New insights on the mechanism of oxidation of d-galacturonic acid by hypervalent chromium

Article abstract of DOI:10.1039/c1dt10214a

The pollutant Cr^VI is known to be very carcinogenic. In conditions of excess of Cr^VI, oxidation of d-galacturonic acid (Galur), the major metabolite of pectin, yields d-galactaric acid (Galar) and Cr^III. The redox reaction takes place through a multistep mechanism involving formation of intermediate Cr^II/IV and Cr^V species. The mechanism combines one- and two-electron pathways for the reduction of Cr^IV by the organic substrate: Cr^VI → Cr ^IV → Cr^II and Cr^VI → Cr^IV → Cr^III. This is supported by the observation of the optical absorption spectra of Cr^VI esters, free radicals, CrO ₂²⁺ (superoxoCr^III ion) and oxo-Cr^V complexes. Cr^IV cannot be directly detected; however, formation of CrO₂²⁺ provides indirect evidence for the intermediacy of Cr^II/IV. Cr^IV reacts with Galur much faster than Cr ^V and Cr^VI do. The analysis of the reaction kinetics via optical absorption spectroscopy shows that the Cr^IV-Galur reaction rate inversely depends on [H⁺]. Nevertheless, high [H⁺] still does not facilitate accumulation of Cr^IV in the Cr ^VI-Galur mixture. Cr^VI and the intermediate Cr^V react with Galur at comparable rates; therefore the build-up and decay of Cr^V accompany the decay of Cr^VI. The complete rate laws for the Cr^VI, Cr^V and Cr^IV-Galur redox reaction are here derived in detail. Furthermore, the nature of the five-co-ordinated oxo-Cr^V bischelate complexes formed in Cr^VI-Galur mixtures at pH 1-5 is investigated using continuous-wave and pulsed electron paramagnetic resonance (EPR) and density functional theory (DFT).

Full text of DOI:10.1039/c1dt10214a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2011, 40, 7033
www.rsc.org/dalton
PAPER
New insights on the mechanism of oxidation of D-galacturonic acid by
hypervalent chromium†
a
a
a
a
b
c
M. F. Mangiameli, J. C. Gonz a´ lez,* S. I. Garc ´ı a, M. I. Frascaroli, S. Van Doorslaer, J. M. Salas Peregrin
and L. F. Sala*a
Received 8th February 2011, Accepted 14th April 2011
DOI: 10.1039/c1dt10214a
VI
VI
The pollutant Cr is known to be very carcinogenic. In conditions of excess of Cr , oxidation of
III
D-galacturonic acid (Galur), the major metabolite of pectin, yields D-galactaric acid (Galar) and Cr .
The redox reaction takes place through a multistep mechanism involving formation of intermediate
II/IV
V
Cr
and Cr species. The mechanism combines one- and two-electron pathways for the reduction of
IV
VI
IV
II
VI
IV
III
Cr by the organic substrate: Cr → Cr → Cr and Cr → Cr → Cr . This is supported by the
VI
2+
III
observation of the optical absorption spectra of Cr esters, free radicals, CrO
2
(superoxoCr ion) and
V
IV
2+
oxo-Cr complexes. Cr cannot be directly detected; however, formation of CrO
evidence for the intermediacy of Cr . Cr reacts with Galur much faster than Cr and Cr do. The
analysis of the reaction kinetics via optical absorption spectroscopy shows that the Cr –Galur reaction
rate inversely depends on [H ]. Nevertheless, high [H ] still does not facilitate accumulation of Cr in
the Cr –Galur mixture. Cr and the intermediate Cr react with Galur at comparable rates; therefore
the build-up and decay of Cr accompany the decay of Cr . The complete rate laws for the Cr , Cr
and Cr –Galur redox reaction are here derived in detail. Furthermore, the nature of the
five-co-ordinated oxo-Cr bischelate complexes formed in Cr –Galur mixtures at pH 1–5 is
investigated using continuous-wave and pulsed electron paramagnetic resonance (EPR) and density
functional theory (DFT).
2
provides indirect
II/IV
IV
V
VI
IV
+
+
IV
VI
VI
V
V
VI
VI
V
IV
V
VI
Introduction
galacturonic acid to evaluate their capacity to interact with
22,26
hypervalent chromium.
Of particular interest is D-galacturonic
VI
Cr is a very important pollutant, and its derivative compounds
acid (Galur), the major low molecular-weight metabolite of pectin
substances. The determination of the ability of Galur to reduce or
stabilize high oxidation states of chromium will certainly help to
unravel its potential role in chromium (bio)chemistry. In an earlier
represent a potential environmental hazard because of their
mammalian toxicity and carcinogenicity. The observation of the
formation of Cr and Cr
dation of various organic substrates by Cr and their involvement
in Cr-induced cancers
of interest in chromium chemistry and biochemistry. Through
1
–4
V
II/IV
intermediates during the selective oxi-
VI
27
IV
work by some of us, a slow Cr oxidation step was considered to
describe the mechanism of the hypervalent chromium oxidation
of D-galacturonic acid in acidic media.
1
,5–7
has generated a considerable amount
8–13
VI
the examination of reactions of Cr with low-molecular-weight
In the current work, we present new kinetics and spectroscopic
data that shed a new light on the oxidation process, Scheme 1.
The general redox reaction mechanism proposed in this work is
entirely based on experimental spectroscopic evidence, avoiding
the use of complicated mathematical models that try to involve
undetected hypervalent chromium intermediates, which may lead
14–21
22–24
neutral
and acidic
saccharides, we were able to probe the
VI
V
IV
interaction of Cr , Cr and Cr with different sugars present
in biological systems. Since sugars possess 2-hydroxycarboxylato
and vic-diolato functional sites that can potentially chelate Cr
naturally occurring acidic saccharides offer suitable ligands for
the stabilization of Cr
an abundant component of the primary cell walls of plants,
has also been studied along with D-glucuronic, di- and tri-
V/IV
,
V/IV
oxidation states. Poly-galacturonic acid,
to unwieldy results. Furthermore, we explored the ability of Galur
25
V/IV
to coordinate Cr
generated in the redox reaction. In particular,
continuous-wave (CW) and pulsed EPR (electron paramagnetic
resonance) and ENDOR (electron nuclear double resonance)
a
28
29
Departamento de Qu ´ı mica, Facultad de Ciencias Bioqu ´ı micas y Far-
techniques and density functional theory (DFT) computations
V
mac e´ uticas, UNR, Suipacha 531, 2000, Rosario, Argentina. E-mail: sala@
iquir-conicet.gov.ar, gonzalez@iquir-conicet.gov.ar
were used to study the oxo-Cr complexes. Although the unique
potential of pulsed EPR/ENDOR techniques in the study of para-
magnetic molecules has long been recognized,
have, to our knowledge, not been exploited in the analysis of Cr
biochemistry.
b
Department of Physics, SIBAC Laboratory, University of Antwerp, Belgium
28,30
these methods
c
Departamento de Qu ´ı mica, Universidad de Granada, Granada, Spain
V
†
1
Electronic supplementary information (ESI) available. See DOI:
0.1039/c1dt10214a
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7033

View Article Online
VI
Scheme 1 Redox reaction of Cr with excess Galur.
-
1
-1
Results
(e = 29 M cm ). The time evolution of the spectra was followed
for 24 h and resulted in the observation of the typical bands of
VI
Time evolution of the reaction of Galur–Cr
3+
+
-1
-1
[
Cr(OH ) ] ion in the same [H ] range (410 nm, e = 18.2 M cm
1 III
2
6
-
-1
and 574 nm, e = 15.2 M cm ). These results suggest that a Cr –
The UV/Vis absorption spectrum taken directly after mixing
Galur with Cr in a HClO
Galur complex is formed as a final redox product, which then
hydrolyses to Cr (aq)
VI
4
/H O medium shows an absorbance
2
III
.
band at 350 nm and a shoulder at 420–500 nm, characteristic of
VI
Cr in acidic medium, Fig. 1. The absorbances at 350 nm and 420–
II
4
70 nm decay with time, while the absorbance at 570 nm increases,
Intermediacy of Cr
Fig. 1.
IV
It is well known that Cr oxidizes alcohols, acting as a two-
II
electron oxidant to produce Cr and the corresponding oxidized
organic product. As shown earlier, the involvement of Cr in
the oxidation mechanism of several alcohols by Cr and Cr in
HClO can be demonstrated by its conversion into a superoxo-
, upon reaction with dioxygen.
appropriate experimental conditions, such as [H ] (0.1–0.3 M),
33
II
IV
VI
4
III
2+
33–35
Cr ion, CrO
2
Indeed, under
+
VI
-5
high [O ] (1.26 mM) and low [Cr ] (6 ¥ 10 M), the reaction of
2+ 33–35 II
2
II
Cr (if any) with O
2
gives CrO
2
.
If Cr is an intermediate
2
+
35
species in the current redox reaction, CrO
Therefore, the presence of intermediate Cr in the reaction of
Galur with Cr , was examined by monitoring the formation of
2
should be detected.
II
VI
2
+
CrO
2
. A time-dependent optical-absorption analysis of the O -
2
VI
saturated solution of a Cr –Galur reaction mixture showed the
growth and decay of two absorption bands at 290 and 246 nm.
2
+
These two bands, characteristic of CrO
after all the Cr had been consumed, Fig. 2. These spectroscopic
results reveal that Cr is formed in the redox reaction and can
also be taken as indirect evidence that Cr is implied in the redox
2
, persisted in solution
VI
II
IV
VI
VI
Fig. 1 Time evolution of the Galur–Cr reaction as followed by UV/Vis
absorption spectroscopy. [H ] = 0.1 M; I = 1.0 M; [Cr ]
mechanism of the reaction between Cr and Galur, as has been
+
VI
0
= 0.006 M;
◦
[
Galur] = 0.6 M; T = 33 C. First trace was taken at t = 0 min, last trace at
t = 145 min, the time between each trace is 5.0 min.
The evident lack of an isosbestic point (see area marked in
Fig. 1) indicates that there are two or more competing reactions
at any time and that intermediate chromium species are present
VI
III
in appreciable concentrations in the reduction of Cr to Cr .
The rate constants were determined from the absorbance data
obtained at 350 nm. As expected, the time-dependent decrease in
V
IV
absorbance was not monophasic. Since the Cr and Cr species
both absorb at 350 nm, and since their absorbance therefore may
VI
V/IV
superimpose on the Cr one, Cr
contributions have to be taken
into account when interpreting the absorbance decay, especially
VI
V/IV
31
when the Cr and Cr
decay rates are similar. At the end of the
-
1
-1
redox reaction, two d–d bands at lmax = 426 nm (e = 29 M cm )
and 562 nm (e = 33 M cm ) are observed in the electronic
absorption spectrum, attributable to the octahedral 4A → 4T
and 4A → 4T transitions of Cr in O symmetry, Fig. 1.
After 150 min of the reaction, these bands show a slow decrease in
-
1
-1
2
g
1g
2
g
2g
III
32
VI
h
Fig. 2 Time evolution of the Galur–Cr reaction mixture to determine
IV
the presence of Cr by UV/Vis absorption spectroscopy. [Galur] = 0.02 M,
-
1
-1
+
◦
VI
-5
intensity and shift to lmax = 420 nm (e = 23 M cm ) and 566 nm
[H ] = 1.0 M, [O
2
] = 1.26 mM, T = 25 C, [Cr ] = 6.0 10 M.
7
034 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
VI
observed in the reduction of Cr by D-glucuronic acid and other
saccharides.
Abs
•
and Abs
0
are the absorbance at infinite time and the
26
initial absorbance, respectively. Table S1 (see ESI†) summarizes
the values of k for various concentrations of Galur in HClO . It is
known that Cr can disproportionate into Cr and HCrO , with
4
IV
4
III
-
Detailed reaction rate studies
4
VI
a second-order kinetic on [Cr ] and an inverse dependence on
IV
Reaction of Galur–Cr . As is described in the experimental
section, Cr was generated in situ by reaction of Cr with O
acidic media. The reaction of Galur with Cr in acidic media and
-saturated solutions leads to the formation of Cr . This reaction
+
36
VI
-
[
2
H ]. This should be avoided, because Cr (HCrO
4
) absorbs at
IV
II
2
in
2
+
90 nm, interfering with the measurement of CrO
in a very low concentration of Galur, we were able to observe the
appearance of a typical Cr spectrum, with the characteristic band
at 350 nm. Due to that, the experimental conditions were chosen
such that the Galur–Cr reaction can compete successfully with
the Cr disproportionation reaction. Fig. 3A shows the optical
absorption spectra of the Galur–Cr mixture (inset) and the fitting
of the time-dependent absorbance data at 290 nm using eqn (1).
2
. In absence or
IV
II
O
2
VI
can be indirectly monitored by measuring the absorbance increase
2
+
at 290 nm (formation of CrO
2
), since neither the substrate,
IV
nor the oxidized products absorb at this wavelength. Note that
the intensity ratio between the peaks at 246 and 290 nm is
.2, which confirms the presence of CrO
inset). The monotonic increase in the absorbance at 290 nm
was found to follow first-order kinetics, Fig. 3A. The first-order
dependence of the rate upon [Cr ] was verified in a set of
experiments where the [Cr ] was varied between (3.0–6.0) ¥
IV
IV
2
+
33–35
2
2
species,
Fig. 3A
(
As represented in Fig. 3B, the k values depend linearly on the
concentration of Galur for different [H ]. From this, the k4H values
were determined using:
4
+
IV
IV
0
-
5
10
M, while temperature, [Galur], and ionic strength (I) were
k
4
= k4H [Galur]
(2)
kept constant. The corresponding experimental rate constants, k
4
,
+
-1
The bimolecular rate constant, k4H, varies linearly with [H ]
were determined by a non-linear least-square fit of the absorbance-
time data using 80% of the exponential growth of the experimental
values:
I
-1
-1
II
with a positive intercept, k IV (10.30 ± 0.50 s M ) and slope k IV
-
1
(
2.90 ± 0.14 s ) (inset Fig. 3B):
I
II
+ -1
k
4H = k IV+ k IV [H ]
(3)
k
t
4
Abs
t
= Abs
•
+ (Abs
0
- Abs
•
)e-
(1)
IV
Combining eqn (2) and (3), the rate constant for Cr disappear-
ance, k , is given by
4
I
II
+ -1
k
4
= (k IV+ k IV [H ] ) [Galur]
(4)
The experimental values of rate constants obtained here are
more accurate than those obtained employing complicated rate
law expressions, which might give unwieldy results, when the
II
IV
intermediacy of Cr and Cr are assumed without experimental
37
verification.
VI
Reaction of Galur–Cr . The time evolution of the absorbance
VI
at 350 nm of the Galur–Cr mixtures exhibited a monotonic
decrease, which cannot be described by a single exponential decay.
These kinetics profiles were appropriately described by the set of
consecutive first-order reactions given in Scheme 2.
VI
III
Scheme 2 Cr → Cr reduction pathway used to fit the experimental
data.
V
Taking into account the Cr absorption superimposition, the
absorbance at 350 nm is, at any time during the redox reaction,
given by
VI
VI
V
V
Abs350 = e [Cr ] + e [Cr ]
(5)
Combining eqn (5) with the rate expressions derived for the
consecutive first-order reactions in Scheme 2, gives the following
expression:
2
+
IV
Fig. 3 (A) Formation of CrO
2
2
from Cr –Galur reaction under O as
38
monitored by UV/Vis absorption spectroscopy (inset) and the absorbance
at 290 nm vs. time. [H ] = 0.25 M, [O
Cr ] = 0.075 mM. Fitted lines were calculated using eqn (1). (B) Effect
+
◦
2
] = 1.26 mM, I = 1.0 M, T = 15 C,
IV
[
e-2k6t + k
V
VI
(e-k5t - e-2k6t)/(2k
Abs350 = Abs
0
6
e [Cr ]
0
6
- k
5
)
(6)
-1
◦
of [Galur] on k
4
. Inset: linear dependence of k4H on [H+] . T = 15 C, I =
= 0.075 mM and [H ]: (a) 0.10, (b) 0.15, (c) 0.20, (d) 0.25, and
IV
+
VI
1
.0, [Cr ]
0
Where k
6
and k refer to the rate of disappearance of Cr
5
V
V
(
e) 0.30 M.
and Cr , respectively, and e refers to the molar absorptivity
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7035

View Article Online
V
V
3
-1
-1
of oxo-Cr –Galur at 350 nm (e = (2.3 ± 0.23) ¥ 10 M cm ).
plots of k
5
vs. [Galur] show a linear dependence (Fig. 4B), from
This value was determined by UV-vis spectroscopy of a
which the bimolecular rate constant k5H can be derived:
V
-
ligand-exchange product of [Cr (O)(ehba)
2
] (ehba = 2-ethyl-2-
14
k
5
= k5H [HS]
(10)
hydroxybutanoato(-)) with a large excess of Galur, and the
V
formation of a Cr –Galur complex was confirmed by EPR
+
A plot of k5H vs. [H ] reveals a quadratic dependence (inset
spectroscopy (see Experimental Section).
Fig. 4B):
Parameters k
computer fit using eqn (6). Table S2 (see ESI†) lists the values of k
and k for various concentrations of Galur in HClO . In the range
of pH values employed in this work, k linearly depends on [Galur]
Fig. 4A), from which values of k6H can be determined using:
5
and k were evaluated from a non-linear iterative
6
+
2
6
k
5H = k + kS5 [H ]
(11)
5
4
Hence,
6
(
+
2
k = k [HS] + kS5 [H ] [SH]
5
(12)
k
6
= k6H [HS]
(7)
From Fig. 4B, k = (4.4 ± 0.4) ¥ 10 s-1 M-1 and kS5 = (1.65 ± 0.17) ¥
-
3
-
2
-1
-3
with HS being Galur.
10
s
M .
The rate constants k
obtained using continuous-wave (CW)-EPR spectroscopy at room
The bimolecular rate constant, k6H, quadratically depends on
6
and k can also be independently
5
+
the [H ] as shown in the inset of Fig. 4A:
2
7
V
temperature. As reported earlier, oxo-Cr derivatives of Galur
exhibit a clear room-temperature CW-EPR signal that consists
of a strong central signal at giso ª 1.9785 flanked by four weak
signals (see ESI†). While the central signal stems from the oxo-
Cr complexes with Cr isotopes with zero spin (predominantly
Cr), the quartet originates from the Cr isotope (I = 3/2, natural
abundance 9.5%). The central signal shows substructure due to
resolved proton hyperfine interactions and some of us earlier
interpreted the spectrum in terms of different oxo-Cr complexes.
Before treating the nature of these individual oxo-Cr complexes in
+
2
k6H = kS6 [H ]
(8)
Combination of eqn (7) and (8) allows the determination of k
via
6
V
+
2
k
6
= kS6 [H ] [HS]
(9)
52
53
with kS6 = (3.68 ± 0.40) ¥ 10 s-1 M , determined as the slope of
-
2
-3
+
the linear curve showed in the inset of Fig. 4A. At constant [H ],
V
27
V
V
more detail, the kinetics of oxo-Cr formation from the variation
of the CW-EPR signal with time, were first determined. In a 0.10 M
V
HClO
4
solution, the double integral of the oxo-Cr EPR signal of
VI
◦
a 0.973 M Galur–Cr mixture shows, at 20 C, a decrease after
VI
an initial increase, Fig. 5. Note that higher [Cr ] are needed
0
for EPR than for the UV/Vis absorption experiments. The time
dependence of the EPR double integral, h, which is a measure
V
for the total Cr concentration, was fitted using eqn (13) derived
from the earlier-mentioned consecutive first-order reactions in
Scheme 2:
(e- - e-2k6 t)/(2k
k5 t
- k
)
(13)
h = A k
6
6
5
Fig. 4 (A) Effect of [Galur] on k
6
. Inset: Quadratic dependence of k6H on
+
[H ]. Measurement points are determined from the decay of the absorbance
Fig. 5 Double integral of the EPR signal vs. time for the oxidation
VI
VI
◦
at 350 nm. (B) Effect of [Galur] on k
on [H ]. T = 33 C, I = 1.0 M and [H ]: (a) 0.97; (b) 0.80; (c) 0.63; (d) 0.40
5
. Inset: Quadratic dependence of k5H
of Galur by Cr . [Cr ] = 0.047 M; I = 1.0 M, T = 20 C, [Galur] =
+
◦
+
+
0.973 M, [H ] = 0.10 M. The EPR spectra were recorded with a modulation
and (e) 0.20 M.
amplitude of 0.2 mT.
7
036 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
III
V
III
The parameter A depends on the spectrometer settings such as
gain, power and modulation amplitude.
and e thus extracted are reasonable for Cr and Cr complexes
10,24,26
with saccharides.
(15) were in agreement with those obtained via eqn (9)–(12). The
kinetics profile simulated using k and k showed that the time
of maximum intensity (tmax) at Abs570 is very close to the time
calculated (130 s) for reaching the maximum [Cr ] in the reaction
The values of k
6
and k
5
obtained using eqn
The k
5
and k values obtained by EPR are in good agreement
6
with those obtained from UV/Vis absorption data, taking into
account the experimental error and the difference in temperature
5
6
-
4
-1
-3
-1
V
(
k
6
= 1.8 ¥ 10 s , k
5
= 1.2 ¥ 10 s ) and hence it provides
confidence in the proposed set of consecutive first-order reactions
mixture (up to 34% of the total Cr in the solutions) (Fig. 6).
These results indicate that the intermediate Cr species should
IV
V
shown in Scheme 2. Cr is involved in a fast kinetic step and only
V
VI
Cr accumulates in the Galur–Cr mixtures.
be responsible for the increase and decay of the absorbance at
III
In the kinetics studies based on the time dependence of the
absorbance at 570 nm, we have to consider that besides the
570 nm in the short time range, after which the resulting Cr
III
species decompose slowly to Cr (aq)
.
V
IV
III
III
oxo-Cr –Galur complexes, also Cr (aq), Cr (aq) and Cr -ligand
species absorb at this wavelength. According to the results reported
Generation of oxo-CrV–galacturonic complex
IV
above, Cr does not accumulate, so cannot contribute to the
V
light absorption at 570 nm. In Fig. 6, the time dependence of
the absorbance signal at 570 nm is shown. Initially, the signal
rapidly grows to high values (~within the first 150 s), which could
Oxo-Cr –Galur complex can be generated in different ways: a)
VI
ligand exchange; b) direct reaction Galur–Cr and c) reaction
VI
14,15,22,26,27
Galur–Cr /glutathione (GSH) or cysteine.
In the first
V
III
V
-
be related to underlying contributions of Cr and/or Cr –ligand
complexes. After this, the absorbance decay clearly shows a fast
and slow phase (see inset in Fig. 6). In the first 400 s after reaching
the maximum, the decay is dominated by the fast contribution,
method, an exchange reaction occurs between [Cr O(ehba)
2
] and
V
excess of Galur, in which the first compound acts as Cr donor.
This process is favored by a large excess of Galur in the media,
and it is very important to keep the pH close to 3.0 in which
[Cr O(ehba) ] is more stable. In the second method, oxo-Cr
complexes are generated during the direct reaction Galur–Cr ,
also favored by a large excess of Galur in the media and its rates
depend on the pH and [Galur]. Finally, oxo-Cr –Galur complexes
can be generated using Galur–GSH/Cr mixtures. In a very fast
redox reaction, GSH reduces Cr into Cr and the later can be
successfully trapped by Galur. The main advantage of this last
method is that a fast generation of oxo-Cr –Galur species is seen
V
V
-
10
V
which probably corresponds to the decay of the Cr species. The
2
VI
slow part, dominant at a longer time scale, could correspond
III
to the slow hydrolysis of the Cr –ligand complex. To test this
V
hypothesis, the experimental data were fitted with the following
equation derived from Scheme 2, taking into account that the
VI
III
VI
V
final redox species is a Cr –ligand complex:
V
V
III
III
Abs570 = e [Cr ] + e [Cr –ligand]
(14)
V
VI
with a direct mixture of Galur–Cr .
V
Characterization of the Cr derivatives
27
In our earlier work, the room-temperature CW-EPR spec-
VI
trum of the Cr –Galur mixture at low pH were interpreted
in terms of three contributions, two triplets and one doublet,
V
6
5
which in turn were related to two conformers of [Cr O(O ,O -
-
galactofuranuronate)
2
]
(cis and trans forms, I and II in
V
ring
6
5
6
Scheme 3) and a [Cr O(O ,O -galactopyranuronate)(O ,O -
galactofuranuronate)] complex (III in Scheme 3). Oxo-Cr species
were independent on the [H ] and ligand to metal ratios. In the
presence of molar excess of Galur, at room temperature and pH 1–
, Cr /glutathione mixtures (1 : 1 ratio) produced oxo-Cr EPR
V
+
27
VI
V
5
VI
spectra identical to those obtained by direct reaction of Cr with
Galur. Additionally, EPR spectroscopic parameters from oxo-Cr -
V
Galar species showed that Galar (the oxidation product) was not
involved in the Cr coordination.
Fig. 6 Black solid line: Time dependence of the absorbance at 570 nm for
V
27
VI
◦
the Cr –Galur reaction in 0.50 M HClO
4
, I = 1.0 M, T = 33 C, [Galur] =
This interpretation was based on only limited experimental data,
namely the isotropic proton hyperfine values and the giso values
extracted from a single room-temperature EPR spectrum that were
VI
0
.70 M, [Cr ] = 6.0 mM (Inset: large time range). The deconvoluted kinetic
III V
profiles of Cr and Cr are depicted to compare with the experimental
-3
-1
results. Kinetics profiles were calculated using eqn (15), k = 4.7 ¥ 10 s ,
5
V
10
-3
-3
V
-1
-1
III
-1
-1
compared to EPR data reported for other oxo-Cr complexes. In
order to substantiate the earlier assumptions, further experiments
are now performed. No major spectral changes were observed
k = 7.2 ¥ 10 s , e = 89 M cm and e = 17.5 M cm .
6
In this way, the total absorbance at 570 nm at any time is given
by
VI
when comparing the room-temperature CW-EPR spectra of Cr –
2
Galur mixtures at pH 3.0 in H
2
O and in
2
H O (see ESI†),
VI
III
-2k6t
V
III
(e-2k6t - e-k5t)/
Abs570 = [Cr ]
0
{e (1 - e )+ (e - e )k
6
(15)
indicating that the proton hyperfine splitting resolved in the room-
temperature CW-EPR spectra does not stem from exchangeable
protons. This excludes equatorial water coordination to the
(
k
5
- 2k )}
6
A very good fitting of experimental data was thus obtained with
e = 89 M cm and e = 17.5 M cm (Fig. 6). The values of e
V
-1
-1
III
-1
-1
V
V
oxo-Cr complexes which would lead to considerable hyperfine
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7037

View Article Online
V
Fig. 7 X-band CW-EPR spectrum recorded at 50 K of the oxo-Cr –Galur
VI
complexes present at pH 3.0. [Cr ]
0
= [GSH]
0
= 0.67 mM, [Galur] =
0
0
.33 mM. The modulation amplitude was 0.3 mT. (a) Experimental
spectrum using H
for the ENDOR experiments; (b) Experimental spectrum using H
2
O as solvent. (1–3) indicate the observer position chosen
2
2
O as
solvent; (c) Simulation assuming a single species with the parameters given
in Table 1; (d) Simulation assuming two species in 1 : 1 ratio with the
parameters given in Table 1.
1
Fig. 8b–d show the comparison of the H Davies ENDOR spec-
VI
2
tra of the Cr –Galur mixture at pH 3.0 in H
observer positions (1–3) indicated in Fig. 7a. The spectral width
~4 MHz) was the same for all observer positions and corresponds
2
O taken at the three
V
Scheme 3 Proposed structures of oxo-Cr ligating Galur.
(
to the maximal hyperfine coupling component (in absolute values)
of the non-exchangeable protons. The maximal proton hyperfine
coupling of the exchangeable protons is considerably smaller, as is
V
contributions of exchangeable protons. This means that oxo-Cr –
Galur bis-chelates, rather than mono-chelates are present.
A similar observation was done at 50 K, Fig. 7a,b. Fig. 7c shows
demonstrated by the Davies ENDOR spectrum taken at observer
VI
VI
the simulation of the low-temperature CW-EPR spectra of Cr –
position (1) for the Cr –Galur mixture at pH 3.0 in H O (Fig. 8a).
2
V
Galur mixtures at pH 3.0 (Fig. 7a,b) assuming a single oxo-Cr
species (species 1 in Table 1). The corresponding giso (1.9788) and
isotropic Cr hyperfine ( aiso = 51 MHz) values agree well with
Comparison with the ENDOR spectrum in Fig. 8b shows that the
two spectra only differ by the strong central line with a width of
approximately 1 MHz.
5
3
Cr
the values derived earlier from the room-temperature CW-EPR
In order to substantiate the earlier interpretation of the room-
temperature CW-EPR spectra in terms of three species, DFT
computations of the EPR parameters of the three proposed
structures given in Scheme 3 were performed. The computations
Cr
27
spectra (giso,1 = 1.9786, giso,2 = 1.9784, giso,3 = 1.9785, aiso = 51.5 MHz
27
for the two triplets and doublet, respectively). The observed
chromium hyperfine interaction is typical of a five-coordination
39–41
1
of the Cr center.
Although the simulation in Fig. 7c was
are reported on the complexes having R-chirality on the C
1
quite good, it is not perfect. The simulation can be considerably
improved by assuming at least one additional species with slightly
different EPR parameters (species 2 in Table 1) (Fig. 7d), which
corroborates the earlier derivation that multiple species contribute
center. Similar results are found for the S-C forms. Additionally,
computations were performed on [Cr O(O ,O - galactopyra-
nuronate)(O ,O - galactofuranuronate)] complexes thought to be
very reactive and less stable than the three species mentioned
previously (see ESI†). Although the state-of-the-art DFT methods
V
1
6
6
5
-
27
at room temperature to the EPR spectrum.
5
3
VI
Table 1 Experimental principal g and Cr hyperfine values derived from the low-temperature EPR spectra of the Cr –Galur mixture at pH 3. The data
V
V
6
5
-
2
are compared with the DFT computed values for the three oxo-Cr complexes depicted in Scheme 3. [Cr O(O ,O -galactofuranuronate) ] (structures I
and II) and [Cr O(O ,O -galactopyranuronate)(O ,O -galactofuranuronate)] (structure III). The coordinates of the optimized structures are listed in
V
ring
6
5
6
the ESI†
Cr
Cr
Cr
g
x
g
y
g
z
A
x
/MHz
A
y
/MHz
z
A /MHz
Species 1
1.9853 ± 0.0002
1.9855 ± 0.0002
1.9854
1.9868
1.9865
1.9784 ± 0.0003
1.9791 ± 0.0003
1.9810
1.9828
1.9843
1.9726 ± 0.0003
1.9718 ± 0.0003
1.9784
1.9771
1.9751
22 ± 5
22 ± 5
3.6
1.4
7.4
23 ± 5
23 ± 5
15.6
14.2
21.0
108 ± 2
108 ± 2
89.5
85.6
93.7
EPR
EPR
DFT
DFT
DFT
Species 2
Structure I
Structure II
Structure III
7
038 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
The room-temperature CW-EPR spectrum was found to be best
27
described in terms of two triplets and one doublet signal. The
triplet signals arise from the interaction of the unpaired electron
H
with two equivalent protons with | aiso,1| = 1.8 (±0.1) MHz
H
(
triplet 1) and with | aiso,2| = 2.9 (±0.1) MHz (triplet 2). The
protons were earlier suggested to be the two carbinolic protons
27
in structures I and II. These protons are not exchangeable
2
which matches the results of the H exchange experiment reported
here. The DFT computations confirm that the carbinolic protons
are the only ones in structures I and II that have a significant
isotropic hyperfine coupling (Table 2). All other non-exchangeable
protons have isotropic couplings smaller than 0.15 MHz (in
absolute values). Although the isotropic and the maximum proton
hyperfine couplings (in absolute values) of the carbinolic protons
are somewhat overestimated in the DFT computations, there is a
good agreement between the overall trends predicted by DFT and
observed experimentally.
In order to fit the room-temperature CW-EPR spec-
H
1
V
trum, a doublet signal (| aiso,3| = 2.4 (±0.1) MHz) due
Fig. 8 H Davies ENDOR spectra of the oxo-Cr Galur complexes in
27
2
VI
to a third minority species was also taken into account.
H
2
O (a) and H
2
O (b–d) recorded at 20 K. [Cr ]
0
= [GSH]
0
= 0.67 mM,
V
ring
6
5
6
This was ascribed to [Cr O(O ,O -galactopyranuronate)(O ,O -
galactofuranuronate)] (III), with the doublet stemming from the
carbinolic proton of the galactofuranuronate. The DFT analysis
reveals that besides this proton, the carbinolic proton of the
galactopyranuronate also has a significant isotropic hyperfine
coupling (Table 2), which seems to argue against the presence
of this species in the mixture.
[
Galur] = 0.33 mM. (a,b) Observer position (1), (c) observer position (2)
and (d) observer position (3).
0
for transition-metal complexes are not yet so accurate that a full
quantitative comparison between experiment and computations
is possible, the general trends predicted by DFT are usually
quite good. Table 1 shows the computed principal g and Cr
hyperfine values for the three complexes. The g values agree quite
well with the experimental values and fall within the deviation
between experimental and theoretical g values also found for
other oxo-Cr complexes. The theoretical Cr hyperfine values
are underestimated, but similar underestimations were reported
earlier for other oxo-Cr systems.
29
53
V
1
6
6
5
The
[Cr O(O ,O -galactopyranuronate)(O ,O -galactofura-
nuronate)] complexes are found to be considerably less stable
than the two [Cr O(O ,O -galactofuranuronate)
see ESI†). Moreover, the hyperfine couplings of some of the non-
exchangeable protons were found to be too high in comparison
with the experiment.
-
V
6
5
-
2
]
complexes
(
V
42
53
V
42
The proton hyperfine values are also computed. None of the
exchangeable protons in any of the complexes are found to
have isotropic proton hyperfine values larger than 0.15 MHz (in
absolute value). This agrees with the experimental observation that
deuteration of the solvent does not change the room-temperature
CW-EPR spectrum (see ESI†) and that the resolved isotropic
proton hyperfine splitting (> 1.7 MHz in absolute value) thus
must stem from non-exchangeable protons.
Discussion
VI
Oxidation of Galur by Cr
VI
Oxidation of Galur by Cr is a strongly pH dependent process.
Although at pH > 1 the reaction is slow, Cr is rapidly consumed
when the [H ] > 0.2 M. For this reason, the 0.2–1.0 M [H ]
range was chosen to study the kinetics of this reaction. The time-
VI
+
+
1
In Table 2, the computed H hyperfine data of the non-
exchangeable protons that have a non-negligible isotropic hy-
perfine value are reported for the three computed complexes.
+
dependent UV/Vis spectra of the reaction mixture at 1.0 M [H ]
show that the absorbance at 350 nm and 420–470 nm decays
with time, while the absorbance at 570 nm increases, without
an isosbestic point, Fig. 1. As was mentioned previously, this
lack of an isosbestic point indicates the presence of two or more
Table 2 Computed DFT parameters for the carbinolic protons of
V
V
6
5
VI
the three oxo-Cr complexes depicted in Scheme 3. [Cr O(O ,O -
competing reactions at any time during the Cr reduction to
-
V
ring
6
galactofuranuronate)
galactopyranuronate)(O ,O -galactofuranuronate)] (structure III)
2
]
(structures
I
and II) and [Cr O(O ,O -
III
Cr and several intermediate chromium species are present in
5
6
appreciable concentrations.
A
x
/MHz
A
y
/MHz
A
z
/MHz
aiso /MHz
V
Characterization of oxo-Cr –Galur species
Structure I
Structure II
Structure III
H1
H2
H1
0.5
1.4
1.2
1.2
-0.9
1.9
-5.6
-5.6
-5.5
-5.5
-6.8
-3.1
-5.9
-5.9
-5.8
-5.7
-7.6
-3.4
-3.7
-3.4
-3.4
-3.3
-5.1
-1.7
V
The majority of the literature on the identification of Cr
carbohydrate complexes with EPR is based on the interpretation
of room-temperature CW-EPR experiments. Although relevant
fingerprint information may be obtained from CW-EPR spec-
H2a
10
H1b
H2
a
b
troscopy, there is a significant risk that spectral over-interpretation
Galactofuranuronate side; Galactopyranuronate side.
1
takes place. For this reason, low-temperature CW-EPR,
H
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7039

View Article Online
Davies ENDOR and DFT methods were used to substantiate
are: a) u
4
: u
5
ª 400/1; b) u
4
: u
6
ª 700/1 and c) u
5
: u
6
ª 2/1. This
V
VI
IV
the earlier suggestions for the Cr species resulting from Cr –
Galur mixtures at low pH. The present DFT computations
clearly shows that, even at low temperature, Cr reacts faster than
V
VI
the Cr and Cr species, indicating that in spite of the fact that
1
IV
VI
in combination with low-temperature CW-EPR and H Davies
Cr is formed during Galur oxidation with Cr , this intermediate
does not accumulate and should not be considered in the fitting of
experimental kinetic data. Therefore, the time dependence of the
reaction absorption data at several wavelengths can be fitted with
the sequence proposed in Scheme 2. Besides, the evolution of the
EPR area in time can be fitted with eqn (13) and the first-order
rate constants obtained from these measurements agree perfectly
with those calculated from the electronic spectroscopy data.
By taking into account the current and previous experimental
V
6
5
ENDOR experiments corroborate the presence of [Cr O(O ,O -
-
galactofuranuronate)
2
] (I and II in Scheme 3) species at low pH,
V
ring
6
5
6
but the presence of a [Cr O(O ,O -galactopyranuronate)(O ,O -
galactofuranuronate)] (III in Scheme 3) could not be confirmed.
In fact, the room-temperature CW-EPR data can be simulated
relatively well assuming only two triplets and also the low-
temperature data are well reproduced with two contributions
V
(
Fig. 7). Cr has been reported to favor five-membered over
43–45
27
six-membered chelates.
for the formation of such Cr chelates, since it possesses 2-
Additionally, Galur is well suited
results, more particularly (a) the kinetic results, (b) the poly-
V
merization of acrylonitrile added to the reaction mixture, (c) the
VI
hydroxycarboxylato and vic-diolato sites for potential chelation.
detection of an intermediate Cr ester at high pH, (d) the detection
V
of Cr species by EPR, (e) the absence of an isosbestic point in
2
+
Kinetics analysis
Fig. 1, (f) the observation of CrO
2
, and (g) the formation of
Galar as the only organic reaction product, new light can be shed
on the possible mechanism for the reaction of Galur with Cr ,
V
IV
Formation of Cr and/or Cr intermediates in the redox reaction
VI
VI
with Cr has been previously observed for the chromic oxidation
of organic substrates.
for reactions of Cr with aldoses, alditols, aldonic acids and
methyl glycosides,
first-order steps, as found in the Galur case. For all the oxidation
reactions of these substrates with Cr , the presence of Cr has
being considered, but not experimentally proved. In the case of
Galur, the detection of CrO
the observation of relatively long-lived Cr species in the EPR
experiments indicated that both Cr and Cr intermediate species
are formed in this reaction. The molar extinction coefficients
for the Cr species were experimentally obtained as is indicated
in the experimental section. This parameter is important in the
fitting of the kinetics data. Similar coefficients to those reported
for other Cr compounds.
sugar complex is oxo-Cr –Galur, it was not appropriate to use the
molar extinction coefficient value already reported on literature
for [Cr (O)(ehba)
Due to the lack of accurate data about the stability of Cr –
Galur under experimental conditions, Cr (aq) was generated
Scheme 4.
11,46
It is also known that the kinetic data
+
VI
VI
At the [H ] and [Cr ] used in the kinetics studies, Cr exists
VI
19
47
48
- 50
as HCrO
4
.
It was proposed that this species is the reactive
14,17
generally can be fitted with two consecutive
VI
form of Cr , in agreement with the first-order dependence of
the reaction rate on [Cr ]. Also, oxidation of alcohols by Cr is
VI
VI
VI
IV
26,31,51
preceded by the formation of a chromate ester.
In previous
27
work, the observation of the absorbance bands characteristic of
chromate oxy-esters around 360 and 387 nm 2 min after mixing
2
+
VI
2
in the reaction with Cr along with
V
VI
Galur and Cr , under conditions where the redox reaction was
slow, indicated that at least two intermediate Cr complexes are
IV
V
VI
rapidly formed prior to the redox steps. Thus, the first step of the
V
VI
mechanism consists of the formation of Galur–Cr mono-chelates
(
eqn. a, IV–VI, Scheme 4). Although different coordination modes
VI
VI
are possible for the Cr –Galur species, the complex with Cr
V
10,24,26
V
were obtained. As the Cr (O)–
bound to the anomeric hydroxyl group (VI) is the only redox-
active intermediate, since oxidation of Galur occurs only at this
position. This complex must also be in rapid equilibrium with
the other linkage isomers (IV–VI), because the absorbance bands
V
V
-
2
] (ehba = 2-etyl-2-hydroxybutanoic acid).
IV
27
around 360 and 387 nm from the esters decay with the same rate.
49
IV
VI
The second step proposed in Scheme 4 is the Cr consumption
IV
in situ (see experimental section) to test the Cr –Galur re-
in a slow step that involves intramolecular two-electron transfer
actions instead of using a ligand exchange reaction between
VI
IV
within the active Galur–Cr complex (VI) to yield Cr and Galar,
eqn. b. According to the kinetics data, conversion of the active
IV
2-
[
Cr (O)(ehba)
2
]
and Galur. Because of this, the presence of
the different detected chromium intermediates must be carefully
analyzed to determine whether all or only one species needs to be
considered to fit the experimental kinetics data. Oxidation rates,
(n = 4,5,6), derived for Cr , Cr and Cr (eqn (16), 17 and 18),
were compared using the k values obtained from eqn (4), 12 and 9
IV
intermediate (VI) into Cr and Galar requires two protons. Hence,
the redox reaction should be favored in acid medium, as was indeed
VI
observed, eqn. b. The rate law for Cr consumption derived from
VI
V
IV
u
n
eqn. a and b in Scheme 4 is:
n
VI
+
2
VI
+
(19)
-
d[Cr ]/dt = k6a [H ] [Galur] [Cr ]
T
respectively, in the following conditions: 0.3 M Galur, 0.3 M [H ],
,07 mM Cr and [Cr ]
IV
VI
V
0
T
= [Cr ]
T
= 0.6 mM. It should be noted
VI
VI
where [Cr ] refers to the total [Cr ] in the reaction mixture. This
T
◦
that k
4
was determined at 15 C, while the other rate constants (k
) were determined at 33 C.
5
equation perfectly agrees with the experimental rate law in eqn (9),
◦
and k
6
VI
where k6a = k6S; (if K [HS]ꢁ 1).
IV
IV
As previously mentioned, the Cr reaction was faster than those
n
4
= k
4
[Cr ]
(16)
VI
V
IV
of Cr and Cr , so Cr was not considered in the fitting of the
IV
V
redox system. The value of k and the rate law for the Cr redox
4
n
5
= k
5
[Cr ]
T
(17)
step were experimentally determined. Kinetics data indicates that
IV
the Cr formed in eqn. b can be further oxidized through different
VI
n
6
= k
6
[Cr ]
T
(18)
competitive fast steps to yield different redox products, as is shown
The derived rate values are: u
4
= 4.2 ¥ 10 M s-1 ꢀ u
-
4
5
= 1.1 ¥
in Scheme 4. The first route gives Cr and Galur (radical) (eqn.
III
∑
6
-1
-7
-1
II
1
0- M s > u
6
= 6 ¥ 10 M s . The ratios between those values
c) and the other routes lead to Cr and Galar (eqn. d and e). The
7
040 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
VI
VI
V
V
VI
V
Scheme 4 Proposed mechanism for the reaction of Galur with Cr . K , K
1
and K
2
are formation constants for oxoCr , oxo-Cr -monochelate and
V
∑
oxo-Cr -bischelate respectively. HS = Galur; w = water; P = Galar; R = radical.
first reaction is supported by the observation of a polymerization
of acrylonitrile when added to the reaction mixture. The second
IV
⎛
⎜
⎝
k ′K ⎞
a
d[Cr
]
27
⎜
⎟
+
IV
−
=⎜k +
⎟[Cr ][HS]
(24)
⎟
⎟
T
dt
[H ]⎠
2
+
routes are supported by the formation of CrO
2
, the product of
II
the reaction of Cr with O
2
(eqn. h), which is taken as evidence of
the Cr formation. The bimolecular rate constants were found
to be inversely proportional to [H ] in the proton concentration
IV
34
IV
The thus derived rate law for the disappearance of Cr (eqn
(24)), which was in total agreement with the experimentally derived
rate law (eqn (4)), where k = k IV, k¢ K
[Galur]. Eqn (24) represents a fast step for the intermediacy of
Cr during the chromic oxidation of Galur.
Scheme 4 shows that Cr is produced through two methods:
(1) by a fast reaction between Cr and Cr (eqn. f)
by a rapid reaction of Galur radical with Cr (eqn. g). In the
probing of the former reaction, the [Cr ] used was higher than
+
I
II
range 0.1–0.3 M (Fig. 3B, inset). The rate law can be deduced
using the known acid–base equilibrium between the carboxylic
a
= k IV and [SH]
T
is the total
-
IV
acid (HS) and the conjugate base (S ):
V
K
−
+
a
HS ꢂꢀ ꢀꢀ ꢀꢀꢁꢀ S + H
(
(
(
20)
21)
22)
II
VI
34,35
and (2)
∑
VI
VI
IV
k
II
+
HS +Cr ⎯ ⎯→ P +Cr + H
2
+
VI
the one employed for the detection of CrO
2
, therefore Cr can
for Cr (Fig. 4). The last reaction
Galur /Cr ) is so fast that the organic radicals could not be
II
successfully compete with O
2
−
IV
k ′
II
S +Cr ⎯ ⎯→ P +Cr
∑
VI
(
V
detected by EPR. The kinetics data in Fig. 6 indicate that Cr
formed in the fast steps can further oxidize Galur through two
competitive slow steps to yield Cr and Galar as the final redox
products in both cases (eqn. i). Therefore, on the basis of the
kinetics studies using EPR and optical absorption spectroscopy,
Cr is proposed to react with Galur to form oxo-Cr (Galur) mono-
IV
−
+
⎛
⎜
⎞
d[Cr
]
k[S ][H ]
−
⎟
IV
−
=⎜
+[S ]k ′⎟[Cr
]
III
⎟
⎜
⎜
⎝
Ka
⎟
⎠
dt
(23)
⎛
⎜
⎜
⎜
⎜
⎜
+
⎞
T
kK [HS] [H ] k ′K [HS] ⎟
a
T
a
⎟
+
IV
=
+
⎟[Cr
]
+
⎟
V
V
K +[H ]⎟
⎟
a
⎠
(Ka
+
[H ] K
)
⎝
a
V
V
chelates (K
1
) with rapid further reaction to form oxo-Cr (Galur)
2
+
52
V
V
if [H ] ꢁ K
a
bis-chelates (K
2
). The oxo-Cr (Galur)
2
bis-chelates yield the
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7041

View Article Online
redox products in the presence of another Galur molecule, directly
Methods
DFT computations. Spin-unrestricted DFT computations
V
(
k
5I) or through an acid-catalyzed step (k5II). The Cr (Galur) and
V
Cr (Galur)
2
complexes in eqn. i represent several linkage isomers
56–59
were performed with the ORCA package (version 2.7.b)
on
as earlier described in the EPR section, but the selectivity of the
oxidation of Galur requires that the complex with the anomeric
V
five different Cr –Galur complexes (see main text and ESI†).
To simulate the solvent, a dielectric surrounding with the di-
electric constant of water was used according to the COSMO
V
V
hydroxyl group bound to Cr should be the redox active Cr
species. The fact that Galur was oxidized at a higher rate than
any of the aldoses (either pyranoses or furanoses), which were
also oxidized at the hemiacetalic position forming stable oxo-
60
model.
For the geometry optimizations, the Becke-Perdew density
6
1–63
V
functional (BP86)
identity method.
was used together with the resolution of the
The Ahlrichs split-valence plus polarization
Cr (diolate)
2
species at pH > 1, suggests that the redox-active
6
4,65
V
V
1
6
6
5
oxo-Cr species could be of the type [Cr (O)(O ,O -Galur)(O ,O -
Galur)] , with C –O and the carboxylate bound to Cr . The DFT
computations show that complexes of this type are less stable than
the [Cr (O)(O ,O -galactofuranuronate)
and DFT data show that oxo-Cr bischelates are the major Cr
-
1
V
(SV(P)) basis set was used for all atoms, except for the chromium
6
6,67
atom a more polarized triple-z valence (TZVPP) basis
was
-
8
V
6
5
-
used. The energy was converged to 1 ¥ 10 hartree (Eh), and the
2
] complexes. The EPR
-
4
V
V
convergence tolerances in the geometry optimization were 3 ¥ 10
6
-
VI
Eh/bohr for the gradient and 5 ¥ 10 Eh for the total energy.
No symmetry restrictions were imposed during the geometry
optimization. The coordinates of the optimized structures are
given in the ESI.†
species observed for Galur : Cr ratios > 30 : 1 at pH £ 5. Oxo-
V
Cr (Galur) monochelates are only observed as minor species when
VI
27
the Galur : Cr ratio £ 10 and [H+] > 0.25 M. Therefore, the rate
law for the disappearance of Cr can be described as:
V
68
Single-point calculations with the B3LYP functional and
using the optimized geometries were carried out to predict the
EPR spectral parameters. In these calculations, the same basis
sets were used as in the geometry optimization except for the
chromium basis, which was the triply polarized “Core Properties”
V
+
2
V
-
d[Cr ]/dt = {k5I + (k5II [H ] )} [HS] [Cr ]
T
(25)
with k5I = k and k5II = kS5, which is in total agreement with the
V
V
2
V
experimental rate law (eqn (12) and (15)) for K
I
K
II [HS] ꢀ K
I
(
CP(PPP)) basis. This basis was based on the TurboMole DZ
[
HS].
67
basis developed by Ahlrichs and co-workers. The Ahlrichs
polarization functions were obtained from the Turbo-Mole basis
set library and the Barone “EPR-II” basis set was used for the
protons.
V
The excess of Galur over Cr in all the experiments also favors
the bischelate formation. The redox reaction can still occur at low
H ], through the acid independent paths (k5I), such as observed
69
70
+
[
in the EPR measurements.
◦
EPR measurements. Room-temperature (20 ± 1 C) CW-EPR
spectra were obtained with a Bruker ESP 300 E spectrometer
operating at X-band frequencies (~ 9.44 GHz). Spectra were
recorded as first derivatives of the microwave absorption using
Experimental section
Materials
1
0 mW microwave power and 100 kHz modulation frequency.
a-D-Galacturonic acid monohydrate (Sigma, 98%, mp 161–
Different modulation amplitudes were used as specified in the
text.
◦
162 C), glutathione reduced form, GSH (Sigma, 98–100%),
ehba = 2-ethyl-2-hydroxybutanoic acid (Aldrich 99%), potassium
Low-temperature CW-EPR spectra and pulsed ENDOR exper-
iments were performed using a Bruker E580 Elexsys spectrometer
(9.76 GHz) equipped with a liquid He cryostat from Oxford, Inc.
All pulsed ENDOR experiments were performed using a repetition
rate of 1 kHz. An ENI-3100L rf amplifier was used at 0 dB rf power.
The magnetic field was measured with a Bruker ER 035 M NMR
Gauss meter.
5
3
dichromate (Mallinckrodt), sodium chromate (Merck), Cr
2
O
3
(
Atomic Energy Research Establishment, Harwell, England,
9
5%), perchloric acid (A.C.S. Baker), sodium hydroxide (Cicarelli,
2
2
p.a.),
H
2
O (CortecNet, 99.89% H atom) were used without
V
further purification. Na[Cr (O)(ehba)
cording to the method described in the literature. Water was
deionised (purite Select Analyst HP, Thame Oxon, UK) or doubly
2
]·H
2
O was synthesized ac-
53
71
Davies-ENDOR experiments were performed using the mi-
distilled over KMnO
range, the pH of the solutions was adjusted by addition of 0.5 M
HClO or 1.0 M NaOH. In experiments performed at constant
ionic strength (I = 1.0 M) and different proton concentrations,
mixtures of sodium perchlorate solutions and perchloric acid
solutions were used. Sodium perchlorate solutions were prepared
dissolving the salt in appropriated amount of water, to reach
a 7.12 M solution. The concentration of stock solutions of
perchloric acid was determined by titration employing standard
analytical methods.
4
. For experiments performed in the 1–5 pH
crowave pulse sequence p-T-p/2-t-p-t-echo with tp/2 = 100 ns, t
p
=
200 ns, T = 36.5 ms, and t = 300 ns. An rf pulse of variable frequency
and length 35 ms was applied during time T. The rf increment was
set to 100 kHz.
4
The EPR samples were generated as follows. Long-lived oxo-
V
Cr –Galur complexes were generated by two different methods: by
reaction of K
2
Cr
2
O with an aqueous solution containing Galur
7
VI
and by mixing GSH and K
2
Cr
2
O
7
where [Cr ] = [GSH] at pH 3.
= 0.33 mM. The samples
VI
[Cr ]
0
= [GSH]
0
= 0.67 mM, [Galur]
0
were immediately measured (room-temperature CW-EPR) or
immediately frozen in liquid nitrogen (for all other EPR/ENDOR
experiments).
VI
V
CAUTION. Cr compounds are human carcinogens, and Cr
54
complexes are mutagenic and potential carcinogens. Contact
with skin and inhalation must be avoided. Acrylonitrile is a
carcinogen and must be handled in a well-ventilated fume hood.
The CW-EPR spectra were simulated using EasySpin, a Matlab
55
72
tool.
7
042 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
IV
2+
2+
In situ generation of oxo-Cr , CrO . Aqueous CrO can
440 nm. The goodness of fit at these wavelengths was used to
corroborate the rate expressions used to obtain the rate constants
2
+
be generated by rapid oxygen-oxidation of Cr employing the
following procedure: Zn/Hg amalgam was prepared in a 50 mL
balloon by stirring a mixture of Zn (~ 10.0 g, previously washed
VI
at 350 nm. The first-order dependence of the rate upon [Cr ]
VI
was verified in a set of experiments where the [Cr ] was varied
0
with 3.0 M HCl during 5 min) and HgCl
for about 30 min. The excess of HgCl was eliminated and the
resulting amalgam washed three times with 0.5 M H
finally with distilled water. Then, 35.0 mL of 0.2 M HClO
2
(0.3 M in 1.0 M HCl)
between 0.3–0.6 mM, but T, [Galur] and ionic strength were kept
2
constant.
VI
2
SO
4
and
and
Time evolution of the Galur–Cr reaction. Time-dependent
4
UV/Vis spectra were recorded for a reaction mixture containing:
VI
water were added in order to reach pH 1.0 and the balloon closed
with a rubber septum-cap and stirring with H bubbling for at
least 15 min to ensure a reducing medium. After this time, 2.0 mL
of 6.0 mM Cr(ClO were injected, keeping the H bubbling and
the stirring constant. After 3.0 h, Cr(ClO was quantitatively
reduced into Cr . The [Cr ] was determined by treating a
0.6 M Galur, 0.1 M HClO
4
, I = 1.0 M and 0.6 mM Cr at
◦
2
33 C, between 200–800 nm every 5 min until total consumption
of Cr and to be able to determine if the final redox product of
the reaction was Cr (aq) or Cr –ligand. Time evolution of the
Galur–Cr mixture was monitored by following changes in the
VI
III
III
4
)
3
2
VI
4
)
3
2
+
2+
VI
570 nm absorption band. In these experiments, the [Cr ]
0
was
III
-3
reaction aliquot with an aqueous solution of [Co (NH
3
)
5
Cl]Cl
2
,
kept constant at 6.0 ¥ 10 M, [Galur] was 0.70 M, in 0.50 M
. The concentration of Cr used in these experiments was
10 times higher because [Cr(OH ) ] has a very small e at 570 nm.
VI
under anaerobic atmosphere (Ar); the mixture was then poured
HClO
4
II
3+
into concentrated HCl and the [Co ] content was analyzed by
2
6
2
-
46
measuring the absorbance of [CoCl
4
]
at 692 nm. For the
Rate constants obtained at this wavelength are in agreement with
those calculated from data at 350 nm for these experimental
conditions.
2
+
2+
in situ generation of CrO , a deoxygenated solution of Cr was
injected into an acidic aqueous solution of Galur, saturated with
2
+
IV
2+
O
2
(1.26 mM). In a typical experiment, 100 mL of 6.0 mM Cr
was injected into a septum-capped spectrophotometric quartz
cell, with a 1.0 cm path length, filled with 2.3 mL of an O
saturated solution containing 1.0–2.5 mM Galur and appropriate
Galur–Cr reactions. The oxidation reaction of Galur by CrO
was spectrophotometrically monitored following the appearance
2
+
2
-
of CrO
2
as the final redox product. Kinetic data were collected
2
+
spectrophotometrically following the formation of CrO
290 nm (e = 3000 M cm ) at 15 C. At this wavelength, neither the
2
at
+
-1
-1
◦
concentrations of HClO
4
and NaClO
4
([H ] = 0.1–0.3 M, I = 1.0 M)
◦
2+
at 15 C. Under these conditions, CrO was immediately formed
substrate, Galur, nor the oxidized products absorb. In the kinetics
2
+
2+
IV
and reacts with Galur. At very low Cr /O
was quantitatively formed, while at intermediate Cr /O
2
ratios (< 0.05), CrO
2
measurements, [Cr ], I, [O
2
] and temperature were kept constant
2
+
◦
2
ratios
. The
at 0.07 mM, 1.0 M, 1.26 mM and 15 C respectively. The [HClO
4
]
2
+
2+ 34
(
~ 0.15), the reaction afforded mixtures of CrO and CrO
2
varied between 0.1–0.3 M and the employed [Galur] range was
2
+
2+
2+
reaction between Cr and O
(
2
rapidly produced 0.07 mM CrO
chosen in such a way to avoid CrO disproportionation, between
2
+
28% average based on total [Cr ]), which then reacted with the
organic substrate to produce Cr and oxidized organic products.
0.1–0.4 mM. The experimental pseudo first order rate constants
2
+
(k ), obtained from nonlinear least square fits of absorbance data
4
2
+
These experimental conditions, allowed Cr formed by reaction
at 290 nm were averages of at least six determinations and were
within ± 10% of each other. Data used to calculate the kinetic
2
+
2+
of Galur with CrO , to be quantitatively transformed into CrO
2
2
+
(
Cr /O
2
ratio < 0.05) and no autocatalytic consumption of
constant, k
4
, correspond to 80% of the exponential growth of the
2
+
2+
34,73
2+
CrO
of Cr with O
Cr into 2.3 mL of an O
(
2
by Cr occured.
The [CrO ] generated by reaction
was determined by injection of 100 mL of 6.0 mM
-saturated solution of 0.1 M ehba buffer
pH 3.0), at 15 C. Immediately after, solution turned on a pink
experimental values. The first order dependence of the rate upon
2
+
IV
IV
2
[Cr ] was verified in a set of experiments where the [Cr ] was
0
2
+
2
varied between 0.030–0.062 mM, but T, [Galur] and I were kept
constant.
◦
IV
2-
color and the absorbance of [Cr (O)(ehba)
2
]
at 512 nm (e =
Determination of the molar extinction coefficients (e) of oxo-
380 M- cm ) was measured.
1
-1
49
Cr –Galur species. The adopted method for determining the
V
2
V
e of the oxo-Cr –Galur species was ligand exchange between
V
-
Spectrophotometric measurements. All kinetics measurements
[Cr (O)(ehba)
2
] and Galur: 2.0 mL of Galur 1.0 M (pH = 3.0, I =
-
4
V
-
were made by monitoring the absorbance changes using on
a Jasco V-550 spectrophotometer with fully thermostated cell
1.0 M) were mixed with 10 mL of 1.25 ¥ 10 M [Cr (O)(ehba)
reach a final concentration of (5 ¥ 10 M) in buffer ehba (pH 3.0)
at 25 and 33 C, in order to obtain a Galur : Cr ratio of 2000 : 1.
2
] to
-
4
◦
◦
V
compartments (± 0.2 C). The reactions were followed under
pseudo-first-order conditions, using an excess of Galur over Cr.
Reactant solutions were thermostated prior to the experiment and
transferred into a quartz cell of 1.0 cm path length immediately
after mixing. All kinetics data were fitted exploiting routines
written in the Origin6.0 package.
After that, absorbance changes vs. time were measured at 350 nm
V
-
for 800 s every 0.2 s. Stock solutions of [Cr (O)(ehba)
2
] were
prepared in 0.2 M ehba buffer, pH = 3.0.
III
Detection of superoxo-Cr ion during the reaction of Galur with
VI
VI
VI
2+
VI
Galur/Cr reactions. Disappearance of Cr was followed by
Cr . The possible formation of Cr in Galur–Cr mixtures
monitoring the absorbance at 350 nm until at least 80% conversion.
was examined by UV/Vis periodic scanning (220–500 nm) of
◦
VI
Measurements were made at 33 C. The initial concentration
solutions of 0.015 M Galur, 0.07 mM Cr and 1.0 M HClO
4
2+
VI
-4
◦
of Cr was 6.0 ¥ 10 M, whilst the Galur concentration was
saturated with oxygen ([O
is formed, it is rapidly converted by dioxygen into CrO
2
] = 1.26 mM), at 25 C. Whenever Cr
2
+
varied from 0.18 M to 0.42 M. The observed rate constants
2
3,34
, which
3
(
k
6
and k
5
) were deduced from multiple determinations and
has characteristic absorption bands at 290 and 247 nm
and can
VI
were within ± 10% of each other. The rate constants obtained
at 350 nm were used to fit the absorbance changes at 420–
be detected at low [Cr ]. Periodic scanning of the reaction mixture
VI
showed that the Cr band at 350 nm decreased in intensity, while
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7043

View Article Online
2
+
new peaks at 290 nm and 247 nm, characteristic of CrO
in.
2
, grew
27 J. C. Gonz a´ lez, V. Daier, S. Garc ´ı a, B. A. Goodman, A. M. Atria, L. F.
Sala and S. Signorella, Dalton Trans., 2004, 2288–2296.
2
8 A. Schweiger and G. Jeschke, Principles of Pulse Electron Paramagnetic
Resonance, University Press Oxford, Oxford, 2001.
2
3
9 F. Neese, Coord. Chem. Rev., 2009, 253, 526–563.
Acknowledgements
0 S. Van Doorslaer and E. Vinck, Phys. Chem. Chem. Phys., 2007, 9,
4
620–4638.
We thank the National Research Council of Argentina (CON-
ICET) PIP 0075 and the National University of Rosario (UNR)
BIO45 grants respectively. Van Doorslaer and Sala thank bilat-
eral Mincyt-FWO agreement FWO9/06. Van Doorslaer thanks
support from the Hercules Foundation, Flanders (contract
AUHA013).
3
3
3
1 G. P. Haight, G. M. Jursich, M. T. Kelso and P. J. Merrill, Inorg. Chem.,
1985, 24, 2740–2746.
2 A. B. P. Lever, in Inorganic Electronic Spectroscopy, Elsevier, Amster-
dam, 2nd edn, 1984, pp. 419.
3 S. L. Scott, A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1991, 113,
7
787–7788.
34 S. L. Scott, A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1992, 114,
4
205–4213.
3
3
5 J. C. Gonz a´ lez, S. Garc ´ı a, N. Mamana, L. F. Sala and S. Signorella,
Inorg. Chem. Commun., 2006, 9, 437–440.
References
6 A. Nemes and A. Bakac, Inorg. Chem., 2001, 40, 2720–2724.
1
2
3
4
5
International Agency for Research on Cancer (IARC). Overall Evalu-
ations of Carcinogenicity to Humans, IARC, 203, http//www.iarc.fr.
J. Guertin, J. A. Jacobs and C. P. Avakian, ed. Chromium (VI)
Handbook; CRC Press, Boca Raton 2005.
37 T. W. Swaddle, In Inorganic Chemistry, Academic Press, California,
USA, 1997, ch.2, pp. 24–25.
38 J. H. Espenson, in Chemical Kinetcis and Reactions Mechanism, Mc
Graw Hill, New York, 2nd edn, 2002, pp 70–90.
39 R. Bramley, J. Y. Ji, R. J. Judd and P. A. Lay, Inorg. Chem., 1990, 29,
3089–3094.
40 G. Barr-David, M. Charara, R. Codd, R. P. Farrell, J. A. Irwin, P. A.
Lay, R. Bramley, S. Brumby, J. Y. Ji and G. R. Hanson, J. Chem. Soc.,
Faraday Trans., 1995, 91, 1207–1216.
41 R. P. Farrell, R. J. Judd, P. A. Lay, R. Bramley and J. Y. Ji, Inorg. Chem.,
1989, 28, 3401–3403.
42 R. Kapre, K. Ray, I. Sylvestre, T. Weyherm u¨ ller, S. DeBeer
George, F. Neese and K. Wieghardt, Inorg. Chem., 2006, 45, 3499–
3509.
R. Cood, J. A. Irwin and P. A. Lay, Curr. Opin. Chem. Biol., 2003, 7,
2
13–219.
A. Levina, L. Zhang and P. A. Lay, J. Am. Chem. Soc., 2010, 132,
720–8731.
A. Levina and P. A. Lay, Coord. Chem. Rev., 2005, 249, 281–
98.
8
2
6
7
M. Stefanikou and C. Maravelias, Curr. Top. Toxicol, 2004, 1, 161–167.
G. N. Babu, R. Ranjani, G. Farceda and S. D. S. Murthy, J. Phytol.
Res., 2007, 20, 1–6.
A. Levina, R. Codd, P. A. Lay, In Biological Magnetic Resonance,
G. R. Hanson, L. J. Berliner, Es; Springer Publishers; New York, 2009,
Vol 28. Part 4, pp 551–579.
8
43 M. Branca, G. Micera and A. Dess ´ı , Inorg. Chim. Acta, 1988, 153,
61–65.
9
K. D. Sudgenm, D. M. Stearns and J. Environ, Pathol. Toxicol. Oncol,
44 R. Codd and P. A. Lay, J. Am. Chem. Soc., 1999, 121, 7864–
7876.
2
000, 19, 215–230.
1
1
0 R. Codd, C. T. Dillon, A. Levina and P. A. Lay, Coord. Chem. Rev.,
001, 216–217, 537–582.
1 S. Signorella, C. Palopoli, M. Santoro, S. Garc ´ı a, V. Daier, J. C.
Gonz a´ lez, V. Rold a´ n, M. I. Frascaroli, M. Rizzotto and L. F. Sala,
Trends Inorg. Chem., 2001, 7, 197–207.
45 S. P. Kaiwar, M. S. S. Raghavan and C. P. Rao, Carbohydr. Res., 1994,
2
256, 29–40.
46 O. A. Babich and E. S. Gould, Inorg. Chem., 2001, 40, 5708–
5710.
47 V. P. Rold a´ n, V. A. Daier, B. Goodman, M. I. Santoro, J. C. Gonz a´ lez,
N. Calisto, S. R. Signorella and L. F. Sala, Helv. Chim. Acta, 2000, 83,
3211–3228.
48 S. Signorella, M. Santoro, C. Palopoli, C. Brondino, J. M. Salas-
Peregrin, M. Quiroz and L. F. Sala, Polyhedron, 1998, 17, 2739–
2749.
1
1
2 B. Gyurcsik and L. Nagy, Coord. Chem. Rev., 2000, 203, 81–149.
3 A. Levina, R. Codd, C. T. Dillon and P. A. Lay, Prog. Inorg. Chem.,
2
003, 51, 145–250.
1
4 M. Rizzotto, A. Levina, M. Santoro, S. Garc ´ı a, M. I. Frascaroli, S.
Signorella, L. F. Sala and P. A. Lay, J. Chem. Soc., Dalton Trans., 2002,
3
206–3213.
49 R. Codd, P. A. Lay and A. Levina, Inorg. Chem., 1997, 36, 5440–
1
5 V. Rold a´ n, J. C. Gonz a´ lez, M. Santoro, S. Garc ´ı a, N. Casado, S.
Olivera, J. C. Boggio, J. M. Salas-Peregrin, S. Signorella and L. F.
Sala, Can. J. Chem., 2002, 80, 1676–1686.
6 S. Signorella, R. Lafarga, V. Daier and L. F. Sala, Carbohydr. Res.,
2
7 S. Signorella, M. I. Frascaroli, S. Garc ´ı a, M. Santoro, J. C. Gonz a´ lez,
C. Palopoli, V. Daier, N. Casado and L. F. Sala, J. Chem. Soc., Dalton
Trans., 2000, 1617–1623.
5448.
50 N. E. Brasch, D. A. Buckingham, A. B. Evans and C. R. Clark, J. Am.
Chem. Soc., 1996, 118, 7969–7980.
1
1
51 M. Mitewa and P. R. Bontchev, Coord. Chem. Rev., 1985, 61, 241–
000, 324, 127–135.
272.
52 G. M. Escandar and L. F. Sala, Can. J. Chem., 1992, 70, 2053–
2057.
53 M. Krumpolc and J. Rocek, Inorg. Synth., 20, 63–65.
54 International Agency for Research on Cancer (IARC). Monogr. Eval.
Carcinog. Risks Hum. Monograph 49, Updated 1997, 1990, Lyon,
France.
55 R. Feldam, in: Occupational and Environmental Neurotoxicology,
Lippincott–Raven Publishers, Philadelphia, 1999, pp. 337–338.
56 F. Neese, J. Chem. Phys., 2001, 115, 11080–11096.
57 F. Neese, J. Phys. Chem. A, 2001, 105, 4290–4299.
58 F. Neese, J. Chem. Phys., 2003, 118, 3939–3948.
59 F. Neese, J. Chem. Phys., 2005, 122, 034107–034119.
60 S. Sinnecker, A. Rajendran, A. Klamt, M. Diedenhofen and F. Neese,
J. Phys. Chem. A, 2006, 110, 2235–2245.
1
1
2
2
2
8 M. Rizzotto, V. Moreno, S. Signorella, V. Daier and L. F. Sala,
Polyhedron, 2000, 19, 417–423.
9 S. Signorella, V. Daier, S. Garc ´ı a, R. Cargnello, J. C. Gonz a´ lez, M.
Rizzotto and L. F. Sala, Carbohydr. Res., 1999, 316, 14–25.
0 S. Signorella, S. Garc ´ı a and L. F. Sala, J. Chem. Educ., 1999, 76, 405–
4
08.
1 M. I. Frascaroli, J. M. Salas-Peregrin, L. F. Sala and S. Signorella,
Polyhedron, 2009, 28, 1049–1056.
2 J. C. Gonz a´ lez, S. I. Garc ´ı a, S. Bell u´ , A. M. Atria, J. M. Salas Peregrin,
A. Rockenbauer, L. Korecz, S. Signorella and L. F. Sala, Polyhedron,
2
009, 28, 2719–2729.
2
2
3 S. E. Bell u´ , J. C. Gonz a´ lez, S. I. Garc ´ı a, S. Signorella and L. F. Sala,
J. Phys. Org. Chem., 2008, 21, 1059–1067.
61 J. P. Perdew, Phys. Rev. B, 1986, 33, 8822–8824.
62 J. P. Perdew, Phys. Rev. B, 1986, 34, 7406–7406.
63 A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100.
64 B. I. Dunlap, J. W. D. Connolly and J. R. Sabin, J. Chem. Phys., 1979,
71, 3396–3402.
65 O. Vahtras, J. Alml o¨ f and M. W. Feyereisen, Chem. Phys. Lett., 1993,
213, 514–518.
4 M. F. Mangiameli, J. C. Gonz a´ lez, S. Garc ´ı a, S. Bell u´ , M. Santoro, E.
Caffaratti, M. I. Frascaroli, J. M. Salas Peregr ´ı n, A. M. Atria and L. F.
Sala, J. Phys. Org. Chem., 2010, 23, 960–971.
2
2
5 B. Manunza, S. Deiana, M. Pintore and C. Gessa, Carbohydr. Res.,
1
997, 300, 85–88.
6 J. C. Gonz a´ lez, S. Garc ´ı a, S. Bell u´ , J. M. Salas Peregr ´ı n, A. M. Atria,
66 A. Sch a¨ fer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97, 2571–
2577.
L. F. Sala and S. Signorella, Dalton Trans., 2010, 39, 2204–2217.
7
044 | Dalton Trans., 2011, 40, 7033–7045
This journal is © The Royal Society of Chemistry 2011

View Article Online
6
7 The Ahlrichs (2df,2pd) polarization functions were obtained
from the TurboMole basis set library under ftp.chemie.uni-
karlsruhe.de/pub/basen, R. Ahlrichs and co-workers (unpublished).
8 P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys.
Chem., 1994, 98, 11623–11627.
9 ftp.chemie.uni-karlsruhe.de/pub/basen: 2p functions from A. J. H.
Wachters, Chem. Phys., 1970, 52, 1033 plus one f-function from the
TurboMole library.
70 V. Barone, in Recent Advances in Density Functional Methods, ed.
D. P. Chong, World Scientific Publ. Co., Singapore, 1996, pp. 287–
333.
71 E. R. Davies, Phys. Lett. A, 1974, 47, 1–2.
72 S. Stoll and A. Schweiger, J. Magn. Reson., 2006, 178, 42–
55.
6
6
73 M. E. Brynildson, A. Bakac and J. H. Espenson, J. Am. Chem. Soc.,
1987, 109, 4579–4583.
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 7033–7045 | 7045