First high-nuclearity palladium halide/carbonyl/phosphine cluster(1) (see abstract) containing an octacapped octahedral Pd6(μ 3-Pd)6(μ3-I)2 fragment: Structure-to-synthesis generation from different synthetic routes

Article abstract of DOI:10.1016/j.ica.2004.09.064

The original synthesis and stereochemical characterization of the [Pd ₁₂(μ₃-I)₂(μ₄-I) ₃(μ₂-CO)₆L₆]⁺ monocation (1) (L = PEt₃; [PF₆]^- salt) was an outgrowth of our investigation of the chemical behavior of two unusual thallium-palladium clusters: (μ₆-Tl)[Pd₃(CO) ₃L₃]₂⁺ (2) which possesses a Pd ₃TlPd₃ sandwich framework, and [Tl₂Pd ₁₂(CO)₉L₉]²⁺ (3) which may be viewed as edge-fusions of three Pd₅ trigonal bipyramids to a central Tl₂Pd₃ trigonal bipyramid. Room-temperature reactions of 2 and 3 with I₂ in THF gave rise in each case to small yields (<10%) of 1 along with two square-planar palladium(II) co-products, trans-Pd₂(μ₂-I)₂I₂L₂ (7) and trans-PdI₂L₂ (8) (L = PEt₃). The geometries and compositions of 1, 7, and 8 were unequivocally established from low-temperature CCD X-ray crystallographic determinations. 1 was characterized by solid-state/solution IR and multinuclear (³¹P, ¹³C, ¹H) NMR spectra. The 12-atom metal-core architecture of this geometrically unprecedented Pd₆(μ₃-Pd) ₆(μ₃-I)₂(μ₄-I)₃ kernel of 1 of crystallographic D₃ (32) site symmetry may be envisioned as a distorted hexacapped octahedral Pd(oc)₆Pd(cap) ₆ core with its two metal-uncappedtrans octahedral Pd(oc)₃ faces additionally capped by iodide μ₃-I atoms. The three tetracapping μ₄-I atoms are each coordinated to two Pd(oc) and two adjacent Pd(cap) atoms. A comparative geometrical/qualitative bonding analysis of the hexacapped octahedral Pd(oc)₆Pd(cap)₆ core in 1 with the structurally analogous cores in the recently reported Pd₁₂ clusters, Pd₁₂(μ₂-CO)₆(μ₃-CO) ₆(PR₃)₆ (R = n-Bu (4), Ph (5)), revealed significantly different architectural features but yet emphasized the importance of the Pd(cap) atoms in stabilizing the Pd(oc) octahedra in 1, 4, and 5. In fact, strong bonding interactions of the μ₃-I and μ₄-I atoms to the Pd₁₂ polyhedron in 1 are evidenced by its black-violet crystals being air-stable for at least one month and by 1 dissolved in THF, acetone, or acetonitrile not undergoing decomposition to AgI upon addition of Ag(OAc). An exploration via the structure-to-synthesis approach of possible preparative pathways involving 14 different chemical reactions was carried out in order to isolate 1 in much higher yields. This systematic investigation demonstrated the importance of conproportionation reactions (i.e., Pd(0) + Pd(II) → Pd(1/2) in 1) utilizing the co-products trans-Pd ₂(μ₂-I)₂I₂L₂ (7) and trans-PdI₂L₂ (8), in different chemical reactions as palladium(II) precursors; high yields of 1 (ca. 50% based upon 6) were obtained from conproportionation reactions in THF of palladium(0) Pd₁₀(CO) ₁₂L₆ (6) with 7 in the presence of Pd(OAc)₂ and (NBu4n)(PF6). The square-planar palladium(II) geometries of the iodide-bridged dimeric 7 and monomeric 8 are compared with each other and with those of the previous crystallographically determined 8 (at room temperature) and several crystallographically known analogues (with different phosphine ligands); corresponding molecular parameters were found to be in remarkably close agreement with distinct bond-length variations in bridging Pd-I(b) and terminal Pd-P bonds being readily attributed to the well-documented trans influence.

Full text of DOI:10.1016/j.ica.2004.09.064

Inorganica Chimica Acta 358 (2005) 1557–1570
www.elsevier.com/locate/ica
First high-nuclearity palladium halide/carbonyl/phosphine cluster,
[Pd12(l3-I)2(l4-I)3(l2-CO)6(PEt3)6]+ monocation
containing an octacapped octahedral Pd6(l3-Pd)6(l3-I)2 fragment:
structure-to-synthesis generation from different synthetic routes
*
Evgueni G. Mednikov, Lawrence F. Dahl
Department of Chemistry, University of Wisconsin-Madison, Madison, WI 53706, USA
Received 9 August 2004; accepted 1 September 2004
Available online 4 February 2005
Dedicated to Gordon Stone, a highly creative and exceptionally prolific inorganic chemist, for his many outstanding accomplishments and leadership
in organometallic chemistry during his illustrious scientific career over approximately 50 years at Harvard University, Queen Mary College
(University of London), University of Bristol, and Baylor University
Abstract
The original synthesis and stereochemical characterization of the [Pd12(l3-I)2(l4-I)3(l2-CO)6L6]+ monocation (1) (L = PEt3;
[PF6]ꢀ salt) was an outgrowth of our investigation of the chemical behavior of two unusual thallium–palladium clusters: (l6-
+
Tl)[Pd3(CO)3L3]2 (2) which possesses a Pd3TlPd3 sandwich framework, and [Tl2Pd12(CO)9L9]2+ (3) which may be viewed as
edge-fusions of three Pd5 trigonal bipyramids to a central Tl2Pd3 trigonal bipyramid. Room-temperature reactions of 2 and
3 with I2 in THF gave rise in each case to small yields (<10%) of 1 along with two square-planar palladium(II) co-products,
trans-Pd2(l2-I)2I2L2 (7) and trans-PdI2L2 (8) (L = PEt3). The geometries and compositions of 1, 7, and 8 were unequivocally
established from low-temperature CCD X-ray crystallographic determinations. 1 was characterized by solid-state/solution IR
and multinuclear (31P, 13C, 1H) NMR spectra. The 12-atom metal-core architecture of this geometrically unprecedented
Pd6(l3-Pd)6(l3-I)2(l4-I)3 kernel of 1 of crystallographic D3 (32) site symmetry may be envisioned as a distorted hexacapped octa-
hedral Pd(oc)6Pd(cap)6 core with its two metal-uncapped trans octahedral Pd(oc)3 faces additionally capped by iodide l3-I
atoms. The three tetracapping l4-I atoms are each coordinated to two Pd(oc) and two adjacent Pd(cap) atoms. A comparative
geometrical/qualitative bonding analysis of the hexacapped octahedral Pd(oc)6Pd(cap)6 core in 1 with the structurally analogous
cores in the recently reported Pd12 clusters, Pd12(l2-CO)6(l3-CO)6(PR3)6 (R = n-Bu (4), Ph (5)), revealed significantly different
architectural features but yet emphasized the importance of the Pd(cap) atoms in stabilizing the Pd(oc) octahedra in 1, 4,
and 5. In fact, strong bonding interactions of the l3-I and l4-I atoms to the Pd12 polyhedron in 1 are evidenced by its
black-violet crystals being air-stable for at least one month and by 1 dissolved in THF, acetone, or acetonitrile not undergoing
decomposition to AgI upon addition of Ag(OAc). An exploration via the structure-to-synthesis approach of possible preparative
pathways involving 14 different chemical reactions was carried out in order to isolate 1 in much higher yields. This systematic
investigation demonstrated the importance of conproportionation reactions (i.e., Pd(0) + Pd(II) ! Pd(1/2) in 1) utilizing the co-
products trans-Pd2(l2-I)2I2L2 (7) and trans-PdI2L2 (8), in different chemical reactions as palladium(II) precursors; high yields of
1 (ca. 50% based upon 6) were obtained from conproportionation reactions in THF of palladium(0) Pd10(CO)12L6 (6) with 7 in
the presence of Pd(OAc)2 and ðNBun4ÞðPF6Þ. The square-planar palladium(II) geometries of the iodide-bridged dimeric 7 and
monomeric 8 are compared with each other and with those of the previous crystallographically determined 8 (at room temper-
ature) and several crystallographically known analogues (with different phosphine ligands); corresponding molecular parameters
*
Corresponding author.
E-mail address: dahl@chem.wisc.edu (L.F. Dahl).
0020-1693/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2004.09.064

1558
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
were found to be in remarkably close agreement with distinct bond-length variations in bridging Pd–I(b) and terminal Pd–P
bonds being readily attributed to the well-documented trans influence.
ꢀ 2004 Elsevier B.V. All rights reserved.
1. Introduction
For heteronuclear palladium clusters, which are not
under consideration herein, l3-X halide coordination
Palladium halides (X = Cl, Br, I) and their deriva-
tives constitute a widespread class of compounds, a
number of which were originally utilized for palla-
dium-catalyzed reactions. Although many of these
palladium halide compounds possess two or more
palladium atoms, the largest ones that are ligated pri-
marily by terminal and/or doubly bridging l2-X ha-
lide atoms contain less than eight metal atoms.
Among homopalladium compounds, triply bridging
l3-X halide coordination has been observed for only
a few 48-electron Pd3 triangular clusters that are sup-
ported by bidentate ligands, such as Ph2PCH2PPh2
(dppm) or Ph2AsCH2AsPh2 (dpam); examples are
was found in CoPd3ðl3-XÞðl2-COÞ ðl3-COÞðCOÞ
3
ðPBut3Þ (where X = Cl, Br, I) [4].
3
Among carbonyl-containing halide derivatives of
homonuclear palladium compounds with monodentate
phosphine ligands, only two kinds of clusters are
known: namely, tetranuclear Pd4(l2-I)2(l2-CO)x(PPh3)4
(where x = 2, 3) [5a] and hexanuclear Pd6(l2-X)4(l2-
CO)4(PR3)4 (where R = Ph, X = Cl [5b,5c] Br, [5b,5c]
I [5a]; R = But, X = Br, I [2]). Of the hexanuclear
Pd6 clusters, crystal structures were determined by X-
ray diffraction studies for Pd6(l2-Cl)4(l2-CO)4(PPh3)4
[5c] and Pd6ðl2-BrÞ ðl2-COÞ ðPBut3Þ [4]. Each consists
4
4
4
of two 44-electron triangular Pd3(l2-X)(l2-CO)2(PR3)2
moieties that are only joined together by two addi-
tional doubly bridging (l2-X) atoms (i.e., no Pd–Pd
bond). The two Pd3(l2-Cl)(l2-CO)2(PPh3)2 triangles
in the former centrosymmetric Pd6 cluster are oriented
in an anti conformation with an overall pseudo-C2h
molecular geometry, whereas the two Pd3ðl2-BrÞ
ðl2-COÞ ðPBut3Þ triangles in the latter Pd6 cluster are
[Pd3(l3-X)(l3-CO)(l2-dppm)3]+ (X = Cl [1a,1b],
I
[1c]), [Pd3(l3-I)(l3-PF3)(l2-dppm)3]+ [1d] and [Pd3(l3-
I)(l3-CO)(l2-dpam)3]+ [1e]. To our knowledge, there
is only one recently reported l3-X halide homopalla-
dium cluster, Pd3(l3-Cl)2Cl3(PPh3)3 [1f] that has
monodentate phosphorus ligands. This 49-electron tri-
angular non-carbonyl Pd3 cluster is characterized by
2
2
˚
0.2–0.3 A longer Pdꢁ ꢁ ꢁPd distances (range 3.064(1)–
in a syn conformation with an overall pseudo-C2v
molecular geometry.
˚
3.154(1) A) [1f] indicative of weaker Pd–Pd bonding
interactions. In striking contrast to the above 48-elec-
tron Pd3 clusters with bidentate phosphorus ligands,
Pd3(l3-Cl)2Cl3(PPh3)3 contains much shorter (normal)
Pd-(l3-Cl) distances for the two independent (l3-Cl)
ligands with means of 2.47 and 2.29(1) A (i.e., this
˚
0.18 A difference was readily attributed to a strong
Of particular interest is the reversible chemical trans-
formation, characterized from TLC, IR, and UV-vis
measurements, of the tetranuclear 58-electron Pd4(l2-
I)2(l2-CO)2(PPh3)4 into the 60-electron Pd4(l2-I)2(l2-
CO)3(PPh3)4 that occurs in solution by the rapid
addition of one bridging carbonyl ligand under a CO
atmosphere; however, this bridging CO ligand can then
be removed under an Ar atmosphere to generate the ori-
ginal Pd4(l2-I)2(l2-CO)2(PPh3)4 cluster [5a]. Although
their solid-state molecular structures have not as yet
been established from X-ray crystallographic studies,
analogous crystallographically characterized palladium
carbonyl/phosphine clusters are known; these include
the 58-electron butterfly-shaped Pd4(l2-CO)5(PPh3)4
[6a,6b] and Pd4(l2-CO)5(PPh2Me)4 [6c], and the 60-elec-
tron tetrahedral-shaped Pd4ðl2-COÞ ðPBun3Þ [6d]. Fur-
˚
˚
trans effect) [1f] than those of 2.90 A for [Pd3(l3-
Cl)(l3-CO)(l2-dppm)3]+ ([CF3CO2]ꢀ salt) [1a] and
+
2.86 A for [Pd3(l3-Cl)(l3-CO)(l2-dppm)3] ([BF4]ꢀ
˚
salt) [1b]. The abnormally long triply bridging Pd-
(l3-Cl) distances in each of these latter two monoca-
tions were ascribed to very weak bonding and raised
the question as to whether the linkage of the (l3-Cl)
ligand is entirely ionic or whether partial covalency is
involved. In this connection, two comprehensive re-
views by Puddephatt et al. [2] and Harvey and collab-
orators [3] provide an excellent detailed overview of
the structures and bonding of these clusters together
with their unusual catalytic, optical, and electrochem-
ical properties and chemical reactivities. Both groups
consider these and related clusters as halide adducts
of the coordinatively unsaturated 42-electron [Pd3(l3-
CO)(l2-dppm)3]2+ host; they attribute the abnormally
long distances between the palladium atoms and ha-
lide anions to steric effects caused by the phenyl rings
of the bidentate dppm ligands that prevent close ap-
proaches of the halide adducts to normal distances.
6
4
thermore, conversion of Pd4ðl2-COÞ ðPBun3Þ into
5
4
Pd4ðl2-COÞ ðPBun3Þ under CO atmosphere has been ob-
6
4
served [6d,6e].
Systematic exploration (by E.G.M.) of various syn-
thetic pathways for obtaining triphenylphosphine
Pd4(l2-I)2(l2-CO)x(PPh3)4 (x = 2, 3) and Pd6(l2-I)4(l2-
CO)4(PPh3)4 compounds gave rise to several different
reaction routes including: (a) direct iodination of
Pdn(CO)x(PPh3)y clusters; (b) conproportionation reac-
tions of zerovalent Pdn(CO)x(PPh3)y clusters with
square-planar palladium(II) PdI2(PPh3)2; (c) oxidative

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1559
1, 2
addition of iodophosphorane, R3PI2
[7,8], to
Fig. 1(c) and (d) shows that the principal C3-axis
passes through the two tricapping l3-I atoms, while
each of the three horizontal C2-axes passes through
one of the three tetracapping l4-I atoms and the mid-
points of two trans-oriented Pd(oc)–Pd(oc) bond-edges
of the octahedral Pd(oc)6 kernel. The resulting crystal-
lographically asymmetric unit is one-sixth of the mon-
ocation (1).
Pdn(CO)x(PR3)y (R = Ph); and (d) reduction of
Pd(OAc)2 by CO in the presence of PPh3 and iodide an-
ions [5a].
Our research interests at UW-Madison have unex-
pectedly become intertwined with the above findings.
During systematic studies in chemistry of two
thallium–palladium carbonyl/phosphine clusters, (l6-
+
Tl)[Pd3(l2-CO)3L3]2 (2) [9] and [Tl2Pd12(l2-CO)6(l3-
CO)3L9]2+ (3) [10] (L = PEt3; [PF6]ꢀ salts), we found
that reaction of the labile sandwich 2 with I2 gives (in
small yields) the unusual [Pd12(l3-I)2(l4-I)3(l2-CO)6
(PEt3)6]+ monocation (1), (isolated as the [PF6]ꢀ salt).
As soon as its composition was unambiguously pro-
vided from an X-ray crystallographic determination,
we then successfully modified and applied via a struc-
ture-to-synthesis approach several different preparative
pathways (mentioned above [5a] for previously isolating
the triphenylphosphine ‘‘Pd4(l2-I)2’’ and ‘‘Pd6(l2-I)4’’
clusters), which in most cases produced 1 in reasonable
yields.
This 12-atom metal-core monocation (1) represents
an initial example of a high-nuclearity halide-containing
carbonyl/phosphine cluster of palladium (arbitrarily
designated by us to have a minimum of 10 metal atoms
with direct metal-metal bonding connectivities). Our
investigation reported herein includes not only several
different synthetic routes to 1 but also its stereochemical
and spectroscopic characterization.
2.1.2. Pd12 core
Its 12-atom metal core can be considered as a dis-
torted hexacapped Pd(oc)6Pd(cap)6 octahedron com-
posed of an octahedral Pd(n) kernel (n = 1, 1A, 1B,
1C, 1D, 1E) that is triangularly capped by six Pd(n)
atoms (n = 2, 2A, 2B, 2C, 2D, 2E); each of these six cap-
ping Pd atoms is connected to a PEt3 ligand (Fig. 1(a)
and (b)). The remaining two metal-uncapped triangular
trans-Pd(oc)3 faces of the Pd(n) octahedron (namely,
those with n = 1, 1B, 1D and n = 1A, 1C, 1E) may be
designated as two symmetry-equivalent intralayer (ba-
sal) equilateral triangles with one independent bonding
Pd(oc)–Pd(oc) distance of 2.840(1) A within each basal
Pd(oc)3 triangle and two independent bonding Pd(oc)–
˚
Pd(oc) distances of 2.821(1) and 2.949(1) A between
the basal Pd(oc)3 triangles. The Pd(oc)6Pd(cap)6 core is
formally constructed by the capping of each of the six
interlayer triangular Pd(oc)3 faces connecting the two
basal triangles with Pd(cap) atoms; the resulting 18
bonding Pd(oc)–Pd(cap) distances consist of three inde-
˚
˚
pendent ones of 2.721(1), 2.768(1), and 2.948(1) A. This
hexacapped octahedral Pd12 core of 1 can also be envi-
sioned as two stacked intralayers of non-planar m2 Pd6
triangles (where m2 designates 3 equally spaced atoms
along each edge) composed of Pd(n) with n = 2, 1B,
2B, 1D, 2D, 1 and n = 2A, 1C, 2C, 1E, 2E, 1A; these
two stacked Pd6 layers conform to a trigonal-antipris-
matic geometry with a twist angle about the principal
threefold axis deviating from a regular staggered confor-
mation by 6.3ꢁ (Fig. 1(a)). Each of the three outer Pd(n)-
capping atoms in the two layers (namely, those with
n = 2, 2B, 2D and n = 2A, 2C, 2E) is displaced from
its corresponding basal octahedral plane (defined by
the three inner Pd atoms with n = 1, 1B, 1D and
n = 1A, 1C, 1E) toward the opposite basal octahedral
Pd(oc)3 triangle; the resulting angle between Pd(n),
n = 1, 2, 1B and n = 1, 1B, 1D planes and the other five
symmetry-equivalent corresponding pairs is 5.4ꢁ. These
significant geometrical deviations are ascribed primarily
to the combined steric/electronic effects of the three l4-I
atoms and six l2-CO ligands (vide infra).
2. Results and discussion
2.1. Solid-state structure of [Pd12(l3-I)2(l4-I)3(l2-
CO)6(PEt3)6]+ (1)
2.1.1. General comments
The consecutive build-up scheme of 1 with its crys-
tallographically imposed trigonal D3 (32) molecular
geometry is shown in Fig. 1(a)–(e). A comparison of
its connectivities is given in Table 1, for which a given
connectivity is identified in Fig. 1 from the atom-label-
ing of a specified individual atom-pair; the number of
symmetry-equivalent distances under D3 symmetry for
a given connectivity is also presented in Table 1.
1
Crystallographic X-ray studies of iodophosphorane R3PI2 species
have shown (to date) that phosphorus possesses a tetrahedral-like R3P-
I-I geometry in the solid state; these include R3 = (t-Bu)3 [7a], Ph3 [7b],
PhMe2 [7c], and [2,4,6-(CH3O)3C6H2]3 [7d].
2
Because we used an excess amount of I2, it also should be noted
that iodophosphoranes R3PI2 are able to react with additional free
2.1.3. Pd12(l3-I)2(l4-I)3 fragment
ꢀ
iodine to form iodophosphonium triiodide R3PI+I3 ions and their
The two oppositely directed triangular basal Pd(oc)3
faces (normal to the pseudo-threefold axis) are capped
by two l3-I ligands with the six symmetry-equivalent
dimeric derivatives [(R3PI)2I3]+I3 with different solid-state structures
ꢀ
that are influenced by the nature of the R substituents and solvents.
Crystallographically characterized examples are [(PPh3I)2I3]I3 [8a],
(PPh3I)I3 [8a], ½ðPPri3IÞ2I3ꢂI3 [8b], and ½ðPrn2NÞ3PIꢂI3 [8b].
˚
Pd(oc)-(l3-I) distances being 2.757(1) A.

1560
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
Fig. 1. The consecutive build-up scheme of [Pd12(l3-I)2(l4-I)3(l2-CO)6L6]+ monocation (1) (L = PEt3; [PF6]ꢀ salt) of crystallographic D3 (32) site
symmetry: (a) top and (b) side views, respectively, of the octacapped octahedral Pd6(l3-Pd)6(l3-I)2P6 fragment formally constructed from the central
Pd(oc)6 octahedron by the capping of six of its triangular faces with Pd(cap)P groups and the remaining two trans triangular faces with l3-I atoms; (c)
top view displaying the additional tetracapping of the octacapped octahedral Pd(oc)6Pd(cap)6(l3-I)2P fragment with three l4-I atoms, each having a
localized butterfly-shaped coordination consisting of two hinge Pd(oc) atoms and two wingtip Pd(cap) atoms; (d) and (e) top and side views,
respectively, of the entire structure (without phosphorus-attached ethyl substituents) formed by the addition of the 12 symmetry-equivalent edge-
bridging l2-CO ligands that span Pd(oc)–Pd(cap) edges of the other three unoccupied butterfly-shaped Pd(oc)2Pd(cap)2 fragments. Pd(oc) atoms are
dark green; Pd(cap) atoms, light green; l3-I and l4-I atoms, purple; carbon atoms, black; and oxygen atoms, red.
Each of the threefold-related tetracapping l4-I atoms
is coordinated to localized butterfly-shaped
Pd(oc)2Pd(cap)2 fragment composed of two twofold-
related basal Pd(oc) atoms and two twofold-related
wingtip Pd(cap) atoms; the determined independent
Addition of the three tetracapping l4-I atoms (and
six doubly bridging l2-CO ligands and 18 phosphorus-
attached ethyl substituents) gives rise to a large
deformation of the Pd6(l3-Pd6)(l3-I)2P6 fragment from
a
ꢀ
pseudo-centrosymmetric D3d ð32=mÞ symmetry to the
˚
Pd(cap)-(l4-I) distances are 3.110(1) and 2.634(1) A.
observed crystallographic D3 symmetry (i.e., involving

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1561
Table 1
˚
Comparison of connectivities (A) for [Pd12(l3-I)2(l4-I)3(l2-
CO)6(PEt3)6]+ (1) under crystallographic D3 site symmetry
Connectivitya
Individual atom–pairb Nc Distance
Pd(oc)–Pd(oc)
(within basal triangles)
Pd(oc)–Pd(oc)
(between basal triangles)
Pd(oc)–Pd(oc)
(between basal triangles)
Pd(oc)–Pd(cap)
(interlayer)
Pd(1)–Pd(1B)
Pd(1)–Pd(1E)
Pd(1)–Pd(1A)
Pd(1A)–Pd(2)
Pd(1)–Pd(2)
6
3
3
6
6
6
2.840(1)
2.821(1)
2.949(1)
2.721(1)
2.768(1)
2.948(1)
Pd(oc)–Pd(cap)
(intralayer)
Pd(oc)–Pd(cap)
(intralayer)
Pd(1B)–Pd(2)
Pd(cap)ꢁ ꢁ ꢁPd(cap)
Pd(cap)ꢁ ꢁ ꢁPd(cap)
Pd(oc)–(l3-I)
Pd(oc)–(l4-I)
Pd(cap)–(l4-I)
Pd(cap)–P
Pd(2)ꢁ ꢁ ꢁPd(2A)
Pd(2)ꢁ ꢁ ꢁPd(2E)
Pd(1)–I(1)
3
3
6
6
6
6
6
6
6
4.267(2)
3.334(2)
2.757(1)
3.110(1)
2.634(1)
2.290(3)
1.97(1)
Fig. 2. Centrosymmetric architecture of the Pd(oc)6Pd(cap)6P6 frag-
ment in the two known isostructural Pd12(l2-CO)6(l3-CO)6(PR3)6
clusters (R = n-Bu (4) [11], Ph (5) [12]). Each of their essentially
identical molecular geometries (other than the different R substituents)
Pd(1)–I(2)
Pd(2E)–I(2)
Pd(2)–P
Pd(oc)-(l2-CO)
Pd(cap)-(l2-CO)
C–O
Pd(1)–C(1)
Pd(2)–C(1)
C(1)–O(1)
ꢀ
of crystallographic Ci ð1Þ site symmetry ideally conforms to centro-
1.96(1)
1.15(1)
ꢀ
symmetric D3d ð32=mÞ symmetry.
a
Pd(oc) and Pd(cap) designate the six symmetry-equivalent octa-
hedral and capping palladium atoms, respectively. The two basal tri-
angles denote the two opposite (trans-oriented) triangular Pd(oc)3
faces of the octahedral Pd6 kernel that are not capped by Pd(cap)
atoms (i.e., they are instead capped by the two l3-I atoms).
(or non-bonding) interlayer Pd(oc)–Pd(oc) means which
˚
differ by 0.06 A (vide infra). Particularly noteworthy is
that the molecular geometry for 4 and also for 5 has
ꢀ
b
Atom-labeling is given in Fig. 1.
N denotes the number of symmetry-equivalent connectivities
under D3 symmetry.
crystallographically Ci ð1Þ site symmetry and that addi-
c
tion of the P atoms (without substituents) and CO li-
gands to its Pd12 core preserves the centrosymmetric
ꢀ
pseudo-D3d ð32=mÞ symmetry.
loss of both the inversion center i and the three vertical
rd mirror planes that bisect the pairs of three horizontal
C2-axes). Steric/electronic consequences are evidenced
by the two wingtip Pd(cap) atoms of the (l4-I)-occupied
butterfly-shaped fragment being separated by a much
longer non-bonding Pd(cap)ꢁ ꢁ ꢁPd(cap) distance of
In contradistinction, the Pd6(l3-Pd)6(l3-I)2P6 frag-
ment in 1 is greatly deformed (as previously mentioned)
from pseudo-D3d to crystallographic D3 symmetry be-
cause of the non-conformity of the three l4-I and six
l2-CO ligands to D3d symmetry. Visual evidence for
the much greater geometric distortions of the Pd12 core
in 1 is shown from a comparison of Figs. 1 and 2.
Apparent differences include the two symmetry-identical
intralayers in 4 and 5 being centrosymmetrically ori-
ented in a staggered close-packed array such that the
two interlayer Pd(oc) triangles of the Pd(oc)6 kernel
form a regular trigonal antiprism with the six interlayer
Pd(cap) atoms being equally separated; in sharp con-
trast, the six Pd(cap) atoms in 1 are irregularly separated
(Fig. 1) due primarily to steric effects involving the three
tetracapping l4-I atoms (vide supra). The resulting ma-
jor structural variation between 1 and 4 (5) is that the
average framework of the inner Pd(oc)6 kernels in 4
(5) is much more deformed from an octahedral toward
a regular trigonal-antiprismatic geometry due to a con-
siderably greater elongation of the six interlayer
Pd(oc)–Pd(oc) distances (between basal triangles) with
˚
˚
4.267(2) A compared to that of 3.334(2) A for the two
wingtip Pd(cap) atoms of the other three symmetry-
equivalent unoccupied butterfly-shaped fragments.
2.2. Geometrical/bonding relationship of the hexacapped
octahedral Pd12 core in [Pd12(l3-I)2(l4-I)3(l2-CO)6
(PEt3)6]+ (1) with the hexacapped octahedral Pd12 core
in two isostructural [Pd12(CO)12(PR3)6] clusters
(R = n-Bu (4), Ph (5))
A comparison of the Pd6(l3-Pd)6 core geometry of 1
with those of the recently reported isostructural Pd12
clusters, Pd12(l2-CO)6(l3-CO)6(PR3)6 (R = n-Bu (4)
[11], Ph (5) [12]) is informative. These latter two molec-
ular clusters (represented in Fig. 2 by the triphenylphos-
phine 5), which were isolated from entirely different
synthetic routes, possess essentially identical geometries
(other than the different phosphorus-attached substitu-
ents); corresponding means for their Pd6(l3-Pd)6 cores
˚
resulting means of 3.23 (3.17) A for 4 (5) versus 2.82
˚
and 2.95 A for the corresponding non-equivalent means
(under D3 symmetry) for 1. The bonding 12 intralayer
and six interlayer Pd(cap)–Pd(oc) means of 2.81 (2.82)
˚
vary by less than 0.02 A except for the weakly bonding

1562
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
2.3. Synthesis of [Pd12(l3-I)2(l4-I)3(l2-CO)6L6]+ (1)
(L = PEt3; [PF6]ꢀ salt) from different preparative
pathways
˚
and 2.70 (2.72) A, respectively, for 4 (5) versus 12 intra-
˚
layer ones of 2.77 and 2.95 A and six interlayer ones of
˚
2.72 A for 1 are in accordance with Pd(cap) atoms sta-
bilizing the Pd(oc)6 kernels in both 1 and 4 (5) via
Pd(cap)–Pd(oc) bonding interactions.
2.3.1. General comments
An examination of predicted electron counts for the
octacapped octahedral Pd6(l3-Pd)6(l3-I)2 fragment in 1
compared to the hexacapped octahedral Pd6(l3-Pd)6
core in 4 (5) provides a qualitative indication concerning
their overall bonding differences. If the Pd12 core in 4 (5)
is considered as a hexacapped octahedron, the cluster
valence electron count is 84 CVEs (i.e., 6 · 10
(Pd(oc)) + 6 · 0 (Pd(cap)PR3) + 12 · 2 (CO) = 84). The
corresponding electron count for the octacapped octahe-
dron in 1 is not unambiguous due to the donor contribu-
tion of each l4-I atom (designated ‘‘n’’) not being
straightforward [i.e., 6 · 10 (Pd(oc)) + 6 · 0 (Pd(cap)-
PEt3) + 2 · 5 (l3-I) + 3 · n (l4-I) + 2 · 6 (CO) ꢀ 1
(charge) = 81+3n(l4-I) CVEs]. If each l4-I ligand is as-
sumed to donate five of its seven valence electrons
(i.e., n = 5), the resulting electron count of 96 CVEs is
abnormally high and unrealistic for Pd-capped Pd6
octahedral clusters. In any event, this qualitative elec-
tron-count scheme supports our view that the iodide-
coordinated 1 is an electron-rich cluster.
Based upon the crystallographically determined
structure and composition of 1, which was originally ob-
tained in low yield from the reaction of the labile
þ
ðl6-TlÞ½Pd3ðl2-COÞ L3ꢂ
sandwich (2) (L = PEt3;
3
2
[PF6]ꢀ salt), a systematic exploration of possible syn-
thetic routes via the structure-to-synthesis approach
was carried out in order to isolate 1 in much higher
yields. Table 2 provides a summary of the reactants
and boundary conditions of 14 different reactions car-
ried out at room temperature in THF along with the
estimated yields of the crystalline products. In most
cases 1 was obtained in reasonable yields.
2.3.2. Direct iodination of heterometallic triethylphosphine
palladium/thallium Pd6Tl (2) and Pd12Tl2 (3) clusters
and the homometallic triethylphosphine Pd10 cluster (6)
Entries 1–3 in Table 2 show that 1 as the [PF6]ꢀ salt
was isolatedþ from reactions of I2 with ðl6-TlÞ½Pd3
ðl2-COÞ L3ꢂ ð2Þ [9] and [Tl2Pd12(CO)9 L9]2+ (3) [10],
3
2
and with Pd10(CO)12L6 (6) [6e,13] in the presence of
Table 2
Summary of reactants, reaction conditions, and estimated yields of crystalline products for 14 different reactions performed at room temperature in
THF in an exploration of possible synthetic routes to generate [Pd12I5(CO)6L6]+ monocation (1) as [PF6]ꢀ salt (L = PEt3)
Number Reactants and their stoichiometry
of entry
Quantity of
reactant cluster
(mmol)
Atmosphere
Yields of crystalline products (%)
[Pd12I5(CO)6-
L6]PF6 (1[PF6])
Pd2I4L2 (7) PdI2L2 (8)
1
(l6-Tl)[Pd3(CO)3L3]2[PF6] + 4.5I2
[Tl2Pd12(CO)9L9][PF6]2 + 6I2
Pd10ðCOÞ12L6 þ Pd2dba3 þ 3:5I2 þ 2ðNBun4ÞPF6
Pd10(CO)12L6 + 3Pd2I4L2 + 4Pd2dba3 + 4(NH4)(PF6)
Pd10(CO)12L6 + 3Pd2I4L2 + 4Pd2dba3 + 4(NH4)(PF6)
0.05
0.06
0.05
0.03
0.02
N2
N2
N2
N2
5
32
3
17
17
2a
3a
4
8
3b
35b, 15c
6b
0
6b
d
d
5
CO, 0.5 min; 67b, 28c
then N2
0
6e
7f
Pd10(CO)12L6 + 3Pd2I4L2 + 4Pd2dba3 + 4(NH4)(PF6)
6PdI2L2 + 9Pd2dba3 + 4(NH4)(PF6)
0.02
CO
45b, 19c
3c
0
0
18b
34g
N2, 3 min;
then CO
N2
8
Pd10(CO)12L6 + 2Pd2I4L2 + Pd(OAc)2 + 2(NH4)(PF6)
Pd10(CO)12L6 + 4PdI2L2 + 2(NH4)(PF6)
0.05
0.05
54b
0
21b, 53h
57g
24g
27b, 68h
43b, 110h
69i, 59h
12
9
N2
N2
17b
0
10
11
12
13
14
Pd10(CO)12L6 + 3PdI2L2 + 3Pd(OAc)2 + 2(NH4)(PF6) 0.05
47b
0
Pd10(CO)12L6 + I2 + 2Pd2I4L2 + 2(NH4)(PF6)
Pd10(CO)12L6 + I2 + 2Pd2I4L2 + 2(NH4)(PF6)
2Pd4(CO)5L4 + 4.6Pd2I4L2 + (NH4)(PF6)
Pd(OAc)2 + CO + 0.5L + 0.4KI + 4(NH4)(PF6)
0.05
0.05
0.12
1.00j
N2
CO
15b, 12c
4b, 3c
13i
0
0
N2
CO
33h
0
1.6
a
b
c
Small amounts (<2 mg) of colorless crystals were observed but not further investigated.
Based on Pd10(CO)12L6.
Based on all palladium reactants.
d
e
Observed but not defined quantitatively.
Solution of Pd2dba3 in CHCl3 was added through stainless steel canula (without direct contact with CO gas) into THF solution of other
reactants dissolved under CO.
f
Solvent CHCl3 used instead of THF.
Based on initial amount of PdI2L2.
g
h
Based on Pd2I4L2.
Based on Pd4(CO)5L4.
mmol of Pd(OAc)2.
i
j

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1563
Pd2dba3 and ðNBu4nÞðPF6Þ (L = PEt3; dba = dibenzy-
lideneacetone). In the last reaction (entry 3), Pd2dba3
was used to balance the idealized stoichiometry of the
following reaction:
3) clusters with estimated yields of 90% [5a]. Neverthe-
less, our attempts to decrease molar Pdn/I2 ratios in reac-
tions with 2, 3, or 6 did not result in higher yields of 1.
It is apparent that the counterion [PF6]ꢀ plays a cru-
cial role in the stabilization of the monocation (1). Thus,
previous iodination reactions of the corresponding trib-
utylphosphine Pd10ðCOÞ ðPBun3Þ over a broad range of
Pd10ðCOÞ L6 þ Pd2dba3 þ 3I2 þ ðNBun4ÞðPF6Þ
12
! ½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 6CO þ 3dba þ ðNBun4ÞI
6
12
6
molar Pd10/I2 ratios of 1/1–1/10 without counter ions
gave no evidence for the formation of the analogous
monocation (1) [5a].
ð1Þ
All three reactions are complicated by the formation of
Pd black.
Hypothetically I2 could play the same role as O2 for
enlargement of the Pdn-core sizes of Pdn(CO)x(PR3)y
clusters by oxidative deligation of PR3 ligands with for-
mation of phosphorane R3PI2 species (analogous to the
formation of R3PO species upon reaction with O2 [14]);
this potential synthetic pathway would involve the inter-
mediate generation of coordinatively unsaturated palla-
dium intermediates which would then oligomerize to
larger palladium clusters. But in contrast to O2, reactions
with iodine normally result in the oxidation of Pd(0) into
Pd(II) species that are stabilized by formation of strong
Pd–I bonds. Thus, reactions with iodine (entries 1–3)
gave rise to only small yields of 1 due largely to the for-
mation of the stable square-planar palladium(II) trans-
Pd2I4(PEt3)2 (7) and trans-PdI2(PEt3)2 (8); both 7 and 8
were identified from 31P{1H} NMR and complete X-
ray structural analyses. It should be noted that formation
of 7 and 8 could have occurred not only from direct oxi-
dation by I2 of the zerovalent palladium atoms in the
reactants 2, 3 and 6 but also from secondary in situ oxi-
dation reactions with the expectedly formed iodophos-
phorane Et3PI2 or Et3PI4 species (see footnotes 1 and 2)
[7,8]. In fact, a fast reaction between R3PI2 (R = Ph
and Pd10(CO)12(PPh3)6) under CO atmosphere was pre-
viously reported to give Pd4(l2-I)2(l2-CO)3(PPh3)4 [5a].
Furthermore, McAuliffe and co-workers [15] showed
that reactions of iodophosphoranes with unactivated
metal powders produced a variety of new and existing
transition metal phosphine complexes.
2.3.3. Conproportionation reactions of zerovalent Pd10
(CO)12(PEt3)6 (6), Pd4(CO)5(PEt3)4, and Pd2dba3
with square-planar palladium(II) Pd2I4L2 (7) and
PdI2L2 (8) (L = PEt3)
In further attempts to increase the yield of 1, we
developed another synthetic approach (Table 2, entries
4–6), based on the idealized stoichiometry of the follow-
ing reaction:
Pd10ðCOÞ L6 þ 3Pd2I4L2 þ 4Pd2dba3 þ 2ðNH4ÞðPF6Þ
12
! 2½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 12dba þ 2ðNH4ÞI
6
ð2Þ
In sharp contrast to the preparation of triethylphos-
phine 1 via direct oxidation with I2 (entries 1–3), reac-
tion (2) utilizes as a starting reagent the undesirable
final product of the previous reactions, Pd2I4(PEt3)2
(7). Another major difference between reactions involv-
ing the direct oxidation with I2 is that the role of oxidant
is carried out by palladium(II) (from Pd2I4L2) instead of
I2 (Table 2, entries 4–6). Thus, reaction (2) is a conpro-
portionation process involving changes in formal oxida-
tion states of palladium from 0 and +2 in the precursors
6 and 7, respectively, to that of +0.5 in 1.
In spite of the fact that reaction (2) is again compli-
cated by formation of Pd black, this different chemical
route afforded much greater yields of 1 than that involv-
ing direct iodination of clusters 2, 3, 6 (compare entries
4–6 versus 1–3 in Table 2).
Subsequent possible reactions between reactants 2, 3
and 6 with the generated 7 and 8 must also be consid-
ered. Although a considerable number of reactions be-
tween triphenylphosphine Pdn(CO)x(PR3)y clusters and
palladium(II) halide compounds have previously been
performed [5], analogous reactions are less likely to oc-
cur in the present cases because of the higher reactivities
of 2, 3 and 6 (especially that of 2 [9]) with I2 and their
significant consumption at the initial stage of each of
these reactions.
In general, molar ratios of Pdn(CO)x(PR3)y/I2 should
greatly affect the composition and yields of the products.
For instance, reactions of I2 with triphenylphosphine
Pd4(CO)5(PPh3)4 under Ar or CO atmosphere with a
molar Pd4/I2 ratio of 1:1 afforded the previously men-
tioned triphenylphosphine Pd4I2(CO)x(PPh3)4 (x = 2 or
The other crystalline product of reaction (2) is the
unexpected mononuclear 16-electron square-planar pal-
ladium(II) complex, trans-PdI2(PEt3)2 (8). The observa-
tion of 8 was unanticipated because of the large (four
times) deficiency of PEt3 necessary for its formation
from reaction (2), in which the total molar Pd/PEt3 ratio
of the reactants was 2/1.
The sole source of CO for formation of 1 by reaction
(2) (entry 4) is Pd10(CO)12(PEt3)6 (6). In reality the avail-
able amount of CO from 6 is not sufficient for the best
yield of 1, because carbonyl/phosphine clusters of palla-
dium readily lose CO under an inert atmosphere [14b],
especially in the presence of other reagents [14c]. In fact,
the yield of 1 was found to be twice higher if a short-
time atmosphere of CO was applied to the initial stage
of reaction (2) (entry 5). On the other hand, if reaction

1564
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
(2) was exclusively performed under CO atmosphere
(entry 6), the resulting yield of 1 decreased to an analo-
gous value as that obtained in reactions under N2 atmo-
sphere with the sole source of CO being the Pd10 cluster
(6) (compare entry 4 with entry 6). Thus, the short-time
exposure to a CO atmosphere was found to avoid to a
considerable extent the undesirable competitive reaction
between Pd2dba3 and CO with formation of Pd black
and simultaneously to support the formation of 1 in
reaction (2).
An attempt (entry 7) to use the mononuclear PdI2
(PEt3)2 complex (8) instead of binuclear PdI4(PEt3)2
complex (7), in the absence of Pd10(CO)12(PEt3)6 (6)
gave only a small yield of 1 in reaction (3) mainly due
to the decomposition of Pd2dba3 to Pd black in the pres-
ence of CO:
tion with Pd2I4L2 (7) (entry 4) raised the question as
to whether both synthetic approaches could be com-
bined, as shown (entry 11) in the idealized stoichiometry
of the following reaction:
Pd10ðCOÞ L6 þ I2 þ 2Pd2I4L2 þ ðNH4ÞðPF6Þ
12
! ½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 2PdI2L2 þ 6CO þ ðNH4ÞI
6
ð6Þ
Though reaction (6), and likewise reaction (4), implies
the formation of PdI2L2 (8), the yields of 1 in these reac-
tions are less than those from only the conproportiona-
tion reaction (2) (compare entries 11 and 12 versus
entries 4–6). The yield of 110% for PdI2L2 (8) (based
only upon Pd2I4L2 (7)) from reaction (6) performed un-
der CO (entry 12) clearly indicates that its formation re-
sults from both palladium precursors Pd10(CO)12L6 (6)
and Pd2I4L2 (7).
The use (entry 13) of Pd4(CO)5L4 instead of
Pd10(CO)12L6 (6) in accordance with the idealized stoi-
chiometry of reaction (7) afforded a 3–5 times lower
yield of 1 than those found in reactions (2) and (4) (com-
pare entry 13 versus entries 4–6, 8):
6PdI2L2 þ 9Pd2dba3 þ 2ðNH4ÞðPF6Þ þ 12CO
! 2½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 27dba þ 2ðNH4ÞI
ð3Þ
6
We found that the elimination of Pd2dba3 from the
starting reagents by replacement with Pd(OAc)2 de-
creased the amount of Pd black and gave good yields
of 1 (entry 8) in accordance with the idealized stoichiom-
etry of the following reaction:
4Pd4ðCOÞ L4 þ 8Pd2I4L2 þ ðNH4ÞðPF6Þ
5
! ½Pd12I5ðCOÞ L6ꢂPF6 þ 13PdI2L2 þ 14CO þ 7Pd þ ðNH4ÞI
6
Pd10ðCOÞ L6 þ 2Pd2I4L2 þ PdðOAcÞ þ ðNH4ÞðPF6Þ
ð7Þ
12
2
! ½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ PdI2L2 þ PdðOAcÞ L2
6
2
This observation is of particular interest, because
clearly it indicates that even though 1 may be formally
constructed from three butterfly Pd4 subunits joined by
iodides (see Fig. 1), a direct synthetic pathway involving
this idealized condensation (of three butterfly Pd4 cores)
in the presence of triethylphosphine and iodide did not
give rise to a higher yield. 3 In fact, this systematic inves-
tigation established that the highest yields of 1 were
obtained from conproportionation reactions of Pd10
(CO)12L6 (6) with Pd2I4L2 (7) in the presence of either
Pd2dba3 (entries 4–6) or Pd(OAc)2 (entries 8 and 10).
þ 6CO þ Pd þ ðNH4ÞI
ð4Þ
Reaction (4) implies the formation of PdI2L2 (8) and Pd
black, both of which were undesired products obtained
from either direct iodination or conproportionation
reactions.
Entry 9 in Table 2 reveals that 1 can also be isolated
from reactions of 6 with mononuclear PdI2L2 (8). In this
case, the rate of reaction was observed to be much
slower compared with the rates of reactions (2) and
(4), and needed one day instead of 1 h for completion.
Furthermore, the rate of reaction and yield of 1 were
found to be much higher upon the addition of Pd(OAc)2
in the idealized stoichiometry of the following reaction
(entry 10 versus entry 9):
2.3.4. Reduction of Pd(OAc)2 by CO in presence of KI,
PEt3, and (NH4)(PF6)
This synthetic pathway could conceivably have an
obvious advantage upon comparison with the previous
synthetic routes (presented above), because it does not
require the synthesis of either the Pd10 cluster (6) or
other palladium clusters. However, in complete contrast
to the previous preparation of triphenylphosphine
Pd4I2(CO)3(PPh3)4 in 80% yield by reduction of
Pd(OAc)2 with CO in the presence of iodide and PPh3
Pd10ðCOÞ L6 þ 3PdI2L2 þ 3PdðOAcÞ þ ðNH4ÞðPF6Þ
12
2
! ½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 3PdðOAcÞ L2 þ 6CO
6
2
þ Pd þ ðNH4ÞI
ð5Þ
Reactions of 6 and the PdI2L2 complex (8), carried
out with or without Pd(OAc)2 (entry 10 versus entry
9), gave rise to the formation of a very unstable product,
which was not characterized. In order to avoid contam-
ination of 1 by this product and by products of its
decomposition, recrystallization of 1 from THF/hexane
solutions was effectively twice performed.
3
Although a high yield of Pd12(CO)12(PPh3)6 (5) obtained from
condensation of Pd4(CO)5(PPh3)6 [6a,6b] in the presence of oxygen (see
[14b] and note added in proof in [12]) may be considered as a structure-
to-synthesis appraisal, this approach does not necessarily have ‘‘cure-
all’’ character.
The above isolation of 1 from reactions of 6 via oxi-
dation either with I2 (entry 3) or from conproportiona-

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1565
[5a], 1 was isolated in very small quantity from the ide-
alized stoichiometry of the following reaction (entry 14):
bridging iodide ligands being strongly linked to the
Pd12 polyhedron in 1 is that we did not observe either
their elimination as AgI upon treatment of solutions
of 1 (in THF, Me2CO, or MeCN) with Ag(OAc) or
any signs (within a reasonable time) of reaction be-
tween 1 and TlCo(CO)4. Furthermore, whereas palla-
dium carbonyl/phosphine clusters and their known
halide derivatives are generally sensitive to air, it was
found that 1 in dried form is stable under air for at
least one month.
12PdðOAcÞ þ 15CO þ 6L þ 5KI þ ðNH4ÞðPF6Þ þ 9H2O
2
! ½Pd12I5ðCOÞ L6ꢂ½PF6ꢂ þ 9CO2 þ 5KOAc
6
þ ðNH4ÞðOAcÞ þ 18HOAc
ð8Þ
2.4. Synthesis and characterization of the corresponding
bromide analogue, [Pd12(l3-Br)2(l4-Br)3(l2-CO)6L6]+
monocation (9) (L = PEt3; [PF6]ꢀ salt) from
conproportionation reaction
2.6. Spectroscopic characterization of 1
On the basis of the methodology utilized to prepare 1
from the conproportionation reaction (4), a similar reac-
tion in THF of Pd10(CO)12L6 with trans-Pd2(l2-
Br)2Br2L2 in the presence of Pd(OAc)2 and (NH4)(PF6)
was carried out. A small quantity (<5%) of black-violet
crystals, which readily transformed into oil, was isolated
together with two square-planar palladium(II) bromide
co-products, trans-PdBr2L2 (54%) and cis-PdBr2L2
(11%). The unambiguous formulation of 9 as the bro-
mide analogue of 1 is based upon IR and 31P {1H}
NMR measurements which revealed virtually identical
spectroscopic values: IR spectrum m(CO): THF solution,
An IR spectrum of 1 in THF solution exhibited
one strong carbonyl band at 1826 cmꢀ1, whereas an
IR spectrum of 1 in nujol displayed one broadened
strong carbonyl band at 1810 cmꢀ1. These spectra
suggest the presence of just one type of bridging car-
bonyl ligand.
31P{1H} NMR spectra of 1 in CDCl3 solution under N2
exhibited one signal at 7.4 ppm indicative of six equivalent
phosphorus nuclei. A NMR 13C{1H} spectrum of a
CDCl3 solution of crystals of 1 showed strong signals of
methylene and methyl carbon nuclei from the phospho-
rus-attached ethyl substituents at 18.1 and 8.7 ppm,
respectively; a low-field signal at 215.7 ppm was attrib-
uted to the 13C nuclei of six carbonyl ligands, and two
equal weak signals at 68.2 and 25.8 ppm were ascribed
to nuclei of the original crystal-solvated THF molecules.
The signal at 215.7 ppm was readily assigned to the car-
bonyl 13C nuclei on the basis of its relative intensity and
chemical shift value. The assignment of the two weak sig-
nals to THF was substantiated from a control 13C{1H}
NMR spectrum of THF (recorded at the same conditions)
which consisted of singlets at 67.8 (CH2–O) and 25.5 ppm
1824 (s) cmꢀ1 31P{1H} NMR spectrum, CDCl3 (N2),
;
(121 MHz): d = 10.0 ppm (s), ꢀ145 ppm (septet). Partic-
ularly noteworthy is that the reflected/transmitted colors
of a THF or CDCl3 solution of the black-violet crystals
of 9 are the same as those of 1: namely, black in reflected
light, and violet in transmitted light. Both trans- and cis-
PdBr2L2 co-products were identified from their 31P{1H}
and 13C{1H} NMR spectra; yields are based upon the
Pd2Br4L2 precursor (L = PEt3).
2.5. Attempted syntheses of mixed Pd–Pt analogues of 1
or of other Pd–Pt clusters
(C–CH2–CH2–C). Noteworthy is that the methylene 13
C
signals in 1 consisted of an unresolved multiplet (with
the appearance of a distorted triplet) which could only
arise from 4J(P,P) coupling [16]. In the absence of J(P,P)
Cross mixed-metal conproportionation reactions
were carried out. Analogous to the corresponding hom-
opalladium reaction in entry 10, the first one (a) in-
volved the reaction of palladium(0) Pd10(CO)12L6 with
platinum(II) PtI2L2 in the presence of Pd(OAc)2 and
(NH4)(PF6). The second one (b) involved the reaction
of platinum(0) Pt5(CO)6(PEt3)4 with palladium(II)
Pd2I4L2 in the presence of (NH4)(PF6).
Neither reaction afforded heterometallic Pd–Pt clus-
ters, but instead gave the following identified products
with estimated yields (in parentheses): 1 (57%) and 8
(3%) for reaction (a), and 1 (4%), 8 (6%), Pt2I4L2
(11%), and trans-PtI2L2 (54%) for reaction (b). The ac-
tual transfer in reaction (a) of the iodide ligands from
the platinum(II) PtI2L2 to 1 with each Pd possessing a
formal oxidation state of +0.5 indicates a relatively
strong binding of the l3-I and l4-I ligands to the
Pd12 polyhedron in 1. Other chemical evidence of the
1
interactions, it should be a doublet due to only J(C,P)
coupling (i.e., as was observed for a first-order spectrum
of 7), whereas for the presence only one type of J(P,P) cou-
pling – a triplet (i.e., as was observed for a second-order
spectrum of 8 [17a] and for a number of related com-
2
pounds with only J(P,P) interactions between phospho-
rus atoms [17]). The observation of the above-mentioned
multiplet for the 13C methylene signals in 1 indicates two
possible types of 4J(P,P) coupling realized through both
kinds of PPd(cap)Pd(oc)2Pd(cap)P butterfly fragments
of 1: namely, the three symmetry-equivalent (l4-I)-occu-
pied ones and the three symmetry-equivalent unoccupied
ones with two l2-CO ligands spanning two of the four
4
wingtip-hinge Pd(cap)–Pd(oc) edges. The J(P,P) cou-
pling in the latter butterfly-shaped PPd(cap)(l2-CO)P-
3
d(oc)2(l2-CO)Pd(cap)P fragment may include a J(P,P)
component due to relatively weak, but direct interaction

1566
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
between the two P-attached wingtip Pd(cap) atoms sepa-
˚
Pd–I(b)–Pd0 bond angles (range 94.61(2)–95.68(5)ꢁ)
rated by 3.33 A.
and I(b)–Pd–I(b0) bond angles (range 84.32(5)–
A NMR 1H spectrum of a CDCl3 solution of crystals
of 1 displayed a doublet of triplets (dt) at 1.13 ppm
(–CH3) and a multiplet at 1.94 ppm (–CH2–P), which
could be either a sextet or octet. Reliable definition of its
multiplicity is virtually impossible due to overlapping
of higher field edge of this signal with the complex pat-
tern of signals from the methylene protons, C–CH2–
CH2–C, of the THF molecule at 1.85 ppm. In any event,
the methylene proton signal at 1.94 ppm could not be
explained in terms of a first-order spectrum. In contrast
to this signal, the methyl proton signal at 1.13 corre-
sponds to a first-order spectrum with only 3J(H,H)
85.39(2)ꢁ). The Pd–P bonds (range 2.248–2.263(1) A)
˚
are likewise essentially equivalent.
Table 3 discloses that the corresponding molecular
parameters determined at 100 K for the two centrosym-
metric half-molecules A and B of trans-PdI2(PEt3)2 (8)
are virtually identical with those previously determined
[18c] for 8 at 293(2) K, even though the unit cell volume
and corresponding monoclinic lattice parameters are
significantly different (with the corresponding low-
temperature values being expectedly smaller than those
at room-temperature). 4 Table 3 also shows that the cor-
responding molecular parameters are in remarkably
close agreement with those of the triphenylphosphine
analogue (both unsolvated [18d] and chloroform sol-
vated [18e]). Noteworthy is that the Pd–I(t) distances
3
and J(H,P) splitting. As mentioned above, one group
of proton signals of the THF molecule was observed
at 1.85 ppm, while the second one from the two CH2–
O fragments was displayed at 3.75 ppm with the specific
symmetrical pattern precisely corresponding to that at
3.73 ppm for the THF control spectrum per se. Compar-
ison of its integral intensity with that of the dt signal of
CH3 and the overlapped signals at 1.94 ppm of CH2–P
gave virtually the same ratios of 1/37 and 1/38, respec-
tively, which correspond to the 1/THF ratio of 1/0.4
for crystals of 1 grown from THF/hexane. (A small con-
tribution of THF into the integral intensity of signals
from CH2–P at 1.94 ppm was taken into account by sub-
traction of the integral intensity of signals at 3.75 ppm,
which is equal to the overlapped THF component.)
The determined chemical shifts in the NMR 1H spec-
˚
(range 2.598(1)–2.610(1) A) for these trans-PdI2(PR3)2
˚
complexes compare favorably (within 0.02 A) with both
the terminal Pd–I(t) and bridging Pd–I(b) distances
whose bonds are trans to the terminal I(t) bonds for
the trans-Pd2(l2-I)2I2(PR3)2 dimers. The markedly long-
˚
er mean distance of ca. 0.08 A for the bridging Pd–I(b)
bonds trans to the phosphine ligands in the dimers may
be attributed to the trans influence [19]. The other salient
structural feature illustrated in Table 3 is that the two
symmetry-equivalent terminal Pd–P bonds in the trans-
˚
Pd2(l2-I)2I2(PR3)2 dimers are approximately 0.10 A
shorter than those in the trans-PdI2(PR3)2 monomers;
this distinct bond-length variation may likewise be as-
cribed to an electronic effect primarily involving the
trans ligand in a localized square-planar palladium(II)
environment. This highly significant difference is in
accordance with X-ray crystallographic bond-length
observations which led to the proposal 30 years ago that
‘‘a metal-phosphine bond was less stable trans to an-
other phosphine than when trans to a halide’’ [19].
1
trum of 7 and NMR 31P{1H}, 13C{1H} and H spectra
of 8 are in good agreement with ones reported previ-
ously [17a].
2.7. Comparative structural/bonding analysis of the
co-products 7 and 8
The molecular square-planar palladium(II) structures
of trans-Pd2(l2-I)2I2(PEt3)2 (7) and trans-PdI2(PEt3)2 (8)
were established unequivocally from low-temperature
CCD X-ray crystallography. Fig. 3(a) and (b) displays
their geometries (without Et substituents) with atom-
labeling; selected bond lengths and bond angles for both
the iodide-bridged dimeric 7 and monomeric 8 are listed
in Table 3 along with molecular parameters which agree
extremely well with those given for other phosphine io-
dide analogues [18].
4
The initial high-quality structural determination of trans-PdI2
(PEt3)2 (8) reported by Fehlhammer and co-workers [18c] in 2001 was
based upon X-ray diffractometry data collected with Mo Ka radiation
at room-temperature (293(2) K). Crystal data include the following:
˚
˚
monoclinic; P21/c with Z = 4; a = 17.726(4) A, b = 8.116(2) A,
3
˚
˚
c = 15.569(6) A, b = 113.35(3)ꢁ. V = 2056.4(10) A . Corresponding
crystal data collected at 100(2) K in our independent structural
determination of trans-PdI2(PEt3)2 (8) presented herein are: P21/c with
˚
˚
˚
Z = 4; a = 17.688(2) A, b = 7.899(1) A, c = 15.543(2) A, b = 114.48(1)ꢁ.
V = 1976.3(4) A . Both analogous structural refinements resulted in
A comparison of trans-Pd2(l2-I)2I2(PEt3)2 (7) with
the tri(n-butyl)phosphine analogue [18a] and diphenylvi-
nylphosphine iodide analogue [18b] reveals that the
3
˚
reasonably close agreement of final unweighted and weighted R
indices: namely, unweighted R1(F) = 0.035 versus 0.025 [I > 2r(I)] and
weighted wR2(F2) = 0.102 versus 0.067 (all data). We conclude that the
˚
significantly smaller volume difference of 80 A and smaller b lattice-
˚
length difference of 0.22 A reported in our structural determination are
˚
bridging Pd–I(b) bonds (range 2.660–2.685(1) A) trans
to the phosphine ligands are considerably longer than
3
˚
the Pd–I(b) bonds (range 2.593(2)–2.608(1) A) trans to
the terminal I(t) iodides and the Pd–I(t) bonds (range
˚
2.581(2)–2.595(1) A). The resulting asymmetric Pd2I(b)2
rhombus of each complex possesses distinctly different
primarily a consequence of our low-temperature data collection. Of
prime interest is that the corresponding molecular parameters for the
˚
two structural determinations are amazingly alike (within 0.007 A and
0.35ꢁ).

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1567
Fig. 3. Molecular square-planar palladium(II) geometries (without Et substituents) of: (a) dimeric trans-Pd2(l2-I)2I2(PEt3)2 (7); and (b) monomeric
trans-PdI2(PEt3)2 (8).
Table 3
Comparison of selected bond lengths (A) and angles (ꢁ) for trans-Pd2(l2-I)2I2(PEt3)2 (7) and trans-PdI2(PEt3)2 (8) with several known analoguesa
˚
Complexa
Ref.
Pd(1)–I(1) Pd(1)–I(2) Pd(1)–I(2A) Pd(1)–P(1) P(1)–Pd(1)–I(1) I(2)–Pd(1)–I(2A) Pd(1)–I(2)–Pd(1A)
trans-Pd2I4(PR3)2,
R = Et (7)b
present 2.595(1)
work
2.685(1)
2.666(2)
2.660
2.608(1)
2.593(2)
2.593
2.263(1)
2.256(4)
2.248
90.19(3)
90.0(1)
90.3
85.39(2)
84.32(5)
84.8
94.61(2)
95.68(5)
94.8
trans-Pd2I4(PR3)2,
R = n-Buc
[18a]
2.581(2)
trans-Pd2I4(PR3)2,
R3 = Ph2(CHCH2)d
[18b]
2.583
trans-PdI2(PEt3)2 (8)e present
work
Molecule A
2.608(1)
2.610(1)
2.346(1)
2.343(1)
87.08(2),
92.92(2)
87.15(2),
92.85(2)
Molecule B
f
trans-PdI2(PEt3)2
Molecule A
[18c]
2.602(1)
2.601(1)
2.598(1)
2.603(1)
2.346(2)
2.342(2)
2.350(1)
2.343(2)
87.25(5),
92.75(5)
87.45(5),
92.55(5)
88.45(3),
91.55(3)
87.56(5),
92.32(5)
Molecule B
g
trans-PdI2(PPh3)2
[18d]
[18e]
trans-PdI2(PPh3)2 Æ
CHCl3
h
a
Atom-labeling for both trans-Pd2(l2-I)2I2(PEt3)2 (7) and trans-PdI2(PEt3)2 (8) is given in Fig. 3.
b
c
ꢀ
CCD X-ray data taken at 100(2) K; one crystallographically independent half-molecule with Ci ð1Þ site symmetry.
ꢀ
X-ray data taken at at room temperature; one crystallographically independent half-molecule with Ci ð1Þ site symmetry.
d
X-ray data taken at 290 K; two crystallographically independent molecules with C1 (1) site symmetry. The mean of four individual connectivities
is given for each different (bond/angle)-type.
e
ꢀ
CCD X-ray data taken at 100(2) K; two crystallographically independent half-molecules A and B with Ci ð1Þ site symmetry.
f
ꢀ
X-ray data taken at 293(2) K; two crystallographically independent half-molecules A and B with Ci ð1Þ site symmetry.
g
h
ꢀ
X-ray data taken at 160(2) K; one crystallographically independent half-molecule with Ci ð1Þ site symmetry.
X-ray data taken at 300 K; one crystallographically independent half-molecule with C2 (2) site symmetry.
3. Experimental
Pd10ðCOÞ ðPBun3Þ [6e] and was purified via recrystalli-
12
6
zation from C6H6/heptane. For preparation of trans-
Pd2I4(PEt3)2 (7) and trans-PdI2(PEt3)2 (8) we used the
published procedure [21] but with PdCl2 instead of hy-
drated Na2PdCl4. Other chemicals were purchased and
used without additional purification.
3.1. Materials and methods
Reactions were carried out via standard Schlenk tech-
niques on a preparative vacuum line under nitrogen
atmosphere. All solvents were deoxygenated prior to
use by the passing of N2 through them for at least
20 min at room temperature. [Pd6Tl(l2-CO)6L6]PF6
(2), [Pd12Tl2(CO)9L9][PF6]2 (3), and Pd4(CO)5L4 were
synthesized by known methods [9,10,20], and
Pd10(CO)12L6 (6) was prepared as described for
31P{1H}, 13C{1H} and 1H NMR spectra were ob-
tained for CDCl3 solutions under N2 atmosphere on a
Bruker AM-300 spectrometer. 31P{1H} spectra were ref-
erenced to 85% H3PO4 in D2O as an external standard.
13C chemical shifts were referenced to TMS via the 13C
1
signal of the solvent CDCl3 at 77.23 ppm. H chemical

1568
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
shifts were referenced to internal TMS. Each NMR sam-
ple was prepared by use of a nitrogen/vacuum line.
Infrared spectra were recorded on a Mattson Polaris
FT-IR spectrometer by use of a nitrogen-purged CaF2
cell. Nujol suspensions were prepared under nitrogen
as well.
but X-ray diffraction measurements revealed that all of
them possessed the same unit cell. One block crystal of
size 0.41 · 0.20 · 0.17 mm3 was used for the CCD X-
ray data collection.
IR spectra of 1, m(CO): nujol, 1810 (s) cmꢀ1; THF
solution, 1826 (s) cmꢀ1
.
Yields of crystals 1, 5, 6 were defined by their
mechanical separation under a microscope and/or
from an NMR 31P{1H} technique by use of an inner
standard, as described below. An integral intensity of
31P{1H} NMR spectrum of 1 (121 MHz): d = 7.4
ppm (s), ꢀ145.2 ppm (septet, 1J(P,F) = 715 Hz).
13C{1H} NMR spectrum of 1 (75.4 MHz): d = 215.7
ppm (s, 6C, CO), 68.2 ppm (s, ꢃ0.8C, CH2–O), 25.8
ppm (s, ꢃ0.8C, C–CH2–CH2–C), 18.1 ppm (complex
multiplet with three main signals at 18.242, 18.105 and
a
31P{1H} signal of one of the compounds (e.g., 1,
7, or 8) in an initial NMR spectrum of a mixture
of two or three crystalline species was used as a ref-
erence. Then a weighed amount of the crystals of
one of the defining compounds was added into the
NMR tube. The resulting spectrum was analyzed
via the increased intensity of the signal of the added
compound versus the unchanged reference signal; this
method provided a direct way to calculate the clus-
terꢀs yield. The signals at 7.4 ppm for 1 and 7.6
ppm for 8 were sufficiently resolved due to their nar-
row lines.
1
17.999 ppm, 18C, CH2–P), 8.7 ppm (s, 18C, CH3). H
NMR spectrum of 1 (300 MHz): d = 3.75 ppm (m,
ꢃ1.6H, CH2–O), 1.94 ppm (m, 36H, CH2–P), 1.85
ppm (m, ꢃ1.6H, C–CH2–CH2–C), 1.13 ppm (dt, 54H,
CH3, 3J(H,H) = 7.7 Hz; 3J(H,P) = 15.7 Hz). 31P{1H}
NMR spectrum of 7 (121 MHz): d = 42.4 ppm (s).
13C{1H} NMR spectrum of 7 (75.4 MHz): d = 21.5
1
ppm (d, 6C, CH2, J(C,P) = 32.8 Hz), 9.2 ppm (s, 6C,
CH3). 1H NMR spectrum of 7 (300 MHz): d = 2.23
3
2
ppm (dq, 12H, CH2, J(H,H) = 7.6 Hz, J(H,P) = 10.4
Hz), 1.22 ppm (dt, 18H, CH3, 3J(H,H) = 7.6 Hz,
3J(H,P) = 18.2 Hz). 31P{1H} NMR spectrum of 8 (121
MHz): d = 7.6 ppm (s). 13C{1H} NMR spectrum of 8
3.2. Synthesis of [Pd12(l3-I)2(l4-I)3(l2-CO)6L6]+, 1,
as [PF6]ꢀ salt (L = PEt3)
(75.4
MHz):
d = 20.2
ppm
(t,
6C,
CH2,
þ
3.2.1. Direct iodination of ðl6-TlÞ½Pd3ðCOÞ L3ꢂ
|1J(C,P) + 3J(C,P)| = 29.9 Hz), 9.1 ppm (s, 6C, CH3).
1H NMR spectrum of 8 (300 MHz): d = 2.28 ppm (qt,
3
2
sandwich ([PF6]ꢀ salt)
3
A solution of I2 (30 mg; 0.118 mmol) in 4 mL of
THF was added dropwise under stirring to a solution
of [TlPd6(CO)6L6]PF6 (0.100 g; 0.0536 mmol) in 3 mL
of THF. The solution color immediately changed to
red and then to brown-red. An IR spectrum of the
reaction solution taken after 30 min showed charac-
teristic carbonyl bands of 1, 1827 (m) cmꢀ1; 3, 1873
(m), 1845 (m) cmꢀ1; and an unassigned band at
1899 (m) cmꢀ1. After 1 h of stirring, an additional
amount of I2 (31 mg; 0.122 mmol) was added, which
resulted in an immediate formation of black precipi-
tate. After 2 h an IR spectrum of the reaction solu-
tion displayed only one strong carbonyl band at
1826 cmꢀ1 due to 1. The solution was filtrated from
Pd black, 17 mg; subsequent crystallization from the
THF solution in the presence of hexane gave 4 mg
of black-violet crystals of [Pd12(l3-I)2(l4-I)3(l2-
CO)6L6]PF6 (5%), 50 mg of trans-Pd2I4L2, 7 (32%)
and 32 mg of trans-PdI2L2, 8 (17%). These estimated
yields were defined by the previously described NMR
technique (Table 2, entry 1).
12H, CH2, J(H,H) = 7.8 Hz, |2J(H,P) + 4J(H,P)| = 6.6
Hz), 1.11 ppm (tt, 18H, CH3,3J(H,H) = 7.8 Hz,
|3J(H,P) + 5J(H,P)| = 33.6 Hz).
3.2.2. Conproportionation reactions
The following typical reaction is
a
detailed
description of entry 8 (Table 2). A mixture of
Pd10(CO)12L6 (0.100 g; 0.0474 mmol), trans-Pd2I4L2
(91 mg; 0.0951 mmol), Pd(OAc)2 (10.6 mg; 0.0472
mmol), and (NH4)(PF6) (15.5 mg; 0.0951 mmol)
was stirred in 10 mL of THF. The solution color
changed from red to black-violet within 10–15 min.
After 3 h the solvent was evaporated via N2 flow.
Extraction of the residue with Et2O resulted in the
separation of 60 mg of yellow crystals of 8. The
remaining residue was washed with water and dried
via vacuum under liquid N2, after which 1 was ex-
tracted from the residue by use of 15 mL of THF;
subsequent addition of ca. 12 mL of hexane and
keeping of the solution overnight at ꢀ20 ꢁC gave
63 mg (54%) of black-violet crystals of 1. The final
residue was Pd black, 21 mg. Both crystalline com-
pounds 1[PF6] and 8 were characterized spectroscop-
ically (Table 2, entry 8).
Reactions of 3 and 6 with iodine were carried out
analogously (Table 2, entries 2 and 3). From these reac-
tions single crystals of 1, 7, and 8 were separated
mechanically and used for multinuclear NMR and X-
ray diffraction studies. Particularly noteworthy is that
there were three types of black-violet single crystals of
1: namely, blocks, needles and hexagonal hollow tubes;
All other conproportionation reactions were per-
formed either analogously or under specific conditions
discussed in Section 2 and mentioned in the footnotes
of Table 2.

E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
1569
3.2.3. Reduction of Pd(OAc)2 by CO in the presence of
KI, PEt3 and (NH4)(PF6)
gen atoms were included in structure factor calculations
at idealized positions and were allowed to ride on at-
tached carbon atoms with relative isotropic displace-
ment coefficients. The displacement ellipsoids of the P
and F atoms were elongated along the c-direction, there-
by indicating dominant anisotropic thermal displace-
ment along the threefold axis. The final difference map
revealed around a 32 site two physically meaningful
residual peaks, Q5 and Q8, which could be attributed
either to O or C atoms of a solvated THF molecule dis-
ordered among six equivalent positions. This crystallo-
This reaction is a detailed description of entry 14 (Ta-
ble 2). To a solution of Pd(OAc)2 (0.224 g; 1.00 mmol) in
12 mL of acetone the following reagents were added
consecutively under stirring: HOAc (3 mL), (NH4)(PF6)
(55 mg; 0.34 mmol), PEt3 (73 lL; 0.50 mmol), KI (70
mg; 0.42 mmol) and H2O (3 mL). After 1 day of stirring
under CO atmosphere, 15 mL of H2O were added. The
resulting back precipitate was filtered, washed with
H2O, extracted with 90% EtOH, and then with Et2O.
The ethanol extract gave 4 mg of 1 (1.6%) and 18 mg
of 8. Additionally, 56 mg of 8 were obtained from the
Et2O extract, corresponding to a total yield of 12%.
Extraction of the remaining residue with THF led to
the formation of an extremely unstable brown solution.
The final residue, 70 mg, was Pd black.
1
graphic assignment is completely consistent with a H
NMR spectrum of single crystals dissolved in CDCl3,
which exhibited a proton resonance at 3.75 ppm charac-
teristic of the –CH2–O–CH2– part of the C4H8O mole-
cule. An estimated 1/THF ratio of 1/0.4 (that roughly
corresponds one-half of a solvated molecule) was found
upon the comparison of the integral intensity for this
signal with those for the methylene and methyl signals
of the PEt3 ligands of 1. Because approximately one-half
of a solvated THF molecule in the independent unit dis-
ordered over six equivalent orientations would give rise
to only 1/12th the normal electron density of C or O
atoms at each atomic position, this presumed crystal dis-
order was not modeled.
3.3. X-ray crystallographic determinations
3.3.1. General procedures
X-ray data for crystals of [1]PF6, 7, and 8 were col-
lected at 100(2) K via a Bruker SMART CCD-1000 area
detector diffractometer with graphite-monochromated
˚
Mo Ka radiation (k = 0.71073 A) from a sealed-tube
generator. The crystal structures were determined from
direct methods. Full-matrix least-squares refinements
(based on F2) were carried out with SHELXTL [22].
CCDC reference numbers 246935 (1), 246936 (7), and
246937 (8).
3.3.3. trans-(PEt3)2Pd2I4
M = 956.70, P21/n, a = 8.9860(12) A, b = 14.2391(19) A,
˚
˚
˚
c = 9.3548(13)
A,
a = c = 90ꢁ,
V = 1196.3(3) A , Z = 2; F(000) = 872; Dcalc = 2.656
Mg Æ mꢀ3
10869 reflections were obtained over
5.22 6 h 6 56.62ꢁ. Empirical absorption correction
(
SADABS) was applied [l(Mo-Ka) = 6.790 mmꢀ1, maxi-
b = 91.927(2)ꢁ,
3
˚
.
3.3.2. [Pd12I5(CO)6(PEt3)6][PF6]
ꢀ
˚
M = 2933.2, P31C, a = b = 17.4050(7), c = 15.8014(9) A,
3
˚
a = b = 90ꢁ, c = 120ꢁ, V = 4145.5(3) A , Z = 2;
mum/minimum transmission, 0.3166/0.1720]. Refine-
ment of 2945 independent merged reflections
[Rint = 0.0302] with 91 parameters (0 restraints) con-
verged at wR2(F2) = 0.0678 for all data; R1(F) = 0.0262
F(000) = 2732; Dcalc = 2.350 Mg Æ mꢀ3. 33490 reflec-
tions were obtained over 5.40 6 h 6 52.72ꢁ. Empirical
absorption correction (SADABS) was applied [l(Mo-
Ka) = 4.578 mmꢀ1, maximum/minimum transmission,
0.5100/0.2555]. Refinement of 2836 independent merged
reflections [Rint = 0.0350] with 89 parameters (13 re-
straints) converged at wR2(F2) = 0.1772 for all data;
R1(F) = 0.0511 for I > 2r(I); maximum/minimum resid-
for I > 2r(I); maximum/miniꢀm3um residual electron den-
˚
sity, 1.200/ꢀ0.662 e Æ A
;
goodness-of-fit (on
F2) = 0.942. Crystal size: 0.40 · 0.29 · 0.22 mm3. All
non-hydrogen atoms were refined anisotropically.
Hydrogen atoms were included as described above.
The trans-(PEt3)2Pd2I4 complex was also crystallized in
ual electron density, 1.47/ꢀ1.06 e Æ Aꢀ3; goodness-of-fit
˚
2
˚
ꢀ
(on F ) = 1.127. Under P31c symmetry with Z = 2, the
the orthorhombic crystal system with a = 6.995(2) A,
b = 14.272(3) A, c = 23.730(5) A, a = b = c = 90ꢁ,
two monocations (1) and two [PF6]ꢀ anions each possess
32 site symmetry. The crystallographically independent
unit of 1 consists of one Pd(oc) and one Pd(cap) in 12-
fold general positions, one l4-I on a horizontal twofold
axis, one l3-I on the threefold axis, and one carbonyl C
and O atoms and six ethyl carbon atoms in 12-fold gen-
eral positions; the independent part of the two anions
has one P atom with 32 point symmetry and one F atom
in 12-fold general positions. All non-hydrogen atoms,
except for the ethyl carbon atoms, were refined aniso-
tropically. Most carbon atoms of the ethyl groups
showed multiple disorder, which did not model. Hydro-
˚
˚
3
V = 2369.1(8) A , Z = 4.
˚
3.3.4. trans-(PEt3)2PdI2
M = 596.50, P21/c, a = 17.6876(19) A, b = 7.8987(8) A,
˚
˚
˚
c = 15.5429(17) A, a = c = 90ꢁ, b = 114.4770(10)ꢁ,
V = 1976.3(4) A , Z = 4; F(000) = 1136; Dcalc = 2.005
3
˚
Mg Æ mꢀ3
.
4872 reflections were obtained over
2.52 6 h 6 56.56ꢁ. Empirical absorption correction
SADABS) was applied [l(Mo-Ka) = 4.209 mmꢀ1, maxi-
(
mum/minimum transmission, 0.4074/0.3050]. Refine-
ment of 4872 independent merged reflections

1570
E.G. Mednikov, L.F. Dahl / Inorganica Chimica Acta 358 (2005) 1557–1570
(b) S.M. Godfrey, D.G. Kelly, C.A. McAuliffe, A.G. Mackie,
R.G. Pritchard, S.M. Watson, J. Chem. Soc., Chem. Commun.
(1991) 1163;
[Rint = 0.0297] with 163 parameters (0 restraints) con-
verged at wR2(F2) = 0.0671 for all data; R1(F) = 0.0251
for I > 2r(I); maximum/minimum residual electron
(c) N. Bricklebank, S.M. Godfrey, H.P. Lane, C.A. McAuliffe,
R.G. Pritchard, J.-M. Moreno, J. Chem. Soc., Dalton Trans.
(1995) 2421;
ꢀ3
˚
density, 1.114/ꢀ0.766 e Æ A
;
goodness-of-fit (on
F2) = 1.086. Crystal size: 0.37 · 0.30 · 0.26 mm3. All
non-hydrogen atoms were refined anisotropically.
Hydrogen atoms were included as described for 1.
(d) S.M. Godfrey, C.A. McAuliffe, R.G. Pritchard, J.M. Sheffield,
J. Chem. Soc., Dalton Trans. (1998) 1919.
[8] (a) F.A. Cotton, P.A. Kibala, J. Am. Chem. Soc. 109 (1987)
3308;
(b) W.I. Cross, S.M. Godfrey, C.A. McAuliffe, R.G. Pritchard,
J.M. Sheffield, G.M. Thompson, J. Chem. Soc., Dalton Trans.
(1999) 2795.
Acknowledgments
[9] E.G. Mednikov, L.F. Dahl, J. Chem. Soc., Dalton Trans. (2003)
3117.
This research was supported by the National Science
Foundation. The CCD area detector system was pur-
chased in part in 1995 with a NSF Grant (CHE-
9310428). Color drawings were prepared with Crystal
Maker, Interactive Crystallography. David C. Palmer
(P.O. Box 183 Bicester, Oxfordshire, UK OXG TBS).
[10] S.A. Ivanov, R.V. Nichiporuk, E.G. Mednikov, L.F. Dahl, J.
Chem. Soc., Dalton Trans. (2002) 4116.
[11] E.G. Mednikov, Yu.T. Struchkov, Yu.L. Slovokhotov, J. Orga-
nomet. Chem. 566 (1998) 15.
[12] M. Kawano, J.W. Bacon, C.F. Campana, B.E. Winger, J.D.
Dudeck, S.A. Sirchio, S.L. Scruggs, U. Geiser, L.F. Dahl, Inorg.
Chem. 40 (2001) 2554.
[13] (a) E.G. Mednikov, Ph.D. Thesis, Moscow State University,
1983;
References
(b) D.M.P. Mingos, C.M. Hill, Croat. Chim. Acta 68 (1995) 745.
[14] (a) E.G. Mednikov, N.K. Eremenko, Izv. Acad. Nauk SSSR. Ser.
Khim. (1984) 2781, [Bull. Acad. Sc. USSR, Div. Chem. Sc. 33
(1984) 2547 (Engl. Transl.)];
[1] (a) L. Manojlovic-Muir, K.W. Muir, B.R. Lloyd, R.J. Puddeph-
att, J. Chem. Soc., Chem. Commun. (1985) 536;
(b) P.D. Harvey, R. Provencher, J. Gagnon, T. Zhang, D. Fortin,
K. Hierso, M. Drouin, S.M. Socol, Can. J. Chem. 74 (1996) 2268;
(c) B.R. Lloyd, L. Manojlovic-Muir, K.W. Muir, R.J. Puddeph-
att, Organometallics 12 (1993) 1231;
(b) E.G. Mednikov, N.K. Eremenko, G.A. Hivintseva, Koordin.
Khim. (Sov. J. Coord. Chem.) 13 (1987) 106 (in Russian);
(c) E.G. Mednikov, N.I. Kanteeva, Izv. Acad. Nauk. Ser. Khim.
(1995) 167, [Russ. Chem. Bull. 44 (1995) 163 (Engl. Transl.)] and
references therein.
(d) A.L. Balch, B.J. Davis, M.M. Olmstead, Inorg. Chem. 32
(1993) 3937;
[15] (a) S.M. Godfrey, D.G. Kelly, A.G. Mackie, P.P. Mac Rory,
C.A. McAuliffe, R.G. Pritchard, S.M. Watson, J. Chem. Soc.,
Chem. Commun. (1991) 1447;
(e) T. Zhang, M. Drouin, P.D. Harvey, J. Chem. Soc., Chem.
Commun. (1996) 877;
(f) K. Sugimoto, T. Kuroda-Sowa, S.-G. Yan, M. Maekawa, M.
Munakata, Acta Crystallogr., Sect. C 56 (2000) 414.
[2] R.J. Puddephatt, L. Manojlovic-Muir, K.W. Muir, Polyhedron 9
(1990) 2767, and references therein.
(b) C.A. McAuliffe, S.M. Godfrey, A.G. Mackie, R.G. Pritchard,
Angew. Chem. Int. Ed. 31 (1992) 919.
[16] D.A. Redfield, J.H. Nelson, L.W. Cary, Inorg. Nucl. Chem. Lett.
10 (1974) 727, and references therein.
[3] P.D. Harvey, Y. Mugnier, D. Lucas, D. Evrard, F. Lemaitre, A.
Vallat, J. Cluster Sci. 15 (2004) 63, and references therein.
[4] R. Vilar, S.E. Lawrence, S. Menzer, D.M.P. Mingos, D.J.
Williams, J. Chem. Soc., Dalton Trans. (1997) 3305.
[5] (a) E.G. Mednikov, Metalloorgan. Khim. 2 (1989) 317, [Organo-
met. Chem. in the USSR 2 (1989) 148 (Engl. Transl.)];
(b) E.G. Mednikov, E.A. Grebenschikova, N.K. Eremenko, Izv.
Acad. Nauk SSSR. Ser. Khim. (1986) 2155, [Bull. Acad. Sc.
USSR, Div. Chem. Sc. 35 (1986) 1968 (Engl. Transl.)];
(c) E.G. Mednikov, N.K. Eremenko, Yu.L. Slovokhotov, Yu.T.
Struchkov, Metalloorgan. Khim. 2 (1989) 1289, [Organomet.
Chem. in the USSR 2 (1989) 680 (Engl. Transl.)].
[17] (a) J.E. Fergusson, P.F. Heveldt, Inorg. Chim. Acta 31 (1978)
145;
(b) J.A. Rahn, M.S. Holt, M. OꢀNeil-Johnson, J.H. Nelson,
Inorg. Chem. 27 (1988) 1316.
[18] (a) A.D. Ryabov, L.G. Kuzꢀmina, V.A. Polyakov, G.M. Kazan-
kov, E.S. Ryabova, M. Pfeffer, R. van Eldik, J. Chem. Soc.,
Dalton Trans. (1995) 999;
(b) N.W. Alcock, W.L. Wilson, J.H. Nelson, Inorg. Chem. 32
(1993) 3193;
(c) R. Fra¨nkel, J. Kniczek, W. Ponikwar, H. No¨th, K. Polborn,
W.P. Fehlhammer, Inorg. Chim. Acta 312 (2001) 23;
(d) G.W.V. Cave, W. Errington, J.P. Rourke, Acta Crystallogr.,
Sect. C 55 (1999) 320;
[6] (a) R.D. Feltham, G. Elbaze, R. Ortega, C. Eck, J. Dubrawski,
Inorg. Chem. 24 (1985) 1503;
(e) M. Kubota, S. Ohba, Y. Saito, Acta Crystallogr., Sect. C 47
(1991) 1727.
(b) E.G. Mednikov, N.K. Eremenko, Yu.L. Slovokhotov, Yu.T.
Struchkov, S.P. Gubin, Koordin. Khim. (Sov. J. Coord. Chem.)
13 (1987) 979 (in Russian);
[19] T.G. Appleton, H.C. Clark, L.E. Manzer, Coord. Chem. Rev. 10
(1973) 335.
(c) J. Dubrawski, K.C. Kriege-Simondsen, R.D. Feltham, J. Am.
Chem. Soc. 102 (1980) 2089;
[20] E.G. Mednikov, N.K. Eremenko, Izv. Acad. Nauk SSSR. Ser.
Khim. (1981) 2400, [Bull. Acad. Sc. USSR, Div. Chem. Sc. 30
(1981) 1980 (Engl. Transl.)].
(d) E.G. Mednikov, N.K. Eremenko, S.P. Gubin, Yu.L. Slov-
okhotov, Yu.T. Struchkov, J. Organomet. Chem. 239 (1982) 401;
(e) E.G. Mednikov, N.K. Eremenko, Izv. Acad. Nauk SSSR. Ser.
Khim. 11 (1982) 2540, [Bull. Acad. Sc. USSR, Div. Chem. Sc. 31
(1982) 2240 (Engl. Transl.)].
[21] J. Chatt, L.M. Venanzi, J. Chem. Soc. (A) (1957) 2351.
[22] G. Sheldrick: all crystallographic software and sources of the
scattering factors are contained in the SHELXTL (version 5.1)
program library, Bruker Analytical X-ray Systems, Madison,
WI.
[7] (a) W.-W. du Mont, M. Ba¨tcher, S. Pohl, W. Saak, Angew.
Chem. Int. Ed. 26 (1987) 912;