Gold nanozyme-based paper chip for colorimetric detection of mercury ions

In this study, we developed a facile gold nanozyme-based paper chip (AuNZ-PAD) for Hg2+ detection. This device has the advantages of being simple, rapid, cost effective, sensitive, selective, high throughput, and applicable to onsite detection. The colorimetric mercury assay on the AuNZ-PAD is established based on the enzyme-like catalytic activity of gold nanoparticles promoted by the formation of Au–Hg amalgam, which is correlated to the intensity of the colorimetric response resulting from the catalytic reaction of 3,3′,5,5′-tetramethylbenzidine (TMB) and H2O2. Highly sensitive and selective detection of Hg2+ ions is achieved in both distilled and tap water samples, indicating the feasibility and applicability of our device for the determination of mercury pollution in real samples. Moreover, AuNZ-PAD analysis using a smartphone camera eliminates the need for expensive analytical equipment, thereby increasing the practicality of field monitoring of trace Hg2+ compared with other sensing methods.

printing method to create hydrophilic/hydrophobic patterns for fluid transport via capillary action, and are commonly used for colorimetric analysis owing to the high color contrast on paper substrates 21 . The assay results can be directly interpreted by the naked eye and quantified using a mobile camera instead of bulky instruments, making PADs ideal for onsite monitoring applications 22,23 .
Herein, we present a gold nanozyme-based paper chip (AuNZ-PAD) for simple, rapid, sensitive, selective, and cost-effective detection of trace Hg 2+ in aqueous systems. Gold nanoparticles (AuNPs) were employed as a sensitive and selective probe for colorimetric assays based on their enzyme-like catalytic activity, which could be enhanced remarkably by the formation of Au-Hg amalgam (alloy). The chromogenic enzyme substrates were catalytically oxidized on Au-Hg amalgam in the presence of H 2 O 2 , resulting in a blue stain on the paper chip. The mercury-promoted nanozyme activity of AuNPs was systemically investigated under various conditions (AuNP size, H 2 O 2 concentration, and pH). The analytical performance of the AuNZ-PAD was demonstrated in both distilled and tap water samples. Moreover, coupling this paper chip with a smartphone camera eliminated the need for expensive, sophisticated equipment for readouts.

Results
It is known that several nanomaterials (e.g., gold, platinum, graphene oxide, Fe 3 O 4 , and CeO 2 ) possess intrinsic peroxidase-like catalytic activity [24][25][26][27][28][29][30][31][32] . Among these materials, citrate-stabilized AuNPs exhibit relatively low catalytic activity 33 . However, it was recently reported that the nanozyme activity of AuNPs can be enhanced considerably by the formation of Au-Hg amalgam in aqueous solution 34 . By taking advantage of the mercury-promoted nanozyme activity of AuNPs, colorimetric mercury detection was established, as depicted in Fig. 1. When Hg 2+ ions are introduced on the AuNZ-PAD, they are spontaneously deposited on the surface of AuNPs, resulting in the formation of Au-Hg amalgam by specific metallophilic interactions 35,36 . The presence of Hg atoms around AuNPs causes changes in the surface properties of AuNPs, which accelerate decomposition of H 2 O 2 , leading to considerable improvement of the peroxidase-like catalytic activity of AuNPs 34,37 . As a result, an intense blue stain is produced on the paper chip by the catalytic reaction of TMB (chromogenic peroxidase substrate) with H 2 O 2 . In contrast, the TMB-H 2 O 2 catalytic reaction hardly occurs on bare AuNPs owing to their low catalytic activity, and no color change is observed on the paper chip.
To demonstrate the mercury-promoted nanozyme activity of AuNPs, the TMB-H 2 O 2 catalytic reaction was investigated using UV-vis spectroscopy. As shown in Supplementary Figure S1, negligible color change was observed in the absence and presence of AuNPs (curves a and b), indicating the poor nanozyme activity of bare AuNPs. However, when AuNPs and Hg 2+ ions coexisted in solution, the TMB-H 2 O 2 catalytic reaction proceeded to produce oxidized TMB, which has an intense blue color in solution, with a maximum absorbance at 652 nm (curve c). This result reveals that the TMB-H 2 O 2 catalytic reaction is accelerated in the presence of both AuNPs and Hg 2+ ions, suggesting that the nanozyme activity of AuNPs can be enhanced when Au-Hg amalgam is formed by the addition of Hg 2+ ions. In addition, the Au-Hg amalgam was characterized by X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM). The Au 4 f XPS spectrum shows two peaks at 84.1 and 87.8 eV corresponding to Au 4f 7/2 and Au 4f 5/2 , respectively which were assigned to metallic Au 0 (Supplementary Figure S2A) 38 . In Supplementary Figure S2B, the two peaks observed in the Hg 4 f XPS spectrum at 99.8 eV (Hg 4f 7/2 ) and 103.8 eV (Hg 4f 5/2 ) correspond to the Hg 0 oxidation state. These results reveal that metallic mercury was indeed formed on the surface of the AuNPs 39 . Furthermore, TEM images of the AuNPs in Supplementary Figure S3 show no obvious difference in the particle size of the AuNPs before and after addition of Hg 2+ ions, indicating that only a monolayer or submonolayer of Hg 0 was formed on the surfaces of the AuNPs.
To study the effect of experimental parameters (AuNP size, H 2 O 2 concentration, and pH) on the mercury-promoted nanozyme activity of AuNPs, we systemically investigated the colorimetric response on paper in the absence and presence of Hg 2+ ions under various conditions. Figure   520 nm) were drop-dried on each test zone with a diameter of 10 mm. Experiments were then carried out by adding 1 μL of the test sample, followed by the addition of 2 μL of TMB substrate containing 0.5% H 2 O 2 . The AuNP size was varied from 5 to 40 nm, and the size dependence of the mercury-promoted nanozyme activity of AuNPs was evaluated by comparing the colorimetric intensity ratio, ΔI/I 0 = (I − I 0 )/I 0 , where I and I 0 denote the colorimetric intensity values obtained for 10 ng Hg and a blank solution, respectively. As shown in Fig. 2A, 20 nm AuNPs showed the highest sensitivity to Hg 2+ , with a significant difference in color observed, whereas AuNPs with sizes below 10 nm exhibited a blue stain irrespective of the addition of Hg 2+ owing to strong intrinsic catalytic activities induced by the high surface-to-volume ratio of small particles. These results demonstrated that the AuNP size has an important impact on mercury-promoted nanozyme activity, and 20 nm AuNPs are the most appropriate for Hg 2+ detection. Figure 2B and C show the effect of H 2 O 2 concentration and pH on the colorimetric response, respectively. For these experiments, 20 nm AuNPs (OD of 1/6, 0.194 nM) were applied as the optimum condition. As shown in Fig. 2B, the colorimetric signal ratio gradually increased upon increasing the H 2 O 2 concentration up to 3.0%, as oxidation of TMB is established in the presence of H 2 O 2 . However, further increasing the concentration decreased the signal ratio, indicating that the optimal H 2 O 2 concentration for the TMB-H 2 O 2 catalytic reaction on AuNZ-PAD is 3.0%. Moreover, to examine the pH dependence of the catalytic activity, the Hg 2+ samples were diluted with 10 mM acetate buffer in the pH range of 4.0-10.0, followed by application of the TMB substrate containing 3.0% H 2 O 2 . The maximum mercury-promoted nanozyme activity of AuNPs occurred at approximately pH 6.0 (Fig. 2C). The colorimetric signal ratio increased with increasing pH from 4.0 to 6.0, and then decreased drastically when the pH value exceeded 7.0. At high pH conditions, mercury oxide/hydroxide species (HgO and Hg(OH) 2 ) could be formed through interactions with OH − ions, which would prevent the formation of Au-Hg amalgam, resulting in reduction of the nanozyme activity of AuNPs for the TMB-H 2 O 2 catalytic reaction 40,41 . In addition, the optimum pH for peroxidase-catalyzed oxidation of TMB substrate is typically between 4.0 and 7.0 42 . Therefore, based on these results, further tests on paper chips were performed under the optimized conditions (20 nm AuNPs, 3.0% H 2 O 2 , and pH 6.0).
The design of the AuNZ-PAD is shown in Fig. 3A. The AuNZ-PAD consists of three zones: reagent loading, detection, and absorbent. To perform the colorimetric assay on the AuNZ-PAD, Hg 2+ samples diluted with 10 mM acetate buffer (pH 6.0) were applied on the detection zone, which includes six test zones where 20 nm AuNPs were previously integrated, allowing simultaneous analysis of multiple samples. Subsequently, TMB substrate containing 3.0% H 2 O 2 was added on the reagent loading zone. The TMB substrate then spread to the six test zones by capillary action until reaching the arch-shaped absorbent zones. After 10 min, the colorimetric assay results could be readily observed by the naked eye. Figure 3B shows a photographic image of the AuNZ-PAD taken using a smartphone camera. Six samples with different mercury level were simultaneously tested on a single paper chip. As the mercury level increased, the color of the test zone changed from colorless to dark blue, which indicated that the intensity of color corresponds to the amount of mercury in the sample. The calibration plot in Fig. 3C shows a good linear relationship between the colorimetric responses and the logarithmic values of the mercury levels in the range between 0.1 and 200 ng. The colorimetric signal intensity was saturated above 500 ng Hg, which may be due to the limited amount of AuNPs available for forming Au-Hg amalgam. These results demonstrated that the AuNZ-PAD provides a large linear dynamic range (about three orders of magnitude) for Hg 2+ detection, which is an important factor for quantitative analysis in practical applications. Figure 4A shows an interesting feature of the AuNZ-PAD for colorimetric signal amplification. In this experiment, a 2 μL drop of test sample containing 0.4 ng Hg was repetitively applied on each test zone of the AuNZ-PAD with a time interval of 2 min. As the number of drops of test sample increased, a conspicuous color change was observed, as shown in Fig. 4A. Moreover, the colorimetric signal intensity was found to be proportional to the number of drops (Supplementary Figure S4). These results revealed that multiple applications of test sample allow for in situ preconcentration of trace Hg 2+ on the AuNZ-PAD, thereby leading to an enhancement of the colorimetric response. Therefore, it is reasonable to expect that the detection sensitivity and dynamic range of this device can be easily improved by repeated additions of test sample. Considering the preconcentration factor, the linear regression equation can be expressed as I = 10049 log[n Hg(ng)] + 12780, where n corresponds to the number of drops of test sample, with a correlation coefficient of 0.986. The limit of detection (LOD, 3 SD) was estimated to be 0.06 ng (30 μg L −1 ) for a single drop of test sample, and could be reduced to 0.012 ng (1.2 μg L −1 ) by using five applications of test sample. Moreover, if needed, the LOD can be further decreased by simply increasing the number of drops of test sample. Thus, our device offers an LOD that is low enough for drinking water analysis and has a wider dynamic range (>10-fold) than other paper-based sensors, as shown in Supplementary Table S1. Although ICP-MS analysis provides an excellent LOD (0.001 μg L −1 ) and a broad dynamic range (8-9 orders of magnitude) compared with those of paper-based sensors, it requires expensive instrumentation and highly trained operators, and suffers from severe memory effects owing to adsorption of mercury. Therefore, our paper chip is potentially more applicable for on-site mercury analysis because of its low cost, rapid response, portability, and low labor requirement.  To demonstrate the selectivity of the AuNZ-PAD toward Hg 2+ , 15 other metal/metalloid ions (Ag + , Fe 2+ , Fe 3+ , Ni 2+ , Co 2+ , Cd 2+ , As 3+ , Pb 2+ , Bi 3+ , Zn 2+ , Sb 3+ , Cu 2+ , Mn 2+ , Ca 2+ , and Na + ) were tested. Figure 4B and Supplementary Figure S5 show the colorimetric responses against metal/metalloid species (100 ng) and mercury (1 ng). Although the amounts of the interfering species were 100-times greater than that of mercury, only Hg 2+ ions caused a conspicuous color change of the test zone and negligible colorimetric responses were observed for each of the interfering species (less than 8% of that of Hg 2+ ; Fig. 4B). Thus, our device showed excellent selectivity toward Hg 2+ over other interfering species owing to the amalgam process that occurs specifically between Au and Hg atoms.
To further demonstrate the applicability of the AuNZ-PAD in practical applications, different concentrations of Hg 2+ (0.05, 0.5, 5, and 50 mg L −1 ) were spiked in tap water (1:1 mixed with 10 mM acetate buffer, pH 6.0) and tested on our device. Tap water was collected from the laboratory in our institute. As summarized in Table 1, the obtained recoveries were in the acceptable range of 84.4-112.9% with a relative standard deviation (RSD) of 3.8-8.7% for four different Hg 2+ spiked levels in tap water. The results clearly demonstrate that the AuNZ-PAD is applicable for practical analysis of Hg 2+ in real samples.

Discussion
We report a new paper-based mercury sensor based on the nanozyme activity of AuNPs. This device provides a simple, portable, and cost-effective approach for the determination of Hg 2+ without requiring complicated and time-consuming surface modification with Hg 2+ -specific ligands and bulky and expensive instruments for readouts. Moreover, multiple tests can be simultaneously performed on a single device and readily analyzed using a smartphone camera. The AuNZ-PAD showed high sensitivity with a wide detection dynamic range (about three orders of magnitude) and selective specificity toward Hg 2+ without interference from other metal/metalloid species. The LOD for Hg 2+ was found to be 0.06 ng and 0.012 ng for a single application and five applications of test sample, respectively, and could be further improved by repeated additions of test sample. These results suggest that the proposed AuNZ-PAD is capable of facilitating onsite monitoring of mercury pollution in aqueous environments for the prevention of mercury poisoning.

Fabrication of AuNZ-PAD.
Paper-based devices were fabricated using a commercially available wax printer (Color Cube 8870, Xerox, Norwalk, CT, USA) as previously reported 43 . Hydrophobic patterns, which were designed using graphics software (Adobe Illustrator CS3), were printed on chromatography paper to define the fluidic path and control liquid transport. The paper was heated at 100 °C for 1 min in a dry oven to melt the patterned wax. The patterned paper (6.0 × 6.0 cm) was cut, and the backside of each piece of paper was fully covered by an adhesive PET film. To integrate AuNPs onto the paper chip, 1.5 μL of 20 nm AuNP solution (0.194 nM) was drop-dried on each test zone (with a diameter of 1.0 cm) at ambient temperature.
Assay procedure and signal quantification. A 2 μL volume of sample was applied to each test zone of the paper chip. Then, 100 μL of 0.05% TMB solution containing 3.0% H 2 O 2 was added to the reagent loading zone. This solution then flowed along the fluidic path until it reached the arch-shaped absorbent zones. The assay results were directly visible to the naked eye, amplifying over time. To quantify the results, after 10 min incubation, photographs were taken using a smartphone camera (LG-G5, 16.0 megapixels) with a shutter speed of 1/20 at ISO 50. Subsequently, colorimetric signal analysis was carried out using ImageJ software (National Institutes of Health, Bethesda, MD, USA).