The oxidation of bromide ion by [(FeTPP)2O]+SbF6- In dichloromethanej

Article abstract of DOI:10.1039/b003827j

The oxidation of cetyltrimethylammonium bromide (CTAB) by mono-oxidised iron(m) tetraphenylporphyrin u-oxo-dimer, [(FeTPPOSbF, ; , has been found to be much slower than the corresponding oxidation of iodide ion, but can still be studied by stopped-flow techniques. The stoichiometry of the reaction has been established as 2 [Fe(TPP)2O]+ + 3Br -> 2(FeTPP)2O + Br3-. The rate law for the reaction is d [(FeTPP)2O]/df = [Fe(TPP)2O+] (k_f [CTAB] -A'r). The rate constant for the first term has been identified with the rate determining forward process, the formation of the unstable Br2-, and the second term with the attack of Br2- on the uncharged u-oxo-dimer. A subsequent process involves what appears to be the rapid oxidation of Br2- in the presence of Br- to give Br3 . At 298 K k_f is 737(±30) NT1 s^-1 and k, is 0.80(±0.16) s^-1. For k_f, Δ = 58.8(±1.6) kJ mol^-1, AS = 8.62(10.45) J K^-1 moP1; for k, , ΔH = 43.6(110.8) kJ mol^-1, AS = -97.8(136.0) J K^-1 mol^-1. The addition of cetyltrimethylammonium perchlorate (inert to reaction with the oxidised u-oxo-dimer) slowed the reaction in a manner which indicated competition between perchlorate and bromide for the formation of ion pairs with the oxidised iron(m) dimer, the bromide ion pair being the reactive one. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b003827j

View Article Online / Journal Homepage / Table of Contents for this issue
؊
The oxidation of bromide ion by [(FeTPP)2O]؉SbF6 in
dichloromethane‡
Colin D. Hubbard,*a John G. Jones*†b and John McKnightc
a Institute for Inorganic Chemistry, Friedrich-Alexander University of Erlangen-Nürnberg,
Egerlandstrasse 1, 91058 Erlangen, Germany
b ABCS, University of Ulster, Coleraine, Northern Ireland, UK BT52 1SA.
E-mail: JG.Jones@ulst.ac.uk
c SKB, Worthing, Sussex, UK
Received 15th May 2000, Accepted 14th July 2000
Published on the Web 31st August 2000
The oxidation of cetyltrimethylammonium bromide (CTAB) by mono-oxidised iron() tetraphenylporphyrin
Ϫ
µ-oxo-dimer, [(FeTPP)2O]ϩSbF6 , has been found to be much slower than the corresponding oxidation of iodide
ion, but can still be studied by stopped-flow techniques. The stoichiometry of the reaction has been established as
Ϫ
2 [Fe(TPP)2O]ϩ ϩ 3BrϪ → 2(FeTPP)2O ϩ Br3 . The rate law for the reaction is d [(FeTPP)2O]/dt = [Fe(TPP)2Oϩ]·
(kf [CTAB] Ϫ kr). The rate constant for the first term has been identified with the rate determining forward process,
the formation of the unstable Br2 , and the second term with the attack of Br2Ϫ on the uncharged µ-oxo-dimer.
Ϫ
Ϫ
A subsequent process involves what appears to be the rapid oxidation of Br2Ϫ in the presence of BrϪ to give Br3 .
At 298 K kf is 737( 30) MϪ1 sϪ1 and kr is 0.80( 0.16) sϪ1. For kf, ∆H‡ = 58.8( 1.6) kJ molϪ1, ∆S‡ = 8.62( 0.45)
J KϪ1 molϪ1; for kr, ∆H‡ = 43.6( 10.8) kJ molϪ1, ∆S‡ = Ϫ97.8( 36.0) J KϪ1 molϪ1. The addition of cetyltrimethyl-
ammonium perchlorate (inert to reaction with the oxidised µ-oxo-dimer) slowed the reaction in a manner which
indicated competition between perchlorate and bromide for the formation of ion pairs with the oxidised iron()
dimer, the bromide ion pair being the reactive one.
close to iron() typical values rather than the values for
Introduction
FeIV (where δ typically is equal to zero4c). However, the ESR
Models for and studies on oxidative enzymes which contain
spectrum shows but one narrow peak at g = 2 with no fine
structure, and no peak at g = 6 (characteristic of high spin
iron and porphyrin combined together in the prosthetic group
have been legion since the pioneering studies of George2 assign-
iron() in a porphyrin environment); the overall magnetic
ing formal oxidation states to compounds I and II, intermedi-
moment for the singly oxidised dimer is 3.2 0.2 µB at 270 K,6b
ates formed from the reaction of horseradish peroxidase (HRP)
corresponding to two unpaired electrons on a spin-only basis.
with hydrogen peroxide. The prosthetic group of HRP com-
From this it has been inferred that, as in uncharged D, there is
pound I has since been identified and modelled as a ferryl unit,
considerable antiferromagnetic coupling between the two iron
centres; NMR results indicate that electron transfer between
O᎐FeIV, co-ordinated to a protoporphyrin IX cation radical,3
᎐
and compound II as a ferryl unit co-ordinated to protopor-
the porphyrin round one iron centre within the molecule
phyrin IX (3,7,12,17-tetramethyl-8,13-divinylporphyrin-2,18-
and the radical cation round the other is extremely rapid.6b
dipropanoic acid). Model compounds have been made and
Coupling between unpaired spin on the Fe–O–Fe skeleton and
studied involving porphyrins and ferryl units,4 porphyrins and
on porphyrin is harder to assess. We shall refer to Dϩ without
iron()5 and porphyrin radical cations with iron()6 to name
specifying where the positive charge lies.
but a few. The study of oxidised transition metal porphyrins
In our previous investigation1 the simple rate law (1) applied.
both structurally and as catalysts is still expanding.7
d [(FeTPP)2O]/dt = k[Fe(TPP)2Oϩ][CTAI]
(1)
Owing to our experience in the preparation of one class
of oxidised porphyrin complex8,9 we decided to commence a
kinetic comparison of oxidised model compounds with this
class. In a previous paper1 we described the kinetics of
reduction of iodide (cetyltrimethylammonium iodide, CTAI;
cetyl = hexadecyl) by the singly oxidised form of iron() tetra-
We argued that in dichloromethane and with iodide present
in excess, [Fe(TPP)2O]ϩ (Dϩ) was likely to be present almost
exclusively as the ion pair, DϩIϪ, ion pair saturation being the
prevailing circumstance. Attack on this ion pair of free iodide
was postulated as the rate-determining step, the precursor
phenylporphyrin µ-oxo-dimer, [FeTPP–O–FeTPP]ϩ SbF6
Ϫ
(designated as Dϩ) which appeared stable in the solid state
and reasonably long-lived in dry dichloromethane solution.
Formally Dϩ contains iron in both oxidation states  and ,
but further work6b has inclined investigators to the belief that
the compound has two iron() centres, one of which is co-
ordinated to a porphyrin radical cation. For the singly oxidised
dimer the Mössbauer 57Fe chemical shift, δ, is 0.295 mm sϪ1 at
300 K and the isotopic splitting, ∆EQ, is 0.447 mm sϪ1,6a very
Ϫ
complex being DϩIϪؒIϪ and the successor complex DؒI2 with
Ϫ
electron transfer being inner sphere. The radical ion I2 could
form I3Ϫ either by rapid disproportionation or by further attack
of oxidised dimer ion pair on it, eqn. (2).
Ϫ
DϩIϪ ϩ I2Ϫ → D ϩ I3
(2)
In this paper we describe the oxidation of bromide ion also
in dichloromethane, by the same oxidised iron() porphyrin
dimer, in the presence of the same cetyltrimethylammonium
counter ion. The reaction is considerably slower than the oxid-
ation of iodide, and we have been able to confirm both the
† Current address: Unilever Research, Colworth House, Sharnbrook,
Beds., UK MK44 1LQ. E-mail: Colin.Hubbard@unilever.com
‡ Oxidations by [(FeTPP)2O]ϩSbF6Ϫ. Part II.1
DOI: 10.1039/b003827j
J. Chem. Soc., Dalton Trans., 2000, 3143–3148
This journal is © The Royal Society of Chemistry 2000
3143

View Article Online
stoichiometry and that ion pairing between bromide and posi-
tively charged dimer prior to reaction is crucial.
Experimental
Materials
BDH “Analar” dichloromethane was used after being refluxed
for thirty minutes over calcium hydride and distilled into a pre-
dried flask. Cetyltrimethylammonium bromide (CTAB or QBr)
from Aldrich was recrystallised from dry dichloromethane and
stored in vacuo over phosphorus pentoxide. The perchlorate
(QClO4) was made by adding a saturated solution of sodium
perchlorate in 2:1 ethanol–water to a similarly saturated solu-
tion of QBr in the same solvent. The white precipitate was
collected, recrystallised from dichloromethane–acetone (1:1)
solution and then dried and stored over phosphorus pentoxide
at room temperature. CAUTION: no tendency to explode was
noted, though perchlorate salts should always be handled with
extreme caution. The tribromide (QBr3) was prepared by taking
3.19 g (0.020 mol) of bromine in a sample bottle, dissolving it in
the minimum amount of dry dichloromethane, adding 7.28 g
(0.020 mol) of CTAB dissolved in the minimum of dry dichloro-
methane, and then shaking. The orange-yellow product was
precipitated by adding dry pentane, collected, washed with
pentane, and dried under vacuum.
Fig. 1 Spectra of a [(FeTPP)2O]ϩSbF6Ϫ in dichloromethane at 298 K,
b “a” after reaction with excess of QBr. Main spectrum: absorbance
from 0 to 1.4. Insert spectrum: absorbance from 0 to 0.14.
1.06 × 10Ϫ4 M) of [(FeTPP)2O]ϩSbF6Ϫ solution was diluted 1:1
with dry dichloromethane in the sample cell, the spectrum
Ϫ
was recorded (the concentration of [(FeTPP)2O]ϩSbF6 was
effectively 5.03 × 10Ϫ5 M and the spectrum could be compared
directly with the spectra of each of the reaction sample
mixtures). The spectra of a series of CTAB solutions (e.g.
1.50 × 10Ϫ5 to 4.77 × 10Ϫ4 M) were recorded in which equal
The preparation of peroxolauric acid (CH3(CH2)10CO3H) and
the hexafluoroantimonate of iron() tetraphenylporphyrin10
have been described, as has the production of the oxidised
volumes of 1.06 × 10Ϫ4 M [(FeTPP)2O]ϩSbF6 and bromide
Ϫ
Ϫ
µ-oxo-dimer [(FeTPP)2O]ϩSbF6 from these reactants.1,8,9
solutions of known concentration were mixed. The wavelength
chosen to monitor the reaction was 610 nm, at which sub-
stantial optical density changes were observed with consistency
(Fig. 1).
Experience showed that it was best to prepare [(FeTPP)2O]ϩ-
Ϫ
SbF6 freshly prior to use employing the following slightly
modified procedure. In a typical preparation 0.020 g (3.5 ×
10Ϫ5 mol) of 59% peroxouric acid, dissolved in the mini-
mum amount of dry benzene, was added to 0.0912 g (1.0 ×
Kinetic experiments
Ϫ
10Ϫ4 mol) of FeTPPϩSbF6 dissolved in the minimum amount
Reaction kinetics was studied by the stopped-flow method and
utilised a Hi Tech SF-3L unit suspended over a Gallenkamp
refrigerated thermostat bath filled with 70:30 water–crude
methanol: storage coils and the mixing/observation chamber
were completely immersed in the bath. The bath capacity was
ten litres and its insulation enabled it to retain a virtually con-
stant temperature for the ten minute period of a series of
stopped-flow runs using a particular concentration set. (This
reduces all noise due to refrigeration or heating thermostat
action.) A Bausch and Lomb grating monochromator with a
tungsten–halogen lamp was used in conjunction with a Farnell
stabilised power supply, and detection was via an IP28A
photomultiplier and a Brookdeal or Farnell high-voltage power
supply. The voltage developed to earth across a chosen high
impedance was monitored as described in a recent paper;11 the
voltage/time curve was converted into an absorbance/time
curve and analysed for first-order build up or decay using a
BBC computer, as also described in that paper. The wavelength
chosen was 410 nm.
of dried dichloromethane. The solution changed from brown
to green-black. The addition of light petroleum (bp 40–60 ЊC)
produced a fine blue-black precipitate with a yellow super-
natant fluid. The product was isolated by centrifugation (4500
r.p.m. for approximately 15 minutes) and washed with benzene
and light petroleum until the washings were colourless. The
product was dried at 353 K for several hours: the UV-visible
spectrum in dry dichloromethane was identical to that pub-
lished for the monocation, [(FeTPP)2O]ϩSbF6Ϫ, by Baldwin and
co-workers8 and to that for [(FeTPP)2O]ϩClO4Ϫ 10 The infrared
.
spectrum showed a strong band at 1275 cmϪ1, a characteristic
of the oxidised porphyrin ring10 and bands at 650 and 280
Ϫ
cmϪ1, characteristic of the SbF6 ion (seen previously in the
Ϫ
FeTPPϩSbF6 spectrum). As well as the usual bands due to
the tetraphenylporphyrin unit there was a band at 865 cmϪ1
characteristic of the Fe–O–Fe system. The field desorption
mass spectrum displayed a parent peak at m/z 1587 ([DSbF6]ϩ)
and one for Dϩ at m/z = 1352. In these ways the product was
Ϫ
.
identified as [(FeTPP)2O]ϩSbF6
All apparatus was cleaned and dried in an oven at 373 K
before use.
Results and discussion
Stoichiometry experiments
i Stoichiometry
Ϫ
Making up the dilute solutions employed here involved the
use of an “Oertling” microbalance and small aluminium foil
weighing boats. The mass of a boat was determined and the
appropriate mass of reactant added. The reactant was then
added to the appropriate volumetric flask and the boat and
residual reactant weighed again.
Spectra were obtained using a Hewlett Packard 8451A diode-
array spectrophotometer. In a typical run a reference spectrum
of the solvent, dry dichloromethane, between 340 and 760 nm
was recorded as a baseline, prior to measuring the spectra
of the reaction mixtures. Then a known concentration (e.g.
The stoichiometry of the reaction between [(FeTPP)2O]ϩSbF6
and CTAB was determined spectrophotometrically by the
method of continuous varations.12 The principle of the pro-
cedure is that when a solution of known concentration in
[(FeTPP)2O]ϩSbF6 is treated with a series of solutions of
Ϫ
CTAB of equal volume containing bromide at a series of
known and increasing concentrations the concentration
of [(FeTPP)2O]ϩSbF6 remaining in the reaction mixture
Ϫ
will decrease and the proportion of µ-oxo-dimer product will
increase. The concentration of the product, initially zero, will
not continue to increase indefinitely as the bromide concentra-
3144
J. Chem. Soc., Dalton Trans., 2000, 3143–3148

View Article Online
Table 1 Absorbance data for the determination of the overall stoichio-
Table 2 Rate constants at four temperatures and activation
Ϫ
Ϫ
metry of the reaction of [(FeTPP)2O]ϩSbF6 with cetyltrimethyl-
parameters for the reduction of [(FeTPP)2O]ϩSbF6 by CTAB in
ammonium bromide, in dichloromethane at 25 ЊC
dichloromethane
Absorbance
T/K
kf /MϪ1 sϪ1
Activation parameters, 298 K
105[QBr]/M
460
570
610 nm
273
289
298
303
92( 16)
389( 16)
737( 30)
1400( 78)
Ea = 61.3( 3.2) kJ molϪ1
A = 4.8( 0.2) × 1013
0
2.02
1.91
1.83
1.62
1.50
1.60
1.62
0.802
0.762
0.732
0.637
0.569
0.646
0.662
0.658
0.626
0.600
0.521
0.405
0.349
0.354
∆S‡ = 8.62( 0.45) J KϪ1 molϪ1
∆H‡ = 58.8( 1.6) kJ molϪ1
∆G‡ = 53.9( 2.2) kJ molϪ1
0.74
1.49
2.98
5.96
11.9
23.9
Table 3 Rate constants and activation parameters for the reverse of
the reduction of [(FeTPP)2O]ϩSbF6Ϫ by CTAB in dichloromethane
T/K
kr /sϪ1
Activation parameters, 298 K
273
289
298
303
0.23( 0.10)
0.51( 0.09)
0.80( 0.16)
2.08( 0.20)
Ea = 45.9( 10.7) kJ molϪ1
A = 1.3( 0.3) × 108
∆S‡ = Ϫ98( 36) J KϪ1 molϪ1
∆H‡ = 43.6( 10.8) kJ molϪ1
∆G‡ = 70.2( 7.0) kJ molϪ1
Fig. 2 Plot of absorbance at 610 nm vs. [BrϪ] for the spectrophoto-
metric titration of QBr (CTAB) and [(FeTPP)2O]ϩSbF6 in dichloro-
Ϫ
methane at 298 K.
tion is increased because at a specific bromide ion concentration
Ϫ
the reduction of [(FeTPP)2O]ϩSbF6 to the µ-oxo-dimer will
Ϫ
be complete, i.e. the concentration of [(FeTPP)2O]ϩSbF6 will
be zero and that of the µ-oxo-dimer will be equal to the initial
concentration of [(FeTPP)2O]ϩSbF6Ϫ. If this concentration of
bromide ion is exceeded there will be no additional increase
in the concentration of the µ-oxo-dimer. The stoichiometry of
the reaction will be discovered from the relationship between
the CTAB concentration which represents the complete con-
version of the reactants into products, and the initial concen-
tration of [(FeTPP)2O]ϩSbF6Ϫ. In practical terms the bromide
concentration at which the optical density change arrived at a
limiting value was used to determine the reaction stoichiometry.
In a typical series of experiments (the results are listed
Fig. 3 Rate profiles for the reaction of CTAB with [(FeTPP)2O]ϩ-
SbF6Ϫ in dichloromethane.
progress of reduction of Dϩ; a substantial increase in absorb-
ance occurs here. Kinetic studies were performed at 273, 289,
298 and 303 K; at each temperature a series of CTAB con-
centrations from 10Ϫ4 to 10Ϫ2 M were employed, the initial
concentration of Dϩ being around 10Ϫ5 M for each run. Under
these conditions of excess of bromide the change in optical
density at 410 nm followed a first-order course, from which
eqn. (4) can be deduced. At all temperatures the observed rate
Ϫ
in Table 1) the absorbance values of [(FeTPP)2O]ϩSbF6 in
the presence of CTAB decreased at 610 nm with increasing
bromide ion concentration until a limiting absorbance value
of 0.35 was reached, as illustrated in Fig. 2. The measured
absorbance values of the reaction samples vs. the bromide con-
centration (Fig. 2) were extrapolated to determine the concen-
tration of the bromide ion (7.5 × 10Ϫ5 M) at which the limiting
d[D]/dt = kobs [Dϩ]
(4)
constant was proportional to the sum of a constant term and
a term proportional to the concentration of added CTAB, with
no evidence of a second-order term in bromide concentration
(Fig. 3). Thus eqn. (5) is obtained.13 We have chosen to identify
absorbance was achieved. The ratio of [(FeTPP)2O]ϩSbF6
Ϫ
to BrϪ was found to be 5.03 × 10Ϫ5 M:7.5 × 10Ϫ5 M which
simplifies to 1:1.5. Eqn. (3) is consistent with this ratio. It
kobs = k[CTAB] ϩ kЈ
(5)
Ϫ
2[Fe(TPP)2O]ϩ ϩ 3BrϪ → 2(FeTPP)2O ϩ Br3
(3)
the slope of the plot of kobs vs. [CTAB] as the forward rate
constant k, at 298 K equal to 738( 30) MϪ1 Ϫ1, and the con-
s
stant kЈ as that for the reverse electron transfer, DؒBr2Ϫ →
DϩBrϪؒBrϪ. At 298 K this is equal to 0.8( 0.2) sϪ1. An altern-
ative identification of the kЈ term with a reduction process
not dependent on bromide concentration is deemed unlikely,
as Dϩ is stable in pure dry dichloromethane, and also in the
presence of a salt such as QϩClO4Ϫ. So k = kf and kЈ = kr. The
results are summarised in Tables 2 and 3.
should be noted that this is of the same form as that proposed
in our previous paper,1 but there without confirmation.
ii Kinetics and mechanism
Preliminary experiments showed that in the presence of an
excess of bromide the visible product was the µ-oxo-dimer,
displaying characteristic absorbance maxima at 410, 570 and
610 nm. The wavelength 410 nm was chosen for monitoring the
Several approaches were tried to track down the most likely
mechanism for the reaction. They are summarised in turn
J. Chem. Soc., Dalton Trans., 2000, 3143–3148
3145

View Article Online
Fig. 4 Plot of the concentration of added CTAB vs. the calculated
concentration of the quadrupole {[(FeTPP)2O]ϩBrϪ,QϩBrϪ}.
Fig. 6 Plot of kobs, at 298 K, vs. the product of the ion pair concen-
tration and the total bromide concentration {[BrϪ]total[(FeTPP)2Oϩؒ
BrϪ]}: (᭹) in the absence of perchlorate, (᭺) in the presence of
0.0050 M perchlorate and (ϩ) in the presence of varying perchlorate
concentration and a constant bromide concentration of 0.0050 M.
QϩClO4Ϫ. It should be noted that (iii) is in agreement with the
stoichiometric findings: the reaction sequence is proposed to be
as in Scheme 1.
KIP
a
D
ϩ ϩ BrϪ
{DϩBrϪ}
(very fast)
k1
b {DϩBrϪ} ϩ BrT
Ϫ
kϪ1
ket
kϪet
k2
Ϫ
({DϩBrؒϪBrϪ}
{DؒBr2Ϫ})
D ϩ Br2
kϪ2
(BrTϪ is the sum of the free and ion-paired bromide ion)
Ϫ
Ϫ
c
2Br2
BrϪ ϩ Br3
Fig. 5 Plot of the observed rate constant for the reaction of [(Fe-
TPP)2O]ϩSbF6Ϫ with CTAB vs. ion-triplet concentration for [BrϪ] up to
0.010 M.
(very fast and equilibrium far to the right)
Scheme 1
below, the first sentence being the key assumption of each
paragraph.
It is postulated that the rate-determining steps here concern
electron transfer. Step a involves the formation of an ion
pair. To calculate this one needs to calculate the amount of
free bromide provided by CTAB and then how much Dϩ has
been converted into {DϩBrϪ}. These calculations indicate that
though some 80% of the CTAB is present as ion pairs at
the lowest concentration, 2.5 × 10Ϫ4 M CTAB, 80% of Dϩ is
present as ion pairs, and when [CTAB] is 1 × 10Ϫ2 M, 97% of
Dϩ is present as ion pairs. (The concentration of Dϩ is assumed
to be 1 × 10Ϫ5 M.) So these calculations say that {DϩBrϪ} form-
ation can be considered complete before the reaction starts.
(i) Electron transfer is preceded by prior formation of a
quadrupole {DϩBrϪ,QϩBrϪ}, and the rate of reaction is that
of rate determining electron transfer within the quadrupole to
give D and the transient QϩBr2Ϫ. As a first step the calculation
of the ion-pair concentration for QϩBrϪ was performed using a
value of KIP of 105 MϪ1 (from Beard and Plesch14) for tetra-
methylammonium perchlorate in dichloromethane, relative
permittivity ε = 9. Then the concentration of free bromide
ion was calculated, and from this one was able to calculate the
initial concentration of the ion pair, DϩBrϪ. This concentration
was then combined with the concentration of free bromide and
the ion-triplet constant,15 102 MϪ2, to give the concentration
of the triplet [{BrϪDϩBrϪ}]. This triplet concentration was
combined with the concentration of free Qϩ, using K = 105 MϪ1
to give the quadrupole concentration, treating the triplet as
if it were just another negative ion. The quadrupole concen-
tration was seen to saturate at total added bromide concentra-
tions above 10Ϫ3 M, which does not reflect the observed rate
behaviour (Fig. 4).
Step b concerns the formation of a precursor complex,
Ϫ
}
{DϩBrؒϪBrϪ}, its transition to the successor complex {DؒBr2
and the fission of this to D and Br2Ϫ. As far as the forward
reaction is concerned we shall assume that the precursor com-
plex is a transient intermediate and that the rate determining
step is electron transfer described by ket. The rationale for
this is that, as described later, the differences in the heats of
activation for bromide and iodide reduction are mirrored by the
differences in electron affinities of Br2 and I2. It can be shown
that if kϪ1 ӷ ket, kf = ketk1/kϪ1; if one says that the precursor
complex is ion triplet-like, then as k1/kϪ1 is like the association
constant for ion triplet formation from an ion pair and a free
ion and as this is approximately 1 × 102 MϪ1 one can calculate
the probable value of ket as around 7.37 0.30 sϪ1 at 298 K.
With respect to the reverse reaction the arguments that trans-
fer involves the production of transient Br2Ϫ, a radical anion,
rather than the bromine atom, are very close to those already
advanced in our first paper. Thus, in water Malone and
(ii) Rapid reversible formation of ion triplets, {BrϪDϩBrϪ},
is the essential step before rate determining electron transfer to
give D and the transitory species Br2Ϫ. This approach seemed
promising, but a plot of rate constant vs. triplet concentration
was not of the observed form, so it too was rejected (Fig. 5).
(iii) The rate determining step is attack of bromide in what-
ever form on the ion pair DϩBrϪ to give D and the transient
Br2Ϫ. When the ion-pair concentrations were calculated nearly
all of the oxidised iron() dimer species was seen to be present
as the ion pair. The assumptions used here give rise to agree-
ment with the observed behaviour (Fig. 6, calculations as out-
lined later). It will be seen later that they also account for
the observed kinetics in the presence of the inert electrolyte
Endicott16 calculated that the reduction potential of Br Ϫ–BrϪ
᝽
Ϫ
was 2.2 V and of Br2 –BrϪ 1.9 V; Woodruff and Margerum17
came to a similar conclusion, but reported a slightly larger dif-
ference in potentials. If the same holds true in dichloromethane
one would expect the production of the radical anion to be
3146
J. Chem. Soc., Dalton Trans., 2000, 3143–3148

View Article Online
favoured; a simple way of looking at it is to say that part of the
energy needed to take an electron from the bromide ion is com-
pensated for by the formation of a weak bond with another
bromide to give Br2Ϫ. It is notable that Br2Ϫ is a stronger oxidis-
Ϫ
ing agent than I2Ϫ; the I2 –IϪ potential has been given as around
1.0 V,25 insufficient to be able to observe a back reaction, and
indeed none was noted by us. It should also be noted that we
have made attempts with bromine in dichloromethane and with
QϩBr3Ϫ in dichloromethane to generate the oxidised dimer from
the µ-oxo-dimer. Both attempts failed, in each case the product
having the spectrum of FeTPPؒBr. So it is reasonable to suppose
that the simple back reaction can only be brought about by a
species more oxidising than Br2 or Br3Ϫ. This species is most
probably the bromine molecule radical anion.18 As well as being
able to oxidise the µ-oxo-dimer, two of these radical anions
can react extremely rapidly19 to give BrϪ and Br3Ϫ. In aqueous
solution this has been reported to occur at near the collision
frequency limit.19 Even taking into account that repulsion
between similarly charged ions will be greater in the low relative
permittivity solvent dichloromethane, a rate constant of at least
106 MϪ1 sϪ1 still looks plausible.§
Fig. 7 Plot of kobs, at 298 K, vs. varying perchlorate concentration,
total bromide concentration being kept constant at 0.0050 M.
It should also be noted that this mechanism has close similar-
ities to that proposed by Fudge and Sykes20 for the reaction
of iron() ions with iodide in aqueous solution, and by Nord
et al.21 in reactions where iodide or thiocyanate is oxidised by a
series of metal() complexes. Dihalogen radical anions have
featured as suggested intermediates in one electron reductions
by halides in many cases, especially since their discovery as
reactive transient species in Grossweiner’s flash photolysis
experiments.18
Referring to Scheme 1 once more, by analogy with the
forward reaction the successor complex should lie in a shallow
energy minimum, making it likely that k2 ӷ kϪet and so giving
kr = kϪetkϪ2/k2. As the constant for association between D and
Br2Ϫ is likely to be lower than an ion-triplet formation constant,
kr should be much closer to kϪet than kf is to ket.
Fig. 8 Plot of kobs, at 298 K, vs. varying bromide concentration:
(a) in the absence of added perchlorate, (b) in the presence of 0.0050 M
perchlorate.
There are interesting differences between the activation
parameters here and in our previous paper. The forward rate
constant at 298 K is over one hundred times less than with
iodide as reducing agent, and the enthalpy of activation is
five times greater. The actual difference in heats of activation
is some 48 kJ molϪ1; a large part of that can be ascribed to
the difference in electron affinities of Br2 and I2, some 34 kJ
iii Ion pairing
An important series of findings concerns the effect of added
inert electrolyte QϩClO4Ϫ on the rate of reaction. For a reaction
series with a constant concentration of bromide and a varying
concentration of perchlorate the observed rate constant
declines in a non-linear fashion as the perchlorate concen-
tration is increased (Fig. 7). For a series where there is a
set amount of added perchlorate but a variable amount of
bromide, a linear plot is obtained, but of a lesser slope from
that without added perchlorate (Fig. 8), and with zero
intercept.
molϪ1 22
Iodide reduction has a negative activation entropy
.
whereas the bromide reduction displays a small positive
activation entropy. The reverse reaction has a smaller heat of
activation than the forward reaction; the overall reaction is
endothermic to the extent of 15.2 kJ molϪ1. However, the
reverse reaction has a negative entropy of activation, and the
reaction overall has a negative free energy change of 16.5 kJ
molϪ1 or an equilibrium constant of 780 MϪ1 for the electron
exchange at 298 K. The entropy change for the overall reaction
is 106 J KϪ1 molϪ1, i.e. large and positive, hardly surprising
if one starts off with what approximates to an ion triplet
and finishes with what must be a much looser association
between uncharged D and Br2Ϫ. The difference in the entropies
of activation between the iodide and the bromide reaction is at
first sight surprising, but if it is considered that the iodide
transition state for reduction is early, i.e. like the tight ion-
The treatment used by one23 of us on the solvolysis and
anation of 1,10-phenanthrolineiron() in dmso in the presence
of reactive (ClϪ) and non-reactive (ClO4Ϫ) ions can be applied
to this case too. The concentration of free Qϩ, [Q]Free, was
first calculated from the approximation that it is equal to the
square root of ([Qϩ]Total/105), 1 × 105 being a suitable value for
the ion-pair formation constant.24 Then if [QClO4]Total is xM
and [QBr]Total is yM, the amounts of ion-paired ClO4Ϫ and BrϪ
are given by eqns. (6) and (7). Adding both of these quantities
[QϩClO4
]
IP = 105[Q]Free x/(1 ϩ 105[Q]Free
)
(6)
(7)
Ϫ
triplet precursor complex, and that the bromide transition state
Ϫ
is late, like the loose successor complex between D and Br2
,
then the values are more understandable. The negative entropy
[QϩBrϪ]IP = 105[Q]Free y/(1 ϩ 105[Q]Free
)
Ϫ
of the reverse reaction may concern separated D and Br2
coming together to form the more tightly bound transition state.
and subtracting them from [Qϩ]Total gave a new value of [Q]Free
,
Ϫ
and from this new values of [QϩClO4
]
and [QϩBrϪ]IP were
IP
§ For example, let it be supposed that the reaction has a ∆G‡ of activ-
ation of 4 kJ molϪ1 in water. Then if ∆G‡ varies as 1/ε, in dichloro-
methane the value of ∆G‡ will be 36 kJ molϪ1 (ε = 9 rather than 81).
calculated, leading to a further new value of [Q]Free. The values
typically converged after four cycles.
The converged value of [Q]Free was then used to calculate free
bromide and free perchlorate from eqns. (8) and (9). It should
Using a value of kT/h = 6 × 1012 MϪ1 sϪ1 at 300 K one can calculate
‡
k = 2 × 106 MϪ1 sϪ1 from k = kT/hؒeϪ(∆G /RT)
.
J. Chem. Soc., Dalton Trans., 2000, 3143–3148
3147

View Article Online
[BrϪ]Free = [QϩBrϪ]IP/[Q]Free105
(8)
(9)
rate of reaction, in line with a model involving competitive ion
pairing, with the perchlorate ion pair being inactive. There are
Free = [QϩClO4 ]IP/[Q]Free105
Ϫ
Ϫ
Ϫ
[ClO4
]
good grounds for believing that the back reaction involves Br2
.
The reduction of Dϩ by bromide is slower than by iodide, the
difference being largely accounted for by the difference in the
electron affinities of Br2 and I2. The activation parameters point
to the bromide reaction having a later transition state than that
of the iodide reaction.
be noted that although KIP is given the value of 105 for both
ions, one might more properly use values obtained from
conductivity studies. The concentration of the ion pair DϩBrϪ
was then given by eqn. (10).23 This leads to a revised overall
expression (11) for the forward rate constant.
[DϩBrϪ]IP = 105[Dϩ]Total[BrϪ]Free
{1 ϩ 105[BrϪ]Free ϩ 105[ClO4
/
Acknowledgements
Ϫ
]
}
(10)
Free
J. McK. acknowledges the receipt of a DENI (Department
of Education for Northern Ireland) postgraduate award and
travel grant, J. G. J. the support of the University of Ulster,
Helena Campbell Jones, and hospitality from Brown Uni-
versity, Providence, RI and the University of New Hampshire,
Durham, NH. C. D. H. acknowledges hospitality from
the University of Ulster. We also thank Professors Dwight
Sweigart and John Edwards for fruitful discussions, and NATO
for a travel grant (J. G. J.).
kobs = k1[BrϪ]Total(105[Dϩ]Total[BrϪ]Free
/
Ϫ
{1 ϩ 105[BrϪ]Free ϩ 105[ClO4
]
Free}) (11)
From this it can be seen that if our proposed mechanism
holds the observed rate constant will vary approximately
inversely as the added perchlorate concentration for a set con-
centration of added bromide (Fig. 7). The lessened slope of the
plot of observed rate constant in the presence of a set concen-
tration of perchlorate and a varying amount of added bromide
(Fig. 8) is also accounted for. An interesting feature of the latter
plot is the much smaller value of the reverse rate constant in
the presence of perchlorate (it tends to zero). This is easy to
account for in terms of competition between perchlorate and
Br2Ϫ for the product dimer. Only encounters with the dibromide
radical anion will be involved in the reverse reaction, and since
the concentration of perchlorate is so much higher than that of
the transient radical anion the number of active encounters will
be negligible compared to those which do not lead to reaction.
In the literature of further-oxidised iron() species emphasis
References
1 Part 1: C. D. Hubbard and J. G. Jones, Inorg. Chim. Acta, 1986,
125, 71.
2 P. George, Nature (London), 1953, 169, 612.
3 D. Dolphin, Isr. J. Chem., 1981, 21, 67; a good selection of
references can be found in P. K. Shantha and A. L. Verna, Inorg.
Chem., 1996, 35, 2723.
4 (a) J. T. Groves, R. C. Haushalter, M. Nakamura, T. E. Nemo and
B. J. Evans, J. Am. Chem. Soc., 1981, 103, 2884; (b) G. N. LaMar,
J. S. deRopp, L. Latos-Grazynski, A. L. Bach, R. B. Johnson,
K. M. Smith, D. W. Parish and R.-J. Cheng, J. Am. Chem. Soc.,
1983, 105, 782; (c) A. Gold, K. Jayaraj, P. Doppelt, R. Weiss,
G. Chottard, E. Bill, X. Ding and A. X. Trautwein, J. Am. Chem.
Soc., 1988, 110, 5756.
5 D. L. Hickman, A. Nanthakumar and H. M. Goff, J. Am. Chem.
Soc., 1988, 110, 6384; A. Nanthakunmar and H. M. Goff, J. Am.
Chem. Soc., 1990, 112, 4047.
6 (a) P. Gans, G. Buisson, E. Durée, J.-C. Marchon, B. S. Erler,
W. F. Scholtz and C. A. Reed, J. Am. Chem. Soc., 1986, 108, 1223;
(b) M. A. Phillipi and H. M. Goff, J. Am. Chem. Soc., 1982, 104,
6026.
7 (a) O. Almarsson and T. C. Bruice, J. Am. Chem. Soc., 1995, 117,
4533; (b) M. H. Liu and Y. O. Su, J. Electroanal. Chem., Interfacial
Electrochem., 1998, 452, 113.
8 G. G. Haraldsson, J. E. Baldwin, J. G. Jones and J. Debernardis,
Polyhedron, 1993, 20, 2453.
9 J. E. Baldwin, G. G. Haraldsson and J. G. Jones, Inorg. Chim. Acta,
1981, 51, 29.
10 E. T. Shimomura, M. A. Phillippi, H. M. Goff and W. F. Sholz,
J. Am. Chem. Soc., 1981, 103, 6778.
11 J. G. Jones, D. J. Farrington, F. M. McDonald, P. M. Mooney, Jr.
and M. V. Twigg, Transition Met. Chem., 1998, 23, 693.
12 M. M. Jones, Elementary Coordination Chemistry, Prentice-Hall,
Englewood, NJ, 1964, p. 289.
13 R. G. Wilkins, The Study of the Kinetics and Mechanism of
Reactions of Transition Metal Complexes, Allyn and Bacon, Boston,
MA, 1974, pp. 16–19.
14 J. H. Beard and P. H. Plesch, J. Chem. Soc., 1964, 4075.
15 J. H. Beard and P. H. Plesch, J. Chem. Soc., 1964, 4879.
16 S. D. Malone and J. F. Endicott, J. Phys. Chem., 1972, 76, 2223.
17 W. H. Woodruff and D. W. Margerum, Inorg. Chem., 1973, 12, 962.
18 L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 1957, 61,
1089.
19 M. E. Langmuir and E. Hayon, J. Phys. Chem., 1967, 71, 3808.
20 A. J. Fudge and K. W. Sykes, J. Chem. Soc., 1952, 119.
21 G. Nord, B. Pedersen and O. Farver, Inorg. Chem., 1978, 17, 2233.
22 H. B. Gray and G. P. Haight, Basic Principles of Chemistry, W. A.
Benjamin, Inc., New York, NY, 1967.
has been on oxygen atom transfer from the ferryl group, Fe᎐O,
᎐
to the reducing agent. In this and our previous paper all the
evidence points to the Fe–O–Fe grouping remaining intact in
electron transfer. In the non-donor solvent dichloromethane it
is to be expected that the sites trans to oxygen in Dϩ are vacant,
and bromide ion would be expected to co-ordinate transiently
at one of these in forming the essential ion pair. In ferryl species
there is generally a ligand (e.g. histidine) trans to the oxygen of
the ferryl group blocking direct co-ordination to the iron, and
electron transfer must go via the porphyrin ring (outer sphere)
or by prior transient co-odination to the oxygen atom formally
doubly bonded to the iron (inner sphere).
The formal oxidation state of iron in the oxidising agent Dϩ
is 3.5, and the conclusion from Goff’s experiments6b is that
each ring has a charge of ϩ0.5. However, the blocking effect of
perchlorate ions and the success of the ion-pairing treatment
might suggest that iron is the site of bromide attack, as each
ring provides such an area for attack on its circumference that
effective blocking seems unlikely. On the other hand the effect
of added perchlorate is to reduce the number of DϩBrϪ ion
pairs wherever BrϪ is situated adjacent to the Dϩ species, so
this apparently common-sense argument is not conclusive. It
would be interesting to repeat these experiments for the doubly
oxidised species1,8 D2ϩ. This has a formal oxidation state of 4
for each iron atom, as in the ferryl grouping, and the UV-visible
spectrum in the Soret region could be held to indicate a lot of
porphyin radical cation character in the porphyrin rings, as do
other parameters.6b
Conclusion
The reaction between a formally FeIII–O–FeIV porphyrin dimer
(Dϩ) and a quaternary ammonium bromide in dichloromethane
involves a dimer/bromide ion pair being attacked by bromide
ion in both the form of a free ion and as an ion pair with the
quaternary cation (an inner-sphere reaction). The products
23 D. J. Farrington, J. G. Jones and M. V. Twigg, J. Chem. Soc., Dalton
Trans., 1979, 221.
24 V. Lepentsiotis, R. van Eldik, F. F. Prinsloo and J. J. Pienaar,
J. Chem. Soc., Dalton Trans., 1999, 2759; V. Imper and H. E. van
Wart, Croat. Chem. Acta, 1996, 69, 997.
25 D. Behar, P. L. T. Bevan and D. Scholes, J. Phys. Chem., 1972, 76,
1537.
Ϫ
are Br3 and (FeTPP)2O. Adding quaternary ammonium per-
chlorate reduces the concentration of DϩBrϪ ion pairs, and the
3148
J. Chem. Soc., Dalton Trans., 2000, 3143–3148