1 Introduction

Surface modification methods to synthesize water-dispersible semiconductor nanocrystals (NCs) have been intensively studied in various fields for various application purposes [1,2,3]. However, there are still many challenges remained to synthesize very well-dispersed colloidal NCs since most semiconductor nanocrystalline materials have extremely low solubility in water [4]. In addition, these materials have critical problems such as a high biological toxicity which should be solved to be applied in any biological system [5]. In these regards, there have been great amounts of publications regarding the development of novel types of polar and non-toxic surface-capping ligands to enhance the low solubility of the NCs in water, and to reduce the cytotoxicity of the NCs in a biological system as well [6, 7]. Especially, bio-friendly natural polymer chitosan was found to be proper capping agent which has high solubility in water and low biological toxicity at the same time [8]. Therefore, the use of the chitosan as surface-capping agent to prepare colloidal non-toxic semiconductor NCs for developing bio-labeling agents for medical diagnostic purposes became very popular [9, 10]. In addition, recently, application area of the chitosan-capped NCs are greatly enlarging to photocatalyst [11] and biological photosensor [12] due to their unusual physical, optical and electro-chemical properties in aqueous solution. Previously, we have reported the syntheses of the water-dispersible ZnS:Mn NCs using various organic polymers as surface-capping agents [13, 14]. In this study, the ZnS semiconductor parent materials were especially chosen because they do not contain biologically and environmentally hazardous elements such as Cd and Hg, which are commonly utilized precursor elements in many other nano-sized semiconductor materials studies. In addition, we showed that those polymer coated ZnS:Mn NCs had very low toxicity effects over the growth of E. coli bacteria in situ. Therefore, those organic polymer-capped colloidal ZnS:Mn NCs were considered as highly efficient fluorescence bio-labeling materials with low cytotoxicity, which can replace the toxic cadmium-based traditional semiconductor parent core materials. In the article, we describe the syntheses and application of the colloidal chitosan-capped ZnS:Mn NCs as an efficient photosensor in the detection of certain heavy transition metal ion in an aqueous solution sample. The chitosan-capped ZnS:Mn NCs have been made and applied for organic glucose molecule and iodine sensor [10, 12]. However, they have never been applied for the detection of biologically important heavy transition metal ions in any water sample.

2 Experiments and discussion

2.1 Synthesis of the ZnS:Mn-Chit NCs

Colloidal ZnS:Mn-Chit NCs were prepared and characterized by following previously reported methods to synthesize the chitosan-capped ZnS:Mn NCs in the literature with slight modifications [12]. One of the modifications was made at the final step to control the pH of the colloidal NCs to induce negative surface charge on the surface of the NCs for further sensor application. 5.0 mmol of ZnSO4 solid (Sigma-Aldrich) was dissolved in 25 mL of freshly distilled and filtered deionized water and combined with 25 mL of aqueous solution which contains 10 mmol of chitosan (Sigma-Aldrich, Mw: 120,000 g mol−1; deacetylation ratio: 80%; viscosity: 20–30 cP in 1% acetic acid solution) at 0 °C using an ice-water/salt bath. The solution was naturally warmed to room temperature with slow stirring to presumably form [Zn-chitosan] precursor composites in the mixture solution. In a different Pyrex glass flask, MnSO4 (Sigma-Aldrich, 0.1 mmol) and Na2S (Sigma-Aldrich, 5.0 mmol) solids were dissolved together in 20 mL of diluted sulfuric acid solution (1.0 × 10–6 M). This resulting colorless clear solution was added to the flask containing the [Zn-chitosan] precursor composites with continuous stirring followed by the solution was heated to reflux temperature (ca. 90 ℃). After reflux for 20 h, the electrical heater was taken out and the whole solution was allowed to gradually cool down to the room temperature. A 100 mL of Absolute ethanol was poured to induce the formation of white–gray precipitates at the bottom of the flask. The product NC powders were separated using a centrifuge machine (4000 rpm) followed by decanting the supernatant liquid. The off-white powders were quickly washed with cold ethanol/water (1:1) mixture and dried (ca. 60 ℃) for overnight in a high-vacuum oven. Finally, the isolated and dried NC powders were dissolved in freshly distilled DI water and 1.0 × 10–6 M of diluted sodium hydroxide (NaOH) solution was slowly and carefully added until the pH of the whole mixture solution reaches to 10 checking with a digital pH meter to prepare the colloidal ZnS:Mn-Chit NC samples. The essential experimental data obtained from the colloidal ZnS:Mn-Chit NCs are summarized and listed in Table 1.

Table 1 Data Summary of the synthesized ZnS:Mn-Chit NCs
Full size table

2.2 Instrumentations

The presented UV–Visible absorption spectra (Fig. 1a) were obtained using a Lamda-25 spectrophotometer (Perkin-Elmer), and the solution photoluminescence (PL) spectra (Fig. 1b) were obtained using a LS-45 spectrophotometer (Perkin-Elmer) equipped with a 500 W Xenon lamp light source. The presented high resolution-transmission electron microscopy (HR-TEM) image in Fig. 2a) and the energy dispersive x-ray spectroscopy (EDXS, Fig. 2b) were obtained using a JEOL JEM 1210 electron microscope. The magnification range of the microscope was from 1000 to 800,000 and the accelerating voltage was from 40 to 120 kV. The crystalline X-ray diffraction (XRD) pattern diagrams presented in Fig. 3 were obtained using an X-ray diffractometer (Rigaku 300) which used a Cu–Kα (1.54 Å wavelength) light source. The elemental analysis for the ZnS:Mn-Chit NCs were additionally performed using a Perkin-Elmer Optima-430 ICP-AES spectrometer. The surface-capping ligands, chitosan molecules, were investigated on their vibrational modes using Fourier transform infrared spectroscopy (FT-IR, Fig. 4) recorded by a Perkin-Elmer Spectrum One (resolution of 0.5 cm−1) spectrophotometer having an attenuated total reflection (ATR) unit. Finally, the surface charge and the particle size distribution diagram for the colloidal ZnS:Mn-Chit NCs were obtained using Otsuka ELSZ-2000ZS spectrophotometer equipped with an avalanche photodiodes detector (Fig. 5a and b).

Fig. 1
figure 1

a UV–Visible absorption spectrum and b room temperature photoluminescence (PL) spectrum of the ZnS:Mn-Chit NCs

Full size image
Fig. 2
figure 2

a HR-TEM image and b EDXS diagram of the ZnS:Mn-Chit NCs

Full size image
Fig. 3
figure 3

XRD pattern diagram of the ZnS:Mn-Chit NCs

Full size image
Fig. 4
figure 4

FT-IR spectra of the ZnS:Mn-Chit NCs (red) and the neat chitosan (black)

Full size image
Fig. 5
figure 5

a zeta-potential distribution diagram (− 33.4 mV) and b size distribution diagram (avg. 385 nm) of the ZnS:Mn-Chit NCs

Full size image

2.3 Characterizations of the colloidal ZnS:Mn-Chit NCs

The specific optical properties of the colloidal ZnS:Mn-Chit NCs were investigated using UV–visible (Fig. 1a) and photoluminescence (PL) spectroscopy (Fig. 1b). The UV–visible absorption spectrum showed the band energy gap caused by the conduction and the valence bands separation in the ZnS parent nanocrystals [15], and the band gaps for the chitosan-capped ZnS:Mn NCs (3.75 eV) were slightly higher than that of bulk ZnS:Mn crystals (3.54 eV), which were probably caused by the well-known confinement effects for the nano-sized semiconductor materials [16]. The PL spectrum for the ZnS:Mn-Chit NCs showed broad emission peaks appeared at 597 nm. The presented PL spectrum was obtained by setting the excitation wavelengths of the light source of the spectrophotometer at that of UV–Visible absorption peak (335 nm) obtained from the ZnS:Mn-Chit NCs NCs. The emission peak appeared around at 597 nm from the ZnS:Mn-Chit NCs was resulted from the incorporated dopant metal ion (Mn2+), which can be also expressed as a 4T1–6A1 transition [17]. The large Stokes shift (262 nm) between the absorption and emission wavelengths for the ZnS:Mn-Chit NCs was probably caused by the additional surface defect energy levels, which are formed close to the conduction bands of the ZnS parent crystals [18]. This phenomenon is one of the typical properties appeared in most semiconductor nanocrystalline materials doped with transition metal ions [19].

The relative PL quantum yield for the colloidal ZnS:Mn-Chit NCs was experimentally measured and calculated by following the exactly same method reported by Williams et al. [20]. This method includes a calculation of a relative quantum yield of certain NCs by comparing to a well-characterized standard reference material, which has a similar emission wavelength to the ZnS:Mn-Chit NCs. A 0.1 M H2SO4 aqueous solution containing quinine sulfate (Fluka) was used in this study as the standard reference because it is a commonly used standard solution for ZnS:Mn NCs whose absolute quantum yield was known as 54.6% at 22 °C. The excitation wavelength for the quinine sulfate reference dye was fixed at 335 nm, which was same with that for the ZnS:Mn-Chit NCs shown in the UV–Visible absorption spectrum. The solution photoluminescence spectra for both the ZnS:Mn-Chit NCs and the quinine sulfate dye were recorded at five different molar concentrations. Then, a diagram including the integrated PL intensity versus the absorbance gained from the UV–visible spectrum for the both samples was drawn together. Finally, the relative quantum efficiency for the ZnS:Mn-Chit NCs against the reference was calculated using the following equation:

$${\text{QEz}} = {\text{QEref}}\left( {\frac{{{\text{Grad}}_{{\text{z}}} }}{{{\text{Grad}}_{{{\text{ref}}}} }}} \right)\left( {\frac{{ \in_{{\text{z}}}^{{2}} }}{{ \in_{{{\text{ref}}}}^{{2}} }}} \right)$$

In above equation, QE is abbreviation of quantum efficiency, and ‘ref’ and ‘z’ represent the reference (quinine sulfate dye) and the ZnS:Mn-Chit NCs respectively. Moreover, ‘Grad’ and ‘ɛ’ are the gradient and the refractive index of the used solvent in this experiment, respectively. However, the solvent factor was actually eliminated by performing the analysis using the same solvent (deionized water) for the reference and the ZnS:Mn-Chit NCs. As a result, the obtained relative quantum yields of the colloidal ZnS:Mn-Chit NCs was 6.5%, which are quite similar but slightly higher than that for other water-soluble ligands capped ZnS:Mn NCs (in a range of 2–5%) reported by other research groups [21].

The average size of the ZnS:Mn-Chit NCs was directly measured from the obtained pictures using a high resolution-transmission electron (HR-TEM) microscopy as presented in Fig. 2a. In the picture, by measuring the 20 identifiable spherical shaped particles also having fringe images, we obtained average particle size of the ZnS:Mn-Chit NCs as 7.5 nm. Especially, those clearly observable fringe images in the picture with about 4 Å lattice spacings strongly indicated that the solid-state NC products are majorly made of single crystalline materials.

The analysis of the molar concentration of the dopant Mn2+ ions in the ZnS:Mn-chit NCs was carried combining EDXS (energy dispersive X-ray spectroscopy, Fig. 2b) and ICP-AES (Inductively Coupled Plasma-Atomic Emission Spectrometry) together for better precision. We intended to adjust the concentration of the dopant Mn ions to be approximately 2.0% against the ZnS parent crystals because it has been reported as the condition resulting the best fluorescence quantum efficiencies for many other kinds of ligand-capped ZnS:Mn NCs in the literature [18, 19]. The actually obtained atomic percentages of the Mn2+ activator ions included in the ZnS parent crystals for the ZnS:Mn-Chit NCs were 1.7% (EDXS; 39.8% (Zn): 35.9% (S)) and 1.5% (ICP-AES; 36.8% (Zn): 32.9% (S)), respectively. The other elemental contents, such as carbon and silicon, are probably caused by the sample grids for those analyses.

In Fig. 3, we provided X-ray diffraction (XRD) pattern diagrams recorded from the powder sample of the ZnS:Mn-Chit NCs. The diagram obtained from the NC samples showed broad peaks, whereas those for the known bulk ZnS:Mn solid are very sharp. This peak broadening is one of the well-known characteristics of nano-sized semiconductor materials in the solid state [22]. However, three peaks in those diagrams, which were assigned as (111), (220), and (311) planes, matched to a cubic zinc blende phase (JCPDS 05-0566). Therefore, the ZnS:Mn-Chit NCs have the cubic zinc blende parent crystal structure which is kinetically favored form for the solid-state ZnS crystals [23]. Moreover, using the obtained XRD peak data and the Debye–Scherrer formula [24], we calculated the particle size of the ZnS:Mn-Chit NCs to compare with that manually measured from the HR-TEM image in Fig. 2a. As a result, we obtained the average particle size for the ZnS:Mn-Chit NCs as 7.2 nm, which was very close to that measured from the HR-TEM picture.

The specific vibrational modes of the surface-capping ligand chitosan attached on the surface of the ZnS:Mn-Chit NCs were studied by analyzing the obtained Fourier Transform Infrared (FT-IR) spectral data for the NCs and the neat chitosan molecule for comparison. Figure 4 shows FT-IR spectra obtained from the colloidal ZnS:Mn-Chit NCs. The FT-IR peaks were assigned based on comparison to that for the free chitosan molecules, combining computational calculations as in the reference literature [25]. In the FT-IR spectrum, the primary bands for the chitosan molecules directly attached on the ZnS:Mn nanocrystal surface were assinged as: stretching peaks of O–H and N–H groups (3460 cm−1), symmetric and assymmetric C–H stretching peak (2980 cm−1), bending peaks of the C–N moieties (1460 cm−1), stretching peaks of the C–C–O bonds (1045 cm−1), finally bending peaks of the C–H bonds (878 cm−1).

The surface properties of the ZnS:Mn-Chit NCs were also investigated since they are directly related to their sensor activities in water. We measured surface charges of the ZnS:Mn-Chit NCs using a Zeta-PSA spectrophotometer. The zeta potential of the colloidal nanocrystals at ambient temperature was measured as − 33.4 mV (Fig. 5a). The negative surface charge was probably induced by the adjusting the pH of the solution at 10. Intention of raising up the pH of the solution (e.g. pH 13) by further addition NaOH solution actually caused degradation of the whole ZnS:Mn-Chit NCs producing gray-black precipitates at the bottom of the flask with no emission light from the NCs. In addition, the degree of aggregation of ZnS:Mn-Chit NCs at this pH condition was measured using the same Zeta-PSA spectrophotometer (Fig. 5b), which showed the formation of huge aggregates (approximately 50 NC particles) with an average size of 385 nm, indicating that there are some strong inter-molecular interactions such as hydrogen bonding between the chitosan molecules of the neighboring ZnS:Mn-Chit NCs.

2.4 Photosensor activities of the ZnS:Mn-Chit NCs

The prepared colloidal ZnS:Mn-Chit NCs were applied as a photosensor in the detection of specific cation species among the first-row divalent transition metal cations in a water sample. Figure 6a shows fluorescence images of the ZnS:Mn-Chit NCs upon addition of aqueous solution containing various divalent transition metal cations at the same conditions in terms of concentration, pH and temperature. These images were taken under irradiation of UV light (325 nm) using a conventional 6 W-UV lamp (UV-Tech Inc.), which has a light source with the similar wavelength to the absorption wavelength of the ZnS:Mn-Chit NCs. In the picture, the yellow-orange emission lights of the ZnS:Mn-Chit NCs were not changed by most added metal ions, whereas addition of Cu2+ caused significant quenching the original luminescence light of the NCs. Additionally, Fig. 6b presents the PL diagrams of the ZnS:Mn-Chit NCs upon addition of copper (II) metal ions comparing to the initial blank (NCs only) solutions. As can be seen in those spectra, the intensities of the emission peaks from the ZnS:Mn-Chit NCs decreased and ultimately completely diminished by the gradual addition Cu2+ ions. Therefore, the ZnS:Mn-Chit NCs can be applied as convenient and efficient photosensor materials in the detection of Cu2+ ions in a tap water or a waste water sample at ambient conditions without using advanced sensing devices. Development of a highly sensitive copper ion sensor is important issue since the copper ions are one of the most essential components to human life. For example, it has been known that failing in controlling the concentration of copper ions in human body can cause serious disease such as Menkes and Parkinson disease [26].

Fig. 6
figure 6

a Fluorescence images of ZnS:Mn-Chit NCs upon addition of corresponding divalent transition metal ions, taken under irradiation of UV lamp light (325 nm). (‘BLK’ refers to NCs only), b emission (PL) spectral changes of ZnS:Mn-Chit NCs upon addition of copper (II) ions in water, c linear fitting diagrams showing the Stern–Volmer relationship between the PL intensity changes and the concentrations of the added copper (II) ions to the ZnS:Mn-Chit NCs (k = 1.7 × 107 M−1, and R2 = 0.996)

Full size image

The limit of detection (LOD) [27] using those NCs by the added copper (II) metal ions (in molar concentrations, [MNC]) were measured as 1.8 × 10–8 M (ZnS:Mn-Chit NCs, 10.0 mg L−1), which was fairly close to that for other ligands (such as 4-mercaptophenylacetic acid) capped ZnS:Mn NCs reported in the literature [28]. In addition, as a photosensor for the copper ion detection, other types of semiconductor nanocrystals, such as CdSe NCs (LOD of 5 nM), showed higher sensitivity than the ZnS:Mn-Chit NCs at similar conditions [29]. However, one of the advantages of the chitosan capping for the ZnS:Mn NCs over the other organic ligands is that chitosan has much lower cytotoxicity since it was made from natural sources [30]. Moreover, the ZnS:Mn NCs do not have biologically toxic components comparing to the CdSe NCs, which makes the ZnS:Mn-Chit NCs suitable for a direct bio-medical or environmental application.

To speculate the quenching mechanism for the ZnS:Mn-Chit NCs by the addition of Cu2+ ions, a Stern–Volmer kinetic study [31] was employed, which measures the degree of decrease in the PL intensity of the ZnS:Mn-Chit NCs as the concentration of Cu2+ ions was gradually increased. Figure 6c shows that the resulting data were fitted well into the modified (a log dependent) first-order Stern–Volmer equation. This plot showed a fairly good linear relationship in the molar concentration range from 0 to 140 µM (R2 = 0.996) of the quencher. The experimental range of the concentration of the copper (II) ions was carefully selected to avoid complete quenching (I = 0) in the PL spectrum. The quenching rate constant for the ZnS:Mn-Chit NCs, kr = 1.7 × 107 M−1, was calculated from the slope of the well fitted line in the diagram. The plot of Fig. 6c strongly suggested that the fluorescence quenching process of the ZnS:Mn-Chit NCs probably goes through ionic collisions between the weakly negatively charged ZnS:Mn-Chit NCs and the added Cu2+ positive metal ions because the kinetic study result was quite similar to that for other ligands capped ZnS:Mn NCs with Cu2+ ions, which were known as taking ionic collision pathway in their emission quenching process [32].

Regarding the fluorescence quenching of the ZnS:Mn-Chit NCs by copper (II) ions, we searched some similar cases reported in the literature. First of all, Bo et al. [33] has reported that the colloidal mercaptopropionic acid (MPA)-capped CdTe NCs had shown similar fluorescence quenching effect by the addition of Cu2+ ions among the first-row divalent transition metal ions. Their kinetic studies proposed that the quenching mechanism was initiated by ionic collision between the Cu2+ ions and the negatively charged CdTe-MPA NCs in water. In addition, another similar fluorescence quenching effect was reported by Chen et al. [34], that the luminescence light of the thioglycerol capped CdS NCs was exclusively quenched by addition of Cu2+ ions. In this article, the authors provided an experimental evidence of the electron transfer between the Cu metal ion and the NCs using an electron paramagnetic resonance (EPR) spectroscopy. The initially paramagnetic Cu2+ (d9) ions were consequently reduced to diamagnetic Cu1+ (d10) species after addition to the CdS NCs, which indicated that the electron transfer between the Cu metal ion and the NCs were actually occurred to prevent recombination process of the excitons in the CdS NCs. This EPR study strongly suggested that there was one-equivalent electron transition from the NCs to the added copper (II) ions. Among the first-row transition metal ions, only the Cu1+ ion has been known as stable species in water, whereas other metal ions with a stable (+1) oxidation state are extremely rare in the literature [35]. Therefore, the copper (II) ions could cause the exclusive fluorescence quenching effect for the chitosan-capped ZnS:Mn NCs.

There are some different propositions regarding the fluorescence quenching mechanisms for the semiconductor colloidal NCs upon addition of transition metal ions in the literature. For instance, Xie et al. [36] reported that they observed an exclusive fluorescence quenching for the bovine serum albumin (BSA) capped CdSe/ZnS NCs upon addition of Cu2+ ions in aqueous solution. In this study, the authors proposed that the luminescence light quenching was caused by the formation and growth of additional crystals, such as CuS, on the surface of the NCs. The small Cu2+ ions (ionic radius of 73 pm) can easily penetrate between the steric BSA capping ligands to react with sulfide (S−) ions to form small CuS crystals on the surface of the NCs. The growth of the CuS can gradually destroy the CdSe/ZnS NCs and the CuS crystals have extremely low solubility in water which can cause precipitation of the whole composites at the bottom to rapidly induce the fluorescence quenching of the NCs. However, there was none of experimental evidence indicating the formation of the CuS crystals on the surface of the chitosan-capped ZnS:Mn NCs in their XRD pattern diagrams taken after the addition of Cu2+ ions to those NCs. Moreover, we did not observe any dark precipitation formation after addition of Cu (II) ions to the NCs as described in that paper. Therefore, we concluded that the described fluorescence quenching mechanism for the CdSe/ZnS-BSA NCs cannot be applied for the ZnS:Mn-Chit NCs.

3 Conclusion

In this article, we experimentally proved that the chitosan can be useful and efficient surface-capping agent for the synthesis and the novel application of the colloidal ZnS:Mn NCs. The well-characterized colloidal ZnS:Mn-Chit NCs were applied as a photosensor in the selective detection of transition metal ion species in water sample. As a result, the ZnS:Mn-Chit NCs showed an exclusive fluorescence quenching effect by the addition of Cu2+ ions among other first-row divalent transition metal cations. This result was quite different comparing to that for the other chitosan-capped ZnS:Mn NCs, which suggested that one can control the sensing ion selectivity of the NCs simply by controlling the preparation conditions such as pH of the NCs in water.