Auracarboranes with and without Au-Au interactions: An unusually strong aurophilic interaction

Article abstract of DOI:10.1021/ja953976y

The auracarboranes 1,2-(AuPPh₃)₂-1,2-C₂B₁₀H₁₀ (1) and Ph₃)₂-[2-(1',2'-C₂B₁₀H₁₀)-1,2-C₂B₁₀H₁₀] (2) have been synthesized. Both compounds were characterized by NMR and X-ray crystallography. Compound 2 was found to contain an aurophilic interaction between the gold centers. A variable-temperature NMR investigation indicated that the energy barrier separating the gold-gold bonded state and the nonbonded state is 11 ± 1 kcal/mol. Compound 1 crystallized in the monoclinic space group P2₁/c with a = 18.3380(9) A?, b = 14.1037(6) A?, c = 19.4716(8) A?, β = 112.003(2)°, V = 4669 A?³, and Z = 4. Data were collected using Mo Kα radiation, to a maximum 2θ = 50°, giving 8865 unique reflections, and the structure was solved by heavy atom methods. The final discrepancy index was R = 0.047, R(w) = 0.055 for 3423 independent reflections with 1 > 3σ(I). Compound 2 crystallized in the monoclinic space group P2₁/c with a 14.058(6) A?, b = 18.365(8) A?, c = 20.387(9) A?, β = 109.22(1)°, V = 4970 A?³, and Z = 4. Data were collected using Mo Kα radiation, to a maximum 2θ = 45°, giving 5232 unique reflections, and the structure was solved by heavy atom methods. The final discrepancy index was R = 0.066, R(W) = 0.069 for 2536 independent reflections with I > 3σ(I).

Full text of DOI:10.1021/ja953976y

J. Am. Chem. Soc. 1996, 118, 2679-2685
2679
Auracarboranes with and without Au-Au Interactions: An
Unusually Strong Aurophilic Interaction
David E. Harwell, Mark D. Mortimer, Carolyn B. Knobler, Frank A. L. Anet, and
M. Frederick Hawthorne*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California at
Los Angeles, Los Angeles, California 90095-1569
ReceiVed NoVember 27, 1995X
Abstract: The auracarboranes 1,2-(AuPPh3)2-1,2-C2B10H10 (1) and 1,1′-(AuPPh3)2-[2-(1′,2′-C2B10H10)-1,2-C2B10H10]
(2) have been synthesized. Both compounds were characterized by NMR and X-ray crystallography. Compound 2
was found to contain an aurophilic interaction between the gold centers. A variable-temperature NMR investigation
indicated that the energy barrier separating the gold-gold bonded state and the nonbonded state is 11 ( 1 kcal/mol.
Compound 1 crystallized in the monoclinic space group P21/c with a ) 18.3380(9) Å, b ) 14.1037(6) Å, c )
19.4716(8) Å, â ) 112.003(2)°, V ) 4669 Å3, and Z ) 4. Data were collected using Mo KR radiation, to a maximum
2θ ) 50°, giving 8865 unique reflections, and the structure was solved by heavy atom methods. The final discrepancy
index was R ) 0.047, Rw ) 0.055 for 3423 independent reflections with I > 3σ(I). Compound 2 crystallized in the
monoclinic space group P21/c with a ) 14.058(6) Å, b ) 18.365(8) Å, c ) 20.387(9) Å, â ) 109.22(1)°, V ) 4970
Å3, and Z ) 4. Data were collected using Mo KR radiation, to a maximum 2θ ) 45°, giving 5232 unique reflections,
and the structure was solved by heavy atom methods. The final discrepancy index was R ) 0.066, Rw ) 0.069 for
2536 independent reflections with I > 3σ(I).
Introduction
Scheme 1
Although examples of closo-carboranes bonded to gold by
2c-2e σ-bonds are relatively rare1-3 (as opposed to the more
widely studied nido-carboranes interacting with gold Via cluster
bonding4-8 ), it is known that ortho-carborane derivatives form
unusually strong Au-C σ-bonds. Mitchell and Stone report
that the Au-C bond is unusually stable in the compound 1-Ph3-
PAu-2-Me-1,2-closo-C2B10H10, which when reacted with bro-
mine yields the gold(III) complex 1-Ph3PAu(Br)2-2-Me-1,2-
closo-C2B10H10 (Scheme 1). Since the reaction of gold alkyls
with bromine most often results in cleavage of the Au-C bond,
this unusual stability was attributed to the electron-withdrawing
influence of the carboranyl cage.1 Similar stability is seen in
σ-bonded gold-perfluoroalkyl compounds.9,10
Scheme 2
More recent work by Baukova and by Welch reports the
synthesis and structural characterization of two related com-
pounds, 1-Ph3PAu-1,2-closo-C2B10H113 and 1-CH3OCH2-2-Ph3-
PAu-1,2-closo-C2B10H10.2 In both of these species, the Au-C
bond is observed to be only 2.039(8) Å. In a theoretical and
crystallographic study of model gold-alkyl compounds, Welch
observes that this bond is shorter, and consequently stronger,
than other gold-carbon σ-bonds, thereby supporting Stone’s
X Abstract published in AdVance ACS Abstracts, March 1, 1996.
(1) Mitchell, C. M.; Stone, F. G. A. Chem. Commun. 1970, 1263-1264.
(2) Reid, B. D.; Welch, A. J. J. Organomet. Chem. 1992, 438, 371-
384.
observations. However, it is Welch’s belief that the shortness
of the Au-C bond is due to efficient σ-donor properties of the
carborane and that the carborane does not function as an
electron-withdrawing group. None of the gold derivatives
described above displayed gold-gold bonding interactions.
Aurophilicity is defined by Schmidbaur as “the unprecedented
affinity between gold atoms even with ‘closed-shell’ electronic
configurations and equivalent electrical charges.” Many ex-
amples of this phenomenon, including both intermolecular and
intramolecular complexes, have now been documented. Rep-
resentative examples of intermolecular and intramolecular
complexes are shown in Scheme 2, parts a and b, respec-
tively.11,12 The interatomic distance usually observed in these
aurophilic interactions is on the order of 3.00 ( 0.25 Å,13 and
their strength has been estimated to be in the range of 6 to 8
kcal/mol.14-16 Our intention was to ascertain what effect, if
(3) Baukova, T. V.; Kuz’mina, L. G.; Dvortsova, N. V.; Porai-Koshits,
M. A.; Kravtsov, D. N.; Perevalova, E. G. Metalloorg. Khim. 1989, 2, 1098-
1105.
(4) Howard, J. A. K.; Jeffery, J. C.; Jelliss, P. A.; Sommerfeld, T.; Stone,
F. G. A. J. Chem. Soc., Chem. Commun. 1991, 1664-1666.
(5) Jeffery, J. C.; Jelliss, P. A.; Stone, F. G. A. J. Chem. Soc., Dalton
Trans. 1993, 1073-1081.
(6) Jeffery, J. C.; Jelliss, P. A.; Stone, F. G. A. Inorg. Chem. 1993, 32,
3943-3947.
(7) Jeffery, J. C.; Jelliss, P. A.; Stone, F. G. A. Organometallics 1994,
13, 2651-2661.
(8) Jeffery, J. C.; Jelliss, P. A.; Stone, F. G. A. J. Chem. Soc., Dalton
Trans. 1994, 25-32.
(11) Blumenthal, A.; Beruda, H.; Schmidbaur, H. J. Chem. Soc., Chem.
Commun. 1993, 1005-1006.
(12) Esperas, S. Acta Chem. Scand. 1976, A30, 527.
(13) Schmidbaur, H. Gold Bull. 1990, 23, 11-21.
(9) Vaughan, L. G.; Shepard, W. A. J. Am. Chem. Soc. 1969, 91, 6151-
6156.
(10) Nyholm, R. S.; Royo, P. Chem. Commun. 1969, 421.
0002-7863/96/1518-2679$12.00/0 © 1996 American Chemical Society

2680 J. Am. Chem. Soc., Vol. 118, No. 11, 1996
Harwell et al.
The solution was warmed to room temperature and stirred for 2.5 h to
form 1,2-Li2-1,2-C2B10H10. To this solution was added (triphenylphos-
phine)gold(I) chloride (1.37 g, 2.78 mmol). After stirring at room
temperature for an additional 16 h, solvent was removed in Vacuo and
the residue taken up in CH2Cl2 (50 mL). The solution was filtered to
remove lithium chloride and reduced in volume to ca. 10 mL, and 1,2-
(AuPPh3)2-1,2-C2B10H10 was precipitated by addition of diethyl ether
(0.83 g, 57%). 1H NMR (CDCl3) 7.15-7.39 (m, 30 H, C6H5) ppm;
13C{1H} NMR (CDCl3) 134.2 (C2, Ph), 131.4 (C4, Ph), 129.7 (C1, Ph),
129.0 (C3, Ph) (CAu not observed due to extreme broadening) ppm;
11B{1H} NMR (acetone) -1.8 (2 B), -7.1 (2 B), -9.1 (6 B) ppm;
31P{1H} NMR (CDCl3) 38.5 ppm. High resolution positive ion FAB
mass spectroscopy calculated mass for C38H4010B111B9Au2P2 m/z
1061.2930; observed m/z 1061.2924.
Scheme 3
any, the substitution of a carboranyl cage with its unique
electronic properties would have upon the physical character-
istics of the aurophilic interaction. Consequently, the compound
1,2-(Ph3PAu)2-1,2-closo-C2B10H10 (1, Scheme 3) was synthe-
sized. Because the two gold centers of 1 are in close proximity
to each other on the carboranyl cage surface we predicted that
an aurophilic interaction might be observed. However, the two
gold atoms in 1 are restrained by the scaffolding of the carborane
cage in positions too far apart to allow interaction. With this
in mind, a second system was designed utilizing the free-rotor
property of bis(ortho-carborane), 1,1′-(AuPPh3)2-[2-(1′,2′-
C2B10H10)-1,2-C2B10H10] (2, Scheme 3). Compound 2 is
observed to support an aurophilic interaction and the strength
of this interaction has been estimated using variable temperature
31P NMR and simulations.
Synthesis of 1,1′-(AuPPh3)2-[2-(1′,2′-C2B10H10)-1,2-C2B10H10] (2).
Bis(ortho-carborane), (0.27 g, 0.94 mmol) was dissolved in THF (100
mL) and a 2.5 M solution of n-butyllithium in hexanes (0.75 mL, 1.9
mmol) was added. The solution was stirred for 0.5 h to form 1,1′-
Li2-[2-(1′,2′-C2B10H10)-1,2-C2B10H10]. To this solution was added
(triphenylphosphine)gold(I) chloride (0.933 g, 1.89 mmol). The
reaction flask was then fitted with a condenser and the solution was
heated at reflux for 3 h. After stirring at room temperature for an
additional 12 h, solvent was removed in Vacuo and the residue taken
up in CH2Cl2 (50 mL). The solution was filtered to remove lithium
chloride and reduced in volume to ca. 10 mL, and 1,1′-(Ph3PAu)2-[2-
(1′,2′-C2B10H10)-1,2-C2B10H10] was isolated using column chromatog-
raphy (alumina, 3% CH3OH: 97% CH2Cl2) (0.67 g, 59%). 1H NMR
(CDCl3) 7.36-7.63 (m, 30 H, C6H5), 0.7-3.3 (vbr, 10 H BH) ppm;
13C{1H} NMR (CDCl3) 134.0 (C3, Ph), 131.6 (C1, Ph), 129.3 (C4, Ph),
129.2 (C2, Ph) (CAu not observed due to extreme broadening) ppm;
11B{1H} NMR (acetone) -2.3 (4 B), -7.6 (14 B), -12.6 (2 B) ppm;
31P{1H} NMR (CDCl3) 36.8 ppm. High resolution positive ion FAB
Experimental Section
Materials and Methods. (Triphenylphosphine)gold(I) chloride17,18
and bis(ortho-carborane)19 were prepared according to literature
methods. Standard glovebox, Schlenk, and vacuum line techniques
were employed for all manipulations of air- and moisture-sensitive
compounds. Reaction solvents were reagent grade and were distilled
from appropriate drying agents under nitrogen prior to use. Tetrahy-
drofuran (THF) was distilled from sodium benzophenone ketyl.
Deuterated solvents were obtained from Cambridge Isotope Laboratories
Inc., Knoxville, TN. Consumer Health Research of Los Angeles
provided ortho-carborane. Other reagent grade starting materials were
obtained from Aldrich and used without further purification.
11
mass spectroscopy calculated mass for C40H5010B8 B12Au2P2 m/z
1203.4717; observed m/z 1203.4717.
Crystal Structure Determinations. Computer programs used in
the crystal structure determinations include locally modified versions
of the following programs: CARESS (Broach, Coppens, Becker, and
Blessing), peak profile analysis, Lorentz and polarization correction,
and full-matrix least-squares refinement; ABSCOR, absorption cor-
rection based on psiscans; SHELX76 (Sheldrick), a crystal structure
package; SHELX86 (Sheldrick), a crystal structure package; and
ORTEP (Johnson). X-ray data were collected on a Huber diffractometer
constructed by Professor C. E. Strouse of this department. Data were
collected at 25 °C in the θ-2θ scan mode and were corrected for
Lorentz and polarization effects and for secondary extinction and
absorption. Atoms were located by use of heavy atom methods. All
calculations were performed on a VAX 3100 computer in the J. D.
McCullough X-Ray Crystallography Laboratory. Gold and phosphorus
atoms were refined with anisotropic parameters. Phenyl rings were
refined as rigid C6H5 groups: C-C 1.395 Å, C-H 1.0 Å, angles 120°.
Other hydrogen atoms were included in located positions. H atoms
were assigned isotropic displacement values based approximately on
the value for the attached atom. Scattering factors for H were obtained
from Stewart et al.21 and for other atoms were taken from the
International Tables for X-Ray Crystallography.22
Proton (1H NMR) spectra were obtained with a Bruker AF 200
spectrometer at 200.133 MHz. Carbon (13C NMR) and phosphorus
(31P NMR) spectra were obtained using a Bruker AM 400 spectrometer
at 100.625 and 161.976 MHz, respectively. Boron (11B NMR) spectra
were obtained at 160.46 MHz with a Bruker AM 500 spectrometer.
Chemical shifts for H and 13C NMR spectra were referenced to an
1
internal standard of SiMe4 (0.00 ppm) and measured with respect to
residual protons in deuterated solvents. Chemical shift values for 11
B
spectra were referenced relative to external BF3‚OEt2 (0.00 ppm).
Chemical shift values for 31P spectra were referenced relative to an
internal standard of Ph3PAuCl (33.15 ppm) which was calibrated against
an external standard of P(OMe)3 (141.00 ppm). Temperature calibration
for variable-temperature NMR measurements was obtained by com-
parison of ∆δ values for methanol below 300 K and ∆δ values for
ethylene glycol above 300 K.20 Positive ion FAB mass spectra were
obtained from the UCLA Mass Spectrometry Facility at the University
of California, Los Angeles.
Data Collection and Structure Refinement of 1. A yellow crystal,
obtained from a pentane/THF solution, was mounted on a glass fiber.
Unit cell parameters were determined from a least-squares fit of 89
accurately centered reflections (7.6 < 2θ < 20.1°). These dimensions
and other parameters, including conditions of data collection, are
summarized in Table 1. Three intense reflections (3,-2,0; 3,2,0; 1,0,4)
were monitored every 97 reflections to check stability. Intensities of
these reflections did not decay during the course of the experiment
(79.8 h). Of these 5595 unique reflections measured, 3423 were
considered observed (I > 3σ(I)) and were used in the subsequent
structure analysis. A molecule of THF has been included as five carbon
atoms and no hydrogen atoms were included for this solvent molecule.
The largest peak on a final difference electron density map, near Au-
Synthesis of 1,2-(AuPPh3)2-1,2-C2B10H10 (1). ortho-Carborane
(0.20 g, 1.4 mmol) was dissolved in THF (15 mL) and a 2.0 M solution
of n-butyllithium in hexanes (1.39 mL, 2.78 mmol) was added at 0 °C.
(14) Dziwok, K.; Lachmann, J.; Wilkinson, D. L.; Mu¨ller, G.; Schmid-
baur, H. Chem. Ber. 1990, 123, 423-431.
(15) Schmidbaur, H.; Graf, W.; Mu¨ller, G. Angew. Chem., Int. Ed. Engl.
1988, 27, 417-421.
(16) Schmidbaur, H.; Graf, W.; Mu¨ller, G. HelV. Chim. Acta 1986, 69,
1748-1756.
(17) McAuliffe, C. A.; Panish, R. V.; Randall, P. D. J. Chem. Soc., Dalton
Trans. 1979, 1979, 1730.
(1), was 0.55 e Å-3
.
(18) Bruce, M. I.; Nicholson, B. K.; Shawkataly, O. B. Inorg. Synth.
1989, 26, 324.
(19) Paxson, T. E.; Callahan, K. P.; Hawthorne, M. F. Inorg. Chem. 1973,
12, 708-709.
(20) Variable Temperature Unit B-VT 2000 Manual; USA Bruker
Instrument Inc.
(21) Stewart, R. F.; Davidson, E. R.; Simpson, W. T. J. Chem. Phys.
1965, 42, 3175.
(22) International Tables for X-ray Crystallography; Kynoch Press:
Birmingham, England, 1974; Vol. IV.

Auracarboranes with and without Au-Au Interactions
J. Am. Chem. Soc., Vol. 118, No. 11, 1996 2681
Table 1. Details of Crystallographic Data Collection for
Compounds 1 and 2
parameter
formula
1
2
C38H40B10Au2P2
1061
C40H50B20Au2P2
1202
formula weight
temperature (°C)
crystal size (mm)
normal to faces
appearance
radiation (graphite
monochromator)
wavelength (Å)
space group
a (Å)
25
25
0.08 × 0.2 × 0.4
100, 010, 001
yellow plate
Mo KR
0.05 × 0.05 × 0.3
001, 010, 1h01
pale yellow needle
Mo KR
0.7107
P21/c
0.7107
P21/c
14.056(6)
18.365(8)
20.387(9)
109.22(1)
4970
18.3380(9)
14.1037(6)
19.4716(38)
112.003(2)
4669
b (Å)
c (Å)
â (deg)
V (Å3)
Z
4
1.61
63.6
1.3
1.6
6.
4
1.66
60.3
1.3
1.6
3.
density (g cm-3), calcd
µ (cm-1
)
scan width, below KR1
scan width, above KR2
scan rate (deg min-1
)
no. of unique reflcns
no. of obsd (I > 3σ(I))
reflcns
5595
3423
5232
2536
2θ max (deg)
50
45
data collected
+h,+k,(l
178
+h,+k,(l
222
Figure 1. ORTEP representation of compound 1, 1,2-(AuPPh3)2-1,2-
C2B10H10, showing the numbering scheme. All hydrogens are removed
for clarity. Ellipsoids are drawn at the 0.50 probability level.
no. of parameters refined
R
Rw
GOF
0.047
0.055
1.51
0.066
0.069
1.55
Scheme 4
Data Collection and Structure Refinement of 2. A pale yellow
crystal, obtained from a pentane/CH2Cl2 solution, was mounted on a
fiber. Unit cell parameters were determined from a least-squares fit
of 20 accurately centered reflections (7.3 < 2θ < 16.2°). These
dimensions and other parameters, including conditions of data collec-
tion, are summarized in Table 1. Three intense reflections (0,4,0; 1,0,4;
5,2,-1) were monitored every 97 reflections to check stability.
Intensities of these reflections decayed 6% during the course of the
experiment (128.9 h). Of the 5232 unique reflections measured, 2536
were considered observed (I > 3σ(I)) and were used in the subsequent
structure analysis. A molecule of CH2Cl2 was included as one carbon
1
atom and three chlorine atoms all at
/ occupancy. This solvent
4
molecule must be disordered because it was found to lie near a center
of inversion. The largest peak on a final difference electron density
map, near Au(1), was 0.5 e Å-3
.
Table 2. Selected Bond Parameters for 1
Results and Discussion
bond
distance (Å)
bond
distance (Å)
Synthesis of 1 and 2. Compounds 1 and 2 were prepared
by a method similar to that reported by Mitchell and Stone for
the synthesis of compounds containing a single Au-C σ-bond.1
Reaction of either ortho-carborane or bis(ortho-carborane) with
2 molar equiv of n-butyllithium afforded the respective dilithium
salt, which when reacted with 2 molar equiv of (triphenylphos-
phine)gold(I) chloride gave the diauracarborane species in 57%
(1) and 59% (2) yields (Scheme 4).
Au(1)-P(1)
Au(1)-C(1)
P(1)-C(31P)
C(1)-C(2)
2.270(4)
2.055(14)
1.807(12)
1.71(2)
Au(2)-P(2)
Au(2)-C(2)
P(2)-C(41P)
Au(1)-Au(2)
2.273(5)
2.033(15)
1.794(12)
3.567(1)
atoms
angle (deg)
atoms
angle (deg)
174.2(4)
P(1)-Au(1)-C(1)
178.9(4) P(2)-Au(2)-C(2)
Au(1)-P(1)-C(11P) 112.9(4) C(11P)-P(1)-C(31P) 104.0(5)
B(12)-C(1)-Au(1) 174.0(8) B(9)-C(2)-Au(2)
176.3(7)
Compounds 1 and 2 are yellow crystalline materials which
are readily soluble in THF, acetone, chloroform, and dichlo-
romethane. Both compounds are air-stable; however, 2 was
observed to slowly decompose in solution.
X-ray Structural Analysis of 1. An ORTEP representation
of 1 is presented in Figure 1 and important bond lengths and
angles appear in Table 2. The carboranyl cage of 1 is found to
be a regular icosahedron with gold(triphenylphosphine) groups
bonded perpendicularly to the icosohedral surface at the two
carbon vertices. The interatomic angles B(12)-C(1)-Au(1)
and B(9)-C(2)-Au(2) are 174.0(8)° and 176.3(7)°, respectively.
Bond angles around the gold atoms are also nearly linear with
angles of 178.9(4)° and 174.2(4)° for P(1)-Au(1)-C(1) and
P(2)-Au(2)-C(2), respectively. These values agree well with
those reported for Au(I) centers in similar compounds.11,15,23,24
The lack of distortion around the gold centers coupled with the
Au(1)-Au(2) distance of 3.567(1) Å, well outside the more
usual range of 3.00 ( 0.25 Å,13 precludes an aurophilic
interaction. This is presumably due to the rigidity of the
carborane cage, which prevents the gold(triphenylphosphine)
substituents from closely approaching each other.
(23) Schmidbaur, H.; Graf, W.; Mu¨ller, G. HelV. Chim. Acta 1986, 69,
1748-1756.
(24) Crane, V. S.; Beall, H. Inorg. Chim. Acta 1978, 31, L469-L470.

2682 J. Am. Chem. Soc., Vol. 118, No. 11, 1996
Harwell et al.
C(1′)-Au(1) are 168(1)° and 170(1)°, respectively, showing
significant distortions from linearity. Likewise, the bond angles
P(3)-Au(1)-C(1′) and P(4)-Au(2)-C(1) are 166.2(8)° and
161.4(8)°, respectively, and nonlinear. Significantly, the gold
atoms are displaced away from each other, rather than being
pulled together by their mutual attraction. This implies that
the aurophilic interaction is competing with the mutual steric
repulsion arising from the bulky triphenylphosphine ligands.
Although distortions have been reported for gold(I) centers in
Au-bridging and cluster compounds, the angle of 161.4(8)°
observed at the gold center reported here is the largest distortion
from linearity reported for a gold-phosphine complex of this
coordination type of which we are aware.11,15,16,23,25-27 The
phenyl groups attached to the two phosphorus atoms of 2 are
intermeshed in a fashion resembling the cogs of two gears and
their component atoms are not within van der Waals radii of
each other. The distortions in the structure of 2, concomitant
with the formation of the aurophilic interaction, are in marked
contrast to the undistorted structure of compound 1, which
contains no such interaction. All the relief of steric strain
associated with the formation of the aurophilic interaction is
localized within the gold(triphenylphosphine) groups and their
connections to the bis(ortho-carborane) moiety. The bis(ortho-
carborane) itself shows very little distortion from its expected
geometry. In particular, the angles B(9)-C(2)-C(2′) and
C(2)-C(2′)-B(9′) are 177(2)° and 176(2)°, respectively.
Figure 2. ORTEP representation of compound 2, 1,1′-(AuPPh3)2-[2-
(1′,2′-C2B10H10)-1,2-C2B10H10], showing the numbering scheme. All
hydrogens are removed for clarity. Ellipsoids are drawn at the 0.50
probability level.
Multinuclear NMR Analysis of 1. Only the resonances due
to the phenyl groups associated with the PPh3 ligands of 1 are
observed in the 1H and 13C{1H} NMR spectra. The signals of
the hydrogens attached to the boron atoms of a carborane cage
are invariably quite broad and are seen for compound 1 as a
Table 3. Selected Bond Parameters for 2
bond
distance (Å)
bond
distance (Å)
Au(1)-P(3)
Au(1)-C(1′)
P(3)-C(11P)
C(2)-C(2′)
2.293(8)
2.09(3)
1.82(2)
1.51(4)
Au(2)-P(4)
Au(2)-C(1)
P(4)-C(41P)
Au(1)-Au(2)
2.281(9)
2.13(3)
1.82(2)
3.119(2)
1
slight protuberance in the baseline of the H NMR spectrum
between 0.70 and 3.30 ppm. Signals arising from the carbons
of the carborane cage are not observed, presumably due to line
broadening as a result of their direct attachment to borons in
the cage and coupling to 31P nuclei. The 11B{1H} NMR
spectrum of 1 displays three peaks in a ratio of 2:2:6 with
chemical shifts of -1.8, -7.1, and -9.6 ppm. The pattern and
the shifts observed are consistent with a disubstituted ortho-
carborane cage. The 31P{1H} NMR spectra of 1 exhibits only
a single resonance, at 38.5 ppm.
Multinuclear NMR Analysis of 2. The NMR spectra of 2
are very similar to the corresponding spectra of 1. Again, the
resonance expected from the carbons of the carborane cages
attached to the gold centers is not observed in the 13C{1H} NMR
spectrum. Significantly, the observed resonance for compound
2 in the 31P{1H} NMR spectrum was broad, whereas the
corresponding spectrum of 1 was not, indicating that a dynamic
process was occurring on the NMR time scale. A variable-
temperature NMR study was therefore undertaken with the
objectives of freezing-out the different conformations of 2 and
determining the energy barrier for their interconversion.
angle
(deg)
atoms
angle (deg)
atoms
P(3)-Au(1)-C(1′)
166.2(8) P(4)-Au(2)-C(1)
161.4(8)
Au(1)-P(3)-C(11P) 111.2(8) C(11P)-P(3)-C(31P) 108.2(10)
Au(2)-P(4)-C(41P) 116.9(7) C(41P)-P(4)-C(61P) 105.6(10)
Au(2)-C(1)-C(2)
Au(2)-C(1)-B(6)
C(2′)-C(2)-B(9)
C(2)-C(1)-Au(2)
127(2)
128(2)
177(2)
127(2)
Au(2)-C(1)-B(4)
B(9′)-C(2′)-C(2)
C(2′)-C(1′)-Au(1)
B(12)-C(1)-Au(2)
107(2)
176(2)
130(2)
168(1)
B(12′)-C(1′)-Au(1) 170(1)
dihedral
angle (deg)
atoms
C(1)-C(1′)-Au(1)-Au(2)
C(1′)-C(2′)-C(2)-C(1)
C(2)-C(2′)-Au(1)-Au(2)
C(1′)-midpt. C(2)C(2′)-midpt. Au(1)Au(2)-Au(2)
C(1)-midpt. C(2)C(2′)-midpt. Au(1)Au(2)-Au(1)
C(1′)-midpt. C(1)C(1′)-midpt. Au(1)Au(2)-Au(2)
C(1)-midpt. C(1)C(1′)-midpt. Au(1)Au(2)-Au(1)
31(4)
31(4)
49(2)
144(1)
141(1)
142
142
Variable-Temperature 31P NMR Analysis of 2. The 31P-
{1H} NMR spectrum of compound 2 was measured at roughly
10-deg intervals between 200 and 310 K in CDCl3 (Figure 3a),
using the proton chemical shifts of either methanol or ethylene
glycol as temperature calibrants. At temperatures of 280 K and
above, a major peak is observed which becomes sharper as the
temperature is increased. A small subsidiary peak at 37 ppm
arises from an unidentified phosphorus-containing impurity. The
spectrum taken at 278 K is very close to the coalescence point.
X-ray Structural Analysis of 2. An ORTEP representation
of 2 is presented in Figure 2 and important bond lengths and
angles appear in Table 3. In the solid state the gold-
(triphenylphosphine) moieties are both on the same side of the
molecule, with an Au(1)-Au(2) distance of 3.119 Å. This is
within the range of distances reported for other molecules with
an aurophilic interaction,15,16,23,25,26 and much shorter than that
seen in compound 1. The structure as shown in Figure 2 is
chiral, but the space group is P21/c; therefore, the crystal
contains a 1:1 mixture of both enantiomers. A view down the
long axis of the bis(ortho-carborane) group reveals that the gold
atoms of 2 are not eclipsed, but instead are rotated 49(2)° away
from each other as defined by the dihedral angle Au(1)-C(2)-
C(2′)-Au(2). The angles B(12)-C(1)-Au(2) and B(12′)-
(25) Schmidbaur, H.; Zeller, E.; Weidenhiller, G.; Steigelmann, O.;
Beruda, H. Inorg. Chem. 1992, 31, 2370-2376.
(26) Schmidbaur, H.; Brachtha¨user, B.; Steigelmann, O.; Beruda, H.
Chem. Ber. 1992, 125, 2705-2710.
(27) Scherbaum, F.; Huber, B.; Mu¨ller, G.; Schmidbaur, H. Angew.
Chem., Int. Ed. Engl. 1988, 27, 1542-1544.

Auracarboranes with and without Au-Au Interactions
J. Am. Chem. Soc., Vol. 118, No. 11, 1996 2683
Figure 4. Proposed potential energy diagram for the interconversion
of syn-2 and anti-2. For an explanation see text.
The Nature of the Energy Barrier to Interconversion.
Figure 4 is a diagram depicting the potential energy of 2 with
respect to rotation around the C(2)-C(2′) axis (bold line). In
solution 2 exists in one of two conformations, anti-2 or syn-2
(Scheme 5). Both of the mirror image structures of syn-2 are
possible, and the crystal structure of 2 is observed to consist of
a 1:1 mixture of the two enantiomers. The crystal structure
represents the lowest energy (syn) conformation of 2 in the solid
state, and corresponds to the two deep wells on each side of
the energy diagram (one for each of the enantiomers). The
center well of the energy diagram is broad and corresponds to
the several different configurations of similar energy possible
for the anti-conformation of 2, which are related to one another
by rotation around the C(2)-C(2′) axis. As the Ph3PAu groups
are brought into closer proximity to each other by rotation about
the C(2)-C(2′) axis the energy is expected to rise due to steric
repulsions until the gold atoms become sufficiently close for a
bonding interaction to occur. Formation of the aurophilic
interaction would then result in a lowering of the energy which
is depicted by the narrow wells on each side of Figure 4.
Figure 3. (a) Observed variable-temperature 31P{1H} NMR spectra
of 2 in deuteriochloroform. (b) Simulated variable-temperature 31P-
{1H} NMR spectra of 2.
Scheme 5
Large energy barriers are depicted on the extreme left and
right sides of Figure 4 to indicate that there is no interconversion
between enantiomers by the (triphenylphosphine)gold groups
rotating past each other. The X-ray structure of 2 suggests that
such a rotation is highly unlikely. In fact, the manner in which
the gold atoms are bent away from each other by the steric bulk
of the attached phosphine groups implies that the molecule is
already under considerable steric strain. Narrow wells are
indicated for the syn-conformations of 2 because the X-ray
structure of 2 indicates that the aurophilic interaction is at a
maximum when the dihedral angle Au(1)-C(2)-C(2′)-Au(2)
is 49°. Angles less than 49° would result in excessive steric
strain which could not be compensated by the increase in gold-
gold orbital overlap. Angles much greater than 49° would result
in rotating the gold atoms too far apart to maintain the aurophilic
interaction.
At temperatures below 280 K, two resonances are observed.
Only one resonance can be assigned to the structure observed
in the X-ray analysis of 2; however, the second resonance must
correspond to a closely related structure and the two structures
must be interconverting more or less slowly on the NMR time
scale. We ascribe these resonances to the freezing-out of the
gold-gold bonded (syn) and the non-interactive (anti) confor-
mations of 2 as depicted in Scheme 5. The chemical shift of
one peak (A) is strongly temperature dependent, although the
temperature dependence of peak B is minimal in the temperature
range studied. The relative populations of the two peaks vary
greatly. At 200 K, peak B (90%) is extensively more populated
than peak A (10%), whereas above room temperature peak A
(55 %) is more populated than peak B (45%), as shown by the
dynamic NMR study presented below. These observations
indicate that peak B corresponds to the enthalpically favored
conformation whereas peak A corresponds to the entropically
favored conformation. Therefore, we assigned anti-2, the
species with the larger entropy (more rotational states), to peak
A, and syn-2, the rigid species observed in the crystal structure,
to peak B.
We have calculated a value for the barrier between anti-2
and syn-2 (Vide infra); however, the energy of the aurophilic
interaction does not necessarily correspond to this value. In
reference to Figure 4, the energy of the aurophilic interaction
is measured vertically from the bottom of either of the narrow
wells representing syn-2 up to the surface of the potential energy
plot for 2 in the absence of an aurophilic interaction. (The
transition state at the top of the energy barrier will have a certain
amount of gold-gold bonding character.) Two possibilities
exist for the potential energy of 2 in the absence of an aurophilic
interaction. At one extreme, the potential energy would increase
rapidly on each side of a broad minimum asthe two Ph3PAu

2684 J. Am. Chem. Soc., Vol. 118, No. 11, 1996
Harwell et al.
Table 5. Ground State and Activation Parameters for 2
Table 4. Calculated Parameters for Simulated 31P NMR Spectra of
2
∆H°AfB
∆S°AfB
∆Hq
∆Sq
∆Hq
∆Sq
AfB
AfB
BfA
BfA
temp ∆G°AfB
(K) (kcal/mol) (kcal/mol)
∆Gq
kAfB
(s-1
δA
δB
∆ν
(kcal/mol)
(eu)
(kcal/mol)
(eu)
(kcal/mol)
(eu)
AfB
)
(ppm) (ppm) (Hz) PA
3.0 ( 0.5
10 ( 2
11 ( 1
6 ( 3
8 ( 1
-8 ( 3
200
213
221
234
244
255
267
278
289
300
310
1.01
0.88
0.80
0.68
0.58
0.47
0.35
0.25
0.13
0.03
-0.07
11.0
11.2
11.3
11.5
11.7
11.8
12.0
12.2
12.4
12.6
12.7
4.0
36.9 36.5
69 0.07
77 0.11
81 0.14
87 0.19
92 0.23
97 0.28
1.4 × 101 37.0 36.5
3.0 × 101 37.0 36.5
8.5 × 101 37.0 36.5
1.8 × 102 37.1 36.5
3.7 × 102 37.1 36.5
respect to temperature (eq 3a), a least-squares analysis was used
to obtain a best-fit straight line, and values for PA and PB were
optimized based on the best fit. The improved values for PA
and PB were used to obtain new values for τ. Values for the
kinetic parameter ∆Gq were calculated from the newly obtained
values for τ utilizing the Eyring equation (eq 4)29,30 where k
7.8 × 102 37.1 36.5 103 0.34
1.5 × 103 37.2 36.5 108 0.39
2.6 × 103 37.2 36.5 113 0.44
4.5 × 103 37.2 36.5 119 0.49
7.1 × 103 37.3 36.5 123 0.53
∆G° ) -RT ln K1
∆G° ) ∆H° - T∆S°
∆Gq ) ∆Hq - T∆Sq
(2)
(3a)
(3b)
groups are rotated into one another. This extreme is depicted
by the fine lines across the top of the gray areas in Figure 4.
The other extreme, depicted by the fine lines at the bottom of
the gray areas in Figure 4, arises from possible energy-lowering
interactions between the phosphine groups upon close contact.
These might include, for example, a local energy minimum as
the phenyl goups on the phosphorus atoms intermesh, or
attractive edge face or π-stacking interactions. The actual
potential energy surface for 2 in the absence of the aurophilic
interaction most likely lies somewhere between these two
extremes within the gray areas, and thus the energy of the
interaction may be somewhat stronger or weaker than the height
of the barrier from syn-2 to anti-2. The calculation of this
barrier will be discussed in the next section. It should be noted
that the ambiguity in the assignment of a value for the energy
of the aurophilic interaction should not be confused with the
error in the calculation of the barrier to interconversion between
anti-2 and syn-2. The calculation of this error will also be
discussed later in the paper.
kh
kT
∆Gq ) -RT ln
(4)
(
)
represents the Boltzmann constant, and plotted against temper-
ature. A least-squares analysis was again performed to obtain
a straight line. The equations for both plots (∆G° vs T and
∆Gq vs T) were entered into a spreadsheet and the values of
their gradients (∆S° and ∆Sq, respectively) and y-intercepts (∆H°
and ∆Hq, respectively) were varied in an iterative process until
the best fit for all data points was obtained (Figure 3b, Table
4).
Evaluation of Error in NMR Calculations. Values for ∆G
are related to the values of ∆H and ∆S through the Gibbs
function (eqs 3a and 3b). Insertion of the constant TM, which
represents the mean temperature of the experimental measure-
ments, results in eq 5. The constant, ∆GM, can be defined as
the value for ∆G at TM, and T - TM can be rewritten as ∆T.
Therefore, eq 5 can be simplified to eq 6. Insertion of the
variable ∆∆S, which is defined as the change in ∆S, yields eq
7.
31P NMR Simulations. Simulated spectra were calculated
using the equations of Rogers and Woodbrey,28 and the final
values of the important parameters are given in Table 4. PA is
defined as the fractional population of anti-2, corresponding to
the area under peak A in the spectra, and PB ()1 - PA) is
defined as the fractional population of syn-2, corresponding to
the area under peak B in the spectra (Figure 3a). The rate
constants kAfB and kBfA are defined as the rate of conversion
of anti-2 to syn-2 and the reverse, respectively. The time τ is
defined by eq 1, and ∆ν is the difference in the chemical shifts
of peaks A and B expressed in hertz.
∆G ) ∆H - TM∆S - (T - TM)∆S
∆G ) ∆GM - ∆T∆S
(5)
(6)
(7)
∆G ) ∆GM - ∆T(∆S + ∆∆S)
Values for ∆∆S were varied incrementally and new values for
∆G° and ∆Gq were calculated. Values for τ and the ratio of
PA to PB were calculated from ∆G° and ∆Gq, respectively.
These newly obtained values were used to calculate a set of
simulated spectra. The value of ∆∆S was changed until the
simulated spectra were judged to no longer agree with the
observed spectra. The minimum values of ∆∆S° and ∆∆Sq
where the calculated spectra no longer fit the observed spectra
were taken as the error in ∆S° and ∆Sq, respectively. The error
in ∆H° and ∆Hq can be calculated by the relationship depicted
in eq 8 where ∆∆H is defined as the change in ∆H (Table 5).
-1
τ ) (kAfB + kBfA
)
(1)
The spectral line shape depends on PA, τ, ∆ν, and the T2
values of A and B in the absence of exchange. Values of T2,
the transverse (spin-spin) relaxation time, for anti-2 and syn-2
were estimated by measuring the width at half-height of the
peaks A and B, respectively, in the slow-exchange region (213
K) and these values were held constant in all calculations. Initial
values for PA and PB were determined by integrating peaks A
and B at each temperature and then systematically subtracting
the contribution due to the small impurity at 37 ppm. The
contribution due to the impurity was determined by integration
in the fast-exchange region where the combined peak for A and
B was well differentiated from the impurity peak. Calculated
spectra were fitted to the experimental spectra by changing the
values of τ and the ratio of PA to PB. Values for the
thermodynamic parameter ∆G° were then calculated utilizing
eq 2 at each temperature and plotted against temperature, where
K1 ) PA/PB. Utilizing the linear relationship of ∆G° with
∆∆H ) T∆∆S
(8)
Table 5 shows the final calculated values for the entropies
and enthalpies of transition, and for the process as a whole.
The syn-conformation of 2 was found to be 3 ( 0.5 kcal/mol
more stable than the anti-conformation (Figure 4), with the
added stability of syn-2 attributed to the aurophilic interaction.
The height of the energy barrier going from the anti-conforma-
(29) Glasstone, S.; Laidler, K. J.; Eyring, H. The Theory of Rate
Processes; McGraw-Hill: New York, 1941.
(30) Eyring, H. Chem. ReV. 1935, 17, 65.
(28) Rogers, M. T.; Woodbrey, J. C. J. Phys. Chem. 1962, 66, 540-
546.

Auracarboranes with and without Au-Au Interactions
J. Am. Chem. Soc., Vol. 118, No. 11, 1996 2685
tion to the transition state was calculated as 8 ( 1 kcal/mol,
while the height going from the syn-conformation to the
transition state was calculated as 11 ( 1 kcal/mol. In physical
terms, the anti to transition state barrier represents the significant
structural distortions discussed in the X-ray structural analysis.
Potential energy is bound up in the deviation of the angles
around gold and the carbons attached to gold from their optimal
hybridization in the absence of the aurophilic interaction. The
syn to transition state barrier represents the energy required to
break the aurophilic interaction, plus any energy which may be
associated with disentangling the phosphine groups. However,
this is anticipated to be a minor term and we feel justified in
using the figure of 11 ( 1 kcal/mol as a good estimate of the
strength of the aurophilic interaction.
The value of 11 ( 1 kcal/mol for the aurophilic interaction
is approximately one and a half times that previously estimated
in other systems at between 6 and 8 kcal/mol.14,15,23 We believe
that the explanation for this surprisingly strong interaction is to
be found in the powerful electron-withdrawing nature of the
ortho-carboranyl groups, which make the gold atoms more
electron deficient than in other gold(I) compounds. This is
compensated for by maximizing bonding interactions between
the gold atoms. In this we agree with Stone’s interpretation of
the nature of the carboranyl group as an electron-withdrawing
substituent and disagree with Welch’s alternative view that it
acts as an electron donor. A great deal of independent study
over the past 30 years has led to the characterization of a
C-ortho-carboranyl group as powerfully electron-withdrawing
(as opposed to B-ortho-carboranyl groups which are more or
less electron-donating, depending on the position of B-substitu-
tion). Arguments are based on the high acid strength of
carboranyl carboxylic acids,31-34 Hammett σ-constant determi-
nations35 and Taft analysis of 19F NMR shifts32,36,37 of suitable
derivatives, the ease of metalation of the carboranyl C-H
group,38,39 the relative lack of reactivity of C-halomethyl
carboranes toward nucleophiles,1,40,41 and a host of other
observations regarding the reactivity of these species.42-54
In addition to the electron-withdrawing ability of the carbo-
ranyl cage, we believe that the shortening of the Au-C bond
length observed by Stone, Baukova, and Welch is due to the
hybridization of the atomic orbitals of the carbon in the
carborane cage. This hybridization has more s-character than
traditional alkyl, sp3, substituents. Consequently bonds involv-
ing C-vertices of carborane should be shorter and, in fact, C-C
bonds between the carbon of a carboranyl-cage and an alkyl
substituent are observed to be shorter than C-C bonds in
alkanes.55
Conclusions
We have succeeded in synthesizing both a gold-carborane
compound containing an aurophilic interaction and an electroni-
cally similar model compound having no aurophilic interaction
for comparison. Detailed NMR variable-temperature experi-
ments coupled with crystallographic analyses have allowed us
to propose an explanation of the fluxional behavior of 2 based
on a dynamic equilibrium between the syn and anti forms of
the compound, and to estimate the strength of the aurophilic
interaction based on the calculated height of the barrier to
interconversion. The strength of this interaction, at 11 kcal/
mol, is considerably greater than that previously estimated by
Schmidbaur et al. using similar techniques. We propose that
the electron-deficiency of the gold atoms caused by the strong
negative inductive effect of the carborane cages is responsible
for the stronger interaction, in contradiction to other analyses
of the nature of gold carboranyl compounds.
Acknowledgment. We thank the National Science Founda-
tion for their financial support of this work under Grant No.
NSF-CHE-93-14037.
Supporting Information Available: Tables of bond dis-
tances and angles and position and displacement parameters (13
pages); listing of observed and calculated structure factors for
1 and 2 (35 pages). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, can be ordered from the ACS, and can
be downloaded from the Internet; see any current masthead page
for ordering information and Internet access instructions.
JA953976Y
(31) Zakharkin, L. I. IzV. Akad. Nauk SSSR, Ser. Khim. 1965, 1114-
1116.
(43) Zakharkin, L. I.; Stanko, V. I.; Brattsev, V. A.; Chapovskii, Y. A.;
Klimova, A. I.; Okhlobystin, O. Y.; Ponomarenko, A. A. Dokl. Akad. Nauk
SSSR 1964, 155, 1119-1122.
(32) Zakharkin, L. I.; Kalinin, V. N.; Snyakin, A. P. Zh. Obsch. Khim.
1970, 40, 2424-2431.
(33) Zakharkin, L. I.; Chapovskii, Y. A. IzV. Akad. Nauk SSSR, Ser. Khim.
(44) Zakharkin, L. I. Dokl. Akad. Nauk SSSR 1964, 157, 1149-1152.
(45) Clegg, W.; Coult, R.; Fox, M. A.; Gill, W. R.; MacBride, J. A. H.;
Wade, K. Polyhedron 1993, 12, 2711-2717.
1964, 772.
(34) Potapova, T. V.; Svitsyn, R. A.; Zhigach, A. F.; Laptev, V. T.;
Persianova, I. V.; Sorokin, P. Z. Russ. J. Inorg. Chem. 1965, 10, 1133-
1135.
(46) Yang, X.; Knobler, C. B.; Hawthorne, M. F. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1507.
(35) Hawthorne, M. F.; Berry, T. E.; Wegner, P. A. J. Am. Chem. Soc.
1965, 87, 4746-4750.
(36) Adler, R. G.; Hawthorne, M. F. J. Am. Chem. Soc. 1970, 92, 6174-
6182.
(47) Yang, X.; Knobler, C. B.; Zheng, Z.; Hawthorne, M. F. J. Am. Chem.
Soc. 1994, 116, 7142.
(48) Grimes, R. N. Carboranes, 1st ed.; Academic Press: New York,
1970; p 272.
(37) Piepgrass, K. W.; Stockman, K. E.; Sabat, M.; Grimes, R. N.
Organometallics 1992, 11, 2404-2413.
(49) Papetti, S.; Heying, T. L. Inorg. Chem. 1963, 2, 1105-1107.
(50) Papetti, S.; Schaehher, B. B.; Troscianiec, H. J.; Heying, T. L. Inorg.
Chem. 1964, 3, 1444-1447.
(38) Zakharkin, L. I.; Stanko, V. I.; Klimova, A. I.; Chapovskii, Y. A.
IzV. Akad. Nauk SSSR, Ser. Khim. 1963, 2236-2237.
(39) Aleksandrov, A. Y.; Bregadze, V. I.; Gol’danskii, V. I.; Zakharkin,
L. I.; Okhlobystin, O. Y.; Khrapov, V. V. Dokl. Akad. Nauk 1965, 165,
593-596.
(51) Grafstein, D.; Bobinski, J.; Dvorak, J.; Paustian, J. E.; Smith, H.
F.; Karlan, S.; Vogel, C.; Fein, M. M. Inorg. Chem. 1963, 2, 1125-1126.
(52) Grafstein, D.; Bobinski, J.; Dvorak, J.; Smith, H.; Schwatrz, N.;
Cohen, M. S.; Fein, M. M. Inorg. Chem. 1963, 2, 1120-1125.
(53) Heying, T. L.; Ager, J. W.; Clark, S. L.; Alexander, R. P.; Papetti,
S.; Reid, J. A.; Trotz, S. I. Inorg. Chem. 1963, 2, 1097-1105.
(54) Harmon, K. M.; Harmon, A. B.; Thompson, B. C. J. Am. Chem.
Soc. 1967, 89, 5309-5311.
(40) Zakharkin, L. I.; Kalinin, V. N. IzV. Akad. Nauk. SSSR, Ser. Khim.
1965, 1311.
(41) Zakharkin, L. I.; Podvisotskaya, L. S. IzV. Akad. Nauk SSSR, Ser.
Khim. 1965, 1464-1466.
(42) Zakharkin, L. I.; Stanko, V. I.; Brattsev, V. A.; Chapovskii, Y. A.;
Okhlobystin, O. Y. IzV. Akad. Nauk SSSR, Ser. Khim. 1963, 2238-2239.
(55) Kiveka¨s, R.; Sillanpa¨a¨, R.; Teixidor, F.; Vin˜as, C.; Nun˜ez, R.; Mar
Abad, M. Acta Crystallogr. 1995, C51, 1864-1868.