Nonenzymatic Hydrogen Peroxide Detection Using Surface-Enhanced Raman Scattering of Gold–Silver Core–Shell-Assembled Silica Nanostructures

Hydrogen peroxide (H2O2) plays important roles in cellular signaling and in industry. Thus, the accurate detection of H2O2 is critical for its application. Unfortunately, the direct detection of H2O2 by surface-enhanced Raman spectroscopy (SERS) is not possible because of its low Raman cross section. Therefore, the detection of H2O2 via the presence of an intermediary such as 3,3,5,5-tetramethylbenzidine (TMB) has recently been developed. In this study, the peroxidase-mimicking activity of gold–silver core–shell-assembled silica nanostructures (SiO2@Au@Ag alloy NPs) in the presence of TMB was investigated using SERS for detecting H2O2. In the presence of H2O2, the SiO2@Au@Ag alloy catalyzed the conversion of TMB to oxidized TMB, which was absorbed onto the surface of the SiO2@Au@Ag alloy. The SERS characteristics of the alloy in the TMB–H2O2 mixture were investigated. The evaluation of the SERS band to determine the H2O2 level utilized the SERS intensity of oxidized TMB bands. Moreover, the optimal conditions for H2O2 detection using SiO2@Au@Ag alloy included incubating 20 µg/mL SiO2@Au@Ag alloy NPs with 0.8 mM TMB for 15 min and measuring the Raman signal at 400 µg/mL SiO2@Au@Ag alloy NPs.


Preparation of SiO2@Au@Ag Alloy NPs
SiO2@Au@Ag alloy NPs were prepared according to the steps outlined in a previous study [25]. Briefly, SiO2@Au seed was obtained by incubating 10 mL of Au NP suspension and 2 mL of aminated silica NPs overnight. The SiO2@Au@Ag alloy NPs were prepared by reducing 300 µM Ag+ to Ag on the surface of SiO2@Au in an aqueous medium using ascorbic acid in polyvinylpyrrolidone (PVP). The pellet was obtained by centrifuging the suspension for 15 min at 8500 rpm, washed thoroughly with EtOH, and re-dispersed in absolute EtOH to obtain a 200 µg/mL SiO2@Au@Ag alloy NP suspension.

Peroxidase-Like Activity of SiO2@Au@Ag Alloy NPs in the TMB and H2O2 Mixture in Various Reaction Conditions
To verify the peroxidase-like activity of SiO2@Au@Ag alloy NPs, 10 mM TMB solutions were first prepared in EtOH. Next, 100 µL of TMB solution and 100 µL of SiO2@Au@Ag alloy NPs were added to 100 µL of phosphate-buffered saline (PBS) buffer (pH 7.0) containing freshly prepared H2O2 solution. The mixture was incubated for 15 min at 25 °C and centrifuged at 15,000 rpm for 15 min. The excess reagents were washed thoroughly with PBS containing 0.1% Tween 20 (PBST), and SiO2@Au@Ag@TMB NPs were then re-dispersed in PBST.

TMB Concentration
Solutions of various TMB concentration were prepared in EtOH. Next, 100 µL of TMB solution, 100 µL of 200 µg/mL SiO2@Au@Ag alloy NPs, and 100 µL of freshly prepared 2.0 M H2O2 solution were added to a pH 7.0 buffer (700 µL) to obtain final concentrations of TMB in the range of 0.1 to 1.0 mM. Each mixture was incubated for 15 min at 25 °C and Scheme 1. Illustration of the peroxidase-mimicking nanozyme activity of gold-silver core-shell-assembled silica alloy suspension (SiO 2 @Au@Ag alloy nanoparticles) in the presence of 3,3,5,5-tetramethylbenzidine and H 2 O 2 .

Preparation of SiO 2 @Au@Ag Alloy NPs
SiO 2 @Au@Ag alloy NPs were prepared according to the steps outlined in a previous study [25]. Briefly, SiO 2 @Au seed was obtained by incubating 10 mL of Au NP suspension and 2 mL of aminated silica NPs overnight. The SiO 2 @Au@Ag alloy NPs were prepared by reducing 300 µM Ag + to Ag on the surface of SiO 2 @Au in an aqueous medium using ascorbic acid in polyvinylpyrrolidone (PVP). The pellet was obtained by centrifuging the suspension for 15 min at 8500 rpm, washed thoroughly with EtOH, and re-dispersed in absolute EtOH to obtain a 200 µg/mL SiO 2 @Au@Ag alloy NP suspension.

Peroxidase-like Activity of SiO 2 @Au@Ag Alloy NPs in the TMB and H 2 O 2 Mixture in Various Reaction Conditions
To verify the peroxidase-like activity of SiO 2 @Au@Ag alloy NPs, 10 mM TMB solutions were first prepared in EtOH. Next, 100 µL of TMB solution and 100 µL of SiO 2 @Au@Ag alloy NPs were added to 100 µL of phosphate-buffered saline (PBS) buffer (pH 7.0) containing freshly prepared H 2 O 2 solution. The mixture was incubated for 15 min at 25 • C and centrifuged at 15,000 rpm for 15 min. The excess reagents were washed thoroughly with PBS containing 0.1% Tween 20 (PBST), and SiO 2 @Au@Ag@TMB NPs were then re-dispersed in PBST.

TMB Concentration
Solutions of various TMB concentration were prepared in EtOH. Next, 100 µL of TMB solution, 100 µL of 200 µg/mL SiO 2 @Au@Ag alloy NPs, and 100 µL of freshly prepared 2.0 M H 2 O 2 solution were added to a pH 7.0 buffer (700 µL) to obtain final concentrations of TMB in the range of 0.1 to 1.0 mM. Each mixture was incubated for 15 min at 25 • C and centrifuged at 15,000 rpm for 15 min. The SiO 2 @Au@Ag@TMB NPs were then re-dispersed in PBST (100 µL) to obtain a SiO 2 @Au@Ag@TMB suspension.

Reaction Time
The effect of reaction time on the SERS signal of SiO 2 @Au@Ag alloy NPs was investigated in the mixture containing 8.0 mM TMB solution (100 µL), SiO 2 @Au@Ag alloy NPs (0.2 mg/mL, 100 µL), and 2.0 M H 2 O 2 solution (100 µL) in a pH 7.0 buffer (700 µL). The mixtures were reacted for 5 to 60 min. The prepared NPs were obtained by centrifuging at 15,000 rpm for 15 min, washed thoroughly, and redispersed in PBST.

Amount of SiO 2 @Au@Ag Alloy NPs
To investigate the effect of SiO 2 @Au@Ag alloy NP amount on the SERS signal of SiO 2 @Au@Ag, 8.0 mM TMB solution (100 µL), 2.0 M H 2 O 2 solution (100 µL) and SiO 2 @Au@Ag alloy NPs in the range of 10 to 50 µg were added in a pH 7.0 buffer (700 µL). The mixtures were reacted for 15 min at 25 • C, centrifuged for 15 min, washed thoroughly with PBST, and then re-dispersed in PBST (100 µL).

Concentration of SiO 2 @Au@Ag Alloy NPs for Raman Measurement
A 10 mM TMB solution (100 µL), SiO 2 @Au@Ag alloy NPs (0.2 mg/mL, 100 µL), and freshly prepared 2.0 M H 2 O 2 solution (100 µL) were added to pH 6.0 buffer. Next, the mixtures were reacted for 15 min at 25 • C, centrifuged for 15 min, washed thoroughly with PBST, and then re-dispersed in of PBST. The volume of PBST was adjusted to obtain final concentrations of SiO 2 @Au@Ag alloy NPs in the range of 50 to 400 µg/mL for the Raman measurement.
2.3.6. Detection of H 2 O 2 using SiO 2 @Au@Ag Alloy NPs A quantity of 20 µg SiO 2 @Au@Ag alloy NPs in PBST (100 µL) and 8.0 mM TMB in EtOH (100 µL) were added to pH 6.0 buffer (700 µL). PBS (100 µL) containing different concentrations of H 2 O 2 (0.1 to 120 mM) was added to the above-mentioned mixture and allowed to react for 15 min at 25 • C. This mixture was centrifuged for 15 min at 15,000 rpm, washed thoroughly with PBST, and then re-dispersed in PBST to obtain a 400 µg/mL NP suspension for Raman measurement.

Long-Term Stability of SiO 2 @Au@Ag Alloy NPs
To investigate the long-term stability of SiO 2 @Au@Ag alloy NPs, they were dispersed in EtOH at 200 µg/mL and stored at 4 • C for 60 days. The dispersion of SiO 2 @Au@Ag alloy NPs was shaken and diluted to 20 µg/mL. UV-vis spectroscopy was performed at wavelengths of 300 to 800 nm, and the absorbance of the SiO 2 @Au@Ag alloy NP suspension at 460 nm was recorded.

SERS Measurement
The SERS signals were measured at 10 mW for 5 s at randomly selected sites using a micro-Raman system with a 532 nm diode-pumped solid-state laser excitation source and an optical microscope (Olympus BX41, Tokyo, Japan) with a 10× objective lens (0.90 NA, Olympus, Tokyo, Japan). The laser beam spot was~2.1 µm, and the SERS spectrum in the range of 300-1800 cm −1 was obtained.

Results and Discussion
SiO 2 @Au@Ag alloy NPs were prepared using a protocol reported by Pham et al. [27,[34][35][36][37][38][39]. The transmission electron microscopy (TEM, JEOL, Akishima, Tokyo) images and UV-vis extinction spectra (Mecasys, Seoul, Korea) of the SiO 2 @Au@Ag alloy NPs are shown in Figure  S1. The surface of SiO 2 @Au was effectively coated with the Ag shell. Various tiny Au NPs decorated the surfaces of the SiO 2 NPs. The UV-vis extinction spectra of the SiO 2 @Au@Ag alloy NPs are consistent with the TEM images ( Figure S1b). The as-prepared SiO 2 @Au@Ag alloy NP suspension shows the characteristic spectrum in Figure S1 with a broad band from 320 to 700 nm and a maximum peak at~460 nm. This indicates that Ag shells formed on the SiO 2 @Au NP surfaces, created many hot-spot structures, and led a continuous spectrum of resonant multi-modes of the SiO 2 @Au@Ag alloy NP suspension [25,27]. These results are consistent with Mie's theory, which states that an increase in particle size leads to a shift of the plasmon absorption band to longer wavelengths [60]. The peroxidase-like activity of the SiO 2 @Au@Ag alloy NPs was evaluated through the oxidation of TMB. The oxidation of TMB includes two steps: First, TMB is oxidized to TMB + (oxTMB), and then the clear TMB solution changes to blue in color. However as TMB + is quite unstable, it is oxidized to TMB 2+ in acidic conditions and exhibits a yellow color [61]. In this study, we investigated the peroxidase-mimicking activity of SiO 2 @Au@Ag alloy NPs by mixing 20 µg SiO 2 @Au@Ag alloy NPs (i) in 100 µL of H 2 O 2 (ii), TMB (iii), and TMB-H 2 O 2 mixture (iv) as shown in Figure 1a. The reaction solutions were incubated for 15 min at 25 • C, and the results are shown in Figure 1a. The color of the SiO 2 @Au@Ag alloy NP suspension in TMB solution (left column, (iii)) was dark brown and similar to that of SiO 2 @Au@Ag (left column, (i)), whereas the SiO 2 @Au@Ag alloy NP suspension in H 2 O 2 solution (left column, ii) turned grey in color due to its oxidation by H 2 O 2 solution.
In the presence of TMB and H 2 O 2 , the color of the SiO 2 @Au@Ag alloy NP suspension was light brown (left column, iv). After centrifugation, the supernatant in TMB and H 2 O 2 solution in the presence of SiO 2 @Au@Ag alloy NPs was transparent and colorless (center column, (ii) and (iii)), indicating that peroxidase-mimicking activity did not occur in the absence of either H 2 O 2 or TMB. By contrast, the supernatant in the TMB-H 2 O 2 mixture and SiO 2 @Au@Ag alloy NPs (center column, (iv)) changed in color from transparent to yellow as shown in Figure 1a(i-iv). The colors of the pellets of SiO 2 @Au@Ag alloy NPs after centrifugation and redispersion in PBST are shown in the right-hand column of Figure 1a. Once again, the color of SiO 2 @Au@Ag alloy NP suspension (i) was the same as that of SiO 2 @Au@Ag alloy NPs with TMB solution (iii), whereas the SiO 2 @Au@Ag alloy NP suspensions in H 2 O 2 (ii) and a mixture of TMB and H 2 O 2 (iv) became light grey and dark grey, respectively. These results imply that SiO 2 @Au@Ag alloy NPs possess peroxidase-mimicking activity in the presence of H 2 O 2 during the conversion of TMB to oxTMB.
The UV-vis extinction spectra of the SiO 2 @Au@Ag alloy NP suspension before and after centrifugation are shown in Figure 1b. The SiO 2 @Au@Ag alloy NP suspension without TMB and H 2 O 2 showed a broad band from 320 nm to 700 nm with a maximum peak at 460 nm. The presence of either TMB or H 2 O 2 led the UV-vis extinction spectra of the SiO 2 @Au@Ag alloy NPs to be slightly red-shifted from~460 nm to~500 nm (left column), whereas the suspension of SiO 2 @Au@Ag alloy NPs in TMB-H 2 O 2 mixture showed a broad and strong peak in the range of 350 nm to 800 nm with clear and multiple peaks at 370 nm and 650 nm [62]. This indicated that TMB was converted to oxTMB in the presence of H 2 O 2 ; this reaction was catalyzed by SiO 2 @Au@Ag alloy NPs, as shown in Figure S2. Moreover, the supernatant of the SiO 2 @Au@Ag alloy NPs in TMB-H 2 O 2 mixture also confirmed the presence of oxTMB when the SiO 2 @Au@Ag alloy NPs were removed from the suspension because of the excess oxTMB in the supernatant, which did not adsorb onto the surface of SiO 2 @Au@Ag alloy NPs (center column). Therefore, SiO 2 @Au@Ag alloy NPs possessed an intrinsic peroxidase-mimicking activity that catalyzed the conversion of TMB to oxTMB, as expected. In addition, the zeta potential values of the SiO 2 @Au@Ag alloy NPs in TMB, H 2 O 2 , and a mixture of TMB and H 2 O 2 were also studied, as shown in Figure S3a. SiO 2 @Au@Ag alloy NPs showed a zeta potential of −24.5 ± 0.6 mV due to the rich electron cloud of the Ag layer. The presence of H 2 O 2 or TMB in the suspension of SiO 2 @Au@Ag alloy NPs converted the surface charge to −14.6 ± 0.4 mV or −2.3 ± 0.6 mV, respectively. Thus, the zeta potential of SiO 2 @Au@Ag alloy NPs increased from −24.5 ± 0.6 mV to −14.6 ± 0.1 mV after adding the TMB-H 2 O 2 mixture into the reaction mixture.
face of SiO2@Au@Ag alloy NPs (center column). Therefore, SiO2@Au@Ag alloy NPs possessed an intrinsic peroxidase-mimicking activity that catalyzed the conversion of TMB to oxTMB, as expected. In addition, the zeta potential values of the SiO2@Au@Ag alloy NPs in TMB, H2O2, and a mixture of TMB and H2O2 were also studied, as shown in Figure S3a.
SiO2@Au@Ag alloy NPs showed a zeta potential of −24.5 ± 0.6 mV due to the rich electron cloud of the Ag layer. The presence of H2O2 or TMB in the suspension of SiO2@Au@Ag alloy NPs converted the surface charge to −14.6 ± 0.4 mV or −2.3 ± 0.6 mV, respectively. Thus, the zeta potential of SiO2@Au@Ag alloy NPs increased from −24.5 ± 0.6 mV to −14.6 ± 0.1 mV after adding the TMB-H2O2 mixture into the reaction mixture. In EtOH solution, the Raman spectra of the SiO2@Au@Ag alloy NP suspension in TMB and H2O2 are shown in Figure 1c(i). In the SiO2@Au@Ag alloy NP suspension, SERS bands were obtained at 431, 883, 1049, 1093, 1277, and 1455 cm −1 , which were assigned to the EtOH solution [27]. New SERS bands of the TMB-H2O2 mixture in the presence of SiO2@Au@Ag alloy NPs were also observed at 510, 1191, 1341, 1463, and 1608 cm −1 . The bands of the TMB-H2O2 mixture at 1341, 1463, and 1608 cm −1 were remarkably increased compared to those of the TMB solution. However, the overlap in some EtOH and oxTMB bands and the high background signal of EtOH can hinder analysis and give false results. Therefore, SiO2@Au@Ag alloy NPs must be centrifuged and re-dispersed in PBST, and the Raman spectra must be measured in the PBST solution. Indeed, the SERS bands of EtOH disappeared, and the SERS bands of TMB and oxTMB are clearly observed in Figure 1c(ii). The TMB bands showed typical SERS bands at 1191, 1341, and 1608 cm −1 , which correspond to the CN stretching vibration, CH stretching vibration, and CC stretching vibration, respectively [59,62]. The SERS bands of SiO2@Au@Ag alloy NPs in the TMB-H2O2 mixture were observed at 1191, 1341, 1468, 1563, and 1608 cm −1 . In contrast with the results In EtOH solution, the Raman spectra of the SiO 2 @Au@Ag alloy NP suspension in TMB and H 2 O 2 are shown in Figure 1c(i). In the SiO 2 @Au@Ag alloy NP suspension, SERS bands were obtained at 431, 883, 1049, 1093, 1277, and 1455 cm −1 , which were assigned to the EtOH solution [27]. New SERS bands of the TMB-H 2 O 2 mixture in the presence of SiO 2 @Au@Ag alloy NPs were also observed at 510, 1191, 1341, 1463, and 1608 cm −1 . The bands of the TMB-H 2 O 2 mixture at 1341, 1463, and 1608 cm −1 were remarkably increased compared to those of the TMB solution. However, the overlap in some EtOH and oxTMB bands and the high background signal of EtOH can hinder analysis and give false results. Therefore, SiO 2 @Au@Ag alloy NPs must be centrifuged and re-dispersed in PBST, and the Raman spectra must be measured in the PBST solution. Indeed, the SERS bands of EtOH disappeared, and the SERS bands of TMB and oxTMB are clearly observed in Figure 1c(ii). The TMB bands showed typical SERS bands at 1191, 1341, and 1608 cm −1 , which correspond to the CN stretching vibration, CH stretching vibration, and CC stretching vibration, respectively [59,62]. The SERS bands of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture were observed at 1191, 1341, 1468, 1563, and 1608 cm −1 . In contrast with the results of a previous study [58,59], oxTMB in our study did not induce the aggregation of Ag NPs but was adsorbed on the surface of SiO 2 @Au@Ag alloy NPs and showed the characteristic bands of oxTMB as mentioned above. Therefore, these Raman peaks were all used for the measurement of the SiO 2 @Au@Ag alloy NPs + TMB-H 2 O 2 system.
According to the literature [61,63], peroxidase activity was stopped by adding H 2 SO 4 to convert oxTMB to TMB 2+ , which is stable under acidic conditions and possesses an absorbance at 455 nm ( Figure S2). The colors of the mixtures of SiO 2 @Au@Ag alloy NPs and TMB-H 2 O 2 after adding H 2 SO 4 are shown in Figure 2a (left column). The suspension changed from brown to blue-green at low H + concentration (0.01 M H 2 SO 4 ), indicating that oxTMB can partly convert to TMB 2+ , giving a mix of blue (oxTMB) and yellow (TMB 2+ in acidic conditions). By contrast, the colors of the mixtures of SiO 2 @Au@Ag alloy NPs and TMB-H 2 O 2 changed from brown to light yellow and yellow, respectively, when 0.1 M and 1.0 M H 2 SO 4 were added to the reaction suspension (Figure 2a). This indicates that oxTMB can be completely converted to TMB 2+ . The results were confirmed by the color of supernatant after centrifugation in Figure 2a (center column).
According to the literature [61,63], peroxidase activity was stopped by adding H2SO4 to convert oxTMB to TMB 2+ , which is stable under acidic conditions and possesses an absorbance at 455 nm ( Figure S2). The colors of the mixtures of SiO2@Au@Ag alloy NPs and TMB-H2O2 after adding H2SO4 are shown in Figure 2a (left column). The suspension changed from brown to blue-green at low H + concentration (0.01 M H2SO4), indicating that oxTMB can partly convert to TMB 2+ , giving a mix of blue (oxTMB) and yellow (TMB 2+ in acidic conditions). By contrast, the colors of the mixtures of SiO2@Au@Ag alloy NPs and TMB-H2O2 changed from brown to light yellow and yellow, respectively, when 0.1 M and 1.0 M H2SO4 were added to the reaction suspension (Figure 2a). This indicates that oxTMB can be completely converted to TMB 2+ . The results were confirmed by the color of supernatant after centrifugation in Figure 2a (center column).
The UV-vis extinction spectra of the SiO2@Au@Ag alloy NPs suspensions in TMB-H2O2 mixture + H2SO4 are consistent with the optical images in Figure 2b. At low H + concentrations, the peaks of oxTMB at 370 and 650 nm are clearly observed in Figure 2b(i,ii). These two peaks disappeared after the addition of 0.1 M and 1.0 M H2SO4, indicating the conversion of oxTMB to TMB 2+ . However, the absorbance intensity of the SiO2@Au@Ag alloy NP pellet collected after centrifugation and redispersion in PBST dramatically decreased with the addition of H2SO4 (Figure 2b(iii)). The UV-vis extinction spectra of the SiO 2 @Au@Ag alloy NPs suspensions in TMB-H 2 O 2 mixture + H 2 SO 4 are consistent with the optical images in Figure 2b. At low H + concentrations, the peaks of oxTMB at 370 and 650 nm are clearly observed in Figure 2b(i,ii). These two peaks disappeared after the addition of 0.1 M and 1.0 M H 2 SO 4 , indicating the conversion of oxTMB to TMB 2+ . However, the absorbance intensity of the SiO 2 @Au@Ag alloy NP pellet collected after centrifugation and redispersion in PBST dramatically decreased with the addition of H 2 SO 4 (Figure 2b(iii)).
Furthermore, the SERS signal of SiO 2 @Au@Ag alloy NP suspension in TMB-H 2 O 2 + H 2 SO 4 sharply decreased with an increase in H 2 SO 4 concentration from 0.01 M to 1.0 M (Figure 2c(i)). The decrease in the SERS signal of oxTMB following the addition of H 2 SO 4 was caused by the desorption of oxTMB from the surface of SiO 2 @Au@Ag alloy NPs and/or low enhancement of the Ag substrate due to the formation of Ag 2 SO 4 on the surface of SiO 2 @Au@Ag alloy NPs. To examine the charge of SiO 2 @Au@Ag alloy NPs in TMB-H 2 O 2 + H 2 SO 4 , the zeta potential of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture before and after adding H 2 SO 4 was measured, as shown in Figure S3b. The zeta potential of SiO 2 @Au@Ag alloy NPs increased from −14.6 ± 0.1 mV to −4.4 ± 0.4 mV after the addition of 1.0 M H 2 SO 4 to the TMB-H 2 O 2 mixture. This result indicates that TMB 2+ could still be immobilized on the surface of SiO 2 @Au@Ag alloy NPs. Therefore, we changed the acidic agent from H 2 SO 4 to HCl and HNO 3 , while retaining the H + concentration at 2.0 M in the reaction. Similar to that with H 2 SO 4 , the SERS signal of oxTMB in the SiO 2 @Au@Ag alloy NP suspension also decreased slightly with an increase in HCl concentration, possibly because of the formation of AgCl on the surface of SiO 2 @Au@Ag alloy NPs (Figure 2c(ii)). By contrast, the SERS signal of oxTMB in SiO 2 @Au@Ag alloy NP suspension remained almost the same when 0.2 M and 2.0 M HNO 3 were added. However, the SERS signal of oxTMB slightly decreased when 0.02 M HNO 3 was added to the reaction. This was because the highly water-soluble AgNO 3 was dissolved in the aqueous solution and was not adsorbed on the surface of the SiO 2 @Au@Ag alloy NPs. Thus, the decrease in the SERS signal of SiO 2 @Au@Ag alloy NPs in TMB-H 2 O 2 mixture after adding H 2 SO 4 was caused by the formation of Ag 2 SO 4 , and it decreased the electromagnetic enhancement of the Ag layer on the surface of the SiO 2 @Au@Ag alloy NPs . Therefore, we decided not to use acidic conditions to terminate the peroxidase reaction of SiO 2 @Au@Ag alloy NPs.

Effect of Ag + Concentration on the Detection of H 2 O 2 by SiO 2 @Au@Ag Alloy NPs
According to our previous report, the SERS enhancement of the SiO 2 @Au@Ag alloy NPs depends on the gaps between Ag NPs on the surface of SiO 2 @Au@Ag alloy NPs, as indicated by adjusting the concentration of AgNO 3 in the solution from 50 to 300 µM [27,38,39]. In Figure 3a, the size of the Au@Ag increased with the concentration of Ag + used.

Optimization of SiO2@Au@Ag Alloy NPs for Detecting H2O2
In the literature, the catalytic activity of nanozymes is also affected by reaction conditions [1,[66][67][68][69][70]. The effects of the reaction conditions, including TMB concentration, reaction time, number of SiO2@Au@Ag alloy NPs, and pH of the buffer, on the peroxidasemimicking activity of SiO2@Au@Ag alloy NPs were considered in this study. As mentioned above, the SERS bands of oxTMB at 1191, 1341, 1468, 1563, 1608, and 1628 cm −1 were investigated in the TMB-H2O2 mixture. The effects of these conditions on the SERS signal The UV-vis extinction spectra of the SiO 2 @Au@Ag alloy NPs shown in Figure 3b are in agreement with the results of TEM images. The maximum UV-vis extinction peak was red-shifted from 450 to 530 nm and broadened from 300 to 800 nm with an increase of Ag + Nanomaterials 2021, 11, 2748 9 of 15 because of the generation of hot-spot structures between two adjacent Au@Ag NPs on the SiO 2 @Au@Ag alloy NPs [22,50,64].
The peroxidase-mimicking activities of SiO 2 @Au@Ag alloy NPs prepared at various Ag + concentrations toward the TMB-H 2 O 2 mixture are shown in Figure 3c. The SERS signal intensities at 1314, 1468, 1563, 1608, and 1628 cm −1 were proportional to the AgNO 3 concentration and highest at 300 µM due to the narrow gaps between Ag NPs on the SiO 2 @Au@Ag alloy NP surfaces that created "hot-spots" and strongly enhanced the electromagnetic field surrounding the SiO 2 @Au@Ag alloy NPs [33,65]. At higher Ag + concentration, the SiO 2 @Au@Ag alloy NPs were aggregated in the TMB-H 2 O 2 mixture. Therefore, 300 µM AgNO 3 was the optimal concentration for synthesizing SiO 2 @Au@Ag alloy NPs with the highest peroxidase-mimicking activity.

Optimization of SiO 2 @Au@Ag Alloy NPs for Detecting H 2 O 2
In the literature, the catalytic activity of nanozymes is also affected by reaction conditions [1,[66][67][68][69][70]. The effects of the reaction conditions, including TMB concentration, reaction time, number of SiO 2 @Au@Ag alloy NPs, and pH of the buffer, on the peroxidasemimicking activity of SiO 2 @Au@Ag alloy NPs were considered in this study. As mentioned above, the SERS bands of oxTMB at 1191, 1341, 1468, 1563, 1608, and 1628 cm −1 were investigated in the TMB-H 2 O 2 mixture. The effects of these conditions on the SERS signal of the SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture were examined and optimized, as shown in Figure 4. the TMB-H2O2 mixture decreased when 40 µg of SiO2@Au@Ag alloy NPs was used, indicating that the density of oxTMB gradually decreased with an increase in the number of substrates. The SERS signals at all SERS bands of SiO2@Au@Ag alloy NPs increased with a decrease in alloy quantity, as shown in Figure 4c. However, a small amount of SiO2@Au@Ag alloy NPs reduces the peroxidase-mimicking catalytic efficiency of the SiO2@Au@Ag alloy NPs to convert TMB to oxTMB and lower the SERS signal of oxTMB at 10 µg ( Figure 4c). Therefore, to ensure that the SiO2@Au@Ag alloy NPs can convert sufficient amounts of TMB to oxTMB in the presence of H2O2, we decided to use 20 µg of SiO2@Au@Ag alloy NPs for further study. The effect of the pH on the SERS signal of SiO2@Au@Ag alloy NPs in the TMB-H2O2 mixture is shown in Figures S4 and 4d. Similar to that of Au, the catalytic activity of SiO2@Au@Ag alloy NPs on H2O2 was also pH-dependent. The SERS signal of SiO2@Au@Ag alloy NPs in the TMB-H2O2 mixture showed the highest peroxidase-mimicking activity at pH 6.0. At pH ≤ 4, the SERS signal of the TMB-H2O2 mixture was too weak in all SERS bands. The SERS signal increased in the pH range of 5-6 and then decreased in the pH range of 7-9. In previous reports, the catalytic reaction of TMB-H2O2 was much faster in a weakly acidic solution than in neutral or basic solutions. However, in our study, the SiO2@Au@Ag alloy NPs lost 95% of their maximum activity at pH 3.0 and retained ~70% of their maximum activity in the pH range of 7.0-9.0. This is because SiO2@Au@Ag alloy NPs catalyze the generation of ·OH from the decomposition of H2O2 with the dissolution of Ag to release Ag + under strongly acidic conditions, as shown in the following equation: To establish the effect of TMB concentration on the SERS signal of the SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture, the concentrations of TMB were investigated in the range of 0 to 1.0 mM (Figure 4a). The SERS signal of SiO 2 @Au@Ag alloy NPs was almost insignificant at TMB concentrations lower than 0.4 mM. It increased remarkably when the TMB concentration was higher than 0.4 mM, achieved the highest value at 0.8 mM, but then decreased at higher TMB concentrations because of the poor solubility of TMB in aqueous solution (Figure 4a) [63].
The reaction time or incubation time for detecting H 2 O 2 using SiO 2 @Au@Ag alloy NPs is also shown in Figure 4b. The SERS signals of oxTMB at 1191, 1341, 1468, 1563, 1608, and 1628 cm −1 increased with reaction time and were saturated at 20 min. A gradual decrease in the SERS signal of SiO 2 @Au@Ag alloy NPs occurred after 20 min, owing to the instability of oxTMB [61].
In addition, the effect of the amount of SiO 2 @Au@Ag alloy NPs was observed in the range of 10-40 µg, as shown in Figure 4c. The SERS signal is dependent on the Raman reporter density on the nanomaterial surface [26]. Therefore, when a large amount of SiO 2 @Au@Ag alloy NPs was added, less oxTMB was available on the surface of the SiO 2 @Au@Ag alloy NPs. The SERS signals at all SERS bands of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture decreased when 40 µg of SiO 2 @Au@Ag alloy NPs was used, indicating that the density of oxTMB gradually decreased with an increase in the number of substrates. The SERS signals at all SERS bands of SiO 2 @Au@Ag alloy NPs increased with a decrease in alloy quantity, as shown in Figure 4c. However, a small amount of SiO 2 @Au@Ag alloy NPs reduces the peroxidase-mimicking catalytic efficiency of the SiO 2 @Au@Ag alloy NPs to convert TMB to oxTMB and lower the SERS signal of oxTMB at 10 µg (Figure 4c). Therefore, to ensure that the SiO 2 @Au@Ag alloy NPs can convert sufficient amounts of TMB to oxTMB in the presence of H 2 O 2 , we decided to use 20 µg of SiO 2 @Au@Ag alloy NPs for further study.
The effect of the pH on the SERS signal of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture is shown in Figure S4 and Figure 4d. Similar to that of Au, the catalytic activity of SiO 2 @Au@Ag alloy NPs on H 2 O 2 was also pH-dependent. The SERS signal of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture showed the highest peroxidasemimicking activity at pH 6.0. At pH ≤ 4, the SERS signal of the TMB-H 2 O 2 mixture was too weak in all SERS bands. The SERS signal increased in the pH range of 5-6 and then decreased in the pH range of 7-9. In previous reports, the catalytic reaction of TMB-H 2 O 2 was much faster in a weakly acidic solution than in neutral or basic solutions. However, in our study, the SiO 2 @Au@Ag alloy NPs lost 95% of their maximum activity at pH 3.0 and retained~70% of their maximum activity in the pH range of 7.0-9.0. This is because SiO 2 @Au@Ag alloy NPs catalyze the generation of ·OH from the decomposition of H 2 O 2 with the dissolution of Ag to release Ag + under strongly acidic conditions, as shown in the following equation: The enhancement of the H 2 O 2 -reducing ability at high pH led to an increase in the SERS signal of oxTMB. However, under basic conditions, the Ag layer on the surface of SiO 2 @Au@Ag alloy NPs can be converted to Ag(OH) or Ag 2 O, which lowers the signal enhancement of oxTMB on SiO 2 @Au@Ag alloy NPs.
For Raman measurement in the liquid phase using a capillary tube, the concentration of the SERS substrate during Raman measurement strongly affects the SERS signal [26,36,38]. Figure 4d shows the effect of the concentration of SiO 2 @Au@Ag alloy NPs after incubation in the TMB-H 2 O 2 mixture. In the absence of H 2 O 2 , the SERS signal of the SiO 2 @Au@Ag alloy NP suspension decreased slightly with increasing concentration of SiO 2 @Au@Ag alloy NPs for Raman measurements. Moreover, the SERS signals of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture increased slightly when the concentration of SiO 2 @Au@Ag alloy NPs decreased sharply. The SERS signal achieved the highest value at 400 µg/mL. Thus, the optimal condition for H 2 O 2 detection by SiO 2 @Au@Ag alloy NPs in TMB-H 2 O 2 mixture was achieved at 8 mM TMB for 15 min reaction with 20 µg SiO 2 @Au@Ag alloy NPs, and Raman measurement was performed at 400 µg/mL SiO 2 @Au@Ag alloy NPs.

Calibration Curve for Detecting H 2 O 2
At optimal conditions, the SERS spectra of SiO 2 @Au@Ag alloy NPs in the presence of TMB were recorded at various concentrations of H 2 O 2 . Variation in the SERS signal of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture was obtained in concentrations of H 2 O 2 from 0.1 to 120 mM (Figure 5a). The SERS signals at 1191, 1341, 1468, 1563, 1608, and 1628 cm −1 gradually increased when the concentration of H 2 O 2 was lower than 20 mM ( Figure 5b). However, they increased remarkably in the range of 40 to 100 mM H 2 O 2 . This implies that H 2 O 2 efficiently converted TMB to oxTMB and that oxTMB was immobilized on the surface of the SiO 2 @Au@Ag alloy NPs. The SERS peak reached saturation with an increase in H 2 O 2 concentration. This result indicates a complete coverage of oxTMB on the SiO 2 @Au@Ag alloy NP surfaces. However, the SERS band of oxTMB showed a shift at high H 2 O 2 concentration due to the formation of a dimer, a trimer, or the twist of oxTMB on the surface of the SiO 2 @Au@Ag alloy NPs [71]. Therefore, we concluded that the SERS signal of SiO 2 @Au@Ag alloy NPs in H 2 O 2 solution in the presence of TMB was the result of the catalytic activity involved in the conversion of TMB to oxTMB and the adsorption of oxTMB on the SiO 2 @Au@Ag alloy NP surfaces.

Calibration Curve for Detecting H2O2
At optimal conditions, the SERS spectra of SiO2@Au@Ag alloy NPs in the presence of TMB were recorded at various concentrations of H2O2. Variation in the SERS signal of SiO2@Au@Ag alloy NPs in the TMB-H2O2 mixture was obtained in concentrations of H2O2 from 0.1 to 120 mM (Figure 5a). The SERS signals at 1191, 1341, 1468, 1563, 1608, and 1628 cm −1 gradually increased when the concentration of H2O2 was lower than 20 mM ( Figure  5b). However, they increased remarkably in the range of 40 to 100 mM H2O2. This implies that H2O2 efficiently converted TMB to oxTMB and that oxTMB was immobilized on the surface of the SiO2@Au@Ag alloy NPs. The SERS peak reached saturation with an increase in H2O2 concentration. This result indicates a complete coverage of oxTMB on the SiO2@Au@Ag alloy NP surfaces. However, the SERS band of oxTMB showed a shift at high H2O2 concentration due to the formation of a dimer, a trimer, or the twist of oxTMB on the surface of the SiO2@Au@Ag alloy NPs [71]. Therefore, we concluded that the SERS signal of SiO2@Au@Ag alloy NPs in H2O2 solution in the presence of TMB was the result of the catalytic activity involved in the conversion of TMB to oxTMB and the adsorption of oxTMB on the SiO2@Au@Ag alloy NP surfaces. The calibration of H2O2 detection was performed via linear curve fitting in the experimental data points ranging from 40 to 100 mM ( Figure S5). A significant linear relationship of y = 20.04x + 802.17 was found between the SERS signals and H2O2 concentration, where x is the H2O2 concentration and y is the SERS signal at 1468 cm −1 (R 2 = 0.98). The theoretical limit of detection was 33.3 mM, as estimated using the 3s blank criterion. The calibration of H 2 O 2 detection was performed via linear curve fitting in the experimental data points ranging from 40 to 100 mM ( Figure S5). A significant linear relationship of y = 20.04x + 802.17 was found between the SERS signals and H 2 O 2 concentration, where x is the H 2 O 2 concentration and y is the SERS signal at 1468 cm −1 (R 2 = 0.98). The theoretical limit of detection was 33.3 mM, as estimated using the 3s blank criterion.
The effect of long-term storage of SiO 2 @Au@Ag alloy NPs is shown in Figure S6. SiO 2 @Au@Ag alloy NPs (200 µg/mL) were stored at 4 • C for 60 days. The UV-vis spectra of the SiO 2 @Au@Ag alloy NPs were measured at the desired time, and the absorbance at 450 nm was monitored. As shown in Figure S7, the SERS signal was stable for 60 days.

Conclusions
We developed a SERS-based H 2 O 2 detection method using SiO 2 @Au@Ag alloy NPs in the presence of TMB. In this work, we demonstrated that TMB was converted to oxTMB by the SiO 2 @Au@Ag alloy NPs in the presence of H 2 O 2 and that oxTMB was absorbed on the surface of SiO 2 @Au@Ag alloy NPs. We also provide a calibration curve to evaluate H 2 O 2 species in the range of 40 to 100 mM with a limit of detection of 33.3 mM. Moreover, the optimal conditions for H 2 O 2 detection using SiO 2 @Au@Ag alloy NPs include incubating 20 µg/mL SiO 2 @Au@Ag alloy NPs with 0.8 mM TMB for 15 min and measuring the Raman signal at 400 µg/mL of SiO 2 @Au@Ag alloy NPs. Even though the limit of detection of our structure is not low, it acted as both a nanozyme and a SERS substrate for the adsorption of TMB. This result greatly expands its applicability for the detection of other biologically active targets.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/nano11102748/s1, Figure S1. (a) Transmission electron microscopy images and (b) ultravioletvisible extinction spectra of (i) SiO 2 @Au (1 mg/mL) and (ii) SiO 2 @Au@Ag alloy nanoparticles (NPs) (20 µg/mL) synthesized using 2 mg SiO 2 @NH 2 and 300 mM Ag + . Figure S2. Schematic illustration of the catalytic mechanism of SiO 2 @Au@Ag alloy NPs in the TMB-H 2 O 2 mixture. TMB is oxidized to oxTMB by SiO 2 @Au@Ag alloy NPs that act as peroxidase in the presence of H 2 O 2 . Next, oxTMB is converted to TMB 2+ in the acidic condition. Figure S3. (a) Zeta potential of SiO 2 @Au@Ag alloy NPs alone and SiO 2 @Au@Ag alloy NPs in the presence of H 2 O 2 , TMB, and a mixture of TMB and H 2 O 2 . (b) Zeta potential of SiO 2 @Au@Ag alloy NPs in a mixture of TMB and H 2 O 2 before and after the addition of H 2 SO 4 . Figure S4. Surface-enhanced Raman spectroscopy (SERS) signals of SiO 2 @Au@Ag alloy NPs in various pH solutions, with pH ranging from 3.0 to 9.0 in the TMB-H 2 O 2 mixture. Figure