One-pot preparation of graphene–Ag nano composite for selective and environmentally-friendly recognition of trace mercury(II)

Abstract

A graphene–silver nano composite (G–Ag NPs) has been identified and prepared by a simple one-pot redox method using ascorbic acid as the catalytic reagent. After its structure was characterized by UV-vis spectroscopy, Raman spectroscopy and transmission electron microscopy, the absorption spectral properties were investigated in detail. The characteristic absorption peak of the as-prepared G–Ag NPs at ca. 420 nm could be selectively quenched upon addition of Hg²⁺ in the presence of ascorbic acid, owing to the formation of a specific amalgam between mercury and silver nano particles. Under pH 5.0, the response possessed an excellent selectivity and sensitivity to Hg²⁺ over the linear range of 2.0 × 10⁻⁷ to 3.25 × 10⁻⁶ M with a detection limit of 3.1 × 10⁻⁸ M. Other commonly existing metal ions had almost no influence on the absorption response. Importantly, the proposed low-toxicity G–Ag NPs were successfully applied to detect trace Hg²⁺ in three real environmental water samples with a relative standard deviation (R.S.D.) of less than 4.6% (n = 5) and with little secondary pollution.

---

1. Introduction

With rapid development of modern industry, agriculture and mineral industries, numerous heavy metals have been released from the coal-burning power plants, volcanic emissions, mining, solid-waste incineration, and so on.^1,2 Among heavy metal pollutants, Hg²⁺ is one of the most toxic and dangerous because of its strong affinity to S-containing ligands, which can cause serious dysfunction of living cells and a wide variety of diseases in the brain, kidneys, central nervous system, etc.^3–5 During the past few decades, great efforts have been devoted to efficiently and conveniently detect and separate Hg²⁺ from environmental water.^6–10 By virtue of the special affinity between Ag nanoparticles (Ag NPs) and mercury to form a solid amalgam-like structure, Ag NPs have been widely applied for Hg²⁺ detection. For example, Shen, et al.¹¹ developed a Ag NPs-based sensor for the rapid detection of toxic Hg²⁺ with a good linear relationship and a low detection limit. However, free Ag NPs are toxic and no reports have been published regarding the separation of Hg²⁺ and sensing materials as far as we know.

Since it was discovered in 2004,¹² graphene has been especially useful as an extremely versatile platform for constructing various sensing materials with high performances, owing to its unique two-dimensional honeycomb structure, high specific area and superior electrical conductivity. A plethora of multifunctional graphene-based composites have been designed and constructed for Hg²⁺ detection, including DNA covalently-linked to graphene oxide,¹³ reduced graphene oxide–Fe₃O₄ nanocomposites,¹⁴ thymine functionalized graphene oxide,¹⁵ and polypyrrole-reduced graphene oxide composite.¹⁶ However, to our knowledge, there are no reports on the colorimetric response of Hg²⁺ using Ag NPs-functionalized graphene to detect Hg²⁺ based on an action mechanism for the formation of Hg–Ag alloys on the surface of Ag NPs in the presence of citrate sodium as the reducing agent.

In this manuscript, we describe a sensing system for Hg²⁺ detection with low toxicity and little secondary pollution. First, Ag NPs-functionalized graphene (G–Ag NPs) was identified as the target sensor and prepared by a simple one-pot redox method using ascorbic acid as an environmentally-friendly catalytic reagent. After the structure and properties were characterized in detail, the mutifunctional G–Ag NPs were shown to possess a high selective response to Hg²⁺ by virtue of the formation of an amalgam with silver.^17–19 The high specific area of the graphene scaffold²⁰ and a “molecular wire effect” from the synergistic interaction between graphene's π-conjugated backbone and the d-orbital of Ag NPs^21–23 also contributed to the selectivity of the system. The G–Ag NPs have low toxicity and can be easily separated from the environmental sample after treatment, such that the proposed sensing procedure will be environmentally-friendly with little secondary pollution.

2. Experimental

Mercury(II) chloride, ascorbic acid, silver nitrate and all the other chemical reagents were purchased from Shanghai Chemical Reagent Company and were of analytical grade; all chemicals were used directly as received without any further purification. Throughout the study, water was doubly deionized before use. Graphene oxide (GO) dispersion was prepared according to our previous report.²⁴

Phosphate buffers were prepared by mixing 0.01 mol L⁻¹ solutions of H₃PO₄, K₂HPO₄, KH₂PO₄ or KOH in a proper ratio to acquire the desired pH.

The morphology of the as-prepared G–Ag NPs was characterized by the transmission electron microscopy (TEM) using a JEOL JEM-2100F TEM (operated at 200 kV). Raman measurements were conducted with a Renishaw 2000 laser Raman microscope equipped with a 514 nm argon ion laser of 2 mm spot size for excitation. Absorption spectra were obtained by a lambda-750 UV-vis-NIR spectrophotometer using a 1 cm square quartz cell. Content of Hg in the filtrate was determined using an Inductively Coupled Plasma Mass Spectrometer (ICP-MS; Perkin-Elmer, Elan DRC Plus). All pH measurements were obtained using a basic pH meter PHS-25 (Jingke Scientific Instruments Co., Ltd., Shanghai, China).

Target G–Ag NPs were prepared by thoroughly mixing 2.0 mL of 28% (m/m) ammonia water with 100 mL of 1.0 × 10⁻³ mol L⁻¹ AgNO₃ solution to obtain [Ag(NH₃)₂]⁺ solution. At the same time, 100 mL of 0.1 mg mL⁻¹ GO dispersion liquid was adjusted to pH 10 by the addition of ammonia water to produce GO with negative surface charges (O⁻). Then, the [Ag(NH₃)₂]⁺ solution was added drop-wise into the as-obtained GO dispersion liquid, and the mixture was stirred for 2 h at room temperature to ensure adequate absorption of [Ag(NH₃)₂]⁺ onto the surface of the GO nano sheets via electrostatic attraction. A freshly prepared aqueous solution of ascorbic acid (0.5%, m/m) was then added dropwise into the mixture and stirred for another 2 h in an ice-water bath to reduce [Ag(NH₃)₂]⁺ to Ag NPs. The mixture was further heated at 80 °C for 8 h. After cooling down to room temperature, the dispersion was centrifuged, washed thoroughly with doubly deionized water and dried in an oven at 60 °C overnight to obtain the target G–Ag NPs.

To test the response to metal ions, 1.0 mL phosphate buffer (pH 5.0), 1.0 mL aqueous solution of G–Ag NPs (10 μg mL⁻¹) and 1.0 mL metal ion test solution (3.25 × 10⁻⁴ M) were combined and transferred to a 10 mL volumetric flask. The mixture was diluted to 10 mL with doubly deionized water and its absorption spectrum was measured from 200 nm to 700 nm with the band-slit set at 1.0 nm. The absorption intensity change (ΔA) of the system at 420 nm was used for the quantitative and qualitative analyses, in which ΔA represents the decreased absorption intensity of G–Ag NPs, i.e., ΔA = A₀ − A, where A₀ and A are the absorption intensities of the systems in the absence and presence of Hg²⁺, respectively.

Cytotoxicity was evaluated by 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay (MTT 5 mg mL⁻¹ in PBS). Cells were seeded into 96-well microculture plates under the given conditions (6 × 10⁴ cells per mL, 100 μL per well in complete culture medium) for 24 h. Subsequently, the media were replaced by solutions (100 μL per well) containing various concentrations of G–Ag NPs and Ag NPs in the range from 1 to 10 μg mL⁻¹ for 48 h. After the incubation period, solutions containing different nanoparticles were removed and 100 μL per well DMEM medium (Invitrogen, 11960-044) with 10% MTT stock solution were added to each well, followed by another 1 h of incubation. The mixture was removed and replaced by 100 μL DMSO per well followed by another 60 min incubation period at room temperature. A microplate reader (BioTek, Power Wave XS) was used to determine the absorbance at 570 nm. The estimation was based on one experiment comprising four replicates per concentration level.

3. Results and discussion

3.1 Preparation and characterization of G–Ag NPs

Two solutions bearing opposite charges were combined to form G–Ag NPs via electrostatic attraction as follows: a solution of positively charged silver ions, [Ag(NH₃)₂]⁺, was prepared from a stock solution of AgNO₃, whereas the negatively charged GO (O⁻) was made via adjustment of GO dispersion liquid to pH 10. Thus, [Ag(NH₃)₂]⁺ could be efficiently absorbed on the surface of GO nano sheets by electrostatic attraction during the 2 h stirring process at room temperature. Then a freshly prepared aqueous solution of ascorbic acid (0.5%, m/m) was added dropwise and kept for another 2 h in an ice-water bath to reduce the ionic [Ag(NH₃)₂]⁺ to Ag NPs tightly and evenly grafted onto the GO surface. After further reduction of graphene oxide to graphene for 8 h at 80 °C, the resultant Ag NPs could be uniformly distributed on the surface of the graphene sheets without agglomeration. The as-obtained dispersion was centrifuged after cooling down to room temperature, washed with doubly deionized water and dried overnight in a drying oven at 60 °C. The as-prepared sample was characterized by TEM, Raman spectroscopy and UV-vis spectroscopy as shown in Fig. 1.

Fig. 1 (a) TEM image; (b) Raman spectrum and (c) UV-vis spectrum of the G–Ag NPs.

From the TEM image (Fig. 1a), we could see that the Ag NPs were almost uniformly distributed on the surface of the graphene sheet with an average diameter of ca. 20 nm. The Raman spectrum (Fig. 1b) exhibited both D- and G-mode peaks at ca. 1348 and 1595 cm⁻¹, respectively, which were ascribed to disordered carbon (sp³) with dangling bonds in disordered plane terminations and the ordered vibration of sp²-hybridized carbon atoms in a 2D hexagonal lattice.²⁴ The Raman signals of the D bands increased with the formation and absorption of Ag NPs onto graphene, owing to the chemical interactions between the graphene matrix and the metal surface.²⁵ The D-mode peak was especially blue-shifted by 9 cm⁻¹ i.e., from 1348 cm⁻¹ to 1357 cm⁻¹, and was greatly strengthened compared to the G-mode peaks. This peak shifting provided experimental evidence that Ag NPs were immobilized on the surface of the graphene and that there was a strong synergistic effect between the Ag NPs and the large π-conjugated graphene matrix.²⁶ There were two obvious absorption peaks in the UV-vis spectrum, at ca. 280 nm and 420 nm (Fig. 1c). The absorption peak at 280 nm was attributed to π → π* electron transitions of the graphene matrix, which suggested that GO was successfully reduced and the electronic conjugation within the graphene sheets was thereby restored.²⁷ The other peak at 420 nm was attributed to a characteristic absorption peak of Ag NPs. All the above results confirmed that Ag NPs were effectively adsorbed onto the surface of graphene.

3.2 Special response to Hg²⁺

To investigate the response and selectivity of G–Ag NPs to heavy metals, we investigated the colorimetric response of G–Ag NPs to some common and physiologically or environmentally important alkali, alkaline earth, and transition metal ions, such as Ag⁺, Al³⁺, Ba²⁺, Cd²⁺, Co³⁺, Cu²⁺, Fe³⁺, K⁺, Mg²⁺, Na⁺, Ni²⁺, Pb²⁺, Sr²⁺ and Hg²⁺. As shown in Table 1, except Hg²⁺, addition of all the other abovementioned metal ions in 3.25 × 10⁻⁴ M concentration resulted in a negligible change in the absorption peaks of the proposed G–Ag NPs system at 420 nm. Moreover, all changes in this absorption intensity were less than 5% relative to that of Hg²⁺ at 1/100^th of the concentration, i.e., 3.25 × 10⁻⁶ M for Hg²⁺ versus 3.25 × 10⁻⁴ M for the other abovementioned metals. These results suggested that the common metal ions had almost no influence on the absorption spectrum of the G–Ag NPs system; however, the system exhibited an excellent and selective response to Hg²⁺.

Table 1 Absorption responses of various common metal ions^a

Metal ions	ΔA (%)	Metal ions	ΔA (%)	Metal ions	ΔA (%)
a 1.0 mL phosphate buffer (pH 5.0), 1.0 mL 10 μg mL^−1 G–Ag NPs aqueous solution and 1.0 mL metal ion solution (all 3.25 × 10^−4 M; except Hg^2+ 3.25 × 10^−6 M).
Ag^+	2.8	Al^3+	0.3	Co^3+	2.5
Cd^2+	1.2	Cu^2+	0.7	Fe^3+	2.6
Pb^2+	1.6	Ni^2+	0.6	Mg^2+	0.4
Ba^2+	0.2	Sr^2+	0.4	Hg^2+	100
K^+	0.1	Na^+	0.1

---

3.3 Optimization of recognition conditions for Hg²⁺

The sensing conditions using G–Ag NPs were optimized to ensure that they were practical for the intended application. It is well known that the redox potential of ascorbic acid and Hg²⁺ is sensitive to pH; thus, the effect of pH was investigated over a range of 3.0–9.0. Fig. 2 shows the change (ΔA) in the absorption intensity at 420 nm before and after Hg²⁺ was added to the proposed G–Ag NPs sensing system at different pH levels (ΔA = A₀ − A, where A₀ and A are the absorption intensities of the systems in the absence and presence of Hg²⁺, respectively). Clearly, the pH of the solution played an important role in the interaction between G–Ag NPs and Hg²⁺. ΔA increased gradually as the pH value decreased, reaching its maximum at pH ca. 5.5. This may be attributed to the redox potential of Hg²⁺, which is lower at high pH; moreover, the higher concentration of OH⁻ lowers the solubility of Hg²⁺ in aqueous solution. Both low redox potential and low solubility of Hg²⁺ would result in less Hg²⁺ being reduced at high pH, and, consequently, less amalgam would be formed between mercury and G–Ag NPs. Based on this result, we selected a pH value of 5.0 for all the subsequent tests.

Fig. 2 Effect of pH on the absorption intensity (ΔA) of the proposed system at 420 nm.

We also checked and confirmed that the absorption intensity of the proposed system at 420 nm was not influenced by the system's ionic strength over a NaCl concentration range from 1.0 × 10⁻² to 1.0 × 10⁻⁷ mol L⁻¹. Thus, the stability of the present system was confirmed for application in various salty environments.

3.4 Analytical parameters and samples detection

The response of the proposed G–Ag NPs to Hg²⁺ was further investigated in detail, i.e., the absorption spectra were obtained at different Hg²⁺ concentrations between 0–3.25 × 10⁻⁸ mol L⁻¹, as shown in Fig. 3. We found that the absorption intensity at 420 nm decreased gradually with an increase in Hg²⁺ concentration (Fig. 3a), presenting a good linear relationship between ΔA and Hg²⁺ concentration over the range from 2.0 × 10⁻⁷ to 3.25 × 10⁻⁶ mol L⁻¹ with a correlation coefficient of 0.9758 (Fig. 3b). The regression equation was ΔA = 3.16 × 10⁻³ + 8.51 × 10⁻⁴c (10⁻⁸ mol L⁻¹). Based on the definition of the detection limit (LOD), three times the average deviation of absorbance at 420 nm in 20 blank samples without Hg²⁺ was used in this study. The LOD for Hg²⁺ response was up to 3.1 × 10⁻⁸ mol L⁻¹, which is the an acceptable standard for detecting Hg²⁺ in drinking water for human beings, according to the World Health Organization (limit: 3.0 × 10⁻⁸ mol L⁻¹).

Fig. 3 (a) Absorption spectra of the G–Ag NP composite with different Hg²⁺ concentrations (from top to bottom: 0, 20, 50, 100, 150, 200, 250, 300, 325 × 10⁻⁸ mol L⁻¹); (b) a linear relationship between the ΔA of G–Ag NPs at 420 nm and Hg²⁺ concentrations.

To further evaluate the feasibility of the present method, it has been applied for the determination of Hg²⁺ concentration in 3 environmental water samples from the Pi River, underground water and campus tap water (Table 2).

Table 2 Results for environmental water samples (n = 5)^a

Samples	cHg^2+ in sample^b (nM)	Spiked (nM)	Found (nM)	Recovery (%)	R.S.D. (%)
a Phosphate buffer, pH 5.0.b The environmental water Hg^2+ concentration determined using G–Ag NPs with the proposed method. The real values are the table values ×10^−2 nmol L^−1, since the water samples were concentrated 100 times.
1 (the Pi River)	412.8	500.0	906.1	98.3	2.7
2 (underground water)	868.2	500.0	1380.5	100.9	1.8
3 (tap water)	0.00	500.0	516.1	103.2	4.6

---

All the samples were filtered several times and concentrated 100 times via vacuum distillation. For recovery studies, a known concentration of Hg²⁺ was added to the environmental water samples and the total Hg²⁺ concentrations were determined. The recoveries of different amounts of Hg²⁺ spikes were 98.3% to 103.2% with a satisfying analytical precision (R.S.D. ≤ 4.6%), which further validated the reliability and practicality of the method.

3.5 Cytotoxicity of G–Ag NPs and its separation effect to Hg²⁺

The proposed sensing system was evaluated for toxicity and environmental effects, beginning with the cytotoxicity. A mouse fibroblast cell line 3T3 was incubated in the presence of increasing amounts of G–Ag NPs and Ag NPs for 48 hours, respectively. The cytotoxicities were calculated by normalizing the MTT readings of the corresponding nanoparticle-treated groups against those of the untreated control culture for the same amount of time, as shown in Fig. 4. The results showed that G–Ag NPs possessed lower toxicity than the corresponding Ag NPs. Approximately 87% of cells in each cell line survived, even at the highest G–Ag NP concentration tested (10.0 μg mL⁻¹). However, less than 50% cells survived in the presence of Ag NPs under the same incubation condition. This result suggested that the cytotoxicity of G–Ag NPs was negligible compared to that of its precursor Ag NPs.

Fig. 4 The cytotoxic effect of G–Au NPs on a series of mouse fibroblast cell line 3T3.

To evaluate the secondary pollution effects of G–Ag NPs after the Hg²⁺ sensing procedure, 1.0 mL of aqueous ascorbic acid (0.5%, m/m), 1.0 mL phosphate buffer (pH 5.0), and 1.0 mL G–Ag NPs or Ag NPs dispersion (10.0 μg mL⁻¹) were added dropwise into a 1.0 mL aqueous solution of Hg²⁺ (1.0 × 10⁻⁵ mol L⁻¹). After the mixture was diluted to 10 mL with doubly deionized water and was stirred for 30 min, it was simply filtered using 0.45 μm microporous membrane. The Hg²⁺ content that remained in the filtrate was measured using ICP-MS. Less than 2% Hg²⁺ (1.9 × 10⁻⁸ mol L⁻¹) remained in the filtrate of the G–Ag NPs system, whereas ca. 11% (10.8 × 10⁻⁸ mol L⁻¹) remained in the Ag NPs system. This suggested that G–Ag NPs system offered a more efficient enrichment and separation effect for Hg²⁺ pollution. After Hg²⁺ determination, both the pollutant and the sensing materials can be easily separated and extracted from the environment by a simple filtration, which will efficiently reduce the possible secondary pollution.

4. Conclusions

In conclusion, a kind of graphene-based nano composite (G–Ag NPs) was designed and prepared by a simple one-pot redox reaction using ascorbic acid as the reducing agent. By virtue of their high affinity and super-high specific surface area, the resultant G–Ag NPs possessed an excellent and selective response to Hg²⁺ in the linear range of 2.0 × 10⁻⁷ to 3.25 × 10⁻⁶ M (R² = 0.9758) with a detection limit (3σ, n = 20) of 3.1 × 10⁻⁸ M. The G–Ag NPs were low toxicity and most of the detected Hg²⁺ could be efficiently separated and extracted from the environment by a simple filtration using a 0.45 μm microporous membrane with little secondary pollution. This work will provide a generic and effective strategy to construct environmentally-friendly sensing materials for analyzing and separating heavy metal ions with little secondary pollution.

Acknowledgements

The authors gratefully acknowledge the financial supports from National Natural Science Foundation of China (No. 21277103) and Anhui Provincial Natural Science Foundation (No. 1508085QB35) and Shandong Provincial Natural Science Foundation (No. ZR2016BQ13).

Notes and references

1. S. K. Sharma, N. Kaur, J. Singh, A. Singh, P. Raj, S. Sankar, D. Y. Kim, N. Singh and H. Singh, Sens. Actuators, B, 2016, 232, 712–721.
2. Z. Q. Yan, L. Hu and J. M. You, Anal. Methods, 2016, 8, 5738–5754.
3. P. K. Khanna, S. Bhanoth, V. Dhanwe, A. Kshirsagar and P. More, RSC Adv., 2015, 5, 92818–92828.
4. Z. Q. Yan, M.-F. Yuen, L. Hu, P. Sun and C.-S. Lee, RSC Adv., 2014, 4, 48373–48388.
5. Y. L. Hu, D. M. Wang and G. K. Li, Anal. Methods, 2015, 7, 6103–6108.
6. S. P. Li, J. P. Lai, L. M. Qi, M. Saqib, S. Majeed, Y. J. Tong and G. B. Xu, Analyst, 2016, 141, 2362–2366.
7. P. She, Y. X. Chu, C. W. Liu, X. Guo, K. Zhao, J. G. Li, H. J. Du, X. Zhang, H. Wang and A. P. Deng, Anal. Chim. Acta, 2016, 906, 139–147.
8. L. Hu, Y. F. Zhang, L. Nie, C. G. Xie and Z. Q. Yan, Spectrochim. Acta, Part A, 2013, 104, 87–91.
9. Z. Q. Yan, L. Hu, L. Nie and H. Lv, Spectrochim. Acta, Part A, 2011, 79, 661–665.
10. M. Li, Y. Sun, L. Dong, Q. C. Feng, H. Xu, S. Q. Zang and T. C. W. Mak, Sens. Actuators, B, 2016, 226, 332–341.
11. Z. G. Shen, Y. Q. Luo, Q. Wang, X. Y. Wang and R. C. Sun, ACS Appl. Mater. Interfaces, 2014, 6, 16147–16155.
12. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
13. C. Lu, P. J. J. Huang, Y. B. Ying and J. W. Liu, Biosens. Bioelectron., 2016, 79, 244–250.
14. S. Q. Xiong, B. Y. Yang, D. Q. Cai, G. N. Qiu and Z. Y. Wu, Electrochim. Acta, 2016, 188, 928.
15. D. Dinda, B. K. Shaw and S. K. Saha, ACS Appl. Mater. Interfaces, 2015, 7, 14743–14749.
16. V. Chandra and K. S. Kim, Chem. Commun., 2011, 47, 3942–3944.
17. C. Y. Lin, C. J. Yu, Y. H. Lin and W. L. Tseng, Anal. Chem., 2010, 82, 6830–6837.
18. A. P. Fan, Y. Ling, C. W. Lau and J. Z. Lu, Talanta, 2010, 82, 687–692.
19. G. Q. Wang, Z. P. Chen, W. H. Wang, B. Yan and L. X. Chen, Analyst, 2011, 136, 174–178.
20. X. Zhang, Q. Liang, Q. Han, W. Wan and M. Ding, Analyst, 2016, 141, 4219–4226.
21. Q. Zeng, J. W. Y. Lam, C. K. W. Jim, A. J. Qin, J. G. Qin, Z. Li and B. Z. Tang, J. Polym. Sci., Polym. Chem., 2008, 46, 8070–8080.
22. Z. Q. Yan, S. Y. Guang, H. Y. Xu and X. Y. Liu, Dyes Pigm., 2013, 99, 720–726.
23. X. H. Cheng, S. Li, H. Z. Jia, A. S. Zhong, C. Zhong, J. Feng, J. G. Qin and Z. Li, Chem.–Eur. J., 2012, 18, 1691–1699.
24. Z. Q. Yan, H. T. Xue, K. Berning, Y.-W. Lam and C.-S. Lee, ACS Appl. Mater. Interfaces, 2014, 6, 22761–22768.
25. Z. X. Xu, H. Y. Gao and H. Guoxin, Carbon, 2011, 49, 4731–4738.
26. M. H. Jin, T. H. Kim, S. C. Lim, D. L. Duong, H. J. Shin, Y. W. Jo, H. K. Jeong, J. Chang, S. S. Xie and Y. H. Lee, Adv. Funct. Mater., 2011, 21, 3496–3501.
27. Y. J. Han, Z. M. Luo, L. H. Yuwen, J. Tian, X. R. Zhu and L. H. Wang, Appl. Surf. Sci., 2013, 266, 188–193.

---