Synthesis and molecular structure of gold triarylcorroles

Article abstract of DOI:10.1021/ic202023r

A number of third-row transition-metal corroles have remained elusive as synthetic targets until now, notably osmium, platinum, and gold corroles. Against this backdrop, we present a simple and general synthesis of β-unsubstituted gold(III) triarylcorroles and the first X-ray crystal structure of such a complex. Comparison with analogous copper and silver corrole structures, supplemented by extensive scalar-relativistic, dispersion-corrected density functional theory calculations, suggests that "inherent saddling" may occur for of all coinage metal corroles. The degree of saddling, however, varies considerably among the three metals, decreasing conspicuously along the series Cu > Ag > Au. The structural differences reflect significant differences in metal-corrole bonding, which are also reflected in the electrochemistry and electronic absorption spectra of the complexes. From Cu to Au, the electronic structure changes from noninnocent metal(II)-corrole(?2-) to relatively innocent metal(III)-corrole(3-).

Full text of DOI:10.1021/ic202023r

Article
pubs.acs.org/IC
Synthesis and Molecular Structure of Gold Triarylcorroles
Kolle E. Thomas,† Abraham B. Alemayehu,† Jeanet Conradie,†,‡ Christine Beavers,§ and Abhik Ghosh*,†
†Department of Chemistry, University of Tromsø, 9037 Tromsø, Norway
‡Department of Chemistry, University of the Free State, 9300 Bloemfontein, Republic of South Africa
§Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California 94720-8229, United States
S
* Supporting Information
ABSTRACT: A number of third-row transition-metal corroles have remained elusive as synthetic targets until now, notably
osmium, platinum, and gold corroles. Against this backdrop, we present a simple and general synthesis of β-unsubstituted
gold(III) triarylcorroles and the first X-ray crystal structure of such a complex. Comparison with analogous copper and silver
corrole structures, supplemented by extensive scalar-relativistic, dispersion-corrected density functional theory calculations, suggests
that “inherent saddling” may occur for of all coinage metal corroles. The degree of saddling, however, varies considerably among the
three metals, decreasing conspicuously along the series Cu > Ag > Au. The structural differences reflect significant differences in
metal−corrole bonding, which are also reflected in the electrochemistry and electronic absorption spectra of the complexes. From Cu
to Au, the electronic structure changes from noninnocent metal(II)−corrole(•2−) to relatively innocent metal(III)−corrole(3−).
coupled with density functional theory (DFT) calculations,6 have
permitted a detailed comparative study of a full triad of coinage
metal corroles. Despite superficial resemblance, the results
indicate major variations in electronic character among the
three coinage metals, when complexed with corroles.
INTRODUCTION
Despite the brisk pace of corrole research over the past decade,
■
corroles continue to serve as platforms for novel coordination
chemistry.1 A number of third-row transition-metal corroles
remain to be synthesized. Thus, although rhenium2 and iridium3
corroles have been synthesized, osmium, platinum, and gold
corroles remain largely unexplored. In our laboratory, we have
focused particularly on the problem of gold insertion into
corroles.4 Although β-octabromo-meso-triarylcorroles have
proven amenable to gold insertion,4,5 gold complexes of
β-unsubstituted meso-triarylcorroles have remained elusive until
now; thus, reagents such as chloroauric acid, H[AuCl4]·H2O,
were found to lead to uncontrolled β-chlorination and little
actual gold insertion. Obliged to search for new routes to gold
corroles, we decided to examine gold salts that lacked oxida-
tively transferable ligands such as chloride, a strategy that
proved effective. Using gold(III) acetate, we have succeeded
in synthesizing a family of β-unsubstituted gold(III) meso-
triarylcorroles with systematically varying meso-aryl groups.
Furthermore, single-crystal X-ray structure determinations
could be obtained for the silver and gold complexes of meso-
tris(p-fluorophenyl)corrole, [T(p-F)PC], permitting a detailed
comparison with existing structural studies on copper corroles.
The structural, electrochemical, and electronic absorption data,
RESULTS AND DISCUSSION
■
a. Synthesis and Proof of Composition. Gold(III)
triarylcorroles were obtained by interaction of the free ligands
and a moderate excess of gold(III) acetate in pyridine. The
yields of about 30% were lower than those obtained for copper
and silver corroles, which are typically >70%.7 The complexes
yielded mass and 1H NMR spectra consistent with their
expected structures (Figure 1). Temperature-dependent 1H
NMR spectra revealed no evidence of paramagnetic states, as
are found for copper triarylcorroles.8 Figure 1 depicts the room-
1
temperature H NMR spectrum of gold(III) meso-triphenyl-
corrole, Au[TPC], as a representative example.
b. Molecular Structures. X-ray-quality crystals proved
difficult to obtain for the microcrystalline gold triarylcorroles.
Received: September 15, 2011
Published: November 23, 2011
© 2011 American Chemical Society
12844
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
1
Figure 1. Room-temperature H NMR spectrum of Au[TPC]. See the Experimental Section for assignments.
One of the complexes, Au[T(p-F)PC], however, lent itself to
crystallographic analysis (Figure 2), providing the first such
characterization for a simple β-unsubstituted gold corrole. The
analogous silver structure could also be solved, and both these
structures may be compared with a number of copper
triarylcorrole crystal structures in the literature.9 Key structural
details, both experimental- and DFT-based, are presented in
Table 1. A comparison of the gold and silver10 structures shows
that the Ag/Au−N distances are essentially identical (1.95−
1.96 Å), but they are about 0.5−0.6 Å longer than the Cu−N
distances.11 Like copper corroles,12,13 the silver and gold
corrole structures reported herein are also saddled. Although
common for metalloporphyrins, saddling is rare for metal-
locorroles and, as far as is known, unique to coinage metal
The degree of saddling varies considerably among corrole
complexes formed by the three coinage metals. As shown in
Table 1 and Figure 2, whereas β-unsubstituted copper tri-
arylcorroles are very significantly saddled,9a,12 with the saddling
dihedral χ3 ∼ 45−55°, Ag[T(p-F)PC] is less so (χ3 ∼ 38.8°) and
Au[T(p-F)PC] is even more flattened (χ3 ∼ 24.5°). DFT
(BP86-D22,23/STO-TZP)24 calculations of saddling potentials
(Figure 3) qualitatively reproduce these observations,
confirming that gold corroles are much less saddled than
their copper and silver analogues. Consistent with this picture,
the crystal structure of Au[Br8TPFPC] reported by Gross et al. also
reveals an essentially planar corrole.5
An examination of the Kohn−Sham molecular orbitals (MOs;
Figure 4) affords further insight into the comparative stereo-
2
2
chemistry of the three coinage metals. The formally empty dx −y
corroles. In the case of metallocorroles, the saddling is believed
orbital of the “Cu(III)” center has a low orbital energy and avidly
absorbs electron density from the corrole “a2u” highest occupied
molecular orbital (HOMO); copper corroles thus have
significant copper(II)−corrole(•2−) character.12,25 Observe
from Figure 4a that both the HOMO and lowest unoccupied
14
2
2
to result from a metal(dx −y )−corrole(“a2u”) orbital inter-
action, which is symmetry-forbidden for a planar macrocycle
geometry. Before proceeding further, however, it is useful to
̀
contextualize the issue of saddling vis-a-vis metallocorroles.
For metalloporphyrins, saddling most commonly occurs as a
result of peripheral overcrowding;15 it is thus particularly com-
mon for dodecasubstituted porphyrins and is essentially un-
2
2
molecular orbital (LUMO) of Cu[TPC] involve dx −y −“a2u”
2
2
interaction; they differ only in the relative phases of the dx −y
2
2
and “a2u” components. The low energy of the Cu dx −y orbital
translates into a low HOMO−LUMO gap for copper
triarylcorroles, consistent with an increasing population of
paramagnetic states in the 1H NMR spectra above room
temperature.8,9a In the gold case, the high energy of the Au
2
2
known for sterically unhindered porphyrins. The metal(dx −y )−
porphyrin(a2u) orbital interaction is well-established in such com-
plexes. It explains, for instance, why saddled copper porphyrins
are more easily oxidized relative to analogous saddled nickel
porphyrins.16 The same orbital interaction also accounts for the
diamagnetism of saddled copper porphyrin π-cation-radical
derivatives.17 However, in none of these cases is the saddling
a result of this orbital interaction. For metalloporphyrins, the
2
2
5dx −y orbital discourages significant mixing with the corrole
a2u HOMO; thus, as shown in Figure 4b, the HOMO of
2
2
Au[TPC] has little Au 5dx −y character. It follows that the
HOMO−LUMO gap of Au[TPC] is much higher than that of
Cu[TPC] and is comparable to that of a non-transition-metal
porphyrinoid such as a typical zinc porphyrin; this may be seen
from the energy level diagram shown in Figure 4c. In fact, the
2
2
dx −y −a2u interaction is undoubtedly strengthened by saddling;
however, without exception, the saddling arises primarily as a
result of peripheral steric crowding. Metallocorroles are dif-
ferent in this regard.
2
2
Au 5dx −y orbital is so high in energy that it does not even
correspond to the LUMO; for Au[TPC], it is actually the
LUMO+2.
2
2
First, the dx −y −“a2u” interaction in coinage metal corroles
results in saddling even in the absence of peripheral steric
crowding. Second, even strong steric hindrance on the macro-
cycle periphery generally does not result in saddling in the
Like the majority of porphyrin and corrole crystal structures,
coinage metal corroles show stacked macrocycles. For C2/c, the
stacking is along the c axis. For the silver and gold structures
reported herein, nearest-neighbor metal−metal distances are
3.8768(5) and 4.2294(5) Å for silver and 3.9296(4) and
4.3380(4) Å for gold. The corroles stack as inversion-related
pairs, with a 2-fold axis and glide plane between two sets of
pairs. The inversion-related pairs have the longer metal−metal
distance, and the 2-fold/c-glide generates the shorter distance.
2
2
absence of a suitable dx −y −“a2u” interaction. This is why
saddling is generally not observed for noncoinage metal
corroles. For example, whereas copper β-octabromo-meso-
triarylcorroles are strongly saddled,18,19 the analogous group 9
(cobalt, rhodium, and iridium) corroles are relatively
planar.3a,20 For metallocorroles, saddling thus directly reflects
2
2
the mechanical power of the dx −y −“a2u” interaction and
provides a unique window into ligand noninnocence.21
12845
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
Figure 2. (a) Thermal ellipsoid (50%) diagram of Au[T(p-F)PC]. The disordered solvent was omitted for clarity. (b) Side-on view of copper (red),
silver (gray), and gold (gold) corrole cores overlaid on one another. (c) Side view of corrole stacking.
The Au···Au distances are too long to be considered aurophilic
interactions, which occur commonly for gold(I) compounds.26
c. Electrochemistry. Table 2 presents redox potentials for
the various M[T(p-X)PC] (M = Cu, Ag, Au; X = CF3, F, H,
Me, OMe) complexes studied. Figure 5 presents cyclic
voltammograms of the three M[T(p-F)PC] (M = Cu, Ag,
Au) complexes. The results provide striking support for the
MO picture presented above. Thus, whereas the first oxidation
potentials vary little across the coinage metal triad, the
reduction potentials vary dramatically, becoming increasingly
negative down the triad. Stated differently, the electrochemical
“HOMO−LUMO gaps” (defined as the algebraic difference
between the first oxidation and reduction potentials) widen
dramatically down the triad, in excellent agreement with our
12846
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
a
Table 1. Selected Distances (Å) and Dihedral Angles (deg) for M[T(p-X−P)C] (M = Cu, Ag, Au) Crystal Structures
b
b
b
complex
M−N1
M−N2
χ1
χ2
χ3
source
c
c
Cu[TPC]
1.891 (1.889)
1.943 (1.969)
1.939 (1.960)
1.891 (1.901)
1.955 (1.982)
1.955 (1.978)
27.5 (25.3)
19.9 (20.8)
13.8 (7.3)
53.4 (47.7)
38.8 (39.0)
48.7 (42.7)
38.8 (39.6)
24.5 (21.3)
ref 9
Ag[T(p-F−P)C]
Au[T(p-F−P)C]
this work
this work
23.1 (18.3)
a
b
c
BP86-D values are shown in parentheses. Average of two values for each experimental structure. The metrical parameters reported in ref 12 are
similar.
DFT calculations. A detailed spectroelectrochemical study has
established that one-electron reduction of copper corroles is
significantly metal-centered.27 This is consistent with the nature
of the LUMO shown in Figure 4a. According to our DFT
calculations, Ag[TPC] too should undergo a similar, signifi-
cantly metal-centered reduction. By contrast, our calculations
indicate that, in the case of gold triarylcorroles, the reduction is
exclusively corrole-centered.28,29 Detailed electron paramag-
netic resonance and Raman spectroelectrochemical studies are
in progress to confirm these predictions, but the general agree-
ment between the electrochemical and DFT-derived HOMO−
LUMO gaps leaves little doubt about the qualitative correctness
of our MO picture.
CONCLUDING REMARKS
■
In summary, we have presented a synthetic route to
β-unsubstituted gold triarylcorroles and the first crystal
structure of such a complex. Macrocycle saddling appears to
be a common, structural feature of coinage metal corroles.7−9,17
However, substantial variations in the degree of saddling,
HOMO−LUMO gaps, and substituent effects on electronic
absorption spectra reflect significant differences in metal−
ligand bonding among the three coinage metals.17 The
experimental data, supported by DFT calculations, indicate
that, as one moves down the triad, the electronic structure
changes from a relatively noninnocent metal(II)−corrole(•2−)
to an innocent metal(III)−corrole(3−) description.
d. Electronic Absorption Spectra. Many metallotri-
arylcorrole families such as Cu, MnCl, and FeCl corroles
exhibit strong meso-substituent effects on the Soret maxima
(see Figure 6, inset, for the copper case). These have been
analyzed in some detail for copper triarylcorroles and ascribed
to so-called hyper character, i.e., phenyl-to-metal charge-transfer
(CT) character mixing into the Soret transitions.14 We have
previously reported that such substituent effects are not
observed for silver triarylcorroles,14 and as shown in Figure 6,
they are not observed for gold triarylcorroles either. The high
HOMO−LUMO gaps in the silver and gold cases provide a
natural explanation for the lack of CT character in their Soret
bands.16
EXPERIMENTAL SECTION
■
Instrumentation. UV−vis spectra were recorded on an HP 8453
spectrophotometer.
Cyclic voltammetry was performed with an EG&G model 263A
potentiostat having a three-electrode system: a glassy carbon working
electrode, a platinum wire counter electrode, and a saturated calomel
reference electrode (SCE). Tetra(n-butyl)ammonium perchlorate
(TBAP), recrystallized from absolute ethanol and dried in a desiccator
for at least 1 week, was used as the supporting electrolyte. Anhydrous
CH2Cl2 was used as the solvent. The reference electrode was
separated from bulk solution by a fritted-glass bridge filled with the
solvent/supporting electrolyte mixture. Pure argon was bubbled
through the sample solutions for at least 5 min before the experiments
were run. An argon blanket was maintained over the solutions while
the experiments were in progress. All potentials were referenced to the
SCE.
NMR spectra were recorded on a Mercury Plus Varian spectrometer
1
(400 MHz for H and 376 MHz for 19F) at room temperature in
1
CDCl3. H and 19F shifts (δ) in ppm were referenced to residual
chloroform (δ = 7.26) and 2,2,2-trifluoroethanol-d3 (δ = −77.8),
respectively.
Matrix-assisted laser desorption ionization time-of-flight (MALDI-
TOF; α-cyano-4-hydroxycinnamic acid used as the matrix) and laser
desorption ionization (LDI) mass spectra were recorded on a Waters
Micromass MALDI micro MX mass spectrometer.
Materials. All reagents and solvents were used as purchased. Silica
(DAVISIL LC150A 35−70 μm) was used for flash chromatography.
Silica gel 60 preparative thin-layer chromatography (TLC) plates
(20 cm × 20 cm, 0.5 mm thick, Merck) were used for further
purification of some gold corroles for characterization purposes. Free
base corroles H3[T(p-X−P)C], where X = CF3, F, H, CH3, and OCH3,
were synthesized as described.30
Figure 3. BP86-D saddling potentials for selected coinage metal
corroles. Except for the saddling dihedral χ, all other internal
coordinates have been optimized.
Synthesis of Gold Triarylcorroles. A detailed procedure is given
below for gold(III) 5,10,15-tris(4-fluorophenyl)corrole. Exactly the
12847
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
Figure 4. BP86/STO-TZP frontier MOs of (a) Cu[TPC] and (b) Au[TPC]. Also shown are C2 irreducible representations and orbital energies.
The MOs of Ag[TPC] are not shown because they are visually almost indistinguishable from those of Cu[TPC]. Part c shows the Kohn−Sham
MO energy levels (eV) for the same two molecules, with the HOMO−LUMO gaps indicated.
same procedure is applicable for the other gold triarylcorroles reported
herein, except for certain details of chromatographic purification,
which varied for the different complexes and are therefore indicated
separately.
Gold 5,10,15-(4-Fluorophenyl)corrole. To a 100-mL round-
bottomed flask with H3[T(p-F−P)C] (0.086 mmol, 50 mg) dissolved
in 5 mL of pyridine was added 5 equiv of gold(III) acetate (161 mg,
0.43 mmol). The reaction mixture was stirred overnight for at least
16 h. The resulting reddish-brown mixture was rotary-evaporated, and
the brown residue obtained was chromatographed on a silica gel
column with 1:1 n-hexane/CH2Cl2. The gold corrole eluted as the first
red band and was obtained as a dark-red solid after rotary evaporation.
Table 2. Half-Wave Potentials vs SCE (E1/2, V) and
Hammett ρ Values (mV) for M[T(p-X−P)C] in CH2Cl2
Containing 0.1 M TBAP (Scan Rate = 0.1 V s−1)
oxidation
first second
reduction
first
M
X
3σ
ρox
ρred
a
Cu
CF3
F
1.62
0.18
0.89
0.86
0.76
0.70
0.65
0.91
0.82
0.73
0.69
0.66
0.94
0.85
0.80
0.78
0.76
−0.08
−0.13
−0.20
−0.23
−0.24
−0.78
−0.87
−0.86
−0.88
−0.91
−1.29
−1.37
−1.38
−1.42
−1.57
97
69
b
a
H
0.00
a
CH3
OCH3
CF3
F
−0.51
−0.81
1.62
a
a
Ag
Au
105
75
51
95
b
0.18
1.27
a
H
0.00
a
CH3
OCH3
CF3
F
−0.51
−0.81
1.62
a
b
b
0.18
1.38
1.35
1.35
1.32
b
H
0.00
b
CH3
OCH3
−0.51
−0.81
b
a
Alemayehu, A.; Conradie, J.; Ghosh, A. Eur. J. Inorg. Chem. 2011,
Figure 5. Cyclic voltammograms of M[T(p-F)PC] (M = Cu, Ag, Au)
in CH2Cl2. See the Supporting Information for experimental details.
b
1857−1864. This work.
12848
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
7.62 (d, J = 8 Hz, 4H, 5,15-m or -o, Ph), 7.58 (d, 2H, J = 8 Hz, 10-m or -o,
Ph), 2.71 (s, 6H, 5, 15-p-CH3), 2.70 (s, 3H, 10-p-CH3). MS (MALDI-
TOF, major isotopomer): M+ = 762.65 (expt), 762.65 (calcd for
C40H29N4Au). Elem anal.: 62.69% C (calcd 62.99%), 3.83% H (calcd
3.83%), 7.27% N (calcd 7.35%).
Gold 5,10,15-Tris(4-methoxyphenyl)corrole. Silica gel chromatog-
raphy with 2:3 n-hexane/CH2Cl2 gave the complex as the first red
eluate. Yield: 16 mg (24%). UV−vis (CH2Cl2): λmax, nm (ε × 10−4,
1
M−1 cm−1) 420 (8.34), 560 (1.26), 580 (1.92). H NMR: δ 9.20 (d,
3
3
2H, JHH = 4.4 Hz, β-H), 9.05 (d, 2H, JHH = 5.2 Hz, β-H), 8.85 (d,
3
3
2H, JHH = 4.4 Hz, β-H), 8.82 (d, 2H, JHH = 4.8 Hz, β-H), 8.22 (d,
4H, 3JHH = 8.8 Hz, 5,15-m or -o, Ph), 8.11 (d, 2H, 3JHH = 8.8 Hz, 10-m
3
or -o, Ph), 7.36 (d, 4H, JHH = 8.8 Hz, 5,15-o or -m, Ph), 7.32 (d, 2H,
3JHH = 8.8 Hz, 10-o or -m, Ph), 4.11 (s, 6H, 5,15-p-OCH3, Ph), 4.10 (s,
3H, 10-p-OCH3, Ph). MS (MALDI-TOF, major isotopomer): M+ =
810.14 (expt), 810.65 (calcd for C40H29N4O3Au). Elem anal.: 58.98%
C (calcd 59.26%), 3.55% H (calcd 3.61%), 6.89% N (calcd 6.91%).
Copper and Silver Corrole Syntheses. Copper and silver 5,10,
15-tris(4-fluorophenyl)corroles were synthesized according to reported
procedures.7,9a The latter was crystallized as dark-red needles by the
slow evaporation of its chloroform solution within 1 week.
Figure 6. Electronic absorption spectra of Au[T(p-X−P)C] in
CH2Cl2. Inset: spectra of Cu[T(p-X−P)C].
The complex was purified with preparative TLC using 13:7 n-hexane/
CH2Cl2 as the eluent. Yield: 18 mg (27%). Dark-red needles suitable
for X-ray analyses were obtained by the slow evaporation of a
chloroform solution of the complex within 8 days. UV−vis (CH2Cl2):
λmax, nm (ε × 10−4, M−1 cm−1) 419 (13.89), 530 (0.87), 559 (2.64),
573 (3.06). 1H NMR: δ 9.20 (d, 2H, 3JHH = 4.4 Hz, β-H), 9.01 (d, 2H,
Copper 5,10,15-Tris(4-fluorophenyl)corrole. Yield: 84.6%. UV−vis
(CH2Cl2): λmax, nm (ε × 10−4, M−1 cm−1) 413 (9.18), 539 (0.69),
615 (0.10). 1H NMR: δ 7.92 (d, 2H, 3JHH = 4 Hz, β-H), 7.75 (dd, 4H,
3JHH = 8 Hz, 3JHF = 5.6 Hz, 5,15-o or -m, Ph), 7.68−7.62 (m, 4H, β-H
3
3JHH = 4.8 Hz, β-H), 8.84 (d, 2H, JHH = 4.4 Hz, β-H), 8.78 (d, 2H,
3
3JHH = 5.2 Hz, β-H), 8.24 (dd, 4H, 3JHH = 8.4 Hz, 3JHF = 5.6 Hz, 5,15-o
or -m, Ph), 8.14 (dd, 2H, 3JHH = 8.4 Hz, 3JHF = 5.6 Hz, 10-o or -m, Ph),
7.56 − 7.44 (6H, 5,15-m or -o and 10-m or -o, Ph, overlapping
triplets). 19F NMR: δ −115.52 (m, 2F, 5,15-p-F, Ph), −115.60 (m, 1F,
10-p-F, Ph). MS (MALDI-TOF, major isotopomer): M+ = 774.09
(expt), 774.54 (calcd for C37H20N4F3Au). Elem anal.: 57.19% C (calcd
57.38%), 2.48% H (calcd 2.60%), 7.10% N (calcd 7.23%).
and 10-o or -m, Ph), 7.34 (d, 2H, JHH = 4 Hz, β-H), 7.24−7.25 (m,
8H, β-H, 10-m or -o, Ph and 5,15-m or -o, Ph, overlapping). 19F NMR:
δ −112.64 (m, 2F, 5,15-p-F, Ph), −112.95 (m, 1F, 10-p-F, Ph). MS
(MALDI-TOF, major isotopomer): M+ = 640.13 (expt), 641.12 (calcd
for C37H20N4F3Cu). Elem anal.: 69.66% C (calcd 69.32%), 3.28% H
(calcd 3.14%), 8.89% N (calcd 8.74%).
Silver 5,10,15-Tris(4-fluorophenyl)corrole. Yield: 70%. UV−vis
(CH2Cl2): λmax, nm (ε × 10−4, M−1 cm−1) 422(11.53), 522 (0.75),
Gold 5,10,15-Triphenylcorrole. Column chromatography on silica
gel with 4:1 n-hexane/CH2Cl2 gave the complex as the first red eluate.
Final purification was accomplished with preparative TLC, where 1:1
n-hexane/CH2Cl2 was used as the eluent, yielding the pure complex
as the first red band. Yield: 17 mg (25%). UV−vis (CH2Cl2): λmax, nm
1
3
563 (0.32), 581 (3.30). H NMR: δ 9.21 (d, 2H, JHH = 4 Hz, β-H),
3
3
8.93 (d, 2H, JHH = 4.8 Hz, β-H), 8.73 (d, 2H, JHH = 4.8 Hz, β-H),
8.71 (d, 2H, JHH = 4.4 Hz, β-H), 8.26 (dd, 4H, JHH = 8.8 Hz, JHF
3
3
3
=
3
5.6 Hz, 5,15-o or -m, Ph), 8.16 (dd, 2H, JHH = 8.8 Hz, 3JHF = 5.6 Hz,
10-o or -m, Ph), 7.56−7.44 (6H, 5,15-m or -o and 10-m or -o, Ph,
overlapping triplets). 19F NMR: δ −115.38 (m, 2F, 5,15-p-F, Ph),
−115.53 (m, 1F, 10-p-F, Ph). MS (LDI-TOF, major isotopomer):
M+ = 686.08 (expt), 685.45 (calcd for C37H20N4F3Ag). Elem anal.: 65.10%
C (calcd 64.83%), 3.05% H (calcd 2.94%), 8.35% N (calcd 8.17%).
Crystallographic Analysis of Au[T(p-F−P)C]·0.5C6H14. A dark-
red needle of dimensions 0.35 × 0.04 × 0.04 mm3 was mounted in the
100(2) K nitrogen cold stream provided by an Oxford Cryostream
low-temperature apparatus on the goniometer head of a Bruker D85
diffractometer equipped with an Apex II CCD detector, on beamline
11.3.1 at the Advanced Light Source in Berkeley, CA. Diffraction data
were collected using synchrotron radiation monochromated with
silicon(111) to a wavelength of 0.774 90(1) Å. A full sphere of data
was collected using 0.3° ω scans. A multiscan absorption correction
was applied using the program SADABS 2008/1. The data consist of
48 038 reflections collected, of which 10 431 were unique [R(int) =
0.0563] and 9238 were observed [I > 2σ(I)]. The structure was solved
by direct methods (SHELXS) and refined by full-matrix least squares
on F2 (SHELXL-97) using 455 parameters and 28 restraints. The
solvent, hexane, is disordered. Two orientations were identified and
refined. Because of the overlapping of the two orientations, distance
1−2 and 1−3 restraints were used to control the hexane geometries.
The hydrogen atoms on carbon atoms were generated geometrically
and refined as riding atoms with C−H = 0.95−0.99 Å and Uiso(H) =
1.2Ueq(C) for CH and CH2 groups and Uiso(H) = 1.5Ueq(C) for CH3
groups. The maximum and minimum peaks in the final difference Fourier
1
(ε × 10−4, M−1 cm−1) 418 (12.31), 560 (2.37), 575 (2.99). H NMR:
δ 9.18 (d, 2H, 3JHH = 4 Hz, β-H), 9.04 (d, 2H, J = 4 Hz, β-H), 8.87 (d,
3
3
2H, JHH = 4 Hz, β-H), 8.81 (d, 2H, JHH = 4 Hz, β-H), 8.29 (d, 4H,
3JHH = 8 Hz, 5,15-o or -m, Ph), 8.20 (d, 2H, JHH = 8 Hz, 10-o or- m,
3
Ph), 7.83−7.75 (m, 9H, 5,15-m or -o; 10-m or -o, and 5,10,15-p, Ph,
overlapping). MS (MALDI-TOF, major isotopomer): M+ = 720.57
(expt), 720.67 (calcd for C37H23N4Au). Elem anal.: 61.38% C (calcd
61.67%), 3.10% H (calcd 3.22%), 7.65% N (calcd 7.78%).
Gold 5,10,15-Tris(4-trifluoromethylphenyl)corrole. Two succes-
sive chromatographic separations on silica gel columns with 3:2
n-hexane/CH2Cl2 yielded the complex as the first red eluate. Final
purification was accomplished with preparative TLC, where 3:2
n-hexane/CH2Cl2 was used as the eluent, giving the complex as the
first red band. Yield: 15 mg (24%). UV−vis (CH2Cl2): λmax, nm (ε ×
1
10−4, M−1 cm−1) 419 (16.79), 530 (0.84), 571 (3.16). H NMR: δ
3
3
9.17 (d, 2H, JHH = 4.4 Hz, β-H), 9.00 (d, 2H, JHH = 4.4 Hz, β-H),
3
3
8.81 (d, 2H, JHH = 4.8 Hz, β-H), 8.78 (d, 2H, JHH = 4.8 Hz, β-H),
3
3
8.38 (d, 4H, JHH = 8.0 Hz, 5,15-m or -o, Ph), 8.29 (d, 2H, JHH = 7.6
Hz, 10-m or -o, Ph), 8.09 (d, 4H, 3JHH = 7.6 Hz, 5,15-o or -m, Ph), 8.06
(d, 2H, JHH = 8.0 Hz, 10-o or -m, Ph). 19F NMR: δ −62.53 (s, 9F,
3
5,10, 15-p-CF3, Ph). MS (MALDI-TOF, major isotopomer): M+
=
924.06 (expt), 924.57 (calcd for C40H20N4F9Au). Elem anal.: 52.12%
C (calcd 51.96%), 2.44% H (calcd 2.18%), 5.77% N (calcd 6.06%).
Gold 5,10,15-Tris(4-methylphenyl)corrole. Silica gel chromato-
graphy with 4:1 n-hexane/CH2Cl2 gave the complex as the first red
eluate. Final purification was accomplished with preparative TLC,
where 1:1 n-hexane/CH2Cl2 was used as the eluent, giving the com-
map were 1.599 and −1.291 e Å−3. Crystal data: C40H27F3N4Au, Mw
=
817.62, monoclinic, C2/c, a = 21.0234(17) Å, b = 19.9239(16) Å, c =
15.8500(13) Å, α = 91.4200(10)°, V = 6637.0(9) Å3, T = 100(2) K, Z = 8,
R1 [I > 2σ(I)] = 0.0311, wR2 (all data) = 0.0955, GOF (on F2) = 1.030.
Crystallographic Analysis of Ag[T(p-F−P)C]·CHCl3. A dark-red
block of dimensions 0.05 × 0.03 × 0.02 mm3 was mounted as
described above, on beamline 11.3.1 at the Advanced Light Source in
plex as the first red band. Yield: 24 mg (35%). UV−vis (CH2Cl2): λmax
,
1
nm (ε × 10−4, M−1 cm−1) 420 (13.16), 560 (2.31), 576 (3.25). H
NMR: δ 9.16 (d, 2H, J = 4 Hz, β-H), 9.05 (d, 2H, J = 8 Hz, β-H), 8.85
(d, 2H, J = 8 Hz, β-H), 8.81 (d, 2H, J = 8 Hz, β-H), 8.17 (d, 4H,
J = 8 Hz, 5,15-o or -m, Ph), 8.08 (d, 2H, J = 8 Hz, 10-o or -m, Ph),
12849
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
Berkeley, CA. Diffraction data were collected and processed in the
method described above. The data consist of 47 883 reflections
collected, of which 9866 were unique [R(int) = 0.0558] and 7634 were
observed [I > 2σ(I)]. The structure was solved by direct methods
(SHELXS) and refined by full-matrix least squares on F2 (SHELXL-97)
using 442 parameters and 0 restraints. The hydrogen atoms on carbon
atoms were generated geometrically and refined as riding atoms as
described above. The maximum and minimum peaks in the final
difference Fourier map were 1.283 and −2.000 e Å−3. Crystal data:
C38H21F3N4Cl3Ag, Mw = 804.81, monoclinic, C2/c, a = 20.2684(9) Å,
b = 20.2432(9) Å, c = 15.6451(7) Å, α = 90.478(3)°, V = 6418.9(5) Å3,
T = 100(2) K, Z = 8, R1 [I > 2σ(I)] = 0.0516, wR2 (all data) = 0.1437,
GOF (on F2) = 1.029.
DFT Calculations. All calculations were carried out with the ADF
2009 program system. The dispersion-corrected BP86-D functional
was used throughout, and scalar relativistic effects were taken into
account with the ZORA Hamiltonian and ZORA STO-TZP basis sets.
Fine integration grids and tight criteria for self-consistent-field and
geometry optimization were used to ensure accurate geometries and
saddling potential curves.
Z. Eur. J. Inorg. Chem. 2004, 1724−1732. (c) Broring, M.; Bregier, F.;
̈
Tejero, E. C.; Hell, C.; Hell, C.; Holthausen, M. C. Angew. Chem., Int.
Ed. 2007, 46, 445−448.
(10) For other examples of silver corrole X-ray structures, see:
(a) Pacholska, E.; Espinosa, E.; Guilard, R. Dalton Trans. 2004, 3181−
3183. (b) Bruckner, C.; Barta, C. A.; Brinas, R. P.; Bauer, J. A. K. Inorg.
̈
Chem. 2003, 42, 1673−1680. (c) Stefanelli, M.; Mastroianni, M.;
Nardis, S.; Licoccia, S.; Fronczek, F. R.; Smith, K. M.; Zhu, W.; Ou, Z.;
Kadish, K. M.; Paolesse, R. Inorg. Chem. 2007, 46, 10791−10799.
(11) The fact that these distances are all fairly similar is dictated by
the steric constraints of the corrole macrocycle. The fact that the bond
distances involving silver and gold are similar is explained by their
similar ionic radii, 0.89 and 0.99 Å, respectively, for the 3+ ions, which
may be compared with significantly smaller values of 0.54 Å for CuIII
and 0.73 Å for CuII: Shannon, R. D. Acta Crystallogr. 1976, A32, 751−
767.
(12) The fact that copper corroles are inherently11 saddled was first
noted in: Alemayehu, A. B.; Gonzalez, E.; Hansen, L.-K.; Ghosh, A.
Inorg. Chem. 2009, 48, 7794−7799.
(13) The term “inherent” or “intrinsic” refers to the fact that the
saddling is driven primarily by metal−ligand orbital interactions and
not by peripheral substituents, although the latter can accentuate the
saddling. The term “intrinsic ruffling” has been used for nickel
porphyrins: Kozlowski, P. M.; Rush, T. S.; Jarzecki, A. A.; Zgierski, M.
Z.; Chase, B.; Piffat, C.; Ye, B.-H.; Li, X.-Y.; Pulay, P.; Spiro, T. G.
J. Phys. Chem. A 1999, 103, 1357−1366.
(14) The two HOMOs of corrole, a2 and b1 in C2v symmetry, are very
similar in shape to the porphyrin a1u and a2u HOMOs, respectively. We
will therefore continue to use the D4h porphyrin irreps to describe the
corrole HOMOs: Ghosh, A.; Wondimagegn, T.; Parusel, A. B. J.
J. Am. Chem. Soc. 2000, 122, 5100−5104.
ASSOCIATED CONTENT
* Supporting Information
■
S
DFT-optimized coordinates and a combined CIF file. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: abhik.ghosh@uit.no.
■
(15) For reviews on nonplanar porphyrins, see: (a) Senge, M. Chem.
Commun. 2006, 243−256. (b) Shelnutt, J. A.; Song, X. Z.; Ma, J. G.;
Jia, S. L.; Jentzen, W.; Medforth, C. J. Chem. Soc. Rev. 1998, 27, 31−41.
(16) Ghosh, A.; Halvorsen, I.; Nilsen, H. J.; Steene, E.;
Wondimagegn, T.; Lie, E.; van Caemelbecke, E.; Guo, N.; Ou, Z.;
Kadish, K. M. J. Phys. Chem. B 2001, 105, 8120−8124.
(17) (a) Renner, M. W.; Barkigia, K. M.; Zhang, Y.; Medforth, C. J.;
Smith, K. M.; Fajer, J. J. Am. Chem. Soc. 1994, 116, 8582−8592.
(b) Renner, M. W.; Barkigia, K. M.; Fajer, J. Inorg. Chim. Acta 1997,
263, 181−187.
(18) Alemayehu, A.; Hansen, L. K.; Ghosh, A. Inorg. Chem. 2010, 49,
7608−7610.
(19) The most strongly saddled metallocorrole reported to date is a
copper β-octakis(trifluoromethyl)-meso-triarylcorrole: Thomas, K. E.;
Hansen, L. K.; Conradie, J.; Ghosh, A. Eur. J. Inorg. Chem. 2011,
1865−1870.
(20) Paolesse, R.; Nardis, S.; Sagone, F.; Khoury, R. G. J. Org. Chem.
2001, 66, 550−556.
ACKNOWLEDGMENTS
■
This work was supported by the Research Council of Norway.
The Advanced Light Source is supported by the Director,
Office of Science, Office of Basic Energy Sciences, of the U.S.
Department of Energy under Contract No. DE-AC02-
05CH11231. J.C. acknowledges the National Research Fund
of the Republic of South Africa.
REFERENCES
■
(1) For a recent review, see: Aviv-Harel, I.; Gross, Z. Chem.Eur. J.
2009, 15, 8382−8394.
(2) Tse, M. K.; Zhang, Z.; Mak, T. C. W.; Chan., K. S. J. Chem. Soc.,
Chem. Commun. 1998, 1199−1200.
(3) (a) Palmer, J. H.; Day, M. W.; Wilson, A. D.; Henling, L. M.;
Gross, Z.; Gray, H. B. J. Am. Chem. Soc. 2008, 130, 7786−7787.
(b) Palmer, J. H.; Durrell, A. C.; Gross, Z.; Winkler, J. R.; Gray, H. B.
J. Am. Chem. Soc. 2010, 132, 9230−9231.
(21) Ligand noninnocence is ubiquitous for metallocorroles. For key
studies, see: (a) Cai, S.; Walker, F. A.; Licoccia, S. Inorg. Chem. 2000,
39, 3466−3478. (b) Steene, E.; Wondimagegn, T.; Ghosh, A. J. Phys.
Chem. B 2001, 105, 11406−11413; Addition/correction: J. Phys.
(4) Alemayehu, A.; Ghosh, A. J. Porphyrins Phthalocyanines 2011, 15,
106−110.
(5) While this paper was under review, a report of gold(III)
β-octabromo-meso-tris(pentafluorphenyl)corrole, Au[Br8TPFPC], includ-
ing an X-ray crystal structure, appeared in the literature: Rabinovitch, E.;
Goldberg, I.; Gross, Z. Chem.Eur. J. 2011, 17, 12294−12301.
(6) Reviews on DFT studies of porphyrins and related compounds:
(a) Ghosh, A. Acc. Chem. Res. 1998, 31, 189−198. (b) Ghosh, A. In
The Porphyrin Handbook; Kadish, K. M., Guilard, R., Smith, K. M.,
Eds.; Academic: San Diego, 1999; Vol. 7, Chapter 47, pp 1−38.
(c) Ghosh, A.; Steene, E. J. Biol. Inorg. Chem. 2001, 6, 739−752.
(d) Ghosh, A. J. Biol. Inorg. Chem. 2006, 11, 712−724.
(7) Wasbotten, I. H.; Wondimagegn, T.; Ghosh, A. J. Am. Chem. Soc.
2002, 124, 8104−8116.
Chem. B 2002, 106, 5312. (c) Zakharieva, O.; Schunemann, V.;
̈
Gerdan, M.; Licoccia, S.; Cai, S.; Walker, F. A.; Trautwein, A. X. J. Am.
Chem. Soc. 2002, 124, 6636−6648. (d) Nardis, S.; Paolesse, R.;
Licoccia, S.; Fronczek, F. R.; Vicente, M. G. H.; Shokhireva, T. K.; Cai,
S.; Walker, F. A. Inorg. Chem. 2005, 44, 7030−7046. (e) Walker, F. A.;
Licoccia, S.; Paolesse, R. J. Inorg. Biochem. 2006, 100, 810−837.
(f) Roos, B. O.; Veryazov, V.; Conradie, J.; Taylor, P. R.; Ghosh, A.
J. Phys. Chem. 2008, 112, 14099−14102. (g) Kadish, K. M.; Shen, J.;
Fremond, L.; Chen, P.; Ojaimi, M. E.; Chkounda, M.; Gros, C. P.;
Barbe, J.-M.; Ohkubo, K.; Fukuzumi, S.; Guilard, R. Inorg. Chem. 2008,
47, 6726−6737.
(22) (a) Becke, A. D. Phys. Rev. 1988, A38, 3098−3100. (b) Perdew,
J. P. Phys. Rev. 1986, B33, 8822. Erratum: Perdew, J. P. Phys. Rev.
1986, B34, 7406.
(8) Steene, E.; Wondimagegn, T.; Ghosh, A. J. Am. Chem. Soc. 2003,
125, 16300−16309.
(9) For key early examples of copper corrole X-ray structures, see:
(a) Bruckner, C.; Brinas, R. P.; Bauer, J. A. K. Inorg. Chem. 2003, 42,
4495−4497. (b) Luobeznova, I.; Simkhovich, L.; Goldberg, I.; Gross,
(23) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104−154119.
̈
12850
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851

Inorganic Chemistry
Article
(24) The ADF program system uses methods described in: Velde, G. T.;
Bickelhaupt, F. M.; Baerends, E. J.; Guerra, C. F.; Van Gisbergen, S. J. A.;
Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001, 22, 931−967.
(25) The noninnocent nature of copper corroles has been confirmed
by ab initio CASPT2 calculations: Pierloot, K.; Zhao, H.; Vancoillie, S.
Inorg. Chem. 2010, 49, 10316−10329.
(26) For a recent study, see: Assadollahzadeh, B.; Schwerdtfeger, P.
Chem. Phys. Lett. 2008, 462, 222−228.
(27) Ou, Z.; Shao, J.; Zhao, H.; Ohkubo, K.; Wasbotten, I. H.;
Fukuzumi, S.; Ghosh, A.; Kadish, K. M. J. Porphyrins Phthalocyanines
2004, 8, 1236−1247.
(28) The Hammett analyses included in Table 2 do not appear to be
particularly valuable in this regard. The various ρ’s are all rather
similar,24 suggesting that the various redox processes have comparable
amounts of corrole-centered character. Our MO picture is consistent
with this conclusion, even though in the copper and silver cases the
LUMOs have significant metal character as well and the reductions are
partially metal-centered.
(29) Much higher ρ’s have been measured for copper β-octakis-
(trifluoromethyl)-meso-triarylcorroles:17 Thomas, K. E.; Wasbotten,
I. H.; Ghosh, A. 2008, 47, 10469−10478. Addition/correction: Inorg.
Chem. 2009, 48, 1257−1257.
(30) Koszarna, B.; Gryko, D. T. J. Org. Chem. 2006, 71, 3707−3717.
12851
dx.doi.org/10.1021/ic202023r | Inorg. Chem. 2011, 50, 12844−12851