Benzonitrile adducts of terminal diarylphosphido complexes: Preparative sources of "ru=PR2"

Article abstract of DOI:10.1021/om200822e

Dehydrohalogenation of secondary diarylphosphine ruthenium complexes in the presence of benzonitrile yields stable, isolable nitrile adducts of the formula [Ru(η⁵-indenyl)(PAr₂)(NCPh)(PPh₃)], in which the terminal phosphid

Full text of DOI:10.1021/om200822e

Article
pubs.acs.org/Organometallics
Benzonitrile Adducts of Terminal Diarylphosphido Complexes:
Preparative Sources of “RuPR2”
Marc-Andre M. Hoyle,† Dimitrios A. Pantazis,‡ Hannah M. Burton,† Robert McDonald,§
́
and Lisa Rosenberg*,†
†Department of Chemistry, University of Victoria, P.O. Box 3065, Victoria, British Columbia, Canada V8W 3V6
‡Max Planck Institute for Bioinorganic Chemistry, Stiftstr. 34-36, 45470 Mulheim an der Ruhr, Germany
̈
§X-ray Crystallography Laboratory, Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
S
* Supporting Information
ABSTRACT: Dehydrohalogenation of secondary diarylphos-
phine ruthenium complexes in the presence of benzonitrile
yields stable, isolable nitrile adducts of the formula [Ru(η 5-
indenyl)(PAr2)(NCPh)(PPh3)], in which the terminal phos-
phido ligand is pyramidal at P and contains a stereochemically
active lone pair. Unlike the analogous carbonyl adducts
[Ru(η5-indenyl)(PAr2)(CO)(PPh3)], these benzonitrile complexes behave as masked sources of the highly reactive planar
phosphido complexes [Ru(η5-indenyl)(PAr2)(PPh3)], which contain a RuPAr2 π bond. This is illustrated by the addition
(or cycloaddition) reactions of the benzonitrile adducts with dihydrogen, methyl iodide, and 1-hexene, as well as their
thermal decomposition via orthometalation of the PPh3 ligand. Enthalpies of CO vs NCPh dissociation from the [Ru(η 5-
indenyl)(PR2)(PPh3)] fragments (R = alkyl, aryl) have been calculated, as has the trajectory of addition of H2 to the model
planar phosphido complex [Ru(η5-indenyl)(PMe2)(PMe3)]. The latter study shows the intermediacy of an η2-H2 adduct,
[Ru(η5-indenyl)(PAr2)(η2-H2)(PPh3)], in the formation of [RuH(η5-indenyl)(HPMe2)(PMe3)], a further indication of the
importance of the variable binding modes of the terminal phosphido ligand in this system.
catalysts for hydrophosphination.2 So far, we have been able to
INTRODUCTION
■
isolate these very reactive complexes only for bulky alkyl
In our exploration of the coordination chemistry of secondary
phosphines, we have prepared unusual five-coordinate phos-
phido complexes via the dehydrohalogenation reaction
substituents at phosphorus (R = Cy (2a), Pri (2b)). We have
identified the analogous diaryl species (R = Ph (2c), Tolp (2d))
at low temperature in solution by 31P{1H} NMR, but these
illustrated in Scheme 1.1a These dark blue complexes exhibit
complexes decompose to a mixture of unidentified products as
they are warmed to room temperature. We have also trapped
Scheme 1
the diarylphosphido complexes as their CO adducts (3c,d in
Scheme 1);1a however, preliminary reactivity and photolysis
studies3 indicated that none of the carbonyl ligands in 3a−d is
sufficiently labile to yield a reactive five-coordinate species able
to participate in 1,2-addition or [2 + 2]-cycloaddition
chemistry. We report here the synthesis and isolation of the
corresponding benzonitrile adducts (4c,d, eq 1), from which
the benzonitrile ligands dissociate with relative ease. We also
present a survey of the reactivity of both the carbonyl and
RuPR2 π bonding and undergo a range of facile 1,2-addition
2]-cycloaddition reactions of both alkenes1c and alkynes,1d
chemistry, for both polar and nonpolar addenda,1a,b and [2 +
Received: September 1, 2011
which are of particular interest in the development of new
Published: November 7, 2011
© 2011 American Chemical Society
6458
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
benzonitrile adducts of these terminal diarylphosphido
complexes, to show how the benzonitrile adducts behave as a
source of the coordinatively unsaturated complexes [Ru(η 5-
indenyl)(PAr2)(PPh3)] (2c,d) while the more inert carbonyl
adducts do not.
indicated by infrared spectroscopy, with νCN 2198 cm−1 for
both complexes. This is lower than for free benzonitrile (ν CN
2232 cm−1),8 which is consistent with a decrease in the nitrile
bond order due to π back-bonding from the relatively electron-
rich Ru to the nitrile π* orbitals.9 The structure of 4c in the
solid state was confirmed by X-ray crystallography (Figure 1,
RESULTS AND DISCUSSION
■
The addition of toluene to a mixture of solid KOBut, [RuCl(η5-
indenyl)(PR2H)(PPh3)] (R = Ph (1c), Tolp (1d)), and neat
benzonitrile gives purple-red solutions from which the
benzonitrile complexes [Ru(η5-indenyl)(PR2)(NCPh)(PPh3)]
(4c,d) can be isolated (eq 1).4 31P{1H} NMR signals for these
2
complexes (Table 1) do not show JPP coupling: only slightly
Table 1. 31P{1H} NMR Data for New Compounds: δ (ppm)
2
(Multiplicity, JPP or ω1/2 (Hz))
PR2, HPR2, or
complex
MePR2
PPh3
a
[Ru(η5-indenyl)(NCPh)(PPh2)(PPh3)] (4c)
26.1 (s, 4)
25.6 (s, 4)
52.3
51.8
[Ru(η5-indenyl)(NCPh)(PTolp )(PPh3)]
2
b
(4d)
[Ru(η5-indenyl){κ2-(o-C6H4)PPh2}(HPPh2)]
46.5 (d, 30)
44.6 (d, 29)
−18.7
−18.5
c
(5c)
Figure 1. Molecular structure of 4c. Non-hydrogen atoms are
represented by Gaussian ellipsoids at the 20% probability level, only
the ipso carbons are shown for phenyl rings bound to phosphorus;
hydrogen atoms are not shown. Selected interatomic distances (Å) and
bond angles (deg) (C* denotes the centroid of the plane defined by
C(7A)−C(1)−C(2)−C(3)−C(3A)): Ru−P(1) = 2.3824(6), Ru−
P(2) = 2.3122(6), Ru−N = 2.0059(19), Ru−C* = 1.916, N−C(10) =
1.150(3), C(10)−C(11) = 1.438(3); P(1)−Ru−P(2) = 92.59(2),
P(1)-Ru−N = 88.87(5), P(2)-Ru−N = 88.88(6), P(1)−Ru−C* =
123.4, P(2)−Ru−C* = 126.0, N−Ru−C* = 126.1, Ru−N−C(10) =
172.40(18), N−C(10)−C(11) = 174.3(2). Indenyl crystallographic
slip distortion:11 Δ = d(Ru−C(7A),C(3A)) − d(Ru−C(1),C(3)) =
0.161 Å.
[Ru(η5-indenyl){κ2-(o-C6H4)PPh2}
c
(HPTolp )] (5d)
2
c
[Ru(η5-indenyl)H(HPTolp )(PPh3)] (6d)
47.0 (d, 33)
33.7 (d, 36)
70.6
54.0
2
[Ru(η5-indenyl)(NCPh)(PPh2Me)(PPh3)]I
d
(7c)
d
[Ru(η5-indenyl)I(PPh2Me)(PPh3)] (8c)
41.1 (d, 42)
30.8 (d, 38)
45.1
53.9
[Ru(η5-indenyl)(NCPh)(PTolp Me)(PPh3)]I
2
d
(7d)
d
[Ru(η5-indenyl)I(PTolp Me)(PPh3)] (8d)
38.9 (d, 42)
29.0 (d, 25)
26.8 (d, 26)
44.2
46.8
48.8
2
d
[Ru(η5-indenyl)(CO)(PPh2Me)(PPh3)]I (9c)
[Ru(η5-indenyl)(CO)(PTol2Me)(PPh3)]I
d
(9d)
[Ru(η5-indenyl)(κ2-BunCH2CH2PPh2)
−19.3 (d, 24)
−21.1 (d, 25)
−21.4 (d, 24)
−23.2 (d, 25)
63.3
62.5
63.4
62.7
c
(PPh3)] (syn-10c)
[Ru(η5-indenyl)(κ2-BunCH2CH2PPh2)
Table 2). Features associated with the pyramidal geometry at
the PPh2 ligand are as expected: the sum of angles at P1 is
315.42°, and the Ru−PPh2 distance is 0.07 Å longer than the
Ru−PPh3 distance, due to the transition-metal “gauche” effect
associated with the pyramidal phosphorus lone pair.10 The
benzonitrile ligand is bent slightly away from the phosphido
and triphenylphosphine ligands (Ru−N−C10 = 169.4°, N−
C10−C11 = 174.3°), illustrating both the overall steric
crowding in this structure and the stereochemical requirements
of the phosphido lone pair.
c
(PPh3)] (anti-10c)
[Ru(η5-indenyl)(κ2-BunCH2CH2PTolp )
2
c
(PPh3)] (syn-10d)
[Ru(η5-indenyl)(κ2-BunCH2CH2PTolp )
2
c
(PPh3)] (anti-10d)
a
Conditions: at 202.46 MHz, samples in d8-toluene, recorded at 240
b
c
K. Conditions: at 202.46 MHz, samples in d6-benzene. Conditions:
at 121.49 MHz, samples in d6-benzene. Conditions: at 121.49 MHz,
samples in d-chloroform.
d
The diarylphosphido benzonitrile adducts 4c,d are more
stable than the analogous dialkylphosphido complexes 4a,b,
which we previously generated in situ from the addition of
benzonitrile to solutions of the planar phosphido complexes
2a,b and characterized by 31P{1H} NMR.1b Small amounts
(≤20%) of 2a,b remain in equilibrium with 4a,b in these
solution samples, even in the presence of a large excess of
benzonitrile. Attempts to isolate the diisopropylphosphido
nitrile adduct (4b) by removal of solvent under vacuum
revealed that this equilibrium shifts further toward 4b as the
sample becomes more concentrated (and presumably more
benzonitrile-rich (bp 190 °C)): we were able to obtain a red
paste, which, when redissolved in pentane, rapidly turned blue,
indicating the loss of benzonitrile to give 2b. Subsequent
removal of the pentane under vacuum regenerated the red
paste. Several repetitions of this trituration gave an oily dark
reddish solid, an infrared spectrum of which confirmed the
broadened singlets (ω1/2 ≈ 4 Hz) are observed for the PPh3
2
and PAr2 ligands. Such reduced or negligible JPP values are
diagnostic of pyramidal geometry at the terminal phosphido
ligands,5 and we previously reported similar, very low values for
the analogous diarylphosphido carbonyl adducts (Scheme 1;
2JPP = 8 Hz (3c), 7 Hz (3d)).1a The aromatic regions of the 1H
and 13C NMR spectra of 4c,d are quite complex, with
considerable broadening of signals due to PPh3, a feature of
this crowded ruthenium indenyl system that we frequently
observe.6 However, spectra of 4c at low temperature (240 K)
and of 4d at room temperature are sufficiently well resolved to
1
allow assignment of the H NMR signals corresponding to
coordinated benzonitrile,7 which are supported by correlations
in the 1H/13C-HMBC spectra between signals due to Hortho and
the quaternary nitrile carbon (PhCN, δ 123.2 ppm). The
presence of the coordinated nitrile ligand in 4c,d is also
6459
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
Table 2. Crystallographic Data for [(η5-indenyl)Ru(NCPh)-
Scheme 2
(PPh2)(PPh3)] (4c)
formula
C46H37NP2Ru
formula wt
cryst color, habit
cryst dimens (mm)
cryst syst, space group
a (Å)
766.78
orange rod
0.10 × 0.12 × 0.46
orthorhombic, Pbcn (No. 60)
19.9629(6)
19.0706(6)
19.6393(6)
7476.8(4)
8
b (Å)
c (Å)
V (Å3)
Z
ρcalcd (g cm−3
)
1.362
μ (mm−1
)
0.538
temp (°C)
−100
data collection 2θmax (deg)
52.80
total no. of data collected
57 326 (−24 ≤ h ≤ 24, −23 ≤ k ≤ 23,
stable to thermal decomposition, showing no change by
31P{1H} NMR even with prolonged heating (5 days) at 60 °C.
This is consistent with the higher π acidity of CO relative to
that of benzonitrile, which leads to stronger M−L bonds in
these electron-rich ruthenium complexes. Computed ligand
dissociation enthalpies (Table 3) for the conversion of 3a−d to
2a−d plus CO and for 4a−d to 2a−d plus NCPh, respectively,
indicate that CO binds approximately twice as strongly as
benzonitrile to Ru throughout this series. The calculated Ru−
CO dissociation enthalpies for the dialkylphosphido complexes
3a,b are similar to experimental CO dissociation energies of
41−44 kcal/mol reported for Ru(CO)5, Ru3(CO)12, and
Ru(dmpe)2CO.13 We have found no absolute bond
dissociation data for ruthenium (or other) nitrile complexes,
but our calculated Ru−NCPh values are close to those
determined by solution calorimetry for Ru−PR3 bonds in
half-sandwich complexes, which fall in the range of 9−24
kcal/mol.14 Most striking is the approximately 7 kcal/mol
increase in calculated ligand dissociation enthalpies for the
diarylphosphido complexes 3c,d and 4c,d relative to the
corresponding dialkylphosphido complexes 3a,b and 4a,b,
consistent with our observation of the instability of 4b relative
to 4c,d (vide supra).
−24 ≤ l ≤ 24)
no. of indep rflns (Rint)
7675 (0.0595)
5918
2
no. of obsd rflns (Fo
≥
2
2σ(Fo ))
range of transmissn factors
0.9496−0.7883
7675/0/451
no. of data/restraints/
params
goodness of fit (S) (all
1.024
a
data)
b
2
2
R1 (Fo ≥ 2σ(Fo ))
0.0293
c
wR2 (all data)
0.0705
largest diff peak, hole (e
0.312, −0.499
Å−3
)
a
S = [∑w(Fo − Fc2)2/(n − p)]1/2 (n = number of data; p = number
2
of parameters varied; w = [σ2(Fo ) + (0.0247P)2 + 6.0877P]−1, where
2
b
c
P = [Max(Fo , 0) + 2Fc24]/3). R1 = ∑||Fo| − |Fc||/∑|Fo|. wR2 =
2
[∑w(Fo − Fc2)2/∑w(Fo )]1/2
.
2
presence of coordinated benzonitrile, with ν CN at 2199 cm−1.
However, attempts to characterize this sample by solution
NMR inevitably resulted in decomposition, including
dissociation of the benzonitrile ligand and regeneration of
2b, accompanied by thermal degradation to the orthometa-
lated product 5b1a (vide infra). It is not yet clear whether the
greater stability of the diarylphosphido benzonitrile adducts
4c,d in solution, relative to the dialkylphosphido analogues
4a,b, is due to the lesser bulk of the diaryl-substituted
phosphido ligands in 4c,d or to their diminished donor
ability.
Table 3. Computed Ligand Dissociation Enthalpies ΔH
a
(kcal mol−1)
Complexes 4c,d are sufficiently stable to allow their isolation
and characterization, but they are highly air-sensitive even in
the solid state12 and, in the absence of air or moisture, they
undergo thermal decomposition in solution. Heating sealed
samples of 4c,d in d6-benzene to 60 °C gives a slow color
change from deep purple-red to a clear orange-red,
corresponding to the formation of the orthometalated
products [Ru(η 5-indenyl){κ2-(o-C6H4)PPh2}(PR2H)] (5c,d
in Scheme 2), as determined by 31P{1H} NMR spectroscopy.
Similarly to the previously characterized complexes 5a,b,1a the
31P{1H} signals for the orthometalated PPh3 ligand in 5c,d
show a diagnostic upfield shift, from ∼52 ppm in 4c,d to −18.7
ppm (5c) and −18.5 ppm (5d). 31P NMR confirms the
presence of the secondary phosphine ligands in 5c,d: signals at
R
L = CO (3)
L = PhCN (4)
Cy (a)
Pri (b)
Ph (c)
Tolp (d)
45.4
45.1
52.0
52.3
18.1
17.9
24.6
25.3
a
PBE/DKH2-TZVP calculations.
The fact that complexes 4c,d give the same thermal
decomposition product as the five-coordinate dialkylphosphido
complexes 2a,b suggests that these adducts cleanly dissociate
benzonitrile to generate the diarylphosphido complexes 2c,d in
solution. This is further supported by addition of dihydrogen to
the adducts in d6-benzene, which generates the secondary
1
46.5 and 44.6 ppm, respectively, show a large JPH value of 342
Hz, and similar coupling constants are observed for the P−H
1
signals in the corresponding H NMR spectra. In contrast,
samples of the carbonyl adducts 3c,d in d8-toluene are quite
6460
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
Figure 2. PBE/DKH2-TZVP free energy reaction profile for a simplified model (PPh3 replaced by PMe3; Cy or Pri replaced by Me) of the addition
of H2 to 2a,b.
phosphine hydrido complexes [RuH(η 5-indenyl)(PR2H)-
(PPh3)] (6c,d) within seconds (eq 2), on the basis of the
state, the H on Ru has a calculated charge of −0.1, while the H
on P has a charge of +0.1, a very small charge separation. It is
easy to visualize the ready dissociative substitution of
benzonitrile by dihydrogen as an entry to this essentially
homolytic pathway of addition across the RuPAr2 bond
in 4c,d.
We showed previously that the dialkylphosphido complexes
2a,b undergo 1,2-addition of the polar bonds in reagents such
as HCl and methyl iodide,1a as well as a range of protic
reagents.1b As shown for methyl iodide in Scheme 3, these
reactions invariably place the electrophilic portion of the
addendum at P and the nucleophilic portion at the co-
ordinatively unsaturated Ru. Addition of MeI to the
benzonitrile adducts 4c,d ultimately gives the analogous
diarylmethylphosphine iodide complexes 8c,d (Scheme 3),
Scheme 3
color change from deep purple-red to clear yellow-orange, and
as confirmed by 1H and 31P{1H} NMR.15 The same overall 1,2-
addition product 6a results from the addition of H2(g) to
complex 2a (eq 3),1a and similar hydrido amine or phosphine
complexes have resulted from the addition of dihydrogen to
other metal amido or phosphido complexes, respectively.16
This apparent outer-sphere activation of dihydrogen is
analogous to that proposed to occur for the Noyori-type
ruthenium amide hydrogenation catalysts.17 We have put some
effort into identifying the trajectory of hydrogen addition,
computationally, to better understand this process. DFT results
for a simplified dialkylphosphido ruthenium indenyl complex
are shown in Figure 2. The intermediate η2-H2 adduct
highlights the importance to this reaction of the ready access
to coordinative unsaturation at Ru resulting from disruption of
the Ru−P double bond.18 Although the overall reaction is
formally a heterolytic cleavage of dihydrogen, the hydrogens in
the η2 adduct carry zero charge, and in the cyclic transition
but our studies of this reaction indicate a stepwise mechanism
in which nucleophilic attack of the terminal phosphido ligand at
6461
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
Meδ+ generates the cationic intermediate 7, which then under-
goes nucleophilic substitution of the benzonitrile at Ru by I−.
The addition of MeI to 4c,d in toluene results in the immediate
formation of a bright yellow precipitate (1−2 s), which, when
dissolved in d-chloroform, slowly converts from clear yellow to
clear orange-red. The identity of the yellow precipitate as
[Ru(η5-indenyl)(NCPh)(PPh3)(PR2Me)]I (7c,d) was con-
firmed by observation of νCN (2239 cm−1, 7c; 2227 cm−1,
7d) in the IR spectra and observation of the cation mass peaks
by ESI-MS, with predicted isotopic distributions (see the
Scheme 4
1
Supporting Information). H NMR spectra show a diagnostic
2
doublet with JPH = 9 Hz at 1.2−1.3 ppm due to the new
P−CH3 group. 31P{1H} NMR spectra of these d-chloroform
solutions initially show major signals due to the benzonitrile
adducts 7c,d, but we also see small signals due to [RuI(η 5-
indenyl)(PPh3)(PR2Me)] (8c,d).19 These gradually increase in
intensity (over 5 days) to become the major product signals. In
the 1H NMR spectra the P−Me signals remain but shift slightly
(∼0.5 ppm).20 Cationic complexes analogous to the benzoni-
trile-containing 7c,d, [Ru(η5-indenyl)(CO)(PPh3)(PR2Me)]I
(9c,d), result from the addition of excess MeI to the carbonyl
complexes 3c,d (eq 4), again as diagnosed by 1H NMR, IR, and
amounts of the anti products appearing only after 1 day or
more. This surprising difference in the 1-hexene concentration
dependences of the formation of the two diastereomers is not
out of line with our ongoing studies of the cycloaddition
reactions of 1-hexene at 2a,b, for which we observe a kinetic
preference for the formation of the syn isomer, facile
cycloreversion, and slow epimerization.21
CONCLUSIONS
■
With the synthesis of these benzonitrile-stabilized terminal
diarylphosphido complexes, we have expanded the scope of a
series of highly reactive five-coordinate ruthenium half-
sandwich fragments. The reactivity and computational studies
described above illustrate the importance of ready access to a
vacant coordination site at ruthenium, via facile ligand
dissociation or variable phosphido binding mode, to allow the
addition of nonpolar H−H or C−H bonds, or [2 + 2]-
cycloaddition of alkenes, at the Ru−PR2 bond. Meanwhile, the
stepwise reactions with MeI confirm that the nucleophilicity at
the terminal phosphido ligands in these complexes is not
compromised by the presence of less donating aryl substituents
at P or by the ready dissociation of the benzonitrile ligands. We
are currently studying the mechanism(s) of cycloaddition in
these complexes more closely and examining their activity in
catalytic phosphination and hydrophosphination reactions.
ESI-MS. These carbonyl cations are stable in d-chloroform
indefinitely at room temperature, showing no sign of sub-
stitution of the CO ligand by the I− counterion.
As described in the Introduction, the dialkylphosphido
complexes 2a,b participate in [2 + 2]-cycloaddition reactions
with alkenes and alkynes.1c,d Of particular note is the fact that
even the simple olefins ethylene and 1-hexene participate in this
apparently concerted cycloaddition. Our final test of the ability
of 4c,d to act as masked sources of the coordinatively
unsaturated 2c,d involved the addition of 1-hexene to these
benzonitrile adducts. We screened the behavior of both 3c,d
and 4c,d in the presence of an excess of 1-hexene, in sealed
NMR tube experiments. The carbonyl adducts 3c,d showed no
reaction with the alkene, while the benzonitrile adducts 4c,d
gradually changed color from dark purple-red to red-orange,
and 31P{1H} NMR showed that first one and then a second,
minor product formed, with signals corresponding closely in
EXPERIMENTAL SECTION
■
General Considerations. Unless otherwise noted, all reactions
and manipulations were performed under nitrogen in an MBraun
Unilab 1200/780 glovebox or using conventional Schlenk techniques.
All solvents were sparged with nitrogen for 25 min and dried using an
MBraun Solvent Purification System (SPS). Deuterated solvents were
purchased from Canadian Isotopes Laboratory (CIL), freeze−pump−
thaw-degassed, and vacuum-transferred from sodium/benzophenone
(d6-benzene, d8-toluene) calcium hydride (d-chloroform) before use.
Benzonitrile was predried with K2CO3 and fractionally distilled from
P2O5 under dynamic vacuum. Potassium tert-butoxide was purchased
from Aldrich Chemical Co. and used as received without further
purification. Carbon monoxide and hydrogen gas were purchased from
Praxair Canada Inc. [Ru(η5-indenyl)Cl(PR2H)(PPh3)] (R = Ph (1c6),
Tolp (1d1a)), [Ru(η5-indenyl)(PPri2)(PPh3)] (2b1a), and [Ru(η5-
indenyl)(CO)(PR2)(PPh3)] (R = Ph (3c), Tolp (3d))1a were pre-
pared as previously reported in the literature. (All secondary phos-
phines were purchased from Strem Chemicals as 10 wt % solutions in
hexanes; concentrations were checked against a known quantity of
triphenylphosphine oxide by 31P{1H} NMR before use.)
2
shift and JPP values to those previously observed for the syn
and anti diastereomers of the [2 + 2]-cycloadducts [Ru(η 5-
indenyl)(κ2-BunCH2CH2PR2)(PPh3)] (10a,b).1c The high-field
signals (−19.3 to −23.2 ppm) are particularly diagnostic of the
incorporation of “PR2” into the four-membered metallacycle.
We presume that the cycloaddition of 1-hexene at 4c,d is
proceeding with similar regiochemistry to give the analogous
syn and anti diastereomers of 10c,d, where syn and anti refer to
the orientation of the Bun substituent on the α-carbon with
respect to the Ru−indenyl bond (Scheme 4). Upon complete
consumption of 4c,d in these reactions the diastereomer ratio
(by analogy to assignments for 10a,b) is about 80:20 syn/anti,
which is slightly lower than the >95:5 ratio observed for 10a,b.
Interestingly, a second set of sealed NMR tube experiments, in
which slightly less than 1 equiv of 1-hexene was added to 4c,d,
gave only the putative syn diastereomers as products, with trace
6462
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
31P NMR data are given in Table 1. 1H and 13C NMR data for 4c,d
and 1H NMR data for products of the NMR scale reactions are in the
Supporting Information. NMR spectra were recorded on a Bruker
For 3c,d, samples remained dark red, and initial 31P{1H} spectra
showed no changes. After 5 days, 31P{1H} spectra still showed
principally unreacted 3c,d, along with small amounts of PPh3 (<5%)
and an unidentified product (in 3c, singlet at −56.7 ppm (<5%); in 3d,
singlet at −58.9 ppm (<5%).
For 4c,d, samples gradually (4c, 90 h; 4d, 70 h) lightened from
deep purple-red to clear orange-red, at which point 31P{1H} NMR
showed conversion to the orthometalated complexes [Ru(η 5-
indenyl){κ2-(o-C6H4)PPh2}(PR2H)] (5c,d, by analogy to shifts and
coupling constants previously observed for the orthometalated
dialkylphosphine complex 5a1a).
1
AVANCE 500 operating at 500.13 MHz for H, 125.77 MHz for 13C,
and 202.46 MHz for 31P or on a Bruker AVANCE 300 operating at
1
300.13 MHz for H, 74.47 MHz for 13C, and 121.49 MHz for 31P.
Chemical shifts are reported in ppm at ambient temperature unless
otherwise noted. 1H chemical shifts are referenced against residual
protonated solvent peaks at 7.16 (C6D5H), 2.09 (PhCD2H), and 7.26
ppm (CHCl3). 13C chemical shifts are referenced against d6-benzene at
128.4 ppm and d8-toluene at 20.4 ppm. All 1H and 13C chemical shifts
are reported relative to tetramethylsilane, and 31P chemical shifts are
reported relative to 85% H3PO4(aq).
(b). Reaction with H2(g). Solid [Ru(η 5-indenyl)(CO)(PR2)-
(PPh3)] (3c, 5 mg, 0.007 mmol; 3d, 5 mg, 0.006 mmol) or [Ru(η 5-
indenyl)(PhCN)(PR2)(PPh3)] (4c, 5 mg, 0.007 mmol; 4d, 5 mg,
0.007 mmol) was dissolved in d6-benzene (3c, 4d, 0.6 mL) or d8-
toluene (3d, 4c, 0.6 mL) and added to a J. Young NMR tube. The
sample was degassed by three freeze−pump−thaw cycles before approxi-
mately 1 atm of hydrogen was introduced. The sealed sample was
inverted five times, and the contents were then analyzed by 31P{1H}
NMR spectroscopy.
IR spectra were recorded for KBr pellets under a nitrogen atmo-
sphere on a Perkin-Elmer FTIR Spectrum One spectrometer. UV−vis
data were obtained using a Varian Cary 5 UV−vis spectrophotometer.
Microanalysis was performed by Canadian Microanalytical Service
Ltd., Delta, BC, Canada. Low-resolution mass spectra were collected at
the University of Victoria on a Micromass Q-ToF micro hybrid
quadrupole/time-of-flight mass spectrometer in positive-ion mode
using electrospray ionization. High-resolution mass spectra were
acquired by Dr. Yun Ling at the University of British Columbia,
Vancouver, BC, Canada, on a Water/Micromass LCT-ToF mass
spectrometer in positive-ion mode using electrospray ionization.
Synthesis of Benzonitrile Adducts 4c,d. [Ru(η5-indenyl)-
(PPh2)(PhCN)(PPh3)] (4c). To a Schlenk flask containing [Ru(η 5-
indenyl)Cl(HPPh2)(PPh3)] (1c; 423 mg, 0.604 mmol) and KOBut
(132 mg, 1.18 mmol) was added benzonitrile (0.31 mL, 0.31 g, 3.0 mmol)
to form a dark red sludge. Toluene (25 mL) was added, and the
resulting purple-red mixture was stirred for 3 h, after which it was
filtered through Celite to remove solid and gelatinous byproducts KCl
and HOBut. The toluene was removed under vacuum, and the
resulting red paste was precipitated from a minimum volume of
toluene (3 mL) layered with pentane (25 mL). The dark red powder
was then washed with pentane (4 × 10 mL) to give [Ru(η5-
indenyl)(PhCN)(PPh2)(PPh3)] (4c; 272 mg, 0.355 mmol, 59% yield).
IR (KBr, cm−1): 2198 (s, νCN). UV−vis (toluene): λmax 523 nm, ε
3000 M−1 cm−1. HR-ESI-MS ([M + H]+, m/z): calcd for
C46H38NP299Ru 765.1539, found 765.1558 (error 2.5 ppm). Anal.
Calcd for C46H37NP2Ru: C, 72.03; H, 4.87. Found: C, 71.99; H, 4.93.
Dec pt: 148−150 °C.
For 3c,d, samples retained their red color, and 31P{1H} spectra
showed no changes.
For 4c,d, the dark purple-red solutions turned clear yellow-orange
after the sample was inverted (∼10 s), and 31P{1H} NMR showed
complete conversion to the hydrido phosphine complexes [Ru(η 5-
indenyl)(H)(PR2H)(PPh3)] (6c,d).13
(c). Reaction with MeI. Solid [Ru(η5-indenyl)(CO)(PR2)(PPh3)]
(3c, 28 mg, 0.040 mmol; 3d, 25 mg, 0.035 mmol) or [Ru(η5-
indenyl)(NCPh)(PR2)(PPh3)] (4c, 37 mg, 0.046 mmol; 4d, 30 mg,
0.038 mmol) was dissolved in toluene (1−5 mL). MeI (50 μL, 0.8
mmol) was added to the dark red solutions.
For 3c,d, mixtures immediately formed a yellow precipitate along
with a solution color change to clear yellow (1−2 s). Toluene was
removed under vacuum, and analysis of the resulting yellow
precipitates by NMR, IR, and ESI-MS indicated they were the iodide
salts [Ru(η5-indenyl)(CO)(PR2Me)(PPh3)]I (9c,d). 31P{1H} NMR
spectra of the samples in d-chloroform showed that 9c,d was the major
product, although there were minor P-containing impurities.
Continued monitoring of these clear yellow samples by 31P{1H}
NMR showed no changes over 3−5 days.
Data for 9c: LR-ESI-MS (20 V, CH2Cl2, m/z) 707.2 (M+, 100%);
IR (KBr, cm−1) 1959 (s, νCO).
[Ru(η5-indenyl)(PTolp )(PhCN)(PPh3)] (4d). This complex was pre-
2
pared by a method similar to that for 4c, using [Ru(η5-indenyl)Cl-
Data for 9d: LR-ESI-MS (20 V, CH2Cl2, m/z) 735.2 (M+, 100%);
IR (KBr, cm−1) 1980 (s, νCO).
(HPTolp )(PPh3)] (1d; 270 mg, 0.371 mmol), KOBut (84 mg,
2
0.75 mmol), and benzonitrile (0.21 mL, 0.21 g, 2.0 mmol). The product
For 4c,d, mixtures formed a yellow precipitate within 1−2 s, with light
orange supernatant solutions. Toluene was removed under vacuum, and
analysis of the resulting yellow precipitates by NMR, IR, and ESI-MS
indicated they were the iodide salts [Ru(η 5-indenyl)(NCPh)(PR2Me)-
(PPh3)]I (7c,d). The NMR samples in d-chloroform slowly darkened
from yellow-orange to red-orange. Continued monitoring by 31P{1H}
NMR showed complete consumption of 7c,d after 11 days, at which
point 8c,d were the major products in solution.
[Ru(η5-indenyl)(PhCN)(PTolp )(PPh3)] (4d) was isolated as a dark
2
red powder (188 mg, 0.237 mmol, 64% yield). IR (KBr, cm−1): 2198
(s, νCN). UV−vis (toluene): λmax 523 nm, ε 3000 M−1 cm−1. HR-ESI-
MS ([M + H]+, m/z): calcd for C48H42NP299Ru 793.1852, found
793.1827 (error −3.1 ppm). High air sensitivity precluded the
satisfactory elemental analysis of this complex (see the Supporting
1
Information for H and 31P{1H} spectra of isolated product). Dec pt:
148−150 °C.
Data for 7c: LR-ESI-MS (20 V, CH2Cl2, m/z) 782.1 (M+, 50%),
679.0 ([M − PhCN]+, 100%); IR (KBr, cm−1) 2239 (s, νCN).
Data for 7d: ESI-MS (20 V, CH2Cl2, m/z) 810.2 (M+, 40%), 707.1
([M − PhCN]+, 100%); IR (KBr, cm−1) 2227 (s, νCN).
(d). Reaction with Excess 1-Hexene. Solid [Ru(η5-indenyl)(CO)-
(PR2)(PPh3)] (3c, 7 mg, 0.01 mmol; 3d, 7 mg, 0.01 mmol) or
[Ru(η5-indenyl)(PhCN)(PR2)(PPh3)] (4c, 8 mg, 0.01 mmol; 4d,
8 mg, 0.01 mmol) was dissolved in d6-benzene (1.0 mL) and added to
a sealable NMR tube. The sample was freeze−pump−thaw-degassed
three times. 1-Hexene (0.10 mL, 0.81 mmol) was added by syringe,
and the tube was flame-sealed. The thawed sample was inverted five
times to mix the reagents and then monitored by 31P{1H} NMR
spectroscopy.
Attempted Isolation of [Ru(η5-indenyl)(PPri2)(PhCN)(PPh3)]
(4b). To a Schlenk flask containing [Ru(PPri2)(η5-indenyl)(PPh3)]
(2b; 82 mg, 0.14 mmol) in toluene (5 mL) was added an excess of
benzonitrile (720 mg, 7 mmol, ∼50 equiv), which caused the dark blue
solution to turn deep red. The mixture was stirred for approximately
10 min, and then the solvent was removed under vacuum. Trituration
of the resulting bright red oily residue with pentane (4 × 5 mL) gave a
tacky reddish solid (crude yield 79 mg, 0.11 mmol, 81%). IR (KBr
pellet, cm−1): 2199 (w, νCN).
NMR-Scale Reactions of the Adducts [Ru(η 5-indenyl)(PR2)-
(L)(PPh3)] (L = CO (3c,d), PhCN (4c,d)). (a). Thermolysis. Solid
[Ru(η5-indenyl)(CO)(PR2)(PPh3)] (3c, 6 mg, 0.008 mmol; 3d, 5 mg,
0.006 mmol) or [Ru(η5-indenyl)(PhCN)(PR2)(PPh3)] (4c, 5 mg,
0.007 mmol; 4d, 8 mg, 0.011 mmol) was dissolved in d8-toluene (3c,d,
0.7 mL) or d6-benzene (4c,d, 0.7 mL) and added to a J. Young NMR
tube. The NMR sample was heated to 60 °C in an oil bath and was
removed periodically for monitoring by 31P{1H} NMR spectroscopy.
For 3c,d, samples remained dark red, and 31P{1H} NMR spectra
showed no changes, even after 24 h.
For 4c,d, both dark red samples gradually changed to clear red-
orange (18−21 h).
6463
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
(e). Reaction with 1 equiv of 1-Hexene. Solid [Ru(η5-indenyl)-
(PhCN)(PR2)(PPh3)] (4c, 10 mg, 0.013 mmol; 4d, 10 mg, 0.013
mmol) was dissolved in d6-benzene (0.7 mL) and added to a sealable
NMR tube. 1-Hexene (0.161 M in toluene, 0.08 mL, 0.01 mmol) was
added by syringe, and the tube was flame-sealed. The sample was
inverted five times to mix the reagents and then monitored by 31P{1H}
NMR spectroscopy.
For 4c,d, for both samples, the dark red color persisted over several
days, despite changes in the 31P{1H} NMR spectra, due to the
presence of residual, darkly colored starting material. (1H NMR
confirmed the complete consumption of 1-hexene.)
(4) 31P{1H} NMR of the crude reaction mixtures indicate complete
consumption of 1c,d. In some trials we see traces (<5%) of an
unidentified impurity containing a single phosphorus environment
(singlet at 36.2 ppm (R = Ph) or 36.5 ppm (R = Tolp)).
2
(5) Similarly low JPP values have been reported for other mixed
PR3/pyramidal PR′2 complexes: (a) Planas, J. G.; Gladysz, J. A. Inorg.
Chem. 2002, 41, 6947. (b) Planas, J. G.; Hampel, F.; Gladysz, J. A.
Chem. Eur. J. 2005, 11, 1402.
(6) Derrah, E. J.; Marlinga, J. C.; Mitra, D.; Friesen, D. M.; Hall, S.
A.; McDonald, R.; Rosenberg, L. Organometallics 2005, 24, 5817.
1
(7) The relative H shifts and multiplicities for peaks due to the
X-ray Structure Determination. Crystals of 4c were grown via
slow diffusion of hexanes into a toluene solution of the compound.
Data were collected using Mo Kα radiation (λ = 0.710 73 Å) on a
Bruker APEX II CCD detector/PLATFORM diffractometer22 with the
crystals cooled to −100 °C. The data were corrected for absorption
through Gaussian integration from indexing of the crystal faces. The
structure was solved using a Patterson search for heavy atoms followed
by structure expansion (DIRDIF-200823). Refinements were com-
pleted using full-matrix least squares on F2 (SHELXL-9724). All non-
hydrogen atoms were refined with anisotropic displacement
parameters. All hydrogen atoms were assigned positions based on
the idealized sp2 or sp3 geometries of their attached atoms and were
given thermal parameters 20% greater than those of the parent
carbons. Further crystallographic experimental data and refinement
details can be found in Table 2.
Computational Details. For all calculations the Perdew−Burke−
Ernzerhof density functional (PBE)25 was employed. Scalar relativistic
effects were explicitly treated using the second-order Douglas−Kroll−
Hess Hamiltonian,26 in combination with all-electron TZVP basis sets
of the Karlsruhe group,27 appropriately recontracted to satisfy the
requirements of the scalar relativistic Hamiltonian.28 The resolution of
the identity approximation was used with decontracted auxiliary TZV/
J basis sets. Harmonic vibrational frequencies and thermodynamic
corrections were obtained at the same level of theory. All calculations
were performed with the ORCA program29 using tight convergence
criteria and enhanced integration grids.
aromatic protons on coordinated benzonitrile are very similar to those
we observe for free benzonitrile in d6-benzene: 7.11−6.97 (m, Hp),
6.97−6.81 (m, Ho), and 6.81−6.64 (m, Hm) ppm.
(8) Mrozek, M. F.; Wasileski, S. A.; Weaver, M. J. J. Am. Chem. Soc.
2001, 123, 12817.
(9) Values of νCN for transition-metal nitrile complexes are often
higher than for the free nitrile: (a) Storhoff, B. N.; Lewis, H. C. Coord.
Chem. Rev. 1977, 23, 1. (b) Kuznetsov, M. L.; Dement’ev, A. I.;
Nazarov, A. A. Russ. J. Inorg. Chem. 2005, 50, 731. This can be
rationalized by σ donation of the lone pair at N (which renders the N
more electropositive and strengthens the N−C σ-bond) and the low π
acidity of the N-bound nitrile (consider the sizes of the π* lobes at N
(small) and C (large), relative to coordinated CO). However, there are
other examples of Ru(II) nitrile complexes that show red-shifted νCN
values similar to those of 4c,d: (c) Ford, P. C. Coord. Chem. Rev. 1970,
5, 75. (d) Alves, J. J. F.; Franco, D. W. Polyhedron 1996, 15, 3299.
(10) (a) Barre, C.; Boudot, P.; Kubicki, M. M.; Moise, C. Inorg.
Chem. 1995, 34, 284. (b) Buhro, W. E.; Zwick, B. D.; Georgiou, S.;
Hutchinson, J. P.; Gladysz, J. A. J. Am. Chem. Soc. 1988, 110, 2427.
(11) Faller, J. W.; Crabtree, R. H.; Habib, A. Organometallics 1985, 4,
929. Typically a slip distortion of less than 0.25 Å indicates η5-
indenyl coordination.
(12) The deep red-purple solids turn light orange-brown within
seconds upon exposure to air.
(13) Ru(CO)5: (a) Behrens, R.G. J. Less-Common Met. 1977, 56, 55.
Ru3(CO)12: (b) Connor, J.; et al. Top. Current Chem. 1977, 71, 71.
(c) Housecroft, C. E.; Wade, K.; Smith, B. C. J. Chem. Soc., Chem.
Commun. 1978, 765. Ru(dmpe)2CO: (d) Belt, S. T.; Scaiano, J. C.;
Whittlesey, M. K. J. Am. Chem. Soc. 1993, 115, 1921.
(14) Luo, L.; Zhu, N.; Zhu, N. J.; Stevens, E. D.; Nolan, S. P.; Fagan,
P. J. Organometallics 1994, 13, 669.
(15) Complex 6c was prepared previously by the addition of NaOMe
to complex 1c,6 and complex 6b was identified by NMR as the product
of addition of MeOH to 2b.1b
ASSOCIATED CONTENT
■
S
* Supporting Information
1
Tables and figures giving H and 13C{1H} NMR data for 4c,d,
1H NMR data for products of NMR scale reactions of 3c,d and
4c,d, 31P{1H} NMR spectra of the progress of NMR scale
reactions of 4c,d, ESI-MS data for 7c,d and 9c,d, and Cartesian
coordinates of all optimized structures and a CIF file giving
crystallographic data for 4c. This material is available free of
charge via the Internet at http://pubs.acs.org.
(16) See, for example: (a) Fryzuk, M. D.; MacNeil, P. A.; Rettig, S. J.
Organometallics 1985, 4, 1145 (Ir amido). Fryzuk, M. D.;
Montgomery, C. D.; Rettig, S. J. Organometallics 1991, 10, 467 (Ru
amido). Fryzuk, M. D.; Bhangu, K. J. Am. Chem. Soc. 1988, 110, 961
(Ir phosphido). Roddick, D. M.; Santarsiero, B. D.; Bercaw, J. E. J. Am.
Chem. Soc. 1985, 107, 4670 (Hf phosphido). Dahlenburg, L.; Hock,
N.; Berke, H. Chem. Ber. 1988, 121, 2083 (Rh phosphido).
(17) Hartwig, J. F., Organotransition Metal Chemistry: From Bonding
to Catalysis; University Science Books: Mill Valley, CA, 2010; p 602.
(18) We searched carefully for minima and saddle points containing
η3-indenyl structures in this system and found none. The first
transition state in Figure 2 shows considerable slip distortion (Δ =
0.26 Å) relative to the intermediate (Δ = 0.16 Å) and the second
transition state (Δ = 0.13 Å), but it is still well outside the range for
formal η3 coordination (Δ = 0.69−0.80 Å): Cadierno, V.; Diez, J.;
Gamasa, M. P.; Gimeno, J.; Lastra, E. Coord. Chem. Rev. 1999, 193−5,
147.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: lisarose@uvic.ca.
ACKNOWLEDGMENTS
■
We thank the Natural Sciences and Engineering Research
Council (NSERC) of Canada for funding.
REFERENCES
■
(1) (a) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg, L.
Organometallics 2007, 26, 1473. (b) Derrah, E. J.; Giesbrecht, K. E.;
McDonald, R.; Rosenberg, L. Organometallics 2008, 27, 5025.
(c) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg, L.
Angew. Chem., Int. Ed. 2010, 49, 3367. (d) Derrah, E. J.; McDonald, R.;
Rosenberg, L. Chem. Commun. 2010, 46, 4592.
2
(19) 31P{1H} shifts and JPP values for 8c,d are similar to those
previously observed for 8a.1a
(20) Due to the complexity of the aromatic region of the H NMR
1
spectra, we were unable to identify signals due to coordinated (7c,d)
(2) (a) Glueck, D. S. Top. Organomet. Chem. 2010, 31, 65.
(b) Waterman, R. Dalton Trans. 2009, 18.
(3) Derrah, E. J. Ph.D. Thesis, University of Victoria, Victoria, 2009.
or free (8c,d) benzonitrile in these samples.
(21) Morrow, K. M. E.; Pantazis, D. A.; McDonald, R.; Rosenberg, L.
Manuscript in preparation.
6464
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465

Organometallics
Article
(22) Programs for diffractometer operation, unit cell indexing, data
collection, data reduction and absorption correction were those
supplied by Bruker.
(23) Beurskens, P. T.; Beurskens, G.; de Gelder, R.; Smits, J. M. M.;
Garcia-Granda, S.; Gould, R. O. The DIRDIF-2008 Program System;
Crystallography Laboratory, Radboud University, Nijmegen, The
Netherlands, 2008.
(24) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.
(25) (a) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996,
77, 3865. (b) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett.
1997, 78, 1396.
(26) (a) Douglas, M.; Kroll, N. M. Ann. Phys. 1974, 82, 89.
(b) Jansen, G.; Hess, B. A. Phys. Rev. A 1989, 39, 6016.
(27) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
3297.
(28) Pantazis, D. A.; Chen, X. Y.; Landis, C. R.; Neese, F. J. Chem.
Theory Comput. 2008, 4, 908.
(29) Neese, F. ORCA-an ab initio, Density Functional and Semi-
empirical Program Package, 2.8.0; Universitat Bonn, Bonn, Germany,
̈
2010.
6465
dx.doi.org/10.1021/om200822e|Organometallics 2011, 30, 6458−6465