Introduction

Heavy metal ions are one type of the most serious contaminants in waters. In recent years, the presence of heavy metals in ambient environment has been sharply increased, owing to the rapid development of modern industry, agriculture and transportation, and in particular the uncontrolled battery manufacturing, metal melting, and old ship demolition (Zhang 2011). Lead (Pb) is a kind of heavy metal with significant environmental and biological toxicities. It is nondegradable in the natural environment for a long time, and is harmful to organisms even at a very low concentration (Rogers et al. 2003). The sensitive detection of heavy metals and successful prevention from damaging effects at the very earliest stage is a proper way to minimize the pollution of heavy metals. Conventional analytical techniques for heavy metals, such as atomic absorption spectrometry, emission spectrometry, chemiluminescence method, X-ray fluorescence spectrometry, stripping voltammetry, and polarography, have been used for the detection of metal ions for many years. However, most of these techniques suffer the disadvantages of high costs of testing, precise and heavy instruments, and long process for preconcentration and separation of samples, limiting their use in real-time and on-site monitoring of metal ions. So the development of supersensitive, low-cost, and fast methods to detect heavy metal has attracted increasing attention in recent years.

In the last two decades, semiconductor quantum dots (QDs) have attracted much interest due to their outstanding fluorescence properties (Aldana et al. 2001; Jiang et al. 2009; Kanelidis et al. 2011; Lin et al. 2004; Pan et al. 2006), such as broad excitation bands, narrow luminescence peaks, high quantum yields, size- and composition- tunable emission, and high stability against photobleaching (Guo et al. 2003; Mulvihill et al. 2010; Siskova et al. 2011). With the development of techniques in synthesis and surface modification, high-quality QDs have been used to detect a number of substances (Duong and Rhee 2007; Geszke-Moritz et al. 2012; Gan et al. 2012; Zhao et al. 2003). Taking advantage of the rapid fluorescence response, semiconductor QDs are becoming one of the simplest, fastest, and most efficient sensors. Heavy metals in waters usually exist in the form of metallic cations. QD fluorescent probes have shown greater promise for the qualitative and quantitative analysis of heavy metal ions (Ali et al. 2007; Generalova et al. 2011; Koneswaran and Narayanaswamy 2012; Page et al. 2011; Shen et al. 2012; Shete and Benson 2009). Aiming at this application, the QDs should be water-soluble and possess stable fluorescent properties in water. It has been reported the detection of Cu2+, Pb2+, and Hg2+ ions by ZnE and CdE QDs (E = S, Se, or Te) as fluorescent probes. These QDs are mostly coated with organic reagents to improve their water solubility and surface functionalization, and dispersed in buffer solution to obtain stable luminescence in testing.

In this paper, we report the facile synthesis of silica-coated ZnS (ZnS@SiO2) QDs. ZnS is the most environmentally friendly and harmless material in II–VI compounds. Compared with the previous reports that prepared QDs with toxic precursors at high temperature, our strategy is a moderate and green method. So the preparation and application of ZnS@SiO2 QDs will hardly pollute the environment. Subsequently, the ZnS@SiO2 QDs were used as fluorescent probes for detection of heavy metal ions, and they showed fluorescence sensitivity to Pb2+ ions. In addition, the luminescence of ZnS@SiO2 QDs is stable in a wide pH range. All these tests were carried out in water without any buffer solution.

Experiment

Materials

The zinc acetate dihydrate (Zn(CH3COO)2·2H2O), oleic acid, tetraethyl orthosilicate (TEOS), triton X-100, n-hexanol, and ammonium hydroxide solution (NH3·H2O, 25.0–28.0 %) were purchased from Sinopharm Chemical Reagent CO., Ltd. The sodium oleate was purchased from Shanghai Kefeng Industry CO., Ltd. The sodium sulfide nonahydrate (Na2S·9H2O) was purchased from Beijing Chemical Works. The ethanol and cyclohexane were purchased from Tianjin Bodi Chemical Co., Ltd. All the above reagents were analytical grade and used without further purification.

Preparation of ZnS, ZnS@SiO2 and SiO2 QDs

The ZnS cores were prepared via a liquid–solid-solution (LSS) process (Wang et al. 2005). Typically, 0.1317 g (0.6 mmol) of Zn(CH3COO)2·2H2O and 10 mL of deionized water were added to a 100 mL Teflon cup. A colorless transparent solution was obtained after stirring. Then 15 mL of oleic acid, 2 g of sodium oleate, and 25 mL of ethanol were added into the autoclave in order. The mixture was stirred for 10 min to form a uniform emulsion, and then 10 mL of 0.06 M Na2S solution was added rapidly. The mixture was stirred for 5 min at room temperature. The Teflon-lined autoclave was sealed tightly and maintained at 200 °C for 12 h, and then cooled to ambient temperature naturally. The product containing ZnS QDs was stored in a well-closed container and used within 1 week.

The ZnS@SiO2 QDs were prepared through the hydrolysis of TEOS in the presence of ZnS cores at room temperature. In a typical synthesis, 1 mL of ZnS product was rinsed with ethanol and cyclohexane to remove the excess oleic acid and other impurities and the ZnS QDs were dispersed in 10 mL of cyclohexane. Then 100 μL of TEOS was injected to ZnS suspension under agitation. After 10 min, 1 mL of triton X-100 and 1 mL of n-hexanol were added sequentially and the mixture turned transparent immediately. The mixture was stirred for another 30 min to form the microemulsion system. Then 0.3 mL of NH3·H2O was introduced to initiate the polymerization. After stirring for 24 h, the growth of silica shells completed. The SiO2-coated nanoparticles were separated by centrifugation and washed with ethanol and deionized water. Finally, the ZnS@SiO2 QDs were ultrasonically dispersed in 5 mL of deionized water to form a suspension.

The SiO2 QDs were obtained by the hydrolysis of TEOS using the similar procedure with that of ZnS@SiO2 QDs, except without addition of ZnS QDs as cores.

Characterization

The ZnS and ZnS@SiO2 QDs were centrifuged and rinsed with cyclohexane/deionized water and ethanol for several times. After drying under vacuum overnight, the powders were obtained for the X-ray powder diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR) examinations. XRD studies were performed with a Rigaku D/max-rB instrument (Cu Kα1 radiation, λ = 1.54060 Å). FTIR measurements were carried out on a Thermo Fisher Scientific Nicolet 6700 FTIR spectrometer.

Photoluminescence (PL) spectra were recorded on a Jobin Yvon Fluorolog3-P spectrophotometer. All the spectra were monitored at room temperature in ambient atmosphere. The oil-soluble ZnS QDs and water-soluble ZnS@SiO2 QDs were ultrasonically dispersed in cyclohexane and deionized water, respectively, and placed in a 1 cm quartz cuvette for PL measurements. Photomicrographs were obtained from a JEM-1200 EX transmission electron microscope (TEM), using an accelerating voltage of 100 kV. The Zeta potential of ZnS@SiO2 QDs was tested with a Malvern Zetasizer NANO ZS system.

Detection of metal ions

The suspension of ZnS@SiO2 QDs was ultrasonically dispersed for several minutes before use. In a typical testing, 63 μL of ZnS@SiO2 QDs suspension was added to 2.44 mL of deionized water, and then 10 μL of metal ions solution with a certain concentration was injected to the diluted QDs suspension. The mixture was shaken by hand and left to stand for 10 min, then placed in a quartz cuvette for PL measurements.

Results and discussion

Figure 1 gives the XRD patterns of ZnS and ZnS@SiO2 QDs. The diffraction peaks at 2θ = 28.6°, 47.5°, and 56.3° in curve a could be indexed as the (111), (220), and (311) planes of ZnS in a sphalerite phase (JCPDS Card, File No. 05-0566), while the broad dispersal peaks at about 2θ = 22° in curve b could be attributed to the amorphous SiO2 (Hessel et al. 2006). The characteristic peaks of ZnS are scarcely visible in curve b, which is dissimilar to some earlier reports about silica-coated nanoparticles (Ding et al. 2012; Dong et al. 2009). Although only a small amount of ZnS cores were used in experiments, the existence of them in the final sample is confirmed by the PL results, so the peaks of ZnS are considered to be concealed by the diffraction peaks of abundant SiO2.

Fig. 1
figure 1

Powder XRD patterns of the (a) ZnS and (b) ZnS@SiO2 QDs

Full size image

Figure 2 shows the TEM images of ZnS and ZnS@SiO2 QDs. From Fig. 2a, it can be seen that the as-prepared ZnS are spherical particles, with a diameter about 6–7 nm. The ZnS QDs show good dispersion stability in cyclohexane, which is the foundation for the growth of SiO2 shells. The TEM images of ZnS@SiO2 QDs are presented in Fig. 2b, c. Each core@shell nanoparticle contains about 5 ZnS cores, and the size of ZnS@SiO2 QDs ranges from 36 to 40 nm. A great effort has been made to obtain the ZnS@SiO2 QDs with single ZnS core, but it has not been fabulously successful to this day. While the good PL performance of samples after being transferred from cyclohexane to water, which is quite different from other reports, may benefit from this imperfect structure.

Fig. 2
figure 2

TEM images of the a ZnS and b, c ZnS@SiO2 QDs

Full size image

The FTIR spectra of the bare and silica-coated ZnS QDs are illustrated in Fig. 3. The ZnS QDs were synthesized with oleic acid (OA) as oil phase solvent, and the results confirm the presence of OA ligands on the surface of ZnS QDs. The vibrational band assignment for OA (Siskova et al. 2011) is as follows: The peak at 3,009 cm−1 is from O–H stretching vibration and the peak at 2,956 cm−1 is assigned to the CH3 asymmetric stretching, the peaks at 2,919 and 2,850 cm−1 are assigned to the CH2 asymmetric and symmetric stretching, respectively, the peaks at 1742, 1596, and 1552 cm−1 originate from C=O stretching vibration, the peak at 1,466 cm−1 comes from CH3 asymmetric bending, and the peak at 1,411 is from C–H deformation of CH2 adjacent to C=O, the peak at 721 cm−1 generates from the (CH2)7 in-plane deformations rocking. It is generally known that the QDs modified with OA ligands can be easily dispersed in nonpolar solvents. However, it is quite difficult to transfer them to water and keep steady luminescence performance. For the spectrum of ZnS@SiO2 QDs, the broad -OH peak at 3,441 cm−1 indicates that a fraction of the Si atoms exist on the surface has been hydroxylated, which gives the QDs very good dispersion in water. The peak at 1,628 cm−1 is ascribed to H–O–H deformation vibration, and the peaks at 1,101 and 800 cm−1 are attributed to the Si–O–Si asymmetric and symmetric stretching, respectively. The remaining bands are related to the network bending modes of SiO2 lattice (Jeong et al. 2000; Lee and Wachs 2007).

Fig. 3
figure 3

FTIR spectra of (a) ZnS and (b) ZnS@SiO2 QDs

Full size image

The emission spectra of ZnS, ZnS@SiO2 and SiO2 QDs are shown in Fig. 4. Excited by 323 nm, the emission spectrum of ZnS QDs shows a peak at about 440 nm. This emission is ascribed to the native trap state emission of ZnS nanocrystals, corresponding to a donor–acceptor type transition (Joo et al. 2003; Murase et al. 1999). Our experimental results on the synthesis of ZnS QDs show a correlation between the emission intensity and the number of sulfur vacancies, so the self-activated centers are suggested to associate with sulfur vacancies, which is correspondent with some classic reporting examples (Becker and Bard 1983; Karar et al. 2004). While the emission peak of ZnS@SiO2 QDs is at about 450 nm, and the maximum intensity is slightly less than the uncoated sample. It is worthwhile to mention that although the number of ZnS QDs for curve a is theoretically controlled equal to the number of ZnS cores in ZnS@SiO2 QDs for curve b, the latter must be less than the former because of the unavoidable loss during experiments. That is to say, the total ZnS amount in the sample of ZnS@SiO2 is less than in the sample of ZnS QDs. So the emission intensity of ZnS@SiO2 QDs in water is considered to be satisfactory.

Fig. 4
figure 4

Emission spectra of the (a) ZnS, (b) ZnS@SiO2 and (c) SiO2 QDs (λ ex = 323 nm)

Full size image

Even the SiO2 QDs synthesized under the same conditions are not luminous (curve c), the SiO2 shells play a prominent part in the formation and application of ZnS@SiO2 QDs. First, the oil-soluble ZnS QDs are transferred to water by coating with SiO2 shells, which has been discussed in the discussion of FTIR spectra. Second, the favorable luminescence properties of ZnS@SiO2 QDs are related with SiO2 shells. As is stated above, about 5 ZnS cores are coated in a 36–40 nm sized SiO2 shell. The confinement from core@shell structure may lead to an effect similar to the aggregation of QDs, which cause the red shift of emission peak and the increasement of emission intensity (An et al. 2002; Guo et al. 2014). The ZnS cores are coated with SiO2 to prevent from direct contact with water, so the fluorescence quenching by water could be avoided to some extent. Last but not least, the SiO2 shells have much to do with the tolerance of ZnS@SiO2 for the application as fluorescent probes, which will be elaborated in more detail in the following paragraphs.

Because of the quite good dispersion and luminescent properties in water, ZnS@SiO2 QDs would be expected as fluorescent probes for the detection of heavy metal ions. However, there are some challenges to face in the application of QD fluorescent probes.

First, the water samples in environmental sciences are usually acidic or alkaline, which may greatly influence the luminescence of QDs and even lead to the corrosion of QDs. Dispersing the fluorescent QDs in a buffer solution to keep pH stable at about 7 is an effective method to solve this problem in most cases (Cai et al. 2006; Liang et al. 2010; Liu et al. 2012a, b; Luan et al. 2012). Accordingly, the experimental procedures must be increased and the test cost may be difficult to be reduced. In our study, the luminescence spectra of ZnS@SiO2 QDs with addition of HCl or NaOH in some concentration are shown in Fig. 5a, b, and the changes of PL intensity at 450 nm with dependent on the pH of water samples are presented in Fig. 5c. It is clear that the fluorescence of ZnS@SiO2 QDs is basically stable in the water sample pH range of 3–12. On the one hand, the high endurance of QDs comes from the protection of SiO2 shells, and on the other hand, it may relate with the large volume ratio between QDs systems and test samples. Here the ZnS@SiO2 QDs were dispersed in 2.5 mL deionized water and the volume of water sample is 10 μL. In a word, the ZnS@SiO2 QDs have the potential to be directly used as fluorescent probes for water samples in a broad pH range without the help of buffer solution.

Fig. 5
figure 5

PL spectra of ZnS@SiO2 QDs upon addition of different concentrations of a HCl and b NaOH solution. c Emission intensity at 450 nm of PL spectra upon addition of water sample in different pHs (λ ex = 323 nm)

Full size image

Second, the water samples in environmental generally contain a variety of metal ions, such as alkali metal ions (Na+ and K+), alkaline earth metals ions (Mg2+ and Ca2+), transition metal ions (Zn2+, Cd2+, Mn2+, Co2+, Fe3+ and Cr6+, etc.), and Al3+ which is the most abundant metal ions in crust. Some of them usually quench the emission of QDs. Selectivity is a very important parameter to evaluate the performance of a type of fluorescent probe. In this sense, several environmentally relevant metal ions were chosen to test the specific response of ZnS@SiO2 QDs fluorescence. The PL spectra of ZnS@SiO2 QDs upon addition of 12 common metal ions are shown in Fig. 6a, and the changes of emission intensity at 450 nm are given in Fig. 6b. It can be seen that Pb2+ ions induce a prominent fluorescence quenching compared to the blank sample (no metal ion was added), whereas some other metal ions led to slight or even no fluorescence change. Although the luminescence from ZnS@SiO2 QDs is suppressed by Cd2+ ions to some extent, the red shift of emission peak is the most obvious characteristic that can distinguish Cd2+ ions from Pb2+ ions. It can be seen that the ZnS@SiO2 QDs show better selectivity toward Pb2+ ions than the other metallic ions, and they may serve as fluorescent sensors for the detection of Pb2+ ions in aqueous solution.

Fig. 6
figure 6

a PL spectra and b emission intensity at 450 nm for ZnS@SiO2 QDs upon addition of different metal ions, respectively. (λ ex = 323 nm, concentration of metal ions solution: 10−4 M)

Full size image

Figure 7a illustrates the effect of the different amounts of Pb2+ ions on the luminescence spectra of ZnS@SiO2 QDs. The spectra are closely monitored from a very trace concentration (Pb2+ ions in water sample: 10−9 M) until a quite high concentration (Pb2+ ions in water sample: 2.6 × 10−4 M). With the increase of Pb2+ in water samples added, the fluorescence of ZnS@SiO2 QDs is gradually reduced until completely quenched. The PL intensities at 450 nm corresponding to the Pb2+ concentrations in water samples are listed in Fig. 7b. On the whole, all data points follow the hyperbolic distribution. Here we use the Stern–Volmer equation to analyze the PL quenching. The Stern–Volmer equation is a convenient analysis tool for photophysical and photochemical data (Bhalla et al. 2013; Carraway et al. 1991; Chen and Rosenzweig 2002; Gunnlaugsson et al. 2006; Guo et al. 2009; He et al. 2008; Landes et al. 2001). It can describe the dependence of a quenching effect on the concentration of quenchers. In our research, the fluorescence quenching by Pb2+ ions is well described by the Stern–Volmer equation

$$ {{F_{0} } \mathord{\left/ {\vphantom {{F_{0} } F}} \right. \kern-0pt} F} = 1 + K_{SV} \left[ Q \right], $$
(1)
Fig. 7
figure 7

a PL spectra and b fluorescence intensity at 450 nm of ZnS@SiO2 QDs upon addition of Pb(NO3)2 solution at various concentrations. (λ ex = 323 nm) (c, d, e, f) Stern–Volmer plots representing the quenching effects of Pb2+ ions on the fluorescence emission of ZnS@SiO2 QDs at 450 nm

Full size image

where F and F 0 are the luminescence intensities of ZnS@SiO2 QDs in the presence and absence of Pb2+ ions, K SV is the Stern–Volmer constant, and [Q] is the concentration of Pb2+ ions in QDs system. It must be noted that the concentration of Pb2+ ions in QDs system is only about 0.4 % of that in water samples, which is resulted from the experiment itself. Figure 7c shows the Stern–Volmer plots representing the quenching effects of Pb2+ ions on the fluorescence emission of ZnS@SiO2 QDs. A quite good linear relationship between F 0/F and the concentration of Pb2+ ions can be observed in the Pb2+ concentration range from 4 × 10−12 to 1.04 × 10−6 M (corresponding to 10−9 to 2.6 × 10−4 M in water samples), and K SV is determined to be 1.51 × 107 M−1 (R 2 = 0.9794). The data were analyzed and processed for several times to find the sensitive response range, and the results of data fitting are shown in Fig. 7c–f and listed in Table 1. There are quite little changes in K SV and R 2 based on different intervals, implying an even detection sensitive in the whole experimental range. An equation very similar to the Stern–Volmer equation is used to represent the fluorescence quenching by Pb2+ ions with greater facility in the potential application

$$ {{F_{0} } \mathord{\left/ {\vphantom {{F_{0} } F}} \right. \kern-0pt} F} = 1 + k\left[ {Q'} \right], $$
(2)
Table 1 The results of data fitting
Full size table

where k is the quenching constant and [Q’] is the concentration of Pb2+ ions in water samples. The fitting results are also expressed in Table 1.

The wide detection range from 10−9 to 2.6 × 10−4 M made the ZnS@SiO2 QDs-based sensing system competent for Pb2+ analysis at low to high concentrations. There was no obvious decrease in the luminescence intensity of ZnS@SiO2 QDs in the experimental condition within 2 h. Furthermore, the fluorescence from ZnS@SiO2 QDs stored as undiluted suspension was with great consistency for longer than 2 months. It is indicated that the ZnS@SiO2 QDs displayed a relative luminescent stability, making them suitable for analytical application.

There are two main types of explanations for the mechanism of fluorescence quenching by heavy metal ions in previous reports. First, the quenching may be caused by the displacement of Zn2+ or Cd2+ in ZnE or CdE QDs (E = S, Se, or Te) by heavy metal ions (Chan et al. 2010; Liu et al. 2012a, b; Luan et al. 2012; Shen et al. 2012; Sung and Lo 2012; Wang et al. 2011; Yang et al. 2011). It is due to the higher binding affinity to E2− ions and indicated by the lower solubility product constant (K sp) than ZnE or CdE. The formed ultrasmall particles of heavy metal sulfides can quench the fluorescence of QDs by facilitating electron–hole nonradiative annihilation (Dong et al. 2006; Isarov and Chrysochoos 1997).

Second, the competitive binding of molecules between QDs and heavy metal ions make the ligands strip from the surface of QDs and sometimes may cause the aggregation of QDs (Ali et al. 2007; Cai et al. 2012; Liang et al. 2010; Wang and Guo 2009; Wang et al. 2012a, b; Wu et al. 2010; Xie et al. 2012). The QDs are usually capped with some modifiers, such as cysteine, glutathione, and biological molecules etc., containing free –NH2, –COOH, –SH, or some other groups to selectively interact with certain heavy metal ions. It is apparent that the second theory is not applicable here, because the ZnS@SiO2 QDs were directly used without any further surface modification. While the inadaptability of the first theory for our results will be expounded in the following.

The hydroxylated SiO2 shells are negatively charged, which has been affirmed by Zeta potential (−47.6 mV). The free metallic cations can be attracted to surfaces of SiO2 shells by the electrostatic interaction. The SiO2 shells prepared by microemulsion method under the low temperature are generally porous (Darbandi et al. 2007; Shajesh et al. 2009; Xu et al. 2013). It is possible for the metal ions bounded on QDs surface to pass through SiO2 shells and interact with ZnS cores, and the Zn2+ ions on the surface of ZnS cores can be exchanged by some metal ions, resulting in a fluorescence response. It is reasonable that the Cu2+ ions are more easily deposited than Pb2+ ions judging from the solubility product constants (Ksp(CuS) = 6.3 × 10−36, Ksp(PbS) = 8.0 × 10−28), hence the fluorescence of QDs should be more dramatically quenched by Cu2+ ions than by Pb2+ ions in accordance with the above first theory. However, the facts are just the opposite. The fluorescence of ZnS@SiO2 QDs is most significantly quenched by Pb2+ ions in our study. So here, we propose an extended hypothesis to clarify the problem.

Undoubtedly, there are ultrasmall particles of CuS or PbS formed on the surfaces of ZnS cores. These ultrasmall particles can quench the fluorescence of ZnS cores and act as electron–hole nonradiative recombination centers (Isarov and Chrysochoos 1997). As we all known, ZnS is an important semiconductor materials with a bulk band gap of 3.65 eV, which is the largest value of all II–VI compound semiconductors. CuS and PbS are both narrow gap semiconductors. CuxS as bulk material exists in five stable phases at room temperature, with the band gap varying from 1.05 to 1.7 eV and the transitions being direct or indirect (Nascu et al. 1997; Rodriguez-Lazcano et al. 2009). Bulk PbS is a direct band gap semiconductor with a band gap of 0.41 eV (Machol et al. 1993; Wang et al. 1987). Quantum confinement in the QDs results in discrete energy levels and an increase in the effective band gap. The band gap of CuS nanocrystals has been reported to be in the range between 1.25 and 2.81 eV (Mageshwari et al. 2011; Nair et al. 1998; Raevskaya et al. 2004; Yildirim et al. 2009). The band gap of PbS nanocrystals has been reported to be in the range 0.5–2.32 eV (Joshi et al. 2004; Mukherjee et al. 1994; Nanda et al. 2002) and even as high as 5.2 eV (Thielsch et al. 1998). Although the exact band gaps of CuS and PbS ultrafine grains formed on the surface of ZnS cores are unknown, we assume that the electron and hole transfer from the excited ZnS to the energy levels of PbS may be faster than to that of CuS, bringing about the quicker fluorescence quenching by Pb2+ ions.

Conclusions

In summary, we have prepared ZnS@SiO2 QDs by a facile and green process. The ZnS cores were synthesized by a moderate LSS method, and the oil-soluble ZnS QDs were successfully transferred to water through the coating of SiO2 shells. The photoluminescence properties of the core@shell structural nanoparticles were investigated. With the help of SiO2 shells, the ZnS@SiO2 QDs presented an acceptable fluorescence in water, and the luminescence was quite stable after mixing with acidic or alkaline water samples. The QDs were used as fluorescent probes for detection of heavy metal ions without the need for additional buffer solution, and they showed a sensitive fluorescence response to Pb2+ ions. The fluorescence quenching was well described by the Stern–Volmer equation, and an even quenching constant was obtained when the Pb2+ ions concentration in water samples ranging from 10−9 to 2.6 × 10−4 M. The ZnS@SiO2 QDs are expected to become a potential sensor for Pb2+ analysis at low to high concentrations.