Substituted metal carbonyls. 27.1 Synthesis, structures, and metal-metal bonding of a ferrocenylphosphine exo-bridged cluster with two heterometallic triangles, [AuMn2(CO)8(μ-PPh2)]2(μ-dppf), and a twisted-bowtie cluster, PPN+[Au{Mn2(CO)8(μ-PPh2)} 2]-

Article abstract of DOI:10.1021/om950797g

Redox condensation of PPN[Mn₂(CO)₈(μ-PPh₂)] (1; (PPN = N(PPh₃)₂) with Au₂Cl₂(μ-P-P) (P-P = (C₅H₄PPh₂)₂Fe (dppf), Ph₂PC₂H₄PPh₂ (dppe)) gives two hexanuclear Au-Mn clusters [AuMn₂(CO)₈(μ-PPh₂)]₂(μ-P-P) (P-P = dppf, (2), dppe (4)), both of which contain a diphosphine bridging two Mn₂Au triangles. Complex 2 is formed via an intermediate, AuCl-(μ-dppf)[AuMn₂(CO)₈(μ-PPh₂)], (3), which was isolated. Bridge cleavage of 2 occurs at thf reflux with PPh₃ and room temperature with P(OEt)₃ to give the triangular clusters [(PR₃)-AuMn₂(CO)₈(μ-PPh₂)] (R = Ph (5), OEt (6)), respectively. The latter exchange of dppf with P(OEt)₃ is reversible in solution. Condensation of 1 with AuCl(SMe2) gives an anionic pentanuclear cluster, PPN[Au{Mn₂(CO)₈(μ-PPh₂)}₂] (7). Complexes 2 and 7 were structurally characterized by single-crystal X-ray diffractometry. Complex 2, which is centrosymmetric with Fe in dppf at a crystallographic inversion center, consists of a ferrocenylphosphine bridging two heterometallic triangles (Au-Mn = 2.660(1) and 2.776(1) A?; Mn-Mn = 3.049(2) A?). Complex 7 is made up of two planar AuMn2P metallacycles fused at Au at an angle of 85.50(4)°. With crystallographic C₂ symmetry, a twisted-bowtie skeleton resulted with gold at its center. Both Au-Mn (mean 2.806(1) A?) and (PPh₂-bridged) Mn-Mn (3.105(2) A?) lengths are significantly longer than those in 2. The Mn-Mn bond of 2 is also significantly longer than that of 1. Fenske-Hall MO calculations on 1, 2, and 7 together with Mn₂(CO)₈(μ-H)(μ-PPh₂) (8) and (PPhMe₂)AuMn₂(CO)₈(PPh₂) (9) indicate that aside from 1, all the complexes, including 2 and 7, give a negative overlap population in the Mn-Mn interactions. The Mn-Mn distance appears to be determined by the strength of the AuMn₂ interaction and/or the size of H compared to Au. The weaker Mn-Mn and Au-Mn interactions in 7 (as compared to those in 2 and 9, respectively) are likely to be caused by the absence of Au orbital reinforcement in the direction of the Mn₂ moiety as a consequence of symmetry.

Full text of DOI:10.1021/om950797g

Organometallics 1996, 15, 2595-2603
2595
Su bstitu ted Meta l Ca r bon yls. 27.1 Syn th esis, Str u ctu r es,
a n d Meta l-Meta l Bon d in g of a F er r ocen ylp h osp h in e
exo-Br id ged Clu ster w ith Tw o Heter om eta llic Tr ia n gles,
[Au Mn 2(CO)8(µ-P P h 2)]2(µ-d p p f), a n d a Tw isted -Bow tie
Clu ster , P P N+[Au {Mn 2(CO)8(µ-P P h 2)}2]- (d p p f )
1,1′-Bis(d ip h en ylp h osp h in o)fer r ocen e)
Pauline M. N. Low, Agnes L. Tan,† and T. S. Andy Hor*
Department of Chemistry, Faculty of Science, National University of Singapore,
Kent Ridge, Singapore 119260
Yuh-Sheng Wen and Ling-Kang Liu*
Institute of Chemistry, Academia Sinica, Taipei 11529, Taiwan, ROC
Received October 10, 1995X
Redox condensation of PPN[Mn2(CO)8(µ-PPh2)] (1; (PPN ) N(PPh3)2) with Au2Cl2(µ-P-P)
(P-P ) (C5H4PPh2)2Fe (dppf), Ph2PC2H4PPh2 (dppe)) gives two hexanuclear Au-Mn clusters
[AuMn2(CO)8(µ-PPh2)]2(µ-P-P) (P-P ) dppf, (2), dppe (4)), both of which contain a
diphosphine bridging two Mn2Au triangles. Complex 2 is formed via an intermediate, AuCl-
(µ-dppf)[AuMn2(CO)8(µ-PPh2)], (3), which was isolated. Bridge cleavage of 2 occurs at thf
reflux with PPh3 and room temperature with P(OEt)3 to give the triangular clusters [(PR3)-
AuMn2(CO)8(µ-PPh2)] (R ) Ph (5), OEt (6)), respectively. The latter exchange of dppf with
P(OEt)3 is reversible in solution. Condensation of 1 with AuCl(SMe2) gives an anionic
pentanuclear cluster, PPN[Au{Mn2(CO)8(µ-PPh2)}2] (7). Complexes 2 and 7 were structurally
characterized by single-crystal X-ray diffractometry. Complex 2, which is centrosymmetric
with Fe in dppf at a crystallographic inversion center, consists of a ferrocenylphosphine
bridging two heterometallic triangles (Au-Mn ) 2.660(1) and 2.776(1) Å; Mn-Mn )
3.049(2) Å). Complex 7 is made up of two planar AuMn2P metallacycles fused at Au at an
angle of 85.50(4)°. With crystallographic C2 symmetry, a twisted-bowtie skeleton resulted
with gold at its center. Both Au-Mn (mean 2.806(1) Å) and (PPh2-bridged) Mn-Mn
(3.105(2) Å) lengths are significantly longer than those in 2. The Mn-Mn bond of 2 is also
significantly longer than that of 1. Fenske-Hall MO calculations on 1, 2, and 7 together
with Mn2(CO)8(µ-H)(µ-PPh2) (8) and (PPhMe2)AuMn2(CO)8(PPh2) (9) indicate that aside from
1, all the complexes, including 2 and 7, give a negative overlap population in the Mn-Mn
interactions. The Mn-Mn distance appears to be determined by the strength of the AuMn2
interaction and/or the size of H compared to Au. The weaker Mn-Mn and Au-Mn
interactions in 7 (as compared to those in 2 and 9, respectively) are likely to be caused by
the absence of Au orbital reinforcement in the direction of the Mn2 moiety as a consequence
of symmetry.
In tr od u ction
of these clusters. One notable example is Os3(CO)10(µ-
AuPPh3)(µ-H),6a which represents an isolobal replace-
Heterometallic gold clusters have attracted attention
because of their unusual bonding and structural proper-
ties.2 The catalytic enhancement of gold in some of
these carbonyl clusters3 provides an added impetus for
research in this area. Many of these species are
synthesized by nucleophilic attack of metal carbonyl
anions on AuCl(PR3)4-6 or condensation of metal car-
bonyl hydrides with AuMe(PR3) or AuCl(PR3), eliminat-
ing CH4 or HCl.7,8 The isolobal relationship of [AuPR3]
with [H] also enables the designed syntheses of many
(2) Mingos, D. M. P.; Wales, D. J . Introduction to Cluster Chemistry;
Prentice-Hall: Englewood Cliffs, NJ , 1990; Chapter 4, p 154. Mingos,
D. M. P.; Watson, M. J . Adv. Inorg. Chem. 1992, 39, 327. Bott, S. G.;
Mingos, D. M. P.; Watson, M. J . J . Chem. Soc., Chem. Commun. 1989,
1192. Teo, B. K.; Keating, K. J . Am. Chem. Soc. 1984, 106, 2224. Teo,
B. K.; Zhang, H.; Shi, X. J . Am. Chem. Soc. 1993, 115, 8489. Teo, B.
K.; Shi, X.; Zhang, H. Inorg. Chem. 1993, 32, 3987. Teo, B. K.; Zhang,
H.; Shi, X. Inorg. Chem. 1994, 33, 4086. Contel, M.; J imenez, J .; J ones,
P. G.; Laguna, A.; Laguna, M. J . Chem. Soc., Dalton Trans. 1994, 2515.
Kappen, T. G. M. M.; Schlebos, P. P. J .; Bour, J . J .; Bosman, W. P.;
Smits, J . M. M.; Beurskens, P. T.; Steggerda, J . J . Inorg. Chem. 1994,
33, 754. J ohnson, A. J . W.; Spencer, B.; Dahl, L. F. Inorg. Chim. Acta
1994, 227, 269. Gould, R. A. T.; Craighead, K. L.; Wiley, J . S.; Pignolet,
L. H. Inorg. Chem. 1995, 34, 2902.
(3) Evans, J .; Gao, J . J . Chem. Soc., Chem. Commun. 1985, 39.
Braunstein, P.; Rose, J . Gold Bull. 1985, 18, 17.
(4) Haines, R. J .; Nyholm, R. S.; Stiddard, M. H. B. J . Chem. Soc. A
1968, 46. Green, M.; Orpen, A. G.; Salter, I. D.; Stone, F. G. A. J . Chem.
Soc., Dalton Trans. 1984, 2897.
(5) (a) Coffey, C. E.; Lewis, J .; Nyholm, R. S. J . Chem. Soc. 1964,
1741. (b) Kasenally, A. S.; Lewis, J .; Manning, A. R.; Miller, J . R.;
Nyholm, R. S.; Stiddard, M. H. B. J . Chem. Soc. 1965, 3407.
* To whom correspondence should be addressed.
† Computational Science Programme, National University of
Singapore.
X Abstract published in Advance ACS Abstracts, May 1, 1996.
(1) Part 26: Low, P. M. N.; Yong, Y. L.; Hor, T. S. A.; Lam, S.-L.;
Chan, K. W.; Wu, C.; Au-Yeung, S. C. F.; Wen, Y.-S.; Liu, L.-K.
Organometallics 1996, 15, 1369.
S0276-7333(95)00797-7 CCC: $12.00 © 1996 American Chemical Society

2596 Organometallics, Vol. 15, No. 11, 1996
Low et al.
9
Sch em e 1
ment of hydride in Os3(CO)18(µ-H)2 by an [Au(PPh3)]
fragment. As a result, pioneering work by Nyholm et
al. in 1964 in which various simple bimetallic complexes
were made by coupling metal carbonylate anions (e.g.
[Mn(CO)5]- and [Co(CO)4]-) with AuCl(PPh3)5a can
simply be rationalized by the isolobal principle. Our
recent isolation of the Au-M-bonded (M ) Mn, Re)
complexes [AuM(CO)5]2(µ-dppf)10 proved that this syn-
thetic methodology can be extended to Au2Cl2(µ-dppf).
In this case, the complex behaves as a useful Au(I)
analogue of a dibasic acid. The synthetic value of this
“acid” is that, since the “protons” are linked by an
electroactive ferrocenyl group, upon “protonation” of two
di-, tri-, or polynuclear metal carbonylates, one can
accomplish a one-step access to ferrocenyl-bridged clus-
ters11 which are molecular analogues of electroactive
materials.12 This novel yet simple approach to ferro-
cenylphosphine exo-bridged heterometallic metal tri-
angles is described in this paper. The precursor PPN-
[Mn2(CO)8(µ-PPh2)] is chosen on the basis of our
experience with [AuMn(CO)5]2(µ-dppf) and the well-
known auration of this anion with [Au(PR3)] fragments
across the Mn-Mn bond without Mn-Mn fragmenta-
tion.13,14 A similar approach in the literature using a
mononuclear carbonylate resulted in the isolation of
simple bimetallics such as (OC)5MnAu(PR3)5 (R ) Ph,
PhO, p-MeO-C6H4) or closed clusters such as [(Ph3-
15
PAu)4Mn(CO)4]BF4 and (Ph3PAu)3Mn(CO)3.15,16 We
shall also describe the use of a labile Au(I) complex such
as AuCl(SMe2) to synthesize a heterometallic compound
which is similar but which has an open structure. The
effects of different isolobal fragments, viz. [Au2(dppf)],
[Au(PR3)], and [H], on the strength of the Mn-Mn bonds
in relation to that of the Au-Mn bonds in the hetero-
metallic complexes synthesized will also be examined
with the aid of theoretical calculations.
(µ-H)(µ-PPh2)]17 by NaBH4), redox condensation resulted
and [AuMn2(CO)8(µ-PPh2)]2(µ-dppf) (2) and AuCl-
(µ-dppf)[AuMn2(CO)8(µ-PPh2)] (3) were isolated in 51%
and 13% yields, respectively. If AgBF4 was omitted and
the addition of the two substrates reversed, viz. by
adding PPN[Mn2(CO)8(µ-PPh2)] to Au2Cl2(µ-dppf), the
yield of 2 remained at 50% but that of 3 increased to
20%. When 1 was added to [Au2(µ-dppf)](BF4) in a 2:1
mole ratio, the yield of 2 increased further to 65% and
that of 3 was negligible. These observations suggested
that 3 is formed as a result of incomplete substitution
of Cl- in Au2Cl2(µ-dppf) by the anionic 1 (Scheme 1)
Both 2 and 3 have been fully characterized by IR and
1H and 31P NMR spectroscopy. The structure of 2 has
also been established by X-ray crystallography.
The 31P NMR spectrum of 2 shows a sharp singlet at
δ 58.02 ppm attributed to the phosphine attached to the
gold atom and a very low-field resonance (δ 212.31 ppm)
characteristic of the bridging phosphido group. The
latter is broadened by the neighboring 55Mn quadrupo-
lar nuclei. Similar features are found in 3 but with an
additional resonance at δ 27.82 ppm, which is consistent
with a pendant Au-P site (cf. Au2Cl2(µ-dppf), δ 28.27
ppm).
Resu lts a n d Discu ssion
When an equimolar amount of [Au2(µ-dppf)](BF4)2
(generated in situ from Au2Cl2(µ-dppf) and AgBF4
(1.0:2.6)) was added to 1 (from reduction of [Mn2(CO)8-
(6) (a) J ohnson, B. F. G.; Kaner, D. A.; Lewis, J .; Raithby, P. R. J .
Organomet. Chem. 1981, 215, C33. (b) J ohnson, B. F. G.; Kaner, D.
A.; Lewis, J .; Raithby, P. R.; Rosales, M. J . J . Organomet. Chem. 1982,
231, C59. (c) J ohnson, B. F. G.; Lewis, J .; Nelson, W. J . H.; Raithby,
P. R.; Vargas, M. D. J . Chem. Soc., Chem. Commun. 1983, 608. (d)
Briant, C. E.; Hall, K. P.; Mingos, D. M. P. J . Chem. Soc., Chem.
Commun. 1983, 843.
(7) Luke, M. A.; Mingos, D. M. P.; Sherman, D. J .; Wardle, R. W.
M. Transition Met. Chem. 1987, 12, 37. Casalnuova, A. L.; Laska, T.;
Nilsson, P. V.; Olofson, J .; Pignolet, L. H. Inorg. Chem. 1985, 24, 233.
(8) Green, M.; Mead, K. A.; Mills, R. M.; Salter, I. D.; Stone, F. G.
A.; Woodward, P. J . Chem. Soc., Chem. Commun. 1982, 51. Bateman,
L. W.; Green, M.; Howard, J . A. K.; Mead, K. A.; Mills, R. M.; Salter,
I. D.; Stone, F. G. A.; Woodward, P. J . Chem. Soc., Chem. Commun.
1982, 773.
(9) Orpen, A. G.; Riera, A. V.; Bryan, E. G.; Pippard, D.; Sheldrick,
G.; Rouse, K. D. J . Chem. Soc., Chem. Commun. 1978, 723.
(10) Low, P. M. N.; Yan, Y. K.; Chan, H. S. O.; Hor, T. S. A. J .
Organomet. Chem. 1993, 454, 205.
(11) Fang, Z.-G.; Wen, Y.-S.; Wong, R. K. L.; Ng, S.-C.; Liu, L.-K.;
Hor, T. S. A. J . Cluster Sci. 1994, 5, 327.
The IR bands in the carbonyl region of 2 and 3 are
similar to those of (PPhR2)AuMn2(CO)8(µ-PPh2) (R )
Ph,13 Me14) and Mn2(CO)8(µ-PPh2)(µ-H).17 This suggests
the maintenance of a [Mn2(CO)8] moiety and the similar
coordination behavior of the [Au2(dppf)] as compared to
(12) Togni, A. In Ferrocenes; Togni, A., Hayashi, T., Eds.; VCH:
Weinheim, Germany, 1995; Chapter 8, p 433.
(13) Haupt, H.-J .; Heinekamp, C.; Flo¨rke, U.; J u¨ptner, U. Z. Anorg.
Allg. Chem. 1992, 608, 100.
(14) Iggo, J . A.; Mays, M. J .; Raithby, P. R.; Hendrick, K. J . Chem.
Soc., Dalton Trans. 1984, 633.
(15) Nicholson, B. K.; Bruce, M. I.; bin Shawkataly, O.; Tiekink, E.
R. T. J . Organomet. Chem. 1992, 440, 411.
(16) Ellis, J . E.; Faltynek, R. A. J . Chem. Soc., Chem. Commun.
1975, 966.
(17) Iggo, J . A.; Mays, M. J .; Raithby, P. R.; Hendrick, K. J . Chem.
Soc., Dalton Trans. 1983, 205.

Substituted Metal Carbonyls
Organometallics, Vol. 15, No. 11, 1996 2597
F igu r e 1. ORTEP drawing of the structure of [AuMn2(CO)8(µ-PPh2)]2(µ-dppf) (2).
discussed by us recently.20 An alternative view of 2 is
Ta ble 1. Cr ysta llogr a p h ic Da ta for
[Au Mn 2(CO)8(µ-P P h 2)]2(µ-d p p f) (2) a n d
that two Mn-Mn-bonded dimers are bridged by a
heterometallic fragment, [Au-dppf-Au]. The phos-
phide bridge is symmetrically disposed (∆(Mn-P) )
0.005(2) Å) but not the gold bridge (∆(Au-Mn )
0.116(1) Å). The asymmetric disposition of the latter
is manifested in the uneven Au-Mn(1,2)-Mn(2,1)
(∆ ) 3.62(3)°) and Au-Mn(1,2)-C(1,5) (∆ ) 6.5(3)°)
angles.
In order to illustrate that this approach to form
ligand-bridged cluster triangles is applicable to other
diphosphines such as Ph2P(CH2)nPPh2, we have simi-
larly obtained [AuMn2(CO)8(µ-PPh2)]2(µ-dppe) (4) from
Au2Cl2(µ-dppe). The isolation of 4 may suggest that
these clusters could tolerate bridges of different chain
lengths.
In a study to examine the strength of the cluster
bridge, a representative phosphine (PPh3) or phosphite
(P(OEt)3) was added in excess to complex 2. Phosphine
displacement readily occurs to give 2 equiv of (PR3)-
AuMn2(CO)8(µ-PPh2) (R ) Ph (5), P(OEt)3 (6)). The
expected ease of bridge cleavage can be explained on
the basis of steric and entropic grounds. The ability for
the incoming PR3 to undergo associative displacement
on a 14-electron gold center also explains the kinetic
facility. The observed displacement of dppf is in con-
trast to the dissociation of [Au(PEt3)]+ (identified as [Au-
(PEt3)2]+) from (PEt3)AuMn2(CO)8(µ-PPh2) in a similar
reaction of the latter with PEt3.14
P P N[Au {Mn 2(CO)8(µ-P P h 2)}2] (7)
2
7
chem formula
fw
C74H48Au2FeMn4O16P4 C76H50AuMn4NO16P4
1987
1774
cryst syst
space group
a, Å
monoclinic
P21/c
monoclinic
C2/c
14.592(2)
14.675(2)
16.658(2)
91.46(1)
3565.9(8)
2
24.332(6)
17.517(4)
17.245(6)
104.29(24)
7123(3)
4
b, Å
c, Å
â, deg
V, Å3
Z
d
T, K
calcd, g cm-3
1.850
1.654
298
298
λ(Mo KR), Å
0.709 30
5.10
0.710 69
2.86
µ, mm-1
no. of obsd rflns 3462
3169
Ra
R′
0.030
0.035
0.031
0.033
b
a
b
2
R ) ∑||Fo| - |Fc||/∑|Fo|. R′ ) (∑w(|Fo| - |Fc|)2/∑w|Fo| )1/2
.
two [H] or [Au(PR3)] fragments. The X-ray single-
crystal diffraction study confirmed the solution findings
and showed that with both gold atoms of Au2Cl2(dppf)
inserted into the Mn-Mn bonds in two anions of 1,
complex 2 is a dimeric analogue of (PR3)AuMn2(CO)8-
(µ-PPh2)13,14 (Figure 1). It is centrosymmetric with the
Fe atom at the center of inversion. The ferrocenyl C5
rings are exactly staggered (τ ) 180°) as a condition of
this inversion. This staggered arrangement of the C5
rings is often but not always found in many open-
bridged dppf complexes.18 Complex 2 can thus be
viewed as two heterometallic AuMn2 triangles singly
bridged by dppf. This concept of linked clusters has
been described elsewhere in which ligands (“fillers”) are
used to link two identical or different cluster frag-
ments.19 Complex 2 is, however, the only example of
heterometallic clusters which are joined similarly. Het-
eropolymetallic oligomers bridged by dppf have been
The isolation of a pure sample of 6 from 2 was plagued
by the ease of phosphite dissociation and the reentry of
adventitious dppf. This reversibility demonstrates a
fine balance among the entropy factors, the kinetic
lability of P(OEt)3, and the preference of gold for a
(19) Cullen, W. R.; Rettig, S. J .; Zheng, T. Organometallics 1992,
11, 928. Le Grand, J .-L.; Lindsell, W. E.; McCullough, K. J .; McIntosh,
C. H.; Meiklejohn, A. G. J . Chem. Soc., Dalton Trans. 1992, 1089.
Tunik, S. P.; Osipov, M. V.; Nikolskyi, A. B. J . Organomet. Chem. 1992,
426, 105. Sappa, E. J . Organomet. Chem. 1988, 352, 327.
(20) Phang, L.-T.; Gan, K.-S.; Lee, H. K.; Hor, T. S. A. J . Chem. Soc.,
Dalton Trans. 1993, 2697.
(18) Gan, K.-S.; Hor, T. S. A. In Ferrocenes; Togni, A., Hayashi, T.,
Eds.; VCH: Weinheim, Germany, 1995; Chapter 1, p 3

2598 Organometallics, Vol. 15, No. 11, 1996
Low et al.
Ta ble 2. Selected Bon d Len gth s (Å) a n d An gles (d eg)
(a) [AuMn2(CO)8(µ-PPh2)]2(µ-dppf) (2)
Au-Mn(1)
2.660(1)
3.049(2)
2.280(2)
1.78(1)
Au-Mn(2)
2.776(1)
Mn(1)-Mn(2)
Mn(1)-P(2)
Mn(1)-C(1)
Mn(1)-C(3)
Mn(2)-C(5)
Mn(2)-C(7)
mean Fe-C
mean C-O(5-8)
Au-P(1)
2.318(2)
2.285(2)
1.829(9)
1.83(1)
1.841(8)
1.850(8)
1.15(1)
Mn(2)-P(2)
Mn(1)-C(2)
Mn(1)-C(4)
Mn(2)-C(6)
Mn(2)-C(8)
mean C-O(1-4)
1.829(9)
1.787(8)
1.842(8)
2.049(7)
1.14(1)
Mn(1)-Au-Mn(2)
Au-Mn(2)-Mn(1)
Mn(1)-Mn(2)-P(2)
Mn(1)-Au-P(1)
Au-Mn(1)-P(2)
Au-Mn(1)-C(1)
Au-Mn(1)-C(3)
Au-Mn(2)-C(5)
Au-Mn(2)-C(7)
Mn(1)-C(1)-O(1)
Mn(1)-C(3)-O(3)
Mn(2)-C(5)-O(5)
Mn(2)-C(7)-O(7)
68.20(3)
54.09(3)
48.02(6)
149.43(5)
104.86(6)
161.7(3)
81.0(2)
155.2(3)
75.1(2)
176.8(9)
178.9(7)
177.1(7)
175.5(6)
Au-Mn(1)-Mn(2)
Mn(1)-P(2)-Mn(2)
Mn(2)-Mn(1)-P(2)
Mn(2)-Au-P(1)
Au-Mn(2)-P(2)
Au-Mn(1)-C(2)
Au-Mn(1)-C(4)
Au-Mn(2)-C(6)
Au-Mn(2)-C(8)
Mn(1)-C(2)-O(2)
Mn(1)-C(4)-O(4)
Mn(2)-C(6)-O(6)
Mn(2)-C(8)-O(8)
57.71(3)
83.82(7)
48.16(6)
142.34(5)
101.17(6)
71.3(3)
90.5(3)
68.2(2)
105.8(3)
170.9(8)
177.5(8)
172.4(6)
173.4(8)
C(3)-Mn(1)‚‚‚Mn(2)-C(7)
C(4)-Mn(1)‚‚‚Mn(2)-C(7)
7.5(3)
-166.0(4)
C(3)-Mn(1)‚‚‚Mn(2)-C(8)
C(4)-Mn(1)‚‚‚Mn(2)-C(8)
-172.1(4)
14.5(4)
(b) PPN[Au{Mn2(CO)8(µ-PPh2)}2] (7)
Au-Mn(1)
Au-Mn(2)
Mn(1)‚‚‚Mn(2)
Mn(2)-P(1)
Mn(1)-C(2)
Mn(1)-C(4)
Mn(2)-C(6)
Mn(2)-C(8)
C(2)-O(2)
2.800(1)
2.812(1)
3.105(2)
2.275(2)
1.766(8)
1.801(8)
1.850(8)
1.828(8)
1.16(1)
1.17(1)
1.14(1)
1.144(9)
1.596(3)
5.200(2)
Au-Mn(1)a
2.780(1)
2.812(1)
2.279(2)
1.842(8)
1.852(8)
1.770(7)
1.865(8)
1.13(1)
1.14(1)
1.157(9)
1.13(1)
1.596(3)
5.143(2)
5.111(2)
Au-Mn(2)a
Mn(1)-P(1)
Mn(1)-C(1)
Mn(1)-C(3)
Mn(2)-C(5)
Mn(2)-C(7)
C(1)-O(1)
C(3)-O(3)
C(4)-O(4)
C(5)-O(5)
C(6)-O(6)
C(7)-O(7)
C(8)-O(8)
N-P(2)
N-P(2)a
Mn(1)‚‚‚Mn(1)a
Mn(2)‚‚‚Mn(2)a
Mn(1)‚‚‚Mn(2)a
Mn(1)-Au-Mn(2)
Mn(1)-Au-Mn(2)a
Au-Mn(1)-P(1)
Mn(1)-P(1)-Mn(2)
Au-Mn(1)-C(2)
Au-Mn(1)-C(4)
Au-Mn(2)-C(6)
Au-Mn(2)-C(8)
P(1)-Mn(1)-C(2)
P(1)-Mn(1)-C(4)
P(1)-Mn(2)-C(6)
P(1)-Mn(2)-C(8)
C(6)-Mn(2)-C(7)
Mn(1)-C(2)-O(2)
Mn(1)-C(4)-O(4)
Mn(2)-C(6)-O(6)
Mn(2)-C(8)-O(8)
67.19(3)
135.85(3)
103.51(6)
85.99(7)
161.6(2)
69.6(2)
97.9(2)
68.9(2)
93.2(2)
170.2(3)
85.3(2)
168.0(2)
177.0(3)
177.7(7)
169.7(6)
175.2(6)
169.0(6)
Mn(1)-Au-Mn(1)a
Mn(2)-Au-Mn(2)a
Au-Mn(2)-P(1)
Au-Mn(1)-C(1)
Au-Mn(1)-C(3)
Au-Mn(2)-C(5)
Au-Mn(2)-C(7)
P(1)-Mn(1)-C(1)
P(1)-Mn(1)-C(3)
P(1)-Mn(2)-C(5)
P(1)-Mn(2)-C(7)
C(1)-Mn(1)-C(3)
Mn(1)-C(1)-O(1)
Mn(2)-C(3)-O(3)
Mn(2)-C(5)-O(5)
Mn(2)-C(7)-O(7)
133.42(4)
130.70(4)
103.26(6)
95.9(2)
79.6(2)
160.1(2)
79.1(2)
86.6(2)
94.5(2)
94.6(2)
95.3(2)
175.5(3)
175.8(7)
175.7(6)
178.4(6)
175.0(6)
C(1)-Mn(1)‚‚‚Mn(2)-C(6)
C(3)-Mn(1)‚‚‚Mn(2)-C(6)
165.7(3)
-18.2(3)
C(1)-Mn(1)‚‚‚Mn(2)-C(7)
C(3)-Mn(1)‚‚‚Mn(2)-C(7)
-16.7(3)
159.4(3)
stronger σ donor. On the other hand, 4 shows no
activity toward free dppf at room temperature.
1, as a precursor not only should allow a higher
nuclearity cluster to form but also should stabilize the
product, as there is less electron charge on the gold
center as a result of its higher formal oxidation state
Since the two AuMn2 triangles are not expected to
communicate when separated by a phosphine bridge, a
natural question one would ask is “Could we remove
the bridge and fuse the triangles?”. The observed
lability of P(OEt)3 in 5 prompted us to use AuCl(SMe2)
as a source of “naked” gold which would achieve this
fusion. Although a similar reaction of AuCl(SMe2) with
[Mn(CO)5]- did not give the expected linear cluster
[(OC)5Mn-Au-Mn(CO)5]- 21 but significant decomposi-
tion instead,22 the use of a dimanganese monoanion, i.e.
1
(+1). The spectroscopic data (IR and H and 31P NMR)
suggested the formation of the expected anionic cluster
PPN[Au{Mn2(CO)8(µ-PPh2)}2] (7) (Scheme 1). An X-ray
crystallographic study revealed four pseudo-octahedral
[Mn(CO)4] moieties fused at a gold center, which has a
(21) Braunstein, P.; Dehand, J . J . Organomet. Chem. 1975, 88, C24.
Braunstein, P.; Schubert, U.; Burgard, M. Inorg. Chem. 1984, 23, 4057.
(22) Low, P. M. N. and Hor, T. S. A. Unpublished results.

Substituted Metal Carbonyls
Organometallics, Vol. 15, No. 11, 1996 2599
significantly weaker than that in 1 (2.867(2) Å).14 This
bond lengthening upon protonation and auration is
generally explained on geometric and electronic
grounds.14 The Au-Mn bonds (mean 2.718(1) Å) appear
to be consistent with those in AuMn2 (2.68-2.84 Å) and
higher nuclearity clusters, e.g. [{(PPh3)Au}4Mn(CO)5]-
BF4 (2.658(4)-2.766(5) Å),15 PPh3AuMn2(CO)8(µ-PPh2)
(2.678(3)-2.710(3) Å),13 {PPh3Au}2Mn2(CO)6(µ-C(Ph)O)-
(µ-PPh2) (2.702(2)-3.084(2) Å),13 {(PPh3)Au}2Mn2(CO)8-
(µ4-PCy) (2.762(2) Å),27 Mn2(µ-AuPCy3)(µ3-PCy-
(AuPCy3))(CO)8 (2.688(1)-2.726(1) Å),27 and {PPh3Au}3-
Mn2(CO)6(µ-PPh2)‚CH2Cl2 (2.698(2)-2.742(3) Å).13 All
these Au-Mn bonds are longer than those in the
bimetallic system (e.g. PPh3Au-Mn(CO)5 (2.52(3) Å)28
and PPh3Au-Mn(CO)4{P(OPh)3} (2.57(1) Å).29
For 7, both the phosphido-bridged Mn-Mn (3.105(2)
Å) and Au-Mn (mean 2.806(1) Å) lengths are signifi-
cantly longer than those of their counterparts in 2. The
Mn-Mn length is longer than all the Mn-Mn bonds
found in the known AuMn2 trinuclear species (2.748-
3.080 Å; mean 2.926 Å).13-15,27,30,31 These unexpectedly
long M-M distances in 7 compared to those in 1 and 2
raised a number of questions regarding the M-M bonds;
two important issues are “Are there Mn-Mn bonds in
7?” and “Why are the Mn-Mn and Au-Mn bonds
weaker in 7 compared to 2?”.
F igu r e 2. ORTEP drawing of the anion of PPN[Au{Mn2-
(CO)8(µ-PPh2)}2] (7).
crystallographic C2 symmetry, giving a bowtie config-
uration of the metals in the cluster (Figure 2). A similar
configuration was proposed for the isoelectronic Hg[Mn2-
(CO)8(µ-PPh2)]2 although its crystallographic structure
has not been established.23 This bowtie is twisted with
two AuMn2P planes (maximum deviation 0.04 Å) fused
at an angle of 85.50(4)°. The endo Mn(1)-Au-Mn(2)
angle ((67.19(3)°) is significantly smaller than the
tetrahedral angle, whereas the exo Mn(1)-Au-Mn-
(2a) and (135.85(3)°) is significantly larger. This distor-
tion of Au from an ideal tetrahedral geometry is
required for its participation in a {PMn2Au} ring
system. The Mn atoms are also highly distorted from
an ideal octahedron; e.g., the Au-Mn(1) vector is tilted
away from the normal of the plane defined by Mn(1)-
P(1)-C(1)-C(3)-C(4) and the axial COs (C(1,3/6,7)-
O(1,3/6,7)) are twisted at ca. 17.5(3)° from parallel. The
The opening up of Mn-P-Mn from 2 (83.82(7)°) to
7 (85.99(7)°) despite the statistically identical Mn-P
bonds in these molecules (2.282(2) Å in 2 and 2.277(2)
Å in 7) appears to suggest some repulsive forces between
the Mn atoms in 7. However, the Mn-Mn bond in 7 is
only 0.056(2) Å longer than that found in 2 and Mn-
Mn σ bonds have been reported to be as long as 3.23
Å.24 The bond length found in the crystal structure
therefore does not provide a conclusive answer to
whether there are direct Mn-Mn bonds in 7. The EAN
rule provides a guide but not necessarily an answer.32
Complex 7, being a 78-electron cluster, would require
six M-M bonds if all five metal atoms are in adherence
latter distortion is less severe in 2 (ca. 11.0(4)°).
A
carbonyl on each Mn group is bent ( Mn(1)-C(4)-O(4)
) 169.7(6)° and Mn(2)-C(8)-O(8) ) 169.0(6)°) in
order to avoid nonbonding contacts between the two
metallacyclic planes. The PPh2 group symmetrically
bridges the two Mn centers (Mn-P ) 2.279(2),
2.275(2) Å). As expected, without the influence of dppf
as in 2, the bite of gold on the phosphido-bridged Mn
33
with the 18-electron configuration (e.g. Os5(CO)19 ).
This is, however, unlikely for gold34 and is not consistent
with the observed structure. On the basis of a simple
localized orbital approach assuming 18-electron Mn, one
could propose the presence of 5 (I), 4 (II), or even 3 (III)
M-M bonds based on 16-, 14- and 12-electron configu-
rations, respectively, for gold. In each of these, a series
atoms is essentially symmetrical (∆(Au-Mn)
0.012(1) Å).
)
An important question on the structures of 2 and 7
is the strength of the Mn-Mn bonds compared to those
in 1, Mn2(CO)8(µ-H)(µ-PPh2), and (PR3)AuMn2(CO)8(µ-
PPh2) and the Au-Mn bonds compared to those in
(PR3)AuMn2(CO)8(µ-PPh2) and other Au-Mn clusters
and whether these bonds are anticipated from the EAN
rule or other bonding theories. The Mn-Mn distance
in complex 2 (3.049(2) Å) seems without doubt to
indicate direct Mn-Mn bonding when compared to
other Mn-Mn-bonded systems24 (e.g. Mn-Mn in
Mn2(CO)10 is 2.9038(6) Å25 ). It is very similar to those
(27) Haupt, H.-J .; Schwefer, M.; Flo¨rke, U. Inorg. Chem. 1995, 34,
292.
(28) Powell, H. M.; Mannan, K.; Kilbourn, B. T.; Porta, P. Proc. Int.
Conf. Coord. Chem. 1964, 8, p 155.
(29) Mannan, Kh. A. I. F. M. Acta Crystallogr. 1967, 23, 649.
(30) Riera, V.; Ruiz, M. A.; Tiripicchio, A.; Camellini, M. T. J . Chem.
Soc., Dalton Trans. 1987, 1551. Carren˜o, R.; Riera, V.; Ruiz, M. A.;
Bais, C.; J eannin, Y. Organometallics 1992, 11, 2923.
(31) The only other AuMn4 bowtie cluster known is AuMn4(CO)12(µ-
tedip)2(µ-H)5 (tedip ) (EtO)2POP(OEt)2), which is a 76-electron cluster
with Mn-Mn ) 2.860(2) and 3.080(2) Å and mean Au-Mn ) 2.822(2)
Å (see: Carren˜o, R.; Riera, V.; Ruiz, M. A.; Tiripicchio, A.; Tiripicchio-
Camellini, M. Organometallics, 1994, 13, 993). It, however, cannot be
compared directly with complex 7 because the bridging ligands are
entirely different and the electronic saturations are not the same.
(32) Bernal et al. have drawn a similar conclusion after studying
the Mn-Mn distances of a series of dimanganese complexes. The
authors found an unpredictably wide range of Mn-Mn lengths, which
led them to conclude that there is “a need for caution in using simple
rules (such as the EAN rule) to explain metal-metal distances”. Refer
to: Creswick, M.; Bernal, I.; Reiter, B.; Herrmann, W. A. Inorg. Chem.
1982, 21, 645.
found in (PPhR2)AuMn2(CO)8(µ-PPh2) (R
)
Ph
(3.067(3) Å),13 Me (3.066(8) Å)14). The EAN rule sup-
ports the Mn-Mn bond formation on the basis of the
one-electron contribution from Au(PPhR2) as its isolobal
counterpart H. It is, however, significantly longer, and
presumably weaker, than that in Mn2(CO)8(µ-H)-
(µ-PPh2) (2.951(1)14 and 2.937(5) Å26 ), which in turn is
(23) Iggo, J . A.; Mays, M. J . J . Chem. Soc., Dalton Trans. 1984, 643.
(24) Bernal, I.; Creswick, M.; Hermann, W. A. Z. Naturforsch., B:
Anorg. Chem., Org. Chem. 1979, B34, 1345.
(25) Churchill, M. R.; Amoh, K. N.; Wasserman, H. J . Inorg. Chem.
1981, 20, 1609.
(26) Doedens, R. J .; Robinson, W. T.; Ibers, J . A. J . Am. Chem. Soc.
1967, 89, 4323.
(33) Farrar, D. H.; Nicholls, J . N.; J ackson, P. F.; J ohnson, B. F. G.;
Lewis, J . J . Chem. Soc., Chem. Commun. 1981, 273. Nicholls, J . N. J .
Chem. Soc., Dalton Trans. 1982, 1395.
(34) Hall, K. P.; Mingos, D. M. P. Prog. Inorg. Chem. 1984, 32, 237.

2600 Organometallics, Vol. 15, No. 11, 1996
Low et al.
F igu r e 3. Energy levels of the filled Mn-based molecular
orbitals in 1. The HOMO is Mn-Mn σ bonding, while the
LUMO is antibonding.
Ta ble 3. Ca lcu la ted Tota l Over la p P op u la tion s
2
2
2
a n d Mn -Mn d z -d z a n d Mn d z -Au /H (a ll or bita ls)
P a r tia l Over la p P op u la tion s for Mn -Mn a n d
Mn -Au /H In ter a ction s in Mod els of Com p ou n d s 1,
2, 7, 8, a n d 9
of “no-bond” resonance structures (I-III) can then be
constructed which explain the weakness of both Mn-
Mn and Au-Mn bonds. In this bowtie arrangement,
all four Mn atoms are interconnected through the Au
center; electron delocalization thus helps to stabilize the
two metal planes and keep the cluster intact. These
localized-bond concepts can give a qualitative idea of
the M-M bonds but could not illustrate the nature and
relative importance of the metal orbitals in these bonds.
As an attempt to address such issues, Fenske-Hall
molecular orbital calculations were carried out on
models of 1, 2, 7, Mn2(CO)8(µ-H)(µ-PPh2) (8), and (PMe2-
Ph)AuMn2(CO)8(µ-PPh2) (9). For comparison, their
Mn-Mn distances are 2.867(2), 3.049(2), 3.105(2),
2.951(1), and 3.066(8) Å, respectively. For simplicity,
compd overlap pop.
Mn-Mn Mn(1)-Au/H Mn(2)-Au/H
1
2
7
8
9
total
dz
0.00127
0.0142
-0.0091
0.0078
-0.0142
0.0044
-0.0177
0.0055
2
total
0.0899
0.0060
0.0500
0.0016
0.1078
0.0294
0.0727
0.0050
0.0559
0.0040
0.0461
0.0016
0.1073
0.0293
0.0862
0.0051
2
dz
total
2
dz
total
2
dz
total
-0.0084
0.0079
2
dz
Mn-Mn, Au-Mn, and Mn-H overlap populations in
these compounds (Table 3). With regard to Mn-Mn
interactions, it is striking to observe that only in the
case of 1 is the overlap population positive; the rest are
all negative.
2
let the Mn-Mn line be along the z axis, so that the dz
orbitals are directed at one another.
Upon inspection of the structures of the compounds,
it is readily apparent that in the anion 1,14 the Mn-
Mn bond is a component of the σ-bonding framework
(IV). In the other compounds, the orientation of the
From a qualitative viewpoint, accounting for all the
electrons necessitates the filling of three “nonbonding”
orbitals36 per Mn, leaving two electrons (four for com-
2
pound 7) for Mn-Mn/Au/H interactions. In 1 the dz
orbitals do not belong to the nonbonding set; their
interaction yields a bonding and an antibonding orbital,
and the two electrons available for Mn-Mn interaction
fill the bonding orbital (Figure 3).37 On the other hand,
in 2, 7, 8, and 9, Mn-Mn is not a dominant σ interac-
2
tion; therefore, the dz orbitals are part of the nonbond-
ing set. Under normal circumstances one would expect
2
filled dz bonding and antibonding orbitals, resulting in
a net antibonding interaction.
We focus on the electronic structure of 9 as a
prototype of the compounds with Au or H. The HOMO
describes a two-electron-three-center bond (as de-
scribed by Hall and Mingos34) involving Au and the two
Mn atoms (V). It is interesting to note that the Mn-
Mn-C bonds with respect to the Mn-Mn and Mn-
Au/H bonds35 indicates that the latter constitutes the
sixth bond in the Mn pseudo-octahedron;14 it can thus
be inferred that Mn-Au/H interactions are dominant
over Mn-Mn interactions. This is reflected in the total
2
Mn dz combinations mix with the Mn dyz and Au
(35) In [Mn2(CO)8(µ-PPh2)]-, the molecular planes containing the
metals and phosphorus have two carbonyls at approximately right
angles (and another two trans) to the Mn-Mn bond. In both 2 and 7,
these carbonyls move away from the Mn-Mn axis so as to accom-
modate the bridging gold atom. As a result, these carbonyls are trans
to phosphorus and the sites opposite the Mn-Mn axis are vacant.
(36) These orbitals participate in π interactions. It should be noted
in 1 that this set has σ-antibonding contributions due to severe
distortion from a perfect octahedron.
(37) Contributions to the seven filled Mn-based MO’s as well as the
LUMO are given as Supporting Information.

Substituted Metal Carbonyls
Organometallics, Vol. 15, No. 11, 1996 2601
Ta ble 4. Ma jor Or bita l Con tr ibu tion s to th e
HOMO2, HOMO, a n d LUMO of 9a
atomic orbital
HOMO2
HOMO
LUMO
Au
dyz
s
py
pz
0.1461
0.0235
-0.0344
0.2111
0.0662
-0.3342
0.2623
0.0397
-0.0495
0.0268
-0.0085
-0.3352
Mn2
2
dz
-0.5116
-0.2841
-0.2452
0.1888
-0.4338
0.1003
orbitals: the bonding combination occurs in the HO-
MO6, HOMO5, and HOMO, while the antibonding
combination occurs in the HOMO2 and LUMO. The
major contributions to the HOMO2, HOMO, and LUMO
are listed in Table 4.38 (Note that Mn dyz is involved in
a σ interaction with Au.) Despite its presence in the
dyz
Mn3
2
dz
0.2243
-0.2238
-0.3841
-0.1596
0.4046
0.0678
dyz
a
See V for relative positions of atoms. The energies of these
molecular orbitals are -8.04, -6.97, and -4.41 eV, respectively.
2
HOMO, Mn dz does not play a major role in Mn-Au
2-
bonding, as can be inferred from the partial dz Au
interaction, which accounts for an average of 14% of the
total Au-Mn overlap populations in 9, is present in all
the compounds with Au. In a sense, it could be
considered as π back-donation from the filled Au dyz
orbital into the Mn σ orbitals, somewhat analogous to
the behavior of Au toward phosphine ligands.
overlap populations listed in Table 3. From consider-
ations of “orbital conservation”, it can be inferred that
mixing allows the antibonding combination to be par-
tially unoccupied (as part of the LUMO), and thus
2
results in a weak dz bonding interaction, as shown by
their partial overlap populations (Table 3). This allows
the two Mn atoms to approach each other more closely,
even though the total overlap population is negative.39
As described by the HOMO (Table 4), in 9 Au
interacts with the Mn2 moiety mainly through a hybrid
of its s and py orbitals (V; the same is true of 2). In
contrast, of the two MO’s in 7 which are the counter-
parts of the HOMO in 9, one utilizes only the Au s
orbital and the other only the py orbital (shown quali-
tatively in VI); since Au is situated on a symmetry axis,
The isolobal principle prompted some intensive work
on [AuPR3]+ and H+ in edge-bridging a variety of M-M-
bonded anions.6a-c,8,34,41,42 This study showed that, by
stripping off the phosphine, one can have a direct access
to a bowtie cluster. The strength of the homo and hetero
metal-metal bonds appears to be dependent not only
on the electronic state of the metals but also on their
ligand arrangements. Future directions of our project
are charted by these findings.
Exp er im en ta l Section
Gen er a l P r oced u r es. All reactions were performed under
pure dry argon using standard Schlenk techniques. Mn2(CO)8-
(µ-H)(µ-PPh2)17 and PPN[Mn(CO)8(µ-PPh2)]14 were prepared by
literature methods. Au2Cl2(µ-dppf)43 was prepared by a modi-
45
fied procedure from AuCl(SMe)44 and dppf in a 2:1 molar
ratio. Au2Cl2(µ-dppe) was prepared from HAuCl4 and dppe
in a 1:3 ratio. All other reagents were commercial products
and were used as received. FT-IR spectra were recorded on a
Perkin-Elmer 1710 spectrometer. All NMR spectra were
recorded in CDCl3 solution on a Bruker AC 300F spectrometer.
The 31P NMR chemical shifts are externally referenced to 85%
H3PO4. Elemental analyses were performed by the analytical
service of this department. The presence of solvate molecules
in the sample was confirmed by NMR analysis. The solvent
inclusion in the crystal lattice of many dppf complexes has
been noted in the literature.46 Molecular weight determination
was carried out by vapor-pressure osmometry in a Knauer-
Dampfdruck osmometer by Galbraith Laboratories, Inc., in
Knoxville, TN, using toluene as solvent.
all the MO’s must be either symmetric or antisymmetric
with respect to Au, thus precluding sp mixing. Conse-
quently, there is no Au orbital reinforcement in the
direction of any Mn2 moiety. This is a likely cause for
weaker Au-Mn interactions, as reflected in the longer
Au-Mn bond lengths.
With regard to the longer Mn-Mn distance in 7, the
same symmetry constraints apply to the mixing of Mn
40
2
“dz ”-type orbitals with Au orbitals. However, the
dominance of the Au-Mn interactions and the foregoing
results suggest that the length of the Mn-Mn bond is
largely determined by the strength of the Au-Mn
interactions, especially since the latter involve mainly
the Au s and py orbitals. This conclusion can be
generalized to include compounds 2, 8 (Mn-H), and 9;
in 8 the size of H could also be important.
The models of the molecules used for the calculations are
based on their crystal structures (see ref 14 and this work).
All the phenyl rings and methyl groups on phosphine ligands
were replaced by hydrogen atoms, and the P-H bond lengths
(41) Ferrer, M.; Reina, R.; Rossell, O.; Seco, M.; Alvarez, S.; Ruiz,
E. Organometallics 1992, 11, 3753. Reina, R.; Rossell, O.; Seco, M.;
Ros, J .; Yaˆn˜ez, R.; Perales, A. Inorg. Chem. 1991, 30, 3973. Low, A.
A.; Lauher, J . W. Inorg. Chem. 1987, 26, 3863.
(42) Hor, T. S. A.; Tan, A. L. C. J . Coord. Chem. 1989, 20, 311.
(43) Hill, D. T.; Girard, G. R.; McCabe, F. L.; J ohnson, R. K.; Stupik,
P. D.; Zhang, J . H.; Reiff, W. H.; Eggleston, D. S. Inorg. Chem. 1989,
28, 3529.
The interaction described by HOMO2 of 9 deserves
some comment. This is actually Au-Mn antibonding
(with mixing to minimize the “bad” interaction), but
there are corresponding lower energy bonding molecular
orbitals involving Au dyz and Mn σ orbitals. The Au dyz
(44) Bonati, F.; Minghetti, G. Gazz. Chim. Ital. 1973, 103, 373.
(45) Marr, G.; Hunt, T. J . Chem. Soc. C 1969, 1070. Bishop, J . J .;
Davison, A.; Katcher, M. L.; Lichtenberg, D. W.; Merrill, R. E.; Smart,
J . C. J . Organomet. Chem. 1971, 27, 241.
(46) Butler, I. R.; Cullen, W. R.; Kim, T. J .; Rettig, S. J .; Trotter, J .
Organometallics 1985, 4, 972.
(38) Contributions to all filled Au/Mn-based orbitals as well as the
LUMO of 9 are given as Supporting Information.
(39) The overall negative overlap population is due to net antibond-
ing π interactions.
(40) Only one Mn-Mn line can be parallel to the z axis.

2602 Organometallics, Vol. 15, No. 11, 1996
Low et al.
were adjusted to the sum of covalent radii. The models were
aligned such that one Mn-Mn line is parallel to the z axis
and the Mn-Au-Mn plane is parallel to the yz plane. In 9,
the bridging hydride, whose coordinates were not given in the
crystal structure, was placed in the same plane as the Mn-
P-Mn triangle, with Mn-H distances slightly longer than the
sum of covalent radii.
Calculations were performed using the self-consistent Fen-
ske-Hall approximate molecular orbital method.47 Except for
hydrogen, the basis functions were obtained by fitting the
results of XR (Herman-Skillman) calculations48 for M(+1)
(transition metals) and X(0) (main-group metals) to Slater
orbitals.49 Double-ú functions were used for transition-metal
valence d orbitals and main-group valence p orbitals; all the
rest were single-ú functions. An orbital exponent of 1.2 was
used for hydrogen.
was refluxed for ca. 4 h. The reaction, followed by TLC, was
stopped when there were no further changes on the TLC
plates. The solvent was removed in vacuo and the orange
residue extracted with ether to give an orange-red solution.
Applying the ether extract to silica TLC plates and eluting
with CH2Cl2/hexane (1:4) gave two products corresponding to,
in order of decreasing Rf value, (PPh3)AuMn2(CO)8(µ-PPh2) (5;
0.020 g, 31%) and unreacted 2 (0.008 g, 12%). Anal. Calcd
for 5: C, 44.74; H, 2.43; Au, 19.83; Fe, 2.81; Mn, 11.06; P, 6.23.
Found: C, 44.35; H, 2.20; Au, 17.41; Fe, 2.52; Mn, 9.60; P,
6.30. 1H NMR (δ): 7.28-7.36 (m, Ph, 5H); 7.44-7.60 (m, Ph,
16H); 7.82-7.88 (m, Ph, 4H). 31P NMR (δ): 63.01 (s, PPh3);
212.54 (s, br, PPh2). IR (CO, cm-1): 2021 s, 1983 vs, 1955 s,
1930 s (CHCl3).
Rea ction of Com p lex 2 w ith P (OEt)3. P(OEt)3 (0.5 cm3,
2.92 mmol) was injected by a syringe into an orange solution
of 2 (0.066 g, 0.033 mmol) in thf (8 cm3), which readily turned
brownish yellow. The reaction was monitored by TLC. Two
spots, in order of decreasing Rf value, corresponding to a major
yellow product identified as (P(OEt)3)AuMn2(CO)8(µ-PPh2) (6)
and a minor compound identified as 2 were observed. After
the mixture was stirred at room temperature for 2 h, the
solvent was removed in vacuo and the residue extracted by a
minimum amount of CH2Cl2 and chromatographed on silica
TLC plates. Elution with CH2Cl2/hexane (1:1) and recrystal-
lization with CH2Cl2/hexane gave band 1 as complex 6 (0.012
g, 21%) and band 2 as complex 2 (0.005 g, 8%). Anal. Calcd
for 6: C, 36.54; H, 3.18; Au, 21.78; Mn, 12.16; P, 6.85.
Found: C, 36.14; H, 3.20; Au, 17.25; Mn, 9.57; P, 5.98. 1H
Rea ction of P P N[Mn 2(µ-P P h 2)(CO)8] (1) w ith Au 2Cl2-
(µ-d p p f). A solution of AgBF4 (0.050 g, 0.26 mmol) in thf (5
cm3) was added to that of Au2Cl2(µ-dppf) (0.101 g, 0.098 mmol),
which led to an immediate precipitation of AgCl. The solution,
shielded from direct light, was stirred for ca. 1 h before filtering
dropwise into a flask containing a yellow solution of PPN-
[Mn2(CO)8(µ-PPh2)] (1; (0.100 g, 0.095 mmol) in CH2Cl2 (5 cm3).
The resultant orange-red mixture was further stirred at room
temperature for ca. 3 h. The solvent was removed and the
orange-red residue obtained redissolved in a minimum amount
of CH2Cl2 and chromatographed on silica TLC plates. Elution
with CH2Cl2/hexane (2:5) gave two major bands, which were
extracted with CH2Cl2 and recrystallized from a CH2Cl2/
hexane mixture to give, in order of decreasing Rf value,
[AuMn2(CO)8(µ-PPh2)]2(µ-dppf) (2; 0.050 g, 51%) and AuCl(µ-
dppf)[AuMn2(CO)8(µ-PPh2)] (3; 0.020 g, 13%). A third minor
band near the base line corresponds to unreacted Au2Cl2(µ-
dppf). Omitting AgBF4 in a subsequent reaction and reversing
the addition of 1 to Au2Cl2(µ-dppf) (1:1) gave 2 in 50% yield
but 3 in 20% yield. Anal. Calcd for 2: C, 44.74; H, 2.43; Au,
19.83; Fe, 2.81; Mn, 11.06; P, 6.23. Found: C, 44.37; H, 2.38;
Au, 17.41; Fe, 2.54; Mn, 10.03; P, 6.34. 1H NMR (δ): 4.22 (q,
CpHR, 4H); 4.74 (m, CpHâ, 4H); 7.32-7.46 (m, Ph, 32H); 7.81-
7.88 (m, Ph, 8H). 31P NMR (δ): 58.02 (s, PAu); 212.31 (s, br,
3
NMR (δ): 1.31 (t, CH3, 9H, J HH′ ) 7.08 Hz); 4.13 (dq, CH2,
3
3
6H, J HH′ ) 7.08 Hz J PH′ ) 9.52 Hz); 7.19-7.27 (m, Ph, 6H);
7.64-7.89 (m, Ph, 4H). 31P NMR (δ): 182.67 (s, P(OEt)3; 217.54
(s, br, PPh2). IR (CO, cm-1): 2029 s, 1985 vs, 1957 s, 1931 s
(CHCl3).
In the process of applying the mixture to the plates followed
by elution, it was found that the relative intensities of the first
(complex 6) and second (complex 2) bands changed with time.
The earlier plates gave a higher proportion of 6 compared to
2, while the later plates gave more 2 than 6. This was the
first indication that dppf replacement of 1 by P(OEt)3 is
reversible.
P
Mn). IR (CO, cm-1): 2021 s, 1983 vs, 1956 s, 1930 s (CHCl3).
Mol wt (toluene): 2058, calcd 1987. Anal. Calcd for 3: C,
43.16; H, 2.55; Au, 26.21; Cl, 2.36; Fe, 3.72; Mn, 7.31; P, 6.18.
Found: C, 43.26; H, 2.76; Au, 25.52; Cl, 2.32; Fe, 3.45; Mn,
Rea ction of Com p lex 1 w ith Au Cl(SMe2). A solution of
1 (0.155 g, 0.147 mmol) in thf (20 cm3) was cooled in an ice
bath before it was transferred by a teflon delivery tube to a
solution of AuCl(SMe2) (0.038 g, 0.129 mmol) in dimethyl
6.93; P, 6.28. 1H NMR (δ): 4.22 (m, CpHR(Au-Mn ), 2H); 4.20
2
sulfide (2 cm3) at 0 °C to give an orange-red solution. The
(m, CpHR(Au), 2H); 4.72 (m, CpHâ, 4H); 7.33-7.54 (m, Ph, 26H);
7.82-7.88 (m, PhPPh2, 4H). 31P NMR (δ): 27.82 (s, FcPAuCl);
1
mixture was stirred in an ice bath for ca.
/ h followed by
2
58.04 (s, FcPAuMn ); 212.31 (s, br, PPh2). IR (CO, cm-1): 2022
2
stirring at room temperature for ca. 4 h. The solvent was
removed in vacuo and the dark brown residue washed with
benzene (ca. 3 × 10 cm3) and then MeOH (2 × 10 cm3). The
residue was extracted with acetone and filtered through celite
to remove the fine Au particles. Hexane was added to the
filtrate to give a red precipitate which was redissolved in CH2-
Cl2; hexane was then added. When this mixture stood for ca.
7 days, a red crystalline solid was isolated (at times, it may
be necessary to remove some minor foreign crystals, which are
orange-yellow, of 1) to give orange-red crystals of PPN[Au-
{Mn2(CO)8(µ-PPh2)}2] (7; 0.025 g, 11%). The yield is low, as it
is calculated on the basis of the amount of added AuCl(SMe2),
some of which inevitably decomposed in the course of reaction.
Anal. Calcd for 7: C, 51.46; H, 2.84; Au, 11.10; Mn, 12.38; N,
0.79; P, 6.98. Found: C, 51.41; H, 2.77; Au, 9.90; Mn, 11.77;
N, 0.83; P, 6.89. 1H NMR (δ): 7.35-7.89 (m, Ph, 50H). 31P
NMR (δ): 22.54 (s, N(PPh3)2+); 190.71 (s, br, Mn-PPh2). IR
(CO, cm-1): 2041 w, 2014 s, 1984 vs, 1971 m, sh, 1910 s, br
(acetone).
s, 1983 vs, 1957 s, 1930 s (CHCl3). A 2:1 stoichiometric ratio
of 1 to Au2Cl2(µ-dppf) in the presence of AgBF4 gave 2 in 65%
yield but a negligible amount of 3.
Rea ction of Com p lex 1 w ith Au 2Cl2(µ-dppe). A solution
of 1 (0.109 g, 0.103 mmol) in thf (25 cm3) was transferred
dropwise to a solution of Au2Cl2(µ-dppe) (0.109 g, 0.126 mmol)
in CH2Cl2 (25 cm3). The mixture, which turned from colorless
to orange, was stirred at room temperature for ca. 3 h and
stripped of solvent and the orange residue was redissolved in
a minimum amount of CH2Cl2. Development on silica TLC
plates with CH2Cl2/hexane (2:5) gave [AuMn2(µ-PPh2)(CO)8]2-
(µ-dppe) (4; (0.026 g, 12%). Anal. Calcd for 4: C, 43.31; H,
2.42; Au, 21.52; Mn, 12.06; P, 6.77. Found: C, 43.12; H, 2.45;
Au, 20.15; Mn, 10.93; P, 7.26. 1H NMR (δ): 2.97 (s, CH2, 4H);
7.29-7.36 (m, Ph, 12H); 7.41-7.48 (m, Ph, 12H); 7.53-7.57
(m, Ph, 8H); 7.82-7.85 (m, Ph, 8H). 31P NMR (δ): 62.63 (s,
P
Au); 215.46 (s, br, PPh2).
Rea ction of Com p lex 2 w ith P P h 3. A solution of 2 (0.065
g, 0.033 mmol) and PPh3 (0.020 g, 0.076 mmol) in thf (25 cm3)
Cr ysta llogr a p h ic An a lysis. Red crystals of complexes 2
and 7 were grown at room temperature (25 °C) by a layering
method, with hexane on a sample solution in CH2Cl2, and slow
evaporation of a sample solution in a mixture of CH2Cl2 and
hexane, respectively. Single crystals of dimensions 0.19 × 0.31
× 0.38 mm for 2 and 0.50 × 0.09 × 0.22 mm for 7 were chosen
(47) Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768.
(48) Herman, F.; Skillman, S. Atomic Structure Calculations; Pren-
tice-Hall: Englewood Cliffs, NJ , 1963.
(49) Bursten, B. E.; J ensen, J . R.; Fenske, R. F. J . Chem. Phys. 1978,
68, 3320.

Substituted Metal Carbonyls
Organometallics, Vol. 15, No. 11, 1996 2603
for the X-ray diffraction study with graphite-monochromated
Mo KR (λ ) 0.709 30 Å for 2 and λ ) 0.710 69 Å for 7) radiation
at room temperature. The crystal data and experimental
conditions are given in Table 1. Atomic coordinates and
equivalent isotropic displacement coefficients are given in the
Supporting Information. The unit cell constants were deter-
mined by a least-squares fit to the setting parameters of 25
independent reflections in the range of 17.26° < 2θ < 36.08°
for 2 and 15.77° < 2θ < 62.48° for 7. Totals of 4611 for 2 (2θmax
) 44.8°) and 4642 unique reflections for 7 (2θmax ) 44.9°) were
measured by scanning on a Nonius diffractometer. The range
of indices was h ) -15 to +15, k ) 0-15, l ) 0-17 for 2 and
h ) -26 to +25, k ) 0-18, l ) 0-18 for 7. After corrections
for absorption effects, 3462 data with I > 3.0σ(I) were used
for 2 and 3169 data with I > 2.5σ(I) for 7. The minimum and
maximum transmission factors were 0.368 and 0.999 for 2 and
0.781 148 and 0.999 825 for 7. The structures were solved by
direct methods and MULTAN.50 Refinement was by full-
matrix least-squares calculations with all non-hydrogen atoms
allowed anisotropic motion. All hydrogen atoms were fixed
as isotropic ellipsoids in the final cycles of least-squares
refinement. The last least-squares cycle was calculated with
75 atoms, 457 parameters, and 3462 reflections and converged
with RF ) 0.030, R′ ) 0.035 (GOF 1.43) (for all reflections RF
) 0.052, R′ ) 0.036) for 2 and with 77 atoms, 461 parameters,
and 3169 reflections and converged with RF ) 0.031, R′ ) 0.033
(GOF 1.10) (for all reflections RF ) 0.063, R′ ) 0.038) for 7.
Weights based on counting statistics were used. The weight
modifier was 0.000 100. The maximum shift/error ratio was
0.001 for 2 and 0.004 for 7. In the last difference map the
deepest hole was -0.590 e Å-3 and the highest two peaks of
1.580 and 1.102 e Å-3 were at 1.11 Å from the gold atom in 2.
The next positive peak height was <0.4 e Å-3
.
For 7, the
deepest hole and highest peak were -0.320 and 0.350 e Å-3
.
Computations were carried out on a microVAX 3600 computer
with the NRCC package. Selected bond distances and angles
are listed in Table 2. For 7, both PPN+ and [Au{Mn2(CO)8-
(µ-PPh2)}2]- are required to possess a rigorous C2 molecular
symmetry in the solid state because the central Au and N
atoms are located on crystallographic 2-fold axes. Only half
of the atoms in the formula are thus crystallographically
independent.
Ack n ow led gm en t. Financial assistance from the
National University of Singapore (NUS; Grant No.
RP850030) and technical help from our department at
NUS are gratefully acknowledged. We especially thank
Y. L. Yong and S. L. Boo for experimental assistance
and Y. P. Leong for stenographic work. A.L.T. would
like to acknowledge the National Science and Technol-
ogy Board for financial support through a Postdoctoral
Fellowship.
Su p p or tin g In for m a tion Ava ila ble: Tables giving crys-
tal data and refinement details, positional and thermal
parameters, bond distances and angles, and torsion angles for
2 and 7, least-squares plane data for 7, and MO data for 1
and 9 (26 pages). Ordering information is given on any current
masthead page.
(50) Main, P.; Fiske, S. E.; Hull, S. L.; Germain, G.; Declerq, J . P.;
Woolfson, M. H. In MULTAN, a System of Computer Programs for
Crystal Structure Determination from X-ray Diffraction Data; Uni-
versities of York and Louvain, 1980.
OM950797G