Low temperature stopped-flow studies in inorganic chemistry

Article abstract of DOI:10.1039/b107927c

Low temperature stopped-flow methods together with diode array instrumentation have become extremely useful for studying reactions that are too fast at ambient temperatures and/or for detecting reactive intermediates that can only be observed at low tempe

Full text of DOI:10.1039/b107927c

View Article Online / Journal Homepage / Table of Contents for this issue
Low temperature stopped-flow studies in inorganic chemistry†
Markus Weitzer,a Markus Schatz,a Frank Hampel,b Frank W. Heinemanna
and Siegfried Schindler*a
a Institute of Inorganic Chemistry, University of Erlangen-Nürnberg, Egerlandstrasse 1,
91058 Erlangen, Germany. E-mail: schindler@chemie.uni-erlangen.de
b Institute of Organic Chemistry, University of Erlangen-Nürnberg, Henkestrasse 42,
91054 Erlangen, Germany
Received 3rd September 2001, Accepted 14th November 2001
First published as an Advance Article on the web 25th January 2002
Low temperature stopped-flow methods together with diode array instrumentation have become extremely useful for
studying reactions that are too fast at ambient temperatures and/or for detecting reactive intermediates that can only
be observed at low temperatures. Furthermore, global analysis fitting methods are described that allow the analysis of
complex reaction mechanisms. To illustrate the method the reaction of dioxygen with the tripodal copper() complex
[Cu(Me2-uns-penp)(CH3CN)]ClO4 (Me2-uns-penp = (2-dimethylaminoethyl)bis(2-pyridylmethyl)amine) is discussed
as an example.
copper() complexes react with dioxygen.7,19 The course of these
Introduction
reactions depends on the temperature, ligand and solvent.7,19
In the last decade, mainly as a consequence of the advances in
Therefore, it is of great interest to elucidate the reaction
computer technology, instrumentation for kinetic investigations
mechanisms of the binding and activation of dioxygen by
has improved dramatically. For example a stopped-flow unit
copper() complexes.
in the late eighties consisted of a large number of different
The first example of a structurally characterized copper
parts, which included: the syringe system, a lamp together with
a monochromator, an oscilloscope for viewing the data and a
transient recorder for collecting the data and for their transfer
to a computer for data fitting. Many problems occurred during
these measurements and they were time consuming because
programs for analysing the data were still slow once absorbance
vs. time traces were observed that could not be fitted to a single
exponential function. Furthermore, only measurements at
single wavelengths were possible, therefore making it often
necessary to repeat the analysis at different wavelengths.
peroxo complex was obtained by Karlin and co-workers from
the reaction of [Cu(tmpa)(CH3CN)]PF6 (tmpa = tris[(2-pyridyl)-
methyl]amine, Scheme 1) with O2 at low temperatures.21,22
Today a stopped-flow instrument consists of only the unit
itself combined with a diode array setup and a computer allow-
ing fast kinetic measurements of time resolved UV-vis spectra
under anaerobic conditions, high pressure and/or low temper-
atures. Improvements have been made here as well: e.g. syringes
are installed vertically instead of horizontally (to avoid prob-
lems with gas bubbles) and peek (polyetheretherketone) is used
instead of Teflon for valves and flow tubes to improve the
anaerobic capabilities of the instrument. Furthermore, the
syringe drives setup was optimized. Companies offering
modern stopped-flow equipment are Hi-Tech Scientific
(Salisbury, UK), Applied Photophysics Ltd. (Leatherhead,
UK), Bio-Logic-Science Instruments SA (Claix, France), Olis
Incorporated (Bogart, GA, USA) and KinTek Corporation
(Austin, TX, USA).
Scheme 1
A detailed kinetic investigation of the reversible reaction of
[Cu(tmpa)(CH3CN)]ϩ with O2 was performed at low temper-
atures in propionitrile, which allowed the spectroscopic
observation of a superoxo complex prior to the formation of
the peroxo complex.23 Furthermore, the influence of steric
hindrance as well as different donor atoms in the ligand on the
formation of the peroxo complexes was studied.23,24 Recently,
we have investigated the effect of the chelate ring size on the
properties of these complexes.8
We discovered earlier that similar to the reaction of
[Cu(tmpa)(CH3CN)]ϩ with O2, the copper() complex of the
aliphatic ligand Me6tren (= tris(2-dimethylaminoethyl)amine,
Scheme 1) also supports the formation of a copper superoxo
We employ extensively the stopped-flow technique for the
investigation of the reaction of dioxygen with copper() com-
plexes.1–9 Such compounds can be regarded as low molecular
weight model complexes for a large number of copper proteins
involved in the redox processing of dioxygen.7,10–19 They not
only provide a better understanding of the biological mole-
cules but also assist in the development of new homogeneous
catalysts for selective oxidations under mild conditions.20
A whole variety of copper dioxygen adducts can form when
† Based on the presentation given at Dalton Discussion No. 4, 10–13th
January 2002, Kloster Banz, Germany.
686
J. Chem. Soc., Dalton Trans., 2002, 686–694
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b107927c

View Article Online
Table 1 Crystal data for [Cu(Me2-uns-penp)(CH3CN)]ClO4 (1) and [Cu(Me2-uns-penp)Cl]ClO4 (2)
1
2
Empirical formula
Formula weight
Temperature/K
Crystal system
Space group
a/Å
b/Å
C18H25ClCuN5O4
474.42
173(2)
C16H22Cl2CuN4O4
468.82
295
Monoclinic
C2/m
13.723(1)
19.224(2)
7.776(1)
101.90(1)
2007.3(3)
4
Triclinic
¯
P1
7.66920(10)
11.1356(2)
13.1908(2)
78.0720(10)
1037.39(3)
2
c/Å
β/Њ
Volume/Å3
Z
Dcalc./Mg mϪ3
1.519
1.216
1.551
1.383
µ/mmϪ1
F(000)
492
964
Crystal size/mm3
Scan technique
θ range for data collection/Њ
Index ranges
0.40× 0.35 × 0.30
0.60 × 0.35 × 0.30
ω-scan
3.7 to 29.0
Ϫ4 ≤ h ≤ 18, Ϫ4 ≤ k ≤ 26, Ϫ10 ≤ l ≤ 10
4486
2756, 0.0237
1931
2756/0/141
1.047
0.0478
0.1288
1.94 to 27.47
Ϫ9 ≤ h ≤ 9, Ϫ14 ≤ k ≤ 14, Ϫ17 ≤ l ≤ 17
8224
Reflections collected
Independent reflections, Rint
Observed reflections [Fo > 4σ(F)]
Data/restraints/parameters
Goodness of fit on F 2
R1 [Fo > 4σ(F)]
wR2 (all data)
4737, 0.0174
4737/0/362
1.085
0.0303
0.0924
0.376 and Ϫ0.498
Largest diff. peak and hole/e ÅϪ3
0.577 and Ϫ0.322
and peroxo complex at low temperatures.3,4,7 In a detailed
mechanistic study we observed clear differences in the kinetic
behaviour during the oxidation of the two complexes.25 There-
fore, we became interested in the investigation of the copper()
complexes of the “mixed ligands” Me2-uns-penp (= (2-dimethyl-
aminoethyl)bis(2-pyridylmethyl)amine) and Me4-apme (= bis(2-
dimethylaminoethyl)(2-pyridylmethyl)amine) shown in Scheme
1. Herein we present our first results on the kinetic investigation
of the copper() complex [Cu(Me2-uns-penp)(CH3CN)]ClO4
with dioxygen in regard to the application of low temperature
stopped-flow techniques and advanced data fitting methods.
Further examples of this technique can be found in the
literature.26–35
As discussed above, the ligand Me2-uns-penp can be regarded
as a “mixture” of the two ligands tmpa and Me6tren and there-
fore a comparison of the crystal structures of the copper com-
plexes of these ligands with [Cu(Me2-uns-penp)(CH3CN)]ClO4
is interesting. [Cu(tmpa)(CH3CN)]ϩ crystallizes in the same way
as [Cu(Me2-uns-penp)(CH3CN)]ϩ with an acetonitrile mole-
cule as additional ligand in the axial site (Cu–N: 1.99(1) Å
in [Cu(tmpa)(CH3CN)]ϩ,36 Cu–N: 2.038(2) Å in [Cu(Me2-uns-
penp)(CH3CN)]ϩ). The coordination geometry of both
complexes is best described as slightly distorted trigonal bipyr-
amidal. The copper() ion is displaced 0.508 Å out of the
N5–N12–N22 trigonal plane in [Cu(Me2-uns-penp)(CH3CN)]ϩ
and 0.55 Å in [Cu(tmpa)(CH3CN)]ϩ.
In contrast, the copper() complex of Me6tren shows a quite
different crystal structure.4 Here no acetonitrile molecule is
coordinated in the solid state, however, a perchlorate ion shows
a weak interaction with the copper ion. The copper ion in
[Cu(Me6tren)]ϩ only lies slightly below the trigonal plane
(0.191(8) Å) towards the perchlorate ion. The reason for this is
that the soft aromatic nitrogen atoms of the pyridines form
stronger bonds with the soft copper() ion than the hard
aliphatic amine donor atoms. As a result the bond lengths to
the axial and the equatorial nitrogens in [Cu(Me6tren)]ϩ are
very similar (2.122(7) and 2.200(14) Å) in contrast to the other
two complexes where the bond length to axial nitrogen is
Results and discussion
Synthesis and characterization of copper complexes
The copper() complex of Me2-uns-penp was prepared by
mixing stoichiometric amounts of the amine and [Cu(CH3-
CN)4]ClO4 in acetonitrile under argon. An ORTEP55 plot of
the cation of [Cu(Me2-uns-penp)(CH3CN)]ClO4 (1) is shown
in Fig. 1 and a summary of the crystallographic data and
refinement parameters can be found in Table 1. Selected bond
lengths and angles are reported in Table 2.
significantly longer (Cu–Naxial = 2.424(2) Å and Cu–Nequ.
=
2.061(2)–2.208(2) Å). Therefore, the crystal structure of
[Cu(Me2-uns-penp)(CH3CN)]ϩ more closely resembles the
structure of [Cu(tmpa)(CH3CN)]ϩ than that of [Cu(Me6tren)]ϩ.
The copper() complex [Cu(Me2-uns-penp)Cl]ClO4 (2) was
prepared by mixing stoichiometric amounts of Cu(ClO4)2ؒ
6H2O, CuCl2ؒ2H2O and Me2-uns-penp in a mixture of meth-
anol and water (1 : 1). A summary of crystal parameters and
refinement results for this compound is given in Table 1. An
ORTEP plot of this complex is presented in Fig. 2 with selected
bond distances and angles in Table 2.
The molecule is situated on a crystallographic mirror plane
and the coordination geometry is best described as distorted
square pyramidal. This is surprising because the related com-
plexes [Cu(tmpa)Cl]ClO4 and [Cu(Me6-tren)Cl]ClO4 are both
trigonal bipyramidal in the solid state.
The square pyramidal geometry is not completely retained
Fig. 1 View of the cation of [Cu(Me2-uns-penp)(CH3CN)]ClO4.
J. Chem. Soc., Dalton Trans., 2002, 686–694
687

View Article Online
Table 2 Selected bond distances (Å) and angles (Њ) for [Cu(Me2-uns-penp)(CH3CN)]ClO4 (1) and [Cu(Me2-uns-penp)Cl]ClO4 (2)
1
Cu(1)–N(1)
Cu(1)–N(5)
Cu(1)–N(12)
2.4236(14)
2.2608(15)
2.0932(15)
Cu(1)–N(22)
Cu(1)–N(100)
2.0613(14)
2.0380(17)
N(1)–Cu(1)–N(5)
N(1)–Cu(1)–N(12)
N(1)–Cu(1)–N(22)
N(1)–Cu(1)–N(100)
N(5)–Cu(1)–N(12)
76.53(5)
76.10(5)
76.04(5)
173.79(6)
101.64(5)
N(5)–Cu(1)–N(22)
N(5)–Cu(1)–N(100)
N(12)–Cu(1)–N(22)
N(12)–Cu(1)–N(100)
N(22)–Cu(1)–N(100)
124.98(5)
97.76(6)
116.46(6)
107.76(7)
105.86(6)
2
Cu(1)–N(1)
Cu(1)–N(2)
Cu(1)–N(3)
2.320(3)
2.058(3)
2.000(3)
Cu(1)–N(3)A
Cu(1)–Cl(1)
2.000(3)
2.252(2)
N(1)– Cu(1)–N(2)
N(1)–Cu(1)–N(3)
N(1)–Cu(1)–N(3)A
N(1)–Cu(1)–Cl(1)
N(2)–Cu(1)–N(3)
83.94(13)
96.32(7)
96.32(7)
100.94(9)
82.67(7)
N(2)–Cu(1)–N(3)A
N(2)–Cu(1)–Cl(1)
N(3)–Cu(1)–N(3)A
N(3)–Cu(1)–Cl(1)
N(3)A–Cu(1)–Cl(1)
82.67(7)
175.12(10)
159.5(2)
96.69(7)
96.69(7)
Fig. 2 View of the cation of [Cu(Me2-uns-penp)Cl]ClO4.
Fig. 3 Low temperature stopped-flow module.
in solution because in the UV-vis spectrum a very broad
band with an absorbance maximum at 780 nm was observed.
Square pyramidal complexes typically have an absorbance
maximum around 650 nm while for trigonal bipyramidal
copper complexes absorbance maxima between 850 and 900 nm
are observed.37,38
all tubing and connections. This avoided almost completely the
contact of the solutions with air (only a very small risk of air
contamination remained at the valves of the syringes). This
setup has been used successfully for many years by Zuberbühler
and coworkers. Later on Hi-Tech Scientific replaced the glass
connections with Teflon tubing. This reduced the risk of glass
breaking but introduced the problem of oxygen diffusion
through the Teflon tubes. In the beginning of our investigations
our instrument was equipped with such Teflon tubing and,
being aware of this problem, we tried to exclude contamination
with dioxygen in different ways. We could successfully follow
the oxidations of the copper() complexes with dioxygen, but we
had to learn that diffusion of air into the system still could not
be completely avoided. This was extremely unfortunate because
as a consequence it was not possible to perform quantitative
kinetic studies with this instrument. Therefore (in cooperation
with the company J&M, Aalen, Germany) we decided to replace
the Teflon tubing by glass in a different manner than in the
original instrument.
On top of the stopped flow unit commercially available
5 ml glass syringes (Fortuna Optima, Graf & Co. GmbH,
Wertheim, Germany) are connected by three-way Hamilton
valves (Hamilton Company, Reno, NV, USA) to the glass tubes
which then spiral downwards to the cuvette. Between the glass
tubing and the optical 1 cm quartz cuvette (specially manu-
factured for this instrument by Hellma GmbH & Co. KG,
Müllheim, Germany) a mixing jet is installed. The connections
are made tight with O-rings and flexible light fibers (J&M) are
attached with Teflon sealings at the cuvette. With this setup
contamination of the solutions with dioxygen can be com-
pletely avoided. An excellent test for the system was our kinetic
investigation of the reactions of the extremely dioxygen
sensitive Ni() complex [(bipy)Ni(COD)] with unsaturated
The low temperature stopped-flow technique
Recently, “copper dioxygen adducts” have been reported which
are quite stable at room temperature.39–42 Commonly such com-
pounds formed by the reaction of copper() complexes with
dioxygen are stable only at low temperatures.7,10,24,39 Typically,
for copper() complexes these reactions are usually quite fast
and need to be investigated using stopped-flow techniques.7,10,24
As a consequence several problems need to be overcome in
order to successfully study the kinetics of these reactions. The
instrument must allow measurements at temperatures as low as
Ϫ100 ЊC to be possible and simultaneously the connection and
tubing system has to be air tight and prevent the leaking of
the cooling media into the solutions of the reactants. Sealings,
O-rings, etc., are especially vulnerable because of the effects of
organic solvents and the large temperature changes involved.
Stopped-flow instruments commercially available today do not
incorporate these features and therefore need to be modified to
perform such kinetic studies. Unfortunately these modifications
usually do not allow work below temperatures of Ϫ40 ЊC. The
low temperature stopped-flow instrument SF-40 (latest model
number) manufactured by Hi-Tech Scientific (based on an
original design by Caldin, Crooks and Queen)43 suitable for
such measurements is, however, no longer commercially
available due to its production being stopped recently.
In our laboratory we employ a modified Hi-Tech SF-3 L
low temperature stopped-flow unit (Fig. 3). In the original
instrument only glass was used as the material for the syringes,
688
J. Chem. Soc., Dalton Trans., 2002, 686–694

View Article Online
substrates.26 A possible alternative to the glass tubing for
anaerobic measurements had been introduced by Hi-Tech
earlier, using either stainless steel or Teflon coated stainless steel
tubing. Here dioxygen contamination can also be excluded
completely after the oxygen in the tubing reacted (this setup has
been used successfully by other research groups).27,35
The two glass coils and the mixing chamber with the
cuvette are immersed in an ethanol bath which is placed in a
Dewar that is filled with liquid nitrogen for low temperature
measurements. The ethanol bath is cooled by liquid nitrogen
evaporation, and its temperature is measured by using a Pt
resistance thermometer and can be maintained to within 0.1
ЊC using a temperature-controlled thyristor power unit (both
Hi-Tech). With our diode array instrumentation complete
spectra can be collected between 190 and 620 nm or 350 and
1100 nm (both J&M).
At this point it should be pointed out that there are altern-
atives for the diode array instrumentation: for example auto-
matic repetitive measurements (scanning over each single
wavelength in a selected range) can be performed using an
Applied Photophysics Ltd. stopped-flow instrument. Most of
the time the diode array technique is superior to these altern-
atives but a problem can arise if photochemistry is observed
(filters can be used in some cases to suppress this).5,44
Fig. 4 Time resolved spectra of the reaction of [Cu(Me2-uns-
penp)(CH3CN)]PF6 with dioxygen in propionitrile at Ϫ90.1 ЊC
([complex] = 2.44 × 10Ϫ4 M, [O2] = 4.4 × 10Ϫ3 M, total time = 149.83 s,
four time bases (number of measurements 20/50/80/106, number of
intermissions 0/20/100/1000), integration time = 1.3 ms). Only every
seventh spectrum is shown.
oxo [(Me2-uns-penp)CuO2]ϩ (λmax = 412 nm) and peroxo
[(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ (λmax = 528 nm)
complexes accordingly.
Even at Ϫ90.1 ЊC the formation of [(Me2-uns-penp)CuO2]ϩ is
too fast to be followed spectroscopically under these conditions.
Only very few spectra showing the increase of the absorbance at
412 nm can be obtained using very short overall reaction times
(0.33 s) and therefore the majority of this reaction is missing
during the measurement. In contrast the formation of [(Me2-
uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ can be followed nicely by
observing the increase of the absorbance maximum at 528 nm.
This increase is accompanied by the decrease of the absorb-
ance of the superoxo complex at 412 nm. The reaction pathway
follows eqns. (1)–(3) as described previously for [(tmpa)Cu-
(CH3CN)]ϩ and [Cu(Me6tren)(CH3CN)]ϩ (R = CH3, C2H5; L =
tripodal ligand).4,23
Furthermore, it is worthwhile mentioning that it is possible
to perform stopped-flow measurements at low temperatures
and high pressures. However, it is a challenging task to study
dioxygen sensitive samples under these conditions.5,45
As dilute solutions of the copper() complexes are very
reactive towards dioxygen, solutions for the kinetic measure-
ments (usually 0.1–0.5 mM) were always prepared in an argon
glove box. Transfer of these solutions to the stopped-flow unit
was achieved by using 50 ml glass syringes (Fortuna Optima)
connected to a three-way Hamilton valve. No leakage of
dioxygen into the solutions in the syringes was observed.
Rinsing the instrument 3 to 4 times with the complex solution
prior to the measurements allowed the removal of all the
remaining oxygen.
Saturated solutions of dioxygen were obtained by bubbling
dioxygen through the required solvent for about 20 minutes
(while avoiding contamination with water) and then thermo-
statting the solution at 20 ЊC.23 The solution can then be
transferred to the stopped-flow instrument in a syringe with a
Hamilton valve. Dilution was accomplished by mixing the
dioxygen saturated solvent with the argon saturated solvent,
with the two syringes being connected by Hamilton valves.
Alternatively a gas mixing setup could be used.34 Air saturation
has been used in the past to obtain lower concentrations of
dioxygen but here it should be made certain that no reaction of
the reactants with carbon dioxide or nitrogen can occur.23
Kinetic data are obtained by measuring absorbance vs. time
data at a single wavelength or, much more preferable (especially
for more complex systems like the one described herein; see
below), by obtaining complete time resolved spectra using a
diode array system. Modern diode array systems allow a full
spectrum to be recorded every 0.8 ms. For our studies we use a
system from J&M.
(1)
(2)
[(L)Cu(O2)Cu(L)]2ϩ
irreversible decay
(3)
A comparison of the stability of the peroxo complexes
[(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ
[(tmpa)CuO2Cu-
,
(tmpa)]2ϩ and [(Me6tren)CuO2Cu(Me6tren)]2ϩ clearly shows
that, interestingly enough, [(Me2-uns-penp)CuO2Cu(Me2-uns-
penp)]2ϩ is more persistent at higher temperatures in contrast
to the other two peroxo complexes. As shown in Fig. 5a [(Me2-
uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ is still observable at 5 ЊC
for a few minutes in contrast to [(tmpa)CuO2Cu(tmpa)]2ϩ (Fig.
5b). [(Me6tren)CuO2Cu(Me6tren)]2ϩ is stable for a few seconds
under these conditions.
Prior to the experiment we expected to find the stability
of the peroxo complexes in the order of [(Me6tren)CuO2Cu-
The analysis of the reaction of [Cu(Me2-uns-penp)(CH3CN)]-
ClO4 with dioxygen
(Me6tren)]2ϩ > [(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ
>
[(tmpa)CuO2Cu(tmpa)]2ϩ. An explanation for the fact that
[(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ is more persistent
at higher temperatures than the other two peroxo complexes
might be a consequence of their different geometries. In con-
trast to the trigonal bipyramidal geometries of [(tmpa)CuO2-
Cu(tmpa)]2ϩ and [(Me6tren)CuO2Cu(Me6tren)]2ϩ with the
peroxo ligand occupying the axial position, the geometry of
[(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ is distorted square
pyramidal with the peroxo ligand occupying an equatorial
position. No crystal structure of the peroxo complex is available
To illustrate the data collection with such a low temperature
stopped-flow unit, the reaction of [Cu(Me2-uns-penp)(CH3-
CN)]ClO4 with dioxygen is presented below.46 Time resolved
spectra of this reaction in propionitrile at Ϫ90.1 ЊC are shown
in Fig. 4. These spectra provide clear evidence for the formation
of superoxo and peroxo complexes because the spectral
features are very similar to the UV-vis spectra of the copper
superoxo and peroxo complexes of the parent ligands tmpa and
Me6tren.4,23 We assigned the absorbance maxima of the super-
J. Chem. Soc., Dalton Trans., 2002, 686–694
689

View Article Online
and absorbance vs. time traces can be fitted to a single exponen-
tial function using programs such as Igor or Origin. These
programs immediately calculate kobs and also provide inform-
ation on the goodness of fit (plots of the residuals). Methods
that linearize the kinetic data by log plots instead of the
iterative methods are still applied but are much less accurate
and should no longer be used.47 According to eqn. (1) a plot of
kobs vs. [O2] would allow the calculation of the second order rate
constant for the forward reaction from the slope and the first
order rate constant for the back reaction from the intercept. In
principle this approach could be used for the reaction of
[Cu(Me2-uns-penp)(CH3CN)]ClO4 with dioxygen. Unfortun-
ately the formation of the superoxo complex is too fast to be
followed spectroscopically under these conditions (this fitting
method has been nicely described for the reaction of cobalt()
complexes with dioxygen34).
However, while the approach of using exponential functions
(to fit absorbance vs. time traces at single wavelengths) could
be applied to the formation of the superoxo complex this is
no longer possible for the formation of the peroxo complexes.
In this case a global analysis fitting routine is necessary and
programs specifically suitable for the analysis of kinetic data
have been commercially available for some years, often distrib-
uted together with a stopped-flow instrument (Pro/Kineticist
(Applied Photophysics Ltd.), SPECFIT (Hi-Tech Scientific)).
These programs represent a major advance in the capability of
data analysis and are extremely powerful. Their application in
kinetic studies is highly recommended. Most importantly they
allow one to globally analyse the complete set of wavelength/
absorbance/time data according to a proposed reaction scheme.
The mathematical background of these programs is described
in detail in the literature.48–52 In the following the fitting
procedure with SPECFIT/32 (authors R. A. Binstead, B. Jung
and A. D. Zuberbühler) is illustrated (Pro/Kineticist is based on
the same mathematical routines).
As discussed in the Introduction only recently has instru-
mentation become available for generating high quality sets
of time dependent spectra comprising of many thousands of
individual absorbance measurements. For stopped-flow kinetics
the use of this method only became possible due to the
improvement of diode array instrumentation, as discussed
above, allowing these data to be obtained on a very fast time
scale. SPECFIT provides the global analysis of equilibrium
and kinetic systems with expanded SVD (singular value
decomposition) and non-linear regression modelling by the
Levenberg–Marquardt method.48–52 These fitting routines make
it possible to fully analyse complete data sets.
Fig. 5 (a) Time resolved spectra of the reaction of [Cu(Me2-uns-
penp)(CH3CN)]PF6 with dioxygen in propionitrile at ϩ4.9 ЊC
([complex] = 2.44 × 10Ϫ4 M, [O2] = 4.4 × 10Ϫ3 M, total time = 177.66 s,
four time bases (number of measurements 20/40/80/116, number of
intermissions 0/10/250/1000), integration time = 1.3 ms). Only every
tenth spectrum is shown. (b) Time resolved spectra of the reaction of
[Cu(tmpa)(CH3CN)]ClO4 with dioxygen in propionitrile at ϩ5.1 ЊC
([complex] = 2.0 × 10Ϫ4 M, [O2] = 4.4 × 10Ϫ3 M, total time = 177.66 s,
one time bases (number of measurements 256), integration time = 2.0
ms). Only every seventh spectrum is shown.
to date but the structure can be inferred from [Cu(Me2-uns-
penp)Cl]ClO4 (described above) as has been shown for copper
tmpa complexes.4,21–23 This structural difference is import-
ant because a peroxo ligand in the axial position of a square
pyramidal copper complex would be very labile towards
substitution. Most likely for this reason we could not detect
oxygenation intermediates spectroscopically during the oxid-
ation of [Cu(trien)]ϩ, where the peroxo ligand would be axially
coordinated in a square pyramidal product complex.6
Further support for this explanation comes from the fact that
oxygenation of copper() complexes with bispidine-type ligands
leads to unusually stable end-on µ-peroxo dicopper() com-
pounds. Bispidine-type ligands are very rigid and enforce a
square-pyramidal structure with the additional ligand being
coordinated in an equatorial position.39
During the global analysis the non-linear reaction parameters
(rates) at all measured wavelengths are simultaneously opti-
mised. Herein individual rates contribute to each kinetic record
to a different degree over the measured wavelength range and
the global analysis of all records reduces the correlation of
these parameters (see below). This allows one to fully analyse
all data according to a proposed reaction scheme and a reliable
identification of minor intermediates. In addition spectra of
transient intermediates can be calculated.
Application of global analysis for data fitting
The procedure of data fitting generally starts by importing all
measured data into the program (for example data obtained for
the measurements shown in Fig. 4). The import menu provides
access to a number of instrument-specific file import converters,
and in addition, there are different ASCII file converters for
both ASCII XY (single scan) and specialized ASCII XYZ
spreadsheet (multi-scan) file converters to import multi-
wavelength scans from unsupported data sources. Typically
such data sets consist of individual scans from a spectro-
photometer (UV-vis, FTIR etc.) that have been collected
as a function of time. In kinetic studies, SPECFIT provides
specialized procedures to import diode array and rapid-scan
spectra from the most commonly used stopped-flow kinetic
spectrometers.
The investigation of the reactions of these tripodal copper()
complexes with dioxygen not only requires sophisticated
techniques for the actual measurements (as illustrated above)
but also the analysis of the kinetic data is only possible by using
advanced data fitting methods. This is a consequence of the
combined consecutive equilibrium reactions (eqns. (1)–(3)).
Today the majority of researchers still use the method of fitting
data with exponential functions by the application of iterative
fitting procedures.47 For example the first step (eqn. (1)) in the
reaction scheme described above is the formation of a superoxo
complex. Using pseudo-first order conditions ([O2] ӷ [copper()
complex]) leads to the simple rate law:
d[superoxo complex]/dt = kobs × [copper() complex]
690
J. Chem. Soc., Dalton Trans., 2002, 686–694

View Article Online
The SVD procedure can be carried out during the process of
the file import. The user provides additional information (e.g.
concentration of reactants, nature of the solvent etc.) either for
each individual file or for a complete file series. Furthermore,
the program calculates automatically the correct concentrations
of the reactants according to the density corrections at low
temperatures.
Experimental 3D data sets obtained in such a manner
contain measurements at many more wavelengths than the
number of coloured components that are represented by the
colourimetric changes for the equilibrium or kinetic system
under study. The goal of the data reduction procedure is to
determine a solution for Beer’s Law,
A = C × E
where
Fig. 6 Plot of the eigenvectors vs. time.
A = matrix of experimental absorbance readings Nm × Nw)
C = matrix of calculated concentration profiles (Nm × Nc)
E = matrix of molar absorptivity spectra (Nc × Nw)
and
A ϩ B
A ϩ B
A ϩ C
C
(a)
(b)
(c)
(d)
(e)
C
Nc = number of coloured species
Nm = number of measurements
Nw = number of wavelengths in scans
D
D
A ϩ C
This is an example of an over-determined system of linear
equations, which has no unique mathematical solution. There-
fore, it is necessary to employ special methods in order to solve
simultaneously for both unknown spectra and concentration
profiles. For this the singular value decomposition (SVD)
matrix methodology is used to obtain a basis set of eigen-
vectors, whose matrix product YЈ (U × S = concentration
eigenvectors; S = diagonal matrix, whose elements are the
square roots of the Eigenvalues; V = spectroscopic eigenvector):
D
E
Calculations for the present kinetic experiments are based on
this scheme, with one or more simplifications, depending on the
experimental observations. Some of the kinetic parameters can
only be calculated in a restricted temperature range. Outside the
range, either extrapolated values may have to be used or the
mechanistic model scheme may be appropriately simplified as
discussed earlier for the [Cu(tmpa)(CH3CN)]ϩ system.23
Furthermore, information is provided here by the user as to
whether the species is coloured or not. If the spectrum of an
intermediate is known, the fixed spectrum of this compound
will be used in the calculation. The resulting system of differen-
tial rate equations will be solved via numerical integration
methods (rate constants can be fixed or variable). After the fit is
performed according to the model, the results of the calculation
can be compared to the measured data (the correctness of
the model can then be either confirmed or proven incorrect).
For our example of the reaction of [Cu(Me2-uns-penp)(CH3-
CN)]ClO4 with dioxygen at Ϫ90.1 ЊC only eqns. (a)–(d) need to
be used (decomposition of the peroxo complex according
to eqn. (e) is not relevant under these conditions). Measured
data and the fit together with the residuals for the absorbance
vs. time traces at 412 and 528 nm are shown in Figs. 7a and b.
They clearly demonstrate the goodness of the fit according to
the model used.
YЈ = U × S × V
is the least-squares best estimator of the original 3D data set
(Y). The SVD method produces a linearly independent set of
eigenvectors with all of the colourimetric information for the
experiment, plus some additional noise eigenvectors, that can
be excluded from further consideration. The matrix product
(U × S) contains the information of the coloured components
and they are often called the concentration eigenvectors which
are used as the input data for the global fitting procedure. If the
number of significant eigenvectors is known (which represent
the number of coloured species), it is much easier to enter
a certain kinetic model.
The prediction of relevant eigenvectors should only be used
as a guide. To ensure that the number of calculated eigenvectors
is correct, the plots of the eigenvectors are normally more
definitive. A relevant eigenvector should show a structured
behaviour, while
a
non-relevant eigenvector only shows
Rate constants for the fit according to the model described
statistical noise. For the example of the reaction of [Cu(Me2-
uns-penp)(CH3CN)]ClO4 with dioxygen we observed two
relevant eigenvectors at Ϫ90.1 ЊC (shown in Fig. 6). All the
other eigenvectors are non-relevant and only show statistical
noise (only one of these is shown in Fig. 6). Two relevant
eigenvectors in our example mean that the decay of the
species [Cu(Me2-uns-penp)(CH3CN)]ClO4 has been completed
within the mixing time and only the superoxo- and peroxo-
complex are involved in the observed reaction under these
conditions.
above were calculated to be: k1 = 1.5 × 103 MϪ1 sϪ1, kϪ1 = 1.2 sϪ1
,
k2 = 2.6 × 103 MϪ1 sϪ1 and kϪ2 = 1.0 × 10Ϫ4 sϪ1 (see eqns. (1) and
(2)).46 These values compare well with the results for the
copper() tmpa system (see below) and the preliminary rate con-
stants obtained for the reaction of [Cu(Me6tren)(CH3CN)]ϩ
with dioxygen (k1 = 9.4 × 104 MϪ1 sϪ1, kϪ1 = 0.07 sϪ1, k2 = 1.5 ×
104 MϪ1 sϪ1 and kϪ2 = 5.4 × 10Ϫ5 Ϫ1).25 However, it should be
s
pointed out that Zuberbühler and coworkers strongly discour-
age comparisons of kinetic or equilibrium constants at a given
temperature because these can sometimes be very mislead-
ing.23,53 Therefore, discussion of activation or thermodynamic
parameters should be preferred wherever possible.23,25,46,53
For a complete analysis of such a kinetic system more than a
hundred time resolved spectra at different temperatures and
concentrations need to be recorded and analysed.23 This allows
the calculation of the rate constants and the activation param-
eters for all individual steps. Furthermore, the equilibrium
The Kinetic Model Editor menu provides a group of controls
that allow the user to define complex kinetic models in the form
of elementary reaction steps. The reactions are entered in a
manner that is familiar to chemists (e.g., A ϩ B
C) and
are automatically translated to differential equations as each
equation is completed. For our example the complete reaction
scheme is given by:
J. Chem. Soc., Dalton Trans., 2002, 686–694
691

View Article Online
Table 3 Kinetic and thermodynamic parameters for the reaction of O2
with [(tmpa)Cu(CH3CN)]ϩ 23,53
Parameter
183 K
(1.8 0.1) × 104
298 K
k1/MϪ1 sϪ1
8 × 107
∆H#/kJ molϪ1
∆S#/J KϪ1 molϪ1
32
14 18
4
k
Ϫ1/sϪ1
8
1
2 × 108
∆H#/kJ molϪ1
∆S#/J KϪ1 molϪ1
k2/MϪ1 sϪ1
66
137 18
(3.2 0.2) × 104
4
(1.8 0.1) × 106
(1.2 0.3) × 103
(0.34 0.08)
∆H#/kJ molϪ1
∆S#/J KϪ1 molϪ1
14
1
Ϫ78
2
k
Ϫ2/sϪ1
(1.5 0.8) × 10Ϫ4
∆H#/kJ molϪ1
∆S#/J KϪ1 molϪ1
K1/MϪ1
14
1
Ϫ78
2
(1.9 0.1) × 103
∆H0/kJ molϪ1
∆S0/J KϪ1 molϪ1
K2/MϪ1
Ϫ34
1
Ϫ123
(2.2 0.7) × 108
Ϫ47
4
(1.5 0.4) × 103
∆H0/kJ molϪ1
∆S0/J KϪ1 molϪ1
3
Ϫ97 10
characterization of [Cu(tmpa)(CH3CN)]ClO4 is described
1
elsewhere.22 The H-NMR measurements were recorded on a
Bruker DXP 300 AVANCE spectrometer. Elemental analyses
were performed on a Carlo-Erba Element Analyzer (type 1106)
at the University of Erlangen-Nürnberg. The amine uns-penp
was synthesized according to the literature.54
CAUTION: Perchlorate salts of metal complexes are
potentially explosive and should be handled with care.
Fig. 7 (a) Measured data (ϩ), fit (solid line) and residuals for the
absorbance vs. time trace at 412 nm. (b) Measured data (ϩ), fit (solid
line) and residuals for the absorbance vs. time trace at 528 nm.
Syntheses
constants can be derived for the reactions shown in eqns. (1)
and (2). So far, there are only a few systems analysed in such a
detailed way, e.g. the reaction of [(tmpa)Cu(CH3CN)]ϩ with
dioxygen.23,53 The kinetic and thermodynamic results for this
reaction (obtained by a low temperature stopped-flow study
using global analysis fitting methods) are presented in Table 3.
A good overview on the kinetics and thermodynamics of
copper()–dioxygen interaction has been reported earlier by
Karlin, Kaderli and Zuberbühler.53
Me2-uns-penp. The ligand was obtained by a reductive meth-
ylation of the amine uns-penp. To uns-penp (4 g, 16.5 mmol) in
a flask a 98% solution of formic acid (25 ml, 0.66 mol) and a
38% solution of formaldehyde (15 ml, 0.54 mol) were added.
The solution was stirred at 80 ЊC for 5 hours and the stirring
was continued without heating overnight. HCl was added until
pH 2 was reached and the volume of the solution was reduced
to dryness. The residue was dissolved in a small amount of
water and NaOH (25%) was added until a pH of 11 was
reached. The solution was then extracted with CH2Cl2, and
after drying over Na2SO4 the solvent was removed on a rotary
evaporator. The crude product was obtained as a dark brown
viscous liquid. The amine was further purified by column
chromatography (alumina) with ethyl acetate–methanol (9 : 1)
yielding a brown oil in 55% yield. 1H-NMR (CDCl3): δ
8.54–7.12 (m, 8 H, aromatic), 3.83 (s, 4 H, methylene), 2.7 (t,
2 H, –CH2–CH2–NMe2), 2.5 (t, 2 H, –CH2–CH2–NMe2), 2.18
(s, 6 H, –CH3)
Summary
The advances in low temperature stopped-flow techniques
(together with diode array instrumentation) as well as in com-
puter programs for data fitting (using numerical integration
methods, and factor analysis) have enabled deduction of multi-
step complex kinetic mechanisms of reactions that are too fast
at ambient temperatures. This includes determination of kinetic
and thermodynamic parameters and elucidation of complete
UV-vis spectra of reactive intermediates that can only be
observed at low temperatures.
To illustrate this method we have described herein the
investigation of the reaction of [Cu(Me2-uns-penp)(CH3CN)]-
ClO4 with dioxygen. Clear differences between this reaction and
the analogous oxidation of [Cu(Me6tren)]ClO4 and [Cu(tmpa)-
(CH3CN)]ClO4 were observed. Interestingly enough, it was
found that [(Me2-uns-penp)CuO2Cu(Me2-uns-penp)]2ϩ is more
persistent at higher temperatures than the peroxo complexes
obtained from the other two copper() complexes. This might be
a consequence of their different geometries as discussed above.
[Cu(Me2-uns-penp)(CH3CN)]ClO4. [Cu(CH3CN)4]ClO4 (0.18
g, 0.55 mmol) was added with stirring to a solution of Me2-uns-
penp (0.15 g, 0.55 mmol) in a small amount of acetonitrile
under argon. A large amount of diethyl ether was added to the
orange solution to precipitate the complex. The solution was
filtered through a medium porosity frit, and the solid was
washed with diethyl ether. The yellow powder was dissolved in a
small amount of acetonitrile. Crystals suitable for structural
characterization were obtained by diffusion of diethyl ether
into this solution (yield: 0.1 g, 40%). Anal. calcd. for C18H25-
ClCuN5O4: C, 45.57; H, 5.31; N, 14.76. Found: C, 45.77; H,
5.50; N, 14.65%.
Experimental
[Cu(Me2-uns-penp)Cl]ClO4. To a solution of Me2-uns-penp
(0.45 g, 1.66 mmol) in 15 ml of methanol was added a solution
of Cu(ClO4)2ؒ6 H2O (0.31 g, 0.83 mmol) and CuCl2ؒ2 H2O
(0.14 g, 0.83 mmol) in 10 ml of water. The solution was stirred
for 10 min and filtered. After the recrystallization of the crude
product from methanol blue crystals formed which were suit-
General
Preparation and handling of air-sensitive compounds was
carried out in a glove box filled with argon (Braun, Garching,
Germany; water and dioxygen less than 1 ppm). Synthesis and
692
J. Chem. Soc., Dalton Trans., 2002, 686–694

View Article Online
7 S. Schindler, Eur. J. Inorg. Chem., 2000, 2311.
able for structural characterization (yield: 0.54 g, 70%). Anal.
calcd. for C16H22Cl2CuN4O4: C, 40.99; H, 4.73; N, 11.95.
Found: C, 41.07; H, 4.91; N, 11.89%.
8 M. Schatz, M. Becker, F. Thaler, F. Hampel, S. Schindler, R. R.
Jacobsen, Z. Tyeklár, N. N. Murthy, P. Ghosh, Q. Chen, J. Zubieta
and K. D. Karlin, Inorg. Chem., 2001, 40, 2312.
9 M. Schatz, M. Becker, O. Walter, G. Liehr and S. Schindler, Inorg.
Chim. Acta, 2001, 324, 173.
Kinetic measurements
10 K. D. Karlin and A. D. Zuberbühler, in Bioinorganic Catalysis, 2nd
edn., ed. J. Reedijk and E. Bouwman, Marcel Dekker, New York,
1999.
11 W. Kaim and J. Rall, Angew. Chem., 1996, 108, 47.
12 N. Kitajima, Adv. Inorg. Chem., 1992, 39, 1.
13 N. Kitajima and Y. Moro-oka, Chem. Rev., 1994, 94, 737.
14 Bioinorganic Chemistry of Copper, ed. K. D. Karlin and Z. Tyeklár,
Chapman and Hall, New York, 1993.
Propionitrile used for the kinetic measurements was purified
according to published procedures. Preparation and handling
of air-sensitive compounds was carried out in a glove box.
Dioxygen saturated solutions for the kinetic measurements
were obtained by bubbling dioxygen (Linde, Germany) through
the solvent for 20 minutes as described earlier (solubility of
dioxygen in propionitrile is (8.8 1.0) × 10Ϫ3 M atmϪ1).23 Time
resolved spectra of the reactions of dioxygen with copper()
complexes were recorded on a modified Hi-Tech SF-3 L low
temperature stopped-flow unit (Hi-Tech, Salisbury, UK)
equipped with a J&M TIDAS 16–500 diode array spectro-
photometer (J&M, Aalen, Germany) as described above. Data
fitting was performed using the integrated J&M software Kin-
spec, Origin (OriginLab Corporation, Northampton, MA,
USA) or Igor (WaveMetrics, Inc., Lake Oswego, OR, USA) for
simple exponential functions or the program Specfit (Spectrum
Software Associates, Marlborough, MA 01752, USA) as dis-
cussed above.
15 E. Spodine and J. Manzur, Coord. Chem. Rev., 1992, 119, 171.
16 H.-C. Liang, M. Dahan and K. D. Karlin, Curr. Opin. Chem. Biol.,
1999, 3, 168.
17 K. D. Karlin, Science, 1993, 261, 701.
18 S. Fox and K. D. Karlin, in Active Oxygen in Biochemistry, ed. J. S.
Valentine, C. S. Foote, A. Greenberg and J. F. Liebman, Blackie
Academic and Professional, Chapman & Hall, Glasgow, 1995.
19 A. G. Blackman and W. B. Tolman, in Structure and Bonding,
ed. B. Meunier, Springer, Berlin, 2000.
20 Bioinorganic Catalysis, 2nd edn., ed. J. Reedijk and E. Bouwman,
Marcel Dekker, New York, 1999.
21 R. R. Jacobson, Z. Tyeklár, A. Farooq, K. D. Karlin, S. Liu and
J. Zubieta, J. Am. Chem. Soc., 1988, 110, 3690.
22 Z. Tyeklár, R. Jacobson, R., N. Wei, N. Murthy, J. Zubieta and
K. D. Karlin, J. Am. Chem. Soc., 1993, 115, 2677.
23 K. D. Karlin, N. Wei, B. Jung, S. Kaderli, P. Niklaus and
A. D. Zuberbühler, J. Am. Chem. Soc., 1993, 115, 9506.
24 K. D. Karlin, S. Kaderli and A. D. Zuberbühler, Acc. Chem. Res.,
1997, 30, 139.
25 E. Hörmann, M. Weitzer, B. Jung, S. Kaderli, A. D. Zuberbühler
and S. Schindler, manuscript in preparation.
26 C. Geyer and S. Schindler, Organometallics, 1998, 17, 4400.
27 A. L. Feig, M. Becker, S. Schindler, R. van Eldik and S. J. Lippard,
Inorg. Chem., 1996, 35, 2590.
28 R. W. Cruse, S. Kaderli, K. D. Karlin and A. D. Zuberbühler, J. Am.
Chem. Soc., 1988, 110, 6882.
29 R. W. Cruse, S. Kaderli, C. J. Meyer, A. D. Zuberbühler and
K. D. Karlin, J. Am. Chem. Soc., 1988, 110, 5020.
30 J. A. Halfen, S. Mahapatra, E. C. Wilkinson, S. Kaderli,
V. G. Young, L. Que, A. D. Zuberbühler and W. B. Tolman, Science,
1996, 271, 1397.
31 D.-H. Lee, N. Wei, N. N. Murthy, Z. Tyeklar, K. D. Karlin,
S. Kaderli, B. Jung and A. D. Zuberbühler, J. Am. Chem. Soc., 1995,
117, 12498.
32 H.-C. Liang, K. D. Karlin, R. Dyson, S. Kaderli, B. Jung and
A. D. Zuberbühler, Inorg. Chem., 2000, 39, 5884.
33 S. Mahapatra, V. G. Young, S. Kaderli, A. D. Zuberbühler and
W. B. Tolman, Angew. Chem., Int. Ed. Engl., 1997, 36, 130.
34 E. V. Rybak-Akimova, W. Otto, P. Deardorf, R. Roesner and
D. H. Busch, Inorg. Chem., 1997, 36, 2746.
Crystal structure analyses
Crystal data and experimental conditions for the two complexes
are listed in Table 1. The molecular structures are illustrated in
Figs. 1 and 2. Selected bond lengths and angles with standard
deviations in parentheses are presented in Table 2. Three-
dimensional X-ray data were collected either on a Nonius-
KappaCCD (1) or on a Siemens P4 diffractometer (2) using
graphite monochromated Mo-Kα radiation (λ = 0.71069 Å).
Single crystals were coated with polyfluoroether oil and
mounted on a glass fiber. Lorentz, polarization, and empirical
absorption corrections were applied. Space groups were deter-
mined from systematic absences and subsequent least-squares
refinement. The structures were solved by direct methods.
The parameters were refined using SHELXTL 5.03.56 Non-
hydrogen atoms were refined with anisotropic thermal param-
eters. The hydrogen atoms were localized and isotropically
refined for 1 and fixed in idealized positions using a riding
model for 2.
CCDC reference numbers 173001 and 173002.
See http://www.rsc.org/suppdata/dt/b1/b107927c/ for crystal-
lographic data in CIF or other electronic format.
35 S. V. Kryatov, E. V. Rybak-Akimova, V. L. MacMurdo and L. Que,
Jr., Inorg. Chem., 2001, 40, 2200.
36 B. S. Lim and R. H. Holm, Inorg. Chem., 1998, 37, 4898.
37 A. B. P. Lever, Inorganic Electronic Spectroscopy, 2nd edn., Elsevier,
Amsterdam, 1984.
38 F. Thaler, C. D. Hubbard, F. W. Heinemann, R. van Eldik,
S. Schindler, I. Fabian, A. M. Dittler-Klingemann, F. E. Hahn and
C. Orvig, Inorg. Chem., 1998, 37, 4022.
39 H. Börzel, P. Comba, K. S. Hagen, H. Pritzkow, M. Schatz,
S. Schindler and O. Walter, Inorg. Chem., submitted for publication.
40 M. Kodera, K. Katayama, Y. Tachi, K. Kano, S. Hirota, S. Fujinami
and M. Suzuki, J. Am. Chem. Soc., 1999, 121, 11006.
41 Z. Hu, R. D. Williams, D. Tran, T. G. Spiro and S. M. Gorun, J. Am.
Chem. Soc., 2000, 122, 3556.
42 K. D. Karlin, D.-H. Lee, S. Kaderli and A. D. Zuberbühler, Chem.
Commun., 1997, 475.
43 E. F. Caldin, J. E. Crooks and A. Queen, J. Phys. E: Sci. Instrum.,
1973, 6, 930.
44 K. D. Karlin, M. S. Nasir, B. I. Cohen, R. W. Cruse, S. Kaderli and
A. D. Zuberbühler, J. Am. Chem. Soc., 1994, 116, 1324.
45 A commercial low temperature high pressure stopped-flow unit
based on the design of Merbach and coworkers is commercially
available from Hi-Tech Scientific.
Acknowledgements
The authors gratefully acknowledge financial support from
the DFG. Furthermore, they thank Prof. Rudi van Eldik
(University of Erlangen-Nürnberg) for his support of this
work. Markus Weitzer acknowledges the support from
the DAAD (scholarship) and Prof. Andreas Zuberbühler
(University of Basel, Switzerland) for allowing him to spend six
very fruitful months in the research group of Prof. Zuberbühler.
References
1 M. Becker, S. Schindler and R. van Eldik, Inorg. Chem., 1994, 33,
5370.
2 S. Ryan, H. Adams, D. Fenton, M. Becker and S. Schindler, Inorg.
Chem., 1998, 37, 2134.
3 S. Schindler, C. D. Hubbard and R. van Eldik, Chem. Soc. Rev.,
1998, 27, 387.
4 M. Becker, F. Heinemann and S. Schindler, Chem. Eur. J., 1999, 5,
3124.
5 M. Becker, S. Schindler, K. D. Karlin, T. K. Kaden, S. Kaderli,
T. Palanché and A. Zuberbühler, Inorg. Chem., 1999, 1989.
6 M. Becker, F. Heinemann, F. Knoch, W. Donaubauer, G. Liehr,
S. Schindler, G. Golub, H. Cohen and D. Meyerstein, Eur. J. Inorg.
Chem., 2000, 719.
46 M. Weitzer, M. Schatz and S. Schindler, work in progress.
47 J. H. Espenson, Chemical Kinetics and Reaction Mechanisms,
McGraw-Hill, Inc., New York, 1995.
48 H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbühler, Talanta,
1985, 32, 95.
J. Chem. Soc., Dalton Trans., 2002, 686–694
693

View Article Online
53 K. D. Karlin, S. Kaderli and A. D. Zuberbühler, Acc. Chem. Res.,
49 H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbühler, Talanta,
1985, 32, 257.
1997, 30, 139.
50 H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbühler, Talanta,
54 J. Mandel, C. Maricondi and B. Douglas, Inorg.Chem., 1988, 27,
2990.
1985, 32, 1133.
51 H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbühler, Talanta,
55 M. N. Burnett and C. K. Johnson, ORTEP3, Report ORNL-6895,
Oak Ridge National Laboratory, Oak Ridge, TN, 1996.
56 SHELXTL, version 5.03, Siemens Analytical X-Ray Instruments,
Madison, WI, 1997.
1986, 33, 943.
52 M. Maeder and A. D. Zuberbühler, Anal. Chem., 1990, 62,
2220.
694
J. Chem. Soc., Dalton Trans., 2002, 686–694