Toward a square grid polymer: Electrochemistry of tentacled tetragonal star connectors, C4R4-Co-C5(HgY)5, on mercury

Article abstract of DOI:10.1021/jp983242i

Electrochemical reduction of cyclobutadienecyclopentadienylcobalt derivatives, C₄R₄-Co-C₅(H_gY)₅ [R = C₆H₅-, Y = -SMe (3), -SCOMe (4), -SCH₂CH₂SCH₃ (5), -SCH₂CH₂SCOCH₃ (6); R= p-EtOCOC₆U₄-, Y = -SEt (7), -SCOMe (8), -SCH₂CH₂SCH₃ (9), -SCH₂CH₂SCOCH₃ (10)], and the parent C₄R₄-Co-C₅H₅ [R = C₆H₅- (11) and R = p-EtOCOC₆H₄- (12)] has been examined. Reversible one-electron metalcentered reduction of 11 (-2.24 V vs Ag/AgCl) yields a stable radical anion. In 12, it takes place at -1.70 V. In 3-10, similar metal-centered reduction occurs irreversibly, accompanied by an adsorption pre-wave at -0.7 to -1.15 V. We propose that the radical anion loses a -HgSX substituent soon after it is formed. Electrochemical cleavage of the substituents occurs before the metal-centered reduction in 4 and 8 and afterward in the others. The cleaved thiolate fragments yield an anodic peak. In all compounds with ester groups, ester reduction is observed at -2.1 to -2.5 V.

Full text of DOI:10.1021/jp983242i

10062
J. Phys. Chem. B 1998, 102, 10062-10070
Toward a Square Grid Polymer: Electrochemistry of Tentacled Tetragonal Star
Connectors, C4R4-Co-C5(HgY)5, on Mercury
Thierry Brotin,† Lubom´ır Posp´ısˇil,‡ Jan Fiedler,‡ Benjamin T. King,† and Josef Michl*,†
Department of Chemistry and Biochemistry, UniVersity of Colorado, Boulder, Colorado 80309-0215, and
J. HeyroVsky´ Institute of Physical Chemistry, DolejsˇkoVa 3, 18223 Prague 8, Czech Republic
ReceiVed: July 31, 1998
Electrochemical reduction of cyclobutadienecyclopentadienylcobalt derivatives, C4R4-Co-C5(HgY)5 [R )
C6H5-, Y ) -SMe (3), -SCOMe (4), -SCH2CH2SCH3 (5), -SCH2CH2SCOCH3 (6); R ) p-EtOCOC6H4-,
Y ) -SEt (7), -SCOMe (8), -SCH2CH2SCH3 (9), -SCH2CH2SCOCH3 (10)], and the parent C4R4-Co-
C5H5 [R ) C6H5- (11) and R ) p-EtOCOC6H4- (12)] has been examined. Reversible one-electron “metal-
centered” reduction of 11 (-2.24 V vs Ag/AgCl) yields a stable radical anion. In 12, it takes place at -1.70
V. In 3-10, similar metal-centered reduction occurs irreversibly, accompanied by an adsorption pre-wave at
-0.7 to -1.15 V. We propose that the radical anion loses a -HgSX substituent soon after it is formed.
Electrochemical cleavage of the substituents occurs before the metal-centered reduction in 4 and 8 and afterward
in the others. The cleaved thiolate fragments yield an anodic peak. In all compounds with ester groups,
ester reduction is observed at -2.1 to -2.5 V.
Introduction
and comparison with the parent structures 11 and 12 seemed
advisable before an investigation of adsorption properties. The
results are reported presently, and the adsorption behavior will
be described separately.5
The overall scheme proposed1 for the synthesis of square grid
polymers by interfacial synthesis calls for monomeric modules
consisting of a cross-shaped connector mounted on a pedestal
that carries three or more tentacles with functionalities that have
high affinity for the surface of a liquid such as mercury. The
structures that we have initially chosen for this purpose are
organometallic sandwich complexes and, among others, tetra-
p-substituted tetraphenylcyclobutadienecyclopentadienylcobalt
complexes. The five mercurio substituents are introduced onto
the cyclopentadienyl (Cp) ring of these complexes with mercuric
trifluoracetate to afford 1 and 2 in good yield, and the trifluoro-
acetate moiety is subsequently easily displaced by thiol-
terminated chains to attach almost any kind of desired structure.2
We have exploited this facile access for the preparation of a
series of pentatentacled derivatives of 1 and 2 and hope to find
the potential ranges within which they adhere firmly to a
mercury surface. The simpler model compounds 3-6 with
unsubstituted phenyl groups on the cyclobutadiene (Cb) ring
are useless for future grid formation, but should help us unravel
the electrochemical behavior of the more complex compounds
7-10. Of the two choices of adhesive groups, the alkylthio
group, present in 3, 5, 7, and 9, was previously found to adsorb
firmly on gold,3 and on mercury at positive potentials.4 The
potential range in which the acetylthio groups of 4, 6, 8, and
10 will adsorb on mercury is unknown. In principle, the
acetylthio group also offers an opportunity for hydrolytic
conversion into a thiol group, which will surely adhere firmly
to both metals, but the free thiols appear to be unstable with
respect to thiol ligand exchange on mercury. They will probably
have to be generated from adsorbed acetylthio derivatives in
situ, and we have not examined them so far.
Little is known about the electrochemical behavior of
cyclobutadienecyclopentadienylcobalt complexes. The unsub-
stituted parent compound, CbCoCp, undergoes two irreversible
oxidation processes at +0.79 and +1.14 V vs saturated calomel
electrode (SCE),6 but its reduction has apparently not been
examined. A derivative with four alkyl substituents on the Cb
ring shows a reversible oxidation wave at +0.46 V vs SCE,7
and some more complicated structures with two Co centers have
also been investigated,8 but once again, the reduction seems
not to have been examined. A related cation, tetraphenylcy-
clobutadiene(benzene)cobalt(I),9 is reduced reversibly at -1.22
V vs SCE, then irreversibly at -1.75 V and finally again
reversibly at -2.06 V, and the product is oxidized irreversibly
at -0.23 V. A replacement of the four phenyl substituents with
Because of the multitude of electroactive groups in the
molecules, an examination of oxidation-reduction properties
† University of Colorado.
‡ J. Heyrovsky´ Institute of Physical Chemistry.
10.1021/jp983242i CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/05/1998

Tentacled Tetragonal Star Connectors
J. Phys. Chem. B, Vol. 102, No. 49, 1998 10063
four methyl groups changes the reduction pattern; only one
irreversible reduction at -1.92 V and one irreversible oxidation
at -0.80 V are observed. The formation of a protonated
cyclobutenyl complex was postulated as a reason for the
irreversibility and for the appearance of an oxidation wave.9
2-Acetylthioethanethiol. To a stirred solution of KHCO3
(10.63 g, 0.106 mol) in water (100 mL), a solution of
1,2-ethanedithiol (10 g, 0.106 mol) in diethyl ether (60 mL)
was added. The mixture was cooled to 0 °C, and acetic
anhydride (10 mL) was slowly added in one portion. The
mixture was stirred at room temperature for 45 min, cooled at
0 °C, and quenched with 5 N HCl (30 mL). The organic
material was extracted with diethyl ether, and the extract was
washed with water and dried over Na2SO4. Evaporation of
diethyl ether gave rise to 2 g (11%, mp 68-69 °C, lit.12 68.5
°C) of 1,2-bisacetylthioethane, which was collected on a frit.
The residual oil was chromatographed on silica gel (ether/
pentane, 10/90). The desired compound was eluted first. After
the solvent was evaporated, the oily product was distilled twice
to give the pure product as a colorless oil (1.2 g, 8%): bp 87-
Experimental Section
Methods. Electrochemical measurements were done using
a home-built electrochemical system for cyclic voltammetry,
phase-sensitive AC polarography, and DC polarography. It
consisted of a fast rise-time potentiostat and a lock-in amplifier
(EG&G, model 5209). The instruments were interfaced to a
486-type personal computer via an IEEE interface card (PC-
Lab, AdvanTech, model PCL-848) and a data acquisition card
(PCL-818) using 12-bit precision. The source of the AC signal
for the cell was derived from the internal oscillator of the lock-
in amplifier. The amplitude was attenuated to 5 mV (p-p).
The impedance spectra were measured in the range 1 Hz to
100 kHz using a frequency spectrum analyzer (Stanford
Research, model SRS760) connected to a potentiostat (EG&G
Princeton Applied Research, model 263). A three-electrode
electrochemical cell was used. Ag/AgCl/1M LiCl served as a
reference electrode and was separated from the test solution by
a salt bridge. The half-wave potential of the ferrocene couple
vs our reference electrode was +0.562 V. The working
electrode was a valve-operated static mercury electrode (SM-
DE2, Laboratorn´ı Prˇ´ıstroje, Prague) with an area of 1.32 × 10-2
cm-2. The auxiliary electrode was a cylindrical platinum net.
Oxygen was removed from the solution by passing a stream of
argon. THF was distilled under the inert atmosphere prior to
use. Preparative electrolysis of 5 was performed on a Hg pool
electrode of a 3.8 cm2 area. The anode and cathode spaces
were separated by a horizontal salt bridge of 8 cm terminated
on both sides with a glass frit. The solution was stirred with a
magnetic stirrer. Potential was controlled with a EG&G model
273 potentiostat that also measured the charged that passed.
Polarographic or voltammetric curves were recorded prior to
and after the electrolysis. After completed electrolysis, the
solvent was removed under reduced pressure and the resulting
solid was triturated with benzene. The extract was evaporated
to dryness and again triturated with benzene, and the process
was repeated one more time. Final purification of the reduction
product was by chromatography on silica gel with hexanes/
benzene. Product identification was done by comparison of 1H
NMR and of TLC Rf values with those of authentic standards.
Melting points were recorded using a Melt temp II apparatus
1
90 °C/17 Torr (lit.13 92 °C/17 Torr); H NMR (CDCl3) δ 1.59
(t, 1H, J ) 8.4 Hz), 2.32 (s, 3 H, COCH3), 2.655 (m, 2 H,
CH2), 3.07 (m, 2 H, CH2); 13C {1H} NMR (CDCl3) δ 24.60
(CH2), 30.60 (CH2), 33.10 (CH3), 195.13 (CO); IR (neat, cm-1
)
2927 (ν -CH, w), 2564 (ν SH, w), 1691 (ν CdO, s), 1424 (δ
CH3, CH2, s), 1354 (δ CH3, s), 1273 (m), 1215 (m), 1134-
1105 (s), 954 (s), 702 (w), 624 (ν S-CO, m); MS (EI+) m/z
(rel int.) 136 (M, 7%), 77 (14%), 59 (32%), 43 (100%, CH3-
CO+), 27 (28%, CH2dCH+), 15 (29%, CH3+). Anal. Calcd
for C4H8S2O: C, 35.27; H, 5.92. Found: C, 35.42; H, 6.03.
Tetraphenylcyclobutadienepentakis(methylthiomercurio)-
cyclopentadienylcobalt (3). A solution of sodium thiomethox-
ide (0.035 g, 0.45 mmol) was dissolved in a minimum of water
and added under argon in one portion to a solution of 1 (0.152
g, 0.074 mmol) in THF (6 mL) cooled to 0 °C. After being
stirred for 1 min at 0 °C, a mixture of EtOH (5 mL) and hexanes
(50 mL) was added, and stirring was continued for 5 min at
room temperature. The precipitate was collected on a frit,
washed with copious EtOH and then hexanes, and dried in air
1
(85 mg, 66%): mp >250 °C (dec); H NMR δ 7.64 (m, 8 H,
phenyl), 7.20 (m, 12 H, phenyl), 2.12 (s, 15 H, CH3); 13C {1H}
NMR δ 10.01 (CH3), 73.34 (Cb), 125.30 (Cp), 126.33, 128.21,
128.29, 137.34 (Ph); IR (KBr, cm-1) 3054 (ν dCH, vw), 2916
(ν -CH, w), 1630 (ν Ph, vw), 1596 (ν Ph, w), 1497 (m), 1441
(w), 1311 (w), 1118 (w), 695 (m), 559 (m). Useful FAB MS
could not be obtained due to low solubility. Anal. Calcd for
C38H35S5Hg5Co: C, 26.63; H, 2.06. Found: C, 26.88; H, 2.08.
General Procedure for Reactions of 1 and 2 with Thiols.
A solution of thiol dissolved in a minimum of THF was added
in one portion to a solution of 1 or 2 in THF cooled to 0 °C
under argon. The mixture was stirred for 1 min, and a mixture
of hexanes was added to quench the reaction and to precipitate
the desired product. Stirring was continued for 5 min at room
temperature. The compounds were collected on a frit, washed
with copious hexanes, and purified by precipitation (3, 5, 6) or
recrystallization (4, 7, 8-10). Chromatography on silica gel
or basic alumina under ordinary conditions caused partial
decomposition. NMR spectra gave no indication of hindrance
to rotation of the two decks relative to each other.
(Laboratory Devices). Unless specified otherwise, 1H and 13
C
NMR spectra were obtained in DMSO-d6 solution with a Varian
VXR-300 S spectrometer at 299.949 and 75.428 MHz respec-
tively, using Me4Si as an external reference. IR spectra were
obtained with a Nicolet 800 FT-IR spectrometer. Electron
impact (EI) and fast atom bombardment (FAB) mass spectra
were obtained with a VG 7070 EQ-HF hybrid tandem mass
spectrometer. Microanalyses were performed by Desert Ana-
lytics, Tucson, AZ, and the Microanalytical Laboratory at the
University of California, Berkeley.
Materials. All reagents were purchased from Aldrich, except
for ethyl 4-iodobenzoate (TCI America), and were used as
supplied. 2-Methanethioethanethiol was prepared by a re-
ported10 procedure. 4,4′-Di(ethoxycarbonyl)phenylacetylene2
was purified by column chromatography on silica gel. Tetra-
substituted tetraphenylcyclobutadienecyclopentadienylcobalt com-
plexes were prepared by the method of Rausch11 as reported
previously,2 using freshly distilled CpCo(CO)2. All solvents
were analytical grade and were freshly distilled before use.
Tetraphenylcyclobutadienepentakis(acetylthiomercurio)-
cyclopentadienylcobalt (4). Thiolacetic acid (0.19 g, 2.5 mmol)
in THF (2 mL) was added to 1 (0.5 g, 0.25 mmol) in THF (8
mL). After precipitation with hexanes (50 mL), the product
was purified by recrystallization from THF/EtOH (dark-red
crystals, 0.25 g, 55%): mp 165-170 °C (dec); 1H NMR δ 7.715
(m, 8 H, Ph), 7.19 (m, 12 H, Ph), 2.27 (s, 15 H, CH3); 13C
{1H} NMR δ 25.095 (CH3), 74.28 (Cb), 123.19 (Cp), 126.42,
128.39, 128.58 (Ph), 137.10, 201.34 (CO); IR (KBr, cm-1) 3056
(ν dCH, vw), 2926 (ν -CH, vw), 1624 (br, ν CdO), 1497

10064 J. Phys. Chem. B, Vol. 102, No. 49, 1998
Brotin et al.
(m), 1350 (m), 1114 (s), 952 (m), 708 (m), 635 (ν S-CO, s);
MS (FAB+, 3-NOBA) m/z maximum calcd at 1854, found
1856. Anal. Calcd for C43H35S5Hg5Co‚THF: C, 29.30; H,
2.25. Found: C, 29.15; H, 2.18.
found 2075. Anal. Calcd for C55H61S5O8Hg5Co: C, 31.88;
H, 2.96. Found: C, 31.77; H, 2.86.
Tetra(4-ethoxycarbonylphenyl)cyclobutadienepentakis-
(acetylthiomercurio)cyclopentadienylcobalt (8). 8 was syn-
thesized from 2 (0.2 g, 0.086 mmol) in THF (5 mL) and
thiolacetic acid (0.045 g. 0.59 mmol). Hexanes (60 mL)
precipitated 8 (0.13 g, 71%), which was recrystallized from
THF/EtOH (dark-red crystals): mp 200-205 °C (dec); 1H NMR
δ 7.86 (d, 8 H, J ) 8.4 Hz, Ph), 7.79 (d, 8 H, J ) 8.4 Hz, Ph),
4.31 (q, 8 H, J ) 7.2 Hz, CH2), 2.26 (s, 15 H, COCH3), 1.32
(t, 12 H, J ) 7.2 Hz, CH3); 13C {1H} NMR δ 14.29 (CH3),
35.24 (CH3), 60.50 (CH2), 74.05 (Cb), 124.40 (Cp), 127.65,
128.74, 129.43, 142.32 (Ph), 165.58 (SCO), 201.05 (COO); IR
(KBr, cm-1) 2980 (ν CH, w), 1716 (ν CdO, s), 1637 (s), 1603
(ν Ph, s), 1273 (ν CO, s), 1176 (m), 1111 (s), 1017 (m), 952
(m), 861 (m), 772 (vw), 757 (vw), 726 (vw), 703 (vw), 636 (ν
S-CO, m); MS (FAB+) m/z maximum calcd at 2142, found
2144. Anal. Calcd for C55H51S5O13Hg5Co: C, 30.84; H, 2.40.
Found: C, 31.12; H, 2.40.
Tetra(4-ethoxycarbonylphenyl)cyclobutadienepentakis(2-
methylthioethanethiomercurio)cyclopentadienylcobalt (9). 9
was synthesized from 2 (0.2 g, 0.086 mmol) in THF (5 mL)
and 2-methylthioethanethiol (0.056 g, 0.52 mmol) in THF (1
mL). Hexanes (50 mL) precipitated 9 (0.13 g, 66%), which
was recrystallized from DMSO/EtOH at -20 °C (dark-red
crystals): mp 130-135 °C (dec); 1H NMR δ 7.81 (d, 8 H, J )
8.4 Hz, Ph), 7.75 (d, 8 H, J ) 8.4 Hz, Ph), 4.30 (q, 8 H, J )
6.9 Hz, OCH2), 1.96 (s, 15 H, SCH3), 1.31 (t, 12 H, J ) 6.9
Hz, SCH3); 13C {1H} NMR δ 14.22 (CH3), 14.48 (CH3), 26.70
(CH2), 39.15 (CH2), 60.53 (OCH2), 73.10 (Cb), 127.08 (Cp),
127.72, 128.32, 129.33, 142.30 (Ph), 165.34 (COO); IR (KBr,
cm-1) 3050 (ν dCH, vw), 2977-2911 (ν -CH, w), 1713 (ν
CdO, s), 1602 (ν Ph, s), 1418-1405 (δ CH2, vw), 1384-1365
(δ CH3, m), 1272 (ν CO, s), 1175 (m), 1111 (s), 1017 (m), 859
(m), 782 (vw), 772 (vw), 756 (w), 725 (w), 703 (w); MS
(FAB+) m/z maximum calcd at 2302, found 2302. Anal. Calcd
for C60H71S10O8Hg5Co: C, 31.31; H, 3.11. Found: C, 31.41;
H, 3.14.
Tetra(4-ethoxycarbonylphenyl)cyclobutadienepentakis(2-
acetylthioethanethiomercurio)cyclopentadienylcobalt (10). 10
is synthesized from 2 (0.2 g, 0.086 mmol) in THF (5 mL) and
2-acetylthioethanethiol (0.08 g, 0.59 mmol) in THF (1 mL).
Hexanes (60 mL) precipitated 10 (0.11 g, 53%), which was
recrystallized from acetone (red crystals): mp 97 °C (dec); 1H
NMR δ 7.81 (d, 8 H, J ) 8.4 Hz, Ph), 7.74 (d, 8 H, J ) 8.4
Hz, Ph), 4.30 (q, 8 H, J ) 7.0 Hz, OCH2), 2.83 (m, 20 H, C2H4),
2.30 (s, 15 H, COCH3), 1.31 (t, 12 H, J ) 7.0 Hz, CH3); 13C
{1H} NMR δ 14.20 (CH3), 26.99, 30.63, 34.35, 60.52 (OCH2),
73.04 (Cb), 126.35 (Cp), 127.71, 128.28, 129.32, 142.38 (Ph),
165.32 (COO), 194.89 (SCO); IR (KBr, cm-1) 2978 (ν CH,
vw), 1713 (ν CdO, s), 1693 (ν CdO, s), 1602 (ν Ph, s), 1273
(ν CO, s), 1175 (w), 1112 (s), 1018 (w), 861 (vw), 782, (vw),
772 (vw), 756 (w), 725 (w), 703 (w) 626 (ν S-CO, m); MS
(FAB) m/z maximum calcd at 2443, found 2442. Anal. Calcd
for C65H71S10O13Hg5Co‚(CH3)2CO: C, 32.51; H, 3.05. Found:
C, 32.66; H, 3.10.
Tetraphenylcyclobutadienepentakis(2-methylthioethaneth-
iolmercurio)cyclopentadienylcobalt (5). 5 was synthesized
from 1 (0.2 g, 0.098 mmol) in THF (5 mL) and 2-methylthio-
ethanethiol (0.064 g. 0.58 mmol) in THF (1 mL). Hexanes (50
mL) were added to precipitate 5, which was then redissolved
in a minimum of THF, the solution was filtered through a filter
paper, and addition of hexanes (40 mL) precipitated 5 as a
yellow solid (0.119 g, 60%): mp 73-74 °C (dec); 1H NMR δ
7.66 (m, 8 H, Ph), 7.20 (m, 12 H, Ph), 3.04 (broad t, 10 H, J )
6.6 Hz, CH2), 2.63 (broad t, 10 H, J ) 6.6 Hz, CH2), 1.96 (s,
15 H, CH3); 13C {1H} NMR δ 14.46 (CH3), 26.76 (CH3), 39.14
(CH2), 73.42 (Cb), 126.20 (Cp), 126.38, 128.28, 128.35, 137.19
(Ph); IR (KBr, cm-1) 3054 (ν dCH, vw), 2910 (ν -CH, w),
1629-1596 (ν Ph, w), 1497 (m), 1428 (w), 1260 (w), 1197
(w), 1118 (w), 705 (m), 558 (m); MS (FAB+, NOBA) m/z
maximum calcd at 2014, found 2015. Anal. Calcd for
C48H55S10Hg5Co: C, 28.62; H, 2.75. Found: C, 28.88; H, 2.80.
Tetraphenylcyclobutadienepentakis(2-acetylthioethanethi-
omercurio)cyclopentadienyl-cobalt (6). 6 was synthesized
from 1 (0.2 g, 0.098 mmol) in THF (5 mL) and 2-acetylthio-
ethanethiol (0.09 g, 0.66 mmol) in THF (1 mL). Addition of
hexanes (50 mL) precipitated 6 (0.19 g, 90%): mp 65-70 °C
(dec); 1H NMR δ 7.63 (m, 8 H, Ph), 7.195 (m, 12 H, Ph), 2.85
(m, 20 H, C2H4), 2.31 (s, 15 H, CH3); 13C {1H} NMR δ 27.10
(CH2), 30.75 (CH2), 34.42 (CH3), 73.42 (Cb), 125.39 (Cp),
126.43, 128.25, 128.36, 137.31 (Ph), 195.00 (CO); IR (KBr,
cm-1) 3050 (ν dCH, vw), 2917 (ν -CH, vw), 1690 (ν CdO,
s), 1596 (ν Ph, w), 1497 (m), 1352 (w), 1262 (w), 1202 (w),
1132 (s), 961 (w), 700 (m), 625 (ν S-CO, m), 559 (w); MS
(FAB+, 3-NOBA) m/z maximum calcd at 2154, found 2154.
Anal. Calcd for C53H55S10O5Hg5Co: C, 29.55; H, 2.57.
Found: C, 29.40; H, 2.47.
Thiol Exchange. An attempt to remove the acetyl groups
from 6 with 0.2 N KOH in aqueous THF at room temperature
gave complicated mixtures, presumably because the liberated
thiol groups exchanged with those already attached to mercury.
Evidence for such exchange was obtained from reaction of 6
with 20 equiv of ethanethiol. After 5-10 min at room
temperature, the product was separated, and its NMR clearly
showed that some of the initially present tentacles had been
replaced by ethanethiol.
Tetra(4-ethoxycarbonylphenyl)cyclobutadienepentakis-
(ethylthiomercurio)cyclopentadienylcobalt (7). 7 was syn-
thesized from 2 (0.2 g, 0.086 mmol) in THF (5 mL) and
ethanethiol (0.032 g, 0.515 mmol). Hexanes (50 mL) precipi-
tated 7, which was redissolved in a minimum of THF, filtered
through a filter paper, precipitated with hexanes (50 mL) as a
yellow-orange powder (0.112 g, 63%) and recrystallized from
DMSO (dark-red crystals): mp 185-190 °C (dec); 1H NMR δ
7.82 (d, 8 H, J ) 9.0 Hz, Ph), 7.75 (d, 8 H, J ) 9.0 Hz, Ph),
4.30 (q, 8 H, J ) 6.9 Hz, OCH2), 2.69 (q, 10 H, J ) 7.2 Hz,
SCH2), 1.30 (t, 12 H, J ) 6.9 Hz, CH3), 1.12 (t, 15 H, J ) 7.2
Hz, CH3); 13C {1H} NMR δ 14.21 (CH3), 21.42 (CH3), 22.31
(SCH2), 60.51 (OCH2), 72.95 (Cb), 127.16 (Cp), 127.61, 128.28,
129.25, 142.45 (Ph), 165.36 (COO); IR (KBr, cm-1) 2975-
2863 (ν CH, w), 1716 (ν CdO, s), 1602 (ν CdC, s), 1446 (δ
CH2, w), 1384-1367 (δ CH3, m), 1273 (ν CO, s), 1175 (m),
1111.0 (s), 1017 (m), 859 (w), 782 (vw), 772, (w), 756 (w),
725 (w), 702 (vw); MS (FAB) m/z maximum calcd at 2072,
Results and Discussion
The CbCoCp core structure of the cobalt complexes 3-12 is
the same, and one might expect some underlying commonality
in their electrochemical oxidation-reduction processes. How-
ever, both the ethoxycarbonyl substituents introduced onto the
Cb phenyl groups in 7-10 and 12 and the various thiomercurio
substituents introduced in 3-10 will not only interact with the

Tentacled Tetragonal Star Connectors
J. Phys. Chem. B, Vol. 102, No. 49, 1998 10065
TABLE 1: Potentials of Voltammetric Peaks in THF vs
Normal Ag/AgC1 Electrodea
compd
A
B
C
D
E
11
12
3
4
5
6
7
8
9
-2.24
-1.696
-2.16, -2.48
-0.80b
-0.69
-1.14b
-0.88b
-1.02
-0.75
-1.10
-1.08b
-2.42c
-0.52
-0.40
-0.80
-0.40b
-0.77
-0.42
-0.67
-0.50
-(1.9-2.30)c
-2.05
-2.27
-1.72
-2.60
-2.50
d
-2.10, -2.42
-2.33, -2.40
-2.20, -2.54
-2.06, 2.66
-1.88c
-1.89
-1.80
d
d
10
a Ferrocene redox potential was +0.562 V vs this reference electrode.
b The highest of two or three closely spaced maxima. c Peaks B and C
are believed to overlap. d Peaks C and E are believed to overlap.
Figure 1. DC polarogram of 1 mM 11 and 0.1 M TBAPF6 in THF
(full points) at 2 s drop time: dashed line, blank solution; inset, log
plot analysis.
organocobalt core and modify its electrochemical behavior but
will also be electroactive on their own. A complicated set of
properties must be expected.
To acquire sufficient understanding of the oxidation-reduc-
tion properties to permit a future investigation of the adsorption
of these compounds on mercury, we have examined the
compounds by DC and AC polarography, cyclic voltammetry
(CV), UV-vis spectroelectrochemistry, and preparative elec-
trolysis. Experiments were performed in two solvents. Solubil-
ity in tetrahydrofuran (THF) is adequate for routine electro-
chemistry in millimolar concentrations. The low dielectric
permittivity of THF may suppress adsorption to some extent.
In acetonitrile, only 0.1 mM solutions could be prepared. This
solvent is likely to promote adsorption and might permit a
specification of the potential regions in which adsorbed layers
are formed. Presently, we deal only with the elucidation of
the fundamentals of the faradaic processes involved.
To facilitate the comparison of related faradaic processes
across the series, we denote the electron-transfer reactions by
the letters A, B, C, D and E. The last of these reactions
comprises two subsequent steps, E1 and E2. Roughly speaking,
we shall find that these processes correspond to an adsorption-
confined reduction (A), “metal-centered” reduction (B), reduc-
tive cleavage of tentacles (C), anodic reaction of released
tentacles (D), and ester group reduction (E), respectively, but
in many instances, these waves and processes cannot be
separated neatly and the labels are only approximate. Therefore,
the summary of electrochemical potentials of the processes A-E
in the 10 compounds given in Table 1 is only an approximate
guide.
Parent Complexes 11 and 12. The electrochemical char-
acterization of these two simplest structures is fairly complete.
The behavior of 11 is the simplest of all. Both in THF and
acetonitrile, DC polarography shows a reversible one-electron
reduction at -2.240 V (Figure 1). At 1 mM in THF and 2 s
drop time, the limiting current of the well-developed wave is
3.65 µA. Comparison with the wave of ferrocene confirms the
consumption of one electron and gives an estimate of the
diffusion coefficient, D ) 2.14 × 10-5 cm2‚s-1. The wave
shape is characterized by a 60 mV/decade slope of the log plot.
The corresponding voltammetric curve shows chemical and
electrochemical reversibility. The reduction process is ac-
companied by a substantial change of the UV-vis spectrum
observed on a transparent Pt electrode; the 272 nm absorption
band of 11 disappears in the fully reduced form, and two new
bands at 328 and 412 nm gradually intensify (Figure 2).
Phase-sensitive AC polarography (160 Hz) yields a single
reversible faradaic maximum at the half-wave potential, with
real and imaginary admittance components of equal height. In
Figure 2. UV-vis spectra of 3 mM 11 in 0.1 M TBAPF6 in THF
(thin-layer optically transparent cell, Pt mesh electrode) recorded during
a potential scan from -2.0 to -2.5 V: Ox, reactant; Red, fully reduced
form.
THF, the presence of 11 changes the double layer capacity Cd
only negligibly around 0 V (Cd ) Y′′/ωA), where Y′′ is the
imaginary component of the admittance vector Y, ω is the
angular frequency of the sine wave perturbation, and A is the
electrode area). A small hump at -0.65 V suggests a desorption
process that is evidenced also by a discontinuity in Figure 1. In
acetonitrile, a substantial lowering of Cd is observed at potentials
more positive than -0.5 V. The observed surface activity at
positive potentials is attributed to the presence of π-electron-
rich ligands on the metal that can interact with a positively
charged electrode surface.
We conclude that the observed process involves the uptake
of one electron and involves a formation of the radical anion
of 11. This is formally a 19-electron cobalt complex and in
principle, might have an 18-electron “slipped-ring” structure14
in which the Cp ring is only a 4-electron donor (diene-like, with
the fifth carbon acting as a free radical center) and the Cb ring
acts as a dianion and a 6-electron donor. Since none of our
observations demand that such a complication occurs, we
assume for the time being that the redox change does not change
the molecular geometry substantially.
Formally, the reduction corresponds to going from cobalt(I)
to cobalt(0). The degree to which the additional electron
actually ends up on the metal center is unclear. In the absence
of high-level calculations, we have attempted to obtain a rough
idea by performing extended Hu¨ckel calculations15 for several

10066 J. Phys. Chem. B, Vol. 102, No. 49, 1998
Brotin et al.
TABLE 2: Extended Hu1ckel LUMO Energies (eV) and
Distributions
compd
E (LUMO)
% on Coa % on Cbb % on Cpc
11
12
3
7
4
-9.41
23
12
3
12
0
62
78
13
59
0
16
10
84
-9.79
-9.77
-9.93
29
(-9.75)d -10.07
99e
99e
8
(-9.94)d -10.08
0
0
a Contribution of an electron in the LUMO to the total charge on
the cobalt atom, in the radical anion (averaged over the two nearly
degenerate LUMOs, except in 4 and 8). b Contribution of an electron
in the LUMO to the total charge on the cyclobutadiene ring and its
substituents, in the radical anion (averaged similarly). c Contribution
of an electron in the LUMO to the total charge on the cyclopentadienyl
ring and its substituents, in the radical anion (averaged similarly).
d Energy of the orbital that corresponds to the LUMO in 3, 7, 11, and
12, with dxz and dyz contributions on the Co atom. e All concentrated
on a single HgSCOMe substituent; there are altogether five ap-
proximately degenerate LUMOs spread over the five HgSCOMe
substituents.
Figure 3. Cyclic voltammograms of 1 mM 12 in 0.1 M TBAPF6 in
THF at three different switching potentials; scan rate ) 0.5 V/s.
members of the series at geometries derived from published2
X-ray structures of CbCoCp complexes and an unpublished5
X-ray structure of 8, assuming no Cp ring slipping14 in the
reduced form. The resulting energies of the approximately
degenerate lowest unoccupied molecular orbital (LUMO) and
their distribution in the molecule are listed in Table 2.
According to the structure of the LUMO in these simple
calculations, averaged over the members of the nearly degenerate
pair, in 11 the additional electron spends about a quarter of its
time on the cobalt atom, most of its time on the tetraphenylated
Cb ring, and only a little on the Cp ring. It is therefore likely
to be a considerable oversimplification to refer to the reduction
as “metal-centered”, but in the absence of a better name, we
shall adhere to this traditional label. The metal orbitals that
participate in the nearly degenerate LUMO are of the dxz and
dyz types, where z is the ring-center-metal-ring-center direc-
tion. These orbitals also participate heavily in the highest
occupied molecular orbital (HOMO), and the LUMO corre-
sponds to a metal-ring antibonding orbital.
The tetraester 12 differs from 11 by the presence of an
ethoxycarbonyl group in the para position of each phenyl
substituent on the Cb ring. The electron-withdrawing COOEt
substituents, which themselves are susceptible to reduction, have
the expected marked influence on the reduction potential of the
metal complex, and 12 yields a well-developed DC polaro-
graphic wave at -1.696 V, shifted by 0.540 V to more positive
potentials relative to 11. The limiting current is 2.27 µA,
slightly lower than that of 11, and the estimated diffusion
coefficient is D ) 8.3 × 10-6 cm2‚s-1. The log plot slope is
60 mV/decade and confirms the reversible nature of the
reduction. The LUMO energy of 12 obtained from the extended
Hu¨ckel calculation is 0.38 eV lower than that of 11, in
surprisingly good semiquantitative agreement with the observed
0.54 eV. According to Table 2, in the radical anion of 12, the
fractional density of the additional electron on the cobalt atom
has been reduced somewhat, that on the Cp ring virtually
vanished, and the electron now resides almost entirely on the
heavily substituted upper deck of the sandwich. This would
suggest that substituent shifts on the Cp ring will be less
pronounced in 12 than in 11.
Figure 4. Phase-sensitive AC polarogram of 1 mM 12 in 0.1 M
TBAPF6 in THF at 2 s drop time; 160 Hz, amplitude ) 5 mV (p-p).
Real (empty points, full blank curve) and imaginary (full points, dotted
blank curve) components of the admittance are shown.
to the reduction of the ester groups, based on the known
reduction of PhCOOCH3 at -2.11 V.16 When the switching
potential is in the range of ester reduction, the back-scan
voltammetric curve becomes rather complicated and the revers-
ible peak corresponding to the metal-centered oxidation is almost
hidden in overlapping peaks.
The adsorption of 12 from THF solutions is not pronounced.
Nevertheless, small humps, similar to those mentioned for 11,
appear at -0.9 V on the DC, AC, and CV curves. This may
indicate that the presence of the four ester groups extends the
adsorption toward more negative potentials by 0.25 V. The
comparison of Figure 3 with the phase-sensitive AC polarogram
(Figure 4) demonstrates how, at a suitably chosen frequency,
the AC technique differentiates irreversible from reversible
faradaic processes. The two admittance peaks at -2.05 and
-2.4 V (E1 and E2, respectively) are much smaller than the
reversible admittance peak B at -1.70 V, although their
voltammetric peak currents show just the opposite relation. This
feature will help us to locate a reversible reduction overlapped
by adjacent irreversible processes for other compounds examined
below.
Tentacled Complexes 3-10. The redox behavior of all
complexes that carry tentacles on the Cp ligand is much more
complicated and not nearly as well understood. All the
compounds exhibit certain common features, which have been
tentatively assigned to waves A-E in Figure 5 (Table 1).
Cathodic faradaic currents observed by DC polarography or
cyclic voltammetry occur at about -1 V (wave A), in the
The reversible reduction of 12 at -1.696 V is followed by
much larger irreversible polarographic and voltammetric currents
at more negative potentials (Figure 3). Irreversible peaks E1
and E2 at about -2.15 and -2.45 V, respectively, are assigned

Tentacled Tetragonal Star Connectors
J. Phys. Chem. B, Vol. 102, No. 49, 1998 10067
Wave A. This wave corresponds to reduction in the adsorbed
state (“adsorption current”). This follows from its nonlinear
dependence on the bulk concentration, linear dependence on
the scan rate, narrow and symmetrical shape at fast scans, and,
for concentrations in the micromolar range, from the dependence
of its peak height on the resting time at the initial potential.
The resulting reduction product remains strongly adsorbed at
potentials between waves A and B, as can be inferred from the
low value of the double layer capacity Cd (see below).
Given the only weakly negative potential at which wave A
appears, it appears likely that the chemical nature of the
reduction is the same as that in wave B of 11 and 12, i.e., a
metal-centered reduction that produces the radical anion of the
reactant. This attribution of wave A is supported by the results
of preparative electrolysis of 5 at a potential of -1.300 V. After
the passage of five electrons per molecule contained in the
solution, wave A disappeared and drawn-out anodic waves at
-0.4 and -0.1 V were observed. However, 90% of the starting
material was subsequently recovered with tentacles intact, and
only traces of unidentified products were isolated.
The variation in the position of peak A throughout the series
3-10 is undoubtedly related to some degree to the electronic
effects of the substituents, but differences in their adsorption
properties are likely to play a decisive role. Peak A usually
has fine structure in THF and exhibits shoulders or additional
CV maxima mainly at the slower scan rates and especially in
compounds that contain two sulfur atoms in each of their
tentacles. This is most likely related to a time evolution of the
adsorbed structure and will be more thoroughly discussed in a
future article.5
Figure 5. Cyclic voltammograms of 3-12 in 0.1 M TBAPF6 in THF
at 0.5 V/s scan rate. Concentration: 4-7, 11, 12, 1 mM; 8, 10, 0.53
mM; 9, 0.43 mM; 3, saturated solution, ∼0.04 mM. The 50 µA bar
applies to all compounds except for 3, where the 5 µA bar applies.
Wave B. Solely on the basis of the general similarity of the
position of wave B to those found in 11 and 12, including a
substantial shift to less negative potentials in complexes 7-10
that carry ethoxycarbonyl groups at the phenyls, it is likely that
the reduction process in the other compounds is the same as
that in 11 and 12. The shift due to the ester groups is smaller
than that in the parent compounds with unsubstituted phenyl
groups, as expected from the above argument based on extended
Hu¨ckel calculations. Quantitative comparisons are impossible,
both because wave B is doubled in some instances and because
it is irreversible and its position is not simply tied to equilibrium
properties alone.
In marked contrast to the behavior of 11 and 12, the reduction
process at wave B of 3-10 has irreversible character. The
waveform is drawn out, the anodic counterpart expected upon
scan reversal is completely missing, and CV measurements in
THF performed at fast scan rates and/or with different scan
switching potentials failed to detect any corresponding anodic
reoxidation peak. Spectroelectrochemistry yielded no indication
that a stable radical anion is formed, and the UV-vis spectrum
resembles that of the reactant at all times (Figure 6). If the
reduction process is indeed the same as that in 11 and 12, metal-
centered reduction, we must conclude that the radical anions
of the tentacled compounds 3-10 are unstable. A likely
decomposition path is a loss of one of the tentacles as a thiolate
anion in a homogeneous process that could be as simple as a
unimolecular dissociation but could also be quite complex.
Either way, this would ultimately regenerate the chromophore
of the original reactant and account for the spectral results.
Figure 6. Cyclic voltammetry of 1 mM 5 in 0.1 M TBAPF6 in THF
at 0.5 V/s scan rate. Inset shows irreversible changes in the UV-vis
spectrum recorded in a spectroelectrochemical cell during the voltage
scan of peaks B and C.
potential range of -1.7 to -2.0 V (wave B), and at potentials
more negative than -2.2 V (a large wave C or E). A distinct
anodic peak (D) common to all tentacled complexes is located
in the range -0.4 to -0.8 V.
The behavior is illustrated in results for a typical example, 5
(Figure 6). Irreversible reduction is accompanied by minor
changes in the UV-vis spectrum (inset in Figure 6) that are
irreversible as well. Complexes with ethoxycarbonyl substit-
uents on the phenyl groups (7-10, 12) have an additional
complex pattern of anodic peaks, usually dominated by a
symmetric peak near -1 V. Since this peak is present in 12
(Figure 3), it is not specifically associated with the presence of
tentacles on the Cp ring. Most of these anodic peaks are
undoubtedly due to the reduction products of the ester groups.
These are of limited interest in this study and we do not attempt
to analyze them.
In agreement with the assignment of the polarographic wave
B to the same metal-centered reduction in all compounds, its
height corresponds to a one-electron reduction in 5 and 6. In
4 and 8, a much larger current starts already at somewhat less
negative potentials, and DC polarographic limiting currents are

10068 J. Phys. Chem. B, Vol. 102, No. 49, 1998
Brotin et al.
CbCoC5(Hg-SX)5 + 10 e- + 5 H+ f
approximately 5 times higher. It seems clear that wave C is
greatly shifted to less negative potentials and covers up wave
B in both of these two cases. The rather facile heterogeneous
reduction of the -HgSCOMe moiety is in accord with expecta-
tions based on literature data for structurally related com-
pounds.17-19 Waves B and C also overlap in 3 but at much
more negative potentials. In the remaining compounds, 7, 9,
and 10, wave B is not covered up by C, but its polarographic
limiting current could not be measured reliably.
CbCoC5H5 + 5 Hg + 5 XS-
This is analogous to the previously investigated17 reduction of
(C6H5S)2Hg. After a charge corresponding to 10 electrons per
solute molecule is passed, preparative electrolysis of 5 at -2.300
V caused the disappearance of the polarographic wave C.
Separation of a product from the supporting electrolyte yielded
a sample that was identical with 11 by comparison of NMR
spectra.
The position of wave C is most clearly defined in 3-6, since
the presence of ethoxycarbonyl groups in 7-10 causes the
appearance of the more complicated overlapping wave or waves
E. This also prevents the determination of the number of
electrons involved in wave C. Where the position of wave C
can be determined, it seems quite constant as would be expected,
with the exception of 4 and 8, already discussed above.
Wave D. This anodic wave is common to all tentacle-
carrying complexes but is observed only when the voltammetric
switching potential is chosen in the region of wave C, and we
assign it to the oxidation of the thiol groups of the free tentacles
generated in wave C. This interpretation is supported by the
occurrence of peak D also in complexes 3-6, which carry no
ester groups; hence, it certainly is not due to their reduction
products. The oxidation peak D, though not as sharp, is
observed on a gold microelectrode as well.
Waves E. The reduction of the ester groups encountered in
7-10 resembles their reduction in 12 and requires no further
comment, as it lies outside our present range of interest.
Waves B and C: A Closer Look. The striking irreversibility
of the CV wave B in 3-10 on stationary electrodes contrasts
with the perfectly reversible one-electron metal-centered reduc-
tion of 11 and 12 and calls for an explanation. The results of
phase-sensitive AC polarography and impedance spectroscopy
certainly support the view that the electron transfer at wave B
is inherently fast.
In acetonitrile, 5 is about 10 times less soluble than in THF.
Acetonitrile enhances adsorption effects, but the behavior in
THF is similar. In a CV taken in acetonitrile (Figure 7, top),
peak A is higher and sharper, peaks B and D are shifted toward
less negative potentials, and peak B is much smaller than C.
Wave B is much clearer in AC polarography (Figure 7, bottom).
The imaginary and the real AC admittance peaks B at -1.78 V
are of equal height and are higher than the AC peaks C at -2.25
V. The imaginary component of peak C is smaller than its real
counterpart. According to the theory of the faradaic admittance,
equal admittance peaks are expected for a reversible case,
whereas for a slow redox process, the imaginary part is smaller
than the real part; hence, the irreversibility of wave B observed
by CV is due to an EC mechanism and not to a slow electron
transfer. Figure 7 also shows that the adsorption of 5 is not a
simple process, and four potential regions, marked I-IV, are
identified.5
The effect of the substitution with five -HgSX groups on
the ease of the reduction of 11 and 12 shows no clear-cut pattern,
and shifts by up to 0.2 V in either direction are observed. This
apparent absence of systematic behavior may be due to poor
guesses concerning the assignment of wave B in some cases
and the irreversible nature of the processes involved in others.
In general, it seems that the effect of the tentacles on the
reduction of 11 is small and that they make the reduction of 12
a little harder.
It is not obvious a priori what shifts of the half-wave potential
should be expected when the five -HgSX substituents are
introduced onto the Cp ring. The electropositive nature of the
metal atom and the facility of thiol exchange suggest that
resonance structures of the type -Hg+-SX are important and
that the Hg atoms carry considerable positive charge and cause
inductive electron withdrawal. In MO terms, the vacant σH/ gS
orbital has a large amplitude on the Hg atom and acts as an
electron density acceptor. Indeed, the inductive substituent
constants for this general type of group with an electronegative
group X suggest an electron-withdrawing power comparable
to that of the ethoxycarbonyl group.20 Results obtained with
the extended Hu¨ckel method suggest that the five -HgSMe
substituents do not change the nature of the nearly degenerate
LUMO, with its characteristic dxz, dyz metal contribution, and
stabilize it by 0.36 eV in 11 and 0.15 eV in 12 (Table 2). The
difference between the two corresponds to the above qualitative
expectations, but the former magnitude appears excessive. The
-HgSCOMe substituent is computed to be just as strongly
electron withdrawing. Pentasubstitution in 11 to yield 4
stabilizes the metal-centered (dxz, dyz) orbital pair by 0.35 eV,
while pentasubstitution in 12 to yield 8 stabilizes the analogous
orbitals by 0.15 eV. The similarity of the two substituent effects
is not surprising if the effect is primarily due to the polarity of
the HgS bond.
The calculations for 4 and 8 are in agreement with the above
conclusion that in these two compounds wave C precedes and
covers up wave B. According to Table 2, the lowest unoccupied
orbitals of the five -HgSCOMe substituents lie well below the
metal-centered (dxz, dyz) orbital pair, and they now constitute
the several LUMOs of 4 and 8, at virtually the same energy in
both compounds. This is best seen in the distribution of the
additional density upon occupying the LUMO, which is
calculated to reside exclusively on a single tentacle in both 4
and 8. It is calculated to lie below the metal-centered orbital
by a full 0.32 eV in 4 and by 0.14 eV in 8, where the metal-
centered orbital is stabilized by the four ester groups. This then
not only produces a nominal stabilization of the LUMO by 0.66
V upon going from 11 to 4 and by 0.29 V upon going from 12
to 8 but also changes its nature altogether.
The admittance data observed in single-frequency AC experi-
ments were further confirmed by inspection of frequency spectra
of electrode impedance (Z) or admittance (Y ) 1/Z). Peak B is
characterized by a frequency spectrum typical of fast electron
transfer coupled with the adsorption of reactants.21 The coupling
of redox and adsorption processes is most conveniently dis-
played as the complex capacitance plot of Y′′/ω vs Y′/ω, where
Y′′ and Y′ are the imaginary and the real component of the
electrode admittance, respectively (Figure 8). The coupling of
the reduction of the metal center with adsorption of the complex
yields a capacitance plot composed of part of a quarter circle
Wave C. This also represents an irreversible heterogeneous
process. In 5, the polarographic limiting current corresponds
to the overall uptake of 10 ( 1 electrons, suggesting that the
process is a complete cleavage of all five tentacles.

Tentacled Tetragonal Star Connectors
J. Phys. Chem. B, Vol. 102, No. 49, 1998 10069
Figure 9. Complex impedance plot for the solution of Figure 7 at DC
potentials in the vicinity of peak C: (9) -2.140 V, ([) -2.160 V,
(4) -2.180 V, (O) -2.200 V, (2) -2.220 V, (×) -2.240 V, (b)
-2.260 V, and (3) -2.280 V.
Figure 7. Cyclic voltammogram (top, scan rate 0.5 V/s) and phase-
sensitive AC polarogram [bottom, 160 Hz; amplitude ) 5 mV (p-p)]
of 0.1 mM 5 in 0.1 M TBAPF6 in acetonitrile.
Figure 10. Faradaic phase angle as a function of frequency for the
solution of Figure 7 at the two DC potentials shown, close to the half-
wave potential (-2.20 V). Corrected for solution resistance (308.9 Ω)
and double layer capacitance (4.66 µF‚cm-2). For procedure validation,
cf. the reversible behavior of 5 mM ferrocene ([) in the same medium
at its half-wave potential.
frequency dependence of the faradaic peak C from Figure 7 is
shown in Figure 9 as the complex impedance plot (Z′′ vs Z′,
where Z′′ and Z′ are the imaginary and real components of the
electrode impedance, respectively). The semicircle shape is
different than that observed for peak B and indicates a slow
kinetic process limited by diffusion at the lowest frequencies
and at the most negative potentials.
The standard treatment23 of data for peaks B and C involves
the evaluation of the imaginary and real component of the
faradaic impedance (Z′′and Z′, respectively) by performing a
correction for solution resistance (in our cell, typically about
40 Ω in acetonitrile and 330 Ω in THF) and subtraction of the
interfacial capacitance ωCd (where Cd is estimated from the
high-frequency limit of the complex capacitance plot). The
resulting frequency dependence of the faradaic phase angle φ
Figure 8. Complex capacitance plot for the solution of Figure 7 at
the DC potential of peak B, -1.750 V (corrected for solution resistance,
308.9 Ω, obtained from the high-frequency limit of the complex
impedance plot).
f
f
and the low-frequency straight line with slope ) 1. A simple
redox exchange should yield a straight line, either horizontal
or tilted at 45°. Pure adsorption/desorption phenomena in the
absence of electron transfer would yield only a quarter circle
or a semicircle.
The sensitivity with which adsorption effects coupled to the
electron-transfer reaction are detected depends on the value of
the heterogeneous rate constant k0. Our previous simulations22
have shown that frequency dependence such as that in Figure
8 is obtained only for relatively fast redox reactions. The
(defined as cot φ ) Z′f′/Z′ and shown in Figure 10) permits the
f
estimation of the heterogeneous rate constant from the slope of
the high-frequency linear asymptote and yields the value k0 )
1.5 × 10-4 cm‚s-1. The cot plot also indicates the presence of
a coupled homogeneous reaction as a low-frequency linear
asymptote. This is believed to be the cleavage of tentacles, as
confirmed above by preparative electrolysis, and all of the data
for wave C thus are in agreement with the proposed interpreta-

10070 J. Phys. Chem. B, Vol. 102, No. 49, 1998
Brotin et al.
tion. Similar large differences in the observed AC reversibility
of peaks B and C are observed for other compounds as well.
References and Notes
(1) Michl, J. In Applications of Organometallic Chemistry in the
Preparation and Processing of AdVanced Materials; Harrod, J. F.; Laine,
R. M., Eds.; Kluwer: Dordrecht, The Netherlands, 1995; p 243. Harrison,
R. M.; Magnera, T. F.; Vacek, J.; Michl, J. In Modular Chemistry; Michl,
J., Ed.; Kluwer: Dordrecht, The Netherlands, 1997; p 1.
Conclusions
The parent tetraphenyl-substituted CbCoCp complex 11
undergoes a reversible one-electron reduction at -2.24 V vs
Ag/AgCl to yield a radical anion with a characteristic UV-vis
spectrum. According to extended Hu¨ckel calculations, the
additional electron resides in a nearly degenerate metal-ring
antibonding orbital, primarily located on the Cb-containing deck
(“metal-centered reduction”). Introduction of four ester groups
into the para positions of the phenyl substituents on this deck
to yield 12 facilitates the reduction by 0.54 V, and the
calculations suggest that now the additional electron favors the
Cb deck even much more strongly. In addition, at more negative
potentials, new reduction waves appear, undoubtedly related to
the reduction of the ester groups but not investigated further.
The electrochemical behavior of the derivatives 3-10, which
carry five -HgSX tentacles on the Cp ring, is far more
complicated. It is characterized by the appearance of an
adsorption pre-wave at much less negative potentials, in which
the metal-centered reduction takes place and which will be the
subject of a separate study. The diffusion-controlled metal-
centered reduction still takes place at similar potentials as before,
but is now irreversible, and no radical anion spectra have been
observed. Presumably, the radical anions of the tentacled
species are unstable with respect to the loss of a tentacle in a
homogeneous solution process. When the tentacle is -Hg-
SCOMe, its heterogeneous reductive cleavage is so facile that
it takes place before the metal-centered reduction can, but the
cleavage of tentacles of other structures is observed only at more
negative potentials. The cleaved tentacle fragments yield an
anodic peak common to all tentacled complexes. Reduction of
ester groups is observed at very negative potentials, as in 12.
(2) Harrison, R. M.; Brotin, T.; Noll, B. C.; Michl, J. Organometallics
1997, 16, 3401.
(3) Scho¨berl, U.; Magnera, T. F.; Harrison, R.; Fleischer, F.; Pflug, J.
L.; Schwab, P. F. H.; Meng, X.; Lipiak, D.; Noll, B. C.; Allured, V. S.;
Rudalevige, T.; Lee, S.; Michl, J. J. Am. Chem. Soc. 1997, 119, 3907.
(4) Posp´ısˇil, L.; Heyrovsky´, M.; Pecka, J.; Michl, J. Langmuir 1997,
13, 6294.
(5) Posp´ısˇil, L.; Fiedler, J.; Brotin, T.; Noll, B. C.; Michl, J.
Unpublished results.
(6) Rosenblum, M.; North, B.; Wells, D.; Giering, W. P. J. Am. Chem.
Soc. 1972, 94, 1239.
(7) Gleiter, R.; Ro¨ckel. H.; Pfla¨sterer, G.; Treptow, B.; Kratz, D.
Tetrahedron Lett. 1993, 50, 8075.
(8) Schimanke, H.; Gleiter, R. Organometallics 1998, 17, 275.
(9) Herberich, G. E.; Klein, W.; Ko¨lle, U.; Spiliotis, D. Chem. Ber.
1992, 125, 1589.
(10) Meade, E. M.; Woodward, F. N. J. Chem. Soc. 1948, 1895.
(11) Rausch, M. D.; Genetti, R. A. J. Org. Chem. 1970, 35, 3888.
(12) Petukhova, N. P.; Prilezhaeva, E. N. Zh. Org. Khim. 1966, 11, 1947.
(13) Okuyama, T. Tetrahedron Lett. 1982, 23, 2665. Okuyama, T. J.
Am. Chem. Soc. 1984, 106, 7134.
(14) Wakatsuki, Y.; Aoki, K.; Yamazaki, H. J. Chem. Soc., Dalton Trans.
1982, 1, 89. Richards, T. C.; Geiger, W. E. J. Am. Chem. Soc. 1994, 116,
2028.
(15) Parameters for cobalt, see: Saillard, J.-Y.; Hoffmann, R. J. Am.
Chem. Soc. 1984, 106, 2006. For mercury, see: Underwood, D. J.; Hofmann,
R.; Tatsumi, K.; Nakamura, A.; Yamamoto, Y. J. Am. Chem. Soc. 1985,
107, 5968.
(16) Baizer, M. M., Lund, H., Eds. Organic Electrochemistry: An
Introduction and a Guide, 3rd ed.; M. Dekker: New York, 1991; p 419.
(17) Person, B.; Nygård, B. J. Electroanal. Chem. 1974, 56, 373.
(18) Bond, A. M.; Cassey, A. T.; Thackeray, J. R. J. Electroanal. Chem.
1973, 48, 71.
(19) Toropova, V. F.; Budnikov, R. G. K.; Ulakhovich, N. A. Talanta
1975, 25, 263.
(20) Exner, O. In Correlation Analysis in Chemistry; Chapman, N. B.,
Shorter, J., Eds.; Plenum Press: New York, 1978; p 439.
(21) Timmer, B.; Sluyters-Rehbach, M.; Sluyters, J. H. J. Electroanal.
Chem. 1967, 15, 343. De Levie, R.; Posp´ısˇil, L. J. Electroanal. Chem. 1969,
22, 277.
(22) Posp´ısˇil, L. J. Electroanal. Chem. 1976, 74, 369.
(23) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals
and Applications; Wiley: New York, 1980; p 338.
Acknowledgment. This work was supported by the Depart-
ment of Energy (FG03-94ER12141), the National Science
Foundation (CTS-9727165), the NATO Scientific and Envi-
ronmental Division (SRG 951530), and the Grant Agency of
the Czech Republic (203/97/1048). T.B. is grateful for a NATO
fellowship.