Kinetic study of the oxidation of ascorbic acid by aqueous copper(II) catalysed by chloride ion

Article abstract of DOI:10.1039/a701574g

Two methods of monitoring the chloride-catalysed oxidation of ascorbic acid by aqueous Cu^II have been developed that allow the reaction conditions to be varied more widely than previously and thereby permit a fuller elucidation of the behaviour and rate law for this system. It has been possible to study the reaction over a wide concentration range from 1.6 to 500 mM Cl^- and 4 to 100 mM Cu^II. For the first time, it is shown that the ascorbic acid (H₂asc)-Cu^II-chloride ion-Cu^I-dehydroascorbic acid (dha) system comes to equilibrium, under anaerobic conditions, with all species present and is driven towards products by chloride complexation of Cu^I. The results are consistent with the following reaction scheme (i = 0-2). The full rate law has been elucidated and values for various k_1i, k_-2i, and k_2i/k_-1i have been determined. (Chemical Equation Presented).

Full text of DOI:10.1039/a701574g

View Article Online / Journal Homepage / Table of Contents for this issue
Kinetic study of the oxidation of ascorbic acid by aqueous copper(II)
catalysed by chloride ion*
Margaret J. Sisley and Robert B. Jordan
Department of Chemistry, University of Alberta, Edmonton, Alberta, T6G 2G2, Canada
Two methods of monitoring the chloride-catalysed oxidation of ascorbic acid by aqueous CuII have been
developed that allow the reaction conditions to be varied more widely than previously and thereby permit a fuller
elucidation of the behaviour and rate law for this system. It has been possible to study the reaction over a wide
concentration range from 1.6 to 500 m ClϪ and 4 to 100 m CuII. For the first time, it is shown that the ascorbic
acid (H2asc)–CuII–chloride ion–CuI–dehydroascorbic acid (dha) system comes to equilibrium, under anaerobic
conditions, with all species present and is driven towards products by chloride complexation of CuI. The results
are consistent with the following reaction scheme (i = 0–2). The full rate law has been elucidated and values for
various k1i, kϪ2i and k2i/kϪ1i have been determined.
k1i
HascϪ ϩ CuIICliϩ2Ϫi
CuICliϩ1Ϫi ϩ asc Ϫ ϩ Hϩ
᝽
kϪ1i
k2i
kϪ2i
asc Ϫ ϩ CuIICliϩ2Ϫi
CuICliϩ1Ϫi ϩ dha
᝽
There have been several studies of the copper()–ascorbic acid
system, and the earlier work has been summarized by Xu and
Jordan.1 In the latter study the dependence of the rate on
chloride-ion concentration was quantified, but the mechanistic
source of this catalysis was not established. The present study
was undertaken to extend the results of Xu and Jordan using
different monitoring methods, and to expand the experimental
concentration range of the reactants in order more fully to
expose the rate law with the hope that this would provide
insight into the mechanism. An understanding of the mechan-
ism should allow one to determine whether chloride-ion cata-
lysis of oxidations by copper() can be expected to be a general
phenomenon or is unique to ascorbic acid.
HO
O
O
O
O
HOCH
O
O
–2e–
HOCH2
0.39 V
OH
O
HO
OH
OH
OH
H2A
dhae
–H+
–2e–
K = 6x106 +H2O
pKa
4
0.27 V
O
HOCH
HOCH
O
O
–2e–
HOCH2
HOCH2
0.47 V
O–
O
OH
O
HA–
dhak
–e–
–e–
0.74 V
0.24 V
The oxidation pathways of ascorbic acid are summarized in
Scheme 1, where the reduction potentials are taken from the
recent summary of TurЈyan and Kohen.2 The process involves
not only an electron transfer but also cyclization finally to give
bicyclic enol (dhae), commonly called dehydroascorbic acid.
The enol is presumed to form by hydration of the keto inter-
O
HOCH
HOCH
O
O
pKa –0.4
–H+
•
•
HOCH2
+
HOCH2
O–
O–
OH
O
mediate (dhak). The one-electron oxidation radical (A Ϫ) has
•
• –
᝽
HA
A
been detected in pulse radiolysis–ESR studies,3 but the speci-
ation is somewhat ambiguous because of the formation of sev-
eral hydroxy radical adducts and their rearrangement. How-
ever, the same dominant radical also has been produced by
Scheme 1
oxidation reaction in the presence of dioxygen is driven to
products by the oxidation of CuI by O2. Under the anaerobic
conditions of the present study, chloride ion serves an analo-
gous purpose by complexing with the CuI, but then the extent
of reaction depends on the chloride-ion concentration. This
gives some control of the equilibrium position and allows a
more detailed kinetic characterization of the system.
2ϩ
photosensitized oxidation of sodium ascorbate by Ru(bpy)3
(bpy = 2,2Ј-bipyridine) in dimethyl sulfoxide.4 The stages at
which the cyclization and hydration occur are still rather
uncertain. In the ESR spectrum of A Ϫ the hyperfine coupling
᝽
to the H at C6 (at the end of the sidechain) is larger than to that
at C5, and this led to the suggestion5 that the radical is cyclized.
Kirino and Kwan6 found the same coupling for the radical
formed from the 5,6-isopropylidene derivative of ascorbic acid.
This derivative cannot cyclize, but the 5,6-isopropylidene sub-
stituent is easily hydrolysed,7 and the observations of Kiro and
Kwan may actually pertain to ascorbic acid. Further back-
ground chemistry has been summarized by Sisley and Jordan.8
The reduction potentials show that the oxidation of H2A by
aqueous CuII is thermodynamically unfavourable. The auto-
Results
In the present study the oxidation of ascorbic acid by copper()
has been followed by two monitoring techniques. In the first
method the direct disappearance of CuII has been observed at
750 nm. This method is limited to concentrations of
CuII у 3 × 10Ϫ3  because of the low molar absorptivity of
CuII(aq), and to concentrations of ClϪ у 0.03  to prevent pre-
cipitation of CuCl by forming CuCl2Ϫ. In the second method
the disappearance of Co(NH3)5(N3)2ϩ due to reduction by the
copper() product has been observed at 518 nm. Parker and
* Dedicated to Professor R. G. Wilkins on the occasion of his 70th
birthday.
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888
3883

View Article Online
Fig. 1 Variation with time of the absorbance of CuII during reduction
by ascorbic acid for runs with varying [Hϩ]; all 0.12  ClϪ and 6.0 × 10-3
 ascorbic acid in 1.00  HClO4–LiClO4 at 25 ЊC. The concentrations
of Hϩ and CuII are 0.010 and 4.01 × 10Ϫ3 (᭺), 0.0250 and 8.00 × 10Ϫ3
(ᮀ), 0.0502 and 8.00 × 10Ϫ3 (᭛) and 0.0750 and 8.00 × 10Ϫ3 (᭿)
Fig. 3 Variation with time of the absorbance of CuII during reduction
by ascorbic acid for runs with varying [CuII] and added CuI; all 0.0100 
Hϩ and 6.0 × 10Ϫ3  ascorbic acid in 1.00  HClO4–LiClO4 at 25 ЊC.
The copper() dependence is shown for runs with 0.120  ClϪ and the
copper() concentrations are 4.01 × 10Ϫ3 (᭺), 8.00 × 10Ϫ3 (ᮀ) and
10.01 × 10Ϫ3 (᭛). The copper() effect is shown for runs with Hϩ, ClϪ,
and copper() concentrations of 0.025, 0.120 and 8.00 × 10Ϫ3 ,
respectively and with initial CuI of 0 (᭿) and 4.42 × 10Ϫ3 (ϩ)
H2asc
Cu2ϩ ϩ ClϪ
Cu2ϩ ϩ 2 ClϪ
Cuϩ ϩ ClϪ
Cuϩ ϩ 2 ClϪ
HascϪ ϩ Hϩ;
Ka = 1 × 10Ϫ4 
β1II = 2.36 Ϫ1
CuIIClϩ;
CuIICl2;
CuICl;
β2II = 1.48 Ϫ2
β1I = 5.1 × 102 Ϫ1
β2I = 1.15 × 105 Ϫ2
CuICl2
Ϫ
;
Scheme 2
k1i
HascϪ ϩ CuIICliϩ2Ϫi
CuICliϩ1Ϫi ϩ asc Ϫ ϩ Hϩ
ؒ
Fig. 2 Variation with time of the absorbance of CuII during reduction
by ascorbic acid for runs with varying [ClϪ]; all 0.0100  Hϩ,
8.00 × 10Ϫ3 CuII and 6.0 × 10Ϫ3  ascorbic acid in 1.00  HClO4–
LiClO4 at 25 ЊC. The chloride concentrations are 0.0320 (᭺), 0.0599
(ᮀ), 0.120 (᭛), 0.250 (᭿) and 0.500 (ϩ)
kϪ1i
k2i
kϪ2i
asc Ϫ ϩ CuIICliϩ2Ϫi
CuICliϩ1Ϫi ϩ dha
ؒ
Scheme 3
Espenson9 observed that the reaction Co(NH3)5(N3)2ϩ
ϩ
show an initial rapid absorbance change, followed by a slow
almost linear change of absorbance with time. Qualitatively,
Figs. 1 and 2 show that the rate decreases with increasing
[Hϩ] and increases with increasing [ClϪ]. The three lower
curves in Fig. 3 show that the rate increases with increasing
[CuII], and the upper curves show that addition of CuI decreases
the rate.
CuI(aq) is fast (k = 1.5 × 103 Ϫ1 sϪ1 at 25 ЊC in 0.2  LiClO4),
so that Co(NH3)5(N3)2ϩ is an effective scavenger for CuI(aq).
The Co(NH3)5(N3)2ϩ has the advantage of a relatively high
molar absorptivity compared to other cobalt() ammine com-
plexes. This allows the study of lower copper() concentrations,
and the low steady-state [CuI] permits lower chloride-ion
concentrations to be used.
After consideration of several generic reaction schemes, it
was determined that the absorbance vs. time profiles could be
fitted if the slowing of the reaction is caused by the increasing
copper() product concentration. Since the speciation in this
system is rather well defined by previous studies, we will
proceed to a general mechanism that has been found to be con-
sistent with the observations. In developing such a scheme the
species and reactions in Scheme 2 have been considered as being
in rapidly maintained equilibria, with the values of βiII from
Ramette10 and of βiI from Arhland and Rawthorne,11 Fritz12
and Sharma and Millero.13
The overall results from the two different methods of moni-
toring the reaction are in good agreement, but the methods
differ substantially in kinetic complexity and will be discussed
separately. For the conditions of this study in which the
chloride-ion concentration is larger than the total copper(, )
concentration, the overall reaction may be represented by equa-
tion (1), where H2asc and dhae are ascorbic acid and its two-
H2asc ϩ 2 CuII ϩ 4 ClϪ
dhae ϩ 2 CuCl2Ϫ ϩ 2 Hϩ (1)
electron oxidation product dehydroascorbic acid. Represent-
ation of the copper species is based on the fact that CuII is
weakly complexed by ClϪ,10 but CuI is strongly complexed11–13
and is mainly present as CuCl2Ϫ. From the reduction potentials
of 0.17 V for CuII–CuI and 0.4 V for dhae–H2asc, it is apparent
that the reaction of aqueous CuII and H2asc is thermodynamic-
ally unfavourable. However, the overall formation constant for
CuCl2Ϫ, β2I = 1.15 × 105 Ϫ2, allows one to calculate the equi-
librium constant for equation (1) as K1 = 2.2 × 102 Ϫ2. The
reaction is exoergic and driven to the right by the complexation
of CuI.
Then the general mechanism in Scheme 3 can be formulated,
with i = 0, 1, 2. In Scheme 3, the anion HascϪ is shown as the
reactive form of the reducing agent and this will be confirmed
by the [Hϩ] dependence of the experimental reaction rate. The
product of the initial electron transfer in Scheme 3 is given as
ؒ
the ascorbate radical anion because the radical Hasc is a strong
acid with pKa ≈ Ϫ0.45, based on the pulse radiolysis–ESR
observations of Laroff et al.3
The rate law for Scheme 3 can be developed with a steady-
state assumption for the radical asc Ϫ. It also will be assumed at
᝽
this stage that the chloride-ion concentration is in pseudo-first-
order excess, a feature that is true for most but not all of these
experiments. With this assumption, the concentrations of the
copper species are given by equations (2) and (3), where
Direct monitoring of copper(II) disappearance
Some representative absorbance vs. time profiles are shown in
Figs. 1–3. The majority of these profiles are unusual in that they
3884
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888

View Article Online
[CuII]tot
DII
β1II[ClϪ][CuII]tot
[Cu2ϩ] =
;
[CuClϩ] =
;
DII
Table 1 Kinetic results for oxidation of ascorbic acid by copper()
while observing the disappearance of aqueous copper() in 1.00 
HClO4–LiClO4 at 25 ЊC
β2II[ClϪ]2[CuII]tot
(2)
(3)
[CuCl2] =
DII
[Hϩ]/
[ClϪ]/
103 [Cu2ϩ
]
tot/
103 [asc]tot/
k
Ϫ2/Ϫ1 sϪ1
0.0100
0.0100
0.0100
0.0100
0.0100
0.0100
0.0100
0.0100
0.0100
0.0100
0.0251
0.0250
0.0302
0.0601
0.0601
0.0599
0.0599
0.1200
0.1200
0.1200
0.2500
0.5000
0.0600
0.1200
8.00
5.94
5.94
8,00
8.00
4.01
8.00
10.01
8.00
8.00
8.00
8.00
8.00
8.00
8.00
8.00
8.00
8.00
6.00
2.66
2.66
3.50
6.00
5.99
6.00
5.99
6.00
6.00
6.00
5.98
6.00
3.49
6.00
5.99
5.99
6.01
0.34
1.45
1.45
1.40
1.35
6.8
6.5
5.7
28
120
1.40
6.5
5.5
1.60
1.55
6.5
1.60
6.5
I
[CuI]tot
DI
β1 [ClϪ][CuI]tot
[Cuϩ] =
;
[CuCl] =
;
DI
I
β2 [ClϪ]2[CuI]tot
[CuCl2Ϫ] =
DII
DII = 1 ϩ β1II[ClϪ] ϩ β2II[ClϪ]2 andDI = 1 ϩ β1 [ClϪ] ϩ β2 [ClϪ]2.
I
I
Then the rate of disappearance of CuII is given by equation (4),
Rate =
0.0253* 0.120
k1Ka[Cu ]tot2[Hasc]tot k1[CuI]tot [H ][dha]
II
2
ϩ








0.0502
0.0502
0.0502
0.0753
0.0750
0.0599
0.0599
0.120
0.0600
0.120

Ϫ
[Hϩ](DII)2
K00(DI)2

(4)
Ϫ2 
k1[Hϩ][CuI]tot [CuII]tot



ϩ
k
Ϫ2K00DI
DII
* Contains initially [CuI] = 4.42 × 10Ϫ3 
where the rate and equilibrium constants are defined in terms
of Scheme 2 as shown in equations (5)–(7), where K00
=
determined in the earlier study. In this case the numerical
analysis has only two adjustable parameters, k12 and kϪ2. If the
assumptions are correct, then k12 and kϪ2 should not change as
[Hϩ] is changed in different runs, and k12 should be independent
of [ClϪ]. However, kϪ2 may have a [ClϪ] dependence as given by
equation (6), and will actually vary during a given run because
the free [ClϪ] varies somewhat due to complexation with the
copper() product.
k1 = k10 ϩ k11β1II[ClϪ] ϩ k12β2II[ClϪ]2
(5)
(6)
I
I
k
Ϫ2 = kϪ20 ϩ kϪ21β1 [ClϪ] ϩ kϪ22β2 [ClϪ]2
[Cuϩ]2[Hϩ][dha]
[Cu2ϩ]2[HascϪ]
k10k20
(7)
K00
=
=
kϪ10kϪ20
In the penultimate analysis the effect of [ClϪ] variation on kϪ2
was ignored, and the apparent kϪ2 values in Table 1 were deter-
mined. These values are generally consistent with the expect-
ation described above in that they are independent of [Hϩ] and
have a 10% range for a particular initial [ClϪ]. Although not
normally unusual, such a range of values is disturbing because
kϪ2 is so well defined by the data and it turns out simply to be
due to the [ClϪ] variation. This is qualitatively apparent from
Table 1, where it can be seen that smaller kϪ2 values are
obtained at the higher [CuII]tot and [asc]tot values; these are runs
in which more CuI is produced to complex with the ClϪ. In the
final analysis it was found that kϪ2 is given by equation (10), and
4.1 × 10Ϫ4 . The first term in the denominator of equation (4)
has been obtained by substitution using the relationship in
equation (8). Simple rearrangement of equation (4) yields (9).
kϪ1/k2 = k1/kϪ2K00
(8)
d[CuII]tot
dt
=
II
2
ϩ
k1Ka[Cu ]tot2[Hasc]tot k1DII[CuI]tot [H ][dha]









Ϫ
[Hϩ]DII
K00(DI)2

(9)
Ϫ2 
k1DII[Hϩ][CuI]tot
Ϫ2K00DI



ϩ [CuII]tot
Ϫ2 = (5.1 0.1) × 102 [ClϪ]2
(10)
k
k
Qualitatively, this rate law may be expected to fit the observ-
ations because the [CuI] term in the denominator becomes larger
as the reaction proceeds and will cause the rate to decrease if
this term is significant relative to [CuII]tot. It should be noted
that during the more rapid initial change the rate will be sub-
stantially controlled by k1, while during the slower, nearly linear
change the rate will be controlled by kϪ2. Since the latter region
is dominant for many runs the value of kϪ2 is quite well defined.
The absorbance vs. time profiles were analysed by numerical
integration of equation (9) because the total ascorbic acid con-
centration was not always in pseudo-first-order excess over the
total copper(), and free [ClϪ] decreases somewhat due to
complexing with CuI as predicted by Scheme 3. The overall
stoichiometry and increase in [Hϩ] predicted by equation
(1) were assumed. For each run, the initial absorbance was
determined from blank solutions of CuII and ClϪ at the
appropriate concentrations and the variation of this with free
[ClϪ] during a run was taken into account. The deadtime
between mixing and start of observation was treated as an
adjustable parameter for each run, and was in the expected
range of 10 to 15 s.
this was included in the final numerical integration model to fit
the data. Comparison to equation (6) indicates that
k
Ϫ22β2I = 5.1 × 102 Ϫ3
s
Ϫ1, so that kϪ22 = 4.4 × 10Ϫ3 Ϫ1 sϪ1
.
I
There is no detectable contribution from kϪ20 or kϪ21β1 [ClϪ]
over the range of 0.03 to 0.50  ClϪ, and one can estimate an
upper limit of kϪ21β1I < 5 Ϫ2 sϪ1
.
The value of k1 is consistent with the earlier values of k10 and
k11 except at the two highest chloride-ion concentrations, where
the predicted rate is too slow. This indicates the presence of the
[ClϪ]2 term in equation (5) with k12β2II = (6.0 0.5) × 103 Ϫ3
s
Ϫ1, so that k1 is given by equation (11).
k1 = 4.0 ϩ 5.8 × 103 [ClϪ] ϩ 6.0 × 103 [ClϪ]2 (11)
Monitoring of Co(NH3)5(N3)2ϩ disappearance
The principle of this method is that the CuI produced by ascor-
bic acid reduction of CuII will be oxidized back to CuII by the
Co(NH3)5(N3)2ϩ. Then the system can be described by equa-
tions (12) and (13), and the disappearance of Co(NH3)5-
H2asc ϩ 2 CuII ϩ 2nClϪ
During the numerical analysis for each run k1 was assumed to
follow equation (5), and k10 and k11 were fixed at the values
dhae ϩ 2 CuCln1Ϫn ϩ 2 Hϩ (12)
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888
3885

View Article Online
1.10
0.90
1.40
1.20
0.90
0.70
0.70
0.50
0.30
0.10
1.00
0.80
0.60
0.40
0.50
0.30
0.10
0
250 500 750 1000 1250 1500 1750 2000 2250
t/s
0
30 60 90 120 150 180 210 240 270
t/s
Fig. 4 Variation with time of the absorbance of Co(NH3)5(N3)2ϩ dur-
ing reduction of CuII by ascorbic acid for runs with 0.0500  Hϩ in 1.00
 HClO4–LiClO4 at 25 ЊC. The concentrations of CuII, ClϪ, ascorbic
Fig. 5 Variation with time of the absorbance of Co(NH3)5(N3)2ϩ dur-
ing reduction of CuII by ascorbic acid for runs with varying [Hϩ] in 1.00
 HClO4–LiClO4 at 25 ЊC. The concentrations of Hϩ, CuII, ClϪ, ascor-
acid and Co(NH3)5(N3)2ϩ respectively are 0.0501, 2.42 × 10Ϫ3
,
bic acid and Co(NH3)5(N3)2ϩ respectively are 0.050, 5.01 × 10Ϫ2
,
,
,
4.64 × 10Ϫ4, 12.1 × 10Ϫ4 (᭺) offset down by 0.30 absorbance units,
31.6 × 10Ϫ3
,
3.55 × 10Ϫ4
,
8.10 × 10Ϫ4 (᭺), 0.0252, 5.00 × 10Ϫ2
0.0250, 1.604 × 10Ϫ3
, ,
3.48 × 10Ϫ4 8.02 × 10Ϫ4 (ᮀ) and 0.0126,
32.3 × 10Ϫ3, 3.39 × 10Ϫ4, 7.72 × 10Ϫ4 (ᮀ) and 0.0120, 5.00 × 10Ϫ2
31.5 × 10Ϫ3, 3.46 × 10Ϫ4, 7.85 × 10Ϫ4 (᭛)
1.602 × 10Ϫ3, 3.48 × 10Ϫ4, 8.01 × 10Ϫ4 (᭛)
CuCln1Ϫn ϩ Co(NH3)5(N3)2ϩ ϩ 6 Hϩ
CuII ϩ nClϪ ϩ CoII ϩ 5 NH4ϩ ϩ HN3 (13)
1.00
0.80
(N3)2ϩ can be used to monitor the CuII–ascorbic acid reaction.
It was established that Co(NH3)5(N3)2ϩ does not react with
ascorbic acid under our reaction conditions.
This method appears to provide several technical and kinetic
0.60
0.40
0.20
advantages. The electronic spectrum of Co(NH3)5(N3)2ϩ (ε 260

Ϫ1 cmϪ1 at 518 nm) makes its disappearance easy to monitor at
modest concentrations. The oxidation of CuI [equation (13)]
should maintain a low steady-state concentration of CuI so that
the inhibition effect described in the previous section should be
minimized. Substantial Hϩ is consumed in equation (13), and
this places some limit on the reactant concentrations if [Hϩ] is
to remain reasonably constant. The chloride-ion concentration
will stay constant in a run because very little is complexed by
the small amount of CuI. In practice, these advantages proved
to be somewhat illusory for two reasons. The rate of reaction
(13) is substantially inhibited by ClϪ while reaction (12) is cata-
lysed by ClϪ. As a result, the rate of reaction (13) influences the
observations, and even dominates the kinetics for [ClϪ] у 0.04
. At intermediate [ClϪ] the copper() concentration builds to
small but significant levels (10Ϫ4–10Ϫ5 ) and the copper()
inhibition is significant. At low [ClϪ] (<1 × 10Ϫ2 ) the copper()
concentration is small (10Ϫ7–10Ϫ8 ) because reaction (12)
is slower than (13), and no copper() inhibition is observed.
It is doubtful if the complexities of this system could have
been unravelled without the knowledge gained from the direct
monitoring of copper() disappearance.
30 60 90 120 150 180 210 240 270
t/s
Fig. 6 Variation with time of the absorbance of Co(NH3)5(N3)2ϩ dur-
ing reduction of CuII by ascorbic acid for runs with varying [ClϪ] in
0.050  [Hϩ], 5.0 × 10Ϫ2  CuII and 1.00  HClO4–LiClO4 at 25 ЊC. The
concentrations of ClϪ, ascorbic acid and Co(NH3)5(N3)2ϩ are
11.7 × 10Ϫ3, 3.54 × 10Ϫ4, 7.85 × 10Ϫ4 (᭺), 31.6 × 10Ϫ3, 3.55 × 10Ϫ4
,
8.10 × 10Ϫ4 (ᮀ) and 62.2 × 10Ϫ3, 3.51 × 10Ϫ4, 8.02 × 10Ϫ4 (᭛). The data
for the upper (᭺) and lower (ᮀ) curves have been displaced by 0.07
absorbance units, up and down, respectively, for clarity. The three
curves for the 62 × 10Ϫ3  ClϪ data (᭛) use k13 values of 10, 8 and 6 
sϪ1
data illustrate the complication that arises at high [ClϪ] due to
the slowness of reaction (13). The rate clearly increases as [ClϪ]
changes from 11.7 to 31.6 m, but then decreases at 62.2 m.
This apparent anomaly results because reaction (13) is inhibited
by ClϪ and has become entirely rate limiting at 62.2 m ClϪ. In
fact the rate of reaction (13) has an influence on all the curves in
Fig. 6. The run at 62.2 m clearly shows the initial induction
period due to build-up of CuI, and this feature is apparent but
less obvious for other runs in Figs. 5 and 6.
It seems simplest to note at this stage that the rate of reaction
(13) seems to be adequately described by equation (14), with
Rate = k13[Cuϩaq][Co(NH3)5(N3)2ϩ] =
k13[CuI]tot[Co(NH3)5(N3)2ϩ]/DI (14)
Owing to the complexities described above, the absorbance
vs. time profiles have been modelled by numerical integration.
The stoichiometry described by reactions (12) and (13) was
assumed, and the predicted change in [Hϩ] was also taken into
account. The expected ascorbic acid:Co(NH3)5(N3)2ϩ stoichi-
ometry was confirmed by the observed total absorbance
changes. The concentration of ClϪ stays essentially constant
because of the low copper() concentration, but CuClϩ and
CuCl2 formation were included in the analysis.
The numerical analysis was carried out by treating kϪ2 [equa-
tion (9)] and k13 [equation (14)] as fitting variables with k1 given
by equation (11). Of course, it was expected that kϪ2 would be
given by equation (10), and k13 was adjusted accordingly to fit
the absorbance vs. time curves. However, k13 is quite well estab-
k13 = 2.6 × 103 Ϫ1 sϪ1 (1.00  HClO4–LiClO4), in reasonable
agreement with the value of 1.5 × 103 Ϫ1 sϪ1 (0.20  LiClO4)
determined by Parker and Espenson.9 The absence of [ClϪ] or
[ClϪ]2 terms in the rate law indicates that CuCl and CuCl2Ϫ are
not kinetically active reducing agents for Co(NH3)5(N3)2ϩ
.
This is not surprising because of their substantially lower
reducing power compared to Cuϩ(aq).
Representative absorbance vs. time curves are shown in Figs.
4–6. For the runs in Fig. 4 the [ClϪ] is low (1.6 m), and reac-
tion (13) is fast enough so that k1i [Scheme 3 and equation (11)]
is rate controlling. In Fig. 5, the increase in rate with decreasing
[Hϩ] is shown, with other concentrations essentially constant.
The variation of the rate with [ClϪ] is shown in Fig. 6; these
3886
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888

View Article Online
Table 2 Kinetic runs for oxidation of ascorbic acid by CuII in the presence of Co(NH3)5(N3)2ϩ in 1.00  HClO4–LiClO4 at 25 ЊC
[Hϩ]/
103[ClϪ]/
102[Cu2ϩ
]
tot/
104[asc]tot/
104[Co(NH3)5(N3)2ϩ]/
k
Ϫ2/Ϫ1 sϪ1
10Ϫ2k13/Ϫ1s Ϫ 1
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0500
0.0252
0.0252
0.0252
0.0120
1.602
1.604
1.606
1.606
2.42
6.50
11.7
11.7
20.7
31.6
61.9
62.2
62.2
32.3
31.6
31.1
31.5
1.26
2.50
5.00
5.01
5.00
5.01
5.01
5.00
5.00
5.01
5.01
5.01
5.01
5.00
10.4
10.4
5.00
3.48
3.48
3.49
2.46
4.64
3.55
3.47
3.54
3.49
3.55
3.52
3.47
3.51
3.39
3.49
2.18
3.46
8.01
8.02
8.03
8.03
12.1
8.02
7.99
7.85
7.88
8.10
13.9
8.02
8.02
7.72
7.95
4.99
7.85
(13)*
(13)*
(13)*
(13)*
(9.9)*
(4.0)*
2.0
2.2
0.55
0.35
0.080
0.080
0.080
0.30
0.38
0.30
0.045
0.048
0.15
0.32
1.44
1.40
1.42
0.33
0.32
0.35
0.32
0.35
* Calculated from equation (14), k1 is rate controlling under these conditions.
lished independently for runs with [ClϪ] у 0.02  for reasons
described above. For [ClϪ] р 0.01  as in Fig. 4 kϪ2 has no
influence and the rate of reduction of Co(NH3)5(N3)2ϩ [equa-
tion (13)] is fast enough so that the rate is controlled by k1
[equation (9)].
The values of kϪ2 and k13 for different runs are given in Table
2, and representative fits for the various conditions are shown in
Figs. 4–6. The values and [ClϪ] dependence of kϪ2 are consistent
with the results from the direct observation of copper() disap-
pearance (Table 1).
of K00 and βs can be used to calculate the desired ratios. The
results are given by equations (16) and (17). From these ratios
I
I
k22
(kϪ22β2 )β1
(16)
(17)
= K00
II = 0.011
kϪ11
(k11β1II)β2
I
I
k22
(kϪ22β2 )β2
= K00
II = 2.3
kϪ12
(k12β2II)β2
and the above estimates of kϪ11 and kϪ12 one can calculate that
k22 is 2 × 108 or 2.3 × 108 Ϫ1 sϪ1 and these values are self-
consistent and reasonable for reaction with a radical as dis-
Discussion
The results of this study are consistent with the earlier observ-
ations of Xu and Jordan1 on the rate of oxidation of ascorbic
acid by CuII at modest chloride-ion and copper() concentr-
ations. It has been established here that the reaction is inhibited
by copper() species, and details of the chloride-ion catalysis and
reactivity patterns are elucidated by studies at higher [ClϪ]. All
the observations can be accounted for by the reaction sequence
in Scheme 3. The large effect of ClϪ on the rate results from the
strong complexation of CuI by ClϪ and weak complexation by
CuII. This gives the order of driving force for the oxidant as
CuIICl2 ӷ CuIIClϩ ӷ Cu2ϩ and this is reflected in the reactivity
as expected from Marcus theory for electron-transfer reactions.
The kinetically determined values of k11β1II and k12β2II [equa-
tion (5)] can be used to calculate that k11 = 2.5 × 103 and
cussed for kϪ11 and kϪ12
The values of
[CuCl2][asc Ϫ])([dhae]/[dhak]) = 5.3 × 103 × 6 × 106 can be
.
k22
and
K22 = ([CuCl2Ϫ][dha]/
Ϫ1. This can be
᝽
combined to calculate kϪ22 = 0.6 × 10Ϫ3 Ϫ1
s
compared to kϪ22 = 4.4 × 10Ϫ3 Ϫ1 sϪ1 from the experimental
value of kϪ22β2I = 5.1 × 102 Ϫ3 Ϫ1. The factor of 7 difference
s
between these values can be ascribed to uncertainties in the
parameters used to calculate K22.
The simplified Marcus relationship is given by equation (18)
¹
²
krxn = (kI/IIkr/pKeq)
(18)
where kI/II and kr/p are the self-exchange rate constants for the
appropriate CuI/CuII and radical/parent reactions, respectively,
and Keq is the equilibrium constant for the reaction. This equa-
tion predicts that the ratio of k11 :k12 can be used to estimate the
ratio of kI/II(CuII/ICl) to kI/II(CuII/ICl2) from the reduction poten-
tials of the appropriate CuII–CuI half-reactions because the kr/p
for ascorbate cancels when the ratio is taken. The calculation
gives kI/II(CuII/ICl):kI/II(CuII/ICl2) = 1.4 × 102 :1. The smaller
value for kI/II(CuII/ICl2) could be due to a larger inner-sphere
k12 = 4.0 × 103 Ϫ1
s
Ϫ1. The appropriate reduction potentials
ؒ
and the Ka of Hasc can be used to calculate the respective
equilibrium
constants
of
K11 = 5 × 10Ϫ8 × 2.5
and
K12 = 1.8 × 10Ϫ5 × 2.5. The forward rate constants and equi-
librium constants can be used to calculate the reverse rate con-
stants of kϪ11 = 2 × 1010 and kϪ12 = 1 × 108 Ϫ1 Ϫ1. It is not
s
unexpected that the reverse reactions should be close to diffu-
sion controlled, given the large driving force for the reverse
reaction and that one of the reactants is a radical. Although k-11
seems about 5 times larger than expected for diffusion control,
this could be due to uncertainty in the Ka or to a change of
≈0.03 V in the EЊ used to calculate K11.
For the purpose of the numerical analysis, it was convenient
to express equation (9) with kϪ2 as a variable, but the ratios
k22 :kϪ11 and k22 :kϪ12 are useful in providing estimates of k22.
From the general expression given by equation (15) it can be
Ϫ
reorganization energy between the linear CuICl2 and tet-
ragonal CuCl2(OH2)4.
It is interesting that our studies with Co(NH3)5(N3)2ϩ have a
formal analogy to the much studied autooxidation of ascorbic
acid by CuII in which O2 oxidizes CuI back to CuII. The substan-
tial kinetic disagreements between some of the autooxidation
studies have been described previously.1 Without labouring the
point, it is clear from equation (9) that the apparent order of the
reaction can depend on the relative reagent concentrations and
whether initial rates or the full reaction were used to determine
the rate constant.
k1ik2jβiIIβkII/kϪ1ikϪ2jβiIβkI = K00
seen that the kinetically determined values of k11β1II and k12β2
(15)
In general, the kinetic complexity of this system results
because the initial electron transfer to produce the ascorbate
radical is highly unfavourable and therefore subject to inhib-
II
[equation (5)] and kϪ22β2I [equation (10)] and the known values
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888
3887

View Article Online
3 G. P. Laroff, R. W. Fessenden and R. H. Schuler, J. Am. Chem. Soc.,
ition by the copper() product. Similar effects can be expected
for many organic reductants that give unstable radical inter-
mediates. The chloride-ion catalysis results in large part from
the strong complexation of CuI by chloride ion which increases
Keq in equation (18) and thereby increases the rate constant. It
appears that any species that complexes with CuI much more
strongly than with CuII should have an analogous catalytic
effect, as long as the self-exchange rate kI/II is not proportion-
ately decreased.
1972, 94, 9062; M. Schöneshöfer, Z. Naturforsch., Teil B, 1972, 27,
649; R. W. Fessenden and N. C. Verma, Biophys. J., 1978, 24, 93;
B. H. J. Bielski, A. O. Allen and H.A. Schwartz, J. Am. Chem. Soc.,
1981, 103, 3516.
4 J. N. Gex, C. Daul and A. von Zelewsky, Chem. Phys. Lett., 1986,
132, 276.
5 H. Sapper, A. Pleyer-Weber and W. Lohman, Z. Naturforsch., Teil
C, 1982, 37, 129.
6 Y. Kirino and T. Kwan, Chem. Pharm. Bull., 1972, 20, 2651.
7 J. Chatlas and R. B. Jordan, unpublished work.
8 M. J. Sisley and R. B. Jordan, Inorg. Chem., 1992, 31, 2137.
9 O. J. Parker and J. H. Espenson, J. Am. Chem. Soc., 1969, 91, 1968.
10 R. W. Ramette, Inorg. Chem., 1986, 25, 2481.
Acknowledgements
11 S. Arhland and J. Rawthorne, Acta Chem. Scand., 1970, 24, 157.
12 J. J. Fritz, J. Phys. Chem., 1984, 88, 4358.
13 V. K. Sharma and F. J. Millero, Inorg. Chem., 1988, 27, 3257.
The authors thank the Natural Sciences and Research Council
of Canada for financial support for this work.
References
1 J. Xu and R. B. Jordan, Inorg. Chem., 1990, 29, 2933.
2 Y. I. TurЈyan and R. Kohen, J. Electroanal. Chem. Interfacial
Electrochem., 1995, 380, 273 and refs. therein.
Received 6th March 1997; Paper 701574G
3888
J. Chem. Soc., Dalton Trans., 1997, Pages 3883–3888