Replacement of imidazoles by azide at an iron(III) porphyrin center: Part 1 replacement of N-methylimidazole in bis-(N-methylimidazole) tetraphenylporphyrinato iron(III)

Article abstract of DOI:10.1080/00958972.2010.503274

Addition of tetra-n-butylammonium azide to acetone solutions of the tetraphenylporphyrinato iron(III) complex [Fe(TPP)(MeIm)₂] ⁺N^-₃ formed in situ from the reaction of N-methylimidazole and [Fe(TPP)N₃] afforded equilibrium amounts of [Fe(TPP)(MeIm)₂]⁺ and [Fe(TPP)(MeIm)N₃]. Equilibrium experiments made using a range of known concentrations of added tetra-n-butylammonium azide and N-methylimidazole using stopped-flow apparatus gave an estimate of about 50 for the equilibrium constant for the formation of [Fe(TPP)(MeIm)N₃] from the addition of azide to [Fe(TPP)(MeIm) ₂]⁺. Kinetic studies indicated that the substitution of azide ion by N-methylimidazole is a dissociative process, and the results were interpreted using a stationary state approach in which [Fe(TPP)(MeIm)] ⁺ was the transient intermediate that discriminated in favor of the reaction with azide as opposed to the reaction with N-methylimidazole by about a factor of two. Loss of N-methylimidazole from [Fe(TPP)(MeIm)₂] ⁺ is at least 50 times faster than that from the product [Fe(TPP)(MeIm)N₃]. Using calculated values of free azide concentrations from experimentally determined ion-pair formation constants led to success in rationalizing results obtained under different conditions.

Full text of DOI:10.1080/00958972.2010.503274

Journal of Coordination Chemistry
Vol. 63, Nos. 14–16, 20 July–20 August 2010, 2766–2778
Replacement of imidazoles by azide at an iron(III) porphyrin
center: part 1 replacement of N-methylimidazole in
bis-(N-methylimidazole)tetraphenylporphyrinato iron(III)
JOHN G. JONESy, JOHN MCKNIGHTz and MARTYN V. TWIGG*x
yBiomedical Science Research Institute, Centre for Molecular Biosciences, University of
Ulster, Coleraine, BT52 1SA, UK
zGlaxo SmithKline Plc, New Frontiers Science Park, Third Avenue, Harlow,
Essex CM19 5AW, UK
xOrchard Laboratory, Johnson Matthey Plc, Orchard Road, Royston,
Hertfordshire SG8 5HE, UK
(Received 30 March 2010; in final form 27 April 2010)
Addition of tetra-n-butylammonium þazide to acetone solutions of the tetraphenylporphyrinato
iron(III) complex [Fe(TPP)(MeIm)2] Nꢀ3 formed in situ from the reaction of N-methylimida-
zole and [Fe(TPP)N3] afforded equilibrium amounts of [Fe(TPP)(MeIm)2]þ and
[Fe(TPP)(MeIm)N3]. Equilibrium experiments made using a range of known concentrations
of added tetra-n-butylammonium azide and N-methylimidazole using stopped-flow apparatus
gave an estimate of about 50 for the equilibrium constant for the formation of
[Fe(TPP)(MeIm)N3] from the addition of azide to [Fe(TPP)(MeIm)2]þ. Kinetic studies
indicated that the substitution of azide ion by N-methylimidazole is a dissociative process,
and the results were interpreted using a stationary state approach in which [Fe(TPP)(MeIm)]þ
was the transient intermediate that discriminated in favor of the reaction with azide as opposed
to the reaction with N-methylimidazole by about a factor of two. Loss of N-methylimidazole
from [Fe(TPP)(MeIm)2]þ is at least 50 times faster than that from the product
[Fe(TPP)(MeIm)N3]. Using calculated values of free azide concentrations from experimentally
determined ion-pair formation constants led to success in rationalizing results obtained under
different conditions.
Keywords: Azide; N-methylimidazole; Tetraphenylporphyriniron(III); [Fe(TPP)(MeIm)2]þ;
[Fe(TPP)(MeIm)N3]; Ion-pair formation; Stopped-flow
1. Introduction
Previously, we reported [1] the results of a study on the kinetics and equilibria involved
in the two-step formation of [Fe(TPP)(MeIm)2)]þN3ꢀ from [Fe(TPP)N3] and MeIm*,
equations (1) and (2). This differed from the earlier studies [2–11] using [Fe(TPP)X]
(X ¼ F, Cl) in that the intermediate in this study, [Fe(TPP)(MeIm)N3], is stable at room
temperature and can be prepared as a crystalline solid [12]. It contains iron in the
low-spin Fe(III) state, whereas intermediates in the earlier studies, [Fe(TPP)(N-base)Cl]
*Corresponding author. Email: martyn.twigg@matthey.com
Journal of Coordination Chemistry
ISSN 0095-8972 print/ISSN 1029-0389 online ß 2010 Taylor & Francis
DOI: 10.1080/00958972.2010.503274

Iron(III) porphyrin
2767
and [Fe(TPP)(N-base)F], are only transiently stable at ꢀ40ꢁC and have optical spectra
characteristic of high-spin Fe(III) compounds [2–11, 13–16]. In our previous work [1],
we found that the second stage reaction, equation (2), was rate determining, and kinetic
and thermodynamic parameters were determined for this reaction.
½FeðTPPÞN3ꢂ þ MeIm ¼ ½FeðTPPÞðMeImÞN3ꢂ
ð1Þ
½FeðTPPÞðMeImÞN3ꢂ þ MeIm ¼ ½FeðTPPÞðMeImÞ2Þꢂþ þ N3ꢀ
ð2Þ
Equilibrium parameters were also determined for the fast interconversion of
[Fe(TPP)N3] and [Fe(TPP)(MeIm)N3] in the presence of excess MeIm. A free-energy
profile was constructed for the two-stage formation of [Fe(TPP)(MeIm)2)]þN3ꢀ from
[Fe(TPP)N3] and the results compared with those for related Fe(III) porphyrin systems.
We have now found that the second stage reaction, equation (2), can be reversed
in the presence of MeIm by the addition of an excess of tetra-n-butylammonium azide
(N(n-C4H9)4N3) using the same solvent that was used previously, dry acetone. The
product is [Fe(TPP)(MeIm)N3] and under the conditions of this study, which were
essentially the same as those used previously [1]. But with added azide present, there was
no evidence for the transformation of [Fe(TPP)(MeIm)N3] to [Fe(TPP)N3]. However,
now both the concentration of MeIm and Nꢀ3 were in excess and effectively constant
and well defined throughout the course of the reactions because each was substantially
more than the total concentration of the porphyrin complexes present. In our earlier
study, Nꢀ3 in solution was only derived from the dissociation of [Fe(TPP)N3] and so its
concentration varied during the reactions from around zero at the beginning to about
2 ꢃ 10ꢀ5 mol Lꢀ1 at the end of the reaction.
Our finding that reaction (2) is reversible enabled a thorough investigation of this
reaction, particularly of the thermodynamic equilibrium, by varying the concentration
of both Nꢀ3 and MeIm. Using higher concentrations of Nꢀ3 meant that the possibility of
ion pairing might have been important, and Nꢀ3 in acetone (dielectric constant 20.7) is
likely to be partitioned between free Nꢀ3 and ion-paired Nꢀ3 , thereby reducing the
effective concentration of Nꢀ3 . These effects are considered in this article, along with the
results obtained by keeping the total concentration of electrolyte constant using tetra-n-
butylammonium perchlorate (N(n-C4H9)4ClO4). Full details of the conductance studies
that established the ion-pairing constants used in this article will be described in a
subsequent publication.
2. Experimental
2.1. Liquids
Dimethylformamide (DMF, reagent grade BDH) was distilled from calcium hydride
under a reduced pressure of nitrogen and used immediately. Acetone (AnalaR) was
dried over molecular sieves (Linde 4A, activated at 200ꢁC for 4 h at 10 mm Hg). The dry
acetone was filtered through an oven-dried sintered-glass filter and kept in a sealed flask
in a dry box. N-methylimidazole (Aldrich) was dried over crushed KOH and distilled
from calcium hydride and kept under dry nitrogen over activated molecular sieves
(Linde 4A).

2768
J.G. Jones et al.
2.2. Solids
Iron(III) tetraphenylporphyrin azide, [Fe(TPP)N3], was prepared from spectroscopi-
cally pure tetraphenylporphyrin via iron insertion to give [Fe(TPP)Cl], followed by
alkali treatment to form [(Fe(TPP))2O], and final conversion to the azide using a dilute
aqueous solution of hydrazoic acid as detailed below.
2.2.1. Preparation of [Fe(TPP)Cl]. Spectroscopically pure tetraphenylporphyrin
(2.94 g), obtained from the Rothenmund reaction [17, 18], that had been treated with
2,3-dichloro-5,6-dicyanobenzoquinone (DDQ) [19] to remove traces of chlorin, was
dissolved in dry DMF (550 mL). The solution was brought to reflux and anhydrous
FeCl2 (2.0 g) dissolved in the minimum amount of dry DMF was added, and refluxing
continued for 0.5 h. After allowing the solution to cool, it was filtered through sintered
glass and the retained crystals extracted with dichloromethane. The remaining filtrate
was poured into an equal volume of distilled water and the product extracted with
dichloromethane. The combined dichloromethane extracts were evaporated in the
presence of a small amount of added methanol and crystallization ensued. The product
obtained by filtration was washed with pentane and dried at 60ꢁC for 2 h. If the visible
spectrum of the product showed the characteristic narrow bands of unconverted
(TPP)H2, the purification treatment was repeated to give complete conversion. The
yield of spectroscopically pure [Fe(TPP)Cl] was typically 80%.
2.2.2. Preparation of [(Fe(TPP))2O] (l-oxo-dimer). Dichloromethane (500 mL,
AnalaR) was deacidified by shaking with aqueous NaOH 100 mL (2 mol Lꢀ1) in a
separating funnel, and the lower layer dried with two successive treatments of
anhydrous CaCl2. In a typical preparation, [Fe(TPP)Cl] (800 mg) in dry dichloro-
methane (50 mL) was shaken with three successive portions of aqueous NaOH (100 ml,
2 mol Lꢀ1). The dark brown lower layer was separated and dried over anhydrous
Na2CO3. The spectrum of the ꢀ-oxo-dimer had distinguishing peaks in the visible
spectrum at 408, 570, and 610 nm, and its final purification was by column
chromatography. The dried dichloromethane extracts were transferred to a chroma-
tography column (60 mL) containing alkaline alumina (30 mL) and eluted with dry
de-acidified dichloromethane. The main band near the top of the column was preceded
by a pale green band. The brown eluate solution had the spectrum of [Fe(TPP)2O], and
its volume was reduced by 75% on a rotary evaporator and the product precipitated by
the addition of methanol; the crystals were collected by filtration and dried at 80ꢁC
for 2 h. The yield was typically greater than 90%.
2.2.3. Preparation of [Fe(TPP)N3]. The method employed was a modification of that
of Summerville and Cohen [20]. Hydrazoic acid solution was prepared by passing
a saturated aqueous NaN3 solution (150 mL) through an acid cation ion-exchange
column (50W-X8, 125 mL). The resulting solution was then shaken with [Fe(TPP)2O]
(400 mg) in freshly distilled benzene (300 mL) and the mixture allowed to stand for
0.5 h. The aqueous layer was then discarded and the benzene layer washed three times
with an equal volume of distilled water, dried over anhydrous MgSO4, and filtered.
After reducing the volume of the filtrate to one-third on a rotary evaporator (with
exclusion of moisture), pentane was added to give crystals of the product, which were

Iron(III) porphyrin
2769
separated by filtration and dried (4 h at 60ꢁC). The yield was 92%. The product was
discarded if its visible spectrum in dry acetone had bands characteristic of [Fe(TPP)2O]
(570 and 610 nm), resulting from incomplete conversion of the reactant or hydrolysis of
[Fe(TPP)N3]. None of the [Fe(TPP)N3] used in the present stopped-flow experiments
contained [(Fe(TPP))2O], and elemental analyses were similar to or better than those
reported previously.
2.2.4. Tetra-n-butylammonium perchlorate. N(n-C4H9)4ClO4 was prepared by adding
a concentrated aqueous solution of N(n-C4H9)4OH (Aldrich) to a saturated solution of
NaClO4. The resulting white precipitate was harvested by filtration, washed, and
recrystallized from aqueous methanol, and then dried over phosphorus pentoxide in
a vacuum desiccator.
2.2.5. Tetra-n-butylammonium azide. N(n-C4H9)4N3 was prepared from a concen-
trated aqueous solution of N(n-C4H9)4OH and saturated NaN3 according to the
procedure outlined by Brandstrom et al. [21]. In a typical preparation, N(n-C4H9)4OH
solution (Aldrich) (10 mL 40%) was diluted with an equal volume of distilled water and
extracted three times with deacidified chloroform (shaken with sodium hydroxide
solution) to remove traces of N(n-C4H9)4Br that was often present in the commercial
N(n-C4H9)4OH solutions. The purified aqueous N(n-C4H9)4OH solution was mixed
with a saturated solution of NaN3 (2.0 g). A white precipitate formed that re-dissolved,
and the desired product was isolated by the extraction with chloroform, evaporation of
the solution to dryness, and removal of remaining traces of chloroform in vacuo over
phosphorus pentoxide. The pure product was obtained by recrytallization from dry
benzene or toluene. Analyses were typically C ¼ 67.2%, H ¼ 13.0%, and N ¼ 20.1%;
C16H36N4 requires: C ¼ 67.53%, H ¼ 12.78%, and N ¼ 19.69%. It was confirmed that
no bromide was present in the product and it was observed by the oxidation of azide
with acidic NaNO2 in a procedure based on that of Vogel [22, 23] followed by titration
with mercuric nitrate at pH 3 to estimate bromide. Total azide was estimated using
Arnold’s method with ammonium ceric sulfate [24].
2.3. Kinetic and equilibrium measurements
Kinetic and equilibrium measurements were made using stopped-flow equipment and
procedures described previously [1].
3. Results and discussion
3.1. Conductance studies
Â
Ãþ
Ion-pair fþormation constants for FeðTPPÞðMeImÞ2 Nꢀ3 , Nðn-C4H9Þ4þClO4ꢀ, and
Nðn-C4H9Þ4 Nꢀ3 in dry acetone (with rigorous exclusion of moisture) were determined
from conductance measurements. All of the compounds behaved as weak electrolytes,
and the values of the determined ion-pair formation constants collected in table 1 were

2770
J.G. Jones et al.
Table 1. Ion-pair formation constants (KIP) for selected
salts in dry acetone at 25.0ꢁC.
Salt
KIP
Source
Nðn-C4H9Þþ4 ClOꢀ4
Nðn-C4H9Þþ4 Nꢀ3
104
335
335 est.
[32]
[33]
[33]a
[Fe(TPP)(MeIm)2]þN3ꢀ
aMeasurements made in the presence of 5 mol Lꢀ1 MeIm to ensure
the complete formation of [Fe(TPP)(MeIm)2]þ.
Figure 1. Visible spectrum of [Fe(TPP)(MeIm)2]þ in dry acetone at 25ꢁC (–––––––) and in the presence of
excess added N(n-C4H9)4N3 (– – – –). The inset has the absorbance scale increased by a factor of five.
used to calculate the proportion of Nꢀ3 present as free ionic and ion-paired species in
the reactant solutions used in this study.
3.2. Equilibrium measurements
Stopped-flow equipment was used to monitor the progress of the reaction when an
acetone solution of [Fe(TPP)N3] was mixed with a solution of MeIm. The disparity in
the fast formation of [Fe(TPP)(MeIm)N3] and the much slower reaction to give the final
product [Fe(TPP)(MeIm)2]N3 enabled the equilibrium constant for the first reaction,
equation (1), to be estimated. Thus addition of a solution of N(n-C4H9)4N3 to one of
[Fe(TPP)(MeIm)2]N3 generated in situ from [Fe(TPP)N3] and MeIm gave a distinct
change in the optical spectrum, as shown in figure 1. The absorbance changes were
greatest at 445 and 550 nm.
Addition of a solution of [Fe(TPP)N3] containing different amounts of N(n-
C4H9)4N3 at 25ꢁC to one of MeIm so that the final concentration of MeIm was
0.01 mol Lꢀ1 resulted in the absorbance changes listed in table 2, where AS is the
absorbance of the intermediate [Fe(TPP)(MeIm)N3] immediately after the first reaction,
AE the absorbance when the reaction reached equilibrium, and AF the absorbance if the

Iron(III) porphyrin
2771
Table 2. Absorbance changes following mixing equal volumes of dry acetone solutions
of [Fe(TPP)N3] containing different amounts of N(n-C4H9)4N3 to one of MeIm
(0.02 mol Lꢀ1) so that the final concentration of MeIm was 0.01 mol Lꢀ1 at 25.0ꢁC.
[Nꢀ3 ]
AS
AE
AF
DA1
DA2
0
0.19
0.19
0.19
0.19
0.19
0.19
0.00125
0.0025
0.0045
0.005
0.01
0.645
0.592
0.641
0.650
0.664
0.578
0.575
0.623
0.641
0.655
0.067
0.017
0.018
0.009
0.009
0.388
0.385
0.433
0.451
0.465
See text for definitions of the absorbance symbols.
Figure 2. Absorbance changes at 445 nm when N(n-C4H9)4N3 was added to an acetone solution of
[Fe(TPP)(MeIm)2]þ containing 0.01 mol Lꢀ1 MeIm at 25ꢁC with the total porphyrin concentration of about
5 ꢃ 10ꢀ5 mol Lꢀ1
.
reaction had gone to completely transform [Fe(TPP)N3] into [Fe(TPP)(MeIm)2]N3.
The term AF was measured by mixing the porphyrin solution containing no added
N(n-C4H9)4N3 to the MeIm solution. It can be shown that DA1 ¼ (AS ꢀ AE) is
proportional to the amount of [Fe(TPP)(MeIm)2]N3 present at equilibrium, and that
DA2 ¼ (AE ꢀ AF) is proportional to the amount of [Fe(TPP)(MeIm)N3] at equilibrium.
Then the equilibrium constant for the reaction in equation (1) is given by equation (3),
provided [MeIm] ꢄ [Fe(TPP)N3], and [N(n-C4H9)4N3] ꢄ [Fe(TPP)N3].
K ¼ DA1½Nðn-C4H9Þ4N3ꢂ=DA2½MeImꢂ
ð3Þ
This treatment depends heavily on the actual values of AS and AF and can be subject to
considerable error; similarly DA1 values become small as [N(n-C4H9)4N3] increases
relative to [MeIm], and all of these effects contribute to the total error associated with K.
Averaging the five values of K calculated from the data in table 2 gave K ¼ 67 ꢅ 23 for
reaction (4).
þ
½FeðTPPÞðMeImÞ2Þꢂ þ Nꢀ
½FeðTPPÞðMeImÞN3ꢂ þ MeIm
ð4Þ
ꢀ!
ꢀ
3
Figure 2 shows these results graphically in the form of a plot of DA1 against
increasing [N(n-C4H9)4N3] indicating the extent to which the amount of

2772
J.G. Jones et al.
Figure 3. Variation of the observed first-order rate constant (kobs) for the conversion of [Fe(TPP)(MeIm)2]þ
to [Fe(TPP)(MeIm)N3] in the presence of N(n-C4H9)4N3 (2.75 ꢃ 10ꢀ3 mol Lꢀ1) and varying concentrations
of MeIm in dry acetone.
[Fe(TPP)(MeIm)2]þ decreased as the reaction proceeded, entirely in line with the
equilibrium nature of the reaction.
3.3. Kinetic studies
Adding azide, as N(n-C4H9)4N3, to a solution of [Fe(TPP)(MeIm)2]þ gave the neutral
product [Fe(TPP)(MeIm)N3] and this reaction could be studied by the stopped-flow
method. In these experiments, an acetone solution of [Fe(TTP)N3] that had been
converted to [Fe(TPP)(MeIm)2]þ by addition of a known excess of MeIm was mixed
with a solution containing a known amount of N(n-C4H9)4N3. The resulting
absorbance increases at 445 nm were analyzed as accurate first-order processes to
give observed first-order rate constants. In figure 3, results are presented which show
the effect of adding the same amount of azide (2.75 ꢃ 10ꢀ3 mol Lꢀ1 in the mixed
solution) to solutions containing varying amounts of MeIm. The observed first-order
rate constants decreased as the concentration of MeIm increased, and in figure 4, it
is shown that their values were inversely proportional to the concentration of MeIm
present. This behavior is characteristic of a dissociative ligand-replacement reaction
involving a transient intermediate [25], and previously high pressure techniques, refined
and extensively used by van Eldik and his coworkers [26–28], demonstrated the
dissociative nature of other axial ligand exchange reactions at iron porphyrin centers
[29]. Equation (5) characterizes this process in terms of the stationary state intermediate
[Fe(TPP)(MeIm)]þ, formed by the fission of one Fe–N bond in the bis-MeIm complex.
Since the concentration of added azide is over 100-fold greater than that from
[Fe(TPP)N3], only the amount added is considered in deriving equation (6) for the
observed rate constant based on the mechanism in equation (5).
k3½Nꢀ3 ꢂ
k1
þ
½FeðTPPÞðMeImÞ2ꢂþ
½FeðTPPÞðMeImÞꢂ þ MeIm
½FeðTPPÞðMeImÞN3ꢂ
ꢀꢀꢀꢀ!
ꢀꢀꢀꢀ
k2½Melmꢂ
ꢀꢀꢀꢀ!
ꢀꢀꢀꢀ
k4
ð5Þ

Iron(III) porphyrin
2773
Figure 4. Plot of the observed rate constant (kobs/sꢀ1) against 1/[MeIm] for the addition of azide
(N(n-C4H9)4N3, 2.75 ꢃ 10ꢀ3 mol Lꢀ1) to [Fe(TPP)(MeIm)2]þ in the presence of varying concentrations of
MeIm in dry acetone.
Figure 5. Plots of kobs (sꢀ1) for the addition of azide (N(n-C4H9)4N3) to [Fe(TPP)(MeIm)2]þ in the presence
of a constant concentration MeIm (open points 0.02 mol Lꢀ1, solid points 0.01 mol Lꢀ1) at 25ꢁC with total
ionic strength 0.01 mol Lꢀ1 in dry acetone.
kobs ¼ ðk1k3½Nꢀ3 ꢂ þ k2k4½MeImꢂÞ=ðk2½MeImꢂ þ k3½Nꢀ3 ꢂÞ
kobs ¼ k1k3½Nꢀ3 ꢂ=k2½MeImꢂ þ k4
ð6Þ
ð7Þ
Only by assuming that k2[MeIm] ꢄ k3[Nꢀ3 ] is a simplified expression of the correct
form, equation (7), obtained which explains the dependence of observed rate constant
on added MeIm in the presence of constant added azide. Equation (7) also explains the
linear dependence of observed rate constant on added azide in the presence of a
constant concentration of MeIm as described below.
In experiments where the concentration of MeIm was kept constant but the
concentration of azide was varied, the observed rate constants changed in an almost
linear fashion with total added azide concentration, as shown in figure 5 for
[MeIm] ¼ 0.01 and 0.02 mol Lꢀ1
, respectively. The linearity was better when

2774
J.G. Jones et al.
Figure 6. Plots of calculated [Nꢀ3 ]free against total added [N(n-C4H9)4N3] for azide/perchlorate mixtures
(constant ionic strength, solid points) and [N(n-C4H9)4N3] alone (open points) in dry acetone at 25ꢁC.
0.02 mol Lꢀ1 MeIm was present. The total ionic strength was kept constant for the
results shown in figure 5 at 0.01 mol Lꢀ1 with N(n-C4H9)4ClO4, and there were also
experiments performed without added N(n-C4H9)4ClO4.
In figure 4, the slope 0.066 represents k1k3[Nꢀ3 ]/k2 with [N3ꢀ] ¼ 0.00275 mol Lꢀ1
,
giving k1k3/k2 ¼ 24 with a relative standard error of 5.5%. However, the intercept k4,
0.314 sꢀ1, was ill defined having a relative standard error of 72%. Thus, the equilibrium
constant (K ¼ k1k3/k2k4) was also defined poorly as 76 ꢅ 55 sꢀ1. The series of
experiments represented by figure 5 provided slightly better estimates of the kinetic
and thermodynamic parameters. With [MeIm] ¼ 0.02 mol Lꢀ1
, the slope k1k3/
(k2[MeIm]) was 1525 sꢀ1 with a relative standard error of 6.4%, and the intercept,
k4, was 0.969 sꢀ1 with a relative standard error of 51% giving K ¼ 30 ꢅ 16. With
[MeIm] ¼ 0.01 mol Lꢀ1, the slope k1k3/(k2[MeIm]) was 2215 sꢀ1 with a relative standard
error of 3.3%, and an intercept, k4, of 1.579 sꢀ1 with a relative standard error of 28%
gave a better value for K ¼ 14.0 ꢅ 4.0.
In this study, the absorbance changes monitored were small, and in different
experiments the media were different with different concentrations of MeIm or Nꢀ3 . So
it is not surprising that the numerical agreement was not perfect. Thus, comparing the
results in figure 5, the slopes should be inversely proportional to the concentration of
MeIm but they are only very approximately so. However, it is instructive to compare
the two sets of results where the concentration of MeIm was the same, namely
0.02 mol Lꢀ1 MeIm. In figure 4 the interpolated value of kobs at [MeIm] ¼ 0.02 mol Lꢀ1
is 3.61 sꢀ1, whereas in figure 5 the corresponding value for kobs is 5.16 sꢀ1. This
difference reflects a higher ionic strength in the latter case (0.01 compared with
0.00275 mol Lꢀ1) that facilitates the ionic reaction by raising the dielectric of the
medium. It therefore seemed appropriate, instead of using the formula concentration of
azide in solution, to calculate the amount of ‘‘free’’ ionic azide present using the
measured ion-pair association constants. Such an approach worked well previously
when ion pairs were involved in substitution and oxidation reactions of iron complexes
[30, 31].
Figure 6 shows that when an ionic strength of 0.01 mol Lꢀ1 is maintained with azide/
perchlorate mixtures (solid points), the concentration of ‘‘free’’ azide is diminished

Iron(III) porphyrin
2775
Figure 7. Variation of observed rate constant (kobs) with [Nꢀ3 ]free in the presence of excess MeIm (solid
points 0.01 mol Lꢀ1, open points 0.02 mol Lꢀ1) in dry acetone at 25ꢁC with a total constant ionic strength
0.01 mol Lꢀ1
.
Figure 8. Plot of 1/kobs (sꢀ1) against 1/[Nꢀ3 ]free for the reaction of N(n-C4H9)4N3 with [Fe(TPP)(MeIm)2]þ
in the presence of excess MeIm (0.02 mol Lꢀ1) with constant total ionic strength (0.01 mol Lꢀ1) in dry
acetone at 25ꢁC.
because of ion-pair formation, but the concentration of free azide ion in the
concentration range considered is a linear function of the concentration of total
azide. In contrast, the relationship is more complex when azide is the sole electrolyte
and the ionic strength varies (open points). The calculated free azide concentrations
were used in the plot of kobs against the ‘‘free’’ azide concentration shown in figure 7,
and also in the plot of 1/kobs against 1/[Nꢀ3 ]free shown in figure 8. In figure 7 excess
MeIm concentrations were used (solid points 0.01 mol Lꢀ1, open points 0.02 mol Lꢀ1
)
that gave two different curved lines, but there was good linearity in the ‘‘double
reciprocal’’ plot in figure 8.
There is a theoretical basis for this observed behavior. The curvature in figure 7 can
be accounted for if the k2[MeIm] term is not much greater than the value of k3[Nꢀ3 ]
as the concentration of Nꢀ3free is increased while that of MeIm is constant.

2776
J.G. Jones et al.
Figure 9. Plot of 1/kobs (sꢀ1) against 1/[N3ꢀ]free for the reaction of N(n-C4H9)4N3 with [Fe(TPP)(MeIm)2]þ in
the presence of excess MeIm (0.01 mol Lꢀ1) in dry acetone at 25ꢁC. Open points with constant azide/
perchlorate ionic strength ¼ 0.01 mol Lꢀ1, solid points with ionic strength varying with [N3ꢀ]total
.
Taking recipꢀrocals of both sides of equation (5) gives equation (8). Then, with
increasing [N3free] if k1k3[Nꢀ3 ] ꢄ k2k4[MeIm], equation (9) results.
1=kobs ¼ ðk2½MeImꢂ þ k3½Nꢀ3 ꢂÞ=ðk1k3½N3ꢀꢂ þ k2k4½MeImꢂÞ
ð8Þ
ð9Þ
1=kobs ¼ ðk2½MeImꢂ=k1k3½Nꢀ3 ꢂÞ þ 1=k1
Equation (9) is of the right form to give the linear plot of 1/kobs against 1/[N3ꢀfree
]
shown in figure 8, with a constant MeIm concentration of 0.02 mol Lꢀ1. The slope
(k2[MeIm]/k1k3) had a value of 1.86 ꢃ 10ꢀ4 with a relative standard error of 9.7%, and
an intercept (1/k1) was 0.0140 ꢅ 0.0194. These values gave an estimate of the
discrimination ratio (k2/k3) of 0.645 with a large associated error, but this lends some
support that rate constant for the reaction of azide with the steady state intermediate
[Fe(TPP)(MeIm)]þ is probably greater than that for its reaction with MeIm, and that
[Fe(TPP)(MeIm)]þ only slightly discriminates between the two incoming ligands, azide,
and MeIm.
The situation was convincing when 0.01 mol Lꢀ1 MeIm was used as shown in figure 9,
that is a graph of 1/kobs against 1/[Nꢀ3free] combining both reactions with azide alone
(filled points) and with azide/perchlorate mixtures maintaining a constant ionic strength
of 0.01 mol Lꢀ1 (open points). The points comprise a common data set showing the
concentration of free azide along with that of MeIm are the important factors
determining the observed rate constants at 25ꢁC. A regression analysis gave a slope
of 106 ꢅ 4 ꢃ 10ꢀ6 (relative standard error of 3.6%) and a significant intercept of
1.95 ꢅ 0.45 (relative standard error of 23%) that when combined gave a discrimination
ratio of 0.543 for k2/k3, confirming that the intermediate [Fe(TPP)(MeIm)]þ is almost
twice as reactive toward free azide than toward MeIm. Comparing the value of k1 (51.3)
with that of k4 (between 0.36 and 0.97), determined from the earlier experiments,
suggested the fission of MeIm from [Fe(TPP)(MeIm)2]þ to give the intermediate
[Fe(TPP)(MeIm)]þ is much easier than the fission of azide from the non-electrolyte
[Fe(TPP)(MeIm)N3]. A simple explanation for this is that the latter reaction involves

Iron(III) porphyrin
2777
charge separation in a low dielectric solvent, whereas the former involves no charge
separation, though each case involves the breaking of a Fe–N bond.
Comparing the results of this study with those from our earlier study [1] in which the
reaction proceeded in the opposite direction and the concentration of ionic species was
very low, not constant, and entirely from the starting material [Fe(TPP)N3], the derived
value of k4 was 0.017 sꢀ1, more than an order of magnitude less than that determined
in this study. This difference can be attributed to the more polar environment provided
by the relatively high ionic concentration that promotes the ionization of coordinated
azide from [Fe(TPP)(MeIm)N3].
4. Conclusions
Addition of azide in the form of N(n-C4H9)4N3 to [Fe(TPP)(MeIm)2]þ in dry acetone
led to displacement of MeIm to form equilibrium amounts of the starting material
along with the nonelectrolyte product [Fe(TPP)(MeIm)N3]. Both equilibrium and
kinetic measurements were made using stopped-flow apparatus involving varying
known concentrations of azide and N-methylimidazole. The kinetic results were
accounted for by the loss of N-methylimidazole from [Fe(TPP)(MeIm)2]þ being a
dissociative process described by a stationary state approach involving the transient
intermediate [Fe(TPP)(MeIm)]þ. The results enabled an estimate to be made for the
equilibrium constant for the formation of [Fe(TPP)(MeIm)N3] by the addition of azide
to [Fe(TPP)(MeIm)2]þ and to establish the intermediate [Fe(TPP)(MeIm)]þ discrimi-
nated in favor of the reaction with azide over reaction with MeIm by nearly a factor of
two, and that fission of MeIm from the starting material is at least 50 times easier than
from [Fe(TPP)(MeIm)N3]. The approach of using calculated values of free azide in
solution using experimentally determined ion-pairing formation constants put results
obtained under different conditions on a common basis.
Acknowledgments
JGJ gratefully acknowledges the support from the University of Ulster, and a joint
NATO grant with Professor Dwight A. Sweigart, Department of Chemistry, Brown
University, Providence, Rhode Island, USA, and also many fruitful discussions. JMcK
acknowledges the receipt of a quota award from the Department of Education for
Northern Ireland and a grant to visit Professor Sweigart’s laboratory.
References
[1] W. Byers, J.A. Cossham, J.O. Edwards, A.T. Gordon, J.G. Jones, E.T.P. Kenny, A. Mahmood,
J. McKnight, D.A. Sweigart, G.A. Tondreau, T. Wright. Inorg. Chem., 25, 4767 (1986).
[2] J.M. Duclos. Bioinorg. Chem., 2, 263 (1973).
[3] C.L. Coyle, P.A. Rafson, E.H. Abbott. Inorg. Chem., 12, 2007 (1973).

2778
J.G. Jones et al.
[4] P.R. Ciacco, J.V. Ellis, M.E. Munson, G.L. Kedderis, F.X. McConville, J.M. Duclos. J. Inorg. Nucl.
Chem., 38, 1885 (1976).
[5] F.A. Walker, M.W. Lo, M.T. Ree. J. Am. Chem. Soc., 98, 5552 (1976).
[6] P.A. Adams, D.A. Baldwin, C.E. Hepner, J.M. Pratt. Bioinorg. Chem., 9, 1949 (1978).
[7] T. Yoshimura, T. Ozaki. Bull. Chem. Soc. Jpn., 52, 2268 (1979).
[8] M. Momentau. Biochem. Biophys. Acta, 304, 814 (1973).
[9] M. Momentau, M. Rougee, B. Loock. Eur. J. Biochem., 71, 63 (1976).
[10] M. Momentau, J. Mispelter, D. Lexa. Biochem. Biophys. Acta, 320, 652 (1973).
[11] (a) D. Burdige, D.A. Sweigart. Inorg. Chim. Acta, 28, L131 (1978); (b) W. Fiske, D.A. Sweigart. Inorg.
Chim. Acta, 36, L429 (1979).
[12] K.M. Adams, P.G. Rasmussen, W.R. Scheidt, K. Hatano. Inorg. Chem., 18, 1892 (1979).
[13] M.M. Doeff, D.A. Sweigart. Inorg. Chem., 21, 3699 (1982).
[14] G.A. Tondreau, D.A. Sweigart. Inorg. Chem., 23, 1060 (1984).
[15] J.G. Jones, G.A. Tondreau, J.O. Edwards, D.A. Sweigart. Inorg. Chem., 24, 296 (1985).
[16] Q. Meng, G.A. Tondreau, J.O. Edwards, D.A. Sweigart. J. Chem. Soc., Dalton Trans., 2269 (1985).
[17] P. Rothemund, A.R. Menotti. J. Am. Chem. Soc., 63, 267 (1941).
[18] A.D. Adler, F.R. Longo, J.D. Finarelli, J. Goldmacher, J. Assour, L. Korsakov. J. Org. Chem., 32, 476
(1967).
[19] G.H. Bennett, M.F. Hudson, K.M. Smith. Tetrahedron Lett., 2887 (1983).
[20] D.A. Summerville, I.A. Cohen. J. Am. Chem. Soc., 98, 1747 (1976).
[21] A. Brandstrom, B. Lamm, I. Palmertz. Acta Chem. Scand., B28, 699 (1974).
[22] A.I. Vogel. Textbook of Quantitative Inorganic Analysis Including Elementary Instrumental
Analysis, 3rd Edn, p. 274, Longmans, London (1974).
[23] W.R. Briggs, R.W. Gaver. Anal. Chem., 44, 1870 (1972).
[24] J.W. Arnold. Ind. Eng. Chem. Anal. Ed., 17, 215 (1945).
[25] (a) See for example discussions in: R.G. Wilkins. The Study of Kinetics and Mechanism of Reactions of
Transition Metal Complexes, VCH Publishers Inc., New York (1991); (b) J.D. Atwood, Inorganic and
Organometallic Reaction Mechanisms, pp. 79–85, Brooks-Cole Publishing, Monterey, CA (1985).
[26] R. van Eldik, J. Jonas (Eds). In High Pressure Chemistry and Biochemistry, Vol. 197, NATO ASI Series
C, Mathematical and Physical Sciences, D. Reidel Publishing Co, Holland (1987).
[27] R. van Eldik, C.D. Hubbard (Eds). In Chemistry under Extreme and Non-Classical Conditions, p. 317,
John Wiley, New York (1997).
[28] R. van Eldik, F.-G. Klarner (Eds.). In High Pressure – Synthetic, Mechanistic, and Supercritical
¨
Applications, John Wiley, New York (2002).
[29] R. van Eldik. Coord. Chem. Rev., 251, 1649 (2007).
[30] D.J. Farrington, J.G. Jones, M.V. Twigg. J. Chem. Soc., Dalton Trans., 221 (1979).
[31] C.D. Hubbard, J.G. Jones, J. McKnight. Royal Soc. Chem., Dalton Trans., 3143 (2000).
[32] M.B. Reynolds, C.A. Kraus. J. Am. Chem. Soc., 70, 1709 (1948).
[33] J. McKnight. The reactions of iron porphyrins in the ferric and higher oxidation states. DPhil thesis,
University of Ulster (1989).