CoO, Cu, and Ag Nanoparticles on Silicon Nanowires with Photocatalytic Activity for the Degradation of Dyes

Abstract

Photocatalytic semiconductors require maintaining stability and pursuing higher efficiencies. The studied systems were silicon nanowires (Si_NWs), silicon nanowires with cobalt oxide nanoparticles (Si_NWs-CoO_NPs), and silicon nanowires with copper nanoparticles (Si_NWs-Cu_NPs). Si_NWs were synthesized by metal-assisted chemical etching (MACE) from silicon wafers keeping the remaining silver nanoparticles for all three sample types. The nanowires were about 23–30 µm in length. CoO_NPs and Cu_NPs were deposited on Si_NWs by the autocatalytic reduction processes (electroless). There were many factors in the process that affect the resulting structures and degradation efficiencies. This work shows the degradation of methyl orange (MO) together with the chemisorption of methylene blue (MB), and rhodamine 6G (Rh6G) by direct illumination with visible radiation. The MO degradation kinetics were in the sequence Si_NWs-Cu_NPs (88.9%) > Si_NWs (85.3%) > Si_NWs-CoO_NPs (49.3%), with the Si_NWs-Cu_NPs having slightly faster kinetics. However, Si_NWs-CoO_NPs have slow degradation kinetics. The chemisorptions of MB and Rh6G were Si_NWs-Cu_NPs (87.2%; 86.88%) > Si_NWs (86%; 87%) > Si_NWs-CoO_NPs (17.3%; 12%), showing dye desorptions together with lower chemisorption capacities. This work shows iridescence in optical microscopy images by the visible light interference caused by the spaces between the nanowire bundles.

1. Introduction

In recent years, soil, air, and water bodies have been the primary concern due to the alarming increment of their contamination. Pollution comes mainly from pesticides, heavy metals, and dyes, causing severe damage to human health and altering the ecological equilibrium. Thus, quantifying these pollutants and their removal is important for protecting human wellness and environmental conservation [1]. Nowadays, the trend is to develop new technologies and materials capable of facing environmental contamination and being at the same time eco-friendly and sustainable.

The textile industry is important for many nations because materials, assembled parts, or final products are generally exported to the United States of America and the European Union. However, the textile industry contributes the most to environmental deterioration due to a large amount of water used, later converted into wastewater with a high content of harmful chemicals to humans and the ecosystems. In this industry, dyes are challenging compounds to remove from sewage because they are present in ppm concentrations and are hardly biodegradable. Azo dyes are some of the priority dyes to remove from wastewater due to their release of chemicals. They have already been banned in the European Union. Therefore, the treatment of colored effluents becomes an issue of environmental concern. The techniques used to treat dyes in wastewater are the Fenton processes [2,3], ozonation [4], adsorption [4,5,6,7,8], membrane technology [4,9,10], electrochemical techniques [10], photocatalysis [11,12,13], and others [14,15,16].

Heterogeneous photocatalysis with nanostructured semiconductors has become an up-and-coming innovative technique due to its unique structures and properties [11]. This is the case with silicon nanowires [12,17,18,19,20,21], a nanostructure proving to be an effective photocatalyst for the oxidation of organic dyes and aromatic molecules, which exhibit higher efficiency if they are decorated with metallic nanoparticles. Here, efficiency is assumed as the conversion rate of radiative energy into breakage of atomic bonds of the pollutant molecules by transforming photons into electron–hole pairs and then into reactive oxygen species. Previous works have demonstrated the efficacy of silicon nanowires decorated with copper nanoparticles in degrading organic dyes, such as methyl orange, and reducing other chemicals, such as chromium, using visible light [12]. This, even considering that silicon in bulk is not a photocatalyst because it does not have the conduction band potential to generate OH radicals or even hydrogen peroxide. Another drawback of this nanostructured silicon with Schottky barriers is its efficiency reduction in consecutive photocatalysis cycles attributed to the increased copper oxide layer around the nanoparticles. Nonetheless, silicon is the most used semiconductor in the electronic industry and is broadly used in photovoltaics. Improving its efficiencies could make photocatalysis a feasible method of water and air treatments.

The cobalt oxides CoO and Co₃O₄ [22,23], among other oxides such as TiO₂, ZnO, or CuO, have been profusely investigated for their catalytic activities in the oxidative degradation of organic pollutants [11,13,24,25,26,27,28]. In addition, their catalytic activities have been analyzed for water splitting [24,25,29,30,31], showing an oxygen evolution reaction and CO₂ reduction, conducting to obtaining species, such as CH₄ [32,33].

Aligned with the above topics, methyl orange (MO) [1,34,35], methylene blue (MB) [15,36,37,38,39], and rhodamine 6G (Rh6G) [16,34,40,41,42,43] are dyes frequently used to evaluate the photocatalytic activity of some semiconductors and other compounds, such as TiO₂, ZnO, or reduced graphene oxide [11,13,26].

This work consists of depositing cobalt to obtain cobalt oxide nanoparticles on the silicon nanowires to cause a reduction of organic dyes using photocatalysis with visible irradiation. The proposal for cobalt oxide nanoparticles instead of copper is that cobalt oxide has been reported as photocatalytic and is already stable when exposed to oxygen from the environment, making it an excellent candidate to be stable in consecutive photocatalysis cycles. The organic dyes studied were methyl orange, methylene blue, and rhodamine 6G in a concentration of 20 ppm. The silicon nanowires were synthesized by metal-assisted etching (MACE), while the cobalt oxide nanoparticles were deposited by the electroless technique on the nanowires.

2. Materials and Methods

2.1. Preparation of the Silicon Wafers

The monocrystalline p-type silicon c-Si wafers (12.5 cm in diameter and (100) preferential crystalline direction) were cut into 2.5 × 3 cm plates with a diamond point cutter. The estimated electrical resistivity for the silicon wafers was about 0.01–0.02 Ω cm. Then, with sandpaper of SiC (average grain size of 77 µm), the silicon plates were polished on one or both sides. A deionized water drop was put before sanding and the polishing was in only one direction. The silicon plates were highly brittle, so care was put into avoiding breaking them or scratching the mirror finishing part in the polishing procedure. Finally, c-Si plates were rinsed with plenty of deionized water to remove most of the free sanding residues (incrusted residual particles).

2.2. Silicon Plate Cleaning

The sanded silicon plate was placed with the mirror finishing side up in a 100 mL plastic beaker. Acetone was added to the beaker until the plate was covered and sonicated for 5 min in the ultrasonic cleaning bath. The plate was rinsed with plenty of deionized water. Subsequently, ethanol was added to another glass beaker and sonicated for 5 min, followed by rinsing. Deionized water was added to another glass beaker until the plate was covered and sonicated for 10 min in the ultrasonic cleaner. Finally, the plate was left to dry at room temperature or dried with a hairdryer.

2.3. Synthesis of Si_NWs by the MACE Method

The metal-assisted chemical etching (MACE) method was used to generate silicon nanowires (Figure 1) [44]. Solutions of 10% HF, Ag (5M HF/0.035 M AgNO₃), and OX (14.1 M HF/1.9 M H₂O₂) were prepared. The recirculating bath was set at 58 °C and the OX solution was heated for 1 h. After heating the OX solution, the cleaned silicon plates were immersed in the 10% HF solution for 20 s. Immediately after, the silicon plates were immersed in the Ag solution for 10 s. Then, they were rinsed with ultra-pure water (18 MΩ٠cm, Millipore Milli-Q) to remove residues of the previous treatments. Finally, the silicon plates were immersed in the OX solution for 5 min. After that time, they were rinsed with ultra-pure water avoiding the breakage of the nanowires.

Si_NWs were not further treated with a nitric acid solution to remove the Ag nanoparticles located at the bottom of the nanowires. All the samples of Si_NWs, Si_NWs-CoO_NPs, and Si_NWs-Cu_NPs used in this experimentation have Ag nanoparticles at the bottom in a similar size and number.

2.4. CoO_NPs Deposit on Si_NWs (Si_NWs-CoO_NPs)

Figure 2a represents the stage of the Si_NWs-CoO_NPs sample preparation. The preparation of the electroless bath consisted in mixing CoCl₂ (0.1M)/HCHO (0.2M)/NaOH (9M) [45]. The plate with Si_NWs was immersed in the electroless bath for 5 s and rinsed with plenty of ultrapure water to remove possible residues on the plate. Subsequently, it was left to dry at room temperature or dried with a hairdryer and stored in a Petri dish without water.

2.5. Synthesis of Si_NWs with Copper Electroless (Si_NWs-Cu_NPs)

The absorption spectra of the solutions were obtained by a Shimadzu UV-2600 spectrophotometer with a step size of 0.5 nm.

The electroless copper deposits were made with a 0.1 M HF/CuSO₄ solution (Figure 2b). The plate with Si_NWs was immersed for 5 s in the Cu solution. Then, it was rinsed with plenty of ultra-pure water. Finally, the Si_NWs-Cu_NPs plate was stored in a Petri dish with deionized water to reduce copper oxidation.

2.6. Photocatalysis with MO, MB, Rh6G

Figure 3 represents the stages of photocatalysis experiments. Dye solutions of MO, MB, and Rh6G were prepared at 20 ppm. Then, 26 mL of the dye solution were measured and deposited in a glass crystallizer. The pH of the solutions was measured. The Si_NWs-Cu_NPs contributed to pH reduction, with the pH of the MO solution passing from 6.2 to 4.3 on average. The glass crystallizer was placed inside the photocatalysis box, and the previously synthesized plate (Si_NWs, Si_NWs-CoO_NPs, or Si_NWs-Cu_NPs) was inserted.

An initial sample (1.5 mL) of the dye solution was taken. Then, the photocatalysis box was covered and the lamp was turned on. Aliquots of 1.5 mL were taken from the dye solutions at 5, 15, 30, 90, 120, and 150 min. The photocatalytic box had visible LED lights installed within, with a total power of 18 W/m². Each time the sample was taken, the LEDs were turned off and the photocatalysis box was open. The box was covered again after the samples were taken out, and the LEDs were turned on again. After 150 min, the plate was removed from the solution and rinsed with plenty of deionized water. Finally, the plate was stored in dry Petri dishes.

2.7. Specifications of the Characterization Techniques and Experimental Setup

After being obtained, the plates with Si_NWs, Si_NWs-Co_NPs, or Si_NWs-Cu, were segmented with a diamond tip cutter. The plates were placed transversely with a metal holder to obtain images of the nanowires with the KEYENCE VHX-5000x digital microscope, using the compound lens objective VH-Z500R/Z500T, which has a zoom range of 500×–5000×.

A JEOL, model JSM-6510 LV, scanning electron microscope (SEM) was used, coupled to a Bruker, model Quantax 200, energy dispersive X-ray spectrometer (EDS). Plates were placed transversely with a metal holder and adhered with conductive carbon tape. Secondary electron images, backscattered electrons, punctual EDS analysis, and element mapping were obtained.

A Bruker AXS diffractometer, model D8Advance, was used with a copper X-ray source, using the CuKα1 spectral line with a wavelength of 1.5406 Å (8.0 keV). The measurements were performed in the detector/X-ray tube continuous coupled-mode with a step interval of 0.04 steps/° 2θ, a sweep speed of 2.0 s/step, a voltage of 40 kV, a current of 40 mA, and various intervals in the range of 10–120° 2θ, were used.

The Si_NWs, Si_NWs-Co_NPs, and Si_NWs-Cu samples were used in photocatalysis tests with the different dye solutions. The photocatalysis cell was a black PVC box with 12 V (20 W/m² luminous intensity) white LEDs connected in series (λ > 450 nm, 960 ± 10 lx) [12]. The illuminated area of the samples was about 5.5 cm² at a distance of about 8 cm. Samples of the solutions were taken during the tests at the following times 0, 5, 15, 30, 90, 120, and 150 min. A Shimadzu, model UV-2600, UV-Vis spectrometer was used to obtain the dye degradation spectra.

X-ray photoelectron spectroscopy (XPS) was conducted using Thermo Scientific brand equipment, model K-Alpha™ +. The pressure in the chamber was about 10-9 mbar, with a monochromatic X-ray source of Al Kα (1.4866 keV). The spot size of the analyses was about 400 μm. A step size of 20.0 eV and a total of 10 scans were used. The Advantage^® software was used, fitting with the Gaussian–Lorentzian function and Shirley’s type background correlation, referenced to the C1s bond at 284.8 eV (NIST Standard Reference Database 20, version 4.1). The high-resolution spectra were taken in the ranges of C1s, O1s, Si2p, Ag3d, and Co2p.

2.8. Reactions of the MACE Anisotropic Wet Etching Process for Si_NWs Formation and Nanometric Metallic Decoration

Some etching mechanisms have been proposed to explain the MACE process. A reduction reaction for the cathodic process implies the decomposition of H₂O₂ at the metallic surface in acidic conditions [46,47,48,49,50].

$$H_2{O}_2+2H^{+}\to 2H_{2}O+2h^{+} \mathrm{or} \frac{n}{2}{H}_2{O}_2+n{H}^{+}\to n{H}_{2}O+n{h}^{+}$$

(1)

$$2{\mathrm{H}}^{+}\to {\mathrm{H}}_2\uparrow +2{\mathrm{h}}^{+}$$

(2)

The electronegativity of Ag is higher than Si, therefore, the metal ions Ag⁺ can withdraw electrons from silicon atoms, or equivalently the holes are generated into the silicon, taking place the deposition of metallic silver as nanoparticles on the silicon surface [49,51]:

$$A{g}^{+}+e_{VB}^{-}\to A{g}^0$$

(3)

$$S{i}^0+2H_{2}O\to Si{O}_2+4H^{+}+4e_{VB}^{-}$$

(4)

$$Si{O}_2+6HF\to H_{2}Si{F}_6+2H_{2}O$$

(5)

The deposited silver into the silicon surface can promote the formation of pores. It has been proposed that pore formation is caused mainly by the decomposition of nitrate ions from AgNO₃, which acts as an oxidant and metal source. Besides this, another proposed reaction includes the generation of holes due to the silver ions reduction [52]:

$$N{O}_3^{-}+3H^{+}\to HN{O}_2+H_{2}O+2h^{+}$$

(6)

$$N{O}_3^{-}+4H^{+}\to NO+2H_{2}O+3h^{+}$$

(7)

$$A{g}^{+}\to A{g}^0+h^{+}$$

(8)

The next step is the etching of the Si layer. In this process, the silicon is oxidized to SiO₂ and $Si{F}_6^{2-}$. The holes h⁺ generated are consumed by the silicon substrate and the anodic process occurring under the metallic deposited silver could be described as follows [46,47,50,52]:

Silicon dissolution model 1:

$$S{i}^0+4h^{+}+4HF\to Si{F}_4+4H^{+}$$

(9)

$$Si{F}_4+2HF\to H_{2}Si{F}_6$$

(10)

Silicon dissolution model 2:

$$S{i}^0+4H{F}_2^{-}\to Si{F}_6^{2-}+2HF+{\mathrm{H}}_2\uparrow +2e^{-}$$

(11)

Silicon dissolution model 3:

$$S{i}^0+n{h}^{+}+6HF\to H_{2}Si{F}_6+n{H}^{+}+\frac{4-n}{2}{H}_2$$

(12)

$$S{i}^0+\frac{n}{2}{H}_2{O}_2+6HF\to n{H}_{2}Si{F}_6+H_{2}O+\frac{4-n}{2}{H}_2$$

(13)

Silicon dissolution model 4:

$$S{i}^0+2H_{2}O+n{h}^{+}\to Si{O}_2+4H^{+}+\left(4-n\right)e^{-}$$

(14)

$$Si{O}_2+6F^{-}+4H^{+}\to Si{F}_6^{2-}+2H_{2}O$$

(15)

If the number n of holes equals 4, the main process will involve the production of SiO₂ followed by its etching and the absence of H₂. In addition, if n = 2, there is a ratio of 1:1 between the etched silicon and one molecule of H₂ formed. While the former mechanism is associated with the silicon/metal interface producing a straight etched profile, the latter can occur outside the interface and allows obtaining nanoporous silicon [47].

On the other hand, the proposed reactions for the electroless copper deposit on silicon nanowires are as follows [53]:

Copper reduction model 1:

$$Si+hv\to Si+e^{-}+h^{+}$$

(16)

$$C{u}^{2+}+2e^{-}\to C{u}^0\hspace{1em}{\mathrm{E}}^0=+0.340 \mathrm{V}$$

(17)

Copper reduction model 2:

$$Si+2C{u}^{2+}+2H_{2}O\to Si{O}_2+2Cu+4H^{+}$$

(18)

$$Si{O}_2+6HF\to H_{2}Si{F}_6+2H_{2}O$$

(19)

Furthermore, the proposed reactions for the electroless cobalt deposit and its oxidation CoO on silicon nanowires are as follows [22,23]:

$$C{o}^{2+}+2e^{-}\to C{o}^0\hspace{1em}{\mathrm{E}}^0=-0.280 \mathrm{V} (\mathrm{vs}. \mathrm{SHE})$$

(20)

$$Si+2C{o}^{2+}+2H_{2}O\to Si{O}_2+2Co+4H^{+}$$

(21)

$$2C{o}^0+O_2\to 2CoO$$

(22)

$$Si{O}_2+6HF\to H_{2}Si{F}_6+2H_{2}O$$

(23)

3. Results and Discussion

3.1. Characterization of the Si_NWs-Cu_NPs and Si_NWs-CoO_NPs

The topography and distribution of the synthesized nanowires were observed (Figure 4). In these visualizations, the central part of the silicon plates had a more homogeneous distribution of nanowires on the silicon plate. So, the micrographs of the differently prepared samples were taken in these areas. The explanation for observing the structure using optical or electronic microscopies is that those images show bundles or conglomerates of nanowires and only such groups are visible. The diameter of each nanowire only has slight fluctuations depending on the synthesis procedure (about 70–90 nm), different from their length which depends on many process variables (about 10–80 μm). Therefore, the MACE method indeed generates silicon nanowires.

Figure 4 shows the micrographs of Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs. Figure 4a,b show the micrographs of a silicon nanowire plate (Si_NWs). In Figure 4a, the Si_NWs plate is seen in its cross-section, having an approximate length of 42.7 µm. Nanowires formed vertically on the silicon plate, which were abundant and not so far apart. Figure 4b shows the top view of the Si_NWs plate. This micrograph shows the homogeneous distribution of the nanowires and the agglomerations that occur between them, forming small bundles. Figure 4c,d show the micrographs of a silicon nanowire plate decorated with copper nanoparticles (Si_NWs-Cu_NPs). Figure 4c shows the cross-sectional view of the nanowires with a vertical orientation on the surface of the silicon plate and a length of about 44.1 µm.

Figure 4d shows the surface view of the Si_NWs-Cu_NPs plate. There is a homogeneous distribution of nanowires and the islands formed by their agglomeration. Moreover, there is a color change on the surface due to the deposition of the copper nanoparticles. Figure 4e,f show the micrographs of a Si_NWs plate decorated with cobalt nanoparticles (Si_NWs-CoO_NPs). Figure 4e shows the cross-sectional view of this Si_NWs -CoO_NPs plate. It should be noted that the formation of these nanowires occurred differently from the Si_NWs shown in Figure 4a. In this case, the nanowires were formed vertically on the silicon surface but in a highly spaced distribution, having a length of about 23.48 µm. Figure 4f shows the superficial view, which shows the spacing between the nanowires (the bright points or areas correspond to nanowires). The micrograph has different color compared with that of the Si_NWs and this is caused by the cobalt nanoparticles deposited on the nanowire surfaces.

In the optical images, there is the phenomenon of interference, observed as iridescent colors, because of the size of the filaments resulting from the MACE method and the space between them, which, as far as we know, this work is the first or between the few that show interference of visible light in the nanostructured bundles (Figure 4a,c,e,f), when observing them by Digital Optic Microscopy using 1000×–5000×. The rainbow-colored images resulted from the visible light of the microscope, causing interference when passing across the groups and between the nanowires according to the interference Equations (1) and (2) [54], maximum and minimum, respectively.

$$\mathrm{d}=\frac{\left( 2 \mathrm{m} +1 \right)\mathsf{\lambda } }{4 \mathrm{n}}$$

(24)

$$\mathrm{d}=\frac{\mathrm{m} \mathsf{\lambda }}{2 \mathrm{n}}$$

(25)

where, d is the optical thickness, n is the refractive index, λ is the wavelength, and m is an integer number (order of interference).

The silver, copper, and cobalt metallic particles were grown by the electroless method, which implies no use of an imposed external potential. Even more, no activation agents were used to seed the silicon surface and grow the metallic crystals. Therefore, the metallic ions were reduced both by the HF-treated silicon [17] and by electrons on the silicon surface, which were promoted from the valence band when absorbing light since the silicon had p-type doping and had no electrons in excess. The electroless method has a limited capability of increasing the mass of deposited metals beyond the nanometric scale (1–100 nm) because once the nucleation sites were covered, the potentials on these areas change and the active sites for further reductions diminish. Usually, the galvanic replacement reaction is followed by one or two electrolytic depositions to thicken the coatings. Consequently, the size of electroless grown metallic particles in about 10 s for silver, 5 s for cobalt, and 5 s for copper were nanometric in size [12,17,55].

Using scanning electron microscopy (SEM), the formation and structure of the synthesized silicon nanowires were observed in more detail. Moreover, the samples were analyzed using energy dispersive X-ray spectroscopy (EDS) to obtain the elemental mapping.

Figure 5 shows the SEM micrographs with a 10.00 µm scale bar and elemental point analysis on the cross-sectional area by EDS. Moreover, the figure shows the elemental mapping over the cross-sectional of the Si_NWs-CoO_NPs sample. Figure 5a shows the SEM micrograph of the Si_NWs sample at 1500×. In this secondary electron image, the length of the nanowires of about 32 µm is shown. Moreover, this image shows how the nanowires were agglomerated, forming small bundles. Some nanowires were broken, and some can have flections during the process of silicon etching. The concentration of H₂O₂ directly influences the morphology of the nanowires. On the other hand, Figure 5b, a micrograph made with backscattered electrons, gives a change in contrasts that allows observing the silver deposited at the base of the nanowires (bright white areas and dots). In some cases, the etching with HF, used to form nanowires, resulted in the silver nanoparticle agglomeration or dendritic forms. Part of the photocatalysis was attributed to these silver structures at the bottom of the nanowires. The silver–silicon junction, a Schottky barrier, is proposed as a higher efficient electron transferor when shaped as dendrites because the sharp shape has a lower work function, which means that it is easier to extract electrons from the pick than from a flat or round shape. In some other works, the silver or metal used to shape the structures is washed away with a nitric acid solution. Nonetheless, we agree with the works that propose that such a Schottky barrier contributes to photocatalysis.

EDS analysis was performed on the cross-sectional area of the Si_NWs sample on the yellow rectangle of Figure 5a. The elements present were Si, C, O, and Ag in percentages of 80.14%, 12.71%, 3.96%, and 3.19%, respectively. C and O presence in the sample was attributed to the contamination when in contact with the atmosphere. Moreover, the oxygen present was related to the formation of oxides, in this case, with the Si. The samples of Si_NWs-Cu_NPs were also characterized by SEM at 2000× and the micrographs are shown in Figure 5d,e. Figure 5d is an image of secondary electrons showing the nanowires with a length of about 48 µm. The nanowires were assembled into small bundles. Figure 5e was made with backscattered electrons showing contrasts between the silver deposited at the base of the nanowires in the form of dendrites (bright white areas) and the copper nanoparticles that decorate the tips of the nanowires (bright points). Figure 5f shows the elemental EDS analysis of Si_NWs-Cu_NPs on the yellow rectangle area in Figure 5h. In this case, the elements present in the sample were Si, O, Cu, and Ag in a percentage of 83.27%, 11.41%, 3.28%, and 2.03%, respectively. The presence of copper in the sample indicates the deposition of copper nanoparticles on the nanowires. In this case, in addition to a SiO₂ layer on top of the silicon, the oxygen can be attributed to a layer of copper oxides on top of the nanoparticles.

The 2200× SEM micrographs of the synthesized Si_NWs-CoO_NPs are shown in Figure 5g,h. Figure 5g is an image generated by secondary electrons, where the length of the nanowires is about 30 µm. It is observed that the clusters of agglomerated nanowires at the bottom have a different thickness than at the top, so they end in a tip. On the other hand, Figure 5h shows the contrast of the silver in the form of dendrites (bright white areas in the SEM micrograph of backscattered electrons), that, after the formation of nanowires, remained in the bottom. These silver structures were not removed by washing with a nitric acid solution because they can contribute to photocatalysis.

In the reviewed literature, there were no reports of silver dendrites at the bottom of nanowires. In the elemental EDS analysis of the Si_NWs-CoO_NPs (Figure 5i), corresponding to the area of the yellow rectangle of Figure 5h, the elements present in the sample were Si, O, C, Ag, Al, Na, and Co in a percentage of 62.59%, 20.25%, 14.27%, 1.30%, 1.12%, 0.41, and 0.06%, respectively. The presence of cobalt in the sample is an indicator of the deposition of cobalt on the nanowires, although its percentage is low due to the concentration of 0.01 M of CoCl₂ in the electroless bath. However, the EDS analysis did not confirm the oxidation state of cobalt, identifying it as metallic or oxidized. The presences of Al and Na were because of residues generated in the Si_NWs formation process. Na was part of the composition of the electroless bath and Al₂O₃ was an inlaid residue of the plate sanding. The distributions by zones of each element are shown in Figure 5j, where the elemental mapping of the Si_NWs-CoO_NPs is shown.

The growth process of silver dendritic structures in the nanowire formation process was observed with SEM. Figure 6 shows the Ag dendritic structures observed on p-type monocrystalline silicon (c-Si) substrate immersed for 10 s into a mixture of HF (5M)/AgNO₃ (0.035M) in an aqueous solution. Figure 6a shows a cross-sectional view of the surface composed of Ag_NPs and Si_NWs. Figure 6b shows that the silicon substrate was covered with nanoparticles and dendrites of Ag. Moreover, this image gives evidence of a diversity of shapes of the crystalline dendrites and branches. The uniformity of a large number of such structures is illustrated in Figure 6c, showing that the products consist almost entirely of spike-shaped Ag dendrite crystals. Figure 6d shows the random orientation of the Ag crystals obtained by transforming the nanoparticles used initially for shaping the nanowires.

The Ag dendrite formation process is caused by the galvanic displacement reaction consisting of redox reactions that occur simultaneously on the silicon surface. The cathodic reaction reduces Ag⁺ on the Si wafer surface by injecting holes into the valence band of Si and generating Ag metallic particles. The anodic reaction results in the oxidation and dissolution of Si adjacent to Ag particles [56]. During the displacement reaction, the amorphous phase leads to rapid and continuous deposition on the surface of the growing particles, and several randomly oriented Ag particles spontaneously crystallizing from the amorphous phase. The Ag particles randomly oriented realign and grow in preferential directions, eventually forming single crystal dendrites. The growth rate, morphologies, and structures of Ag dendrites are highly dependent on Ag⁺ ion concentration (AgNO₃), reaction time, and deposition temperature.

Figure 7a shows the diffractogram of the Si_NWs surface. The diffractogram shows diffraction peaks at 33.038° and 69.13°, both peaks belong to the planes (2 1 1) and (4 0 0) of the Si (JCPS 01-072-1088 and JCPS 01-75-0589). The plane (4 0 0) of cubic silicon located at 69.13° of 2θ coincides with that reported by articles that use the same synthesis method [57]. Although, in this case, that peak is not the preferential direction since the 33.038° of 2θ peak corresponding to the Si plane (2 1 1) has a higher intensity. The change in the intensity of the peaks of silicon was attributed to the presence of the nanostructures. Moreover, there is a 25° peak associated with the presence of SiO₂ generated by synthesizing the nanowires. The noise generated in the peak was due to the lack of a well-defined crystal structure.

Figure 7b shows the diffractogram of the surface of Si_NWs-CoO_NPs performed in grazing-incidence diffraction mode with a tube position at 1°, in a range from 2° to 120° of 2θ, using a voltage and current of 40 kV and 40 mA, respectively. The diffractogram shows different diffraction peaks, the peaks at 28.44° and 56.12° belong to the planes (1 1 1) and (3 1 1) of the Si (JCPS 01-75-0589). The peaks at 38.12°, 44.3°, and 77.399° belong to the planes (1 1 1), (2 0 0), and (3 1 1) of the cubic Ag (JCPS 01-065-2871), respectively. The peak at 44.23° is associated with the plane (1 1 1) of the Co (JCPS 01-089-4307). Compared to the coupled mode diffractogram, the change in Si diffraction peaks can be observed on the surface of Si_NWs-CoO_NPs. The preferential orientation of Si was in the plane (1 1 1) but the (3 1 1) shows an interference phenomenon [58,59] with the higher intensity occurring at 56.12° and not in the (2 1 1), as in the pattern of silicon powder. This interference can have possible additional peaks around the value of the central position. The appearance of the plane (1 1 1) was seen at 28.44°. There were probably slight alterations in the unit cells that the flexing of the Si_NWs can cause. The stress caused by these flexes would make the unit cell smaller on one side and larger on the opposite.

Figure 7c was in grazing-incidence diffraction mode with a tube position at 4° 2θ, in a smaller range of 26° to 50° of 2θ. The higher resolution allows appreciating better the peaks already seen in Figure 7b. Moreover, it will enable observing the appearance of a new peak corresponding to the Si plane (2 2 0) at 47.305°. However, the plane (1 1 1) of the Co and the plane (2 0 0) of the Ag overlapped in the same peak. So, to determine if both elements were present, the Rietveld refinement was used (Figure 7d). By using the Rietveld refinement, the fitting was 100% with the Ag pattern and because the intensity of the peak in the adjustment (red line) was over that of the experimental diffractogram, it was not considered that Co (1 1 1) was present.

X-ray photoelectron spectroscopy (XPS) was used to determine the surface composition and oxidation states of a Si_NWs-CoO_NPs sample. Figure 8a shows the complete XPS or survey spectrum and the identified elements were Si, C, O, Ag, and Co. In the Si2p signal (Figure 8b), two peaks were found at 99.9 eV and 104 eV corresponding to Si and SiO₂, respectively. On the other hand, in the C1s, two signals were observed at 285.9 eV and 289.9 eV (Figure 8c). These signals are typical of adventitious carbon contamination, so carbon was not part of the sample. Figure 8d shows the O1s region with a single peak at 533.6 eV, corresponding to the CO bonds. It has a small shoulder at about 530.5 eV attributed to metal oxides, in this case, Co. Likewise, Figure 8e shows the Ag3d region having two signals at 368.2 eV and 374.2 Ev. These two peaks are within the range 378–366 eV that can be associated with metallic silver, while the strongest signal occurs at 368.2 eV and corresponds to the characteristic signal of metallic silver. Finally, in the Co2p region (Figure 8f), there are four peaks at 782.48 eV, 787.0 eV, 798.08 eV, and 803.78 eV. It can be noted that none of these peaks correspond to metallic Co, so all the deposited Co by electroless was oxidized in the process. The obtained spectrum in this sample can coincide with Co(OH)₂ and CoO, but the difference between these signals is at most 1 eV, so the analysis was not conclusive, although it is very likely that both were present.

The reason for the Co deposit not being in its metallic form was attributed to the fact that during the synthesis process, the pH of the electroless solution was 14. The Pourbaix diagram of Co in aqueous solutions [34] indicates that Co²⁺ ions in an aqueous solution are below pH 9. At pH 14, there is a high concentration of OH ions forming Co(OH)₂.

3.2. Heterogeneous Photocatalysis

For heterogeneous photocatalysis experiments, solutions having 20 ppm of MO, MB, and Rh6G were used as organic molecules for degradation. The photocatalysts used were Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs. After photocatalysis, the collected samples were analyzed using a UV-Vis spectrophotometer, which showed the degradation of the MO and the chemisorption of MB and Rh6G.

Firstly, the dye calibration curves were experimentally obtained to know the changes made in the degradation of MO, MB, and Rh6G (Figure 9). Solutions with different concentrations were made and they were 2, 4, 6, 8, 10, 12, 14, 16, 18, and 20 ppm for each dye. Three different aliquots were taken for each solution. Then, the absorption spectra of the solutions were taken. From the obtained curves, the highest absorption peaks of the dyes were taken as reference points, which were 464 nm, 663 nm, and 526 nm for MO, MB, and Rh6G, respectively. In addition, an average was made with the maximum absorbance values of the three aliquots taken for each concentration that were plotted with the dye concentrations. Finally, a linear adjustment was made to the experimental data to generate the calibration curves of the three dyes, estimating the reduction in the concentration of the dyes after photocatalysis with Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs.

3.3. Photocatalysis with MO

Recurrently have been observed that Si_NWs cause water splitting, which collaterally causes the pH reduction from 6.2, in the MO solution, to 4.3 on average. Figure 10 shows the photocatalytic activities of the Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs with the MO using a white LED light. In this figure, the characteristic absorption peak of the MO at 464 nm decreases over time with the photocatalytic treatment, which is indicative of MO degradation. The maximum photocatalysis time was 150 min. This decrease in intensity occurs with the three photocatalysts (Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs), suggesting a similar photocatalytic degradation process. However, it was observed that in the case of the Si_NWs-Cu_NPs samples, the degradation of the MO was higher when reaching 150 min.

Figure 10d compares the photocatalytic activity with Si_NWs and Si_NWs-CoO_NPs have efficiencies at 150 min of 85.3% and 49.3%, respectively. The degradation efficiency of Si_NWs-Cu_NPs ranges from 5.5% at 5 min to 88.9% at 150 min, which makes it the most efficient photocatalyst.

The degradation mechanisms of MO, MB, and Rh6G have been extensively investigated [27,28], which altogether with their other characteristics make them usual targets for testing catalysts.

In the spectra of Figure 10a, while the absorbance intensity decreases at 464 nm (λ corresponding to the azo bond), there was an increase in intensity at 448 nm (λ related to the amino group). This indicates a reduction in the concentration of the azo bonds, promoting the formation of the amino groups [60].

When comparing the efficiency of the three photocatalysts, the degradation kinetics of the Si_NWs-Cu_NPs exhibit a behavior almost parallel to the Si_NWs. However, the kinetics of Si_NWs-Cu_NPs was slightly faster, having a rate constant (k) of about −0.14 compared with -0.012 of Si_NWs, attributed to the contribution of Cu metallic nanoparticles deposited on their surface. Contrarily, when comparing Si_NWs-CoO_NPs with Si_NWs, there was a higher reduction in the rate of degradation kinetics. This was attributed to the CoO nanoparticles, which were not separating electron–hole pairs to contribute to the degradation of the organic molecule of the MO.

Table 1. Calculations of the reaction order, R², and rate constant (k) for the tests with methyl orange, methylene blue, and rhodamine 6G.

		Methyl Orange		Methylene Blue		Rhodamine 6G
	Order	R2	k	R2	k	R2	k
SiNWs-CuNPs	1	0.983	−0.01416	0.9719	−0.014	0.983	−0.0134
	2	0.927	774.45	0.9738	820.43	0.8706	623.185
SiNWs	1	0.997	−0.0129	0.694	−0.0129	0.829	−0.013
	2	0.94	583.75	0.4742	625.78	0.655	570.47
SiNWs-CoNPs	1	0.978	−0.00435	0.746	−0.0016	6.88 × 10−5	−7.26 × 10−6
	2	0.953	99.238	0.743	27.80	6.99 × 10−6	0.040

shows that all the reactions in this study were of order 1 for the three semiconductor systems with each of the three dyes.

3.4. Photocatalysis with MB

Figure 11 shows the photocatalytic activity in the MB of the Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs with LED light. In this figure, the MB chemisorption was observed as the intensity decreases at its characteristic absorption peak, at 663 nm. However, the intensity of that peak does not decrease over time. Similar to the UV-Vis spectrum of Si_NWs, the intensity of the peak increases after 90 min instead of decreasing. The same happened at different times in the absorption spectrum of Si_NWs-CoO_NPs.

Brik et al. [61], using silicon nanowires for MB degradation, found a 21% reduction attributed to molecule absorption on the surface. They show significant changes in the degradation percentages even using a UV source. Ameen et al. [62] attribute the dye reduction in the dark, using Si_NWs as a photocatalyst, to the retention of the molecules by the large surface area of the nanowires.

Figure 11d compares the photocatalytic activities of the three photocatalysts. The highest chemisorption efficiency of Si_NWs-Cu_NPs was 87.2% at 150 min. This was the only photocatalyst with different behavior. The highest chemisorption efficiency of Si_NWs was 86% at minute 90 because, by minute 50, their efficiency decreased to 73%. In the case of the Si_NWs-CoO_NPs photocatalyst, its maximum chemisorption efficiency was 17.3% at minute 90 and by the end of the period, its efficiency decreased to 12.6%.

The UV-Vis absorption spectra of the photocatalysts show that the peak at 663 nm decreased and that there were no new peaks. So, in the case of MB, it was attributed that the phenomenon was not the degradation but the adsorption of organic molecules [60]. Moreover, the desorption of the molecules was shown, which explains why the absorbance increases at certain times of photocatalysis [16]. This process was identified as chemisorption and that behavior was attributed to a combination of absorption/desorption regulated by the changing parameters of the solution. The absence of degradation of the organic molecules was attributed to the lack of the potential necessary to break the bonds of that molecule with three aromatic rings. As with MO, the highest decoloration was obtained with that of Si_NWs-Cu_NPs, followed by Si_NWs, and Si_NWs-CoO_NPs as the lowest.

3.5. Photocatalysis with Rh6G

Figure 12 shows the photocatalytic activity in the Rh6G with the Si_NWs, Si_NWs-CoO_NPs, and Si_NWs-Cu_NPs with LED light. The Rh6G has its characteristic absorption peak at 526 nm, so as the intensity at that peak decreases with increasing photocatalysis time, dye chemisorption occurs as in the Si_NWs-Cu_NPs spectrum. However, in the Si_NWs and Si_NWs-CoO_NPs spectra, this peak intensity does not decrease over time.

Figure 12d compares the photocatalytic activities of the three photocatalysts. In this way, the Si_NWs-Cu_NPs had a chemisorption efficiency of 86.88% at 150 min, while the highest chemisorption efficiency of the Si_NWs was 87% at 120 min. In the case of the Si_NWs-CoO_NPs photocatalyst, its maximum chemisorption efficiency was approximately 12% at minute 5.

On the other hand, the UV-Vis absorption spectra of the photocatalysts show that as the peak decreases at 526 nm, there was no creation of new peaks, so it was attributed that the phenomenon present in Rh6G photocatalysis was not degradation, but the adsorption of organic molecules similarly to MB [60]. Again, similarly to MB, the desorption of the molecules was observed, explaining the increases at certain times of the photocatalysis [16]. The degradation of the Rh6G organic molecule was not achieved due to the lack of potential in the photocatalysts linked to the ability to concentrate charges by the resistivity of the semiconductor. The molecular structure of Rh6G has multiple aromatic rings, which need a sufficient amount of energy to break their bonds.

As with MB, the adsorption and desorption of organic dye molecules occurred, with the most efficient photocatalyst being Si_NWs-Cu_NPs. In the case of Si_NWs and Si_NWs-CoO_NPs exhibiting dye desorption, there was a reduction in chemisorption efficiency, which was very large in Si_NWs-Co_NPs because the deposited Co nanoparticles probably were oxidized and could be as Co(OH)₂/CoO deposited on the surface. It should be noted that the difference in efficiency between Si_NWs and Si_NWs-CoO_NPs was minimal.

The reason for the desorption of the dye molecules was attributed to the fact that, in the cathodic regions present on the surface of the nanowires, where there was a considerable accumulation of electrons (negative charge), molecular adsorption was carried out preferably until saturation. However, when these regions lose their charge (electric discharge), the process of desorption occurs because the charge of the molecule was no longer related to the charge on the surface of the nanowires.

3.6. Review and Comparison of Photodegradations of MO, MB, and Rh B Dyes by Si_NWs

Table 2. List of reported degradation efficiencies of MO, MB, and Rh B dyes using Si_NWs.

Silicon Type and Orientation	Photocatalyst Composite	Light Source	Target Molecule	Concentration (mg/L)	Degradation Efficiency (%)	Ref.
p	Cu- SiNWs	Visible light LEDs	MO	20.0	92 (150 min)	[12]
p (100)	Pd- SiNWs	Visible light	MO	25.0	84.5 (36 h in a fuel cell)	[63]
n (100)	H-SiNWs	UV	MO	0.327	93 (200 min)	[18]
p (111)	H-SiNWs	UV	MB	0.32	49.35 (200 min)	[18]
p (100)	SiNWs/GO	UV	MB	6.4	93 (280 min)	[64]
p (100)	Ag-SiNWs	Visible	MB	3.199	~98 (200 min)	[65]
	SiNWs	580 nm	MB	31.985	18.4 (120 min)	[53]
	Cu2O-SiNWs	580 nm	MB	15.99	53.8 (120 min)	[53]
p (100)	Cu-SiNWs/NaBH4	None	MB	15.99	~96 (10 min)	[66]
	SiNWs/GO + H2O2			6.4	96 (120 min)	[64]
	SiNWs/GO			6.4	93 (280 min)	[64]
p (100)	CuO2-SiNWs	580 nm	MB	0.2 × 10−3 M (5 mL)	(100 min)	[67]
p & n (100), (111), (110); 0.7 × 1.5 cm2; 0.02, 5–10, 100 Ω cm	H-SiNWs	>420 nm	MB MO	10−6 M (4 mL)	(200 min)	[18]
p (100), 1–10 Ω cm	SiNWs—AuNPS	>420 nm	Rh B	0.5 × 10−6 M (3 mL)	(240 min)	[17]
p (100), 1.2 × 0.8 cm2, 0.009–0.01 Ω cm	SiNWs—AgNPS SiNWs—CuNPS	>420 nm	Rh B	0.5 × 10−5 M (2 mL)	(120 min)	[52]
p (100), 2 × 2 cm2, 5–10 Ω cm	SiNWs	400 nm	Rh B	10	(90 min)	[62]
p (100) 2 × 2 cm2	SiNWs	>420 nm	Rh B	1 (50 mL)	(300 min)	[68]
p (111)	SiNWs	Visible light LEDs	MO, MB, Rh6G	20.0	85.3 (150 min), 86 (90 min), 87 (120 min)	This work
	Cu-SiNWs				88.9%, 87.2%, 86.88 (150 min)
	CoO-SiNWs				49.3 (150 min), 17.3 (90 min), 12 (5 min)

shows a brief review of the literature that use Si_NWs as photocatalysts with different crystalline orientation, heterojunction with other semiconductors, light sources, target dye molecule, dye concentration in aqueous solutions, treatment time, and degradation efficiencies.

The three dyes (MO, MB, and Rh 6G) have been used as target molecules to evaluate many different photocatalysts. In the case of Si_NWs, this table resumes some process parameters but many more cause differences in the degradation efficiencies. Moreover, there is a growing tendency to use heterojunctions with other semiconductors in addition to the variations of metallic decorating nanoparticles. In some cases, there is missing information about these or other parameters, such as the power of illumination, the heating of the aqueous solution, the atmospheric pressure, pulsed illumination, potential in case of photoelectrocatalysis, volume of the samples, area of lighting, and so forth.

The MO degradation kinetics were in the sequence Si_NWs-Cu_NPs (88.9%) > Si_NWs (85.3%) > Si_NWs-CoO_NPs (49.3%), with the Si_NWs-Cu_NPs having slightly faster kinetics. However, Si_NWs-CoO_NPs have slow degradation kinetics. In the references of Table 2, there are some of them with higher efficiencies than in this work, included in the last rows, especially using Cu_NPs. The proposal in this work points to CoO as a more stable photocatalyst than CuO₂. This does not imply that it has higher degradation efficiencies, which are lower in the conditions tested here. One of the reasons for such results was a lower obtained deposited amount of CoO. According with the EDS quantifications, Si_NWs (Ag 3.98%), Si_NWs-Cu_NPs (Ag 2.43%, Cu 3.94%), Si_NWs-CoO_NPs (Ag 2.01%, Co 0.01%), normalizing relative to the silicon content. The deposited nanoparticles, including Cu and Co, were by electroless without adding an activator, which means that the nanowires own potentials and electron supply regulates, in part, the kinetics and amount of depositions.

In the case of working with oxides, CuO₂ and CoO, they should not require further processing to maintain stability on multiple runs. Nonetheless, all the known reports show decrements in consecutive tests even considering that the solutions have neither other pollutants nor salts. This is another main reason, together with electron–hole recombinations, for photocatalysis has not been used broadly. One explanation for this is that this process is based on redox reactions, which generate ionic compounds as byproducts. Such charged species increase the double layer around the active sites having cathodic or anodic potentials, which blocks them, in part, causing diminish in efficiencies.

4. Conclusions

The Si_NWs were synthesized using the MACE method. The Si_NWs-Cu_NPs were synthesized by depositing the Cu nanoparticles auto-catalytically. Similarly, the Si_NWs-CoO_NPs were synthesized firstly as Co and oxidized in the same process. The nanowires had a length of about 23 µm to 30 µm with a vertical structure on the wafer. They were peak-shaped and spaced.

Silicon wafers have no photocatalytic effect lacking the potential for splitting water or causing redox reactions. Si_NWS, without Schottky barriers or further doping, have a higher area but keep the silicon characteristics. Si_NWs, Si_NWs-CoO_NPs, and Si_NWs-Cu_NPs, prepared in the described procedure and having Ag at the bottom, manifest the effects shown in this and other works. There were dendritic conformations of the silver nanoparticles at the bottom, originally used for generating the nanowire bundles. This structure type can contribute to faster electron transfer.

In the optical images, there is the phenomenon of interference, observed as iridescent colors, because of the size of the filaments resulting from the MACE method and the space between them, which, as far as we know, this work is the first or between the few that show interference of visible light in the nanostructured bundles. Moreover, there was interference in the plane (3 1 1) for the X-ray diffractograms caused by the coincidence of the Cukα1 (1.5604 Å) under grazing incidence with the silicon plane distance.

The EDS analysis confirmed the presence of Co in the samples (oxidized). In the case of Si_NWs-Cu_NPs, the approximate length of the nanowires was between 44 µm and 48 µm. They showed a vertical structure on the plate, homogeneity, and agglomerations between nanowires. The EDS analysis confirmed the presence of Cu.

The XPS analyses on the Si_NWs-CoO_NPs samples show that the deposited of Co was not in metallic form but as CoO_NPs or as Co(OH)₂/CoO. The pH bath conditions allowed obtaining CoO_NPs instead of metallic cobalt.

In the case of MO, the degradation kinetics of Si_NWs-Cu and Si_NWs were almost parallel, with Si_NWs-Cu_NPs having slightly faster kinetics due to the contribution of Cu in photocatalysis. Si_NWs-CoO_NPs were found to have very slow degradation kinetics. This was attributed to the cobalt oxide that, contrary to Cu nanoparticles, did not contribute to the electron–hole separation.

MB and Rh6G showed no degradation of the dye molecules during the tests; instead, there was adsorption and desorption of them. The most efficient photocatalyst was Si_NWs-Cu_NPs. Nonetheless, it should be noted that Si_NWs and Si_NWs-Cu_NPs show similar efficiencies. Si_NWs-CoO_NPs show substantially less efficiency.

The photocatalytic activity of this work’s materials, the Si_NWs, Si_NWs-Cu_NPs, and Si_NWs-CoO_NPs, depend not only on the materials themselves but also on multiple factors, which alter the kinetics and final efficiencies. Under the conditions reported, with the characteristics described, all the materials were photocatalytic, but in the sequence Si_NWs-Cu_NPs > Si_NWs > Si_NWs-CoO_NPs. Even more, their capability of degrading methyl orange was not replicated for methylene blue and rhodamine 6G, which shows that, in addition to the factors to that sequence, there are others to address when changing the target for decomposing. The tests can be repeated several times, but there is a decrement caused by sorption on the surfaces of ionic byproducts of the decomposition, changes in the oxidation state of Cu_NPs, chemical reactions, and various other processes.