Solvent effect and amine interference on colorimetric changes of azobenzene-conjugated dithiaazadioxo crown ether mercury sensor

Article abstract of DOI:10.1016/j.tetlet.2013.10.013

The colorimetric behavior of an azobenzene-based dye containing dithiaazadioxo ring for Hg²⁺-detection was investigated in diverse solvents. The mercury specific ligand shows hypsochromic changes in pure and aqueous MeCN upon Hg²⁺-bi

Full text of DOI:10.1016/j.tetlet.2013.10.013

Tetrahedron Letters 54 (2013) 6841–6847
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Solvent effect and amine interference on colorimetric changes
of azobenzene-conjugated dithiaazadioxo crown ether mercury
sensor
Chang Hoon Jeon a,b, Jayeon Lee a,b, Sang Jung Ahn c, Tai Hwan Ha a,b,
⇑
a Nanobiotechnology (Major), School of Engineering, University of Science & Technology Centre, Yuseong-gu, Daejeon 305-350, Republic of Korea
b Research Center of Integrative Cellulomics, Korea Research Institute of Bioscience and Biotechnology (KRIBB), Yuseong-gu, Daejeon 305-806, Republic of Korea
c Center for Nano-Imaging Technology, Division of Industrial Metrology, Korea Research Institute of Standards and Science (KRISS), Yuseong-gu, Daejeon 305-340, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
The colorimetric behavior of an azobenzene-based dye containing dithiaazadioxo ring for Hg2+-detection
was investigated in diverse solvents. The mercury specific ligand shows hypsochromic changes in pure
and aqueous MeCN upon Hg2+-binding while strong bathochromic shift was observed in chloroform. In
both solvent systems, dithiaazadioxo ring is revealed to be indispensable to the selective detection of
mercury ion, and several binding schemes in different solvents were proposed on the selective Hg2+ bind-
ing. Grafted on cationic polymer (PEI) through amide coupling, it was observed that the ligand basically
maintains the Hg2+ detection capability, but substantially hindered by amine moieties in PEI.
Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved.
Article history:
Received 30 August 2013
Revised 30 September 2013
Accepted 3 October 2013
Available online 11 October 2013
Keywords:
Colorimetric mercury sensor
Dithiaazadioxo crown ether
Azobenzene
Heavy metal ions
Polyethylene imine
Push–pull dipole
Introduction
alternative routes for mercury detection with high sensitivity.
Most of fluorogenic or colorimetric dyes largely consist of
Highly sensitive detection of mercury ion (Hg2+) is of great
importance since diverse mercury compounds are emitted from
many industries to surrounding milieu as byproducts or wastes,
thereby being an environmentally threatening agent for the public
health even in extremely low concentrations. Due to its highly
toxic nature toward human body, mercury contents of foods and
water are strictly regulated in most countries; it is reported to be
related with many health problems in the central nerve systems,
kidney, and endocrine systems.1 Therefore, sensitive and cost-
effective detection methods to monitor the concentrations of
Hg2+ ion have been highly demanding and have attracted much re-
search efforts, which might eventually enable ordinary consumers
to measure the content of the hazardous ions. Spurred by these de-
mands, colorimetric mercury sensors have been enthusiastically
pursued, and many specialized organic ligands were devised,
showing some diversities in terms of specific affinity for mercury
ions, cost-effectiveness, and accessibility of synthetic routes.2
Aside from many organic ligands, oligonucleotides tethered on
gold nanoparticles3 or electrochemical detections4 have provided
mercury-binding sites including azathiacrown,5 dithiacarbamate,6
or thiooxaaza macrocycles7 and chromogenic centers like
squarine,8 azobenzene,5c,9 or rhodamine moieties.10 An excellent
review Letter can be found elsewhere, which is helpful for system-
atic understanding of recent developments in optical metal ion
detections.2 Herein, a synthesis of azobenzene-based and colori-
metric mercury sensor is reported and its distinct color changes
on solvent variation are investigated; dithiaazadioxo crown ether,
as a mercury capture ligand, is linked to the core chromophore. The
sensing molecules are also attempted to covalently tether onto a
polymer backbone, which might enable one to reuse the colorimet-
ric indicator on polymer resins. Upon immobilization, the colori-
metric indicator showed substantially different behaviors and the
effects of polymer grafting are discussed (see Fig. 1).
Results and discussion
Dye design and synthesis for Hg2+ detection
As shown in Scheme 1, initial N-phenyl compound 1 was con-
verted into dithiaazadioxo crown ether 2 through the coupling
reaction between the dimesylate and corresponding dithiol, which
⇑
Corresponding author.
E-mail address: taihwan@kribb.re.kr (T.H. Ha).
0040-4039/$ - see front matter Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.10.013

6842
C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
Figure 1. The absorbance spectra and optical images of various metal ions (100 lM) with L3 (50 lM) (a) in a mixed solvent of acetonitrile/water (1:1, v/v), (b) in neat
acetonitrile, and (c) in chloroform.
Figure 2. The absorbance spectra and optical images of L3 (50
lM) upon the gradual addition of Hg(ClO4)2 (a) in acetonitrile/water (4:1, v/v), (b) in neat acetonitrile, and (c)
in chloroform.
dithiaazadioxo ring were also provided. The dyes adopting
azobenzenes are mostly colorimetric types and their studies were
performed in organic solvents like MeCN or methanol because of
the low solubility in aqueous solution. In the meantime, some con-
troversial observations are noteworthy in those dyes with the
dithiaazadioxo rings as Hg2+ capture ligand, which motivated us
to investigate them in detail. For instance, an anthracene-based
dye (D4 in Supporting information) showed preferential selectivity
for Ag+ ions instead of Hg2+ ions,11 or some dyes with the same ring
showed additional responses for other metal ions like Cu2+ or Fe2+
in conjunction with Hg2+ ions while others did not.9,12 Moreover,
similar fluorogenic dyes containing the same tetrathiaaza crown
ether (NS4 ring) demonstrated different selectivity for heavy metal
ions. For instance, NS4 ring with a xanthene derivative showed a
preference for Hg2+ ions5b while the same ring with a naphthali-
mide chromophore preferred to interact with Ag+ ions.5e
Scheme 1. A schematic cartoon depicting the synthesis of dithiaazadioxo crown
ether, subsequent conjugation with 4-aminobenzoic acid as a mercury sensor, and
further conjugation with ethanolamine via amide formation.
Solvatochromic observations upon Hg2+ addition
In the first place, the colorimetric changes of the L3, responding
was followed by the in-situ intramolecular ring-closure reaction as
described in the literature; the overall yield was ꢀ28%.5d In the fol-
lowing synthesis of azobenzene chromophore, diazonium salt that
had been prepared in advance was coupled to the compound 2,5c
yielding the final carboxylated Hg2+-binding chromophore L3
(ꢀ65% yield). The major difference with previous reports from
other groups is the fact that the current chromophore bears a car-
boxylate group in the para-position instead of nitro group, enabling
it to be tethered onto solid or polymeric substrates as a platform
for advanced mercury sensor in the future.
to diverse metal ions, were examined by dissolving it (50 lM) in
excessive metal solutions (ꢀ2 equiv). The internal charge transfer
(ICT) band of L3 centered at 450 nm shifted to 350 nm only for
Hg2+ ion in aqueous MeCN solution, which yields apparently losing
the typical yellow color of L3 upon binding with Hg2+ ions. The
hypsochromic shift is largely coincident with previous works
employing azobenzene chromophores,5c and is mostly attributed
to the mitigation of electron donating ability with the amine moi-
ety in the dithiaazadioxo ring. As for Ag+ and Cu2+ ions, the band
intensity slightly diminished, suggesting some interactions with
the ligand. In fact, it was previously reported that anthracene-
based fluorometric dye (D4) containing the same dithiaazadioxo
ring showed highly selective fluorescence toward Ag+ ion, but
In the Table S1 of the Supporting information, several azoben-
zene dyes were listed as the chromophore for Hg2+ detection,
and more fluorogenic or colorimetric dyes employing the

C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
6843
not to Hg2+ ions. However, the ICT band of L3 that had already
shifted by chelation of Hg2+ ion (at 350 nm) did not restore the ori-
ginal spectrum (toward 450 nm) in the addition of Ag+ or Cu2+ ions,
indicating preferential binding with Hg2+ over Ag+ or Cu2+ ions.
These controversial observations might be attributed to a new
combination of the dithiaazadioxo ring with azobenzene instead
of anthracene dye, and suggests some co-operative interaction of
Hg2+ ion with the diazo group.5c,13
to diazo chelation of Hg2+ ion as is depicted in Figure 6. The bath-
ochromic shift from protonation or mercury chelation of diazo
group could be elaborated by enlarging ICT dipole moment
through fortifying acceptor capability of the push–pull dipole
scheme via diazo chelation.13 In fact, azobenzene groups grafted
on silica matrix were reported to have an affinity toward Hg2+
ion co-operatively and illustrated bathochromic changes.15 How-
ever, the bright pink color of L3 quickly disappeared by addition
of aqueous organic solvents like methanol or MeCN, suggesting
that the diazo chelation of Hg2+ occurs only in the absence of water
(data not shown). Overall, all these solvatochromic observations
with different solvents suggest that the binding between Hg2+ ions
and L3 is strongly dependent on the natures of solvents. Moreover,
to our best knowledge on Hg2+-chelating dyes, these quite distinct
color changes (colorless in MeCN vs bright pink in CHCl3) toward
Hg2+ binding have never been observed (see Fig. 1).
On the other hand, when the working solvent was changed into
MeCN only, the overall trends were maintained except two minor
changes: First, the hypsochromic shift in response to Hg2+ ion was
widened into a band at 330 nm (
D
m
ꢀ120 nm), indicative of
changes in push–pull dipole in the absence of water. Second,
Cu2+ ions additionally developed comparable bathochromic bands
as well as the small hypsochromic changes mentioned above; sim-
ilarly Pb2+ solution also exhibited small bathochromic shoulder
band. Although the hypsochromic band height was similar with
the bathochromic one in intensity for Cu2+ ions, the apparent color
change in the visible range seems dominated by the bathochromic
change (i.e., pink color). Nevertheless, L3 still demonstrated prefer-
ential binding with Hg2+ ions in pure MeCN, even in the presence of
other metal ions as observed in MeCN/H2O mixture solvent (data
not shown).
UV/vis titrations upon Hg2+ addition
In order to evaluate the Hg2+-binding ability of L3 in detail,
titrations with various concentrations of Hg2+ ion were attempted
in a fixed concentration of L3 (50 lM). As shown in Figure 2a and b,
Hg2+ ions were observed to interact with L3 through a one to one
correspondence in MeCN solution (aqueous or pure), which was
well coincident with previous studies adopting dithiaazadioxo-
based chromophores.5d,13 On the other hand, when L3 was titrated
Intriguingly, another dramatic change was observed when the
working solvent was further changed into CHCl3. The ICT band at
450 nm substantially red-shifted into ꢀ550 nm for Hg2+ ions, sug-
gesting that Hg2+ ion is differently interacting with the L3 in CHCl3
since the changing direction in the ICT band is quite opposite.
Additionally, small decrease in intensity was observed as for Pb2+
ions, which had shown a bathochromic shoulder band with L3 in
MeCN. Resultantly, only in response to Hg2+ ions, the color of L3
turned to bright pink instead of being colorless as observed in
MeCN or MeCN/H2O solvents. As noticed above, Cu2+ ions in pure
MeCN exhibited a similar color change, albeit to less extent. The
pink color indicates that a similar binding scheme observed in
Hg2+ ions in CHCl3 might partially contribute to Cu2+-L3 interac-
tions in pure MeCN. At this stage, it seems noteworthy that the
bright pink color of L3 is usually observed in acidic aqueous
solution (lower than pH 1), which is well known to be related with
the protonation of diazo group in the push–pull structure of ICT
chromophore (carboxylic acid as electron withdrawing group).13,14
Therefore, the bathochromic shift in CHCl3 is tentatively attributed
Figure 3. 1H NMR spectra of L3 (1 mM) in CDCl3 solution in the presence of
Figure 4. 1H NMR spectra of L3 (1.0 mM) (a) in CD3CN/D2O solution and (b) in
different amounts of Hg(ClO4)2.
CD3CN by gradual addition of Hg(ClO4)2.

6844
C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
in CHCl3 as shown in Figure 2c, the interaction ratio changed to 2:1
correspondence, indicating that two molecules of L3 are interact-
ing with a single Hg2+ ion. Considering the alleged participation
of diazo group in binding with Hg2+ in CHCl3, a new binding
scheme in CHCl3 should be proposed for Hg2+–L3 interaction where
a single Hg2+ ion is sandwiched between two diazobenzene groups
of the two L3 ligands. At this moment, dithiaazadioxo ring appears
not to bind with Hg2+ ions in CHCl3, while the same ring captures
Hg2+ ion in MeCN or MeCN/H2O solution, which is intriguing in
that the dithiaazadioxo ring that was known for a strong mercury
chelator is unable to bind mercury in CHCl3.
On the other hand, a commercially obtained diazo-compound,
without any specific mercury binder, did not show any mercury
selectivity, nor did any spectral changes over other metal ions. This
suggests the participation of dithiaazadioxo ring upon discriminat-
ing Hg2+ ion in CHCl3 through an unknown mechanism. Interest-
ingly, the involvement of two ligands against a single Hg2+ ion
was already observed with a similar azobenzene dye with L3 along
with bathochromic shift, but in a different solution.9 Although any
specific scheme was not presented for the interaction with Hg2+
ions, we assume that the same diazo chelation might be applied
to the report. In the next section, the proposed binding schemes
in different solvents were further investigated by NMR observa-
tions during the Hg2+-titrations.
cyclic ring plays the expected and major role in capturing Hg2+ ions
as designed. The growth of new bands might indicate stronger
binding with Hg2+ ions while the band broadening observed in
CHCl3 shows weaker or dynamic interactions. Since similar
changes in the NMR titration of compound 2 (without diazo group)
were observed in the same solvent system (d-MeCN/D2O), the
major role of the dithiaazadioxo ring and changes in aromatic
backbone could be reasonably supported (see Supporting informa-
tion, Fig. S6).
The band shifts of the dithiaazadioxo ring were more clearly ob-
served in pure d-MeCN solvent since the interfering broad and
intensive HOD band at ꢀ3 ppm is absent in d-MeCN only
(Fig. 4b). In this titration, only 1st protons in the dithiaazadioxo
ring maintained the characteristic sharp peak, albeit in a shifted
position, while other protons (2nd–5th) were both broadened
and shifted, implicating that Hg2+ ion is a little bit biased toward
the amine atom. Besides the dithiaazadioxo ring protons, protons
in the aromatic diazo group also suffered changes in the positions;
6th proton most drastically shifted down-field while the other
bands mildly moved to down-field. Compared with the band shift
in d-MeCN/D2O, the band shifts in the aromatic group (7–9th pro-
tons) were reverted into down-field shifting. These observations
suggest that hydration layer of Hg2+ or L3 seems to affect the bind-
ing in d-MeCN/D2O compared to pure d-MeCN.
NMR titrations upon Hg2+ addition
Colorimetric behavior upon polymer immobilization
In NMR spectra of L3 in CDCl3, protons in the core benzene ring
(near the diazo group, 6–8th protons) gradually decreased and
broadened as the content of Hg2+ ion increased (see Fig. 3). Among
the aliphatic protons in the dithiaazadioxo rings, only protons next
to the amine group (5th protons) responded to the Hg2+ ions, in-
deed indicating the diazo chelation of Hg2+ ions that was proposed
above.
Although the current L3 has shown very interesting features
upon Hg2+-binding in organic solvents like CHCl3 and MeCN, or
in aqueous mixtures, most frequent test samples for Hg2+ ions
might be an aqueous solution from food, tap water, or river in
the neighborhood. Therefore, most colorimetric sensors for metal
ions may be preferable to be designed for water sample. In this
regard, several attempts were made to immobilize the mercury
chelating ligand onto silica matrix and demonstrated good perfor-
mances.15 Instead, we attempted to tether the colorimetric Hg2+
sensors onto a polymer backbone, which enables L3 to dissolve
easily in aqueous solution and thereby to widen the applicability
of the diazo compound, for instance, by layer-by-layer deposition
protocol. In Figure 5a, titration curves of L3 tethered on cationic
polymer PEI are shown; roughly 1.0% among the primary amines
present in PEI (M.W. 25,000) was decorated with L3 and the
Recalling the stoichiometry of L3 to Hg2+ ion (2:1) in CHCl3, a
chelation scheme of L3 could be presented where two diazo-
groups are interacting with a single Hg2+ ion, but is relatively lean-
ing toward the dithiaazadioxo ring as will be seen below (Fig. 6a).
The diazo chelation of L3 was further confirmed in a control exper-
iment where only the dithiaazadioxo ring (compound 2) was
titrated with Hg2+ and monitored by NMR in CDCl3. In this titration,
no apparent change was observed in NMR spectra, suggesting
diazo-chelation in L3 is the dominant interaction in CHCl3 solvent
(see Supporting information, Fig. S5).
On the other hand, the spectral shifts of the protons of L3 in
d-MeCN/D2O are quite different compared to the changes in CHCl3
(Fig. 4a). Instead of band broadening, new bands (black arrow)
emerged and original bands diminished as the titration proceeded.
These NMR spectra seem to definitely show that the Hg2+ binding
modes in two solvents (CHCl3 vs MeCN/H2O) are different and that
the dithiaazadioxo rings are dominantly involved in MeCN/H2O. In
detail, the bands of protons near two sulfur atoms in the dithiaaza-
dioxo ring (3rd and 4th protons) quickly diminished in intensity
and shifted down-field, while other bands in the ring were
substantially broadened and new bands developed simulta-
neously; the individual identifications for the newly developed
bands were difficult, but were tentatively assigned at our best.
In addition, protons in the aromatic backbone of L3 also appear
to be affected by the addition of Hg2+ ions so that a new band for
6th protons developed down-field while the original one disap-
pears, and other aromatic protons (7–8th) developed up-field; ori-
ginal peak intensities of all the protons in the aromatic backbone
diminished and new bands were developed (albeit also broad-
ened), which are different from observations in CDCl3 (i.e., band
broadening without position shifts). These drastic changes in
protons of the dithiaazadioxo ring suggest that the dithiaazadioxo
Figure 5. UV/vis spectra of the L3 (ꢀ10
lM) conjugated on cationic polymer PEI
upon gradual addition of Hg(ClO4)2 in water; (a) PEI1.0-L3 (b) PEI65-L3. (c) UV/vis
spectra of the L4 (50 l
M) on the addition of Hg2+ ions in aqueous MeCN solution
and (d) UV/vis spectral changes of L3 with one equivalent of Hg2+ restoring original
band upon the addition of free amine molecules.

C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
6845
concentration of the grafted L3 was adjusted to be 10 lM
(PEI1.0-L3). Unlike in the state of free L3, PEI1.0-L3 showed sub-
stantially decreased sensitivity; for full spectral changes, twenty
equivalents or more Hg2+ ions were needed. In order to see the
effect of tagging ratio, it was attempted that more primary amines
(25% and 65%) in polymeric PEI were decorated with L3 (PEI25-L3
and PEI65-L3, respectively). Interestingly, although PEI25-L3
demonstrated improved sensitivity to some extent, the responses
toward Hg2+ ions still lagged, indicating that PEI grafting still inter-
fering to the Hg2+ ion capture. Further increase in L3 grafting ratio
up to 65% showed similar responses with PEI25-L3 and still failed
to revamp the sensitivity of free L3 ligand, as shown in Figure 5b.
With PEI65-L3, about ten equivalents of Hg2+ ions (ꢀ100
lM) were
needed to saturate the spectral changes of the tethered L3
(ꢀ10
lM). These observations revealed that simply high dose of
L3 in cationic PEI cannot improve its binding affinity toward Hg2+
ions.
In order to clarify the grafting effect, L3 was conjugated with a
simple amine like ethanolamine (L4), which is also converting the
carboxylic group into an amide one and might affect the push–pull
dipole of azobenzene in the same way as in PEI grafting. Intrigu-
ingly, L4 demonstrated almost the same responses as in L3 on
Hg2+ additions without decreasing the sensitivity, but L4 showed
two differences as shown in Figure 5c. First, the Hg2+-titration of
L4 in aqueous MeCN solution revealed very similar responses with
L3 (i.e., hypsochrmic changes and one to one correspondence to
Hg2+ ions) except one aspect; the hypsochromic changes of L4 in
aqueous MeCN are widened (ꢀ330 nm) while L3 moved less in
the same MeCN/H2O solvent (ꢀ350 nm). The larger hypsochromic
shifts were also observed with PEIx-L3 in aqueous solution and
with L3 in pure MeCN. Second, the titration of the synthesized L4
with Hg2+ ion in CHCl3 demonstrated the same pink color change
(bathochromic shifts) and the same spectral features as in L3, but
L4 showed 1:1 reaction ratio, suggesting different behaviors of L4
from L3. Since the same bathochromic shift of ICT band was ob-
served in L4, it seems plausible to assume the same diazo chelation
with Hg2+ ions in CHCl3, but the observed stoichiometry (L4 to
Hg2+, 1:1) seems to require further elaboration. Considering the
only difference between L3 and L4, the distinct stoichiometry in
CHCl3 and widened hypsochromic band shift in aqueous MeCN
could be plausibly attributed to loss of acidic proton from carbox-
ylic acid, owing to the amide conjugation.14
In the subsequent NMR titration, L4 showed similar changes in
d-MeCN with L3, fortifying that similar binding scheme could be
applied (see Supporting information, Fig. S7); 1st protons drasti-
cally shifted down-field with sharp characteristics while other
protons in the dithiaazadioxo ring were shifted and broadened.
While additional aliphatic protons from ethanolamine showed no
changes indicating less involvement upon Hg2+ binding, 6th pro-
tons in aromatic backbone showed largest shift and other protons
in azobenzene chromophore suffered minor changes. Upon the
titration in CDCl3 solvent, no apparent change was observed
among protons in the dithiaazadioxo ring except 5th protons as ob-
served in L3. Overall, we could not find any significant changes
from NMR titration, which suggests a similar binding scheme on
L3 and L4. Therefore, a similar but slightly different binding
scheme was proposed for L4 in CHCl3 as is shown in Figure 6c. In
spite of the detailed investigation on L4, we largely failed to ac-
count for the origin of the fouled sensitivity in the PEI conjugation
of L3.
Figure 6. Proposed binding schemes of L3 with Hg2+ (a) in CHCl3, (b) in MeCN, and
for L4 with Hg2+ ions, (c) in CHCl3 and (d) in MeCN.
ring) from binding with Hg2+ ions. Although it is hard to pin-point
the origin of this lagging, it seems that free amine molecules might
intercept the added Hg2+ ions as an alternative chelator or might
co-ordinate with the dithiaazadioxo rings preferentially. In either
condition, the polymeric nature of PEI seems to exacerbate the
binding efficiencies of grafted L3. Moreover, it is noteworthy that
secondary and tertiary amines (not to mention primary amines)
are still massively present even in PEI65-L3. This is quite an inter-
esting observation since the amine-hindering is unprecedented in
the colorimetric mercury detection, while interferences from other
metal ions or anions have been intensively investigated, hith-
erto.7,9 In addition, this finding might have technical implications
in treating biological samples with these synthetic dyes when eval-
uating mercury level through colorimetric methods.
Combining the overall NMR titration data and UV/vis spectra
upon gradual addition of Hg2+ ions, two major binding schemes
were proposed for L3 in MeCN/H2O and CHCl3 solvents as shown
in Figure 6. In CHCl3, it seems that a single Hg2+ ion resides be-
tween two azobenzenes of two L3 ligands (Fig. 6a), while a single
Hg2+ ion does in a dithiaazadioxo ring of L3 in MeCN (pure or aque-
ous, Fig. 6b). Although the NMR shifts of aromatic protons in pure
MeCN were slightly different from those in aqueous MeCN, the
proposed binding scheme was tentatively maintained due to the
drastic changes of the dithiaazadioxo ring in NMR titrations and
similar UV/vis spectral changes in two solvent systems. As noticed
before, quite different stoichiometry between L3 and L4 in CHCl3 is
ascribed to the loss of the acidic proton in L3, so that two Hg2+ ions
are sandwiched between two L4 as shown in Figure 6c. In aqueous
MeCN, L4 is assumed to have the same binding scheme with L3
(Fig. 6d), along with the absence of protonation at the diazo group
As such, another control experiment was performed in which
primary amine molecules were simply added into Hg2+-saturated
L3 solution (MeCN/H2O) in advance. As shown in Figure 5d, once
shifted ICT band of L3 (ꢀ350 nm) gradually restored the original
spectrum by the incremental addition of ethanolamine, revealing
that free amine groups are preventing L3 (or the dithiaazadioxo

6846
C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
that widens the hypsochromic shift in UV/vis spectra as shown
above.
was combined and dried over anhydrous MgSO4 and filtered. After
evaporation of solvent in vacuo, the crude residue was chromato-
graphed on silica gel with CHCl3/MeOH mixture (9:1), yielding L3
as an orange solid (37 mg, 65% yield, Rf = 0.42 with CHCl3/MeOH
mixture, 9:1). Absorption maximum of the product is 450 nm. 1H
NMR (400 MHz, CDCl3): d = 2.74–2.77 (t, 4H, –CH2S), 2.87–2.91 (t,
4H, –CH2S), 3.60–3.65 (m, 8H, –CH2N, –CH2O), 3.78–3.82 (t, 4H, –
CH2O), 6.63–6.70 (d, 2H, ArH), 7.86–7.92 (d, 4H, ArH), 8.18–8.22
(d, 2H, ArH).
In summary, the colorimetric behavior of azobenzene-based
chromophore containing the dithiaazadioxo ring for Hg2+-detec-
tion was investigated in various solvent systems. The Hg2+-specific
ligand showed hypsochromic changes in pure and aqueous MeCN
upon Hg2+ binding while a strong bathochromic shift was observed
in CHCl3. For the bathochromic shift, the absence of water was
essential and these observations were unprecedented in previous
studies employing dithiaazadioxo crown ether for a specific ligand
of mercury ion. In both solvent systems, dithiaazadioxo ring is
revealed to be indispensable to the selective detection of mercury
ion. Guided by these observations, a few binding schemes in
different solvents were proposed. Grafted on a cationic polymer
(i.e., PEI) through amide coupling, it was observed that L3 largely
maintained the Hg2+ detection capability, but was substantially
hindered by amine moieties in PEI.
Coupling of L3 with ethanolamine (L4)
L3 (250.0 mg, 0.53 mmol), N-(3-dimethylaminopropyl)-N0-eth-
ylcarbodiimide hydrochloride (EDC, 132.3 mg, 0.69 mmol), and
N-hydroxysuccinimide (NHS, 79.4 mg, 0.69 mmol) were dissolved
in methylene chloride/MeOH (10:1). After stirring at rt for 3 h, eth-
anolamine (48.3 mg, 0.79 mmol) in minimal amount of CHCl3 was
added. After stirring for another 12 h at rt, solvent was evaporated
in vacuo. The crude residue was chromatographed on silica gel
with CH2Cl2/MeOH mixture, 10:1, yielding L4 as an orange solid
(120.1 mg, 43.7% yield, Rf = 0.4 with mixture of CH2Cl2 and MeOH,
10:1).
Experimentals
Materials
PEI25-L3
N-Phenyldiethanolamine, methanesulfonyl chloride, 3,6-dioxa-
1,8-octanedithiol, 4-aminobenzoic acid, ethanolamine, and deute-
rium oxide were purchased from Sigma–Aldrich. Chloroform-d,
acetonitrile-d3, and dimethylsulfoxide-d6 were purchased from
Cambridge isotope laboratories (CIL). 1H NMR spectra were mea-
sured with a Varian HFT 80 spectrometer or with WH300, and
UV/vis spectra were obtained from Perkin Elmer VT 3000
spectrophotometer.
L3 (25.0 mg, 53.0
addition of dicyclohexylcarbodiimide (DCC, 10.9 mg, 58.3
and NHS (6.0 mg, 58.3 mol). After stirring at rt for 12 h, 16.5 mg
of polyethyleneimine (PEI, Mw 25,000) was added to thus-formed
NHS ester of L3 solution. After stirring for another 6 h, the reaction
mixture was washed with water three times. The whole aqueous
layer was dialyzed in dialysis bag (MWCO 7000) with water. After
dialysis for 60 h, 100 ml of PEI25-L3 aqueous solution (0.2 mg/ml)
was obtained (49.3% yield).
l
mol) in 10 mL of CHCl3 was followed by the
lmol)
l
(Phenylazanediyl)bis(ethane-2,1-diyl)dimethanesulfonate (1)
N-Phenyldiethanolamine (10.00 g, 55.18 mmol) and triethyl-
amine (22.33 g, 220 mmol) were dissolved in 400 mL of CHCl3 at
0 °C, and then methanesulfonyl chloride (13.91 g, 120 mmol) was
added dropwise. After stirring for 1 h at 0 °C, the reaction mixture
was washed with ice water, 5% HCl, saturated NaHCO3, and 5%
brine sequentially. The organic layer was combined and dried over
anhydrous MgSO4, filtered, and evaporated in vacuo to give 1 as a
solid (18.54 g, 99.6% yield).
PEI65-L3
Typically, L3 (122.0 mg, 0.26 mmol) in 10 ml of chloroform was
followed by the addition of DCC (70.0 mg, 0.34 mmol) and NHS
(40.0 mg, 0.35 mmol) in minimal amount of CHCl3. After stirring
at rt for 12 h, PEI (Mw 25,000, 80.0 mg) dissolved in CHCl3 was
added to the solution of NHS ester of L3. After stirring for addi-
tional 6 h, the reaction mixture was washed with water three
times through a separatory funnel. The whole aqueous layer was
dialyzed in dialysis bag (MWCO 3500) with water. After dialysis
for 72 h, 80 ml of aqueous PEI65-L3 solution (0.55 mg/ml) was
obtained.
1,4-Dioxa-7,13-dithia-10-phenyl-10-azacyclopentadecane (2)
3,6-Dioxa-1,8-octanedithiol (9.00 g, 49.37 mmol) was dissolved
in 300 mL of dry MeCN, followed by addition of K2CO3 (15.00 g,
108.53 mmol). This suspension was stirred and heated up to
80 °C. Then, 1 (18.54 g, 55.18 mmol) dissolved in 150 mL dry
MeCN, was added dropwise for 5 h. After another 12 h, the reaction
mixture was cooled down to room temperature, white precipitate
was filtered off, and the solvent evaporated in vacuo. The crude
residue was chromatographed on silica gel with CHCl3, yielding 2
as a white crystalline solid (1.56 g, 28.4%, Rf = 0.84 with CHCl3/
MeOH mixture (9:1)). 1H NMR (400 MHz, CDCl3): d = 2.74–2.77
(t, 4H, –CH2S), 2.87–2.91 (t, 4H, –CH2S), 3.60–3.65 (m, 8H, –
CH2N, –CH2O), 3.78–3.82 (t, 4H, –CH2O), 6.63–6.70 (m, 3H, ArH),
7.19–7.24 (t, 2H, ArH).
NMR titration experiments
For the NMR titration studies in CDCl3, a stock solution of
Hg(ClO4)2 (5.0 mM) dissolved in CDCl3 was added incrementally
by using a micropipette (20–200
l
L) to the L3, L4, and compound
L) in the NMR tube.
2 (typically ꢀ1.0 mM) dissolved in CDCl3 (700
l
For other solvent systems, previously dissolved Hg(ClO4)2 in corre-
sponding solvents was gradually added to L3 or L3 derivatives.
Acknowledgements
This work was supported by the Grants of the Korea Research
Foundation (NRF-2012R1A1A2020074 and NRF-2010-0020840)
and KRIBB initiative programs.
4-((4-(1,4-Dioxa-7,13-dithia-10-azacyclopentadecan-10-
yl)phenyl)diazenyl)benzoic acid (L3)
Sodium nitrite (8.33 mg, 0.12 mmol) in 10 mL of water pre-
cooled at 0 °C was added dropwise into a solution of 4-aminoben-
zoic acid (16.48 mg, 0.12 mmol) in 20 mL of 6 M HCl at 0 °C and
stirred for 20 min. The diazonium salt solution was added drop-
wise into a solution of 2 (40 mg, 0.12 mmol) in 50 mL of DMF at
0 °C. The solution was stirred for additional 6 h at 0 °C. Then,
20 mL of CHCl3 and 10 mL of water were added. The organic layer
Supplementary data
Supplementary data (additional NMR data for L3 and L4 synthe-
ses, and Hg2+-titrations of L4 and compound 2 in organic solvents,
in which UV/vis titration curves of L4 in organic solvents and UV/

C. H. Jeon et al. / Tetrahedron Letters 54 (2013) 6841–6847
6847
7. Lee, H.; Lee, S. S. Org. Lett. 2009, 11, 1393–1396.
8. Cheng, X.; Li, Q.; Li, C.; Qin, J.; Li, Z. Chem. Eur. J. 2011, 17, 7276–7281.
9. Sancenón, F.; Martínez-Máñez, R.; Soto, J. Angew. Chem., Int. Ed. 2002, 41, 1416–
1419.
vis spectra of simple azobenzene responding to the addition of
various metal ions) associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2013.
10.013. These data include MOL files and InChiKeys of the most
important compounds described in this article.
10. (a) Yang, Y.-K.; Ko, S.-K.; Shin, I.; Tae, J. Org. Biomol. Chem. 2009, 7, 4590–4593;
(b) Yan, F.; Cao, D.; Wang, M.; Yang, N.; Yu, Q.; Dai, L.; Chen, L. J. Fluoresc. 2012,
22, 1249–1256; (c) Ma, Q.-J.; Zhang, X.-B.; Zhao, X.-H.; Jin, Z.; Mao, G.-J.; Shen,
G.-L.; Yu, R.-Q. Anal. Chim. Acta 2010, 663, 85–90; (d) Lee, H. Y.; Swamy, K. M.
K.; Jung, J. Y.; Kim, G.; Yoon, J. Sens. Actuators, B 2013, 182, 530–537.
11. Park, C. S.; Lee, J. Y.; Kang, E.-J.; Lee, J.-E.; Lee, S. S. Tetrahedron Lett. 2009, 50,
671–675.
12. Ros-Lis, J. V.; Martínez-Máñez, R.; Sancenón, F.; Soto, J.; Spieles, M.; Rurack, K.
Chem. Eur. J. 2008, 14, 10101–10114; (b) Berdnikova, D. V.; Fedorov, Y. V.;
Fedorova, O. A. Dyes Pigments 2013, 96, 287–295; (c) Abalos, T.; Jimenez, D.;
Moragues, M.; Royo, S.; Martinez-Manez, R.; Sancenon, F.; Soto, J.; Costero, A.
M.; Parra, M.; Gil, S. Dalton Trans. 2010, 39, 3449–3459.
13. Ábalos, T.; Moragues, M.; Royo, S.; Jiménez, D.; Martínez-Máñez, R.; Soto, J.;
Sancenón, F.; Gil, S.; Cano, J. Eur. J. Inorg. Chem. 2012, 2012, 76–84.
14. Hofmann, K.; Brumm, S.; Mende, C.; Nagel, K.; Seifert, A.; Roth, I.;
Schaarschmidt, D.; Lang, H.; Spange, S. New J. Chem. 2012, 36, 1655–1664.
15. Sanchez, G.; Curiel, D.; Ratera, I.; Tarraga, A.; Veciana, J.; Molina, P. Dalton
Trans. 2013, 42, 6318–6326.
References and notes
1. (a) McKeown-Eyssen, G. E.; Ruedy, J.; Neims, A. Am. J. Epidemiol. 1983, 118, 470;
(b) Harris, H. H.; Pickering, I. J.; George, G. N. Science 2003, 301, 1203.
2. Isaad, J.; El Achari, A. Tetrahedron 2013, 69, 4866–4874.
3. Lin, Y.-W.; Huang, C.-C.; Chang, H.-T. Analyst 2011, 136, 863–871.
4. Gong, J.; Zhou, T.; Song, D.; Zhang, L.; Hu, X. Anal. Chem. 2009, 82, 567–573.
5. (a) Yuan, M.; Li, Y.; Li, J.; Li, C.; Liu, X.; Lv, J.; Xu, J.; Liu, H.; Wang, S.; Zhu, D. Org.
Lett. 2007, 9, 2313–2316; (b) Liptay, W. Angew. Chem., Int. Ed. Engl. 1969, 8,
177–188; (c) Yan, Y.; Hu, Y.; Zhao, G.; Kou, X. Dyes Pigments 2008, 79, 210–215;
(d) Descalzo, A. B.; Martínez-Máñez, R.; Radeglia, R.; Rurack, K.; Soto, J. J. Am.
Chem. Soc. 2003, 125, 3418–3419; (e) Huang, J.-H.; Wen, W.-H.; Sun, Y.-Y.;
Chou, P.-T.; Fang, J.-M. J. Org. Chem. 2005, 70, 5827–5832.
6. Meng, X.-M.; Liu, L.; Hu, H.-Y.; Zhu, M.-Z.; Wang, M.-X.; Shi, J.; Guo, Q.-X.
Tetrahedron Lett. 2006, 47, 7961–7964.