Inner- and Outer-Sphere Roles of Ruthenium Phosphido Complexes in the Hydrophosphination of Alkenes

Article abstract of DOI:10.1021/acs.organomet.5b00835

An inner-sphere synthetic cycle for the hydrophosphination of alkenes is proposed, based on observed [2 + 2] cycloaddition of a wide range of alkenes at a coordinatively unsaturated Ru=PR₂ complex. Key intermediates in the cycle were prepared, and their reactions with various organic acid/base pairs were examined to identify both new ruthenium precursors and base cocatalysts that allow turnover of the proposed cycle. Two new cationic ruthenium indenyl phosphine complexes were isolated and structurally characterized. Although preliminary screening studies show the moderate activity of these and related neutral phosphido complexes for catalytic hydrophosphination of acrylonitrile by both HPPh₂ and HPCy₂, and comparable activity for the hydrophosphination of tert-butyl acrylate by HPPh₂, no activity was observed for the analogous hydrophosphination of 1-hexene. This is attributed to strong binding of the substrate phosphine to the unsaturated, planar Ru - PR₂ fragment generated in situ, which inhibits the inner-sphere, alkene cycloaddition mechanism. An alternative, outer-sphere Michael addition process, involving a saturated complex with a strongly nucleophilic pyramidal Ru-PR₂ ligand, is proposed to rationalize the observed selectivity for catalytic hydrophosphination of activated, but not simple, alkenes. Implications for further catalyst development are discussed.

Full text of DOI:10.1021/acs.organomet.5b00835

Article
pubs.acs.org/Organometallics
Inner- and Outer-Sphere Roles of Ruthenium Phosphido Complexes
in the Hydrophosphination of Alkenes
Roman G. Belli,† Krista M. E. Burton,†,§ Stephanie A. Rufh,† Robert McDonald,‡ and Lisa Rosenberg*,†
†Department of Chemistry, University of Victoria, P.O. Box 3065, Victoria, British Columbia, Canada V8W 3V6
‡X-ray Crystallography Laboratory, Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada, T6G 2G2
S
* Supporting Information
ABSTRACT: An inner-sphere synthetic cycle for the hydro-
phosphination of alkenes is proposed, based on observed [2 +
2] cycloaddition of a wide range of alkenes at a coordinatively
unsaturated RuPR2 complex. Key intermediates in the cycle
were prepared, and their reactions with various organic acid/
base pairs were examined to identify both new ruthenium precursors and base cocatalysts that allow turnover of the proposed
cycle. Two new cationic ruthenium indenyl phosphine complexes were isolated and structurally characterized. Although
preliminary screening studies show the moderate activity of these and related neutral phosphido complexes for catalytic
hydrophosphination of acrylonitrile by both HPPh2 and HPCy2, and comparable activity for the hydrophosphination of tert-butyl
acrylate by HPPh2, no activity was observed for the analogous hydrophosphination of 1-hexene. This is attributed to strong
binding of the substrate phosphine to the unsaturated, planar RuPR2 fragment generated in situ, which inhibits the inner-
sphere, alkene cycloaddition mechanism. An alternative, outer-sphere Michael addition process, involving a saturated complex
with a strongly nucleophilic pyramidal Ru−PR2 ligand, is proposed to rationalize the observed selectivity for catalytic
hydrophosphination of activated, but not simple, alkenes. Implications for further catalyst development are discussed.
We previously described the synthesis of a series of highly
reactive terminal phosphido complexes Ru(η5-indenyl)(
PR2)(PPh3) (1), which are prepared via dehydrohalogenation
of the corresponding secondary phosphine complexes Ru(η5-
indenyl)Cl(PR2H)(PPh3) (2) using the strong base KOBut.5
We have investigated particularly thoroughly the reactions of 1
for R = Cy, which include the intriguing [2 + 2] cycloaddition
of a range of both simple and activated alkenes, giving
metallacycles 3.6 The mechanism by which this P−C bond-
forming reaction occurs is of particular interest, given the wide
scope of alkenes that can participate, and the limited examples
in the literature of such well-defined insertions of alkenes into
metal−heteroatom bonds. A manuscript describing our kinetic
and computational investigation of this process is currently in
preparation.7 However, the cycloaddition of alkenes at this
RuP double bond also represents the basis of a synthetic,
potentially catalytic, cycle leading to hydrophosphination
(Figure 1). The cycle features the use of an external base to
deprotonate the metal-coordinated secondary phosphine,
similar to some mechanisms that have been proposed for
late-metal-catalyzed hydrophosphination.8,9 The resulting con-
jugate acid can then deliver the proton to complete the
hydrophosphination cycle: in this case it is essential for
protonolysis of the metallacycle formed from the apparent
inner-sphere cycloaddition step.
INTRODUCTION
■
The hydrophosphination of alkenes by primary or secondary
phosphines represents an atom-economical route to the
synthesis of secondary or tertiary phosphines containing new
alkyl substituents. Unlike the analogous N−H addition
reaction, hydroamination, a catalyst is not necessarily required
for hydrophosphination: the reaction can occur thermally or
photochemically or can be initiated by acids, bases, or radicals.1
However, catalytic hydrophosphination can potentially intro-
duce regio- and stereoselectivity into this process that may
obviate the tedious and reagent-wasting separations currently
required in the production of structurally sophisticated and
value-added phosphine ligands. Catalytic hydrophosphination
may also allow the introduction of P-alkyl groups that are
challenging to obtain via noncatalyzed hydrophosphination
processes or classic, stoichiometric salt metathesis routes
employing MPR2 or ClPR2 reagents.
Accordingly, interest in metal-catalyzed hydrophosphination
has been growing,2 and among recent examples, the importance
of metal phosphido (M−PR2) intermediates has been
increasingly highlighted. In early, or electron-deficient, metal
systems the insertion of alkenes into the M−P bond appears to
be critical to P−C bond formation,3 while for many examples of
late, or electron-rich, metal centers, the substrate scope seems
to indicate the importance of a Michael-type addition of the
highly nucleophilic phosphido P at activated, electrophilic
alkenes.4 Thus, an apparent dichotomy exists between catalytic
systems participating in inner-sphere, versus outer-sphere, P−C
bond-forming reactions.
We present here reactions of complexes 1 and 2 confirming
the viability of this synthetic cycle, including studies to identify
Received: October 7, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
respectively12) will cleave the Ru−C bond in all three
metallacycles, to give the new mixed phosphine products
Ru(η5-indenyl)Cl{PCy2(CH2CH2R′)}PPh3 (4a−c) (eq 2).
Figure 1. Synthetic cycle for the inner-sphere hydrophosphination of
alkenes by the secondary phosphine HPCy2, mediated by
coordinatively unsaturated phosphido complex 1 ([Ru] = Ru(η5-
indenyl)PPh3).
The identity of 4a (R′ = Bun), isolated from a larger-scale
reaction of 3a with pyridine hydrochloride, was confirmed by
LIFDI-MS analysis (see the Supporting Information). 31P{1H}
NMR analysis of a d6-benzene sample of 4a showed that ligand
redistribution occurs readily in solution: singlets due to the
symmetric bis(phosphine) complexes Ru(η5-indenyl)Cl(PPh3)2
and (presumably) Ru(η5-indenyl)Cl{PCy2(CH2CH2Bun)}2
slowly replaced signals for 4a over 24 h. Further ligand
substitution reactions occurred with the addition of excess
HPCy2 to a solution of 4a generated in situ from the addition
of pyridine hydrochloride; the product mixture included the
regenerated mixed phosphine complex Ru(η5-indenyl)Cl-
(PCy2H)(PPh3) (2). These observations are an encouraging
indication of the substitutional lability of the product
phosphine, which is critical to a putative catalytic cycle in the
presence of excess HPCy2 (step D, Figure 1). The weakest acid
able to participate in the protonolysis, observed only for 3a, is
triethylamine hydrochloride (pKa = 10.7 (H2O)13). However,
neither pyridine nor triethylamine, conjugate bases of
sufficiently acidic salts, are capable of dehydrohalogenating
the secondary phosphine precursor 2 (step A in Figure 1), as
determined by 31P{1H} NMR; only signals due to unreacted 2
were observed.
a base capable of acting as a “proton shuttle” as shown in Figure
1. We report also our preliminary assessment of a series of
related Ru indenyl complexes for their activity as catalyst
precursors for alkene hydrophosphination. This points to the
operation of an alternative, outer-sphere cycle for hydro-
phosphination under catalytic conditions and provides insight
into the challenges of developing late-metal catalysts for the
hydrophosphination of unactivated alkenes.
RESULTS AND DISCUSSION
■
Identifying a Proton Shuttle. To identify the optimal base
cocatalyst for hydrophosphination using this ruthenium indenyl
system, we first screened various organic acids for their ability
to effect protonolysis of metallacycles 3 (step C in Figure 1).
Although we previously generated a series of metallacycles from
the addition of alkenes to isolated samples of complex 1,6 for
these reactivity studies we implemented a “one-pot” procedure
to obtain the metallacycles directly from the secondary
phosphine precursor complex 2 (eq 1). The addition of excess
These results suggested that starting from halide-free,
cationic analogues of complex 2 would simplify the role of
the added base in a putative catalytic cycle (Figure 2). This
alkene to 2 in toluene in the presence of KOBut allows the
formation of 1 in situ at room temperature and its subsequent
reaction to form metallacycle products. This is a cleaner and
higher-yielding alternative to isolating the highly soluble and
very reactive 1. We prepared both of the previously
characterized metallacycles 3a,b, where R = Cy and R′ = Bun,
CN, respectively, and a new analogue 3c derived from styrene
(R′ = Ph) using this method. The cleanliness of this one-pot
procedure bodes well for the ultimate adaptation of this
chemistry to catalysis: the presence of excess unsaturated
reagent as required in step B in Figure 1 does not interfere with
the base-mediated phosphine deprotonation in step A.
Possible protonolysis reactions of the three representative
metallacycles 3a−c were monitored by 31P{1H} NMR. These
experiments indicated that neither the secondary phosphine
substrate HPCy2 (pKa = 35.7 (THF)10) nor ButOH (pKa =
19.2 (H2O)11), the conjugate acid of the base we use to prepare
1, is sufficiently acidic to effect such protonolysis at the isolated
metallacycles. However, stronger acids such as pyridine and
lutidine hydrochlorides (pKa = 5.25 and 6.70 (H2O),
Figure 2. Possible catalytic cycle for the inner-sphere hydro-
phosphination of alkenes, in which phosphido complex 1 forms
from halide-free, cationic precursors ([Ru] = Ru(η5-indenyl)PPh3)
that contain labile nitrile ligands L.
would also prevent competition of the Cl− ligand with the
alkene and secondary phosphine substrates for coordination at
Ru. Furthermore, cationic secondary phosphine complexes
should exhibit P−H bond acidity higher than that of their
neutral chloride analogues,14 which might allow the use of
weaker bases as cocatalysts (providing stronger conjugate acids
for the eventual protonolysis step). We assessed the possible
participation of the cationic complex [Ru(η5-indenyl)(NCMe)-
B
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(PCy2H)(PPh3)2][PF6] (5) in the revised synthetic cycle
shown in Figure 2. Previous studies indicated the high lability of
the acetonitrile ligand in 5 and also the ready in situ generation
of 1 from this complex in the presence of the strong base
KOBut.5b However, 5 did not react with NEt3 to generate
complex 1 (step A in Figure 2), as determined by 31P{1H}
NMR. The cation did react with the stronger base 1,8-
diazabicycloundec-7-ene (DBU; pKa of the conjugate acid 13.5
(H2O)15) to give 1 in situ; addition of acrylonitrile “trapped”
the reactive phosphido complex as the metallacycle 3b.
The above results demonstrate the viability of the synthetic
hydrophosphination cycle shown in Figure 2 for R = Cy,
although we have not yet identified conditions/reagents to
allow a full cycle in a one-pot reaction.16 At this point we
decided to switch our focus to the use of the more acidic
phosphine HPPh2. This secondary phosphine is by far the most
commonly used substrate for catalytic hydrophosphination,
perhaps because of its ready availability, but probably also
because of its relatively low pKa.17 The greater acidity of this
phosphine, relative to HPCy2, should encourage a larger
equilibrium constant for deprotonation by DBU, an inter-
mediate-strength base, in step A of the cycle shown in Figure 2,
providing higher concentrations of both the phosphido
complex 1 (with R = Ph) and the conjugate acid DBU·H+.
Thus, as described below, we screened a series of halide-free
ruthenium indenyl complexes for their activity in hydro-
phosphination reactions of HPPh2.
helps to ensure the higher solubility of these ionic complexes in
less polar solvents such as benzene. We presumed that facile
substitution of one or more ligands at 6 would occur in the
presence of excess HPPh2, allowing the entry of this complex
into the putative catalytic cycle shown in Figure 2 prior to step
A.
The solution structures of complexes 6 and 7 were confirmed
1
by H, 31P{1H}, and 13C NMR spectroscopy and ESI mass
spectrometry (see the Experimental Section and the Supporting
Information). IR spectroscopy (KBr) also confirmed the
presence of the benzonitrile ligands in these complexes,
showing νCN 2228 and 2237 cm−1 for 6 and 7, respectively.18
Single crystals suitable for X-ray diffraction analysis were also
obtained for both complexes: the solid-state molecular
structures are shown in Figure 3, and selected interatomic
distances and bond angles are shown in Table 1. Crystallo-
Table 1. Selected Interatomic Distances (Å) and Bond
Angles (deg) for the Molecular Structures of the Cations
from [Ru(η5-indenyl)(NCPh)(PPh3)2][B(C6F5)4]·2.5C6H6
(6) and [Ru(η5-indenyl)(NCPh)(PPh2H)(PPh3)][B(C6F5)4]
a
(7)
6
7
Ru−P1
Ru−P2
Ru−N
Ru−C*
2.3123(5)
2.3664(5)
2.0426(16)
1.904
2.3148(5)
2.2889(5)
2.0310(16)
1.893
Catalyst Precursor Complexes. We prepared the two new
cationic complexes [Ru(η5-indenyl)(NCPh)(PPh3)2][B-
(C6F5)4] (6) and [Ru(η5-indenyl)(NCPh)(PPh2H)(PPh3)][B-
(C6F5)4] (7) through the addition of K[B(C6F5)4] to the
corresponding neutral chloride precursors (eq 3). These
Ru−H1P
Δ
1.34(2)
0.11
0.16
P1−Ru−P2
P1−Ru−N
P2−Ru−N
Ru−P1−H1P
Ru−N−C10
P1−Ru−C*
P2−Ru−C*
N−Ru−C*
98.240(17)
96.82(5)
91.84(5)
92.932(19)
91.07(5)
89.61(5)
117.4(9)
178.66(17)
124.4
169.95(16)
120.4
121.2
124.1
121.9
124.8
a
C* denotes the centroid of the plane defined by C(7A)−C(1)−
C(2)−C(3)−C(3A), and Δ is the “indenyl crystallographic slip
distortion”, defined as d(Ru−C(7A),C(3A)) − d(Ru−C(1),C(3)).19
complexes both contain the labile benzonitrile ligand, which
addresses a possible interference in the putative catalytic cycle
of competing metalation of an acetonitrile ligand by the highly
basic phosphido ligand in 1.5b The [B(C6F5)4]− counterion
graphic indenyl slip factors (Δ) of 0.16 and 0.11 Å for 6 and 7,
respectively, are well within the normal range for η5
coordination of the indenyl ring in these pseudo-octahedral
Figure 3. Molecular structures of the cations from [Ru(η5-indenyl)(NCPh)(PPh3)2][B(C6F5)4]·2.5C6H6 (6; left) and [Ru(η5-indenyl)(NCPh)-
(PPh2H)(PPh3)][B(C6F5)4] (7; right). Non-hydrogen atoms are represented by Gaussian ellipsoids at the 30% probability level. The hydrogen atom
attached to P1 in complex 7 is shown with an arbitrarily small thermal parameter; all other hydrogens are not shown.
C
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
complexes;19 the slightly larger value for complex 6 is
consistent with more steric crowding in this bis-
(triphenylphosphine) complex, although this value is still
much smaller than the corresponding slip factor of 0.21 Å
reported for the crystal structure of its neutral chloride-
containing precursor.20 The greater steric crowding in 6,
relative to 7, can also be seen from the wider angles at Ru
between the two phosphine and nitrile donor atoms and a
slight loss of linearity in the Ru−nitrile linkage: the Ru−N−
C10 bond angle in 6 is 169.95(16)°, while the nitrile in 7 is
almost linear, with an analogous angle of 178.66(17)°.
The cationic complexes 6 and 7 both provide entry points
prior to step A in the putative cycle shown in Figure 2. We also
chose to investigate the catalytic activity of two neutral, halide-
free complexes containing the terminal PPh2 ligand, whose
structures are related to the coordinatively unsaturated complex
1 by simple equilibria (Scheme 1); these give an entry point
Table 2. Reactions Screening the Activity of Complexes 6−9
a
in Catalytic Hydrophosphination
overall
Ru:DBU
(mol %)
early TOF
conversion
b
c
entry complex
(h−1
)
(%)
linear:branched
HPPh2 + H2CCH(CN):
1
6
7
8
9
10:10
10:10
10:10
10:10
0:0
2(0.6)
76(1)
90(4)
92(6)
72(5)
94:6
2
0.9(0.1)
7(0.6)
98:2
3
87:13
92:8
4
2 (0.5)
d
5
8(2)
99:trace
89:11
87:13
99:trace
99:trace
75:25
90:10
6
0:10
5:5
0.4(0.2)
13(1)
41(10)
73(10)
7(3)
7
8
6
7
8
9
8
10:0
10:0
10:0
10:0
Scheme 1
9
1(0)
10
11
5(0)
74(1)
66(3)
2(0.6)
HPPh2 + H2CCH(Bun)
d
12
13
8
8
10:10
10:0
trace
trace
HPPh2 + H2CCH(CO2But)
14
15
16
17
8
9
10:0
10:0
0:0
2.9(0.2)
1.7(0.3)
74(3)
75(4)
99:trace
99:trace
99:trace
99:trace
e
8
e
0:10
36
HPCy2 + H2CCH(CN)
d
18
19
20
6
10:10
0:0
0.7 (0.1)
43(12)
99:trace
trace
trace
0:10
a
Reactions were carried out in 3/1 THF/d6-benzene in NMR tubes at
prior to step B in the catalytic cycle in Figure 2. We previously
reported the dark purple-red benzonitrile complex Ru(η5-
indenyl)(PPh2)(NCPh)(PPh3) (8), which behaves as a masked
source of the RuPPh2 moiety;21 the thermal decomposition
of 8 to the ortho-metalated complex Ru(η5-indenyl){κ2-(o-
C6H4)PPh2}(PPh2H) provides strong evidence for the facile
dissociation of benzonitrile to give 1 (R = Ph) in situ, and this
was further supported by reactions with hydrogen and 1-
hexene, which gave 1,2-addition and [2 + 2] cycloaddition
products, respectively, analogous to those observed for
reactions of the same substrates with 1 (R = Cy, Pri).5 The
facile loss of benzonitrile from 8 allows this precursor to enter
the synthetic cycle at step B (Figure 2). The benzonitrile ligand
in 8 is readily substituted by 1 equiv of secondary phosphine to
give the deep purple mixed phosphine/phosphido complex
Ru(η5-indenyl)(PPh2)(PPh2H)(PPh3) (9), which is also related
to 1 by a simple equilibrium ( PPh2H; Scheme 1).22 We have
isolated complex 9 and included it in the catalyst screening
reactions described below.
Catalysis. We screened the activity of complexes 6−9 for
hydrophosphination of the activated substrate acrylonitrile by
Ph2PH (entries 1−4, Table 2), in the presence of the base
cocatalyst DBU, in NMR-tube reactions monitored by 31P{1H}
NMR. All four complexes showed moderate activity at room
temperature and 10% catalyst loading, giving conversions of
72−92% over 24 h to the linear, anti-Markovnikov product
Ph2PCH2CH2CN and small amounts (≤13%) of the branched
product Ph2PCH(CN)CH3. A control reaction (entry 5) with
no added Ru precatalyst or base cocatalyst indicated a thermal
room temperature, with periodic shaking to ensure mixing. For each
entry, alkene and secondary phosphine were present in equimolar
concentrations, and the results shown are the average of triplicate runs.
b
Quantitative 31P{1H} NMR experiments (128 scans, relaxation delay
55 s) were initiated within 15 min of sample preparation. These
experiments required 2 h 44 s, so the resulting spectra provide time-
averaged signal intensities corresponding approximately to the first 1 h
of reaction time. Unless otherwise noted, maximum deviations from
c
the average for triplicate runs are shown in parentheses. Unless
otherwise noted, conversions were calculated for spectra obtained after
24 h, from relative integration of 31P signals for hydrophosphination
product isomers26 and free HPR2. Other signals appeared in the region
due to Ru−P-containing complexes. This formula necessarily under-
estimates the conversions obtained for catalysts 7−9, all of which
contain 1 equiv or more of “PPh2” that may be incorporated in one or
more of the products or residual HPPh2. Maximum deviations from
d
the average for triplicate runs are shown in parentheses. Conversion
e
calculated for 48 h of reaction, instead of 24 h. Results from just one
experiment.
background reaction giving less than 10% conversion to almost
exclusively the linear regiosiomer over the same time period, as
previously observed by Glueck23 and Morris14 for the same
reaction catalyzed by Pt and Ru, respectively.24 Complicating
interpretation of these results is our observation that this
hydrophosphination reaction is also catalyzed by DBU itself
under these conditions: entry 6 in Table 2 shows that, in the
absence of Ru, 10 mol % DBU gives 41% conversion, with
regioselectivity similar to that observed for the Ru/base
cocatalyst mixtures. This was not entirely unanticipated; the
D
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 4. Inner- and outer-sphere catalytic cycles for the hydrophosphination of alkenes ([Ru] = Ru(η5-indenyl)PPh3). The two cycles are linked by
the equilibrium between intermediate 1 and its HPPh2 adduct, complex 9, which apparently competes effectively with cycloaddition of alkene
substrates at 1.
use of 1 mol % KOBut in this reaction under similar conditions
(THF, room temperature) was recently reported to give “less
than 10%” conversion after 24 h,14 which works out to
approximately the same overall activity we are observing for
DBU.25
containing species whose signals are attributable to mixed
phosphine and mixed phosphine/phosphido complexes of
varying composition.29 Further support for the relevance of this
outer-sphere catalytic cycle comes from the fact that we do not
observe signals due to the putative metallacycle intermediate 3
that would form from cycloaddition of acrylonitrile or 1-hexene
in spectra of the reaction mixtures represented by entries 3, 4,
and 12 in Table 2,30 although this does not rule out the
possibility that such intermediates are forming and are rapidly
consumed in subsequent protonolysis reactions. We previously
showed that this metallacycle is formed in the reaction of 8 with
1-hexene in d6-benzene31 but that the analogous carbonyl
adduct Ru(η5-indenyl)(PPh2)(CO)(PPh3), which forms irre-
versibly and is not a source of the RuPPh2 fragment, does not
react with this simple alkene.32 Thus, the fact that we observe
no signals for metallacycle formation and that we observe no
catalytic turnover for 1-hexene suggests that the formation of 9
from 8 under our catalyst screening conditions may be similarly
irreversible: in other words, if the equilibria shown in Scheme 1
lie heavily toward complex 9, complex 8 cannot be regarded as
a source of “RuPPh2” in the presence of excess HPPh2. A
thermolysis experiment monitoring the decomposition of
complex 9 to the ortho-metalated complex Ru(η5-indenyl)-
{κ2-(o-C6H4)PPh2}(PPh2H) via dissociation of HPPh2 suggests
that the barrier to this dissociation is higher than that for the
corresponding dissociation of benzonitrile from complex 8; the
half-life for decomposition of complex 9 at 60 °C is about 96 h,
in comparison to 4 h for complex 8.21 It is likely that, in the
presence of excess HPPh2, loss of HPPh2 from complex 9 to
regenerate the vacant coordination site at 1 is not sufficiently
favored to allow competitive alkene cycloaddition.
The outer-sphere catalytic cycle shown on the right side of
Figure 4 includes a zwitterionic intermediate that results from
nucleophilic attack of the phosphido ligand in complex 9 at the
electrophilic carbon of the acrylonitrile double bond in step E.
We had presumed, on the basis of the inner-sphere cycle
proposed in Figure 2, that external base would act as a proton
shuttle in these reactions; [DBU·H]+ present in the reaction
mixture would be required to protonolyze the proposed
metallacyclic intermediate. In the outer-sphere cycle in Figure
4, the role of this proton source is simply to “quench” the
carbanion in the zwitterionic intermediate (step F); the base is
then available to deprotonate another 1 equiv of Ru-bound
secondary phosphine (step G). However, this carbanion is
likely to be extremely basic14 and thus could be protonated by
species far less acidic than [DBU·H]+. For example, Glueck has
shown that either Pt−H or added ButOH or water can transfer
a proton to a similar carbanion during Pt-catalyzed hydro-
phosphination,4a,b while Morris has proposed that even free
We calculated “initial” turnover frequencies (TOF) to gauge
the progress of these reactions after approximately 1 h (Table
2), to get a sense of how efficiently each of the catalyst
precursors was being activated for catalysis. Since we obtained
both the highest initial TOF and the highest overall conversion
for the neutral phosphido complex 8 (entry 3), we also assessed
the activity of this complex for a lower catalyst loading of 5 mol
% (entry 7, Table 2), finding a higher initial TOF and only
slightly reduced conversion over 24 h. However, despite the
relatively high catalytic activity of complex 8 for the
hydrophosphination of acrylonitrile by HPPh2, this phosphido
complex shows essentially no activity for the hydrophosphina-
tion of 1-hexene under comparable conditions (Table 2, entry
12), even after 48 h. If we assume that the catalytic cycle shown
in Figure 2 is in effect for the acrylonitrile reaction, the absence
of catalysis for 1-hexene could be ascribed to a much lower rate
of cycloaddition of 1-hexene, relative to acrylonitrile. For
example, kinetic studies of alkene cycloadditions at the Ru
PCy2 analogue of 1 show that the rate constant for
cycloaddition of 1-hexene is at least 2 orders of magnitude
smaller than that for cycloaddition of acrylonitrile.7 Such a rate
difference might be sufficient to allow other reactions of the Ru
precatalyst to compete effectively with productive catalysis.
Another possible explanation for the lack of activity of
complex 8 in the hydrophosphination of 1-hexene by HPPh2 is
that the cycle shown in Figure 2 is not accessible for either
alkene substrate, under catalytic conditions. Although we know
from previous studies that benzonitrile dissociation occurs
readily from complex 8 to generate the coordinatively
unsaturated “RuPPh2” in solution,21 in the presence of
excess HPPh2 complex 8 reacts rapidly to give the secondary
phosphine/phosphido complex 9.27 Figure 4 illustrates how
this complex represents an entry point into an alternative
catalytic cycle that relies on the outer-sphere Michael addition
of the pyramidal phosphido ligand in this coordinatively
saturated complex at the electron-deficient, activated acryloni-
trile substrate.28 Such nucleophilic addition would be unlikely
to occur at the electron-rich 1-hexene. Tellingly, the only
signals observed in the Ru−P region of 31P{1H} NMR spectra
for the reactions of 1-hexene with HPPh2 in the presence of
complex 8 (Table 2, entries 12 and 13) are those due to
complex 9, whereas for the reactions of acrylonitrile with
HPPh2 mediated by either 8 or 9 (Table 2, entries 3, 4, 7, 10,
and 11) we do not see complex 9 but do see a variety of Ru-
E
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
HPPh2 can play this role in the hydrophosphination of
acrylonitrile catalyzed by a Cp*Ru phosphido complex,
generating free [PPh2]−, which competes for binding to Ru
in the reaction mixture.14 Consistent with this hypothesis, we
find that both phosphido complexes 8 and 9 are active for the
hydrophosphination of acrylonitrile in the absence of DBU
cocatalyst (entries 10 and 11 in Table 2).33 As for the Morris
system, the most likely proton source in these mixtures is the
substrate HPPh2; we have not yet undertaken experiments to
determine whether the protonation step is intermolecular (i.e.,
from free phosphine in the mixture) or intramolecular (from a
more acidic Ru-bound HPPh2). The intramolecular pathway
(F′ in Figure 4) is an appealing one, though, since it directly
generates not only the product tertiary phosphine, which can
undergo substitution by incoming HPPh2, but also a new PPh2
phosphido ligand, capable of intermolecular nucleophilic attack
at acrylonitrile.
Although the proposed outer-sphere hydrophosphination
mechanism shown in Figure 4 is consistent with the observed
reactivity of activated acrylonitrile versus unactivated 1-hexene,
an alternative outer-sphere mechanism involving the inter-
molecular attack of free HPPh2 at a putative N-bound
acrylonitrile−ruthenium adduct would also account for this
substrate selectivity.34 Precedent for this mechanism comes
from Togni’s studies of the Ni-catalyzed enantioselective
hydrophosphination of methacrylonitrile by a range of
secondary phosphines.35 However, unlike the Togni system,
formation of the acrylonitrile adduct in our system does not
seem to be favored. In screening reactions of the four catalyst
precursors examined in this study, we found evidence for the
formation of such an adduct only for complex 6: a new 31P
singlet at 46.9 ppm, consistent with the formula [Ru(η5-
indenyl)(NCCHCH2)(PPh3)2]+, appears adjacent to the
signal due to unreacted 6, when excess acrylonitrile is added to
6 in the absence of HPPh2 (Figure S21 in the Supporting
Information). The same signal appears only transiently, and at
much lower intensity, when both excess acrylonitrile and excess
HPPh2 are added to 6. This signal does not show up at all
under catalytic conditions, when DBU is also present (entry 1
in Table 2), or in any reactions of the other catalyst precursors.
More convincing evidence for the importance of the outer-
sphere phosphido mechanism shown in Figure 4, instead of this
putative N-bound acrylonitrile pathway, is our observation of
comparable catalytic activity for the hydrophosphination of tert-
butyl acrylate using either complex 8 or 9 (entries 14 and 15 in
Table 1). We would not expect the relatively sterically
encumbered carbonyl group in this substrate to form an O-
bound adduct, yet catalysis proceeds.
CONCLUSIONS
■
These results highlight the importance of coordinative
unsaturation at the metal in encouraging inner-sphere, alkene
insertion pathways for hydrophosphination and demonstrate
the challenges in broadening the alkene substrate scope for late-
metal catalysts. Currently we are examining the controlled
introduction of secondary phosphine substrate in these
reactions to identify conditions that may allow less activated
alkenes to compete for binding and insertion at Ru in these
systems, in addition to refining our choice of proton shuttle to
encourage an inner-sphere mechanism. We also continue to
probe the activity of this system for the hydrophosphination of
activated alkenes and have begun to explore the reactions of
other phosphines and other electron-deficient substrates in this
context.
EXPERIMENTAL SECTION
■
General Details. Unless otherwise noted, all reactions and
manipulations were performed under nitrogen in a glovebox or
using conventional Schlenk techniques. Methanol was distilled from
calcium hydride. THF was distilled from sodium under nitrogen and
then stored over sodium/benzophenone, and was degassed by three
freeze−pump−thaw cycles and vacuum-transferred before use.
Dichloromethane, toluene, pentane, and hexane were degassed by
sparging and then passed through columns of activated alumina in a
solvent purification system. Deuterated solvents (Sigma-Aldrich) were
stored over sodium/benzophenone (d6-benzene) or calcium hydride
(d1-chloroform) and then freeze−pump−thaw degassed three times
and vacuum-transferred before use, except for d2-dichloromethane and
d8-THF, which were used as received in 1 g ampules. Unless otherwise
specified, reagents were purchased from Sigma-Aldrich Canada and
used as received (secondary phosphines, or dried and degassed using
established procedures (other reagents). Diphenylphosphine was also
purchased from Strem Chemicals as a 10% solution in hexanes; the
concentration was checked against a known concentration of
triphenylphosphine oxide by 31P{1H} NMR before use. K[B(C6F5)4]
was purchased from Boulder Scientific Co. and was used as received.
Pyridine hydrochloride was recrystallized (CHCl3/ethyl acetate)
before use. Ru(η5-indenyl)Cl(PHR2)PPh3 (2; R = Cy, Ph),36 2,6-
lutidine hydrochloride,37 Ru(η5-indenyl)Cl(PPh3)2,38 and Ru(η5-
indenyl)(PPh2)(NCPh)PPh3 (8)21 were prepared using literature
procedures.
NMR spectra were acquired on a Bruker AMX 360 spectrometer
operating at 360.13 MHz for 1H and 145.78 MHz for 31P and a Bruker
1
AVANCE 500 spectrometer operating at 500.13 MHz for H, 202.43
MHz for 31P, and 125.77 MHz for 13C. Chemical shifts are reported in
1
ppm at ambient temperature. H chemical shifts are referenced to
residual protonated solvent peak at 7.16 ppm (C6D5H). 13C chemical
shifts are referenced to C6D6 at 128.4 ppm and CDCl3 at 77.5 ppm. 1H
and 13C chemical shifts are reported relative to tetramethylsilane, and
31P chemical shifts are reported relative to 85% H3PO4(aq).
Melting/decomposition temperatures were recorded using a
Gallenkamp apparatus and are uncorrected. Microanalysis was
performed by Canadian Microanalytical Services Ltd., Delta, BC,
Canada. IR spectra were recorded for KBr pellets on a PerkinElmer
FTIR Spectrum 1000 spectrophotometer. Electrospray ionization mass
spectrometry (ESI-MS) was carried out by Rhonda Stoddard (group
of Prof. Scott McIndoe) and Dr. Jingwei Luo at the University of
Victoria on Waters QTOF Micromass and QTOF II instruments,
using the following conditions: capillary voltage 3000 V, sample cone
voltage 15 V, extraction voltage 0.5 V, source temperature 60 °C,
desolvation temperature 120 °C, cone gas 100 L/h, desolvation gas
100 L/h, collision energy 2 V.
As mentioned above, in the growing literature of metal-
catalyzed hydrophosphination of alkenes, the use of phosphine
substrates that contain aryl substituents predominates. In this
context, it is notable that the hydrophosphination of
acrylonitrile by HPCy2 is catalyzed by 10 mol % of complex
6 with DBU, giving 43% conversion to a single product after 48
h at room temperature (Table 2, entry 18). While this
conversion is suspiciously close to that which we observed for
the HPPh2 reaction catalyzed by DBU alone, control reactions
without Ru, and without either Ru or DBU, indicate that
neither the base-catalyzed nor the thermal hydrophosphination
reaction is occurring appreciably for this dialkyl substrate
(Table 2, entries 19 and 20).
One-Pot Synthesis of Metallacycles 3a−c. General Procedure.
Toluene (10 mL) was placed in a Schlenk flask containing RuCl(η5-
indenyl)(HPCy2)(PPh3) (2) and KOBut (1.2 equiv) to give an orange
suspension. With stirring the solution quickly became brown with a
F
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
blue meniscus, and unsaturated substrate was added by gastight
syringe. The mixture was stirred for 18 h to give a clear yellow or
orange solution and then worked up as described below. 31P{1H}
equiv) was added, and the mixture was monitored by 31P{1H} NMR.
Complex 4a was completely consumed after 24 h, giving a mixture of
products that includes Ru(η5-indenyl)Cl(PCy2H)(PPh3) (2), Ru(η5-
indenyl)Cl{PCy2(CH2CH2Bun)}2, and several other phosphine-
containing products that have not yet been identified (see Supporting
Information).
1
NMR data for all of the metallacycles and H and 13C NMR data for
3c are given in the Supporting Information.
Ru(η5-indenyl)(κ2-BunCHCH2PCy2)(PPh3) (3a). Reagents: 2 (563
mg, 0.790 mmol), KOBut (137 mg, 1.22 mmol), 1-hexene (400 mL,
3.20 mmol). Solvent was removed under vacuum to give a yellow
residue that was extracted with a 5/1 mixture of hexanes and toluene
(25 mL) and filtered through Celite. The solvent volume was reduced
under vacuum to 5 mL, and acetonitrile (15 mL) was added via
cannula, resulting in the immediate precipitation of an orange powder.
The product was isolated by filtration, washed with cold hexanes (3 ×
5 mL), and dried under vacuum to give 3a (393 mg, 0.52 mmol, 65%).
1H and 31P{1H} NMR spectra indicate the exclusive formation of syn-
Reactions of [Ru(η5-indenyl)(NCMe)(PCy2H)(PPh3)2][PF6]
(5).39 Reaction of 5 with DBU. DBU (350 mg, 2.3 mmol, 100
equiv) was added to a solution of finely ground complex 5 (20 mg,
0.023 mmol) in d1-chloroform in a J. Young NMR tube. The solution
turned dark green-blue, and an initial 31P{1H} NMR showed signals
due to complex 15a as the major product. The complex subsequently
decomposed to a variety of unidentified products in this solvent.
Reaction of 5 with DBU and Acrylonitrile. DBU (18 mg, 0.12
mmol, 5 equiv) and acrylonitrile (6 mg, 0.1 mmol, 5 equiv) were
added to a solution of finely ground complex 5 (20 mg, 0.023 mmol)
in a 70/30 mixture of d6-benzene and PhF in a J. Young NMR tube.
The solution gradually changed color from yellow to orange, and
31P{1H} NMR confirmed the slow formation of metallacycle 3b.6
3a. Our previous preparation of 3a from the addition of 1-hexene to
the isolated phosphido complex 1 gave 9% anti-3a upon workup.6
Ru(η5-indenyl)(κ2-NCCHCH2PCy2)(PPh3) (3b). Reagents: 2 (317
mg, 0.445 mmol), KOBut (65 mg, 0.52 mmol), acrylonitrile (60 μL,
1.3 mmol). Solvent was removed under vacuum to give an orange
residue that was extracted with a 5/1 mixture of hexanes and toluene
(25 mL) and filtered through Celite. The solvent volume was reduced
under vacuum to 1 mL, and acetonitrile (10 mL) was added via
cannula, resulting in the immediate precipitation of a bright orange
powder, 3b. The complex was isolated by filtration, washed with cold
hexanes (3 × 5 mL), and dried under vacuum (210 mg, 0.29 mmol,
66% yield). 1H and 31P{1H} NMR spectra indicate the exclusive
formation of syn-3b. Our previous preparation of 3b from the addition
of acrylonitrile to the isolated phosphido complex 1 gave 8% anti-3b
upon workup.6
Synthesis of Catalyst Precursors 6, 7, and 9. 1H and 13C NMR
assignments are given in the Supporting Information.
[Ru(η5-indenyl)(NCPh)(PPh3)2][B(C6F5)4] (6). Benzonitrile (0.35 mL,
3.4 mmol) was added to an orange suspension of 2 (0.875 g, 1.13
mmol) and KB(C6F)4 (0.809 g, 1.13 mmol) in methanol (25 mL).
The resulting orange mixture was stirred for 1 day, during which time
it turned yellow with a yellow precipitate. The solution was filtered,
and the solid was dried under vacuum. It was then dissolved in
dichloromethane, filtered to remove KCl, and layered with methanol,
which gave 6 as an orange crystalline solid (1.25 g, 70% yield). Anal.
Found (calcd for C76H42NP2RuBF20): C, 59.81 (59.94); H, 2.64
(2.78); N, 0.85 (0.92). IR (KBr, cm−1): 2228 (w, vCN). LR-ESI-MS
(CH2Cl2, m/z): 844.35 (M+, 100%), 741.34 ([M − NCPh]+, 40%).
Dec pt: 172−175 °C. 31P{1H} NMR (202.43 MHz, d2-dichloro-
methane): δ 46.9 (s, PPh3).
Ru(η5-indenyl)(κ2-PhCHCH2PCy2)(PPh3) (3c). Reagents: 2 (307 mg,
0.431 mmol), KOBut (60 mg, 0.50 mmol), styrene (0.16 mL, 1.4
mmol). The solution was filtered through Celite, and the solvent was
removed from the filtrate to give a red-orange powder (>99% syn-3c,
273 mg, 81% crude yield). A portion of the crude sample (120 mg,
0.15 mmol) was recrystallized from dichloromethane (5 mL) by slow
layer diffusion of acetonitrile (∼50 mL) to give analytically pure 3c (94
mg, 80% yield, > 98% syn isomer). Anal. Found (calcd for
C47H52P2Ru): C, 71.97 (72.38); H, 6.63 (6.72). Mp: 164 °C. Note:
unlike the case for 3a,b, gentle heating of syn-3c in solution gave rapid
epimerization, yielding a 1/1 syn/anti mixture within minutes at 60 °C.
Reactions of Metallacycles 3a−c with Protic Reagents. NMR-
Scale Protonolysis Experiments. The reactions of complexes 3a−c
with pyridine hydrochloride, lutidine hydrochloride, and triethylamine
hydrochloride were monitored by 31P{1H} NMR. Formation of the
new mixed phosphine product Ru(η5-indenyl)Cl{PCy2(CH2CH2R′)}-
PPh3 (4, where R′ = Bun (a), CN (b), Ph (c)) was diagnosed by the
appearance of new pairs of doublets in the 31P{1H} NMR spectra. In
some cases, subsequent ligand redistribution reactions were observed,
as diagnosed by the appearance of the signal due to Ru(η5-
indenyl)Cl(PPh3)2 (1). Full details of these experiments and
31P{1H} NMR data for complexes 4a−c are included in the Supporting
[Ru(η5-indenyl)(NCPh)(PPh2H)(PPh3)][B(C6F5)4] (7). The procedure
described for 6 was followed, using benzonitrile (0.87 mL, 8.40
mmol), Ru(η5-indenyl)(Cl)(PPh2H)(PPh3) (0.982 g, 1.40 mmol), and
KB(C6F5)4 (1.00 g, 1.40 mmol) in methanol (25 mL). The reaction
mixture was stirred for 3 days before workup, which gave 7 as an
orange crystalline solid (1.28 g, 63% yield). Anal. Found (calcd for
C70H38NP2RuBF20): C, 58.29 (58.11); H, 2.46 (2.65); N, 0.90 (0.97).
IR (KBr, cm−1): 2322 (w, vPH), 2237 (w, vCN). LR-ESI-MS (CH2Cl2,
m/z): 768.07 (M+, 100%). Dec pt: 173−175 °C. 31P{1H} NMR
(202.43 MHz, d-chloroform): δ 56.3 (d, 42, PPh3), 35.0 (d, PPh2H).
Ru(η5-indenyl)(PPh2)(PPh2H)(PPh3) (9). KOtBu (44 mg, 0.40
mmol) was added to an orange solution of Ru(η5-indenyl)(Cl)-
(PPh2H)(PPh3) (232 mg, 0.33 mmol) and HPPh2 (0.17 mL, 0.99
mmol, 3 equiv) in toluene (30 mL) in a Schlenk flask. The solution
turned deep purple and was stirred for 1 h. The solution was then
filtered through Celite to remove KCl and HOtBu. The solvent was
removed under vacuum to give a purple solid, which was washed with
pentane (3 × 10 mL) and dried (190 mg, 68% yield). We have so far
been unable to obtain this complex free of minor impurities (see the
Supporting Information for NMR spectra); thus, it has not been
submitted for microanalysis. IR (KBr, cm−1): 2310 (w, vPH). Dec pt:
178−180 °C. 31P{1H} NMR (202.43 MHz, d8-THF): δ 49.1 (d, 35,
PPh3), 42.6 (dd, PPh2H), 40.5 (d, 15, PPh2).
X-ray Diffraction Studies of Complexes 6 and 7. Crystals of 6
were grown from a solution of the compound in benzene, while
crystals of 7 were grown by layering methanol onto a dichloromethane
solution of the compound. Suitable crystals were coated with a thin
layer of Paratone-N and then mounted on a Bruker PLATFORM
diffractometer equipped with an APEX II CCD area detector,40 with
the crystals cooled to −80 °C under a cold nitrogen gas stream.
Diffraction measurements were made using Mo Kα radiation (λ =
0.71073 Å). The structure of 6 was solved with the direct methods
program SHELXTL-2014,41a while the structure of 7 was determined
using a Patterson search for heavy atoms followed by structure
expansion, as implemented in the DIRDIF-200842 program system.
The structure refinements were completed using the least-squares
Information.
Synthesis of Ru(η5-indenyl)Cl(PCy2(hexyl))PPh3 (4a). Pyridine
hydrochloride (5.84 M in acetonitrile, 50 μL, 0.29 mmol) was added
to a solution of Ru(η5-indenyl)(κ2-BunCHCH2PCy2)PPh3] (3a; 212
mg, 0.279 mmol) in toluene (10 mL) in a Schlenk flask. The resulting
red solution was stirred for 24 h at room temperature. Removal of the
solvent under vacuum gave a red solid, which was extracted with
pentane (3 × 5 mL) and transferred by cannula to another Schlenk.
The solvent was removed under vacuum, leaving pure 4a (50 mg, 23%
yield). 31P{1H} NMR analysis showed that ligand redistribution occurs
readily in solution; LIFDI-MS analysis of a freshly prepared toluene
solution of the product confirmed the identity of 5d and also showed
small amounts of 1 and the free PCy2(hexyl) ligand (see the
Supporting Information).
Reaction of 4a with HPCy2. Pyridine hydrochloride (5.84 M in
acetonitrile, 4 μL, 0.023 mmol) was added to 12 mg (0.016 mmol) of
3a dissolved in d6-benzene (0.4 mL) to generate 4a in situ, as
determined by 31P{1H} NMR. HPCy2 (0.4 M in hexane, 0.2 mL, 5
G
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
refinement program SHELXL-2014.41b The phosphorus-bound hydro-
gen atom in 7 was located from a difference Fourier map, and its
coordinates and isotropic displacement parameter were freely refined.
All other hydrogens were generated in idealized positions on the basis
of the sp2 or sp3 geometries of their attached carbons and given
isotropic displacement parameters 120% of the Ueq values for their
parent C atoms.
Catalyst Screening and Associated Control Reactions.
General Procedure for Catalytic Runs. To 1 equiv of catalyst (0.02
mmol) in a J. Young NMR tube was added 1 equiv of DBU (0.2 mL,
0.1 M in THF), 10 equiv of alkene (0.2 mL, 1.0 M in THF), and 10
equiv of HPR2 (0.2 mL, 1.0 M in THF) and d6-benzene (0.2 mL).
Each reaction was performed in triplicate; conversions and product
ratios were determined from relative integrations of the product and
free secondary phosphine signals in 31P{1H} NMR spectra obtained
using a gated decoupled experiment with a relaxation delay of 55 s.
General Procedure for Control Reactions. To 0.2 mL of C6D6 in a
J. Young NMR tube was added two or more of 0.02 mmol of a Ru
catalyst, 0.2 mL of HPR2, alkene (1.0 M in THF, 10 equiv), or DBU
(0.1 M in THF, 1 equiv). In reactions where one or two reagents were
left out, the corresponding volume of THF (0.2 or 0.4 mL) was added
to maintain the same concentrations as would be present in the
catalytic reactions described above. The reactions were monitored by
31P{1H} NMR, using the quantitative pulse sequence where necessary.
(d) Wauters, I.; Debrouwer, W.; Stevens, C. V. Beilstein J. Org.
Chem. 2014, 10, 1064−1096.
(2) Recent reviews of catalytic hydrophosphination include:
(a) Rosenberg, L. ACS Catal. 2013, 3, 2845−2855. (b) Koshti, V.;
Gaikwad, S.; Chikkali, S. H. Coord. Chem. Rev. 2014, 265, 52−73.
(c) Greenhalgh, M. D.; Jones, A. S.; Thomas, S. P. ChemCatChem
2015, 7, 190−222. (d) Rodriguez-Ruiz, V.; Carlino, R.; Bezzenine-
Lafollee, S.; Gil, R.; Prim, D.; Schulz, E.; Hannedouche, J. Dalton
Trans. 2015, 44, 12029−12059.
(3) This mechanism was originally delineated by Marks: (a) Douglass,
M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 2001, 123, 10221−
10238. (b) Motta, A.; Fragala, I. L.; Marks, T. J. Organometallics 2005,
24, 4995−5003. Recent examples of alkene hydrophosphination
catalyzed by electron-deficient metals include: (c) Liu, B.; Roisnel, T.;
Carpentier, J. F.; Sarazin, Y. Chem. - Eur. J. 2013, 19, 13445−13462.
(d) Ghebreab, M. B.; Bange, C. A.; Waterman, R. J. Am. Chem. Soc.
2014, 136, 9240−9243. (e) Basalov, I. V.; Dorcet, V.; Fukin, G. K.;
Carpentier, J. F.; Sarazin, Y.; Trifonov, A. A. Chem. - Eur. J. 2015, 21,
6033−6036. (f) Kissel, A. A.; Mahrova, T. V.; Lyubov, D. M.;
Cherkasov, A. V.; Fukin, G. K.; Trifonov, A. A.; Del Rosal, I.; Maron,
L. Dalton Trans. 2015, 44, 12137−12148.
(4) This mechanism has been studied in-depth by Glueck for Pt-
based catalysts: (a) Scriban, C.; Kovacik, I.; Glueck, D. S.
Organometallics 2005, 24, 4871−4874. (b) Scriban, C.; Glueck, D.
S.; Zakharov, L. N.; Kassel, W. S.; DiPasquale, A. G.; Golen, J. A.;
Rheingold, A. L. Organometallics 2006, 25, 5757−5767. Recent
examples of alkene hydrophosphination catalyzed by electron-rich
metals include: (c) Chew, R. J.; Teo, K. Y.; Huang, Y. H.; Li, B. B.; Li,
Y. X.; Pullarkat, S. A.; Leung, P. H. Chem. Commun. 2014, 50, 8768−
8770. (d) Gallagher, K. J.; Webster, R. L. Chem. Commun. 2014, 50,
12109. (e) Hao, X. Q.; Huang, J. J.; Wang, T.; Lv, J.; Gong, J. F.; Song,
M. P. J. Org. Chem. 2014, 79, 9512−12111. (f) Isley, N. A.; Linstadt, R.
T. H.; Slack, E. D.; Lipshutz, B. H. Dalton Trans. 2014, 43, 13196−
13200. (g) Brown, C. A.; Nile, T. A.; Mahon, M. F.; Webster, R. L.
Dalton Trans. 2015, 44, 12189−12195. (h) Geer, A. M.; Serrano, A. L.;
de Bruin, B.; Ciriano, M. A.; Tejel, C. Angew. Chem., Int. Ed. 2015, 54,
472−475. (i) Xu, Y.; Yang, Z. H.; Ding, B. Q.; Liu, D. L.; Liu, Y. G.;
Sugiya, M.; Imamoto, T.; Zhang, W. B. Tetrahedron 2015, 71, 6832−
6839. (j) Yang, X. Y.; Gan, J. H.; Li, Y. X.; Pullarkat, S. A.; Leung, P. H.
Dalton Trans. 2015, 44, 1258−1263.
The results of these reactions are shown in Table 2 and/or
described in the text, and representative spectra are included in the
Supporting Information.
Thermolysis of Complex 9 Monitored by 31P{1H} NMR. A
solution of 9 (20 mg, 0.024 mmol) in d8-toluene was placed in a flame-
sealable NMR tube, degassed by three freeze−pump−thaw cycles, and
sealed under vacuum. An initial 31P{1H} NMR spectrum was obtained.
The solution was heated at 60 °C in an oil bath and removed
periodically for monitoring by 31P{1H} NMR. After 96 h,
approximately 50% of 9 had decomposed, giving the ortho-metalated
complex [Ru(η5-indenyl){κ2-(o-C6H4)PPh2}(PPh2H)] as well as free
PPh3 and a corresponding complex tentatively identified as Ru(η5-
indenyl)(PPh2)(PPh2H)2 (see the Supporting Information).
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.5b00835.
(5) (a) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg, L.
Organometallics 2007, 26, 1473−1482. (b) Derrah, E. J.; Giesbrecht, K.
E.; McDonald, R.; Rosenberg, L. Organometallics 2008, 27, 5025−
5032.
Crystallographic data for complexes 6 and 7 (CIF)
Additional experimental details, including NMR spectra
of new compounds and representative product mixtures,
and crystallographic data for complexes 6 and 7 (PDF)
(6) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg, L.
Angew. Chem., Int. Ed. 2010, 49, 3367−3370.
(7) Morrow, K. M. E., M.Sc. Thesis, University of Victoria, 2012.
(8) Malisch, W.; Klupfel, B.; Schumacher, D.; Nieger, M. J.
Organomet. Chem. 2002, 661, 95−110.
AUTHOR INFORMATION
Corresponding Author
*E-mail for L.R.: lisarose@uvic.ca.
(9) See: Chew, R. J.; Li, X. R.; Li, Y. X.; Pullarkat, S. A.; Leung, P. H.
Chem. - Eur. J. 2015, 21, 4800−4804 as well as refs 4a and 4j and
references therein..
■
(10) Issleib, K.; Kummel, R. J. Organomet. Chem. 1965, 3, 84−91.
(11) Stewart, R. The Proton: Applications to Organic Chemistry;
Academic Press: Orlando, FL, 1985; Chapter 2.
Notes
The authors declare no competing financial interest.
§Nee
́
Morrow.
(12) Rodima, T.; Kaljurand, I.; Pihl, A.; Maemets, V.; Leito, I.;
Koppel, I. A. J. Org. Chem. 2002, 67, 1873−1881.
ACKNOWLEDGMENTS
(13) Kolthoff, I. M.; Chantooni, M. K., Jr; Bhowmik, S. J. Am. Chem.
Soc. 1968, 90, 23−28.
■
We thank the NSERC of Canada for financial support
(Discovery Grant to L.R., CGS-M to R.G.B.) and Bochao
(Chris) Huang for preparing complex 5 and carrying out its
reactions with DBU.
(14) Sues, P. E.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2014,
136, 4746−4760.
(15) Kaupmees, K.; Trummal, A.; Leito, I. Croat. Chem. Acta 2014,
87, 385−395.
(16) Our continued assessment of reagents for this synthetic cycle
includes the addition of organic acids (pKa ≈ 11−14) containing
weakly coordinating counterions to isolated metallacycles 3 in the
presence of excess secondary phosphine. Preliminary experiments
REFERENCES
■
(1) (a) Quin, L. D. A Guide to Organophosphorus Chemistry; Wiley:
Hoboken, NJ, 2000. Other leading references include: (b) Crepy, K.
V. L.; Imamoto, T. Top. Curr. Chem. 2003, 229, 1−40. (c) Delacroix,
O.; Gaumont, A. C. Curr. Org. Chem. 2005, 9, 1851−1882.
−
using [lutidine·H]+BF4 in CH2Cl2 in the absence of secondary
phosphine gave mixed and somewhat messy results, which may be
H
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
attributed to the high reactivity of the coordinatively unsaturated
cationic product resulting from step D in Figure 2 and resulting
noninnocence of the solvent. Future experiments will exploit the
benzene-soluble B(C6F5)4 counterion.
(17) Literature pKa values for HPPh2 vary considerably, ranging from
(26) Literature δ(31P) data for product phosphines in d6-benzene:
−15.3 ppm (Ph2PCH2CH2CN, ref 4b), −20.1 ppm (Ph2PCH(CN)
CH3, ref 14), −14.7 (Ph2PCH2CH2CO2But, ref 4b). Under our
conditions (3/1 THF/d6-benzene), we observe signals within 1 ppm
of these values. We found no literature value for Ph2PCH(CO2But)
CH3 but do observe a very tiny peak at −23.3 ppm that we have
tentatively assigned to this branched isomer. For the reaction of
acrylonitrile with HPCy2 we observe one major product with δ(31P) at
−2.0 ppm that we have assigned to the linear product; 1H/31P HMBC
NMR confirms the connectivity of this tertiary P to Cy and other alkyl
groups (see the Supporting Information). We also see three tiny
signals in the range −2 to −10 ppm; we presume one of these is due to
the branched product Cy2PCH(CN)CH3 but have not identified the
other minor impurities.
(27) We carried out a control reaction under conditions comparable
to those for the catalyst screening reactions shown in Table 2, in which
10 equiv of HPPh2 was added to complex 8 in the absence of
acrylonitrile or DBU. This reaction shows complete conversion to
complex 9 within minutes of preparing the sample (see Figure S27 in
the Supporting Information).
(28) As noted in the Introduction, other late metal phosphido
complexes also undergo such Michael additions at activated alkenes.
See for examples refs 4a−c, 8, 9, and 14.
(29) See the Supporting Information, including Figures S29−S34, for
further discussion of, and tentative structural assignments for, the
spectroscopically observed “catalyst resting states” for the catalytic
reactions of acrylonitrile with HPPh2.
(30) We also do not see signals due to metallacycles 3 in control
reactions that omit the DBU cocatalyst (entries 10, 11, and 13 in Table
2).
−
21.7 (THF, ref 10) to 22.9 (calculated for DMSO, ref 17a) to 38
4
(THF, ref 17b). Importantly, where values for HPCy2 are also noted,
they are inevitably at least 7−14 pKa units higher (vide supra). (a) Li,
J. N.; Liu, L.; Fu, Y.; Guo, Q. X. Tetrahedron 2006, 62, 4453−4462.
(b) Abdur-Rashid, K.; Fong, T. P.; Greaves, B.; Gusev, D. G.; Hinman,
J. G.; Landau, S. E.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2000,
122, 9155−9171.
(18) These values are very close to that for free benzonitrile (νCN
=
2232 cm−1, ref 18a) as is commonly observed for metal-coordinated
nitriles (refs 18b and c). (a) Mrozek, M. F.; Wasileski, S. A.; Weaver,
M. J. J. Am. Chem. Soc. 2001, 123, 12817−12825. (b) Storhoff, B. N.;
Lewis, H. C. Coord. Chem. Rev. 1977, 23, 1−29. (c) Kuznetsov, M. L.;
Dement’ev, A. I.; Nazarov, A. A. Russ. J. Inorg. Chem. 2005, 50, 731−
739.
(19) Faller, J. W.; Crabtree, R. H.; Habib, A. Organometallics 1985, 4,
929−935. Typically a slip distortion of less than 0.25 Å indicates η5-
indenyl coordination.
(20) Kamigaito, M.; Watanabe, Y.; Ando, T.; Sawamoto, M. J. Am.
Chem. Soc. 2002, 124, 9994−9995.
(21) Hoyle, M. A. M.; Pantazis, D. A.; Burton, H. M.; McDonald, R.;
Rosenberg, L. Organometallics 2011, 30, 6458−6465.
(22) Similarly to complex 8 and other L adducts of complex 1 that
we have prepared previously (refs 5 and 21), the 31P{1H} spectrum of
complex 9 shows negligible 2JPP coupling between the PPh3 ligand and
the pyramidal PPh2 ligand (see the Experimental Section). This is the
first example we have prepared in which L is a phosphine; we expected
(31) We added 10 equiv of acrylonitrile to complex 8 in 3/1 THF/
d6-benzene in the absence of HPPh2 or DBU, in an attempt to observe
the corresponding metallacycle, which had not previously been
prepared. However, the NMR sample solidified. We suspect this
basic phosphido complex may promote polymerization (or telomeri-
zation, see refs 4a and b) of the acrylonitrile under these conditions.
We were unable to obtain NMR spectra for this solid to see if any
metallacycle forms under these conditions.
2
that JPP for coupling between the secondary phosphine HPPh2 and
pyramidal PPh2 ligands would also be suppressed, but it is actually
larger (15 Hz). We attribute these differences to conformational
preferences in this adduct, which affect the relative dihedral angles
between the stereochemically active lone pair at PPh2 and each of the
tertiary phosphines. For more on the “transition metal gauche effect”
see: (a) Buhro, W. E.; Zwick, B. D.; Georgiou, S.; Hutchinson, J. P.;
Gladysz, J. A. J. Am. Chem. Soc. 1988, 110, 2427−2439. (b) Barre, C.;
Boudot, P.; Kubicki, M. M.; Moise, C. Inorg. Chem. 1995, 34, 284−
291.
(32) As described in ref 21, heating a sealed sample of Ru(η5-
indenyl)(PPh2)(CO)(PPh3) in d8-toluene to 60 °C for up to 90 h gave
no reaction; none of the ortho-metalated complex Ru(η5-indenyl){κ2-
(o-C6H4)PPh2}(PPh2H) that would result from CO dissociation was
observed.
(33) For complexes 6 and 7, the DBU is needed to initiate catalysis
(either inner or outer sphere) by deprotonating coordinated secondary
phosphine; control reactions shown in entries 8 and 9 in Table 2
indicate that hydrophosphination of acrylonitrile is not catalyzed by
these complexes in the absence of DBU.
(34) Yet another mechanistic possibility has been proposed for the
Pd-catalyzed hydrophosphination of alkynes and for Rh-catalyzed
hydrophosphination of dimethyl fumarate and ethylene (ref 4h), which
involves oxidative addition of the P−H bond, followed by insertion of
the unsaturated substrate into the M−H bond, and subsequent
reductive elimination of a P−C bond. Such a mechanism is unlikely to
be occurring in this Ru system, since we see no evidence in any of our
reactions for the oxidative addition of a P−H bond (cycles involving
such Ru(II)/Ru(IV) oxidation state changes are rare), and no Ru−H
intermediates are observed. Kazankova, M. A.; Efimova, I. V.;
Kochetkov, A. N.; Afanas’ev, V. V.; Beletskaya, I. P.; Dixneuf, P. H.
Synlett. 2001, 497−500. Ananikov, V. P.; Makarov, A. V.; Beletskaya, I.
P. Chem. - Eur. J. 2011, 17, 12623−12630.
(23) Wicht, D. K.; Kourkine, I. V.; Kovacik, I.; Glueck, D. S.;
Concolino, T. E.; Yap, G. P. A.; Incarvito, C. D.; Rheingold, A. L.
Organometallics 1999, 18, 5381−5394.
(24) The structurally similar half-sandwich complexes Ru(η5-Cp*)
(PPh2)(PP), where PP is a chelating diphosphine ligand, either cis-1,2-
bis(diphenylphosphino)ethylene or 1,2-bis(diphenylphosphinoben-
zene), reported by Morris in ref 14 show initial TOFs of 30−60
h−1, about 1 order of magnitude higher than the values we observe, but
the catalysts deactivate after 1 h. The higher initial activity could be
due simply to the stronger donating ability of Cp* relative to indenyl,
which should render the phosphido ligand more basic/nucleophilic
(vide infra). Glueck’s Pt diphosphine complexes (ref 23) show
activities closer to those which we observe (10 h−1), for reactions
carried out at 50 °C instead of room temperature.
(25) Other literature examples of the base-mediated reaction of
HPPh2 with acrylonitrile have exploited bases much stronger than
DBU in stoichiometric quantities. Two reports of the use of 50%
NaOH(aq) (refs 25a and b) cite the much earlier use of these
conditions for the reactions of H2PPh with acrylonitrile.25 Another
example (ref 25d) uses NEt4OH for the reaction of HPPh2 with
activated alkenes other than acrylonitrile. (a) Li, Y. Z.; Li, Z. M.; Li, F.;
Wang, Q. R.; Tao, F. G. Tetrahedron Lett. 2005, 46, 6159−6162.
(b) Dibowski, H.; Schmidtchen, F. P. Tetrahedron 1995, 51, 2325−
2330. (c) Rauhut, M. M.; Hechenbleikner, I.; Currier, H. A.; Schaefer,
F. C.; Wystrach, V. P. J. Am. Chem. Soc. 1959, 81, 1103−1107.
(d) Wiese, B.; Knuhl, G.; Flubacher, D.; Priess, J. W.; Ulriksen, B.;
Brodner, K.; Helmchen, G. Eur. J. Org. Chem. 2005, 2005, 3246−3262.
(35) Sadow, A. D.; Togni, A. J. Am. Chem. Soc. 2005, 127, 17012−
17024.
(36) Derrah, E. J.; Marlinga, J. C.; Mitra, D.; Friesen, D. M.; Hall, S.
A.; McDonald, R.; Rosenberg, L. Organometallics 2005, 24, 5817−
5827.
(37) Nuss, H.; Nuss, J.; Jansen, M. Z. Kristallogr. - New Cryst. Struct.
2005, 220, 95−96.
I
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(38) Oro, L. A.; Ciriano, M. A.; Campo, M.; Focesfoces, C.; Cano, F.
H. J. Organomet. Chem. 1985, 289, 117−131.
(39) Huang, B. B.Sc. (Honours) Thesis, University of Victoria, 2014.
(40) Programs for diffractometer operation, data collection, data
reduction and absorption correction were those supplied by Bruker.
(41) (a) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Adv. 2015,
71, 3−8. (b) Sheldrick, G. M. Acta Crystallogr. 2015, C71, 3−8.
(42) 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.
J
DOI: 10.1021/acs.organomet.5b00835
Organometallics XXXX, XXX, XXX−XXX