Synthesis and ruthenium coordination complexes of the chelating phosphine phosphonium-1-indenylide 1,1-bis(diphenylphosphino)methane-1-indenylide,1-C 9H6Ph2PCH2PPh2

Article abstract of DOI:10.1021/om300715a

Bis(diphenylphosphino)methane-1-indenylide, 1-C₉H ₆PPh₂CH₂PPh₂ (1-C₉H ₆dppm, I), has been synthesized and characterized by NMR spectroscopy and X-ray crystallography. I reacts with [CpRu(MeCN)₃]PF₆ to form the conventional sandwich complex [CpRu(η⁵-I)]PF ₆ (II), which contains a dangling -PPh₂ group. Complex II isomerizes to the 18-electron species IIIc, in which the dangling -PPh ₂ group coordinates to the ruthenium, forcing slippage of the five-membered ring to an unanticipated 1,9,8-η³-exo mode of coordination.

Full text of DOI:10.1021/om300715a

Article
pubs.acs.org/Organometallics
Synthesis and Ruthenium Coordination Complexes of the Chelating
Phosphine Phosphonium-1-indenylide 1,1-
Bis(diphenylphosphino)methane-1-indenylide,1‑C9H6Ph2PCH2PPh2
Riaz Hussain, Kevin G. Fowler, Francoise Sauriol, and Michael C. Baird*
Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada
Peter H. M. Budzelaar*
Department of Chemistry, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada
S
* Supporting Information
ABSTRACT: Bis(diphenylphosphino)methane-1-indenylide, 1-
C9H6PPh2CH2PPh2 (1-C9H6dppm, I), has been synthesized and characterized
by NMR spectroscopy and X-ray crystallography. I reacts with [CpRu-
(MeCN)3]PF6 to form the conventional sandwich complex [CpRu(η5-I)]PF6
(II), which contains a dangling −PPh2 group. Complex II isomerizes to the 18-
electron species IIIc, in which the dangling −PPh2 group coordinates to the
ruthenium, forcing slippage of the five-membered ring to an unanticipated 1,9,8-η3-exo mode of coordination.
riphenylphosphonium cyclopentadienylide, C5H4PPh3,
1
Thas been known since 1956, when its unusual stability
was noted and attributed to the electron delocalization implied
by the zwitterionic resonance structure b of Figure 1.
Figure 2. Resonance structures of phosphonium-1-indenylides.
indenylide, was reported in the 1960s, although it was not
characterized by NMR spectroscopy or crystallography.6a In
2004 two PHIN analogues, 1-C9H6P(CH2Ph)Ph2 and 1-
C9H6P(CH2C6F5)Ph2, were reported and characterized.6b
Adding to their interest, PHIN ligands are planar prochiral,
with the result that their coordination compounds exhibit
planar chirality (Figure 3). Thus, as an example, we reported
Figure 1. Resonance structures a and b of C5H4PPh3.
Phosphonium cyclopentadienylides of this type are thus
isoelectronic with neutral arene and anionic cyclopentadienyl
ligands, although, as has been noted, the coordination
chemistry of this class of very interesting ligands has received
little systematic attention in spite of its clear potential.2−4a In an
effort to expand the coordination chemistry of the phospho-
nium cyclopentadienylides, we earlier carried out an inves-
tigation of transition-metal complexes of phosphonium cyclo-
pentadienylides and reported the synthesis and reactivity of the
group 6 compounds M(η5-C5H4PMePh2)(CO)3 (M = Cr, Mo,
W) of the ligand C5H4PMePh2, methyldiphenylphosphonium
cyclopentadienylide.2,3 We found, inter alia, that the donor
properties of C5H4PMePh2 fall between those of benzene and
the cyclopentadienyl (Cp) ligand, consistent with the presence
of a partial negative charge on the five-membered ring and
suggesting that there would be interesting comparisons and
contrasts to be made with the chemistry of similar arene and
cyclopentadienyl complexes.
Figure 3. Enantiomers of the chiral complex Cr(η5-1-C9H6PMePh2)-
(CO)3.
the synthesis and characterization, including crystallographic, of
the planar chiral chromium compound Cr(η5-1-C9H6PMePh2)-
(CO)3.5 On the basis of the IR spectrum in the carbonyl region,
the indenylide ligand was found to exhibit donor properties
very similar to those of its cyclopentadienylide analogue.5
In order to broaden the chemistry of this general class of
ligand, we subsequently began an investigation of the
phosphonium-1-indenylide (PHIN) ligands 1-C9H6PR3 (Figure
2).5 The first such compound, triphenylphosphonium-1-
Received: July 27, 2012
Published: September 26, 2012
© 2012 American Chemical Society
6926
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
More recently we reported new and/or improved syntheses
of the PHIN ligands triphenylphosphonium-1-indenylide (1-
C9H6PPh3), methyldiphenylphosphonium-1-indenylide (1-
C9H6PMePh2), and dimethylphenylphosphonium-1-indenylide
(1-C9H6PMe2Ph), as well as syntheses of the corresponding
planar chiral ruthenium(II) complexes [Ru(η5-C5H5)(η5-1-
C9H6PPh3)]PF6, [Ru(η5-C5H5)(η5-1-C5H4PMePh2)]PF6, and
[Ru(η5-C5H5)(η5-1-C9H6PMe2Ph)]PF6.7 The ruthenium com-
plexes were characterized by 1H, 13C, and 31P NMR
spectroscopy, by X-ray crystallography, and by extensive DFT
calculations, which provided insights into the nature of the
metal−ligand bonding. The PHIN−Ru bond strengths were
calculated to be ∼20 kcal/mol greater than the corresponding
benzene−Ru bond strength of [Ru(η5-C5H5)(η6-C6H6)]+,
compatible with the observed configurational stability of the
chiral complexes.
The 1H NMR spectrum of the product mixture was complex
because of the presence of the two regioisomers and was not
fully interpreted. However, since isomer A contains a chiral
center at C(1), the P−CH2−P hydrogens are diastereotopic
and distinct one-hydrogen triplet resonances could be readily
identified at δ 4.53 and 4.94 (JHH ≈ JHP ≈ 14.6 Hz for both). In
contrast, the corresponding two-hydrogen resonance of isomer
B was reasonably assigned to a doublet at δ 4.49 (JHP ≈ 14.4
Hz).
Interestingly, a 1H−31P HMBC spectrum of a mixture
showed that both methylene protons in isomer A couple to the
phosphonium resonance at δ 30.4 and not to the resonance of
the dangling −PPh2 group at δ −25.6. Similarly, the methylene
protons in isomer B couple strongly to the phosphonium
resonance at δ 14.4 and relatively weakly to the resonance of
the dangling −PPh2 at δ −25.5. The reasons for the great
differences in values of 2JHP are attributed to differences in the s
character of the two types of P−CH2 bonds in isomers A and B.
It is well-known that values of JHP of trigonal phosphines
(essentially p3 bonding) are generally much lower than
corresponding values of JHP of analogous tetrahedral
phosphonium compounds (sp3 bonding),8a−c and the same
correlation appears to apply here also.
We have now extended the scope of PHIN ligands to an
example involving a diphosphine and report here the synthesis
and characterization of the potentially chelating phosphonium-
1-indenylide bis(1,1-diphenylphosphino)methane-1-indenylide,
1-C9H6dppm (I), containing a dangling −PPh2 tertiary
phosphine moiety.
X-ray-quality crystals of isomer A were obtained by layering a
CH2Cl2 solution with hexanes, and the structure is shown in
Figure 5. Full crystallographic information is given in the
The coordination chemistry of this ligand is explored, and we
have characterized the sandwich complex [Ru(η5-C5H5)(η5-1-
C9H6dppm)]PF6 (II), which isomerizes to a species which
appears to be the unprecedented chelate complex [CpRu(1,9,8-
η3-1-C9H6PPh2CH2-κPPh2]PF6 (IIIc).
RESULTS AND DISCUSSION
■
We previously prepared PHIN ligands by reaction of the
appropriate phosphine with 1-bromoindene to form the
corresponding phosphonium salt as mixtures of regioisomers;
these, on deprotonation with NaH in THF, afforded the
corresponding PHIN ligands.7 We have accordingly found that
reaction of 1-bromoindene with dppm in a 1:1 ratio gives in
good yield the expected phosphonium salt, (1-C9H7dppm)Br,
as a mixture of regioisomers A and B, as indicated in Figure 4.
The 31P NMR spectrum of the mixture exhibited two pairs of
doublets, those at δ 30.4 and −25.6 (JPP = 62.0 Hz) being
assigned on the basis of 1D and 2D NMR experiments as the
Figure 5. X-ray structure of isomer A of (1-C9H7dppm)Br.
Supporting Information, and we note here only that the
chirality at C(1) and the presence of the C(2)−C(3) CC
double bond (1.335(3) Å) are confirmed.
Deprotonation of (1-C9H7dppm)Br with NaH in THF
afforded the green, air-sensitive PHIN ligand 1-C9H6dppm (I).
1H and 13C NMR assignments can be found in Table 1 and
resonances of isomer A and those at δ 14.4 and −25.5 (JPP
=
63.6 Hz) as resonances of isomer B. The resonances with
negative chemical shifts are assigned to the dangling −PPh2
groups on the basis of their very close similarity and their
similarity to the 31P chemical shift of free dppm (δ −23.6).
Figure 4. Synthetic route to I. Also indicated is the numbering scheme utilized henceforth for spectroscopic and crystallographic purposes.
6927
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
1
Table 1. H and 13C NMR Data for I−IV
I
II
III
H, C
IV
Position δ(1H); J (Hz) δ(13C); J (Hz)
δ(1H); J (Hz)
δ(13C); J (Hz)
δ(1H); J (Hz)
δ(13C); J (Hz) δ(1H)
1
2
3
4
66.7 (dd, JCP
59.6 (d, JCP
46.4 (d, JCP
2.3, 120.1)
102.5)
81.5)
a
6.91 (t, JHH
JHP 4.8)
,
,
126.8 (d, JCP
∼14)
107.0 (d, JCP
15.4)
4.96 (t, JHH 2.5, JHP 2.5)
5.79 (t, JHH 2.5, JHP 2.5)
78.8 (dd, JCP
13.7, 6.9)
7.63 (s)
7.17 (s)
142.1 (d, JCP
8.7)
NA
6.57 (t, JHH
JHP 4.8)
71.9 (d, JCP
9.3)
123.6 (d, JCP
10.7)
6.46
138.3 (d, JCP
15.8)
95.2 (d, JCP
12.0)
148.3 (d, JCP
6.5)
5
6
7.59 (m)
120.6 (s)
117.3 (s)
7.54 (m)
126.7 (s)
126.0 (s)
7.84 (m)
7.49 (m)
122.8 (s)
125.2 (s)
6.96
6.64
6.87 (t, JHH
7.4)
7.03 (ddd, JHH 8.7, 6.5, 0.8)
6.87 (ddd, JHH 8.9, 6.4, 1.2)
6.75 (dd, JHH 8.9, 0.6)
7
8
9
6.70 (t, JHH
7.4)
117.9 (s)
117.8 (s)
125.8 (s)
7.60 (m)
129.0 (s)
70.4 (s)
6.21
6.50
6.90 (m)
124.4 (d, JCP
9.4)
4.45 (dd, JHH 6.4 JHP(2) 10.5)
135.9 (d, JCP
14.1)
97.4 (d, JCP
10.0)
98.0 (d, JCP
6.3)
PCH2P 3.73 (d, JHP
27.2 (dd, JCP
57.4, 35.4)
3.83 (m, JHH 15.0, JHP ∼14) 3.87
(m, JHH 15.0, JHP ∼14)
27.5 (dd, JCP
36.9, 27.3)
2.81 (dt, JHH 16.25 JHP 6.9), 4.10 30.6 (dd, JCP
2.81,
4.10
13.5)
(dt, JHH 16.15, JHP 9.3
70.2, 7.6)
Ph
7.20−7.75
128−136
7.30−7.83
125−135
see Figure 7
125−135
(m)
C5H5
4.34 (s)
73.8 (s)
3.58 (s)
80.0 (s)
4.16
a
NA = not assigned.
correspond well with values observed for previously reported
PHIN ligands.7 The 31P NMR spectrum exhibited doublet
resonances, at δ 7.5 and −28.3 (JPP = 61 Hz); these are
assigned to the ylidic and phosphine phosphorus atoms,
respectively, on the basis of comparisons with data for
previously reported PHIN ligands7 and free dppm (δ −23.6).
Of note, as with the isomeric phosphonium cations, the two
2JHP values for the P−CH2 group were quite different: 13.5 Hz
to the ylidic phosphorus and ∼0 Hz for the phosphine
phosphorus atom.
X-ray-quality crystals of I were obtained by layering a CH2Cl2
solution with hexanes, and the structure is shown in Figure 6.
The bond lengths of I are very similar to the corresponding
bond lengths of 1-C9H6PPh3, 1-C9H6PMePh2, and 1-
C9H6PMe2Ph.7 Thus, for instance the P(1)−C(1) bond length
of I is 1.715(3) Å in comparison with an average of 1.72 Å for
the other PHIN ligands and is shorter than the P(1)−CH2 and
P(2)−CH2 bond distances, 1.812(2) and 1.859(2) Å,
respectively, and the average of the P−Ph distances, 1.82 Å.
Thus, the P−ylide bond possesses considerable double-bond
character. The dimensions of the five- and six-membered rings
also compare well with those of 1-C9H6PPh3, 1-C9H6PMePh2,
and 1-C9H6PMe2Ph,7 with some shortening of C(2)−C(3)
(1.365(4) Å), C(5)−C(6) (1.378(4) Å), and C(7)−C(8)
(1.371(3) Å) relative to the other ring C−C distances
(1.401(4)−1.441(3) Å). Full crystallographic information is
given in the Supporting Information.
Compound I was found to react with the labile ruthenium-
9
(II) complex [CpRu(MeCN)3]PF6 in CH2Cl2 to give bright
red solutions containing two major products, the relative
amounts of which varied with time. Reactions in CD2Cl2 were
therefore monitored by NMR spectroscopy, with the finding
that the 31P NMR spectrum of the kinetic product (II)
exhibited doublets at δ 23.0 and −29.1 (JPP = 69 Hz) while that
of the thermodynamic product (III) exhibited doublets at δ
37.7 and 58.1 (JPP = 71 Hz) (see the Supporting Information,
Figure S1); the conversion of II to III was complete within 10
days at room temperature. Spectra sometimes also contained
resonances of a very minor product, indicated by an AB quartet
at δ 21.63 and 21.24 (JPP = 12.1 Hz); this will be discussed
below.
By analogy with previous work which showed that reactions
of [CpRu(MeCN)3]PF6 with other PHIN ligands under similar
conditions give sandwich complexes of the type [CpRu(η5-
PHIN)]PF6,7 we anticipated that one of the products would be
[CpRu(η5-I)]PF6 (II).
Figure 6. X-ray structure of 1-C9H6dppm (I).
6928
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
Such a structure contains a dangling −PPh2 group, and it was
therefore gratifying to note that the chemical shift of the
resonance at δ −29.1 in the 31P spectrum of the kinetic product
is very similar to that of the dangling −PPh2 in the
uncoordinated PHIN-dppm ligand, δ −28.3. The resonance
at δ 23.0 is therefore to be attributed to the ylide phosphorus
and, by analogy with previously studied sandwich complexes of
the type [CpRu(η5-PHIN)]PF6,7 η5 coordination of I to Ru(II)
to give II should result in a downfield shift of the ylide 31P
resonance of about 14−21 ppm. This is as found (δ 23.0 vs δ
7.5; Δδ = 15.5 ppm), and thus the NMR data are all consistent
with assignment of the kinetic product as II. All attempts to
grow crystallographically useful crystals unfortunately failed but,
as with the aforementioned complexes of monodentate PHIN
ligands,7 a freshly prepared electrospray mass spectrum (ES-
MS, MeCN) exhibited solely a multiplet centered at m/e 665
with an isotopic pattern in excellent agreement with the
calculated spectrum.
ring hydrogens of II are very similar to those reported
previously.7
Interestingly, both of the 31P resonances of the thermody-
namic product (δ 37.7, 58.1) lie well downfield of the 31P
resonances of the free ligand. There are no precedents as yet for
such coordination shifts for coordination complexes of PHIN
ligands,5,7 but the apparently very large downfield coordination
shift from δ −28.3 of the dangling −PPh2 of free I calls to mind
previous observations of anomalously large 31P downfield
coordination shifts in the case of five-membered chelate rings.10
For instance, 31P chemical shifts of chelated 1,2-bis-
(diphenylphosphino)ethane (dppe) complexes are routinely
some 20−40 ppm downfield of the corresponding chemical
shifts of 1,2-bis(diphenylphosphino)propane (dppp) and 1,2-
bis(diphenylphosphino)butane (dppb) complexes, although the
31P chemical shifts of the free phosphines are similar.10 It thus
seemed quite likely that the thermodynamic product contains I
chelated to the ruthenium in such a manner that a five-
membered ring is formed, and we initially considered the η3-
allylic species [CpRu(2,1,9-η3-1-C9H6PPh2CH2-κPPh2]PF6
(IIIa) and [CpRu(1,2,3-η3-1-C9H6PPh2CH2-κPPh2]PF6 (IIIb).
Since II could not be characterized crystallographically, the
structure was of necessity inferred from a series of 1D and 2D
1
1
NMR experiments (31P, H, H{31P}, HMBC (H−C and H−
1
P), NOESY, COSY, HSQC). The H NMR spectrum of a
mixture of II and III (see the Supporting Information, Figure
S2) is very complex and challenging to interpret since, for
example, II and III both contain four nonequivalent phenyl
groups and, of course, the relative amounts of the two species
were continually changing. However, by monitoring relative
intensities as a function of time, it was possible to identify many
of the resonances of II, as indicated in Table 1; a number of the
phenyl resonances were obscured, and therefore these are not
listed. These assignments were strengthened to an extent by an
H−P HMBC experiment (Figure S5, Supporting Information),
which showed that the 31P resonance of II at δ −29.1 correlates
Structures IIIa,b both obey the 18-electron rule and thus
appear to be reasonable, although η3-allylic structures for
phosphonium cyclopentadienylides have not previously been
reported. Ring slippage as in IIIb would, however, be analogous
to the well-known indenyl effect11 in which an η5-indenyl
complex rearranges to an η3 mode of coordination, freeing up a
coordination site and permitting associative substitution
reactions of neighboring ligands. The process is made possible
because of stabilization of the six-membered aromatic ring.11
Since III could not be characterized crystallographically, the
structure was of necessity inferred from a series of ES-MS and
NMR experiments, as with II, supported by DFT calculations.
Experiments involved the utilization of somewhat aged samples
which contained predominantly III (by NMR), and it was
gratifying to note that the ES-MS spectrum of the product
mixture exhibited the same multiplet centered at m/e 665.
Thus, the kinetic and thermodynamic products were indeed
isomeric.
1
with multiplets in the H NMR spectrum at δ 3.85, 7.30, and
7.4, while that at δ 23.0 correlates with the multiplet at δ 3.85
and with others at δ 7.55, 7.67, and 7.82. In addition, the 31P
1
resonance of III at δ 37.8 correlates with multiplets in the H
NMR spectrum at δ 2.81, 4.10, 7.17, 7.41, 7.65, 7.84, and 8.20,
while that at δ 58.1 correlates with the multiplets at δ 2.81, 4.10,
and 4.45 and with others at δ 6.98, 7.19, 7.65, and 8.09.
Readily identified on the basis of relative intensities and a
lack of correlations with any other protons in a COSY spectrum
of a mixture (Figure S6, Supporting Information) were the Cp
and PCH2P resonances of II at δ 4.34 (s) and 3.85 (m),
respectively. The latter multiplet is actually composed of two
doublet of doublets because the protons at position 10 are not
1
equivalent and appear as an AB quartet in the H{31P} NMR
spectrum (Figure S3, Supporting Information). The COSY
spectrum also showed that the resonance at δ 4.96 correlated
only with that at δ 5.79 and these were assigned to H(2) and
H(3), respectively, of II. In confirmation of these assignments,
a NOESY spectrum (Figure S7, Supporting Information)
exhibited through-space interactions of the resonance at δ 5.79
with that at δ 4.96 and also with a resonance at δ 7.54, which is
therefore assigned to H(5). Similar correlations allowed
assignment of resonances at δ 7.03, 6.87, and 6.75 to H(6),
H(7), and H(8), respectively. 13C assignments for II, based on
HSQC (Figure S8, Supporting Information) and HMBC
(Figure S9, Supporting Information) experiments, can be
found in Table 1. As anticipated, since II is similar in structure
to previously studied sandwich complexes of the type
[CpRu(η5-PHIN)]PF6,7 the chemical shifts of the indenyl
As with II, the Cp (δ 3.58 (s)) and nonequivalent PCH2P (δ
2.81 (dt), 4.10 (dt)) resonances were readily identified on the
basis of relative intensities and a lack of correlations with any
other protons in a COSY spectrum of a mixture (Figure S6,
Supporting Information) (although, of course, the resonances
at δ 2.81 and 4.10 were mutually coupled). The COSY
spectrum also exhibited correlation between resonances at δ
7.17 and 7.63, neither of which exhibited correlations to other
protons. As discussed above for II, this type of correlation is as
expected for the resonances of H(2) and H(3) and it was
subsequently found that NOESY correlations existed between
these two resonances, between the resonance at δ 7.17 with
another at δ 7.84 and between resonances at δ 4.43 and 7.60.
Thus, the resonances at δ 7.63 and 7.17 are assigned to H(2)
and H(3), respectively, and those at δ 7.84, 7.60 and 4.46 are
6929
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
assigned to H(5), H(7), and H(8), respectively. These
assignments were confirmed by detailed considerations of the
COSY spectrum, which also suggested that a resonance at δ
7.49 may be assigned to H(6).
Interestingly, an H−P HMBC spectrum exhibited a weak
correlation between the resonance of H(3) at δ 7.17 and the
phosphorus resonance at δ 37.8, and thus the latter can be
assigned to the ylidic phosphorus P(1) and that at δ 58.1 to the
coordinated phosphorus P(2). (A stronger correlation between
P(1) and H(2) is likely, but the presumed cross-peak is
obscured.)
atoms in η5-1-indenyl complexes of ruthenium7 and is observed
here for the corresponding carbon atoms of II.
These observations seem most consistent with the structure
of [CpRu(1,9,8-η3-1-C9H6PPh2CH2-κPPh2]PF6 (IIIc), in which
the ruthenium center is bonded in an allylic fashion through
C(1), C(9), and C(8).
In an attempt to obtain more information about the structure
of III, it was noted that the 1H resonances at δ 7.41, 8.20, 6.98,
and 8.09 were all greatly simplified upon phosphorus
decoupling (compare Figures S2−S4); it follows that these
resonances are to be attributed to the ortho protons of the four
P-phenyl rings. Further detailed considerations of the COSY
and H−P HMBC spectra resulted in conclusions that the
resonances at δ 7.41 and 8.20 may be attributed to the phenyl
rings bonded to the ylidic phosphorus P(1) and those at δ 6.98
Corroborating this structure, as noted above, spin−spin
coupling is observed between the P(2) resonance at δ 58.1 and
the resonance of H(8) at δ 4.45, three-bond coupling which
requires coordination of C(8) to the ruthenium as in IIIc.
Rationalization of the aforementioned NOEs in terms of this
structure will be presented below in the context of DFT
calculations.
1
and 8.09 to the coordinated phosphorus P(2). Indeed, all H
chemical shifts on the four phenyl groups were ultimately
determined from COSY and NOESY spectra and are shown in
Figure 7.
Interestingly, the NOESY spectrum of IIIc showed a
significant number of cross peaks due to exchange in addition
to through-space interactions. In fact, every peak assigned to
IIIc has an exchange cross peak with an unknown compound
IV, with the exception of the methylene protons. The peaks in
the 1H NMR spectrum corresponding to IV are too weak to be
analyzed meaningfully, and the equilibrium between these
1
species is strongly in favor of IIIc; however, most of the H
chemical shifts for IV can be assigned and are given in Table 1.
Isomerization of IIIc to IV results in shielding of all the
protons on the indenylide ring, with the exception of H(8),
which is deshielded considerably to a value more typical for a
phenyl proton. Coordination of I in IV therefore does not
involve C(8). Several of the phenyl protons also shift
considerably upon isomerization; for example, the ortho proton
at δ 8.10 shifts to δ 6.35. The dramatic change in chemical
environment may be caused by shifting from the deshielding to
the shielding zone of the Cp ring.
1
Figure 7. H chemical shifts of the phenyl groups in III: Ph(1) and
Ph(2) bonded to the ylidic phosphorus P(1) and Ph(3) and Ph(4)
bonded to P(2).
The NOESY spectrum also provided insight into the relative
orientations of the indenylide and the groups of the dppm
moiety in III. Strong NOEs were observed between the
resonance of H(8) and the CH2 resonance at δ 2.81
(henceforth proton H(10a)), and between the Cp resonance
and the CH2 resonance at δ 4.10 (henceforth proton H(10b)).
Thus, H(10b) is oriented toward the Cp ligand while H(10a)
lies relatively close to H(8) and thus must point toward the six-
membered ring of the indenylide ligand. The fact that the
resonance of proton H(10a) does not exhibit an NOE with the
resonances of H(2) or H(3) would seem to rule out either IIIa or
IIIb as the structure of the thermodynamic product.
An NOE was also observed between the resonance of the Cp
ligand and that of the o-H of Ph(3), as were medium NOEs
between the resonance of the Cp ligand and those of the o-H
atoms of Ph(2) and Ph(4). However, no NOE was observed
between the resonance of the Cp ligand and that of the o-H of
Ph(1), and thus the Cp ligand must be oriented toward Ph(2),
Ph(3), and Ph(4) but away from Ph(1).
Finally, as noted above, synthesis of mixtures of II and IIIc
sometimes resulted in the formation of a relatively minor
species which exhibited an AB quartet at δ 21.63 and 21.24 (JPP
= 12.1 Hz) in the 31P NMR spectrum. This species was not
always present in reaction mixtures, and we ignored it until we
happened to run a 31P NMR spectrum of a sample which had
aged somewhat in an NMR tube. The 31P NMR spectrum of
the sample was now found to exhibit, in comparable amounts,
the resonances of IIIc and the AB quartet at δ 21.63, 21.24,
while an ES-MS spectrum exhibited both the multiplet of IIIc
centered at m/e 665 and a second multiplet centered at m/e
681, with an isotopic pattern in excellent agreement with that
anticipated for a monoruthenium species. The difference in
mass between IIIc and the new species corresponds to one
oxygen atom, and it seems likely that the phosphine
phosphorus atom had undergone slow oxidation by adventi-
tious oxygen to give the corresponding η5 complex of 1-
C9H6PPh2CH2POPh2, containing a dangling phosphine oxide
A series of 13C, HSQC, and HMBC spectra (Figures S8 and
S9, Supporting Information) were used to assign 13C
resonances of III, and the resulting chemical shifts and
coupling constants can be found in Table 1. Of note are the
unusual chemical shifts of C(8) (δ 70.4) and C(9) (δ 98.0),
shielded by ∼48 and 38 ppm, respectively, relative to the
corresponding chemical shifts of the free ligand. Similar
shielding has been reported previously for bound carbon
1
group. While assignments of neither the H nor the 31P NMR
spectra have been possible, we note that the two 31P chemical
shifts are very similar to both that of the phosphonium 31P
resonance of II, as would be expected, and also those in the
compounds Ph2P(O)CH2PPh2 (δ 27.76)8d and Ph2P(O)-
CH2P(O)Ph2 (δ 24.5).8e
6930
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
Computational Results. Neither II nor IIIc could be
recrystallized satisfactorily, and computational studies were
therefore carried out for the η5-1-indenyl complex II and for
several η3-κPPh2 chelating isomers such as IIIa−c. Complex II
was found to exhibit a sandwich structure very similar to those
of analogous structures containing simple 1-indenylides,7 and
although the calculated Ru−C distances are slightly longer than
analogous crystallographically determined Ru−C distances,7 the
metal−indenyl bonding was found to exhibit the expected three
short (∼2.22 Å) and two long (∼2.32 Å) Ru−C bond
distances. In addition, the six-membered ring of the 1-
indenylide ligand exhibits short C(5)−C(6) and C(7)−C(8)
bonds. As indicated above, the NMR data for II are also
consistent with the anticipated sandwich structure, which now
seems confirmed.
A number of discrete local minima, differing in which
carbons of the indenyl ligand interact with the ruthenium atom,
were found for the η3-κPPh2 chelate isomers but, to our initial
surprise, neither IIIa nor IIIb was the lowest energy isomer.
The most stable structure located is in fact essentially that
discussed above as IIIc, albeit with the ylidic phosphorus P(1)
lying well above the five-membered ring, which is essentially
planar (Figure 8).
the Cp ring but that Ph(1) is oriented well away from Cp is also
borne out by the structure shown in Figure 8.
The energy difference between the η5 and η3-κPPh2 chelate
structures is rather small, 1−3 kcal/mol, depending on the level
of calculation, and our final “best estimate” for the free energy
difference in solution is actually 3.2 kcal/mol in favor of the η5
structure. The difference is well within the limits of computa-
tional uncertainty, and we can only conclude that the two
structures are close in energy. The η5 structure, with its larger
flexibility of the dangling −PPh2 group, is slightly favored by
entropy contributions (the enthalpy difference is only 1.4 kcal/
mol). For the simplified model system bearing PH2 groups, we
calculate a very similar energy difference (ΔΔG = 3.8 kcal/mol,
ΔΔH = 2.6 kcal/mol) in favor of the η5 structure.
Lying about 13 kcal/mol higher than the η3-κPPh2 chelate
structure discussed above is an η3 structure which, surprisingly,
is neither IIIa nor IIIb. Rather, it is the complex [CpRu(2,3,4-
η3-1-C9H6PPh2CH2-κPPh2]PF6 (IIId), which contains a six-
membered ring. In this the average of the Ru−C(2), Ru−C(3),
and Ru−C(4) bonds is 2.52 Å, while the Ru−C(1) and Ru−
C(9) distances are both calculated to be >3.2 Å. Attempts were
made to investigate the slipped 1,2,3-η3 isomer IIIb, but
surprisingly, a local minimum could not be found and we
suspect none exists because of steric crowding and ring strain.
CONCLUSIONS
■
The ylidic ligand 1-C9H6PPh2CH2PPh2, readily prepared from
1-bromoindene and dppm, can potentially bind to metals in a
variety of coordination modes. NMR experiments strongly
suggest that the kinetic product from reaction of 1-
C9H6PPh2CH2PPh2 with [CpRu(MeCN)3]PF6 is the sandwich
complex [CpRu(η5-I)]PF6 (II), containing the 1-
C9H6PPh2CH2PPh2 coordinated via the five-membered ring
in an η5 fashion and with a dangling −PPh2 group. In solution,
slow isomerization occurs to a thermodynamic product in
which the phosphine moiety coordinates to the ruthenium. A
combination of NMR spectroscopy and DFT calculations has
allowed the elucidation of an unexpected 1,9,8-η3 coordination
mode of the 1-C9H6PPh2CH2PPh2. The preference for this
structure over the expected 1,2,3-η3 coordination mode is most
likely caused by constraints imposed by the ligand framework.
EXPERIMENTAL SECTION
■
All syntheses were carried out under dry, deoxygenated argon using
standard Schlenk line techniques. Argon was deoxygenated by passage
through a heated column of BASF copper catalyst and then dried by
passing through a column of activated 4A molecular sieves. NMR
spectra were recorded using Bruker AV500 and AV600 spectrometers,
1H and 13C NMR data being referenced to TMS via the residual
proton signals of the deuterated solvent and 31P data to external
Figure 8. Computed structure of the lowest energy η3-allylic isomer,
corresponding to IIIc.
In this structure, the Ru−C(1), Ru−C(8), and Ru−C(9)
distances are calculated to be in the range 2.27−2.55 Å, while
the Ru−C(2), Ru−C(3), and Ru−C(4) distances are all
calculated to be >3.2 Å and thus nonbonding. Interestingly, the
C(2)−C(3), C(4)−C(5), and C(6)−C(7) distances are all
relatively short, consistent with the structure shown.
The geometry of structure IIIc is also consistent with subtle
details of the NOESY experiments discussed above. For
instance, the NOESY experiment suggested strongly that
H(10a) lies close to H(8) rather than to H(2) or H(3), a
conclusion which ruled out structures IIIa,b. The calculated
H(8)−H(10a) and H(2)−H(10a) distances are 2.54 and 5.37
Å, respectively, very consistent with the NOE data and the
structural conclusion drawn therefrom. The conclusion, drawn
from the NOE data, that Ph(2), Ph(3), and Ph(4) all lie near
9
phosphoric acid. [Ru(η5-C5H5)(MeCN)3]PF6 was purchased from
Strem Chemicals, while 1-bromoindene was prepared via cleavage of
the carbon−silicon bond of 1-trimethylsilylindene by dioxane
dibromide.12
Synthesis of (1-C9H7PPh2MePPh2)Br. To a solution of 1.24 g of
dppm (3.22 mmol) in 50 mL of toluene was added dropwise 0.88 g of
1-bromoindene12 (4.52 mmol). The resulting reaction mixture was
stirred at room temperature for 24 h as an off-white precipitate
formed. The precipitate was filtered off, washed with toluene (3 × 5
mL) and hexanes (2 × 5 mL), and then dried in vacuo for 1 h to give
1.2 g (74% yield) of off-white product. 1H NMR of isomer A (CD2Cl2,
600 MHz): δ 4.94 (t, J = 14.4 Hz, Ph2P+CHaHbPPh2), 4.53 (t, J = 14.4
Hz, Ph2P+CHaHbPPh2), 6.60 (m, CCHCHP+Ph2), 6.86 (m, CH
CHCHP+Ph2), 7.23 (overlap, CCHCHP+PH2), 7.05−8.11 (m,
aromatic). 31P NMR of isomer A (CD2Cl2): δ 30.4 (d,
6931
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932

Organometallics
Article
1
P+Ph2CH2PPh2), −25.6 (d, P+Ph2CH2PPh2). H NMR of isomer B
(CD2Cl2, 600 MHz): δ 4.48 (d, J = 14.3 Hz, CHP+Ph2CH2PPh2), 3.80
(s, CH2CHCP+Ph2CH2PPh2) 8.05 (d, CH2CHCP+PH2), 7.05−
8.11(m, aromatic). 31P NMR of isomer B (CD2Cl2): δ 14.4 (d,
P+PH2CH2PPh2), −25.5 (d, P+Ph2CH2PPh2).
REFERENCES
■
(1) (a) Ramirez, F.; Levy, S. J. Org. Chem. 1956, 21, 488. (b) Ramirez,
F.; Levy, S. J. Org. Chem. 1956, 21, 1333. (c) Ramirez, F.; Levy, S. J.
Am. Chem. Soc. 1957, 79, 67. (d) Ramirez, F.; Dershowitz, S. J. Org.
Chem. 1957, 22, 41. (e) Ramirez, F.; Levy, S. J. Am. Chem. Soc. 1957,
79, 6167. (f) Ramirez, F.; Levy, S. J. Org. Chem. 1958, 23, 2035.
(2) Brownie, J. H.; Baird, M. C.; Schmider, H. L. Organometallics
2007, 26, 1433.
(3) Brownie, J. H.; Baird, M. C.; Laws, D. R.; Geiger, W. E.
Organometallics 2007, 26, 5890.
(4) For a comprehensive review of phosphonium cyclopentadienides
and their coordination chemistry, see: (a) Brownie, J. H.; Baird, M. C.
Coord. Chem. Rev. 2008, 252, 1734. For more recent, related studies,
Synthesis of 1-C9H6PPh2MePPh2 (I). A mixture of 2.23 g of (1-
C9H7PPh2MePPh2)Br (4.46 mmol) and 0.12 g of NaH (5 mmol) in
40 mL of THF was stirred for 12 h, turning green and then brown.
The resulting mixture was filtered through Celite, and the solvent was
removed under reduced pressure to give a brown solid which was
redissolved in a minimum amount of CH2Cl2 and reprecipitated by
layering with hexanes (hexanes/CH2Cl2 4/1). The resulting
precipitate was filtered, washed with methanol, and dried to give
1.76 g (79% yield) of green, air-sensitive II. X-ray-quality crystals were
obtained by layering a saturated solution of II in CH2Cl2 with hexanes.
1H and 13C NMR data are given in Table 1. 31P NMR (CD2Cl2): δ
7.52 (d, JPP = 61 Hz, P+PH2CH2PPh2), −28.3 (d, P+Ph2CH2PPh2).
Synthesis of Isomers of [CpRu(1-C9H7PPh2MePPh2)]PF6. A
solution containing 0.01 g of [CpRu(MeCN)3]PF6 and 0.012 g of I in
see also (b) Serrano, E.; Navarro, R.; Soler, T.; Carbo,
Urriolabeitia, E. B. Inorg. Chem. 2009, 48, 6823. (c) Petrov, A. R.;
́ ́
J. J.; Lledos, A.;
Elfferding, M.; Mobus, J.; Harms, K.; Rufanov, K. A.; Sundermeyer, J.
̈
Eur. J. Inorg. Chem. 2010, 4157. (d) Xu, T.; Ye, J.; Gong, W.; Lin, Y.;
Ning, G. Acta Crystallogr., Sect. E: Struct. Rep. Online 2010, E66,
m1391. (e) Xu, T.; Gong, W.; Ye, J.; Lin, Y.; Ning, G. Organometallics
2010, 29, 6744. (f) Lichtenberg, C.; Hillesheim, N. S.; Elfferding, M.;
Oelkers, B.; Sundermeyer, J. Organometallics 2012, 31, 4259.
(5) Brownie, J. H.; Baird, M. C. J. Organomet. Chem. 2008, 693, 2812.
(6) (a) Crofts, P. C.; Williamson, M. P. J. Chem. Soc. 1967, 1094.
(b) Rufanov, K. A.; Ziemer, B.; Hummert, M.; Schutte, S. Eur. J. Inorg.
Chem. 2004, 4759.
0.6 mL of CD2Cl2 quickly turned red. Monitoring the H and 31P
1
NMR spectra over several days revealed two products, the apparent
kinetic product II and the thermodynamic product III. Initially the
ratio of II to III was very nearly 1:1 but, after 10 days, II was no longer
present and the amount of III had increased relative to the solvent
peak. H and 13C NMR data can be found in Table 1.
1
Computational Methods. Geometries of complexes II and III
and of simplified versions bearing H instead of Ph substituents at P
were optimized at the b3-lyp13/TZVP14 level using the Turbomole
program15 coupled to an external optimizer.16 For the chelate isomer,
several discrete local minima were located, differing in the indenyl
carbons bound to the Ru atom; only the lowest energy structure is
discussed in the text. Final geometries were checked by a vibrational
analysis (no imaginary frequencies), which was also used to calculate
thermal corrections (enthalpy and entropy). Single-point calculations
were then carried out with a larger basis set (TZVPP17) and inclusion
of an electronic correction for solvent effects using COSMO18 (ε =
9.1, CH2Cl2). These improved energies were combined with the above
thermal corrections (entropy scaled by 0.67 to account for reduced
freedom of motion in solution19) to obtain the final free energies
mentioned in the text.
(7) Fowler, K. G.; Littlefield, S. L.; Baird, M. C.; Budzelaar, P. H. M.
Organometallics 2011, 30, 6098.
(8) (a) Kuhl, O. Phosphorus-31 NMR Spectroscopy, A Concise
̈
Introduction for the Synthetic Organic and Organometallic Chemist;
Springer: Berlin, 2008. (b) Gorenstein, D. G. Phosphorus-31 NMR;
Academic Press: New York, 1984; Chapter 2. (c) Pregosin, P. S.;
Kunz, R. W. 31P and 13C NMR of Transition Metal Phosphine
Complexes; Springer-Verlag: New York, 1979. (d) Faller, J. W.; Friss,
T.; Parr, J. J. Organomet. Chem. 2010, 695, 2644. (e) Marchetti, F.;
Pettinari, C; Pizzabiocca, A.; Drozdov, A. A.; Troyanov, S. I.;
Zhuravlev, C. O.; Semenov, S. N.; Belousov, Y. A.; Timokhin, I. G.
Inorg. Chim. Acta 2010, 363, 4038.
(9) Trost, B. M.; Older, C. M. Organometallics 2002, 21, 2544.
(10) Garrou, P. E. Chem. Rev. 1981, 81, 229.
(11) (a) Basolo., F. Coord. Chem. Rev. 1982, 43, 7. (b) Rerek, M. E.;
Ji, L.-N.; Basolo, F J. Chem. Soc., Chem. Commun. 1983, 1208.
ASSOCIATED CONTENT
■
(c) Calhorda, M. J.; Romao, C. C.; Veiros, L. F. Chem. Eur. J. 2002, 8,
̃
S
* Supporting Information
4.
(12) Murphy, J. A.; Patterson, C. W. J. Chem. Soc., Perkin Trans. 1
1993, 405.
(13) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Becke, A. D.
J. Chem. Phys. 1993, 98, 1372. (c) Lee, C. T.; Yang, W. T.; Parr, R. G.
Phys. Rev. B 1988, 37, 785.
(14) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
5829.
(15) (a) Treutler, O.; Ahlrichs, R. J. Chem. Phys. 1995, 102, 346.
Tables, figures, and CIF files giving crystallographic details for
1-C9H6PPh2CH2PPh2 (I) and (1-C9H7dppm)Br, including
figures showing complete numbering schemes, thermal ellipsoid
figures, and tables of positional and thermal parameters, bond
lengths, and bond angles, and energy calculations for
compounds II and IIIc,d (Table S1) and of the xyz coordinates
used in the theoretical calculations (Table S2). This material is
available free of charge via the Internet at http://pubs.acs.org.
The crystallographic data for 1-C9H6PPh2CH2PPh2 (CCDC
892305) and (1-C9H7dppm)Br (CCDC 892303) may also be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
(b) Ahlrichs, R.; Bar, M.; Baron, H.-P.; Bauernschmitt, R.; Bocker, S.;
̈
̈
Ehrig, M.; Eichkorn, K.; Elliott, S.; Furche, F.; Haase, F.; Haser, M.;
̈
Hattig, C.; Horn, H.; Huber, C.; Huniar, U.; Kattannek, M.; Kohn, A.;
Kolmel, C.; Kollwitz, M.; May, K.; Ochsenfeld, C.; Ohm, H.; Schafer,
̈
̈
̈
̈
A.; Schneider, U.; Treutler, O.; Tsereteli, K.; Unterreiner, B.; Von
Arnim, M.; Weigend, F.; Weis, P.; Weiss, H.; Turbomole, version 5;
Theoretical Chemistry Group, University of Karlsruhe, Karlsruhe,
Germany, 2002.
(16) (a) Baker, J. J. Comput. Chem. 1986, 7, 385. (b) Baker, J. PQS,
version 2.4; Parallel Quantum Solutions, Fayetteville, AR, 2001.
(17) Weigend, F.; Furche, F.; Ahlrichs, R. J. Chem. Phys. 2003, 119,
12753.
(18) Klamt, A.; Schuurmann, G. J. Chem. Soc., Perkin Trans. 2 1993,
799.
(19) (a) Tobisch, S.; Ziegler, T. J. Am. Chem. Soc. 2004, 126, 9059.
(b) Raucoules, R.; de Bruin, T.; Raybaud, P.; Adamo, C. Organo-
metallics 2009, 28, 5358.
AUTHOR INFORMATION
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the Natural Sciences and Engineering Research
Council for financial support (Discovery Grants to M.C.B. and
P.H.M.B.), the Higher Education Commission of Pakistan for a
Fellowship to R.H., and Johnson-Matthey for a generous loan
of RuCl3·3H2O.
6932
dx.doi.org/10.1021/om300715a | Organometallics 2012, 31, 6926−6932