Reversible cyclometalation of silyl ligands. First X-ray structure of an iridium(I) silyl that is not stabilized by chelation

Article abstract of DOI:10.1021/om960214i

The iridasilacycle fac-(PMe₃)₃Ir(o-C₆H₄SiMe ₂)(H) (1) results from the ortho metalation of the transient (PMe₃)₃Ir(SiMe₂Ph) (7), generated from Ir(PMe₃)₄Cl and PhMe₂SiLi at ambient temperature. Complex 1 can also be obtained from fac-(PMe₃)₃Ir(CH₃)(H)(SiMe₂Ph) (6), which, when heated, reductively eliminates methane. 1 and its analog fac-(PMe₃)₃Ir(o-C₆-H₄SiPh ₂)(H) (2), react with hydrogen at 80 °C to yield the dihydridosilyliridium(III) complexes fac-(PMe₃)₃Ir(SiR₂Ph)(H)₂ (3, R = Me; 4, R = Ph). The reaction of 1 with deuterium demonstrates that initial opening of the iridasilacycle takes place, followed by D-D oxidative addition to 7. The reaction of 1 with CO allows us to trap the iridium(I) species and provides a synthetic route to Ir(PMe₃)₂(CO)₂(SiMe₂Ph) (5). Complex 5 was identified by spectroscopy and by X-ray crystallography and was synthesized independently from 3 and CO. The chemistry presented shows that metalation of silyl ligands can be a facile reversible process that takes place under mild conditions. It also indicates that iridasilacycles can be used as a masked form of electron-rich Ir(I) silyl complexes, which are not readily accessible species.

Full text of DOI:10.1021/om960214i

Organometallics 1996, 15, 3317-3322
3317
Rever sible Cyclom eta la tion of Silyl Liga n d s. F ir st X-r a y
Str u ctu r e of a n Ir id iu m (I) Silyl Th a t Is Not Sta bilized by
Ch ela tion
Michael Aizenberg and David Milstein*
Department of Organic Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel
Received March 21, 1996X
The iridasilacycle fac-(PMe3)3Ir(o-C6H4SiMe2)(H) (1) results from the ortho metalation of
the transient (PMe3)3Ir(SiMe2Ph) (7), generated from Ir(PMe3)4Cl and PhMe2SiLi at ambient
temperature. Complex 1 can also be obtained from fac-(PMe3)3Ir(CH3)(H)(SiMe2Ph) (6),
which, when heated, reductively eliminates methane. 1 and its analog fac-(PMe3)3Ir(o-C6-
H4SiPh2)(H) (2), react with hydrogen at 80 °C to yield the dihydridosilyliridium(III) complexes
fac-(PMe3)3Ir(SiR2Ph)(H)2 (3, R ) Me; 4, R ) Ph). The reaction of 1 with deuterium
demonstrates that initial opening of the iridasilacycle takes place, followed by D-D oxidative
addition to 7. The reaction of 1 with CO allows us to trap the iridium(I) species and provides
a synthetic route to Ir(PMe3)2(CO)2(SiMe2Ph) (5). Complex 5 was identified by spectroscopy
and by X-ray crystallography and was synthesized independently from 3 and CO. The
chemistry presented shows that metalation of silyl ligands can be a facile reversible process
that takes place under mild conditions. It also indicates that iridasilacycles can be used as
a masked form of electron-rich Ir(I) silyl complexes, which are not readily accessible species.
In tr od u ction
available due to C-H reductive elimination which
occurs under mild conditions, and (iii) they can be
trapped by H2 or CO. We also report the first X-ray
structural characterization of an iridium(I) silyl com-
plex, Ir(PMe3)2(CO)2(SiMe2Ph), which is not stabilized
by chelation. Part of these results has been com-
municated previously.5
Intramolecular activation of ligand C-H bonds lead-
ing to the formation of metallacycles is a common
reaction of transition-metal complexes.1 It is well-
established for different ligand types such as hydrocar-
byls, as well as N-, P-, As-, Sb-, O-, and S-bound
ligands.1 However, there are only a few examples of
metallacycles resulting from the metalation of silyl
Exp er im en ta l Section
ligands. These, to our knowledge, include (dcpe)Pt-
(C6H4)Si(SiMe3)2, prepared by Fink and co-workers,2 fac-
(PMe3)3Ir(o-C6H4SiPh2)(H) and fac-(PMe3)3Ir(CH2CH2-
OSi(OEt)2)(H), which we reported earlier,3 and fac-
Gen er a l Con sid er a tion s. Methods of handling air-sensi-
tive compounds which were employed in this work are
described in detail elsewhere.3b Deuterium gas (Matheson,
99.9% D) and carbon monoxide (99.9%, from local sources) were
used without further purification. HSiMe2Ph (98%, Aldrich)
was degassed by purging with dry N2 before use. The NMR
spectra were measured using a Bruker AMX 400 spectrometer
at 400 MHz (1H), 162 MHz (31P), and 100 MHz (13C) and
referenced as described earlier.3b 2H NMR spectra (61.4 MHz)
were externally referenced to the signal of C6D6 at δ 7.15 ppm.
FD-MS spectra were obtained from the Institute of Mass
Spectrometry, University of Amsterdam, Amsterdam, The
Netherlands. Microanalyses were performed by the Mi-
croanalysis Laboratory of The Hebrew University of J erusa-
lem.
(P M e 3 )3 I r (S i M e 2 S i M e (S i M e 3 )S i M e 2 C H 2 )(H ), 4
communicated very recently by Tilley et al. The reac-
tivity of these complexes has been studied very little,
the only accessible information being hydrolytic cleav-
age of the Si-Ph bond in Fink’s compound2 and hydro-
gen-catalyzed equilibration of the diastereomeric mix-
ture of isomers of Tilley’s iridasilacycle.4 We report here
our results on the reversibility of cyclometalation of
arylsilyl ligands in (trimethylphosphine)iridium com-
plexes. We show that: (i) this electron-rich system does
not require elevated temperatures for the ortho meta-
lation to occur, ( ii) reactive (PMe3)3IrSiR2Ph species are
P r ep a r a tion of fa c-(P Me3)3Ir (o-C6H4SiMe2)(H) (1). To
a stirred suspension of 52 mg (0.098 mmol) of Ir(PMe3)4Cl6 in
5.4 mL of THF was dropwise added 0.21 mL of a 0.62 M
solution of Me2PhSiLi7 in THF. The mixture was stirred for
10 min at room temperature, after which the solvent was
X Abstract published in Advance ACS Abstracts, J une 15, 1996.
(1) Reviews: (a) Parshall, G. W. Acc. Chem. Res. 1970, 3, 139. (b)
Bruce, M. I. Angew. Chem., Int. Ed. Engl. 1977, 16, 73. (c) Ryabov, A.
D. Chem. Rev. 1990, 90, 403.
(2) Chang, L. S.; J ohnson, M. P.; Fink, M. Organometallics 1991,
10, 1219.
(3) (a) Aizenberg, M.; Milstein, D. Angew. Chem., Int. Ed. Engl.
1994, 33, 317. (b) J . Am. Chem. Soc. 1995, 117, 6456.
(4) Mitchell, G. P.; Tilley, T. D.; Yap, G. P. A.; Rheingold, A. L.
Organometallics 1995, 14, 5472.
(5) (a) Aizenberg, M.; Milstein, D. Abstracts of the 9th International
Symposium on Homogeneous Catalysis. J erusalem, Israel, Aug 21-
26, 1994; p 297. (b) Abstracts of the 60th Annual Meeting of the Israel
Chemical Society, Rehovot, Israel, Feb 7-8, 1995; p 136.
(6) Herskovitz, T. Inorg. Synth. 1982, 21, 99.
(7) (a) Gilman, H.; Peterson, D. J .; Wittenberg, D. Chem. Ind. 1958,
1479 (Nov 8). (b) George, M. V.; Peterson, D. J .; Gilman, H. J . Am.
Chem. Soc. 1960, 82, 403.
S0276-7333(96)00214-2 CCC: $12.00 © 1996 American Chemical Society
+
+

3318 Organometallics, Vol. 15, No. 15, 1996
Aizenberg and Milstein
removed under vacuum. The residue that formed was ex-
tracted with 5 mL of pentane. The extract after filtration was
evacuated to dryness to produce a yellow oily material. The
oil was recrystallized from a minimum amount of pentane to
yield after 2 days at -20 °C 26 mg (48%) of pure 1 in the form
of colorless crystals. Anal. Calcd for C17H38IrP3Si: C, 36.74;
H, 6.89. Found: C, 36.97; H, 6.73. FD-MS (Ir-193/191): m/z
556/554. 31P{1H} NMR (C6D6): δ -62.4 (dd with 29Si satellites,
of 10 (C6D6): δ -45.4 (d of pseudo t, 2J P-P,cis ) 20.3 Hz, 2J P-D,cis
2
) 3.1 Hz, 2P, mutually trans), -54.4 (t of pseudo t, J P-P,cis
)
2
1
20.3 Hz, J P-D,trans ) 22.3 Hz, 1P trans to D). The H NMR of
10 is essentially the same as that reported3 for its non-
deuterated isotopomer, except for the absence of the signals
due to the hydride and the C6H5 groups.
Rea ction of 1 a n d 2 w ith H2. Solutions of 1 and 2 (0.01
mmol) in C6D6 (0.4 mL) were pressurized with 80 psi of
hydrogen in a high-pressure 5 mm NMR tube. The tubes were
placed in an oil bath and were heated at 80 °C for 20 h (1) or
40 h (2), after which quantitative formation of 3 and 4
respectively was detected by 31P{1H} and 1H NMR.
Rea ction of 1 w ith D2. This reaction was carried out
completely analogously to the reaction of 1 with H2. Charac-
terization of 3-D2: 31P{1H} NMR (C6D6) δ -61.7 (m, 1P trans
to Si), -57.4 (m, 2P trans to D); 1H NMR is the same as for 3,
except for the absence of the multiplet at δ -12.22 due to the
hydrides; 2H{1H} NMR (C6H6) δ -12.2 (br symmetrical m, J 1
) 14.4 Hz, J 2 ) 3.2 Hz, Ir-D).
2
J 1 ) 21.3 Hz, J 2 ) 14.9 Hz, J P-Si,trans ) 127 Hz, 1P trans to
Si); -59.1 (dd, J 1 ) 17.6 Hz, J 2 ) 14.9 Hz, 1P trans to C);
-55.2 (dd, J 1 ) 21.3 Hz, J 2 ) 17.7 Hz, 1P trans to H). 1H
NMR (20 °C, C6D6): δ -10.32 (d of pseudo t, 2J H-P,trans ) 124.5
Hz, 2J H-P,cis ) 19.4 Hz, 1H, Ir-H); 0.72 (d, 4J H-P,trans ) 3.0 Hz,
3H, Si-CH3; the signal of another CH3 group overlaps with
the signal of PMe3 at δ 1.01 ppm but shifts upfield and is easily
observed at 60 °C as a doublet at δ 0.95 ppm (4J H-P,trans ) 2.5
2
2
Hz, 3H); 1.01 (d, J H-P ) 7.4 Hz, 9H, P(CH3)3); [1.27 (d, J H-P
2
) 7.1 Hz, 9H, P(CH3)3); 1.29 (d, J H-P ) 6.8 Hz, 9H, P(CH3)3);
these two signals appear as a pseudotriplet but are clearly
distinguished as two doublets at δ 1.32 and 1.33 ppm at 60
°C]; [7.18 (m), 7.29 (m), 4H, C6H4]. 13C{1H} NMR ((CD3)2CO):
δ 1.0 (dd, J 1 ) 9.2 Hz, J 2 ) 1.1 Hz; Si-CH3); 7.3 (m, J 1 ) 7.3
Hz, J 2 ) 3.8 Hz, J 3 < 1 Hz, unresolved; Si-CH3); 20.3 (ddd,
J 1 ) 27.8 Hz, J 2 ) 5.8 Hz, J 3 ) 3.0 Hz; P(CH3)3); 23.5 (dt, J d
) 22.6 Hz, J t ) 2.9 Hz; P(CH3)3); 25.0 (ddd, J 1 ) 29.3 Hz, J 2
) 5.0 Hz, J 3 ) 3.5 Hz; P(CH3)3); [119.1 (br), 126.1 (br), 127.0
Rea ction of 1 w ith CO. F or m a tion of Ir (P Me3)2(CO)2-
(SiMe2P h ) (5). A solution of 1 (0.02 mmol) in C6D6 (0.4 mL)
was pressurized with 40 psi of CO in a high-pressure 5 mm
NMR tube. The tube was heated in an oil bath at 85 °C for
40 h, after which 31P{1H}, 1H, and 13C{1H} NMR indicated
quantitative formation of 5 and liberation of 1 equiv of PMe3.
The solvent was removed under vacuum to afford 5 as a
colorless solid in essentially quantitative yield. Colorless
needles suitable for X-ray analysis were obtained by slow
evaporation of a concentrated pentane solution of 5 at room
temperature. FD-MS (Ir-193/191): m/z 536/534. IR (film): νCO
(cm-1) ) 1896 vs, 1905 s, 1932 sh, 1945 vs. 31P{1H} NMR
(br), 134.6 (br), 137.6 (ddd, J 1 ) 84.2 Hz, J 2 ) 10.2 Hz, J 3
7.6 Hz), 168.9 (dt, J t ) 10.7 Hz, J d ) 2.2 Hz); C6H4].
)
P r ep a r a tion of fa c-(P Me3)3Ir (SiMe2P h )(H)2 (3). To a
8
solution of 35 mg (0.07 mmol) of HIr(PMe3)4 in 2 mL of
benzene was added 11.5 µL (0.075 mmol) of HSiMe2Ph. The
yellow solution was left overnight at room temperature, after
which it became colorless. The solvent was removed under
vacuum to yield 43 mg (90%) of pure 3 as a waxy white solid.
Anal. Calcd for C17H40P3SiIr: C, 36.61; H, 7.23. Found: C,
36.54, H, 7.48. 31P{1H} NMR (C6D6): δ -61.7 (t with 29Si
2
(C6D6): δ -64.6 (s with 29Si satellites, J P-Si ) 30.5 Hz). 1H
NMR (C6D6): δ 1.06 (second-order pseudo d, J ) 8.3 Hz, 18H;
4
2P(CH3)3), 1.13 (t, J H-P ) 1.1 Hz, 6H; Si(CH3)2), [7.18 (m,
1Hpara), 7.32 (t, J ) 7.5 Hz, 2Hmeta), 8.03 (m, 2Hortho); Si-C6H5].
13C{1H} NMR (C6D6): δ 8.05 (t, J ) 3.8 Hz; Si(CH3)2); 23.16
(symmetrical six-line pattern of X part of an ABX spin system;
2P(CH3)3); [127.57 (s), 127.59 (s), 134.55 (s), the fourth signal
2
satellites, J ) 20.4 Hz, J P-Si,trans ) 128 Hz, 1P trans to Si);
-57.4 (d, J ) 20.4 Hz, 2P trans to H). 1H NMR (C6D6): δ
-12.22 (symmetrical second-order m, J 1 ) 95.5 Hz, J 2 ) 20.3
2
was not found; Si-C6H5], 188.57 (t, J C-P ) 5.5 Hz; Ir-CO).
4
Hz, 2H; Ir-H); 0.99 (d, J H-P,trans ) 2.3 Hz, 6H; Si(CH3)2; 1.19
Rea ction of 3 w ith CO. The procedure and amounts were
2
2
(d, J H-P ) 7.4 Hz, 18H; 2P(CH3)3, trans to H); 1.26 (d, J H-P
) 7.6 Hz, 9H; P(CH3)3, trans to Si); [7.20 (tt, J 1 ) 7.3 Hz, J 2
) 1.3 Hz, 1Hpara), 7.38 (t, J ) 7.4 Hz, 2Hmeta), 8.08 (m, 2Hortho);
Si-C6H5].
exactly the same as for the reaction of 1 with CO. The reaction
1
was much slower and less selective. The 31P{1H} and H NMR
spectra were periodically checked to show after 312 h, when
the reaction was interrupted, that the reaction mixture
contained 67% of 5, 18% of (PMe3)2Ir(CO)2(H) (8), and 9% of
(PMe3)2Ir(CO)(H)2(SiMe2Ph) (9), along with 6% of unreacted
3. Complexes 8 and 9 were tentatively assigned on the basis
of the following signals in the NMR spectra. 8: 31P{1H} NMR
P r ep a r a tion of fa c-(P Me3)3Ir (CH3)(H)(SiMe2P h ) (6).
This complex was prepared in complete analogy with its
triethyl-, triphenyl-, and triethoxysilyl analogs, for which we
reported full characterization data, including X-ray crystal
structures of the first two of them.3 6 was isolated from its
benzene solution in quantitative yield as a white powder. FD-
MS (Ir-193/191): m/z 572/570. 31P{1H} NMR (C6D6): δ -63.4
1
2
(C6D6) δ -62.0 (br s); H NMR (C6D6) δ -11.31 (t, J H-P,cis
)
28.4 Hz, 1H, Ir-H). 9: 31P{1H} NMR (C6D6) δ -62.6 (d, J )
21 Hz, 1P), -60.2 (d, J ) 21 Hz, 1P); H NMR (C6D6) δ -10.9
1
2
2
2
(dd with 29Si satellites, J 1 ) 20.5 Hz, J 2 ) 17.6 Hz, J P-Si,trans
(d of br m, J H-P,trans ≈ 111 Hz, J H-P,cis ≈ 22 Hz, 1H, Ir-H),
2
-10.46 (symmetrical five-line m (in fact, ddd), J H-P,cis ) 23
) 153 Hz, 1P trans to Si); -58.9 (pseudo t, J ) 17.4 Hz, 1P
trans to CH3); -58.5 (dd, J 1 ) 20.5 Hz, J 2 ) 17.2 Hz, 1P trans
2
2
Hz, J H-P,cis ) 19 Hz, J H-H,cis ) 4 Hz, 1H, Ir-H).
2
to H). 1H NMR (C6D6): δ -11.91 (d of pseudo t, J H-P,trans
)
X-r a y Cr ysta llogr a p h ic An a lysis of 5. The structure of
compound 5 was determined using a Rigaku AFC5R four-circle
diffractometer with Mo KR radiation (graphite monochroma-
tor; λ ) 0.710 73 Å). Unit cell dimensions were determined
from 25 reflections. Details of crystal parameters and data
collection are given in Table 1. Three standards were collected
19 times each with a 4% change of intensity. The structure
was solved by the Patterson method (SHELXS-929a) and
refined using full-matrix least-squares refinement based on
F2 (SHELXL-939b). The data were not corrected for absorption.
Hydrogens were calculated from a difference Fourier map and
refined in a riding mode. For 207 parameters and 3384 data
with no restraints, the following final discrepancy factors were
obtained: R1 ) 0.0360 (based on F2), wR2 ) 0.0577 (I > 2σ-
(I)); R1′ ) 0.0923, wR2′ ) 0.0751 (all data); GOF (on F2) )
2
3
135.3 Hz, J H-P,cis ) 19.4 Hz, 1H, Ir-H); 0.26 (m, J H-H ) 1.1
Hz, 3H, Ir-CH3); 0.85 (d, 4J H-P,trans ) 1.7 Hz, 3H, Si-CH3); 0.88
4
2
(d, J H-P,trans ) 2.4 Hz, 3H, another Si-CH3); 1.09 (d, J H-P
)
2
7.0 Hz, 18H; 2P(CH3)3, coincidence); 1.11 (d, J H-P ) 7.4 Hz,
9H; P(CH3)3); [7.21 (tt, J 1 ) 7.3 Hz, J 2 ) 1.4 Hz, 1Hpara), 7.37
(t, J ) 7.5 Hz, 2Hmeta), 8.07 (m, 2Hortho); Si-C6H5].
Th er m olysis of 6. A solution of 12 mg (0.02 mmol) of 6 in
0.5 mL of C6D6 was heated in a screw-capped NMR tube at 95
1
°C in an oil bath. 31P{1H} and H NMR spectra of the reaction
mixture were periodically checked to show complete disap-
pearance of the starting material after 2 days. The resulting
mixture contained ∼70% of 1, ∼15% of 3, and ∼15% of
mer,trans-(PMe3)3Ir(C6D5)2(D) (10), which were identified by
their NMR spectra. The 1H NMR spectrum exhibited also
signals due to CH4 (δ 0.15 ppm), (CH3)3SiPh (δ 0.19 ppm), and
CH3D (δ 0.14 ppm, 1:1:1 triplet with J ) 2 Hz). 31P{1H} NMR
(9) (a) Sheldrick, G. M. SHELXS92, Program for crystal structure
solution; University of Go¨ttingen, Go¨ttingen, Germany, 1992. (b)
Sheldrick, G. M. SHELXL93, Program for crystal structure refinement;
University of Go¨ttingen, Go¨ttingen, Germany, 1993.
(8) Thorn, D. L.; Tulip, T. H. Organometallics 1982, 1, 1580.
+
+

Reversible Cyclometalation of Silyl Ligands
Organometallics, Vol. 15, No. 15, 1996 3319
Ta ble 1. Cr ysta llogr a p h ic P a r a m eter s for th e
Str u ctu r e of 5
Ta ble 3. Selected Bon d An gles (d eg) in a Molecu le
of 5
empirical formula
formula mass
color and habit
cryst size
cryst system
space group
a, Å
C16H29O2P2SiIr
535.62
C(10)-Ir(1)-C(11) 137.0(5)
C(10)-Ir(1)-Si(2)
C(11)-Ir(1)-Si(2)
P(3)-Ir(1)-Si(2)
P(4)-Ir(1)-Si(2)
84.7(2)
84.8(2)
94.25(10)
166.91(10)
C(10)-Ir(1)-P(3)
C(11)-Ir(1)-P(3)
C(10)-Ir(1)-P(4)
C(11)-Ir(1)-P(4)
P(3)-Ir(1)-P(4)
112.4(3)
109.9(3)
88.5(2)
92.8(2)
colorless needles
0.2 × 0.2 × 0.4
monoclinic
P21/c (No.14)
10.456(2)
8.104(2)
25.028(3)
100.10(3)
2087.9(8)
4
O(10)-C(10)-Ir(1) 178.0(9)
98.67(7) O(11)-C(11)-Ir(1) 177.7(11)
b, Å
c, Å
produce 1. However, this route is complicated by the
competitive C-Si and H-Si reductive eliminations,3
which lead to the formation of about 30% of undesired
products (see Experimental Section for details). Still,
crude 1 can be recrystallized from cold pentane and
obtained pure.
â, deg
V, Å3
Z
d(calcd), g/cm3
µ, mm-1
1.704
6.609
diffractometer
radiation (wavelength, Å)
monochromator
temp, K
Rigaku AFC5R
Mo KR (λ ) 0.710 73)
graphite
A better way to obtain 1 involves a nucleophilic attack
of (dimethylphenylsilyl)lithium at Ir(PMe3)4Cl (eq 2).
110
mode
ω
θmax, deg
27.53
scan speed, deg/min
scan width, deg
collcn range
8
1.2
-13 e h e -3, 0 e k e 10,
-29 e l e 32
no. of reflcns
collcd
3445
indepdt
no. of variables
R1
3389
207
0.0360
0.0570
wR2
This reaction presumably generates the intermediate
7 at room temperature in THF.10 Removal of the
solvent under vacuum followed by the pentane workup
affords pure 1 in the form of colorless crystals.
Ta ble 2. Selected In ter a tom ic Bon d Len gth s (Å) in
a Molecu le of 5
Ir(1)-C(10)
Ir(1)-C(11)
Ir(1)-P(3)
Ir(1)-P(4)
Ir(1)-Si(2)
Si(2)-C(27)
Si(2)-C(21)
Si(2)-C(28)
1.873(10)
1.887(10)
2.340(3)
2.353(2)
2.442(3)
1.863(12)
1.890(10)
1.890(10)
P(3)-C(33)
P(3)-C(31)
P(3)-C(32)
P(4)-C(42)
P(4)-C(43)
P(4)-C(41)
O(10)-C(10)
O(11)-C(11)
1.806(11)
1.829(9)
1.831(7)
1.794(12)
1.801(11)
1.798(10)
1.149(11)
1.146(12)
The composition and stereochemistry of 1 were elu-
1
cidated from its 31P, H, and 13C NMR spectra and FD-
MS and microanalysis data. Three doublets of doublets
(one of them having 29Si satellites11 ) in the 31P{1H}
NMR spectrum together with appropriately split signals
1
in H and 13C{1H} NMR spectra indicate facial arrange-
1.042. The largest residual peak and hole had values of 0.875
Selected interatomic bond lengths and
angles are presented in Tables 2 and 3.
ment of the PMe3 ligands in the complex. The octahe-
dral coordination around the iridium center is completed
by the silicon, the ortho-metalated aryl fragment, and
and -0.958 e Å-3
.
1
the hydride, which gives rise in the H NMR spectrum
Resu lts a n d Discu ssion
to a widely spaced doublet of triplets at δ -10.32 ppm.
Importantly, such a stereochemistry makes the two
silicon-bound methyl groups inequivalent, as clearly
seen in 1H (doublets at δ 0.72 and 1.02 ppm) and 13C-
{1H} (doublet of doublets at δ 1.0 and a multiplet at 7.3
ppm) NMR spectra. As expected, once the rigid sila-
metallacycle is formed, the six carbons of the aromatic
ring become chemically and magnetically inequivalent
as well and give rise to six signals in the 13C{1H} NMR
spectrum. The signal due to the metalated carbon atom
appears at δ 137.6 ppm as a doublet of doublets of
Syn th esis of Ir id a sila cycle 1. The compound fac-
(PMe3)3Ir(o-C6H4SiMe2)(H) (1) can be obtained in two
different ways. The first one is completely analogous
to the synthesis which we reported for its analog 23 (eq
1). When complex 6, which has a methyl, a hydride,
2
doublets with a large J C-P,trans value of 84.2 Hz and
2
two different small J C-P,cis values of 10.2 and 7.6 Hz.
The facility of the ortho metalation that leads to 1 is
noteworthy. It clearly indicates that the heating used
in the synthesis of 1 from 6 is necessary not for the
intramolecular C-H activation of the dimethylphenyl-
silyl group in 7 but, rather, for inducing reductive
elimination of methane from 6. We note that an
analogous metalation (preceded by a series of rear-
(10) The analogous reaction of Rh(PMe3)4Cl leads to the stable
nonmetalated Rh(I) analog of 7, namely (PMe3)3Rh(SiMe2Ph); see:
Aizenberg, M.; Milstein, D. Science 1994, 265, 359.
(11) For the range of coupling constants J P-Si,cis and J P-Si,trans see,
for example: (a) Ozawa, F.; Hikida, T.; Hayashi, T. J . Am. Chem. Soc.
1994, 116, 2844. (b) Zarate, E. A.; Kennedy, V. O.; McCune, J . A.;
Simons, R. S.; Tessier, C. A. Organometallics 1995, 14, 1802.
and a dimethylphenylsilyl group in mutually cis posi-
tions, is heated to 95 °C for 2 days, C-H reductive
elimination occurs. It leads to the intermediate (PMe3)3-
Ir(SiMe2Ph) (7; not observed), which ortho-metalates to
2
2
+
+

3320 Organometallics, Vol. 15, No. 15, 1996
Aizenberg and Milstein
rangements) of an aliphatic C-H bond of the tris-
(trimethylsilyl)silyl ligand at the same (PMe3)3Ir center
was also shown to proceed at ambient and even lower
temperatures.4
Rea ctivity of Com p lexes 1 a n d 2. (a ) Rea ction s
w ith H2 a n d D2. The iridasilacycles 1 and 2 are easily
isolable, very stable compounds. They are not sensitive
to air and moisture in the solid state. Moreover,
complex 1 remains unchanged after prolonged (several
days) heating at 100-110 °C in cyclohexane. However,
these silametallacycles react with hydrogen. When
benzene solutions of 1 or 2 are placed under 80 psi of
hydrogen at 80 °C, the dihydridosilyliridium(III) com-
plexes 3 and 4 are quantitatively formed (eq 3). Com-
2
two broad, heavily distorted multiplets due to H-31P
couplings. Their chemical shift values exactly match
those of the non-deuterated 3. (2) In the 1H NMR
spectrum signals of the ortho protons of the phenyl ring
at δ 8.08 ppm were not affected with respect to both
the line shape and integrated intensity. At the same
time the signal due to the hydrides at δ -12.22 ppm
2
was negligible. (3) The H{1H} NMR spectrum showed
essentially no incorporation of the deuterium label into
the ortho positions of the phenyl ring, while a broad
symmetrical multiplet appeared at δ -12.2 ppm. Its
chemical shift as well as the apparent coupling con-
stants (J 1 ) 14.4, J 2 ) 3.2 Hz) fit those expected for
the dideuteride 3-D2. (4) The reaction between 3 and
D2 carried out under exactly the same conditions did
not result in any appreciable Ir-(H)2/D2 exchange.
After 20 h at 80 °C the signals of 3 at δ -61.7 and -57.4
ppm in the 31P{1H} NMR spectrum remained a sharp
triplet and doublet, respectively, the signal at δ -12.22
plex 4 is known. It was synthesized from HIr(PMe3)4
and HSiPh3 and was characterized by spectroscopy and
by X-ray crystallography.11b For the purposes of iden-
tification, complex 3 was synthesized independently
using a completely analogous reaction of HIr(PMe3)4
with HSiPhMe2 and was fully characterized.
Formation of the dihydrides 3 and 4 from the sila-
cycles 1 and 2 may be rationalized in two ways. The
first one invokes C-H reductive elimination, i.e., open-
ing of the cycle, with subsequent H-H oxidative addi-
tion to the intermediate 7 (path A). The second one
involves H-H oxidative addition, i.e., formation of an
Ir(V) intermediate, prior to the opening of the cycle
(path B). Iridium(V) complexes with hydrides and silyls
as ligands are not very common but are known.12 On
the other hand, very electron-rich octahedral complexes
of Ir(III) with tightly bound small ancillary ligands, like
the ones described herein, seem not to be very well
suited for very facile reductive eliminations,3 which
would result in the formation of Ir(I) species. This
mechanistic dilemma was resolved by using deuterium
instead of hydrogen. If the silacycle first opens up and
deuterium adds to the resulting Ir(I), then nonlabeled
Si-C6H5 and iridium dideuteride would be formed. If
the D-D oxidative addition precedes the C-H reductive-
elimination step, then one would expect scrambling of
the deuterium label between the ortho position of the
aromatic ring and the iridium hydrides. The latter case
requires also that there should be no fast secondary
exchange between deuterium gas and the dihydridosi-
lyliridium complex.13 Examination of 31P{1H}, 1H, and
2H{1H} NMR spectra of the product of the reaction
between 1 and D2 unequivocally demonstrated that path
A is operative and the dideuteridosilyl species 3-D2 is
formed (eq 4).
1
ppm in the H NMR spectrum due to the hydrides did
not diminish in integration, and the intensity of Ir-D
2
signals in the H{1H} NMR spectrum was negligible.
The described facility of C-H reductive elimination
leading to the opening of the iridasilacycles 1 and 2 is
somewhat surprising. For example, it is known that the
closely related complex fac-(PMe3)3Ir(CH2C(CH3)2CH2),
which has a four-membered Ir-C-C-C ring, is ex-
tremely stable.14 It is nonreactive with C6D6 at 90 °C
and with CO at moderate pressures14a and survives
when heated at 150 °C for days.14b The difference
between this complex, on one hand, and complexes 1
and 2, on the other, is even more striking, since in the
former only sp3 carbons are involved in metalation.
Activation of C-H bonds of these carbons is expected
to be both kinetically and thermodynamically less
favorable than that of sp2 carbons of 1 and 2. Still, the
metalation of (PMe3)3Ir(CH2E(CH3)3) (E ) C, Si) is
essentially irreversible up to 80 °C,15 while 1 and 2 are
quite reactive at this temperature. It was noted that
an increase of steric bulk at a tris(phosphine)iridium
center can facilitate both cyclometalation and opening
of metallacycles. The latter was also associated with
the strain relief of the four-membered ring.14,15 We
believe that 1 and 2 (as well as their precursors) are
not extremely sterically congested. Their four-mem-
bered rings include silicon which is larger in volume
than carbon, probably resulting in less strain. Thus,
the reasons for the easy C-H reductive eliminations
This conclusion is based on the following observations.
(1) The 31P{1H} NMR spectrum of the product exhibited
(12) (a) Fernandez, M.-J .; Maitlis, P. J . Chem. Soc., Dalton Trans.
1984, 2063. (b) Tanke, R. S.; Crabtree, R. H. Organometallics 1991,
10, 415. (c) Loza, M.; Faller, J . W.; Crabtree, R. H. Inorg. Chem. 1995,
34, 2937. (d) Esteruelas, M. A.; Lahoz, F. J .; On˜ate, E.; Oro, L. A.;
Rodriguez, L. Organometallics 1996, 15, 823.
(13) For example, Ir(H)2(Et3Si)(CO)(dppe) was reported to undergo
75% exchange of hydrides with D2 at 60 °C over 24 h: Hays, M. K.;
Eisenberg, R. Inorg. Chem. 1991, 30, 2623.
(14) (a) Tulip, T. H.; Thorn, D. L. J . Am. Chem. Soc. 1981, 103, 2448.
(b) Calabrese, J . C.; Colton, M. C.; Herskovitz, T.; Klabunde, U.;
Parshall, G. W.; Thorn, D.; Tulip, T. H. Ann. N.Y. Acad. Sci. 1983,
415, 302.
(15) Harper, T. G. P.; Desrosiers, P. J .; Flood, T. C. Organometallics
1990, 9, 2523.
+
+

Reversible Cyclometalation of Silyl Ligands
Organometallics, Vol. 15, No. 15, 1996 3321
that 1 and 2 undergo remain unclear and may have to
do with an electronic effect of the silyl group.
(b) Rea ction of 1 w ith CO. F or m a tion of th e Ir -
SiMe2P h Com p lex 5. To make use of the above
described ability of 1 to open up, generating the tran-
sient iridium(I) silyl intermediate 7, we carried out the
reaction of 1 with carbon monoxide with the hope of
obtaining a stable Ir(I) silyl complex. We note that there
are very few reported examples of isolated silyl com-
plexes of iridium(I).16 Subjecting a benzene solution of
1 to 40 psi of CO gas at 85 °C in a high-pressure NMR
tube resulted in liberation of 1 equiv of PMe3 and
quantitative formation of bis(phosphine)dicarbonyl(dim-
ethylphenylsilyl)iridium(I) (5) as the sole organometallic
product (eq 5). It was identified using IR and NMR
spectroscopy, and its structure was verified by X-ray
diffraction.
Complex 5 exhibits in 31P{1H} NMR a sharp singlet
resonance with 29Si satellites at δ -64.6 ppm which
remains sharp even at -70 °C. Since the P(CH3)3
1
protons do not give rise in H NMR to a virtual triplet
signal,17 this is indicative of a very fast intramolecular
dynamic process which equilibrates the two phosphines.
The 31P chemical shift value is typical for other satu-
rated 5-coordinate Ir(I) trimethylphosphine complexes.
The 29Si satellites indicate the presence of a silyl ligand
bound to the iridium center. The value of 2J P-Si ) 30.5
Hz is much smaller than the ones observed for rigid
hexacoordinate Ir(III) complexes with the same silyl
F igu r e 1. Kinetic followup of the reaction of 3 (0.05 M)
with 40 psi of CO in C6D6 at 85 °C: (a) plot of ln([3]) versus
time; (b) plots of [3], [5], [8], and [9] versus time.
disappearance of 3 is given in Figure 1. It is first order
in [3] with kobs ) 9.18 × 10-3 h-1 (R ) 0.999). After
312 h of monitoring by NMR the reaction was inter-
rupted, resulting in a mixture of 67% of 5, 6% of
unchanged 3, and two other products which on the basis
of their NMR signals were tentatively assigned as
(PMe3)2Ir(CO)2(H) (8; 18%) and (PMe3)2Ir(CO)(H)2(SiMe2-
Ph) (9; 9%) (Figure 1b). Free PMe3, H2, and HSiMe2Ph
2
ligand (compare J P-Si,trans ) 127 Hz for 1, 128 Hz for
3, and 153 Hz for 6). This may be a result of 5 being in
solution a fluxional molecule without a permanent trans
arrangement of the silyl and one phosphine ligand.
There is only one signal in the carbonyl region of 13C-
{1H} NMR of 5 at δ 188.57. It is split into a triplet with
2J C-P ) 5.5 Hz, indicating again the apparent equiva-
lence of the two phosphorus nuclei. The IR spectrum
exhibits two strong bands for the coordinated carbonyl
groups with νCO values of 1896 and 1945 cm-1 (the
former is additionally split at its top into two bands,
while the latter has a shoulder at 1932 cm-1). This
definitely rules out a structure with a single carbonyl
ligand as well as a configuration with two carbonyls
arranged trans to each other. The final proof for the
structure of 5 was obtained from an X-ray crystal-
lographic analysis (see below).
were detected in 31P and 1H NMR. These results
together with the lack of exchange between 3 and D2
imply that H-H and H-Si reductive eliminations from
3 do not occur under the conditions studied. Only after
predissociation of one of the tightly bound PMe3 ligands
from 3 and CO coordination can these reactions take
place (leading to 5 and 8), the overall process being
much slower than the corresponding reactions of the
iridasilacycles 1 and 2.
Descr ip tion of th e Str u ctu r e of 5. Slow evapora-
tion of a concentrated pentane solution of 5 yielded
colorless needles suitable for an X-ray crystallographic
study. A perspective view of the solid-state structure
of 5 is presented in Figure 2. As seen from Figure 2
and from Tables 2 and 3, the coordination geometry
around the iridium center is trigonal bipyramidal with
the silyl and one of the phosphine ligands in the axial
positions, the second phosphine and two carbonyls
Alter n a tive Rou te to 5. Having synthesized 5 from
1, one would suppose that 5 could also be formed in a
formally analogous reaction of H-H reductive elimina-
tion from 3 under a CO atmosphere. Indeed, this
reaction is feasible, but under the same conditions it is
much slower and less selective than reaction of 1 with
CO. The kinetic followup by 31P{1H} NMR of the
(16) (a) Auburn, M. J .; Grundy, S. L.; Stobart, S. R.; Zaworotko, M.
J . Am. Chem. Soc. 1985, 107, 266. (b) Esteruelas, M. A.; Lahoz, F. J .;
Oliva´n, M.; On˜ate, E.; Oro, L. A. Organometallics 1994, 13, 4246.
(17) The closely related Rh analog of 5, namely Rh(PMe3)2(CO)2-
(SiPh3), is known. According to its 1H NMR the P(CH3)3 ligands are
mutually trans and give rise to a virtual triplet. See: Thorn, D. L.;
Harlow, R. L. Inorg. Chem. 1990, 29, 2019.
+
+

3322 Organometallics, Vol. 15, No. 15, 1996
Aizenberg and Milstein
temperatures and exhibits no reactivity. Hydrogen does
not add to 5 as well. For a comparison, related square-
planar bis(trimethylphosphine) complexes of Ir(I) that
have only one CO ligand, such as trans-(PMe3)2Ir(CO)X
(X ) Cl, CH3, Ph), all rapidly add H2 at room temper-
ature and 1 atm, resulting in the corresponding dihy-
drides.19 Cis-phosphine analogs (dppe)Ir(CO)X20 and
(dfepe)Ir(CO)H21 do so as well. Importantly, even the
coordinatively saturated (PCy3)Ir(CO)3(SiPh3), (dfepe)-
Ir(CO)2(H), and (PMe3)2Ir(CO)2Cl, which have phos-
phines or another ligand of lower σ-donating capacity,
readily lose one CO in solution16b,21,22 and are, therefore,
capable of reacting with H2. The observed lack of
reactivity of 5 is undoubtedly a consequence of its
retaining the saturated 18-electron structure in solution.
The synergism between the very strong σ-donor capabil-
ity of both PMe3 and SiMe2Ph groups and the π-acceptor
character of CO ligands results in the very high stability
of 5 with respect to ligand dissociation and, hence, in
its inertness in oxidative-addition reactions.23
Con clu sion s
In this paper we have presented an example of
metalation of a silyl ligand which occurs at the (PMe3)3-
Ir center under very mild conditions. Moreover, we
have provided the first conclusive demonstration of
reversibility of such a process. In contrast to the very
well-established reversal of metalation of phosphine
ligands,24 this reaction was unknown for silyls. We have
also described synthetic ways to, and measured the
X-ray structure of, a coordinatively saturated Ir(I) silyl
complex not stabilized by chelation, which is remarkable
in its stability with respect to ligand dissociation. We
continue to explore possible ways of synthesizing coor-
dinatively unsaturated iridium(I) silyls. We also con-
tinue investigating the chemistry of such complexes
generated in situ.
F igu r e 2. Perspective view of the coordination geometry
of 5. Hydrogen atoms are omitted for clarity.
occupying three equatorial sites. Although in the solid-
state structure of 5 the two phosphines are cis to each
other and are obviously different, such a molecule is
expected to be stereochemically nonrigid in solution,
which results in the NMR spectral data described above.
There are several important features of the structure
of 5. It is worthy of note that the arrangement of the
ligands is the same as that reported by Stobart for the
chelate-stabilized complex [Ir(PPh2CH2CH2SiMe2)(PPh3)-
(CO)2].16a The Ir-Si distance of 2.442(2) Å is long. It
is longer than typical Ir-Si bonds in Ir(III) complexes,
both octahedral and pentacoordinate,18 as well as those
in seven-coordinate Ir(V) silyls.12c The aforementioned
compound of Stobart has a long Ir-Si bond as well
(2.454(6) Å16a). The anchoring of the silyl ligand in it
results in an almost linear arrangement of Ptrans-Ir-
Si (the corresponding angle has a value of 175.7(2)°16a),
while in 5 the P(4)-Ir(1)-Si(2) angle equals 166.91(10)°.
This may decrease the degree to which the trans
influence of the phosphorus is transmitted, leading to
a shorter Ir-Si bond. To our knowledge, 5 is the second
Ir(I) silyl to be structurally characterized. It is the first
structure of an Ir(I) silyl that is not stabilized by
chelation.
Ch em ica l P r op er ties of 5. One would expect that
5, if capable of ligand dissociation in solution, should
be quite reactive in oxidative-addition processes, as is
the case for other Ir(I)-trimethylphosphine complexes.
We were quite surprised to find out that a cyclopentane
solution of 5 is not sensitive to either atmosperic oxygen
or moisture. Moreover, a 30-fold excess of MeI did not
react with 5 in cyclopentane over 1 day. When 5 is
produced from 1, it coexists in solution with 1 equiv of
free PMe3 for prolonged periods of time at elevated
Ack n ow led gm en t. We thank Dr. Linda J . W. Shi-
mon of the X-ray diffraction facility of the Weizmann
Institute for solving the crystal structure of 5. This
work was supported by the US-Israel Binational Science
Foundation, J erusalem, Israel. M.A. thanks The Clore
Foundations, J erusalem, Israel, for a scholarship. D.M.
is the holder of the Israel Matz Professorial Chair of
Organic Chemistry.
Su p p or tin g In for m a tion Ava ila ble: Tables of atomic
coordinates, bond distances and angles, and anisotropic dis-
placement parameters for all non-hydrogen atoms and hydro-
gen atom coordinates for 5 (5 pages). Ordering information
is given on any current masthead page.
OM960214I
(19) Burk, M. J .; McGrath, M. P.; Wheeler, R.; Crabtree, R. H. J .
Am. Chem. Soc. 1988, 110, 5034.
(20) Deutsch, P. P.; Eisenberg, R. Chem. Rev. 1988, 88, 1147 and
references therein.
(21) Schnabel, R. C.; Carroll, P. S.; Roddick, D. M. Organometallics
1996, 15, 655.
(22) Burk, M. J .; Crabtree, R. H. Inorg. Chem. 1986, 25, 931.
(23) Relevant to our observations is the reported facile reductive
elimination of H2 from IrH3(CO)(PEt2Ph)2 under a CO atmosphere,
which leads to the similarly stabilized 18-electron species IrH(CO)2-
(PEt2Ph)2. See: Mann, B. E.; Masters, C.; Shaw, B. L. J . Chem. Soc.
D 1970, 846.
(18) Information on Ir-Si bond lengths in five- and six-coordinate
complexes of Ir(III) can be found in: (a) Tilley, T. D. In The Silicon-
Heteroatom Bond; Patai, S., Rappoport, Z., Eds.; Wiley: New York,
1991; pp 264, 343. (b) Goikhman, R.; Aizenberg, M.; Kraatz, H.-B.;
Milstein, D. J . Am. Chem. Soc. 1995, 117, 5865. See also refs 3b, 11d,
and 12.
(24) Bennett, M. A.; Milner, D. L. J . Am. Chem. Soc. 1969, 91, 6983.
+
+