Fabrication of Black Silicon Microneedle Arrays for High Drug Loading

Silicon microneedle (Si-MN) systems are a promising strategy for transdermal drug delivery due to their minimal invasiveness and ease of processing and application. Traditional Si-MN arrays are usually fabricated by using micro-electro-mechanical system (MEMS) processes, which are expensive and not suitable for large-scale manufacturing and applications. In addition, Si-MNs have a smooth surface, making it difficult for them to achieve high-dose drug delivery. Herein, we demonstrate a solid strategy to prepare a novel black silicon microneedle (BSi-MN) patch with ultra-hydrophilic surfaces for high drug loading. The proposed strategy consists of a simple fabrication of plain Si-MNs and a subsequent fabrication of black silicon nanowires. First, plain Si-MNs were prepared via a simple method consisting of laser patterning and alkaline etching. The nanowire structures were then prepared on the surfaces of the plain Si-MNs to form the BSi-MNs through Ag-catalyzed chemical etching. The effects of preparation parameters, including Ag+ and HF concentrations during Ag nanoparticle deposition and [HF/(HF + H2O2)] ratio during Ag-catalyzed chemical etching, on the morphology and properties of the BSi-MNs were investigated in detail. The results show that the final prepared BSi-MN patches exhibit an excellent drug loading capability, more than twice that of plain Si-MN patches with the same area, while maintaining comparable mechanical properties for practical skin piercing applications. Moreover, the BSi-MNs exhibit a certain antimicrobial activity that is expected to prevent bacterial growth and disinfect the affected area when applied to the skin.


Introduction
Transdermal drug delivery (TDD) is an attractive strategy for delivering therapeutic agents due to its ability to avoid the challenges associated with traditional oral administration and hypodermic injections, such as inducing gastric irritation and liver injury, low bioavailability, and poor patient compliance [1,2]. One of the challenges in the context of TDD lies in overcoming the stratum corneum (SC), the outermost protective barrier of the skin. This barrier poses a significant hindrance to the effective delivery of active ingredients and limits the range of drugs that can be efficiently administered through TDD [3]. Various methods have been developed to enhance permeability through the stratum corneum barrier to improve TDD, including the use of chemical enhancers, iontophoresis, microdermabrasion, laser ablation, and electroporation. However, these techniques entail the [HF/(HF + H 2 O 2 )] ratio during Ag-catalyzed chemical etching, on the morphology and properties of BSi-MNs were investigated in detail. Furthermore, the BSi-MN patch was preliminarily tested for drug loading and antibacterial activity.

Materials and Chemicals
In this study, a <100>-oriented Cz-grown 8-inch single-side polished B-doped wafer with a starting thickness of 725 µm (Kunshan Xiaofei Photovoltaic Technology Co., Kunshan, China) was used as the substrate. The chemical reagents used in this study were potassium hydroxide (KOH, Aladdin, Shanghai, China, 99.9%), hydrofluoric acid (HF, Aladdin, Shanghai, China, 49%), silver nitrate (AgNO 3 , Macklin, Shanghai, China, 99%), hydrogen peroxide (H 2 O 2 , Macklin, Shanghai, China, 30%), nitric acid (HNO 3 , CrystalClear, Suzhou, China, 69%), propidium iodide (PI, Absin, Shanghai, China, 50 µg/mL), and 4 ,6-diamidino-2-phenylindole (DAPI, Biogo, Shanghai, China, 10 µg/mL). None of the above reagents required further purification. Deionized water (DIW resistivity of~18 MΩ·cm) and phosphate-buffered solution were used throughout the whole experimental process. Figure 1 shows a schematic diagram of the process flows to fabricate the BSi-MN arrays. The <100>-oriented silicon wafer was first oxidized in an oxidation furnace tube through a wet oxidation process to prepare a protective layer of silicon oxide (SiO x ) with a thickness of 1~1.5 µm on both sides. The protective layer on one side was opened by using a fiber laser (RPP12, DelphLaser, Suzhou, China) to form an orthogonal grid pattern. The detailed laser parameters were as follows: a speed of 100 mm/s, a power output of 25 W, a frequency of 5 KHz, and a duty cycle of 35%. Subsequently, alkaline etching was carried out in a 20 wt% KOH solution at 8 • C to prepare plain silicon MNs. The 2 × 2 cm 2 Si-MN patches were obtained from the 8-inch Si-MN wafer through laser cutting. The Si-MN patches were cleaned using a 5 wt% HF solution to remove the surface oxide layer. Then, these Si-MN patches were immersed in a AgNO 3 /HF aqueous solution at room temperature for 20 s to deposit Ag-NPs on their surfaces. The subsequent formation of BSi-MNs was performed in an HF/H 2 O 2 aqueous solution through a Ag-catalyzed chemical etching process. Finally, the BSi-MN patches were sequentially cleaned in 30 wt% HNO 3 and dilute HF aqueous solutions to remove Ag residues and the oxide layer, and then dried. Throughout the wet process, the patches were thoroughly rinsed with DIW between all steps.

Materials and Chemicals
In this study, a <100>-oriented Cz-grown 8-inch single-side polished B-doped wa with a starting thickness of 725 µm (Kunshan Xiaofei Photovoltaic Technology Co., Ku shan, China) was used as the substrate. The chemical reagents used in this study we potassium hydroxide (KOH, Aladdin, Shanghai, China, 99.9%), hydrofluoric acid (H Aladdin, Shanghai, China, 49%), silver nitrate (AgNO3, Macklin, Shanghai, China, 99% hydrogen peroxide (H2O2, Macklin, Shanghai, China, 30%), nitric acid (HNO3, Cry talClear, Suzhou, China, 69%), propidium iodide (PI, Absin, Shanghai, China, 50 µg/m and 4′,6-diamidino-2-phenylindole (DAPI, Biogo, Shanghai, China, 10 µg/mL). None the above reagents required further purification. Deionized water (DIW resistivity of ~ MΩ·cm) and phosphate-buffered solution were used throughout the whole experimen process. Figure 1 shows a schematic diagram of the process flows to fabricate the BSi-M arrays. The <100>-oriented silicon wafer was first oxidized in an oxidation furnace tu through a wet oxidation process to prepare a protective layer of silicon oxide (SiOx) w a thickness of 1~1.5 µm on both sides. The protective layer on one side was opened using a fiber laser (RPP12, DelphLaser, Suzhou, China) to form an orthogonal grid patte The detailed laser parameters were as follows: a speed of 100 mm/s, a power output of W, a frequency of 5 KHz, and a duty cycle of 35%. Subsequently, alkaline etching w carried out in a 20 wt% KOH solution at 8 °C to prepare plain silicon MNs. The 2 × 2 c Si-MN patches were obtained from the 8-inch Si-MN wafer through laser cutting. The MN patches were cleaned using a 5 wt% HF solution to remove the surface oxide lay Then, these Si-MN patches were immersed in a AgNO3/HF aqueous solution at room te perature for 20 s to deposit Ag-NPs on their surfaces. The subsequent formation of B MNs was performed in an HF/H2O2 aqueous solution through a Ag-catalyzed chemi etching process. Finally, the BSi-MN patches were sequentially cleaned in 30 wt% HN and dilute HF aqueous solutions to remove Ag residues and the oxide layer, and th dried. Throughout the wet process, the patches were thoroughly rinsed with DIW b tween all steps.

Quantification of Drug Loading
To determine the actual drug loading capacity of the BSi-MN patches, weighing experiments were conducted on the Si-MN patches and BSi-MN patches with an area of 2 × 2 cm 2 . Specifically, normal saline was first added to the surfaces of both patches using a syringe until excess, and the patches were then gently vertically lifted to remove excess normal saline. The weight of the MN sample is subtracted from the weight obtained by weighing to obtain its actual loading capacity.

Live/Dead Staining Experiments
In order to determine the antimicrobial capacity of the patches, live-dead bacterial staining using the fluorescent dyes PI and DAPI was performed in this study [43,44]. The standard strain Escherichia coli (E. coli) ATCC 25922 (ATCC, Manassas, VA, USA) was used in this experiment. The strain was incubated in a bacterial incubator at 37 • C until it reached the midlogarithmic phase. The optical density of the bacterial solution was adjusted to OD 600 nm = 0.1. The Si-MN patches, including the plain Si-MN patches and BSi-MN patches, were immersed in 1 mL of the bacterial solution and incubated for 30 min in the dark. After incubation, all the patches were rinsed three times with phosphate-buffered solution and then stained successively with PI and DAPI. The patches were observed under an inverted confocal microscope (LSM 710, Zeiss, Jena, Germany) with a 63× oil objective. All images were taken under the same instrument settings. The fluorescence images were analyzed using ImageJ (version 1.51j8) software. We performed at least three independent experiments in each case.

Characterization
The morphologies of the patches were observed using a field emission scanning electron microscope (FE-SEM, Hitachi, S-4700). A contact angle measuring instrument (Shanghai Huafu Information Technology Co., Ltd., Shanghai, China, A23-605L) was employed to determine the contact angle. The reflectance spectra in the wavelength range of 400-1000 nm were measured using an ultraviolet-visible near-infrared spectrophotometer (Shimadzu, Kyoto, Japan, UV-3600). A precision mechanical tester (Tengba, Shanghai, China, Universal TA) was used to determine the mechanical properties of the MNs by using a probe to vertically compress the MN patch until the needle broke/collapsed. Fluorescence staining tests for live/dead bacteria were performed on Si-MNs and BSi-MNs using a laser scanning confocal microscope (LSCM, Zeiss, LSM 710). The size distributions of Ag-NPs deposited on the surfaces of Si-MNs were analyzed using NanoMeasure1.2. The coverage of Ag-NPs and the porosity of black silicon nanowire structures were calculated using ImageJ (version 1.51j8) software.

Fabrication of Si-MNs
Mask patterning is essential to make Si-MNs adequately reach target shapes. In the fabrication process, laser grooving was used to realize the orthogonal grid pattern opening of the SiO x protective layer on the wafer surface and to form deep kerfs in the wafer. Figure 2a shows the surface morphology of the wafer with the SiO x protective layer after laser grooving. The spacing between any two adjacent grooves is 500 µm. The corresponding enlarged image (Figure 2b) shows that the width of each groove is approximately 50 µm. This indicates that the mask pattern consists of small squares with an area of approximately 450 × 450 µm 2 . The depth of the kerfs directly affects the aspect ratio of the Si-MNs, which can be controlled by adjusting several parameter settings of the infrared fiber laser device, such as output power, scanning speed, duty ratio, and several laser scans. The depth of the laser kerfs in this study is approximately 90 µm, as shown in Figure 2c. After laser grooving, the Si-MN array was then fabricated through anisotropic etching in a 20 wt% KOH solution at 85 • C. In this procedure, the heights of the Si-MNs could be regulated by varying the depth of the laser kerfs and the duration of the alkaline etching. Figure 2d shows the morphology (30 • -tilted SEM view) of the Si-MN array prepared after etching for 2 h. The spacing between two adjacent Si-MNs is also 500 µm, which corresponds to the spacing of the two adjacent laser grooves. A single Si-MN is shown in the enlarged SEM image, as shown in Figure 2e. The Si-MN exhibits good sharpness, which is conducive to penetrating the skin. The cross-sectional SEM image in Figure 2f shows that the Si-MN has an aspect ratio of approximately 2.4. The superior morphology of the Si-MNs serves as an excellent foundation for the subsequent production of BSi-MN arrays.
J. Funct. Biomater. 2023, 14, x FOR PEER REVIEW 5 of 13 regulated by varying the depth of the laser kerfs and the duration of the alkaline etching. Figure 2d shows the morphology (30°-tilted SEM view) of the Si-MN array prepared after etching for 2 h. The spacing between two adjacent Si-MNs is also 500 µm, which corresponds to the spacing of the two adjacent laser grooves. A single Si-MN is shown in the enlarged SEM image, as shown in Figure 2e. The Si-MN exhibits good sharpness, which is conducive to penetrating the skin. The cross-sectional SEM image in Figure 2f shows that the Si-MN has an aspect ratio of approximately 2.4. The superior morphology of the Si-MNs serves as an excellent foundation for the subsequent production of BSi-MN arrays.

Influence of AgNO3/HF Ratio on BSi-MN Arrays
Subsequently, black silicon nanostructures were fabricated on the Si-MN arrays to yield BSi-MN arrays featuring hydrophilic surfaces through Ag-catalyzed chemical etching. To comprehensively understand the formation of the nanostructures on the Si-MN, we first investigated the influence of the deposition behavior of the Ag-NPs on the surface of the Si-MN patch on the formation of the nanostructures. The morphologies of the Ag-NPs deposited on the surface of the Si-MN patch in AgNO3/HF aqueous solution under different Ag + concentrations and the corresponding black silicon nanostructures etched in HF/H2O2 aqueous solution are shown in Figure 3. To facilitate the study, we introduce the parameter where ( ) and ( ) represent the molar concentrations of HF and H2O2 in the HF/H2O2 aqueous solution, respectively. At this stage, we fixed the HF concentration at 0.005 mol/L and the ρ value at 68%. When the Ag + concentration was relatively low (0.005 mol/L), the size of the Ag-NPs was relatively dispersed (Figure 3(a1)), and typical porous black silicon structures appeared on the surface of the Si-MNs ( Figure 3(a2,a3)). The corresponding contact angle of the surface of the BSi-MN array is 105.8° (Figure 3(a4)), indicating that it is still hydrophobic. When increasing the Ag + concentration to 0.075 mol/L, the Ag-NPs grew and became dense (Figure 3(b1)), the black silicon nanostructures changed from porous structures to wispy nanowires (Figure 3(b2,b3)), and the corresponding surface contact angle decreased to 39.3° (Figure 3(b4)). A further increase in the Ag + concentration to 0.01 mol/L resulted in larger and denser Ag-NPs (Figure 3(c1)) and more uniformly arranged nanowires (Figure 3(c2,c3)). The corresponding surface contact angle decreased to 0° (Figure 3(c4)), indicating an ultra-hydrophilic surface. When the Ag + concentration reached 0.015 mol/L, the Ag-NPs continued to become larger and started to where c(HF) and c(H 2 O 2 ) represent the molar concentrations of HF and H 2 O 2 in the HF/H 2 O 2 aqueous solution, respectively. At this stage, we fixed the HF concentration at 0.005 mol/L and the ρ value at 68%. When the Ag + concentration was relatively low (0.005 mol/L), the size of the Ag-NPs was relatively dispersed (Figure 3(a1)), and typical porous black silicon structures appeared on the surface of the Si-MNs (Figure 3(a2,a3)). The corresponding contact angle of the surface of the BSi-MN array is 105.8 • (Figure 3(a4)), indicating that it is still hydrophobic. When increasing the Ag + concentration to 0.075 mol/L, the Ag-NPs grew and became dense (Figure 3(b1)), the black silicon nanostructures changed from porous structures to wispy nanowires (Figure 3(b2,b3)), and the corresponding surface contact angle decreased to 39.3 • (Figure 3(b4)). A further increase in the Ag + concentration to 0.01 mol/L resulted in larger and denser Ag-NPs (Figure 3(c1)) and more uniformly arranged nanowires (Figure 3(c2,c3)). The corresponding surface contact angle decreased to 0 • (Figure 3(c4)), indicating an ultra-hydrophilic surface. When the Ag + concentration reached 0.015 mol/L, the Ag-NPs continued to become larger and started to aggregate (Figure 3(d1)). However, the Si-MNs were severely etched away, destroying the original needle shape (Figure 3(d2,d3)). Although the surface contact angle was also 0 • , the BSi-MNs were not suitable for TDD. aggregate (Figure 3(d1)). However, the Si-MNs were severely etched away, destroyi original needle shape (Figure 3(d2,d3)). Although the surface contact angle was a the BSi-MNs were not suitable for TDD. Previous studies have shown that the HF concentration in AgNO3/HF soluti another important factor affecting the deposition behavior of Ag-NPs on silicon su [45][46][47]. As well as removing the oxide layer, HF can form Si-F bonds on the silico face, which is conducive to the capture of holes and release of electrons [45], thus pr ing the deposition of Ag-NPs. Figure 4a-c shows the morphologies of the Ag-NPs d ited on the surface of the Si-MNs produced by varying the HF concentration AgNO3/HF aqueous solution while maintaining the Ag + concentration at the optim mol/L. The corresponding statistical analysis based on SEM images was obtained by ImageJ (version 1.51j8) software, as shown in Figure 4d-f. The Ag-NPs deposited at concentration of 0.005 mol/L were uniformly distributed, with a size distribution ra from 10 to 150 nm. The corresponding mean value and standard deviation (SD) w nm and 36 nm, respectively, indicating that the particle sizes were mainly distribut tween 37 and 109 nm. With regard to the Ag-NPs deposited at an HF concentrat 0.009 mol/L, the particles started to agglomerate slightly, and the sizes were distri from 50 to 170 nm. The corresponding mean value and SD were 103 nm and 35 n spectively, indicating that the particle sizes were mainly concentrated in the 68-1 range. When increasing the HF concentration to 0.014 mol/L, the particles exhibite nificant agglomeration. In this case, the particle size distribution ranged from 15 nm with a mean value of 94 nm and SD of 43 nm. The coverage rate of Ag-NPs decr with increasing HF concentration due to the agglomeration of particles.
Similarly, the Si-MN samples with the above three types of silver particles etched in the HF/H2O2 aqueous solution with a ρ value of 68% to prepare the BS patches. The surface morphologies (30°-tilted SEM view) of the corresponding BS are shown in Figure 4g-i. As can be seen, the tip and edge of the corresponding S were more severely etched in the HF/H2O2 aqueous solution with increasin Previous studies have shown that the HF concentration in AgNO 3 /HF solutions is another important factor affecting the deposition behavior of Ag-NPs on silicon surfaces [45][46][47]. As well as removing the oxide layer, HF can form Si-F bonds on the silicon surface, which is conducive to the capture of holes and release of electrons [45], thus promoting the deposition of Ag-NPs. Figure 4a-c shows the morphologies of the Ag-NPs deposited on the surface of the Si-MNs produced by varying the HF concentration in the AgNO 3 /HF aqueous solution while maintaining the Ag + concentration at the optimal 0.01 mol/L. The corresponding statistical analysis based on SEM images was obtained by using ImageJ (version 1.51j8) software, as shown in Figure 4d-f. The Ag-NPs deposited at an HF concentration of 0.005 mol/L were uniformly distributed, with a size distribution ranging from 10 to 150 nm. The corresponding mean value and standard deviation (SD) were 73 nm and 36 nm, respectively, indicating that the particle sizes were mainly distributed between 37 and 109 nm. With regard to the Ag-NPs deposited at an HF concentration of 0.009 mol/L, the particles started to agglomerate slightly, and the sizes were distributed from 50 to 170 nm. The corresponding mean value and SD were 103 nm and 35 nm, respectively, indicating that the particle sizes were mainly concentrated in the 68-138 nm range. When increasing the HF concentration to 0.014 mol/L, the particles exhibited significant agglomeration. In this case, the particle size distribution ranged from 15 to 195 nm with a mean value of 94 nm and SD of 43 nm. The coverage rate of Ag-NPs decreased with increasing HF concentration due to the agglomeration of particles.
Similarly, the Si-MN samples with the above three types of silver particles were etched in the HF/H 2 O 2 aqueous solution with a ρ value of 68% to prepare the BSi-MN patches. concentration in the AgNO3/HF aqueous solution. This is because the agglomerat Ag-NPs is more severe at the tip and edge (Figure S1, Supplementary Materials cross-sectional SEM images of the corresponding nanowire structures on the surfa the BSi-MNs are shown in the insets. The nanowire structures prepared by etching HF/H2O2 aqueous solution became longer and looser with increasing HF concentrat the AgNO3/HF aqueous solution. In the interim, it was observed that the water c angles of the aforementioned BSi-MN patches were 0° ( Figure S2, Supplementary M als). Such an observation implies that the surface of BSi-MN evinces a remarkable d of hydrophilicity.

Influence of H2O2/HF Ratio on BSi-MN Arrays
Through the above experiments and analyses, it is found that a suitable Ag + co tration and low HF concentration are conducive to the uniform deposition of Ag which is beneficial to better nanostructures for BSi-MNs. The next objective is to co a further investigation into the characteristics of Ag-catalyzed chemical etching o surface of Si-MNs in the HF/H2O2 system. To do so, Ag-NPs were first deposited o surfaces of the Si-MNs under the same conditions of a Ag + concentration of 0.01 mol/ HF concentration of 0.005 mol/L. The nanowires were then prepared through Ag lyzed chemical etching on the surfaces of Si-MNs in HF/H2O2 aqueous solutions wi ferent ρ values. The top-view and cross-sectional SEM images of the as-prepared owires are shown in Figure 5. When the ρ value was 88%, nanopores rather than owires were observed on the surface, and the nanopores were perpendicular to the face (Figure 5a). When decreasing the ρ value to 68%, distinct micropores appeared surface, corresponding to the "collapse" of some nanowires observed in the cros tional SEM image (Figure 5c). Further decreasing the ρ value to 48%, more and larg cropores appeared on the silicon surface, and the "collapse" phenomenon of nano was even more pronounced (Figure 5c).

Influence of H 2 O 2 /HF Ratio on BSi-MN Arrays
Through the above experiments and analyses, it is found that a suitable Ag + concentration and low HF concentration are conducive to the uniform deposition of Ag-NPs, which is beneficial to better nanostructures for BSi-MNs. The next objective is to conduct a further investigation into the characteristics of Ag-catalyzed chemical etching on the surface of Si-MNs in the HF/H 2 O 2 system. To do so, Ag-NPs were first deposited on the surfaces of the Si-MNs under the same conditions of a Ag + concentration of 0.01 mol/L and HF concentration of 0.005 mol/L. The nanowires were then prepared through Ag-catalyzed chemical etching on the surfaces of Si-MNs in HF/H 2 O 2 aqueous solutions with different ρ values. The top-view and cross-sectional SEM images of the as-prepared nanowires are shown in Figure 5. When the ρ value was 88%, nanopores rather than nanowires were observed on the surface, and the nanopores were perpendicular to the Si surface (Figure 5a). When decreasing the ρ value to 68%, distinct micropores appeared on the surface, corresponding to the "collapse" of some nanowires observed in the cross-sectional SEM image (Figure 5c). Further decreasing the ρ value to 48%, more and larger micropores appeared on the silicon surface, and the "collapse" phenomenon of nanowires was even more pronounced (Figure 5c). occurs only in a relatively small region in contact with the Ag particle. In a low ρ value (H2O2-rich) system, the concentration of H2O2 is relatively high, which can provide more hole injection. Apart from preferentially injecting silicon near the Ag particle through the Ag/Si interface [48], excess holes can diffuse into nearby silicon substrates and surfaces, causing these areas to be etched laterally to form stain layers [49]. The stain layers can dissolve in the HF/H2O2 mixture, resulting in the formation of inverted cone-shaped micropores (Figure 5e).

Properties of BSi-MN Arrays
Finally, the properties of the BSi-MNs prepared at a ρ value of 68% were evaluated and compared with typical Si-MNs. Photographs of the BSi-MN patch and Si-MN patch are shown in Figure S3 (Supplementary Materials). The BSi-MN patch shows a completely black appearance, indicating an extremely low surface reflection due to the strong lighttrapping effect of the nanowire structures. As a comparison, the reflectance curves of the BSi-MN patch and Si-MN patch are shown in Figure 6a, and the corresponding SEM images are shown in the inset. To determine whether the black silicon nanowire structures could potentially compromise the mechanical strength of the MNs, compression experiments were conducted on both plain Si-MNs and BSi-MNs. It was found that the mechanical properties of the BSi-MNs exhibited similar mechanical strength to that of the plain Si-MNs (Figure 6b), and no fracture was observed during the test. The results indicate that the black silicon nanowire structures have no significant effect on the mechanical strength of Si-MNs. Figure 6c presents the water contact angles of the plain Si-MN patch and BSi-MN patch. The water contact angle of the Si-MN patch is approximately 82.82°, while that of the BSi-MN patch is 0°. This indicates that the BSi-MN patch has an ultra-hydrophilic Based on the aforementioned results, we proposed a mechanism for the Ag-catalyzed chemical etching of silicon in the HF/H 2 O 2 system. It is well known that H 2 O 2 functions as an oxidizing agent and preferentially injects holes into silicon through the Ag/Si interface. In a high ρ value (HF-rich) system, the concentration of H 2 O 2 is relatively low, and the holes are only preferentially injected into the silicon at the bottom of the Ag-NPs. The silicon oxide at the bottom of the Ag particle is then etched by HF, and this process is repeated over and over again, which causes the Ag particle to move forward in the silicon (Figure 5d). In this process, the diameter of the etching path (pore) is comparable to the size of the Ag particle, which is due to the fact that the Ag-catalyzed chemical etching occurs only in a relatively small region in contact with the Ag particle. In a low ρ value (H 2 O 2 -rich) system, the concentration of H 2 O 2 is relatively high, which can provide more hole injection. Apart from preferentially injecting silicon near the Ag particle through the Ag/Si interface [48], excess holes can diffuse into nearby silicon substrates and surfaces, causing these areas to be etched laterally to form stain layers [49]. The stain layers can dissolve in the HF/H 2 O 2 mixture, resulting in the formation of inverted cone-shaped micropores (Figure 5e).

Properties of BSi-MN Arrays
Finally, the properties of the BSi-MNs prepared at a ρ value of 68% were evaluated and compared with typical Si-MNs. Photographs of the BSi-MN patch and Si-MN patch are shown in Figure S3  subtracted from the weight obtained by weighing to obtain its actual loading The weights of the two types of MN patches loaded with normal saline are sho ure 6d. The drug loading of the BSi-MN patch is twice that of the plain Si-MN the same area. To ensure the accuracy of weighing, each sample was weighed a times, and the average value of the final result was taken.

Preliminary Antibacterial Experiments on BSi-MN Arrays with E.coli
In addition, we evaluated the antibacterial activity of the BSi-MNs against G ative E. coli using plain Si-MNs as the control group. To quantitatively compar microbial performance of the two types of MNs, E. coli was cultured separately patches. After staining with live/dead cells, fluorescence images were obtained ser scanning confocal microscopy, as shown in Figure 6e,f. The first channel channel, which corresponds to all bacteria (both live and dead) on the sample. T channel is the red channel, which corresponds to dead bacteria. For the BSi-M more red fluorescence was observed compared to the reference Si-MN sample, that more bacteria were killed. This suggests that the nanowire structures on t  To determine the actual drug loading capacity of the BSi-MN patch, weighing experiments were conducted on the Si-MN patches and BSi-MN patches with an area of 2 × 2 cm 2 . Specifically, normal saline was first added to the surfaces of both patches using a syringe until excess, and the patches were then gently vertically lifted to remove excess normal saline. The weight of the MN sample is subtracted from the weight obtained by weighing to obtain its actual loading capacity. The weights of the two types of MN patches loaded with normal saline are shown in Figure 6d. The drug loading of the BSi-MN patch is twice that of the plain Si-MN patch with the same area. To ensure the accuracy of weighing, each sample was weighed at least five times, and the average value of the final result was taken.

Preliminary Antibacterial Experiments on BSi-MN Arrays with E.coli
In addition, we evaluated the antibacterial activity of the BSi-MNs against Gramnegative E. coli using plain Si-MNs as the control group. To quantitatively compare the antimicrobial performance of the two types of MNs, E. coli was cultured separately on the two patches. After staining with live/dead cells, fluorescence images were obtained under laser scanning confocal microscopy, as shown in Figure 6e,f. The first channel is the blue channel, which corresponds to all bacteria (both live and dead) on the sample. The second channel is the red channel, which corresponds to dead bacteria. For the BSi-MN sample, more red fluorescence was observed compared to the reference Si-MN sample, indicating that more bacteria were killed. This suggests that the nanowire structures on the surface of the BSi-MNs can enhance the antibacterial activity, which may be due to the stretching effect of the nanowire structure on bacterial cell membranes. The stretching of the cell wall adsorbed on the nanowire structures is due to the interaction between the elastic deformation of the cell wall and the intrinsic attraction of the cell wall to the surface. The mechanical bactericidal effect on the surfaces of the BSi-MNs is associated with the stretching of the cell membrane beyond their elastic limit. The cell membrane ruptures in the suspension region between the nanowires when the bacterial cells are adsorbed onto the nanowires [50,51]. The high-aspect ratio nanostructures possess a unique capability of storing elastic energy through their flexibility. When these nanostructures come into contact with bacterial cells, the elastic energy that was previously accumulated in the nanowires is discharged, causing the nanowires to bend and physically perturb the cell membrane by stretching it, ultimately resulting in cell death [52,53]. It is conceivable that the same mechanism of action would occur when applied to the skin to prevent bacteria from growing in the pierced skin area. The results indicate that the BSi-MNs also have potential in antibacterial applications.

Conclusions
In summary, we propose a method to prepare a novel BSi-MN array with ultrahydrophilic surfaces for high drug loading by using simple laser patterning, standard alkaline etching, and well-established Ag-catalyzed chemical etching. The preparation parameters of black silicon nanostructures on plain Si-MNs were investigated in detail. The experimental results show that a suitable Ag + concentration, relatively low HF concentration, and moderate ρ value are suitable for the preparation of ideal BSi-MNs. The mechanical strength of the as-prepared BSi-MNs is close to that of plain Si-MNs, which is beneficial for practical skin piercing applications. Moreover, the BSi-MNs exhibit excellent drug-loading capability and certain antimicrobial activity. Taken together, our fabrication method for BSi-MNs is simple and can facilitate manufacturing in a large area and on a large scale, which has great application potential in the TDD field.

Conflicts of Interest:
The authors declare no conflict of interest.