Removal of As(V) Based on Amino-Group Surface-Functionalized Porous Silicon Derived from Photovoltaic Silicon Cutting Powder

In this study, amino group surface-functionalized porous silicon adsorbent was successfully prepared for the first time using diamond wire saw silicon powder (DWSSP) as raw material through copper-assisted chemical etching (Cu-ACE) and organic functional group grafting. Amino-functionalized porous silicon adsorbent (TEPA-GTS-NPSi) can be used for removing As(V) from water. The properties and mechanism of the new adsorbent were characterized by infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (FE-SEM), Brunauer–Emmett–Teller analysis (BET), and thermogravimetric analysis (TGA). The concentration of metal ions in the solution was determined by inductively coupled plasma spectrometry. Meanwhile, the effects of initial pH, adsorption time, initial concentration and adsorbent dosage on the removal of As(V) in an aqueous solution were studied by intermittent adsorption experiments. The results showed that the adsorption equilibrium could be reached rapidly after 30 min soaking. Under the optimized pH of 7, the maximum adsorption capacity was 13.2 mg/g, and the minimum adsorption limit was 3 mg/L. The adsorbent shows good adsorption performance after five successive regenerated cycles. Based on the density functional theory (DFT) analysis results, the adsorption mechanism is attributed to hydrogen bond interaction between the NH2 group and As(V) ions.


Introduction
Solar energy has become one of the most important clean energy sources in recent years. As we know, the silicon solar cell, one of the most important photovoltaic devices, is widely acclaimed for its advantages, such as easy installation, low maintenance cost and strong adaptability [1]. As a result, silicon ingots are the main raw material of the solar energy industry [2]. Metallurgical grade silicon (MG-Si, <99%, 2 N) is purified into solar grade silicon (SoG-Si, >99.9999%, 6 N) ingots by a high consumption process. During the slicing process, about 35% of the silicon ingot becomes cutting silicon powder [3]. With the innovations in slicing technology, however, it is roughly estimated that more than 160,000 tons of diamond wire saw silicon powder (DWSSP) will be discarded in 2019 alone, as shown in Figure S1. If a large amount of micron and sub-micron cutting powder is directly discharged, it will not only cause the waste of resources but also further increase pollution of the environment. Therefore, the value-added recovery of cutting waste will bring considerable environmental and economic benefits [4][5][6].
Arsenic is widely found in nature, and arsenic pollution mainly comes from the exploitation, smelting and extensive use of arsenic compounds, such as the production and use of arsenic-containing pesticides, and many industries, all of which led to arsenic contamination in the environment [7]. If the human body is chronically exposed to arsenic, it can lead to a range of chronic diseases and cancer [8]. The abundant existing species of arsenic in pollutants are inorganic and organic arsenic, and the toxicity of the inorganic form is worse than that of the organic form. The main existing forms of inorganic arsenic are As(III) and As(V). In the process of treating wastewater polluted by arsenic, the oxidation of As(III) into As(V) is mostly adopted for subsequent treatment [9]. To date, diverse treatment technologies, such as chemical precipitation [10], ion exchange [11] and electrocoagulation [12], as well as ultrafiltration (UF) [13], nanofiltration (NF) and reverse osmosis (RO) [14], have been used for removal of arsenic. However, these technologies have disadvantages, such as high energy costs and excessive use of toxic substances, and they can also easily cause secondary pollution; hence, they have certain limitations in practical applications. Adsorption technology has been widely studied because of its advantages of high efficiency, flexibility, and simple operation [15][16][17]. Therefore, this is an environmentally friendly method for removing high concentrations of arsenic. Among various adsorbents, silica gel or magnetic nanoparticles were used as the substrate and modified by organic functional groups as adsorbents of arsenic [18,19]. Porous silicon can be used for modification of organic functional molecules due to its large specific surface area and more active groups (Si-OH) on the surface [20][21][22], which can also be used as a potential adsorbent substrate material. According to a research report by Dariusz M.
Jarzabek, et al., the friction performance of the silicon surface was greatly improved when the porous silicon was exposed to an aqueous solution of alkali metal ions [23]. On the other hand, the organic modification of the hydrogen-terminated silicon surface summarized by Danial D. M. Wayner et al. also provides more possibilities for the application of modified porous silicon [24].
Herein, we focus on the resource recycling and the removal of arsenic in aqueous solution based on porous silicon derived from the DWSSP combined with copper nanoparticleassisted chemical etching [21,22], and modification of 3-GTS and TEPA. The effects of initial pH value, initial concentration of As(V), contact time and amount of adsorbent on the adsorption property of the resulting organic-inorganic hybrid were carefully investigated. Moreover, the uptake performance for arsenic of TEPA-GTS-NPSi was appraised by isothermal adsorption, thermodynamics, and kinetics studies. To further clarify the arsenic adsorption process, the experimental data of kinetics were fitted by using a pseudo-firstorder model and a pseudo-second-order model. Finally, the adsorption mechanism was further confirmed by DFT analysis. This work not only solves the problem of silicon waste recovery in the photovoltaic industry but also provides a potentially feasible strategy for As(V) adsorption in aqueous solution.

Chemicals and Reagent
Diamond-wire saw cutting silicon powder was provided by LONGi Silicon Material Co. Ltd. (Chuxiong, China). (3-glycidyloxypropyl)  (Tianjin, China). The reagents and solvents used are of analytical grade and need no further purification unless otherwise specified.

Preparation of Nanoporous Silicon (NPSi)
The silicon raw material was pre-washed with ammonia, hydrogen peroxide and deionized water (volume ratio 2:2:5) to remove surface organic impurities, and then washed with hydrochloric acid, hydrogen peroxide and deionized water (volume ratio 2:2:5) to remove oxides. Finally, a 5% HF solution was used for washing to obtain purified silicon powder [25].
The copper nanoparticles were chemically deposited on the silicon surface for 1 min and then placed in a mixed solution of HF (40%) and H 2 O 2 (5 mM) for chemical etching [22]. After 1 h reaction, the etched silicon powder was filtered and cleaned with 30% nitric acid to remove residual nano-Cu particles. After filtration, the porous silicon powder was washed with ethanol and deionized water repeatedly until the pH was neutral. Finally, the filtered porous silicon powder was dried under vacuum at 60 • C for 12 h.
To increase the active group on the surface of porous silicon powder, the dried silicon powder was stirred for 3 h in a solution with the H 2 SO 4 /H 2 O 2 volume ratio of 3:1 [26]. After activation, it was filtered and washed repeatedly with ethanol and deionized water until the pH was neutral. Finally, the filtered porous silicon powder was dried under vacuum at 60 • C for 12 h.

Preparation of Nanoporous Silicon (NPSi)
The silicon raw material was pre-washed with ammonia, hydrogen peroxide and deionized water (volume ratio 2:2:5) to remove surface organic impurities, and then washed with hydrochloric acid, hydrogen peroxide and deionized water (volume ratio 2:2:5) to remove oxides. Finally, a 5% HF solution was used for washing to obtain purified silicon powder [25].
The copper nanoparticles were chemically deposited on the silicon surface for 1 min and then placed in a mixed solution of HF (40%) and H2O2 (5 mM) for chemical etching [22]. After 1 h reaction, the etched silicon powder was filtered and cleaned with 30% nitric acid to remove residual nano-Cu particles. After filtration, the porous silicon powder was washed with ethanol and deionized water repeatedly until the pH was neutral. Finally, the filtered porous silicon powder was dried under vacuum at 60 °C for 12 h.
To increase the active group on the surface of porous silicon powder, the dried silicon powder was stirred for 3 h in a solution with the H2SO4/H2O2 volume ratio of 3:1 [26]. After activation, it was filtered and washed repeatedly with ethanol and deionized water until the pH was neutral. Finally, the filtered porous silicon powder was dried under vacuum at 60 °C for 12 h.

Characterization of Materials
FT-IR (Nicoletis10) measurements were used to investigate the functional groups on the surface of the adsorbent. The FTIR spectrophotometer (Nicoletis10, Thermo Fisher Scientific, Waltham, MA, USA) had a wave number range from 4000 to 400 cm −1 . The element composition and the existing form of the element were determined by X-ray photoelectron spectroscopy (XPS, K-Alpha+, Thermo Fisher Scientific, Waltham, MA, Scheme 1. The synthesis process of TEPA-GTS-NPSi. First, (3-glycidyloxypropyl) trimethoxy-silane (3-GTS) (5 mL), NPSi (2 g) and anhydrous toluene (30 mL) were mixed and stirred for reflux at 60 • C for 24 h. After being filtered, this was washed with ethanol 5 times and dried under vacuum at 60 • C for 12 h [27]. The products obtained in the first step are denoted as GTS-NPSi. Secondly, GTS-NPSi (1.5 g), TEPA (tetraethylenepentamine) and anhydrous toluene (30 mL) were mixed and stirred for reflux at 60 • C for 24 h [28]. The products obtained in the second step are denoted as TEPA-GTS-NPSi.

Characterization of Materials
FT-IR (Nicoletis10) measurements were used to investigate the functional groups on the surface of the adsorbent. The FTIR spectrophotometer (Nicoletis10, Thermo Fisher Scientific, Waltham, MA, USA) had a wave number range from 4000 to 400 cm −1 . The element composition and the existing form of the element were determined by X-ray photoelectron spectroscopy (XPS, K-Alpha+, Thermo Fisher Scientific, Waltham, MA, USA) with a µ-spot monochromatic AlKα source (AlKα, pass energy = 50.0 eV (wide scan) and 30 eV (narrow scan), Voltage = 10 kV, Current = 20 mA). The specific surface area and pore size of the TEPA-GTS-NPSi were determined utilizing a N 2 adsorption/desorption specific surface area meter (Quadrasorb-evo, Quantachrome Corporation, Boynton Beach, FL, USA). TGA (NETZSCH, Selbu, Germany) were used to investigate the number of organic functional groups. The morphology of silicon powder was characterized by FE-SEM (Nova NanoSEM 450, 10 kV, FEI, Hillsboro, OR, USA). ICP-OES (Leeman prodigy 7, Teledyne Leeman Labs, Inc., Hudson, NH, USA) was used to measure the concentration of metal ions in the adsorbed solution.

Batch Adsorption Studies
The effect of pH on the adsorption capacity was investigated with 20 mL of 100 mg/L As(V) solutions in the pH range of 1-8 at 25 • C. The effect of time on the adsorption capacity was analyzed at pH = 7, As(V) concentration of 100 mg/L and 25 • C for different adsorption times (0.5-120 min). The effect of initial As(V) concentration on the adsorption capacity was researched at pH = 7 and 25 • C for 2 h with different initial concentrations (10-150 mg/L). The effect of dosage on the adsorption capacity was investigated with 20 mL of 20 mg/L As(V) solutions at 25 • C for 2 h. The dosage range of adsorbent was 10-35 mg. The removal efficiency (R) and the adsorption capacity (q e ) were calculated by the following Equations (1) and (2), respectively: where C 0 and C e are initial and equilibrium concentration of As(V) in aqueous solution (mg/L), respectively. V is the volume of solution, and m is the mass of absorbent.
To study the recycling properties of adsorbents, 1 mol/L of HCl solution was used as the desorption agent. After the adsorption, the adsorbent was desorbed in desorption agent for 3 h to study the cyclic usability of the adsorbent. Then, the adsorbent was washed with distilled water several times and vacuum dried at 25 • C for 24 h for the next round of adsorption. The above procedures were repeated five times.

Characterization of the Amino-Functionalized Nanoporous Silicon
The morphology and microstructure of the prepared porous silicon and organic functionalized porous silicon were characterized by SEM, as shown in Figure 1. As seen in Figure 1a, the raw material of DWSSP is mainly in the form of debris, with relatively flat particle surface and no obvious hole structure. Compared with the morphology of the unetched sample (Figure 1b), the morphology of amino-functionalized nanoporous silicon is clearly changed, pore structures start to appear on the particle surface, and the pore structure mainly appears on the particle surface after etching. This may be attributed to the fact that Cu nanoparticles accelerate the rate of local electrochemical reactions on the surface of silicon particles in the etching solution system. Meanwhile, the size of silicon particles decreased, indicating that the nanoporous silicon with pore structure has been prepared by diamond wire-cut waste.
Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) were used to characterize functional groups on the surface of organic functionalized porous silicon, and the infrared spectrum of GTS-NPSi and TEPA-GTS-NPSi was obtained as shown in Figure 2.
The ligands of the two adsorbents are trimethoxy-silane and tetraethylenepentamine, respectively. Therefore, it was found that the peak at 470 cm −1 is corresponding to the bending vibration of Si-O-Si, the peak at 803 cm −1 is ascribed to the symmetric stretching of Si-O-Si, and 1101 cm −1 is corresponding to the stretching vibration of Si-O-Si [29]. The peaks at 3440 and 1630 cm −1 were corresponding to O-H stretching vibration and deformation vibration, respectively. Moreover, in the spectrum of GTS-NPSi and TEPA-GTS-NPSi, the peaks at 2934 and 2834 cm −1 were corresponding to the C-H stretching vibration, indicating that organic silane was successfully grafted onto porous silicon nanoparticles' surface [30]. In the spectrum of TEPA-GTS-NPSi, the band at 1384 cm −1 is due to the -NH 2 and -NH bending vibrations [31]. These indicate that organic functional groups have successfully grafted on the surface of porous silicon. Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) were used to characterize functional groups on the surface of organic functionalized porous silicon, and the infrared spectrum of GTS-NPSi and TEPA-GTS-NPSi was obtained as shown in Figure 2. The ligands of the two adsorbents are trimethoxy-silane and tetraethylenepentamine, respectively. Therefore, it was found that the peak at 470 cm −1 is corresponding to the bending vibration of Si-O-Si, the peak at 803 cm −1 is ascribed to the symmetric stretching of Si-O-Si, and 1101 cm −1 is corresponding to the stretching vibration of Si-O-Si [29]. The peaks at 3440 and 1630 cm −1 were corresponding to O-H stretching vibration and deformation vibration, respectively. Moreover, in the spectrum of GTS-NPSi and TEPA-GTS-NPSi, the peaks at 2934 and 2834 cm −1 were corresponding to the C-H stretching vibration, indicating that organic silane was successfully grafted onto porous silicon nanoparticles' surface [30]. In the spectrum of TEPA-GTS-NPSi, the band at 1384 cm −1 is due to the -NH2 and -NH bending vibrations [31]. These indicate that organic functional groups have successfully grafted on the surface of porous silicon.
To determine the grafting amount of organic functional groups, thermogravimetric analysis (TGA) was used to verify it. Thermogravimetric diagrams of NPSi, GTS-NPSi and TEPA-GTS-NPSi are shown in Figure 3. Among them, the mass loss of NPSi, GTS-NPSi and TEPA-GTS-NPSi was 6.6%, 13.6%, and 20.7%, respectively. Considering the decomposition of the organic group and evaporation of water, the amounts of grafted GTS and TEPA groups were 7.0% and 7.1%, respectively.  Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) were used to characterize functional groups on the surface of organic functionalized porous silicon, and the infrared spectrum of GTS-NPSi and TEPA-GTS-NPSi was obtained as shown in Figure 2. The ligands of the two adsorbents are trimethoxy-silane and tetraethylenepentamine, respectively. Therefore, it was found that the peak at 470 cm −1 is corresponding to the bending vibration of Si-O-Si, the peak at 803 cm −1 is ascribed to the symmetric stretching of Si-O-Si, and 1101 cm −1 is corresponding to the stretching vibration of Si-O-Si [29]. The peaks at 3440 and 1630 cm −1 were corresponding to O-H stretching vibration and deformation vibration, respectively. Moreover, in the spectrum of GTS-NPSi and TEPA-GTS-NPSi, the peaks at 2934 and 2834 cm −1 were corresponding to the C-H stretching vibration, indicating that organic silane was successfully grafted onto porous silicon nanoparticles' surface [30]. In the spectrum of TEPA-GTS-NPSi, the band at 1384 cm −1 is due to the -NH2 and -NH bending vibrations [31]. These indicate that organic functional groups have successfully grafted on the surface of porous silicon.
To determine the grafting amount of organic functional groups, thermogravimetric analysis (TGA) was used to verify it. Thermogravimetric diagrams of NPSi, GTS-NPSi and TEPA-GTS-NPSi are shown in Figure 3. Among them, the mass loss of NPSi, GTS-NPSi and TEPA-GTS-NPSi was 6.6%, 13.6%, and 20.7%, respectively. Considering the decomposition of the organic group and evaporation of water, the amounts of grafted GTS and TEPA groups were 7.0% and 7.1%, respectively. To determine the grafting amount of organic functional groups, thermogravimetric analysis (TGA) was used to verify it. Thermogravimetric diagrams of NPSi, GTS-NPSi and TEPA-GTS-NPSi are shown in Figure 3. Among them, the mass loss of NPSi, GTS-NPSi and TEPA-GTS-NPSi was 6.6%, 13.6%, and 20.7%, respectively. Considering the decomposition of the organic group and evaporation of water, the amounts of grafted GTS and TEPA groups were 7.0% and 7.1%, respectively.

Effect of pH on As(V) Adsorption
The effect of initial pH values on the As(V) adsorption capacity of TEPA-GTS-NPSi is shown in Figure 4. In the experimental pH range of 1.0-8.0, with the increase of pH value, the adsorption capacity of TEPA-GTS-NPSi shows a trend of first increasing and then stabilizing with the increase of pH. After pH = 4, the increasing trend weakened and reached stability near pH = 7. However, the adsorption capacity of the NPSi remains at a low level in all pH ranges.
The different As(V) species at different pH values are also shown in Figure 4. At the low pH value, part of As(V) was electrically neutral (H 3 AsO 4 ). As the pH increased, the main form of As(V) was H 2

Effect of pH on As(V) Adsorption
The effect of initial pH values on the As(V) adsorption capacity of TE is shown in Figure 4. In the experimental pH range of 1.0-8.0, with the i value, the adsorption capacity of TEPA-GTS-NPSi shows a trend of first i then stabilizing with the increase of pH. After pH = 4, the increasing trend reached stability near pH = 7. However, the adsorption capacity of the NP low level in all pH ranges.

Effect of pH on As(V) Adsorption
The effect of initial pH values on the As(V) adsorption capacity of TEPA-GTS-NPSi is shown in Figure 4. In the experimental pH range of 1.0-8.0, with the increase of pH value, the adsorption capacity of TEPA-GTS-NPSi shows a trend of first increasing and then stabilizing with the increase of pH. After pH = 4, the increasing trend weakened and reached stability near pH = 7. However, the adsorption capacity of the NPSi remains at a low level in all pH ranges.  The different As(V) species at different pH values are also shown in Figure 4. At the low pH value, part of As(V) was electrically neutral (H3AsO4). As the pH increased, the main form of As(V) was H2AsO4 − and HAsO4 2− , At this point, As(V) species are electronegative; however, the main form of amino in TEPA-GTS-NPSi is -NH3 + (pKa(NH3·H2O) = 9.3) [29,32], which is electropositive. In the adsorption process, H2AsO4 − and HAsO4 2− capture the H + ion in the -NH3 + ; thus, the As(V) species are adsorbed by TEPA-GTS-NPSi through electrochemical action.

Effect of the Contact Time and Adsorption Kinetics
Adsorption rate is an important parameter used to characterize the properties of adsorbent; fast adsorption is more beneficial to practical application. The effect of the adsorption time on the As(V) adsorption by TEPA-GTS-NPSi is presented in Figure 5. Obviously, As(V) adsorption by the adsorbent can be viewed in two stages as follows: a rapid adsorption stage and slow adsorption stage. The whole adsorption process consists of the first 10 min rapid phase, and the reaction reaches equilibrium at around 60 min. In the initial rapid phase, the adsorption rate is very fast, and the adsorption capacity is greatly increased, which is because adsorbent surfaces are rich in active sites, the mass transfer impetus caused by the concentration difference and the high chemical potential. After 30 min, the adsorption tended to be stable and the adsorption of adsorbent reached saturation. rapid adsorption stage and slow adsorption stage. The whole adsorption process consists of the first 10 min rapid phase, and the reaction reaches equilibrium at around 60 min. In the initial rapid phase, the adsorption rate is very fast, and the adsorption capacity is greatly increased, which is because adsorbent surfaces are rich in active sites, the mass transfer impetus caused by the concentration difference and the high chemical potential. After 30 min, the adsorption tended to be stable and the adsorption of adsorbent reached saturation. To evaluate the kinetic mechanism, two different kinetic models were employed to describe the adsorption process. The pseudo-first-order and pseudo-second-order models are represented as Equations (3) and (4): where the k1 (min −1 ) is the rate constant of pseudo-first-order, k2 [g/(mg·min)] is the rate constant of pseudo-second-order, qe (mg/g) and qt (mg/g) are the adsorption capacity at equilibrium time and adsorption capacity at time (t), respectively. The results of nonlinear fitting are also shown in Figure 5, and the kinetic constants and correlation coefficients (R 2 ) were listed in Table 1. Table 1. Comparison of the pseudo-first and pseudo-second-order constants.

Pseudo-First-Order Model
Pseudo-Second-Order Model k1 = 0.18633 min −1 k2 = 0.01953 g·mg −1 ·min −1 R 2 = 0.9909 R 2 = 0.9952 To evaluate the kinetic mechanism, two different kinetic models were employed to describe the adsorption process. The pseudo-first-order and pseudo-second-order models are represented as Equations (3) and (4): where the k 1 (min −1 ) is the rate constant of pseudo-first-order, k 2 [g/(mg·min)] is the rate constant of pseudo-second-order, q e (mg/g) and q t (mg/g) are the adsorption capacity at equilibrium time and adsorption capacity at time (t), respectively. The results of nonlinear fitting are also shown in Figure 5, and the kinetic constants and correlation coefficients (R 2 ) were listed in Table 1. Table 1. Comparison of the pseudo-first and pseudo-second-order constants.

Pseudo-First-Order Model
Pseudo-Second-Order Model k 1 = 0.18633 min −1 k 2 = 0.01953 g·mg −1 ·min −1 R 2 = 0.9909 R 2 = 0.9952 According to the fitting results, the correlation coefficient of the pseudo-second-order model is significantly higher than that of the pseudo-first-order model (R 2 = 0.9909). The pseudo-second-order model assumed that the determining step of the adsorption rate is a chemical reaction. This indicates that the adsorption of As(V) on TEPA-GTS-NPSi is mainly due to the chemical adsorption.

Adsorption Mechanism
To gain further insight into the As(V) adsorption mechanism of TEPA-GTS-NPSi, XPS analysis of TEPA-GTS-NPSi before and after As(V) adsorption was performed, as shown in Figure 6. According to the comparison of XPS spectra before and after adsorption, in the As3d spectra (Figure 6d) after adsorption, a spectrum peak of binding energy of 45.2 eV was found, which was corresponding to the peak of As3d 5/2 [33]. The appearance of this spectrum peak initially indicates that As(V) was successfully adsorbed on the TEPA-GTS-NPSi functionalized porous silicon adsorbents. As can be seen from the TEPA-GTS-NPSi N1s map (Figure 6c), after adsorption, the original -NH 2 corresponding to 398.5 eV and -NH 3 + corresponding to 400.5 eV were also detected [34]. In addition,-the strong peaks of NH 3 + and the significantly lower intensity peaks of -NH 2 after adsorption can clearly be seen (Figure 6c); this also shows that in the process of adsorption, TEPA-GTS-NPSi adsorbent of amino was the main form of -NH 3 + . Based on the above results, it is suggested that the main reason for the adsorption of As 5+ by TEPA-GTS-NPSI functionalized porous silicon under weak acidic conditions is the interaction between -NH 3 + and H 2 AsO 4 − to achieve the adsorption effect.
To gain further insight into the As(V) adsorption mechanism of TEPA-GTS-NPSi, XPS analysis of TEPA-GTS-NPSi before and after As(V) adsorption was performed, as shown in Figure 6. According to the comparison of XPS spectra before and after adsorption, in the As3d spectra (Figure 6d) after adsorption, a spectrum peak of binding energy of 45.2 eV was found, which was corresponding to the peak of As3d5/2 [33]. The appearance of this spectrum peak initially indicates that As(V) was successfully adsorbed on the TEPA-GTS-NPSi functionalized porous silicon adsorbents. As can be seen from the TEPA-GTS-NPSi N1s map (Figure 6c), after adsorption, the original -NH2 corresponding to 398.5 eV and -NH3 + corresponding to 400.5 eV were also detected [34]. In addition,-the strong peaks of NH3 + and the significantly lower intensity peaks of -NH2 after adsorption can clearly be seen (Figure 6c); this also shows that in the process of adsorption, TEPA-GTS-NPSi adsorbent of amino was the main form of -NH3 + . Based on the above results, it is suggested that the main reason for the adsorption of As 5+ by TEPA-GTS-NPSI functionalized porous silicon under weak acidic conditions is the interaction between -NH3 + and H2AsO4 − to achieve the adsorption effect.

Effect of Initial As(V) Concentration and Adsorption Isotherms
The effect of initial concentration on As(V) removal by TEPA-GTS-NPSi was investigated in Figure 7.

Effect of Initial As(V) Concentration and Adsorption Isotherms
The effect of initial concentration on As(V) removal by TEPA-GTS-NPSi was investigated in Figure 7. The equilibrium adsorption capacity increased with the increase of the concentration of As(V) at first. When the concentration was lower than 100 mg/L, the adsorption capacity increased with the increase of the concentration, but the trend of increase gradually slowed. When the concentration was higher than 100 mg/L, the saturated adsorption capacity finally reached (13.2 mg/g). This is because when the concentration of As(V) is low, the functional groups are not fully occupied by As(V), and there are still excess ligand functional groups that can continue to react with As(V). Therefore, the The equilibrium adsorption capacity increased with the increase of the concentration of As(V) at first. When the concentration was lower than 100 mg/L, the adsorption capacity increased with the increase of the concentration, but the trend of increase gradually slowed.
When the concentration was higher than 100 mg/L, the saturated adsorption capacity finally reached (13.2 mg/g). This is because when the concentration of As(V) is low, the functional groups are not fully occupied by As(V), and there are still excess ligand functional groups that can continue to react with As(V). Therefore, the adsorption capacity is relatively low. As the initial concentration of As(V) continued to increase, the active site of TEPA-GTS-NPSi continued to adsorb with As(V) through electrostatic interaction, so the adsorption capacity increased. By comparing with some previous reports in Table 2, the TEPA-GTS-NPSi hybrid materials derived from cutting waste show an acceptable adsorption capacity. Moreover, to evaluate the isotherms mechanism, Langmuir and Freundlich isotherm models were employed to describe the interaction between As(V) and TEPA-GTS-NPSi. The Langmuir isotherms model can be described by the Equation (5).
where q e (mg/g) is the equilibrium absorption capacity, C e (mg/L) is the concentration at equilibrium, q m (mg/g) is the maximum adsorption capacity and K L (L/g) is the equilibrium constant of the Langmuir isotherms model. The Freundlich isotherms model can be described by the Equation (6): where C e (mg/L) is the concentration at equilibrium, K F and n are the constants of Freundlich isotherms models. The results of nonlinear fitting are also shown in Figure 8, and the fitting parameters and correlation coefficients were listed in Table 3. Compared with the fitting coefficient (R 2 ) of the Langmuir and Freundlich adsorption isotherms models, the Langmuir model was more in line with the adsorption process of TEPA-GTS-NPSi adsorbent on As(V) than the Freundlich model. Moreover, the maximum adsorption capacity obtained by the Langmuir isotherms model was 13.87 mg/g, which was close to the experimental adsorption capacity of 13.2 mg/g. It also indicated that the adsorption of As(V) by TEPA-GTS-NPSi was mostly monolayer adsorption. Table 3. Constants of Langmuir and Freundlich isotherms.

Effect of Dosage and Adsorption Limit
To study the lower limit of adsorption of TEPA-GTS-NPSi adsorbent for As(V), with the same amount of As(V), different amounts of adsorbent were used for the experiment. The result is shown in Figure 8. At the low amount of adsorbent, with the increase of the amount of TEPA-GTS-NPSi adsorbent, the adsorption capacity slightly increased, but when the amount of adsorbent exceeded 25 mg, the adsorption capacity decreased. However, in the process in which the amount of adsorbent increased, the removal rate increased steadily and remained stable after the amount of adsorbent exceeded 25 mg. This is because at the low dosage, As(V) concentration is relatively high, and the concentration results in a high mass transfer impetus and high chemical potential; as a result, the adsorption quantity increases slightly. When dosage exceeds 25 mg, the residual concentration of As(V) in the solution was not enough, and the mass transfer impetus and chemical potential was low; thus, the adsorbent cannot absorb the remaining As(V). At this point, the concentration of As(V) in the solution was the lowest adsorption limit of TEPA-GTS-NPSi adsorbent, and the lower adsorption limit was 3 mg/L. (R 2 ) of the Langmuir and Freundlich adsorption isotherms models, the Langmuir model was more in line with the adsorption process of TEPA-GTS-NPSi adsorbent on As(V) than the Freundlich model. Moreover, the maximum adsorption capacity obtained by the Langmuir isotherms model was 13.87 mg/g, which was close to the experimental adsorption capacity of 13.2 mg/g. It also indicated that the adsorption of As(V) by TEPA-GTS-NPSi was mostly monolayer adsorption.  To study the lower limit of adsorption of TEPA-GTS-NPSi adsorbent for As(V), with the same amount of As(V), different amounts of adsorbent were used for the experiment. The result is shown in Figure 8. At the low amount of adsorbent, with the increase of the amount of TEPA-GTS-NPSi adsorbent, the adsorption capacity slightly increased, but when the amount of adsorbent exceeded 25 mg, the adsorption capacity decreased. However, in the process in which the amount of adsorbent increased, the removal rate increased steadily and remained stable after the amount of adsorbent exceeded 25 mg. This is because at the low dosage, As(V) concentration is relatively high, and the concentration results in a high mass transfer impetus and high chemical potential; as a result, the adsorption quantity increases slightly. When dosage exceeds 25 mg, the residual concentration of As(V) in the solution was not enough, and the mass transfer impetus and chemical potential was low; thus, the adsorbent cannot absorb the remaining As(V). At this point, the concentration of As(V) in the solution was the lowest adsorption limit of TEPA-GTS-NPSi adsorbent, and the lower adsorption limit was 3 mg/L.

Regeneration and Reuse of TEPA-GTS-NPSi
The reusability of adsorbent is the key factor to judge the value of adsorbent. Therefore, in this study, the adsorption-desorption in the removal of As(V) was investigated. The adsorbents were initially loaded with As(V), then eluted with HNO3 solution (0.1 mol·L −1 ), and then repeatedly utilized for As(V) capture. The results of the recyclability studies are shown in Figure 9.

Regeneration and Reuse of TEPA-GTS-NPSi
The reusability of adsorbent is the key factor to judge the value of adsorbent. Therefore, in this study, the adsorption-desorption in the removal of As(V) was investigated. The adsorbents were initially loaded with As(V), then eluted with HNO 3 solution (0.1 mol·L −1 ), and then repeatedly utilized for As(V) capture. The results of the recyclability studies are shown in Figure 9. The adsorption efficiency of TEPA-GTS-NPSi adsorbent was still good after five cycles. The adsorption efficiency was 95.6% after the first cycle and 91.8% after the fifth cycle. This is mainly because some functional groups on the surface were lost or became inactivated in the circulation. The results indicate that TEPA-GTS-NPSi has good regeneration capability for As(V) adsorption. The adsorption efficiency of TEPA-GTS-NPSi adsorbent was still good after five cycles. The adsorption efficiency was 95.6% after the first cycle and 91.8% after the fifth cycle. This is mainly because some functional groups on the surface were lost or became inactivated in the circulation. The results indicate that TEPA-GTS-NPSi has good regeneration capability for As(V) adsorption.

Practical Application of Industrial Wastewater
The application of adsorbent in actual wastewater is the most intuitive way to test the effect of an adsorbent. Thus, the TEPA-GTS-NPSi adsorbent was used for adsorption in actual wastewater which was obtained from a factory in Yunnan, China. The results were shown in Table 4. Most industrial wastewater contains As, Cu, and Zn. As can be seen from Figure 4, the adsorption capacity of TEPA-GTS-NPSI adsorbent for As was 12.37 mg/g, which reached 93.71% of the theoretical maximum adsorption capacity at about pH 7 (13.87 mg/g). The results show that the adsorption effect of TEPA-GTS-NPSI adsorbent on As is better than that on copper and zinc, and it has certain practicability in the treatment of industrial wastewater containing arsenic.

DFT Analysis
The geometric optimizations were conducted by density function theory (DFT) at B3LYP level with 6-311++G(d, p) basis. All optimization of electronic and geometrical structures was performed by Gaussian 09 suite of program. According to the result of pH and XPS, the optimized geometries of TEPA-GTS, H 3 AsO 4 and As-TEPA-GTS are shown in Figure 10.
It is can be seen that the H 3 AsO 4 group mainly adsorbed by -NH 2 group on the functional molecule after the adsorption process. Table 5 presents the selected structural parameter of optimized geometries on bond length.
According to the bond length, it is obvious that after the adsorption process, the bond lengths of N-H and O-H are getting longer; however, the bond length of O-H is shorter than that of before adsorption. This phenomenon can be explained by the charge action between TEPA-GTS functional molecule and H 3 AsO 4 , resulting in the bond length changes. Also, the bond length of N1-H6 is 1.64556 Å; this atomic distance indicates a strong hydrogen bond between them [39].
In addition, the NBO charge distributions are shown in Table 5 and the contour maps of As-TEPA-GTS are shown in Figure 11.
According to Table 6, the NBO charge of selected atoms has increased after the adsorption process; this indicates that there is a transfer of charge in these atoms which causes the adsorption. The electron density spread over N and O-H from Figure 10, which confirms the strong interaction of TEPA-GTS and H 3 AsO 4 [40]. Also, second-order perturbation theory analysis was applied to explore the interaction between donor atom and metal ion, and the corresponding E(2) stabilization energy was employed to measure the interaction between the donor and acceptor. During the adsorption process, the lone pair electrons of N are the donor, the antibonding orbital of H6-O5 is the acceptor, and the stability energy of E(2) of LP (N1) → BD × (H6−O5) is 40.75 kcal/mol. The results indicate that the interaction between the N atom and the H-O bond dominated the adsorption. All the results show that the adsorption mechanism is mainly due to the strong interaction between the terminal amino group and arsenic species. It is can be seen that the H3AsO4 group mainly adsorbed by -NH2 group on the functional molecule after the adsorption process. Table 5 presents the selected structural parameter of optimized geometries on bond length. According to the bond length, it is obvious that after the adsorption process, the bond lengths of N-H and O-H are getting longer; however, the bond length of O-H is shorter than that of before adsorption. This phenomenon can be explained by the charge action between TEPA-GTS functional molecule and H3AsO4, resulting in the bond length changes. Also, the bond length of N1-H6 is 1.64556 Å; this atomic distance indicates a strong hydrogen bond between them [39].
In addition, the NBO charge distributions are shown in Table 5 and the contour maps of As-TEPA-GTS are shown in Figure 11.   According to Table 6, the NBO charge of selected atoms has increased after the adsorption process; this indicates that there is a transfer of charge in these atoms which causes the adsorption. The electron density spread over N and O-H from Figure 10, which confirms the strong interaction of TEPA-GTS and H3AsO4 [40]. Also, second-order perturbation theory analysis was applied to explore the interaction between donor atom and metal ion, and the corresponding E(2) stabilization energy was employed to measure

Conclusions
A new amino-functional adsorbent was prepared by copper-assisted chemical etching of kerf loss silicon waste, and shows effective removal of As(V) from aqueous solution. The TEPA-GTS-NPSi adsorbent can quickly reach adsorption equilibrium after 30 min and the maximum adsorption capacity of As(V) is 13.2 mg/g at room temperature. The adsorption kinetics could be well described with the pseudo-second-order kinetic model, which indicates the adsorption process was controlled by chemical interactions. The adsorption isotherms could be well described with the Langmuir isotherms model, which revealed the adsorption process was mono-layer adsorption. TEPA-GTS-NPSi could be regenerated with 1 mol/L HCl washing, which resulted in a low adsorption capacity attenuation after five cycles. The XPS and DFT analysis showed that the mechanism of As(V) removal by TEPA-GTS-NPSi is mainly due to the strong hydrogen bonding between the -NH 2 group and arsenic species. This work could provide novel insight into new absorbent fabrication and shows extremely practical significance to recycle kerf-loss silicon.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/coatings11070850/s1, Figure S1: Silicon solar cell industry chain form solar grade silicon to solar module.