3.1 Fabrication and characterization of GSH-coated CuNCs
The as-prepared GSH-CuNCs were characterized by Fourier transfer infrared (FT-IR) spectra, UV-Vis spectroscopy, XPS, DLS, and TEM. As seen in Fig. 1a, there was an obvious contrast between the FT-IR spectra of pure GSH and GSH-CuNCs. The peak at 2522 cm-1 which was ascribed to the S-H stretching vibration of GSH disappeared notably in GSH-CuNCs, indicating the new bond of S-Cu was formed in GSH-CuNCs[42]. The UV-Vis spectra of CuSO4, GSH and GSH-CuNCs were illustrated in Fig. 1b, a broad absorption peak was observed in the range of 250-300 nm but no absorption peaks appeared in the visible region resulted from the surface plasmon resonance (SPR) of big metal nanoparticles[43-45], implying that the prepared CuNCs possessed ultra-small size. These results were consistent with those of formerly reported.
XPS was performed to identify the elementary composition and valence states of GSH-CuNCs, as seen in Fig. 2. Fig. 2a showed that C, N, O, S, and Cu elements in the precursors were still present in the CuNCs. XPS results of sulfur element in Fig. 2b illustrated that the binding energy of S2p3/2 and S2p1/2 showed 1.2 eV spin-orbit splitting with S2p3/2 at 162.1 eV, which was negative shifted owing to S-metal interaction by contrast to elemental S with binding energy at 164.2 eV of S2p3/2[46]. To verify valence state of copper, its XPS spectrum was performed in Fig.2c. Obviously, two tense peaks at 932.3 eV and 952.2 eV emerged in the high-resolution spectrum of Cu2p, which were belonged to Cu2p3/2 and Cu2p1/2 respectively, that was in accordance with the oxidation state of Cu (0) or Cu (I)[47, 48]. However, no peaks at 942 eV of Cu2p3/2 were found implying the absence oxidation state of Cu (II). Thus, it was deduced that Cu (II) was thoroughly reduced by GSH during the synthesis process of GSH-CuNCs.
To investigate the size and morphology of GSH-CuNC, the TEM image and size distribution histogram were shown in Fig. 3. Notably, the GSH-CuNCs were dispersed excellent uniform with an ultra-small size of 2.3 ± 0.2 nm. Overall, these results revealed that the protective ligand of GSH was successfully anchored onto the surface of CuNCs.
3.2 AIE effect of Ce3+ and other rare-earth metal ions (RE3+) on GSH-CuNCs
Although several rare-earth metal ions like Gd (III)[49], Sm (III)[50] have been used to induce the AuNCs aggregated leading to increasement of fluorescence intensity (FI), however, their AIE effect on the CuNCs have not been studied to date. On the basis of strong coordination between trivalent lanthanide ions and the carboxyl groups of GSH[49], herein, we first systematically researched AIE function of GSH-CuNCs caused by RE3+. Fluorescent emission spectra upon different excitation wavelength were demonstrated in Fig. 4a, their emission spectra were independent of excitation wavelength and the optimum excitation wavelength was determined to be 270 nm firstly. Then fluorescence emission spectra of GSH-CuNCs in the absence and presence of RE3+ were recorded. All measurements were tested under the same conditions that were set as follows: HAc-NaAc buffer with pH=5.0, with excitation wavelength at 270 nm, with the content of RE3+ and GSH-CuNCs at 18 µM and 0.5 mg/mL respectively. As performed in Fig. 4b, the GSH-CuNCs solution itself did not emit fluorescence under UV light. Interestingly, when a series of different RE3+ were added into the GSH-CuNCs solutions, FI of 610 nm immediately enhanced to varying degrees, in the order of Ce3+, Gd3+, Sm3+, Tb3+, La3+, Y3+, Yb3+, Pr3+, Dy3+, Er3+, Tm3+, Nb3+, Ho3+, Eu3+. Qian et al[51] have reported that RE3+ exhibited the ability to drive carbon quantum dots for distinct aggregation via coordination with carboxylic groups of the surface ligands. In addition, Jyun-Guo et al[52] also concluded that RE3+ were able to coordinate with carboxylic groups of GSH. Hence, it was deduced that different enhancement of FI may be resulted from different ability for binding interaction of RE3+ with two carboxylic and amino groups of GSH. Remarkably, Ce3+ produce the strongest FI in comparison with other lanthanides, that was about 1×103 fold stronger than the original FI of individual GSH-CuNCs. By contrast, Er3+, Tm3+, Nb3+, Ho3+, Eu3+ produce rather weakly luminescence, Other RE3+ caused FI enhancement of GSH-CuNCs at a moderate extent. As consequence, Ce3+ was selected as a modulator throughout the subsequent sensing work.
After comparison of fluorescence emission spectra with Ce3+, Ce3+/GSH, GSH-CuNCs and Ce3+/GSH-CuNCs (Fig. S1), only Ce3+/GSH-CuNCs combine emitted bright fluorescence. To figure out luminescence enhancement mechanism of GSH-CuNCs triggered by Ce3+ clearly, UV–vis absorption spectra of Ce3+/GSH-CuNCs were monitored with gradually increasing concentration of Ce3+ from 0 µM to 20 µM, as seen in Fig. 5a. Background scattering was gradually enhanced with the increase of Ce3+ concentration at wavelengths below 400 nm. It was also obvious to see that GSH-CuNCs solution changed from transparent to turbidity after addition of Ce3+ as shown inset of Fig. 5a. 2 (under daylight), indicating the formation of aggregated complexes. The aggregated complexes emitted ultra-bright luminescence when placed under an UV light, as photographed inset of Fig. 5a. 1. Thus, it was preliminary speculated that the ultra-bright emission upon addition of Ce3+ was attributed to AIE property of GSH-CuNCs.
GSH is a tripeptide containing two carboxylic groups with pKa1 and pKa2 at 2.12 and 3.53 respectively[53]. So ion forms of GSH varied with the change of solution pH by protonation or deprotonation. In particular, isoelectric point of GSH was at 2.85[50]. When solution pH was over 3.53, the capping ligand of GSH on the surface of CuNCs became deprotonated, existed in the mainly form of COO- with negative charges. If Ce3+ was introduced to the solution at the same time, we reason electrostatic attraction between Ce3+ and COO- would work efficiently, thereby aggregating individual GSH-CuNCs into larger particles. Evidently, our supposition was proved based on the change of zeta potential from -13.83 mV of GSH-CuNCs to - 4.13 mV by the introduction of Ce3+, as presented in Fig. 5b. It implied that Ce3+ indeed neutralized partial negative charge of GSH-CuNCs through electrostatic attraction and then coordinated with the ligand of GSH. Since Ce3+ is a hard Lewis acid, it is supposed to be easily bound to amino nitrogen atom and carboxylate oxygen atom of GSH. In order to obtain further evidence for aggregation behavior of GSH-CuNCs driven by Ce3+, TEM and DLS experiments were further conducted in the presence of Ce3+, as characterized in Fig. 5c and 5d. DLS was used to obtain the average size of the aggregated CuNCs. Compared to individual GSH-CuNCs (Fig. 3), it was clearly to observe that GSH-CuNCs aggregated into larger nanoparticles when Ce3+ was added. The size also changed from 2.3 ± 0.2 nm to 182 ± 61 nm after aggregation. In summary, it was no doubt to draw a conclusion that remarkably fluorescence enhancement of GSH-CuNCs induced by Ce3+ was attributed to AIE mechanism. In addition, it was more noteworthy that only a tiny of Ce3+ content about 18 µM could enhance the fluorescence by 1×103 times. As far as we all know, such high efficient AIE performance hasn’t been reported yet.
In conclusion, two forces of electrostatic interaction and coordination action between Ce3+ and GSH-CuNCs dominated the aggregation behavior of GSH-CuNCs. The stronger of the two forces, the denser of the aggregation and the brighter fluorescence will follow. That was because the resulting aggregation can significantly restrict inter- or intra-molecular vibration and rotation of GSH, leading to reduction of nonradiative decay. Then, the return of a large number of excited electrons to the ground state became the predominant form. Consequently, FI was greatly enhanced.
Furthermore, the fluorescence quantum yields(QYs) of GSH-CuNCs and Ce3+/GSH-CuNC s were calculated by the following formula
using quinine sulfate as the standard reference (QY=0.54)[54], the result were presented in Table S1. Here,
represents QY, F implies the integrated area of fluorescence emission spectrum with excitation wavelength at 340 nm, A refers to UV absorption intensity at 340 nm, η indicates the refractive index of the solvent. Additionally, the subscript x and D were test sample and standard sample respectively.
3.3 Optimization of experimental conditions and DPA sensing
As mentioned above, FI of GSH-CuNCs was well tuned by the degree of aggregation based on the mechanism of AIE. While three crucial parameters including solution pH, the content of Ce3+ and reaction time were capable of controlling the aggregation degree. In order to maximize FI, experimental conditions were optimized firstly before DPA sensing. According to previously studies, the charge of the GSH-CuNCs surface was positive at low pH, instead, the charge became negative when pH was increasing[55]. Further, the nucleation of the GSH-CuNCs would be disrupted when solution pH > 5.0, leading to reduction of FI[25]. Hence, in the first step, solution pH was optimized under acidic conditions using HAc-NaAc as buffer in the range of 3.5 - 6.0, as seen in Fig. 6a and 6b. FI increased gradually when pH was from 3.5 to 5.0, however, FI reached saturation and maintained balance when solution pH was at 5.0. Thus, pH=5.0 was selected as the optimal acidity to sense DPA. In the second step, the optimal concentration of Ce3+ was explored. Fluorescence spectra of GSH-CuNCs without and with gradual increase concentrations of Ce3+ were performed as shown in Fig. 6c. It was found that individual GSH-CuNCs emitted no luminescence. On the contrary, their FI began to continuously increase along with increasing Ce3+ concentration until the maximum was reached and kept. The FI of CuNCs was collected and plotted vs concentration of Ce3+, as displayed in Fig. 6d. Their FI was proportionally enhanced when the concentration of Ce3+ increased consecutively from 4 µM to 18 µM. While the FI remained unchanged even if concentration of Ce3+ was up to 20 µM. So, concentration of Ce3+ was optimized at 20 µM. In the third step, incubation time was investigated to ensure coordination and electrostatic reaction completely, as shown in Fig. S2, 2 min was employed as the reaction time. Finally, the reaction time of Ce3+/GSH-CuNCs against DPA was plotted in Fig. S3, 1 min was applied.
Under optimization conditions, Ce3+/GSH-CuNCs ensemble was utilized as a probe to quantitatively detect DPA concentrations, as indicated in Fig. 7a. The FI of Ce3+/GSH-CuNCs decreased monotonically with increasing concentrations of DPA from 0 µM to 22 µM. It was obviously to find that DPA had the strongest quenching efficiency about 99.5 % when DPA concentration was 22 µM, calculated by the following equation: Effq% = (F0 - F)/F0. Even if GSH and DPA have the same electron donor of carboxyl oxygen atom and nitrogen atom, while DPA is a tridentate ligand and can form two stable five-membered rings with Ce3+. The stability constant of [Ce(DPA)3]3- complexes was logβ = ~22, which was much higher than that of Ce-GSH with logβ = ~ 8.22[56, 57]. The higher binding affinity between Ce3+ and DPA than Ce3+ and GSH caused detachment of Ce3+ from GSH-CuNCs to form more stable [Ce(DPA)3]3- complexes. Such changes resulted in disaggregation of GSH-CuNCs at a fast rapid, thereby decreasing the FI of GSH-CuNCs. Comparison of ultraviolet absorption spectra of Ce3+ and DPA with Ce-DPA complexes were monitored, as shown in Fig. S4. The absorption peak of Ce-DPA became sharply when Ce3+ and DPA were combinedin comparison with those of Ce3+ and DPA separately. This result further proved that Ce3+ can strongly coordinate with DPA and form stable Ce-DPA complexes. As a result, linear regression equation for (F0 - F)/F0 versus DPA concentration was described as y = 0.0518 x + 0.0042 (R2 = 0.997) ranging from 0.3 µM to 18 µM, seen in Fig. 7b. F0 and F in the above-mentioned equation indicated the FI of Ce3+/GSH-CuNCs when DPA was absent and present. The LOD for DPA was estimated to be 0.19 µM on the basis of the formula LOD = 3 σ / k, which was 315 times lower than 6×10-5 M, the infectious dosage of anthrax spores.
3.4 Selectivity and anti-interference
The properties of selectivity and anti-interference of Ce3+/GSH-CuNCs ensemble nanoprobe for DPA detection was further evaluated by conducting fluorescence response of Ce3+/GSH-CuNCs to nineteen possible coexisting interfering cations and anions: Ba2+, Mg2+, Zn2+, Ca2+, Mn2+, NH4+, Na+, K+, HCO3-, CO32-, F-, SO32-, SO42-, NO3-, Br-, I-, ClO-, COO-, NO2-, as shown in Fig. 8. Ce3+/GSH-CuNCs ensemble nanoprobe exhibited the specific interaction with DPA. Under the same experimental conditions, no significant change of the relative FI was found after addition of 19 kinds of potential substances mentioned above separately to the fluorescence detecting system without and with DPA. All the concentration of these used interferents were 200 µM as 20 times as that of DPA. The relative FI was denoted as F/F0. F and F0 were FI of Ce3+/GSH-CuNCs with and without different interfering substances when DPA was absent and present. The result was probably ascribed to the highly stability constant of Ce (III) and DPA, so Ce (III) was more inclined to coordinate with DPA rather than other coexisting cations and anions. Thus, the Ce3+/GSH-CuNCs sensor exhibited specifically identifying ability towards DPA among the other coexisting materials, suggesting that this proposed probe had great potential to be employed for detection of DPA. Besides, specific comparison of this method with other previously reported DPA sensing methods were summarized in table S2.
3.5 Analysis of real samples
With the aim of assessing the feasibility and practicality of the constructed Ce3+/GSH-CuNCs ensemble nanoprobe, DPA detection was performed using tap water and river water. In advance, two kinds of water samples were treated with a syringe filter (0.22 µm). Subsequently, the standard addition method was utilized to analyze the amount of DPA in the water samples. The recoveries of DPA determination were obtained via labeling recovery experiments, as listed in Table 1.
Table.1 Recoveries and analysis of DPA in the water samples
Sample
|
Found in sample
|
Add
|
Found
|
Recovery
|
RSD(n=3)
|
|
(µM)
|
(µM)
|
(µM)
|
(%)
|
(%)
|
Tap water
|
None
|
1
|
0.99±0.04
|
98.8
|
4.08
|
|
|
5
|
5.03±0.16
|
100.6
|
3.18
|
|
|
10
|
10.11±0.13
|
101.1
|
1.29
|
River water
|
None
|
1
|
0.91±0.05
|
91.3
|
5.49
|
|
|
5
|
5.18±0.10
|
103.6
|
1.93
|
|
|
10
|
9.78±0.07
|
97.8
|
0.72
|