Complexes of gold(I), silver(I), and copper(I) with pentaaryl[60]fullerides

Article abstract of DOI:10.1021/ja201297r

Gold(I), silver(I), and copper(I) phosphine complexes of 6,9,12,15,18-pentaaryl[60]fullerides 1a and 1b, namely, [(4-MeC ₆H₄)₅C₆₀]Au(PPh₃) (2a), [(4-t-BuC₆H₄)₅C₆₀]Au(PPh ₃) (2b), [(4-MeC₆H₄)₅C ₆₀]Ag(PCy₃) (3a), [(4-t-BuC₆H₄) ₅C₆₀]Ag(PPh₃) (3b), [(4-t-BuC₆H ₄)₅C₆₀]Ag(PCy₃) (3c), [(4-MeC ₆H₄)₅C₆₀]Cu(PPh₃) (4a), and [(4-t-BuC₆H₄)₅C₆₀]Cu(PPh ₃) (4b), have been synthesized and characterized spectroscopically. All complexes except for 3c were also characterized by single-crystal X-ray diffraction. Several coordination modes between the cyclopentadienyl ring embedded in the fullerene and the metal centers are observed, ranging from η¹ with a slight distortion toward η³ in the case of gold(I), to η²/η³ for silver(I), and η⁵ for copper(I). Silver complexes 3a and 3b are rare examples of crystallographically characterized Ag(I) cyclopentadienyls whose preparation was possible thanks to the steric shielding provided by fullerides 1a and 1b, which stabilizes these complexes. Silver complexes 3a and 3b both display unexpected coordination of the cyclopentadienyl portion of the fulleride anion with Ag(I). DFT calculations on the model systems (H₅C ₆₀)M(PH₃) and CpMPH₃ (M = Au, Ag, or Cu) were carried out to probe the geometries and electronic structures of these metal complexes.

Full text of DOI:10.1021/ja201297r

ARTICLE
pubs.acs.org/JACS
Complexes of Gold(I), Silver(I), and Copper(I) with
Pentaaryl[60]fullerides
Merissa Halim, Robert D. Kennedy, Mitsuharu Suzuki, Saeed I. Khan, Paula L. Diaconescu, and Yves Rubin*
Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095-1569, United States
S Supporting Information
b
ABSTRACT: Gold(I), silver(I), and copper(I) phosphine com-
plexes of 6,9,12,15,18-pentaaryl[60]fullerides 1a and 1b, namely,
[(4-MeC6H4)5C60]Au(PPh3) (2a),[(4-t-BuC6H4)5C60]Au(PPh3)
(2b), [(4-MeC6H4)5C60]Ag(PCy3) (3a), [(4-t-BuC6H4)5C60]
Ag(PPh3) (3b), [(4-t-BuC6H4)5C60]Ag(PCy3) (3c), [(4-
MeC6H4)5C60]Cu(PPh3) (4a), and [(4-t-BuC6H4)5C60]Cu-
(PPh3) (4b), have been synthesized and characterized spectro-
scopically. All complexes except for 3c were also characterized by
single-crystal X-ray diffraction. Several coordination modes be-
tween the cyclopentadienyl ring embedded in the fullerene and
the metal centers are observed, ranging from η1 with a slight distortion toward η3 in the case of gold(I), to η2/η3 for silver(I), and η5 for
copper(I). Silver complexes 3a and 3b are rare examples of crystallographically characterized Ag(I) cyclopentadienyls whose preparation
was possible thanks to the steric shielding provided by fullerides 1a and 1b, which stabilizes these complexes. Silver complexes 3a and 3b
both display unexpected coordination of the cyclopentadienyl portion of the fulleride anion with Ag(I). DFT calculations on the model
systems (H5C60)M(PH3) andCpMPH3 (M = Au, Ag, or Cu) were carried out to probe the geometries andelectronic structures ofthese
metal complexes.
’ INTRODUCTION
In contrast, only seven X-ray structures of neutral gold(I)
cyclopentadienyl complexes have been reported.8 They display
atypical coordination modes for cyclopentadienyl transition metal
complexes, denoted as η1/η3 in this study, and are stabilized
exclusively with phosphine ligands. Even more intriguing is the
fact that there is only a single example of a crystallographically
characterized phosphine-stabilized silver(I) cyclopentadienyl
complex, that of [C5H2(SiMe3)3]Ag[P(n-Bu)3]. It displays both
η5 and η3 coordination modes in the crystal.9
Over the past decade, the group of Nakamura hasdevelopedthe
organometallic chemistry of the 6,9,12,15,18-pentaarylfulleride
(1) family, in which coordination to the metal center occurs via
a cyclopentadienyl ring embedded within the fullerene cage
(Chart 1a).10,11 The thallium(I) complex (η5-Ph5C60)Tl was
the first example of a pentahapto metalꢀfullerene complex.12a
Pentamethyl- or pentaarylfulleride complexes with elements of
groups 6,12b 7,12c 8,12cꢀi 9,12c,j,k and 1012l have since been
investigated. Many of these complexes have been characterized
crystallographically, revealing η5 coordination in all cases. Re-
cently, Bouwkamp and Meetsma isolated the discrete ion pair
[Zr(NMe2)3(THF)2][Ph5C60] in an attempt to extend the series
to include early transition metals.12m
Cyclopentadienyl complexes of the coinage metals are dra-
matically underrepresented in comparison to those of earlier
transition metals. Although only a handful of group 11 metal
cyclopentadienyl complexes have been isolated and character-
ized crystallographically, a range of coordination modes between
the cyclopentadienyl ligand and the metal center has nonetheless
been observed.1,3ꢀ9 The characterization of group 11 metal
cyclopentadienyl compounds, especially of copper and silver,
has often been hampered by their instability and high sensitivity
to air. These issues have been addressed in some instances by
using bulky phosphine or cyclopentadienyl ligands.
The synthesis of a compound with the formula CpCu 2[P(n-Pr)3]
3
was reported in a 1938 patent and describes possibly the first example
of a cyclopentadienyl complex of a transition metal element.1 This
copper cyclopentadienyl complex was characterized further and
assigned a monohapto-bonded structure after the discovery of
ferrocene.2,3 Pentahapto bonding was suggested following a reexami-
nation of the infrared spectra of CpCu(PEt3) and CpCu[P(n-Bu)3],4
whose structure was eventually confirmed by X-ray crystallography.5
All copper(I) cyclopentadienyl complexes are η5 coordinated
to the Cp ligand, with the exception of the anionic copper(I)
cyclopentadienyl complexes [CuCp2]ꢀ and [Cu2Cp3]ꢀ, which
have slipped-sandwich structures.6 Neutral cyclopentadienyl
copper(I) complexes with a general structure (C5R5)Cu(I)L
possess stabilizing ligands (L) such as phosphines,7a,7b small
organometallic clusters,7c carbodiphosphoranes,7d N-heterocyc-
lic carbenes,7e alkynes,7f and isocyanides.7g
We have become intrigued by the prospect of stabilizing
reactive metal complexes or other species with bulky ligands such
as the 6,9,12,15,18-pentaarylfulleride anion (1). In particular,
Received: February 10, 2011
Published: April 08, 2011
r
2011 American Chemical Society
6841
dx.doi.org/10.1021/ja201297r J. Am. Chem. Soc. 2011, 133, 6841–6851
|

Journal of the American Chemical Society
ARTICLE
Chart 1. (a) Structure of Pentaarylfulleride Anion 1 with
Specific Anions 1a and 1b Used in This Study; (bꢀd) Complexes
of Gold(I) (2a and 2b), Silver(I) (3a, 3b, and 3c), and Copper(I)
(4a and 4b) Prepared by Reaction with Corresponding Metal
Phosphine Halides; Experimentally Observed Bonding Modes
between Fulleride Ligands 1a and 1b and Gold(I), Silver(I), and
Copper(I)
there have been only two reports on the isolation and character-
ization of pentaarylfulleride complexes of group 11 elements,
namely, (Ph5C60)Cu(PEt3) reported by Sawamura et al.12a and
the gold(I) complexes [(4-MeC6H4)5C60]Au(PPh3) (2a) and
[(4-t-BuC6H4)5C60]Au(PPh3) (2b) reported by us.13 The latter
preliminary communication reported the only example of a
crystallographically characterized group 11 metal pentaarylfuller-
ide complex in which a fullerene cage is bonded directly to the
metal. The present work extends our investigations and describes
the synthesis and crystallographic characterization of the pentaar-
ylfulleride tricyclohexylphosphine complex of silver(I) (3a), the
corresponding triphenylphosphine complex 3b, as well as the
pentaarylfulleride triphenylphosphine complexes of copper(I) 4a
and 4b. We make a comparative structural and computational
study of these new complexes along with the previously reported
phosphine gold(I) complexes 2a and 2b.
Figure 1. Representations of the crystal structures of 2a (ODCB)2
3
(aꢀd) and 2a (CHCl3)2(CS2) (eꢀh). (a) ORTEP drawing (thermal
3
ellipsoids at 50% probability) of compound 2a (ODCB)2. (b) Partial
3
structure of 2a (ODCB)2 showing the ligandꢀmetal bonding.
3
(c, d) Space-filling representations of compound 2a (ODCB)2, viewed
3
approximately perpendicular and parallel to the molecular quasi-C5 axis.
(e) ORTEP drawing (thermal ellipsoids at 50% probability) of com-
pound 2a (CHCl3)2(CS2). (f) Partial structure of 2a (CHCl3)2(CS2)
3
3
showing the metalꢀligand bonding. (g, h) Space-filling representations
of solvate 2a (CHCl3)2(CS2), viewed approximately perpendicular and
3
’ RESULTS AND DISCUSSION
parallel to the molecular quasi-C5 axis. Hydrogen atoms are omitted for
clarity in panels a and e.
Preparation and Characterization. Gold(I) Complexes. To
simplify the characterization of reaction products, the pentaaryꢀl-
[60]fulleride anions (4-MeC6H4)5C60ꢀ (1a) and(4-t-BuC6H4)5C60
(1b) were chosen as ligands owing to the simplicity of the 1H and
13C NMR spectra of their protonated precursors, as well as the
spectra of their corresponding metal complexes. Also, we have
found that the corresponding pentaarylfullerenes and their deriva-
tives are particularly crystalline, facilitating characterization by
single-crystal X-ray diffraction. Addition of ClAu(PPh3) to a
solution of K[(4-MeC6H4)5C60] (1a Kþ) in THF followed by
3
purification by chromatography on silica gel afforded [(4-
MeC6H4)5C60]Au(PPh3) (2a) in 70% yield. The tert-butyl deri-
vative [(4-t-BuC6H4)5C60]Au(PPh3) (2b) was prepared in a
ꢀ
similar fashion from (4-t-BuC6H4)5C60 (1b) in 95% yield. The
6842
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
Table 1. Crystallographic Interatomic Distances and Bond Angles for Gold(I) Complexes 2a (ODCB)2, 2a (CHCl3)2(CS2), and
ARTICLE
3
3
2b (CHCl3)3; Silver(I) Complexes 3a (C5H12)0.6(C2H2Cl4)3 and 3b (C4H8O)0.5(C6H6)0.5 (Molecules A and B); and Copper(I)
3
3
3
Complexes 4a (CS2)3 and 4b (CHCl3)3
3
3
compound
2a
η1/η3
2a
η1/η3
2b
3a
3b(A)
η2/η3
3b(B)
η2/η3
4a
4b
coordinationa
metal
η1/η3
Au
η2/η3
η5
η5
Au
Au
Ag
Ag
Ag
Cu
Cu
phosphine ligand
included solvent
PPh3
PPh3
PPh3
PCy3
PPh3
PPh3
PPh3
PPh3
2 ODCB
2 CHCl3 CS2
3 CHCl3
0.6 C5H12
3 C2H2Cl4
0.5 THF
0.5 C6H6
0.5 THF
0.5 C6H6
3 CS2
3 CHCl3
Interatomic Distances (Å)
M(1)ꢀP(1)
M(1)ꢀC(1)
M(1)ꢀC(2)
M(1)ꢀC(3)
M(1)ꢀC(4)
M(1)ꢀC(5)
M(1)ꢀC(9)
C(1)ꢀC(2)
C(2)ꢀC(3)
C(3)ꢀC(4)
C(4)ꢀC(5)
C(5)ꢀC(1)
2.2602(15)
2.160(6)
2.823(7)
3.452(6)
3.439(4)
2.782(5)
3.066(7)
1.469(7)
1.381(8)
1.441(8)
1.388(7)
1.467(8)
2.2602(11)
2.176(4)
2.590(4)
3.377(4)
3.592(4)
3.050(4)
3.039(4)
1.474(5)
1.380(5)
1.423(5)
1.383(5)
1.477(5)
2.254(2)
2.3669(12)
2.259(4)
2.525(4)
3.180(4)
3.359(4)
2.894(4)
3.198(4)
1.451(5)
1.407(6)
1.416(5)
1.399(5)
1.434(5)
2.3687(14)
2.288(4)
2.464(5)
3.224(4)
3.503(5)
3.051(5)
3.157(5)
1.459(6)
1.409(6)
1.413(6)
1.386(6)
1.434(6)
2.3830(15)
2.278(5)
2.548(5)
3.330(5)
3.583(5)
3.063(5)
3.085(5)
1.460(7)
1.405(6)
1.419(6)
1.395(7)
1.442(6)
2.148(8)
2.215(2)
2.171(2)
2.246(2)
2.352(2)
2.337(3)
2.1570(12)
2.224(4)
2.318(4)
2.345(4)
2.281(4)
2.190(4)
2.155(7)
2.729(9)
3.379(8)
3.449(8)
2.854(8)
3.084(8)
1.494(10)
1.374(10)
1.425(11)
1.381(10)
1.477(10)
1.418(3)
1.422(3)
1.429(3)
1.423(3)
1.430(3)
1.414(5)
1.423(5)
1.411(5)
1.431(5)
1.424(5)
Bond Angles (deg)
P(1)ꢀM(1)ꢀC(1)
M(1)ꢀC(1)ꢀC(2)
M(1)ꢀC(1)ꢀC(5)
M(1)ꢀC(1)ꢀC(9)
176.97(16)
100.4(4)
98.3(4)
172.12(10)
88.2(2)
177.3(2)
174.46(10)
82.7(2)
168.68(12)
78.8(3)
170.67(12)
82.9(3)
95.1(5)
69.5(1)
76.4(1)
139.2(2)
68.3(3)
73.3(3)
141.0(3)
111.8(2)
108.4(2)
101.9(5)
111.7(4)
100.8(2)
113.8(3)
107.9(3)
110.0(3)
108.8(3)
106.7(3)
110.6(3)
Pyramidalization Angles (θp)b (deg)
C(1)
13.5
5.8
5.2
4.5
5.3
6.9
13.4
5.6
4.3
5.3
6.1
6.9
13.4
5.9
5.3
4.3
6.1
7.0
10.2
7.9
5.8
4.8
6.4
7.0
10.1
7.8
5.5
5.2
6.3
7.0
10.2
7.9
6.1
5.2
6.4
7.2
7.3
5.5
6.7
6.8
6.6
6.6
6.3
6.3
6.8
6.3
7.3
6.6
C(2)
C(3)
C(4)
C(5)
C(1)ꢀC(5) av
a Coordination of the cyclopentadienide unit in the fulleride ligands 1a or 1b with the metal M(1). b Pyramidalization angles calculated with the π-orbital
axis vector (POAV) model.14
1H and 13C NMR spectra of 2a and 2b display signal patterns and
intensities corresponding to C5 symmetry down to ꢀ80 °C, which
indicates either static pentahapto bonding or fast metallotropic
isomerization. Fluxional behavior involving η1 f η2 f η1
isomerization has been reported for other cyclopentadienyl gold(I)
phosphine complexes.8
Single crystals of 2a were obtained by slow diffusion of
n-pentane into a 1,2-dichlorobenzene (ODCB) solution. X-ray
diffraction analysis shows that compound 2a forms a solvate with
ODCB and crystallizes in the monoclinic space group C2/c with
distances of 2.823(7) and 2.782(5) Å. This coordination mode
is characteristic of cyclopentadienyl gold(I) phosphine com-
plexes and is also observed in the related carbaborane anion
[10-endo(7,8-nido-C2B9H11)Au(PPh3)]ꢀ.15 The principal Au-
(1)ꢀC(1) bond (2.160(6) Å) is elongated in comparison to a
typical AuꢀC(sp3) σ bond with no π-bonding contribution such
as MeAu(PPh3) (2.06ꢀ2.10 Å).16 The angles Au(1)ꢀC(1)ꢀ
C(2) and Au(1)ꢀC(1)ꢀC(5) (100.4(4) and 98.3(4)°, re-
spectively) indicate the approximately symmetrical position of
the Au(PPh3) fragment above carbons C(1), C(2), and C(5) of
the fullerene cyclopentadienyl ring. Accordingly, the Au(1)ꢀ
C(1)ꢀC(9) angle of 110.6(3)° is significantly greater than the
Au(1)ꢀC(1)ꢀC(2) and Au(1)ꢀC(1)ꢀC(5) angles (100.4(4)
and 98.3(4)°, respectively). Atoms P(1), Au(1), and C(1) are
approximately collinear with a P(1)ꢀAu(1)ꢀC(1) angle of
176.97(16)°. The cyclopentadienyl ring is markedly nonplanar
with a fold angle of 7.6(6)° between the plane defined by C(1),
C(2) and C(5) and the mean plane defined by C(2), C(3), C(4)
2a (ODCB)2 stoichiometry. Representations of the solid-state
3
structure are shown in Figure 1 and crystallographic dimensions
as well as diffraction parameters for all compounds except 3c are
presented in Tables 1 and 2. The embedded fullerenyl cyclo-
pentadienyl ring in 2a (ODCB)2 displays slightly distorted η1
3
coordination to the Au(PPh3) fragment leaning toward η3
(denoted η1/η3), with a principal Au(1)ꢀC(1) distance of
2.160(6) Å and secondary Au(1)ꢀC(2) and Au(1)ꢀC(5)
6843
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
Table 2. Selected Single-Crystal X-ray Diffraction Parameters
complex
2a
2a
2b
3a
3b
4a
3(CS2)
4b
solvent
2(C6H4Cl2)
2(CHCl3)
CS2
3(CHCl3)
0.6(C5H12
)
0.5(C4H8O)
0.5(C6H6)
3(CHCl3)
3(C2H2Cl4)
formula
C
125H58AuCl4P C116H52AuCl6PS2 C131H83AuCl9P C122H81.2AgCl12P C133H87AgO0.5P
C116H50CuPS6 C131H83Cl19CuP
formula weight
description
crystal size (mm3)
1929.45
red platelet
0.13 ꢁ 0.07
ꢁ 0.04
1950.33
red block
0.40 ꢁ 0.30
ꢁ 0.15
2203.96
2111.31
red prism
0.20 ꢁ 0.12
ꢁ 0.06
1831.87
red block
0.20 ꢁ 0.18
ꢁ 0.18
1730.43
red platelet
0.50 ꢁ 0.18
ꢁ 0.10
2070.53
red needle
0.10 ꢁ 0.05
ꢁ 0.02
red block
0.14 ꢁ 0.08
ꢁ 0.03
T (K)
100(2)
100(2)
100(2)
100(2)
100(2)
100(2)
100(2)
radiation
crystal system
space group
a (Å)
Mo KR
Mo KR
Mo KR
Mo KR
Mo KR
monoclinic
P21/c
Mo KR
Mo KR
triclinic
monoclinic
C2/c
triclinic
triclinic
monoclinic
P21/n
triclinic
P1
P1
P1
P1
15.9259(14)
16.3684(14)
18.8575(16)
89.7740(10)
69.3100(10)
63.8850(10)
4061.5(6)
2
37.699(6)
20.948(4)
20.166(3)
90.00
15.1223(16)
15.9245(17)
23.844(3)
88.4870(10)
86.7880(10)
72.517(2)
5467.9(11)
2
14.227(3)
14.836(3)
24.446(5)
73.894(3)
76.732(3)
68.275(3)
4558.6(17)
2
26.000(4)
24.254(4)
28.202(4)
90.00
17.674(4)
23.521(5)
18.549(4)
90.00
15.9694(13)
16.1495(14)
20.9835(18)
81.0180(10)
83.4150(10)
63.6490(10)
4783.5(7)
2
b (Å)
c (Å)
R (deg)
β (deg)
γ (deg)
V (Å3)
Z
97.111(2)
90.00
94.136(2)
90.00
96.942(3)
90.00
15803(5)
8
17738(5)
8
7655(3)
4
F
calcd (Mg mꢀ3
)
1.578
1.640
1.339
1.538
1.372
1.502
1.438
3
μ (mmꢀ1
)
2.024
2.198
1.630
0.648
0.305
0.530
0.554
2θmin, 2θmax (deg)
no. refl (unique)
no. refl [I > 2σ(I)]
Rint
7.34, 56.68
19762
7.60, 58.24
21137
7.28, 56.04
25781
7.56, 58.28
33453
7.40, 52.80
36164
7.48, 58.36
20562
7.28, 56.68
23414
12821
17161
16015
24295
19816
13246
9104
0.0984
0.0497
0.1013
0.0
0.0
0.0920
0.0971
R1, wR2 (all data)
0.1164, 0.1316
0.0615, 0.1228
0.0464, 0.1146
1.101
0.1403, 0.2422
0.0859, 0.2193
1.000
0.1162, 0.2422
0.0879, 0.2190
1.030
0.1470, 0.1983
0.0692, 0.1615
1.019
0.1021, 0.1502
0.0546, 0.1262
1.009
0.1747, 0.1761
0.0661, 0.1453
0.827
R1, wR2 [I > 2σ(I)] 0.0616, 0.1113
GOF
0.989
Δ, e Åꢀ3
ꢀ1.401, 1.583
ꢀ2.759, 2.731
ꢀ2.020, 2.412
ꢀ2.469, 1.081
ꢀ1.205, 0.788
ꢀ0.734, 0.507
ꢀ1.021, 1.055
3
and C(5). Atoms C(2), C(3), C(4) and C(5) are sp2-hybridized
and have little pyramidalization, as judged from their π-orbital
axis vector angles (POAV, θp = θσπ ꢀ 90, Table 1),14 whereas
C(1) has appreciable sp3 character and pyramidalization (13.5°).
The η1/η3 coordination and increase in π-bonding relative to a σ-
bonded system is further illustrated by comparison of the structure of
compound 2a with that of the previously reported pentaaryl-
[60]fullerene [4-(EtO2C)C6H4]5C60Me, in which a methyl group
is purely σ-bonded to the cyclopentadienyl ring.17 Importantly, the
crystal structure of [4-(EtO2C)C6H4]5C60Me is atypical in that it
shows very little crystallographic disorder, allowing a meaningful
analysis of bond lengths and angles within the cyclopentadienyl ring.
There is significant bond-length alternation in the cyclopentadienyl
ring of compound 2a (Table 1). The alternation is intermediate
between that of a conjugated diene and an aromatic cyclopentadienyl
anion. On the other hand, the bond-length alternation within the
cyclopentadiene ring of C60[4-C6H4(CO2Et)]5Me is significantly
larger, with formal single and double bonds of 1.523 and 1.349 Å,
respectively.17 Furthermore, the pyramidalization angle of the sp3
cyclopentadienyl carbon atom is 17.4° in the latter, which is 3.9°
larger than the corresponding carbon atom in compound 2a, and
accordingly reflects its greater sp3 character. The angles C(2)ꢀC-
(1)ꢀC(Me) and C(5)ꢀC(1)ꢀC(Me) of C60[4-C6H4(CO2Et)]5-
Me are significantly larger (102.7° and 105.5°, respectively) than
those of 2a (100.4° and 98.3°, respectively), reflecting the contribu-
tion of π bonding between the cyclopentadienyl system and the
Au(PPh3) fragment in the gold(I) complex 2a.
X-ray diffraction analysis was also performed on a single crystal of
compound 2a grown from a CS2/CHCl3 mixed-solvent system.
Under these conditions, compound 2a crystallizes in the C2/c space
group with 2a (CHCl3)2(CS2) stoichiometry. Similarly to the
3
ODCB solvate, the cyclopentadienyl ring displays η1/η3 coordina-
tion to gold(I) (Figure 1f). However, there is significant deviation
toward η2 coordination in this case. Although the principal Au-
(1)ꢀC(1) distance of 2.176(4) Å is nearly identical to that of the
ODCB solvate, the secondary interactions Au(1)ꢀC(2) and Au-
(1)ꢀC(5) are 2.590(4) and 3.050(4) Å, respectively, and the
corresponding angles Au(1)ꢀC(1)ꢀC(2) and Au(1)ꢀC(1)ꢀC-
(5) are 88.2(2)° and 111.8(2)°, illustrating the asymmetry of the
metalꢀligand coordination. Additionally, a carbon disulfide mole-
cule resides in a narrow cavity between the Au(PPh3) moietyandan
adjacent fullerene-anchored tolyl group, which may be responsible
for the observed η1 f η2 distortion. Although there is significant
distortion toward η2 coordination, the cyclopentadienyl ring in
2a (CHCl3)2(CS2) has partially localized 1,3-diene character, and
3
the CꢀC distances are similar to those of solvate 2a (ODCB)2
3
(Table 1). This indicates that the η1fη2 distortion does not
involve a fundamental change in the ligandꢀmetal bonding.
6844
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
Figure 2. Representations of the crystal structure of 2b (CHCl3)3.
3
(a) ORTEP drawing (thermal ellipsoids at 50% probability) of the
partial structure of 2b showing metalꢀligand bonding. (b) Space-filling
representation of 2b viewed along the molecular quasi-C5 axis.
Figure 3. Expansion of the proton-decoupled 31P NMR spectrum of
silver(I) complex 3b showing two doublets arising from 109Agꢀ P (647
31
Finally, X-ray diffraction analysis was also performed on a
single crystal of the tert-butyl derivative 2b grown from a CS2/
CHCl3 mixed-solvent system (Figure 2). Under these condi-
Hz) and 107Agꢀ P (560 Hz) one-bond coupling.
31
tions, compound 2b forms crystals with 2b (CHCl3)3 stoichi-
with K[(4-t-BuC6H4)5C60] (1b Kþ) were carefully carried out,
3
3
ometry. The coordination of the fullerene cyclopentadienyl ring
resulting in the formation of the silver(I) phosphine complexes
[(4-MeC6H4)5C60]Ag(PCy3) (3a), [(4-t-BuC6H4)5C60]Ag-
(PPh3) (3b), and [(4-t-BuC6H4)5C60]Ag(PCy3) (3c),
respectively.23 Compounds 3aꢀc were too air-sensitive for
purification by chromatography on silica gel. Consequently, they
were isolated by removal of THF from the reaction mixture in
vacuo, followed by extraction with CS2 or benzene, and subse-
quent filtration under argon. Using this method, fractional
amounts of (triphenylphosphine)silver(I) chloride were present
in the final products before their crystallization, as determined by
NMR spectroscopy.
to gold(I) is similar in 2b (CHCl3)3 and 2a (ODCB)2, i.e., a
3
3
nearly symmetrical η1/η3 coordination is observed (Figure 2a).
The “secondary” interactions Au(1)ꢀC(2) and Au(1)ꢀC(5)
(2.729(9) and 2.854(8) Å, respectively) confirm a slight distor-
tion toward η2 coordination.
The isolobal relationship18 between the proton and the
cationic Au(PR3)þ fragment can be invoked to outline the
bonding observed in complexes 2a and 2b.19 Whereas the vacant,
degenerate gold-centered 6px and 6py orbitals of the Au(PR3)þ
fragment are too high in energy to interact strongly with the filled
cyclopentadienyl-based orbitals of appropriate symmetry, the
hybrid 6s/6pz orbital behaves similarly to the vacant 1s orbital of
the proton. Pentahapto bonding is disfavored as this would
involve destabilizing interactions between the gold(I) 5d orbitals
and filled cyclopentadienyl π orbitals. Rather, an energetic
compromise is reached with the η1/η3 geometry. On the other
hand, pentahapto bonding is found in the anionic gold
heteroborane clusters [3-(PPh3)-closo-2,1-AuTeB10H10]ꢀ and
[3-(PPh3)-closo-3,1,2-AuAs2B9H9]ꢀ, which possess cyclopenta-
dienyl-type ligands with energetically higher-lying molecular
orbitals.20
Similarly to the gold and copper complexes, the simple 1H and
13C NMR spectra indicate molecular C5 symmetry for all three
complexes 3aꢀc arising from either static η5 coordination or fast
metallotropic isomerization. Complexes 3aꢀc are unusual in that
two doublets are clearly visible in the 31P{1H} NMR spectra
1
31
1
31
(Figure 3), arising from J(109Agꢀ P) and J(107Agꢀ P) cou-
plings (3a, 637 and 552 Hz, respectively; 3b, 647 and 560 Hz,
respectively; 3c, 633 and 548 Hz, respectively). In general, the
phosphorus resonances in unhindered complexes appear as broad
singlets at room temperature because of fast ligand exchange.24 A
further indication that the observed doublets are due to silverꢀpho-
1
31
31
Silver(I) Complexes. In the silver(I) series of cyclopentadienyl
complexes, the compound [C5H2(SiMe3)3]Ag[P(n-Bu)3] is the
only “pristine” CpꢀAg(I) complex present in the literature.9 The
related complexes {[(MeO2C)5C5]Ag(PPh3)}2 and [(MeO2
C)5C5]Ag(PPh3)2 possess η2-bonded pentakis(methoxycar
bonyl)cyclopentadienyl ligands,21 with silver centers having ester
carbonyl groups and water molecules within their coordination
sphere that compete with the cyclopentadienyl ligand for binding
and interfere with a higher hapticity metalꢀcyclopentadienyl
bonding.
sphorus coupling is shown by the J(109Agꢀ P)/1J(107Agꢀ P)
ratio, which corresponds to the quotient of the gyromagnetic ratios
for 109Ag and 107Ag (1.15).24a Even with complex 3b, where the
triphenylphosphine ligand is not as bulky as tricyclohexylphosphine,
the crowded environment around the metal center impedes fast
ligand exchange.
Interestingly, compounds 3aꢀc are relatively air stable and
can be manipulated at 20 °C. This is in stark contrast to the
reported instability of parent complex CpAg(PPh3), which
decomposes within minutes at 20 °C under inert atmosphere.22
The enhanced stability is presumably due to steric shielding
provided by the large aryl groups of fulleride ligands 1a and 1b
and is consistent with reports that silver(I) phosphine complexes
can be stabilized with substituted cyclopentadienyls.22b Accord-
ingly, single crystals of 3a and 3b could be grown under argon at
20 °C in the dark over several days.
We envisioned that the steric encumbrance of the pentaar-
ylfulleride ligand 1 would provide a particularly stabilizing
environment for enhanced metalꢀcyclopentadienyl bonding
by preventing the coordination of other ligands to silver. In
addition, silver(I) cyclopentadienyl complexes are notoriously
unstable, and the aryl groups of 1a or 1b should protect the
resulting complexes and allow isolation and characterization.22
Accordingly, a series of low-temperature reactions using ClAg-
Representations of the crystal structures of 3a and 3b are
shown in Figure 4, and crystallographic dimensions are collected
in Table 1. Slow diffusion of n-pentane into a solution of 3a in
(PCy3) with K[(4-MeC6H4)5C60] (1a Kþ) and ClAg(PPh3)
3
6845
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
of coordination geometry is apparent from the small differences in
interatomic distances (0.266(4) and 0.635(4) Å, respectively)
between the principal metalꢀligand interaction Ag(1)ꢀC(1)
(2.259(4) Å) and the secondary interactions Ag(1)ꢀC(2)
(2.525(4) Å) and Ag(1)ꢀC(5) (2.894(4) Å). Furthermore, in
comparison to the gold complexes 2a and 2b (Table 1), a reduction
in the metalꢀligand bond angles Ag(1)ꢀC(1)ꢀC(2) (82.7(2)°)
and Ag(1)ꢀC(1)ꢀC(5) (100.8(2)°), together with a correspond-
ing increase in the angle Ag(1)ꢀC(1)ꢀC(9) (113.8(3)°), illus-
trates an increase in hapticity of metalꢀligand bonding from η1 to
η3. The bond alternation within the cyclopentadienyl ring of the
silver(I) complex 3a is markedly reduced in comparison to the gold
complexes 2a and 2b, lying between a bond-localized butadiene
system and a π-delocalized cyclopentadienyl system with local C5
symmetry. However, there is stronger preference for π-delocaliza-
tion in 3a compared to the gold(I) complexes 2a and 2b (Table 1).
The POAV angles for the cyclopentadienyl carbon atoms C-
(1)ꢀC(5) indicate that the curvature is distributed more equally
between the five carbon atoms of this convex ligand compared to
gold complexes 2a and 2b. Carbon C(1) has less sp3 character
(POAV = 10.2°), and the remaining atoms C(2) to C(5) have
slightly higher sp3 character (POAV = 4.8ꢀ7.9°). These angles
indicate a geometry intermediate between the η1/η3 coordination
of gold complexes 2a and 2b and the η5 coordination of copper
complexes 4a and 4b.
Slow diffusion of n-pentane into a THF/benzene solution of the
tert-butyl derivative [(4-t-BuC6H4)5C60]Ag(PPh3) (3b) at room
temperature in the dark resulted in the growth of crystals suitable for
single-crystal X-ray diffraction analysis (Figure 4eꢀh).23 Under these
conditions, complex 3b forms a solvate (3b (C4H8O)0.5(C6H6)0.5)
3
with two independent molecules in the unit cell, which are referred to
as 3b(A) and 3b(B) in the following discussion. For each complex,
an occluded solvent molecule resides within a narrow, shallow cavity
between the PPh3 unit and an adjacent 4-tert-butylphenylene group.
In 3b(A), a benzene molecule is occluded, whereas in 3b(B) the
THF occluded molecule is disordered over two positions. The
geometries of the two molecules are similar and only 3b(A) is
described (Table 1, Figure 4eꢀh). As found for the silver complex 3a
described above, a situation between η2 and η3 coordination is
observed. In the case of 3b(A) however, the distortion toward η2 is
more severe. There is a difference of 0.587(5) Å between the
interatomic distances Ag(1)ꢀC(2) and Ag(1)ꢀC(5) (2.464(5)
and 3.051(5) Å, respectively), which is significantly greater than the
corresponding difference of 0.369(5) Å in 3a. The distortion from η3
to η2 in 3b(A) is also apparent from the large difference in bond
angle (29.1(3)°) between the ligandꢀmetal angles Ag(1)ꢀC-
(1)ꢀC(2) and Ag(1)ꢀC(1)ꢀC(5) (78.8(3)° and 107.9(3)°,
respectively), compared to the corresponding difference of
18.1(2)° in 3a. Thus, the hapticity is more accurately described as
η2 with a slight distortion toward η3. Indeed, the Ag(1)ꢀC(1) and
Ag(1)ꢀC(2) distances (2.288(4) and 2.464(5) Å, respectively)
differ only by 0.176(5) Å, and the Ag(1)ꢀC(3) and Ag(1)ꢀC(5)
distances (3.224(4) and 3.051(5) Å, respectively) only by 0.173(5)
Å. However, although the metalꢀligand distances indicate η2
coordination, bond lengths and pyramidalization angles within the
cyclopentadienyl ring are consistent with η3 coordination and are
similar to those of 3a.
Figure 4. Representations of the crystal structures of 3a (C5H12)0.6
-
3
(C2H2Cl4)3 (aꢀd) and 3b (C4H8O)0.5(C6H6)0.5, molecule 3b(A)
3
(eꢀh). (a) ORTEP drawing (thermal ellipsoids at 50% probability) of
compound 3a. (b) Partial structure of 3a showing the metalꢀligand
bonding. (c, d) Space-filling representations of compound 3a, viewed
approximately parallel and perpendicular to the bonding cyclopentadienyl
ring. (e) ORTEP drawing (thermal ellipsoids at 50% probability) of
compound 3b(A). (f) Partial structure of 3b(A) showing the ligandꢀmetal
bonding. (g, h) Space-filling representations of compound 3b(A), viewed
approximately parallel and perpendicular to the bonding cyclopentadienyl
ring. Hydrogen atoms are omitted for clarity in panels a and e.
1,1,2,2-tetrachloroethane resulted in the formation of single
crystals of a solvate with 3a (C5H12)0.6(C2H2Cl4)3 stoichiom-
3
etry. The extended crystal packing structure of 3a has no eye-
catching features and there are no short intermolecular contacts.
Importantly, a distorted η3 coordination with a significant
deviation toward η2 (i.e., η2/η3) is observed for complex 3a, which
is unexpected and different from the η5 and η3 coordination modes
reported for complex [C5H2(SiMe3)3]Ag[P(n-Bu)3].9 The change
The coordination modes observed in the crystal structures of 3a
and 3b are different from those previously reported for the complex
[C5H2(SiMe3)3]Ag[P(n-Bu)3] (Figure 5).9 In this complex, the
asymmetric unit comprises two isomers with different coordination
modes; one isomer displays η5 coordination, the other η3. The η3
6846
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
Figure 5. Representations of the crystal structure of [C5H2(SiMe3)3]
Ag[P(n-Bu)3].9 Views perpendicular to the (a) η3- and (b) η5-coordi-
nated cyclopentadienyl rings of the complex.
isomer has higher hapticity ligandꢀmetal bonding than is observed
in compounds 3a and 3b in that it is distorted toward η5
coordination with a longer, principal Ag(1)ꢀC(1) interatomic
distance (2.293(5) Å) and shorter, secondary interactions for
Ag(1)ꢀC(2) (2.421(5) Å) and Ag(1)ꢀC(5) (2.644(5) Å). As
discussed in the theoretical section below, the coordination pre-
ference of silver(I) in these complexes is intermediate between that
of gold(I) and copper(I).
Copper(I) Complexes. Copper pentaarylfulleride complexes are
of particular interest since the group of Nakamura has proposed
them as intermediates in the copper(I)-mediated synthesis of
pentaaryl hydrofullerenes.10a Sawamura et al. reported the complex
(Ph5C60)Cu(PEt3) synthesized via the reaction between Ph5C60H
1
and (t-BuO)Cu(PEt3). The H and 13C NMR spectra indicated
apparent C5 symmetry and the authors suggested η5-coordination.
However, considering the dearth of crystal structures of copper(I)
cyclopentadienyl complexes and the possibility for fluxional pro-
cesses showing higher apparent symmetry than that of the actual
coordination geometry, we decided to investigate the (triphenyl
phosphine)copper(I) pentaarylfulleride complexes derived from the
ligand precursors 1a and 1b. Thus, the copper complex [(4-
MeC6H4)5C60]Cu(PPh3) (4a) was synthesized via the reaction
of [ClCu(PPh3)]4 with K[(4-MeC6H4)5C60] (1a Kþ) in THF
3
at ꢀ60 C.23 HPLC analysis23 of the reaction mixture indicated
complete consumption of the fulleride starting material with con-
comitant formation of a single eluting product. Purification by
chromatography on silica gel was not possible because of the high
sensitivity of compound 4a to air while bulk purification by crystal-
lization was unsatisfactory. Rather, compound 4a was purified by
removal of THF in vacuo, followed by extraction with CS2 and
subsequent anaerobic filtration to remove salts and polar side
products.
Figure 6. Representations of the crystal structures of 4a (CS2)3 (aꢀd)
3
and 4b (CHCl3)3 (eꢀh). (a) ORTEP drawing (thermal ellipsoids at
3
50% probability) of compound 4a. (b) Partial structure of 4a showing
the ligandꢀmetal bonding. (c, d) Space-filling representations of
compound 4a, viewed approximately parallel and perpendicular to the
bonding cyclopentadienyl ring. (e) ORTEP drawing (thermal ellipsoids
at 50% probability) of compound 4b. (f) Partial structure of 4b showing
the metalꢀligand bonding. (g, h) Space-filling representations of
compound 4b, viewed approximately parallel and perpendicular to the
bonding cyclopentadienyl ring. Hydrogen atoms are omitted for clarity
in panels a and e.
Single crystals of copper(I) complex 4a were obtained by
diffusion of n-pentane into a CS2 solution at 20 °C in the dark
over 7 days under argon. A similar reaction between K[(4-t-
BuC6H4)5C60] (1b Kþ) and [ClCu(PPh3)]4 afforded the com-
3
plex [(4-t-BuC6H4)5C60]Cu(PPh3) (4b). In this case, it was
possible to isolate compound 4b in 65% yield via fast chroma-
tography on a short silica gel column under argon and to obtain
single crystals by slow diffusion of n-pentane into a solution of 4b
in CHCl3. Single-crystal X-ray diffraction analyses were possible
for both 4a and 4b (Figure 6).23 Compounds 4a and 4b form
solvates with 4a (CS2)3 and 4b (CHCl3)3 stoichiometry, re-
cyclopentadienyl ring vary only to a small degree (max deviation of
0.012 Å in the case of 4a (CS2)3) and the pyramidalization angles of
3
3
3
spectively. Representations of the crystal structures are shown in
Figure 6 and crystallographic dimensions are collected in Table 1.
The bonding in complexes 4a (CS2)3 and 4b (CHCl3)3 is
the carbon atoms of the cyclopentadienyl ring are nearly equal. These
variations in bond lengths and angles are comparable to those of
pentahapto “buckyferrocenes”, e.g., (η5-C60Me5)FeCp,12d andalsoto
those of other copper(I) cyclopentadienyl phosphine complexes.7
Comparison of the whole series of crystal structures
(Table 1) reveals that the shortest AgꢀC interatomic distances
(2.259(4)ꢀ2.288(4) Å) are significantly longer than the corre-
sponding shortest AuꢀC distances (2.155(4)ꢀ2.176(4) Å).
3
3
essentially identical and can be described as pentahapto with a very
slight deviation toward η3 coordination. Thus, in 4a (CS2)3 the
difference between the shortest and longest metalꢀligand interatomic
distances, Cu(1)ꢀC(2) (2.171(2) Å) and Cu(1)ꢀC(4) (2.352(2)
Å), respectively, is 0.181(2) Å. The bond lengths within the
3
6847
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
Table 3. DFT-Calculated Interatomic Distances and Bond Angles for Simplified Models of Fullerenyl and Cyclopentadienyl
Gold(I), Silver(I), and Copper(I) Phosphine Complexes Using the PBE Hybrid Functional with the 6-31G(d) Basis Set for H, C,
and P, and the Stuttgart/Dresden ECP Plus DZ (SDD) Basis Set for Au, Ag, and Cu
metal
Au
Au
Ag
Cu
coordination
Cp ligand
η1
η2
H5C60
η1
η2
Cp
η3
η5
η5
η5
H5C60
PH3
Cp
PH3
H5C60
PH3
Cp
PH3
H5C60
PH3
Cp
PH3
PH3
PH3
phosphine
Interatomic Distances (Å)
M(1)ꢀP(1)
M(1)ꢀC(1)
M(1)ꢀC(2)
M(1)ꢀC(3)
M(1)ꢀC(4)
M(1)ꢀC(5)
C(1)ꢀC(2)
C(2)ꢀC(3)
C(3)ꢀC(4)
C(4)ꢀC(5)
C(5)ꢀC(1)
2.299
2.126
2.775
3.408
3.408
2.775
1.481
1.374
1.430
1.374
1.481
2.272
2.276
2.978
3.318
2.978
2.276
1.423
1.401
1.401
1.423
1.472
2.296
2.115
2.770
3.467
3.465
2.768
1.467
1.374
1.434
1.374
1.467
2.257
2.278
2.910
3.229
2.910
2.278
1.420
1.404
1.404
1.420
1.458
2.350
2.212
2.562
3.007
3.008
2.565
1.448
1.405
1.410
1.405
1.448
2.300
2.460
2.465
2.468
2.465
2.460
1.423
1.423
1.423
1.423
1.423
2.140
2.223
2.226
2.223
2.226
2.223
1.423
1.423
1.423
1.423
1.423
2.128
2.210
2.211
2.212
2.211
2.211
1.423
1.423
1.423
1.423
1.423
Bond Angles (deg)
P(1)ꢀM(1)ꢀC(1)
M(1)ꢀC(1)ꢀC(2)
M(1)ꢀC(1)ꢀC(5)
179.4
99.0
99.0
161.1
104.9
71.1
179.7
161.3
101.2
71.3
176.8
86.2
86.3
150.8
73.4
73.2
147.3
71.5
71.3
147.1
71.3
71.2
99.7
99.6
Because the coordination of the cyclopentadienyl ring to each of
these two metal centers is similar, the differences in interatomic
distances reflect the greater covalent radii of silver (1.45 Å) relative
to gold (1.36 Å).25 However, although the covalent radius of
copper (1.32 Å) is less than that of gold, the shortest CuꢀC
distances (2.171(2) and 2.190(4) Å) are slightly longer than those
of the Au complexes because the higher coordination of the
cyclopentadienyl ring to copper(I) (η5) results in an elongation
of the shortest metalꢀcarbon bonds.
H, C, P and the Stuttgart/Dresden ECP plus DZ (SDD) basis set
for Au, Ag, and Cu because this level of theory has been studied
comparatively and this method reproduces the geometry of third
row transition metal complexes reasonably well.27b We were also
interested in determining whether the convexity of the cyclo-
pentadienyl ligand in the fulleride complexes has an influence on
metalꢀligand bonding interactions. The simplified fulleride
model (H5C60)M(PH3) (M = Au, Ag, Cu) avoids potential
steric influences by larger substituents on the coordination
geometry, which is important in view of the small energy
differences that can exist between different coordination modes
in these systems.27a With the larger model (Ph5C60)M(PPh3)
(M = Au, Ag, Cu), the bulky phenyl groups can have a dramatic
influence on the calculated geometry, resulting in unexpected
geometries.28
Gold(I) Complexes. The minimum for CpAuPH3 corresponds
to a symmetrical η1 coordination with a marked bond alternation
within the cyclopentadienyl ring (Table 3).23 Additionally, an η2-
coordinated species with a relative energy of þ2.4 kcal/mol after
zero-point energy correction was found as a transition state
linking the two equivalent η1-coordinated minima. These results
are consistent with previous computational results,27a crystal-
lographic data,8 and the results of our NMR spectroscopy
experiments, which indicate a time-averaged C5 symmetry re-
sulting from fast η1 f η2 f η1 metallotropic rearrangement.
Analogously, the energy difference between the η1 minimum and
the η2 transition state of the “bare” fulleride model
(H5C60)Au(PH3) is only þ2.3 kcal/mol after zero-point energy
correction, indicating that, for this gold(I) model, the convex
nature of the cyclopentadienyl ligand does not have a visible
influence on the distortion of the complex between the η1 and η2
coordination states.
Pentahapto bonding in cyclopentadienyl systems has been
rationalized using molecular orbital theory.19 Although the
cationic Au(PPh3)þ, Ag(PPh3)þ and Cu(PPh3)þ fragments
are isolobal, they vary greatly with respect to the relative energies
of their orbitals and the extent to which they can interact with the
molecular orbitals of the anionic cyclopentadienyl fragment.
Thus, unlike the 6px and 6py orbitals of the Au(PR3)þ fragment
described above, the vacant, degenerate 4px and 4py orbitals of
the Cu(PPh3)þ fragment are energetically well matched with the
filled cyclopentadienyl-based orbitals, and together they partici-
pate in stabilizing bonding interactions. Furthermore, relative to
the Au(PPh3)þ fragment, the filled d orbitals of the Cu(PPh3)þ
fragment are low in energy and do not engage in destabilizing
interactions with the filled cyclopentadienyl-based orbitals. To-
gether, these factors stabilize pentahapto bonding relative to
lower coordination modes.
DFT Calculations. We carried out density functional theory
(DFT) calculations using Gaussian 09,23,26 examining the model
system (H5C60)M(PH3) (M = Au, Ag, Cu) to evaluate energetic
and geometric aspects of the bonding modes in these complexes.
The cyclopentadienyl system CpMPH3 (M = Au, Ag, Cu) was
also calculated for comparison. We used the exact-exchange-
incorporated PBE hybrid functional with a 6-31G(d) basis set for
6848
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
geometry, but with slightly elongated bonds (0.012ꢀ0.015 Å)
between Cu(I) and the five carbon atoms C(1) to C(5)
compared to CpCuPH3. The calculated structures for both
models reproduce the experimental interatomic distances for
complexes 4a and 4b. For example, the five interatomic distances
between Cu(I) and carbon atoms C(1) to C(5) in (H5C60)
Cu(PH3) average 2.224 Å (Table 3), compared to the experi-
ment averages of 2.2642(2) and 2.2716(4) Å for 4a and 4b,
respectively (Table 1). The distances between the carbons within
cyclopentadienyl units for both models are nearly identical
(1.423 Å). The calculations reproduce the averaged experimen-
tally observed values for 4a and 4b (1.4244(3) and 1.4206(5) Å,
respectively), and underline the delocalized aromatic nature of
the fulleride-based cyclopentadienyl ligand observed experimen-
tally in these two complexes.
Frontier Molecular Orbitals. The frontier molecularorbitals of
the gold(I) and silver(I) model complexes share similar features
(Figure 7 and Supporting Information).23 In particular, the
HOMOs of the fulleride-based models (H5C60)Au(PH3) and
(H5C60)Ag(PH3) have strong contributions from the correspond-
ing cyclopentadienide π orbital, with small antibonding contribu-
tions from the 5dx2-y2 gold orbital and the 4dx2-y2 silver orbital. The
HOMO-4 for (H5C60)Au(PH3) is mainly a σ orbital between gold
and the cyclopentadienyl carbon C(1).23 The HOMO-2 and
HOMO-4 for the silver(I) model (H5C60)Ag(PH3) are similar,
while for the gold(I) model there is much less σ bond contribution
in the HOMO-2 compared to the HOMO-4.
In contrast, the fulleride copper model complex (H5C60)Cu-
(PH3) shows contributions of copper 4p and 3d orbitals to bonding
in the HOMO-1 and HOMO-2, as one would expect from the η5-
coordination geometry of this complex (Figure 7). The HOMO
of the model copper complex (H5C60)Cu(PH3) is a quasi-C5-
symmetric fullerene-based orbital.
Figure 7. Selected molecular orbitals for (H5C60)Au(PH3),
(H5C60)Ag(PH3), and (H5C60)Cu(PH3) plotted with an isovalue
of 0.015.
’ CONCLUSION
The pentaarylfulleride ligand 1, as represented by its specific
derivatives 1a and 1b, is highly effective in stabilizing group 11
metal complexes (2a, 2b, 4a, 4b, 3a, 3b, and 3c). It is clear that
the steric shielding provided by the aryl substituents on ligands
1a and 1b, in addition to that from triphenylphosphine or
tricyclohexylphosphine, is an important factor in the successful
isolation and crystallographic characterization of an underrepre-
sented family of phosphine cyclopentadienyl complexes. These
new compounds are also the first examples of fullerene com-
plexes of gold(I), silver(I), and copper(I) to be crystallographi-
cally characterized. The gold(I) complexes 2a and 2b coordinate
the fulleride-based cyclopentadienyl anion to the metal in an η1
Silver(I) Complexes. Previously reported computational
studies27a on the simplified system CpAgPH3 indicated that,
although η5 coordination is a minimum, the potential energy
surface with respect to ring slippage is broad and shallow, and a
range of coordination modes are accessible within a few kcal/mol
including η1 and η2. Since there is a discrepancy between the
optimized structure of CpAgPH3 and the solid-state structures of
3a and 3b, the fullerene-based model system (H5C60)Ag(PH3)
was also calculated. Confirming the earlier findings,27a the
minimum for CpAgPH3 has η5 coordination (Table 3). In
contrast, the minimum for (H5C60)Ag(PH3) has η3 coordina-
tion with a principal AgꢀC interaction of 2.212 Å and two
secondary AgꢀC interactions of 2.562 and 2.565 Å, indicating
that the convexity of the fulleride-based cyclopentadienyl ligand
may play a role in the coordination of this model system.
fashion, whereas in the case of 2a (CHCl3)2(CS2) there is
3
distortion of the complex toward η2 coordination that is pre-
sumably due to crystal packing forces. On the other hand,
pentahapto bonding characterizes the fulleride copper complexes
4a and 4b, as is typical for phosphine copper(I) cyclopentadienyl
complexes. The η2/η3-type bonding observed in the silver
complexes 3a and 3b, together with the previously reported
η3 and η5 bonding in the structure of [C5H2(SiMe3)3]Ag-
[P(n-Bu)3],9 shows that a range of coordination modes is
easily accessible in the phosphine silver(I) cyclopentadienyl
family due to the small energy changes accompanying these
distortions.
Copper(I) Complexes. In previous computational work,27a the
global minimum located for the model system CpCuPH3 had η5
coordination. The potential energy curve with respect to ring
slippage was found to be steep and narrow, contrary to that of the
analogous silver(I) complex. Our DFT calculations reproduce
these results.23,27a The minimum for CpCuPH3 has η5 coordina-
tion, with nearly equal bond distances between Cu(I) and the five
carbon atoms C(1) to C(5) of the cyclopentadienyl ligand
(Table 3). The fulleride model (H5C60)Cu(PH3) has similar
The DFT calculations reproduce interatomic distances and
bond angles observed experimentally, and provide insight into
6849
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
the electronic structures of these complexes. Bonding to the
fulleride ligand is dominated by orbital availability, as has been
previously shown, with a subtle influence from the pyramidaliza-
tion of the cyclopentadienyl carbons on the coordination of the
Au(I), Ag(I), and Cu(I) cations. The influence of pyramidaliza-
tion is strongest in the silver(I) complexes. All these effects work
synergistically to modulate orbital overlap between gold(I),
silver(I), or copper(I) and the fullerene cyclopentadienyl
orbitals.
bonding. This combination of factors results in the intermediate η2/
η3 coordination.
Accordingly, the greatly increased stability of metal complexes
resulting from the steric shielding imparted by the aryl groups of
fulleride anions 1a and 1b bodes well for the isolation of other
unstable transition metal complexes and even some reactive
species.31 Further work in this direction will be reported in
due time.
Thus, the inherent angle strain affecting all carbons of the
fullerene-embedded cyclopentadienyl ring may have significant
implications regarding the coordinating properties of ligands 1a
and 1b. This strain is imparted primarily by the spherical nature
of the fullerene cage and results in the pyramidalization of the
cyclopentadienyl carbons in all pentaarylfulleride complexes
(Chart 1b). Interestingly, for unstrained nonfullerenyl cyclopen-
tadienyl complexes, the substituents tilt either toward the metal
(hydrogens of Cp) as a result of a slight rehybridization of the
trigonal carbons leading to increased ligandꢀmetal orbital over-
lap or away for methyl groups and other larger substituents due
to supplementary steric factors.29 On the other hand, the
cyclopentadienyl π orbitals in the fulleride ligands 1a and 1b
are aligned radially with respect to the C60 cage because of the
convex shape of the embedded cyclopentadienyl moiety, rather
than perpendicularly to the plane of the Cp ring (Chart 1). As a
result of this unusual orientation, the 2p orbitals cannot interact
as efficiently with the d orbitals of a metal center in the η5
coordination geometry. This is most clearly seen experimentally
with unusually distinctive ligandꢀmetal distances in the “dual”
fulleroferrocene complex (η5-C60Me5)Fe(η5-C5H5), which has
both fulleride-based and Cp cyclopentadienyl ligands: the aver-
aged CꢀFe bond length between the Me5C60 carbons and
iron(II) is 2.089 Å, whereas it is significantly shorter (2.033 Å)
between the Cp carbons and iron(II).12d,30
’ ASSOCIATED CONTENT
S
Supporting Information. Spectroscopic characterization
b
data for compounds 3aꢀc and 4a,b. Crystallographic informa-
tion files in CIF format for 3a (C5H12)0.6(C2H2Cl4)3, 3b
3
3
(C4H8O)0.5(C6H6)0.5, 4a (CS2)3, and 4b (CHCl3)3. Compu-
3
3
tational details, frontier orbital representations, and atom co-
ordinates for the minimized structures. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
rubin@chem.ucla.edu
’ ACKNOWLEDGMENT
We are grateful to the National Science Foundation, the Office
of Naval Research, and the U.S. Department of Energy, Office of
Basic Energy Sciences as part of an Energy Frontier Research
Center, for financial support with research grants NSF-CHE-
0527015, NSF-CHE-0911758, ONR-N00014-04-1-0410, and
DOE-BES (EFRC DE-SC0001342) for Y.R., the Sloan Founda-
tion for P.L.D., and the NSF for instrumentation grants NSF-
CHE-9871332 (X-ray) and NSF-CHE-9974928 (NMR).
It is important to note that the crystallographically characterized,
noncoordinated metal fulleride salts [Li(THF)4][Ph5C60] or
[Zr(NMe2)3(THF)2][Ph5C60]12m have a similar curvature, as
shown by the averaged POAV values for the cyclopentadienyl
carbons (6.2 and 6.2, respectively). These values are comparable
to the POAV values of the sp2-hybridized carbons C(2)ꢀC(5) in
complexes 2a,b, 3a,b, or4a,b, whichrangefrom4.3to7.9(Table1).
Interestingly, this curvature does not change significantly upon
metal ligation.
However, at a localized level, the angle strain resulting from
pyramidalization of the fulleride cyclopentadienyl ring carbons is
distributed between all five atoms in the case of symmetrical η5
bonding (copper(I)). On the other hand, in the case of σ bonding
(η1-coordination), for example, in C60Ph5Me or (H5C60)Au(PH3),
four formally sp2-hybridized cyclopentadienyl carbons are now
significantly planarized, while the remaining sp3-hybridized C(1)
carbon atom is significantly pyramidalized, which reduces its
hybridization strain and stabilizes the overall complex. Within the
series of complexes described in this work, the copper(I) complexes
4a and 4b represent one extreme of the bonding spectrum, in which
the increase in stability associated with pentahapto bonding over
σ/η1 bonding more than compensates for the destabilization
associated with the greater angle strain at carbons C(1)ꢀC(5).
The gold complexes 2a and 2b represent the other extreme, in
which no stabilization is gained through pentahapto bonding and
the complex adopts η1 coordination. The silver complexes 3a and
3b are interesting intermediate cases, in which the relief of angle
strain is in close competition with the stabilizing effect of pentahapto
’ REFERENCES
(1) Van Peski, A. J.; Melsen, J. A. U.S. Patent 2,150,349, 1938.
(2) Kealy, T. J.; Pauson, P. L. Nature 1951, 168, 1039–1040.
Wilkinson, G.; Rosenblum, M.; Whiting, M. C.; Woodward, R. B.
J. Am. Chem. Soc. 1952, 74, 2125–2126.
(3) Wilkinson, G.; Piper, T. S. J. Inorg. Nucl. Chem. 1956, 2, 32–37.
(4) Cotton, F. A.; Marks, T. J. J. Am. Chem. Soc. 1969, 91, 7281–7285.
(5) (a) Cotton, F. A.; Takats, J. J. Am. Chem. Soc. 1970, 92, 2353–2358.
(b) Delbaere, L. T. J.; McBride, D. W.; Ferguson, R. B. Acta Crystallogr.
1970, B26, 515–521.
(6) Jutzi, P.; Wieland, W.; Neumann, B.; Stammler, H.-G. J. Orga-
nomet. Chem. 1995, 501, 369–374.
(7) (a) Hanusa, T. P.; Ulibarri, T. A.; Evans, W. J. Acta Crystallogr.
1985, C41, 1036–1038. (b) Anderson, Q. T.; Erkizia, E.; Conry, R. R.
Organometallics 1998, 17, 4917–4920. (c) Carriedo, G. A.; Howard,
J. A. K.; Stone, F. G. A. J. Chem. Soc., Dalton Trans. 1984, 1555–1561.
(d) Zybill, C.; M€uller, G. Organometallics 1987, 6, 2489–2494. (e) Ren,
H; Zhao, X.; Xu, S.; Song, H.; Wang, B. J. Organomet. Chem. 2006,
691, 4109–4113. (f) Olbrich, F.; Schmidt, G.; Weiss, E.; Behrens, U.
J. Organomet. Chem. 1993, 456, 299–303. (g) Johnson, A. L.; Willcocks,
A. M.; Raithby, P. R.; Warren, M. R.; Kingsley, A. J.; Odedra, R. Dalton
Trans. 2009, 922–924.
(8) (a) Perevalova, E. G.; Grandberg, K. J.; Dyadchenko, V. P.;
Baukova, T. V. J. Organomet. Chem. 1981, 217, 403–413. (b) Baukova,
T. V.; Slovokhotov, Yu. L.; Struchkov, Yu. T. J. Organomet. Chem. 1981,
220, 125–137. (c) Bruce, M. I.; Walton, J. K.; Skelton, B. W.; White,
A. H. J. Chem. Soc., Dalton Trans. 1983, 809–814. (d) Werner, H.; Otto,
H.; Ngo-Khac, T.; Burschka, Ch. J. Organomet. Chem. 1984,
262, 123–136. (e) Struchkov, Yu. T.; Slovokhotov, Yu. L.; Kravtsov,
6850
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851

Journal of the American Chemical Society
ARTICLE
D. N.; Baukova, T. V.; Perevalova, E. G.; Grandberg, K. J. J. Organomet.
Chem. 1988, 338, 269–280. (f) Schumann, H.; G€orlitz, F. H.; Dietrich, A.
Chem. Ber. 1989, 122, 1423–1426. (g) Bruce, M. I.; Humphrey, P. A.;
Williams, M. L.; Skelton, B. W.; White, A. H. Aust. J. Chem. 1989,
42, 1847–1857. (h) Kuz’mina, L. G.; Churakov, A. V.; Grandberg, K. I.;
Kuz’min, V. S. Koord. Khim. 1997, 23, 170–176.
(28) DFT calculations carried out on the series of metal complexes
(Ph5C60)M(PPh3) (M = Au, Ag, Cu) using ADF 2009 (Local Density
Approximation by Becke-Perdew with Triple-ζ STA Basis Sets for Au, Ag,
Cu, and P) gave geometries similar to those referred to in Table 3 in the
case of (Ph5C60)Au(PPh3) (η1) and (Ph5C60)Ag(PPh3) (η2/η3). How-
ever, the geometry optimization of (Ph5C60)Cu(PPh3) resulted in η1
coordination; see ref 23.
(9) Stammler, H.-G.; Jutzi, P.; Wieland, W.; Neumann, B. Acta
Crystallogr. 1998, C54, IUC9800064.
(29) See for example: (a) Haaland, A. Acc. Chem. Res. 1979,
12, 415–422. (b) Gl€ockner, A.; Tamm, M.; Arif, A. M.; Ernst, R. D.
Organometallics 2009, 28, 7041–7046.
(30) Nakamura, E. Pure Appl. Chem. 2003, 75, 427–434.
(31) (a) Warmuth, R.; Yoon, J. Acc. Chem. Res. 2001, 34, 95–105.
(b) Hof, F.; Craig, S. L.; Nuckolls, C.; Rebek, J., Jr. Angew. Chem., Int. Ed.
2002, 41, 1488–1508.
(10) (a) Matsuo, Y.; Nakamura, E. Chem. Rev. 2008, 108, 3016–3028.
(b) Sawamura, M.; Iikura, H.; Ohama, T.; Hackler, U. E.; Nakamura, E.
J. Organomet. Chem. 2000, 599, 32–36.
(11) The majority of metal complexes of fullerenes involve eta-2
coordination; see: (a) Balch, A. L.; Olmstead, M. M. Chem. Rev. 1998,
98, 2123–2165. (b) Yeh, W.-Y.; Tsai, K.-Y. Organometallics 2010,
29, 604–609 and references therein.
(12) (a) Sawamura, M.; Iikura, H.; Nakamura, E. J. Am. Chem. Soc.
1996, 118, 12850–12851. (b) Matsuo, Y.; Iwashita, A.; Nakamura, E.
Organometallics 2008, 27, 4611–4617. (c) Matsuo, Y.; Kuninobu, Y.;
Muramatsu, A.; Sawamura, M.; Nakamura, E. Organometallics 2008,
27, 3403–3409. (d) Sawamura, M.; Kuninobu, Y.; Toganoh, M.; Matsuo,
Y.; Yamanaka, M.; Nakamura, E. J. Am. Chem. Soc. 2002, 124, 9354–9355.
(e)Nakamura, E. J. Organomet. Chem. 2004, 689, 4630–4635. (f) Matsuo,
Y.; Kuninobu, Y.; Ito, S.; Nakamura, E. Chem. Lett. 2004, 33, 68–69.
(g) Matsuo, Y.; Nakamura, E. Organometallics 2003, 22, 2554–2563.
(h) Herber, R. H.; Nowik, I.; Matsuo, Y.; Toganoh, M.; Kuninobu, Y.;
Nakamura, E. Inorg. Chem. 2005, 44, 5629–5635. (i) Matsuo, Y.; Mitani,
Y.; Zhong, Y.-W.; Nakamura, E. Organometallics 2006, 25, 2826–2832.
(j) Matsuo, Y.; Iwashita, A.; Nakamura, E. Organometallics 2005,
24, 89–95. (k) Sawamura, M.; Kuninobu, Y.; Nakamura, E. J. Am. Chem.
Soc. 2000, 122, 12407–12408. (l) Kuninobu, Y.; Matsuo, Y.; Toganoh,
M.; Sawamura, M.; Nakamura, E. Organometallics 2004, 23, 3259–3266.
(m) Bouwkamp, M. W.; Meetsma, A. Inorg. Chem. 2009, 48, 8–9.
(13) Halim, M.; Kennedy, R. D.; Khan, S. I.; Rubin, Y. Inorg. Chem.
2010, 49, 3974–3976.
(14) (a) Haddon, R. C. J. Am. Chem. Soc. 1997, 119, 1797–1789.
(b) Haddon, R. C. J. Phys. Chem. A 2001, 105, 4164–4165.
(15) Hamilton, E. J. M.; Welch, A. J. Polyhedron 1990, 9, 2407–
2412.
(16) (a) Gavens, P. D.; Guy, J. J.; Mays, M. J.; Sheldrick, G. M. Acta
Crystallogr. 1977, B33, 137–139. (b) Liu, L.-K.; Luh, L.-S.; Wen, Y-S;
Eke, U. B.; Mesubi, M. A. Organometallics 1995, 14, 4474–4482.
(17) Zhong, Y.-W.; Matsuo, Y.; Nakamura, E. Org. Lett. 2006, 8,
1463–1466.
(18) Hoffmann, R. Angew. Chem., Int. Ed. 1982, 21, 711–724.
(19) (a) Mingos, D. M. P.; Forsyth, M. I.; Welch, A. J. J. Chem. Soc.,
Dalton Trans. 1978, 1363–1374. (b) Evans, D. G.; Mingos, D. M. P.
J. Organomet. Chem. 1982, 232, 171–191.
(20) Ferguson, G.; Gallagher, J. F.; Kennedy, J. D.; Kelleher, A.-M.;
Spalding, T. R. Dalton Trans. 2006, 2133–2139.
(21) Bruce, M. I.; Williams, M. L.; Skelton, B. W.; White, A. H.
J. Chem. Soc., Dalton Trans. 1983, 799–808.
(22) (a) Hofstee, H. K.; Boersma, J.; van der Kerk, G. J. M.
J. Organomet. Chem. 1976, 120, 313–317. (b) Lettko, L.; Rausch,
M. D. Organometallics 2000, 19, 4060–4065.
(23) See Supporting Information.
(24) (a) Dias, H. V. R.; Jin, W.; Kim, H.-J.; Lu, H.-L. Inorg. Chem.
1996, 35, 2317–2328. (b) Dias, H. V. R.; Singh, S. Inorg. Chem. 2004,
43, 7396–7402.
(25) Cordero, B.; Gꢀomez, V.; Platero-Prats, A. E.; Revꢀes, M.;
Echeverría, Cremades, E.; Barragꢀan, F.; Alvarez, S. Dalton Trans.
2008, 2832–2838.
(26) Frisch, M. J. et al. Gaussian 09, Revision A.02; Gaussian Inc.:
Wallingford, CT, 2009. See Supporting Information for the full citation
of Gaussian 09 and computational details.
(27) (a) Budzelaar, P. H. M.; Engelberts, J. J.; van Lenthe, J. H.
Organometallics 2003, 22, 1562–1576. (b) B€uhl, M.; Reimann, C.;
Pantazis, A. D.; Bredow, T.; Neese, F. J. Chem. Theory Comput. 2008,
4, 1449–1459.
6851
dx.doi.org/10.1021/ja201297r |J. Am. Chem. Soc. 2011, 133, 6841–6851