Hexaplatinum clusters with carbonyl and diphosphine ligands and the trapping of mercury(0) and thallium(I)

Article abstract of DOI:10.1021/om960148k

Hexaplatinum cluster complexes have been prepared by reduction of [PtCl₂(SMe₂)₂] with NaBH₄ in the presence of CO and the diphosphine ligands Ph₂P(CH₂)_nPPh₂, n = 2 (dp

Full text of DOI:10.1021/om960148k

Organometallics 1996, 15, 3115-3123
3115
Hexa p la tin u m Clu ster s w ith Ca r bon yl a n d Dip h osp h in e
Liga n d s a n d th e Tr a p p in g of Mer cu r y(0) a n d Th a lliu m (I)
Leijun Hao, J agadese J . Vittal, and Richard J . Puddephatt*
Department of Chemistry, University of Western Ontario,
London, Ontario, Canada N6A 5B7
Received February 28, 1996X
Hexaplatinum cluster complexes have been prepared by reduction of [PtCl2(SMe2)2] with
NaBH4 in the presence of CO and the diphosphine ligands Ph2P(CH2)nPPh2, n ) 2 (dppe), 3
(dppp), 4 (dppb), 5 (dpppe), and 6 (dpph). The products are the closed cluster [Pt6(µ-CO)6-
(µ-dppe)3], 2b, and the open cluster [Pt6(µ-CO)6(µ-dppp)2(dppp)2], 9, but the reactions failed
to give isolable products when n ) 4-6. Cluster 9 can encapsulate mercury(0) or thallium(I)
with loss of dppp to give the closed clusters [Pt6(µ6-Hg)(µ-CO)6(µ-dppp)3], 10a , or [Pt6(µ6-
Tl)(µ-CO)6(µ-dppp)3]+, 11a , respectively. The Pt6(µ-CO)6(µ-dppp)3 unit acts as a cryptand
and evidence is presented that the cluster 10a must open to allow reaction with thallium(I)
to give 11a . Reduction of [PtCl2(SMe2)2] with NaBH4 in the presence of CO, the diphosphine
ligands Ph2P(CH2)nPPh2, and mercury can give [Pt6(µ6-Hg)(µ-CO)6{µ-Ph2P(CH2)nPPh2}3],
10a -d , n ) 3-6, respectively, directly, but clusters [Pt6(µ6-Hg)(µ-CO)6(CO)2{µ-Ph2P(CH2)n-
PPh2}2], 12, are also formed in some cases. Cluster 2b reacts with [Ph3PAu]+ to give [Pt6(µ-
CO)6(µ-dppe)3(µ3-AuPPh3)2]2+.
Ch a r t 1a
In tr od u ction
The encapsulation of a metal ion by a cryptand ligand
has dramatic effects on both the thermodynamic stabil-
ity and kinetic reactivity of the resulting cryptate
complex since the metal ion is physically trapped and
cannot easily escape.1 This article describes a study of
the synthesis of cluster cryptate complexes, in which
two trinuclear platinum clusters of the type [Pt3(µ-
CO)3L3], 1,2 are held together by using bidentate phos-
phine ligands LL of the type Ph2P(CH2)nPPh2 to give
[{Pt3(µ-CO)3}2(µ-LL)3], 2 (Chart 1). If n is large enough,
the space between the clusters should be great enough
to encapsulate a guest metal atom or ion. In related
research, the electrochemical reduction at a mercury
electrode of a platinum(II) precursor in the presence of
a diphosphine and an isocyanide ligand has been
studied in detail.3,4 Among a rich diversity of products
were the complexes [Pt6(Hgn)(µ-CNR)6(RNC)2(µ-LL)2] [3,
n ) 1, LL ) dppb; 4, n ) 2, LL ) dpppe, dpph] and
[Pt6(Hg2)(µ-CNR)6(µ-LL)3] (5, LL ) dpppe, dpph) (Chart
1).3 Synthetic reactions could be carried out using
sodium amalgam as reducing agent in some cases.3
Since the reactions must be carried out in the presence
of mercury, only the mercury inclusion complexes could
be isolated, and no mercury inclusion complexes were
detected when the shorter methylene chain diphos-
phines dppm, dppe, and dppp were used; other interest-
ing Pt-Pt-bonded binuclear and trinuclear complexes
a
Complex 8 was originally formulated ad [Pt{Pt(PP)L}2]2+
in ref 3.
were formed instead (Chart 1, complexes 6-8).3,4 Fol-
lowing earlier studies of the cluster [Pt6(µ-CO)6(µ-
dppm)3], 2a ,5 it was of interest to determine if analogous
carbonyl clusters with longer methylene chain diphos-
phines could be synthesized and if they might display
interesting inclusion chemistry as the space between the
Pt3 triangles in 2 increased. The reducing agent used
was Na[BH4], since it can yield mercury-free com-
X Abstract published in Advance ACS Abstracts, J une 15, 1996.
(1) (a) Lehn, J .-M. Angew. Chem., Int. Ed. Engl. 1988, 27, 89. (b)
Mertes, K. B.; Lehn, J .-M. In Comprehensive Organometallic Chem-
istry; Wilkinson, G., Ed.; Pergamon: Oxford, U.K., 1987; Vol. 2,
Chapter 21.3. (c) Muller, A.; Reuter, H.; Dillinger, S. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2328.
(2) For reviews, see: (a) Imhof, D.; Venanzi, L. M. Chem. Soc. Rev.
1994, 185. (b) Cross, R. J . In Comprehensive Organometallic Chemistry,
2nd ed.; Pergamon: Oxford, U.K., 1995; Vol. 9, Chapter 7. (c) Mingos,
D. M. P.; Wardle, R. W. M. Transition Met. Chem. 1985, 10, 441.
(3) Tanase, T.; Ukaji, H.; Kudo, Y.; Ohno, M.; Kobayashi, K.;
Yamamoto, Y. Organometallics 1994, 13, 1374.
(5) (a) Hao, L.; Spivak, G. J .; Xiao, J .; Vittal, J . J .; Puddephatt, R.
J . J . Am. Chem. Soc. 1995, 117, 7011. (b) Spivak, G. J .; Hao, L.; Vittal,
J . J .; Puddephatt, R. J . J . Am. Chem. Soc. 1996, 118, 225.
(4) Tanase, T.; Horiuchi, T.; Yamamoto, Y.; Kobayashi, K. J .
Organomet. Chem. 1992, 440, 1.
S0276-7333(96)00148-3 CCC: $12.00 © 1996 American Chemical Society
+
+

3116 Organometallics, Vol. 15, No. 14, 1996
Hao et al.
plexes.5,6 A preliminary account of the chemistry with
the ligand dppp has been published.7
Ch a r t 2
Resu lts
A Hexa n u clea r Clu ster w ith Bis(d ip h en ylp h os-
p h in o)eth a n e, d p p e. Reduction of [PtCl2(SMe2)2] with
NaBH4 in the presence of CO and dppe led directly to
the closed cluster [Pt6(µ-CO)6(µ-dppe)3], 2b, which was
isolated as a red-brown solid which was slowly decom-
posed on exposure to air. The cluster was characterized
spectroscopically by comparison to the well-known
clusters [Pt3(µ-CO)3(PR3)3], 1.2 The IR spectrum con-
tained only bridging carbonyl bands at ν(CO) ) 1840
(m), 1796 (vs), 1780 (vs, sh), and 1731 (s) cm-1, in the
same region as found for 1.2 The 31P NMR spectrum of
2b is also similar to those reported for 1.2 It contained
only a singlet at δ(31P) ) 53.2, with satellites showing
1
the couplings J (PtP) ) 4814 Hz, 2J (PtP) ) 527 Hz, and
3J (PP) ) 64 Hz. No coupling 3J (PCH2CH2P) and no long
range 195Pt31P coupling between triangles was observed.
In 13C-enriched samples, the bridging carbonyl ligands
gave a single resonance in the 13C NMR at δ(13C) )
248.1 which appeared as a 1:8:18:8:1 quintet due to the
coupling 1J (PtC) ) 709 Hz, this intensity pattern being
characteristic of a Pt2(µ2-CO) ligand.2 In the presence
of free 13CO, this resonance broadened and the 195Pt
satellite spectra were lost, indicating easy exchange
between free and coordinated CO at room temperature.
No long-range 195Pt13C coupling was resolved. The 31P
NMR spectrum was not affected by the presence of free
CO, indicating that no adducts between 2b and CO are
formed in significant amounts. The NMR spectra of 2b
were essentially unchanged at -90 °C, indicating that
the apparent high symmetry is probably not a function
of fluxionality. All evidence therefore indicates that 2b
has the proposed structure and that there is little Pt‚‚‚Pt
bonding between the two triangles.5
to the chelating [δ(31P) ) 25.5, 1J (PtP) ) 3425 Hz,
2J (PtP) ) 294 Hz, 3J (PaPb) ) 19 Hz] dppp ligands. The
13C NMR spectrum contained two resonances due to the
carbonyls C1O [δ(13C) ) 233.2, 1J (Pt2C1) ) 790 Hz,
1J (Pt1C1) ) 628 Hz, 2J (Pt2C1) ) 24 Hz] and C2O [δ(13C)
1
) 256.1, J (Pt2C2) ) 791 Hz]. Finally, the 195Pt NMR
spectrum contained two resonances, due to the non-
equivalent atoms Pt1 [δ(195Pt) ) -2220, 1J (Pt1Pa) )
2
1
3425 Hz, J (Pt1Pb) ) 536 Hz, J (Pt1Pt2) ) 386 Hz] and
Pt2 [δ(195Pt) ) -2424, J (Pt2Pb) ) 4682 Hz, J (Pt2Pb′)
1
2
) 445 Hz, J (Pt2Pa) ) 290 Hz, J (Pt1Pt2) ) 386 Hz].
The major coupling 1J (PtP) gave rise to a triplet for Pt1
and a doublet for Pt2, thus immediately confirming the
number of directly bonded phosphorus atoms. These
spectra clearly define the structure 9, having effective
D2h symmetry.
2
1
A Hexa n u clea r Clu ster w ith Bis(d ip h en ylp h os-
p h in o)p r op a n e, d p p p . Reduction of [PtCl2(Me2S)2]
with excess NaBH4 in the presence of dppp ) Ph2P(CH2)3-
PPh2 and CO gave the new cluster complex [Pt6(µ-CO)6-
(µ-dppp)2(dppp)2], 9, Chart 2, in 85% yield as a brown-
red solid. Cluster 9 can be considered to contain two
separate Pt3 clusters bridged by two µ-dppp ligands, and
it is therefore an analog of the known trinuclear
complexes [Pt3(µ-CO)3L4]2 and in particular of the dppp
derivative [Pt3(µ-CO)3(PCy3)2(dppp)].8 The IR and room-
temperature NMR data for 9 and [Pt3(µ-CO)3(PCy3)2-
(dppp)]8 are very similar, and so 9 was readily charac-
terized by spectroscopic methods as outlined below. At
room temperature, the NMR spectra indicate an effec-
tive plane of symmetry containing all platinum atoms,
the carbonyl ligands, and phosphorus atoms of the
µ-dppp ligands and bisecting the chelating dppp ligands.
Thus, for example, only two 31P resonances of equal
intensity are observed, one due to the bridging [δ(31P)
However, at low temperature, two resonances are
observed in the 31P NMR spectrum for the phosphorus
atoms of the chelating dppp ligand. This is clearly a
result of fluxionality and shows that the true symmetry
of 9 is lower than D2h, probably C2v or C2. There are
two reasonable explanations. First, it is probable that
the structure is not planar but angular, and the flux-
ionality might then involve angular to planar to angular
equilibration as shown in the upper part of Scheme 1.
Second, it is possible that the structure is effectively
planar but that the coordination of the chelating dppp
ligand is equatorial-axial and that the fluxionality
involves equatorial-axial exchange as shown in the
lower part of Scheme 1. Of course, some combination
of both processes is also possible.
The open structure 9 is clearly favored over the closed
structure 2, and attempts to prepare 2, with LL ) dppp,
have been unsuccessful. For example, if a deficiency of
dppp is used during the reduction of [PtCl2(SMe2)2], 9
is still the only detected product with some of the
precursor being reduced to metallic platinum.
1
2
2
) 56.5, J (PtP) ) 4682 Hz, J (Pt1P) ) 534 Hz, J (Pt2P)
) 445 Hz, 3J (PbPb′) ) 60 Hz, 3J (PaPb) ) 19 Hz] and one
(6) (a) Dahmen, K. H.; Moor, A.; Naegeli, R.; Venanzi, L. M. Inorg.
Chem. 1991, 30, 4285. (b) Holah, D. G.; Hughes, A. N.; Krysa, E.;
Magnuson, V. R. Organometallics 1993, 12, 4721.
(7) Hao, L.; Vittal, J . J .; Puddephatt, R. J . Inorg. Chem. 1996, 35,
269.
(8) (a) Hallam, M. F.; Howells, N. D.; Mingos, D. M. P.; Wardle, R.
W. M. J . Chem. Soc., Dalton Trans. 1985, 845. (b) Briant, C. E.; Wardle,
R. W. M.; Mingos, D. M. P. J . Organomet. Chem. 1984, 267, C49.
Hexan u clear Clu ster s with Diph osph in e Ligan ds
d p p b, d p p p e, a n d d p p h . The analogous reactions to
those described above were unsuccessful in these cases.
For example, in the reduction with dppb a red-brown
solid was obtained which gave ν(CO) at 1995 and 1790
cm-1, indicating the presence of both terminal and
+
+

Hexaplatinum Clusters
Organometallics, Vol. 15, No. 14, 1996 3117
Sch em e 1. P ossible Mech a n ism s of F lu xion a lity
Sch em e 2
Ta ble 1. Selected NMR (p p m , Hz) a n d IR Da ta
(cm -1) for th e Closed P t6 Clu ster s
bridging carbonyl ligands. In the 31P NMR spectrum,
there were very broad resonances at δ ) 35 and 50 and
a sharper resonance at δ ) 17, and no 195Pt satellites
were resolved either at room temperature or at -90 °C
in dichloromethane solution. The nature of these
cluster complexes, including the question of whether or
not they are homogeneous, is therefore still unknown.
They are clearly platinum(0) clusters as shown by the
reaction chemistry described below.
Tr a p p in g Rea ction s w ith Mer cu r y(0) a n d Th a l-
liu m (I). Attempts to prepare clusters containing mer-
cury(0) or thallium(I) were made either by direct
reaction with the above clusters or by carrying out the
reduction of [PtCl2(SMe2)2] in the presence of CO, the
diphosphine ligand, and Hg(0) or Tl(I).
2a a
2b
10a
11a
10b
11b
10c
10d
δ(31P)
37.6
53.2 58.6 56.7
46.6 47.0 53.2 52.9
1J (PtP) 5289 4814 5037 4810 4919 5040 4987 4961
2J (PtP) 274
3J (PP)
527
64
403
55
360
50
396
56
c
c
403
55
397
56
b
δ(13CO) 240.5 251
235
714
232.5 238
696 715
236
705
235
720
233
744
1J (PtC) 757
697
ν(CO)
1790 1840 1837 1871 1864 1808 1830 1823
1773 1796 1799 1818 1828 1776 1788 1781
1780 1782
1731
1802
1791
a
b
From ref 5. A broad doublet splitting of 216 Hz is probably
due to 2J (PP). c Broad spectrum, not resolved. Complete data are
in the experimental section.
are not possible due to the general decomposition which
competes with the desired reaction, but it is immediately
obvious that the conversion of 10a to 11a is much slower
than the conversion of 9 to 11a . It is proposed that
direct loss of mercury from 10a is impossible and that
opening up of one of the µ-dppp bridges (aided by free
dppp) must occur prior to the metal for metal exchange,
as shown schematically in Scheme 2. Complex 10a was
also prepared directly by the reduction of [PtCl2(SMe2)2]
in the presence of CO, dppp, and excess mercury.
The complexes 10a and 11a were characterized
spectroscopically, and 11a [BPh4] was also characterized
by an X-ray structure determination. Each complex
gave a single resonance in the 31P NMR spectrum due
to the diphosphine ligands and in the 13C NMR due to
the bridging carbonyl ligands, thus confirming the high
symmetry of the products (Table 1). Complex 10a gave
a single resonance in the 195Pt NMR which appeared
as a doublet due to 1J (PtP) ) 5037 Hz, clearly indicating
that there is just one phosphorus atom bound to each
No pure adducts of 2b with mercury(0) or thallium(I)
could be obtained by the above methods. For example,
when mercury was added to a solution of 2b in CD2Cl2,
the 31P NMR spectrum after 2 h showed that no reaction
had occurred, though a change in color was observed.
After 2 days, much decomposition had occurred but no
mercury adduct was detected by 31P NMR. Molecular
modeling shows that the cavity in 2b is too small to
accommodate a central mercury(0) atom or thallium(I)
ion, and so complex formation would either occur at an
external face or by opening up of the cluster. Neither
appears to give rise to pure clusters.
With the longer chain diphosphines, mercury(0) or
thallium(I) inclusion is possible and gives rise to stable
complexes. This is clearly illustrated by the reaction
of 9 with mercury(0) or Tl[PF6], which occurs with loss
of 1 equiv of dppp to form the closed carcerand clusters
10a or 11a , respectively (Scheme 2). These reactions
occur rapidly and quantitatively. In this sense, the open
cluster 9 acts as a kind of “Venus flytrap” since, once
the heavy metal is complexed and the third µ-dppp
ligand is in place, the incarcerated metal cannot easily
escape. This is demonstrated by attempted metal-for-
metal replacement reactions. Thus, addition of mercury
to 11a gives no reaction after either 1 h or 1 day (some
decomposition occurs over 1 day). Addition of Tl[PF6]
to 10a gives no reaction after 1 h but about 20% 11a is
formed after 1 day (along with some decomposition); the
reaction is accelerated in the presence of free dppp and
significant conversion of 10a to 11a is observed after 1
h under these conditions. Quantitative kinetic studies
1
platinum. The metal-metal couplings were J (HgPt)
2
) 3350 Hz, 1J (PtPt) ) 2200 Hz, and J (PtPt) ) 100 Hz.
1
The coupling 1J (HgPt) is smaller, and J (PtPt) is larger
than in the related clusters [Pt3(µ-CO)3L3(µ3-HgX)2] (L
) phosphine, X ) halide), for which the ranges are
1
1J (HgPt) ) 5784-6723 Hz and J (PtPt) ) 1609-1860
Hz, respectively.9 The 31P NMR spectrum of 11a was
broader than for 10a , and the weaker satellite peaks
needed to extract PtPt and PtTl couplings were not fully
(9) Albinati, A.; Dahmen, K. H.; Demartin, F.; Forward, J . M.;
Longley, C. J .; Mingos, D. M. P.; Venanzi, L. M. Inorg. Chem. 1992,
31, 2223.
+
+

3118 Organometallics, Vol. 15, No. 14, 1996
Hao et al.
angle are in the range 2.658(3)-2.682(3) Å, well within
the range found for Pt-Pt single bonds in related
clusters,2-13 but the intertriangle Pt‚‚‚Pt distances are
greater than 5 Å and so are clearly nonbonding. The
thallium atom is bonded in sandwich fashion to all six
platinum atoms with Pt-Tl distances in the range
2.860(3)-2.992(3) Å. There are no other Pt6Tl clusters
for comparison, but the related cluster [Pt3(µ3-Tl)(µ-
CO)3(PCy3)3]+ has Pt-Tl ) 3.034(1)-3.047(1) Å, longer
than in 11a .11 It is relevant to note that the mean angles
PtPC, PCC, and CCC for the dppp ligands in 11a are
113, 113, and 114°, respectively, all greater than the
ideal tetrahedral angle, suggesting that the dppp ligands
are stretched slightly in order to span the two Pt3
triangles whose centroids are separated by 4.98 Å.
Perhaps there is a balance between these two effects
and the PtTl distances are slightly smaller than their
ideal values in 11a to minimize this stretching of the
dppp ligands.
In the case of the longer chain diphosphine ligands,
the mercury inclusion complexes 10 were best prepared
by reduction of [PtCl2(SMe2)2] in the presence of CO and
the corresponding diphosphine dppb, dpppe, or dpph,
followed by addition of excess mercury. In the case of
dppb, the complex 10b was prepared in high yield by
use of the stoichiometric amount of dppb. However, in
the other cases, formation of 10c (dpppe) or 10d (dpph)
was accompanied by formation of 12, which are analo-
gous to the known isocyanide cluster 3. The clusters
12 could not be isolated in pure form, and their
identification is based primarily on NMR and IR data
of their mixtures with 10. In addition, reaction of 12
with more diphosphine occurred easily to give 10 by
displacement of the terminal carbonyl ligands (eq 1) and
the clusters 10 could then be isolated in analytically
pure form.
F igu r e 1. View of the structure of the cluster cation 11a .
The closest intertriangle Pt‚‚‚Pt distance is Pt(2)Pt(6) )
5.036(3) Å, which is clearly nonbonding. The angle between
the planes Pt(1)Pt(2)Pt(3) and Pt(4)Pt(5)Pt(6) is 2.0(1)°.
Ta ble 2. Selected Bon d Dista n ces (Å) a n d An gles
(d eg) in 11a [BP h 4]
Tl(1)-Pt(1)
Tl(1)-Pt(3)
Tl(1)-Pt(5)
Pt(1)-Pt(2)
Pt(2)-Pt(3)
Pt(4)-Pt(6)
Pt(1)-P(1)
Pt(3)-P(3)
Pt(5)-P(5)
Pt(1)-C(3)
Pt(2)-C(1)
Pt(3)-C(3)
Pt(4)-C(4)
Pt(5)-C(5)
Pt(6)-C(6)
2.860(3)
2.920(3)
2.967(3)
2.682(3)
2.667(3)
2.658(3)
2.23(2)
2.25(1)
2.27(1)
1.91(4)
2.02(5)
2.08(4)
1.95(4)
2.00(4)
1.98(4)
Tl(1)-Pt(4)
Tl(1)-Pt(2)
Tl(1)-Pt(6)
Pt(1)-Pt(3)
Pt(4)-Pt(5)
Pt(5)-Pt(6)
Pt(2)-P(2)
Pt(4)-P(4)
Pt(6)-P(6)
Pt(1)-C(1)
Pt(2)-C(2)
Pt(3)-C(2)
Pt(4)-C(6)
Pt(5)-C(4)
Pt(6)-C(5)
2.909(3)
2.942(3)
2.992(3)
2.673(3)
2.677(3)
2.666(3)
2.25(2)
2.25(1)
2.25(1)
2.11(6)
2.03(4)
2.08(5)
2.02(4)
2.06(5)
2.15(4)
Pt(1)-Tl(1)-Pt(4) 136.5(1)
Pt(1)-Tl(1)-Pt(3)
55.08(7)
55.03(7)
54.12(7)
54.20(6)
Pt(4)-Tl(1)-Pt(3) 121.29(9) Pt(1)-Tl(1)-Pt(2)
Pt(4)-Tl(1)-Pt(2) 165.3(1)
Pt(3)-Tl(1)-Pt(2)
Pt(1)-Tl(1)-Pt(5) 124.32(9) Pt(4)-Tl(1)-Pt(5)
Pt(3)-Tl(1)-Pt(5) 173.9(1)
Pt(1)-Tl(1)-Pt(6) 168.2(1)
Pt(2)-Tl(1)-Pt(5) 131.4(1)
Pt(4)-Tl(1)-Pt(6) 53.52(7)
Pt(3)-Tl(1)-Pt(6) 128.91(9) Pt(2)-Tl(1)-Pt(6) 116.5(1)
Pt(5)-Tl(1)-Pt(6)
Pt(3)-Pt(1)-Tl(1)
Pt(3)-Pt(2)-Pt(1)
Pt(1)-Pt(2)-Tl(1)
Pt(2)-Pt(3)-Tl(1)
Pt(6)-Pt(4)-Pt(5)
Pt(5)-Pt(4)-Tl(1)
Pt(6)-Pt(5)-Tl(1)
Pt(4)-Pt(6)-Pt(5)
Pt(5)-Pt(6)-Tl(1)
53.15(6) Pt(3)-Pt(1)-Pt(2)
63.60(8) Pt(2)-Pt(1)-Tl(1)
59.97(8) Pt(3)-Pt(2)-Tl(1)
60.93(8) Pt(2)-Pt(3)-Pt(1)
63.36(8) Pt(1)-Pt(3)-Tl(1)
59.96(8) Pt(6)-Pt(4)-Tl(1)
63.99(8) Pt(6)-Pt(5)-Pt(4)
63.91(7) Pt(4)-Pt(5)-Tl(1)
60.38(8) Pt(4)-Pt(6)-Tl(1)
62.94(7)
59.74(8)
64.04(8)
62.52(8)
60.29(8)
61.32(7)
64.83(8)
59.67(8)
61.81(7)
61.65(8)
The NMR data for 10b-d are similar to those for 10a
discussed earlier and clearly define the symmetrical
Pt6(µ-CO)6(µ-PP)3 unit (Table 1). The platinum-
diphosphine-carbonyl skeleton of 12 is also clearly
defined. Thus, a 13CO-enriched sample of 12a gave
three carbonyl resonances in the 13C NMR spectrum,
and the characteristic chemical shifts and 1J (PtC)
couplings define one terminal CO resonance, one un-
symmetrical µ2-CO resonance [2 × Pt1Pt2(µ-CO)], and
one symmetrical µ2-CO resonance [Pt22(µ-CO)] at each
Pt3 triangle. The 31P NMR spectrum contained a
singlet, with three sets of 195Pt satellites on either side
due to isotopomers containing a single 195Pt atom,
resolved. However, the presence of thallium was clearly
shown by the doublet appearance of the central reso-
2
nance due to J (TlP) ) 335 Hz. Attempts to obtain a
195Pt NMR spectrum of 11a were unsuccessful, but 11a
was characterized structurally. A view of the cation is
given as Figure 1 and selected bond distances and
angles are given in Table 2.
The six platinum atoms have a geometry inter-
mediate between the trigonal prism and antiprism [the
dihedral angles typified by P(1)Pt(1)Pt(5)P(5) and
P(1)Pt(1)Pt(4)P(4) fall in the ranges 32.4-34.4° and
-79.0 to -86.8°; ideal values would be 0 and -120° for
the trigonal prism and 60 and -60° for the trigonal
antiprism]. The Pt-Pt distances within each Pt3 tri-
(10) Yamamoto, Y.; Yamazaki, H.; Sakurai, T. J . Am. Chem. Soc.
1982, 104, 2329.
(11) Ezomo, O. J .; Mingos, D. M. P.; Williams, I. D. J . Chem. Soc.,
Chem. Comm. 1987, 924.
(12) Yamamoto, Y.; Yamazaki, H. Inorg. Chim. Acta 1994, 217, 121.
(13) Schoettel, G.; Vittal, J . J .; Puddephatt, R. J . J . Am. Chem. Soc.
1990, 112, 6400.
+
+

Hexaplatinum Clusters
Organometallics, Vol. 15, No. 14, 1996 3119
2
appearing as two doublets due to 1J (Pt2P) and J (Pt2′P),
the low-temperature NMR spectrum and indicates easy
exchange of the PPh3 ligands or easy reversible dis-
sociation of the Ph3PAu+ groups.
Neither complex 9 nor 10a -d gave pure products on
reaction with Ph3PAu+. Thus, encapsulation of a mer-
cury atom appears to reduce the tendency to form a
stable adduct with the electrophile Ph3PAu+.
3
with the doublet splitting due to J (PP), and a singlet
due to 2J (Pt1P). In the 195Pt NMR spectrum of 12a , only
the Pt1 resonance was observed and it appeared as a
1:8:18:8:1 quintet of 1:2:1 triplets, as a result of the
couplings 1J (Pt1Pt2) and 2J (Pt1P), respectively. For 12b,
both 195Pt resonances were observed, the Pt1 resonance
being sharp and the Pt2 resonance very broad with only
the J (PtP) coupling resolved. No 199Hg satellites were
1
Discu ssion
resolved for either 12a or 12b.
This work has shown that cluster cryptands can be
prepared based on the triangular platinum clusters
[Pt3(µ-CO)3(PR3)3] by substituting diphosphine ligands
Ph2P(CH2)nPPh2 for the monodentate phosphines. There
are similarities but also major differences between these
clusters and those based on the unit [Pt3(µ-CNXy)3-
(PR3)3], where Xy ) 2,6-dimethylphenyl.3,4 A trend can
be seen in which the two Pt3 triangles tend to be closer
together for the µ-CO than for the µ-CNXy clusters.
Thus, for the isocyanide clusters, there is no analog of
2b or of 10a perhaps because steric hindrance between
XyNC ligands would prevent the necessary close ap-
proach of the two Pt3 triangles. A slightly different
argument is needed to rationalize why there is no XyNC
analog of 10b (reaction of 3 with dppb failed to give the
closed cryptand cluster).4 The (CH2)4 chain would allow
a greater distance between Pt3 triangles, but if there is
only space for one mercury atom, the Pt3 triangles still
need to be about 5 Å apart to allow PtHg bonding to
both triangles. Similarly, the dpppe and dpph ligands
give 5, containing two mercury atoms, for the XyNC
system but 10, containing only one Hg atom, for the CO
system. Naturally, the Pt3 triangles will be much
further apart in 5 and so the problem of steric hindrance
between µ-XyNC groups is eliminated.
The bonding in the clusters 10 and 11 is of interest.
First consider the experimental evidence of the carbonyl
stretching frequencies. For complexes with µ-dppp
ligands, the values of ν(CO) are 2b ∼ 10a < 11a (note
that 2 does not exist with dppp ligands so the closest
analog with dppe ligands is chosen for comparison).
These data indicate that mercury is neither a net donor
nor acceptor with respect to the Pt3 triangles but that
Tl+ is a net acceptor. Thus as charge is donated from
the Pt3 triangle to Tl+ there is correspondingly less back-
bonding to the bridging carbonyl ligands and the
observed frequencies ν(CO) are higher. Within the
series of mercury-encapsulated clusters, the sequence
is ν(CO) for 10a < 10b > 10c > 10d , but the range of
values is small and the significance of the series is not
clear.
The chief problem in characterizing 10b-d and
12a -b is the determination of the number of mercury
atoms present. In the analogous isocyanide clusters,
there is one mercury present when LL ) dppb in 3 but
two when LL ) dpppe or dpph in 4.4 It was therefore
surprising to find that the C,H analyses for all com-
plexes 10b-d indicated the presence of only one mer-
cury atom. The ratio of Pt:Hg was then determined by
energy dispersive X-ray (EDX) analysis on crystalline
samples, and each gave a Pt:Hg ratio of 6:1, as expected
for structure 10, but clearly inconsistent with the
presence of two mercury atoms in each cluster. There
is a marked difference between the stoichiometries of
the otherwise similar µ-CO and µ-CNR cluster com-
plexes.4 Since the clusters 12 were not obtained in
analytically pure form, their composition is less certain
and a structure with two mercury atoms is not excluded.
The thallium sandwich cluster 11b was prepared by
reduction of [PtCl2(SMe2)2] in the presence of CO, dppb,
and thallium(I), but the similar reactions with longer
chain diphosphines dpppe and dpph failed to give pure
clusters. The 31P NMR spectrum of 11b was particu-
larly broad, and no fine structure was observed, except
1
for the satellites due to J (PtP). Otherwise the char-
acterization was similar to that described for 11a .
Ad d ition of P h 3P Au + to 2b. Since it was argued
that 2b could not incorporate a mercury atom because
the space between the two Pt3 triangles was too small,
the question arose as to whether or not external
complexation might occur. No reaction occurred with
mercury(0). However, the reagent Ph3PAu(thf)+, pre-
pared by reaction of Ph3PAuCl with AgPF6 in tetrahy-
drofuran, reacted with 2b to give [Pt6(µ-CO)6(µ-dppe)3-
(µ3-AuPPh3)2][PF6]2, 13, according to eq 2. This reaction
Next consider isolobal analogies. The frontier donor
orbitals of a Pt3 triangle have a + e symmetry, and so
- 14
the Pt3 triangle can be considered isolobal with C5H5
.
Hence the complexes 10 and 11 can be considered
isolobal to Cp2Hg2- (unknown) or Cp2Tl-.15 Now salts
of Cp2Tl- have an angular structure with a stereo-
chemically active lone pair of electrons on thallium(I),
although high-level MO calculations indicate that the
linear structure is only slightly higher in energy.15 In
the structure of 11, the Pt3 triangles are constrained to
is analogous to the known reaction with the trinuclear
clusters 1 and shows that 2b is electron rich.8 The 31P
NMR spectra at both room temperature and at -90 °C
contained singlet resonances for the Ph3PAu groups and
for the dppe ligands, as expected for 13. The triply
bridging nature of the AuPPh3 groups is shown by the
intensities of the satellites due to the long-range cou-
(14) Xiao, J .; Hao, L.; Puddephatt, R. J .; Manojlovic´-Muir, Lj.; Muir,
K. W. J . Am. Chem. Soc. 1995, 117, 6316.
(15) Armstrong, D. R.; Herbst-Irmer, R.; Kuhn, A.; Moncrieff, D.;
Paver, M. A.; Russell, C. A.; Stalke, D.; Steiner, A.; Wright, D. S.
Angew. Chem., Int. Ed. Engl. 1993, 32, 1774.
2
pling J (PtAuP).8 This coupling was observed only in
+
+

3120 Organometallics, Vol. 15, No. 14, 1996
Hao et al.
Ch a r t 3
having platinum pz and carbonyl π* character. 2a1′ has
bonding character in this respect while 3a1′ is antibond-
ing (Chart 3, C, D; note that the CO ligands and orbital
contributions are omitted for the sake of simplicity).
Thus it is the mercury 6s to platinum 6pz overlap that
leads to a net bonding interaction between mercury and
platinum, and this transfers charge from mercury to
platinum.
F igu r e 2. Simplified energy correlation diagram for
formation of [Hg{Pt3(µ-CO)3(PH3)3}2] from {Pt3(µ-CO)3-
(PH3)3}2 and a Hg atom. Complications due to mixing of
the Pt-Pt bonding MO’s with d-levels and mixing of Pt pz
with π*(CO) orbitals are not shown for the sake of clarity.
be almost parallel by the presence of the three µ-dppp
ligands, but the Tl atom is actually slightly displaced
from the ideal position between the centroids of the Pt3
triangles toward Pt(1), Pt(3), and Pt(4) (distances Pt-
Tl ) 2.860, 2.920, and 2.909 Å, respectively) and away
from Pt(2), Pt(5), and Pt(6) (distances Pt-Tl ) 2.942,
2.967, and 2.992 Å, respectively). The distortion is
small, and it is not clear if it arises to minimize
repulsions with the lone pair on thallium(I). EHMO
calculations were carried out on the model clusters
[M{Pt3(µ-CO)3(PH3)3}2]n+, (n ) 0, M ) Hg; n ) 1, M )
Tl), in order to gain further insight.
First, calculations were carried out for geometries
between the trigonal prism and trigonal antiprism while
keeping all bond distances constant. The energy was
marginally lower for the trigonal antiprism but by only
0.04 eV, and there was very little barrier to rotation of
the Pt3 triangles. The stereochemistry adopted by such
sandwich clusters in the solid state is therefore likely
to be determined by steric and packing factors. The
geometry of 11a is intermediate between the two
idealized structures, but most calculations were carried
out in the trigonal prismatic structure with D3h sym-
metry for simplicity. A simplified energy correlation
diagram for formation of the mercury cluster is given
in Figure 2. It is convenient to discuss the bonding in
terms of the interactions of the mercury 6s orbital
(filled) and 6p orbitals (empty) with the platinum cluster
orbitals, whose nature has been discussed extensively
in earlier publications.2,16
A bonding interaction of comparable magnitude oc-
curs by overlap of the filled unsymmetrical combination
of Pt3 bonding MO’s of a2′′ character (again mixed with
2
more dz and dxz, dyz character for better overlap) with
the empty mercury 6pz orbital (Chart 3, E). There is
weaker π-overlap of several filled orbitals having plati-
num 5d-character and e′ symmetry with the empty
mercury 6px, 6py orbitals. Both effects lead to charge
transfer from platinum to mercury.
Both the MO calculations (charge on mercury +0.05
e) and carbonyl stretching frequencies (see above dis-
cussion) suggest that the two bonding effects are about
equal, and so there is very little net transfer of electron
density from mercury to platinum. As the charge on
the central metal increases in the series Hg(0) to Tl(I)
to the hypothetical Pb(II), the corresponding 6s and 6p
orbitals will move to lower energy and the charge
transfer from the Pt3 clusters to the central metal 6p
orbitals will increase while the reverse 6s to Pt3 charge
transfer will decrease, but otherwise, of course, there
are strong similarities. The lower contribution of the
6s versus 6p bonding accounts for the higher values of
ν(CO) for the thallium(I) complex. Our efforts to
prepare the isoelectronic Pb(II) cluster complex have
been unsuccessful.
In all cases, the calculations indicate that the orbital
2a1′ is at relatively high energy and that the HOMO-
LUMO gap is relatively small, the HOMO having Pt-
Pt bonding but Pt-M antibonding character and the
LUMO having mostly CO π* character. It is the small
HOMO-LUMO gap which gives rise to the unusual
blue or green colors of 10 or 11. The orbital 2a1 is
mostly localized on the platinum clusters and has
relatively little 6s character of the central metal. Two
consequences may be envisioned. First, it may be
possible to oxidize the clusters by two electrons, for
example to give the mercury(II) analog with the orbital
2a1′ vacant. We note in this respect that the isoelec-
tronic gold(I) cluster [Au{Pt3(µ-CO)3L3}2]+, L ) PPh3,
The 6s orbital interacts with the symmetrical combi-
nation of Pt3 bonding MO’s of a1′ symmetry (mixed with
2
combinations of orbitals of same symmetry having dz
and dxz, dyz character to improve overlap) to give bonding
and antibonding combinations 1a1′ and 2a1′, respectively
(Chart 3, A, B). Since both orbitals are filled, this would
normally be unfavorable. However, there is also overlap
of the mercury 6s orbital with empty cluster MO’s
(16) Mingos, D. M. P.; Slee, T. J . Organomet. Chem. 1990, 394, 679.
+
+

Hexaplatinum Clusters
Organometallics, Vol. 15, No. 14, 1996 3121
Ta ble 3. Atom ic Ch a r ge on th e Cen tr a l Meta l M
timum stability of the cryptand clusters [M{Pt3(µ-
CO)3(µ-Ph2P(CH2)xPPh2)3]n+ containing a single “guest”
atom or ion Mn+ appears to occur when the bridging
diphosphine is dppp or dppb (x ) 3 or 4).
[Z(M)/e] a n d P t-M a n d P t-P t Over la p P op u la tion s
fr om EHMO Ca lcu la tion s on P t6Mn
)
[P t6(µ6-M)(µ-CO)6(P H3)6]n
CVEa
cluster
Z(M)
OP(Pt-M)
OP(Pt-Pt)
96
94
96
94
96
94
96
94
Pt6Au-
P t6Au +
P t6Hg0
Pt6Hg2+
P t6Tl+
0.02
0.17
0.05
0.18
0.95
1.06
1.83
1.89
0.11
0.12
0.16
0.17
0.16
0.18
0.18
0.20
0.18
0.17
0.17
0.15
0.16
0.15
0.16
0.14
Exp er im en ta l Section
All reactions were carried out under N2 atmosphere using
Schlenk techniques. The compound [PtCl2(Me2S)2] was pre-
pared by the literature method.18 IR spectra were recorded
by using a Perkin-Elmer 2000 spectrometer; the NMR spectra
were recorded by using a Varian Gemini-300 spectrometer. The
proton and carbon chemical shifts are referenced to external
Me4Si, and 31P{1H} and 195Pt{1H} NMR spectra are externally
referenced to 85% H3PO4 and a saturated D2O solution of
K2PtCl4.
Pt6Tl3+
Pt6Pb2+
Pt6Pb4+
a
This gives the cluster valence electron (CVE) counts, counting
12e and 10e for Mn having electron configurations 5d106s2 and 5d10
,
respectively. If the 5d10 electrons are considered core electrons,
the corresponding electron counts would be 86 and 84. The entries
in bold correspond to known complexes. In the calculations, the
distances Pt-M and Pt-Pt are constant at 2.84 and 2.63 Å,
respectively.
EDX analyses were conducted for 10, 11, and 13. The
results showed that the ratio Pt:Hg ) 6 in 10a -d , Pt:Tl ) 6
in 11a ,b, and Pt:Au ) 3:1 in 13.
[P t6(µ-CO)6(µ-d p p e)3], 2b. To a solution of [PtCl2(Me2S)2]
(210 mg, 0.527 mmol) in THF (80 mL) was added dppe (105
mg, 0.27 mmol). CO was bubbled through the solution 20 min,
and excess NaBH4 (279 mg) was then added to the stirred
solution. The color of the solution changed through yellow-
green, green, dark green, and finally to deep red over ca. 1 h.
After 8 h of stirring under a steady CO flow, MeOH (20 mL)
was added dropwise and effervescence occurred. The color of
the solution turned to bright red. The solution was allowed
to stir overnight under a CO atmosphere, and a deep red-
purple solution was obtained. The solvent was then removed
by vacuum, and the red-purple residue was extracted with
CH2Cl2 (20 mL × 8), the solution was filtered through Celite
to remove suspended material, and the solvent was evaporated
to give the product as a brown-red solid in 80% yield. Well-
shaped crystals could be obtained from CH2Cl2/diethyl ether
or CH2Cl2/hexane. Anal. Calcd for C84H72O6P6Pt6: C, 39.8;
H, 2.9. Found: C, 40.2; H, 2.4. IR (Nujol): v(CO) ) 1840 (m),
1796 (vs), 1780 (vs, sh), 1731 (s) cm-1
is already known.17 Perhaps the auride ion Au- might
be stabilized by two-electron reduction in this case.
Results from the EHMO calculations for 96-electron
(2a1′ filled) and 94-electron clusters (2a1′) vacant are
given in Table 3. Note that, for each central metal M,
the charge on M increases only 0.06-0.15 e on going
from the 96e to 94e cluster, as expected if the orbital
2a1 has relatively little M character, and that the Pt-M
and Pt-Pt overlaps are higher and lower, respectively,
for the 94e clusters. The data in Table 3 do suggest
that, perhaps when M ) Au or Hg where the charges
are not excessive, both 96e and 94e clusters might be
prepared. Unfortunately, the Pt3(µ-CO)3 units are not
very robust and chemical oxidation/reduction has not
given isolable products. Electrochemical studies would
be valuable for further study. It is noteworthy that the
mean Pt-Pt and Pt-M distances are slightly shorter
and considerably longer for the 96e 11a [mean Pt-Pt
) 2.670, Pt-Tl ) 2.931 Å] compared to the 94e cluster
[Au{Pt3(µ-CO)3L3}2]+, L ) PPh3 [mean Pt-Pt ) 2.683,
Pt-Au ) 2.728 Å],17 respectively, as expected if the 2a1′
MO is Pt-Pt bonding but Pt-M antibonding. Second,
there is still a question of whether the “linear” or
“angular” M(Pt3)2 units is intrinsically more stable. In
the real complexes, the linear structures are probably
necessary on steric grounds, but the smaller substitu-
ents in the model compounds allow a theoretical study.
To minimize steric effects on going to the angular
structure, we have used the staggered trigonal anti-
prism as the base structure and then changed the angle
between the centroids of the Pt3 triangles and the
central mercury(0) or thallium(I) atom from 180° through
to 150°. The calculations do indicate that the HOMO
is stabilized by this deformation as the “lone pair” has
less antibonding character (compare B and F , Chart 3).
However, this effect is more than offset by several other
minor but unfavorable changes and the calculations
indicate that the linear structure is most stable (the
calculated energy difference is about 0.3 eV for a 20°
distortion).
. NMR in CD2Cl2 at 22
oC: δ(1H) ) 3.43 [s, br, 12H, CH2CH2 of µ-dppe]; δ(13C) ) 246.6
[s, br, µ-CO] in the presence of free 13CO; after removal of
1
excess CO, δ(13C) ) 248.1 [quintet, 1:8:18:8:1, J (PtC) ) 709
1
2
Hz, µ-CO]; δ(31P) ) 53.2 [s, J (PtP) ) 4814 Hz, J (PtP) ) 527
Hz, J (PP) ) 64 Hz, dppe]. NMR in CD2Cl2 at -90 C: δ(1H)
3
o
) 4.35 [s, br, 12H, CH2CH2 of µ-dppe]; δ(13C) ) 250.6 [quintet,
1:8:18:8:1, J (PtC) ) 697 Hz, µ-CO]; δ(31P) ) 55.0 [s, J (PtP)
1
1
2
3
) 4796 Hz, J (PtP) ) 516 Hz, J (PP) ) 64 Hz, dppe].
[P t6(µ-CO)6(µ-d p p p )2(d p p p )2], 9. To a solution of [PtCl2-
(Me2S)2] (238 mg, 0.610 mmol) in THF (80 mL) was added dppp
(168 mg, 0.407 mmol). The mixture was then treated with
CO and NaBH4 (323 mg) as above. The color of the solution
changed through green, dark green, black, and finally to red.
After 8 h of stirring under CO, MeOH (20 mL) was added, the
solution was allowed to stir overnight under CO, the solvent
was evaporated, and the brown-red solid was extracted with
CH2Cl2. The dark-red solution was filtered through Celite and
the solvent evaporated under vacuum to give the product as a
brown-red solid in 85% yield. Anal. Calcd for C114H104O6P8-
Pt6: C, 45.8; H, 3.5. Found: C, 45.7; H, 3.5. IR (Nujol): v(CO)
) 1835 (m), 1783 (vs, sh), 1772 (vs, sh), 1761 (vs) cm-1. NMR
o
3
in CD2Cl2 at 22 C: δ(1H) ) 3.41 [s, br, 8H, J (PtH) ) 51 Hz,
CH2a CH2CH2 of µ-dppp], 3.09 [s, br, 4H, CH2CH2bCH2 of
a
dppp], 2.66 [s, br, 8H, 3J (PtH) ) 50 Hz, CH2cCH2CH2 of
c
chelating dppp], 1.96 [s, br, 4H, CH2CH2d CH2 of dppp]; δ(13C)
) 256.1 [quintet, 1:8:18:8:1, J (Pt2C2) ) 791 Hz, µ-CO], 233.2
1
[quintet of triplet, 1:8:18:8:1, J (Pt2C1) ) 790 Hz, J (Pt1C1) )
1
1
In conclusion, the linear M(Pt3)2 structure appears to
be preferred for metal atoms or ions M with both the
5d10 and 5d106s2 electron configurations and easy redox
reactions between these states may be predicted. Op-
628 Hz, 2J (Pt2C) ) 24 Hz, µ-CO]; δ(31P) ) 56.5 [t, Pb, 1J (Pt2Pb)
) 4682 Hz, J (Pt1Pb) ) 534 Hz, J (Pt2Pb) ) 445 Hz, J (PbPb′)
2
2
3
) 60 Hz, J (PaPb) ) 19 Hz], 25.5 [t, Pa, J (Pt1Pa) ) 3425 Hz,
3
1
2J (Pt2Pa) ) 294 Hz, J (PaPb) ) 19 Hz]; δ(195Pt) ) -2220 [tm,
3
Pt1, 1J (Pt1Pa) ) 3425 Hz, 2J (Pt1Pb) ) 536 Hz, 1J (Pt1Pt2) ) 386
(18) Roulet, R.; Barbey, C. Helv. Chim. Acta 1973, 56, 2179.
(17) Hallam, M. F.; Mingos, D. M. P.; Adatia, T.; McPartlin, M. J .
Chem. Soc., Dalton Trans. 1988, 335.
+
+

3122 Organometallics, Vol. 15, No. 14, 1996
Hao et al.
Hz], -2424 [tm, Pt2, 1J (Pt2Pb) ) 4682 Hz, 2J (Pt2Pb′) ) 445 Hz,
(m), 1838 (s, sh), 1815 (vs, br) cm-1. The NMR data are the
same as for 11a [P F 6].
1
2J (Pt2Pa) ) 290 Hz, J (Pt1Pt2) ) 386 Hz]. NMR in CD2Cl2 at
o
-90 C: δ(1H) showed five broad and overlapped resonances
Rea ction of 10a w ith TlP F 6. To a solution of 10a (45 mg,
0.016 mmol) in CD2Cl2 (0.5 mL) was added TlPF6 (5.7 mg).
The yellow-green solution turned to blue over several hours.
The reaction took ca. 1 day to complete as monitored by 31P
NMR spectroscopy. The analogous reaction of 11a [PF6] with
Hg gave no detectable reaction over a period of 8 h as
monitored by 31P NMR spectroscopy, and only slow decomposi-
tion occurred over 2 days.
[P t6(µ6-Hg)(µ-CO)6(µ-d p p b)3], 10b. To a solution of [PtCl2-
(Me2S)2] (192 mg, 0.492 mmol) in THF (90 mL) was added dppb
(105 mg). The solution was saturated with CO, and then
NaBH4 (260 mg) was added to the stirred solution. The color
of the solution changed slowly to yellow-brown, then brown,
and finally red in ca. 1 h. After 8 h of stirring under CO,
MeOH (20 mL) was added and the solution was stirred
overnight under a CO atmosphere. Excess mercury (8 drops)
was added to the dark-red solution, and the color changed to
yellow-green. The mixture was stirred for another 48 h under
CO. The solvent was then evaporated under vacuum, and the
residue was extracted with CH2Cl2 (20 mL × 8), the solution
was filtered through Celite to give a yellow-green solution, and
the solvent was evaporated under vacuum to give the product
as a yellow-green solid in 76% yield. Well-shaped crystals
could be obtained from CH2Cl2/diethyl ether. Anal. Calcd for
at 3.82, 2.92, 2.78, 2.15, and 2.07; δ(31P) ) 57.6 [s, br, Pc,
1J (Pt2Pc) ) 4562 Hz, 2J (Pt1Pc) ) 476 Hz], 32.3 [s, br, Pa,
1J (Pt1Pa) ) 3618 Hz, 2J (Pt2Pa) ) 362 Hz], 23.1 [s, br, Pb,
2
1J (Pt1Pb) ) 3230 Hz, J (Pt2Pb) ) 206 Hz].
[P t6(µ3-Au P P h 3)2(µ-CO)6(µ-d p p e)3][P F 6]2, 13. This clus-
ter was synthesized by reaction of [Pt6(µ-CO)6(µ-dppe)3] (0.3
mmol) in CH2Cl2 (10 mL) with Ph3PAuPF6 (0.6 mmol) in THF
(10 mL). The dark-green crystalline product 13 was obtained
in 46% yield. Well-shaped black crystals could be obtained
from CH2Cl2/diethyl ether. Anal. Calcd for C120H102Au2-
F
12O6P10Pt6: C, 38.5; H, 2.75. Found: C, 38.7; H, 2.5. IR
(Nujol): v(CO) ) 1833 (vs) cm-1
.
NMR in CD2Cl2 at 22 oC:
δ(1H) ) 3.06 [s, br, CH2CH2 of µ-dppe]; δ(31P) ) 67.2 [s, br,
PPh3], 44.8 [m, br, 1J (PtP) ) 4844 Hz, dppe]. NMR in CD2Cl2
at -90 C: δ(31P) ) 70.0 [quintet, 1:4:7:4:1, 2J (PtP) ) 272 Hz,
o
PPh3], 45.7 [s, 1J (PtP) ) 4792 Hz, 2J (PtP) ) 320 Hz, dppe].
R ea ct ion of [P t6(µ-CO)6(µ-d p p p )2(d p p p )2] w it h P h 3-
P Au P F 6. The reaction was conducted as above, but a complex
mixture of products was formed.
[P t6(µ6-Hg)(µ-CO)6(µ-d p p p )3], 10a . To a solution of 9 (65
mg, 0.025 mmol) in CH2Cl2 (30 mL) was added excess mercury
(1 drop). After 24 h of stirring, a black-green solution was
obtained. The solution was filtered to remove unreacted
mercury, and the product was obtained by evaporation of the
solvent. It was purified by washing with acetone (0.5 mL) and
diethyl ether (6 mL). Yield: 75%. Well-shaped black-green
crystals could be obtained from CH2Cl2/diethyl ether or
C
90H84HgO6P6Pt6: C, 38.35; H, 3.0. Found: C, 39.1; H, 3.2.
EDX: Pt:Hg ) 6:1. IR (Nujol): v(CO) ) 1864 (s), 1828 (s),
1802 (s), 1791 (s, sh), 1778 (s, sh) cm-1. NMR in CD2Cl2 at 22
oC: δ(1H) ) 2.93 [s, br, 12H, 3J (PtH) ) 55 Hz, CH2a CH2bCH2-
a
a
bCH2 of µ-dppb], 2.18 [s, br, 12H, CH2aCH2bCH2bCH2 of
ClCH2CH2Cl/diethyl ether. Anal. Calcd for C87H78
-
µ-dppb]; δ(13C) ) 238.0 [m, 1J (PtC) ) 715 Hz, 2J (PtC) ) 41
HgO6P6Pt6: C, 37.6; H, 2.8. Found: C, 38.1; H, 2.9. EDX:
Hz, µ-CO]; δ(31P) ) 46.6 [s, J (PtP) ) 4919 Hz, J (PtP) ) 396
Hz, 3J (PP) ) 56 Hz, 2J (HgP) ) 56 Hz, 1J (PtPt) ) 1500 Hz,
1J (PtHg) ) 3100 Hz, dppb].
1
2
Pt:Hg ) 6:1. IR (Nujol): v(CO) ) 1837 (s), 1799 (vs), 1782
o
(vs) cm-1. NMR in CD2Cl2 at 22 C: δ(1H) ) 2.94 [s, br, 12H,
a
3J (PtH) ) 52 Hz, CH2a CH2CH2 of dppp], 2.37 [s, br, 6H,
CH2CH2bCH2 of dppp]; δ(13C) ) 235.1 [quintet of triplet, 1:8:
[P t6(µ6-Tl)(µ-CO)6(µ-d p p b)3][P F 6], 11b. The same proce-
dure as above was followed with the use of 1 equiv TlPF6
instead of excess mercury. The green product was isolated in
78% yield. Anal. Calcd for C90H84F6O6P7Pt6Tl: C, 34.1; H,
2.7. Found: C, 33.5; H, 2.3. IR: ν(CO) ) 1808 (s), 1776 (s)
1
2
3
18:8:1, J (PtC) ) 714 Hz, J (PtC) ) 41 Hz, J (PtC) ) 22 Hz,
2J (HgC) ) 75 Hz, CO]; δ(31P) ) 58.6 [s, 1J (PtP) ) 5037 Hz,
2J (PtP) ) 403 Hz, J (HgP) ) 41 Hz, J (PP) ) 55 Hz, µ-dppp];
2
3
δ(195Pt) ) -2532 [dm, J (PtP) ) 5037 Hz, J (PtP) ) 403 Hz,
1
2
1J (HgPt) ) 3350 Hz, J (PtPt) ) 2200 Hz, J (PtPt) ) 100 Hz].
[P t6(µ6-Tl)(µ-CO)6(µ-d p p p )3][P F 6], 11a [P F 6]. To a solu-
tion of 9 (47 mg, 0.018 mmol) in CH2Cl2 (20 mL) was added
TlPF6 (6.4 mg). The red solution changed to dark green
immediately. After 8 h of stirring, the solution was filtered
and the solvent was removed under vacuum. The crude
product was recrystallized from CH2Cl2/diethyl ether and
washed with diethyl ether (6 mL) to give the pure product as
a blue powder in 82% yield. Anal. Calcd for C87H78F6O6P7-
Pt6Tl: C, 35.7; H, 2.7. Found: C, 35.4; H, 2.8. EDX: Pt:Tl )
6:1. IR (Nujol): v(CO) ) 1871 (m), 1818 (vs, br) cm-1. NMR
cm-1. NMR in CD2Cl2: δ(31P) ) 47 [v br, J (PtP) ) 5050 Hz,
1
2
1
dppb].
[P t6(µ6-Hg)(µ-CO)6(CO)2(µ-d p p p e)2], 12a . The same pro-
cedure was followed as above with the use of ligand dpppe
instead of dppb. The product obtained is a green solid, which
contained ca. 5% impurity of 10c, in 76% yield. IR: ν(CO) )
2030 (m), terminal CO, 1826 (m), 1794 (s), bridging CO. NMR
in CD2Cl2: δ(13C) ) 196 [m, 1J (PtC) ) 2295 Hz, 2J (PtC) ) 126
Hz, t-CO]; 225 [m, 1J (PtC) ) 886, 595 Hz, µ-CO]; 235 [m,
1J (PtC) ) 735 Hz, µ-CO]. δ(31P) ) 50.2 [s, 1J (PtP) ) 4960 Hz,
2J (Pt1P) ) 447 Hz, 2J (Pt2P) ) 371 Hz, 3J (PP) ) 46 Hz, dpppe];
δ(195Pt) ) -2324 [m, 1J (Pt1Pt2) ) 2225 Hz, 2J (Pt1P) ) 447 Hz,
Pt1]; Pt2 resonance not observed.
in CD2Cl2 at 22 C: δ(1H) ) 3.05 [s, br, 12H, 3J (PtH) ) 53 Hz,
o
CH2a CH2CH2 of dppp], 2.23 [s, br, 6H, CH2CH2bCH2 of dppp];
a
δ(13C) ) 232.5 [quintet, 1:8:18:8:1, 1J (PtC) ) 696 Hz, CO];
δ(31P) ) 56.7 [d, 1J (PtP) ) 4808 Hz, 2J (PtP) ) 360 Hz, 2J (TlP)
) 335 Hz, µ-dppp].
[P t6(µ6-Hg)(µ-CO)6(µ-d p p p e)3], 10c. To a solution of 12a
(79 mg, 0.031 mmol) in CH2Cl2 (20 mL) was added dpppe
(0.035 mmol). The green solution turned to yellow-green
immediately. After 0.5 h of stirring, the solvent was removed
to obtain a green solid which was recrystallized from CH2Cl2/
hexane to give the product as an green solid in 82% yield. Anal.
Calcd for C93H90O6P6Pt6Hg: C, 39.05; H, 3.2. Found: C, 39.4;
H, 3.3. EDX: Pt:Hg ) 6:1. IR (Nujol): v(CO) ) 1830 (m),
[P t6(µ6-Tl)(µ-CO)6(µ-d p p p )3][BP h 4], 11a [BP h 4]. To a
solution of 11a [P F 6] (105 mg, 0.036 mmol) in a mixture of
CH2Cl2/acetone (5 mL, ratio 1:2) was added a solution of
NaBPh4 (74 mg, 0.215 mmol) in methanol (5 mL). The green
precipitate of the product which formed immediately was
collected by filtration and washed with methanol (1 mL) and
ether (5 mL). Yield: 84%. Well-shaped crystals could be
obtained from CH2Cl2/diethyl ether. Anal. Calcd for
o
1788 (vs) cm-1. NMR in CD2Cl2 at 22 C: δ(1H) ) 3.02 [s, br,
12H, J (PtH) ) 53 Hz, CH2a CH2bCH2cCH2bCH2 of µ-dpppe],
3
a
2.04 [s, br, 12H, CH2aCH2bCH2cCH2bCH2 of µ-dpppe], 1.53 [s,
a
a
C
111H98BO6P6Pt6Tl: C, 43.0; H, 3.2. Found: C, 43.3; H, 3.4.
br, 6H, CH2aCH2bCH2cCH2bCH2 of µ-dpppe]; δ(13C) ) 235
IR (Nujol): v(CO) ) 1870 (m), 1825 (vs, br) cm-1. The NMR
spectroscopic data are the same as for 11a [P F 6].
[1J (PtC) ) 720 Hz, µ-CO]; δ(31P) ) 53.2 [s, 1J (PtP) ) 4980 Hz,
3
2
2J (PtP) ) 403 Hz, J (PP) ) 55 Hz, J (HgP) ) 76 Hz, dpppe];
δ(195Pt) ) -2497 [m, 1J (PtP) ) 4980 Hz, 2J (PtP) ) 400 Hz].
[P t6(µ6-Hg)(µ-CO)6(CO)2(µ-d p p h )2], 12b. The same pro-
cedure was followed as for 2a with the use of ligand dpph
instead of dpppe. The product was obtained as a green solid,
containing the impurity of 10d . The yield of 12b is ca. 30%
[P t6(µ6-Tl)(µ-CO)6(µ-d p p p )3][AsF 6], 11a [AsF 6]. The same
procedure was followed as above with the use of KAsF6 instead
of NaBPh4, to give the product as a black-green solid. Yield:
71%. Anal. Calcd for C87H78AsF6O6P6Pt6Tl: C, 35.2; H, 2.65.
Found: C, 35.0; H, 2.4. IR (Nujol): v(CO) ) 1888 (m), 1869
+
+

Hexaplatinum Clusters
Organometallics, Vol. 15, No. 14, 1996 3123
yield (based on the integration of the 31P NMR spectra). IR:
ν(CO) ) 2030 (m), terminal CO; 1830 (m), 1790 (s), bridging
CO. NMR in CD2Cl2: δ(13C) ) 196 [m, 1J (PtC) ) 2160 Hz,
Ta ble 4. Cr ysta l Da ta a n d Exp er im en ta l Deta ils
for 11a [BP h 4]
formula
fw
C112H100BCl2O6P6Pt6Tl
3184.36
3
1
2J (PtC) ) 125 Hz, J (PC) ) 16 Hz, t-CO]; 225 [m, J (PtC) )
860, 540 Hz, µ-CO]; 235 [m, 1J (PtC) ) 705 Hz, µ-CO]; δ(31P) )
temperature
wavelength
cryst system
space group
cell dimens
23 °C
0.710 73 Å
monoclinic
P21/n
a ) 16.571(4) Å
b ) 26.030(6) Å
c ) 27.447(6) Å
â ) 95.80(2)°
11778(4) Å3, 4
1.796, 1.90(5) g‚cm-3
8.635 mm-1
1
2
2
49.2 [s, J (PtP) ) 4975 Hz, J (Pt1P) ) 430 Hz, J (Pt2P) ) 370
Hz, 3J (PP) ) 50 Hz, dpppe]; δ(195Pt) ) -2271 [m, 1J (Pt1Pt2) )
2070 Hz, J (Pt1P) ) 460 Hz, Pt1]; -2695 [br, J (PtP) ) 5100
2
1
Hz, Pt2].
[P t6(µ6-Hg)(µ-CO)6(µ-d p p h )3], 10d . The same procedure
as for 10c was followed. The product was obtained as a green
solid in 84% yield. Anal. Calcd for C96H96HgO6P6Pt6: C, 39.7;
H, 3.3. Found: C, 39.9; H, 3.7. EDX: Pt:Hg ) 6:1. IR
cell vol, Z
density: calcd, obsd
abs coeff
(Nujol): v(CO) ) 1823 (m), 1781 (vs) cm-1. NMR in CD2Cl2
c
at 22 oC: δ(1H) ) 3.25 [s, br, 12H, CH2a CH2bCH2cCH2
-
-
data/restraints/params
goodness-of-fit (GooF)a on F2
final R indices [I > 2σ(I)]a
12440/61/443
1.000
R1 ) 0.0999, wR2 ) 0.2312
a
b
CH2bCH2 of µ-dpph], 2.15 [s, br, 24H, CH2aCH2bCH2cCH2cCH2
a
1
CH2 of µ-dpph]; δ(13C) ) 232.7 [quintet, 1:8:18:8:1, J (PtC) )
744 Hz, CO]; δ(31P) ) 52.9 [s, 1J (PtP) ) 4961 Hz, 2J (PtP) )
397 Hz, 3J (PP) ) 56 Hz, 2J (HgP) ) 70 Hz, dpph]; δ(195Pt) )
a
2
4 1/2
R1 ) ∑(||Fo| - |Fc||)/∑|Fo|; wR2 ) [∑w(Fo - Fc2)2/∑wFo
] ;
GooF ) [∑w(Fo - Fc2)2/(n - p)]1/2, where n is the number of
reflections and p is the number of parameters refined.
2
1
2
1
-2476 [m, J (PtP) ) 4960 Hz, J (PtP) ) 400 Hz, J (PtPt) )
1
2100 Hz, J (PtHg) ) 3100 Hz].
SHELXL-93 software programs.21 Anisotropic thermal pa-
rameters were refined for all the platinum, thallium, phos-
phorus, and chlorine atoms. The rest of the atoms were refined
isotropically. The phenyl rings were treated like ideal hexa-
gons with C-C ) 1.395 Å. The dichloromethane solvent was
found to occupy two different positions in the crystal lattice
with occupancies 0.6/0.4. Soft constraints were applied for the
distances with similar C-Cl, B-C, and aliphatic C-C bonds
using SADI. No attempt was made to locate the hydrogen
atoms. However all the hydrogen atoms were placed in the
calculated positions in the cation and their isotropic thermal
parameters were assigned 20% more than the atoms they are
attached to. In the final least-squares refinement cycles on
F2, the model converged at R1 ) 0.0999, wR2 ) 0.2312, and
Goof ) 1.000 for 4825 observations with Fo g 4σ(Fo) and 443
parameters and R1 ) 0.2652 and wR2 ) 0.3336 for all 12 440
data. Poor agreement factors are attributed to weak reflec-
tions caused by crystal decay. In the final difference Fourier
synthesis the electron density fluctuates in the range 2.13 to
Deta ils of th e X-r a y Str u ctu r e Deter m in a tion . Very
dark crystals (small crystals are green by reflected light but
red by transmitted light) of 11a [BPh4] were grown from
CH2Cl2/ether by slow diffusion at room temperature. After
many attempts with weakly diffracting crystals, a suitable
crystal fragment with dimensions 0.35 × 0.25 × 0.20 mm was
mounted inside a Lindemann capillary tube and flame sealed.
The diffraction experiments were carried out using a Siemens
P4 diffractometer with XSCANS software package19 using
graphite monochromated Mo KR radiation at 23 °C. The cell
constants were obtained by centering 25 reflections (15.4 e
2θ e 25°). The Laue symmetry was determined by merging
symmetry equivalent reflections. The data were collected in
shells, in the θ range 2.0-21° (-1 e h e 16, -1 e k e 26, -27
e l e 27) in ω scan mode at variable scan speeds (2-20 deg/
min). Background measurements were made at the ends of
the scan range. Four standard reflections were monitored at
the end of every 297 reflection collection. It was found that
the data crystal was very unstable to X-rays and the intensity
of the standards fell during the data collection (correction
factors: minimum, 0.547; mean, 0.851). The data collection
was therefore stopped near the θmax of 21°. At the end of the
data collection, the faces of the data crystal were indexed, and
the distances between the faces were measured. A Gaussian
absorption correction was applied to the data (µ ) 8.635 mm-1).
The maximum and minimum transmission factors are 0.2868
and 0.1629, respectively. The space group P21/n was deter-
mined from the systematic absences. The data processing,
solution, and initial refinements were done using SHELXTL-
PC programs.20 The final refinements were performed using
-2.34 e Å-3
. Most of peaks in the difference Fourier map
were found near the heavy atoms. The mean and the
maximum shift/esd values in the final cycles were 0.004 and
-0.040. The experimental details and crystal data are in
Table 4. Positional and thermal parameters, complete bond
distances and angles, anisotropic thermal parameters, hydro-
gen atom coordinates, and selected torsion angles have been
included in the Supporting Information.
Ack n ow led gm en t. We thank the NSERC (Canada)
for financial support.
Su p p or tin g In for m a tion Ava ila ble: Tables of data col-
lection parameters, atomic coordinates, thermal parameters,
and bond distances and angles (14 pages). Ordering informa-
tion is given on any current masthead page.
(19) XSCANS, Siemens Analytical X-Ray Instruments Inc., Madison,
WI. 1990.
(20) Sheldrick, G. M. SHELXTL-PC Software, Siemens Analytical
X-Ray Instruments Inc., Madison, WI, 1990.
(21) Sheldrick, G. M. SHELXL-93, Institut fur Anorg. Chem.,
Goettingen, Germany, 1993.
OM960148K
+
+