INTER~IEDIATES OF ELECTRODE REACTIONS

OF COORDINATION COMPOUNDS
ANTONiN A. V L<JEK

J. Heyrovskj Polaragraphie Institute,

Czechoslovak Academy of Science, Prague, C.S.S.R.

The redox properties of coordination compound.s, i.e. their oxidation-

reduction ability (given by the standard equilibrium potential), their
oxidation-reduction reactivity (given by the rate with which they are a'ble to
react with various substances) and finally the mechanism of the particular
reaction (given by the nature ofthe intermediates and productc;) represent a
very important group of chemical properties of coordination compounds,
closely related to their atomic composition and especially to their electronic
structure. The redox properties of coordination compounds of transition
metals have attracted a lot of interest since, unlike most organic and many
inorganic compounds, they are able to undergo stepwise redox changes,
passing through various more or less stable intermediates. The stability and
configuration of these intermediates reflect the changes induced in the
electronic configuration of the particle by the change of the nurober of
electrons in the course of the redox process.
There are two main ways of studying the redox properties:
(I) The redox change in a compound is induced by the addition of some
other substance, the redox agen t, and the properties-kinetic or thermo-
dynamic--of the system are studied by various physico-chemical methods.
In this case the reaction with the electron donating or accepting agent
proceeds in the whole volume of the solution, i.e. a homogeneaus redox
reaction is actually being followed.
(2) The redox change is induced by an inert polarized electrode which
is able to donate or accept the electrons from the components of the solution.
In this case the redox reaction proceeds a t the boundary between the metal
and solution only, so that a heterogeneaus redox process is actually followed.
If the specific properties of a heterogeneaus process are omitted for the
present, a polarized electrode can be regarded as a redox agent and the
homogeneaus and electrolytic redox processes can be correlated on the same
basis. In fact, there are very striking parallels between the ra tes and
mechanisms of homogeneaus and electrolytic redox reactionsl.
An electrode as the redox agent has, however, some specific properties in
which it differs from the "homogeneous" redox agents and which must be
taken into consideration when an homogeneaus and an electrolytic process
are compared. First of all it can be supposed, and all the experiments seem
to confirm it, that there is almost no activation energy connected with the
transport of the electron to or from the electrode; so far purely electrostatic
effects connected with the potential drop at the interfacearenot considered~
61
 ANTONIN A. VLCEK
This means that the rate and mechanism of the redox process, taking place
at a certain potential, depends solely upon the structure and properties of
the particles reacting with the electrode.
Secondly, the redox ability of a polarized electrode depends upon the
potential impressed to it from an outer source. This means that the redox
"strength" of an electrode can be varied in rather broad range, and that the
process taking place at the electrode is in most cases the minimal redox
change possible for the given compound. As an example let us compare the
reduction of the dipyridyl(dipy) complex Mn(dipy)32+, at the d.ropping
mercury electrode 2 and by various reducing agents in non-aqueous solvents
(for summary see Herzog3) : at the dropping mercury electrode a reduction to
Mn(dipy)s+ can be seen quite clearly at -1·5 V using 1M Iithium chloride
in ethanol as solvent. This reduction is followed by the reduction of d.ipyridyl
which starts at about -1·6 V. The homogeneaus reduction has resulted in
the formation of the 0 or -1 formal state of manganese only. This is due to
the fact that there is no homogeneaus reducing agent available, the strength
of which would be exactly that needed for the formation of the + 1 state.
However, as the polarographic behaviour shows, the 1.1n(dipy) 3 + complex
is quite stable and might be prepared, e.g. by controlled preparative elec-
trolysis.
The third, and most important specific feature of an electrode is connected
with the heterogeneaus nature of its action: the redox process at the electrode
is manifested by the charge transfer across the electrode-solution boundary.
The current measured at a given potential depends upon the nurober of
electrons which cross this boundary in a unit of time: if this rate is low, no
electrode process is detected even if the redox ability of the electrode is
sufficient from the thermodynamic point of view. The redox process then
takes place, if at all, at a morenegative (or positive) potential at which its
rate increases to such an extent that the number of electrons transferred in a
unit of time is considerable. [This statement is rather oversimplified.
Actually, the measure of the extent to which the electrode reaction proceeds
at a given potential is given by the ratio of the actual current to the hypo-
thetical current which would flow through the electrode at the given
potential for an infinitely rapid electrode reaction. If this ratio is small, the
electrode reaction proper is slow and the potential of the process (expressed
mostly as the half-wave potential) is more negative (or positive for oxida-
tions) than the corresponding standard equilibrium potential of the system.]
In homogeneous reactions, on the other hand, when ·sufficient time is given
to the reaction mixture, most processes take place in accordance with
thermodynamic requirements. This difference between the homogeneaus
and electrolytic reactions might result in great differences in the mechanism
of these two types of processes. As an example the reduction of VOZ+ can
be quoted: under homogeneaus conditions it is possible to reduce V0 2+ to
V(III) quite easily. At the dropping mercury electrode, on the other hand,
the reduction of V02+ proceeds very slowly at the potential corresponding
to the V02+ -7 V(III) standard potential; it takes place mainly at potentials
which are more negative than that of the V (III) -7 V (II) reduction so that
divalent vanadium is the product of the V0 2 + reduction at the electrode.
However if the redox inertness of V02+ is removed (by increasing the
62
 INTERMEDIATES OF ELECTRODE REACTIONS OF COORDINATION
acidity of the solution in the presence of some anionic complexing agents,
e.g. Cl- ions4) the COUrses ofthe homogeneaus and the electrolytic reductions
are identical4.
So far only the so-called direct electrode reaction has been considered,
i.e. a reaction which involves merely a change in the nurober of electrons in
the particle, and no simultaneaus change in the atomic configuration.
Such a reaction proceeds at electrodes with a short time of electrolysis
(as when using a dropping or streaming mercury electrode) only if a potential
can be reached at which the rate of the process is sufficiently high. At
large electrodes with prolonged electrolysis, such a reaction takes place with
a very low current efficiency.
Most of the redox processes are, however, not direct electrode reactions
but consist of a series of chemical reactions, preceding or following the
electrode reaction proper. If a direct electrode reaction of D (depolarizer)
giving P (primary product) is symbolized as:

D
V
p

a general electrode process might then be represented by a scheme:

D1 ~n2 ...... Dn
{}
P~Fl~F2 ... Fn

where D1, D2 . .. Dn are various species interconverted by chemical changes,

but without any redox change. Dn is the species with the greatest redox
reactivity, i.e. a species which at the given potential reacts with the greatest
rate at the electrode. F1, F2 ... Fn are various final products of the electrode
process, interconverted by chemical changes which rnight include homo-
geneous redox changes such as disproportionation or redox reaction~ with
some component of the solution.
vVhat has been said for the direct electrode reaction holds also for all these
chemical changes: "A chemzcal reaction can influence the electrode process only if its
rate is sujficiently high". At a dropping mercury electrode the lowest limit for
the monomolecular reaction is at k = 101 to 1021. mol-1 sec-1. If a reaction
has a lower monomolecular rate constant it does not influence the mechanism
ofthe electrode process at all. From this it follows that an electrode reaction
might proceed along a different path from that of a homogeneaus redox
process. This is true especially for the interconversions of the products and
intermediates: if these reactions are slow the electrode process (at least when
studied by means of electrochemical methods) either stops a t a product
which is not thermodynamically the most stable one, or, at least the reaction
of this product becomes the rate determining step and can be detected by the
analysis of the experimental data. lf it is remernbered that a reaction is
regarded from this point of view as "slow", even if it proceeds with a rate
constant of the order 1()3 or 1041. mol-1 sec-1, it can easily be seen that
63
 ANTONiN A. VLÖEK
many unstable intermed.iates, mostly those which are inaccessible by homo-
geneous methods, can be detected as products of electrode processes.
From many examples which might be quoted, the reduction ofNi(CN) 4 2-
and Ni(d.ipy)32+ demonstrate this situation best (see Figure 1): homogeneous
reduction ofNi(CN) 4 2- in aqueous solution results in all cases in the focma-
tion of Ni2(CN)64 -. In liquid ammonia, however, the reduction with
solutions of alkali metals gives the zero-valent complex, Ni(CN) 44-. The
study of the electrode process shows5 (see Figure 1) that even in aqueous

Ni (CN)f <==::::::;>

Ni(m)
+2 0

Ni (dipy)~+<:===> Ni(dipy)~

+
Ni (f_p)
+2 0
Figure 1. Simplified mechanisms for the reduction ofNi(CN)42- and Ni(dipy)s2+

sol ution the zero-valent complex is formed as the primary product, but is
converted very rapidly into Ni2(CN)64- by reaction with water.
~i(d.ipy)3 2 + is reduced at the electrode to the complex of zero-valent
nicke!, which rapidly decomposes to metallic nickel6. Homogeneous

h: 39 53 76 104 130cm

Figure 2. Oxidation of V(benzene)z at a dropping mercury electrode (dependence upon the

height of mercury reservoir). Solution: 1M LiCl04 in ethanol-benzene mixture (3:1),.
1·5 X 10-:1 M V(C6H6)z; curves start at +0·4 V (aqueous 4 M NaCl calomel electrode);
100 mVfabsc.
64
 INTERMEDIATES OF ELECTRODE REACTIONS OF COORDINATION
reduction in aqueous or alcoholic solutions results in all cases in the formation
of metallic nickel only.
The oxidation of V (benzene)2 and V (diphenyl)2 complexes represents
examples of complicated reactions of intermediates7. Both complexes
mentioned give three waves, the sum ofwhich corresponds to a four electron
oxidation of the vanadium. In fact, V0 2 + is detected as the main final
product. At low temperatur es the first wave corresponds to a two electron
oxidation (see Figure 2). At high temperatures the height of the wave
increases above the value for two electrons, which shows that the product
of the two electron oxidation undergoes a redox process, part of vanadium
being oxidized to a higher state than + 2. The complex formed by the two
electron oxidation undergoes some further chemical change-its nature is
not known as yet-in which a new species in the same oxidation state
( +3) is formed and which is more easily oxidized than the primary product.
The analysis of the polarographic behaviour leads to a scheme for the
mechanism of oxidation which is summarized in Figure 3. The reactions of

(0
.....
6
I

a.
Qj
ti

~·0·25
..... • k_, ~· 0 05

~--------~----------~
------vo2•

•composition unknown k
X-anion a -e

Figure 3. Oxidation mechanism of V(benzene)2 in ethanol-benzene mixtures (preliminary)

the divalent species are rather slow (the disproportionation reaction is

actually at the detectable Iimit).
An even more complicated mechanism, involving various intermediates
and products which are rapidly interconverted, has been found for the
reduction of [Co(CN)sX]3- complexesS, 9. These complexes give a series of
waves (see Figure 4), the most positive one corresponding always to a one
electron reduction to give the divalent cobalt complex. The behaviour
of this divalent species does not depend upon the nature ofX (when X = Cl-,
65
P.A.C.-F
 ANTONIN A. VLCEK
Br-, I-), but changes considerably with the pH ofthe solution. Analysis of
this behaviour shows that a complex with four CN- ions is formed, which is
more easily reduced than the species [Co(CN) 5]3-. At low pH values, a

pH: 7·13 8·21 11·49

A
8 A A
c B

[Co (CN)s Br] 3 ~[.Co (CN lJ 3"+ Br-

A [Co (CN)51 3
-~ [Co(CN)s] 4" ~ ß:o(CN)sH] 3"+ OH-
B ß.:o (CN)sJ3·- [Co(CN)4] 2• ===- [Co(CN) 4] 3:.Q![co(CN)5]4•
c [Co(CN) 5]3---':!.:.. [Co cNr ==>Co(m)
Figure 4. Schematic representation of the dependence of reduction waves of [Co(CN)sX]S-
complexes upon pH. (The processes taking place in various steps are indicated below)

further splitting off of CN- ligands takes place, resulting in the formation of
the [Co(CN)]+ complex. No intermediate stages with two or three CN- ions
have as yet been detected. The species with four and five CN- ligands are
both reduced to a complex of univalent cobalt which, however, reacts
with hydrogen ions giving an electrolytically inactive compound. We
know that this complex contains five CN- ions and one proton, so that it
2e
============?
lke
Co(m)

+3 +2 +1
Figure 5. Simplified schemes for the reduction of [Co(CN)sX]3- complexes (X= Cl-, Br, I-)
66
 INTERMEDIATES OF ELECTRODE REACTIONS OF COORDINATION
can bc assumed to be identical with the complex [Co(CN)sH]3- which is
supposed to be formed by reduction of Co(II)-cyanide complexes with
borohydride or hyd.rogenlO. Figure 5 shows the simplified scheme of the
reduction of [Co(CN) 5 X]3- complexes. (The full scheme involves some
other intermediates in the +2 state and the oxidation of the Co(CN)s3 -
complex.) Using this scheme, and the properties ofvarious species detected
and studied polarographically, it is possible to explain most ofthe reactions
of Co(II)-cyanide complexes.
Many redox reactions can be supposed to proceed via a very active inter-
mediate which is being inactivated in a reaction following the redox reaction
proper. Let us consider the following system:

B
0
?I t
EB
/ K = [B]
A k
[C]
~

~Ic
Eco

the properties ofwhich are: ( -EB 0 ) > ( -Ec 0 ) , K < 1; Bke ~ cke. Band C
are either various electronic isomers, or differ in atomic composition, etc.
For such a system the equation holds: -Eco = -EB 0 - (RTfnF) In K.
In homogeneaus reaction this system always reacts as a system with
standard potential equal to Ec 0 , i.e. A can be reduced even with redox
agents having lower standacd potentials than Eß 0 • At the electrode, how-
ever, the reduction will proceed via B (due to the difference in rate constants
of the two possible reduction paths), and the total behaviour will depend
upon the rate of the interconversion of B into C. If k is small, the process
A => B can be followed directly.
Such a behaviour, complica ted however by other reactions, has been
observed where A is [Cr(CN) 5N0]3-. This ion shows a one electron
reduction a t -1·18 (this reduction was first observed by Gri:ffith, Lewis and
Wilkinsonll). However, as the pH ofthe solution decreases, the height ofthe
wave increases above the value for one electron, reaching the value for a two
electron reduction at pH ~ 212. At the samc time, the half wave potential
shifts to morepositive potentials (see Figure 6). In acidic solutions a second
three electron wave is obtained at morenegative potentials. The reduction
mechanism of the whole system is shown on Figure 712 • As the most impor-
tant feature of this mechanism, it must be emphasized that the species
[Cr(CN)sN0]4-, the primary product of the one electron reduction, is not
reducible as long as it does not accept two protons. The polarographic
behaviour shows that the primary product might be stable in solutions
having low hydrogen ion concentration, and in fact we succeeded in pre-
paring the compound [Cr(CN)5N0]4- by controlled electrolysis in 0·1 M
sodium hydroxide solution (Figure 8). At pH <9·5 no reduction product
can be detected in the solution, decomposition taking place.
67
 ANTONiN A. VLCEK
In the course ofthc study ofthe homogeneaus reduction of [Cr(CN) 5 N0]3-
by Co( dipy)a+ 1 it was, however, found that the univalent cobalt is easily
oxidized by the chromium complex, in spite of the fact that the redox
potential of the chromium coupleis about 250 mV morenegative than that
of the Co(II)-Co(I) dipyridyl couple. Another striking result was found

3
[er (CN) 5 N0] -_ pH dependence

2 o-n
• -E,,2

1·2

·---.--·-
1·1

0 2 4 6

Figure 6. Dependence of the number of accepted electrons (n) and of the half-wave potential
(E1 12) upon pH for the reduction of [Cr(CN)sN0] 3 -

[Cr(CN) 5 NO·H]3-

Figure 7. Mechanism for the reduction of [Cr(CN)5N0] 3 -

when studying this reaction. The Co(I) complex is completely oxidized

even by sub-stoichiometric amounts of the [Cr(CN)sN0]3- compound
(100:1 !), and after the reaction the original amount of [Cr(CN)sN0]3-
can still be detected in the solution. This mea 1sthat the chromium complcx
actually c·atalyses the oxidation of the Co(I) dipyridyl complex at the
68
 INTERMEDIATE S OF ELECTRODE REACTIONS O:F COORDINATIO N
expense of hydrogen ions. The study of the mechanism of this process has
shown that the reaction actually proceeds via the [Cr(CN)5N0] 4- species
according to a mechanism which (simplified and omitting intermediate
steps) can be represented as:
Co(dipy)3+ + [Cr(CN)5N0]3-~Co(dipy)32+ + [Cr(CN)5N04] - (I)

[Cr(CN)5N0]4 - + H+-+ [Cr(CN)5N0]3 - + H (2)

The equilibrium (1) is strongly shifted to the Ieft, but as the reduced
chromium species is removed by protonization and subsequent re-oxidation,

1 I
Reduction of [Cr~CN) 5 N0] 3 -
i =0

Figure 8. Electrolytic reduction of [Cr(CN)5NQ]3- in 0·1 M NaOH solutions; curve 1-10-3 M

[Cr(CN)5N0]3-; curve 2-After 30 min of electrolysis at a mercury pool electrode, E = 1·4 V;
curve 3-CN- ions liberated by decomposition of the product

all the Co(I)-dipyridyl complex is consumed without any actual change in

the concentration of [Cr(CN)5N0]3 -. The protonization of the
[Cr(CN) 5N0]4- species proceeds rather rapidly, but the intramolecular
redox process which follows this protonization is relatively slow and is
"cut off" from the electrode reaction which thus proceeds by a completely
different path from that of the homogeneous redox reaction. However,
when very strong reducing agents are used, the homogeneous reaction
results in the same products as the electrode reaction12,
The reduction of [Cr(CN)sNOJ3 - thus represents a typical example ofthe
general type of redox process discussed above and, to my knowledge, the
only one known at present.
It would be possiole to give many more examples of systems in which the
69
 ANTONIN A. VLCEK
paths of electrolytic and homogeneaus redox reactions differ. I hope,
however, that the examples quoted here suffice to Blustrate the most im-
portant ways in which electrode redox processes differ from homogeneaus
ones. As has been seen, when the specific properties of an electrode as a
redox agent are taken into account, the differences between the two types of
processes can be rationalized.

References
1 A. A. Vlcek. Proc. Inorg. Chem. (Ed. F. A. Cotton). Vol. 5, p. 211, Interscience, New
York (1963).
2 A. A. Vlcek. Unpublished results.
3 S. Herzog and R. Taube. Z. Chem. 2, 208 (1962).
4 J. Suchomel. Thesis, Polaragraphie Institute, Prague 1958.
5 A. A. Vlcek. Collection Czech. Chem. Commun. Suppt. 22, 948, 1736 (1957).
6 A. A. Vlcek. Z. Elektrochem. 61, 1014 (1957).
7 A. A. Vlcek and H. P. Schröer. Naturwiss. In the press.
8 H. Matschiner. Thesis, Polaragraphie Institute, Prague 1962.
9 A. A. Vlcek, H. Matschiner, and V. Kuznecov. Nature. In the press.
10 W. P. Griffith and G. Wilkinson. J. Chem. Soc. 1959, 2757.
11 W. P. Griffith, J. Lewis, and G. Wilkinson. J. Chem. Soc. 1959, 872.
12 D. Bustin and A. A. Vlcek. Proceedings of the Symposium on Structure and Properlies qf Co-
ordination Compounds, Bratislava 1964; Collection C;:,ech. Chem. Commun. In the press.

70