Output power density enhancement of an intermittently contacted metal–semiconductor junction with a water interlayer

Abstract

It has been reported that an intermittently contacted metal–semiconductor junction could harvest vibration energy. However, how to increase the energy conversion efficiency is still challenging. Herein, we find that the output power density of an intermittently contacted gold–silicon (Au–Si) junction can be increased by about 3 orders of magnitude after introduction of a water (H₂O) interlayer between the Au and Si friction layers. Our results demonstrate that the electric double layer and built-in electric field in the Au–H₂O–Si junction induce large values of output current and transferred charges and thereby contribute a high output power. Through studies of the output electric characteristics of the intermittently contacted Au–H₂O–Si junctions with different wettabilities of the Si surfaces, we reveal that the hydrophobic Si surface has the best energy conversion efficiency due to the continual contact electrification at the dynamic H₂O–Si interface.

---

1 Introduction

Due to the great demand for integration of electronic devices with self-power sources or/and rechargeable energy storage devices, micro-electric generators have emerged as a promising self-power source.¹ In the past decade, tremendous research interest has been dedicated to triboelectric nanogenerators (TENGs) as a new technology for mechanical-to-electric power conversion.^2,3 A conventional TENG is composed of a pair of friction layers (at least one of them contain(s) an insulating frictional layer). The contact of the two friction layers induces electron transfer from one to the other contacting the surface due to their electronegativity difference, so that the two contacting surfaces are charged with equally opposite charges. In the ensuing separation and approaching stages, electrons transfer back and forth between the backside metallic pad of the two contacting surfaces through the external circuit due to electrostatic induction, realizing the mechanical-to-electric power conversion.⁴ As electrons could not pass across the insulating surface, the current between the two friction layers is a displacement current, rather than a conduction current.⁵

In 2018, the intermittently contacted semiconductor–semiconductor or metal–semiconductor junction using a pair of friction layers with distinct chemical potentials was reported to harvest vibration energy.⁶ When the two friction layers are brought into contact, electrons diffuse from the higher chemical potential surface to the lower one passing across the contact surfaces, leading to the formation of a space charge region. Once the friction layers are separated by mechanical power, these diffused electrons are pumped out to the external circuit, yielding a transient current and, at the same time, a displacement current in between the two friction layers to maintain the current continuity. While, in the approaching stage, apart from a displacement current in between the two friction layers, the space charge regions can also be restored through diffusion of electrons across the contact surfaces of the two friction layers. Thus, the output current of the intermittently contacted semiconductor junction is dominated by discharging the diffused electrons to the external circuit caused by separation of the two surfaces, while the space charge restoration occurs through both electrostatic induction and electron diffusion processes. This characteristic is distinct from that of the conventional TENG. Although the intermittently contacted semiconductor junctions could harvest vibration energy, a low output power density of about 0.1 μW cm⁻² is still challenging.^6,7 Thus, how to improve the energy conversion efficiency is worth studying in-depth.⁸

Recently, a lot of studies have been carried out to improve the performance of TENGs using dynamic liquid–solid interface(s).^9–13 Thus, in this work, we purposely introduced deionized (DI) water as an interlayer between the gold (Au) and silicon (Si) friction layers to construct an intermittently contacted Au–H₂O–Si junction. It was found that the output power density of an intermittently contacted Au–Si junction can be increased to about 0.3 mW cm⁻² after introduction of a H₂O interlayer between the Au and Si surfaces, which is about 3 orders of magnitude higher than the prior reported values.^6,7 In addition, our work has the following three major findings. Firstly, the impact of the electric double layer (EDL) at the Au–H₂O–Si junction is similar to that of the space charge region at the Au–Si junction. Secondly, the EDL and built-in electric field in the Au–H₂O–Si junction induce large output currents or large amounts of transferred charges, hence a higher output power. Thirdly, the hydrophobic Si surface has the best energy conversion efficiency due to the continual contact electrification at the dynamic H₂O–Si interface. This work focuses on the physical understanding of mechanical-to-electric power conversion with the intermittently contacted Au–H₂O–Si junction. Our results could serve as a guide to design dynamic semiconductor junction-based sensors and electric generators in the future.

2 Results and discussion

2.1 Electric output of the intermittently contacted Au–Si junctions

In our experiments, intermittently contacted n-type Si and Au friction layers were employed as an experimental platform, as shown in Fig. 1a. All Si wafers were purchased from Nanjing MKNANO Tech. Co., Ltd (https://www.mukenano.com). The doping concentration and thickness of the n-type Si friction layer were about 1 × 10¹⁸ cm⁻³ and 541 μm, respectively. The back side of the Si friction layer was coated with a bilayer of 20 nm titanium and 50 nm gold as a back electrode. In order to fix the Si friction layer on the holder, a 2 mm thick acrylic layer was attached to the back of the back electrode. The Au friction layer was fabricated by coating a 1 μm thick Au layer onto a 2 mm thick acrylic substrate using a magnetron sputtering system. The effective contact area of the two friction layers was about 2 cm × 2 cm. The translational stage was precisely controlled by a computer through a linear motor. For hydrophilic and hydrophobic treatment of the Si surface, we purposely chose two pieces of newly cleaned Si wafers for coating a layer of hydroxide radical and fluoroalkyl silane, respectively.^14,15 Fluoroalkyl silane is an insulating material. The thickness of the fluoroalkyl silane layer should not be thicker than 2 nm on the Si surface. If the thickness of the fluoroalkyl silane layer is too high, it may be difficult for the electrons to tunnel through it and then the current generated becomes very small. The Kelvin Probe Force Microscopy (KPFM) images of the Si and Au surfaces are shown in Fig. 1b. The hydroxide radical on the Si surface could introduce a layer of positive charges, attracting electrons from the Si bulk to its surface, making the potential of the hydrophilic Si surface more negative. However, a layer of fluoroalkyl silane on the Si surface could introduce a layer of negative charges, which repulsed electrons from the Si surface to its bulk. As a result, the potential of the hydrophobic Si surface was more positive. From our KPFM measurement results and the work function of Au (W_Au ∼ 5.1 eV), the work functions of the hydrophilic, ordinary and hydrophobic Si surfaces (W_Si) were measured to be about 4.4 eV, 4.8 eV and 5.2 eV, respectively, as shown in Fig. S1 (ESI).†Fig. 1c shows the current–voltage curves measured at the contacted stage of the Au–Si junctions. To fit the current–voltage characteristics, the thermionic emission (TE) model was applied:¹⁶where the Au–Si junction contact area A = 4 cm², elementary charge q = 1.6 × 10⁻¹⁹ C, effective Richardson constant A* = 252 A cm⁻² K⁻², Boltzmann constant k = 1.38 × 10⁻²³ J K⁻¹, Planck constant h = 6.6 × 10⁻³⁴ J s and temperature T = 300 K. According to the TE model, the values of the ideality factor n, Schottky barrier height ϕ_B and contact resistance R_C from the three types of Au–Si junctions were fitted, as shown in Table 1. From the Au–Si junctions, the fitting results reveal that the values of n and ϕ_B were about 1.1 and 0.8 eV, which suggests that a good Schottky contact was established. It should be noticed that the value of R_C is just slightly larger than 1 kΩ. This small value of R_C makes the electrons diffuse easily across the contact surfaces of the Au–Si junction. The gap width d between the Au and Si surfaces in response to the time is shown in Fig. S2 (ESI†). In the approaching stage, the two surfaces got in touch from d ∼ 52 mm down to ∼0 in 0.3 s ① and remained in contact for 1 s in the contacted stage ②. Then, the two surfaces were separated from d ∼ 0 up to ∼52 mm within 0.3 s in the separating stage ③.Fig. 1d and e show the open-circuit voltage and short-circuit current of three Au–Si junctions. In the measurement, the Au and Si surfaces were connected with the positive and negative probes of the system source meter, respectively. It can be seen in Fig. 1d that the open-circuit voltages from the hydrophilic and ordinary Si surfaces were about −0.70 V and −0.59 V, while it became about 0.22 V from the hydrophobic Si surface. This explains that the polarity and magnitude of the open-circuit voltage were dominated by the chemical potential difference between the Au and Si surfaces. Fig. 1e exhibits the short-circuit currents of the three types of Au–Si junctions. The output current at the approaching and separating stages was denoted as I_A and I_S, respectively. The charges collected during I_A and I_S could be calculated as follows:where t_A and t_S are the time duration of I_A and I_S, respectively. As shown in Table S1 (ESI),† |Q_A| was lower than |Q_S| for all Au–Si junctions, indicating that the amount of electrons pumped out to the external circuit during the separating stage was higher than that electrostatically induced in the space charge region during the approaching stage. This result suggests that the space charge region was not fully restored in the approaching stage and then the remaining space charge region restoration would proceed with electron diffusion after the two surfaces were contacted, while the number of electrons diffused was not measurable in the external circuit.^6,7 Taking the Au and Si surfaces as an ideal contact, the number of stored charges was calculated aswhere N is the doping concentration, ε₀ is the vacuum dielectric constant, ε_r is the relative dielectric constant of the Si surface and V_B is the chemical potential difference between the Au and Si surfaces. Clearly, in Table S1 (ESI),† the value of |Q_S| from the intermittently contacted Au–Si junction was about 4 orders of magnitude less than the Q from the ideal Au–Si junction. In the nonideal solid–solid contact, a small air gap of several nanometres in the contacted surfaces could result in a sharp decrease of charge stored in the space charge region.¹⁷ In addition, the high surface charge density could significantly affect the space charge region width.¹⁷ As a result, the intermittently contacted Au–Si junctions usually have a low energy conversion efficiency. The output currents of the Au–Si junctions with different external loads are shown in Fig. S3 (ESI).† Obviously, the output currents were dominated by discharging the diffused electrons to the load; in other words, it was one polarity dominated, especially under a large external load. Fig. 1f shows the peak power density of the Au–Si junctions under different external loads. It can be determined that these Au–Si junctions had the maximum peak power density of about 0.49 μW cm⁻² and an impedance of about 50 MΩ.

Fig. 1 Experimental schematic illustration and output electric characteristics of the intermittently contacted Au–Si junctions. (a) The schematic setup for the experiments of the intermittently contacted Au–Si junctions. (b) The Kelvin probe force microscopy images of the Si and Au surfaces. (c) The current–voltage curves and fitting results of the Au–Si junctions. (d–f) The open-circuit voltage, short-circuit current and peak power density of the Au–Si junctions.

Table 1 Fitting results of the ideality factor (n), Schottky barrier height (ϕ_B) and contact resistance (R_C) from the Au–Si and Au–H₂O–Si junctions

Junctions		n	ϕ B (eV)	R C (Ω)
Au–Si	Hydrophilic	1.09	0.85	1028.01
	Ordinary	1.11	0.80	1431.69
	Hydrophobic	1.07	0.81	1705.26
Au–H2O–Si	Hydrophilic	1.30	0.83	3987.50
	Ordinary	1.16	0.52	2664.60
	Hydrophobic	1.03	0.79	2019.75

---

The pristine current generation mechanism of the intermittently contacted Au–Si junction can be illustrated through the energy band diagrams shown in Fig. 2. The Au surface has a lower chemical potential (or a larger work function) and the n-type Si surface has a higher chemical potential (or a smaller work function). If the surface charges, surface electronic states, and adsorbed impurities of the Si surface can be ignored, the energy band diagram of Si and Au before connection is sketched in Fig. 2a. In the approaching stage (Fig. 2b), electrons must move from Si to Au due to the chemical potential difference, causing a positive transient current I_A flowing from Au to Si through the external circuit. After the two surfaces come in contact, the space charge region can be formed through diffusion of electrons from Si to Au, accompanied by establishment of the Schottky junction till thermal equilibrium is reached. Under thermal equilibrium, no current flows in the external circuit, and the energy band diagram is shown in Fig. 2c. Once the two surfaces are separated with mechanical power, the accumulated electrons on the Au surface are pumped out to the external circuit, yielding a negative transient current I_S (Fig. 2d). For the hydrophilic Si surface, the energy band bends downward as some positive charges are introduced onto the surface by the hydroxide radical layer (Fig. 2e). Hence, near the Si surface, the Fermi level is closer to the bottom of the conduction band. As the chemical potential of the Au surface is still lower than that of the hydrophilic Si surface, the direction of output current from the hydrophilic Au–Si junction is consistent with the ordinary Au–Si junction (Fig. 2f and h). However, for the hydrophobic Si surface, the energy band bends upward and the Fermi level is closer to the top of the valence band because the fluoroalkyl silane layer could introduce some negative charges (Fig. 2i). As a result, the hydrophobic Si surface has a lower chemical potential than the Au surface, which causes the direction of output current from the hydrophobic Au–Si junction to reverse with the ordinary and hydrophilic Au–Si junctions (Fig. 2j and l).

Fig. 2 Energy band diagrams of the intermittently contacted Au–Si junctions. (a–d) The ordinary Si surface. (e–h) The hydrophilic Si surface. (i–l) The hydrophobic Si surface (E₀ is the vacuum level, E_C is the bottom of the conduction band, E_V is the top of the valence band, and E_F1 and E_F2 are the Fermi levels of the Au and Si layers, respectively).

2.2 Electric output of the intermittently contacted Au–H₂O–Si junction

In order to investigate the influence of water on the electric output characteristics of the intermittently contacted Au–Si junction, 1 μL of DI water (conductivity less than 10 μS cm⁻¹) was placed onto the middle of the Si surface to construct an Au–H₂O–Si junction, as shown in Fig. 3a. The DI water was tested to be electrically neutral (Fig. S4, ESI†). The water contact angle (WCA) was measured to study the wettability of the Au and Si surfaces. As shown in Fig. S5 (ESI),† the WCAs on the ordinary Si and Au surfaces were about 80° and 93°, respectively. Water was spread to form a film on the hydrophilic Si surface, and the WCA was almost too small to measure, an indication of strong hydrophilicity. However, the WCA was increased to about 120° on the hydrophobic Si surface, suggesting strong hydrophobicity. The current–voltage curves measured at the contacted stage of the Au–H₂O–Si junctions are shown in Fig. 3b. Using the TE model in eqn (1),¹⁶ the fitted values of n, ϕ_B and R_C from the three types of Au–H₂O–Si junctions are listed, as shown in Table 1. Compared with the Au–Si junctions, the values of n and ϕ_B were nearly unchanged, but R_C was increased apparently after water was introduced. This finding can be interpreted as follows. When the Au surface was pressed onto the Si surface, there was a thin water interlayer in between the two surfaces as the surfaces of Au and Si were not ideally smooth (Fig. S6, ESI†). Hence, the Au–H₂O–Si junctions had a larger R_C value than the Au–Si junctions. Fig. 3c and d show the open-circuit voltages and short-circuit currents as a function of number of cycles from the three types of Au–H₂O–Si junctions. The real-time video of the testing process is shown in Video S1 (ESI).† Output current results from the switching polarity test confirm that the currents come from the intermittently contacted Au–H₂O–Si junctions (Fig. S7, ESI†). It was clearly seen that the positive voltages of about 0.021 V, 0.28 V and 0.20 V were recorded from the hydrophilic, ordinary and hydrophobic Au–H₂O–Si junctions at the 1st cycle, respectively, as shown in Fig. 3c. However, for the hydrophilic and ordinary Au–Si junctions, the open-circuit voltage was negative, as shown in Fig. 1d. Before water was introduced, the KPFM results in Fig. 1b show that the chemical potential difference between the hydrophilic, ordinary, and hydrophobic Si surfaces and the Au surface was about −0.71 V, −0.31 V, and 0.13 V, respectively. Whereas, after water was introduced, the chemical potential difference from the above three friction layer pairs turned out to be about 0.21 V, 0.38 V, and 1.22 V, respectively (Fig. S8, ESI†). This demonstrates that the introduction of water significantly affects the chemical potential of Au and Si surfaces. This result is easy to understand because the bonding states at the liquid–solid interfaces cause the angle between the dipole plane of adsorbed water molecules and the solid surface vary.¹⁸ As a result, the adsorbed water's dipole moment leads to changes in the work function of the Au and Si surfaces under the influence of water. In addition, the magnitude of the open-circuit voltage is dominated by the absolute value of the chemical potential difference between the Au and Si surfaces. After water was introduced, the absolute values of the chemical potential differences between the hydrophilic, ordinary, and hydrophobic Si surfaces and the Au surface increased just by about 0.5 V, 0.1 V, and 1.1 V, respectively. Hence, the introduction of water cannot significantly increase the open-circuit voltage. With the increase of the number of cycles, water got expelled out or evaporated gradually. As a result, the open-circuit voltages of the Au–H₂O–Si junctions turned into the corresponding Au–Si junctions, respectively, see Fig. 3c. The short-circuit current as a function of number of cycles for the Au–H₂O–Si junctions is shown in Fig. 3d. The best output currents from the hydrophilic, ordinary, and hydrophobic Si surfaces with a water interlayer were up to about 15.31 μA, 3.62 μA, and 38.49 μA, which increased by a factor of 77, 24 and 257, respectively. Moreover, when the number of cycles increased, the current peaks from the Au–H₂O–Si junctions all decreased. After more than 60 cycles, the short-circuit currents of the Au–H₂O–Si junctions were also similar to those of the Au–Si junctions (Fig. S9, ESI†). The transferred charges Q_A and Q_S at the approaching and separating stages, calculated from the short-circuit current, are shown in Fig. 3e. The values of Q_A and Q_S increased by about 1–2 orders of magnitude than the corresponding Au–Si junctions (Table S1, ESI†). The above results show that regardless of the wettability of the Si surface, a higher output current and transferred charge could be collected after water was introduced. With the increase of the number of cycles, the output electric characteristics from the hydrophilic Si surface degraded more rapidly than the ordinary and hydrophobic Si surfaces. This may be caused by water evaporation on the hydrophilic Si surface more quickly than on the hydrophobic surface.¹⁹ Fig. S10 (ESI†) shows the stability test of the open-circuit voltage, short-circuit current and transferred charges using a humidifier to spray DI water mist between the Au and Si surfaces rather than introducing water directly (Video S2, ESI†). The three types of Au–H₂O–Si junctions could output stable electric characteristics up to 10 000 cycles. This result confirms that the Au–H₂O–Si junctions could retain a stable electric output in a high humidity environment. The output currents of the Au–H₂O–Si junctions with different external loads are shown in Fig. S11 (ESI).† The output currents decreased with increasing external loads. Fig. 3f shows the peak power density of the Au–H₂O–Si junctions under different external loads. It can be determined that all Au–H₂O–Si junctions displayed an impedance of 60 MΩ, slightly larger than that of the Au–Si junctions (Fig. 1f). More importantly, the hydrophobic Au–H₂O–Si junction had a maximum peak power density of about 0.32 mW cm⁻², about 3 orders of magnitude higher than that of the Au–Si junctions (Fig. 1f).

Fig. 3 Experimental schematic illustration and output electric characteristics of the intermittently contacted Au–H₂O–Si junctions. (a) The schematic setup for the experiments of the intermittently contacted Au–H₂O–Si junctions. (b) The current–voltage curves and fitting results of the Au–H₂O–Si junctions. (c–e) The open-circuit voltage, short-circuit current and transferred charge acquired from the Au–H₂O–Si junctions. (f) The peak power density of the Au–H₂O–Si junctions under humidifier spraying DI water mist.

When water is placed onto the Si surface, electron transfer and ionization reaction occur simultaneously on the H₂O–Si interface due to the overlapping of electron clouds from silicon atoms and water molecules.^12,20 When a water molecule transfers an electron to a silicon atom, it becomes cationic (H₂O⁺) in a short lifetime.²¹ Then, the H₂O⁺ joins with a neighboring water molecule to yield an OH radical and a H₃O⁺.²² Meanwhile, due to the ionization reaction on the Si surface, the oxygen atoms in the water molecules are also likely to form covalent bonds with the silicon atoms and generate O⁻ on the Si surface.^23,24 This ionization reaction also contributes H⁺ in the water.²⁴ Hence, the Si surface becomes negatively charged after coming in contact with water at first. Then, the cations in the water would be attracted to migrate toward the negatively charged Si surface by the electrostatic interactions, forming an EDL at the H₂O–Si interface. Prior studies have demonstrated that the interfacial energy between the hydrophilic Si surface and water is larger than that between the hydrophobic Si surface and water,²⁵ and oxygen atoms in water molecules are more likely to form covalent bonds with the hydrophilic Si surface. In other words, the ionization reaction dominates the CE on the hydrophilic Si surface.²³ In contrast, the ionization reaction between the hydrophobic Si surface and water is less likely to occur, meaning that the CE on the hydrophobic Si surface is electron transfer-dominated.²³

Videos S3–S5 (ESI)† show the real-time videos of the process of squeezing water by the hydrophilic, ordinary and hydrophobic Si surfaces with the Au surface captured using a high-speed camera, respectively. Based on the CE process at the dynamic H₂O–Si interface, the working mechanism of the intermittently contacted Au–H₂O–Si junction is proposed, as shown in Fig. 4. After water comes in contact with the hydrophilic Si surface, an ionization reaction-dominated EDL forms at the H₂O–Si interface, as shown in Fig. S12a (ESI).† The inner Helmholtz plane (IHP) of the Stern layer is mainly constructed by the anions. In contrast, when water comes in contact with the hydrophobic Si surface, an electron-dominated IHP of the Stern layer is formed (Fig. S12b, ESI†). When the Au surface gradually approaches the Si surface and without contact with water, electrons must move from the higher chemical potential surface to the lower chemical potential surface through the external circuit and generates a current I_A. Current output in this process is similar to that in the Au–Si junctions in the approaching stage, as shown in Fig. 2. Hence, the direction of I_A from the hydrophilic Au–H₂O–Si junction is opposite to that of the hydrophobic Au–H₂O–Si junction, as shown in Fig. 4a. Once the Au surface is in contact with water, the cations in the diffuse layer could rapidly migrate under the impact of the built-in electric field . From the work function measurement results in Fig. S1 (ESI),† the direction of is from the hydrophilic Si to the Au (Fig. 4a(i)), and some cations are adsorbed on the Au surface, which causes the potential of the Au surface to increase and then produce a positive current I_A. However, for the hydrophobic Si, the direction of is reversed (Fig. 4a(ii)), and the cations are absorbed on the Si surface and then increases the potential of the Si surface. As a result, a negative current I_A is recorded in the external circuit. When the Au and Si surfaces are in contact, a Schottky contact is formed. Meanwhile, a thin water interlayer could be trapped between the two surfaces, as shown in Fig. 4b. In this stage, the Au–H₂O–Si junction reaches a thermal equilibrium. Once the two surfaces are separated, for hydrophilic Si, water remains on the Si surface, and some cations are absorbed on the Au surface (Fig. 4c(i)). However, for hydrophobic Si, water is absorbed on the Au surface rather than on the Si surface in the separating stage (Fig. 4c(ii)). According to the chemical potential difference measurement results shown in Fig. S8 (ESI),† after introducing water, the chemical potential of the Au surface is higher than that on the Si surface, regardless of the wettability of the Si surface. Hence, the electrons are transferred from Si to Au through the external circuit, yielding a positive current (I_S) until the chemical potential difference between the Au and Si surfaces is balanced (Fig. 4d). Let us revisit Fig. 3e, wherein the value of Q_S from hydrophobic Si is found to be far larger than that of hydrophilic Si. This is because the EDL on the hydrophobic Si surface would collapse in the separating stage, leading to discharging of electrons from the inner IHP of the Stern layer to the external circuit. In contrast, for the hydrophilic Si surface, the EDL remains on the Si surface, and the IHP of the Stern layer is nearly unchanged. As a result, the electrons in the IHP cannot discharge to the external circuit. Hence, the continual forming and collapsing of EDL is the key to retain the high output current of the Au–H₂O–Si junction. By the way, the current output from the ordinary Au–H₂O–Si junction considered to be generated by the hydrophilic and hydrophobic Au–H₂O–Si junctions.

Fig. 4 Mechanism of current output from the intermittently contacted Au–H₂O–Si junctions. (a–d) The approaching, contacted, separating and separated stages, respectively.

From above discussion, it can be deduced that the impact of the EDL at the Au–H₂O–Si junction is similar to that of the space charge region at the Au–Si junction. The transferred charge from the Au–H₂O–Si junction is enhanced by the formation of the EDL and the electromotive force of the built-in electric field. Fig. 5 gives the effects of the volume of water and the solutions on the short-circuit current from the Au–H₂O–Si junctions. It can be seen that the magnitude of the short-circuit current was nearly independent of the volume of water (Fig. 5a). As shown in Fig. 5b, as the concentration of NaCl aqueous solution was increased from 0 to 1.0 mol L⁻¹, the output currents from the hydrophilic and ordinary Au–H₂O–Si junctions were converted from direct current (DC) to an alternating current (AC), and the output currents from the Au–H₂O–Si junctions were all decreased. Although a higher concentration of NaCl aqueous solution could provide more free ions, Na⁺ and Cl⁻ ions in the solution could quickly screen the built-in electric field. As a result, a lower AC current output was recorded when a higher concentration NaCl liquid was introduced. Moreover, it was also found that the output currents were decreased sharply after the dielectric oil was introduced between the Au and Si surfaces. This result suggests that the oil–Si interface could not establish an EDL, so the Au–oil–Si junction functioned as a conventional TENG.

Fig. 5 Effects of different liquids on the short-circuit current of Au–H₂O–Si junctions. (a) Different volumes of water. (b) Different solutions.

3 Conclusion

In this work, we have presented the intermittently contacted Au–H₂O–Si junctions with a high output power density of about 0.32 mW cm⁻², which was about 3 orders of magnitude larger than that of the intermittently contacted Au–Si junctions (without water). This significantly large output power density can be attributed to the interaction between the EDL and the built-in electric field at the Au–H₂O–Si junction. The output electric characteristics of the intermittently contacted Au–H₂O–Si junctions with different wettabilities of the Si surface suggest that hydrophobic treatment of the Si surface is a feasible approach to increase the efficiency of the mechanical-to-electric power conversion. This work not only provides a new physical understanding of the CE at the dynamic liquid–semiconductor interface but also serves as a guide to design dynamic semiconductor junction-based sensors and electric generators in the future.

3.1 Materials and methods

3.1.1 Choosing and fabrication of Si friction layers. Fig. S13 (ESI)† shows the short-circuit current of ordinary Au–H₂O–Si junctions with different doping concentrations of Si friction layers. Obviously, the doping concentration with 1 × 10¹⁸ cm⁻³ showed the best current output. Hence, we chose this type of Si as the friction layer in our experiments. In addition, we did not change the thickness of the Si wafer as we believed it would not significantly affect our results.

The Si wafers were first cleaned in an ultrasonic bath with acetone, isopropyl alcohol and deionized water for 5 minutes in sequence. They were then immersed in 1 : 10 dilute hydrofluoric acid (49%) for 3 minutes to remove any native silicon oxide layer from their surfaces. A bilayer of 20 nm titanium and 50 nm gold was coated on the backside of the Si friction layer in turns to fabricate the back electrode.

3.1.2 Hydrophilic and hydrophobic treatments. For the hydrophilic treatment, the Si samples were treated by inductively coupled plasma-reactive ion etching (ICP-RIE), in which oxygen acts the reaction gas for 10 minutes at a radiofrequency power of 200 W. During the treatment, the Si samples were kept below 100 °C. For the hydrophobic treatment, the Si samples were first subjected to hydrophilic treatment. Subsequently, a 2 nm thick fluoroalkylsilane hydrophobic layer was prepared on these Si sample surfaces by the chemical vapor deposition (CVD) method.

3.1.3 Electric output measurements. The dynamic Au–H₂O–Si junction CPGs were tested normally with contact forces supplied by a linear motor (PL0119x600/520, LinMot, Switzerland). The short-circuit current and open-circuit voltage were measured using an SR570 (Stanford Research Systems, USA) low-noise preamplifier and a 6514 (Keithley, USA) system source meter, respectively. The current–voltage curves were measured using a 2601B (Keithley, USA) system source meter.

3.1.4 AFM measurements. The topographic and Kelvin probe force microscopy images were obtained with a Bruker Dimension Icon AFM system from Shiyanjia Lab (https://www.shiyanjia.com/).

Data availability

The authors declare that the data supporting this article have been included as part of the ESI.†

Author contributions

X. Fan, S. Zhang, and S. Deng designed the experiments. X. Fan, S. Zhang, and Q. Chen contributed to sample preparation. X. Fan, S. Zhang, and Q. Chen performed the experiments. M. Li, H. Lu, S. Deng, and Q. Zhang contributed to data analysis. Q. Chen, S. Deng, and Q. Zhang wrote the paper.

Conflicts of interest

The authors declare no competing interests.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (62304161), Fundamental Research Funds for the Central Universities (2024-LXY-A1-01), Academic Research Fund Tier2 (MOE-T2EP50223-0008) and Tier 1 (RG131/22) for the Ministry of Education, Singapore. The authors would like to thank Dr Chen Ma and Dr Xuanyu Huang from Tsinghua University for their valuable comments. The authors extend their gratitude to Dr Qin from Shiyanjia Lab (https://www.shiyanjia.com) for providing invaluable assistance with the AFM tests.

References

1. X. Guo, L. liu, Z. X. Zhang, S. Gao, T. Y. Y. He, Q. F. Shi and C. G. Lee, Technology evolution from micro-scale energy harvesters to nanogenerators, J. Manuf. Syst., 2021, 31(9), 093002.
2. D. Choi, Y. Lee, Z. H. Lin, S. M. Cho, M. Kim, C. K. Ao, S. Soh, C. Sohn, C. K. Jeong, J. W. Lee, M. B. Lee, S. A. Lee, J. Ryu, P. Parashar, Y. J. Cho, J. Ahn, I. Kim, F. Jiang, P. S. Lee, G. Khandelwal, S. J. Kim, H. S. Kim, H. C. Song, M. Kim, J. Nah, W. Kim, H. G. Menge, Y. T. Park, W. Xu, J. H. Hao, H. Park, J. H. Lee, D. M. Lee, S. W. Kim, J. Y. Park, H. X. Zhang, Y. L. Zi, R. Guo, J. Cheng, Z. Yang, Y. N. Xie, S. M. Lee, J. H. Chung, I. Oh, J. S. Kim, T. H. Cheng, Q. Gao, G. Cheng, G. Q. Gu, M. Shim, J. H. Jung, C. W. Yun, C. Zhang, G. X. Liu, Y. F. Chen, S. Kim, X. Y. Chen, J. Hu, X. Pu, Z. H. Guo, X. D. Wang, J. Chen, X. Xiao, X. Xie, M. Jarin, H. L. Zhang, Y. C. Lai, T. Y. Y. He, H. Kim, I. Park, J. Ahn, N. D. Huynh, Y. Yang, Z. L. Wang, J. M. Baik and D. Choi, Recent Advances in Triboelectric Nanogenerators: From Technological Progress to Commercial Applications, ACS Nano, 2023, 17(12), 11087–11219.
3. H. Lei, Y. F. Chen, Z. Q. Gao, Z. Wen and X. H. Sun, Advances in self-powered triboelectric pressure sensors, J. Mater. Chem. A, 2021, 9(36), 20100–20130.
4. Z. L. Wang, On the first principle theory of nanogenerators from Maxwell’s equations, Nano Energy, 2020, 68, 104272.
5. B. B. Sun, X. Guo, Y. Y. Zhang, Z. L. Wang and J. J. Shao, A generalized model for a triboelectric nanogenerator energy harvesting system, Nano Energy, 2024, 126, 109637.
6. Q. Zhang, R. Xu and W. F. Cai, Pumping electrons from chemical potential difference, Nano Energy, 2018, 51, 698–703.
7. S. X. Li, S. Deng, R. Xu, D. Liu, Y. Nan, Z. W. Zhang, Y. K. Gao, H. F. lv, M. Li, Q. Zhang, J. Wang and Z. L. Wang, High frequency mechanical energy harvester with direct current output from chemical potential difference, ACS Energy Lett., 2022, 7, 3080–3086.
8. R. Z. Yang, R. Xu, W. J. Dou, M. Benner, Q. Zhang and J. Liu, Semiconductor-based dynamic heterojunctions as an emerging strategy for high direct-current mechanical energy harvesting, Nano Energy, 2021, 83, 105849.
9. K. R. Kaja, S. Hajra, S. Panda, M. A. Belal, U. Pharino, H. Khanbareh, N. Vittayakorn, V. Vivekananthan, C. Bowen and H. J. Kim, Exploring liquid-solid interface based triboelectrification, structures, and applications, Nano Energy, 2024, 131, 110319.
10. A. Kulandaivel, S. Potu, A. Babu, N. Madathil, M. Velpula, R. K. Rajaboina and U. K. Khanapuram, Advances in ferrofluid-based triboelectric nanogenerators: Design, performance, and prospects for energy harvesting applications, Nano Energy, 2024, 120, 109110.
11. A. Kulandaivel, S. Potu, R. K. Rajaboina and U. K. Khanapuram, Exploring Wettability: A Key to Optimizing Liquid-Solid Triboelectric Nanogenerators, ACS Appl. Mater. Integr, 2024, 16(43), 58029–58059.
12. K. Q. Wang and J. J. Li, Electricity generation from the interaction of liquid-solid interface: a review, J. Mater. Chem. A, 2021, 9(14), 8870–8895.
13. S. Y. Qin, P. Yang, Z. Q. Liu, J. Hu, N. Li, L. M. Ding and X. Y. Chen, Triboelectric sensor with ultra-wide linear range based on water-containing elastomer and ion-rich interface, Nat. Commun., 2024, 15(1), 10640.
14. T. Suni, K. Henttinen, I. Suni and J. Mäkinen, Effects of plasma activation on hydrophilic bonding of Si and SiO, J. Electrochem. Soc., 2002, 149(6), G348–G351.
15. K. Teshima, H. Sugimura, Y. Inoue, O. Takai and A. Takano, Transparent ultra water-repellent poly(ethylene terephthalate) substrates fabricated by oxygen plasma treatment and subsequent hydrophobic coating, Appl. Surf. Sci., 2005, 244(1–4), 619–622.
16. S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, John Wiley & Sons, New Jersey, 2007.
17. S. Deng, R. Xu, M. Li, L. J. Li, Z. L. Wang and Q. Zhang, Influences of surface charges and gap width between p-type and n-type semiconductors on charge pumping, Nano Energy, 2020, 78, 105287.
18. Z. Q. Liu, S. Q. Lin, P. Yang, S. Y. Qin, J. Hu and X. Y. Chen, A wiping-type semiconductor-liquid generator utilizing water-bearing solid materials and hydrated biological tissues, Energy Environ. Sci., 2024, 17(18), 6582–6593.
19. X. Yan, F. P. Chen, C. Y. Zhao, X. Wang, L. N. Li, S. Khodakarami, K. F. Rabbi, J. Q. Li, M. J. Hoque, F. Chen, J. Feng and N. Miljkovic, Microscale Confinement and Wetting Contrast Enable Enhanced and Tunable Condensation, ACS Nano, 2022, 16(6), 9510–9522.
20. Z. L. Wang and A. C. Wang, On the origin of contact-electrification, Mater. Today, 2019, 30, 34–51.
21. Z. H. Loh, G. Doumy, C. Arnold, L. Kjellsson, S. H. Southworth, A. Al Haddad, Y. Kumagai, M. F. Tu, P. J. Ho, A. M. March, R. D. Schaller, M. S. B. Yusof, T. Debnath, M. Simon, R. Welsch, L. Inhester, K. Khalili, K. Nanda, A. I. Krylov, S. Moeller, G. Coslovich, J. Koralek, M. P. Minitti, W. F. Schlotter, J. E. Rubensson, R. Santra and L. Young, Observation of the fastest chemical processes in the radiolysis of water, Science, 2020, 367(6474), 179.
22. Y. Gauduel, S. Pommeret, A. Migus and A. Antonetti, Some evidence of ultrafast H2O+-water molecule reaction in femtosecond photoionization of pure liquis water: influence on geminate pair recombination dynamics, Chem. Phys. Lett., 1990, 149, 1–10.
23. S. Q. Lin, L. Xu, A. C. Wang and Z. L. Wang, Quantifying electron-transfer in liquid-solid contact electrification and the formation of electric double-layer, Nat. Commun., 2020, 11(1), 399.
24. T. A. Gmür, A. Goel and M. A. Brown, Quantifying Specific Ion Effects on the Surface Potential and Charge Density at Silica Nanoparticle-Aqueous Electrolyte Interfaces, J. Phys. Chem. C, 2016, 120(30), 16617–16625.
25. X. M. Li, P. Bista, A. Z. Stetten, H. Bonart, M. T. Schür, S. Hardt, F. Bodziony, H. Marschall, A. Saal, X. Deng, R. Berger, S. A. L. Weber and H. J. Butt, Spontaneous charging affects the motion of sliding drops, Nat. Phys., 2022, 18(6), 713.

---

Footnotes

† Electronic supplementary information (ESI) available: Fig. S1–S13, Table S1 and Videos S1–S5. See DOI: https://doi.org/10.1039/d5ta00099h

‡ Xinru Fan, Shuo Zhang, and Qihan Chen contributed equally to this work.

---