5972
K. Tsukamoto et al. / Tetrahedron Letters 54 (2013) 5971–5973
Figure 2. Effects of metal ions (5 mM of Ca2+ and Mg2+, and 5
Mn2+, Fe3+, Co2+, Ni2+, Ag+, Pb2+, Cu2+ and Hg2+) on the fluorescence intensity ratio
(F/F0) from the addition of 1 equiv of Cd2+ into 1 (5
M). F0: fluorescence intensity of
1 (5 M) without any metal ions. As metal reagents, CdCl2, Zn(NO3)2, CaCl2, MgCl2,
l ,
M of Zn2+, Cr3+
l
l
CrCl3, MnCl2, FeCl3, CoCl2, NiCl2, AgNO3, Pb(NO3)2, CuCl2, and HgCl2 were used. All
measurements were taken in a pH 7.2 HEPES buffer (50 mM, I = 0.1 M (KNO3)),
kex = 370 nm, and kem = 416 nm. Error bars represent standard deviations (n = 4).
Figure 1. (a) Fluorescence spectra of 1 (5 l
M) upon the addition of Cd2+. (b) A plot
of the difference between the fluorescence intensities (F – F0) at kem = 416 nm
obtained before and after the addition of Cd2+ into a solution of 1 (5
lM) as a
than pH 6, the background response became high, that would be
attributed to protonation of the N atom in 1, to prevent the PET ef-
fect in 1 from working. With these results in hands, it was con-
cluded that a buffer around pH 7.2 is a solvent of choice for high
selective detection of Cd2+ over Zn2+ using 1.
function of equiv of Cd2+. CdCl2 was used as a Cd2+ reagent. All measurements were
taken in a pH 7.2 HEPES buffer (50 mM, ionic strength (I) = 0.1 M (KNO3)) and
kex = 370 nm. Error bars represent standard deviations (n = 4).
To investigate the detailed mechanism of the reaction of 1 with
Cd2+, N-methylated derivative 2 was prepared and subjected to
fluorometry in the absence or presence of Cd2+ (Figs. S8 and S9).
To our surprise, addition of 1 equiv of Cd2+ to a solution of 2 in
any pH conditions gave rise to little or no enhancement in the fluo-
rescence intensity. The apparent KdCd of 2 at pH 7.2 was estimated
to be 2.1 Â 10–4 M, which was larger by three orders of magnitude
than that of 1 (Fig. S3). On 1H NMR spectra in D2O, all signals for 1
or 2 were shifted downfield in the same manner by the addition of
Cd2+ (Fig. 3). The downfield shift in all the region would be caused
by the shielding effect of Cd2+ chelated with the 1,4,7-triazanonyl
group in 1 or 2.9 Moreover, 8-(anthracen-9-yl)-1,4,7-triazaoctane
8, lacking three 2-hydroxyethyl groups on 1, showed a lower fluo-
pound 1 exhibited a suppressed fluorescent spectrum in a pH 7.2
HEPES buffer (50 mM, I = 0.1 M (KNO3)), and its quantum yield
(a) was estimated to be as low as 0.24 at kex = 370 nm (Fig. 1).6 This
would be ascribed to photo-induced electron transfer (PET) from
the N atom adjacent to the anthracene moiety of 1 (Scheme 2).7
Upon the addition of Cd2+, the fluorescent response from 1 was en-
hanced, and a linear relationship was observed between the re-
sponses and the concentration of Cd2+ up to 0.5 equiv (2.5
lM)
(Fig. 1b). By further addition of Cd2+, the fluorescence intensity in-
creased nonlinearly, and reached a plateau. When 1.0 or 2.0 equiv
of Cd2+ was added, the quantum yield of 1 was recovered to 0.43 or
0.47, respectively (Table S1). A Job’s plot of an increase in the fluo-
rescence intensity due to the reaction of 1 with Cd2+ showed a
maximum at a mole fraction ([1]/([1] + [Cd2+])) of 0.5, which
clearly indicated the formation of a 1:1 complex (Fig. S2). The
apparent dissociation constant of 1 for Cd2+ (KdCd) was estimated
to be 1.0 Â 10–7 M, which is so small as to realize highly sensitive
detection for Cd2+ (Fig. S3).8
In contrast with the case of Cd2+, 1 equiv of Zn2+ was added to a
solution of 1, bringing about negligible increment in the fluores-
cence intensity (Fig. 2 and Fig. S5). The apparent dissociation con-
stant of 1 for Zn2+ (KdZn) was estimated by fluorometry to be
5.6 Â 10–5 M (Fig. S3). This value was 560 times larger than the
KdCd, demonstrating that 1 can distinguish Cd2+ from Zn2+ with
high selectivity. The effects by other metal ions (5 mM of Ca2+
and Mg2+, and 5 M of Cr3+, Mn2+, Fe3+, Co2+, Ni2+, Ag+, Pb2+, Cu2+
l
and Hg2+) were also examined. The fluorescence response induced
by the reaction of 1 with Cd2+ was not disturbed by the examined
metal ions, except for Cu2+ and Hg2+ (Fig. 2 and Fig. S5). The pres-
ence of Cu2+ or Hg2+ quenched the fluorescence response of 1 to-
ward Cd2+. Since practical samples might include Cu2+, the effect
by Cu2+ must be solved.
Next, the effect of pH on the fluorescence intensity of 1 itself
was examined (Fig. S7). The intensity was decreased from pH 6.0
to 9.2, probably due to an increase in the PET quenching effect
through the deprotonation of the N atom adjacent to the anthra-
cene in 1. Although the background signal was significantly low
in these basic conditions, the fluorescent response for the reaction
of 1 not only with Cd2+ but also with Zn2+ increased greatly
(Fig. S7), that is, the selectivity of 1 for Cd2+ over Zn2+ became
poorer with an increase in basicity. Under acidic conditions less
Figure 3. 1H NMR spectra (400 MHz) of 1 and 2 (10 mM) in the absence or presence
of 1 equiv of Cd2+ measured in D2O: (a) 1; (b) 1 with CdCl2 (10 mM); (c) 2 with CdCl2
(10 mM); (d) 2.