Methylene Blue, n-hexanol, Triton X-100, and tetraethyl orthosilicate were purchased from Sigma-Aldrich Korea, Ltd. Cyclohexane (99.5%), acetone (99.5%), and EtOH (99.9%) were bought from DaeJung Chemical & Metals (Siheung-Si, Korea). Ammonia hydroxide was purchased form Supelco, and arsenic standard solution was purchased from Kanto Chemical Ltd. (Tokyo, Japan).

To synthesize MB-Silica, 75 mL of cyclohexane, 18 mL of n-hexanol, and 18 mL of Triton X-100 were mixed until the mixture became transparent. Thereafter, 4 mL of 20 mM MB solution (in DI water) and 1 mL of tetraethyl orthosilicate solution were slowly injected into the mixture by using a syringe pump (NE-1000-ES, New Era Pump Systems, Inc.). After 20 min of reaction, 3 mL of ammonia hydroxide solution was added into the mixture, followed by stirring for 24 h to complete the chemical reaction. Upon completion of the stirring process, the compounds of silica and MB-silica were separated using an ultrasonic cleaner (SD-200H, Korea) and acetone. The separated components were washed with EtOH and deionized (DI) water at least three times. The silica and MB-silica obtained from this process were dried using a freeze dryer (FDU-2110, EYELA) for 24 h. To further improve the sensitivity and accuracy of arsenic detection, MB-silica was synthesized using different concentrations of MB solution in accordance with target arsenic concentrations. The MB concentrations used for MB-silica synthesis were 2 mM (for arsenic detection of 0–0.1 ppm), 5 mM (for detection of 0.1–1 ppm), and 20 mM (for detection of 1–100 ppm).

To characterize the surface properties of the developed MB-silica, we utilized FT-IR, XPS, and FE-SEM/EDS. The FT-IR (Cary 630 FT-IR, Agilent Technology, USA) measurement was conducted in the wavelength range of 650–4000 cm⁻¹, and the attenuated total reflection (ATR) method was used to examine the change in functional group during the synthesis of MB-silica.

To investigate the surface modification of silica through the introduction of methylene blue, XPS signal analysis of C_1S, O_1S and Si_2P was performed using an ULVAC PHI XPS instrument equipped with Al source (24.1 W, 15 kV, Al = K_a =1486.6 eV) in order to observe the changes in surface functionalities. The pressure in the analysis chamber was maintained at 7.49 × 10⁻⁷ Pa during the XPS measurement. All binding energies were referred to the C_1s peak to identify the peak of the main component of the sorbents.

FE-SEM was used to examine the surfaces of silica and MB-silica. The synthesized compounds were coated with platinum for 60 s (4-nm-thick coating) by means of sputtering in vacuum (Auto Fine Coater, JFC-1600, JEOL, Tokyo, Japan), and then analyzed using the FE-SEM (Quanta 250 FEG, FEI, USA). Additionally, the Pt-modified synthesized compounds were subjected to EDS analysis to confirm the distribution of elements on their surfaces.

The MB-silica synthesized using 20mM MB was suspended in distilled water (DW) at a concentration of 0.005g/mL. Subsequently, the solution was dispensed into a 96-well plate. After the completion of uniform dispensing, the sample to be used for measurement was dried in an oven at 60 °C for one day to evaporate the distilled water. To proceed with absorbance-based arsenic detection, 200 μl of arsenic standard solution of a specific concentration (0, 1, 5, 10, 25, 40, 60, and 100 ppm) was added to the 96-well plate at room temperature. Then, the arsenic reaction was carried out using a shaker operated at 100 rpm for 30 min. After the reaction, absorbance was measured at the wavelength of 660 nm by using a microplate reader (Hidex Multi-Technology Microplate Reader, Finland). The performance of the MB-silica method was compared to the MB solution method. The detection of arsenic using the MB solution was carried out according to the previously reported method [28].

To address the challenge of distinguishing measurements below 1 ppm arsenic concentration in 20mM MB-Silica, the following experiments were conducted. The preparation of MB-Silica involved the utilization of 2 mM and 5 mM MB to extend the range of bleaching during arsenic measurement. The manufacturing process of 2mM and 5mM MB-Silica was carried out following the previously mentioned procedure. The 2mM MB-Silica aimed to measure arsenic concentrations in the range of 0–0.1 ppm, while the 5mM MB-Silica targeted the measurement of arsenic concentrations in the range of 0.1–1 ppm, and both experiments were conducted following the same procedure as mentioned earlier, with three repetitions for each experiment.

The arsenic detection performance of the proposed MB-silica method was compared with inductively coupled plasma-optical emission spectroscopy (ICP-OES, Agilent, Santa Clara, CA, USA), a conventionally widely used arsenic detection method. For the ICP-OES analysis, a calibration curve was prepared using an arsenic standard solution. To compare the accuracy between MB-Silica and the widely used arsenic detection method, ICP-OES analysis, solutions containing arsenic concentration of 0.03, 0.05, 0.3 and 0.7 ppm were prepared and subjected to comparative analysis. All comparative analysis were conducted with three repetitions. To validate the arsenic-detection capacity of the MB-Silica-based method, the analysis was performed using MB-Silica, which is detectable at the concentration employed in ICP-OES detection and repeated over three times at 660 nm using a Microplate reader (Hidex Multi-Technology Microplate Reader, Finland).

To determine the functional groups of the synthesized compounds, FT-IR analyses were performed using MB, silica, and MB-silica, which results are presented in Fig. 1. In the FT-IR spectrum of the MB (Fig. 1A), peaks were observed in the wavelength bands of 2816, 1635, 1600, 1390, 1347, and 879 cm⁻¹. The peak at 2816 cm⁻¹ corresponds to the -CH-aromatic bond in MB. The peaks at 1635, 1600, 1390, and 1347 represent the -N⁺, C=N, C-H, and C-S bonds, respectively [29–32]. In the FT-IR spectrum of silica (Fig. 1B), peaks were observed at 1064, 947, and 800 cm⁻¹. The peak at 1064 cm⁻¹ was attributed to Si-O-Si conjugation [33]. The absorption band at 947 cm⁻¹ was ascribed to Si-OH [34]. The peak at 800 cm⁻¹ was assigned to Si-CH₃ stretching [35]. In the FT-IR spectrum of the MB-silica, as illustrated in Fig. 1C, the main peaks representing both MB and silica were all observed. For example, the main peaks of the silica (1064, 947, 800 cm−1) and the MB (1635, 1600, 1390, 1347, and 879 cm⁻¹) were all found in the MB-silica. Comparing the silica spectrum, the MB-silica displays the characteristic peaks of the MB, clearly indicating that the MB was successfully synthesized into the silica. From the FT-IR analysis of the synthesized MB-silicas, it was confirmed that MB was chemically conjugated with the silica surface, and MB functional groups were introduced on the surface. Lee et al. [36] reported the FT-IR spectrum of MB solution, where MB contains many C=N bonds and S functional groups that do not exist in silica. This report was consistent with the FT-IR spectrum of MB obtained herein. From these FT-IR spectra, it can be inferred that MB was conjugated with the silica surface because C=N (1600 cm⁻¹) and C=S (1347 cm⁻¹) bonds were formed.

To further confirm the change in functional groups in detail, XPS was utilized to analyze the C_1s, O_1s, and Si_2p signals of the silica and MB-silica. Table 1 summarizes the types of chemical bonds and the degrees of bonding of each element on the basis of Si_2p, C_1s, and O_1s in the XPS results. The two peaks observed at the binding energies of 101.9 and 102.4 eV in the Si_2p XPS spectrum of silica (Fig. 2A) were assigned to the Si-O and Si-O₂ bonds, respectively [37]. However, the peak at 102.8 eV (Fig. 2B) was newly assigned to the Si-O₃ bond on the synthesized MB-silica [38–39]. Furthermore, the newly introduced peak at 101.1 eV in the Si_2P measurement of MB-Silica confirms the presence of methylene blue introduced to the surface of silica [40]. No other peaks were found in the C_1s XPS spectrum of the silica (Fig. 2C) while the MB-Silica exhibits several peaks at 282.6, 284.2, and 285.9 eV owing to the chemical reaction of MB with silica (Fig. 2D). The peaks at the binding energies of 282.6 and 284.2 eV were newly assigned to the C=C and C-C bonds, respectively [41–42]. Moreover, the peak at the binding energy of 285.9 eV was newly assigned to the C-S and C-N bonds on MB-Silica [43]. Lastly, in the O_1s XPS spectra of silica and MB-silica (Fig. 2E, F), the peak at 531.7 eV (silica, Fig. 2E) shifted to 532.1 eV (MB-silica, Fig. 2F), and it was assigned to the Si-O bond [43]. The XPS C_1s and O_1s spectral analysis results confirmed that MB and silica were chemically combined, indicating that the properties of MB were introduced to silica.

Finally, the surfaces of the silica and MB-silica were analyzed through FE-SEM/EDS to investigate and compare the distribution of elements (Fig. S1). As shown in Table S1, the proportion of S atoms on the MB-silica surface increased while the proportion of O atoms decreased, compared to those of silica surface. The change on the MB-silica surface resulted from FE-SEM/EDS observed that MB molecules are bonded to the O in silica, resulting in decrease in the proportion of O, while the proportion of S increases with the introduction of MB molecules.

In sum, the FT-IR, XPS, and FE-SEM/EDS results confirmed that MB and Silica were chemically combined, where the chemical properties of MB were successfully introduced on the surface of MB-Silica [43]. From the above characterization results, we hypothesized that the developed MB-silica can be used as a tool for detecting and characterization the arsenic concentration.

MB-solution arsenic detection method has been reported extensively in the literature [16], [24], [44]. Kundu et al. [16] reported a limit of detection (LOD) of 1.3 ppm when using MB solution to detect arsenic. Based on this previous study, we synthesized 20 mM MB-silica. When conducting the arsenic measurement using the MB solutions with a microplate reader, the detection results were like those reported in the literature, but the arsenic concentrations lower than 25 ppm were hardly distinguishable (Fig. 3A). Furthermore, the rapid decrease in absorbance values at arsenic concentrations ranging from 0 to 20 ppm poses challenges for precise measurement, making it difficult to achieve accurate measurement even at concentration below 20 ppm. This issue leads to a significant margin of error in measurements, resulting in decreased accuracy of the measured concentrations. Moreover, in the MB solution measurement, various other chemical solutions, such as sodium dodecyl sulfate and sodium borohydride must be prepared. These chemicals need to be mixed with the MB solution right before the measurement, which requires additional sample preparation steps and limits its on-site applicability. Additionally, significant impact on the accurate measurement of arsenic concentrations can arise when encountering contamination of chemicals or changes in concentration during the preparation process when they need to be mixed with MB solution, as well as issues regarding the chemical reactivity of the prepared chemicals, which can hinder precise arsenic concentration measurement.

To address these issues in MB solution, a MB-silica method was conceived and proposed in this study with the development of MB-Silica, achieved by introducing MB molecules onto the surface of silica, eliminate the need for additional preparation and usage of chemical solutions. The proposed MB-silica method can simplify the measurement process and be applied to on-site arsenic detection since this method requires less preparation and its measurement is easier than the MB solution. Based on the measurement results (Fig. 3B), it was observed that unlike the conventional MB solution method, the R² value exhibits a high linear relationship, measuring at 0.9836. By establishing a stronger linear relationship, it was possible to overcome the limitation of the original MB-Solution, where the bleaching of the oxidation-reduction reaction ceased beyond a concentration of 10 ppm, ultimately enabling more precise measurement of a wide range of arsenic concentrations. Furthermore, it was observed that the accuracy within the detection range of the measurement was significantly improved, as the errors between the measured values were minimal during repeated measurements.

Taking these results into consideration, the newly developed MB-Silica arsenic detection method can be proposed to improve accuracy and detection resolution compared to the conventional MB solution. Furthermore, it was observed that, like the MB solution method, bleaching of absorbance occurs through arsenic and oxidation-reduction reactions, ultimately confirming the ability to detect arsenic by introducing MB molecules onto the silica surface without compromising their chemical properties.

The reason why the MB-Silica exhibits higher accuracy compared to the methylene blue-based detection can be attributed to the following factors. First, methylene blue molecules vital for detecting arsenic are dispersed within the methylene blue solution, whereas on MB-Silica, methylene blue is densely concentrated on the silica surface. As a result, the sensitivity of the reaction between essential methylene blue and arsenic ions for arsenic detection was increased. Additionally, it can be anticipated that silica with methylene blue imprinted on the surface as a catalyst would enhance the chemical reaction, thereby improving reactivity. According to previous literature [45], silica has been reported as an oxidizing agent, and several reports [46–47] have highlighted that the large surface area and stability of silica can enhance chemical reactions. Furthermore, several previous studies have reported that specific chemical molecules were introduced onto the surface for use as catalysts in chemical reactions such as oxidation-reduction reactions [48–49]. Consequently, by imprinting methylene blue molecules onto the silica surface, the reactivity with arsenic is amplified, and it is expected to boost the stability oxidation-reduction reactions owing to silica’s physical and chemical properties, ultimately resulting in enhanced accuracy in arsenic detection.

When the arsenic ions in aqueous solutions were analyzed by the previously reported MB-solution method (16, 24), it was difficult to distinguish the arsenic concentrations of 1 ppm or lower (Fig. S2). In this low arsenic concentration range, bleaching according to the concentration gradient was not observed, and thus the trend of absorbance measurement could not be confirmed. These results highlighted the difficulty associated with the MB-solution measurement of arsenic at concentrations of 1 ppm or lower. Similarly, the 20 mM MB-Silica exhibited increased accuracy and detection resolution in arsenic detection; however, it faces the challenge of difficult measurement below 1 ppm arsenic concentration. To overcome these issues, experiments were conducted to enhance the detection accuracy and resolution of the detection method for low arsenic concentrations by expanding the range of bleaching caused by oxidation-reduction reactions. In order to expand the range of bleaching, the concentration of MB used in the synthesis of MB-Silica was reduced from 20 mM to 2 mM and 5 mM. After the synthesis of MB-Silica, experiments were conducted with the aim of measuring arsenic concentration below 1ppm, which was a technological hurdle to differentiate in the previous 20mM MB-Silica. The experiment was conducted by following the aforementioned procedure, involving the reaction of the arsenic solution and measurement of wavelength changes at 660 nm.

Various concentrations of arsenic ions (0–10 ppm) were measured using the both 2mM and 5mM MB-Silica, and the detection results are summarized in Fig. 4 and Supplementary Fig. S3. When using the 2mM MB-silica, a linear relationship was derived at concentrations of 0–0.1 ppm, and for the 5mM MB-silica, a linear relationship was observed at concentrations of 0–1 ppm. The measurement of arsenic solution with a concentration range of 0–0.1 ppm using 2mM MB-Silica (Fig.4A) showed a more linear relationship compared to 5mM MB-Silica, resulting in improved detection accuracy and resolution. On the other hand, when measuring arsenic solution with a concentration range of 0.1–1 ppm, the use of 5mM MB-Silica resulted in a more linear relationship compared to 2mM MB-Silica, leading to improved detection accuracy and resolution in the measurement (Fig.4B).

In summary, based on the combined results, it was confirmed that the arsenic solution with concentrations of 1 ppm or lower, which could not be measured using 20 mM MB-Silica, were successfully measured using 2mM and 5mM MB-Silica. In the measurement of arsenic concentrations less than 1 ppm, 2mM MB-Silica exhibited a linear relationship within the range of 0–0.1 ppm, while 5mM MB-Silica exhibited a linear relationship within the range of 0.1–1 ppm, both allowing for clear differentiation of the tested arsenic concentrations. Therefore, it has been confirmed that MB-Silica can be measured differently based on the arsenic concentration levels, with a linear relationship observed at concentrations ranging from 0 to 0.1 ppm for 2 mM MB-Silica, 0 to 1 ppm for 5 mM MB-Silica, and 0 to 100 ppm for 20 mM MB-Silica. The MB-Silica method is expected to improve the accuracy and sensitivity of the arsenic detection by applying different MB-silica depending on the arsenic concentrations to be measured.

To assess the potential applicability of MB-Silica in environmental conditions, experiments were conducted to evaluate its arsenic detection reactivity and interference factors. Initially, arsenic is present in environmental conditions (both arsenic (III) and arsenic (V) ions), to verify their reactivity in arsenic detection. Given that the arsenic standard solution used in aforementioned experiments consists solely of arsenic (III), we prepared an arsenic (V) solution to confirm the reactivity of MB-Silica toward arsenic (V) ions. The preparation of the arsenic (V) solution was conducted following the established method outlined in previous research [50].

Again, the prepared 2 Mm MB-Silica was used for assessing the potential for detecting various ranges of arsenic (V) ions. The measurement results (Fig. 5A) revealed that, when investigating the reactivity of As (V) using 2 mM MB-Silica, a linear relationship was affirmed within the 0 to 0.1 ppm range, similar to As (III). Furthermore, within this range, a linear relationship equation was established as Y=1.8308−11.7248*X, with an R2 value of 0.9649, indicating a higher linearity compared to As (III) measurements. Following the same procedure as previously detailed, we conducted an evaluation of the reactivity of As (V) within the extended range of 0 to 1 ppm with the use of 5 mM MB-Silica, and in addition, we also affirmed the reactivity of arsenic at a further higher concentration, specifically 100 ppm. The analysis results (Fig. 5B) revealed that when measuring As (V) using 5 mM MB-Silica, a linear relationship within the concentration range of 0 to 1 ppm, similar to that of As (III), was observed. Moreover, within the linear range, we verified the linear relationship equation (Y=2.5815−0.9198*X) and obtained an R2 value of 0.9178. signifying a superior linearity compared to As (III) measurements. As the last step, measurements were carried out using 20 mM MB-Silica within further extended range of 0 to 100 ppm (Fig. 5C). Notably, unlike those in As (III), we were able to establish linear relationship within the range of 0 to 40 ppm, with a linear equation of Y=2.6141−0.0209*X and R2 value of 0.9278. This confirms that 20 mM MB-Silica is more efficient for As (III) detection but also capable of detecting As (V) as well. Upon comprehensive consideration of the results, it can be affirmed that MB-Silica enables the concurrent measurement of arsenic ions in their coexisting forms, As (III) and As (V), within aqueous solution.

Through this approach, we affirmed the reactivity of arsenic detection and, in addition, tested the responsiveness of MB-Silica towards various metal ions that might pose interference when quantifying arsenic levels in environmental sample and similar conditions. To verify the reactivity towards interfering factors, mercury (Hg) and cadmium (Cd) were selected from various metal ions, which can cause significant harm to aquatic ecosystems and the environment. For the purpose of assessing the interference caused by metal ions, we employed 20 mM MB-Silica to validate its reactivity, and our preliminary analysis (Fig. 6A) of its response to Hg indicated the absence of a linear relationship and reactivity when exposed to Hg concentrations ranging from 0 to 25 ppm. Moreover, we verified the reactivity of Cd by employing 20 mM MB-Silica, and resembling Hg, we noticed that Cd did not exhibit a linear correlation or reactivity.

While there are differences in reactivity depending on the oxidation-reduction reaction extent of MB-Silica, it has been confirmed that reactivity exists for both As (III) and As (V) in aqueous solution. Furthermore, it is likely that the difference in reactivity would be originated from the distinct chemical properties inherent to As (III) and As (V) [51]. In the case of As (III), it possesses the characteristic of being more readily reducible in its oxidized state, whereas As (V) has the feature of having a high oxidation rate, making it more challenging to reduce. Hence, at lower concentration (0 to 0.1 ppm, 0 to 1 ppm) of both As (III) and As (V), linear relationships could be established within the same range, indicating reactively for both ion forms, but at higher concentration range (0 to 100 ppm), differences in reactivity arise, resulting in slightly altered linear relationship, possibly due to chemical reaction disparities.

Furthermore, although the potential for interference by all metal ions has not been verified, it has been confirmed that reactivity and linear relationship do not exist when reacting with representative environmental pollutant such as Hg and Cd, indicating that they do not act as interfering factors. Thus, reactivity concerning the arsenic ion forms and metal ions that can potentially acts as interfering factors has been verified, confirming the feasibility of employing MB-Silica for arsenic detection. However, for practical application in environmental samples, it is necessary to verify the potential interference not only from metal ions but also from various organic contaminants, solid materials [52–53]. Therefore, in future research, additional verification of reactivity towards various interfering factors and optimization for detection performance improvement will be conducted to facilitate the detection of total arsenic from environmental sample using MB-Silica.

The accuracy of the measurement values was verified by comparing the measured arsenic concentration obtained using MB-Silica with those obtained through ICP-OES method, which is widely used for the measurement of heavy metals and other water contaminants, using prepared samples. As an application strategy of MB-silica for arsenic detection, different MB concentrations were set, and the arsenic detection concentration ranges were derived for each synthesized material and applied. In the arsenic concentration range of 0–0.1 ppm, for the arsenic concentration of 0.03 ppm, the ICP-OES method yielded an arsenic concentration 0.0301 ± 0.0018. For the same concentration, when the measurement was performed using MB-silica prepared with 2 mM MB, the measured arsenic concentration was 0.0300 ± 0.0017. At the arsenic concentration of 0.05 ppm, the results of the ICP-OES- and MB-silica-based methods were 0.0529 ± 0.0006 and 0.0519 ± 0.002, respectively (Fig. 7A). At arsenic concentrations of 0–1 ppm, measurement was performed using MB-silica synthesized using 5mM MB, and the result was compared to those obtained using the ICP-OES-based method. When the arsenic concentration of the measured solution was 0.3 ppm, the ICP-OES- and MB-silica-based method yielded concentration values of 0.3748 ± 0.0013 and 0.3252 ± 0.027 ppm, respectively. In addition, for the 0.7 ppm arsenic solution, the ICP-OES- and MB-silica-based method yielded concentration values of 0.7365 ± 0.0023 and 0.7025 ± 0.0223, respectively (Fig. 7B). These results confirmed that the concentration values obtained using the proposed MB-silica-based method were closer to the standard solution than those obtained using the ICP-OES-based method. Thus, the MB-silica-based method can be considered more accurate than the ICP-OES-based method.

As described above, when arsenic was detected using the MB-silica, the detection capability similar to the conventional method, ICP-OES was achieved. In addition, a higher arsenic detection sensitivity was achieved using the proposed method compared to the existing MB-solution method. In conclusion, the possibility of arsenic detection was successfully verified through the use of the solidified MB-silica arsenic detector. Moreover, the possibility of detecting arsenic in a more precise range was confirmed by varying the MB concentration used to synthesize MB-silica.