Triboelectric Hydrogen Gas Sensor with Pd Functionalized Surface

Palladium (Pd)-based hydrogen (H2) gas sensors have been widely investigated thanks to its fast reaction and high sensitivity to hydrogen. Various sensing mechanisms have been adopted for H2 gas sensors; however, all the sensors must be powered through an external battery. We report here an H2 gas sensor that can detect H2 by measuring the output voltages generated during contact electrification between two friction surfaces. When the H2 sensor, composed of Pd-coated ITO (indium tin oxide) and PET (polyethylene Terephthalate) film, is exposed to H2, its output voltage is varied in proportion to H2 concentration because the work function (WF) of Pd-coated surface changes, altering triboelectric charging behavior. Specifically, the output voltage of the sensor is gradually increased as exposing H2 concentration increases. Reproducible and sensitive sensor response was observed up 1% H2 exposure. The approach introduced here can easily be adopted to development of triboelectric gas sensors detecting other gas species.


Introduction
Hydrogen (H 2 ) has been claimed to be an alternative to replace fossil fuels [1][2][3][4]. Since H 2 is flammable, explosive, colorless, and odorless gas, precise detection of its concentration is important for prevention of potential accidents and its safe use [5]. Thus, many H 2 sensors, exploiting different sensing mechanisms, have been introduced to date [6][7][8][9]. Among them, palladium (Pd)-based H 2 sensor is one of the most widely used H 2 sensors due to its fast reaction to H 2 , high sensitivity, and selectivity [10][11][12][13][14][15][16][17][18][19]. When Pd is exposed to H 2 molecules, Pd is reacted with H 2 by absorbing H 2 molecules inside its lattice, expanding the volume: Pd is transformed as either α-phase PdH x or β-phase PdH x depending on exposed H 2 concentrations. Consequently, the electrical and optical properties of Pd are changed as it is exposed to H 2 . Using this mechanism, several H 2 sensors have been reported. For instance, H 2 sensors, utilizing nanogap structures [10,11], surface acoustic wave (SAW) [13,14], Schottky diodes structure [15,16], and surface plasmon effect [17,18], have been reported. Recently, triboelectric energy harvesting has gained much attention as a route to supply necessary power to operate electronic devices [20][21][22][23][24][25][26]. Further, few attempts have been made to develop chemical sensors utilizing triboelectric effect [27,28]. Specifically, detection is usually made by measuring variable triboelectric output as a result of friction surface reaction with target gas molecules [27]. However, only limited chemical species have been probed so far and further investigation is necessary to expand its application.
Here, we report a self-powered H 2 gas sensor by adopting triboelectric effects. When two dielectric surfaces on metal electrode are periodically contacted and released, each surface is electrified due to charge separation between two surfaces. The surface that gained electrons is charged negatively, while the other surface that lost electrons is positively charged, depending on their difference in electron affinity. These surface charges electrostatically induce counter charges on the facing metal electrode, developing the potential difference between two metal electrodes. Therefore, when the two electrodes are connected through an external load, it drives the electrons to flow between them. In this work, based on this triboelectric generation mechanism, we report a self-powered H 2 gas sensor, demonstrating reliable sensing performance from 0% to 1% H 2 exposures.

Preperation of Self-Powered H 2 Gas Sensor
The schematic representation of the self-powered Pd H 2 gas sensor is shown in Figure 1a. Two indium tin oxide (ITO) (130 nm)/polyethylene Terephthalate (PET) (130 µm) substrates (dimension: 1 cm × 1 cm × 130.13 µm) are first cleaned with IPA acetone and DI water, respectively. On one substrate, a 10 nm-thick Pd is thermally evaporated on the ITO side of the substrate as an active H 2 sensing layer. The other substrate is used as the contacting substrate without any chemical treatment or coating, where the PET side of the substrate is employed as a counter contact surface. Lastly, ITO electrodes of the two substrates are connected through an external circuit. Figure 1b shows the cross sectional transmission electron microscopy (TEM) of the evaporated Pd. The Pd layer id co-deposited on a Si/SiO 2 substrate in order to measure the thickness of the Pd layer. It can be noticed that~10 nm evaporated Pd is deposited on the substrate. Next, energy-dispersive X-ray spectroscopy (EDS) was also examined to verify the evaporated Pd layer on the ITO surface ( Figure 1c). Nanomaterials 2016, 6, 186 2 of 8 electrified due to charge separation between two surfaces. The surface that gained electrons is charged negatively, while the other surface that lost electrons is positively charged, depending on their difference in electron affinity. These surface charges electrostatically induce counter charges on the facing metal electrode, developing the potential difference between two metal electrodes. Therefore, when the two electrodes are connected through an external load, it drives the electrons to flow between them. In this work, based on this triboelectric generation mechanism, we report a selfpowered H2 gas sensor, demonstrating reliable sensing performance from 0% to 1% H2 exposures.

Preperation of Self-Powered H2 Gas Sensor
The schematic representation of the self-powered Pd H2 gas sensor is shown in Figure 1a. Two indium tin oxide (ITO) (130 nm)/polyethylene Terephthalate (PET) (130 μm) substrates (dimension: 1 cm × 1 cm × 130.13 μm) are first cleaned with IPA acetone and DI water, respectively. On one substrate, a 10 nm-thick Pd is thermally evaporated on the ITO side of the substrate as an active H2 sensing layer. The other substrate is used as the contacting substrate without any chemical treatment or coating, where the PET side of the substrate is employed as a counter contact surface. Lastly, ITO electrodes of the two substrates are connected through an external circuit. Figure 1b shows the cross sectional transmission electron microscopy (TEM) of the evaporated Pd. The Pd layer id co-deposited on a Si/SiO2 substrate in order to measure the thickness of the Pd layer. It can be noticed that ~10 nm evaporated Pd is deposited on the substrate. Next, energy-dispersive X-ray spectroscopy (EDS) was also examined to verify the evaporated Pd layer on the ITO surface ( Figure 1c).

Self-Powered H 2 Gas Sensor Test Set up
For self-powered H 2 gas sensor measurement, the test setup shown in Figure 2 was prepared. The test setup is composed of two parts: gas flow control unit, exposing gas concentration using mass flow control units (MFCs) by introducing 5% H 2 and dry air at specific ratios, and measurement unit, measuring voltage between two electrodes by periodic contact. Experimental details of gas exposure are described in the Materials and Methods section. A specific concentration of H 2 , prepared in the gas control part, flowed into an enclosed gas chamber (length × width × height: 6 cm × 6 cm × 6 cm), within which the custom-built pushing machine is installed. Two surfaces of the H 2 sensor were mounted on the pushing machine and were contacted periodically during pressing and releasing motions.

Self-Powered H2 Gas Sensor Test Set up
For self-powered H2 gas sensor measurement, the test setup shown in Figure 2 was prepared. The test setup is composed of two parts: gas flow control unit, exposing gas concentration using mass flow control units (MFCs) by introducing 5% H2 and dry air at specific ratios, and measurement unit, measuring voltage between two electrodes by periodic contact. Experimental details of gas exposure are described in the Materials and Methods section. A specific concentration of H2, prepared in the gas control part, flowed into an enclosed gas chamber (length × width × height: 6 cm × 6 cm × 6 cm), within which the custom-built pushing machine is installed. Two surfaces of the H2 sensor were mounted on the pushing machine and were contacted periodically during pressing and releasing motions. H2 gas sensor test setup consisting of gas control part and measurement part. A specific concentration of H2 is precisely controlled by mixing dry air and H2 at different ratio and the mixed gas is then introduced into the measurement chamber with a custom-made pushing machine installed inside. The output signals from the sensor were measured using the data acquisition instrument. Figure 3a shows the sensing mechanism of a self-powered H2 gas sensor, driven by triboelectric induction mechanism [29]. At the initial state, both Pd and PET surfaces have no surface charges. As the two surfaces are compressed, the electrons from donor interface states of the PET surface move toward Pd-functionalized ITO electrode through an external circuit, negatively charging the Pd-functionalized surface and positively charging the PET surface (Figure 3(ai)) [20]. As the two surfaces start to become separated, the surface charges on both contacting surfaces induce counter charges on the facing electrodes, developing the potential difference between the top and bottom electrode (Figure 3(aii)). At the fully separated state, charge equilibrium state is reached between the contacting surfaces and facing electrodes (Figure 3(aiii)). Lastly, as the two surfaces make contact again, induced charges on the electrodes disappear again due to a break of the charge neutral state (Figure 3(aiv)). To demonstrate the roles of Pd-functionalization, output voltages of the sensor without Pd-functionalized surface were first investigated (Figure 3b). When the non Pd-functionalized ITO and PET surfaces are contacted, it exhibits constant output voltage for varying H2 gas concentration from 0% to 5%. Since both PET and ITO do not react with H2 molecules, the output voltage signals were solely determined by the initial triboelectric sequence difference between them. Next, the output signals between Pd-functionalized ITO surface and PET were examined as a function of different exposing H2 concentrations. Differently from the non-functionalized device, it Figure 2. H 2 gas sensor test setup consisting of gas control part and measurement part. A specific concentration of H 2 is precisely controlled by mixing dry air and H 2 at different ratio and the mixed gas is then introduced into the measurement chamber with a custom-made pushing machine installed inside. The output signals from the sensor were measured using the data acquisition instrument. Figure 3a shows the sensing mechanism of a self-powered H 2 gas sensor, driven by triboelectric induction mechanism [29]. At the initial state, both Pd and PET surfaces have no surface charges. As the two surfaces are compressed, the electrons from donor interface states of the PET surface move toward Pd-functionalized ITO electrode through an external circuit, negatively charging the Pd-functionalized surface and positively charging the PET surface (Figure 3(ai)) [20]. As the two surfaces start to become separated, the surface charges on both contacting surfaces induce counter charges on the facing electrodes, developing the potential difference between the top and bottom electrode (Figure 3(aii)). At the fully separated state, charge equilibrium state is reached between the contacting surfaces and facing electrodes (Figure 3(aiii)). Lastly, as the two surfaces make contact again, induced charges on the electrodes disappear again due to a break of the charge neutral state (Figure 3(aiv)). To demonstrate the roles of Pd-functionalization, output voltages of the sensor without Pd-functionalized surface were first investigated (Figure 3b). When the non Pd-functionalized ITO and PET surfaces are contacted, it exhibits constant output voltage for varying H 2 gas concentration from 0% to 5%. Since both PET and ITO do not react with H 2 molecules, the output voltage signals were solely determined by the initial triboelectric sequence difference between them. Next, the output signals between Pd-functionalized ITO surface and PET were examined as a function of different exposing H 2 concentrations. Differently from the non-functionalized device, it can be clearly noticed that the output voltages of the sensor are increased in proportion to H 2 exposure (Figure 3c). The sensor output is more linearly increased up to 1% H 2 exposure. However, the sensor response is rather saturated when the H 2 concentration exceeds 2%. The observed behavior is consistent with other Pd-based hydrogen gas sensor [11,14]. Thus, Pd-functionalization plays a key role in triboelectric H 2 detection. In Figure 3d, the energy band structure of the sensor is drawn to explain the increase of triboelectric output voltage with exposure to H 2 . The work function (WF) of Pd is~5.2 eV and thus its Fermi level is located slightly above the midgap of PET [30]. As a result, when the two surfaces are contacted, carriers in valence and conduction bands of PET cannot be responsible for observed triboelectric charge exchange between the two surfaces. Therefore, the surface states must play an important role in contact electrifications between the two materials. Here, when the Pd-coated ITO surface is contact-electrified against the PET surface, Pd is negatively charged while PET is positively charged. This indicates that there must be donor states (0.55 and 0.85 eV below conduction band) near the midgap of PET, located slightly above the Fermi level of Pd [31]. As Pd is being exposed to H 2 , H 2 molecules are dissociated and absorbed in the Pd lattice, forming PdH x (palladium hydride). When PdH x is formed, its WF becomes smaller than that of Pd [32]. Consequently, as the WF of PdH x moves more close to donor states in PET, the charge exchange between two materials can be more effective, which enhances surface charge exchange and leads to increase in triboelectric output voltage of the sensor. Therefore, triboelectric output voltage will be continuously increased until H 2 concentration reaches a certain level. can be clearly noticed that the output voltages of the sensor are increased in proportion to H2 exposure (Figure 3c). The sensor output is more linearly increased up to 1% H2 exposure. However, the sensor response is rather saturated when the H2 concentration exceeds 2%. The observed behavior is consistent with other Pd-based hydrogen gas sensor [11,14]. Thus, Pd-functionalization plays a key role in triboelectric H2 detection. In Figure 3d, the energy band structure of the sensor is drawn to explain the increase of triboelectric output voltage with exposure to H2. The work function (WF) of Pd is ~5.2 eV and thus its Fermi level is located slightly above the midgap of PET [30]. As a result, when the two surfaces are contacted, carriers in valence and conduction bands of PET cannot be responsible for observed triboelectric charge exchange between the two surfaces. Therefore, the surface states must play an important role in contact electrifications between the two materials. Here, when the Pd-coated ITO surface is contact-electrified against the PET surface, Pd is negatively charged while PET is positively charged. This indicates that there must be donor states (0.55 and 0.85 eV below conduction band) near the midgap of PET, located slightly above the Fermi level of Pd [31]. As Pd is being exposed to H2, H2 molecules are dissociated and absorbed in the Pd lattice, forming PdHx (palladium hydride). When PdHx is formed, its WF becomes smaller than that of Pd [32]. Consequently, as the WF of PdHx moves more close to donor states in PET, the charge exchange between two materials can be more effective, which enhances surface charge exchange and leads to increase in triboelectric output voltage of the sensor. Therefore, triboelectric output voltage will be continuously increased until H2 concentration reaches a certain level.  To verify reproducible output performance of the H 2 sensor, the device was measured while repeatedly being exposed to different concentrations of H 2 (up to 1% H 2 ) (Figure 4a). Similar output voltages were measured during the repeated measurement. The magnified output signals in Figure 4a show alternating peaks due to the triboelectric induction mechanism described in Figure 3a. Next, sensor response and recovery time were measured while exposing under 1% hydrogen, exhibiting the response and recovery time of~30 min and~15 min, respectively (Figure 3b). In addition, the effect of relative humidity (RH) on the sensor output was also measured. It can be noticed that the output voltage is obviously reduced when the RH is increased from 47% to 65%, negatively affecting triboelectric output voltage. Lastly, the sensor responses were calculated from multiple devices and plotted as a function of exposing H 2 concentrations (Figure 4d). Here, the sensor responses were calculated using the equation ∆V/V 0 , where V 0 is the output voltage measured while being exposed to dry air, V s is the output voltage under exposure to a specific H 2 concentration, and ∆V (= V s − V 0 ) is a relative output voltage change between two states. A gradual increase of sensor response can be noticed as H 2 concentration increases. We also note that the sensor response was more widely varied when the H 2 concentration exceeds 2%. Using the obtained sensor responses, the sensitivity of the hydrogen sensor is subsequently extracted, where the sensitivity is calculated using the equation, To verify reproducible output performance of the H2 sensor, the device was measured while repeatedly being exposed to different concentrations of H2 (up to 1% H2) (Figure 4a). Similar output voltages were measured during the repeated measurement. The magnified output signals in Figure 4a show alternating peaks due to the triboelectric induction mechanism described in Figure 3a. Next, sensor response and recovery time were measured while exposing under 1% hydrogen, exhibiting the response and recovery time of ~30 min and ~15 min, respectively (Figure 3b). In addition, the effect of relative humidity (RH) on the sensor output was also measured. It can be noticed that the output voltage is obviously reduced when the RH is increased from 47% to 65%, negatively affecting triboelectric output voltage. Lastly, the sensor responses were calculated from multiple devices and plotted as a function of exposing H2 concentrations (Figure 4d). Here, the sensor responses were calculated using the equation ∆V/V0, where V0 is the output voltage measured while being exposed to dry air, Vs is the output voltage under exposure to a specific H2 concentration, and ∆V (= Vs − V0) is a relative output voltage change between two states. A gradual increase of sensor response can be noticed as H2 concentration increases. We also note that the sensor response was more widely varied when the H2 concentration exceeds 2%. Using the obtained sensor responses, the sensitivity of the hydrogen sensor is subsequently extracted, where the sensitivity is calculated using the equation, ∆ /

Mechanism and Output Performance of Self-Powered H 2 Gas Sensor
. Up to 1% H2 exposures, higher sensitivity, 0.75, was obtained. Above 2% H2 exposures, on the other hand, the sensitivity is reduced to 0.2, indicating the saturation of H2 molecules absorption in Pd lattice above this level.

Measurement and Characterization
Triboelectric output voltages were measured while exposing the device in different H2 gas concentrations. We maintained total flux of dry air and H2 as 200 sccm. For each H2 concentration, the sample was exposed to each H2 concentration for 30 min, which is sufficient reaction time for Pd in the presence of H2 gas. Different concentrations of H2 were introduced into the chamber by mixing with dry air at the specific ratios using mass flow control (MFC) units. For all measurement, a constant force of 0.1 MPa was applied using the custom-made pushing machine installed inside of a gas chamber at a frequency of 0.5 Hz and the distance between two electrodes was kept as 3.5 cm. The output voltages from the sensor were measured by connecting to the oscilloscope (Waverunner

Measurement and Characterization
Triboelectric output voltages were measured while exposing the device in different H 2 gas concentrations. We maintained total flux of dry air and H 2 as 200 sccm. For each H 2 concentration, the sample was exposed to each H 2 concentration for 30 min, which is sufficient reaction time for Pd in the presence of H 2 gas. Different concentrations of H 2 were introduced into the chamber by mixing with dry air at the specific ratios using mass flow control (MFC) units. For all measurement, a constant force of 0.1 MPa was applied using the custom-made pushing machine installed inside of a gas chamber at a frequency of 0.5 Hz and the distance between two electrodes was kept as 3.5 cm.
The output voltages from the sensor were measured by connecting to the oscilloscope (Waverunner 2 LT 354, Lecroy, Chestnut Ridge, NY, USA). The Pd-coated layer was characterized by using both TEM (JEM-ARM200F, JEOL, Tokyo, Japan) and EDS (HD-2300A, Hitachi, Tokyo, Japan) before gas sensing measurement.

Fabrication of Triboelectric H 2 Gas Sensor
For fabrication of triboelectric H 2 gas sensors, the surface of ITO (130 nm)-coated PET (130 µm) film (Sigma Aldrich, Darmstadt, Germany) was first cleaned with IPA, acetone, and DI water, respectively, followed by N 2 blow dry. After cleaning the film, Pd (10 nm) was thermally evaporated on an ITO surface of the film. Next, the counter contact surface, consisting of PET/ITO film, was prepared. Here, Pd-coated ITO surface was contact-electrified with the PET side of the counter contact surface. Electrical connection between ITO electrodes of the two friction surfaces was made using copper wires. For electrical connection of the ITO surface and copper line as an electrode, the electrode was soldered by soldering with silver paste. Each prepared surface was then mounted on the custom-made pushing machine installed inside a gas test chamber.

Conclusions
In summary, we demonstrated self-powered H 2 gas sensor utilizing triboelectric effects. Due to WF change of Pd-coated surface in the presence of H 2 , triboelectric output voltage of the sensor was varied proportionally. Using this mechanism, different concentrations of H 2 were detected during contact electrification between two friction surfaces without using an external battery. Specifically, the output voltage of H 2 sensor was increased in proportion to H 2 concentration. The WF of PdH x moves close to a donor state of PET, and the charge transfer between the two friction surfaces can be facilitated. Triboelectric H 2 sensor demonstrated reproducible and sensitive responses up to 1% H 2 . The approach introduced here can be easily adopted to development of triboelectric gas sensors detecting other gas species and other self-powered electronics devices.