Characterization of surface changes on silicon and porous silicon after interaction with hydroxyl radicals

2.1 Electrolytic cell and electrodes

The electrodes used in this research were monocrystalline n-Si (1 0 0) with a resistivity of between 0.1 and 3°Ω cm (N_D = 5 × 10¹⁵ cm⁻³); the electrodes had a thickness of 500–550 μm and polished/etched surfaces (Silicon Valley Microelectronics, CA, USA), and they were doped with phosphorous. The silicon wafer was cut into rectangles (3.0 × 1.0 cm²), which were first degreased with acetone by boiling for 10 min. Afterward, the electrodes were treated in the following sequential order: ultrasonication for 10 min in acetone, ethanol and water. Then, they were treated for 10 min with a mixture of H₂SO₄:H₂O₂ 3:1 at 80 °C, which eliminates any trace of heavy metals or organic species on the substrates. Next, the treated substrates were immersed in 2 M HF solution for 2 min to remove the oxide formed in the previous stages. Finally, the electrodes were rinsed with ultra-pure water. Once the electrodes were cleaned, they were mounted on a Teflon support and ohmic contact was made on the etched face. The silicon area exposed to the solution was 1.00 cm². Before each experiment, the electrode was submerged for 2 min in 2 M HF solution; the aim of this procedure was to eliminate the native oxide and to obtain a smooth and finished hydrogen-terminated surface (Si–H). For each measurement, a new n-Si (1 0 0) electrode was used, to achieve reproducibility of the initial conditions.

All the electrochemical measurements were made in a 3-electrode electrochemical cell, using platinum wire as the counter electrode and a mercuric sulfate/mercury electrode (MSE) (Hg/Hg₂SO₄, K₂SO₄ (saturated), 0.640 V vs. NHE) as the reference electrode. All the potentials reported in this study are with respect to this reference electrode. The electrolytic solutions were prepared using distilled and deionized water (Millipore) with a resistivity of 18 MΩ cm and analytical grade Merck reagents (Hohenbrunn, Germany) (NH₄)Fe(SO₄)₂ × 12H₂O, H₂O₂, Na₂SO₄, H₂SO₄, and (NH₄)₂Fe(SO₄)₂ × 6H₂O y HF).

2.2 Synthesis of porous silicon (PS)

The synthesis of PS was carried out through anodic etching of n-type crystalline silicon in an aqueous electrolyte of ethanol and hydrofluoric acid (HF). A silicon rectangle (3.0 × 1.0 cm²) was placed on a Teflon support with a piece of copper as the contact on the back, while a platinum wire was used as the counter electrode; both were positioned on a Teflon cell. The cell was filled with a 3:1 (v/v) mixture of absolute ethanol and aqueous HF at 48%. A (positive) current density of 10 mA cm⁻² was applied to both electrodes for 10 min under illumination. Once the etching was finalized, the sample was washed several times with ethanol and dried under an argon flux. This produces a film with a thickness of a few microns. Similar results regarding the formation and growth of pores in silicon have been found by other authors (Lehmann and Rönnebeck, 1999; Labunov et al., 1979; Zhang et al., 1989; Rivera et al., 2002).

2.3 Generation of hydroxyl radicals

The hydroxyl radicals used in this study were generated via 2 methodologies: (i) Fenton’s reagent, which is composed of 4 mM Fe²⁺ (NH₄)₂Fe(SO₄)₂ × 6(H₂O), 4 mM of ethylenediaminetetraacetic acid (EDTA), 40 mM of acetate buffer (pH 4.7; 0.01 M) and 40 mM of H₂O_2, in the tests using this method, Fe²⁺, EDTA and acetate buffer are first added to the system in a primary stage and once the experiment is mounted, the H₂O₂ is added, thus initializing the reaction; (ii) an electro-Fenton treatment, obtained from mixing 0.2 M of (NH₄)Fe(SO₄)₂ × 12H₂O, 0.1 M H₂SO₄, and 0.05 M Na₂SO₄, and saturating the solution with oxygen. The reagents are added to the electrochemical cell, and an electric current with a density of −10 mA cm⁻² is passed through the cell. This results in the electroreduction of oxygen, which generates hydrogen peroxide in situ. Afterward, the generation of radicals occurs as a product of the reaction of hydrogen peroxide with Fe²⁺.

2.4 Contact angle measurements

To study the change in the surface properties of silicon and porous silicon through contact angle measurements, the electrodes were immersed in the ^•OH radical-generating solution, either with Fenton’s reagent or with the electrochemical in situ generation of hydrogen peroxide (the electro Fenton process).

The contact angle measurement was made via 2 procedures: (i) sessile drop, where 1.0 μL of Millipore-Q water in the shape of a droplet was deposited on silicon or porous silicon; (ii) in situ captive bubble, where the electrodes were mounted on epoxy resin and then positioned in the electrolytic cell, and a bubble of argon was added on the surface of the electrodes. For each contact angle measurement, the angle is measured using a goniometer by taking the internal drop-surface angle into consideration. However, for highly hydrophobic (using the sessile drop) or highly hydrophilic (using the argon bubble) surfaces, the external angle was measured, and a correction was made by considering the supplement of the angle measured.

2.5 Cyclic voltammetry (CV)

CV data were obtained at a scan rate of 100 mV/s on an AUTOLAB potentiostat/galvanostat device [Model: PGSTAT 302]. The solution was purged of oxygen by bubbling argon through it for 30 min. The experiment was then maintained under the inert atmosphere of the Ar gas.

2.6 Photoluminescence spectra

The photoluminescence measurements of PS were made in situ in different solutions: air, water, hydrogen peroxide, and Fenton’s reagent. For these measurements, a laser source of 375 nm was used. The spectra were registered using an ALS optical spectrometer [Model: SEC2000-UV/VIS].

2.7 Intensity-modulated photocurrent spectroscopy (IMPS) and intensity-modulated photovoltage spectroscopy (IMVS)

IMVS and IMPS measurements were made in a two-electrode cell. c-Si was used as the working electrode, and a platinum wire was used as the counter electrode. The solutions used in the IMVS and IMPS spectra were H₂O₂ and Fenton’s reagent. The mounting for the IMVS and IMPS measurements consists of an electrochemical cell, a 5 mW light source (LED or laser) [Model: Stocker Yale] modulated through a Global Specialities Instruments signal generator [Model: 2030] and two Signal Recovery dual-phase lock-in amplifiers [Model: 5210]. The latter detected the in- and out-of-phase signals of both the optical excitation (LED or laser) and the photovoltage and photocurrent response of the cell. The IMPS measurement is conducted under short-circuit conditions. The IMVS measurement is conducted with the same light oscillation but under open circuit conditions.

2.8 Atomic Force Microscopy (AFM)

An atomic force Nanoscope Digital microscope [Model: Digital IIIa] was used for the morphological characterization of the samples by employing tapping mode AFM. Silicon nitride tips were used for this measurement, with a resonance frequency between 200 and 400 kHz and a radius of curvature of between 5 and 10 nm.

3.1 Morphological characterization of PS

The PS samples were characterized via Atomic Force Microscopy (AFM). The results obtained are shown in Fig. 1.

Close

Fig. 1 AFM images for the formation of pores in c-Si in a solution of HF:EtOH 1:1, the sample is obtained at J = 10 mA/cm² and the duration of the formation process is 300 s. Ex situ images, (a) 65 × 65 μm zone 1; (b) 10 × 10 μm zone 2; (c) image zoom of (b), 5 × 5 μm; (d) image zoom of (b), 3 × 3 μm.

Export to PPT

From Fig. 1, we observe that the formation of pores is feasible in these conditions and on these substrates. However, there is a random distribution of pores. Through section analysis with the AFM software, the diameter of the pores was obtained with values between 171 and 486 nm. This corresponds to macroporous silicon, according to the classification found in the bibliography (Gorostiza et al., 1996; Peter et al., 1995; Palacios et al., 2005; Gorbanyuk et al., 2006).

3.2 Contact angle measurements

Fig. 2 shows the changes in the drop contact angle values on the c-Si and PS substrates when they were exposed to different solutions, i.e., hydrogen peroxide and Fenton’s reagent.

Close

Fig. 2 Variation of the contact angle as a function of the immersion period for c-Si (□) and PS ( ), in solutions of H₂O₂ 1 mM ((a) and (b)) and Fenton’s reagent ((c) and (d)). The inset shows the variation of the contact angle for c-Si with a water immersion cleaning treatment.

Export to PPT

The inset of Fig. 2a shows that the surface change occurs with silicon because immersion in water produces a decrease of the contact angle from 75° to 62° after 40 min. This indicates a surface transformation from a hydrophobic to a hydrophilic surface due to the thermodynamic instability of silicon substrates in aqueous solutions. It can be observed in Fig. 2a and b that there is a decrease in the values of the contact angles for both substrates after immersion in H₂O₂. After 40 min of immersion in the solution, there is a decrease from 70° to 53° for c-Si, while for PS the decrease is from 132° to 80°, which indicates a larger increase in the degree of hydrophilicity for H₂O₂ in comparison with pure water. Finally, when the substrates are immersed in Fenton’s reagent, there is a decrease of the contact angle for c-Si from 78° to 33°, while the contact angle for PS decreases from 110° to 22° after 40 min of immersion, as observed in Fig. 2c and d, respectively. As is evident, although the same results can be observed for the H₂O and H₂O₂ tests, the variation of the contact angle in the presence of radicals occurs within a short period of time and with an increase in the degree of hydrophilicity; on one hand, this indicates a greater coverage with hydrophilic functional groups, i.e., hydroxyl groups, and on the other hand, this indicates surface modification at a greater speed.

To contrast the results obtained from the interaction of radicals from the Fenton’s reagent with the c-Si and PS substrates, the electro-Fenton procedure was performed to obtain hydroxyl radicals. In an initial stage, the electrochemical behavior of c-Si and PS in the electrolytic medium was observed; this confirmed that hydrogen discharge accompanies the reduction process of O₂ and Fe³⁺. Because of this discharge, a current density of −10 mA cm² was applied for 30 s to ensure the necessary production of hydroxyl radicals in the reaction and to prevent the exhaustion of the oxygen content in the interface; after this period, the contact angle was measured. The methodologies employed were ex situ (sessile drop) and in situ (captive bubble) methods. The results obtained for both substrates are shown below in Fig. 3.

Close

Fig. 3 Contact angle measurements for working electrodes immersed in radicals produced by the electro-Fenton process as a function of time. Ex situ (sessile drop) (a) PS; (b) c-Si. In situ (captive bubble) (c) c-Si; (d) PS; ( ) Blank (no current applied); (□, ). A 10 mA current was applied.

Export to PPT

According to Fig. 3a and b, an important decrease in the contact angle can be observed when the substrates employed are exposed to radicals coming from the electro-Fenton procedure. In effect, for c-Si, a decrease from approximately 80° to 40° occurred when the electrode was immersed for 300 s (5 min). The contact angle on the PS substrate decreased from 115° to 35° after the same exposure time. The rapid decrease of the contact angle indicates that the interaction between the substrate and the hydroxyl radicals occurred in a short period of time. In addition, at these current values, a high quantity of these radicals should be observed. These results are consistent with those obtained with Fenton’s reagent (Fig. 2).

The contact angle values measured through the captive bubble (in situ) methodology can also be observed in Fig. 3c and d. The values obtained when current was not applied for the electroreduction of oxygen (blank) are shown with blue circles. A small variation in the bubble/substrate contact angle can be observed, but this is not a significant surface modification and can be attributed to the instability of silicon in aqueous solutions. When a reduction current of 10 mA was applied (the black and red squares for c-Si and PS, respectively), a significant increase is observed in the contact angle values, which is similar to the results obtained via the sessile drop method.

As has been observed, the interaction of c-Si and PS with hydroxyl radicals produces a surface with a greater degree of hydrophilicity. This is confirmed by the decrease of the contact angle for the sessile drop method and the increase of the contact angle for the captive bubble method. This interaction is determined by the affinity that exists between the water molecule dipoles and the functional groups on the surface of the silicon substrate. Therefore, the following can be proposed:

1. After the cleaning treatment of c-Si and the formation process of PS, the predominant surface coating on both substrates is Si–H. This surface coating is considered to be of a hydrophobic nature (Grundner and Jacob, 1986), which involves high contact angle values for the sessile drop procedure. The contact angle value for c-Si is approximately 80°. This hydrophobic nature is due to the hydrogen coating stabilizing the surface for c-Si and is because of the small difference between the electronegativity of Si (1.8) and H (2.1). There is no polarization produced between the surface Si–H bonds and the bonds of the main Si–Si material. In these conditions, c-Si exhibits stability for some hours. On the other hand, PS has contact angles above 100°. These contact angle values, which are larger than the values for c-Si, are partly due to the previously explained Si–H hydrophobic coating and partly due to the effects of the morphology of the pores obtained, i.e., pores with diameters under 1.0 μm (Ressine et al., 2007). This behavior has been defined in a model proposed by Cassie (Cassie and Baxter, 1944) and is related to the reduction in the contact surface between the liquid and the solid, and so the drop sits on ‘pillows of air’ due to the discontinuity of the surface (the tip of pore/pore). This is known as a slippery state. The topographic measurements of PS obtained via AFM confirm this fact. In effect, according to Fig. 1, the distribution of the sample pore diameters ranges between 400 and 700 nm, which corroborates the presence of a porous structure associated with contact angles above 100° for the sessile drop method.

2. When c-Si and PS are submerged in the hydroxyl radical-generating solutions, i.e., Fenton and electro-Fenton solutions, there is a dynamic modification of the surface groups from Si–H to Si–OH. For crystalline silicon, the surface modification provokes the polarization of the Si–OH and Si–Si surface bonds, according to the following diagram:

Close

Export to PPT

It is understood that the presence of Si–OH groups induces the appearance of dipoles that allow an important interaction with the water molecule dipoles in the sessile drop. Thus, the surface presents a hydrophilic nature (contact angles of approximately 40° after 40 min of exposure to hydroxyl radicals from Fenton’s reagent and 30° after 300 s of the electro-Fenton treatment).

In accordance with the results obtained, the interfacial solid-liquid free energy (ΔG_SL) was determined; this is associated with the degree of wettability of the liquid in the solid used. The Young-Dupre equation (Savov, 1997), W_A = γ_LV[1 − cos (θ)] = −ΔG_SL, was used to evaluate this wettability by relating the contact angle with the interfacial free energy. Using the water/silicon contact angle values before the exposure to hydroxyl radicals and the value obtained after the longest exposure time, the surface tension value for water in the liquid-vapor interphase γ_LV = 72.8 mJ m⁻², the following table includes the results for each test.

According to the results shown in Table 1, the substrates that exhibit greater hydrophobicity (low wettability) correspond to the c-Si and PS surfaces after the cleaning treatment. When the substrates were submerged for 40 min in the hydroxyl radical-generating systems, a higher hydrophilicity was achieved, and with this, higher adhesion energy values. This can be explained by considering that the substrates initially only interact with water molecules through dispersion forces, i.e., London and van der Waals forces, among others. Then, when the substrate surface is modified with hydroxyl groups, apart from interactions due to dispersive forces, dipole-dipole and hydrogen bridge interactions appear, which increases the substrate/liquid adhesion.

3.3 Optical characterization of PS in various media through photoluminescence measurements

Photoluminescence (PL) measurements were conducted for the optical characterization of porous silicon in different media, which allowed the determination of the influence of environment on the intensity and shift of the optical signal and thereby, the nature of the PS surface. To quantify these shifts observed in the PL spectra, the relative changes in the normalized PL intensity area were analyzed as a function of wavelength. The following means were used to assess these changes: air, water, hydrogen peroxide and Fenton solution. Fig. 4 shows the results.

Close

Fig. 4 Normalized spectra of photoluminescence intensity (left graphs) and relative changes in photoluminescence intensity (right graphs) for PS: (a) Air; (B) Water; (C) H₂O₂; (D) Fenton’s reagent, at different exposure times. (—) 0 s; ( ) 30 s; ( ) 90 s; ( ) 300 s; ( ) 600 s; ( ) 900 s; ( ) 1170 s.

Export to PPT

Fig. 4 shows that PS, without exposure to the media used, has a PL maximum at approximately 670 nm. Then, when PS was immersed in the utilized media, there was a change in the PL signal versus the exposure time (Fig. 4a–c), i.e., a weakening PL signal as a result of a surface oxidation process. However, when PS was immersed in Fenton’s reagent, it was also possible to observe a spectrum shift to higher wavelength values as a function of the exposure time. This redshift can be attributed to the presence of interactions between the PS surface and the free radicals generated, which cause surface oxidation from (PS–H) to (PS–OH). This behavior has been reported by other authors (Chuan and Tsu, 1994; Gösele and Lehmann, 1994), who observed a shift toward longer wavelengths after immersion of PS in a propanol solution irradiated in a UV digester. According to the authors, this procedure causes a “mild oxidation” on the PS surface, resulting in the appearance of PS–OH bonds. Meanwhile, a measure of the spectral shift observed after PS immersion in different media was obtained from the relative changes in the photoluminescence intensity normalized by area. These results are obtained from the normalized photoluminescence spectra (right graphs). Therein, we can observe that the PS/Fenton system, in particular, experiences the largest change in the PL signal at longer wavelengths, which indicates a more dynamic modification compared to other systems. The systems submerged in water and H₂O₂ have slight variations to blue emissions, while the PS in air has a slight variation of the spectrum to red emissions. The changes in the PL for each of the studied systems are summarized in Table 2.

As shown in Table 2, when PS was exposed to air and hydrogen peroxide, the total amplitude of the changes in the PL intensity was 0.6 and 0.55, respectively. In the case of PS immersed in water, it was not possible to estimate the change in the amplitude of the PL from the results obtained, which leads to the conclusion that these changes are negligible at the time of the measurement. For the PS/Fenton system, the total amplitude change is 2.04, which indicates a more important surface modification when the substrate is exposed to the attack of radicals than when PS is exposed to the aforementioned media.

3.4 Study of charge collection efficiency by intensity-modulated photocurrent spectroscopy (IMPS) and intensity-modulated photovoltage spectroscopy (IMVS)

One of the parameters that can be determined by the relationship between IMPS and IMVS is the charge collection efficiency (η_cc). This parameter gives information about the fraction of the substrate charge carriers that are not involved in the recombination process. The measure of IMPS provides kinetic information of the charge under short circuit conditions. Meanwhile, the IMVS measurement is performed with the same light disturbance but under open circuit conditions, which is associated with the lifetime of recombination of charge carriers. The electron transport time (τ_t) can be estimated using Eq. (1):

Close

Fig. 5 Charts for ( ) IMPS; ( ) IMPS measurements. c-Si/H₂O₂ systems (a and b); c-Si/Fenton system (c and d).

Export to PPT

According to the results obtained by each technique, the values f_t and f_r enable us to calculate the values of τ_t for the IMPS measurements and values of τ_r for IMVS measurements. With these values and through the use of Eq. (3), it is possible to obtain the charge collection efficiency (η_cc). The values obtained are summarized in Table 3:

As observed from the values obtained in Table 3, the charge collection efficiency is higher for the H₂O₂ solution. As this parameter is associated with the load fraction that does not participate in the recombination process, in the H₂O₂ system, a higher percentage of the photogenerated charge would participate in faradaic processes at the interface, or transport in the circuit. This was not the case for the c-Si/Fenton system, where a higher percentage of the photogenerated charge participates in a recombination process. The former could be related to the results found in the capacitance voltage measurements (not included), i.e., immersing the silicon substrate in radicals causes an increase in the density of the semiconductor surface states (Navarrete et al., 2014). Because these states are load and recombination entrapment centers, a higher density of these states increases the recombination process, which would result in the lowest charge collection efficiency value observed.