Seed-Mediated Synthesis of Thin Gold Nanoplates with Tunable Edge Lengths and Optical Properties

Thin Au nanoplates show intriguing localized surface plasmon resonance (LSPR) properties with potential applications in various fields. The conventional synthesis of Au nanoplates usually involves the formation of spherical nanoparticles or produces nanoplates with large thicknesses. Herein, we demonstrate a synthesis of uniform thin Au nanoplates by using Au–Ag alloy nanoframes obtained by the galvanic replacement of Ag nanoplates with HAuCl4 as the seeds and a sulfite (SO32−) as a ligand. The SO32− ligand not only complexes with the Au salt for the controlled reduction kinetics but also strongly adsorbs on Au {111} facets for effectively constraining the crystal growth on both basal sides of the Au nanoplates for controlled shape and reduced thicknesses. This seed-mediated synthesis affords Au nanoplates with a thickness of only 7.5 nm, although the thickness increases with the edge length. The edge length can be customizable in a range of 48–167 nm, leading to tunable LSPR bands in the range of 600–1000 nm. These thin Au nanoplates are applicable not only to surface-enhanced Raman spectroscopy with enhanced sensitivity and reliability but also to a broader range of LSPR-based applications.

Thin Au nanoplates are important types of plasmonic nanocrystal. Compared with Ag, Au nanoplates show superior resistance to chemical oxidation. The anisotropic shapes of Au nanoplates shift the LSPR to long wavelengths in the visible and NIR regions, well apart from the interband electron-transition region (<450 nm), enabling excellent Ag-like LSPR properties [4,6,9,30]. The synthesis of Au nanoplates dates back to two decades ago, when Liz-Marzán et al. reported the first plate-like Au nanostructure obtained by a chemical-reduction method [31]. Unfortunately, these seedless syntheses usually produce Au nanoplates with many side products, including nanospheres and

Materials and Methods
Synthesis of Ag nanoplates. The Ag nanoplates were synthesized by using a chemicalreduction method [50]. In a typical synthesis, 0.2 mL of AgNO 3  Synthesis of Au-Ag-alloy nanoframes. The Au-Ag-alloy nanoframes were synthesized by the galvanic replacement reaction of Ag nanoplates with HAuCl 4 . In a typical synthesis, the Ag nanoplates were recovered by centrifugation from 20 mL of the stock solution and redispersed in 3 mL of H 2 O. Next, 2.0 mL of HAuCl 4 (0.1 mM) was slowly added to the solution of the Ag nanoplates using a syringe pump at a rate of 5 mL h −1 under vigorous stirring. The resulting Au-Ag-alloy nanoframes were collected by centrifugation and washed with H 2 O.
Preparation of growth solution of Au. A sulfite-coordinated Au precursor, i.e., Na 3 Au(SO 3 ) 2 , was prepared as the growth solution by following our previously reported protocol [30,46]. Typically, 40 µL of HAuCl 4 (0.25 M), 240 µL of NaOH (0.2 M), and 3.00 mL of Na 2 SO 3 (0.01 M) were dissolved in 4.72 mL of H 2 O. The solution was left undisturbed overnight before use.
Synthesis of Au nanoplates. The pre-synthesized Au nanoframes were suspended in 2 mL of H 2 O. To this solution were added 6.72 mL H 2 O, 500 µL of polyvinylpyrrolidone (PVP, 5 wt%, Mw 10000), 60 µL L-ascorbic acid (AA) (0.1 M), 120 µL NaOH (0.1 M), and 0.6 mL of the growth solution of Au. The reaction was then left undisturbed at 30 • C for 12 h. Finally, Au nanoplates (edge length,~48 nm) were collected by centrifugation, washed with H 2 O, and redispersed in 2 mL of H 2 O.
Synthesis of Au nanoplates with different edge lengths. The Ag nanoplates of various sizes were synthesized by a previously reported seed-mediated growth method [50]. The Au-Ag-alloy nanoframes of different sizes were synthesized with these Ag nanoplates. These Au-Ag-alloy nanoframes were then used as the seeds for the synthesis of Au nanoplates following a similar procedure to that described above.
SERS analysis with the Au-nanoplate substrate in oxidative environments. Typically, 25 µL of Au nanoplates (~48 nm) were dried on a clean silicon wafer (8 mm × 8 mm). Next, a specific amount of an aqueous crystal violet solution (10 −6 M, with additional 2 mM Fe(NO 3 ) 3 or 10 mM H 2 O 2 when oxidative species were involved) was dropped and dried on the substrate. The substrate was then washed with H 2 O to remove free-crystal-violet molecules. Raman spectra were then recorded from the substrate with a 633-nanometer He-Ne laser line at room temperature. For all measurements, the laser power was 3 mW and the signal-acquisition time was fixed at 10 s.
Characterizations. Transmission electron microscopy (TEM) was performed with a Hitachi HT-7700 electron microscope equipped with a tungsten filament at an accelerating voltage of 120 kV. High-resolution TEM (HRTEM) and energy-dispersive X-ray spectroscopy (EDS) was performed on a Philips Tecnai F20 FEG-TEM at an accelerating voltage of 200 kV. The UV-Vis spectra were measured on an Ocean Optics HR2000+ES spectrophotometer. Atomic-force microscopy (AFM) was conducted on Cypher through the tapping mode. Elemental analysis was conducted by inductively coupled plasma mass spectrometry (ICP-MS) on an Agilent 7500CE. Raman spectra were collected with a LabRAM HR800 confocal Raman spectrophotometer equipped with a 633-nanometer He-Ne laser.

Results and Discussion
The seed-mediated synthesis of Au nanoplates requires pre-synthesized seeds with multiple planar defects [33,36]. Such planar defects can be found in Ag nanoplates, which are obtainable in a high yield by a well-established synthesis protocol [23]. When these Ag nanoplates were directly used as the seeds, the resulting Au nanoplates were essentially core-shell nanostructures [30] or nanoframes if a galvanic replacement reaction was involved [51]. In order to achieve the well-controlled synthesis of Au nanoplates with minimal Ag component, these Ag nanoplates (edge length,~45 nm; thickness,~6.8 nm; TEM, Figure S1) were converted into Au-Ag-alloy nanoframes through a galvanic replacement reaction with HAuCl 4 prior to the seed-mediated synthesis (Figure 1a; TEM, Figure S2; EDS mapping, Figure S3). In this reaction, Au preferentially grew on the edges of the Ag nanoplates, accompanying an oxidative etching of the Ag cores, leading to a hollow nanostructure. These nanoframes retained the original planar defects in the Ag nanoplates, as evidenced by the formation of many anisotropic nanostructures through the crystal growth on these seeds [44,45,52]. The planar defects clearly satisfied the symmetry-breaking requirement for the growth of Au nanoplates. nanoplates, as evidenced by the formation of many anisotropic nanostructures through the crystal growth on these seeds [44,45,52]. The planar defects clearly satisfied the symmetry-breaking requirement for the growth of Au nanoplates. The Au nanoplates were synthesized by controlled crystal growth on the Au-Agalloy nanoframes. A growth solution of Au was first prepared by mixing sodium hydroxide (NaOH), sodium sulfite (Na2SO3), and chloroauric acid (HAuCl4) in water, which was left undisturbed overnight [30,46]. To this growth solution were added the Au-Ag-alloy nanoframes, PVP, L-AA, and NaOH to initiate the seed-mediated growth. After the reaction system was kept at 30 °C for 12 h, the Au nanoplates were collected by centrifugation and washed with water. Figure 1b shows the TEM image of the intermediate collected during the reaction. By comparing it with the TEM image of the Au-Ag-alloy nanoframes (Figure 1a), one can infer that the Au growth started from the nanoframes and proceeded inward to fill the inner cavity of the nanoframes. As a result, each nanoplate possessed a single hole in its center. After the extensive crystal growth, the Au fully filled the inner cavity of the nanoframes, forming Au nanoplates that resembled the shapes of the Au-Ag-alloy nanoframes or the original Ag nanoplates (Figure 1c, S4 and S5). A triangular "watermark"-like pattern with a brighter image contrast was found in each Au nanoplate, corresponding to the Au-Ag-alloy-nanoframe seed (Figure 1d). The brighter image contrast can be attributed to the presence of Ag, which possesses a smaller atomic number than Au, in these positions. By comparing the shapes of the watermark pattern and the final Au nanoplate, one can conclude that the Au not only grew inward to fill the inner cavity of the Au-Ag-alloy nanoframe, but also grew outward to reshape the Au nanoplate into truncated triangular or hexagonal shapes.
The coordination of the HAuCl4 with SO3 2− greatly reduced its reduction potential, which ruled out the destructive etching of the Au-Ag-alloy nanoframes by the Au salt. In addition, the coordination decreased the reduction rate of the Au salt and, thus, the growth rate of the nanocrystals. As a result, Au nanoplates with well-defined shapes were The Au nanoplates were synthesized by controlled crystal growth on the Au-Ag-alloy nanoframes. A growth solution of Au was first prepared by mixing sodium hydroxide (NaOH), sodium sulfite (Na 2 SO 3 ), and chloroauric acid (HAuCl 4 ) in water, which was left undisturbed overnight [30,46]. To this growth solution were added the Au-Ag-alloy nanoframes, PVP, L-AA, and NaOH to initiate the seed-mediated growth. After the reaction system was kept at 30 • C for 12 h, the Au nanoplates were collected by centrifugation and washed with water. Figure 1b shows the TEM image of the intermediate collected during the reaction. By comparing it with the TEM image of the Au-Ag-alloy nanoframes (Figure 1a), one can infer that the Au growth started from the nanoframes and proceeded inward to fill the inner cavity of the nanoframes. As a result, each nanoplate possessed a single hole in its center. After the extensive crystal growth, the Au fully filled the inner cavity of the nanoframes, forming Au nanoplates that resembled the shapes of the Au-Ag-alloy nanoframes or the original Ag nanoplates (Figures 1c, S4 and S5). A triangular "watermark"-like pattern with a brighter image contrast was found in each Au nanoplate, corresponding to the Au-Ag-alloy-nanoframe seed ( Figure 1d). The brighter image contrast can be attributed to the presence of Ag, which possesses a smaller atomic number than Au, in these positions. By comparing the shapes of the watermark pattern and the final Au nanoplate, one can conclude that the Au not only grew inward to fill the inner cavity of the Au-Ag-alloy nanoframe, but also grew outward to reshape the Au nanoplate into truncated triangular or hexagonal shapes.
The coordination of the HAuCl 4 with SO 3 2− greatly reduced its reduction potential, which ruled out the destructive etching of the Au-Ag-alloy nanoframes by the Au salt. In addition, the coordination decreased the reduction rate of the Au salt and, thus, the growth rate of the nanocrystals. As a result, Au nanoplates with well-defined shapes were been synthesized (average edge length,~48 nm). No self-nucleated side products, such as spherical nanoparticles, were formed in this process. The average thickness was estimated to be~8 nm according to the TEM image of the nanoplates standing vertically on the grid ( Figure 1c, inset). The thickness was close to that of the original Ag nanoplates ( Figure S1), suggesting that the Au growth was mainly in the lateral directions. Thus, the Au growth on the basal sides was greatly restrained, which can be attributed to the strong adsorption of SO 3 2− on the basal Au {111} facets, highlighting the superiority of our design of the synthesis to those used in many previous strategies.
The structure of the Au nanoplates was further verified by HRTEM imaging and electron diffraction (Figure 1e,f). The HRTEM image showed clear lattices corresponding to the [111]-zone axis. The formally forbidden 1/3{422} diffractions were observable in the electron-diffraction pattern, which originated from the stacking faults along the <111> direction. These results suggest that the crystal structures of the original Ag nanoplates (i.e., {111} facets and abundant stacking faults) were successfully inherited by the Au nanoplates through the seed-mediated crystal growth.
The seed-mediated growth of the Au nanoplates was monitored by UV-Vis spectroscopy ( Figure 2). The initial reaction solution (i.e., reaction time, 0 min) appeared near-colorless because the LSPR band of the Au-Ag-alloy nanoframes moved to the infrared region of the spectrum. During the growth of the Au on these seeds, a progressive blueshift of the LSPR was observed, corresponding to the continuous decrease in the size of the inner cavity in the nanoframes. Accompanying the shift in the LSPR band position, its intensity underwent a continuous increase. This can be attributed to the volume expansion of the Au-Ag nanoframes during their growth into nanoplates because the extinction cross-section was proportional to the volume of the nanocrystals, according to the Mie theory [53]. After 8 h of crystal growth, the LSPR band of the Au nanoplates became stable in terms of both the band position and the intensity, indicating the end of the crystal growth. The final solution of the Au nanoplates showed a deep blue color, which was consistent with the LSPR band eventually shifting to a wavelength of 685 nm (Figure 2, inset). It is worth noting that only an in-plane dipolar LSPR band was observed in the UV-Vis spectrum. This is because in-plane quadrupolar LSPR becomes obvious only when the Au nanoplates are sufficiently large and the shapes and sizes are highly uniform. Our experiment produced Au nanoplates with a small average edge length of 48 nm. In addition, they showed variations in both shape and edge length. We expect well-resolved quadrupolar LSPR bands from Au nanoplates with even larger sizes and improved uniformity.
been synthesized (average edge length, ~48 nm). No self-nucleated side products, such as spherical nanoparticles, were formed in this process. The average thickness was estimated to be ~8 nm according to the TEM image of the nanoplates standing vertically on the grid (Figure 1c, inset). The thickness was close to that of the original Ag nanoplates ( Figure S1), suggesting that the Au growth was mainly in the lateral directions. Thus, the Au growth on the basal sides was greatly restrained, which can be attributed to the strong adsorption of SO3 2− on the basal Au {111} facets, highlighting the superiority of our design of the synthesis to those used in many previous strategies.
The structure of the Au nanoplates was further verified by HRTEM imaging and electron diffraction (Figure 1e,f). The HRTEM image showed clear lattices corresponding to the [111]-zone axis. The formally forbidden 1/3{422} diffractions were observable in the electron-diffraction pattern, which originated from the stacking faults along the <111> direction. These results suggest that the crystal structures of the original Ag nanoplates (i.e., {111} facets and abundant stacking faults) were successfully inherited by the Au nanoplates through the seed-mediated crystal growth.
The seed-mediated growth of the Au nanoplates was monitored by UV-Vis spectroscopy ( Figure 2). The initial reaction solution (i.e., reaction time, 0 min) appeared nearcolorless because the LSPR band of the Au-Ag-alloy nanoframes moved to the infrared region of the spectrum. During the growth of the Au on these seeds, a progressive blueshift of the LSPR was observed, corresponding to the continuous decrease in the size of the inner cavity in the nanoframes. Accompanying the shift in the LSPR band position, its intensity underwent a continuous increase. This can be attributed to the volume expansion of the Au-Ag nanoframes during their growth into nanoplates because the extinction cross-section was proportional to the volume of the nanocrystals, according to the Mie theory [53]. After 8 h of crystal growth, the LSPR band of the Au nanoplates became stable in terms of both the band position and the intensity, indicating the end of the crystal growth. The final solution of the Au nanoplates showed a deep blue color, which was consistent with the LSPR band eventually shifting to a wavelength of 685 nm (Figure 2, inset). It is worth noting that only an in-plane dipolar LSPR band was observed in the UV-Vis spectrum. This is because in-plane quadrupolar LSPR becomes obvious only when the Au nanoplates are sufficiently large and the shapes and sizes are highly uniform. Our experiment produced Au nanoplates with a small average edge length of 48 nm. In addition, they showed variations in both shape and edge length. We expect well-resolved quadrupolar LSPR bands from Au nanoplates with even larger sizes and improved uniformity. This synthesis is highly adaptable to afford Au nanoplates of different sizes for extended plasmonic properties in a wide range of the spectrum. Because the growth of the Au preferentially filled the inner cavity of the Au-Ag-alloy nanoframes, the final sizes of the Au nanoplates were consistent with those of the Au-Ag-alloy nanoframes. Therefore, the size of the Au nanoplates could be systematically tuned by employing Au-Ag-alloy nanoframes of different sizes, obtainable through the galvanic replacement of Ag nanoplates of specific sizes with HAuCl 4 , as the seeds. We demonstrated the synthesis of Au nanoplates with varying edge lengths in the range of 48 nm to 167 nm (Figure 3; more TEM images and size distributions can be found in Figures S4-S11). As confirmed by the TEM images, all these Au nanoplates were obtained with high morphological yield and size uniformity. Most of the nanoplates showed truncated triangular shapes with sharp edges after the seed-mediated growth. This synthesis is highly adaptable to afford Au nanoplates of different sizes fo tended plasmonic properties in a wide range of the spectrum. Because the growth o Au preferentially filled the inner cavity of the Au-Ag-alloy nanoframes, the final siz the Au nanoplates were consistent with those of the Au-Ag-alloy nanoframes. There the size of the Au nanoplates could be systematically tuned by employing Au-Agnanoframes of different sizes, obtainable through the galvanic replacement of Ag n plates of specific sizes with HAuCl4, as the seeds. We demonstrated the synthesis o nanoplates with varying edge lengths in the range of 48 nm to 167 nm (Figure 3; m TEM images and size distributions can be found in Figures S4-S11). As confirmed b TEM images, all these Au nanoplates were obtained with high morphological yield size uniformity. Most of the nanoplates showed truncated triangular shapes with s edges after the seed-mediated growth. The thicknesses of the Au nanoplates were measured by atomic-force micros ( Figure 4). The Au nanoplates were first deposited on a clean silicon substrate. A tap mode of the AFM was used to derive the thickness profiles of the Au nanoplates. height difference between the Au nanoplate and the silicon substrate was defined a thickness of the Au nanoplate. The thickness profiles were relatively smooth, sugge uniform thicknesses across all the individual nanoplates. The thickness of the Au n plates with an average edge length of 48 nm was 7.5 nm, close to the value measure TEM imaging (~8 nm). When increasing the edge length to 115, 132, and 167 nm, th erage thickness increased to 12.5 nm, 15 nm, and 22.5 nm, respectively. This indicates although the SO3 2− was strongly adsorbed on the Au {111} facets, the growth of the A both basal sides of the nanoplates was fully restrained, especially for those of large that required extensive crystal growth. Nevertheless, the increase in the edge length The thicknesses of the Au nanoplates were measured by atomic-force microscopy ( Figure 4). The Au nanoplates were first deposited on a clean silicon substrate. A tapping mode of the AFM was used to derive the thickness profiles of the Au nanoplates. The height difference between the Au nanoplate and the silicon substrate was defined as the thickness of the Au nanoplate. The thickness profiles were relatively smooth, suggesting uniform thicknesses across all the individual nanoplates. The thickness of the Au nanoplates with an average edge length of 48 nm was 7.5 nm, close to the value measured by TEM imaging (~8 nm). When increasing the edge length to 115, 132, and 167 nm, the average thickness increased to 12.5 nm, 15 nm, and 22.5 nm, respectively. This indicates that although the SO 3 2− was strongly adsorbed on the Au {111} facets, the growth of the Au on both basal sides of the nanoplates was fully restrained, especially for those of large sizes that required extensive crystal growth. Nevertheless, the increase in the edge length was substantially quicker than the increase in the thickness, which highlights the role of SO 3 2− in synthesizing Au nanoplates with small thicknesses and, therefore, high aspect ratios in the efficient shifting of the LSPR to long wavelengths. Nanomaterials 2023, 13, x FOR PEER REVIEW 7 of 1 substantially quicker than the increase in the thickness, which highlights the role of SO3 2 in synthesizing Au nanoplates with small thicknesses and, therefore, high aspect ratios in the efficient shifting of the LSPR to long wavelengths. The optical properties of plasmonic metal nanoplates depend greatly on their aspec ratios, which allowed us to effectively tune the LSPR band of the Au nanoplates by vary ing their edge lengths. Figure 5 shows the UV-Vis spectra of the Au nanoplates with dif ferent edge lengths. The in-plane dipolar LSPR band of the Au nanoplates with an averag edge length of ~48 nm was at 685 nm. When increasing the edge length to 115 nm, 132 nm and 167 nm, the dipolar LSPR band underwent a continuous redshift to 755 nm, 855 nm and 955 nm, respectively. This observation revealed the clear dependence of the LSPR properties of the Au nanoplates on their size. In general, the in-plane dipolar LSPR band underwent a continuous redshift in line with the increasing edge lengths of the Au nano plates, which can be attributed to the increasing aspect ratios of the nanoplates. It is wort noting that a wet-chemical synthesis usually produces noble metal nanoplates with mixed triangular and hexagonal shapes, which was true in our synthesis. In addition, the nano plates showed an edge-length distribution. Both may have affected the LSPR propertie of the Au nanoplates. As a result, the dipolar LSPR bands of the Au nanoplates becam broad, covering a range of wavelengths. Nevertheless, the LSPR bands were well defined and resolved. This suggests that the shape and edge-length uniformity of our material was within a satisfactory threshold. This can be attributed to the well-designed seed-me diated synthesis, which enabled the simultaneous growth of the nanoplates at close rates In addition, in-plane quadrupolar LSPR bands were also observed in the Au nanoplate with large edge lengths. They appeared at ~580 nm of the wavelength as a shoulder of th dipolar LSPR band. The appearance of the quadrupolar LSPR bands also indicated satis factory uniformity in the shapes and sizes of the Au nanoplates obtained from our syn thesis. With varying sizes, the Au nanoplates showed the LSPR band spanning from th The optical properties of plasmonic metal nanoplates depend greatly on their aspect ratios, which allowed us to effectively tune the LSPR band of the Au nanoplates by varying their edge lengths. Figure 5 shows the UV-Vis spectra of the Au nanoplates with different edge lengths. The in-plane dipolar LSPR band of the Au nanoplates with an average edge length of~48 nm was at 685 nm. When increasing the edge length to 115 nm, 132 nm, and 167 nm, the dipolar LSPR band underwent a continuous redshift to 755 nm, 855 nm, and 955 nm, respectively. This observation revealed the clear dependence of the LSPR properties of the Au nanoplates on their size. In general, the in-plane dipolar LSPR bands underwent a continuous redshift in line with the increasing edge lengths of the Au nanoplates, which can be attributed to the increasing aspect ratios of the nanoplates. It is worth noting that a wet-chemical synthesis usually produces noble metal nanoplates with mixed triangular and hexagonal shapes, which was true in our synthesis. In addition, the nanoplates showed an edge-length distribution. Both may have affected the LSPR properties of the Au nanoplates. As a result, the dipolar LSPR bands of the Au nanoplates became broad, covering a range of wavelengths. Nevertheless, the LSPR bands were well defined and resolved. This suggests that the shape and edge-length uniformity of our materials was within a satisfactory threshold. This can be attributed to the well-designed seed-mediated synthesis, which enabled the simultaneous growth of the nanoplates at close rates. In addition, in-plane quadrupolar LSPR bands were also observed in the Au nanoplates with large edge lengths. They appeared at~580 nm of the wavelength as a shoulder of the dipolar LSPR band. The appearance of the quadrupolar LSPR bands also indicated satisfactory uniformity in the shapes and sizes of the Au nanoplates obtained Nanomaterials 2023, 13, 711 8 of 12 from our synthesis. With varying sizes, the Au nanoplates showed the LSPR band spanning from the visible to the near-infrared range of the spectrum. In particular, the LSPR bands of the Au nanoplates clearly covered the tissue optical window, which offers promise for many LSPR-based biological applications [4,9,54]. nanoplates clearly covered the tissue optical window, which offers promise for many LSPR-based biological applications [4,9,54]. We believe that Au nanoplates are excellent substrates for SERS applications, with clear advantages (Figure 6). First, the LSPR bands of the Au nanoplates are in the longwavelength region of the spectrum, which is well apart from the electron interband transitions, leading to a high figure of merit and excellent Ag-like light absorption and scattering for near-field enhancement. Second, the nanoplates possess sharp edges and corners, which may serve as antennas to enable strong local electromagnetic fields [3,30,55]. Third, Au nanoplates are chemically inert and, thus, they can produce reliable Raman signals without interference, with many external chemical species. Therefore, we expected a highly sensitive and reliable SERS performance from the Au nanoplates. To demonstrate this, a substrate was prepared by drying Au nanoplates and a low concentration (10 −6 M) of model organic molecules of interest for detection, i.e., CV, on a silicon wafer. Under laser irradiation with a wavelength of 633 nm, strong Raman signals characteristic of crystal-violet molecules were detected. The enhancement factor was calculated to be ~1.8 × 10 4 (Supporting Information). When oxidative species, such as H2O2 and Fe 3+ , were introduced onto the substrate, no noticeable changes in the Raman signals of the crystal violet were detected. The intensities of the SERS signals were almost identical to those obtained in the absence of the interfering species. No obvious structural change in the Au nanoplates was detected by the TEM imaging, which accounted for the high SERS stability (Figures S12 and S13). These results suggest the high sensitivity and excellent reliability of the SERS detection of molecules of interest when using Au nanoplates as the substrates. We expect these Au nanoplates to be particularly useful in practical SERS applications, in which interfering species are inevitably present. We believe that Au nanoplates are excellent substrates for SERS applications, with clear advantages (Figure 6). First, the LSPR bands of the Au nanoplates are in the longwavelength region of the spectrum, which is well apart from the electron interband transitions, leading to a high figure of merit and excellent Ag-like light absorption and scattering for near-field enhancement. Second, the nanoplates possess sharp edges and corners, which may serve as antennas to enable strong local electromagnetic fields [3,30,55]. Third, Au nanoplates are chemically inert and, thus, they can produce reliable Raman signals without interference, with many external chemical species. Therefore, we expected a highly sensitive and reliable SERS performance from the Au nanoplates. To demonstrate this, a substrate was prepared by drying Au nanoplates and a low concentration (10 −6 M) of model organic molecules of interest for detection, i.e., CV, on a silicon wafer. Under laser irradiation with a wavelength of 633 nm, strong Raman signals characteristic of crystal-violet molecules were detected. The enhancement factor was calculated to be~1.8 × 10 4 (Supporting Information). When oxidative species, such as H 2 O 2 and Fe 3+ , were introduced onto the substrate, no noticeable changes in the Raman signals of the crystal violet were detected. The intensities of the SERS signals were almost identical to those obtained in the absence of the interfering species. No obvious structural change in the Au nanoplates was detected by the TEM imaging, which accounted for the high SERS stability (Figures S12 and S13). These results suggest the high sensitivity and excellent reliability of the SERS detection of molecules of interest when using Au nanoplates as the substrates. We expect these Au nanoplates to be particularly useful in practical SERS applications, in which interfering species are inevitably present.

Conclusions
In summary, we demonstrated a robust method for synthesizing thin Au nanoplate with tunable sizes and excellent surface-plasmon-resonance properties, starting from Au Ag-alloy nanoframe seeds. The sulfite ligand played a pivotal role in this synthesis. First its coordination with the Au(III) salt reduced the reduction potential, leading to the con trolled crystal growth of Au on the Au-Ag-alloy-nanoframe seeds. Second, the sulfite ad sorbed strongly on the basal Au {111} facets, which not only guided the formation of th plate-like crystal shape but also restrained the crystal growth on the basal sides of the Au nanoplates, leading to significantly reduced thicknesses compared with those reported in the literature. The Au nanoplates can be as thin as 7.5 nm, although their thickness may increase with the edge length. The edge length can be customized in the range of 48-16 nm, leading to tunable surface-plasmon-resonance bands in the range of 600-1000 nm Due to their high stability and remarkable optical properties, the Au nanoplates devel oped in this work are particularly useful in many LSPR-based applications, in which in terfering species are usually involved. As a demonstration of this, Au nanoplates showed superior performance in the SERS detection of molecules of interest in the presence of Fe 3 and H2O2. We believe this strategy opens a new route for the synthesis of Au nanoplate with controlled LSPR properties for many biological, analytical, and catalytic applications Supplementary Materials: The following supporting information can be downloaded a www.mdpi.com/xxx/s1. Figure S1: Characterization of Ag nanoplates. (a, b) TEM images of Ag na noplates. (c) Size distribution of Ag nanoplates. (d) Thickness histogram of Ag nanoplates. The av erage size of the Ag nanoplates was measured as ~45 nm. The average thickness was calculated a ~6.8 nm. Figure S2: Low-magnification TEM image of the Au-Ag-alloy nanoframes. Figure S3: ED mapping of the Au-Ag-alloy nanoframe. Figure S4: Low-magnification TEM image of the Au nano plates (edge length, ~48 nm). Figure S5: Size histogram of the Au nanoplates (edge length, ~48 nm Figure S6: Low-magnification TEM image of the Au nanoplates (edge length, ~115 nm). Figure S7 Size histogram of the Au nanoplates (edge length, ~115 nm). Figure S8: Low-magnification TEM image of the Au nanoplates (edge length, ~132 nm). Figure S9: Size histogram of the Au nanoplate (edge length, ~132 nm). Figure S10: Low-magnification TEM image of the Au nanoplates (edg length, ~167 nm). Figure S11: Size distribution of the Au nanoplates (edge length, ~167 nm). Figur S12: TEM image of the Au nanoplates after treatment in H2O2 (10 mM) for 10 h. Figure S13: TEM

Conclusions
In summary, we demonstrated a robust method for synthesizing thin Au nanoplates with tunable sizes and excellent surface-plasmon-resonance properties, starting from Au-Ag-alloy nanoframe seeds. The sulfite ligand played a pivotal role in this synthesis. First, its coordination with the Au(III) salt reduced the reduction potential, leading to the controlled crystal growth of Au on the Au-Ag-alloy-nanoframe seeds. Second, the sulfite adsorbed strongly on the basal Au {111} facets, which not only guided the formation of the plate-like crystal shape but also restrained the crystal growth on the basal sides of the Au nanoplates, leading to significantly reduced thicknesses compared with those reported in the literature. The Au nanoplates can be as thin as 7.5 nm, although their thickness may increase with the edge length. The edge length can be customized in the range of 48-167 nm, leading to tunable surface-plasmon-resonance bands in the range of 600-1000 nm. Due to their high stability and remarkable optical properties, the Au nanoplates developed in this work are particularly useful in many LSPR-based applications, in which interfering species are usually involved. As a demonstration of this, Au nanoplates showed superior performance in the SERS detection of molecules of interest in the presence of Fe 3+ and H 2 O 2 . We believe this strategy opens a new route for the synthesis of Au nanoplates with controlled LSPR properties for many biological, analytical, and catalytic applications.

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