Donor-free phosphenium-metal(0)-halides with unsymmetrically bridging phosphenium ligands

Article abstract of DOI:10.1021/ic400886b

Reactions of (cod)MCl₂ (cod = 1,5 cyclooctadiene, M = Pd, Pt) with N-heterocyclic secondary phosphines or diphosphines produced complexes [(NHP)MCl]₂ (NHP = N-heterocyclic phosphenium). The Pd complex was also accessible from a chlorophosphine precursor and Pd₂(dba) ₃. Single-crystal X-ray diffraction studies established the presence of dinuclear complexes that contain μ-bridging NHP ligands in an unsymmetrical binding mode and display a surprising change in metal coordination geometry from distorted trigonal (M = Pd) to T-shaped (M = Pt). DFT calculations on model compounds reproduced these structural features for the Pt complex but predicted an unusual C_2v-symmetric molecular structure with two different metal coordination environments for the Pd species. The deviation between this structure and the actual centrosymmetric geometry is accounted for by the prediction of a flat energy hypersurface, which permits large distortions in the orientation of the NHP ligands at very low energetic cost. The DFT results and spectroscopic studies suggest that the title compounds should be described as phosphenium-metal(0)-halides rather than conventional phosphido complexes of divalent metal cations and indicate that the NHP ligands receive net charge donation from the metals but retain a distinct cationic character. The unsymmetric NHP binding mode is associated with an unequal distribution of σ-donor/π-acceptor contributions in the two M-P bonds. Preliminary studies indicate that reactions of the Pd complex with phosphine donors provide a viable source of ligand-stabilized, zerovalent metal atoms and metal(0)-halide fragments.

Full text of DOI:10.1021/ic400886b

Article
pubs.acs.org/IC
Donor-Free Phosphenium−Metal(0)−Halides with Unsymmetrically
Bridging Phosphenium Ligands
Daniela Forster,§ Jan Nickolaus,§ Martin Nieger,† Zoltan Benko,‡,⊥ Andreas W. Ehlers,∥
́
̋
̈
and Dietrich Gudat*,§
§Institut fur Anorganische Chemie, University of Stuttgart, Pfaffenwaldring 55, 70550 Stuttgart, Germany
̈
†Laboratory of Inorganic Chemistry, Department of Chemistry, University of Helsinki, P.O Box 55 (A.I. Virtasen aukio 1), 00014
Helsinki, Finland
‡Department of Inorganic and Analytical Chemistry, Budapest University of Technology and Economics, Szent Geller
́ ́
t ter 4, 1111
Budapest, Hungary
∥Division of Organic Chemistry, VU University Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands
S
* Supporting Information
ABSTRACT: Reactions of (cod)MCl2 (cod = 1,5 cyclo-
octadiene, M = Pd, Pt) with N-heterocyclic secondary
phosphines or diphosphines produced complexes [(NHP)-
MCl]2 (NHP = N-heterocyclic phosphenium). The Pd
complex was also accessible from a chlorophosphine precursor
and Pd2(dba)3. Single-crystal X-ray diffraction studies estab-
lished the presence of dinuclear complexes that contain μ-
bridging NHP ligands in an unsymmetrical binding mode and
display a surprising change in metal coordination geometry from distorted trigonal (M = Pd) to T-shaped (M = Pt). DFT
calculations on model compounds reproduced these structural features for the Pt complex but predicted an unusual C2v-
symmetric molecular structure with two different metal coordination environments for the Pd species. The deviation between
this structure and the actual centrosymmetric geometry is accounted for by the prediction of a flat energy hypersurface, which
permits large distortions in the orientation of the NHP ligands at very low energetic cost. The DFT results and spectroscopic
studies suggest that the title compounds should be described as phosphenium−metal(0)−halides rather than conventional
phosphido complexes of divalent metal cations and indicate that the NHP ligands receive net charge donation from the metals
but retain a distinct cationic character. The unsymmetric NHP binding mode is associated with an unequal distribution of σ-
donor/π-acceptor contributions in the two M−P bonds. Preliminary studies indicate that reactions of the Pd complex with
phosphine donors provide a viable source of ligand-stabilized, zerovalent metal atoms and metal(0)−halide fragments.
ligands”10). The L → M contribution may here be considered
INTRODUCTION
■
to be insignificant, and the interaction becomes essentially a
dative M → P bond.11 It has recently been pointed out that the
variable electron demand (and the associated structural
changes) of metal-bound phosphenium ions allows us to
establish an analogy to the “noninnocent” behavior of
nitrosyls.12
A unique example of electron-rich metal fragments are halide
complexes of zerovalent metal atoms of the type [LnM(0)X]− (L
= neutral ligand), which lack the strong electrostatic
stabilization by M(n+)···X− Coulomb interactions. Despite
their elusive nature, such species may be important reaction
intermediates; for example, Amatore and Jutand suggested that
anionic complexes [Pd(PR3)2X]− (PR3 = tertiary phosphine, X
= halide) are active participants in the widely applied Pd-
catalyzed C−C bond forming reactions.13−16 The successful
stabilization of metal(0)halides by phosphenium ligands has
Diaminophosphenium ions [(R2N)2P]+ are Lewis ambiphiles
that act as versatile π-acceptor ligands toward transition
metals.1−3 They resemble in this respect archetypes like CO
and NO+ yet have the advantage that their electronic and steric
properties are easily tunable.4 Computational studies suggest
description of the metal−ligand bonding in phosphenium
complexes with electron poor metal fragments as superposition
of a dominating dative M → L π-bond and a weaker dative L →
M σ-bond.2,3 This bonding situation is qualitatively similar to
that in metal carbonyls and nitrosyls or Fischer-type carbene
complexes,4−6 and the M−P bonds exhibit thus pronounced
double bond character. In rare cases, phosphenium ions were
also found to interact, like CO or NO+, with two metals as μ-
bridging ligands.7,8 Although the electronic structure of such
complexes has not been studied in detail, it seems reasonable to
assume a partition of L → M and M → L interactions between
the cationic ligand and both metal centers.
In complexes with electron-rich metals, phosphenium ions
Received: April 9, 2013
Published: June 18, 2013
can alternatively act as essentially pure electrophiles9 (“Z-type
© 2013 American Chemical Society
7699
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
analyses were carried out with the program AOMIX,21,22 and
MOLDEN23 was used for visualization. The NBO analysis was
performed with the NBO 5.0 program.24 Basis sets with electrostatic
potentials (ECP) on the inner shells (cc-pVDZ(-PP), cc-pVTZ(-PP),
aug-cc-pVTZ(-PP), def2-TZVP, def2-TZVPD, def2-TZVPPD) were
obtained from the basis set exchange home page.25,26 The topological
analysis of the electron density was performed using different
functionals (B3LYP, BP86), valence double- and triple-ζ ANO-RCC
basis on the Pd/Pt atoms, and cc-pVDZ and cc-pVTZ basis set on all
the other atoms. The contraction scheme applied for double-ζ all
electron basis function is 6s 5p 3d 1f for Pd and 7s 6p 4d 2f 1g for Pt.
The contraction scheme for triple-ζ all electron basis function is 7s 6p
4d 2f 1g for Pd and 8s 7p 5d 3f 2g for Pt. In order to take relativistic
effects into consideration, Douglas−Kroll−Hess second-order
(DKH2) scalar relativistic calculations27−29 have been performed.
The calculation of NMR properties was carried out using density
functional theory (DFT) as implemented in the Amsterdam Density
Functional (2009.01)30 code in which an all-electron, triple-ζ, double-
polarization TZ2P Slater basis was used including functions up to 5p
and 4f for Pd and 6p and 5f for Pt. Relativistic two-component zero-
order regular approximation (ZORA)31−34 calculations including
spin−orbit coupling35−38 have been performed with the local density
approximation (LDA) in the Vosko−Wilk−Nusair parametrization39
with nonlocal corrections for exchange (Becke88)40 and correlation
(Perdew86)41 included in a self-consistent manner. Isotropic absolute
magnetic shieldings for the phosphorus nuclei in 6″ and 7″ were
calculated using both “experimental” geometries (based on the
crystallographic coordinates of 6 and 7 with hydrogen atom positions
being readjusted using experimental bond and torsional angles and C−
H distances of 1.09 Å) and optimized (at the BP86/cc-pVTZ-(PP)
level) and converted to the common chemical shift scale using the
reference data by Jameson42 (δ(31P) = 328 − σ(31P)).
been realized in readily isolable complexes 1 and 2 (Scheme
1),17,18 and a similar description might also be invoked for 3
a
Scheme 1
a
1, bpy = 2,2′-bipyridine; 3, M = Pd, Pt, X = Cl, Br, I; 4, R = NiPr2,
NCy2.
and the trimeric complexes 4, which had originally described as
Pt(II)-phosphido complex12 or triangulo-clusters with bridging
phosphido ligands,19 respectively. The phosphenium metal(0)-
halides draw their stability, in simple terms, from the fact that
the back-donation to the phosphenium cation and additional π-
acceptor ligands decreases the electron density at the metal
atom, which in turn allows for a favorable interaction with the
strongly electron donating halide anion.15,18
In this Article, we report on dimeric complexes [(NHP)-
MCl]2 (NHP = N-heterocyclic phosphenium; M = Pd, Pt),
which consist, in a similar manner as 4, of an array of metal
atoms supported by μ2-PR2 units and terminal halides.
Computational and spectroscopic studies were carried out to
probe whether these products must be addressed as phosphido-
bridged complexes composed of anionic R2P-ligands and metal
dications or as “donor-free” phosphenium−metal(0)−halides
where the cationic ligand stabilizes an elusive metal(0)−halide
fragment in the absence of ancillary donor ligands. Finally, we
will present preliminary chemical reactivity studies that reveal
that the dimers serve as a viable source of mononuclear
metal(0) complexes.
Synthesis of Complex 6. From 5 and (cod)PdCl2. Diphosphine
543 (200 mg, 0.25 mmol) and 1 equiv of (cod)PdCl2 (70 mg, 0.25
mmol) were dissolved in anhydrous THF (5 mL) and stirred for 15
min at room temperature. A dark green microcrystalline precipitate
formed, which was filtered off, washed two times with diethyl ether,
and dried under reduced pressure. The crude product was purified by
recrystallization from CH2Cl2 at −20 °C. Yield 46%, mp >250 °C.
From 9 and (cod)PdCl2. Secondary phosphine 944 (200 mg, 0.49
mmol) and 1 equiv of (cod)PdCl2 (140 mg, 0.49 mmol) were
dissolved in anhydrous THF (5 mL) and stirred for 15 min at room
temperature. The formed dark green, microcrystalline precipitate was
filtered off, washed two times with diethyl ether, and dried under
reduced pressure. Yield 43%.
From 9 and (cod)PdCl2 in the Presence of NEt3. Secondary
phosphine 9 (15 mg, 37 μmol), 1 equiv of (cod)PdCl2 (10 mg, 37
μmol), and 1 equiv of NEt3 were dissolved in anhydrous d8-THF (0.7
mL). The solution was irradiated in an ultrasonic bath for
approximately 2 min. Quantitative formation of 6 and the absence
of chlorodiazaphospholene 8 as byproduct were established by 31P
NMR spectroscopy. No attempts toward isolation of the product were
1
made. 31P{1H} NMR (C6D6): δ = 225.1 (s). H NMR (C6D6): δ =
EXPERIMENTAL SECTION
■
3
3
7.30 (t, 4 H, JHH = 7.7 Hz, p-CH), 7.03 (d, 8 H, JHH = 7.7 Hz, m-
CH), 6.82 (pseudo-t, 4 H, 5JPH = 5.9 Hz, NCH), 3.03 (sept, 8 H, 3JHH
= 6.9 Hz, CH(CH3)2), 1.04 (d, 24 H, 3JHH = 6.7 Hz, CH3), 0.95 (d, 24
H, 3JHH = 6.7 Hz, CH3). 13C{1H} NMR (C6D6): δ = 147.2 (pseudo-t,
3JPC = 2.0 Hz, o-C), 132.3 (s, i-C), 129.7 (s, p-C), 129.0 (s, NCH),
124.3 (s, m-C), 28.9 (s, CH(CH3)2), 25.9 (s, CH3), 24.8 (s, CH3). (+)
ESI MS: m/e (%): 1059.35(100) [C52H72N4P2Pd2 − OCH3]+. UV−vis
(CH2Cl2): λmax/nm (εmax/(103 L mol−1 cm−1)): 600 (3.3), 515 (35.1),
410 (21.2). C52H72Cl2N4P2Pd2 (1098.85): calcd C 56.84, H 6.60, N
5.10; found C 56.24, H 6.68, N 5.06; the unsatisfactory low C value is
due to the inclusion of small, nonstoichiometric amounts of solvent
(CH2Cl2, verified by 1H NMR), which could not be completely
removed.
All manipulations were carried out under an atmosphere of dry argon
using standard vacuum line techniques. Solvents were dried by
standard procedures. NMR spectra were recorded on Bruker Avance
400 (1H, 400.1 MHz; 13C, 100.5 MHz; 31P, 161.9 MHz; 195Pt, 86.7
MHz) and Avance 250 (1H, 250.1 MHz; 13C, 62.8 MHz; 31P, 101.2
MHz) NMR spectrometers at 303 K; chemical shifts are referenced to
external TMS (1H, 13C), 85% H3PO4 (Ξ = 40.480 747 MHz, 31P), or
1.2 M Na2PtCl6 (Ξ = 21.496 784 MHz, 195Pt). Coupling constants are
given as absolute values. (+)-ESI mass spectra were recorded in
MeOH solution on a Bruker Daltonics MicroTOF Q instrument.
Elemental analyses were determined on a Perkin-Elmer 24000CHN/O
analyzer. Melting points were determined in sealed capillaries with a
Buchi B-545 melting point apparatus. UV−vis spectra were recorded
̈
on a Varian Cary 50 spectrometer.
Synthesis of Complex 7. A solution of diphosphine 5 (200 mg,
0.25 mmol) and 1 equiv of (cod)PtCl2 (92 mg, 0.25 mmol) in
anhydrous THF (5 mL) was stirred for 30 min at room temperature.
The solvent was removed in vacuum. The dark red residue was
All computational studies except the calculations of NMR
properties were performed with the Gaussian0920 suite of programs.
Fragment analysis of molecular orbitals and charge decomposition
7700
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
extracted two times with diethyl ether and dried under reduced
pressure. The crude product was purified by recrystallization from
CH2Cl2/Et2O (2:1) at −20 °C. Yield 41%, mp >250 °C. 31P{1H}
Table 1. X-ray Details for 6 and 7
6
7
1
NMR (C6D6): δ = 26.9 (s, JPPt = 5000 Hz). 195Pt{1H} NMR (C6D6,
empirical formula
C52H72Cl2N4P2Pd2
C52H72Cl2N4P2Pt2
1276.16
303 K): δ = −3078 (t, 1JPPt = 5000 Hz). 1H NMR (C6D6): δ = 7.32 (t,
4 H, 3JHH = 7.8 Hz, p-CH), 7.03 (d, 8 H, 3JHH = 7.8 Hz, m-CH), 6.03
fw (g mol−1
)
1098.78
100(2)
T (K)
123(2)
3
(pseudo-t, 4 H, 5JPH = 3.5 Hz, NCH), 3.44 (sept, 8 H, JHH = 6.9 Hz,
cryst size (mm3)
space group
a (Å)
0.83 × 0.57 × 0.35 0.08 × 0.04 × 0.02
3
CH(CH3)2), 0.98 (d, 48 H, JHH = 6.8 Hz, CH3). 13C{1H} NMR
P21/n (No. 14)
12.7928(3)
14.4161(4)
14.5891(4)
104.2790(10)
2607.43(12)
2
P21/n (No. 14)
12.916(2)
13.470(2)
14.903(2)
104.63(1)
2508.7(7)
2
(C6D6): δ = 148.1 (pseudo-t, 3JPC = 2.0 Hz, o-C), 133.4 (s, i-C), 129.3
(s, p-C), 124.7 (s, m-CH), 122.3 (s, NCH), 29.0 (s, CH(CH3)2), 25.9
(s, CH3), 24.8 (s, CH3). (+) ESI MS: m/e (%): 1235.47(100)
[C52H72N4P2Pt2 − OCH3]+. UV−vis (CH2Cl2): λmax/nm (εmax/(103 L
mol−1 cm−1)): 495 (2.7, sh), 425 (26.9), 375 (sh). C52H72Cl2N4P2Pt2
(1276.18): calcd C 48.94, H 5.69, N 4.39; found C 48.70, H 5.84, N
4.34.
b (Å)
c (Å)
β (deg)
V (Å3)
Z
Dc (Mg m−3
)
1.40
1.69
Reaction of 6 with PPh3. Complex 6 (100 mg, 90 μmol) and 4
equiv of PPh3 (95 mg, 0.36 mmol) were dissolved in CH2Cl2 (6 mL)
to give an orange solution. A 31P NMR spectrum recorded after 20
min showed broad signals attributable to 13 as major reaction product
besides traces of 8, Pd(PPh3)3, and a decomposition product, which
was later identified by NMR spectroscopy and single-crystal X-ray
diffraction as trans-(PPh3)2PdCl2 (12). The signals of 13 sharpened
into resolved multiplets at −20 °C. The amount of 8 and 12 increased
continuously when the reaction mixture was stored for longer time in
order to allow crystallization, up to the point that the initial product 13
disappeared completely, and 12 remained the only product that could
be isolated in pure form by crystallization. 31P NMR (CH2Cl2, 303 K):
δ = 222.7 (br; 13), 24.7 (br; 13). 31P NMR (CH2Cl2, 253 K): δ =
μ (mm−1
)
0.709
5.780
F(000)
1136
1264
θ range (deg)
reflns collected
unique reflns
3.2−30.5
28697
2.5−25.0
23540
7912
4405
Rint
0.017
0.114
max/min transm
data/restraints/params
GOF on F2
0.7426/0.525
7912/0/280
1.05
0.888/0.601
4405/0/280
1.04
R1 [I > 2σ(I)]
wR2 (F2)
0.021
0.061
0.052
0.159
2
2
222.1 (t, JPP = 129 Hz; 13), 24.3 (d, JPP = 129 Hz; 13).
largest diff. peak and hole (e
0.530/−0.450
2.892/−3.564
Å−3
)
Reaction of 6 with HPPh2. Complex 6 (50 mg, 46 μmol) and an
excess of diphenylphosphine were dissolved in CH2Cl2 to give a red
solution. A 31P NMR spectrum recorded after 15 min showed signals
attributable to 14 as main reaction product besides small amounts of 8
and Pd(HPPh2)x and excess diphenylphosphine. Storage of the
reaction mixture for crystallization was accompanied by eventual
decomposition of 14 to give 8 and Pd(HPPh2)x and a few crystals of
Pd(HPPh2)4, which were identified by a single-crystal X-ray diffraction
study, were obtained as the only isolable product. 31P NMR (CH2Cl2,
303 K): δ = 209.9 (br, 14), −16.3 (br; 14). 31P NMR (CH2Cl2, 253
a
Scheme 2
2
2
K): δ = 209.4 (q, JPP = 51 Hz; 14), −13.6 (d, JPP = 51 H; 14).
Crystallography. Crystal structures of 6 and 7 were determined
on a Nonius Kappa-CCD diffractometer at 100(2) K (6) or 123(2) K
(7) using Mo Kα radiation (λ = 0.71073 Å). Direct methods
(SHELXS-97) were used for structure solution. Refinement was
carried out using SHELXL-97 (full-matrix least-squares on F2), and
hydrogen atoms were refined using a riding model. A semiempirical
absorption correction was applied. Details of the crystal structure
determinations are listed in Table 1. Crystallographic data (excluding
structure factors) have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication nos. CCDC-905918
(6) and CCDC-905010 (7). Copies of the data can be obtained free of
charge on application to: The Director, CCDC, 12 Union Road, GB-
Cambridge CB2 1EZ (fax int. +1223/336033); E-mail: deposit@ccdc.
cam-ak.uk.
RESULTS AND DISCUSSION
■
Treatment of N-heterocyclic diphosphine 5 with (cod)MCl2
(cod = 1,5-cyclooctadiene; M = Pd, Pt) at ambient temperature
in THF produced dimeric complexes [(NHP)MCl]2 (6, M =
Pd; 7, M = Pt) and a stoichiometric amount of N-heterocyclic
chlorophosphine 8 (Scheme 2). A mixture of palladium
complex 6 and 8 was also obtained from reaction of
(cod)PdCl2 with secondary N-heterocyclic phosphine 9.
Formation of 8 as byproduct is avoided if the reaction is
carried out in the presence of an acid scavenger like NEt3, and 6
is accessible under these conditions with higher atom efficiency
from equimolar amounts of 9 and (cod)PdCl2 (Scheme 2). The
reaction of 9 with (cod)PtCl2 yielded spectroscopically
a
R = 2,6-iPr2C6H3.
detectable platinum hydride complexes whose isolation and
full characterization is under way and will be reported
elsewhere. Analysis of the 31P NMR spectra of reaction
mixtures revealed in all cases the absence of signals of further
phosphorus-containing products and suggests thus that all
reactions are quantitative. Finally, 6 is also formed, in analogy
to 4,19 as main reaction product upon treatment of
chlorophosphine 8 with an appropriate amount of Pd2(dba)3
(dba = dibenzylidene acetone).
7701
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
Complex 6 precipitated from the solution, whereas isolation
of 7 required removal of volatiles and extraction of byproducts.
Both complexes were obtained as deeply colored, EPR-silent
(and thus presumably diamagnetic), air-stable solids. Complex
6 is moderately soluble in CH2Cl2, sparingly soluble in
CH3CN, and insoluble in hydrocarbon and ether solvents
(Et2O, THF, DME); 7 showed better solubility in all solvents
except Et2O and hydrocarbons. The dimeric nature of both
products was inferred from ESI-mass spectra and the
1
pseudotriplet splittings of the H NMR signals of the protons
in the N-heterocyclic rings. In connection with the chemical
equivalence of the two phosphenium moieties, these features
suggest that the metal atoms are symmetrically bridged by two
NHP units to form a cyclic P2M2 array, and each metal atom
carries an extra terminal chloride (an inverse arrangement with
terminal NHP and bridging chloride ligands should exhibit a
very small, if not negligible, JPP coupling constant and a
1
different splitting pattern for the H NMR signal of the ring
protons). The 31P chemical shift of 225.1 ppm for 6 is
19
substantially lower than that for [{μ-(R2N)2P}PdCl]3 (4, δ31P
= 346.1 for R = iPr, 351.1 for R = Cy), but the difference is of
similar magnitude as in case of the free phosphenium ions (cf.,
1
δ31P = 313 ppm for (iPr2N)2P+ vs 205 ppm for 1,3-bis-
C6H3iPr2-diazaphospholenium45). The much lower chemical
shift of 26.9 ppm for 7 introduced initial doubts about the
structural similarity with the Pd complex, but the presence of an
analogous M2P2 ring was established from analysis of 31P,195Pt
spin couplings: the 195Pt NMR signal is split into a triplet (1JPt,P
= 5000 Hz) indicating attachment of the metal atom to two
equivalent P-donor ligands, and the 31P NMR signal is
accompanied by two sets of 195Pt satellites due to coupling
(1JPt,P = 5000 Hz) of the phosphorus atoms to one (P2(195Pt)1
isotopomer, A2X subspectrum) or two chemically equivalent
metal spins (P2(195Pt)2 isotopomer, A2X2 subspectrum),
respectively.
The structural assignment was confirmed by single-crystal X-
ray diffraction studies. Both complexes crystallize in the
monoclinic space group P21/n and contain isolated molecules
with crystallographic Ci-symmetry. Complex 6 (Figure 1)
exhibits a planar P2Pd2 ring in which adjacent bonds differ
slightly in length (P1−Pd1 2.2738(4) vs P1−Pd1#1 2.2401(4)
Å) but opposite bonds remain equal. The Pd−P−Pd angles are
much more acute (71.61(1)°) than the P−Pd−P angles
(108.39(1)°). The Pd−P bonds are naturally longer than
those to the terminal phosphenium ligands in 218 (2.117(2) Å)
or [(NHP)Pd(PPh3)2]OTf18 (Pd−P 2.1229(11) Å, NHP =
1,3-dimesityl-1,3,2-diazaphospholenium) and lie between the
two distinctly different P···Pd distances in 1012 (Scheme 3) of
2.162(2) and 2.498(2) Å. The planar diazaphospholene rings in
6 are perpendicular to the P2Pd2 ring (dihedral angles 89.1°
between P2Pd2 and PN2C2 planes) and mutually parallel
(dihedral angle 0° between both PN2C2 planes). The angles
between the P1−Pd1/P1−Pd1#1 bonds and the bisecting NHP
ring plane are 50° and 22°, and the diazaphospholene units
adopt thus a distinctly asymmetrical coordination (Figure 1b).
The chlorine ligands lie in the plane of the P2Pd2 ring, but the
Cl−Pd−P angles show likewise a marked dissymmetry (Cl1−
Pd1−P1#1 134.35(2)° vs Cl1−Pd1−P1 117.27(2)°). Taking
all disparities in Pd−P distances, P−Pd−Cl angles, and the
alignment of diazaphospholene rings together, the observed
molecular structure can be conceived to picture an early stage
in the dissociation of a hypothetical dinuclear complex with
ideal D2h symmetry into two monomeric Cl−Pd−NHP units
Figure 1. Representation of the molecular structure of 6 (top) and the
Pd2P2(NHP)2 core (bottom) in the crystal. H atoms were omitted for
clarity; thermal ellipsoids were drawn at the 50% probability level.
Selected bond lengths (Å) and angles (deg): Pd1−P1 2.2738(4),
Pd1−P1#1 2.2401(4), Pd1−Cl1 2.3361(4), Pd1−Pd1#1 2.6410(2),
P1−N2 1.680(1), P1−N5 1.685(1), N2−C3 1.384(2), C3−C4
1.349(2), C4−N5 1.384(2), P1#1−Pd1−P1 108.39(1), Pd1#1−P1−
Pd1 71.61(1), P1−Pd1−Cl1 117.27(2), P1#1−Pd1−Cl1 134.35(2),
Cl1−Pd1−Pd1#1 170.87(1), N2−P1−N5 89.55(5).
Scheme 3
featuring linear dicoordinate metal and planar tricoordinate
phosphorus atoms.
The coordination modes of chloride and phosphenium
ligands in 6 are qualitatively similar to those in the triangulo-
cluster 4,19 whose structural disorder precludes a more
quantitative comparison; the different degree of aggregation
in both compounds is presumably due to different steric
requirements of the phosphenium moieties. The unsymmetrical
7702
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
μ-bridging coordination of the aminophosphenium ligands
resembles the semibridging coordination mode of carbonyls
and has precedence in dipalladium and dicobalt complexes 1012
and 117 (Scheme 3) where the geometrical distortion may be
facilitated by different metal coordination environments (11) or
the constraints imposed by the formation of chelate rings (10).
A further interesting feature of 6 is the presence of a short
metal−metal distance (2.6410(2) Å), which is only slightly
larger than the distance of 2.58 Å calculated from covalent
radii46 and falls well into the range of Pd−Pd bonds in dimeric
Pd(I) complexes (2.72 0.1 Å47). The P−N, C−N, and C−C
bond lengths in the PN2C2 rings match those of known chloro-
diazaphospholenes and NHPs45 and indicate that 1 shows, like
these compounds, a certain degree of π-electron delocalization
in the N-heterocyclic rings.
The molecular structure of platinum complex 7 (Figure 2)
displays the same topology as 6 but exhibits some remarkable
deviations in the coordination environment of the phosphorus
and platinum atoms. The shape of the central ring is conserved,
but the alternation of adjacent bonds (Pt1−P1 2.239(3), P1−
Pt1#1 2.179(3) Å) is now even more pronounced. The
bridging NHP units show a comparable tilt as in 6 (angles of
the P1−Pt1 and P1−Pt1#1 vectors with the local PN2 plane
containing the atoms P1, N2, and N5 are 57° and 42°,
respectively), but the difference in Cl−Pt−P bond angles is
much larger than in 6 (P1−Pt1−Cl1 107.3(1)°, P1#1−Pt1−
Cl1 171.7(1)°). The metal coordination geometry remains
planar but is best described as T-shaped (with the nearly
collinear P1#1−Pt1 and Pt1−Cl1 bonds representing the bar of
the T) rather than distorted trigonal. As a consequence of this
distortion, endocyclic P−Pt−P angles are much smaller
(81.0(1)°) and Pt−P−Pt angles larger (99.0(1)°) than the
corresponding angles in 6 (P−Pd−P 108.39(1)°, Pd−P−Pd
71.61(1)°), and the Pt···Pt distance across the ring increases to
3.360(1) Å, which rules out any direct bonding interaction. The
bonds in the diazaphospholene ring of 7 are generally slightly
shorter than those in 6 and closer to appropriate bond lengths
in phosphenium ions.45 In contrast to the free cations, the
PN2C2 ring in 7 is not planar but displays a distinct envelope
conformation. The Pt1−Cl1 bonds (2.268(3) Å) are slightly
shorter than normal Pt−Cl distances (2.33 0.04 Å48).
The observed molecular structures of 6 and 7 allow for a
similar ambiguity in the interpretation of metal−ligand
interactions as had been discussed for other phosphenium
complexes.12 In principle, two limiting descriptions can be
given: (1) the NHP units act as anionic phosphido ligands that
bind as μ-bridging four-electron σ/π-donors to two metal(II)
centers, and (2) the NHP units represent phosphenium cations
and bind as σ-donor/π-acceptor ligands to two zerovalent metal
atoms. Alternatively, complex 6 may, in view of the short
intermetallic distance, also be conceived to contain two σ-
bonded Pd(I) centers that bind two chlorides and two neutral
diazaphospholenyl radicals (which are then antiferromagneti-
cally coupled to give a diamagnetic product). The depiction as
metal(0) complexes with formally cationic NHP ligands
adheres more closely to the common perception of N-
heterocyclic phosphenium ions as paramount electrophiles,1,3
but recent reports on the noninnocence of some chelating N-
heterocyclic phospheniums12 and the proven generation of
neutral diazaphospholenyl radicals43,49 implies that the
alternative descriptions may not be a priori ruled out.
In order to shed further light on the electronic structure of
complexes 6 and 7, we performed DFT calculations20 on
suitable model compounds. Anticipating that the peripheral
isopropyl groups in 6 and 7 should have negligible impact on
the electronic structure of the P2M2 core, initial computations
were carried out on N-phenyl-substituted model compounds 6′
and 7′. These studies revealed that the molecular core interacts
only weakly with the substituents at nitrogen and that the
computed molecular geometries and electron populations are
quite similar to those of the still simpler NH-substituted species
6″ and 7″. We therefore chose these compounds, which are
much easier to handle computationally, as models for a more
comprehensive electronic structure investigation.
In view of the marked structural deviation between 6 and 7,
we searched first for possible stationary points on the potential
energy hypersurfaces of model complexes 6″ and 7″. To this
end, we performed energy optimizations in different restricted
symmetries (C2v, C2h, D2h) and tested several combinations of
density functionals (BP86, B3LYP, M06-2X, PBE1PBE) and
basis sets (cc-pVDZ(-PP), aug-cc-pVDZ(-PP), cc-pVTZ(-PP),
def2-TZVP, def2-TZVPD, def2-TZVPPD) to establish an
appropriate computational level. The results of these
calculations are all qualitatively similar even if relative energies
Figure 2. Representation of the molecular structure of 7 (top) and the
Pt2P2(NHP)2 core (bottom) in the crystal. H atoms were omitted for
clarity; thermal ellipsoids were drawn at the 50% probability level.
Selected bond lengths (Å) and angles (deg): Cl1−Pt1 2.268(3), Pt1−
P1 2.239(3), Pt1−P1#1 2.179(3), Pt1−Pt1#1 3.360(1), P1−N5
1.654(10), P1−N2 1.668(11), N2−C3 1.375(15), C3−C4 1.342(17),
C4−N5 1.373(16), P1#1−Pt1−P1 81.0(1), P1#1−Pt1−Cl1 171.7(1),
P1−Pt1−Cl1 107.3(1), N5−P1−N2 91.2(5), Pt1#1−P1−Pt1 99.0(1).
7703
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
of isomeric structures may change with the functional applied
(a complete listing is given as Supporting Information), and we
will only discuss the results obtained at the BP86/cc-pVTZ-
(PP) level. The geometries, relative energies, and numbers of
imaginary vibrational modes for stationary states of 6″ and 7″
located under different symmetry constraints (C2v, C2h, D2h) at
this level are displayed in Figure 3.
For the platinum complex 7″, three stationary states were
found: the structures with C2h and C2v symmetry represent local
minima, and the one with D2h symmetry represents a transition
state. The C2h-isomer is more stable by some 4 kcal mol−1 than
the C2v-isomer, and the energy difference between the C2v-
isomer and the D2h-transition state is also larger than in the case
of 6″. The computed molecular structure of the stable C2h-
isomer of 7″ is in reasonable agreement with the found
structure of solid 7 (selected distances are given in Table 1 for
7″ and in Figure 2 for 7). In addition, the low energy of the
D2h-transition state implies that dynamic inversion of the
bridging NHP units is easily feasible and serves thus to explain
the A2X2-type coupling patterns in 31P and 195Pt solution NMR
spectra (which are at odds with a static C2h-structure). The
perceptible energy difference between the C2h- and C2v-isomers
of 7″ suggests further that the special metal coordination
geometry in 7 is not imposed by crystal packing effects. A
similar supremacy of T-shaped over regular trigonal metal
coordination had also been predicted for mercury halide dimers
(HgX2)2 and was convincingly explained as arising from
relativistic contraction (and concomitant energetic stabiliza-
tion) of the mercury 6s-orbital.51 Since this effect becomes even
more important for the 6s-orbital at Pt but not for the 5s-orbital
at Pd,52 we conclude that it may also explain the deviating metal
coordination geometries in 6″ and 7″. This view is supported
by the finding that the calculated natural electron config-
urations (see below) reveal that the contribution of the valence
s-orbital of the metal atoms is much higher for Pt than for Pd.
The consideration of relativistic effects helps also in the
explanation of the remarkably large difference in 31P NMR
chemical shifts of 6 and 7. Relativistic two-component zero-
order regular approximation (ZORA)31−34 calculations with
inclusion of spin−orbit coupling35−38 on the model compounds
6″ and 7″ produced values of δ31P (232.8 ppm for 6″ and 37.4
ppm for 7″ based on experimental geometries; cf. Table 3) that
closely match the experimental data of 6 (δ31P = 225.1 ppm)
and 7 (δ31P = 26.9 ppm), respectively. Analysis of individual
contributions reveals that the stronger shielding of the 31P
nuclei in the Pt complex is due to two major effects, namely, a
reduction in the (nonrelativistic) paramagnetic shielding
contribution (Δσpara = 86.68 ppm) and a shift from a small
negative to a large positive value of the (relativistic) spin−orbit
contribution (ΔσSO = 106.94 ppm). Comparison of shieldings
calculated with different sets of atomic coordinates (derived
from the crystal structure data of 6 and 7 and the energy-
optimized molecular geometries of 6″ and 7″) indicates that
the numerical values of calculated shieldings are, not
unexpectedly, strongly sensitive to variations in the molecular
Figure 3. Geometries, relative energies (kcal mol−1), and numbers of
imaginary vibrational modes (in parentheses) of the isomeric
structures of complexes 6″ and 7″ calculated under different symmetry
constraints at the BP86/cc-pVTZ-(PP) level.
In the case of the palladium complex 6″, the C2v-structure
was identified as the only local minimum. The D2h-structure is a
transition state at negligibly higher energy (0.1 kcal mol−1), and
optimization attempts in C2h-symmetry (or still lower C2- and
Ci-symmetry) converged always to the D2h-structure, even with
small optimization steps. The computed bond distances (Table
2) agree reasonably with the experimental ones, but the overall
molecular structure of 6″ differs subtly from the observed
structure of 6: the pairs of shorter and longer Pd−P bonds
occupy adjacent rather than opposite edges of the P2Pd2
quadrangle, and the tips of the PN2C2 pentagons of both
phosphenium ligands point to the same rather than different
metal atoms.50 The negligible energy differences between the
C2v- and D2h-isomers as well as the energy profiles produced
during the course of the geometry optimization imply that the
molecular structure of 6″ tolerates large distortions in the
alignment of the phosphenium ligands with hardly any
energetic cost. We conclude that a reliable prediction of the
molecular structure is therefore beyond the accuracy of the
computational model and that the observed geometry of 6 is
presumably enforced by the large steric demand of the N-aryl
substituents or intermolecular (“crystal packing”) forces.
Table 2. Selected Bond Lengths (d, in Å), Wiberg Bond Indices (WBI), and Mayer Bond Indices (MBI) for 6″ (C2v), 7″ (C2h),
a
+
and the Free Diazaphospholenium Cation (PN2C2H4 ) at the BP86/cc-pVTZ(-PP) Level
+
6″ (C2v)
7″ (C2h)
PN2C2H4
M−M
M−Cl
2.696 (0.307/0.252)
2.317 (0.723/0.823)
2.298 (0.726/0.839)
2.244 (0.735/0.897)
2.296 (0.654/0.847)
1.723 (0.887/1.097)
1.384 (1.172/1.181)
1.361 (1.631/1.595)
3.442 (0.146/0.148)
2.294 (0.748/0.949)
M−P
2.214 (0.821/1.250)
2.300 (0.692/0.905)
1.718 (0.846/1.076)
1.393 (1.137/1.141)
1.356 (1.673/1.627)
P−N
N−C
C−C
1.695 (1.077/1.291)
1.369 (1.246/1.234)
1.374 (1.536/1.501)
a
Entries are given in the form d (WBI/MBI).
7704
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
Table 3. Computed 31P NMR Chemical Shifts (δ(31P)) and
Diamagnetic (σdia), Paramagnetic (σpara), and Spin-Orbit
(σSO) Contributions to the Absolute Isotropic Shielding
array contains two types of dative bonds of opposite polarity.
The situation in 6″ is less easily interpreted since the NBO
results imply a delocalized structure, which must obviously be
depicted in terms of resonance between several canonical
formulas. Still, the leading Lewis structure contains an array of
two chloride ions, a Pd(0) atom, and a dicationic fragment
[Pd(NHP)2]2+, which is itself composed of two phosphenium
ions binding to a zerovalent metal atom, and may thus likewise
be considered to support a description as phosphenium−
metal(0)−halide. The unsymmetrical coordination of the NHP
fragments in both 6″ and 7″ contrasts further with the
symmetrical arrangement of bridging phosphides.
(iii) The decrease in P−N and N−C bond orders and
concomitant increase of C−C bond orders upon complexation
(see Table 2) suggest that the transfer of extra electron density
to the NHP cations (as revealed by the lower partial charge for
the bound ligand) raises the population of the phosphenium π4-
orbital (the LUMO of the free cation, which is PN and CN
antibonding and CC bonding in character), which emphasizes
the π-acceptor property of the ligand.
a
(σiso) of 6″ and 7″
6″
7″
σdia
σpara
σSO
σiso
δ(31P)
960.51 (959.36)
−853.41 (−888.43)
−11.85 (−17.23)
95.25 (53.70)
962.28 (961.14)
−766.73 (−828.60)
95.09 (57.70)
290.64 (190.24)
37.36 (137.76)
232.75 (274.30)
a
Data are based on the experimental geometries of crystalline 6 and 7;
values derived from calculated geometries of 6″ and 7″ are given in
parentheses.
geometries, whereas the relative magnitudes of the relativistic
and nonrelativistic contributions to the total magnetic shielding
variation remain relatively unaffected.
To get further insight into the bonding situation, we
performed natural bonding orbital53 (NBO) calculations on
the lowest energy structures of 6″ and 7″. Analysis of bond
lengths, Wiberg bond indices (WBI) and Mayer bond indices
(MBI) of 6″ and 7″ (Table 1) revealed that the M−Cl and M−
P bonds exhibit sizable covalent contributions, which are
generally slightly higher for the Pt than the Pd complex;
furthermore, the shorter M−P bonds in each species exhibit
slightly larger covalent contributions than the longer ones. The
P−N and N−C bonds in the NHP ligands are somewhat longer
(weaker) and the C−C bonds shorter (stronger) than in the
free phosphenium ion.
Even if the participation of each phosphorus atom in the
formation of two M−P bonds does not a priori rule out a
description of 6″ and 7″as phosphido complexes (with anionic
σ-donor/π-donor ligands that bind to two M2+ cations), the
alternative picture of phosphenium−metal(0)-species with
cationic NHP ligands and metal oxidation numbers close to
zero seems more appropriate for several reasons:
(iv) A charge decomposition analysis54,55 (CDA) indicates
that M2Cl2 → NHP electron donation is larger in 6″ (M = Pd)
by 0.25 electrons and in 7″ (M = Pt) by 0.50 electrons than
electron donation in the other direction. The observed figures
are in line with the larger M−P bond indexes in 7″ compared
with 6″ (Table 2) and further accentuate that the NHP
fragments in both complexes retain substantial phosphenium
(σ-donor/π-acceptor) character.
The rather short Pd1−Pd1#1 distance in 6 suggests the
possibility of direct Pd···Pd bonding. Analysis of the Wiberg
and Mayer bond indexes reveal that the appropriate values in
6″ are indeed larger than those in 7″, and an AIM (atoms in
molecules56,57) analysis of the electron density of 6″ allowed us
to locate, in addition to the bond critical points representing
the cyclic array of M−P bonds, a further bond critical point
between the palladium atoms (Figure 4). In contrast, no direct
bond path between the metal atoms was found in the M2P2
region of 7″ (Figure 5).
(i) Analysis of NPA (natural population analysis) charges
(Table 4) yields negative partial charges for the metal atoms
Even if these topological differences seem to support the idea
of direct Pd···Pd bonding in 6″, some critical issues remain: the
Pd−Pd bond order (Table 2) is still far from that expected for a
full bond, the position of the Pd···Pd bond critical point in 6″ is
close to the adjacent ring critical points, and the electron
density at this point (ρ = 0.050 au) is only marginally higher
than that at the P2Pt2 ring critical point of 7″ (0.037 au), which
represents the electron density minimum in this region. All
features together imply that the valence charge concentration
on the Pd···Pd “bond path” of 6″ is rather petty, and disfavor
thus a description of the complex as containing two genuinely
σ-bonded Pd(I) centers. We cannot rule out that the additional
bond critical point in 6″ reflects simply the effect of trailing
electron density from two metal atoms that are coerced into
close positions and conclude that neither population nor
topological analysis allows a clear-cut assessment of the metal−
metal interaction. Two further bond critical points were
identified between the CC bonds in the NHP units and
the chlorine atoms of 7″ but are in view of the still lower
electron density (ρ = 0.005 au) likewise attributed to represent
trailing electron density rather than a genuine bonding
interaction.
Table 4. NPA Partial Charges (au) of the Metal Centers (M
= Pd, Pt), Chlorine Atoms, and NHP Units of 6″ (C2v) and
7″ (C2h) at the BP86/cc-pVTZ(-PP) level
6″ (C2v)
7″ (C2h)
M (Pd or Pt)
Cl
−0.32/−0.40
−0.44/−0.43
+0.79
−0.35
−0.38
+0.73
PC2N2H4
and substantial positive partial charges for the diazaphospho-
lene units in both complexes. Likewise, the calculated natural
electron configurations of the metal atoms (6″ Pd1 4d9.31 5s0.51
p0.51, Pd2 4d9.32 5s0.48 p0.59; 7″ Pt 5d9.14 6s0.85 p0.35) agree better
with the picture of zerovalent d10-metal centers, which form
dative bonds via ligand-to-metal σ-donation and metal-to-ligand
π-back-donation, than with the perception of M2+(d8) cations,
which act predominantly as electron acceptors (the description
is also supported by the Mulliken charges, see Supporting
Information).
(ii) In the case of 7″, the ligand contributions dominate in
the NBOs representing the “short” M−P bonds (25.1% Pt vs
74.9% P), and the metal contributions dominate in the “long”
M−P bonds (60.1% Pt vs 39.9% P), suggesting that the M2P2
In order to study the metal−ligand interaction from a
different angle, we recorded UV−vis spectra of 6 and 7. The
visible region of the spectra contains three bands (λmax = 600,
7705
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
Figure 4. AIM analysis of the electron density (calculated at the BP86/cc-pVTZ (light atoms), triple-ζ-ANO-RCC(Pd)//BP86/cc-pVTZ(-PP)
level) of 6″. The lines connecting atoms represent bond paths, and the red and yellow points represent bond and ring critical points, respectively.
Figure 5. AIM analysis of the electron density (calculated at the BP86/cc-pVTZ(light atoms), triple-ζ-ANO-RCC(Pt)//BP86/cc-pVTZ(-PP) level)
of 7″. The lines connecting atoms represent bond paths, and the red, yellow, and green points represent bond, ring, and cage critical points,
respectively.
515, and 410 nm) for 6 and two bands (λmax = 495 and 425
nm) and a shoulder at shorter wavelengths for 7. The first two
absorption bands in both spectra give similar intensity patterns,
but the absorption maxima of 7 are red-shifted by 90−100 nm.
These features were reproduced in simulated electronic spectra
of 6″ and 7″, which had been generated from the results of
TDDFT calculations (see Supporting Information; the weaker
coincidence between observed and calculated transition
energies for 6/6″ is presumably attributable to the larger
deviation between computed and experimentally observed
molecular structures). The TDDFT computations indicate that
a single electronic transition contributes to the first and two
transitions to the second spectral bands and that each transition
involves a combination of several one-electron excitations.
Analysis of the contributions from metal atoms, chlorides, and
NHP ligands to the MOs involved in the individual one-
electron excitations implies that all three electronic transitions
of 6″ and the third lowest excitation of 7″ have perceptible
MLCT character. The first electronic transition in 7″ is not
associated with large shifts in electron population between
fragments, and the second transition exhibits some LMCT
character. In summary, these results can be considered as
further manifestation for the π-acceptor character of the
phosphenium ligands in 6 and 7.
Finally, the view of complex 6 as phosphenium−metal(0)−
halide with a formally zerovalent metal complex is supported by
the outcome of preliminary reactivity studies. Reaction of 6
with 4 equiv of Ph3P produces a darkly colored solution that
contains according to a 31P NMR survey a major reaction
product together with minor amounts of chlorophosphine 8,
Pd(PPh3)n, and cis-(Ph3P)2PdCl2 (12). The signals of the main
product appear at −20 °C as resolved multiplets (AX2 spin
system with δ = 222.1 (PA), 24.3 (PX); JAX = 129 Hz) and
broaden at ambient temperature. The spectral parameters
match those of Pd(0) bromide 2 (δ31P 23.4 and 213.918) and
the ionic Pd(0) triflate, [(N2PMes)(PPh3)2Pd]OTf (δ31P 23.0
and 213.0, JPP = 138 Hz18), and the product is on this basis
assigned as monomeric Pd(0) halide 13 (Scheme 4). The
temperature dependent line broadening had previously been
observed for 2 and was explained as resulting from reversible
halide dissociation and formation of a dynamic equilibrium
between neutral 2 and a cation [(N2PMes)(PPh3)2Pd]+.18
7706
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
Furthermore, the preparation of 6 and 7 from diphosphine 5
emphasizes the nature of 5 as an unusual reducing agent for
metal cations and stimulates further studies to widen its
application.
Scheme 4
Several aspects deserve further consideration: (i) Com-
pounds 6 and 7 extend the limited number of complexes with
unsymmetrically μ-bridging phosphenium units. Computational
studies indicate that this special binding mode, which resembles
the semibridging coordination of carbonyls,7 correlates with an
uneven distribution of σ-donor and π-acceptor interactions in
the P−M bonds and evokes a description of the dinuclear units
as dimers of two [ClM(NHP)] monomers. This view is in
accord with both the structural distortions and the easy
cleavage of the dinuclear unit by donor ligands. This reactivity
opens interesting possibilities to use 6 and 7 (and possibly also
trinuclear species like 4) as synthons for coordinatively and
electronically unsaturated metal centers and should stimulate
studies of their application as stoichiometric or catalytic
reagents. (ii) The computational studies give evidence of a
flat potential energy hypersurface, which allows large distortions
in the alignment of the phosphenium ligands in 6 and 7 at very
low energetic cost (<1 kcal mol−1). Apart from providing a
viable explanation for the apparently higher symmetry of the
molecular structure in solution and the astonishing deviation of
the computed molecular structures of model compound 6′
from that of solid 6, a structural dichotomy between structures
with like and different metal coordination environments may
provide an entry point to a rich chemistry, which stimulates
further studies on these and analogous complexes. (iii) In view
of the fact that, regardless of all structural differences, the
electronic structures of the Pd and Pt complexes 6″ and 7″ bear
great similarities, the large difference in 31P NMR chemical
shifts cannot be considered to reflect simply the differences in
charge transfer, as had been suggested,12 and further efforts to
obtain a satisfactory explanation of observed trends in magnetic
shielding seem necessary.
Complex 13 decomposed eventually to give 8 and 12, which
remained the only species that could be isolated in pure form.
The decomposition can be envisaged to proceed via “non-
reductive” elimination to form 8 and Pd(PPh3)n, which then
reacts with the solvent (CH2Cl2) to produce 12. This scheme
allows us also to explain the formation of the byproducts
observed in the initial reaction mixture. In a similar manner, 6
reacted with excess HPPh2 to produce a solution containing
complex 14 (31P NMR, AX3 spin system with δ = 209.4 (PA),
−13.6 (PX); JAX = 51 Hz; Scheme 4) and traces of 8 and
Pd(HPPh2)n together with unreacted phosphine. Storage of the
reaction mixtures at low temperature resulted in isolation of a
few crystals of [Pd(HPPh2)4],58 which remained the only
product to be isolated.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental and simulated UV−vis spectra of 6/6″ and 7/7″,
electronic transition energies for 6″ and 7″ from TDDFT
calculations at BP86/cc-pVTZ(-PP) level, Mulliken partial
charges (au) of the metal centers (M = Pd, Pt), chlorine atoms,
and NHP units of 6″ and 7″ at the BP86/cc-pVTZ(-PP) level,
relative energies and number of imaginary frequencies for the
different structures of 6″ and 7″ at different levels of theory,
and total energies and coordinates of 6″ and 7″ at different
levels of theory. This material is available free of charge via the
Internet at http://pubs.acs.org.
CONCLUSION
■
Dinuclear complexes [(NHP)MCl]2 (M = Pd (6), Pt (7); NHP
= N-heterocyclic phosphenium fragment) were prepared by
several routes. The products feature terminal chloride ligands
and bridging NHP units and are thus structural analogues of
previously reported trinuclear clusters 4,19 which were, mainly
because of formal reasons, denoted as phosphido complexes. In
contrast, evidence from computational, spectroscopic, and
chemical studies suggests that 6 and 7 should be correctly
addressed as genuine phosphenium−metal(0)−halides with
metal oxidation numbers close to zero and ligands that receive
net back-donation from the metal atoms but retain their nature
as cationic phosphenium ligand. This description is in line with
the recently postulated “noninnocent” behavior of amino-
phosphenium ligands12 and allows us to conclude that formal
clustering of individual electronically unsaturated {ClM-
(NHP)} units renders the metal atoms less electron-rich than
the addition of extra donor ligands. Adopting this description of
phosphenium−metal(0) complexes, the reaction of Pd2(dba)3
with an N-heterocyclic chlorophosphine to give 6 must be
considered another example of a “nonoxidative” addition59 of a
phosphenium−Lewis base adduct to a metal complex.
AUTHOR INFORMATION
Corresponding Author
*Fax: +4971168564241. E-mail: gudat@iac.uni-stuttgart.de.
■
Present Address
⊥ZB: Department of Chemistry and Applied Biosciences, ETH-
Zurich, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland.
̈
̈
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank the Academy of Finland for a Research Fellowship
and the Deutsche Forschungsgemeinschaft (DFG, Grant Gu
■
7707
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708

Inorganic Chemistry
Article
(25) Feller, D. J. J. Comput. Chem. 1996, 17, 1571.
415/11-2) and Deutscher Akademischer Austauschdienst
(DAAD) for financial support. We further thank J. Trinkner
for measurement of mass spectra, Dr. W. Frey (both from
Institute of Organic Chemistry, University of Stuttgart) for the
(26) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L.;
Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. J. Chem. Inf. Model.
2007, 47, 1045. (https://bse.pnl.gov/bse/portal)
(27) Douglas, M.; Kroll, N. M. Ann. Phys. 1974, 82, 89.
(28) Hess, B. A. Phys. Rev. A 1985, 32, 756.
́ ́
collection of X-ray data sets, and Prof. Laszlo Nyulaszi
(Budapest University of Technology and Economics) for
helpful discussions. The bw-grid project60 is acknowledged for
supplying computational resources.
(29) Hess, B. A. Phys. Rev. A 1986, 33, 3742.
(30) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca
Guerra, C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J.
Comput. Chem. 2001, 22, 931. ADF 2002.01: http://www.scm.com.
(31) Schreckenbach, G.; Ziegler, T. J. Phys. Chem. 1995, 99, 606.
(32) Schreckenbach, G.; Ziegler, T. Int. J. Quantum Chem. 1997, 61,
899.
(33) Wolff, S. K.; Ziegler, T. J. Chem. Phys. 1998, 109, 895.
(34) Wolff, S. K.; Ziegler, T.; van Lenthe, E.; Baerends, E. J. J. Chem.
Phys. 1999, 110, 7689.
(35) Dickson, R. M.; Ziegler, T. J. Phys. Chem. 1996, 100, 5286.
(36) Khandogin, J.; Ziegler, T. Spectrochim. Acta A 1999, 55, 607.
(37) Autschbach, L.; Ziegler, T. J. Chem. Phys. 2000, 113, 936.
(38) Autschbach, L.; Ziegler, T. J. Chem. Phys. 2000, 113, 9410.
(39) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1992, 99, 84.
(40) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
REFERENCES
■
(1) Cowley, H.; Kemp, R. A. Chem. Rev. J. 1985, 85, 367.
(2) Gudat, D. Coord. Chem. Rev. 1997, 163, 71.
(3) Nakazawa, H. Adv. Organomet. Chem. 2004, 50, 107.
(4) Abrams, M. B.; Scott, B. L.; Baker, R. T. Organometallics 2000, 19,
4944.
(5) Caputo, C. A.; Jennings, M. C.; Tuononen, H. M.; Jones, N. D.
Organometallics 2009, 28, 990.
(6) Rosenberg, L. Coord. Chem. Rev. 2012, 256, 606.
(7) Hutchins, L. D.; Light, R. W.; Paine, R. T. Inorg. Chem. 1982, 21,
266.
(8) King, R. B.; Fu, W.-K.; Holt, E. M. Chem. Commun. 1984, 1439.
(9) Hutchins, L. D.; Duesler, E. N.; Paine, R. T. Organometallics
1982, 1, 1254.
(41) Perdew, J. P. Phys. Rev. B 1986, 33, 8822.
(42) Jameson, C. J.; De Dios, A.; Jameson, A. K. Chem. Phys. Lett.
1990, 167, 574.
(10) Green, M. L. H. J. Organomet. Chem. 1995, 200, 127.
(43) Forster, D.; Dilger, H.; Ehret, F.; Nieger, M.; Gudat, D. Eur. J.
̈
(11) Burck, S.; Daniels, J.; Gans-Eichler, T.; Gudat, D.; Nattinen, K.;
̈
Inorg. Chem. 2012, 3989.
Nieger, M. Z. Anorg. Allg. Chem. 2005, 631, 1403.
(44) Burck, S.; Gudat, D.; Nieger, M.; Du Mont, W.-W. J. Am. Chem.
Soc. 2006, 128, 3946.
(12) Pan, B.; Xu, Z.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C.
M. Inorg. Chem. 2012, 51, 4170.
(13) Amatore, C.; Azzabi, M.; Jutand, A. J. Am. Chem. Soc. 1991, 113,
(45) Burck, S.; Gudat, D.; Nattinen, K.; Nieger, M.; Niemeyer, M.;
̈
Schmid, D. Eur. J. Inorg. Chem. 2007, 5112−5119.
8375.
(46) Cordero, B.; Gom
Echeverría, J.; Cremades, E.; Barragan
2008, 2832.
́
ez, V.; Platero-Prats, A. E.; Reves
́
, M.;
(14) Amatore, C.; Jutand, A.; Suarez, A. J. Am. Chem. Soc. 1993, 115,
9531.
(15) Amatore, C.; Jutand, A. Acc. Chem. Res. 2000, 33, 314.
(16) Kozuch, S.; Shaik, S.; Jutand, A.; Amatore, C. Chem.Eur. J.
2004, 10, 3072.
(17) Gudat, D.; Haghverdi, A.; Nieger, M. J. Organomet. Chem. 2001,
617, 383.
(18) Caputo, A.; Brazeau, A. L.; Hynes, Z.; Price, J. T.; Tuononen, H.
M.; Jones, N. D. Organometallics 2009, 28, 5261.
(19) Dyer, P. W.; Fawcett, J.; Hanton, M. J.; Mingos, D. M. P.;
Williamson, A.-M. Dalton Trans. 2004, 2400.
(20) Calculations were carried out using the Gaussian program suite:
Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.;
Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.;
Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.
(21) Gorelsky, S. I.; Lever, A. B. P. J. Organomet. Chem. 2001, 635,
187.
(22) Gorelsky, S. I. AOMix: Program for Molecular Orbital Analysis;
University of Ottawa, 2012, http://www.sg-chem.net/.
(23) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000,
14, 123.
(24) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J.
E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO 5.0;
Theoretical Chemistry Institute, University of Wisconsin, Madison,
2001.
́
, F.; Alvarez, S. Dalton Trans.
(47) Mean value and standard deviation of the result of a query in the
CSD database for Pd−Pd distances in dinuclear palladium(I)
complexes.
(48) Mean value and standard deviation of the result of a query in the
CSD database for Pt−Cl distances in complexes with metal
coordination numbers of 4 or less.
(49) Edge, R.; Less, R. J.; McInnes, E. J. L.; Muther, K.; Naseri, V.;
̈
Rawson, J. M.; Wright, D. S. Chem. Commun. 2009, 1691.
(50) A similar molecular structure was recently established by both
X-ray diffraction and computational studies on a nickel complex
Mes
[(PN2Mes)NiCl]2 (PN2 = 1,3-dimesityl1,3,2-diazaphospholidinium)
with analogous composition to 6 and 7: R. T. Baker, personal
communication. Hamilton, C. W.; Morris, D. E.; Blair, M. W. ;
Henson, N. J.; Martin, R. L.; Cross, J. L.; Sutton, A. D.; Jantunen, K.
C.; Scott B. L.; Baker, R. T. INOR 50-When is addition of X-Y to M
not oxidative? Presented at the 235th American Chemical Society
National Meeting, New Orleans, LA, 2008. See also ref 6.
(51) Kaupp, M.; von Schnering, H. G. Inorg. Chem. 1994, 33, 2555.
(52) Pyykko, P. Chem. Rev. 1988, 88, 563.
̈
(53) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735.
(54) Dapprich, S.; Frenking, G. J. Phys. Chem. 1995, 99, 9352.
(55) Frenking, G.; Frohlich, N. Chem. Rev. 2000, 100, 717−774.
̈
(56) Bader, R. F. W. Chem. Rev. 1991, 91, 893.
(57) Bader, R. F. W.; Popelier, P. L. A.; Keith, T. A. Angew. Chem.
1994, 106, 647−659; Angew. Chem., Int. Ed. Engl. 1994, 33, 620−631.
(58) Leoni, P.; Marchetti, F.; Papucci, S.; Pasquali, M. J. Organomet.
Chem. 2000, 593−594, 12.
(59) Abrams, M. B.; Hardman, N. J.; Pribisko, M. A.; Gilbert, T. M.;
Martin, R. L.; Kubas, G. J.; Baker, R. T. Angew. Chem., Int. Ed. Engl.
2004, 43, 1955.
(60) bwGRiD (http://www.bw-grid.de), member of the German D-
Grid initiative, funded by Federal Ministry for Education and Research
and the Ministry for Science, Research and Arts Baden-Wurttemberg.
̈
7708
dx.doi.org/10.1021/ic400886b | Inorg. Chem. 2013, 52, 7699−7708