Solvent and electrolyte effects in enhancing the identification of intramolecular electronic communication in a multi redox-active diruthenium tetraferrocenoate complex, a triple-sandwiched dicadmium phthalocyanine and a ruthenocene-containing β-diketone

Article abstract of DOI:10.1016/j.ica.2010.03.031

Enhanced electrochemical resolution of anodic processes is possible in the presence of [N(ⁿBu)₄][B(C₆F₅) ₄], 1, as supporting electrolyte over that obtained in the presence of [N(ⁿBu)₄

Full text of DOI:10.1016/j.ica.2010.03.031

Inorganica Chimica Acta 363 (2010) 2222–2232
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Solvent and electrolyte effects in enhancing the identification of intramolecular
electronic communication in a multi redox-active diruthenium tetraferrocenoate
complex, a triple-sandwiched dicadmium phthalocyanine and a ruthenocene-
containing b-diketone
Henno J. Gericke , Nicola I. Barnard , Elizabeth Erasmus , Jannie C. Swarts , Michael J. Cook b,
a
a
a
a,*
c
Manuel A.S. Aquino
a
Department of Chemistry, University of the Free State, Bloemfontein 9300, South Africa
School of Chemical and Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK
St. Francis Xavier University, Antigonish, Nova Scotia, Canada B2G 2W5
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Enhanced electrochemical resolution of anodic processes is possible in the presence of
Received 14 December 2009
Received in revised form 3 March 2010
Accepted 11 March 2010
n
n
[
N( Bu)
4
][B(C
6
F
5
)
4
], 1, as supporting electrolyte over that obtained in the presence of [N( Bu)
4
][PF
6
]. By
ꢀ
changing the anion of the supporting electrolyte to a salt having [B(C
cesses of especially cationic analytes can benefit. Thus, the redox chemistry of 0.5 mmol dm solutions
of [Ru
involve four well-resolved ferrocenyl-based electrochemical reversible redox processes as well as reduc-
4 6
tion of Ru –Ru . At 1.0 mmol dm concentrations of 2, or in the presence of [N( Bu) ][PF ], the four ferr-
6
F
)
5 4
] , anions, electrochemical pro-
ꢀ3
Available online 16 March 2010
n
2
(
l-FcCOO)
4
ꢁ(CH
3 2 2 6 2 2 4 6 5 4
CH OH) ][PF ], 2, Fc = ferrocenyl, in CH Cl /[N( Bu) ][B(C F ) ] were found to
Dedicated to Professor W. E. (Bill) Geiger of
Vermont for his contributions to the field of
electrochemistry.
III
II
ꢀ3
n
+
ꢀ
ocenyl processes coalesced into only two waves as a result of (Fc )ꢁ ꢁ ꢁ(PF ) ion paring. Seventeen of the
6
possible 18 one-electron transfer processes of the biscadmium trisphthalocyaninato complex
n
Keywords:
Ferrocene
Ruthenocene
b-Diketone
Cadmium
[Cd {Pc(C H ) } ], 3, could be observed in THF/[N( Bu) ][B(C F ) ], but the electrochemical window of
CH
2
6
13 8
3
4
6 5 4
n
2
Cl
2
/[N( Bu)
4
6 5 4
][B(C F ) ] only allowed detection of 15 of these processes. Although reduction processes
n+
were unaffected, THF solvation leading to species such as (3 )(THF) with 1 6 n 6 4 and x P 1 as well as
x
n+
ꢀ
ion pair formation of the type (3 )ꢁ ꢁ ꢁ(PF6 ) prevented good resolution of oxidation processes. The
n
CH
(
OCH
2
Cl
2
+
/[N( Bu)
4
][B(C
6
F )
5 4
] system also allowed detection of reversible one-electron transfer ferrocenyl
Phthalocyanine
Electrochemistry
+
Fc/Fc ) and ruthenocenyl-based (Rc/Rc ) processes for both enol and keto isomers of the b-diketone FcC-
n
2
CORc, 4, Rc = ruthenocenyl. In CH
3
CN/[N( Bu)
4
][PF
6
], the ruthenocenyl moiety was oxidised to a
IV
Ru species.
Ó 2010 Elsevier B.V. All rights reserved.
1
. Introduction
Cyclic voltammetry (CV) is possibly the simplest and most ver-
voltammetry, better resolution is possible in cases where CV data
does not clearly distinguish between poorly resolved (i.e. near-
overlapping) multiple redox waves [3]. It is important to recognise,
though, that the electrode potentials of closely spaced redox events
are notoriously difficult to assess using CV methods. Taube dis-
cussed this in detail [3a]. In linear sweep voltammetry (LSV), the
dc potential applied to the cell increases linearly very slowly as a
satile electroanalytical technique for the study of electroactive spe-
cies. The effectiveness of CV having a triangular voltage input
signal (i.e. a triangular voltage vs. time plot, see Supplementary
material) is its ability to probe the redox behaviour of an electroac-
tive species at different scan rates (though much faster scan rates
are possible, convenient scan rates are between 50 and
ꢀ
1
function of time (2 mV s maximum compared to 50 mV or faster
for CV, depending on active electrode area). From an analytical
point of view the LSV type of voltammogram is advantageous be-
cause it can be used to accurately determine the relative number
of electrons that flow in a particular electrochemical process com-
pared to other electrochemical processes that take place within the
same molecule (i.e. the equivalent of integration in NMR) [3].
ꢀ
1
5
00 mV s ) over a wide potential range using simple equipment
[
1]. Osteryoung square wave voltammetry (OYSW) utilises a
pulsed voltage input signal in step form [2]. Currents are measured
at various times during the lifetime of these pulses. From OYSW
0
o
The formal reduction potential, E , and a measure of electro-
chemical reversibility or the amount of electrons (n) that flows
per molecule during any single redox process can conveniently
*
Corresponding author. Tel.: +27 (0)51 4012781; fax: +27 (0)51 4446384.
E-mail address: swartsjc@ufs.ac.za (J.C. Swarts).
0
020-1693/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2010.03.031

H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2223
be obtained from Eqs. (1)–(4). By electrochemical reversibility it is
meant that the rate of electron transfer between the electrode and
substrate is fast enough to maintain the concentration of the oxi-
dised and reduced species in equilibrium, according to the Nernst
equation, at the electrode surface at the particular scan rate. In
contrast, chemical reversibility implies a compound can both be
quantitatively oxidised, and the product can again be quantita-
tively reduced back to the original material, but there is no limita-
tion on how fast this must happen. An electrochemically reversible
couple which is not followed by any chemical reaction should obey
Eq. (3) (or its reciprocal – the current ratio of Eq. (3) should be ta-
ken as the current of the reverse scan divided by the current of the
forward scan). An electrochemical process that obeys Eq. (3) must
inevitably also be chemical reversible, but one that does not obey
Eq. (3) can still be chemical reversible, even though it is not any
more electrochemically reversible. From the Randles–Sevcik equa-
tion (Eq. (4) at 25 °C, [1,4]), peak current measurements of electro-
chemically reversible systems are related to diffusion coefficients,
D. If the relationship between i and the square root of scan rate
in V s ) is linear, then the electroactive species is in solution
rather than surface bound. The diagnostic E = 59 mV for electro-
chemical reversible one-electron transfer processes are often diffi-
cult to achieve without instrument compensation for cell
p
ꢀ1
(v
D
resistance and over potentials. Thus, frequently
DE values of up
to 90 mV for perfectly electrochemically reversible processes are
1/2
obtained. In these cases, a linear relationship between i
is more convenient in demonstrating electrochemical reversibility
than E values [5].
p
and v
D
Eo0 ¼ ðEpa þ EpcÞ=2
ð1Þ
ð2Þ
ð3Þ
D
E
p
¼ Epa ꢀ Epc ¼ 59=n mV
i
i
pc=ipa ¼ 1 ðdenominator is always from the “forward”scanÞ
5
3=2
1=2 1=2
p
¼ ð2:69 ꢂ 10 Þn AD
m
C
ð4Þ
n
Fig. 1. Structure of the electrolyte [N( Bu)
4
][B(C
6
F
5
)
4
], 1, as well as complexes 2–4. Green balls = N from the structure of 3 which has been reported elsewhere [15a]. The
groups dissipate the charges on the N and B atoms.
6 5
arrows in 1 indicate how the electron-donating effect of the butyl groups and the electron withdrawing effect of the C F

2
224
H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
It is also important to note that Eo’ should not be affected by dif-
before did not describe the advantage of [N( Bu)
[N( Bu) ][PF
4 6
n
][B(C
] as supporting electrolyte [13]. The isomerisation
kinetics of 4 has recently been studied by us in comparison with
other b-diketones of the type Rc–CO–CH –CO–R by spectroscopic
4
6 5 4
F ) ] over
n
ferences in scan rate in electrochemical reversible processes (it
should remain constant). Accurate results from Eqs. (1)–(4) rely
on measurements of CV data that are independent of analyte-sol-
vent and analyte-supporting electrolyte interactions. Frequently
poor experimental conditions or cell imperfections such as too lit-
tle supporting electrolyte, solvent and/or supporting electrolyte
association with analytes, low solvent dielectric constants, junc-
tion potentials, imperfect or dirty electrode surfaces and electrode
over-potentials distort peak potential and peak current measure-
ments and lead to inaccurate determinations of many of the
above-mentioned data and conclusions that can be drawn from
them, including large scatterings in measured diffusion coefficients
2
and electrochemical means [14]. However, within the context of
this publication, 4 serves especially well to demonstrate how the
] results in the generation of
a Ru product in an electrochemical irreversible process, while
in CH Cl /[N( Bu) ][B(C F ) ], a reversible Ru /Ru couple fol-
2 2 4 6 5 4
lowed by a chemical dimerisation process involving Ru centres
is observed.
n
use of both CH
3
CN and [N( Bu)
4
][PF
6
IV
n
III
II
III
2. Experimental
[
6]. Consequently the choice of solvent, supporting electrolyte and
reference electrodes are crucial for reliable results.
2.1. Materials and apparatus
Traditionally, for non-aqueous experimentation, DMSO or
n
ꢀ
CH
supporting electrolytes. However, both CH
quently co-ordinate to metal complexes to form totally new spe-
3
CN and [N( Bu)
4
][PF
6
] or [ClO
4
]
are favourite solvents and
Lithium tetrakis[pentafluorophenyl]borate etherate (Boulder
Scientific Corporation, USA), ruthenocene, ferrocene (Strem), alu-
minium trichloride and other solid reagents (Aldrich) were used
without further purification. Dichloromethane and acetonitrile
were dried by refluxing under nitrogen over calcium hydride, tet-
rahydrofuran was dried over Na wire and redistilled just prior to
use. Water was double distilled. Complexes 2 [12] and 3 [15] were
3
CN and DMSO fre-
ꢀ
ꢀ
cies [7], and the PF6 or ClO4 anions frequently form ion-pairs
with oxidised species. Due to the explosive nature of the ClO4
salts, use of perchlorate-containing supporting electrolytes is
strongly discouraged. Ion pair formation, especially with positively
charged oxidised species, can be minimised if the negative charge
of the anion of the supporting electrolyte can be spread over a large
ꢀ
1
synthesised as described before. H NMR spectra at 20 °C were re-
corded on a Bruker Avance DPX 300 NMR spectrometer at 300 MHz
n
þ
volume, just as the positive charge on the N atom of the Nð BuÞ
with chemical shifts presented as d values referenced to SiMe
0.00 ppm utilising acid free CDCl as solvent. Melting points were
determined with a Reichert Thermopan microscope with a Kofler
hot-stage and are uncorrected. CDCl was made acid free by pass-
4
at
4
cation are inductively minimised by the four electron-donating bu-
3
tyl groups. Towards supporting electrolytes with limited ion pair-
ing capabilities, the electrolyte [N( Bu)
n
4
6
][B(C F
5
)
4
], 1, Fig. 1, and
3
n
[
N( Bu)
4
][B{C
6
H
3
(CF
3
)
2
}
4
] has been developed by Geiger and Mann,
ing it through basic alumina immediately before use. Elemental
analysis was conducted by the Canadian Microanalytical Service,
Ltd.
ꢀ
respectively, with very good effect [8]. The [B(C
approximate spherical diameter of 10 Å compared to PF6 ’s diam-
6
F
5
)
4
]
anion has an
ꢀ
eter of ca. 3.3 Å and therefore will have a lower charge density than
ꢀ
PF6
.
2.2. Syntheses
2 2
To minimise solvent–analyte interactions, CH Cl is gaining fa-
vour over CH
larger potential window than CH
3
CN as solvent. However, if a need for solvent with a
Cl exists to study analytes under
2.2.1. Synthesis of 1
2
2
An adaption of a published procedure [16] was followed. Sol-
vents used in this synthesis were of the best quality available. Lith-
ium tetrakis[pentafluorophenyl]borate etherate (25 g, 0.046 mol if
one ignores the contribution of the unknown amount of ether to
the relative molar mass) was dissolved in 20 ml methanol. Tetra-
butylammonium bromide (12.75 g, 0.039 mol), dissolved in 10 ml
methanol, was added drop wise at room temperature over
15 min to the lithium solution. An off-white precipitate formed
during this time. The septum-sealed suspension was then placed
in a refrigerator at 0 °C for 30 min, after which it was transferred
to a freezer at ꢀ25 °C where it was stored overnight. The precipi-
tate that formed was removed by filtration from the brown liquid
and washed with 10 ml cold (-25 °C) methanol and air dried before
redissolving in distilled CH Cl (30 ml). MgSO4 was added to fur-
strong reducing conditions (i.e. more negative potentials), THF may
prove to be a beneficial solvent. While THF is nucleophilic and
would interact with electrophilic or positively charged species un-
der oxidising conditions, it can be considered non-interactive to-
wards negatively charged (reduced) species.
Because measured potentials vs. a convenient reference elec-
+
ꢀ3
trode such as Ag/Ag (as a silver wire immersed in a 0.01 mol dm
AgNO solution), or a pseudo reference electrode such as a Ag of Cu
3
wire drift from solvent to solvent or sometimes even from experi-
ment to experiment as conditions such as temperature, cell geom-
etry or concentrations of analytes and supporting electrolytes
differ, IUPAC has issued a directive [9] that all electrochemical data
be reported vs. an internal standard as first suggested by Gagne
et al. [10]. A convenient internal standard is ferrocene or deca-
methylferrocene [11].
2
2
ther dry the solution while stirring for 2 h at room temperature.
The MgSO was filtered off, washed with CH Cl and the combined
4
2
2
In this communication we practically apply and demonstrate all
of the above fundamental aspects by discussing the well-resolved
enhanced voltammograms that can be obtained from a voltammet-
2 2
CH Cl fractions were evaporated to liberate crude 1 as a white so-
lid (21.7 g, 0.024 mol, 60%). Further purification by recrystalliza-
tion was achieved as follows: to a solution of 1 (9 g, 0.01 mol) in
11 ml CH Cl was added 55 ml ether drop wise, while stirring, over
ric study of [Ru
trisphthalocyaninato complex 3, and the mixed-metallocene b-
diketone Fc–CO–CH –CO–Rc, 4, where Fc = ferrocenyl, Rc = ruthen-
ocenyl, in CH Cl /[N( Bu)
obtained in the solvents THF, CH
2
(l
-FcCOO)
4
3 2 2 6
ꢁ(CH CH OH) ][PF ], 2, the biscadmium
2
2
20 min at room temperature. The septum-covered solution was
cooled at 0 °C for an hour and then overnight at ꢀ25 °C. The precip-
itate was filtered off and washed with 30 ml of freshly distilled
hexanes. The solid was air dried for 2 h and recrystallization was
repeated for a second time. The product was then dried in vacuo
for 3 days at 90 °C. 7.7 g of 1 was produced in this way.
2
n
2
2
4
][B(C
6
5
F )
4
]. Electrochemical results
Cl
2 2
and CH CN are compared.
3
The electrochemistry of complex 2 was initially reported utilis-
n
4 6
ing only [N( Bu) ][PF ] as supporting electrolyte [12], which led to
poor resolution of the ferrocenyl peaks. The antineoplastic activity
of 2 has also been studied and related to the initial electrochemical
results [12a]. The detailed electrochemistry of complex 3 described
Mp = 159–161 °C. dH (300 MHz, CDCl )/ppm: 0.98 (t; 12H;
3
4 ꢂ CH ); 1.36 (q, 8H; 4 ꢂ CH ); 1.56 (q, 8H; 4 ꢂ CH ); 3.03 (t,
3
2
2
8H; 4 ꢂ CH ).
2

H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2225
2
.2.2. Synthesis of 4
The reaction vessel was flame-dried and an argon atmosphere
was maintained throughout. To a solution of acetyl ruthenocene
0.500 g; 1.8 mmol) in dry THF (1.2 ml) was added lithium diiso-
3. Results and discussion
3.1. Complex 2
Complex 2 has four redox-active ferrocenyl groups, and a
(
3
ꢀ3
propylamide (1 cm of 1.8 mol dm solution; 1.8 mmol) and the
reaction mixture was stirred for 20 min while cooling on ice.
Methyl ferrocenoate (0.44 g; 1.8 mmol) was added and the reac-
tion mixture stirred for 16 h at room temperature before 35 cm
of ether was added. The precipitate was collected by filtration,
added to an ice-cold HCl solution (0.1 mol dm ; 100 cm ) stirred
for 5 min and the product extracted into ether. The ether layer
was washed with H
and evaporated to dryness under reduced pressure. Chromatogra-
phy over Kieselgel 60 (Merck, grain size 0.063–0.2 mm) of the res-
II
III
II
Ru Ru core that can be irreversibly reduced to (Ru ) . In CH Cl /
2
2
2
ꢀ
3
n
0.1 mol dm [N( Bu) ][PF ], utilising a glassy carbon working elec-
4
6
3
ꢀ3
trode, 2 at 1 mmol dm concentrations showed a weak irreversible
+
cathodic wave at ꢀ503 mV vs. Fc/Fc (not shown). The four ferroce-
ꢀ3
3
nyl waves were located between 200 and 500 mV. This unresolved
peak splitting of the anticipated four ferrocenyl waves is due to
electrostatic effects of the differently charged oxidised ferrocenium
and reduced ferrocenyl groups, and is indicative of intramolecular
communication in 2. However, resolution of the peaks into four sep-
3
2
O (3 ꢂ 100 cm ), dried over MgSO
4
, filtered
idue with hexane:ether (1:4) (R
brown crystals (294 mg; 34%), mp 178 °C; IR (cm , neat) = 1606
f
= 0.53) as eluent gave 4 as red-
arate redox steps was not observed in the PF -containing medium
even though different formal reduction potentials for symmetrical
6
ꢀ1
(
C@O). d
COCHCO), 5.19 (2H; t; C
with Fc enol signal; t; C
H
(300 MHz; CDCl
3
): enol to the Fc side (47%), 5.9 (1H; s;
Rc), 4.81 (2H + 2H Rc enol overlapping
complexes in which mixed-valent (i.e. differently charged) interme-
diates are generated are well known in systems that allow electron
delocalisation, through through-space or through-bond conjugated
paths [13,17,18]. This electrochemical behaviour mirrored very
5
H
4
5
H
4
), 4.60 (5H; s; C
Fc); enol to Rc side (14%), 6.60 (1H; s; COC-
Rc), 4.81 (2H + 2H Rc enol overlapping with
Fc), 4.56 (5H; s; C Rc),
Fc); keto (39%), 5.24 (2H; t; C Rc), 5.10 (2H; t;
Fc), 4.94 (2H + 2H Rc keto overlapping with Fc keto signal; t;
Rc), 4.61 (5H; s; C Rc) 4.23 (5H; s; C Fc) and 2.28 (2H;
CO). Anal. Calc. for C23FeH20 Ru: C, 56.77; H, 4.14. Found:
C, 56.80; H, 4.16%.
5 5 5 4
H Rc), 4.51 (2H; t; C H
Fc), 4.20 (5H; s; C
HCO), 5.14 (2H; t; C
Fc enol signal; t; C
5
H
5
n
5
H
4
much the electrochemical profile in dichloroethane/[N( Bu) ][PF ]
4
6
H
5 4
), 4.72 (2H; t; C
5
H
4
H
5 5
of a similar compound in which only the ethanol axial ligands were
replaced with 1-propanol, i.e. [Ru2(
3
C
C
.99 (5H; s; C
5
H
5
H
5 4
l
-FcCOO) ꢁ(CH CH CH OH) ]
4
3
2
2
2
5
H
H
4
[PF ], 5 [12]. The poorly resolved ferrocenyl electrochemical couples
6
5
4
5
H
5
H
5 5
were also observed when the supporting electrolyte was changed to
ꢀ3
s; COCH
2
O
2
0.10 mol dm 1 (Fig. 2, top). However, upon lowering the concen-
ꢀ3
ꢀ3
tration of 2 to 0.5 mmol dm , in the presence of 0.1 mol dm 1, all
four ferrocene couples 1, 2, 3 and 4, could be resolved. OYSW vol-
tammetry resolved the four ferrocenyl waves the best. Fig. 2 high-
lights the differences of the voltammograms of 2 under these
concentration conditions while Table 1 summarises the electro-
chemical data that can be extracted from these voltammograms.
This result highlights the key observation that higher concentra-
tions (causing increasing non-ideal behaviour) may have a detri-
mental effect on peak resolution. A similar improvement in peak
2
.3. Electrochemistry
Cyclic, linear sweep and Osteryoung square wave voltammetry
were conducted using a computer-controlled BAS model 100 B/G
potentiostat. Data, uncorrected for junction potentials, were col-
lected with standard BAS 100 software and exported to Excel for
manipulation and analyses. Temperature was controlled using a
water bath at 20.0 ± 0.1 °C unless otherwise stated. Experiments
ꢀ3
n
resolution was not observed in CH Cl /0.1 mol dm [N( Bu) ][PF ]
2
2
4
6
were performed under argon on solutions of dried distilled CH
3
CN,
][PF ],
}] as supporting elec-
when the concentration of 2 was halved.
A close inspection of Fig. 2 reveals that the Ru Ru /Ru Ru cou-
ple of also behaved substantially different in CH Cl /
ꢀ3
n
II
III
II
II
CH
2
Cl
2
or THF in the presence of 0.1 mol dm
[N( Bu)
4
6
n
n
[
N( Bu)
4
][B(C
F
6 5
)
4
] or [N( Bu)
4
][B(C
6
3
H (CF
3
)
2
2
2
2
n
trolyte. A three-electrode cell containing a Pt-wire counter electrode
was used. The BAS glassy carbon working electrode with surface
4 6 5 4
[N( Bu) ][B(C F ) ] compared to what was previously described
n
[12] for 2 in (1,2-C H Cl /[N( Bu) ][PF ]. On the cathodic cycle,
2
4
2
4
6
area 3.14 or 7.07 mm2 was pre-treated by polishing on a Buehler
+
the wave labelled ‘‘5” at ꢀ490 mV vs. Fc/Fc represents the irre-
1
II
III
II
II
microcloth first with 1
lm and then
4
lm diamond paste. The ref-
versible Ru Ru /Ru Ru couple. However, this wave is followed
by wave X. The LSV shows these two waves both represent the
same amount of electrons. The insert CV labelled ‘‘CV with i ꢂ 2”
shows that if the switching potential is not at 1000 mV, but at
ꢀ200 mV, then on the second cathodic cycle, peak 5 disappeared.
From this we conclude that 2 loosely coat the surface of the elec-
trode after it was irreversibly reduced to completely modify the
electrode surface. When the potential associated with wave X is
reached, further molecules of 2 from the bulk of the solution are
again reduced, but now at an electrode having markedly different
characteristics than the original clean electrode.
erence electrode was constructed from a silver wire inserted into a
ꢀ
3
ꢀ3
n
solution of 0.01 mol dm AgNO
CH CN in a luggin capillary with a vycor tip, [X] = [PF
or [B(C (CF }]. For measurements in CH Cl bulk solutions, this
luggin capillary was inserted into a second luggin capillary with vi-
3
and 0.1 mol dm [N( Bu)
4
][X]in
F ) ]
6 5 4
3
6
], [B(C
H
6 3
3
)
2
2
2
ꢀ
3
n
cor tip filled with a 0.1 mol dm [N( Bu)
prevent Ag contact with CH Cl and possible precipitation of AgCl.
4
][X] solution in CH
3
CN to
+
2 2
The working electrode was quickly contaminated and as a precau-
tion it was washed (acetone, then methanol, then water) and pol-
ished between successive experimental groupings. It was never
used for more than six successive cycles before it was recleaned.
Successive experiments under the same experimental conditions
showed that all formal reduction and oxidation potentials were
reproducible to within 5 mV. Ferrocene and/or decamethylferro-
cene (Fc*) was added at the end of each experiment as internal stan-
That this adhesion of reduced 2 to the electrode surface is weak,
is obvious from the main CV: if the switching potential is 1000 mV,
by the time the next CV at faster scan rate is performed, the origi-
nal electrode surface has already been regenerated, allowing once
again observation of waves 1–4 in the expected positions.
+
dard to the bulk solution to enable data reporting vs. the Fc/Fc
couple as recommended by IUPAC [9]. Under our experimental con-
0
ditions, the Fc/Fc+ couple exhibited E = 0.087 V vs. Ag/Ag , ipc
o
+
/
3.2. Electrochemistry of 3
0
o
i
pa = 0.98,
D
E
p
= 74 mV in CH
Ag/Ag , ipc/ipa = 0.97,
pa = 0.97,
= 86 mV and E = 0.176 V vs. Ag/Ag . The Fc*/Fc*
couple is at ꢀ0.610 V vs. Fc/Fc in CH
3
CN. In CH
2
Cl
2
, E
þ
¼ 0:263 V vs.
Fc=Fc
+
D
E
p
= 81 mV, while in THF it exhibited ipc
/
The structure of 3 has been solved [15a]. To balance charges,
one would expect 3 to have the formula [Cd H {Pc(C H ) } ].
2 2 6 13 8 3
However, no sign of the two protons to balance the charge of the
three phthalocyanine rings could be found in the single crystal
0
o
+
+
i
D
E
p
+
2
Cl
2
as well as CH
3
CN, while
+
+
in THF the Fc*/Fc* couple is at ꢀ0.515 V vs. Fc/Fc [11].

2
226
H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2
der (tetra decker) species as described elsewhere [15]. Electro-
chemical data from these CV’s is summarised in Table 2.
1
Fc*
There were very noticeable differences in oxidation wave
n
patterns upon moving from CH
2
Cl
2
/[N( Bu)
4
][B(C
6 5 4
F ) ] and –
n
n
[
B{C
[B(C
B{C
6
H
3
(CF
]. With the CH
(CF
3
)
2
}
4
]
to CH
2
Cl
Cl
2
/[N( Bu)
4
][PF
6
]
and THF/[N( Bu)
4
]
]
5
n
n
X
6
F
5
)
4
2
2
/[N( Bu)
4
][B(C F ) ] and –[N( Bu)
6 5 4
4
[
H
6 3
3
)
2
}
4
] solvent/supporting electrolyte systems, the six
Fc
2
3
possible oxidation one-electron transfer processes were detected
in three separate oxidation waves labelled A, B and C (Fig. 3). From
1
4
2
µA
the LSV’s each wave represents a total flow of two electrons. Utilis-
n
OYSW
i x 3
ing [N( Bu)
4
][PF
6
] as the supporting electrolyte, and also in the THF
3
experiments, the six electron oxidation profile manifested in a dif-
ꢀ
2
ferent way. For the PF6 experiments, wave A still represented the
4
first two oxidation processes, but waves B and C coalesced into clo-
sely overlapping peaks that in total represented the flow of four
electrons. In THF, waves A and B each represented a flow of three
electrons, and wave C was absent (Fig. 3). The first oxidation wave,
wave A is in THF also shifted by 80 mV to more negative potentials
1
Fc
CV with i x 2
5
than was found in CH
2 2
Cl solutions. Under oxidative conditions,
X
oxidised products become more positive in character as demon-
CV
1
+
4+
strated by 3 ꢀ3 , the four highest oxidised species of 3 (Scheme
1). Any species that can interact or associate with positively
charged species will interact or associate with it. The oxidation
profiles of 3 as demonstrated by the CV’s of Fig. 3 are consistent
CV
2 µA
5
X
ꢀ
with electrolyte anion PF6 interaction with 3, probably via ion
n+
ꢀ
pair formation of the type 3 ꢁ ꢁ ꢁ(PF ). THF presumably interacts
6
n+
via solvation to generate 3 ꢁ ꢁ ꢁ(THF)
x
, x P 1. The contrasting oxida-
LSV
i x 3
ꢀ
tion patterns highlight how the larger volume of the BðC
6
F
5
Þ
and
4
ꢀ
ꢀ
6
BfC
6
H
3
ðCF Þ g anions over that of PF as well as the electron
3
2
4
withdrawing properties of the F-containing phenyl rings distrib-
ꢀ
utes the negative charge of BðC
6
F
5
Þ
anions over a large volume
4
ꢀ
in comparison with PF . The effect of this lower charge density
6
-
1000
-500
0
500
1000
is to minimise ion pair formation and enhance observation of elec-
tron transfer processes associated with naked cations.
Cyclic voltammetry could not resolve any of the two processes
associated with waves A, B or C in any of the four solvent/electro-
E / mV vs. Fc/Fc+
ꢀ
3
n
2 2 4 6 5 4
Fig. 2. Top: CV of 2 at 1 mmol dm in CH Cl /[N( Bu) ][B(C F ) ] with internal
reference decamethylferrocene, Fc*. The dark line highlights the Ru reduction at
peak 5 in the absence of Fc*. Bottom: LSV and CV’s at 100 (smallest currents), 200,
00, 400 and 500 mV s and a Osteryoung square wave voltammograms at 10 Hz
III
lyte mediums used, but an Osteryoung square wave analyses in
n
CH
2
Cl
2
/[N( Bu)
4
][B{C
6
H
3
(CF
3
)
2
}
4
]
2 2
(not shown) or CH Cl /
ꢀ
1
3
n
[
N( Bu)
4
][B(C
6
F
5
)
4
] (Fig. 3) clearly demonstrated the two compo-
ꢀ
3
of 2 at 0.5 mmol dm
.
nents of wave C. Wave A did not resolve into two separate peaks
with the square wave technique, but a shoulder next to the main
2 2
peak in the presence of 1 is observable. In CH Cl , wave B did not
structure determination. Complex 3 also shows so much radical
content in the resting state [15], that it was impossible to obtain
any NMR spectra. Until further light can be shed on the position
of the two ‘‘missing” protons or the exact radical distribution of
resolve in any form of resolution associated with the two steps
with any of the techniques at our disposal.
The small degree of peak-overlap in wave B is emphasised with
the CV
D
E
½BðC6F5Þ4
ꢃ
6 96 mV. For wave A, the CV
A;½BðC6F5Þ4ꢃ
DE =
value explains why a shoul-
3
, for the electrochemical discussion below, we choose to write this
128 mV. This slightly larger DE
A;½BðC6F5Þ4ꢃ
2ꢀ
2ꢀ
3
compound in the resting state as 3 = [Cd
{Pc(C H
2 6 13
)
8
}
]
. In
der on the main peak of A could be observed in the square wave
general, phthalocyanines with redox-silent metals in the cavity
can exhibit two ring-based one-electron oxidations and four ring-
based one-electron reductions [18]. Since cadmium is not redox-
active, the investigated triple-decker phthalocyanine complex, 3,
can in principle show 2 ꢂ 3 = 6 one-electron transfer oxidation
processes and 4 ꢂ 3 = 12 one-electron reductions. Regardless of
solvent and supporting electrolyte, all six oxidation processes
and CV voltammograms. The portion of the CV’s associated with
n
wave C in CH
2
Cl
2
/[N( Bu)
4
][B(C
6
F
)
5 4
] and –[B{C H
6 3
(CF
3
)
2
}
4
] gave
C
DE P 190 mV and is completely consistent with two near over-
lapping peaks that could be resolved by OYSW voltammetry.
Which two of the three phthalocyanine rings 3 is first oxidised
(or reduced) could not be unambiguously identified. The above-
mentioned relatively small
DE values – each applies to two one-
could be accounted for in waves A, B and C (Fig. 3). In CH
2
Cl
2
, nine
electron transfer processes, not only one – lead us to conclude that
all six one-electron oxidations should be to a large extent electro-
chemically reversible and fast on the CV time scale.
of the twelve reduction processes were identified in waves I
through V. In THF, two further reduction processes could be iden-
tified, implying 17 of the possible 18 one-electron transfer pro-
cesses that 3 can undergo, could be pinpointed. Although each of
these processes should occur at a unique potential, some of them
occurred at potentials so close to each other that they could not
For the reduction processes of 3 the observed electrochemistry
n
in CH
2
Cl
2
/[N( Bu)
4
][B(C
6
F
5
)
4
], –[B{C H
6 3
(CF
3
2
) }
4
] and –[PF
6
] paral-
leled each other very closely implying ion pair formation between
the anions of these electrolyte anions and negatively charged re-
duced forms of 3 is negligible. Only cations would influence reduc-
tion process of anionic species by ion paring of the type (anionic
be resolved. The small CV peaks labelled X at –1.44 V and Y at
+
0
.11 V vs. Fc/Fc are not part of the main redox system associated
ꢀ
+
with 3. They are disassembled and different reassembled products
of 3, leading inter alia to metal free phthalocyanines and higher or-
species) ꢁ ꢁ ꢁ (cation). LSV peaks I, II and III showed each of these
waves are associated with a one-electron transfer process. That

H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2227
Table 1
ꢀ
1
ꢀ3
+
Voltammetry data for 2 and 4 at scan rate = 100 mV s of ca. 1.0 mol dm solutions of compounds at 20 °C. Potentials are vs. Fc/Fc .
0
0
Wave
Epa (V)
D
E
p
(mV)
Eo (V)
i
pa
(
lA)
i
pc/ipaa
Wave
E
pa (V)
D
E
p
(mV)
Eo (V)
i
pa
(lA)
i
pc/ipaa
Ferrocene, Fc
Decamethylferrocene, Fc*
In CH
In CH
In THF
3
CN
Cl
0.037
0.041
0.043
74
81
86
0
0
0
2.32
2.08
2.90
0.98
0.97
0.97
In CH
In CH
In THF
3
CN
Cl
ꢀ0.575
ꢀ0.576
ꢀ0.468
70
68
82
ꢀ0.61
ꢀ0.61
ꢀ0.515
2.12
2.04
2.94
0.99
0.99
0.96
2
2
2
2
n
n
2
1
2
3
4
5
X
(1.0 mM) in CH
2
Cl
2
/[N( Bu)
4
][B(C
60
6
F
5
)
4
]
2 (0.5 mM) in CH
2
Cl
2
/[N( Bu)
4
][B(C
60
74
76
66
—
6 5 4
F ) ]
b
0.9b
0.7b
0.9b
0.292
0.462
0.257
0.415
—
—
0.593
—
0.7
1
2
3
4
5
X
0.324
0.434
0.526
0.626
0.294
0.397
0.488
0.593
—
b
b
b
b
94
0.7
0.9
0.7
0.9
c
c
c
c
c
b
b
—
—
—
—
—
—
—
0.7
0.9
c
c
c
c
c
b
b
—
0.7
0.9
d
d
d
d
ꢀ0.503
—
—
0.4
—
—
ꢀ0.495
0.4
—
—
e
—
—
ꢀ0.787
—
—
0.3
n
n
4
in CH
3
CN/N( Bu)
4
PF
6
2 2 4 6 5 4
4 in CH Cl /[N( Bu) ][B(C F ) ]
0.176f
0.207
0.484
0.636
ca. 74
ca. 74
f
f
0.213
0.244
—
—
—
11.2
10.6
11.4
11.2
b
b
b
b
—
—
—
—
c
Enol Fc, 1
Keto Fc, 2
Enol Rc, 3
Keto Rc, 4
Dimer, 5
0.154
0.203
0.802
0.883
0.515g
ca. 111
ca. 111
ca. 131
ca. 106
0.213
0.258
0.737
0.83
10.6
f
—c
Enol Fc, 1
Keto Fc, 2
Enol Rc, 3
Keto Rc, 4
Dimer, 5
c
f
c
9.4
—
—
—
—
c
c
c
c
c
—
—
—
—
—
—
c
c
c
c
c
c
c
c
c
c
c
c
c
c
—
—
—
—
—
a
b
c
d
e
f
Current ratios were always taken to be (reverse scan current)/(forward scan current).
Estimates only.
Not observed, or poor peak resolution disallowed accurate peak current measurement.
E
pc or ipc (no Epa were detected).
ꢀ3
It is unclear if the shoulder at ꢀ0.83 V actually is peak X as observed when [2] = 0.5 mmol dm
.
Due to poor peak resolution, Epa could not be measured, Epc,1 = 0.176 V. Upon assuming
D
E
p
+
= 74 mV, as was found for free ferrocene under our conditions, Epa could be
0
1
0
2
o
o
estimated, and formal reduction potential estimates are E ꢄ 0.213 and E ꢄ 0.244 V vs. Fc/Fc .
g
III
An Epc value. In CH
2
Cl
2
, a weak reduction process was observed which is attributed to a dimerized Ru species in analogy with what was found for free ruthenocene. At
0
°C this peak was absent and wave 3 became dominant in the cathodic cycle.
Table 2
n
n
Voltammetry data of ca. 0.5 mM solutions of 3 at ꢀ40 °C in CH
2
Cl
2
or THF containing 0.1 M [N( Bu)
4
][B(C
H
6 5
)
4
], [NBu
4
][B{C
6
3
H (CF
3
)
2
}
4
] or [N( Bu)
4
][PF
6
].
LSVa
n
CVc
pc (V)
CV
CV
E
c
OYSWc,d
E (V)
Wave
LSV
n
CVc
CV
CV
E
c
OYSWc,d
E (V)
Wave
0
0
b
o
b
o
E
D
E
p
(mV)
(V)
E
pc (V)
D
E
p
(mV)
(V)
n
n
3
VI
V
, in CH
2
Cl
2
/[N( Bu)
—
4
][B(C
6
F
—
5
)
4
]
3, in THF/[N( Bu)
VI
4
][B(C
6
F
5
)
4
]
e
e
—e
90
—e
—e
2
1
1
1
1
2
1
1
1
3
ꢀ2.89
ꢀ2.51
ꢀ2.36
ꢀ2.23
ꢀ1.96
ꢀ1.82
ꢀ1.12
ꢀ0.85
ꢀ0.65
ꢀ0.20
0.35
140
110
150
80
140
130
139
204
228
100
196
160
ꢀ2.82
ꢀ2.45
ꢀ2.28
ꢀ2.19
ꢀ1.89
ꢀ1.75
ꢀ1.05
ꢀ0.75
ꢀ0.54
ꢀ0.25
0.25
ꢀ2.83
ꢀ2.47
ꢀ2.31
ꢀ2.23
ꢀ1.93
ꢀ1.77
ꢀ1.07
ꢀ0.71
ꢀ0.53
ꢀ0.22
0.3
f
f
2 or 3
ꢀ2.21
ꢀ2.16
ꢀ2.13(ꢀ2.18)
V
V
V
3
e
e
e
e
e
—
—
—
—
—
—
—
—
2
1
e
e
e
e
e
—
—
IV1,2
3
1
1
1
ꢀ1.86
ꢀ1.10
ꢀ0.72
ꢀ0.53
170
62
86
ꢀ1.78
ꢀ1.07
ꢀ0.68
ꢀ0.49
ꢀ1.79(ꢀ1.87)
IV
3
III
II
I
ꢀ1.09
IV1,2
III
II
ꢀ0.7
76
ꢀ0.5
I
g
g
g
f
g
g
A
B
C
2
2
2
ꢀ0.11
0.47
128
92
220
ꢀ0.17
0.42
0.80
ꢀ0.18(ꢀ0.11)
A
h
0.44
B
B
1
3
f
h
0.49g
0.91
0.78(0.91)
2
0.41
0.43
n
3, in CH
2
Cl
2
/[NBu
4
][B{C
6
H
3
(CF
3
2
) }
4
]
3, in CH
2
Cl
2
/[N( Bu)
4 6
][PF ]
V
IV
III
II
I
A
B
C
2 or 3
ꢀ2.18
ꢀ1.77
ꢀ1.08
ꢀ0.72
ꢀ0.52
ꢀ0.08
0.52
90
150
92
92
76
120
82
190
ꢀ2.13
ꢀ1.7
ꢀ2.16
ꢀ1.69
ꢀ1.05
ꢀ0.68
ꢀ0.5
V
IV
III
II
I
A
B
C
2 or 3
ꢀ2.21
ꢀ1.86
ꢀ1.11
ꢀ0.72
ꢀ0.53
ꢀ0.10
90
160
62
88
80
118
96
120
ꢀ2.16
ꢀ1.78
ꢀ1.08
ꢀ0.68
ꢀ0.49
ꢀ0.16
0.42
ꢀ2.17
ꢀ1.81
ꢀ1.07
ꢀ0.65
ꢀ0.48
ꢀ0.15
0.45
3
1
1
1
2
2
2
3
1
1
1
2
4
—
ꢀ1.03
ꢀ0.67
ꢀ0.48
ꢀ0.14
0.48
g
g
g
g
ꢀ0.15
0.48
0.47g
0.82g
f
0.92
0.82
0.77(0.91)
0.76
0.78
a
b
c
ꢀ1
LSV recorded at a scan rate of 2 mV s
.
Amount of electrons that flowed per molecule of 3 for each wave as determined by LSV.
Scan rate = 100 mV s ; potentials vs. Fc/Fc = 0 mV in CH
ꢀ1
+
+
2
Cl
2
or at ꢀ0.095 in THF; under these conditions, the Fc*/Fc* couple is at ꢀ610 mV in both solvents used, see
Table 1.
d
Osteryoung square wave recorded at 10 Hz.
Peak not observed in this solvent.
First-mentioned potentials are those of the of the first identifiable peak, the potential in brackets are for the second identifiable peak or shoulder associated with the main
e
f
peak.
g
E
pa not Epc.
h
The total amount of electrons that flows per molecule at wave BTHF
.
ꢀ1
the region between waves A and I really represents the resting state
DE values for I–III from the CV scans at 100 mV s were all be-
of 3 under the oxygen free conditions in the glove box, is clearly
low 92 mV indicating near electrochemically reversible processes.
Another observation that supported this conclusion was the obser-
vation that peak potentials was independent of scan rate. This
2 2
proved by the CH Cl LSV of Fig. 3. It is only in the potential range
between waves I and A that the LSV experiment drew zero current.

2
228
H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
not
observed
wave VI
[Cd (Pc6-)2(Pc5-)]13-
[Cd (Pc ) ]
2
wave V
3
-
14-
5- 11-
[
Cd (Pc6 ) ]
2
3
2
3
1
e-
wave V2
Cd (Pc5-)2(Pc4-)]10- [Cd (Pc5-)(Pc4-) ]
2e-
9w- ave V1
[Cd (Pc ) ]
1e-
8w- ave IV3
4-
[
2
2
2
2
3
e-
wave IV1,2
Cd (Pc4-)2(Pc3-)]7-
1e-
1e-
1
5w- ave III
wave II
3-
[Cd (Pc3-)2(Pc2-)]4-
[
[Cd (Pc ) ]
2
2
3
2
2
e-
1e-
1e-
3w- ave I
2w- ave A
3e-
wave B
Cd (Pc3-)(Pc2-) ]
2-
1- 1+
3
0
-
5-
4+
[
[Cd (Pc ) ]
[Cd (Pc ) ]
2 3
2 3
C (Pc ) ]
[Cd (Pc ) ]
2
2
2
3
2
3
1
e-
3e-
Scheme 1. Seventeen of the possible one-electron transfer processes for 3 could be
identified in THF according to the reaction sequence above.
In THF, waves I, II and III, are also representative of the first
3ꢀ
4ꢀ
three one-electron transfer processes of 3 to generate 3 , 3
5ꢀ
and 3 , respectively. The set of formal reduction potentials I–III
were within 70 mV of those obtained in CH Cl (Table 2). Although
THF interacted strongly with positively charged oxidised species of
, THF interaction with the large negatively charged reduced forms
of 3 that is generated in waves I through VI was negligible. The re-
2
2
3
3
ꢀ
duced forms of 3 vary between 3 and the fully reduced species
1
4ꢀ
3
(Scheme 1).
In contrast to waves I–III, by virtue of the LSV traces, both waves
IV and V in all CH Cl experiments represent a cluster of three one-
2
2
electron transfer processes. The three one-electron transfer pro-
cesses associated with wave IV are better resolved than in wave
V, because
V
DEIV is 70–80 mV larger than DE = 90 mV (Table 2).
This larger peak separation also allowed the OYSW technique to re-
solve wave IV in the presence of 1 into two of its three components
(Fig. 3). No evidence for the fourth array of reduction processes was
2 2
found in CH Cl as solvent, but evidence for this step was found in
THF.
When wave IVTHF was encountered, it was found to be well
enough resolved to separate two of the three expected one-elec-
tron transfer processes. From the LSV scans, the combined
amount of electrons that was transferred during the overall pro-
cesses associated with wave IV were still three. OYSW voltamme-
try could still not resolve wave IV into three separate peaks.
However, in THF, wave V clearly split into all three its compo-
Fig. 3. CV’s at 100 mV sꢀ1 and ꢀ40 °C highlighting the differences in oxidation
waves A–C due to solvent or electrolyte changes. Reduction processes are labelled I
through VI, numbering start from the resting potential to synchronize wave number
with sequential redox processes. Blue: OYSW and CV of a 0.5 mM solution of 3 in
9ꢀ
10ꢀ
11ꢀ
nents during generation of 3 , 3
and 3 , respectively. The
n
n
CH
in CH
2
Cl
2
/[N( Bu)
4
][B(C
6
F
5
)
4
]. Green: CV in CH
]. Black: CV and OYSW in THF/[N( Bu)
Cl or THF. The experimentally measured LSV currents were
2
Cl
2
/[N( Bu)
4
][B{C
6
H
3
(CF
3
)
2
}
4
]. Red: CV
n
n
OYSW trace shown in Fig. 3 highlights the higher degree of reso-
lution in THF best. Part of the reason why wave V resolved into
its three separate components must centre around a very low de-
2
Cl
2
/[N( Bu)
LSV’s at 2 mV s in CH
4
][PF
6
4
][B(C
F ) ]. Purple:
6 5 4
ꢀ
1
2
2
multiplied by 2 for better display purposes; the dotted lines accentuate the electron
nꢀ
count associated with each wave and pinpoint the resting position of 3. The
gree of interaction between 3 and THF under strong reducing
+
reference potential was taken as ꢀ610 mV for the Fc*/Fc* couple. This implies data
conditions. However, this is not the only reason. A second con-
tributing factor as to why this happened lies in the slower rate
+
+
2 2
is also referenced vs. Fc/Fc = 0 V in CH Cl but not in THF. In THF, the Fc/Fc couple
is at ꢀ80 mV under these conditions, Table 1. The peaks labelled Y and X are not
part of the electrochemical fingerprint of 3, see text. (For interpretation of the
references in colour in this figure legend, the reader is referred to the web version of
this article.)
nꢀ
of electron transfer between 3 , and the working electrode in
THF. From Table 2 it is clear that
smaller in CH Cl than in THF. This means that electron transfer
between 3 and the working electrode is much faster in CH Cl
than in THF (fast electrochemical reversible processes are theoret-
ically characterised by E = 59 mV [1]). It implies that measured
DE values are almost always
2
2
2
2
cluster of the first three reduction processes represents the first
reduction wave on each phthalocyanine ring of 3 and is better re-
solved than any other cluster of three one-electron transfer process
associated with 3. The other peaks in the CV of 3 all represented to
a large degree electron transfer processes consisting of coalesced
peaks. The good resolution between peaks I, II and III is indicative
of an enhanced field effect through through-space communication
D
peak potentials in THF are not necessarily thermodynamically
ideal. Rather, they are shifted in such a way that peak potentials
represent the potentials where the fastest rates of electron trans-
fer occur. This induced quasi electrochemical reversibility in THF
and contributed to the chance observation of three separate one-
electron transfer peaks for wave V and two of three possible
peaks for wave IV.
The final additional feature introduced by the use of THF as sol-
vent was observation of the fourth and final possible reduction
wave set associated with 3 in wave VI. Wave VI could not be
[
3
17] between the differently charged intermediate reduced species
, 3 , 3 and 3 (Scheme 1). Here, the different charges are
2ꢀ
3ꢀ
4ꢀ
5ꢀ
located on the neighbouring phthalocyanine rings which are only
.965 Å apart [15a].
2

H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2229
observed in CH
2
Cl
2
as it occurred outside the potential window of
3.3. Electrochemistry of 4
this solvent. The LSV scans showed wave VI to comprise of at least
two one-electron transfer processes, thus generating 31 and 3
2ꢀ
13ꢀ
.
The b-diketone 4 exists as an equilibrium of enol and keto iso-
mers as described elsewhere [14]. Freshly prepared samples have a
high keto content, but in the solid state, samples of 4 convert with-
in a few weeks exclusively to the enol form. Upon dissolving an
aged, enol-enriched solid sample, it slowly coverts to the equilib-
rium keto-enol position over several hours. The kinetics of this
The onset of solvent decomposition at this low potential made it
uncertain if the final reduction to 31 could be detected. OYSW
analyses could not resolve the two observed processes into sepa-
rate steps. The strength of the cadmium bonds to the phthalocya-
nine macrocycles are put into perspective when one observes that
4ꢀ
1
3ꢀ
3
did not dissociate to monomeric or dimeric species due to
3
conversion have been studied, [14]. In CH CN, the ferrocenyl centre
electrostatic repulsive forces on the CV time scale.
of 4 showed reversible electrochemistry with the formal reduction
In order to write a reaction sequence for the described electro-
chemistry in THF, it is useful to recognising that the charge of Cd is
potentials of the two poorly resolved waves estimated at
0
o
E
= 0.213 for the enol isomer (wave 1, Fig. 4, left, top, from an
2
+ and that of the hydrogen-free phthalocyaninato (Pc) rings is 2ꢀ.
aged sample of 4) and 0.244 V for the keto isomer (wave 2, from
0
ꢀ
2
2ꢀ
a sample of 4 at equilibrium). Eo was calculated by assuming
The resting state of 3 can then be written as [Cd
2
(Pc
)
3
]
. In this
notation, the charge in the bracket represents the charge on an
individual Pc ring while the outer charge represents the total
charge on the molecule. By adopting this writing style, Scheme 1
can be written to describe the 17 observed and 1 unobserved elec-
trochemical processes of 3 in THF as solvent.
p
DE = 74 mV for both waves, the same as that for free ferrocene,
at a scan rate of 100 mV s (Table 1, Fig. 4). In contrast, irrevers-
ꢀ1
ible electrochemistry was found for the ruthenocenyl centre of 4
ꢀ
3
n
in CH
3 4 6
CN in the presence of 0.1 mol dm [N( Bu) ][PF ] support-
ing electrolyte. Only anodic peak potentials for the enol (wave 3)
Fig. 4. Left. Top: Enol-enriched (thin line) and equilibrium enol/keto content CV’s at 100 mV/s of 4 in CH
3
CN at 20 °C. Middle: Enol-enriched (thick line) and equilibrium enol/
keto content CV’s of 4 in CH Cl at 20 °C. The reduction half wave 5red at 515 mV is associated with isomers of a hydride-containing dimer similar to that found for
2
2
III
’
ruthenocene, i.e. [(H)(C
5 5 5 4 2 2 2
H )Ru (C H )] [21]. Bottom: CV s at 0 °C in CH Cl of an equilibrium solution of 4 at scan rates of 100 (smallest currents), 200, 300, 400 and 500 mV.
II
III
The reduction half wave 3red at ca. 662 mV is associated with a simple (C
H
5 5
)Ru (C
R)] . Right: OYSW voltammograms of 4 at 10 Hz and 20 °C of 4 and at equilibrium in CH
thick line for an equilibrium keto-enol solution of 4) or CH Cl (bottom, thick line enriched in the enol form, thin line of keto/enol equilibrium solutions).
5
H
4
R)/(C
5
H
5
)Ru (C
5
H
4
R) couple overlapping the two-electron reduction wave of isomers of
III
III
2+
[
5
(C H
4 5
R)(C H
5
)Ru –Ru (C
5
H
5
)(C
5
H
4
3
CN (top; thin line enriched in the enol form;
2
2

2
230
H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
and keto forms (wave 4) of 4 were observed at 0.484 and 0.636 V,
25.00
20.00
respectively. It is generally accepted that RuIV species arise in CV
experiments involving ruthenocene derivatives in CH
3
CN/
n
[
N( Bu)
4 6
][PF ] [8g,14,19,20]. Here, in analogy to free ruthenocene
IV
2+
[
14,21] and also free osmocene [22], [(C
H
5 4
R)(C H
5 5
3
)Ru (CH CN)]
is probably formed, R = {CO–CH –CO–Fc ꢀ CO–CH@C(OH)–Fc}.
2
1
1
5.00
0.00
5.00
However, Mann and co-workers [8g], and more recently Geiger
and co-workers [8b,21], reported reversible oxidation of free
n
ruthenocene in CH
2
Cl
2
/[N( Bu)
4
][B{(C
6
H
3
)(CF
3
)
2
}
4
] and CH
2
Cl
2
/
n
III
[
N( Bu)
4 6 5 4
][B(C F ) ] to two Ru species, one the simple monomeric
III +
ruthenocenium cation [(C
5 5 2
H ) Ru ] , the other dimeric
III
III
]2+ which exists at low (ꢀ40 to 0 °C) tem-
[
5 5 2 5 5 2
(C H ) Ru –Ru (C H )
perature. At higher temperature this dimer can isomerise revers-
III
ibly to the hydride-containing dimer [(H)(C
5 5 5 4 2
H )Ru (C H )]
0
.00
which also has a Ru–C -bond [21]. The ruthenocene-containing
r
0
.000 0.250 0.500 0.750
ν0.5 / (V s-1)0.5
b-diketone 4 behaved similarly.
3
Upon comparing electrochemical results of 4 from a CH CN
2 2
medium with those obtained in CH Cl (Table 1, Fig 4), a key obser-
+
m
0.5, Eq. (4),
vation is that, although formal reduction potentials vs. Fc/Fc for
ferrocenyl-based waves 1 and 2 are for all practical purposes the
same in both solvents, potentials for waves 3 and 4 of the
ruthenocenyl group of 4 shifted to much more positive values;
from 0.484 and 0.636 V to 0.802 and 0.883 V, respectively. In
Fig. 5. The relationship between ipa and square root of scan rate,
deviates from linearity in waves 3 (dotted line, j) and 4 (broken line, e) of 4 in
CH CN, demonstrating electrochemical irreversibility. In contrast, in CH Cl , this
relation ship is linear. Graph shows wave 3 (d) and the superimposed compliment
of waves 3 and 4 (s; current at Epa,wave 4). Poor wave resolution, prevented
determination of accurate ipa values for wave 4 alone.
3
2
2
CH
nyl fragment, while in CH
almost the same as that of the ferrocenyl group in CH
3
CN,
D
E
pa,Rc = Epa,wave 4 ꢀ Epa,wave 3 = 152 mV for the ruthenoce-
Cl pa,Rc is only 81 mV. This value is
Cl where
2
2
DE
+
tion peak at ca. 515 mV vs. Fc/Fc (Fig. 4, bottom left). In analogy
2
2
to free ruthenocene [8b,21], this reduction wave is considered to
represent the two-electron reduction of isomers of the low
H R)(C H )Ru –Ru (C H )(C H R)]
5 4 5 5 5 5 5 4
overlapping with small amounts of simple monomeric
D
E
pa,Fc = Epa,wave 2 ꢀ Epa,wave 1 = 50 mV. Hence, different rutheno-
cene oxidation products form in CH
2
Cl
2
and CH
3
CN. The closeness
III
III
III
2+
temperature Ru dimer [(C
of the pa,Rc values in CH
D
E
pa,Fc and
D
E
2
Cl is consistent with keto
2
and enol forms of 4 also forming the simple but labile, monomeric
III
+
III
+
5 5 5 4
[(C H )Ru (C H R)] reduction. That wave 3red is associated with
ruthenocenium cations [(C
5 5
H
)Ru (C
5
H
4
R)] at the anodic half-
the reduction of monomeric and dimeric forms of 4 in equilibrium
are also borne out by the shape of this peak and its scan rate
dependence. At low sweep rates peak 3red is rather sharp because
the rate of the equilibration process can keep up with the rate of
the electrochemical experiment. As the sweep rate increases, the
electrochemical process ‘‘outruns” the dimer-monomer equilib-
rium. This renders peak 3red broad, drawn out and weak with the
results that the peak currents associated with back-reduction do
not increase in the same manner as currents in the forward peak.
The OYSW voltammograms of 4 are shown in Fig. 4, right, and bet-
ter resolution of especially the ferrocenyl waves were observed.
waves 3 and 4, just as the ferrocenyl group is involved in the sim-
III
+
II
ple redox couple [(C
5
H
5
)Fc (C
5
H
4
R)] /[(C
5
H
5
)Fe (C
3
CN] forms. For the
5
H
4
R)] at waves
IV
2+
1
and 2. In CH CN, [(C
3
H
5 4
R)(C
5
H
5
)Ru ꢁCH
2 2
ruthenocenyl fragment, Eqs. (1)–(4) are only valid in CH Cl /
n
[
N( Bu)
The ipc/ipa current ratio for waves 3 and 4 in CH
stantially from unity. Results therefore indicate that the oxidised
4 6 5 4
][B(C F ) ].
2
2
Cl deviate sub-
III
+
5 5 5 4
enol and keto [(C H )Ru (C H R)] fragments are prone to further
chemical processes. At 20 °C, the position of the most prominent
III
+
[
(C
5
H
5
)Ru (C
5
H
4
R)] fragment reduction wave at Epa = ca. 0.515 V
+
vs. Fc/Fc in CH
2
Cl (Fig. 4, left middle), is consistent with a two-
2
III
2 2
Utilising the above CH Cl results, Scheme 2 can be written for 4
electron Ru dimer reduction in analogy to free ruthenocene
8b,21], because it is found at a potential ca. 290 mV lower than
to involve electrochemically reversible one-electron oxidation pro-
cesses during the generation of enol and keto isomers E1–E3 and
K1–K3 at waves 1, 2, 3 and 4, respectively. At 20 °C, a two-electron
[
E
pa = 802 mV for peak 3. This peak potential difference is too
III
+
large to be associated with
5 5 5 4
a simple [(C H )Ru (C H R)] /
However, weak peak shoulders 3red
III
II
reduction of isomers of a Ru dimer back to the ruthenocenyl
[
(C
5
H
5
)Ru (C
5
H
4
R)] couple
.
group is associated with wave 5red. With present knowledge, the
dimer is most likely isomers of the hydride-containing dimer
and 4red are observed at ca. Ecathodic = ca. 671 and 777 mV. The esti-
mated
= Epa,peak i ꢀ Epc,shoulder i values of ca. 131 and 106 mV for
peak/shoulder pairs 3 and 4 point to quasi reversible simple
DE
p
III
+
II
[
(C
the square root of scan rate, (
found to be linear for waves 3 and 4 in CH
ruthenocenyl oxidation of 4 in CH Cl obeys Eq. (4), indicating
electrochemical reversibility even though the [(C
5
H
5
)Ru (C
5
H
4
R)] /[(C
5
H
5
)Ru (C
5
H
4
R)] couples, but a plot of
, vs. peak anodic currents was
Cl (Fig. 5). Thus,
0
.5
m
)
H
slow
2
2
_
_
__
Fc-CO CH CO-Rc
Fc-CO-CH -CO-Rc
2
2
2
III
+
E1
K1
5 5 5 4
H )Ru (C H R)]
peak 1
peak 2
cation is chemically converted to a dimeric species in a chemical
process directly after ruthenocene oxidation. This result highlights
the difference between electrochemical reversibility and chemical
reversibility. Chemical irreversibility implies ipc/ipa current ratios
deviate from unity even though the electron transfer process
slow
slow
+
+
Fc -C(OH)=CH-CO-Rc
Fc -CO-CH -CO-Rc
2
E2
K2
peak 3
peak 4
+
+
+
+
Fc -C(OH)=CH-CO-Rc
Fc -CO-CH -CO-Rc
2
may still be electrochemical reversible. In contrast, in CH
3
CN,
plots for waves 3 and 4 deviated from linearity, thereby
confirming ruthenocenyl oxidation in CH CN is not an electro-
chemical reversible process. A repeat CV experiment in CH Cl per-
E3
K3
0.5
chemical dimerisation
ipa–(m)
3
peak 5
III
peak 5
or peak 3
III
III
2+
h
[(
i
C
gh
H
o
R
r
)
l
(
o
C
w
H
t e) R
m
u
p e- Rr a
u
tu
(
r
C
e
H
R u) (C
d
H
im
R
e
)
r
]
s
5
4
5
a
5
nd other form
5
s
5
5
4
2
2
or peak
3
formed not at 20 °C but at 0 °C confirmed a strong reduction peak
+
associated with wave 3 at 662 mV vs. Fc/Fc but no dimer reduc-
Scheme 2. Electron transfer processes associated with 4.

H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
2231
dimerised oxidised product, including [(C
Ru (C H )(C H R)] , was found.
5 5 5 4
5
H R)(C
4
5
H
5
)RuIII–
III
2+
Acknowledgement
The authors acknowledge the UFS for financial support and Dr. I.
Chambrier for supplying samples of 3.
Appendix A. Supplementary material
ꢀ
ꢀ
Fig. 6. Space filling diagram of BðC
6
F
5
Þ
and PF anions. The longest distance
4
6
ꢀ
ꢀ
between diagonal F-atoms on BðC
6
F
5
Þ4 and PF6 anions approach 10 and 3.3 A,
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ica.2010.03.031.
respectively.
III
[
(H)(C
5
H
5
)Ru (C
H
5 4
)]
2
which also has a Ru–C
r
-bond, similar in
References
structure to what was found for free ruthenocene [21] and analo-
gous free osmocene [22]. At 0 °C, a two-electron reduction of iso-
[1] (a) D.H. Evans, K.M. O’Connell, R.A. Peterson, M.J. Kelly, J. Chem. Educ. 60
(1983) 291;
III
III
III
2+
mers of the Ru dimer [(C
5 4 5 5 5 5 5 4
H R)(C H )Ru –Ru (C H )(C H R)]
(
(
b) P.T. Kissinger, W.R. Heineman, J. Chem. Educ. 60 (1983) 702;
c) J.J. Van Benshoten, J.Y. Lewis, W.R. Heineman, J. Chem. Educ. 60 (1983) 772;
back to the ruthenocenyl species is associated with wave 3red
.
The species E1 is considered to be reduced at lower potentials
than K1, because carbonyl (C@O) substituted metallocenes are oxi-
dised at larger (more positive) potentials than alcohol (OH) substi-
tuted metallocenes [17e].
(d) G.A. Mobbot, J. Chem. Educ. 60 (1983) 697.
2] M. Lovric, in: F. Scholz (Ed.), Electroanalytic Methods: Guide to Experiments
and Applications, Springer, Berlin, 2005 (Chapter I.2, p. 23 and Chapter II.3, pp.
[
1
11–133).
[3] (a) D.E. Richardson, H. Taube, Inorg. Chem. 20 (1981) 1287;
b) A. Auger, J.C. Swarts, Organometallics 26 (2007) 102;
(
All of the above described results were made possible by the
n
(c) A. Auger, A.J. Muller, J.C. Swarts, Dalton Trans. (2007) 3623.
4] M. Schnippering, A. Zahn, S.-X. Liu, C. Leumann, S. Decurtins, D.J. Fermin, Chem.
Commun. (2009) 5552.
5] E. Fourie, J.C. Swarts, D. Lorcy, N. Bellec, Inorg. Chem. 49 (2010) 952.
6] D.G. Leaist, Electrochim. Acta 36 (1991) 309.
uses of the CH
2
Cl
2
/[N( Bu)
4
][B(C
F
6 5
)
4
] solvent electrolyte system
[
to minimise ion paring interactions between electrolyte anions
and positively charged oxidised substrates, and by careful choice
of the solvent. Fig. 6 shows a space filling diagram of the
[
[
[
7] J. Conradie, T.S. Cameron, M.A.S. Aquino, G.J. Lamprecht, J.C. Swarts, Inorg.
Chim. Acta 358 (2005) 2530.
ꢀ
ꢀ
BðC
6
F
5
Þ
and PF6 anions generated by the program Chemcraft
4
ꢀ
[
8] Leading publications demonstrating the influence and use of [B(C
6
F
5 4
) ]
and
on the same scale.
ꢀ
ꢀ
[B{(C
6
H
3
)(CF
3
2
) }
4
]
ions, the complimentary role of associated cations with
Upon visualising the size and structure of the ½BðC
6
F
5
Þ ꢃ anion,
4
different charge densities, as well as some solvent effects may be found in (a) F.
Barriere, R.U. Kirss, W.E. Geiger, Organometallics 24 (2005) 48. and references
therein;
ꢀ
it is easy to comprehend how the negative charge on ½BðC
6
F
5
Þ ꢃ
4
ꢀ
can be distributed over a much larger volume than in ½PF
6
ꢃ
be-
(
(
b) S. Trupia, A. Nafady, W.E. Geiger, Inorg. Chem. 42 (2003) 5480;
c) F. Barriere, N. Camire, W.E. Geiger, U.T. Mueller-Westerhoff, R. Sanders, J.
6 5
cause of the electron withdrawing effects of each of the C F
groups, thereby leading to a much smaller charge density on the
anion.
Am. Chem. Soc. 124 (2002) 7262;
(d) F. Barriere, W.E. Geiger, J. Am. Chem. Soc. 128 (2006) 3980;
(
(
(
e) A. Nafady, T.T. Chin, W.E. Geiger, Organometallics 25 (2006) 1654;
f) D.S. Chong, J. Slote, W.E. Geiger, J. Electroanal. Chem. 630 (2009) 28;
g) M.G. Hill, W.M. Lamanna, K.R. Mann, Inorg. Chem. 30 (1991) 4687.
4
. Conclusions
[
9] G. Gritzner, J. Kuta, Pure Appl. Chem. 56 (1984) 461.
[
10] R.R. Gagne, C.A. Koval, G.C. Lisensky, Inorg. Chem. 19 (1980) 2855.
In this paper we have clearly shown how the use of CH
2
Cl
2
/
[11] Leading references describing the electrochemical activity and behaviour of
ferrocene and decamethylferrocene in a multitude of organic solvents is (a) I.
Noviandri, K.N. Brown, D.S. Fleming, P.T. Gulyas, P.A. Lay, A.F. Masters, L.
Phillips, J. Phys. Chem. B 103 (1999) 6713;
n
[
N( Bu)
4
][B(C
F
6 5
)
4
] as solvent/supporting electrolyte results in bet-
ter resolution and a clearer electrochemical picture of multiple re-
dox processes associated with positively charged analytes. Under
strong reducing conditions, the larger negative potential window
(b) N.G. Connelly, W.E. Geiger, Chem. Rev. 96 (1996) 877.
[
12] (a) C.E.J. Van Rensburg, E. Kreft, J.C. Swarts, S.R. Dalrymple, D.M. Macdonald,
M.W. Cooke, M.A.S. Aquino, Anticancer Res. 22 (2002) 889;
of THF can append results from CH
comes positively charge upon reduction. In CH
B(C the complex [Ru -FcCOO) CH OH)
2 2
Cl , especially if the analyte be-
(
b) M.W. Cooke, C.A. Murphy, S.T. Cameron, J.C. Swarts, M.A.S. Aquino, Inorg.
Chem. Comm. 3 (2000) 721;
c) M.W. Cooke, T.S. Cameron, K.N. Robertson, J.C. Swarts, M.A.S. Aquino,
n
2
Cl
2
/[N( Bu)
4
]
(
[
6
F
)
5 4
]
2
(l
4
ꢁ(CH
3
2
2
][PF ], 2,
6
Organometallics 21 (2002) 5962.
+
exhibited all four Fc/Fc couples in well-resolved form, while in
[
13] M.J. Cook, I. Chambrier, G. White, E. Fourie, J.C. Swarts, Dalton Trans. (2009)
1136.
ꢀ
the presence of PF6 , they coalesced into two groups of electron
[
[
14] K.C. Kemp, E. Fourie, J. Conradie, J.C. Swarts, Organometallics 27 (2008) 357.
15] (a) I. Chambrier, D.L. Hughes, J.C. Swarts, B. Isare, M.J. Cook, Chem. Commun.
transfer processes. For the biscadmium trisphthalocyaninato com-
plex 3, all six possible oxidation processes could be identified. Sol-
vent coordination by THF and ion pair formation between
3
504 (2006);
(b) I. Chambrier, G. White, M.J. Cook, Chem. Eur. J. 13 (2007) 7608.
[16] R.J. Lesuer, C. Buttolph, W.E. Geiger, Anal. Chem. 76 (2004) 6395.
17] (a) C. Creutz, H.J. Taube, J. Am. Chem. Soc. 91 (1969) 3988;
b) W.E. Geiger, N. Van Order, D.T. Pierce, T.E. Bitterwolf, A.L. Reingold, N.D.
ꢀ
positively charged oxidised species of 3 and PF6 distorted the oxi-
[
dation wave pattern. The most meaningful oxidation studies with
(
n
positively charged complexes were performed with [N( Bu)
4
]
Chasteen, Organometallics 10 (1991) 2403;
(c) N. Van Order, W.E. Geiger, T.E. Bitterwolf, A.L. Reingold, J. Am. Chem. Soc.
n
[
6
B(C F
)
5 4
] and [N( Bu)
4
][B{C
6
3
H (CF
3
)
}
2 4
] as supporting electrolyte.
109 (1987) 5680;
The wide range of twelve reduction processes was best resolved
in THF.
(
(
d) D.T. Pierce, W.E. Geiger, Inorg. Chem. 33 (1994) 373;
e) W.L. Davis, R.F. Shago, E.H.G. Langner, J.C. Swarts, Polyhedron 24 (2005)
The ferrocenyl group of b-diketone 4 exhibited reversible
1611;
f) W.C. Du Plessis, J.J.C. Erasmus, G.J. Lamprecht, J. Conradie, T.S. Cameron,
M.A.S. Aquino, J.C. Swarts, Can. J. Chem. 77 (1999) 378;
g J. March, Advanced Organic Chemistry, fourth ed., John Wiley and Sons, New
York, 1992, pp. 17–20, 263–269, 273–275.
(
electrochemistry in CH
3
CN as well as in CH
2
Cl
2
. The ruthenocenyl
CN/
fragment exhibited irreversible electrochemistry in CH
3
n
IV
[
N( Bu)
4
][PF
6
]. Results were consistent with formation of the Ru
IV
2+
n
[18] (a) J.C. Swarts, E.H.G. Langner, N. Krokeide-Hove, M.J. Cook, J. Mater. Chem. 11
species [(C
5
4 3 2 2 4 6 5 4
H R)(Cp)Ru (CH CN)] . In CH Cl /[N( Bu) ][B(C F ) ],
(
(
2001) 434;
however, clear evidence for the electrochemically reversible gener-
b) E. Fourie, J.C. Swarts, I. Chambrier, M.J. Cook, Dalton Trans. (2009) 1145;
III
+
ation of [(C
H
5 5
)Ru (C
5 4
H
R)] in chemical equilibrium with a
(c) M. L’Her, A. Pondaven, in: K.M. Kadish, R. Guillard, K.M. Smith (Eds.), The

2
232
H.J. Gericke et al. / Inorganica Chimica Acta 363 (2010) 2222–2232
Porphyrin Handbook, vol. 16, Academic Press, Elsevier Science, USA, pp. 117–
69.
19] T. Kuwana, D.E. Bublitz, G. Hoh, J. Am. Chem. Soc. 82 (1960) 5811.
(1996) 115;
1
(c) T.P. Smith, D.J. Iverson, M.W. Droege, K.S. Kwan, H. Taube, Inorg. Chem. 26
(1987) 2882.
[
[
20] (a) M. Watanaba, I. Motoyama, T. Takayama, M. Sato, J. Organomet. Chem. 549
[21] J.C. Swarts, A. Nafady, J.H. Roudebush, S. Trupia, W.E. Geiger, Inorg. Chem. 48
(2009) 2156.
[22] M.W. Droege, W.D. Harman, H. Taube, Inorg. Chem. 26 (1987) 1309.
(
(
1997) 13;
b) M. Watanaba, I. Motoyama, M. Shimoi, H. Sano, J. Organomet. Chem. 517