Gold and silver cations in the procrustean bed of the bis[2-(diphenylphosphino)phenyl]phenylphosphine ligand. Observations and conclusions ?

Article abstract of DOI:10.1039/a808266i

Reaction of [Au(Me₂S)Cl] or AgCl with equimolar quantities of bis[2-(diphenylphosphino)phenyl]phenylphosphine (TP) gave isomorphous crystalline 1:1 complexes with a greatly distorted tetrahedral environment for Au^I and Ag^I. The large differences in the metal-to-phosphorus interactions are also obvious from the NMR parameters in solution. The complex [Au(TP)Cl] is converted into the tetrafluoroborate [Au₂(TP)₂][BF₄]₂ when treated with AgBF₄. This ionic compound has a dinuclear dication with grossly different environments for the two gold atoms: one is trigonally planar three-co-ordinated with a short intramolecular Au...Au bond [2.8776(4) A]. The second gold atom is five-co-ordinated with the other gold atom as an equatorial contact. By contrast, the dicationic part [Ag₂(TP)₂]²⁺ of the product obtained by reaction of TP with AgBF₄ has equivalent silver atoms (related by a center of inversion) which are in a tetrahedral environment including a very short intramolecular Ag...Ag distance [2.8569(8) A]. The temperature dependence of the NMR spectra of [M₂(TP)₂][BF₄]₂ (M = Au or Ag) reflects the fluxional behaviour of these highly strained units in solution.

Full text of DOI:10.1039/a808266i

View Article Online / Journal Homepage / Table of Contents for this issue
Gold and silver cations in the “Procrustean Bed” of the
bis[2-(diphenylphosphino)phenyl]phenylphosphine ligand.
Observations and conclusions†
Johann Zank, Annette Schier and Hubert Schmidbaur*
Anorganisch-chemisches Institut der Technischen Universität München, Lichtenbergstrasse 4,
D-85747 Garching, Germany
Received 26th October 1998, Accepted 8th December 1998
Reaction of [Au(Me2S)Cl] or AgCl with equimolar quantities of bis[2-(diphenylphosphino)phenyl]phenylphosphine
(TP) gave isomorphous crystalline 1:1 complexes with a greatly distorted tetrahedral environment for AuI and AgI.
The large differences in the metal-to-phosphorus interactions are also obvious from the NMR parameters in
solution. The complex [Au(TP)Cl] is converted into the tetrafluoroborate [Au2(TP)2][BF4]2 when treated with AgBF4.
This ionic compound has a dinuclear dication with grossly different environments for the two gold atoms: one is
trigonally planar three-co-ordinated with a short intramolecular Au ؒ ؒ ؒ Au bond [2.8776(4) Å]. The second gold
atom is five-co-ordinated with the other gold atom as an equatorial contact. By contrast, the dicationic part
[Ag2(TP)2]2ϩ of the product obtained by reaction of TP with AgBF4 has equivalent silver atoms (related by a center
of inversion) which are in a tetrahedral environment including a very short intramolecular Ag ؒ ؒ ؒ Ag distance
[2.8569(8) Å]. The temperature dependence of the NMR spectra of [M2(TP)2][BF4]2 (M = Au or Ag) reflects the
fluxional behaviour of these highly strained units in solution.
Now we report the results of our studies with univalent gold
Introduction
and silver as their chloride and tetrafluoroborate salts. Very
Most of the current research in co-ordination chemistry is gen-
unusual co-ordination geometries involving intriguing metal–
erally focused on complexes with “tailor-made” ligands, which
metal interactions have been observed, that are generated out
of the dilemma that the individual preferences of the metals
and the ligands cannot be reconciled in the given system.
perfectly fit a given metal or its clusters.1–5 In standard cases this
“fit” is beneficial to the stability of the complex and may help
to generate certain chemical and physical properties.6 Much
less interest has been paid to cases where a ligand does not
Results and discussion
The TP complexes of AuCl and AgCl
fit a metal at all, although this may frequently generate an
“active centre” owing to an “open” geometry and a much lower
stability of the system.7 The situation there is reminiscent of
a chapter in ancient Greek mythology, where the monster Pro-
crustes made his victims (metals) “fit” to his bed (ligand),
the infamous “Procrustean Bed”, by very cruel methods.8
The ligand geometry is most crucial for the stability of a
complex if, on the one hand, the metal is rather inflexible in its
preferences for a certain co-ordination number and geometry.
This is true in particular for metals with low oxidation states,
low co-ordination numbers, and high atomic numbers associ-
ated with strong relativistic effects.9–13 Gold() is a prototype for
this group of metals, and it comes as no surprise that linear
two-co-ordination is by far the dominant mode for this heaviest
of the coinage metals.14 For gold() this geometry is rarely real-
ized with bidentate ligands, because this would require exceed-
ingly long loops between the donor atoms, especially so if the
ligand does not tolerate any bending of its framework.15,16
Bis[2-(diphenylphosphino)phenyl]phenylphosphine (TP), on
the other hand, is a very inflexible ligand with a rigid backbone
based solely on arene substituents with very little freedom for
internal molecular motion.17,18 Only a few rotational move-
ments are possible and simple models clearly show the groups
to be always in each others way. Recently we have demonstrated
for TP complexes of gallium and indium that there is virtually
no strain-free arrangement for the TP tripod to co-ordinate
with all three donor centres to a tetrahedral metal centre.19
Treatment of chloro(dimethyl sulfide)gold() with an equimolar
quantity of TP in dichloromethane at room temperature gives
a yellow solution, from which a yellow crystalline product [mp
332 ЊC (decomp.)] can be precipitated in 84% yield by addition
of diethyl ether. The silver analogue is prepared similarly from
silver chloride and TP in the same solvent as green crystals in
68% yield [mp 343 ЊC (decomp.)]. Both products give satisfac-
tory elemental analysis data. The mass spectra (CI) show the
molecular ions of the monomers in low intensity and with the
1
correct isotope patterns. The H NMR spectra contain only
complex multiplets of C6H5/C6H4 resonances which have not
been analysed any further.
Cl
M
Ph2PA
PAPh2
PB
Ph
M = Ag, Au
1/2
The 31P-{1H} NMR spectra show A2B spin systems indicat-
ing virtual equivalence of the terminal Ph2P groups in solution
at ambient temperature. There is rapid site exchange of the
metal atom in [Ag(TP)Cl] as indicated by the absence of Ag–P
coupling in the room temperature spectrum of the compound
(Fig. 1). Upon cooling the sample to Ϫ70 ЊC the A2B pattern is
† Dedicated to Professor Ernst Otto Fischer on the occasion of his 80th
birthday.
J. Chem. Soc., Dalton Trans., 1999, 415–420
415

View Article Online
Fig. 2 Packing of the isomorphous compounds [Au(TP)Cl] 1 and
[Ag(TP)Cl] 2 in the crystal lattice.
Fig. 1 31P-{1H} NMR spectra of [Ag(TP)Cl] in CD2Cl2 at 20 (a) and
Ϫ70 ЊC (b).
turned into two closely spaced A2BX spin systems with
X = 107,109Ag: J(107Ag–P) = 280, J(109Ag–P) = 317 Hz for PA
and J(107,109Ag–P) = 123 Hz (average) for PB, while the J(P,P)
coupling constant and the chemical shifts show only minor
variations with temperature (δA Ϫ6.6, δB Ϫ24.6, J = 204 Hz).
For the spectrum of the gold compound no significant temper-
ature-dependence is observed. The two compounds are photo-
luminescent showing a yellow emission at λmax 580 (1) and 503
nm (2), respectively, upon UV excitation at 258 nm. The phos-
phorescence is stronger for M = Ag than for M = Au.
Single crystals of yellow [Au(TP)Cl] and green [Ag(TP)Cl]
were grown by slow diffusion of diethyl ether into solutions in
dichloromethane at 20 ЊC. The crystals are isomorphous
(monoclinic, space group P21/n, Z = 4) with only minor vari-
ations in the cell dimensions. The lattices are composed of
mononuclear complex molecules with no significant sub-van
der Waals contacts (Fig. 2). The individual molecules have no
crystallographically imposed symmetry. The maximum attain-
able (mirror) symmetry is violated mainly by the relative orien-
tation of the phenyl groups (Fig. 3).
The metal atoms are in the centre of a distorted tetrahedral
environment of one chlorine and three phosphorus atoms. The
distortions are obvious from the spread of the angles P–M–P
and P–M–Cl, which range from 85.40(3) to 121.37(3)Њ for
M = Au and from 81.03(2) to 122.98(2)Њ for M = Ag. The two
smallest P–M–P angles (P1–M–P2/P3) are endocyclic angles of
the two fused five-membered rings and require large angles P2–
M–P3 not spanned by the ligand.
The M–P distances average 2.3834 Å for M = Au and 2.5123
Å for M = Ag, making silver the larger atom as compared to
gold. Again, this result is in excellent agreement with recent
redeterminations of the ionic radii of these two elements in
other pairs of isomorphous species.21 In this context it is worth
noting that the volume of the unit cell is greater for the silver
than for the gold compound (Table 1).
Fig. 3 Molecular structure of compound [Au(TP)Cl] 1 (ORTEP20
drawing with 50% probability ellipsoids, H atoms omitted for clarity).
The structure of the silver compound 2 is very similar. Selected bond
lengths (Å) and angles (Њ) (the corresponding values of the isomorph-
ous silver compound 2 are given in square brackets): Au–Cl 2.5119(10)
[2.4663(5)], Au–P1 2.4608(9) [2.5469(5)], Au–P2 2.3206(10) [2.5049(5)]
and Au–P3 2.3687(10) [2.4852(5)]; P1–Au–Cl 118.17(3) [123.09(2)], P2–
Au–Cl 120.68(3) [121.60(2)], P3–Au–Cl 114.00(3) [122.98(2)], P1–Au–
P2 86.77(3) [81.03(2)], P1–Au–P3 85.40(3) [82.32(2)] and P2–Au–P3
121.37(3) [111.57(2)].
for M = Ag (0.06 Å). These data reflect the weaker acceptor
properties of two-co-ordinated gold. This effect is even more
obvious from the two M–Cl distances: Au–Cl 2.5119(10) is
found longer than Ag–Cl 2.4663(5) Å, which is clear evidence
that the chloride anion is drawn towards the gold atom less
strongly than towards the silver atom. This conclusion is even
more obvious since gold is the smaller atom of the two. With
the three M–P bonds longer and the M–Cl bond shorter for
M = Ag, the volume required by the silver compound is still
larger than that needed for the gold compound.
The molecular structure, shown in Fig. 3, indicates that the
distortions of the tetrahedral structure in the complexes
[M(TP)Cl] are mainly due to the restraints imposed by the lig-
and, which reduce two of the three P–M–P angles to values well
below 90Њ. The chloride anion is filling the resulting large gap in
the co-ordination sphere of the metal in a symmetrical way with
all Cl–M–P angles slightly above 120Њ. It is thus clear that in the
absence of this auxiliary ligand (ClϪ) serious consequences for
the stability of the complexes are to be expected.
The distance between the central phosphorus atom of the
ligand (P1 in both structures) and the metal M is larger than the
distance between M and the two terminal phosphorus atoms.
The difference ∆r is more significant for M = Au (0.14 Å) than
The TP complexes of AuBF4 and AgBF4
Treatment of a solution of [Au(TP)Cl] in dichloromethane with
416
J. Chem. Soc., Dalton Trans., 1999, 415–420

View Article Online
a solution of equimolar quantities of silver tetrafluoroborate
in methanol at room temperature leads to a precipitation of
silver chloride. From the supernatant orange solution the prod-
uct can be isolated after evaporation of the filtrate to dryness
and recrystallized from dichloromethane–diethyl ether [orange
crystals, mp 310 ЊC (decomp.), yield 87%]. A suspension of
AgBF4 in dichloromethane becomes clear after addition of
equimolar quantities of TP and stirring of the mixture in the
dark for 4 d. The crude product, isolated after removal of the
solvent, can also be recrystallized from dichloromethane–
diethyl ether [colourless crystals, mp 215 ЊC (decomp.), yield
74%].
The two tetrafluoroborate salts give satisfactory elemental
analysis data after careful drying in a vacuum. Single crystals of
[Au2(TP)2][BF4]2 3 contain two molecules of dichloromethane;
[Ag2(TP)2][BF4]2 4 crystallizes free of solvent molecules. Both
compounds are photoluminescent showing yellow emissions at
λmax 554 nm for M = Au and λmax 460 nm for M = Ag upon UV
excitation at 259 nm, the intensity being higher for the silver
compound.
The mass spectra (FAB) of the compounds show the uni- and
di-positive cations with the expected isotope distributions. The
1H NMR spectra (in CD2Cl2 at 20 ЊC) contain only extremely
complex multiplet resonances in the arene region, which could
not be resolved. The 31P-{1H} NMR spectrum of compound 3,
in CD2Cl2 at 20 ЊC, exhibits a broad peak at δ 26.2 (2 P) and a
pseudo-quintet signal at δ 16.4 (1 P). At Ϫ60 ЊC the broad
hump at δ 26.2 is split into two resonances, one of which is a
virtual triplet at δ 33.8 (1 P), while the other is a multiplet
centred at δ 19.4 (1 P) (Fig. 4). The third resonance at δ 15.6
(1 P) is now part of a complex multiplet together with a new
resonance at δ 19.4. At the high temperature end of the range
investigated (ϩ125 ЊC, in o-dichlorobenzene–DMSO-d6) the
hump of the 20 ЊC spectrum (δ 26.2) is sharpened to a virtual
triplet at δ 25.6 (2 P), while the former quintet (at 20 ЊC) is
largely retained, but broadened at δ 16.9 (1 P).
Considering the solid state structure (see below and Scheme
1), it is possible to assign this spectrum and its changes as a
function of temperature as follows: at the low temperature limit
the six phosphorus atoms exist as three pairs of mutually chem-
ically equivalent atoms, which owing to strong coupling and
concomitant magnetic inequivalence generate an AAЈBBЈCCЈ
spin system. The signal at lowest field (δ 33.8), well separated
from the other two, is assigned to the pair of P atoms (P2 and
P6) at the low-co-ordinate gold centre (Au1, Fig. 5). Owing to
the particularly strong coupling A–AЈ for trans positions at
low-co-ordinate gold() this resonance has the pronounced
coupling pattern of a virtual triplet [J(A–AЈ) Ϫ J(A–B) = ca.
232 Hz]. The upfield resonance (δ 15.6) must be assigned to the
central P atoms (B,BЈ; P1 and P4 in Fig. 5), because this signal
remains largely unaffected in its chemical shift during the
coalescence occurring as the temperature is increased. At low
temperature it represents the complex BBЈ part of the spectrum
with extensive coupling to A,AЈ,C,CЈ and only a small shift
difference relative to C/CЈ (complex multiplet at δ 19.4). It
should be noted that the assignments proposed for [Au2(TP)2]-
[BF4]2 are in full agreement with published data for [(AuCl)3-
(TP)]: this molecule also has the resonance of the central P
atom at highest field (δ 18.5).18
As the temperature is raised the A/AЈ and C/CЈ resonances
collapse at δ 25.6. This result indicates that there is an intra-
molecular “toothbrush movement” of the two ligands at the
metal “teeth” which renders the two peripheral phosphorus
atoms equivalent on the NMR timescale (Scheme 1). By the
same token the two gold atoms become equivalent, but
unfortunately there is no way to observe this consequence dir-
ectly in solution. Clearly, the dynamic process does not involve
a complete dissociation of a ligand from the gold atoms,
because some virtual P,P coupling across the metal atoms is
retained. This effect causes the pseudo-quintet structure of the
Fig. 4 31P-{1H} NMR spectra of [Au2(TP)2][BF4]2 3 in CD2Cl2 at
Ϫ60 ЊC (a), 20 ЊC (b) and in 1,2-dichlorobenzene at 125 ЊC (c).
2+
2+
Ph
PB
Ph
PB
Ph2PA
Ph2PA
PCPh2
PCPh2
Ph2PA
Ph2PA
PCPh2
PCPh2
Au
Au
Au
Au
PB
Ph
PB
Ph
3
2+
2+
Ph
PB
Ph
PB
PCPh2
Ph2PA
PCPh2
Ag
Ph2PA
Ph2PC
Ag
PAPh2
Ag
Ag
Ph2PC
PB
Ph
PB
Ph
PAPh2
4
Scheme 1 Representation of the co-ordination in the dications
of complexes [Au2(TP)2][BF4]2 3 and [Ag2(TP)2][BF4]2 4 and their
fluxionality.
J. Chem. Soc., Dalton Trans., 1999, 415–420
417

View Article Online
Fig. 5 Molecular structure of the dication of compound 3 (ORTEP
drawing with 50% probability ellipsoids, H atoms omitted for clarity).
Selected bond lengths (Å) and angles (Њ): Au1 ؒ ؒ ؒ Au2 2.8776(4),
Au1–P2 2.325(2), Au1–P6 2.329(2), Au2–P1 2.367(2), Au2–P3 2.468(2),
Au2–P4 2.363(2) and Au2–P5 2.469(2); P2–Au1–P6 147.43(6),
P2–Au1 ؒ ؒ ؒ Au2 106.11(4), P6–Au1 ؒ ؒ ؒ Au2 106.44(4), P1–Au2–P3
82.87(6), P1–Au2–P4 150.74(6), P1–Au2–P5 114.61(6), P3–Au2–P4
115.05(6), P3–Au2–P5 107.35(6), P4–Au2–P5 83.26(6), P1–Au2 ؒ ؒ ؒ Au1
75.52(4), P3–Au2 ؒ ؒ ؒ Au1 126.56(4), P4–Au2 ؒ ؒ ؒ Au1 75.22(4) and P5–
Au2 ؒ ؒ ؒ Au1 126.09(4).
Fig. 6 Molecular structure of compound 3 projected along the Au–Au
axis revealing the close approximation of the structure to the symmetry
of a twofold axis.
high-field resonance at room temperature and above. This sig-
nal would have to be reduced to a triplet if the phenomenon
were caused by dissociation and concomitant intermolecular
exchange (scrambling) of ‘free’ and complexed ligands.
The 31P-{1H} NMR spectrum of the silver compound 4 in
CD2Cl2 at 20 ЊC shows only two resonances: a broad pseudo-
quartet at δ 0.6 (2 P, J = ca. 180 Hz) and a complex, symmetrical
multiplet at δ Ϫ14.8 (1 P). At Ϫ75 ЊC the stronger signal (2 P) is
also split and the multiplicity becomes more complex. This
observation requires a ground-state structure with inequivalent
phosphorus atoms (A, B, C) (Scheme 1), which is borne out by
the crystal structure of the compound (Fig. 7). The temperature
dependence of the spectra is due to a reduction of the coupling
of the phosphorus and silver nuclei with increasing temperature
(31P, 107/109Ag). Owing to partial (but not complete) dissociation
of Ag–P co-ordination as the central P atoms (PB) start to oscil-
late between the two silver atoms, the terminal P atoms are
rendered equivalent (AC collapse to A2).
Fig. 7 Molecular structure of the dication of compound 4 (ORTEP
drawing with 50% probability ellipsoids, H atoms omitted for clarity).
Selected bond lengths (Å) and angles (Њ): Ag ؒ ؒ ؒ AgЈ 2.8569(8), Ag–P1
2.5439(12), Ag–P2Ј 2.4301(12), Ag–P3 2.4997(12); P1–Ag–P2Ј
128.79(4), P1–Ag–P3 81.30(4), P2Ј–Ag–P3 116.49(4), P1–Ag ؒ ؒ ؒ AgЈ
73.22(3), P2Ј–Ag ؒ ؒ ؒ AgЈ 100.58(3), P3–Ag ؒ ؒ ؒ AgЈ 142.84(3).
Crystals of the gold compound [Au2(TP)2][BF4]2 3 (from
CH2Cl2–Et2O; monoclinic, space group P21/c, Z = 4) contain
two molecules of dichloromethane solvent per formula unit.
The lattice is composed of dinuclear dications [Au2(TP)2]2ϩ and
structures and the packing of molecules and ions in the solid
state.14 However, the present example is an extreme case of a
trigonally planar three-co-ordinated gold() centre where a
second gold atom appears as a true third (ligating) component.
Gold atom Au2, on the other hand, is five-co-ordinated and
has this gold–gold contact as the fifth ligating interaction which
is part of a distorted trigonal-bipyramidal environment. Both
ligands are chelating Au2 with their central (P1, P4) and
remaining peripheral phosphorus atoms (P3, P5), which form a
distorted tetrahedron. Atom Au1 is inserted over one of the
tetrahedral edges (P1–P4), the angle P1–Au2–P4 is widened to
150.74(6)Њ and a contact Au1 ؒ ؒ ؒ Au2 = 2.8776(4) Å is installed
on a line exactly bisecting this angle: P1–Au2–Au1 75.52(4) and
P4–Au2–Au1 75.22(4)Њ. To the best of our knowledge this is the
first case not only of five-co-ordination at AuI, but also of
aurophilic bonding at a high-co-ordinate metal centre.
Single crystals of the silver compound [Ag2(TP)2][BF4] 4,
(from CH2Cl2–Et2O, monoclinic, space group C2/c, Z = 8) are
not isomorphous with the gold analogue and contain no sol-
vent. The lattice is composed of tetrafluoroborate anions and
binuclear dications which enclose a crystallographic centre of
inversion (Fig. 7). Each of the two equivalent silver atoms is
chelated by two phosphorus atoms of one ligand (P1, P3) and
further co-ordinated to one phosphorus atom (P2) of the
Ϫ
two BF4 anions, with no significant sub-van der Waals con-
tacts between these components. The anions and the solvent
molecules have standard geometries. The dication (Fig. 5) has
no crystallographically imposed symmetry, but a projection
along the axis connecting the two gold atoms reveals a close
approximation of the molecular structure to the symmetry of a
twofold axis (Fig. 6). The two ligands are therefore mutually
equivalent. An inspection of the projection onto the Au–Au
axis shows, however, that the two metal atoms are in completely
different environments, both of which are highly unusual and
merit a detailed discussion.
Gold atom Au1, on the one hand, is co-ordinated to one
peripheral phosphorus atom (P2, P6) of each TP ligand, but the
axis P2–Au1–P6 is strongly bent [147.43(6)Њ] owing to an
approach of the second gold atom (Au2) to an Au1 ؒ ؒ ؒ Au2
contact distance as short as 2.8776(4) Å. This distance must
definitely be considered as representing gold–gold bonding
between closed shell cations. Aurophilic bonding of this kind is
now well established in the chemistry of low-valent gold com-
pounds and is known to influence significantly many molecular
418
J. Chem. Soc., Dalton Trans., 1999, 415–420

View Article Online
second ligand. However, the resulting AgP3 core is not trigonal
planar as expected for AgI, but trigonal pyramidal owing to a
close mutual approach of the two silver atoms to a short con-
tact of only 2.8569(8) Å, which provides each metal with a
pseudo-tetrahedral environment. The Ag–P distances are in the
range from 2.4301(12) to 2.5439(12) Å, in good agreement with
standard Ag–P bond lengths (compare Fig. 3). The P–Ag–P
angles show large deviations from the tetrahedral values, but
their average (108.9Њ) agrees very well with a tetrahedral angle.
The distortions are of course due to the constraints of the
ligand.
Syntheses
[Au(TP)Cl] 1. The compound [Au(Me2S)Cl] (120 mg, 0.4
mmol) was dissolved in dichloromethane (10 ml). Addition of
TP (250 mg, 0.4 mmol) gave a homogeneous yellow solution,
which was stirred for 2 h. A yellow precipitate formed upon
addition of pentane. Yellow needles of complex 1 were
obtained by allowing diethyl ether to diffuse into a dichloro-
methane solution. Yield: 84%, mp 332 ЊC (decomp.) (Found: C,
56.44; H, 3.86; Cl, 6.23. Calc. for C42H33AuClP3: C, 57.0; H,
3.81; Cl, 6.63%). MS (CI): m/z 826.5 [Mϩ], 630.7 [Mϩ Ϫ AuCl]
and 444.8 [Mϩ Ϫ Ph2PAuCl]. 1H NMR (CDCl3, 20 ЊC): δ 6.74,
7.02–7.38, 7.8 and 7.98 (all m, C6H4, C6H5). 31P-{1H} NMR
(CDCl3, 20 ЊC): A2B spin system, δA 21.4 (d, 2 P), δB 5.1 (t, 1 P),
J(P,P) = 190 Hz.
The structure of the dication [Ag2(TP)2]2ϩ in 4 is unique in at
least two features: (1) it is an extreme case of strong “metallo-
philic” bonding, the more general form of “aurophilic” bond-
ing, between closed-shell cations with a higher co-ordination
number;22,23 (2) the Ag ؒ ؒ ؒ Ag contact is significantly shorter
than the intermetallic distance in metallic silver (2.889 Å)24 and
in other cases where silver is involved in similar closed-shell
interactions.22,23,25 This observation shows that this type of
bonding between coinage metals is a much more general phen-
omenon than previously assumed. It is probably a semantic
question if the two silver atoms in 4 are actually drawn together
by a bonding force, or if they are simply pushed and held
together in the “Procrustean Bed” of the ligand. Ample evi-
dence for true bonding in gold chemistry, where several reliable
estimates are now available for the bond energies involved,
make it very unlikely that the two metal atoms in the dications
[M2(TP)2]2ϩ are simply pressed into or onto each other by the
TP ligands.18,26–30 It rather appears that the gain in bond energy
associated with a metallophilic approach of two metal cations is
not insignificant, even for lighter atoms with reduced relativistic
effects, which were held mainly responsible for aurophilic bond-
ing in the early discussions of the phenomenon.13,14
Unfortunately, no model system has been found yet where
the energy of this type of Ag ؒ ؒ ؒ Ag bonding could be meas-
ured, or at least estimated, but all recent observations suggest
that it may not be much smaller than for Au ؒ ؒ ؒ Au
contacts.18,26–30 Since we are dealing generally with weak forces,
there must be a very delicate balance between several contribu-
tions to the overall energy of the system. This may also explain
why the gold and silver tetrafluoroborates are not isostructural
like the two chlorides. The small differences in the ionic radii of
gold and silver may well be sufficient to induce a distortion of
the dication. The observation that the unsymmetrical or sym-
metrical structure, respectively, is retained in solution at low
temperature is convincing evidence that the distortion in
[Au2(TP)2]2ϩ is not caused by crystal packing, solvation or other
external influences, but is an intrinsic feature of the gold com-
plex. For the silver complex the symmetrical structure is main-
tained, but both species have approximately the same M ؒ ؒ ؒ M
distance, in one case between two four-co-ordinate metal atoms,
and in the other between a three- and a five-co-ordinated metal
atom.
Ag(TP)Cl] 2. To a suspension of AgCl (11.4 mg, 0.08
mmol) in dichloromethane (5 ml) the ligand TP (50 mg, 0.08
mmol) was added and the reaction mixture stirred for 48 h.
From the homogeneous green solution the compound was
isolated as a green solid by removing the solvent in a vacuum.
Diffusion of diethyl ether into a dichloromethane solution led
to the formation of green needles. Yield: 68%, mp 343 ЊC
(decomp.), (Found: C, 64.63; H, 4.2. Calc. for C42H33AgClP3: C,
65.18; H, 4.3%); MS (CI): m/z 630.7 [Mϩ Ϫ AgCl] and 445.4
[Mϩ Ϫ Ph2PAgCl]. 1H NMR (CD2Cl2, 20 ЊC) δ 6.94, 7.05, 7.16,
7.24–7.3 and 7.62 (all m, C6H4, C6H5). 31P-{1H} NMR
(CD2Cl2), (20 ЊC) δ Ϫ7.5 (broad d, PPh2), Ϫ26.6 (broad t, PPh),
J(P,P) = 210 Hz; (Ϫ70 ЊC) δ Ϫ6.7 (dd, PPh2 = exP), Ϫ24.2 (dt,
PPh = cP), J(exP,cP) = 204, J(exP109Ag) = 317, J(exP107Ag) = 280,
J(cP109/107Ag) = 123 Hz.
[Au2(TP)2][BF4]2 3. To a solution of [Au(TP)Cl] (140 mg, 0.15
mmol) in dichloromethane (10 ml) a solution of AgBF4 (30 mg,
0.15 mmol) in methanol (10 ml) was added with stirring. A
white solid (AgCl) precipitated immediately. After 1 h the
orange solution was filtered and the solvents were removed. The
product was obtained as an orange solid. Orange crystals were
obtained by recrystallisation from dichloromethane–diethyl
ether. Yield: 87%, mp 310 ЊC (decomp.) (Found: C, 52.35; H,
3.34; P, 9.8. Calc. for C84H66Au2B2F8P6ؒ2CH2Cl2: C, 51.68; H,
3.53; P, 9.3%). MS (FAB): m/z 1655.1 [Mϩ] and 827.1 [M2ϩ]. 1H
NMR (CD2Cl2, 20 ЊC): δ 5.8, 6.09, 6.39, 6.8–7.44 and 7.8 (all m,
C6H4, C6H5). 31P-{1H} NMR: (CD2Cl2), (20 ЊC) δ 16.4 (pseudo-
qnt, 1 P) and 26.2 (br, 2 P); (CD2Cl2, Ϫ60 ЊC) δ 15.6 (m, 1 P),
19.4 (m, 1 P) and 33.8 (pseudo-t, 1 P); (o-dichlorobenzene–
DMSO-d6, 125 ЊC) δ 16.9 (broad pseudo-qnt, 1 P) and 25.6
(pseudo-t, 2 P). 19F NMR (CD2Cl2, 20 ЊC): δ 73.6 (s, BF4 ).
Ϫ
[Ag2(TP)2][BF4]2 4. A mixture of AgBF4 (31 mg, 0.16 mmol)
and TP (100 mg, 0.16 mmol) in dichloromethane was stirred for
96 h. From the homogeneous colourless solution the product
could be isolated as a white solid by removing the solvent.
Diethyl ether was allowed to diffuse into a dichloromethane
solution of compound 4 to precipitate colourless needle-like
crystals. Yield: 74%, mp 215 ЊC (decomp.) (Found: C, 59.9; H,
4.32. Calc. for C42H33AgBF4P3: C, 61.12; H, 4.03%). MS (FAB):
Experimental
General procedure, measurements and materials
737.7 [M2ϩ]. H NMR (CD2Cl2, 20 ЊC): δ 6.62, 6.96 and 7.18–
1
All experiments were carried out under dry, purified nitrogen.
Solvents used for reactions and crystallizations were dried
using appropriate agents, distilled and kept under nitrogen.
Glassware was oven-dried and filled with nitrogen. The NMR
spectra were obtained on JEOL GX 270 and GX 400 spectro-
meters. Tetramethylsilane, phosphoric acid and trifluoro-
7.48 (all m, C6H4, C6H5). 31P-{1H} NMR (CD2Cl2, 20 ЊC):
δ Ϫ0.6 (pseudo-qnt, 2 P) and Ϫ14.8 (m, 1 P).
Crystal structure determinations
Specimens of suitable quality and size of compounds 1–4 were
mounted in glass capillaries and used for measurements of
precise cell constants and intensity data collection on an Enraf
Nonius CAD4 diffractometer (Mo-Kα radiation, λ = 0.71073
Å). During data collection three standard reflections were
measured periodically as a general check of crystal and instru-
ment stability. No significant changes were observed. Lorentz-
polarization correction was applied, and the data of 1, 3 and 4
1
acetic acid served as reference compounds for H, 31P and 19F
NMR, respectively (δ values in ppm). For mass spectroscopic
measurements a Finnigan MAT 90 spectrometer was used.
Microanalyses were performed on in-house equipment (by
combustion). The ligand bis[2-(diphenylphosphino)phenyl]-
phenylphosphine17 and chloro(dimethyl sulfide)gold()31 were
prepared according to literature procedures.
J. Chem. Soc., Dalton Trans., 1999, 415–420
419

View Article Online
Table 1 Crystal data, data collection, and structure refinement
[Au(TP)Cl] 1
[Ag(TP)Cl] 2
[Au2(TP)2][BF4]2ؒ2 CH2Cl2 3
[Ag2(TP)2][BF4]2 4
Formula
Mr
Crystal system
Space group
C42H33AuClP3
863.01
Monoclinic
P21/n
C42H33AgClP3
773.91
Monoclinic
P21/n
C86H70Au2B2Cl4F8P6
1998.59
Monoclinic
P21/c
C42H33AgBF4P3
825.27
Monoclinic
C2/c
a/Å
b/Å
c/Å
13.586(1)
18.422(1)
13.773(1)
96.71(1)
3419.0(4)
1.677
13.606(1)
18.609(1)
13.731
96.06(1)
3457.2(4)
1.487
12.794(2)
25.383(1)
24.788(3)
91.69(1)
8043(3)
1.650
21.129(3)
17.005(2)
21.876(3)
104.83(1)
7598(2)
1.443
β/Њ
V/ÅϪ3
Dc/g cmϪ3
Z
4
1704
73.24
Ϫ85
4
1576
8.30
Ϫ78
4
3936
39.60
Ϫ78
8
3344
7.06
Ϫ70
F(000)
µ(Mo-Kα)/cmϪ1
T/ЊC
Measured reflections
Unique reflections
Reflections used for refinement
Refined parameters
Final R values [I > 2σ(I)]
R1
6802
7848
16458
15748 [Rint = 0.0250]
14177
13606
8260 [Rint = 0.0325]
7325
6691 [Rint = 0.0225]
6103
424
7539 [Rint = 0.0348]
7155
424
893
460
0.0245
0.0563
0.545/Ϫ0.574
0.0237
0.0609
0.545/Ϫ0.574
0.0423
0.0545
wR2
0.1048
0.1554
ρfin(max./min.)/e ÅϪ3
3.321/Ϫ1.728a
3.248/Ϫ0.970b
a Residual electron densities are located around BF4Ϫ and CH2Cl2. b Residual electron densities are located at Ag.
were corrected for absorption (DIFABS32). The structures of 1,
3, and 4 were solved with direct methods (SHELXS 8633) and
completed by full-matrix least squares techniques against F 2
(SHELXL 9334). The coordinates of the solution of 1 were used
for the refinement of the isomorphous compound 2. The ther-
mal motion of all non-hydrogen atoms was treated anisotropic-
ally, except for those of the solvent CH2Cl2 and the BF4 anion
in the lattice of 3, which were refined isotropically. All hydro-
gen atoms were placed in idealized calculated positions and
allowed to ride on their corresponding carbon atoms with fixed
isotropic contributions [Uiso(fix) = 1.5 × Ueq of the attached C
atom]. Crystal data, data collection and structure refinement
are summarized in Table 1. Important interatomic distances
and angles are shown in the corresponding Figure Captions.
CCDC reference number 186/1275.
1996, 35, 5387; A. Görling, N. Rösch, D. E. Ellis and H.
Schmidbaur, ibid., 1991, 30, 3986.
11 P. Pyykkö and N. Runeberg, J. Chem. Soc., Chem. Commun., 1993,
1812.
12 B. Lippert, F. Zamora, M. Sabat, M. Janik and C. Siethoff, Chem.
Commun., 1997, 485.
13 N. Kaltsoyannis, J. Chem. Soc., Dalton Trans., 1997, 1.
14 H. Schmidbaur, Gold Bull., 1990, 23, 11; Chem. Soc. Rev., 1995, 24,
391.
15 W.-H. Chen, T. C. W. Mak and C.-M. Che, J. Chem. Soc., Dalton
Trans., 1998, 2275.
16 M. Barrow, H. B. Bürgi, D. K. Johnson and L. M. Venanzi, J. Am.
Chem. Soc., 1976, 98, 2356.
17 J. G. Hartley, L. M. Venanzi and D. C. Goodall, J. Chem. Soc., 1963,
3930.
18 J. Zank, A. Schier and H. Schmidbaur, J. Chem. Soc., Dalton Trans.,
1998, 323.
19 M. Sigl, A. Schier and H. Schmidbaur, Eur. J. Inorg. Chem., 1998,
203.
20 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National
Laboratory, Oak Ridge, TN, 1976.
21 A. Bayler, A. Schier, G. A. Bowmaker and H. Schmidbaur, J. Am.
Chem. Soc., 1996, 118, 7006; U. M. Tripathi, A. Bauer and H.
Schmidbaur, J. Chem. Soc., Dalton Trans., 1997, 2865.
22 S.-M. Kuang, Z.-Z. Zhang, Q.-G. Wang and T. C. W. Mak, J. Chem.
Soc., Dalton Trans., 1998, 2927.
Acknowledgements
This work was supported by Deutsche Forschungsgemeinschaft
and Fonds der Chemischen Industrie. The authors are grateful
to Mr J. Riede for establishing the X-ray data sets.
23 C.-M. Che, H.-K. Yip, V. W.-W. Yam, P.-Y. Cheung, T.-F. Lai,
S.-J. Shieh and S.-M. Peng, J. Chem. Soc., Dalton Trans., 1992, 427.
24 L. Pauling, The Nature of the Chemical Bond, 3rd edn., Cornell
University Press, Ithaca, NY, 1980, p. 403.
25 L. S. Ahmed, J. R. Dilworth, J. R. Miller and N. Wheatley, Inorg.
Chim. Acta, 1998, 278, 229.
References
1 R. D. Hancock, P. W. Wade, M. P. Ngwenya, A. S. de Sousa and
K. V. Damu, Inorg. Chem., 1990, 29, 1968.
2 F. Gabbaï, A. Schier and H. Schmidbaur, Inorg. Chem., 1995, 34,
3855.
3 H. Rimml and L. M. J. Venanzi, J. Organomet. Chem., 1983, 259,
C6.
26 H. Schmidbaur, W. Graf and G. Müller, Angew. Chem., Int. Ed.
Engl., 1988, 27, 417.
4 A. L. Balch and E. Y. Fung, Inorg. Chem., 1990, 29, 4764.
5 A. J. Blake, J. A. Greig, J. Holder, T. J. Hyde, A. Taylor and M.
Schröder, Angew. Chem., Int. Ed. Engl., 1990, 29, 197.
6 J. V. McArdle and G. E. Bossard, J. Chem. Soc., Dalton Trans., 1990,
2219.
7 A. J. Blake, W.-S. Li, V. Lippolis, A. Taylor and M. Schröder,
J. Chem. Soc., Dalton Trans., 1998, 2931; A. J. Blake, A. Taylor and
M. Schröder, J. Chem. Soc., Chem. Commun., 1993, 1097.
8 Entries of a standard encyclopedia series: Procrustean Bed, “a
scheme or pattern into which someone or something is arbitrarily
forced,”; Procrustean, “marked by arbitrary often ruthless disregard
of individual differences or special circumstances”; Procrustes, “a
legendary robber of ancient Greece noted for stretching and cutting
off the legs of his victims to adapt them to the length of his bed”
(Damastes).
27 H. Schmidbaur, K. Dziwok, A. Grohmann and G. Müller, Chem.
Ber., 1989, 122, 893.
28 K. Dziwok, J. Lachmann, D. L. Wilkinson, G. Müller and H.
Schmidbaur, Chem. Ber., 1990, 123, 423.
29 M. R. M. Bruce, R. Narayanaswany, M. A. Young, E. Parkhust,
M. Ouelette, M. E. Kerr, D. Ho, R. C. Elder and A. E. Bruce, Inorg.
Chem., 1993, 32, 2506.
30 M. F. Hawthorne, D. E. Harwell, M. D. Mortimer, C. B. Knobler
and F. A. L. Anet, J. Am. Chem. Soc., 1996, 118, 2679.
31 K. C. Dash and H. Schmidbaur, Chem. Ber., 1973, 106, 1221.
32 DIFABS, N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983,
39, 158.
33 SHELXS 86, G. M. Sheldrick, University of Göttingen, 1986.
34 SHELXL 93, G. M. Sheldrick, University of Göttingen, 1993.
9 P. Pyykkö, Chem. Rev., 1997, 97, 597.
10 S. Chung, S. Krüger, H. Schmidbaur and N. Rösch, Inorg. Chem.,
Paper 8/08266I
420
J. Chem. Soc., Dalton Trans., 1999, 415–420