The kinetics and stereochemistry of base hydrolysis of the seven isomers of [Co(dien)(ampy)Cl]2+ and [Co(dien)(ibn)Cl]2+

Article abstract of DOI:10.1016/j.ica.2005.11.050

The kinetics and stereochemistry for the base catalysed substitution reactions of all seven isomers (4 mer and 3 fac) of both [Co(dien)(ibn)Cl]²⁺ and [Co(dien)(ampy)Cl]²⁺ have been studied in detail, for water and azide ion as entering groups. The stereochemistry for the azide ion anation of some of the [Co(dien)(diamine)OH]²⁺ species have also been investigated. The mer isomers are of comparable reactivity and amongst the fastest reacting pentaaminechlorocobalt(III) complexes known. They are also much faster to hydrolyse than the fac species. In both the ibn and ampy systems, a common product stereochemistry is observed for the four reactant mer isomers (the product is a mixture of all four mer configurations), for both azide ion and water as nucleophiles, but not for the three fac reactants (H₂O as nucleophile). The kinetic and equilibrium distributions are quite different. For the mer isomer reactions, a common trigonal bipyramidal five-coordinate intermediate deprotonated at the sec-NH of the dien is overwhelmingly implicated. The substitution mechanisms are argued in detail. Other data reported include isomerisation rates and equilibrium distributions for some mer-hydroxo and a mer-aqua complex of exceptional reactivity, equilibrium distributions for the mer-phosphato complexes in the ampy system under different pH conditions, the crystal structure for the isolated m1-[Co(dien)(ampy)OP(OH)₃]Cl₃ · 2H₂O species, and a rationale for its predominance at neutral pH based on internal H-bonding.

Full text of DOI:10.1016/j.ica.2005.11.050

Inorganica Chimica Acta 359 (2006) 2210–2225
www.elsevier.com/locate/ica
The kinetics and stereochemistry of base hydrolysis of the seven
isomers of [Co(dien)(ampy)Cl]2+ and [Co(dien)(ibn)Cl]2+
a
b,
c
c
Tao Zhu , W. Gregory Jackson *, David C.R. Hockless , Anthony C. Willis
a
Institute of Applied Chemistry, Guizhou University, Guiyang 550025, China
School of Chemistry, University College, University of New South Wales, Australian Defence Force Academy, Northcott Drive,
b
Canberra, ACT 2600, Australia
Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia
c
Received 13 November 2005; accepted 28 November 2005
Available online 15 March 2006
Abstract
The kinetics and stereochemistry for the base catalysed substitution reactions of all seven isomers (4 mer and 3 fac) of both [Co(dien)-
(ibn)Cl]2+ and [Co(dien)(ampy)Cl]2+ have been studied in detail, for water and azide ion as entering groups. The stereochemistry for the
azide ion anation of some of the [Co(dien)(diamine)OH]2+ species have also been investigated. The mer isomers are of comparable reac-
tivity and amongst the fastest reacting pentaaminechlorocobalt(III) complexes known. They are also much faster to hydrolyse than the
fac species. In both the ibn and ampy systems, a common product stereochemistry is observed for the four reactant mer isomers (the
product is a mixture of all four mer configurations), for both azide ion and water as nucleophiles, but not for the three fac reactants
(H2O as nucleophile). The kinetic and equilibrium distributions are quite different. For the mer isomer reactions, a common trigonal
bipyramidal five-coordinate intermediate deprotonated at the sec-NH of the dien is overwhelmingly implicated. The substitution mech-
anisms are argued in detail. Other data reported include isomerisation rates and equilibrium distributions for some mer-hydroxo and a
mer-aqua complex of exceptional reactivity, equilibrium distributions for the mer-phosphato complexes in the ampy system under dif-
ferent pH conditions, the crystal structure for the isolated m1-[Co(dien)(ampy)OP(OH)3]Cl3 Æ 2H2O species, and a rationale for its pre-
dominance at neutral pH based on internal H-bonding.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Base hydrolysis; Phosphoric acid complex; Pentacoordinate intermediate; Diethylenetriamine; Aminomethylpyridine; Isobutylenediamine;
X-ray structure
1. Introduction
N
N
H
H
The structurally related [Co(dien)(ibn)Cl]2+ and
[Co(dien)(ampy)Cl]2+ systems are two for which all seven
geometric isomers are known and which have been fully
characterised by 2D NMR techniques [1,2], and single-
crystal X-ray studies [3,4].
NH2
NH2
Cl
Co
NH2
Cl
Co
N
CH3
CH3
NH2
NH2
NH2
NH2
[Co(dien)(ampy)Cl]2+
[Co(dien)(ibn)Cl]2+
*
Corresponding author. Tel.: +61 2 62688078; fax: +61 2 62688090/
The isomers are shown (Fig. 1) for the ampy system, and
a corresponding set exists for the ibn system.
8017.
E-mail address: wgj@adfa.edu.au (W.G. Jackson).
0020-1693/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2005.11.050

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2211
N
H
N
H
Cl
N
H
NH2
NH2
H2N
Co
N
Cl
Co
N
Cl
Co
N
NH2
NH2
f1
NH2
NH2
m1
NH2
NH2
m2
N
H
Cl
H
H2N
Co
N
NH2
N
N
H
NH2
NH2
NH2
NH2
Cl
Co
N
NH2
Cl
Co
N
NH2
NH2
N
H
H2N
f3
H2N
Co
N
NH2
m3
m4
Cl
f2
Fig. 1. The seven geometric isomers of [Co(dien)(ampy)Cl]2+
.
These systems were originally designed to probe the sub-
tler aspects [5] of the base hydrolysis mechanism. While ibn
and ampy are stereochemically similar, there are some sig-
nificant differences. First, the two ‘ends’ of the diamine are
different in the sense that in ampy a terminal NH2 is
blocked by a pyridyl, which removes a potential deproto-
nation site for base hydrolysis. If this were the reactive site
for base catalysed hydrolysis, then two of the four mer iso-
mers would be much more reactive than the other two, if,
for example, deprotonation at –NH2 on the bidentate
ligand trans (or cis) to Clꢀ was generally preferred. Fur-
ther, the ampy chelate ring is flatter than for ibn. Finally,
the stereochemistry of the pyridyl substituted material is
such that the two possible paths for the re-entering group
Y in the favoured trigonal bipyramidal intermediate in this
reaction are very much different (Fig. 2) compared to ibn
(the cis or trans isomers are defined by the position of the
pyridyl or C(CH3)2 group relative to the sec-NH).
The H6 proton blocks one of the pathways for the re-
entering group, leading to a cis isomer preference. On
the other hand, from the viewpoint of relative stability,
the trans isomer might be expected to be the major prod-
uct because there is less steric interaction between H6
and the sec-NH centre. Thus, the cis isomer(s) would
be predicted to be dominant at the beginning of base
hydrolysis (kinetically preferred), with subsequent rear-
rangement to the more stable trans isomers. Finally, if
each of the four mer isomers gave the same kinetic prod-
uct distribution, this would be additional and strong evi-
dence for the suggested intermediate type, which is the
only one for which the cis/trans and syn/anti distinctions
are removed.
Herein, we report a comprehensive study of the base
catalysed substitution reactions for the [Co(dien)(ibn)Cl]2+
and [Co(dien)(ampy)Cl]2+ systems, including water and
azide capture and isomerization reactions of intermediate
hydroxo and azido products.
2. Experimental
All chemicals were analytical or an equivalent grade. 13
C
and 1H NMR spectra were recorded on a Varian XL
300 MHz instrument at 20 ꢁC. Solvents used were D2O
1
with dioxane as an internal reference (13C, d 69.27; H, d
3.75 ppm relative to DSS), and Me2SO-d6 with the central
peak of the CD3 septet as the reference (13C, d 39.37; 1H, d
2.50 ppm relative to SiMe4). UV–Vis absorption spectra
(600–300 nm) and absorbance–time traces were obtained
with use of a HP 8452 A Diode-Array spectrophotometer
thermostated to 25.0 0.1 ꢁC, with Specfitꢂ software
to extract and analyse the kinetic data. The data were
Fig. 2. Decay paths for the favoured five-coordinate trigonal bipyramidal
intermediate in the mer-[Co(dien)(ampy)X]n+ system.

2212
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
routinely checked to ensure absorbances in the wavelength
range chosen for analysis did not exceed 1.5, and that the
data covered at least 2 t1/2 of reaction (more usually 3–
4 t1/2). All kinetic runs were performed in triplicate, using
[Co] in the range 0.001–0.01 M to achieve the desired
absorbance change.
2.3. Product distributions
A specific isomer of [Co(dien)(ibn)X]Cl2 or [Co(dien)-
(ampy)X]Cl2 (X = Cl, OH2 or N3; ca. 50 mg) was dissolved
in D2O (0.2 g), and an equal volume of 1.0 M NaOD
(Pipetmanꢂ) was added while stirring rapidly (0 or
20 ꢁC). The reactions were quenched with cold HClO4
(70%; 0.15 mL). For the reactions at 20 ꢁC, a series of
experiments were performed, at specific reaction times (0–
60 s), and the product distribution data in all cases were
evaluated by 13C NMR spectroscopy. A separate experi-
ment was performed for each reaction time, and involved
a large number of determinations (each in triplicate).
Because there were four isomeric products, in situ NMR
analysis ( 5% for each component) was superior to chro-
matographic separation and analysis for the individual spe-
cies (of the aqua ions, or chloro ions reformed by heating in
HCl (shown to be retentive)). In situ spectrophotometric
analysis is quite inaccurate where there are four compo-
nents and the UV–Vis spectra are quite similar.
2.1. Buffer solutions
The buffer solutions were made up with TES
(0.200 M) (TES = N-[tris(hydroxymethyl)]-2-aminoethane-
sulfonic acid), partly neutralised by addition of Convol
NaOH (1.000 M) and of total ionic strength 1.0 M
(NaCl); the ratio of [TES]/[TESꢀ] ranged from 9/1 to
1/9 to cover about 2 pH units. NaCl was chosen as
the supporting electrolyte in place of the more usual
NaClO4, one for reasons of complex solubility, and
two, because Clꢀ is a poorer competitor than ClO4 in
ꢀ
base catalysed substitution reactions [6,7]. No problems
of reversibility of the hydrolysis reactions were observed
for 1 M Clꢀ. p[H] determinations were made as described
previously [8].
The product distributions for the base hydrolysis of the
fac-chloro and mer-azido species were carried out similarly
(20 ꢁC), as were the experiments on the four mer-chloro
and aqua ions reacted in unbuffered 1 M NaN3 (20 ꢁC).
The NMR tubes from all experiments were retained (in
the dark) for periods sufficient to observe the equilibrium
distributions.
2.2. Synthesis, separation and characterisation of the
[Co(dien)(ibn)X]2+ and [Co(dien)(ampy)X]2+ (X = Cl,
OH2 and N3) isomers
The chloro and azido complexes were completely
described in previous publications [1,2]. The methods used
to synthesise specific stereoisomers of the aqua species
depended on the particular isomer required.
2.4. X-ray structural determination
X-ray diffraction data were collected on an orange crys-
tal of [Co(dien)(ampy)OP(OH)3]Cl3 Æ 2H2O using a Rigaku
AFC-6S diffractometer.
The structure was solved by direct methods (SIR-92) [9]
and refined using the teXsan computer software package
[10]. Difference electron density maps revealed H atoms
Method 1: A solution of Hg(trif)2 (5 mL, ca. 1.5 M; from
yellow HgO in excess aqueous CF3SO3H) was added to one
of the [Co(dien)(diamine)Cl]Cl2 isomers (0.5 g) dissolved in
water (1 mL). The solution was left at ambient temperature
until the colour of the solution changed to orange-yellow,
then cooled in an ice-bath. Concentrated HCl was added
just until the precipitated HgCl2 redissolved, then poured
into 500 mL of acetone. The chloride salt of the aqua spe-
cies was collected by filtration, and recrystallised from
water/HCl by careful dilution with acetone. This method
proved suitable for the facial (f1, f2, f3) and syn–mer
(m2, m4) isomers.
Method 2: This method was convenient for the synthesis
of chloride free mer-aqua species. The pure perchlorate salt
of the appropriate mer-azido isomer (0.10 g) was dissolved
or suspended in 3 M HClO4 or triflic acid (5 mL), cooled in
an ice-bath, and sodium nitrite (0.050 g) was added to the
solution in small proportions. After the evolution of gas
had ceased, the solution was degassed (ultrasonic bath)
and then loaded on to a Biorex MSZ 50 column (Na+
form) and then washed with water. Elution with 0.2 M
NaClO4/0.05 M HClO4 yielded an orange band. The eluate
was reduced in vacuo at <35 ꢁC to a small volume where-
upon the perchlorate salt crystallised. The crystals were
collected, washed with isopropanol and ether, and air
dried.
Table 1
Crystal data for mer–anti-[Co(dien)(ampy)OP(OH)3]Cl3 Æ 2H2O
Formula
M
C10H24Cl3CoN5O4P Æ 2H2O
510.63
triclinic
Crystal system
Space group
ꢀ
P1
˚
a (A)
8.061(3)
10.405(2)
12.739(2)
95.48(2)
93.90(2)
100.01(2)
1043.5(5)
2
296.2
0.71073
1.625
13.18
3032 (I > 3r(I))
257
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
V (A )
Z
T (K)
˚
k(Mo Ka) (A)
Dcalc (g cmꢀ3
l (cmꢀ1
)
)
Number of observed/unique data
Number of refined parameters
R
0.030
0.039
1.70
Rw
S

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2213
on O(2), O(3) and O(4) and on the O atoms of the water
molecules. These H atoms were included in the structure
and their coordinates were refined. Restraints were
imposed on bond distances and angles for the water mole-
cules. H atoms bonded to N and C atoms were included at
calculated position and were not refined. Crystal and
refinement details are given in Table 1.
the base hydrolysis of the chloro complex. The latter was
simply too fast to measure under the conditions. Thus,
the curvature first seen at higher [OHꢀ] was an artefact –
the measured rate constants were not for the same process.
Fig. 3 shows the rate plot for the correct data.
While the data are not especially accurate, the trends are
clear. We do not regard the (small) intercepts, which are
the ks values, as particularly accurate.
3. Results and discussion
This analysis emphasises how one can be easily misled
when dealing with multi-isomer systems and where one
needs to be aware of kinetic possibilities that can mask
the real facts. As noted previously [13], the kinetic behav-
iour of all reactant and product isomers need be known
to be confident that the intrusion of a ‘missing’ isomer
has not skewed the analysis.
The kinetics for the fac isomers was not subject to the
problems just outlined because the reactions are much
slower and the subsequent (fast) mutarotation of mer-OH
is easily separated out in the data analysis. Also, the
products are largely the fac-OH species that are relatively
unreactive ions. Plots of kobsd versus [OHꢀ] are linear
(Fig. 4).
3.1. Kinetics of base hydrolysis of the [Co(dien)(ibn)Cl]2+
isomers
Preliminary plots of kobsd versus [OHꢀ] (0–5 · 10ꢀ6 M)
for the four mer-isomers were markedly curved (not
shown), and each apparently follows an equation of the
form kobsd = k1[OHꢀ]/(1 + K[OHꢀ]). The origin of the cur-
vature for the mer plots was initially unclear. In this [OHꢀ]
range, net deprotonation, which would accommodate the
convex curvature, is not expected. A closer examination
of the spectral data where there were maximal spectral
changes (300–360 nm) revealed clean isosbestic points and
uniphasic behaviour (to ca. 3 t1/2) for the first few buffers
at the lower pH end, consistent with simple [Co(dien)-
(ibn)Cl]2+ hydrolysis to [Co(dien)(ibn)OH]2+. The specific
rates were essentially linear in [OHꢀ] in this region. Yet
around pH 7, biphasic kinetics were clearly evident; a single
exponential did not fit the data well, and there were no
isosbestic points (‘‘shifting isosbestics’’ are observed). A
reasonable fit to the data could be obtained using just the
first two half lives of absorbance–time traces, or fitting
the data to a consecutive first-order reaction scheme. It is
noted, however, that extracting reliable rate constants for
consecutive reactions where the component rate parame-
ters are within a factor of two is intrinsically difficult
[11,12]. Using Specfitꢂ the slow-fast ambiguity was
resolved by calculating the spectra for the first formed spe-
cies. Only one scheme (fast-slow) yielded a reasonable spec-
trum, the other (slow-fast) gave a nonsensical one (negative
absorbances).
The rate constant for the second step(s) (isomerisation
of the m-OH2+ species, studied independently – see ahead),
while not accurately defined in these experiments, proved
to be approximately constant (2 · 10ꢀ3 sꢀ1, 25 ꢁC), except
at lower pH where its value was diminished. This observa-
tion is consistent with the expected pH independence of the
isomerisation rates for m-OH2+, at pH values above the
pKa of its aqua conjugate acid (pKa ca. 6). The reduction
in rate at ca. pH 6 is therefore consistent with its partial
conversion into the relatively unreactive aqua complex.
The subsequent isomerisation process is not single phasic
– all four mer isomers are involved, since base hydrolysis
of any of the mer-chloro ions initially produces all four
mer-hydroxo ions.
3.2. Kinetics of base hydrolysis of the
[Co(dien)(ampy)Cl]2+ isomers
Again the problem that arose with the mer isomer kinet-
ics in the [Co(dien)(ibn)Cl]2+ system was encountered.
Biphasic kinetics were observed, arising from competitive
rearrangement of the product hydroxo species. At higher
pH, the rearrangement is the slower step, and the primary
hydrolysis is too fast to determine accurately. To resolve
the issue, the data were restricted to the lower pH region;
plots of kobsd versus [OHꢀ] for the four mer-isomers are lin-
ear (Fig. 5). Plots of kobsd versus [OHꢀ] for the three facial
isomers are also linear (Fig. 5).
350
m1
m2
m3
m4
slope = 1.18
slope = 4.74
slope = 3.99
slope = 0.87
300
250
200
150
100
50
0
0
10
20
30
40
50
60
[OH-] X 108/ M
Fig. 3. Plot of observed rate constants vs. [OHꢀ] for the base hydrolysis of
the four mer isomers of [Co(dien)(ibn)Cl]2+ in TES buffers (I = 1 M,
NaCl) at 25 ꢁC. The slopes are kOH values (Mꢀ1 sꢀ1), the intercepts ks
values (sꢀ1).
At the highest pH studied, the kinetics were again uni-
phasic, and first order. We discovered, however, that the
measured rate was that for the second step, and not for

2214
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
250
200
150
100
50
The hydrolysis rates for the m1 and m2 isomers were too
slope =0.33
slope =0.17
slope =0.44
fac1
fac2
fac3
fast to determine over the complete pH range shown; we
measured the rates at two low [OHꢀ] values, 9 · 10ꢀ8 and
1.45 · 10ꢀ7 M, and also in acid where the rates were
immeasurable slow compared to very dilute [OHꢀ]. The
average results were kOH = 8.9 · 106 for the m1 and
8.0 · 106 Mꢀ1 sꢀ1 for the m2 form.
3.3. Reactivity comparisons
The rate data for the ibn, ampy and en systems are sum-
marised in Table 2.
0
The mer–cis isomers in the ampy system are 5- to 10-fold
more reactive than the mer–trans isomers, but this is not
apparent for the corresponding ibn system. In the cis iso-
mers there appears to be a steric interaction of the pyridyl
of ampy with the sec-NH of dien, and movement to the
preferred trigonal intermediate (vide infra) does alleviate
this interaction, and this may reflected in the base hydroly-
sis rates to a mild extent. The syn/anti rate differences are
also relatively small (up to a factor of 5), and inconsistent
– the syn can be more or less reactive than the anti form
[14]. The high reactivities of the ampy complexes are typi-
cal for mer-[Co(dien)(diamine)Cl]2+ complexes [15–19].
Also, switching the ampy ligand to alternately block depro-
tonation sites cis and trans to Co–Cl, has no profound
effect. This result is entirely consistent with the finding
(ahead) that the effective deprotonation site is not on the
diamine but rather on the dien triamine. The intercepts in
the plots shown in Fig. 5, which reflect the ‘acid’ hydrolysis
rates, i.e., the rates of aquation (ks), are indistinguishable
from zero; they are negligible compared to the rates for
the base catalysed path.
0
100
200
300
400
500
[OH-] X108/M
Fig. 4. Plot of observed rate constants vs. [OHꢀ] for the base hydrolysis of
the three fac isomers of [Co(dien)(ibn)Cl]2+ in TES buffers (I = 1 M,
NaCl) at 25 ꢁC.
There is a 1000-fold difference between the reactivity of
the mer and facial isomers, for all these diamine systems
[14,20–22]. The consistent difference indicates that for the
fac isomers the effective deprotonation site also resides on
the dien ligand, but clearly it is very much less effective.
3.4. Base catalysed substitution in [Co(dien)(ibn)Cl]2+
At high [OHꢀ] (0.5 M NaOD/D2O), the hydrolysis of all
of the isomers of [Co(dien)(ibn)Cl]2+ is complete within
Table 2
Base hydrolysis rate data for the [Co(dien)(diamine)Cl]2+ systems; I = 1 M
NaCl, 25 ꢁC (TES buffers)
Reactant isomer
kOH (Mꢀ1 sꢀ1
)
ibn
ampy
ena
7800
10500
7800
10500
27
m1
m2
m3
m4
f1
11800
47400
40000
8700
330
890000
800000
48000
24000
230
f2
f3
170
440
110
110
0.7
27
a
The m1 and m3 (anti) forms are not distinguished for symmetrical
Fig. 5. Kinetic results for two of the four mer and the three facial isomers
of [Co(dien)(ampy)Cl]2+ in TES buffers (I = 1 M, NaCl) at 25 ꢁC.
diamines such as en, nor are m2 and m4.

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2215
seconds. Control experiments were carried out on pure
mer- and fac-hydroxo ions reacted under the same condi-
tions, to ascertain the relative rates of subsequent isomeri-
sation (vide infra). For the mer-chloro reactants, two sets
of experiments were performed, one set at 0 ꢁC (1 s
quench), and another at ambient temperature, at intervals
covering a complete reaction time of 60–80 s.
Table 3 shows the kinetic product distributions for the
base hydrolysis of all isomers in each system (ibn, ampy, en).
Table 4 details the cis/trans and syn/anti product ratios
for the mer isomers of [Co(dien)(ibn)Cl]2+. There are no
fac-OH products, but rather a mixture of the m-OH spe-
cies. These are the kinetic distributions; the equilibrium
values are quite different (Table 3). Further, subsequent
isomerization of the first formed hydroxo ions is very much
slower than base hydrolysis of the chloro substrates, at
higher concentrations of OHꢀ. Further, in experiments
where the base hydrolysis of a chloro isomer was acid
quenched prior to completion, recovered reactant proved
to be a single unchanged isomer. Since two of the four
mer-OH products have their sec-NH centre inverted, it is
clear that N-inversion occurs during the act of base hydro-
lysis, thus identifying the site of deprotonation. The second
significant result is that, within experimental error ( 5%),
the product distributions are identical for each of the four
mer substrates (Table 5). The clear implication is that a
common reduced coordination number intermediate is
involved, and the only conceivable structure common to
all four mer substrates is the one shown (Fig. 6), in which
the cis/trans and syn/anti distinctions are lost.
A species deprotonated at a site other than the sec-NH,
for example, can give rise to only two products, not four,
Table 5
Kinetic and equilibrium [Co(dien)(diamine)N3]2+ product distribution
data for the base catalysed substitution of mer-[Co(dien)(diamine)X]2+
(diamine = ibn, ampy, en) in unbuffered 1 M NaN3, 20 ꢁC
Diamine
Reactant
mer1
(%)
mer2
(%)
mer3
(%)
mer4
(%)
syn
(%)
cis
(%)
Kinetic
ibn
Table 3
mer1-Cl
mer2-Cl
mer3-Cl
mer4-Cl
22
17
16
26
27
27
27
24
19
12
15
20
33
44
32
30
60
61
59
54
49
44
43
50
Steric course of base catalysed substitution for individual [Co(dien)(dia-
mine)Cl]2+ isomers
Ligand
system
Reactant
isomera,b
Product distribution (%)
m1
m2
m3
m4
f1
f2
f3
ampy
en
mer1-Cl
mer2-Cl
mer3-Cl
mer4-Cl
42
45
50
50
50
47
41
40
0
0
0
0
8
8
9
58
55
50
50
92
92
91
90
ibn
m1
m2
m3
m4
f1
14
15
16
17
40
0
36
30
34
32
0
16
15
16
15
0
44
40
34
36
0
0
0
0
0
49
24
0
0
0
0
0
0
0
0
10
0
mer1/3-Cl
mer2/4-Cl
45a
45a
55b
55b
55
55
11
76
11
f2
f3
0
0
0
0
49
40
Equilibrium
ibn
ampy
m1
m2
m3
m4
f1
71
80
79
77
8
0
0
0
0
0
0
23
14
17
17
0
6
6
4
8
0
0
0
0
0
0
80
67
0
0
0
0
12
33
0
0
0
0
0
mer1-Cl
mer2-Cl
mer3-Cl
mer4-Cl
27
29
29
21
9
8
8
3
37
39
44
61
27
26
16
15
36
34
24
18
36
37
37
24
ampy
mer1-Cl
mer2-Cl
mer3-Cl
mer4-Cl
mer3-OH
81
83
82
82
82
5
3
4
5
5
0
0
0
0
0
14
14
14
14
13
19
17
18
19
18
86
86
86
87
87
f2
f3
0
0
0
88
12
en
m1/3c
m2/4c
f1/3c
f2d
50c
50c
en
mer-Cl
87a
13b
13
60c
40c
a
anti isomer.
syn isomer.
a
mer reactants, 0 ꢁC.
fac reactants, 20 ꢁC.
b
b
c
m1 and m3 (mer–anti) isomers are not distinguished for the symmet-
rical bidentate en, nor are m2 and m4 (mer–syn), nor f1 and f3 (unsym-fac;
p in the old nomenclature).
syn and anti-trans (m1 and m2)
d
sym-fac (x in the old nomenclature).
NH2
H2
N
CH3
CH3
Y
Table 4
Kinetic product ratios for the base hydrolysis of [Co(dien)(ibn)Cl]2+ at
N
Co
0 ꢁC
N
Product ratios:
Precursor:
trans/cis
anti/syn
H2
Y
m3/m1
m4/m2
m1/m2
m3/m4
NH2
m1-Cl
m2-Cl
m3-Cl
m4-Cl
1.1
1.0
1.0
0.9
1.4
1.3
1.0
1.1
0.4
0.5
0.5
0.5
0.4
0.4
0.5
0.4
syn and anti-cis (m3 and m4)
Fig. 6. The proposed pentacoordinate intermediate for the base catalysed
substitution of mer-[Co(dien)(ibn)X]n+ complexes.

2216
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
and the sec-NH configuration would be retained, contrary
to observation.
not 1:1, and they also depend upon the nucleophile
(Table 6). However, C(CH3)2 versus CH2 in the
backbone of the chelate does not provide a large differ-
ence, with the kinetically preferred total cis product
being close to 50% for both water and azide as nucleo-
philes (for attack from the side away from the gem
dimethyl group). At equiꢀlibrium, there is a tendency to
more trans, for both N3 and OHꢀ as ligands, but it
is not pronounced.
In 1 M NaN3, base hydrolysis of the mer-chloro com-
plexes is accompanied by the production of not only four
mer-hydroxo ions but also some (ca. 30%) azido com-
plexes, also as a mixture of four mer isomeric forms (Table
5). In the pH region studied (ca. 8–9), the azido ions are
considerably more robust than their corresponding
hydroxo ions, and the latter could be observed to anate
and isomerise to yield ultimately all azido complex, without
affecting the kinetic distribution of mer-azido isomers. No
attempt was made to separate out the contributions from
these initial and secondary reactions, both of which are
must faster than the ultimate equilibration via isomeriza-
tion and N-inversion in the azido products. Also, the equi-
librium azido complex values are quite different to the
initial isomer proportions, as they are also for the hydroxo
distributions (Table 6).
The observed azido distribution is both a kinetic one
and a composite, representing the distributions for
hydrolysis of the chloro ion and the subsequent azide
anation of first formed hydroxo ions. However, the prod-
uct distributions derived from all four mer-Cl isomers,
and the hydroxo ions, are identical, reinforcing the con-
cept of a common pentacoordinate intermediate in which
the cis/trans and syn/anti distinctions are lost (Fig. 6). It
is noted that attack in the trigonal plane is different for
the two entry sites, and the product cis/trans ratios are
It is important to recall [23] that the formation of the
pentacoordinate intermediate via deprotonation at the
sec-NH followed by loss of the leaving group will be mir-
rored in the decay path to products, i.e., attack by the
entering group ðH2O or N3ꢀÞ followed by reprotonation.
Given this sequence of events, it can be seen that both
the cis/trans and syn/anti kinetic distributions are deter-
mined in the first of these two product determining steps.
The final reprotonation merely locks in the now reformed
pyramidal configuration at the sec-N centre. Thus, the first
step of nucleophilic attack in the trigonal plane might be
anticipated to be fairly indiscriminate given the ca. 120ꢁ
angle as opposed to the 90ꢁ angle of the octahedron.
Indeed, this is the case for the syn/anti as well as the cis/
trans distributions which are all close to 50:50. Contrast
the equilibrium distributions where the anti forms are
favoured for the octahedral ions.
In strong OHꢀ the azido isomers completely base
hydrolyse, and we have measured both the kinetic and
Table 6
Summary of base hydrolysis kineticc and equilibriumd product distribution data for mer-[Co(dien)(diamine)X]2+ (diamine = ibn, ampy, en)
Diamine
Reactant
Nucleophile
mer1 (%)
mer2 (%)
mer3 (%)
mer4 (%)
syn (%)
cis (%)
Kinetic
ibn
mer-Cl
mer-N3
mer-Cl
H2O
15.5
16
20
33
29
27
15.5
20
17
36
35
36
69
64
62
48.5
45
46
H2Oe
ꢀ
N3
ampy
en
mer-Cl
mer-Cl
mer-OH
H2O
76
46.5
45
0
44.5
47
18
0
0
6
9
8
6
53.5
55
76
91
92
ꢀ
N3ꢀ
N3
mer-Cl
mer-Cl
mer-OH
H2O
46a,b
55a,b
58a,b
54a,b
45a,b
42a,b
54
45
42
ꢀ
N3ꢀ
N3
Equilibrium
ibn
mer-Cl
mer-N3
mer-Cl
H2O
H2O
N3
29
35
26
3
6
7
59
49
46
9
10
21
12
16
28
32
41
33
ꢀ
ampy
en
mer-Cl
mer-Cl
mer-OH
H2O
8
82
82
0
4
5
80.5
0
0
11.5
14
13
11.5
18
18
8
86
87
ꢀ
N3ꢀ
N3
mer-Cl
mer-OH
mer-Cl
H2O
H2O
87a,b
87a,b
89a,b
89a,b
13a,b
13a,b
11a,b
11a,b
13
13
11
11
ꢀ
N3ꢀ
mer-OH
N3
a
The m1 and m3 (anti) isomers are not distinguished for the symmetrical diamine en, nor are the m2 and m4 (syn) isomers.
Ref. [25].
Reaction time 40 min, 0 ꢁC.
Reaction time 2 days, 20 ꢁC.
20 ꢁC.
b
c
d
e

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2217
equilibrium distributions for the hydroxo products, as
described for the chloro reactants. The results are again
identical within experimental error ( 5%), despite the big
difference in the reactivity of the chloro and azido complexes
NH2
Co
NH2
Co
H2
N
H
CH3
CH3
H2
N
N
CH3
CH3
N
OH-
(Table 6). In the [Co(NH3)5X]n+ system, kOH for X = Clꢀ is
Cl
N
ꢀ
almost 800-fold greater [19] than that for X ¼ N3
.
H2
Cl
N
H2
In summary, common kinetic distributions, independent
NH2
NH2
of the starting mer geometry and also the leaving group
(Clꢀ, N3 or H2O) argue for a common intermediate of
ꢀ
the type shown (Fig. 6). The hydroxo ions in these reactive
mer systems substitute via internal conjugate base process
established previously for the dapo [24] and en [5] systems
and involving a reactive substrate of the type shown, anal-
ogous to that derived from the chloro reactants (Fig. 7).
Note that there is no requirement for the NH to be syn
to the coordinated OHꢀ to get internal proton transfer here
since proton transfer is not rate determining.
NH2
NH2
Co
H
H2
CH3
CH3
H2
N
CH3
CH3
N
N
N
Co
HO
N
H2O
N
H2
H2
NH2
NH2
Fig. 7. The top diagram shows the normal deprotonation step in base
catalysed hydrolysis, which requires external OHꢀ and gives rise to the
usual second-order kinetics. The lower one shows the internal proton
Finally, it should be noted that the common ‘‘equilib-
rium’’ distributions do not necessarily imply that the azido
systems are at equilibrium (amongst the mer isomers), since
a common result would flow from the common initial dis-
tributions, if the reaction times were the same (and they
were). This aspect has been checked by remeasuring some
of the systems at much longer reaction times. We found
the same ratio for the four mer isomers, but together with
some of the fac isomers. It is concluded that the equilib-
rium distributions shown in Tables 5 and 6 are correct.
transfer for
a
hydroxo complex, requiring no external OHꢀ, but
generating an identical reactive species, differing only in the leaving group.
even under equilibrium conditions (Table 6). Secondly,
and the major point, the kinetic distributions are indepen-
dent of source, again consistent with an intermediate which
has lost not only Clꢀ but also memory of its origin (syn/
anti and cis/trans) (Fig. 8).
The common product distribution 77%, 0%, 17%, 6%
(m1, 2, 3, 4) observed for the entry of water is quite differ-
ent to the equilibrium distribution – 8%, 0%, 80%, 12%.
For the four mer-Cl isomers reacted in unbuffered NaN3
(1 M, 20 ꢁC), the corresponding mer-azido product distri-
butions were 47, 44, 0, 9 and 82, 4, 0, 14, again independent
of the starting geometry, but results different to water as a
3.5. Base catalysed substitution in [Co(dien)(ampy)Cl]2+
It was anticipated that the major kinetically preferred
products would be largely cis (m1 and/or m2), whereas
the equilibrium distribution should be largely trans. This
is borne out by the results (Tables 2 and 6) for water as
the nucleophile, but not for azide which prefers to be cis
H2N
H2N
N
H2N
Cl
N
Co
Co
H
H
N
N
N
Cl
H2
H2N
H2N
H2N
N
H2N
Co
N
H2N
H2N
H2N
H2N
Cl
N
N
Co
Cl
N
Co
H
N
N
H2N
H2
H
H2N
Fig. 8. The symmetrical p-bonded TBP intermediate derived from the four different mer substrates.

2218
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
nucleophile. No anti–mer–trans isomer (m3) was observed
initially or at equilibrium, whereas the m2 isomer, unob-
served for water as a nucleophile, was now a major kinetic
product. These features were exploited synthetically to pre-
pare the putative m2-Cl and m4-Cl complexes.
Figs. 9 and 10 show the initial and final azido isomer dis-
tributions for three different mer reactants. The azido prod-
ucts are normalised to 100% to show the agreement
between the initial and final values, and the same rates of
decay, independent of source (Clꢀ or OHꢀ as leaving
group, m1 or m3 configuration). The decay of the starting
material is also shown, to indicate the relative rates in the
three systems, and to indicate the difficulty of accurately
defining the product distributions under conditions where
the substrate has not completely decayed before the prod-
ucts have started to rearrange.
3.5.1. Rearrangement of m2-[Co(dien)(ampy)OH2]3+ in
aqueous acid
In the base hydrolysis studies, the m2-OH species (syn)
was not observed, even under equilibrium conditions,
although the expectation was the usual syn (endo) and anti
(exo) isomers first-formed in comparable amounts. We
have also found that the m2- (and m1-) chloro ions
hydrolyse in acid remarkably quickly. We therefore sus-
pected that the m2-OH ion is unusually labile, and isome-
rises to the more stable m1-form to escape detection (the
amount of first-formed m1 isomer is anomalously high).
This can occur via an internal conjugate base process
[24], whereby N-inversion occurs via the hydroxo complex
by transferring an NH proton to generate water as the
leaving group (Fig. 7), or it can occur by direct
N-inversion.
The azido products at very early reaction times were
quite dominant, with little competitive hydrolysis (based
on analysis by NMR), indicating in crudeꢀterms the capture
of >50% total azido product for 1 M N3 (and thus <50%
hydroxo species). This level of anion competition is not
inconsistent with studies for the simpler systems such as
mer-[Co(dien)(en)Cl]2+ where the actual competition val-
ues have been determined [25] and are of the order of
50% azide capture for 1 M NaN3.
The most stable hydroxo complex is the m3 isomer,
whereas for the azido species it is the m1 isomer; each of
these has the favoured anti-NH form. These results show
clearly that the nature of the sixth ligand X in mer-[Co(dien)-
(ampy)X] can significantly affects the relative energies of
the isomers. This has been already observed for the phos-
phate complexes (see 3.7 ahead) where we argue a special
H-bonding effect for the m1 isomer, but it is clear that simply
changing the ligand can change the preferred configuration.
We can see no particular properties of bound hydroxide or
azide ion that brings about these configurational prefer-
ences. Finally, it is noted that the preferred sec-NH configu-
ration for the trans azido species is exclusively endo (m4).
This is unusual and as yet is not understood.
Fig. 10. Evolving [Co(dien)(ampy)N3]2+ product distributions for the m3-
OH2 reactant in 1.0 M NaN3 at 20 ꢁC (note that the reactant is actually
the hydroxo complex, albeit the leaving group is water – see text). The
vertical axis is % total complex.
Fig. 9. Evolving [Co(dien)(ampy)N3]2+ product distributions for the m1-Cl (left) and m3-Cl (right) reactants in 1.0 M NaN3 at 20 ꢁC. The vertical axes are
% total complex.

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2219
The reaction of the m2-aqua isomer in 3 M triflic acid
was followed at two temperatures, and the clean formation
of the m1 isomer was observed (Fig. 11). This reaction
requires N-inversion which is suppressed in acid solution,
but it is noted that the alternative process of ligand folding
[4] leads to an apparently inverted N-centre (endo to exo)
but a different mer product (m3) (Fig. 12). Also, the latter
process is usually very very slow.
cess, but one that also involves N-inversion. In summary,
it would seem that the m2-aqua isomer isomerises via base
catalysed N-inversion, likely via the labile m2-OH
complex.
3.6. Rearrangement of the mer-hydroxo isomers and anti/syn
isomerisation
On repeating the experiments in 0.1 M triflic acid, a
much faster (ca. 20-fold) m2 to m1 rearrangement was
observed, and subsequent events, m1 to m4 and m3, with
the m3 aqua ion predominant (80%) at equilibrium. We
also observed the m4-aqua to m3-aqua isomerization in
3 M triflic acid, a slower reaction than the m2 to m1 pro-
There are two distinct processes – end to end (cis/trans)
isomerisation, and syn–anti isomerisation. Either or both
can occur via substitution, but syn–anti isomerisation can
also occur by base catalysed N-inversion without ligand
substitution. The following results for the syn/anti and
cis/trans hydroxo ion distributions are averages over the
four mer isomers, from two sources (four mer azido and
four mer chloro ions).
Fig. 13 shows the time dependence of the [Co(dien)-
(ibn)(OH)]2+ product distributions at about 20 ꢁC. The
distribution changes rapidly with time, and a common
equilibrium distribution is attained which is quite different
to the common kinetic distribution. These results also sug-
gest that the isomerisations are not single phase exponen-
tial processes, but rather sequential processes, but the
experimental error of 5% for each isomer obscures the
issue.
The corresponding distribution derived from the azido
complexes as reactants is shown in Fig. 14. The middle
stages are slightly different to those shown in Fig. 13 and
probably reflect the greater error here because the primary
hydrolysis and subsequent isomerisation steps are over-
lapped. The hydroxo complex distributions are normalized
to 100%, and form of the decay curves for the azido ions
are not shown.
Fig. 11. Time dependence of the distribution of isomers derived from m2-
[Co(dien)(ampy)OH2]3+ in 3 M aqueous triflic acid at 20 and 30 ꢁC as a
function of time. The vertical axis is % total Co.
60
m1
50
m2
m3
m4
H
N
N
H
NH2
NH2
N
40
30
20
10
0
H2O
Co
N
Co
H2O
NH2
NH2
N-inversion
NH2
NH2
m1
m2
ligand
folding
OH2
NH2
N
Co
N
H
NH2
NH2
0
10
20
30
40
50
60
70
m3
time, sec
Fig. 12. Alternative reaction modes for m2-[Co(dien)(ampy)OH2]3+ in
aqueous acid.
Fig. 13. Time-dependent mer-[Co(dien)(ibn)OH]2+ distributions following
the base hydrolysis of [Co(dien)(ibn)Cl]2+; 0.5 M NaOH, ca. 20 ꢁC.

2220
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
100
60
50
40
30
20
10
0
80
60
m1
40
20
0
m2
m3
m4
0
10
20
30
40
50
60
0
10
20
30
40
50
60
time /s
time (s)
Fig. 16. Cis/trans product [Co(dien)(ibn)OH]2+ distributions for the base
hydrolysis of the mer-[Co(dien)(ibn)Cl]2+ and mer-[Co(dien)(ibn)N3]2+
ions.
Fig. 14. The distribution of isomeric mer-hydroxo products for the base
hydrolysis of the mer-[Co(dien)(ibn)N3]2+ isomers at ca. 20 ꢁC.
The anti/syn isomer ratio for the [Co(dien)(ibn)(OH)]2+
products increases smoothly with time (Fig. 15). The isom-
erization reactions derived from the azido substrates may
be adjudged somewhat faster, but the initial and final val-
ues agree. The apparent rate difference probably reflects the
lack of good temperature control for the different product
distribution experiments; rates, however, were not the issue
here.
The cis/trans isomer distributions as a function of time
(Fig. 16) are less of an issue since the kinetic and equilib-
rium distributions are not greatly different.
Fig. 17 shows the initial (kinetic) product distribution
derived from the m3-[Co(dien)(ampy)Cl]2+ complex and
how the [Co(dien)(ampy)OH]2+ isomer distribution evolves
with time. The syn–mer–cis (m2) isomer was never
observed. The m1-OH isomer was the dominant initial
product (76%), but the more stable m3-OH species became
the major species in time, with 80% at equilibrium. Other
experiments (Fig. 18) were carried out on the m3-OH com-
Fig. 17. Evolving [Co(dien)(ampy)(OH)]2+ product distributions derived
from the m3-chloro complex (0.5 M NaOD, 20 ꢁC). The vertical axis is %
total Co.
100
80
60
40
20
0
plex, as reactant and under the same conditions, to ascer-
tain the relative rates of subsequent epimerization from
these curves it is clear that base hydrolysis of the chloro
reactant is much faster than the subsequent isomerization
of the hydroxo complexes, as found for the ibn system.
Also, the rates of m1-OH/m3-OH equilibration for the
two sets of experiments are the same, as they should be
since the exponentials reflect the sum of the forward and
reverse rate constants in both cases.
3.7. Formation and rearrangement of
[Co(dien)(ampy)OP(O)(OH)]+
0
10
20
30
40
50
60
time (s)
In the earlier synthetic paper [2], we could prepare the
elusive m1-Cl isomer via the equilibrated azido species, or
better, via its phosphate complex. In NaH2PO4/NaHPO4
Fig. 15. Anti/syn product [Co(dien)(ibn)OH]2+ distributions for the base
hydrolysis of the mer-chloro (lower) and mer-azido ions (upper).

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2221
ꢀ
H2PO4ꢀ, whereas at higher pH (8–9) it is largely HPO4
.
In addition, the acidity of the hydrogen phosphates will
increase on O-coordination. Sites for Hbonding involve
the terminal NH2’s of dien, common to all 4 mer isomers,
the sec-NH centre, accessible only for the m2 and m4 iso-
mers, and the terminal –NH2 centre of the ampy ligand,
accessible only for the m1 and m2 isomers. It would seem
that the best prospects exist for the m1 isomer where thꢀe
2ꢀ
ligand is HPO4 and two H-bonds can form. In H2PO4
media, more m4 is observed at equilibrium (50% m1,
20% m4), presumably because only one P–O oxygen is
now accessible to H-bond to an amine, reducing the contri-
bution from the H-bonded structure shown for the m1 iso-
mer (Fig. 19).
3.7.1. X-ray structure of m1-
[Co(dien)(ampy)OP(OH)3]Cl3 Æ 2H2O
Fig. 18. Isomerisation of m3-[Co(dien)(ampy)OH]2+ in 0.5 M NaOD at
20 ꢁC. The vertical axis is % total Co.
The complex crystallised from acid solution as the fully
protonated form, confirmed by the X-ray structure which
reveals three Clꢀ anions per phosphato cation. As far as
we are aware, this is the first reported structure for a metal
complex of phosphoric acid.
buffers (pH 7.0–9.0), phosphate readily displaces Clꢀ from
any of the chloro complexes, and on standing there is
almost a complete rearrangement to the m1 phosphato spe-
cies, Table 7 (cf. 80% m3 in the absence of phosphate;
Table 3).
Fig. 19 shows some idealised H-bonded structures for
each of the four possible mer phosphate complexes. pH
should be important in the H-bonding process in solution;
at lower pH (3–4) phosphoric acid exists largely as
The ORTEP diagram for the molecular cation is shown
in Fig. 20. Selected bond lengths and angles are presented
in Table 8. The dien rings adopt the chiral kk or dd confor-
mations, and the crystal is a racemate. The bond lengths
and angles are unremarkable. The P–O(1)Co length of
˚
1.480(2) A is appreciably shorter than the other P–OH
˚
bond lengths 1.529–1.558 A. Although not symmetry
Table 7
Product distribution for the reaction of mer-[Co(dien)(ampy)Cl]Cl2 in
NaH2PO4 (0.25 M)/NaOD(0.25 M)/D2O buffer; 50 min at 20 ꢁC
Phosphato isomer !
f1
f2
f3
m1
m2
m3
m4
m1-Cl
m2-Cl
m3-Cl
m4-Cl
93
94
92
90
5
5
6
6
2
1
2
4
H
N
H
NH2
N
NH2
O
Co
N
O
Co
N
NH2
H
NH2
H
P
N
P
HO
N
O
O
HO
O
O
H
H
m1
m2
H
O
N
N
OH
O
H
O
NH2
P
NH2
NH2
P
O
HO
Co
N
NH2
O
Co
N
O
NH2
NH2
Fig. 20. ORTEP diagram for the molecular cation m1-[Co(dien)(ampy)-
OP(OH)3]3+. Thermal displacement parameters are shown at the 20%
level.
m4
m3
Fig. 19. Proposed H-bonded phosphate structures.

2222
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
˚
Table 8
of H3PO4 and H3PO4.0.5H2O are 1.49 and 1.52 A for P–
˚
Selected bond length (A) and bond angle (ꢁ) data for m1-[Co(dien)-
(ampy)OP(OH)3]Cl3 Æ 2H2O
˚
O, 1.57 and 1.56 A for P–OH. For a range of phosphates
P–O is typically 1.51 A and P–OH 1.55–1.58 A [26].
˚
˚
Co(1)–O(1)
Co(1)–N(1)
Co(1)–N(2)
Co(1)–N(3)
Co(1)–N(4)
Co(1)–N(5)
P(1)–O(1)
P(1)–O(2)
P(1)–O(3)
P(1)–O(4)
O(2)–H(1)
O(3)–H(2)
O(4)–H(3)
1.936(2)
1.954(2)
1.954(2)
1.962(2)
1.940(2)
1.932(2)
1.480(2)
1.529(2)
1.526(2)
1.558(2)
0.77(4)
3.8. Steric course of base hydrolysis of the
fac[Co(dien)(diamine)Cl]2+ isomers
The base hydrolysis of three facial [Co(dien)(ibn)Cl]2+
isomers finished ‘instantly’ at high [ODꢀ], and the prod-
uct distributions were quite clearly dependent on the
starting fac configuration (f1, f2 or f3), Table 3. There
can be no common intermediate. The f2-Cl substrate
yields f2 and f1 products, and no mer- or f3-OH isomers.
The other two reactants, f1- and f3-Cl, yield two fac iso-
mers along with a single mer product (Table 3). The
product distributions for the f1 and f3 substrates are
very similar in the sense that the 40:49:11 ratios are
indistinguishable, save for the isomer designations, indi-
cating that the gem-dimethyl groups in either of the
two inequivalent positions on the diamine backbone have
no real influence. Overall the results parallel those for
other [Co(dien)(diamine)Cl]2+ systems [14,23], where the
data are less informative because the diamine (en, tn or
dapo) is symmetrical.
These observations can be explained using sets of
intermediates of the kind proposed for mer reactants, tri-
gonal bipyramidal, but where the dien is facially
arranged and the deprotonation site is presumed to
occupy the trigonal plane. The facial isomers are asym-
metric, and therefore all nine NH protons are inequiva-
lent. The simplifying assumption is made that for any
given facial isomer only one site is effectively deproto-
nated, and may be the same or different for each facial
isomer.
0.70(4)
0.82(3)
O(1)–Co(1)–N(1)
O(1)–Co(1)–N(2)
O(1)–Co(1)–N(3)
O(1)–Co(1)–N(4)
O(1)–Co(1)–N(5)
N(1)–Co(1)–N(2)
N(1)–Co(1)–N(3)
N(1)–Co(1)–N(4)
N(1)–Co(1)–N(5)
N(2)–Co(1)–N(3)
N(2)–Co(1)–N(4)
N(2)–Co(1)–N(5)
N(3)–Co(1)–N(4)
N(3)–Co(1)–N(5)
N(4)–Co(1)–N(5)
O(1)–P(1)–O(2)
O(1)–P(1)–O(3)
O(1)–P(1)–O(4)
O(2)–P(1)–O(3)
O(2)–P(1)–O(4)
O(3)–P(1)–O(4)
Co(1)–O(1)–P(1)
P(1)–O(2)–H(1)
P(1)–O(3)–H(2)
P(1)–O(4)–H(3)
85.07(8)
87.81(8)
91.22(8)
89.34(9)
172.79(9)
86.2(1)
171.3(1)
93.21(9)
91.87(9)
85.8(1)
177.12(9)
98.52(9)
94.66(9)
92.69(9)
84.31(9)
108.7(1)
111.7(1)
113.2(1)
109.5(1)
108.4(1)
105.2(1)
147.7(1)
115(3)
120(3)
113(3)
For the f2 reactant, the results are accommodated by a
single trigonal pyramidal (TBP) intermediate (Fig. 21).
The absence of f3 or any mer product can be understood
in terms of the intermediate(s) that would be required
(Fig. 22).
related, the two ‘endo’ P–OH bond lengths are the same
(1.526(2), 1.529(2) A), and shorter than the ‘exo’ P–OH
(1.558(2) A). Corresponding P–O bonds in the structures
˚
˚
N
N
N
Co
N
(+)-fac2
N
OH
N
N
Co
N
N
Cl
N
Co
N
+ H O
2
N
N
-
- Cl
N
N
N
Co
N
N
OH
N
(+)-fac2
(-)-fac1
N
Fig. 21. The TBP derived from the f2 isomer of [Co(dien)(ibn)Cl]2+, and the stereochemical consequences.

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2223
N
N
fac1 + fac2 + mer1
Co
N
N
N
N
N
N
N
N
Co
Cl
N
Co
N
(+)-fac2
N
N
fac3 + fac2 + mer3
N
Fig. 22. The unfavoured TBP intermediates derived from the f2 isomer of [Co(dien)(ibn)Cl]2+ by axial displacement of Clꢀ.
Considerable chelate arm movement is required to
achieve either of the configurations shown, and it
involves axial displacement of Clꢀ. If the two intermedi-
ates shown were produced, they would appear in similar
amounts, given that they differ only in the position of
the gem-methyl groups, and thus would give rise to f3
and mer products.
The rearrangements of the fac isomers provide good
examples of the edge displacement principle that is based
on minimal rearrangement of the octahedron in any one
substitution step [27,28]. Under this principle, and apart
from the retentive processes that are always possible, only
the direct isomerisations shown (Fig. 23) can occur; each
process involves movement of a terminal chelate arm to
the position occupied by the leaving group. The corre-
sponding set for the f3 isomer (not shown) mirrors that
for the f1, where inverted f3, (+)-f2, (ꢀ)-f2 and m3 are
‘allowed’ substitution products. Under these schemes,
N
N
Co
N
N
N
N
N
OH
N
Co
N
Cl
N
-
- Cl
(+)-fac2
(-)-fac1
N
N
N
N
N
N
N
N
Co
N
Co
Cl
N
-
- Cl
OH
(+)-fac2
(-)-fac2
N
N
N
N
N
N
Co
N
Co
-
- Cl
Cl
N
HO
N
(-)-fac1
mer1
N
N
N
N
N
N
N
OH
N
N
N
N
N
N
Co
N
Co
N
-
Co
N
- Cl
Co
N
N
Cl
N
Cl
N
(-)-fac1
(+)-fac1
(-)-fac1
N
Co
N
N
N
N
N
Co
N
N
N
N
Cl
N
N
Co
N
Cl
N
N
N
Co
N
OH
N
N
-
- Cl
N
(+)-fac1
(-)-fac2
Fig. 23. Possible single-edge displacement processes. The set for the f3
reactant of [Co(dien)(ibn)Cl]2+ (not shown) are generated by switching the
gem-dimethyl substituents in the bidentate backbone of the f1 isomer and
its product.
(-)-fac1
Fig. 24. Two feasible TBP intermediates derived from a f1-Cl reactant of
[Co(dien)(ibn)Cl]2+
.

2224
T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
N
N
N
N
Co
N
(-)-fac1
OH
N
Co
N
N
N
N
N
N
N
N
N
N
N
Co
-
- Cl
Co
mer1
Cl
N
HO
N
(-)-fac1
N
Co
N
OH
N
N
(+)-fac1
N
N
N
N
N
Co
(-)-fac2
N
OH
N
Co
N
N
N
Cl
N
N
N
Co
N
N
-
N
- Cl
N
N
N
OH
N
Co
N
(+)-fac1
(+)-fac1
Fig. 25. The two TBP intermediates which account for the observed steric course of substitution of [Co(dien)(ibn)Cl]2+
.
direct interconversion of f1 and f3 is not allowed, and
indeed is not observed.
racemates, and clearly the mirror image structures are
equally accessible.
Thus, far it has been assumed that water enters the
intermediate along the trigonal plane. Microscopic
reversibility requires water to be lost from the equatorial
position for the reverse process regenerating the interme-
diate. It is therefore reasonable that equatorial loss of
chloride is preferred, precluding the pathways above,
and consistent with observation. Fig. 24 shows the same
f1 reactant drawn in two perspectives to show how each
TBP intermediate is formed by loss of equatorial Clꢀ cis
to deprotonated ibn.
The stereochemical results for the fac-ampy complexes
are analogous to the ibn results, with some differences.
The f1-Cl ion gives f1-, f2- and m1-OH products, but no
f3 (nor any other m-OH isomer). There is much less m1
product than for the ibn analogue, and presumably the sec-
ond intermediate with the pyridyl remote from the trigonal
plane is favoured. Similarly, the f3-Cl isomer yields f3-OH
and m3-OH, but unlike the ibn analogue, no f2-OH. The
f2-Cl complex again gives just f2-OH and f1-OH, as in
the ibn system, but now with considerably more of the lat-
ter. The reason for this is not obvious since the pyridyl is
remote from the equatorial plane of the trigonal bipyrami-
dal intermediate, and in any event, it parallels the symmet-
rical en system (Table 3).
These intermediates account for the observed product
distributions (Fig. 25).
Some of the intermediates (Figs. 21 and 25) involve a
five-membered chelate occupying the trigonal plane. In
consideration of isomeric rearrangements in acid solution
these are excluded [11,27,29], but in base hydrolysis they
are clearly attainable (as demonstrated unequivocally for
the mer substrates). Further, chiral forms are shown in
all these figures, but all experiments were carried out on
4. Supporting information
Crystallographic data have been deposited with the
Cambridge Crystallographic Data Centre, CCDC 274124.

T. Zhu et al. / Inorganica Chimica Acta 359 (2006) 2210–2225
2225
[11] W.G. Jackson, A.M. Sargeson, Inorg. Chem. 17 (1978) 1348.
[12] W.G. Jackson, J.M. Harrowfield, P.D. Vowles, Int. J. Chem. Kinet. 9
(1977) 535.
[13] W.G. Jackson, A.M. Sargeson, Org. Chem. 42 (1980) 273.
[14] W.G. Jackson, P.B. Spencer, R.J. Walsh, Inorg. Chim. Acta 284
(1999) 37.
Copies of this information may be obtained free of charge
from The Director, CCDC, 12 Union Road, Cambridge
CB2 1EZ, UK (fax: +44 1223 336 033; e-mail: deposit@
ccdc.cam.ac.uk or www: http://www.ccdc.cam.ac.uk).
Acknowledgements
[15] R.W. Hay, D.A. House, Inorg. Chim. Acta 19 (1976) 23.
[16] T.K. Huan, J.N. Mulvihill, A.R. Gainsford, D.A. House, Inorg.
Chim. Acta 12 (1973) 517.
[17] L.S. Dong, D.A. House, Inorg. Chim. Acta 19 (1976) 23.
[18] D.A. House, A.R. Gainsford, J.W. Blunt, Inorg. Chim. Acta 57
(1982) 141.
We are grateful to the Australian Research Council for
financial support.
References
[19] M.L. TobeAdvances in Inorganic and Bioinorganic Mechanisms, vol.
2, Academic Press, New York, 1983, p. 1.
[20] T.K. Huan, J.N. Mulvihill, A.R. Gainsford, D.A. House, Inorg.
Chem. 12 (1973) 1517.
[21] R.W. Hay, K.B. Nolan, J. Inorg. Nucl. Chem. 38 (1976) 2118.
[22] F.A. Ha, D.A. House, Inorg. Chim. Acta 38 (1980) 167.
[23] P. Comba, W.G. Jackson, W. Marty, L. Zipper, Helv. Chim. Acta 75
(1992) 1147.
[1] T. Zhu, W.G. Jackson, Inorg. Chem. 42 (2003) 88.
[2] T. Zhu, W.G. Jackson, Inorg. Chim. Acta 343 (2003) 147.
[3] T. Zhu, Q.-J. Zhu, S.-F. Xie, W.G. Jackson, Z.-Y. Zhou, X.-G.
Zhou, Polyhedron 22 (2003) 1603.
[4] T. Zhu, Z. Qian-jiang, X. Sai-fieng, W.G. Jackson, Z. Zhong-Yuan,
Z. Xiang-Ge, J. Mol. Struct. 703 (2004) 83.
[24] P. Comba, W.G. Jackson, W. Marty, L. Zipper, Helv. Chim. Acta 75
(1992) 1172.
[5] W.G. Jackson, Inorg. React. Mech. (Abingdon, United Kingdom) 4
(2002) 1.
[25] D.G. Gaswick, J.A. McKeon, W.G. Jackson, results reported in Ref.
[5].
[6] W.G. Jackson, C.M. Begbie, M.L. Randall, Inorg. Chim. Acta 70
(1983) 7.
[26] A.F. Wells, Structural Inorganic Chemistry, Clarendon Press,
Oxford, 1984.
[7] J.M. Harrowfield, A.M. Sargeson, B. Singh, J.C. Sullivan, Inorg.
Chem. 14 (1975) 2864.
[27] F. Basolo, E.G. Pearson, Mechanisms of Inorganic Reactions, second
ed., Wiley, New York, 1967.
[8] A.J. Dickie, D.C.R. Hockless, A.C. Willis, J.A. McKeon, W.G.
Jackson, Inorg. Chem. 42 (2003) 3822.
[28] D.D. Brown, C.K. Ingold, R.S. Nyholm, J. Chem. Soc. (1953)
2673.
[9] SIR-92 A. Altomare, M. Cascarano, C. Giacovazzo, A. Guagliardi, J.
Appl. Crystallogr. 26 (1993) 343.
[29] W.G. Jackson, in: I. Bernal (Ed.), The Stereochemistry of the Bailar
Inversion and Related Substitution Processes, vol. 1, Elsevier,
Amsterdam, 1986, p. 255.
[10] teXsan, Single Crystal Structure Analysis Software, Version 1.7,
Molecular Structure Corporation, MSC, 3200 Research Forest Drive,
The Woodlands, TX 77381, USA, 1992–1997.