Boronic acid appended azo dyes-colour sensors for saccharides

Article abstract of DOI:10.1039/b108896c

A number of boronic acid functionalised azo dye molecules have been prepared using a simple 3-step synthesis and their spectral properties upon complexation with monosaccharides investigated. The dyes undergo visible colour changes in aqueous solution upon addition of monosaccharide.

Full text of DOI:10.1039/b108896c

View Article Online / Journal Homepage / Table of Contents for this issue
Boronic acid appended azo dyes–colour sensors for saccharides†
Christopher J. Ward,a Prakash Patelb and Tony D. James*a
a Department of Chemistry, University of Bath, Claverton Down, Bath, UK BA2 7AY.
E-mail: T.D.James@bath.ac.uk
1
b Avecia Ltd, Hexagon House, PO Box 42, Blackley, Manchester, UK M9 8ZS
Received (in Cambridge, UK) 2nd October 2001, Accepted 7th December 2001
First published as an Advance Article on the web 21st January 2002
A number of boronic acid functionalised azo dye molecules have been prepared using a simple 3-step synthesis and
their spectral properties upon complexation with monosaccharides investigated. The dyes undergo visible colour
changes in aqueous solution upon addition of monosaccharide.
diols to the rigid cis-diols of saccharides. This observed select-
Introduction
ivity order has since been shown to be common to all mono-
boronic acids, not just benzeneboronic acid.1 The suitability of
the boronic acid functionality as a receptor for saccharides has
been well established in fluorescence, circular dichroism and
electrochemical studies.1,3,5,6
A fairly recent development has been the study of the effect
of saccharides on the colour of dyes containing the boronic
acid functionality. Boronic acid azo dyes have been known
for over forty years where they were used for investigations in
the treatment of cancer by the technique of Boron Neutron
Capture Therapy (BNCT).7–9 However, it has only been in
the last decade that related dyes and their interaction with
saccharides have been explored.
A great amount of attention continues to be devoted to the
development of synthetic molecular receptors with the ability
to recognise neutral organic species, including saccharides.1 The
vast majority of these systems have relied upon hydrogen bond-
ing interactions for the purposes of recognition and binding of
guest species. However, there is still no designed, monomeric
receptor that can compete effectively with bulk water for low
concentrations of monosaccharide substrates.2 As the chem-
istry of saccharides and related molecular species plays a sig-
nificant role in the metabolic pathways of living organisms,
detecting the presence and concentration of biologically
important sugars in aqueous solution is necessary in a variety
of medicinal and industrial contexts. The recognition of -
glucose is of particular interest, since the breakdown of glucose
transport in humans has been correlated with certain diseases:
renal glycosuria, cystic fibrosis, diabetes and also human cancer.
Recent research provides clear evidence that tight control of
blood sugar levels in diabetics sharply reduces the risk of
long term complications, which include blindness, kidney fail-
ure, heart attacks and gangrene. Industrial applications range
from the monitoring of fermentation processes to establishing
the enantiomeric purity of synthetic drugs.3
The boronic acid–saccharide interaction can be utilised to
overcome the problem of undesired solvent competition for the
host. Boronic acids readily and reversibly form cyclic boronate
esters with diols in aqueous basic media.3 Saccharides contain a
linked array of hydroxy groups that provide an ideal structural
framework for binding to boronic acids. The most common
interaction is with 1,2- and 1,3-diols of saccharides to form
five- or six-membered rings respectively, via two covalent bonds
(Scheme 1).
Nagasaki et al. observed that chromophores containing
boronic acid moieties that aggregate in water, changed colour
and deaggregated upon addition of saccharides.10 This was
rationalised by the boronic acid–saccharide complexation
increasing the hydrophilicity of the bound species. Takeuchi et al.
have prepared a boronic acid dye that undergoes an absorp-
tion spectral change on addition of nucleosides through bind-
ing of boronic acid to the ribose unit of the nucleosides.11
Strongin has prepared a tetraboronic acid resorcinarene system
1 for the visual sensing of saccharides.12 However, for a colour
change to be observed the saccharide must be heated to 90 ЊC in
DMSO. This reaction is irreversible so cannot be applied as a
sensor system.
In 1994 Sandanayake and Shinkai reported ‘the first known
synthetic molecular colour sensor for saccharides’.13 This
designed molecular internal charge transfer (ICT) sensor 2 is
based on the intramolecular interaction between the tertiary
amine and the boronic acid group. The electron-rich amine
creates a basic environment around the electron-deficient boron
centre that has the effect of inducing the boronic acid–
saccharide interaction and reducing the working pH of the
sensor.14 The electronic change associated with this decrease in
the pKa of the boronic acid moiety on saccharide complexation
was shown to be transmitted to the neighbouring amine. This
creates a spectral change in the connected ICT chromophore,
which can be detected spectrophotometrically. The pKa value
associated with the boron–nitrogen interaction shifted on
saccharide addition. More recently, Shinkai’s group have syn-
thesised a bis(boronic acid)-based saccharide receptor, bearing
a photo-responsive azobenzene group which has shown high
glucose selectivity when the azobenzene unit is switched to the
thermodynamically unfavourable cis conformation.15 They have
also demonstrated that azobenzene derivatives bearing one or
two aminomethylbenzeneboronic acid groups can be useful for
practical colorimetric saccharide sensing in ‘neutral’ aqueous
media.16 Shinkai has applied the boronic acid–amine inter-
Scheme 1 Benzeneboronic acid complexation with diols
Lorand and Edwards determined the first stability trends and
selectivities of various polyols and saccharides towards benzene-
boronic acid.4 The complex stability was shown to increase
from ethylene glycol to -fructose, i.e. from the simple acyclic
† Electronic supplementary information (ESI) available: absorption–
pH titrations and absorption–saccharide titrations. See http://
www.rsc.org/suppdata/p1/b1/b108896c/
462
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b108896c

View Article Online
Scheme 2 Synthesis of boronic acid azo dyes.
action to the molecular design of a visualized intermolecular
sensing system for saccharides.17 3-Nitrobenzeneboronic acid
interacts with the pyridine nitrogen of 4-(4-dimethylamino-
phenylazo)pyridine 3 in methanol and changes its colour from
yellow to orange. Added saccharides form complexes with the
boronic acid and enhance the acidity of the boronic acid group.
As a result, the boron–nitrogen interaction becomes stronger
and the intensified intramolecular charge-transfer band changes
the solution colour to red.
The main drawback in the development of coloured boronic
acid–saccharide sensors, so far, has been the relatively small
shifts achieved in the absorption bands of the chromophores
upon saccharide binding. The aim of this research was to
develop molecular ICT sensors, which would produce sig-
nificant visible colour changes upon saccharide interaction.
Furthermore, by changing the substituent on the azo group’s
aromatic ring to a variety of electron donating or withdrawing
groups (e.g. methoxy or nitro groups), the electronics of the
chromophore will be subtly altered. No study has been con-
ducted on the effect of these substituents on saccharide binding
with boronic acid azo dyes. If a system with a large colour
change can be developed it could be incorporated into a diag-
nostic test paper for -glucose, similar to universal indicator
paper for pH. Such a system would make it possible to
rapidly measure -glucose concentrations without the need for
specialist instrumentation. This would be of particular benefit
to diabetics in developing countries.
ation conditions to yield the boronic acid azo dyes as single
isomers.
The reaction of m-toluidine with 2-formylbenzeneboronic
acid was performed in a refluxing ethanol–toluene mix (10 : 1
v/v), with azeotropic removal of water by a Dean–Stark trap.
The reaction works very well (isolated yield >90%) and always
shows 100% conversion to imine 4 by proton NMR, so there
is no need for any purification. The imine was reduced by
an excess of sodium borohydride in dry methanol at room
temperature. After an aqueous acidic work-up, amine 5 was
obtained in good yield (81%) and excellent purity.
The final step was to perform the azo coupling reaction.
These reactions occur predominately para to the electron-
donating anilinic nitrogen because of steric effects, but ortho
substitution can also be observed. However, it is known that by
incorporating a methyl group meta to the anilinic nitrogen in
the amine coupling agent, exclusive para coupling is observed.18
Hence the inclusion of this design motif in the target dyes
avoids potentially difficult separation of isomers. Azo coupling
reactions were performed using the diazonium salts of 4-nitro-
aniline, sulfanilic acid (4-aminosulfonic acid), 4-aminobenzoic
acid, p-anisidine (4-methoxyaniline) and 3-aminobenzoic acid,
to yield the respective dyes 6, 7, 8, 9 and 10. A methanol–water
solvent system was employed in all of the coupling reactions for
solubility reasons. All dyes have been fully characterised.
This route is synthetically simpler than previously reported
procedures. For example, Shinkai’s route involved a moderately
yielding lithiation reaction to make the boronic acid, followed
by an azo coupling step that afforded two isomers that required
separation and purification.13
Results and discussion
A. Synthesis of boronic acid azo dyes
B. UV–VIS studies
We report herein a novel and facile synthetic route to boronic
acid azo dyes, with little or no need for purification (Scheme 2).
Commercial 2-formylbenzeneboronic acid is reacted with
m-toluidine to form the respective imine. This is then reduced
to the amine using a standard sodium borohydride reduction.
The final step is to couple with the diazonium salt of a series
of anilinic molecules (e.g. 4-nitroaniline) under normal diazotis-
Ultraviolet–Visible (UV–VIS) spectrophotometry was used to
study the boronic acid azo dye molecules’ interaction with
saccharides. The absorption of visible light by azo compounds
can be described through a Linear Combination of Atomic
Orbitals-Molecular Orbital (LCAO-MO) approach.19 A non-
bonding (n) electron from the azo nitrogen is excited, by the
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470
463

View Article Online
Table 1 Calculated pKa values for all dyes with and without -fructose
absorption of a photon of visible light, into the lowest anti-
bonding (π*) orbital in an n π* electronic transition. The
energy, and hence wavelength, of this transition is dependent
upon the substituents of the azo chromophore and the induct-
ive and conjugation effects that these have on the energy of the
n and π* orbitals.
The anilinic nitrogen found in the target molecules is in con-
jugation with the azo nitrogens through the phenyl ring system.
The anilinic nitrogen is also in close proximity to the boronic
acid boron centre. The electronic changes associated with
saccharide binding can be transmitted through a boron–anilinic
nitrogen interaction to the azo chromophore, hopefully creating
a spectral change in the molecule that can be detected in the
UV–VIS absorption spectrum. It is well known that these
electronic changes are often manifested in a pKa change and
it is this that can be studied by an UV–VIS absorption–pH
titration.13
Without -fructose
With -fructose
pKa
Dye
pKa
r2
r2
∆pKa
6
7
10.19 0.03
10.01 0.02
9.93 0.03
10.22 0.01
10.07 0.02
9.68 0.02
0.997
0.999
0.997
0.999
0.999
0.998
7.04 0.04
7.13 0.03
7.24 0.04
7.17 0.02
7.21 0.04
6.99 0.04
0.998
0.998
0.995
0.999
0.996
0.996
3.15
2.88
2.69
3.05
2.86
2.69
8
9
10
13
C. Absorption–pH titrations
Absorption–pH titrations, from pH 2 to 12, of each of the
boronic acid azo dyes in 0.05 mol dmϪ3 NaCl in methanol–
water (1 : 2, w/w), were followed using a UV–VIS spectrometer.
The experiments were then repeated with 0.05 mol dmϪ3 -
fructose also present. The sodium chloride present acts as an
“osmotic buffer” because small amounts of sodium chloride
are formed on adjustment of the pH with sodium hydroxide
and hydrochloric acid. Because the titrations are carried out
in a methanol–water mixture rather than simply water, the
measurement of pH using a standard electrode is not strictly
applicable to this situation. However, De Ligny and Rehbach
have shown that for solutions in 50% methanol the pH is only
changed by 0.1 of a pH unit compared to a 100% water sol-
ution.20 At percentages lower than 84% methanol, the error
does not exceed 0.2 pH units, but for higher percentages it
dramatically increases and for absolute methanol it is equal to
2 pH units.
Fig. 1 Absorption–pH spectral changes of dye 10 (5.34 × 10Ϫ5 mol
dmϪ3). Representative spectra at pH = 6.59, 9.94, 10.49 and 10.97.
Fig. 1 shows the absorption–pH profile of dye 10. Fig. 2
shows the same experiment with the presence of 0.05 mol dmϪ3
-fructose. -Fructose was chosen as the test saccharide as past
studies have shown that it gives the largest pKa shifts on binding
to mono-boronic acids compared to other monosaccharides.4
Fig. 3 is a plot of absorbance versus pH at 450 nm of dye 10
with and without -fructose. This observed pKa shift upon add-
ition of saccharide is believed to be caused by the association of
the boronic acid with the anilinic nitrogen.13 The plot clearly
shows a shift in this equilibrium upon addition of -fructose.
The spectra and absorbance–pH plots for all of the other dyes
can be found in the Supplementary Information.†
Fig. 2 Absorption–pH spectral changes of dye 10 (5.34 × 10Ϫ5 mol
dmϪ3) with 0.05 mol dmϪ3 -fructose. Representative spectra at pH =
6.56, 7.01, 7.46, 7.98 and 10.99.
The pKa of this interaction can be calculated from the above
data by using eqn. (1).
A = (A0 ϩ AlimKa[Hϩ])/(1 ϩ Ka[Hϩ])21,22
(1)
A is the absorbance for a particular concentration of
proton; A0 is the initial absorbance; Alim is the limiting (final)
absorbance; Ka is the acid dissociation constant; [Hϩ] is the
concentration of protons.
The pKa curves were analysed in KaleidaGraph‡ using
non-linear (Levenberg–Marquardt algorithm) curve fitting of
eqn. (1).
The errors reported are the standard errors ( σ/√N) obtained
from the best fit. The calculation was performed for all dyes
‡ KaleidaGraph Version 3.51 for PC, published by Synergy Software
and developed by Abelbeck Software, 2457 Perikiomen Avenue, Read-
ing, PA 19606. A user defined curve fit [1 ϩ m2 m1 (M0)]/[1 ϩ m1 (M0)]
derived from eqns (1) and (2) was used in all calculations. The initial
value of m1 (K) was set to 1 and the initial value of m2 (Alim) was set to
1.1. The variable (M0) was [Hϩ] for eqn. (1) and [guest] for eqn. (2). The
allowable error was set to 0.001%. For all curves the coefficient of
determination (r2) was >0.98.
Fig. 3 Absorption–pH profile of dye 10 with and without -fructose
at 450 nm. (᭿) With no -fructose, (᭹) with 0.05 mol dmϪ3 -fructose.
with and without -fructose. The results of these calculations
are summarised in Table 1.
The results show that the applied equation is valid for calc-
ulating the pKas of the dyes as the r2 values show an excellent fit
464
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470

View Article Online
(r2 > 0.99). This shift on saccharide binding (∆pKa) is in agree-
ment with previous work and can be explained by the decrease
in boron–oxygen bond angle upon saccharide binding, which
increases the acidity of the boron centre.4 All of the dyes show
similar magnitudes of ∆pKa, although dye 6 shows the largest
change.
All of the dyes gave similar absorption spectra to dye 10, with
and without -fructose. However, dye 6 (derived from 4-
nitroaniline) showed an especially marked difference in its
absorption spectra at high pH (Fig. 4 and Fig. 5).
Fig. 6 Colour–absorbance changes upon addition of -glucose to dye
6 (5.66 × 10Ϫ5 mol dmϪ3); from left to right: with no saccharide, with
0.339 × 10Ϫ2 mol dmϪ3 -glucose, with 1.66 × 10Ϫ2 mol dmϪ3 -glucose
and with 22.7 × 10Ϫ2 mol dmϪ3 -glucose.
Fig. 4 Absorption–pH spectral changes of dye 6 (5.66 × 10Ϫ5 mol
dmϪ3). Representative spectra at pH = 8.05, 9.93, 10.46 and 10.98.
Fig. 7 Absorption spectral changes of dye 6 (2.83 × 10Ϫ5 mol dmϪ3
)
with increasing concentration of -glucose at pH 11.32. Representative
spectra at [-glucose] = 0, 1.69, 3.51, 7.06, 14.2, 28.4 and 56.8 × 10Ϫ3
mol dmϪ3
.
changes upon saccharide binding, none gave as large a spectral
shift as dye 6. The stability constants (K) of the boronic acid
dye–saccharide complexes can be calculated using eqn. (2).
A = (A0 ϩ Alim K [guest])/(1 ϩ K [guest])21,22
(2)
A is the absorbance for a particular concentration of guest;
A0 is the initial absorbance; Alim is the limiting (final) absorb-
ance; K is the stability constant of the receptor with the guest;
[guest] is the concentration of the guest.
The stability constant curves were analysed in KaleidaGraph‡
using non-linear (Levenberg–Marquardt algorithm) curve
fitting of eqn. (2). The errors reported are the standard errors
( σ/√N) obtained from the best fit.
Fig. 5 Absorption–pH spectral changes of dye 6 (5.66 × 10Ϫ5 mol
dmϪ3) with 0.05 mol dmϪ3 -fructose. Representative spectra at pH =
5.09, 7.03, 8.00 and 10.97.
D. Absorption–saccharide titrations
Fig. 5 shows that at high pH, in the presence of -fructose, that
dye 6’s absorption maxima do not shift to as long a wavelength
as without saccharide. This is actually a large enough shift
(55 nm) to give a visible colour change from dark purple to red,
upon the binding of -fructose. To investigate this phenomenon
further it was decided to perform a -fructose titration with dye
6 in pH 11.32 aqueous methanolic buffer solution [52.1 wt%
methanol (KCl, 0.01000 mol dmϪ3; NaHCO3, 0.002771 mol
dmϪ3; Na2CO3, 0.002771 mol dmϪ3)].23
Fig. 8 shows the curve fit using eqn. (2) for dye 6’s interaction
This same colour change was observed. The experiment was
repeated with -glucose and ethylene glycol. The colour–
absorbance changes upon addition of -glucose to 6 are shown
in Fig. 6 and the absorption spectra of this titration are shown
in Fig. 7.
In all three cases there is a maximum wavelength shift of
approximately 55 nm upon complexation. The concentration of
the guest required to produce the change is different in each
case, which is due to the different stability constants of the
binding species. The wavelength shift obtained upon addition
of saccharide with dye 6 is the largest achieved to date in such a
system.
Fig. 8 Curve fit to determine the stability constant of dye 6 with
-glucose at 509 nm.
with -glucose. Table 2 gives the results of these calculations for
all of the boronic acid azo dyes.
The results show that dye 6, which gives the largest spectral
-Fructose and -glucose titrations at pH 11.32 were also
performed on the other dyes. Although all showed colour
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470
465

View Article Online
Table 2 Stability constants (K) for polyol–boronic acid azo dye com-
plexes at pH 11.32
Dye
Polyol
K
r2
σ-valuea
6
6
-Fructose
-Glucose
Ethylene glycol
-Fructose
-Glucose
-Fructose
-Glucose
-Fructose
-Glucose
-Fructose
-Glucose
-Fructose
-Glucose
2550 250
0.991
0.983
0.999
0.995
0.982
0.992
0.997
0.998
0.993
0.999
0.993
—
0.78
0.78
123
9
6
5.1 0.1
3630 274
184 40
0.78
7
0.38b
0.38b
0.13c
0.13c
7
8
4160 352
8
121
9
9
5180 229
163 16
Ϫ0.27
9
Ϫ0.27
0.10c
0.10c
0.78
10
10
13
13
5590 201
198 20
d
—
—
d
—
0.78
25
Ϫ
Ϫ
a See ref.
. . .
b Value for SO3 c Value for CO2 d Could not be deter-
mined from absorbance data.
change, has the lowest binding constant with -fructose, com-
pared to the other dyes. If a comparison is made to the relevant
Hammett σ-values for the phenylazo-ring substituents, a trend
is apparent for the para-substituted dyes. The binding constant
is seen to increase as the σ-value becomes less positive, i.e. the
ring substituent becomes less electron-withdrawing. Dye 10
shows the highest binding constant. Dye 10 is the only meta-
substituted dye in the series and, as such, will have an inductive
effect on the chromophoric system, rather than the conjugative
resonance influence of the para-substituted dyes.
E. Proposed equilibria
At intermediate pH values Shinkai proposes that a boron–
nitrogen interaction is prevalent, whereas at high and low pH
this bond is broken.13 What makes dye 6’s equilibria more inter-
esting is the presence of the anilinic hydrogen, which gives rise
to different species at high pH. This apparently simple modifi-
cation in the molecular structure seems to be responsible for the
enhanced response of these dyes relative to those previously
reported. Scheme 3 shows the proposed equilibrium species for
dye 6 responsible for the observed colour change, consistent
with the experimental results to date.
In the absence of saccharide, at pH 11.32, the observed col-
our is purple and in the presence of saccharide the colour is red
(Fig. 6). From previous work it is known that when saccharides
form cyclic boronate esters with boronic acids, the Lewis acidity
of the boronic acid is enhanced and therefore the Lewis acid–
base interaction between the boronic acid and the amine is
strengthened.1,6 This stronger B–N interaction will favour the
red species over the equivalent saccharide bound purple species.
The reason for this can be understood by considering species W
and X from Scheme 3. In the presence of saccharide the B–N
interaction in species X is stronger than that in species W. The
increased B–N interaction of species X will make the N–H
proton of species X more acidic than the corresponding bond in
species W. Therefore, at higher pH species X will deprotonate to
form the red species Z, whereas the weaker B–N bond in species
W is broken by hydroxide ion to form the purple coloured
species Y.
The colour change arises from the different electronic
environment of the anilinic nitrogen. As mentioned earlier, the
anilinic nitrogen is conjugated to the azo chromophore. A
change in the environment of this nitrogen leads to changes in
the energy levels of the n and π* orbitals of the azo chromo-
phore and hence to a change in the absorption energy and wave-
length. The proposal of these equilibrium species may also
explain why Shinkai’s dye did not give a visible spectral shift
on saccharide binding. Because the anilinic nitrogen is tertiary
in nature rather than secondary, there is no possibility of
deprotonation, so the high pH boron–nitrogen bond cannot be
Scheme 3 Equilibria species.
formed. Hence there is no differentiation between the equi-
librium species at high pH and consequently no spectral shift
is observed.
F. Further evidence of proposed equilibria species
It must be stressed again that the equilibria discussed here
for dye 6 are only proposals that fit the current experimental
data. It was decided to use Electrospray Ionisation Mass
Spectrometry (ESIMS) to see if any of the proposed equi-
librium species could be observed. The fructose-bound high pH
red species (Z in Scheme 3) was indeed detected (m/z 533) by
negative ion ESIMS.
If removal of the anilinic proton in the saccharide-bound
high pH species is the key to the spectral change, then replacing
this proton with an alkyl group should prevent this phen-
omenom from occurring. It was decided to synthesise the N-
methylated derivative of the nitro-substituted boronic acid azo
dye to try to prove this hypothesis. The synthetic route to dye 13
is outlined in Scheme 4. Model dye 15 (with no boronic acid
binding site) was also synthesised, for comparison purposes,
using the route shown in Scheme 1 starting from benzaldehyde
rather than 2-formylbenzeneboronic acid.
N-Methyl-m-toluidine was reacted with 1-bromo-2-bromo-
methylbenzene to form the tertiary amine 11. This was lithiated
and then reacted with trimethylborate to generate boronic acid
12, after an acidic work-up. Coupling of 12 with the diazonium
salt of 4-nitroaniline, as before, afforded the N-methylated
boronic acid azo dye molecule 13. The overall yield for the
reaction sequence was below average. However, the route has
not been optimised in any way.
The previously performed pH and saccharide titrations were
repeated with dyes 13 and 15. Unfortunately, dye 15 showed
solubility problems in the aqueous environment used for the
titrations, but, as expected, showed no spectral changes upon
addition of saccharide under these conditions. The absorption
466
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470

View Article Online
Table 3 Stability constants (K) for polyol–boronic acid azo dye com-
plexes at pH 8.21.
Dye
Polyol
K/dm3 molϪ1
r2
6
6
13
13
-Fructose
-Glucose
-Fructose
-Glucose
291 24
0.994
0.981
0.999
0.998
26
371
23
6
7
1
0.01000 mol dmϪ3; KH2PO4, 0.002752 mol dmϪ3; Na2HPO4,
0.002757 mol dmϪ3)].23 The pH of the buffer was chosen so that
it was below the pKa’s of the unbound dyes, which are typically
around 10 (Table 1). This should mean that, at the beginning
of the titrations, the major solution dye species should be that
containing the boron–nitrogen interaction. Fig. 10 and Fig. 11
Scheme 4 Synthesis of N-methylated boronic acid azo dye.
Fig. 10 Absorption spectral changes of dye 6 (2.83 × 10Ϫ5 mol dmϪ3
)
with increasing concentration of -fructose at pH 8.21. Representative
spectra at [-fructose] = 0, 0.23, 1.97, 7.17, 28.5 and 113 × 10Ϫ3 mol
spectra of the -fructose titration at pH 11.32 with 13 are
shown in Fig. 9.
dmϪ3
.
In this case, there is no shift in the wavelength upon sac-
charide complexation, only a small increase in the intensity of
absorption. Consequently dye 13 does not change colour with
added saccharide. The anilinic nitrogen of 13 is tertiary in
nature, rather than secondary as in the case of 6, so there is no
possibility of deprotonation from this nitrogen. Consequently,
Fig. 11 Absorption spectral changes of dye 13 (2.00 × 10Ϫ5 mol dmϪ3
)
with increasing concentration of -fructose at pH 8.21. Representative
spectra at [-fructose] = 0, 1.10, 2.50, 5.63, 15.8 and 112 × 10Ϫ3 mol
dmϪ3
.
show the respective absorption–spectral changes for the titra-
tions with -fructose. In both cases, as the concentration of
saccharide was increased, a spectral shift to longer wavelength
was observed. The shifts, and subsequent colour changes were
not as pronounced (< 20 nm) as the high pH titrations with
dyes 6–10.
The stability constants (K) of the boronic acid dye–saccharide
complexes can again be calculated using eqn. (2), assuming a
1 : 1 boronic acid–saccharide binding event. Table 3 shows the
results of these calculations for the two dyes. An interesting
point to note about the stability constants at pH 8.21 (Table 3)
is that they are approximately a factor of ten smaller than
those at pH 11.32 (Table 2). It is known that stability constants
measured at near neutral pH are lower than the corresponding
anionic stability constants at higher pH.24
Fig.
concentration of -fructose at pH 11.32. Representative spectra at
[-fructose = 0 and 55.6 × 10Ϫ3 mol dmϪ3
9
Absorption spectral changes of dye 13 with increasing
.
the high pH B–N bond cannot form and there can be no elec-
tronic differentiation between the high pH saccharide bound
species and the unbound species. Hence no colour change is
observed.
It was decided to carry out some saccharide titrations at
near neutral pH to investigate further the effect of pH on the
equilibria. UV–VIS absorbance titrations of dyes 6 and 13
with -fructose and -glucose were performed in a pH 8.21
aqueous methanolic buffer solution [52.1wt% methanol (KCl,
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470
467

View Article Online
Dye 6’s spectral shift at pH 8.21 is in the opposite direction to
that seen at pH 11.32. What is apparent from this titration is
that when dye 6 is complexed to saccharide, it does not matter
what the starting pH is, the final solution saccharide-bound
species is always the same, with an absorbance maximum at
509 nm. In contrast, dye 13 has only one high pH species, no
matter if it is saccharide-bound or not. As saccharide is added
to dye 13 at pH 8.21, it must form the tetrahedral boronate
anionic species. This is shown in Fig. 9 and Fig. 11 for dye
13, where the absorbance maximum of the saccharide bound
species is at 521 nm in both instances.
distilled from sodium benzophenone ketyl under a nitrogen
atmosphere. All other reagents and solvents were used as
supplied by the Aldrich Chemical Co. Ltd., Lancaster Synthesis
Ltd., Fisher Scientific Ltd., Acros Organics, Tokyo Kasei
Kogyo (TCI) Co. Ltd., Merck Ltd. and Frontier Scientific
Europe Ltd.
N-(2-Dihydroxyborylbenzylidene)-m-tolylamine (4)
m-Toluidine (0.72 g, 6.67 mmol) and 2-formylbenzeneboronic
acid (1.00 g, 6.67 mmol) were mixed in absolute ethanol (45 ml)
and toluene (5 ml). A Dean–Stark trap was fitted to permit the
azeotropic removal of water, and the reaction mixture heated
at reflux overnight (18 hours). After cooling, the solvent was
removed in vacuo to obtain 4 (1.35 g, 92%) as a cream coloured
Conclusions
The development of a facile and clean synthesis of boronic acid
azo dyes has been achieved. The boronic acid azo dyes show
visible colour change upon saccharide complexation. The
absorbance maximum of boronic acid azo dye 6 shifts by 55 nm
to shorter wavelength upon addition of diols at high pH–the
largest observed to date in such a system. Comparison with
N-methylated dye 13 confirms that removal of the anilinic
nitrogen proton in 6, when bound with saccharide, gives rise to
the pronounced colour change.
solid: mp 170–172 ЊC (dec.); νmax(Nujol)/cmϪ1 1732 (CH᎐N);
᎐
δH(300 MHz; CD3OD; Me4Si) 2.41 (3 H, s, Ar–CH3), 7.24 (1 H,
d, J 6.0, Ar–H ), 7.34–7.46 (2 H, m, Ar–H ), 7.51–7.66 (4 H, m,
Ar–H ), 7.70 (1 H, d, J 6.0, Ar–H ), 9.12 (1 H, s, ArN᎐CHAr);
᎐
δC(75 MHz; CD3OD; Me4Si) 21.9, 120.3, 123.8, 129.6, 129.9,
130.9, 131.1, 131.5, 134.9, 135.1, 140.0, 141.1, 143.9, 166.6; m/z
(LSIMS) 509 ([M Ϫ 2 H2O ϩ 2 NOBA]ϩ, 22%), 357 ([M ϩ 1
NOBA Ϫ H2O]ϩ, 100) and 222 ([M Ϫ H2O]ϩ, 17).
N-(2-Dihydroxyborylbenzyl)-m-tolylamine (5)
Experimental
Sodium borohydride (0.79 g, 20.9 mmol, 5 equiv.) was added
slowly to 4 (1.00 g, 4.18 mmol) in dry methanol (50 ml). The
reaction was left to stir at room temperature for 2 hours and
then poured into ice–water (20 ml). Hydrochloric acid (1 M,
10 ml) was added slowly and the mixture left to stir for 30 min-
utes, whereupon the precipitate formed was collected by suction
filtration to afford 5 (0.82 g, 81%) as a white solid (Found: C,
69.6; H, 6.4; N, 5.8. C14H16BNO2 requires C, 69.7; H, 6.7; N,
5.8%); mp 72–73 ЊC (dec.); δH(300 MHz; CD3OD; Me4Si) 2.20
(3 H, s, Ar–CH3), 4.29 (2 H, s, ArCH2–NHAr), 6.54–6.66 (4 H,
m, Ar–H ), 6.99 (1 H, t, J 7.5, Ar–H ), 7.15–7.32 (4 H, m,
Ar–H ); δC(75 MHz; CD3OD; Me4Si) 22.3, 52.2, 115.1, 118.6,
122.9, 127.1, 128.2, 130.0, 130.3, 133.0, 140.3, 146.1, 149.0; m/z
(LSIMS) 241 ([M]ϩ, 40%), 223 ([M Ϫ H2O]ϩ, 100).
1H and 13C NMR spectra were recorded on Bruker AC-300 and
Bruker AC-400 spectrometers. All spectral data are reported
relative to tetramethylsilane as the internal standard using the δ
scale. The multiplicities of the spectroscopic data are presented
in the following manner: s-singlet, d-doublet, m-multiplet,
t-triplet, br-broad. J-values are given in Hertz. Due to quadra-
polar relaxation, the aryl carbon atoms attached directly to
the boronic acid boron atoms are not observed by 13C NMR.
Infrared spectra were recorded on a Perkin–Elmer Paragon
1000 FT-IR spectrometer and a Perkin–Elmer 1600 Series
FT-IR spectrometer. Samples were run either as a Nujol mull or
as pressed KBr disks.
Electron Impact (EI) mass spectra were recorded on a VG
ProSpec mass spectrometer. Liquid Secondary Ion (LSI) mass
spectra were recorded on a VG ZabSpec instrument. A Micro-
mass LCT mass spectrometer was used for Electrospray Ionis-
ation (ESI) mass spectra. Fast Atom Bombardment (FAB) and
EI mass spectra were also recorded on a double focussing,
magnetic sector Micromass Autospec instrument, using a
DEC Alpha data system running Opus V3.4X software,
overlaid on an Open VMS operating system. High Resolution
Mass Spectra (HRMS) were obtained from all of the above
instruments.
Elemental analyses were performed at the University of North
London, the University of Birmingham and the University of
Bath. Melting points were measured on Gallenkamp and Büchi
535 melting point apparatus and are uncorrected.
UV–Visible absorption measurements used Sigma Spectro-
photometer Silica (Quartz) Cuvets with 10 mm path lengths
and were recorded on a Perkin–Elmer Lambda 20 UV/VIS
Spectrophotometer. Data was collected via the Perkin–Elmer
UV Winlab software package on a PC running Microsoft
NT v4. All pH measurements were recorded on a Hanna
Instruments HI 9321 Microprocessor pH Meter that was cal-
ibrated using Fisher Chemicals standard buffer solutions (pH
4.0–phthalate, 7.0–phosphate, and 10.0–borate).
N-(2-Dihydroxyborylbenzyl)-[3-methyl-4-(4-nitrophenylazo)-
phenyl]amine (6)
4-Nitroaniline (0.10 g, 0.70 mmol) was mixed in water (1 ml),
methanol (1 ml) and hydrochloric acid (1 ml, 5.0 M) and
then cooled to 0–5 ЊC on an ice-bath. A chilled solution of
sodium nitrite (0.06 g, 0.84 mmol, 0.2 M) was added dropwise.
Excess nitrite was destroyed by the addition of sulfamic acid
after stirring for 5 minutes. N-(2-Dihydroxyborylbenzyl)-m-
tolylamine (0.16 g, 0.66 mmol) was dissolved in methanol
(2 ml) and dilute hydrochloric acid (1 ml, 1 M) then added
dropwise to the reaction mixture, which quickly turned a red
colour. Sodium acetate was added to raise the pH of the solu-
tion to 4 and this was then left to stir at 0–5 ЊC for 3 hours.
Sodium hydroxide (2 M) was slowly added to raise the pH to 7.
The resulting precipitate was collected by suction filtration and
dried in a dessicator overnight to afford 6 (0.19 g, 74%) as a
dark red solid (HRMS: Found 372.1381, [M Ϫ H2O]ϩ.
C20H17BN4O3 requires 372.1392); mp 120–122 ЊC (dec.);
λmax(pH 8.05, 0.05 mol dmϪ3 NaCl, 33% MeOH: 67% H2O
w/w)/nm 486 (ε/dm3 molϪ1 cmϪ1 19000); νmax(KBr disk)/cmϪ1
1602 (N᎐N), 1518 and 1333 (NO ); δ (300 MHz; CD OD;
᎐
2
H
3
Thin-layer chromatography (TLC) was performed on pre-
coated aluminium-backed silica gel plates supplied by Merck
Ltd. (Silica Gel 60 F254, thickness 0.2 mm, Art. 5554). Visualis-
ation was achieved by UV light (254 nm).
Acetonitrile was dried by refluxing with calcium hydride. It
was subsequently distilled and collected by dry syringe as
required. Prior to use, tetrahydrofuran and diethyl ether were
Me4Si) 2.63 (3 H, m, Ar–CH3), 4.45 (2 H, s, ArCH2–NHAr),
6.52–6.58 (1 H, m, Ar–H ), 6.99–7.04 (1 H, m, Ar–H ), 7.18–
7.40 (5 H, m, Ar–H ), 7.63–7.72 (1 H, m, Ar–H ), 7.87–7.93
(1 H, m, Ar–H ), 8.21–8.36 (2 H, m, Ar–H ); δC(125 MHz;
CD3OD; Me4Si) 18.1, 49.7, 113.2, 114.7, 118.3, 122.7, 123.6,
125.7, 127.6, 128.1, 129.8, 130.2, 132.8, 144.3; m/z (EI) 373
([MH ϪH2O]ϩ, 66%), 222 ([M Ϫ H2O–N2C6H4NO2]ϩ, 100).
468
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470

View Article Online
4-[4-(2-Dihydroxyborylbenzylamino)-2-methylphenylazo]-
benzenesulfonic acid (7)
125.4, 128.2, 128.6, 129.0, 130.8, 131.2, 132.2, 133.7, 133.7,
143.8, 145.1, 145.8, 154.6, 155.9, 170.9; m/z (ESI) 418 ([MH Ϫ
2H2O ϩ 2CH3OH]ϩ, 100%).
The above procedure was repeated with sulfanilic acid (0.08 g,
0.44 mmol) in place of p-anisidine (all other reagents were used
in the same mole ratios), to afford 7 (0.05 g, 27%) as a dark red
solid (HRMS: Found 424.1142, [M Ϫ H]ϩ. C20H19BN3O5S
requires 424.1138); mp 219 ЊC (dec.); λmax(pH 8.18, 0.05 mol
dmϪ3NaCl, 33% MeOH: 67% H2O w/w)/nm 421 (ε/dm3 molϪ1
N-(2-Bromobenzyl)-N-methyl-m-tolylamine (11)
2-Bromobenzyl bromide (10.0 g, 40.0 mmol) was mixed in dis-
tilled acetonitrile (100 ml) with potassium carbonate (16.6 g,
120 mmol, 3 equiv.). The system was evacuated and put under
an argon atmosphere. N-Methyl-m-toluidine (4.8 g, 40.0 mmol)
was syringed into solution and the reaction mixture heated to
reflux. After 4 hours, TLC analysis showed the reaction had
gone to completion. The solvent was removed, after filtration,
to yield 11 as a yellow oil (2.22 g, 98%); δH(300 MHz; CDCl3;
Me4Si) 2.53 (3 H, m, Ar–CH3), 3.19 (3 H, s, N-CH3), 4.69 (2 H,
s, ArCH2–NMeAr), 6.60–6.72 (3 H, m, Ar–H ), 7.19–7.39 (4 H,
m, Ar–H ), 7.67–7.71 (1 H, m, Ar–H ); δC(125 MHz; CDCl3;
Me4Si) 22.1, 38.8, 57.5, 109.3, 112.7, 117.7, 122.8, 127.6, 128.0,
128.4, 129.2, 132.9, 137.6, 139.0, 149.4; m/z (EI) 289 ([M]ϩ,
39%), 91 ([C6H5CH2]ϩ, 100).
cmϪ1 19300); νmax(KBr disk)/cmϪ1 1610 (N᎐N), 1350 and 1150
᎐
(SO2); δH(300 MHz; CD3OD; Me4Si) 2.59 (3 H, m, Ar–CH3),
4.41 (2 H, s, ArCH2–NHAr), 6.49–6.59 (2 H, m, Ar–H ), 7.19–
7.41 (4 H, m, Ar–H ), 7.59–7.66 (1 H, m, Ar–H ), 7.81 (2 H, d,
J 9.0, Ar–H ), 7.92 (2 H, d, J 9.0, Ar–H ); δC(125 MHz; CD3OD;
Me4Si) 18.0, 49.7, 113.1, 113.7, 115.1, 117.9, 122.9, 127.5, 127.9,
129.8, 132.7, 146.6, 153.7, 155.7; m/z (ESI) 424 ([M Ϫ H]ϩ,
33%), 406 ([M Ϫ H3O]ϩ, 100).
4-[4-(2-Dihydroxyborylbenzylamino)-2-methylphenylazo]benzoic
acid (8)
The above procedure was repeated using 4-aminobenzoic acid
(0.11 g, 0.83 mmol) instead of 3-aminobenzoic acid (all other
reagents were used in the same mole ratios), to afford 8 (0.10 g,
31%) as an orange–red solid (Found: C, 64.8; H, 5.0; N, 10.7.
C21H20BN3O4 requires C, 64.8; H, 5.2; N, 10.8%); mp 208–
210 ЊC (dec.); λmax(pH 8.42, 0.05 mol dmϪ3 NaCl, 33% MeOH:
67% H2O w/w)/nm 418 (ε/dm3 molϪ1 cmϪ1 16000); νmax(KBr
disk)/cmϪ1 1602 (N᎐N), 1684 (C᎐O); δ (300 MHz; CD OD;
Me4Si) 2.60 (3 H, m, Ar–CH3), 4.42 (2 H, s, ArCH2–NHAr),
6.50–6.61 (2 H, m, Ar–H ), 7.18–7.41 (4 H, m, Ar–H ), 7.61–
7.68 (1 H, m, Ar–H ), 7.82 (2 H, d, J 9.0 Ar–H ), 8.11 (2 H, d,
J 9.0 Ar–H ); δC(125 MHz; CD3OD; Me4Si) 18.1, 49.6, 117.9,
122.8, 127.5, 131.6; m/z (LSIMS) 660 ([MH Ϫ 2H2O ϩ 2
NOBA]ϩ, 55%), 460 ([M Ϫ 2H2O ϩ 2CH3OH ϩ K]ϩ, 100).
N-(2-Dihydroxyborylbenzyl)-N-methyl-m-tolylamine (12)
N-(2-Bromobenzyl)-N-methyl-m-tolylamine (5.00 g, 17.2 mmol)
was mixed in distilled tetrahydrofuran (40 ml) under argon. The
solution was cooled to Ϫ78 ЊC and then n-butyllithium (10.3 ml,
2.5 M in THF, 25.8 mmol, 1.5 equiv.) was syringed slowly into
the reaction mixture. After stirring at Ϫ78 ЊC for 1 hour, the
lithiated solution was transferred dropwise, using a canula, into
dry THF (40 ml) containing trimethylborate (8.9 g, 86.0 mmol,
5 equiv.) at Ϫ78 ЊC under argon. The reaction was allowed to
warm to room temperature before adding hydrochloric acid
(40 ml, 1 M). After stirring for 1 hour, water (50 ml) was added
and the resulting aqueous layer washed with hexane (2 × 50 ml)
and salted with sodium chloride. The aqueous phase was
extracted with dichloromethane (3 × 50 ml) and the organic
layers combined, dried over MgSO4 and filtered. The solvent
was removed in vacuo and 12 (1.08 g, 25%) was afforded by
precipitation from chloroform with hexane as a white solid
(HRMS: Found 256.1501; C15H19BNO2 requires 256.1509);
mp 61–63 ЊC (dec.); δH(300 MHz; CD3OD; Me4Si) 2.41 (3 H, s,
Ar–CH3), 3.28 (3 H, s, Ar–CH3), 4.90 (2 H, s, ArCH2–NHAr),
7.26–7.49 (8 H, m, Ar-H ); δC(125 MHz; CD3OD; Me4Si) 22.0,
45.7, 66.3, 120.3, 123.6, 131.0, 131.8, 132.1, 132.7, 134.4, 136.2,
137.2, 142.9, 143.3; m/z (ESI) 256 ([MH]ϩ, 100%).
᎐
᎐
H
3
N-(2-Dihydroxyborylbenzyl)-[4-(4-methoxyphenylazo)-3-methyl-
phenyl]amine (9)
The previous experiment was repeated using p-anisidine (0.61 g,
4.98 mmol, 6 equiv.) instead of 4-aminobenzoic acid. Purifi-
cation was performed by washing the product (in dichloro-
methane) with 10% sodium hydrogen carbonate solution (w/w)
to remove residual acetic acid, to yield 9 (0.11 g, 35%) as a
orange–brown solid (Found: C, 67.3; H, 5.8; N, 10.2.
C21H22BN3O3 requires C, 67.2; H, 5.9; N, 11.2%); mp 102–
103 ЊC (dec.); λmax(pH 8.42, 0.05 mol dmϪ3 NaCl, 33% MeOH:
67% H2O w/w)/nm 400 (ε/dm3 molϪ1 cmϪ1 14600); νmax(KBr
N-(2-Dihydroxyborylbenzyl)-N-methyl-[3-methyl-4-(4-nitro-
phenylazo)phenyl]amine (13)
disk)/cmϪ1 1602 (N᎐N); δ (300 MHz; CD OD; Me Si) 2.56
᎐
H
3
4
(3 H, m, Ar–CH3), 3.84 (3 H, s, Ar–OCH3), 4.39 (2 H, s,
ArCH2–NHAr), 6.49–6.59 (2 H, m, Ar–H ), 7.00 (2 H, d, J 9.0,
Ar–H ), 7.20–7.40 (4 H, m, Ar–H ), 7.52–7.57 (1 H, m, Ar–H ),
7.76 (2 H, d, J 9.0, Ar–H ); δC(125 MHz; CD3OD; Me4Si) 18.0,
49.7, 56.0, 113.8, 115.2, 115.5, 117.5, 124.8, 127.5, 127.8, 129.7,
132.6, 141.3, 144.3, 145.0, 149.0, 152.5, 162.5; m/z (ESI) 376
([MH]ϩ, 100%).
4-Nitroaniline (0.05 g, 0.39 mmol) was mixed in water (1 ml),
methanol (1 ml) and hydrochloric acid (1 ml, 10.0 M) and
then cooled to 0–5 ЊC on an ice-bath. A chilled solution of
sodium nitrite (0.03 g, 0.43 mmol, 0.2 M) was added dropwise.
Excess nitrite was destroyed by the addition of sulfamic acid.
N-(2-Dihydroxyborylbenzyl)-N-methyl-m-tolylamine (0.10 g,
0.39 mmol) was dissolved in methanol (2 ml) and dilute hydro-
chloric acid (1 ml, 1 M), then added dropwise to the reaction
mixture, which quickly turned a red colour. Sodium acetate was
added to raise the pH of the solution to 4 and this was then left
to stir at 0–5 ЊC for 3 hours. Sodium hydroxide (1 M) was slowly
added to raise the pH to 7. The resulting precipitate was col-
lected by suction filtration and dried in a dessicator overnight to
afford 13 (0.13 g, 82%) as a dark red solid (Found: C, 62.8; H,
5.2; N, 13.1. C21H21BN4O4 requires C, 62.4; H, 5.2; N, 13.8%)
(HRMS: Found 405.1739, [MH]ϩ. C21H21BN4O4 requires
405.1734); mp 115–120 ЊC (dec.); λmax(pH 8.21, 0.01 mol dmϪ3
KCl, 0.002752 mol dmϪ3 KH2PO4, 0.002757 mol dmϪ3
Na2HPO4, 52.1% MeOH: 47.9% H2O w/w)/nm 501 (ε/dm3
3-[4-(2-Dihydroxyborylbenzylamino)-2-methylphenylazo]benzoic
acid (10)
The previous experiment was repeated using 3-aminobenzoic
acid (0.06 g, 0.41 mmol) instead of 4-nitroaniline (all other
reagents were used in the same mole ratios), to yield 10 (0.10 g,
61%) as a red solid (Found: C, 64.9; H, 5.2; N, 10.9.
C21H20BN3O4 requires C, 64.8; H, 5.2; N, 10.8%); mp 133–
135 ЊC (dec.); λmax(pH 8.48, 0.05 mol dmϪ3 NaCl, 33% MeOH:
67% H2O w/w)/nm 406 (ε/dm3 molϪ1 cmϪ1 19000); δmax(KBr
disk)/cmϪ1 1603 (N᎐N); δ (300 MHz; CD OD; Me Si) 2.62
᎐
H
3
4
(3 H, m, Ar–CH3), 4.40 (2 H, s, ArCH2–NHAr), 6.50–6.60 (2 H,
m, Ar–H ), 7.18–7.42 (4 H, m, Ar–H ), 7.52–7.66 (2 H, m,
Ar–H ), 7.94–8.05 (2 H, m, Ar–H ), 8.37–8.41 (1 H, m, Ar–H );
δC(125 MHz; CD3OD; Me4Si) 19.1, 52.5, 114.2, 118.9, 123.1,
molϪ1 cmϪ1 29300); νmax(KBr disk)/cmϪ1 1599 (N᎐N), 1512
᎐
and 1333 (NO2); δH(300 MHz; CD3OD; Me4Si) 2.68 (3 H, s,
Ar–CH3), 3.11 (3 H, s, Ar-CH3), 4.66 (2 H, s, ArCH2–NHAr),
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470
469

View Article Online
6.72–6.79 (2 H, m, Ar–H ), 7.19–7.40 (5 H, m, Ar–H ), 7.70–
7.76 (1 H, m, Ar–H ), 7.91 (2 H, d, J 12.0, Ar–H ), 8.32 (2 H, d,
J 12.0, Ar–H ); δC(125 MHz; CD3OD; Me4Si) 18.5, 40.2, 58.3,
113.0, 115.4, 118.3, 123.5, 125.7, 127.7, 127.9, 128.9, 130.1,
133.0, 142.8, 143.9, 148.8, 155.5, 158.1; m/z (ESI) 405 ([MH]ϩ,
100%).
Absorption–saccharide titrations of 6, 7, 8, 9, 10 and 13 at pH
11.32
The UV–visible absorption spectra of 6 (2.83 × 10Ϫ5 mol dmϪ3),
7 (3.26 × 10Ϫ5 mol dmϪ3), 8 (2.74 × 10Ϫ5 mol dmϪ3), 9 (5.86 ×
10Ϫ5 mol dmϪ3), 10 (4.27 × 10Ϫ5 mol dmϪ3) and 13 (2.09 ×
10Ϫ5 mol dmϪ3) in a pH 11.32 buffer (0.01000 mol dmϪ3 KCl,
0.002771 mol dmϪ3 NaHCO3, 0.002771 mol dmϪ3 Na2CO3
in 52.1 wt% methanol: 47.9 wt% water),23 were recorded as
increasing amounts of various saccharides were added to the
solution.
N-Benzyl-m-tolylamine (14)
Benzaldehyde (1.04 g, 9.84 mmol) and m-toluidine (1.05 g,
9.84 mmol) were mixed in toluene (50 ml) and heated at reflux
for 15 hours in a Dean–Stark apparatus. After cooling, the
solvent was removed in vacuo to afford a yellow oil, which was
used directly in the next reaction step without further purifi-
cation. Sodium borohydride (0.74 g, 19.7 mmol) was added in
small portions to a stirred solution of the imine from the
previous step in methanol (40 ml) at room temperature. TLC
showed completion of the reaction after 2 hours. The reaction
mixture was poured on to ice (5 g) and concentrated hydro-
chloric acid (1 ml) and stirred for 30 minutes, before removing
most of the solvent in vacuo. Water (30 ml) was added and then
the aqueous layer was extracted with dichloromethane. The
organic extracts were combined, dried (MgSO4) and filtered.
The solvent was removed in vacuo to yield 14 (1.87 g, 96%)
as a brown oil (HRMS: Found 197.1197. C14H15N requires
197.1205); δH(300 MHz; CDCl3; Me4Si) 2.37 (3 H, s, Ar–CH3),
4.04 (1 H, br s, NH ), 4.39 (2 H, s, ArCH2–NHAr), 6.49–6.70
(3 H, m, Ar–H ), 7.16 (1 H, t, J 6.0, Ar–H ), 7.33–7.50 (4 H, m,
Ar–H ); δC(75 MHz; CDCl3; Me4Si) 21.7, 48.3, 109.9, 113.7,
118.5, 127.18, 127.5, 128.6, 129.2, 139.0, 139.6, 148.2; m/z (EI)
197 ([M]ϩ, 60%), 91 ([C6H5CH2]ϩ, 100).
Absorption–saccharide titrations of 6 and 13 at pH 8.21
The UV–visible absorption spectra of 6 (2.83 × 10Ϫ5 mol dmϪ3
)
and 13 (2.00 × 10Ϫ5 mol dmϪ3) in a pH 8.21 buffer (0.01000 mol
dmϪ3 KCl, 0.002752 mol dmϪ3 KHPO4, 0.002757 mol dmϪ3
Na2HPO4 in 52.1 wt% methanol: 47.9 wt% water),23 were
recorded as increasing amounts of various saccharides were
added to the solution.
Acknowledgements
T. D. J. wishes to acknowledge the Royal Society, the EPSRC,
and Beckman-Coulter for support. C. J. W. wishes to acknow-
ledge the EPSRC and Avecia Limited for support through the
award of a Studentship. We would also like to acknowledge the
support of the University of Bath.
References
1 J. H. Hartley, T. D. James and C. J. Ward, J. Chem. Soc., Perkin
Trans. 1, 2000, 3155.
N-Benzyl-N-[3-methyl-4-(4-nitrophenylazo)phenyl]amine (15)
2 A. P. Davis and R. S. Wareham, Angew. Chem., Int. Ed., 1999, 38,
The procedure to form 6 was repeated using benzyl-m-
tolylamine (0.20 g, 1.01 mmol) instead of N-(2-dihydroxyboryl-
benzyl)-m-tolylamine (all other reagents were used in the same
mole ratios). Precipitation from chloroform with hexane
afforded analytically pure 15 (0.08 g, 23%) as a dark red solid
(Found: C, 69.1; H, 5.4; N, 16.3. C20H18N4O2 requires C, 69.3; H
5.2; N, 16.2%); mp 147–149 ЊC (dec.); νmax(KBr)/cmϪ1 1604
2978.
3 T. D. James, K. R. A. S. Sandanayake and S. Shinkai, Angew. Chem.,
Int. Ed. Engl., 1996, 35, 1911.
4 J. P. Lorand and J. O. Edwards, J. Org. Chem., 1959, 24, 769.
5 T. D. James, K. R. A. S. Sandanayake and S. Shinkai, Supramol.
Chem., 1995, 6, 141.
6 T. D. James, P. Linnane and S. Shinkai, Chem. Commun., 1996, 281.
7 H. R. Snyder and C. Weaver, J. Am. Chem. Soc., 1948, 70, 232.
8 H. S. Snyder and S. L. Meisel, J. Am. Chem. Soc., 1948, 70, 774.
9 H. Gilman, L. Santucci, D. R. Swayampi and R. O. Ranck, J. Am.
Chem. Soc., 1957, 79, 2898.
10 N. Nagasaki, H. Shinmori and S. Shinkai, Tetrahedron Lett., 1994,
35, 2201.
11 M. Takeuchi, M. Taguchi, H. Shinmori and S. Shinkai, Bull. Chem.
Soc. Jpn., 1996, 69, 2613.
12 C. J. Davis, P. T. Lewis, M. E. McCarroll, M. W. Read, R. Cueto and
R. M. Strongin, Org. Lett., 1999, 1, 331.
13 K. R. A. S. Sandanayake and S. Shinkai, J. Chem. Soc., Chem.
Commun., 1994, 1083.
(N᎐N), 1507 and 1328 (NO ); δ (300 MHz; CDCl ; Me Si) 2.67
᎐
2
H
3
4
(3 H, m, Ar–CH3), 4.43 (2 H, d, J 6.0, ArCH2–NHAr), 4.71
(1 H, t, J 6.0, NH ), 6.47–6.55 (2 H, m, Ar–H ), 7.25–7.38 (5 H,
m, Ar–H ), 7.73–7.78 (1 H, m, Ar–H ), 7.86–7.92 (2 H, m,
Ar–H ), 8.27–8.33 (2 H, m, Ar-H ); δC(75 MHz; CDCl3; Me4Si)
18.3, 47.90, 111.5, 113.3, 117.9, 122.9, 124.9, 127.7, 127.9,
129.1, 138.3, 143.4, 143.7, 147.5, 152.2, 157.2; m/z (EI) 346
([M]ϩ, 42%), 196 ([M Ϫ N2C6H4NO2]ϩ, 38), 91 ([C6H5CH2]ϩ,
100).
14 G. Wulff, Pure Appl. Chem., 1982, 54, 2093.
15 H. Shinmori, M. Takeuchi and S. Shinkai, J. Chem. Soc., Perkin
Trans. 2, 1998, 847.
16 K. Koumoto and S. Shinkai, Chem. Lett., 2000, 856.
17 K. Koumoto, M. Takeuchi and S. Shinkai, Supramol. Chem., 1998,
9, 203.
18 P. Gordon and P. Gregory, Organic Chemistry in Colour, Springer-
Verlag, 1987.
19 H. Zollinger, Color Chemistry, VCH, Weinheim, 1991.
20 C. S. De Ligny and M. Rehbach, Recl. Trav. Chim. Pays-Bas, 1960,
79, 727.
21 B. Valeur, J. Pouget, J. Bourson, M. Kaschke and N. P. Ernsting,
J. Phys. Chem., 1992, 96, 6545.
22 C. R. Cooper and T. D. James, J. Chem. Soc., Perkin Trans. 1, 2000,
963.
23 D. D. Perrin and B. Dempsey, Buffers for pH and Metal Ion Control,
Chapman & Hall, 1974.
24 J. H. Hartley and T. D. James, Tetrahedron Lett., 1999, 40, 2597.
25 Advances in Linear Free Energy Relationships, eds. N. B. Chapman
& J. Shorter, Plenum Press, London, 1972.
Absorption–pH titration of 6, 7, 8, 9, 10 and 13
The UV–visible absorption spectra of 6 (5.66 × 10Ϫ5 mol dmϪ3),
7 (4.66 × 10Ϫ5 mol dmϪ3), 8 (5.47 × 10Ϫ5 mol dmϪ3), 9 (7.33 ×
10Ϫ5 mol dmϪ3) and 10 (5.34 × 10Ϫ5 mol dmϪ3) in a 0.05 mol
dmϪ3 NaCl 33% methanol: 67% water (w/w) solution, were all
recorded as the pH was changed from pH 2 to 12 in approxi-
mate intervals of 0.5 pH units. The pH was controlled using
minimum volumes of sodium hydroxide and hydrochloric acid
solutions. The same experiment was repeated with 13 (1.86 ×
10Ϫ5 mol dmϪ3) in a 0.05 mol dmϪ3 NaCl 67% methanol: 33%
water (w/w) solution.
Absorption–pH titrations of 6, 7, 8, 9, 10 and 13 in the presence
of D-fructose
The previous experiment was repeated with the addition of
0.05 mol dmϪ3 -fructose to the solution in all cases.
470
J. Chem. Soc., Perkin Trans. 1, 2002, 462–470