Potentiometric and spectrophotometric studies of the equilibria between silver(i) ion and crown ethers containing chromophore substituents in propylene carbonate

Article abstract of DOI:10.1039/a908585h

Potentiometric and UV-spectrophotometric methods have been used to study the coordination properties of crown ethers and their derivatives containing 9,10-dioxo-l-anthryl, 8-hydroxy-9,10-dioxo-l-anthryl and 8-tosyloxy-9,10dioxo-1-anthryl substituents with respect to the silver(i) ion in propylene carbonate. It was found that unsubstituted crown ligands were capable of forming two types of complex ion. The silver(i) ion can enter the macrocycle cavity forming 1:1 complexes or can coordinate externally to the nitrogen atom of a ligand as occurs for typical complexes of silver(i) with nitrogen ligands. This results in Ag(crown)²⁺ complexes. Introduction of chromophore groups allows investigation of the mechanism of AgL⁺ and AgL²⁺ complex formation by spectrophotometric methods. Attaching a bulky chromophore to the nitrogen atom of the aza-crown makes the formation of the AgL²⁺ complexes formonoaza-crown-15 strongly restricted or impossible. For 13-(9,10-dioxo-l-anthryl)-l,4,7,10-tetraoxa-13azacyclopentadecane (AjlSCSAQN) the presence of only one substituent leads to an increase in the stability of the external complex (AgL₂) and the formation of other complex structures Ag₂L₃²⁺ and Ag₃L₂³⁺. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a908585h

View Article Online / Journal Homepage / Table of Contents for this issue
Potentiometric and spectrophotometric studies of the equilibria
between silver(I) ion and crown ethers containing chromophore
substituents in propylene carbonate
a
a
a
a
Tadeusz Ossowski,* Jaromir Kira,* Dorota Rogowska, Hanna Warmke
b
and Janusz Młodzianowski
a
Faculty of Chemistry, University of Gda n´ sk, Sobieskiego 18, 80-952 Gda n´ sk, Poland.
E-mail: tedos@chemik.chem.univ.gda.pl; jaromir@chemik.chem.univ.gda.pl
Institute of Experimental Physics, University of Gda n´ sk, Wita Stwosza 57, 80-952 Poland
b
Received 28th October 1999, Accepted 14th January 2000
Potentiometric and UV-spectrophotometric methods have been used to study the coordination properties of crown
ethers and their derivatives containing 9,10-dioxo-1-anthryl, 8-hydroxy-9,10-dioxo-1-anthryl and 8-tosyloxy-9,10-
dioxo-1-anthryl substituents with respect to the silver() ion in propylene carbonate. It was found that unsubstituted
crown ligands were capable of forming two types of complex ion. The silver() ion can enter the macrocycle cavity
forming 1:1 complexes or can coordinate externally to the nitrogen atom of a ligand as occurs for typical complexes
ϩ
of silver() with nitrogen ligands. This results in Ag(crown) complexes. Introduction of chromophore groups
2
ϩ
ϩ
allows investigation of the mechanism of AgL and AgL2 complex formation by spectrophotometric methods.
ϩ
Attaching a bulky chromophore to the nitrogen atom of the aza-crown makes the formation of the AgL2 complexes
for monoaza-crown-15 strongly restricted or impossible. For 13-(9,10-dioxo-1-anthryl)-1,4,7,10-tetraoxa-13-
azacyclopentadecane (A 15C5AQN) the presence of only one substituent leads to an increase in the stability
2
2ϩ
3
3ϩ
of the external complex (AgL ) and the formation of other complex structures Ag L and Ag L2
.
2
2
3
Introduction
Crown ethers are efficient complexing agents for many metal
1
ions. Numerous thermodynamic and kinetic data indicate that
complexes of 1:1 stoichiometry are formed predominantly. The
formation of 1:1 metal ion–coronand complexes does not
necessarily imply that the metal ion is located within the cavity
of the macrocycle. The alkali metal thiocyanates of dibenzo-
Fig. 1 Possible structures of silver–crown complexes.
2–4
1
8-crown-6 (B 18C6) were reported by Pederson et al. to have
2
metal-to-ligand ratio of 1:1 for potassium 1:1 and 1:2 for
rubidium and 1:2 and 2:3 for caesium. A sandwich structure in
which the metal ion is located between two coronand molecules
Several recent studies have reported the complexation
11–15
properties of differently functionalized crown ethers.
In
most cases, polar side-arm groups comprise ethers, amines,
amides or carbocyclic ones. Here, we describe the synthesis and
some complexing properties of N-functionalized monoaza- and
diaza-crown ethers containing anthraquinone, hydroxyanthra-
quinone and tosyloxyanthraquinone base functionalities (Fig.
). Such groups function as spectrophotometric indicators of
complex-formation equilibrium. A number of studies of crown
ethers containing pendant-arm chromophore groups have been
5,6
was suggested as the most probable for the 1:2 complexes.
Similarly, the 2:3 complex can be visualized as being con-
structed of two metal ions between three coronand molecules
4
arranged flat in their tiers.
A similarly complex situation can be expected when
analyzing the interactions of aza-crown ethers with metal ions
that possess high affinity to nitrogen as found for silver() ion.
Usually, the interactions between the macrocycle ring and silver
ion in solution are analyzed in terms of the interactions with a
positively charged ion that leads to the 1:1 inclusion complex.
There are only few reports about more complex sandwich
2
directed in order to find a sensitive and selective metal-ion
16,17
detector.
Unfortunately, only in a few of these studies was
the influence of side-arm groups on the stoichiometry and
18
thermodynamics of complex formation investigated. There-
ϩ
7–9
aggregates with the 1:2 (Ag :ligand) stoichiometry. The
fore we thought it worthwhile to investigate the stoichiometry
of complexes formed by the silver() ion with aza-crown ethers
containing various side-arm groups.
ϩ
ϩ
tendency of the Ag ion to form linear AgL complexes is
2
10
well known in coordination chemistry. Similar structures can
be expected for complexes with macrocycles (examples of
mono- and diaza-crown ethers complexes are shown in Fig. 1).
However, further possibilities exist for extra- and intra-crown
interactions (Fig. 1).
Results and discussion
Spectroscopic studies
When the crown ether can form many types of complexes
with ions, the introduction of a chromophore group as a side-
arm fragment should result in a possibility for a step by step
investigation of the complex-formation equilibrium by spectro-
scopic measurements, particularly for asymmetrically substit-
uted diaza-crowns.
The UV–VIS spectrum of 1-amino-9,10-anthraquinone (AQN)
in the visible region contains a single π→π* redox band (cor-
responding to charge transfer from the amino group to the
19,20
anthraquinone moiety).
The band position is virtually
independent of whether the chromophore fragment is attached
DOI: 10.1039/a908585h
J. Chem. Soc., Dalton Trans., 2000, 689–696
This journal is © The Royal Society of Chemistry 2000
689

View Article Online
band. At the same time, isosbestic points appear in certain
ranges of the ligand:metal molar ratio. The most pronounced
changes are observed for ligands containing hydroxyanthra-
quinone side-arm groups: A15C5AQNOH 5 and A 15C5-
2
AQNOH 9. A typical series of UV–VIS spectra at various
ligand:metal ratios is presented in Fig. 3A. Plots of absorbance
ϩ
vs. Ag :crown molar ratio indicate the formation of a single 1:1
complex (Fig. 3B).
A characteristic feature of the spectrophotometric titrations
ϩ
is that for low CAg /Cligand molar ratios the absorbance changes
are small. This effect does not depend on the concentration of
ligands or silver() perchlorate. This type of pattern always
occurs for ligands containing quinone or hydroxyquinone
moieties connected directly to the nitrogen atom of the crown
and is not observed if the crown moiety contains a bulky
tosyloxy group (A15C5AQNOTs 6) {13-(8-tosyloxy-9,10-
dioxo-1-anthryl]-1,4,7,10-tetraoxa-13-azacyclopentadecane}
(
Fig. 3B). It can therefore be suggested that the introduction of
a bulky tosyl group precludes the possibility of close proximity
of the crown and quinone moieties. Thus, the coordination of
the silver ion will lead to only minimal conformational changes.
On the other hand, metal-ion coordination may cause large
changes in the mutual orientation of the crown and the quinone
moieties for attachment of anthaquinone and hydroxyantha-
quinone moieties to the crown ether resulting in changes of the
spectroscopic characteristics (Fig. 3B).
Fig. 2 Structures of ligands considered in this study.
The influence of the silver() ion on the spectrum of
15C5meAQN is by far the weakest. The position and the inten-
sity of the 510 nm band in the Ag –15C5meAQN system
remains practically unchanged. The changes in absorbance are
of the order of 0.1 log unit. Such changes are too small to be
used for the analysis of coordination interactions. These results
indicate that the anthraquinone moiety does not interact
significantly with silver ions.
Table 1 π→π* band position (λ/nm) and molar absorption coefficients
3
Ϫ1
Ϫ1
ϩ
(
log ε/dm mol cm ) of 1-amino-9,10-anthraquinone and some of its
derivatives and the effect of the presence of silver() ion in propylene
carbonate solution at 25 ЊC
a
Ligand
Ligand–silver()
Ϫ1
Ϫ1
log (ε/mol
dm cm
log (ε/mol
dm cm
3
Ϫ1
3
Ϫ1
For the crown ethers for which the apparent changes in the
spectra caused by the presence of the silver ion were observed,
molar-ratio curves were determined.
λ/nm
)
λ/nm
)
AQN
470
510
510
510
530
506
530
510
3.80
3.60
3.54
3.90
3.89
3.64
3.88
3.70
470
463
467
469
462
463
467
508
3.80
3.31
3.17
3.68
3.69
3.58
3.71
3.70
ϩ
A15C5AQN 4
The resulting data indicate that AgL –type complexes are the
ϩ
A 15C5AQN 8
dominant forms for Ag in propylene carbonate solution. The
2
A 15C5(AQN) 10
2
2
situation is different for mono-substituted diaza-complexes
with A 15C5AQN 8 and A 15C5AQNOH 9. The nature of the
A15C5AQNOH 5
A15C5AQNOTs 6
2
2
absorbance dependence on molar ratio for A 15C5AQN 8 at
2
A 15C5AQNOH 9
2
5
10 nm suggests that there are at least two types of complexes
1
5C5meAQN 2
in propylene carbonate solution (Fig. 4B). This is manifested
a
Concentration of ligand to silver() = 1:5.
ϩ
as inflection points that appear at CAg /CA215C5AQN molar
ratios of 1:1 and 1:1.65.
to the macrocycle moiety by a methylene bridge (15C5meAQN
) {2-[N-(9,10-dioxo-1-anthryl)aminomethyl]-1,4,7,10,13-penta-
2
Determination of equilibrium models
oxacyclopentadecane} or is directly attached to the nitrogen
atom (A15C5AQN 4) [13-(9,10-dioxo-1-anthryl)-1,4,7,10-
tetraoxa-13-azacyclopentadecane] (Table 1). The molar absorp-
tion coefficients of all ligands studied are constant within the
concentration range considered. This demonstrates that ligand
association does not play any significant role in the phenomena
investigated.
When a hydroxy group is present in the molecule, a batho-
chromic shift of the absorption band accompanied by the
increase of molar absorption coefficient is observed. The
presence of two chromophore groups in the molecule doubles
the intensity of the band at 510 nm. This suggests that the two
chromophore groups are fully independent of each other.
The introduction of silver() ion to the solutions of
A15C5AQN 4, A15C5AQNOH 5 [13-(8-hydroxy-9,10-dioxo-1-
ϩ
For all systems, except Ag –A 15C5AQN spectrophotometric
2
and potentiometric data could be fitted assuming the formation
ϩ
ϩ
ϩ
2
of AgL or AgL and AgL species. For A 15C5AQN the
2
21
absorbance–absorbance, obtained from spectrophotometric
measurements, exhibits four inflexion points and five lines as in
contrast to the other systems (Fig. 5) implying more equilibria
than for the other ligands.
ϩ
The potentiometric titration curves of Ag with A15C5 3
(1,4,7,10-tetraoxa-13-azacyclopentadecane) and 15C5meNH 1
[2-(hydroxymethyl)-1,4,7,10-tetraoxa-13-azacyclopentadecane]
in propylene carbonate exhibit characteristic inflection points
indicating the coexistence of at least two equilibria. For A 15C5
2
2
7 (1,4,10-trioxa-7,13-diazacyclopentadecane), only one inflec-
tion point in the titration curve was observed as opposed to
two for A15C5 3, within a wide range of ligand concentrations
anthryl-1,4,7,10-tetraoxa-13-azacyclopentadecane], A 15C5-
2
Ϫ2
Ϫ3
Ϫ5
Ϫ3
AQN 8, A 15C5AQNOH
9
[7-(8-hydroxy-9,10-dioxo-1-
(1.00 × 10 mol dm –1.00 × 10 mol dm ). The best-fitting
2
anthryl]-1,4,10-trioxa-7,13-diazacyclopentadecane] and A2-
equilibrium model for these ligands that contains two equilibria
involves the formation of AgL and AgL2 complexes.
For crown ethers derived with anthraquinone, hydroxy-
anthraquinone and tosyloxyanthraquinone moieties similar
conclusions can be drawn. The analysis of both potentiometric
ϩ
ϩ
1
7
5C5(AQN)2 10 [7,13-bis(9,10-dioxo-1-anthryl)-1,4,10-trioxa-
,13-diazacyclopentadecane] causes dramatic changes in the
positions and intensities of the visible absorption bands. These
changes are manifested by a blue shift of the 510 nm absorption
6
90
J. Chem. Soc., Dalton Trans., 2000, 689–696

View Article Online
Ϫ5
Ϫ3
Fig. 3 A, Absorption spectra recorded in PC solution containing A15C5AQNOH 5 (9.00 × 10 mol dm ) and concentration of silver() perchlorate
Ϫ4
Ϫ3
(
0–4.50 × 10 mol dm ). B, Dependence of absorbance at 506 nm for A15C5AQNOTs 6 (a) and at 530 nm for A15C5AQNOH 5 (b) as a function
ϩ
of molar ratio CAg /C .
L
Ϫ4
Ϫ3
Fig. 4 A, Absorption spectra recorded in PC solutions containing A 15C5AQN 8 (1.20 × 10 mol dm ) and concentration of silver() perchlorate
2
Ϫ4
Ϫ3
(
0–6.00 × 10 mol dm ). B, Dependence of absorbance at 510 nm for A 15C5AQN 8 with silver perchlorate. The experimental points are (a) (open
circles) and the fitting curves for the models are 1 (b) (line), 4 (c) (dotted line), 8 (d) (dashed line).
2
and spectrophotometric data leads to the conclusion that at
least two equilibria exist in the solvent leading to the formation
tration can only be explained assuming the coexistence of three
ϩ
or four equilibria which involve the formation of AgL ,
ϩ
ϩ
ϩ
3ϩ
ϩ
ϩ
3ϩ
2ϩ
ϩ
of AgL or AgL and AgL complexes.
Ag L2 and AgL2 or AgL , Ag L , Ag L3 and AgL2
2
3
3
2
2
2
2,23
For A 15C5AQN 8 the results of fitting various equilibrium
complexes. According to the F-test
the best-fitting model
2
models to the spectrophotometric data are summarized in
Table 2. As shown, the dependence of absorbance on concen-
containing three equilibria is indistinguishable at the 90% sig-
nificance level from a model in which all equilibria are included.
J. Chem. Soc., Dalton Trans., 2000, 689–696
691

View Article Online
Table 2 Results of fitting of various equilibrium models to spectrophotometric data
σ2
a
F
b
Fcrit.
c
αd
σA e
Model
Species considered
ϩ
1
2
3
4
5
6
7
8
AgLϩ
133.8
77.4
30.9
40.6
1.7
1.5
31.8
1.5
527.65
429.80
166.62
221.52
2.62
0.33
323.56
2.26
2.30
2.30
2.30
2.34
1.61
2.34
0.01
0.01
0.01
0.01
0.01
0.10
0.01
0.022
0.016
0.010
0.012
0.002
0.002
0.010
0.002
3ϩ
AgL , Ag L
3
2
3
ϩ
2ϩ
AgL , Ag L
2
ϩ
ϩ
AgL , AgL
2
ϩ
3ϩ
2ϩ
AgL , Ag L , Ag L
3
2
2
3
ϩ
3ϩ
ϩ
AgL , Ag L , AgL2ϩ
3
2
ϩ
2ϩ
AgL , Ag L , AgL2
2
3
ϩ
3ϩ
3ϩ
ϩ
AgL , Ag L , Ag L , AgL2
3
2
2
3
a
2
Residual variance: σ = ω¯ (x)/(n Ϫ p), p is the number of parameters, n is the number of experimental points, ω¯ (x) is the sum of the squares of the
b
differences between measured and calculated quantities (absorbances and concentrations). F = Test value = [ ω¯ (x) Ϫ ω(x*)](p* Ϫ p)/[ω(x*)
c
d
(
n Ϫ p*)], where ω¯ (x*) and p* correspond to the model 8 in which all the simpler models are included. Fcrit. = Critical F value. α = Level of
e
2
¹
significance. σA = Standard deviation in absorbance; σ = (σ )²
.
Table 3 Results of fitting of various equilibrium models to potentiometric data
σ2
a
F
b
Fcrit.
c
αd
σE e
Model
Species considered
ϩ
1
2
3
4
5
6
7
8
AgLϩ
455.0
41.9
72.7
210.2
64.8
49.8
22.0
19.9
940.23
71.78
170.68
612.18
287.14
191.26
14.56
1.55
1.55
1.55
1.55
1.55
1.55
1.55
0.01
0.01
0.01
0.01
0.01
0.01
0.01
40
38
15
25
14
12
6
3ϩ
AgL , Ag L
3
2
3
ϩ
2ϩ
AgL , Ag L
2
ϩ
ϩ
AgL , AgL
2
ϩ
3ϩ
2ϩ
AgL , Ag L , Ag L
3
2
2
3
ϩ
3ϩ
ϩ
AgL , Ag L , AgL2ϩ
3
2
ϩ
2ϩ
AgL , Ag L , AgL2
2
3
ϩ
3ϩ
3ϩ
ϩ
AgL , Ag L , Ag L , AgL2
5
3
2
2
3
a
2
Residual variance: σ = ω¯ (x)/(n Ϫ p), p is the number of parameters, n is the number of experimental points, ω¯ (x) is the sum of the squares of the
b
differences between measured and calculated values (e.m.f. and titrant volumes). F = Test value = [ ω¯ (x) Ϫ ω(x*)](p* Ϫ p)/[ω(x*) (n Ϫ p*)], where
c
d
ω¯ (x*) and p* correspond to the model 8 in which all the simpler models are included. Fcrit. = Critical F value. α = Level of significance.
e
f
2
¹
σA = Standard deviation in absorbance. σ = Standard deviation in e.m.f.; σ = (σ )²
.
E
Fig. 5 Absorbance–absorbance diagrams for measurements of A2-
1
5C5AQN 8 (a) and A15C5AQNOTs 6 (b) with silver() perchlorate.
Fig. 6 Potentometric measurement of silver() perchlorate (6.00 ×
Ϫ4
Ϫ3
Ϫ3
Ϫ3
1
0
mol dm ) with A 15C5AQN 8 (0–2.00 × 10 mol dm ) as a
2
ϩ
function of molar ratio CA215C5AQN/CAg . The experimental points are
The results of fitting various equilibrium models to the
(
(
a) (open circles) and the fitting curves for the models are 1 (b) (line), 4
c) (dotted line), 8 (d) (dashed line).
potentiometric data are summarized in Table 3. In contrast to
spectrophotometry only the model of four equilibria explains
the concentration dependence of the e.m.f. values in terms of
the F-test all other models are distinguishable at the 90%
significance level from the model containing four complexes:
4B) and potentiometric data (Fig. 6) it can be concluded that
ϩ
the minimal plausible model contains four complexes: AgL ,
2
ϩ
3ϩ
ϩ
2
Ag L3 , Ag L2 and AgL . It should be noted that only three
2
3
ϩ
2ϩ
3ϩ
ϩ
2
ϩ
3ϩ
ϩ
AgL , Ag L , Ag L2 and AgL .
equilibria, involving AgL , Ag L and AgL2 are needed to fit
2
3
3
3 2
Based on the results of fitting of spectrophotometric (Fig.
spectrophotometric data. This is caused by the fact that these
6
92 J. Chem. Soc., Dalton Trans., 2000, 689–696

View Article Online
Table 4 Stability constants of silver() complexes with the crown
ethers studied in propylene carbonate at 25 ЊC
sibility of the formation of an intermolecular hydrogen bond
between the hydroxy group of the anthraquinone moiety and
the amino group of the macrocyclic ring. This is supported by
Crown ether
log K1
log K2
log β
the analysis of the IR spectra. For the unsubstituted A 15C5
2
Ϫ1
crown 7, a sharp 3326 cm N–H stretching band is observed.
1
5C5
6.24 ± 0.03
1.77 ± 0.03
The intensity of this band decreases for A 15C5AQN 8 and the
2
2
4
5
.67
band strongly broadens for A 15C5AQNOH 9 suggesting the
2
1
1
5C5meNH 1
9.76 ± 0.07
6.08 ± 0.03
9.75 ± 0.04
4.04 ± 0.05
2.57 ± 0.03
4.77 ± 0.08
4.75 ± 0.03
1.55 ± 0.09
2
formation of an OH ؒ ؒ ؒ NH hydrogen bond.
5C5meAQN 2
For crown ethers containing a nitrogen atom in the side chain
A15C5 3
2
5
25
ϩ
ϩ
2
9
.76 ± 0.04
two complexes: AgL and AgL
are observed. It has been
A15C5AQN 4
5.92 ± 0.03
5.02 ± 0.04
5.90 ± 0.06
13.75 ± 0.04
5.6 [16]
11.71 ± 0.04
suggested that attachment of a non-coordinating side group to
the crown ring diminishes the formation constant owing to
A15C5AQNOH 5
A15C5AQNOTs 6
26
steric hindrance. However, for the example studied in this work
A 15C5 7
3.17 ± 0.03
4.71 ± 0.04
2
(
15C5meAQN 2), it is observed that the introduction of a side
1
a
b
chain to I5C5 only weakly diminishes log K and log K even
A 15C5AQN 8
29.72 ± 0.33
1
2
2
3
1.43 ± 0.18
increases. This suggests that even a relatively bulky anthra-
quinone group separated from the crown by methylene group
does not significantly hinder complex formation.
For A15C5 crown 3, the second formation constant (log
K ) decreases upon the introduction of an anthraquinone
2
A 15C5AQNOH 9
4.98 ± 0.03
5.59 ± 0.05
4.08 ± 0.03
1.85 ± 0.24
2
A 15C5(AQN) 10
2
2
a
3ϩ
b
Determined for Ag L
2
complex formation. Determined for
3
2
2ϩ
Ag L3 complex formation.
substituent in A15C5AQN 4 and for A15C5AQNOTs 6 and
ϩ
A15C5AQNOH 5, AgL2 complexes are not observed. This
two techniques are sensitive to different parameters, i.e. the
concentration of Ag (potentiometry) or to the concentration
of various ligand forms (spectrophotometry).
ϩ
indicates that direct substitution of the aza-crown nitrogen by
the anthraquinone moiety has a major influence on the stability
of different complexes. For diaza-crown-15 the formation of
1:2 complexes is enhanced.
Other structurally reasonable polynuclear complexes, e.g.
2ϩ
Ag L , were also considered in the equilibria models but
2
ϩ
It appears that the formation of the AgL2 complex in this
such models gave much poorer fitting than the complexes
considered.
case may be associated with the coordination of the secondary
nitrogen atom at the macrocycle with the silver() ion. On the
other hand, for 15C5, 15C5meAQN 2 and A15C5AQN 4, the
Discussion of the obtained stability constants
ϩ
AgL2 complex ion can be formed as a sandwich type complex.
As indicated in Table 4, for ligand-to-metal ratio equilibrium
For such a putative sandwich structure, the formation constant
is evidently smaller (2.57–1.55) compared to complexes in
which silver is bonded by a secondary nitrogen atom.
(
1) is dominant.
K
1
ϩ
ϩ
Ag ϩ L
AgL
(1)
For A 15C5AQN 8 we have additionally observed, by spec-
2
2ϩ
3ϩ
troscopic methods, the Ag L3 and Ag L2 complexes. The
2
3
The log K value increases with an increase of the number of
1
equilibrium model obtained from spectrophotometric meas-
urements has been confirmed by fitting of various models to the
potentiometric and spectrophotometric data (Tables 2 and 3).
It is very important that the same conclusions regarding the
correct equilibrium model have been derived from independent
spectrophotometric and potentiometric measurements. Spec-
trophotometry results depend on the concentration of the
ligand whereas potentiometry is dependent on the concen-
tration of the silver ion. The concordance of both models is a
very strong argument for the correctness of the derived
equilibria.
nitrogen atoms from 6.24 for 15C5 (1,4,7,10,13-pentaoxacyclo-
pentadecane) to 9.75 (9.76) for A15C5 3, 9.76 for 15C5meNH2
1
and 13.75 for A 15C5 7, respectively. It should be noted
2
that the increase of the log K value when passing from the
1
exclusively oxygen-containing crown moiety to the aza-crown is
observed regardless of whether the nitrogen atom is incorpor-
ated into the macrocycle ring or is attached as a side-arm.
For all unsubstituted crowns log K values were determined
2
for equilibrium (2).
K
2
ϩ
ϩ
AgL ϩ L
AgL2
(2)
Experimental
1,4,7,10-Tetraoxa-13-azacyclopentadecane (A15C5; Merck),
1,4,10-trioxa-7,13-diazacyclopentadecane (A 15C5; Merck),
These values increase when passing from the oxygen 15C5
crown (1.77) to the aza-crown 15C5meNH 1 (4.04) and to
2
2
A15C5 3 (4.77). For A 15C5 7, an apparent decrease of log K2
2
(
3.17) is observed compared with A15C5 3.
Attaching an anthraquinone substituent to the nitrogen atom
2
-(hydroxymethyl)-1,4,7,10-tetraoxa-13-azacyclopentadecane
(
15C5meNH ; Merck) and 1,4,7,10,13-pentaoxacyclopenta-
2
of A15C5 3 results in the loss of the capability for the form-
ation of other complexes than those of 1:1 type in most cases.
This is confirmed not only by the shape of the potentiometric-
titration curves but also by the results of spectrophotometric
measurements. A bulky aromatic substituent contributes to a
decane (15C5; Merck) were used as supplied without purific-
ation. Silver() perchlorate (Fluka) and tetraethylammonium
perchlorate (TEAP; Fluka) were purified as described prev-
27
iously. Propylene carbonate (PC; Merck) was of the highest
purity available. Silica gel 60 (<0.03 mm; Merck) and aluminium
oxide 90 (Merck) were used for chromatography.
remarkable decrease of log K from 9.75 for A15C5 3 to 5.92 for
1
A15C5AQN 4 which contains an anthraquinone moiety. For
A 15C5 7, the introduction of the first anthraquinone substit-
2
Preparations
uent has a significantly smaller effect on log K compared to the
1
hydroxyquinone substituent.
The preparation of A15C5AQN is shown in Scheme 1(a).
1-Fluoro-9,10-anthraquinone was obtained in 57% yield,
according to a procedure described in the literature.
For A 15C5 7, introduction of the anthraquinone substituent
2
28
diminishes the stability constant by more than two log K units.
1
For the hydroxyanthraquinone substituent the stability con-
stant is diminished by 8.77 logarithmic units, similarly as found
Treatment of this compound with the aza-crown ether A15C5
yields the 1-amino-9,10-dioxo-1-anthryl functionalized crown
ether as a deep red oil. 1-Fluoro-9,10-anthraquinone can be
replaced with 1-chloro-9,10-anthraquinone. However, in this
for the disubstituted A 15C5AQNOH crown 9. This dramatic
2
decrease of the stability constant can by attributed to the pos-
J. Chem. Soc., Dalton Trans., 2000, 689–696
693

View Article Online
1
,4,10-trioxa-7,13-diazacyclopentadecane [A 15C5(AQN) ]. A
2 2
mixture of 1 g (4.59 mmol) of 1,4,10-trioxa-7,13-diazacyclo-
pentadecane (A 15C5) 1.742 g (7.7 mmol), of 1-fluoro-9,10-
2
anthraquinone and 2.55 g (7.8 mmol) of caesium carbonate was
suspended in 15 ml of toluene, stirred and kept at 80 ЊC for 24
h. After cooling to the room temperature the insoluble solids
were removed by filtration and the filtrate concentrated. The
residual crude oil was purified by column chromatography on
aluminum oxide (dichloromethane–methanol 50:4) to give
A 15C5(AQN)2 (red solid) as the main fraction (R = 0.49,
2
f
0
.068 g, 7%) (Found: C, 62.49; H, 5.48; N, 4.53. Calc. for
C H N O : C, 62.38; H, 5.40; N, 4.44%). δ (CDCl ) 3.4–3.8
38
34
2
7
H
3
(
20H, m) and 7.45–8.20 (14H, m). A 15C5AQN was obtained as
2
the second fraction (R = 0.22); (0.662 g, 61%) (Found: C,
f
6
7.79; H, 6.49; N, 6.63. Calc. for C H N O : C, 67.92; H, 6.60;
24 28 2 5
N, 6.60%). δ (CDCl ) 2.9 (4H, s), 3.4–3.8 (19H, m) and 7.45–
H
3
8
.20 (7H, m).
(
iii) 13-[8-Hydroxy-9,10-dioxo-1-anthryl]-1,4,7,10-tetraoxa-
1
3-azacyclopentadecane (A15C5AQNOH). 1 g (4.56 mmol) of
A15C5 and 2.23 g (5.65 mmol) of 1-OH(8-OTs)-AQ were
dissolved in 200 ml of toluene and the mixture was heated to
8
0 ЊC for 24 h. The reaction mixture was cooled to room
temperature and then concentrated in vacuo and the residue
chromatographed on a silica gel [dichloromethane–methanol
(
50:1)]. A15C5AQNOH was obtained as a red solid (R = 0.48,
f
0
.966 g, 48%) (Found: C, 65.52; H, 6.27; N, 3.15. Calc. for
C H NO : C, 65.29; H, 6.16; N, 3.17%). δ (CDCl ) 3.6–3.8
24
27
7
H
3
(
20H, m), 7.4–8.0 (6H, m) and 13.0 (1H, s).
(
iv) 7-(8-Hydroxy-9,10-dioxo-1-anthryl)-1,4,10-trioxa-7,13-
diazacyclopentadecane (A 15C5AQNOH). A 15C5AQN was
2
2
prepared by the same method as used for the synthesis of
A15C5AQNOH. The final product was separated by a column
chromatography on silica gel (dichloromethane–methanol
Scheme 1 Synthesis of aza-substituted crown ethers.
5
0:2) to give A 15C5AQNOH as a red oil (R = 0.15, 21%)
case the yield of the product with A15C5 is lower and more
by-products are obtained. Application of a similar synthetic
procedure to diaza crown ethers yields mono- and di-arm sub-
stituted crown ethers: A 15C5AQN and A 15C5(AQN) . Add-
2
f
(
Found: C, 65.51; H, 6.52; N, 6.35. Calc. for C H N O : C,
24 28 2 6
6
5.44; H, 6.40; N, 6.36%). δ (CDCl ) 2.93 (4H, s), 3.4–3.8 (16H,
H 3
m), 7.45–8.2 (7H, m) and 13.0 (1H, s).
2
2
2
itional chromatographic purification of the crude products
(
v) 13-(8-Tosyloxy-9,10-dioxo-1-anthryl)-1,4,7,10-tetraoxa-
yields an easily soluble oil (A 15C5AQN) or deep red crystals
2
1
3-azacyclopentadecane (A15C5AQNOTs). 0.56 g (2.56 mmol)
of A15C5 and 1.40 g (2.56 mmol) of 1,5-(TsO) (AQN) were
dissolved in 10 ml of toluene. The mixture was stirred and
heated to 100 ЊC for 24 h. The reaction mixture was dissolved in
water–dichloromethane (1:1, 100 ml). After separations, the
organic phase was dried over MgSO and the main product
was purified by column chromatography on aluminum oxide
CH Cl –methanol (50:1)]. A15C5AQNOTs was obtained as
a red solid (R = 0.55, 0.715 g, 47%) (Found: C, 62.72; H, 5.66;
[
A 15C5(AQN) ]. 15C5meAQN was synthesized in a similar
2
2
manner from 15C5meNH and 1-fluoro-9,10-anthraquinone.
2
2
A15C5AQNOH was synthesized through the reaction of 1-
tosyloxy-8-hydroxy-9,10-anthraquinone with the crown ether.
The synthesis always gives two products, A15C5AQNOH as the
main product (78% yield) and A15C5AQNOTs (8% yield) as
the minor product [Scheme 1(b)].
4
[
The synthesis of A15C5AQNOH can also be carried out
using the 1,8-ditosyl derivative of the 1,8-dihydroxyanthra-
2 2
f
N, 2.39. Calc. for C H NO S: C, 62.51; H, 5.58; N, 2.35%).
δ (CDCl ) 2.4 (3H, s), 3.5–3.8 (20H, m) and 7.25–8.25 (10H, m).
quinone 1,8-(TsO) AQN. This gives A15C5AQNOTs which on
31 33
9
2
hydrolysis in alcoholic KOH solution yields A15C5AQOH.
However, this alternative method is less convenient and results
in a lower yield of the final product.
H
3
(
vi)
2-[N-(9,10-Dioxo-1-anthryl)aminomethyl]-1,4,7,10,13-
pentaoxacyclopentadecane (15C5meAQN). 2-(Aminomethyl)-
15-crown-5 (1 g, 4.0 mmol), 1-fluoro-9,10-anthraquinone (1.03
g, 4.56 mmol) and caesium carbonate (1.5 g, 4.6 mmol) were
dissolved in 10 ml toluene and the mixture heated to 80 ЊC for
48 h. After cooling to room temperature the insoluble solid
was removed by filtration and the filtrate was concentrated.
The residue was separated by a column chromatography
on aluminum oxide [dichloromethane–methanol (500:1)].
(
i) 13-(9,10-Dioxo-1-anthryl)-1,4,7,10-tetraoxa-13-azacyclo-
pentadecane (A15C5AQN). 1,4,7,10-Tetraoxa-13-azacyclo-
pentadecane (A15C5) (1.0 g, 4.57 mmol), caesium carbonate
(
1.6 g, 4.90 mmol) and 1-fluoro-9,10-anthraquinone (1.1 g, 4.86
mmol) in 10 ml toluene were stirred and heated at 80 ЊC for
8 h. After cooling to the room temperature, the insoluble
solid was filtered off. The solution was evaporated under
vacuum, purified by column chromatography on silica gel
dichloromethane–methanol (50:2), R = 0.32] to give
A15C5AQN as a red viscous oil (1.32 g, 68%) (Found: C, 67.79;
H, 4.43; N, 3.23. Calc. for C H NO : C, 67.76; H, 4.32; N,
4
15C5meAQN was obtained as a red solid (R
69%) (Found: C, 65.62; H, 3.07; N, 6.42. Calc. for C25
C, 65.45; H, 3.09; N, 6.38%). δ (CDCl ) 2.0–2.2 (2H, m), 3.2–
3.8 (20H, m) and 7.45–8.2 (7H, m).
= 0.21, 1.252 g,
f
[
H29NO :
7
f
H
3
24
27
6
3
.29%); δ (CDCl ) 2.4–3.8 (24H, m) and 7.45–8.20 (7H, m).
H 3
(
vii) 1,8-Bis(tosyloxy)-9,10-anthraquinone (1,8-(TsO) AQN).
2
(
ii) 7-(9,10-Dioxo-1-anthryl)-1,4,10-trioxa-7,13-diazacyclo-
5.09 g (21.1 mmol) of 1,8-dihydroxy-9,10-anthraquinone
and 5.32 g (27.9 mmol) of p-toluenesulfonyl chloride (tosyl
pentadecane (A 15C5AQN) and 7,13-bis(9,10-dioxo-1-anthryl)-
2
6
94
J. Chem. Soc., Dalton Trans., 2000, 689–696

View Article Online
chloride) was dissolved in 200 ml dichloromethane. 15 ml
triethylamine in 100 ml dichloromethane was added dropwise to
the solution over 5 h. The course of the reaction was followed
Spectroscopic measurements
UV–VIS spectra were recorded using a Varian CARY 3E
spectrometer at 25 ЊC. Solvents were used as purchased without
further purification but their purity was verified by measuring
the absorption spectra in the range 220–800 nm. Solutions of
the crown ethers were prepared by solubilization of the crown
by TLC [SiO , dichloromethane–light petroleum (bp 40–60 ЊC)
2
(
1:1)] and was complete after 5–6 h. The reaction mixture was
washed with water (3 × 50 ml) and the organic phase dried with
MgSO . After evaporating the solvent the residue was separated
4
Ϫ5
ether in propylene carbonate at concentrations of 8 × 10 –
by column chromatography on silica gel [dichloromethane–
light petroleum (100:85)]. 1,8-diOTs-AQ was obtained as
orange crystals (3.29 g, 25%) (Found: C, 61.42; H, 3.69; N,
Ϫ4
Ϫ3
4
× 10 mol dm . Solutions of silver() ion were prepared
Ϫ4 Ϫ3
from the perchlorate salt at concentrations of 8 × 10 –6 × 10
Ϫ3
mol dm . The concentration of the ligands was the same in
both solutions. Measurements were performed manually using
a micro-screw with 0.5 ml Hamilton micro-syringe equipped
with a Gauge 30 tube. Magnetic stirrers employing a counter-
shaft inside the spectrometer were used during measurements.
All spectra data were obtained in a digital form and sub-
sequently transferred into a worksheet. Measurement graphs
were plotted and equilibrium constants were calculated if
1
1.72. Calc. for C H O S : C, 61.31; H, 3.67; N, 11.69%).
28 20 8 2
δ (CDCl ) 2.4 (6H, s), 7.4 (8H, dd) and 7.5–8.3 (6H, m).
H
3
(
viii) 1-Hydroxy-8-tosyloxy-9,10-anthraquinone (1-OH-8-
TsOAQN). 5.09 g (21.1 mmol) of 1,8-dihydroxy-9,10-anthra-
quinone and 5.32 g (27.9 mmol) of p-toluenesulfonyl chloride
were dissolved in 200 ml dichloromethane. 15 ml triethylamine
in 100 ml dichloromethane was added dropwise to the solution
over 0.5 h and the reaction continued for 20 h at a room tem-
perature. The reaction mixture was washed with water (3 × 50
32
30,31
possible. The program STOICHIO
based on the non-
linear least squares Gauss–Newton–Marquardt algorithm was
used to fit the parameters of the equilibrium models.
ml) and dried with MgSO . After evaporation of the solvent the
4
residue was separated by a column chromatography on silica gel
with dichloromethane–light petroleum (1:1) and subsequently
with dichloromethane–methanol (10:1). 1-OH-8-OTsAQN
C H O S (MW 394.4) was obtained as a yellow solid (4.93 g,
Other measurements
1
The H NMR spectra were recorded at 400 MHz using a Varian
2
1
14
6
Mercury 400BB spectrometer. Chemical shifts were obtained
on CDCl solution relative to Me Si unless stated otherwise. IR
5
9%). δ (CDCl ) 2.40 (3H, s); 7.3–7.4 (4H, m), 7.8–8.5 (6H, m)
H 3
3
4
and 12.0 (1H, s).
spectra were recorded on a Bruker IFS 66 spectrometer. Spectra
of solid samples were taken as Nujol and hexachlorobutadiene
mulls. Elemental analysis was performed on a Carbo Erba
Elemental Analyzer MOD 1106.
Potentiometric measurements
Potentiometric titrations were performed in propylene carbon-
ate at 25 ЊC using an OP-205 Radelkis pH-meter linked to a
personal computer via a PCL-838 control card. The solvent was
used as purchased without further purification. Silver() ion
solutions were prepared from the perchlorate salt with concen-
Acknowledgements
The authors thank Dr A. Liwo for helpful discussions concern-
ing the modeling of equilibria. This work was supported by
grants BW- 8000-5-0308-9 and PB 026/T09/97/12 from the
Polish State Committee for Scientific Research (KBN).
Ϫ4
Ϫ3
trations in the range 5–8 × 10 mol dm . Concentrations of
Ϫ3
Ϫ2
Ϫ3
ligands were in the range 5.0 × 10 –1.0 × 10 mol dm .
Measurements were performed using a 0.5 ml Hamilton micro–
syringe equipped with a Gauge 30 tin tube. The half-cells were
Ϫ3
connected by a salt bridge filled with 0.1 mol dm tetraethyl-
References
ammmonium perchlorate dissolved in the solvent. The silver
concentration during titrations was determined using silver-
wire electrodes dipped in the solution. For poorly soluble
1 R. M. Izatt, J. S. Bradshaw, S. A. Nelsen, J. D. Lamb and J. J.
Christiansen, Chem. Rev., 1985, 85, 271.
2 C. J. Pedersen, J. Am. Chem. Soc., 1970, 89, 386.
26
ligands an inverted titration technique was used. In this
technique solutions of ligands with concentrations in the range
3
4
C. J. Pedersen, J. Am. Chem. Soc., 1967, 89, 7017.
C. J. Pedersen and H. K. Frensdorf, Angew. Chem., Int. Ed. Engl.,
Ϫ4
Ϫ3
1
–5 × 10 mol dm and a small amount of silver ion were
1
972, 11, 16.
titrated by solutions of silver perchlorate with concentrations in
5
F. P. van Remoortere and F. P. Boerg, Inorg. Chem., 1974, 13, 2071.
Ϫ4
Ϫ3
Ϫ3
the range 8 × 10 –5 × 10 mol dm .
6 N. S. Poonia, J. Am. Chem. Soc., 1976, 96, 107.
7 A. Thaler, R. Bergter, T. Ossowski, B. G. Cox and H. Schneider,
Inorg. Chim. Acta, 1999, 285, 1.
ϩ
ϩ
Stability constants for AgL and AgL2 complexes were
determined as log K and K [eqns. (1) and (2)].
1
2
8
H.-J. Buschmann, in Stereochemical and Stereophysical Behaviour of
Macrocycles, ed. I. Bernal, Elsevier, Amsterdam–Oxford–New
York–Tokyo, 1987, vol. 2, p. 103.
K
1
ϩ
ϩ
Ag ϩ L
AgL
(1)
(2)
9
M. Takagi, in Cation Building by Macrocycles, ed. Y. Inoue and
G. W. Gokel, Marcel Dekker, New York, 1990, p. 464.
K
2
10 F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry,
ϩ
ϩ
AgL ϩ L
AgL2
Wiley, New York, 5th edn., 1988, p. 941.
1
1
1
1 T. A. Kaden, Pure Appl. Chem., 1993, 65, 1473.
2 G. W. Gokel, Chem. Soc. Rev., 1992, 21, 39.
The equilibrium constants for the additional complexes:
3 H. C. Loehr and F. Voegtle, Chem. Ber., 1985, 118, 914.
3ϩ
2ϩ
Ag L2 and Ag L3 were expressed as overall stability con-
14 E. Weber, in Crown Ethers and Analogs, ed. S. Patai and
3
2
stants, [eqns. (3) and (4)]:
Z. Rappoport, Wiley, New York, 1989, p. 305.
5 Y. Katayama, R. Fukuda and M. Takagi, Anal. Chim. Acta, 1986,
1
1
85, 295.
βAg
βAg
3
L
L
2
3
ϩ
3ϩ
2ϩ
3
Ag ϩ 2L
Ag L2
(3)
(4)
16 Y. Katayama, K. Nita, M. Ueda, M. Nakamura and M. Takagi,
3
Anal. Chim. Acta, 1985, 173, 193.
7 Y. Katayama, R. Fukuda, T. Iwasaki, K. Nita and M. Takagi, Anal.
Chim. Acta, 1998, 204, 113.
8 G. W. Gokel and I. E. Trafton, in Cation Building by Macrocycles,
ed. Y. Inoue and G. W. Gokel, Marcel Dekker, New York, 1990,
p. 253.
1
2
ϩ
2
Ag ϩ 3L
Ag L3
2
1
The equilibrium constants were calculated using the program
1
9 H. Inuone, T. Hoshi and Y. Tanizaki, Bull. Chem. Soc. Jpn., 1972,
STOICHIO based on the non-linear least-squares Gauss–
4
5, 1018.
29–31
Newton–Marquardt algorithm.
20 Z. Yoshiba and F. Takabyashi, Tetrahedron, 1968, 24, 933.
J. Chem. Soc., Dalton Trans., 2000, 689–696
695

View Article Online
2
1 J. Polster and H. Lachman, in Spectroscopic Titration. Analysis of
Chemical Equilibria, VCH, Weinheim–Basel–Cambridge, New
York, 1989.
27 D. D. Perrin and W. L. F. Armari, in Purification of Laboratory
Chemicals, Pergamon Press, New York, 1988.
28 G. Valanas and H. Fopff, J. Org. Chem., 1962, 27, 3680.
29 D. W. Marquardt, J. Soc. Ind. Appl. Math., 1963, 11, 431.
30 J. Kostrowicki and A. Liwo, Comput. Chem., 1987, 11, 195.
31 J. Kostrowicki and A. Liwo, Talanta, 1990, 37, 645.
32 P. J. Legett, in Computational Methods for the Determination of
Formation Constants, Plenum Press, New York–London. 1985.
2
2
2
2
2
2 J. Czerminski, in Statistical Methods in Applied Chemistry, Elsevier,
Amsterdam–New York–Tokyo, 1990.
3 G. A. F. Seber and C. J. Wild, in Nonlinear Regression, Wiley, New
York, 1989, p. 228.
4 M.-C. Almasio, F. Arnaud-Neu and M.-J. Swing-Weill, Helv. Chim.
Acta, 1983, 66, 1296.
5 T. Burchard, B. G. Cox, P. Firman and H. Schneider, Ber. Bunsenges
Phys. Chem., 1994, 98, 1526.
6 B. G. Cox and H. Schneider, in Coordination and Transport
Properties of Macrocyclic Compounds in Solution, Elsevier,
Amsterdam, 1992.
Paper a908585h
6
96
J. Chem. Soc., Dalton Trans., 2000, 689–696