354
Can. J. Chem. Vol. 76, 1998
Provided Ki[NH2OH] < < 1 (i = 1–4) and cR > > cFe, the scheme
leads to eq. [10].
K2. On the other hand, the smaller charge borne by I2 should
render an intramolecular electron transfer less facile, and k1 is
expected to be greater than k2. The overall effect of such op-
posing factors is difficult to predict.
Concentration of the dihydroxo species 3 is very small in
the pH range used in kinetics. It does not contain any replace-
able water molecule, and all the ligands present are held more
tightly than H2O. It should also be a weak oxidant due to the
smaller positive charge on its iron(III) centres. Kinetic activity
of 3 is therefore inappreciable in the range pH 3–6.
Conversion of intermediates I1–I3 to products must be mul-
tistep processes. We propose the rate-determining steps to be
one-electron reduction of the intermediates to respective FeII-
O-FeIII dimers, which quickly collapse to FeII and FeIII mono-
mers. The iron(III) monomer [Fe(phen)2(H2O)2]3+ is further
reduced to the iron(II) monomer [Fe(phen)2(H2O)2]2+, which
rapidly forms 4 in the presence of phen. We have verified that
[Fe(phen)3]3+ is reduced by hydroxylamine within the time of
mixing; [Fe(phen)2(H2O)2]3+ is expected to react faster. It is
also known that formation of 4 from Fe2+ and excess phen is
a fast reaction (27) (rate constant ~105 s–1). Inner-sphere re-
duction of [Fe(phen)3]3+ by NH2OH should be slow since sub-
stitution on this iron(III) complex is slow. Such is not the case
for the diaqua derivative. Even with [Fe(phen)3]3+ the outer-
sphere reaction should be fast. We used simple Marcus relation
to calculate the rate constant for outer-sphere oxidation of
NH2OH by [Fe(phen)3]3+ to be 490 M–1s–1. Consequently,
even at the lowest cR (= 0.005 M,) used in kinetics, half life
for this reaction is only 0.28 s. The parameters used in this
calculation are E0 = 1.05 V for the couple [Fe(phen)3]3+/2+, its
self-exchange rate constant (≈107 M–1 s–1) (28), E0 = 0.42 V
for the couple NH2OH+/ NH2OH, and its self-exchange rate
constant, 5 × 10–13 M–1 s–1 (8).
[10] ko{([H+] + Ka1)[phen] + Kd[H+]}/ [NH2OH]
= (k1K1[H+] + k2K2Ka1)[phen] + k3K3Kd[H+ ]
= m1[phen] + c1 (say)
The assumption, Ki[R] < < 1 seems justified because ko and
the absorbance at the isosbestic point increase linearly with
cR. The maximum value of [NH2OH] used in kinetics is
0.015 M (pH 4.50, cR = 0.05 M) up to which we observed no
indication for rate saturation. Hence, an upper limit for Ki is 5
M–1. Stynes and co-workers (26) noted that weakly basic
amines form weak adducts with the oxo-bridged diiron com-
plex [FeIII(dmgBF2)2]2O, (dmgBF2 is difluoro(dimethyl-
glyoximato)borate). Hydroxylamine is a weak base and hence
expected to form weak adducts with 1 and its derivatives. Hy-
droxylamine can form O-bonded complexes, but stability con-
stants for such complexes also are very small (K ≤ 0.1) (21a).
It is not possible for NH3OH+ to coordinate to FeIII without
loss of a proton first. Hence, the small values for Ki and the
weak acidity of NH3OH+ excludes the possibility of
NH3OH+ to act as an active reducing agent.
The ko data at fixed pH (4.50) and cR (0.020 M) but variable
phen yielded a good fit (r = 0.9989; m1 = 8.29 × 10–5; c1 = 2.92
c
× 10–9) into eq. [10]. However, a better fit results (r = 0.9997;
m1 = 8.43 × 10–5; c1 = 1.40 × 10–9) if ko data at the three lowest
c
phen (0.05, 0.10, 0.20 mM) are excluded (Fig. 3). In the pres-
ence of excess phen, [1d] is negligible compared to ([1] + [2]),
and (Ka1 + [H+])[phen] is > > Kd[H+]. Equation [10] may then
be transformed to equation [11].
[11] ko([H+] + Ka1)([H+] + Ka) / cR = a[H+]2 + b[H+] + c
where a = k3K3KdKa/KHphencphen, b = (k1K1Ka + k3K3KdKa/cphen),
Further, the oxo bridge in 1 is stable, and the complex suf-
fers no autodecomposition within the time of kinetic studies.
Martell and co-workers (29) explained that the {Fe2O}4+ core
unit gains stability from superexchange of the d5 high-spin
iron(III) centres across the oxo bridge. But the high-spin FeII
and FeIII are probably less strongly bound to O2–, since both
oxidation states have two anti bonding electrons directed to-
wards the formal bond axes. This would impart weaker
FeII—O—FeIII and FeII—O—FeII bonds, both of which should
be rapidly broken by aquation. In fact the oxo bridge in the
mixed valence system FeII-O-FeIII is rare and in general puta-
tive outside a protein environment (1c). It thus appears that
reduction of {Fe2O}4+ complexes by hydroxylamine involves
several steps, but the rate-determining step is
and c = k2K2Ka1Ka. Values for ko at different pH (≥3.5) but
fixed cphen and cR were fitted into eq. [11] and yielded a = (2.66
± 0.01) × 10–3 M–1 s–1, b = (2.31 ± 0.003) × 10–6 s–1, and c =
(5.46 ± 0.03) × 10–11 M s–1. From these values, one can extract
k1K1 = 0.93 ± 0.005 M–1 s–1, k2K2 = 13.6 ± 0.06 M–1 s–1, k3K3
= 19.1 ± 0.04 M–1 s–1. These parameters and eq. [10] generally
reproduce ko values within ± 10% (see parenthetical values in
Table 1). The calculated values also follow the expected trend
of variation, but they are considerably smaller than the experi-
mental data when [phen] is very low. Trace-metal ion catalysis
may become effective at low [phen] and is a probable cause for
the observed deviations.
The second-order rate constants are composite constants
(Kiki) and both Ki and ki can control the reactivity of the inter-
mediates. A higher value for K3 than either K1 or K2 seems
reasonable, since an iron(III) centre in 1d should be more elec-
tron deficient than that in either 1 or 2. Moreover, 1d is struc-
turally less rigid due to the absence of a bulky phen ligand, and
can easily accommodate structural changes associated with the
electron transfer step,
{Fe2O}4+ → {Fe2O}3+
All other steps are either too fast or too slow to qualify as
the rate-determining step. The immediate oxidation product of
hydroxylamine should be NH2OH+. This is known to generate
N2O via HNO (20a, 23):
NH2OH → H+ + NHOH + e– → H+ + HNO + e–
→ 0.5N2O + 0.5H2O
{FeIII2O}4+ → {FeIIIFeIIO}3+
On this ground, k3 is likely to be larger than both k1 and k2.
The higher reactivity of I3 thus seems justified. It is more dif-
ficult to assess the order of reactivity of I1 with respect to I2.
The water molecule in I1 is ≈4 times a stronger acid than
NH3OH+. So internal proton transfer in I1 may weaken the
adduct. This is not possible for I2. Thus, K1 should be less than
The values for Ki could refer to outer-sphere precursor com-
plex formation. However, hydroxylamine has a record low
self-exchange rate constant, and outer-sphere reactions occur
very rarely with NH2OH in spite of its moderately low E°
value. We observed that 1 does not oxidize [Fe(CN)6]4–,
© 1998 NRC Canada