2.1 Basic structures of direct-current generators and properties of the conductive polymers
Basic structures of all the direct-current generators adopted a metal/conductive polymer/metal sandwich geometry. Typical conductive polymers, i.e. polypyrrole (PPy), polyaniline (PANI) and polyethoxythiophene (PEDOT), and various metals were selected to construct the generators based on the principal of the polymer–metal contact conditions. To overcome the inflexibility of previous direct-current generators, such as the Al/PPy/Au generator in H. Shao et al.28, a method of electrochemical deposition was proposed to prepare conductive polymers on a flexible substrate coated with metal electrodes film to realize the flexibility. The fabrication procedure is shown in Fig.1a. Taking Au/PPy/Al generator as an example, firstly, metal electrode film of Cr/Au was magnetron sputtered on polyimide (PI) prior to electrochemically depositing conductive polymer of PPy and metal-coated PI was exposed to UV-Ultraviolet ozone for a treatment to improve surface property. The PI decorated conductive polymer was assembled with another piece of PI with metal film on it.
As shown in Fig. 1b, the photographic images of three kinds of generators, i.e. Au/PPy/Al, Au/PEDOT/Al and Au/PANI/Al, demonstrate their good conformability. The corresponding circuit diagram is systematically illustrated in Fig. 1c. As shown in Fig. 1c, the working principle of the direct-current generators based on the conductive polymer-metal Schottky contact was that output voltage and current stem from deformation of conductive polymers under the external loadings. In our previous work 38, the energy conversion process when a static compression is presented is discussed. When a direct generator posed with a fixed loading (20% stretching), the related thickness generally varies with time and lasts for 1-2 min.
Different conductive polymers on PI were characterized by the X-ray diffraction (XRD) and scanning electron microscope (SEM) images. Fig. S1 illustrate the XRD spectra of PPy (left panel), PEDOT (middle panel) and PANI (right panel). The characteristic diffraction peak of PPy, PEDOT and PANI film were at 26°, 26°, 26° (the diffraction peak corresponds to C (002)). The similar values of XRD demonstrated that the C (002) common crystal faces exist in PPy, PEDOT and PANI. As shown in Fig. S2, the smooth appearances in SEM images of PPy (left panel), PEDOT (middle panel) and PANI (right panel) suggested our conductive polymer films possess good morphology. Owing to that the growth style of electrochemical was modest stationarity, smooch morphology of these conductive polymers were achieved.
2.2 The output properties of the polymer-metal-Schottky-contact-based direct-current generators
Inspired by the Au/PPy/Al generator in H. Shao’s 28 previous works, to produce direct-current output, one metal layer should form an Ohmic contact with the conductive polymer, while the other should form a Schottky contact. As the HOMO of the PPy 30, PEDOT 31 and PANI 32 were -5.6eV, -5.2eV and -4.84 eV, respectively, the contacts between PPy and Au, PEDOT and Au, PANI and Cu are Ohmic contacts. Thus, direct-current generators based on Au/PPy/Al, Au/PEDOT/Al and Cu/PANI/Al were designed.
As shown in Fig. 2a, all the direct-current generators based on Au/PPy/Al (left panel), Au/PEDOT/Al (middle panel), and Cu/PANI/Al (right panel) produce direct-current outputs with long duration, as measured by an oscilloscope. Furthermore, the distinction magnitude of output voltages (1.85 V, 0.2V and 3.6 V for Au/PPy/Al, Au/PEDOT/Al and Cu/PANI/Al respectively) of the three generators suggests a clear association of the output on both the conductive polymers and the contact metals. It is worth noting that the output voltage (3.6 V) of the Cu/PANI/Ag generator is much higher than that in H. Shao’s previous work (maximum voltage: 0.8-0.9V). Fig. 2b schematically illustrates the energy band diagram of Au/PPy/Al, Au/PEDOT/Al and Cu/PANI/Al. According to the LUMO and the HOMO of the PPy, PEDOT, PANI and the work functions of metal (Au: 5.1eV, Cu: 4.65 eV, Al: 4.28eV and Ag 4.26eV), the contact between Au and PPy, the contact between Au and PEDOT, and the contact between Cu and PANI are Ohmic contact; the contact between Al and PPy, the contact between Ag and PEDOT, and the contact between Al and PANI are Schottky contact. That is, for a direct-current generator with metal/conductive polymers/metal configuration, one metal forms Ohmic contact with the conductive polymer and the other metal forms Schottky contact with the conductive polymer 32, 33.
This energy band diagram illustrates the mechanism underlying the direct-current generator. The direct-current output origins from the mechanical-induced interfacial charge accumulation and the energy-band bending. As a result, the Schottky barrier between the conductive polymer and the metal that forms the Schottky contact reduces and drives the movement of electrons from conductive polymers to the other metal electrodes 34-38.
In our previous work 39, the mechanism of the direct-current output based on Au/PPy/Al generator was investigated. The current density involved is given as:
Where, A* is the effective Richardson’s constant, Φ is the effective barrier potential, T is the absolute temperature, k is the Boltzmann constant. The barrier potential under deformation can be thus estimated based on Equation 1.
According to the above equation, the output currents under the external loadings are associated with the Schottky barrier height between the metal and the conductive polymer that forms the Schottky contact, which is equivalent to that the HOMO value of the conductive polymer subtracts the work function of the contacted metal. Do the LUMO and HOMO of the conductive polymer also vary if the conductive polymers experience deformation? Computational simulations are conducted to address this question, as presented below.
2.3 Computation study on the HOMO and LUMO of the conductive polymers to reveal the mechanism
Fig. 3a shows the HOMO of PPy (top panel), PEDOT (middle panel) and PANI (bottom panel) crystal fragments in normal state (left panel), compressing state (4% deformation, middle panel) and stretching state (4% deformation, right panel), respectively. As shown in the left panel of Fig. 3a, the electronic densities of the HOMO of the PPy in normal state are predominantly localized on one layer of pyridine ring. In the compressing state (4% deformation) and in the stretching state, the electronic densities of the HOMO are predominantly localized on part of pyridine rings in two different layers. That is, the electronic densities of the HOMO of the PPy crystal fragments change with their deformations. The distributions of the electronic densities of the HOMO change in PEDOT (middle panel) and PANI (bottom panel) under the compressing state and increase distance state, as well.
Specially, as the band-gap of PANI is sensitive to the dopant 40-42, acetic acid that served as the dopant was added into the well-prepared PANI film to alter the band-gap. The bandgap (Eg) of PANI between HOMO and LUMO is calculated to be Eg=hc/λ. Where, h is the Planck’s constant, c is the light velocity in vaccum, and λ is the wavelength in the UV-vis absorb spectrum of PANI (as shown in Fig. SI 3a). It can be calculated that the Eg of PANI before being doped and after being doped are 2.57 eV and 2.84 eV, respectively. The HOMO of PANI between HOMO and LUMO is calculated according to the equation: EHOMO=-|eE1/2+4.5eV+ESCE|, where, ESCE is standard electrode potential of the reference electrode. eE1/2 can be extracted from the volt-ampere characteristics (CV) test (as shown in Fig. SI 3b). It can be calculated that the HOMO of PANI before being doped and after being doped are -4.84 eV and -5.24 eV, respectively. Do to ELUMO= Eg- EHOMO, the LUMO of PANI before being doped and after being doped are -2.27 eV and -2.4 eV, respectively. Therefore, doping can increase the bandgap of PANI and further affect the output of the DC generator based on PANI.
The corresponding LUMO of the conductive polymers in their normal state, compressing state (4% deformation) and stretching state (4% deformation) are also studied (illustrated in Fig. SI 3a), showing a clear distinction similar to the HOMOs. The changes of LUMO and HOMO of conductive polymers in deformed state (compressing or stretching) indicate that the output may be related to the variation of frontier molecular orbital of conductive polymers.
The mechanical-induced deformation changes not only the electronic densities, but also the values of LUMO and HOMO of conductive polymers. The calculated HOMO of the conductive polymer: PPy, PEDOT and PANI are shown in Fig. 4 (b). As shown by the blank line in Fig. 3 (b), the 2%, 4%, 6%, 8% and 10% deformation can change the HOMO (ΔHOMO/HOMO) of PPy crystal fragments relatively by 0.204 %, 0.484 %, 0.871 %, 1.400 % and 2.140 %, respectively. (The Schotttky barrier values are close to the HOMO values of conductive polymers and work functions of contacted electrodes). For PEDOT, under the same 2%, 4%, 6%, 8% and 10% deformations, the relative changes in HOMO are 0.013%, 0.016 %, 0.041 %, 0.118% and 0.21%, respectively (blue line in Fig. 4 (b)). For PANI, the relative changes are 0.451 %, 0.946 %, 1.491 %, 2.094 % and 2.763 %, respectively (red line in Fig. 4 (b)). Fig. 3 (c) compares the relative change in HOMO vs. compression curves of PANI, PPy, PEDOT, respectively. Under the same deformation, PANI clearly exhibits a much larger change in HOMO than PPy and PEDOT. Recalling that the direct-current generator that involves PANI also shows a larger voltage output, the analysis indicates an association of the output to the HOMO of the conductive polymers. Similar, LUMO vs. deformations (2%, 4%, 6%, 8% and 10%) for conductive polymer PPy, PEDOT and PANI illustrate in Fig. SI 4b.
2.4 The voltage/current output of direct-current generators with different conductive polymers and different electrodes
To realize the optimized output of the direct-current generators, it is imperative to make clear what is the leading factor related to the magnitude of the output. To this aim, three factors: conductive polymers, the contacted electrodes form the Ohmic contact and the contacted electrodes form the Schottky contact are discussed respectively
Fig. 4a presents the output currents (left panel) and output voltages (middle panel) of three direct-current generators (Au/PPy/Al, Au/PEDOT/Al and Au/PANI/Al) with the same electrodes but different conductive polymers. Under 4% compressive deformation, the output currents of Au/PPy/Al, Au/PEDOT/Al and Au/PANI/Al are 120 μA, 15 μA and 80.5 μA, respectively as shown in the left panel of Fig. 4a. The output voltages are 1.3 V, 0.14 V and 0.46 V, as shown in the middle panel of Fig. 4a. The relationship of the output currents/voltages with varied conductive polymers are intuitively shown in the right panel of Fig. 4a. The output currents and the output voltages basically show the same dependency on the conductive polymers. The current density under deformation is given as:
According to the equation, where -Φdeformation is the Schottky barrier variation after the deformation, the output currents are closely related to the Schottky barrier variation before and after deformation: -Φdeformation = (Wnon-deformation-HOMOnon-deformation)- (Wdeformation-LUMOdeformation)= HOMOdeformation- LUMOnon-deformation, that is ΔHOMO (the work functions of the metals are not changed during deformation). Higher ΔHOMO leads to higher Jdeformation. The relative change of HOMO (ΔHOMO/HOMO) of PPy, PEDOT and PANI under 4% deformation are 0.484 %, 0.06% and 0.946 %, respectively. This trend agrees with the observation that the current/voltage output of the generator based on PPy is higher than that of PEDOT. However, although in the deformed state the relative change of HOMO of PANI is higher than that of PPy, the output of Au/PANI/Al is lower than that of Au/PPy/Al. It can be explained by the nature of contacts between conductive polymers and metals. In the Au/PANI/Al generator, both metal layers form Schottky contact with the conductive polymer; in the Au/PPy/Al generator, one metal layer (Al) forms Schottky contact while the other (Au) forms Ohmic contact. These results and analysis demonstrate that the output properties are closely related to the HOMOs of conductive polymers under deformation.
To explore the effect of the metal electrodes on the output properties, we construct various generators with the conductive polymer fixed to be PANI. To begin with, one metal layer is also fixed to be Cu that forms Ohmic contact with PANI, and various metals are adopted in the other metal layer that forms Schottky contact. Thus, generators of Cu/PANI/Al, Cu/PANI/Ag, Cu/PANI/Au, Cu/PANI/Pt are investigated. For this Equation 2, the output currents have close relationship with the -Φdeformation= (W-Hdeformation), where W and Hdeformation represent the work functions of contacted metal and the HOMO of conductive polymers after deformation, respectively. The values of -Φdeformation(Al-PANI) -Φdeformation(Ag-PANI), -Φdeformation (Au-PANI) and -Φdeformation(Pt-PANI) are 0.56, 0.58, -0.26 and -0.81, respectively. The higher -Φdeformation(metal-PANI) (relative value, not absolute value), the higher output current. As shown in the left panel of Fig. 4b, the output currents of the direct-current generators composed of Cu/PANI/Al, Cu/PANI/Ag, Cu/PPy/Au, Cu/PPy/Pt are 69 μA, 800 μA, 180 μA and 20 μA (4% compressive deformation), respectively. As shown in the middle panel of Fig. 4b, output voltages of these direct-current generators are 0.6 V, 3.5 V, 1 V and 0.26 V (4% compressive deformation), respectively. Whatever output currents or output voltages agree very well with the law: higher -Φdeformation (Pt-PANI) leads to the higher output current. It should be emphasized that both the current output (800 μA) and the voltage output (3.5 V) of the Cu/PANI/Al generator are much higher than those in the previous work of H. Shao et al. (maximum voltage 0.8-0.9 V, maximum current 200 μA). The Cu/PANI/Al generator stands out of the investigated generators because under deformation the HOMO of PANI changes much more than that of PPy and PEDOT, such that the value of -Φdeformation(Al-PANI) is much larger. Meanwhile, this generator also ensures that one metal layer forms Schottky contact with the conductive polymer, while the other layer forms Ohmic contact. The relationship of the output currents/voltages with varied metal electrodes which form Schottky contact with the conductive polymers are intuitively shown in the right panel of Fig. 4b. The variation of the current output and the voltage output show basically the same trend.
The calculated energy conversion efficiency is about 16.8% for Al/PANI-HCl/Au generator (Supporting information).
To further illustrate the effect of the metal that forms Schottky contact with the conductive polymers, the polymer PANI in the above study is replaced by PEDOT (Fig. SI. 5) and PPy (Fig. SI. 6), respectively, while, the other setups are the same. The same trend is found in PANI and PPy: higher -Φdeformation (metal-PEDOT/PPy) leads to a higher output current.
In the above study, as well as in H. Shao’s work, one of the two metal electrodes forms an Ohmic contact with the conductive polymer and the other forms a Schottky contact. Can the generators produce current/voltage outputs if both metal electrodes form the Schottky contacts with the conductive polymer? To answer this question, the direct-current generator based on Ag/PANI/Cu, Ag/PANI/Al, Ag/PANI/Au and Ag/PANI/Pt are prepared. It is worth mentioning that the metal electrodes of Al, Ag and Au form the Schottky contacts with PANI, such that in these generators both metals form Schottky contact. To our surprise, these generators still produce current and voltage outputs under deformation (4% in compressive state), as shown in Fig. 4c. As shown in the middle and right panel of Fig. 4c, though the outputs of the generators with both metal forming the Schottky contact with the conductive polymer are smaller (74μA, 0.48V for Ag /PANI/Al; 158 μA, 0.9 V for Ag /PANI/Au; 6 μA, 0.24 V for Ag /PANI/ Pt) than the counterpart of the generators with one Ohmic contact and one Schottky contact (800 μA, 3.5 V for Ag /PANI/Au), the current outputs are still higher than those in general triboelectric nanogenerators (below 10 μA).
2.5 The relationship between the voltage/current output of the direct-current generators and the external loadings
Besides energy-harvesting, the direct-current generators studied above can also serve as sensors, as their outputs show clear relationships with the external stimuli. Our previous work 39 revealed that the output characteristics of the generator based on PPy were closely related to the deformation of the PPy. The mechanism is that the amount of accumulated spatial charges at the PPy/Al electrode interface depends on the degree of deformations, leading to different amount of reduction of the Schottky barrier, and consequently the change in the output current/voltage (according to Equation 2).
To further illustrate the relationship between the voltage/current output of the direct-current generators and the loadings posing on them, voltage/current vs deformation of the Au/PPy/Al generator is illustrated in Fig. SI 7 with the left panel for output current and the right panel for output voltage. The output currents of Au/PPy/Al generator are 40 μA, 89 μA 150 μA and 170 μA for 2%, 4%, 6% and 8% deformation, respectively, while, the output voltages of the Au/PPy/Al generator are 0.15 V, 0.19 V, 0.42V and 0.77V for 2%, 4%, 6% and 8% deformation, respectively. As shown in the right panel of Fig. SI 7, the output current/voltage exhibits nearly linear relationship with the deformation. Furthermore, this linearity also hold true for the outputs of the direct-current generators based on other conductive polymers (Ag/PEDOT/Au and Cu/PANI/Ag), as shown in Fig. SI8 and Fig. SI9. The linear relationship between current/voltage outputs and strain posed on the generators, combined with the large outputs and flexibility presented in prior sections, demonstrate a great promising of our generators for applications in flexible sensors.
2.6 The flexible direct-current generator matrix array for mechanosensation sensor
Recently, self-powered flexible electronics 43 integrated with energy harvesting units (i. e. direct-current generators) have arose extensive attention. Due to the limitation of the fabrication in H. Shao’s previous work, the reported generator was not flexible and therefore not practical to apply to flexible electronics. The electrochemical deposition method adopted here enables an array of direct-current generators based on conductive polymers to be integrated into flexible electronics, via fabricating them on soft elastomeric substrate such as PDMS. The flexibility and stretchability of such device allow the conformal contact with various target objects, which offers a promising avenue to apply as mechanosensation sensor matrix array for applications.
To demonstrate this application, an array of direct-current generators on a PDMS substrate (4 × 4 pixels in cross- intersected configuration) was fabricated to form a mechanosensation sensor matrix array to sense the strain distribution. The fabrication procedure is illustrated in Fig. 5a, with the process being the same as that for the single unit device in Fig. 1c. The mechanosensation sensor matrix array based on the direct-current generator is feasibly attached on a tennis ball conformally (As shown by the photo image in Fig. 5b).The electrical performance of each unit of the sensor array in the sensation matrix is characterized. As the matrix array is attached on a tennis ball, every unit undergoes a different magnitude of compressing deformation and produces a different voltage singal. The sensing signals are closely related to the level of deformation of the conductive polymers according to the strain exerted on each sensing unit. The output voltages of each unit are measured and plotted in 2D black-and-white coordinates (Fig. 5c). Notably, the circuit design of the matrix guarantees that all the output voltages of the sensing units are read out individually without crosstalk, and meanwhile are recorded simultaneously. The corresponding strains exerted on each unit of the mechanosensation sensor matrix are extracted from the output voltages and shown in Fig. 5d. Only the sensing units have the output signals when the strains are posed on them, suggesting no crosstalk in the mechanosensation matrix.