 2J d ) 18.7 Hz, 1H, Ir-H). 31P{1H} NMR
Organometallics, Vol. 19, No. 22, 2000 4609
the total energies by the use of the default grids are several
orders of magnitude larger than those incurred by using the
default energy convergence criteria.15
H-P,cis
H-P,cis
(pyridine-d5): -45.7 (dt, 2J t
) 18.0 Hz, 2J d
) 11.5
P-P,cis
P-P,cis
Hz, 1P trans to H), -47.5 (dd (apparent t), the weighed average
2
of 2J dP-P, cis and J dP-P, cis ) 19.0 Hz, 2P mutually trans), -54.1
We used a number of exchange-correlation (XC) function-
als: (1) local density approximation (LDA);16,17 (2) the gradient-
corrected Becke-Lee-Yang-Parr (BLYP) functional (the
Becke 1988 exchange functional18 with Lee-Yang-Parr (LYP)
correlation.19); (3) the popular Becke 3-parameter-Lee-Yang-
Parr (B3LYP) functional (Becke’s 3-parameter hybrid exchange
functional20 with LYP correlation); (4) the mPW1PW91 func-
tional recently proposed by Adamo and Barone21 (a 1:3
mixture22 of Hartree-Fock and modified Perdew-Wang 1991
exchange23 with standard Perdew-Wang 1991 correlation23).
The latter appears to yield a better description of long-range
interactions than standard XC functionals.21,24
(td, 2J t
) 20.2 Hz, 2J d
) 11.3 Hz, 1P trans to Cl),
P-P,cis
) 710 Hz, 1P). IR (Nujol): 2065
P-P,cis
-142.0 (heptet, 1J heptet
P-F
(strong, νIr-H trans to a strong σ-donor). Anal. Calcd: C 22.28,
H 5.76. Found: C 22.01, H 5.57.
Isom er iza tion of tr a n s-[HIr Cl(P Me3)4][OSO2-p-tolyl]
(1) to th e cis Isom er 2. (a) At room temperature, in a
pyridine-d5 solution: The isomerization is very slow, and does
not reach equilibrium within a week. (b) At 65 °C: A THF
solution of 1 in a sealed bottle was kept in an oil bath at 65
°C for 48 h. The bottle was quickly cooled to -78 °C and
allowed to warm slowly to room temperature. The solvent was
stripped off under vacuum, and the white residue was redis-
solved in pyridine-d5. The 1:2 ratio obtained was 1:6.
Com p u ta tion a l Meth od s. All the calculations were carried
out on DEC Alpha 500/500 and SGI Octane workstations using
Gaussian 98.7
In addition, we carried out Hartree-Fock calculations for
comparison; these could be considered a special case of hybrid
DFT functionals, with 100% Hartree-Fock exchange and no
correlation. All relevant results can be found in Table 1.
The Hay-Wadt Los Alamos National Laboratory 2-shell
double-ú (LANL2DZ) basis set/relativistic effective core po-
tential (RECP) combination8 was used for the geometry
optimizations and for the harmonic frequency calculations
involved in the zero-point vibrational energies (ZPVE) and in
setting up the partition functions for the rigid rotor-harmonic
oscillator (RRHO) thermal corrections. As is customary, the
Dunning valence double-ú basis set9 was used on hydrogen and
first-row atoms.
The relative energies of the different structures were in
addition evaluated with the same basis set-RECP combina-
tion, augmented with polarization functions taken from ref 10;
this is denoted by the acronym LANL2DZP throughout the
paper. (For hydrogen and first-row atoms, the Dunning cc-
pVDZ basis set11 was used in this case.) Tightened convergence
criteria were used throughout for the SCF/Kohn-Sham itera-
tions and where possible for the geometry optimizations.
To ensure reproducibility of energies, geometries, and
computed thermodynamic functions to the precision tabulated,
it was found necessary to use finer DFT integration grids than
the Gaussian 98 defaults. For all energy and gradient steps,
we employed a pruned (99,590) grid (i.e., a 590-point Lebedev
angular grid12 combined with a 99-point Euler-Maclaurin
radial grid13) rather than the (75,302) default, while the pruned
(50,194) SG1 grid14 (rather than the (35,110) default) was used
in solving the coupled perturbed Kohn-Sham equations. For
third-row transition metal compounds, the errors incurred in
Resu lts a n d Discu ssion
trans-[HIrCl(PMe3)4]+ (1) is obtained within minutes
as a pure white solid by protonation of the previously
reported5 [Ir(PMe3)4]Cl suspended in THF with an
equimolar amount of p-toluenesulfonic acid. Heating of
1 at 65 °C in THF overnight yields a 6:1 mixture of cis-
[HIrCl(PMe3)4]+ (2) and trans-1 (eq 1). Isomerization of
1 to 2 takes place also in pyridine-d5 at room temper-
ature, albeit slowly. Complex 2 can be prepared directly
from the previously reported6 [Ir(PMe3)4]PF6 by bub-
bling gaseous HCl through its red THF solution.
(7) 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, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith,
T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.;
Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen, W.; Wong, M.
W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.; Pople, J . A.
Gaussian 98, Revision A.3; Gaussian, Inc.: Pittsburgh, PA, 1998.
(8) Hay, P. J .; Wadt, W. R. J . Chem. Phys. 1985, 82, 270, 284, 299.
(9) Dunning, T. H.; Hay, P. J . In Methods of Electronic Structure
Theory; Schaefer, H. F., III, Ed.; Plenum Press: New York, 1977.
(10) (a) Ehlers, A. W.; Bo¨hme, M.; Dapprich, S.; Gobbi, A.; Hollwarth,
A.; J onas, V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking,
G. Chem. Phys. Lett. 1993, 208, 111. (b) Hollwarth, A.; Bo¨hme, M.;
Dapprich, S.; Ehlers, A. W.; Gobbi, A.; J onas, V.; Ko¨hler, K. F.;
Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys. Lett. 1993,
208, 237.
(15) Martin, J . M. L.; Bauschlicher, C. W., J r.; Ricca, A. Comput.
Phys. Commun., in press.
(16) Slater, J . C. Phys. Rev. 1951, 81, 385.
(17) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200;
functional III from their paper was used throughout.
(18) Becke, A. D. Phys. Rev. A 1988, 38, 3098; J . Chem. Phys. 1988,
88, 2547.
(19) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(20) Becke, A. D. J . Chem. Phys. 1993, 98, 5648.
(21) Adamo, C.; Barone, V. J . Chem. Phys. 1998, 108, 664
(22) Perdew, J . P.; Emzerhof, M.; Burke, K. J . Chem. Phys. 1996,
105, 9982.
(23) (a) Perdew, J . P. In Proceedings of the 21st Annual International
Symposium on the Electronic Structure of Solids; Ziesche, P., Eschrig,
H., Eds.; Akademie Verlag: Berlin, 1991. (b) Perdew, J . P.; Chevary,
J . A.; Vosko, S. H.; J ackson, K. A.; Pederson, M. R.; Singh, D. J .;
Fiolhais, C. Phys. Rev. B 1992, 46, 6671. (c); Perdew J . P.; Wang, Y.
Phys. Rev. B 1992, 45, 13244.
(24) Martin, J . M. L. In Density Functional Theory: a Bridge between
Chemistry and Physics; Geerlings, P., De Proft, F., Langenaeker, W.,
Eds.; VUB Press: Brussels, 1999.
(11) Dunning, T. H. J . Chem. Phys. 1989, 90, 1007.
(12) Lebedev, V. I. Russ. Acad. Sci. Dokl. Math. 1992, 45, 587.
(13) Murray, C. W.; Handy, N. C.; Laming, G. J . Mol. Phys. 1993,
78, 997.
(14) Gill, P. M. W.; J ohnson, B. G.; Pople, J . A. Chem. Phys. Lett.
1993, 209, 506.

4610 Organometallics, Vol. 19, No. 22, 2000
Ta ble 1. Cis-tr a n s En er gy Differ en ce (k ca l/m ol) for Va r iou s [HIr ClL4]+ Com p lexesa
Blum et al.
4 PH3
4 PMe3
2 H2PCH2CH2PH2
2 Me2PCH2CH2PMe2
∆Ee, LANL2DZ
0.98
1.04
3.31
-0.16
0.68
HF
2.36
1.11
1.07
-0.19
0.60
9.17
7.34
6.84
5.90
6.84
11.87
9.94
9.24
8.46
9.47
mPW1PW91
LDA
BLYP
B3LYP
∆Ee, LANL2DZP
HF
1.96
0.56
0.28
-0.57
0.21
0.64
0.88
3.09
-0.78
0.17
8.88
7.01
6.42
5.75
6.67
11.71
9.68
8.82
8.35
9.30
mPW1PW91
LDA
BLYP
B3LYP
Zero-Point Vibration and Temperature Effects (B3LYP/LANL2DZ)
∆ZPVE
-0.18
-0.06
2.80
-3.23
-0.43
+0.13
-0.04
-0.14
1.40
1.04
2.44
-0.17
-0.02
1.44
-0.95
0.49
-0.21
+0.02
1.39
0.33
1.72
∆(H298-H0)
∆Srot (incl. R ln 2)
∆Svib
∆Stotal
∆(G298-H298) ) -T∆S298
-0.73
-0.15
-0.51
∆G298 (LANL2DZP)b
HF
1.85
0.44
0.17
-0.68
+0.10
-0.26
-0.01
2.19
-1.67
-0.73
8.55
6.67
6.08
5.41
6.33
11.00
8.97
8.10
7.64
8.59
mPW1PW91
LDA
BLYP
B3LYP
a
b
A negative sign implies the cis form to be more stable. ∆Ee[LANL2DZP] + ∆ZPVE[B3LYP/LANL2DZ] + ∆(G298-H0)[B3LYP/
LANL2DZ].
Replacing the four PMe3 ligands by two chelates of
the closely analogous 1,2-bis(dimethylphosphinoethane)
(DMPE) reverses the stability order of the isomers. Behr
and co-workers reported that heating cis-[HIrCl-
(DMPE)2]+ to 100 °C in DMSO yields the trans isomer
as the single product (eq 2).25
OH,34 OCH(CH3)2,34 p-OC6H4(CH3),34 OC(dCH2)C6H5,34,35
oxygen-bound enolates,35 and N(H)C6H5;34 R ) methyl,
X ) OH,35 and Cl.32 cis-HOs(OH)(PMe3)4 probably
belongs to this family of complexes as well.36 Cis to trans
isomerization at room temperature was reported also
with another phosphine (L ) PMe2(C6H5), R ) hydride,
X ) η2-H2).37 Replacing PMe3 by DMPE reverses the
relative stabilities of the isomers for the ruthenium
complexes as well (R ) hydride, X ) OH,35 OC(O)H,38
OC(O)C6H5,39 and Cl 40). For none of these ruthenium
complexes did the authors consider the stability issue
addressed here.
In addition to the electronic (mainly σ-donation by the
ligands) and steric factors, the π-donation capability of
the weak σ-donor is unimportant in reversing the
expected stability order of RMX(PMe3)4, since X can also
be NH3,33 η2-H2,37 CH3CN,33 or CO33 (not π-donors).
Qu a n tu m Ch em ica l Ca lcu la tion s. To explain the
unexpected stability order of the [HIrCl(PMe3)4]+ iso-
mers, we have calculated the energy differences between
the cis and the trans configurations of various [HIr-
The unusual stability of cis-RMX(PMe3)4 is not unique
to iridium. A large series of analogous d6 ruthenium
complexes exhibits the same phenomenon:26 R ) hy-
dride, X ) OH,27 p-OC6H4CH3,28 p-OC6H4(OCH3),29
p-OC6H4Cl,30 p-OC6H4NO2,30 N(H)C6H5,28 P(H)C6H5,30
p-SC6H4(CH3),30 Cl,31,32 and NH3;33 R ) phenyl, X )
(31) J ones, R. A.; Real, F. M.; Wilkinson, G.; Galas, A. M. R.;
Hursthouse, M. B.; Abdul Malik, K. M. J . Chem. Soc., Dalton Trans.
1980, 511.
(32) Statler, J . A.; Wilkinson, G.; Thornton-Pett, M.; Hursthouse,
M. B. J . Chem. Soc., Dalton Trans. 1984, 1731.
(33) Rappert, T.; Yamamoto, A. Organometallics 1994, 13, 4984.
(34) Hartwig, J . F.; Bergman, R. G.; Andersen, R. A. J . Am. Chem.
Soc. 1991, 113, 3404.
(25) Behr, A.; Herdweck, E.; Herrmann, W. A.; Keim, W.; Kipshagen,
W. Organometallics 1987, 6, 2307.
(26) Trans to cis isomerizations were reported for only a few of these
complexes. However, most of the cis products were heated for prolonged
periods in polar and even protic solvents, which would enable isomer-
ization to a putative more stable trans isomer by dissociation of the
anionic ligand X.
(27) Burn, M. J .; Fickes, M. G.; Hartwig, J . F.; Hollander, F. J .;
Bergman, R. G. J . Am. Chem. Soc. 1993, 115, 5875.
(28) Hartwig, J . F.; Andersen, R. A.; Bergman, R. G. Organometallics
1991, 10, 1875.
(29) Burn, M. J .; Bergman, R. G. J . Organomet. Chem. 1994, 472,
43.
(30) Burn, M. J .; Fickes, M. G.; Hollander, F. J .; Bergman, R. G.
Organometallics 1995, 14, 137.
(35) Hartwig, J . F.; Bergman, R. G.; Andersen, R. A. Organometallics
1991, 10, 3326.
(36) Trans O-H oxidative addition products were not observed at
elevated temperatures. Yet, as only apolar solvents were used, it is
possible that a less stable cis product was kinetically trapped: Gotzig,
J .; Werner, R.; Werner, H. J . Organomet. Chem. 1985, 290, 99.
(37) Lough, A. J .; Morris, R. H.; Ricciuto, L.; Schleis, T. Inorg. Chim.
Acta 1998, 270, 238.
(38) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. Organometallics
1996, 15, 5166.
(39) Hartwig, J . F.; Bergman, R. G.; Andersen, R. A. Organometallics
1991, 10, 3344.
(40) Chatt, J .; Davidson, J . M. J . Chem. Soc. 1965, 843.

Geometry of the [HIrCl(PMe3)4]+ Cation
Organometallics, Vol. 19, No. 22, 2000 4611
ClL4]+ complexes using a variety of quantum chemical
approaches. All relevant results are reported in Table 1.
(a) The first term is the rotational symmetry contri-
bution R ln(σtrans/σcis), where σcis and σtrans are the
rotational symmetry numbers of the cis and trans
isomers, respectively. In the case of [HIrCl(PH3)4]+, σcis
) 2 (symmetry Cs) and σtrans ) 4 (symmetry C4v), while
for [HIrCl(PMe3)4]+ (1 vs 2), σcis ) 1 (symmetry C1) and
σtrans ) 2 (symmetry C2v). Either case leads to a
contribution to ∆S of R ln 2 (1.4 eu), and hence of -RT
ln 2 to ∆G (0.41 kcal/mol at room temperature). This
would translate into a 2:1 equilibrium in favor of the
cis isomer in an otherwise fully thermoneutral system.
The same applies to the phosphochelate systems, where
the trans and cis isomers have C2 and C1 symmetry,
respectively. (Quantum chemical packages with inte-
grated RRHO features such as Gaussian 98 will only
include this term correctly if the correct molecular
symmetry is imposed.)
At the B3LYP/LANL2DZ level, trans-[HIrCl(PH3)4]+
is found to have a 4-fold symmetry axis (point group
C4v), while the corresponding cis structure only has a
symmetry plane (point group Cs). Upon introducing the
methyl groups on the phosphines, the symmetry of the
trans isomer drops to C2v, while cis-[HIrCl(PMe3)4]+ has
no symmetry at all. At the HF/LANL2DZ level, trans-
[HIrCl(PH3)4]+ is found to be 2.4 kcal/mol more stable
than its cis counterpart; upon introducing the Me
groups, the difference is reduced to 1.0 kcal/mol. En-
largement of the basis set shifts the equilibrium toward
the cis structures, as it does in the various DFT
calculations.
At the B3LYP level, trans-[HIrCl(PH3)4]+ and trans-
[HIrCl(PMe3)4]+ are found to be 0.6 and 0.7 kcal/mol
more stable than their respective cis counterparts.
Enlargement of the basis set largely eliminates these
differences, leading to the isomers being nearly isoen-
ergetic. The difference between the HF and B3LYP
results is probably mainly due to correlation effects,
since the exchange behavior of HF and B3LYP is
expected to be similar by design.
(b) The second is a rotational entropy term, which
from standard statistical thermodynamics is found to
be (R/2) [ln IA,cisIB,cisIC,cis - ln IA,transIB,transIC,trans], in
which IA, IB, and IC represent the three molecular
principal moments of inertia. Since in the R ) H case
the trans isomer (an oblate symmetric top) is ap-
preciably more “compact” than the cis isomer in terms
of having a smaller IAIBIC product, its rotational entropy
is decreased by an additional 1.4 eu. For the bulkier R
) Me and phosphochelate cases, this particular differ-
ence is not significant enough to be reflected in the
entropy.
LDA exhibits a pronounced shift of the equilibrium
toward the trans form upon Me substitution (∆E chang-
ing from 1.07 to 3.31 kcal/mol with the LANL2DZ and
from 0.28 to 3.09 kcal/mol with the LANL2DZP basis
set), which is quite absent in the gradient-corrected and
hybrid functionals; we conclude that LDA is inadequate
even for this comparatively simple problem. By compar-
ing LDA to BLYP results, we find that the gradient
corrections cause fundamental changes; as expected, the
B3LYP results (relative to BLYP) move slightly in the
direction of the HF values because of the admixture of
HF exchange. We note that trends in the mPW1PW91
results appear to parallel those at the B3LYP level,
although the former exchange-correlation functional
tends to favor the trans form relative to B3LYP. Overall,
we see a sequence HF ≈ mPW1PW91 > B3LYP > BLYP
for the relative energy.
(c) The lower symmetry structures are expected to be
less rigid and hence to have, on average, more low-lying
vibrational levels. This effect contributes about 1 eu in
the R ) Me case; given the limitations of the RRHO
approximation for systems with many low-barrier hin-
dered rotors, this contribution is at best semiquantita-
tive, however. (While most of the error thus incurred
should cancel between the two isomers, complete can-
cellation should not be taken for granted.) In the R )
H case, a contribution of -3.2 eu in the opposite
direction is surprisingly seen. Upon close inspection, this
is found to be due mostly to an anomalously low
frequency for concerted PH3 rotation.
In light of the smallness of the energy changes
involved, it would have been desirable to carry out all
optimization at the B3LYP/LANL2DZP level as well.
However, considering that the B3LYP/LANL2DZ opti-
mization on the largest system required about 2 months
CPU-time-equivalent on an SGI Origin 2000, this does
not appear to be a realistic option at present.
The bottom line of the above considerations is that
even if the cis- and trans-[HIrCl(PMe3)4]+ structures
were perfectly isoenergetic at 0 K, entropy effects would
slightly favor the cis structure.
The 6:1 cis:trans-[HIrCl(PMe3)4]+ ratio found experi-
mentallysif we were to assume that no differential
solvent effects exist between the structuresscorresponds
to a Gibbs energy difference of -1.15 kcal/mol at 298
K. This is in fact intermediate between the large basis
set computed B3LYP (-0.73 kcal/mol) and BLYP (-1.67
kcal/mol) results: From the observed ∆G and the
computed thermal corrections (keeping in mind the
limitations of the RRHO approximation), we obtain a
bottom-of-the-well energy difference of only about -0.4
kcal/mol favoring the cis form. We also note that the
trend in the computed results suggests that basis set
extension favors the cis form, and we expect this trend
to be continued with further basis set enlargement.
Introduction of zero-point energy slightly favors cis-
over trans-[HIrCl(PH3)4]+, but has essentially no effect
for [HIrCl(PMe3)4]+. Introduction of the enthalpy func-
tions H298-H0 has essentially no effect for trans- versus
cis-[HIrCl(PH3)4]+, but slightly favors cis over trans once
the Me groups are introduced, although this small effect
is within the error margin of the calculation.
What is the effect of entropy on the equilibrium?
Entropy contributions to steric interaction are expected
to favor the trans isomer,41 in contradiction with the
observed relative stabilities. Within the RRHO ap-
proximation, the entropy contribution to the equilibrium
consists of three terms:
It should be stressed that the basis sets employed are
still quite some distance removed from the basis set
limit, even for DFT methods, for which basis set
convergence on a variety of properties is much faster
(41) Fernandez, A.; Reyes, C.; Prock, A.; Giering, W. P. Organome-
tallics 1998, 17, 2503.

4612 Organometallics, Vol. 19, No. 22, 2000
Blum et al.
Ta ble 2. Meta l-Liga n d Bon d Dista n ces (Å) a t th e B3LYP /LANL2DZ Level
trans
C4v
,
C2v
,
C2,
C2,
4 PH3
4 PMe3
2 H2PCH2CH2PH2
2 Me2PCH2CH2PMe2
H
1.596
2.544
2.405
2.405
2.405
2.405
1.600
2.576
2.442
2.459
2.442
2.459
1.596
2.552
2.390
2.392
2.390
2.392
1.596
2.568
2.408
2.412
2.408
2.412
Cl
P1
P2
P3
P4
cis
Cs,
C1,
C1,
C1,
4 PH3
4 PMe3
2 H2PCH2CH2PH2
2 Me2PCH2CH2PMe2
H
Cl
1.598
2.471
2.519
2.378
2.402
2.402
1.599
2.515
2.545
2.389
2.457
2.462
1.602
2.479
2.500
2.363
2.382
2.391
1.607
2.500
2.521
2.339
2.368
2.414
P(trans H)
P(trans Cl)
P3
P4
than for conventional electron correlation methods.24
Yet because we are merely comparing energies of two
isomers with like connectivity (same number of bonds
of each formal type) rather than considering the energy
change of a reaction involving bond breaking, consider-
able error cancellation occurs, resulting in the surpris-
ingly good agreement with experiment seen here.
What is the situation for the chelated complex [HIrCl-
(Me2PCH2CH2PMe2)2]+? There too, entropy effects favor
the cis form, but since ∆Ee favors the trans form and is
an order of magnitude larger than T∆S in this case,
essentially 100% trans product is expected (and ob-
tained). The addition of Me groups increases ∆Ee, as
expected, at all levels of theory considered. The most
notable trend among the levels of theory is that Har-
tree-Focksrelative to the various DFT exchange-
correlation functionalssappears to be biased toward the
trans form by 2-3 kcal/mol. It is well known that
neglect of electron correlation usually results in over-
estimation of charge separation (expressing itself, among
other things, in overestimated dipole moments and
overly localized structures). Hence introduction of elec-
tron correlation would be expected to reduce the σ-donor
effect of the substituents, which is consistent with the
difference between Hartree-Fock and the various DFT
results.
it is immediately seen that the geometry perturbation
relative to the trans isomer is substantially larger in
the chelated cases than in the unchelated ones; Me
substitution further enhances the difference. (The
P-Ir-P angles in the chelated complexes are all in the
85-87° range, close to the “bite angle” for the phospho-
chelate ligands.) This would suggest that the geometric
constraints imposed by the chelate would be responsible
for the different behavior of chelated and unchelated
systems.
Summarizing, the computed results, theory, and
experiment are fundamentally in agreement about the
question at hand.
Ack n ow led gm en t. We thank the Israel Science
Foundation, J erusalem, Israel, and the Minerva Foun-
dation, Munich, Germany, for financial support. D.M.
is the holder of the Israel Matz Professorial Chair of
Organic Chemistry. J .M. is a Yigal Allon Fellow and the
Incumbent of the Helen and Milton A. Kimmelman
Career Development Chair. The authors thank Dr. A.
Sundermann for helpful discussions.
Su p p or tin g In for m a tion Ava ila ble: Computed (B3LYP/
LANL2DZ) geometries in Xmol (.xyz) format, harmonic fre-
quencies, and infrared intensities for the species in Table 1
are available in machine-readable form on the World Wide
Web at the location http://theochem.weizmann.ac.il/web/papers/
isom.html. The geometries can be visualized directly in any
Xmol-compatible molecular viewer. This material is also
available free of charge via the Internet at http://pubs.acs.org.
A similar trend is seen for the unchelated systems;
however, both ∆Ee and the changes involved in it are
smaller by almost an order of magnitude there. Why
the difference? The computed metal-ligand bond dis-
tances (Table 2) offer a beginning of an explanation. As
expected, in all cis cases the Ir-P bond trans to the Cl
is shortened and that trans to the H is lengthened. (This
trend is also reflected in the computed Wiberg bond
orders42 from a natural population analysis.43) However,
OM0004848
(42) Wiberg, K. B. Tetrahedron 1968, 24, 1083.
(43) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88,
899.