Smartphone-Based Colorimetric Detection of Chromium (VI) by Maleic Acid-Functionalized Gold Nanoparticles

Because of the significant environmental pollution produced by human activities, there is an ongoing need to develop transportable, simple, and reliable techniques for determining trace contaminants on the spot. This work reported a colorimetric detection method for aqueous Cr(VI) sensing by maleic acid-functionalized gold nanoparticles with high sensitivity and selectivity. The wine-red color of the probe solution can change to gray even in the presence of 1.0 μg L−1 of aqueous Cr(VI). Moreover, with the assistance of a smartphone installed with a commercially available color scan application software, its concentration of could be readily quantified on the spot without the help of UV-Vis spectrometer. The detection limit could reach as low as 0.1 μg L−1 with linear range from 0.2 to 2.0 μg L−1. Most importantly, the coefficient variation of the proposed smartphone-based method was equivalent to that of colorimetry, demonstrating the high accuracy of the proposed method for accurate detection of Cr(VI) in resource-constrained countries. Conclusively, with the help of the smartphone, this nanomaterials-based probe demonstrated the potential in the field of environment monitoring for on-site quantitative detection of any pollutants in resource-constrained countries.

The smartphone, very portable and capable of high-resolution imaging, is a natural candidate for a powerful and convenient point-of-care analytical device. Specifically, smartphone-based colorimetric methods for environmental monitoring of heavy metal ions have been reported [2,11,[43][44][45][46][47][48][49][50][51][52][53]. Herein, we describe an Au NPs-based colorimetric method for the detection of aqueous Cr(VI). Our assay involves an aggregation-induced color change of functionalized Au NPs in the presence of aqueous Cr(VI) ( Figure 1). As we will show, with a color scan application, the smartphone can easily uncover nuances of color in solutions that the naked eye can not observe. Like the UV-Vis spectrophotometer a smartphone can quantify Cr(VI) in aqueous solution. Its detection limit can be as low as 0.1 µg L −1 .
Appl. Sci. 2021, 11, x FOR PEER REVIEW 2 of 10 resonance (SPR) [7,35] offer improved performance compared to traditional methods. Some of them have already been commercialized for practical application. For example, a portable device based on the SERS technique has been developed for on-site detection of pesticides on the surface of fruits with high selectivity and sensitivity [33]. Au NPs with rationally modified surfaces can be the base for colorimetric methods in which elaborate instrumentation is replaced by the naked eye [7,20,[24][25][26][27][28][29][30][31][32]. Furthermore, due to their high extinction coefficient in the visible region, the Au NPs-based colorimetric method can be used to detect heavy metal ions selectively and sensitively [7,20,25,32,[36][37][38][39][40][41][42]. For example, the Mirkin group has reported a highly selective probe for Hg 2+ sensing, based on the strong coordination interaction between Hg 2+ and thymine-rich DNA-functionalized Au NPs. [25] This method has a sensitivity of 20 ppb [25]. Recently, the Gao group developed a Pb 2+ selective detection method using glutathione-modified Au NPs with a detection limit of about 10 nM [38]. However, a UV-vis spectrometer is needed for quantitative colorimetric assays. The associated expense hampers practical use of the Au NPs-based visual detection method. The smartphone, very portable and capable of high-resolution imaging, is a natural candidate for a powerful and convenient point-of-care analytical device. Specifically, smartphone-based colorimetric methods for environmental monitoring of heavy metal ions have been reported [2,11,[43][44][45][46][47][48][49][50][51][52][53]. Herein, we describe an Au NPs-based colorimetric method for the detection of aqueous Cr(VI). Our assay involves an aggregation-induced color change of functionalized Au NPs in the presence of aqueous Cr(VI) ( Figure 1). As we will show, with a color scan application, the smartphone can easily uncover nuances of color in solutions that the naked eye can not observe. Like the UV-vis spectrophotometer a smartphone can quantify Cr(VI) in aqueous solution. Its detection limit can be as low as 0.1 g L −1 .

Preparation and Functionalization of Au NPs
Au NPs were synthesized according to the classic citrate-mediated reduction method [54]. Briefly, 50 mL of water containing 1 mM of HAuCl 4 was heated to a boil under vigorous stirring, and 2 mL of 1% sodium citrate solution was then injected into boiling solution. After another 20 min of reaction, the boiling solution was allowed to cool down to room temperature under stirring. Finally, the obtained wine-red Au NPs colloid solution was stored at 4 • C for further use. For the functionalization, 1 mL of citrate-capped Au NPs was mixed with C maleic acid (MA) under stirring, and the mixture was allowed to stir for 2 h at room temperature under rotation. The MA-functionalized Au NPs were then purified by centrifugation and the purified and functionalized Au NPs were then dissolved in 1 mL of ultrapure water for storage as the probe solution. For storage more than 2 months, the probe should be dispersed in 1 mL of DI water containing 0.1 mM maleic acid at 4 • C. The concentration of obtained MA-capped Au NPs was determined by its absorption at 525 nm with an extinction coefficient of 2.7 × 10 8 L·mol −1 ·cm −1 .

Detection of Aqueous Cr(VI) by MA-Capped Au NPs
Briefly, 10 µL samples of Cr(VI) or other metal ion solutions including Ba 2+ , Ca 2+ , Cd 2+ , Co 2+ , Cu 2+ , Cr(III), Mg 2+ , Mn 2+ , Ni 2+ , Pb 2+ , Zn 2+ , and Al 3+ stock solution prepared in water were added into 1 mL of MA-capped Au NPs and mixed thoroughly for 2 min by a pipette. A UV-Vis spectrophotometer was then used to collect the absorption spectrum of the solution. A smartphone equipped with color scan application software (ColorAssist Ver. 2.4, 2020 from Apple Store) was used to take the MA-capped Au NPs solution photo at room temperature. Typically, for photograph acquiring, the vial should be placed on a white paper together with white background, and glare should be avoided by changing the photographing direction. The camera should be parallel to the vial, and the distance can be adjusted according to the square pattern displayed by the app software; one need not adjust amount. The square pattern should just cover the whole transverse part of the vial. For real sample detection by the smartphone-based colorimetric method, aqueous samples collected from three local lakes (South Lake, Tangxun Lake, and Yezi Lake) were first filtered by a 0.22 µm Super filter. Atomic Absorption Spectroscopy method was then used to measure the absolute concentration of Cr(VI) in the three samples. As for smartphone sensing, 10 µL of the sample was added to 1 mL of MA-capped Au NPs, and the mixture was allowed to react for 2 min before photographing. Each measurement was done three times, and the average was presented with its standard deviation.

Characterization
The aggregation response of Au NPs after interaction with Cr(VI) was obtained by transmission electron microscopy (TEM, JEOL 2010, Tokyo, Japan) at an accelerating voltage of 200 kV. The absorption spectra of the colloid before and after interaction with metal ions were collected using a Shimadzu UV-2550 spectrometer (Shimadzu, Kyoto, Japan). A Canon-350D digital camera (Canon, Tokyo, Japan) takes optical photographs.

Selectivity and Sensitivity Investigation of MA-Capped Au NPs to Aqueous Cr(VI)
The molecules used for Au NPs modification to effect colorimetric assays for heavy metal ions [7,32,[55][56][57] all have carboxylic acid groups. Examples include 5,5-dithiobis(2nitrobenzoic acid), [55] 1,4-dithiothreitol [56] and 5-thio-(2-nitrobenzoic acid) [57] can be used as the capping reagent for Cr(III) or Cr(VI) sensing. We used maleic acid (MA), another organic acid containing two carboxylic groups, was then used as the capping reagent for the functionalization of Au NPs. By incubating MA with citrate-prepared Au NPs for 2 h under rotation, MA-capped Au NPs could be obtained; MA at high concentration displaces the citrate. We compared the absorption spectra and aggregation response for Au NPs before and after MA functionalization, in the presence of a range of metal ions. See Figure 2A-C. While the citrate-bearing Au NPs solution showed poor selectivity to the metal ions, MA-treated Au NPs exhibited a highly selective response to Cr(VI) over other metal ions tested. As can be seen in Figure 2C, Ba 2+ , Ca 2+ , Cd 2+ , Co 2+ , Cu 2+ , Cr(III), Mg 2+ , Mn 2+ , Ni 2+ , Pb 2+ , Zn 2+ , and Al 3+ can not cause a significant effect on the color of MA-capped Au NPs solution, while an obvious color change could occur in the presence of Cr(VI) at a concentration of 1.0 µg L −1 . We further recorded the UV-Vis spectra response of MA-capped Au NPs in the presence of different metal ions, which could further display the excellent selectivity of our probe for aqueous Cr(VI) (Figure 2D,E). Our probe selectivity was found to be working well as well as the methods described by Kim et al., Xue et al., and Mao et al. [6,7,10]. Moreover, TEM images of MA-capped Au NPs in the presence and absence of Cr(VI) further confirmed their aggregation phenomenon induced by Cr(VI) (Figure 2F,G).
See Figure 2A-C. While the citrate-bearing Au NPs solution showed poor selectivit the metal ions, MA-treated Au NPs exhibited a highly selective response to Cr(VI) other metal ions tested. As can be seen in Figure 2C, Ba 2+ , Ca 2+ , Cd 2+ , Co 2+ , Cu 2+ , Cr Mg 2+ , Mn 2+ , Ni 2+ , Pb 2+ , Zn 2+ , and Al 3+ can not cause a significant effect on the color of capped Au NPs solution, while an obvious color change could occur in the presenc Cr(VI) at a concentration of 1.0 µ g L −1 . We further recorded the UV-vis spectra resp of MA-capped Au NPs in the presence of different metal ions, which could further dis the excellent selectivity of our probe for aqueous Cr(VI) (Figure 2D,E). Our probe s tivity was found to be working well as well as the methods described by Kim et al.,et al.,and Mao et al. [6,7,10]. Moreover, TEM images of MA-capped Au NPs in the p ence and absence of Cr(VI) further confirmed their aggregation phenomenon induce Cr(VI) (Figure 2F,G). The selectivity of our probe to Cr(VI) could be ascribed to the hypothesis that Cr existing in the form of Cr2O7 2− , could coordinate with the carboxyl group from maleic through hydrogen bond in an energetically favorable manner. In addition, The selectivity of our probe to Cr(VI) could be ascribed to the hypothesis that Cr(VI), existing in the form of Cr 2 O 7 2− , could coordinate with the carboxyl group from maleic acid through hydrogen bond in an energetically favorable manner. In addition, this discrimination between Cr(VI) and Cr(III) could be ascribed to surface oxidation reaction between maleic acid and Cr 2 O 7 2− because of the strong oxidization ability of Cr 2 O 7 2− . Figure 3A,B show the evolution of color and absorbance spectra after adding 10 µL of different concentrations of aqueous Cr(VI) into MA-capped Au NPs. A change in color already discernable by the naked eye is evident already at a concentration as low as 0.4 µg L −1 by the naked eye. According to Figure 3B, the absorption intensity at 520 nm declines and the intensity at 650 nm grows as the concentration of Cr(VI) increases. Figure 3C shows plots of the absorption intensity ratio A650/A520 against the concentration of aqueous Cr(VI). Below about 2 µg L −1 , the limit beyond which the response reaches saturation, the plot is linear, described by the regression line (y = 0.40089 + 0.31557x) with a 0.94045 correlation coefficient (inset). discrimination between Cr(VI) and Cr(III) could be ascribed to surface oxidation reaction between maleic acid and Cr2O7 2− because of the strong oxidization ability of Cr2O7 2− . Figure  3A,B show the evolution of color and absorbance spectra after adding 10 µ L of different concentrations of aqueous Cr(VI) into MA-capped Au NPs. A change in color already discernable by the naked eye is evident already at a concentration as low as 0.4 µ g L −1 by the naked eye. According to Figure 3B, the absorption intensity at 520 nm declines and the intensity at 650 nm grows as the concentration of Cr(VI) increases. Figure 3C shows plots of the absorption intensity ratio A650/A520 against the concentration of aqueous Cr(VI). Below about 2 g L −1 , the limit beyond which the response reaches saturation, the plot is linear, described by the regression line (y = 0.40089 + 0.31557x) with a 0.94045 correlation coefficient (inset).

Smartphone-Based Colorimetric Assay
The expensive, delicate, and costly UV-vis spectrometer is not appropriate for on-site application. The portable, inexpensive, and sturdy smartphone is more suitable for pointof-care sensors [2,11], especially in resource-constrained countries. Therefore, we carried out the experiment assay with the help of a smartphone equipped with a color scan application software (ColorAssist), specially designed for photo analysis instead of a UV-vis spectrometer. After taking, The software can analyze a photo of the MA-capped Au NPs solution producing a color card and RGB (red, green, and blue) values. As we know, colorimetry is a method that is used to determine the concentration of colored compound in solution by the application of the Beer-Lambert law, which is highly dependent on the use of UV-vis spectrometer. Unlike colorimetry, RGB, as three primary of light, has been widely used in the field of image display since any color could be decoded by RGB value. The RGB components of light can represent any color, as shown in Figure 4. Small differences in the color of solutions of MA-capped Au NPs with different concentration of target could also be reflected by the ratio of green to red values displayed on the screen. Figure  5 shows that the ratio of green to red can be used to establish a correlation with Cr(VI) concentration. A linear calibration curve (y = 0.48297 + 0.49794x) was obtained and showed good linearity with a correlation coefficient of 0.99036 in the concentration range of 0.2 µ g L −1 to 1 µ g L −1 . The detection limit could reach as low as 0.1 µ g L −1 based on the

Smartphone-Based Colorimetric Assay
The expensive, delicate, and costly UV-Vis spectrometer is not appropriate for onsite application. The portable, inexpensive, and sturdy smartphone is more suitable for point-of-care sensors [2,11], especially in resource-constrained countries. Therefore, we carried out the experiment assay with the help of a smartphone equipped with a color scan application software (ColorAssist), specially designed for photo analysis instead of a UV-Vis spectrometer. After taking, The software can analyze a photo of the MA-capped Au NPs solution producing a color card and RGB (red, green, and blue) values. As we know, colorimetry is a method that is used to determine the concentration of colored compound in solution by the application of the Beer-Lambert law, which is highly dependent on the use of UV-Vis spectrometer. Unlike colorimetry, RGB, as three primary of light, has been widely used in the field of image display since any color could be decoded by RGB value. The RGB components of light can represent any color, as shown in Figure 4. Small differences in the color of solutions of MA-capped Au NPs with different concentration of target could also be reflected by the ratio of green to red values displayed on the screen. Figure 5 shows that the ratio of green to red can be used to establish a correlation with Cr(VI) concentration. A linear calibration curve (y = 0.48297 + 0.49794x) was obtained and showed good linearity with a correlation coefficient of 0.99036 in the concentration range of 0.2 µg L −1 to 1 µg L −1 . The detection limit could reach as low as 0.1 µg L −1 based on the definition of three times the deviation of the blank signal (3 σ). Our detection limit was found to be very near to the limitations which were reported by Yingying Qi et al.  [2,7,32]. However, our detection limit was found to be more reliable than the method reported by Sushant D. Bamane et al. (1.3-11.6 μg L −1 ) [11].    [2,7,32]. However, our detection limit was found to be more reliable than the method reported by Sushant D. Bamane et al. (1.3-11.6 μg L −1 ) [11].

Real Sample Detection Based on Smartphone Sensor
Samples collected from three local lakes were used to investigate the performance of our smartphone-based colorimetric method. The absolute concentrations of Cr(VI) in the three samples were measured by the traditional Atomic Absorption Spectroscopy (AAS) method. The samples were filtered through a 0.22 µm Super filter to remove any particulate suspension before the Cr(VI) determination. As revealed by the results in Table 1, one can find that the concentrations of Cr(VI) in all the three lakes were in the detection range of our MA-capped Au NPs sensor, suggesting the samples could be used directly to interrogate the performance of our probe. As expected, the results obtained by the smartphone were consistent with the data obtained by AAS, suggesting the impurities within the samples did not cause serious interference for Cr(VI) detection in real samples collected. Compared to the AAS results, the recovery rate of our smartphone-based method for Cr(VI) detection was in the range of 84.9% to 115.8%. In addition, analytical precision was compared by analyzing the coefficient variations for UV-Vis and the proposed method, which revealed that the precision of smartphone-based assay was equivalent to that of colorimetry, demonstrating the potential of the proposed method for accurate detection of analyte in real world. Table 1. Recovery test of Cr(VI) in a real water sample collected from three local lakes by the smartphone-based method and its coefficient variation comparison with conventional UV-Vis spectrometer a .

Sample b
The Concentration of Cr(VI) (µg L −1 )

Conclusions
In this work, an MA-capped Au NPs based colorimetric method was developed for the rapid detection of aqueous Cr(VI) with high sensitivity and selectivity. The response of our probe to aqueous Cr(VI) could be simply observed with a color change from wine-red to gray, and the limit of detection could reach as low as 0.4 µg L −1 by the naked eye. Moreover, with the help of a smartphone equipped with a color scan application software, Cr(VI) could be quantified on-site in a simple, portable, reliable, and rapid manner and detection liner range from 0.2 to 1.0 µg L −1 with a detection limit as low as 0.1 µg L −1 . The coefficient variations obtained by the proposed method is comparable to that of conventional UV-Vis spectrometer, demonstrating the potential of the proposed method in heavy metal ion detection on site in resource-constrained countries. In the future, it is believed that assays in decentralized studies could be dramatically improved if smartphone were to be integrated with more powerful software for data collection. In addition, the proposed method reported here could potentially be adapted for the detection of any environmental pollutants as long as coordination reagents against it were available, therefore making smartphone-based method a versatile tool for environmental monitoring. Furthermore, this methodology is not confined to colorimetric assay, fluorescence-based assay combined with smartphone seems to be more suitable for on-site detection in real world from the perspective of detection sensitivity, which has been already intensely reported with high accuracy and sensitivity in the past decade.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.