Monitoring Hydrogen Peroxide Using Electrochemically Reduced Graphene Oxide Modied Screen Printed Electrodes During Metal Assisted Chemical Etching Processes

The concentrations of etchant solution substituents in metal assisted chemical etching (MACE) processes control the morphology and reectivity of subsequently etched wafers. In particular, the concentration of hydrogen peroxide (H 2 O 2 ) plays a vital role in the MACE process. Unfortunately, the H 2 O 2 concentration is not stable when prolonging the etching process at higher temperatures. As a result, the commercialization of MACE processes for the production of IP texturization has appeared industrially unattractive. Herein, we proposed an innovative method to monitor hydrogen peroxide during the MACE process with an electrochemical method. Reduced graphene oxide (RGO) prepared through an environmentally benign electrochemical method was used to modify a screen-printed electrode (SPE). Under an optimized condition, the RGO/SPE was used to test etching solutions. The MACE process was conducted and the hydrogen peroxide concentration within the etching solution was checked by the RGO/SPE. The RGO/SPE demonstrated excellent electrochemical performance and could record changes to H 2 O 2 concentrations with cyclic voltammetry (CV). Interestingly, the presence of copper (Cu) in the etching solution catalyzed not only the etching process, but also the electrochemical reduction of H 2 O 2. After etching, the reectivity and structural morphology of the etched wafers were checked. The described modied electrode is disposable, and the fabrication process is rapid and inexpensive, allowing for real time application in, and control of, MACE processes.


Introduction
Texturization of silicon is a crucial step in solar cells fabrication processes that ultimately enhances conversion e ciencies by improving light trapping ability [1,2]. There are currently a variety of techniques that can used for the texturization of single-and multi-crystalline silicon wafers. Anisotropic etching and isotropic etching are the two main techniques used to create upright and inverted pyramid textures on single-crystalline silicon (sc-Si) wafer surfaces [3][4][5]. Notably, alkaline etching techniques exhibit anisotropic behavior, and are therefore not suitable for the etching of multi-crystalline Si wafers [6]. Metal assisted chemical etching (MACE) processes have received enormous attention for its potential use in texturizing silicon wafers. So far, a variety of textures such as nanopores [7], nanowires [8], and inverted pyramid (IP) structures has been produced on the surfaces of sc-Si wafers using the MACE process [9].
Among them, the IP structure has the lowest speci c surface area and has demonstrated excellent light trapping ability. As such, wafers etched in this way have low surface carrier recombination rates and produce highly e ciency solar cells [10]. In the MACE process, reaction parameters, such as reaction temperature, concentration of etchant, and reaction time, play key roles in the formation of IP textures on sc-Si surfaces [11]. Generally, copper (Cu) or silver (Ag) metal is used in the MACE process in combination with hydro uoric acid (HF) and hydrogen peroxide (H 2 O 2 ) [12]. However, evaporation of H 2 O 2 readily occurs during the etching process, especially at high temperatures [13]. Our recent study reveals that the subsequent changes to H 2 O 2 concentration incurs variations in surface morphology and re ectivity of the etched wafers [14]. Hence, it is necessary to monitor the changes in the H 2 O 2 concentration of the etching solution during the throughout process in order to maintain constant concentrations.
Several analytical methods including electrochemical [15], titrimetry [16], spectrophotometry [17], chemiluminiscene [18], and uorimetry [19] have been used for the detection of H 2 O 2 [20]. Of these, the electrochemical method is relatively simple, cost-effective, rapid, and sensitive when compared to the other listed detection methods [21]. However, unmodi ed electrodes are unsuitable for the detection of analytes due to their high electron transfer impedance and poor catalytic activity [22]. Therefore, electrode modi cation with suitable electrode materials is crucial. Graphene as two-dimensional honeycomb sp 2hybridized nano sheets with single atom thickness is a potential candidate [23,24]. To date, graphene has been used for the fabrication of bio and electrochemical sensors due to its high electrical conductivity, speci c surface area (2600 m 2 /g), charge carrier mobility, exibility, and mechanical strength; all of which are more ideal than the properties other carbon nanomaterials [25,26]. Thus, electrochemically reduced graphene oxide (RGO) was used for the fabrication of various modi ed electrodes. The preparation of RGO with an electrochemical method is eco-friendly, rapid, and more conducting than chemical reduction [27]. To date, H 2 O 2 detection methods have only been demonstrated only in biological based samples with no reports for the detection of H 2 O 2 in a harsh environment, such as etching solution [28,29].
In this work, the concentration changes of H 2 O 2 during the MACE process were detected with an electrochemical method. Initially, the copper MACE method was conducted to produce the IP textured wafers. The morphology and re ectivity of the etched wafers were investigated. Then, the stability of spent etching solution was monitored by electrochemical detection using an RGO-modi ed disposable screen-printed electrode. Finally, the solution concentration difference determined by peak current response changes of the electrodes was remade with the introduction of new H 2 O 2 . After makeup, the etching process was conducted again using the same procedure. The morphology and re ectivity changes were recorded and compared with the results of fresh solution.

Materials and methods
Graphene oxide, copper sulfate (Cu(SO 4 ).5H 2 O, 99.99%), sodium phosphate mono basic (NaH 2 PO 4 ) and sodium phosphate di basic (Na 2 HPO 4 ) were received from Echo Chemical Co. Hydro uoric acid (HF, 49%), hydrogen peroxide (H 2 O 2 , 30%), nitric acid (HNO 3 , 70%) were purchased from Renew Chemical Materials Co. The supporting electrolyte, 0.05 M phosphate buffer, was prepared by dissolving Na 2 HPO 4 and NaH 2 PO 4 in DI water. All reagents were used without further puri cation. The screen printed electrodes (SPE) were purchased from Zensor R&D Co.. The surface morphologies of the etched wafers and modi ed electrodes were investigated using an FEI Nova nano scanning electron microscope (SEM) 230 equipped with an EDAX Apollo silicon drift detector energy dispersive X-ray spectroscopy system. A Laser Raman spectrometer (Jasco NRS-5100) was used to examine the defects and disorders of the electrode materials. The electrochemical studies were performed using a CHI1211C instrument. The CHI instruments worked based on a conventional three electrode system using SPE consist of the graphite as a working electrode, saturated Ag/AgCl (saturated KCl) as a reference electrode, and graphite wire as a counter electrode.

Cu assisted chemical etching of sc-Si
Boron doped p-type as-cut sc-Si wafers with thickness of 165 µm, (100)-oriented surface, and resistivity 0.5-1.5 Ω.cm, sliced into 100 mm × 50 mm sheets were used in this study. Prior to etching, the wafers were cleaned using 10% HF for 10 min to remove any native oxides. Thereafter, wafers were immersed in a 0.06M Cu(SO 4 ).

Fabrication of RGO modi ed electrodes
The GO was prepared by a modi ed Hummer's method based on previous reports [30]. Initially, about 7 µL of the GO (0.15 mg/mL) solution was drop cast on the pre-cleaned SPEs and dried using a hot plate. Finally, the GO was reduced electrochemically in the phosphate buffer solution (pH-5) at a 50 mV/s scan rate for 5 cycles in the potential window of 0 to -1.6V. Then, the modi ed electrode was dried and used for the real time detection of H 2 O 2 in the etching solution. Fig. 1 shows the schematic of electrode fabrication and measurement.

Results And Discussion
The surface morphology of the MACE processed wafers were checked by SEM with corresponding images shown in Fig. 2. The raw wafer had grooves and the saw marks on the surface before etching ( Fig. 2(a)). After etching at 60 °C for 7 min in fresh etching solution, random IPs were observed on the surface ( Fig. 2(b)). The surface morphology of the subsequent etched wafers in spent etching solution after 30 min and 60 min were irregular and collapsed, as shown in Figs. 2(c) and 2(d), respectively, due to the evaporation of H 2 O 2 . During the etching process, Cu 2+ ions were reduced as nanoparticles (NP) by electrons donated from the silicon substrate which allowed the Cu-NPs to deposit onto the silicon wafer surface [8]. Then, the Cu-NPs enticing the Cu 2+ ions in the etching solution led to the aggregation of Cu NPs [31]. Hence, the silicon surface was oxidized and dissolved in the etching solution upon reacting with HF.
IP structures formed on the Si wafer surface after the acidic environment reacted with the holes injected by H 2 O 2 via the copper NPs into the silicon [32]. At lower concentrations of H 2 O 2 , the rate of hole injection and the rate of etching in silicon beneath the Cu-NPs simultaneously decreases. The re ectance spectra of the etched wafers were checked and are correspondingly shown in Fig. 3. Due to poor light trapping, the raw wafers exhibited a high re ectivity of about 32.1%. After etching, the re ectivity decreased due to the formation of IP textures on the surface. The formation of IP structures could increase the light travelling path with a triple bounce on the adjoining (111) side walls of IP [33]. However, the re ectivity of the wafers etched after 30 and 60 min increased due to the signi cant changes in the adjoining facet planes of IP. This indicated that the re ectivity and morphology deteriorated when the etching was prolonged for longer periods of time.
3.1 Surface morphology and Raman spectra of the modi ed and unmodi ed electrodes The surface morphologies of the bare, modi ed, and unmodi ed electrodes are shown in Figs. 4(a)-4(c). Flake-like structures were observed for the unmodi ed SPE. The observation of a crumbly thin layer on the electrode con rmed surface modi cation with GO. Further, wrinkles and a shrunken sheet-like structure were attributed the electrochemical reduction of the GO.
Raman spectroscopy is an important tool used to characterize carbon-based nanomaterials, such that the presence of conjugated and double carbon−carbon bonds increases Raman intensities [34,35]. Fig. 5 depicts the Raman spectra of bare SPE, GO/SPE and RGO/SPE. The Raman spectra of the modi ed electrode shows two major peaks assigned for D (disorder) and G (graphitic) bands. The D band (1359 cm -1 ) is associated with the defects and disorder of the sp 2 carbon lattice, and the G band (1587 cm -1 ) is due to the graphitic nature and highly ordered arrangement and of sp 2 carbon [36].
As shown in Fig. 5, owing to the graphitic nature, the intensity of the G band (1587 cm -1 ) is higher than the D band (1359 cm -1 ). When the electrode was modi ed with GO, the intensity of the D band increased due to the disorder in the SP 2 carbon by the formation of oxygen functional groups. The intensity of the D band was further increased past the G band due to the formation of defects and disorders during electrochemical reduction. Moreover, the intensity ratios (I D /I G ) of the bare SPE, GO/SPE, RGO/SPE were calculated to be about 0.26, 0.43, and 1.43, respectively. Signi cantly, the higher intensity ratio of RGO indicated that the reduction process changed the structure of GO by creating more structural defects [37].  Fig. 6(a), the RGO/SPE showed cathodic peaks at -0.4 V (small) and -0.8 V, and an anodic peak at -0.11 V. A well-de ned cathodic peak with a high peak current response at -0.8 V was attributed to the electro reduction of H 2 O 2 . The peaks at -0.4 V and -0.8 V were due to the redox behavior of Cu 2+ ions in the etching solution [38]. To con rm the redox behavior of Cu 2+ ions, the etching solutions (c and d) were checked separately. When adding solution-c without CuSO 4 , the cathodic peak was observed at -0.69 V which corresponded to the reduction of H 2 O 2 and no redox peaks were found in the applied potential window (Fig. 6(b)). The RGO/SPE had redox peaks during the addition of solution-d without H 2 O 2 (Fig. 6(c)). This result con rmed the redox behavior of Cu present in the etching solution. The electrochemical response was poor with the separate addition of 5 mM H 2 O 2 . However, the RGO/SPE exhibited a well-de ned irreversible cathodic peak with high peak current response with the addition of 5 mM H 2 O 2 with 300 µM CuSO 4 (( Fig. 6(d)).
Previously, Cu based nanocomposites were used as an electrode modi er for the detection of H 2 O 2 due to their excellent electrocatalytic properties [39,40]. For instance, Cheng et al. [41] reported the use of the Furthermore, the RGO modi ed electrodes were prepared by CV electrochemically using different cycles such as 5 (RGO/SPE-5) and 10 ( Fig. 7(b)). In comparison, the RGO/SPE-5 showed slightly higher responses than the RGO/SPE-10 cycles. Therefore, 5 cycles were used for the electrochemical reduction of GO in the fabrication process of the modi ed electrode. The electrochemical behavior of the RGO/SPE towards the detection of H 2 O 2 was checked using different pHs values for the electrolyte. Fig. 7(c) shows the electrochemical responses of RGO/SPE at different pH values such as 3, 5, 6 and 7 for the electro reduction of 5 mM H 2 O 2 . Inferior electrochemical response and higher reduction potentials were observed when using pH 6 and 7. Conversely, the RGO/SPE exhibited a higher peak current response and a lower reduction potential. In particular, a well-de ned and sharp cathodic peak was obtained in pH 3 rather than pH 5. Notably, the pH of the etching solution was acidic due to the etching solution containing 2 M HF. Hence, pH-3 was used for the entirety of electrochemical experiments. In order to investigate the electrocatalytic property of the RGO/SPE, different concentrations of H 2 O 2 were added from the etching solution (solution-a) in the 0.05 M phosphate buffer pH 3. As shown in Fig. 7(d), the peak current responses increased gradually for each addition of H 2 O 2 from 50 to 300 mM. The results indicated that the modi ed electrode had desirable electrocatalytic properties and was successful in the detection of the H 2 O 2 in the etching solution without dilution.  Fig. 8(b). The raw wafer showed the highest re ectivity due to poor light trapping ability. After etching, the re ectivity decreased to 5.99 % which con rming the formation of wellde ned IP structures on the silicon wafers surface.
Nevertheless, the re ectivity increased to 7.30 and 8.60 % for after 30  Here, Δm is total mass loss, ΔS is thickness loss, t is time, is mass density, and S is surface area of the silicon wafers, respectively. The thickness lost and etching rate calculated from Eqs. (1) and (2) are tabulated in Table- Funding (information that explains whether and by whom the research was supported) This work was sponsored by the Ministry of Science and Technology (MOST) and the United Renewal Energy Co. through the Advanced Green Energy Project.

Con icts of interest/Competing interests (include appropriate disclosures)
The authors declare that they have no con icts of interest Availability of data and material(data transparency)

Not applicable
Code availability (software application or custom code)

Not applicable
Authors' contributions (optional: please review the submission guidelines from the journal whether statements are mandatory) All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Subbiramaniyan Kubendhiran, Gavin Sison and Hsiao Ping Hsu. The rst draft of the manuscript was written by Chung-Wen Lan and all authors commented on previous versions of the manuscript. All authors read and approved the nal manuscript.
Ethics approval(include appropriate approvals or waivers) The manuscript should not be submitted to more than one journal for simultaneous consideration.