Research Article Cite This: ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
www.acsami.org
n‑Type Ionic−Organic Electronic Ratchets for Energy Harvesting Kenneth Liao, Samuel D. Collins, Viktor V. Brus, Oleksandr V. Mikhnenko, Yuanyuan Hu, Hung Phan, and Thuc-Quyen Nguyen* Center for Polymers and Organic Solids, Department of Chemistry and Biochemistry, University of California at Santa Barbara, Santa Barbara, California 93106, United States
ACS Appl. Mater. Interfaces 2019.11:1081-1087. Downloaded from pubs.acs.org by UNIV OF LOUISIANA AT LAFAYETTE on 01/15/19. For personal use only.
S Supporting Information *
ABSTRACT: Ionic−organic ratchets are three-terminal electronic devices with asymmetric conductivity of the active layer. These devices are capable of generating useful direct current electrical power by converting electromagnetic noise signals available in any environment. In this work, we demonstrate for the first time an n-type ionic−organic ratchet which can generate a current of up to 7.29 μA and power up to 12.5 μW that exceed the values reported for many of the presently state-of-the-art, p-type organic electronic ratchets. We show that n-type ratchets require elimination of electron traps at the SiO2 surface, which is not required in p-type devices. This can be achieved by using a trap-free passivation layer such as benzocyclobutene, where the traditional silane treatment is insufficient. Chemical doping is employed to further fill electron traps in the channel and increase carrier concentration and mobilities. Scanning Kelvin probe force microscopy studies provide evidence of a pn-like rectifying junction in the n-type ratchets fabricated in this work, which inherently differs from the rectification mechanism of previous ionic−organic p-type ratchets. KEYWORDS: PCBM, ions, charge transport, pn-junction, capacitance
■
INTRODUCTION Since the invention and patenting of the rectenna by Brown in 1969,1 the ability to efficiently harness energy from radio frequency (RF) signals has given rise to several important markets and technologies. These include, but are not limited to, wireless sensor networks, RF identification (RFID) tags, and wireless charging of many consumer products such as mobile, medical, and wearable devices.2 Commercially manufactured rectennas employ diodes and capacitors to rectify alternating current (ac) signals collected by antennas and produce direct current (dc) to power electronics. These discrete components are typically fabricated using standard Si photolithography and vapor deposition processes, which normally require high temperatures, high vacuum, and toxic chemicals. A promising alternative to Si-based rectifying circuits is the organic electronic ratchet, which offers several advantages over the traditionally used, inorganic devices. Specifically, the ionic−organic ratchet encompasses the capacitive and rectifying properties of a capacitor and a diode in a single device. This permits a reduction in processing steps, materials, and the number of discrete components required for the rectifying circuit. Organic devices can also be fabricated via solution process methods, allowing flexible, lowcost substrates, and easily scalable production such as inkjet printing and roll-to-roll processing. In 2010, Roeling et al. reported the first room-temperature organic ratchet using a modified organic field-effect transistor © 2018 American Chemical Society
(OFET) that is capable of producing a net source−drain current, even in the absence of a source−drain bias.3 This device is the organic equivalent of inorganic flashing ratchets.4,5 The current is driven by a time-varying, asymmetric potential in the channel. The channel potential is generated by carefully controlled, out-of-phase signals applied to multiple sets of interdigitated electrodes which are embedded within the dielectric material. Thus, although Roeling’s ratchet is capable of producing an output current and a power of 1 μA and 4.5 μW at room temperature, respectively, commercial application in energy harvesters is impractical due to the complexity of managing the input signals and device structure and fabrication. In 2015, Mikhnenko and co-workers reported an ionic− organic ratchet capable of producing currents and powers as large as 2.6 μA and 2.8 μW, respectively.6 This new type of organic electronic ratchet uses a simple OFET structure (see Figure 1a) without the need for additional, embedded electrodes as in the case of flashing ratchets. Shortly after, Brus and co-authors demonstrated the simplicity and low-cost fabrication of ionic−organic ratchets by creating hand-drawn devices using an aluminum foil, a scotch tape, and a graphite pencil with output powers as high as 1.6 μW.7 More recently, Received: September 5, 2018 Accepted: November 27, 2018 Published: November 27, 2018 1081
DOI: 10.1021/acsami.8b15042 ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
Research Article
ACS Applied Materials & Interfaces
Figure 1. (a) Schematic diagram of the ratchet device structure and electrical connections during device operation. The positions of ions after stressing are indicated: as predicted by the electrochemical doping (ECD) model and substantiated by SKPFM measurements (see text). Va is the voltage amplitude of the applied ac signal. (b) Chemical structures of the n-type, small-molecule PCBM and the organic salt/dopant TBABr. (c) Current−voltage measurement of a 15 mol % TBABr-doped PCBM ratchet in the presence of a 5 MHz square wave signal of either 0 V (black, dashed) or 10 V (solid, red) amplitude applied to the gate electrode. Forward and backward scans are shown. The short-circuit current (green dot), open-circuit voltage (blue square), and maximum power point (black diamond) are reported. (d) Short-circuit current and output power of the ratchet device as a function of frequency. The slope is shown for the short-circuit current within the linear fit indicated by the red line.
This could enhance the output currents of solar cells by supplementing the photogenerated current with a ratchetgenerated current. Here, we demonstrate for the first time an n-type ionic− organic ratchet which can generate currents and power that exceed values reported for many of the presently state-of-theart, p-type organic electronic ratchets. Electron trapping at the SiO2 surface of Si/SiO2 substrates is shown to inhibit significant n-type ratchet performance. An organic polymer dielectric is required to eliminate traps, where silane passivation is inadequate. Chemical doping is employed to further fill electron traps in the channel and increase carrier concentration and mobilities. Scanning Kelvin probe force microscopy (SKPFM) studies provide evidence of a pn-like rectifying junction in the n-type ratchets fabricated in this work, which inherently differs from the rectification mechanism of previous ionic−organic p-type ratchets. Furthermore, impedance spectroscopy measurements reveal a frequencydependent device capacitance which is contingent upon the dopant concentration. This effect modifies the frequency dependence of the generated current by reducing the current at higher frequencies. Finally, we discuss how these results impact the optimization procedures of future ionic−organic ratchets to achieve higher currents, power, and efficiencies.
Hu and co-authors have reported an ionic−organic ratchet based on a high mobility polymer capable of producing currents as high as 96.7 μA and powers as high as 169.1 μW, making them the best performing organic electronic ratchets to date.8 The promise of ionic−organic ratchets is evidenced by the increase of more than an order of magnitude in device currents and power in less than 1 year from the date of publication of the first ionic−organic ratchet by Mikhnenko. However, the most important feature of the ionic−organic ratchet is its ability to rectify arbitrary waveforms, including electrical noise. This makes the ionic−organic ratchet versatile and a truly practical and attractive candidate for use as a rectifier in energy-harvesting applications. The devices mentioned above all utilize p-type organic semiconductors (OSCs) for charge transport: Roeling’s ratchet used pentacene, Mikhnenko used poly(3-hexylthiophene-2,5diyl) (P3HT), and both Brus and Hu used poly[4-(4,4dihexadecyl-4H-cyclopenta[1,2-b:5,4-b′]dithiophen-2-yl)-alt[1,2,5] thiadiazolo[3,4-c]pyridine]. There has not yet been a report of an organic electronic ratchet in which an n-type OSC acts as the charge transport layer. There are benefits to exploring n-type organic electronic ratchets that include choice of cheaper metals (p-type ratchets typically require Au electrodes), better control of the energy level alignment between the OSC and electrode work function, and choice of trap-free organic semiconducting materials.9 These factors are of immediate interest as organic electronic ratchets are currently in their infancy and a drive toward improved efficiencies will surely require these degrees of freedom in device engineering. Additionally, new theoretical devices such as the solar ratchet are possible in which common organic photovoltaic materials such as phenyl-C61 butyric acid methyl ester (PCBM) and P3HT are blended together in a single device.10 The solar ratchet would collect energy from absorption of sunlight and RF energy from a coupled antenna.
■
BACKGROUND OF THE WORKING MECHANISM IN P-TYPE RATCHETS The defining feature required in all electronic ratchets is an asymmetry in the device, which when driven out of equilibrium can produce a unidirectional current. In Roeling’s flashing ratchet, this is achieved by an asymmetric, time-varying channel potential. In ionic−organic ratchets, an asymmetric potential is created by rearrangement of ions in the OSC film. The ionic−organic ratchet utilizes an OSC:salt blend as the charge transport layer. The salt is carefully chosen so as to 1082
DOI: 10.1021/acsami.8b15042 ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
Research Article
ACS Applied Materials & Interfaces exhibit some mobility in the OSC film. Applying a voltage stress between the source and drain electrodes rearranges ions in the channel such that injection barriers at the two electrodes are asymmetrically altered. For p-type devices, for example, applying a negative drain bias leads to accumulation of cations at the drain and anions at the source. This leads to enhanced hole injection at the source and creates a large injection barrier at the drain, while extraction is relatively unaffected.6 In effect, the source becomes Ohmic while the drain interface exhibits diode-like behavior. If a negative potential is applied to the gate (of the now asymmetric device), holes are injected to the channel almost exclusively from the source to charge the channel-gate capacitor. When the gate potential is switched to a positive value, holes are extracted equally to the source and drain so that half of the total induced charge has moved from source to drain. This leads to a net source−drain current when the channel-gate capacitor is sequentially charged and discharged by an ac signal applied to the gate. The operating mechanism of this device can thus be described as a charge pump. Isc = 2ηCVaf
device deviation did not exceed 15%. We reported characteristics of the best performing device. Devices are stressed by applying a drain voltage of −50 V for 10 min at 100 °C with source and gate electrodes grounded. An elevated temperature is required for ion rearrangement in our devices because of the relatively low diffusivity of TBABr in the PCBM film at room temperature. As a result, ion relaxation is very slow after cooling to room temperature and devices continue to show asymmetry for several weeks. Current−voltage (I−V) characteristics of a ratchet are obtained by measuring the source− drain current while sweeping the drain voltage from −10 to 10 V and then back to −10 V; the source electrode is always kept at zero potential. Figure 1c shows that when the gate electrode is grounded, the I−V curve crosses the origin (dashed black line). The device thus generates no power in the absence of an applied ac signal. If an ac signal is applied to the gate, a large source−drain current is observed even when there is a zero potential difference between these electrodes (see solid red curve in Figure 1c). With the ratchet-generated current traversing the second quadrant, Isc, open-circuit voltage (Voc), and maximum output power (Pmax) values can be extracted. Isc has a value of 7.29 μA while Pmax a value of 12.5 μW for the best performing device. It is important to note that there is no hysteresis observed in the I−V characteristics in Figure 1c, indicating no substantial ion movement during the I−V scan. The high performance of this n-type device is the result of several important device optimizations. Most importantly, to achieve significant current or power from n-type ionic−organic ratchets, it is necessary to reduce the amount of electron trapping at the semiconductor−dielectric interface. It is wellknown that severe electron trapping readily occurs at interfaces with SiO2.11,12 When electrons are injected into the channel, trap states are first filled until the number of injected charges overcomes the number of trapping states.13−15 Trapped charges are immobile on the time scale of a typical ac period and thus do not contribute to the ratchet current. In this study, BCB was utilized as a passivation layer as simple silane treatments were found to be insufficient at eliminating electron trapping at the SiO2 surface. Synergistically, TBABr was chosen as the ion source for its ability to dope PCBM.16 Extrinsic doping fills residual traps at the PCBM/dielectric interface allowing efficient electron currents during ratchet operation. The combination of BCB passivation and TBABr doping leads to a simultaneous reduction/filling of electron traps and an increase of the electron mobility (see Figures S1 and S2). These important optimizations are required for achieving significant n-type ratchet performance. It was found that an optimum TBABr doping concentration of 15 mol % gave the best ratchet performance. Interestingly, at 15 mol % doping, the device no longer behaves as an FET, but instead acts as a resistor (Figure S2a). It follows that the performance of ratchets is not necessarily tied to FET performance. Instead, optimal devices can be achieved by employing a highly conductive film containing mobile ions. The frequency response of Isc and Pout of the PCBM ratchet are plotted on a log−log scale in Figure 1d for frequencies between 230 kHz and 5 MHz. Measurements above 5 MHz could not be accurately obtained because of limitations of the function generator to provide a well-defined square-wave form of the ac signal. Both Isc and Pout are observed to increase with frequency. According to eq 1, however, the slope of Isc versus
(1)
The generated current, or short-circuit current (Isc), can be described by eq 1, which follows from physical considerations of a charge pump model developed by Mikhnenko.6 Here, η is the charge displacement efficiency, C is the channel-gate capacitance, Va is the voltage amplitude of the applied ac signal, and f is the frequency. η can be thought of as the chargepump efficiency; it is a ratio of the net amount of charge that moves from source to drain in one period of the ac signal and the total charge that can be induced in the channel (CVa). For an ideal device with a single rectifying junction, η has a value of 50%.6,7 The product ηCVa f gives the net amount of charge that moves through the charge pump, per second. This net movement of charge either from source to drain or vice versa is the origin of Isc. The factor of 2 in eq 1 accounts for positive and negative charging of the channel-gate capacitor, for example, in ambipolar devices. Equation 1 is valid up to a maximum frequency ( f max) corresponding to the RC time constant of the device, where R is the total resistance of the device when the rectifying junction is forward biased. At higher frequencies, the channel-gate capacitor can no longer be fully charged or discharged and Isc deviates from a linear relationship with frequency. Peak currents are thus observed at f max.
■
RESULTS AND DISCUSSIONS PCBM Ratchet Characteristics and Performance. Figure 1a shows a schematic of the PCBM ionic−organic ratchet based on a top contact, bottom gate structure. The channel length (L) and width are 60 μm and 1 mm, respectively. The small molecule, PCBM, is used as the ntype OSC while the organic salt, tetrabutyl ammonium bromide (TBABr), acts as the ion source and n-type dopant. The chemical structures of PCBM and TBABr are shown in Figure 1b. The active layer is spun cast from a blend solution containing both the OSC and salt onto the Si/SiO2 substrate, which has been passivated with a thin layer of organic dielectric benzocyclobutene (BCB). The thicknesses of the layers are as follows: SiO2 200 nm, BCB 43 nm, PCBM:TBABr 30 nm, and Ag 80−120 nm. Each of the devices was fabricated on a separate substrate. For each measurement, three independent batches of five devices each were fabricated and tested. The 1083
DOI: 10.1021/acsami.8b15042 ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
Research Article
ACS Applied Materials & Interfaces
Figure 2. (a) Simplified schematic of the device structure and measurement set up used for SKPFM studies. (b) Potential profile from source to drain in the presence of a drain bias of ±5 V before (top) and after (bottom) stressing. Vertical, dashed blue lines demarcate the positions of the source (S) and drain (D) interfaces. (c) Differential resistance profile extracted from the data in (b), using eq 2.
frequency on a log−log plot should yield a value of unity. This is the case with previously reported p-type ionic−organic ratchets but is not observed in PCBM devices fabricated in this study, nor those employing the n-type polymer poly{[N,N9bis(2-octyldodecyl)-naphthalene-1,4,5,8-bis(dicarboximide)2,6-diyl]-alt-5,59-(2,29-bithiophene)} (N2200). The slope of Isc versus frequency in Figure 1d, for example, is only 0.5. This disparity is discussed in the following section. Figure 1d also shows that the current is still increasing at 5 MHz and has not yet reached a point of saturation. Larger currents and power should be attained by driving PCBM ratchets with higherfrequency signals. The maximum generated current is obtained when the device is driven at the frequency f max. In eq 1, C is proportional to L while f max is inversely proportional to L2 (R and C are both proportional to L). Therefore, the maximum Isc of these ratchets can be linearly increased by reducing the channel length. Finally, because f max is inversely proportional to L2, devices can easily be made to operate at an f max equal to the industry standard RFID frequency of 13.56 MHz by simply reducing the channel length below 60 μm. For the device in Figure 1, for example, if we consider 5 MHz to be f max, the extrapolated value of L required to achieve an f max of 13.56 MHz is 36.4 μm. Thus, this type of device should be capable of operating at commercial frequencies. p−n Rectifying Junction and Charge Displacement Efficiency. To better understand ion rearrangement and the rectifying behavior of this n-type ionic−organic ratchet, we employ amplitude-modulated SKPFM. During a SKPFM measurement, a small bias is applied to the drain of the device relative to the source. The potential relative to the source electrode is then measured as a function of position in the channel by a scanning tip as depicted in Figure 2a. A topographical scan is also taken using the same tip operated in tapping mode, giving the position of the source and drain interfaces. Before the device is stressed, the potential drops linearly across the channel for either polarity of the applied drain bias, as shown in Figure 2b (top). This indicates good
Ohmic contacts from the source and drain electrodes to PCBM. The linear potential function also indicates uniformity across the channel, as one would expect before ion rearrangement has been established. Figure 2b (bottom) shows the potential profile in the channel after the device has been stress-annealed at −50 V and 100 °C for 10 min. The potential profile after stressing is dramatically different, with the majority of the potential change confined to a small region in the channel between width 5 and 10 μm . It is important to note that the large potential drop does not occur at either the source or drain interfaces (as for previously reported p-type ratchets), but in the channel away from the interfaces. The dynamics of ion rearrangement in a film by an applied stress voltage has been extensively studied in the light-emitting electrochemical cell (LEC).17 Similar to light-emitting diodes (LEDs), charge recombination is utilized in LECs to release energy in the form of photons. This requires the simultaneous injection of holes and electrons into the twoterminal devices from an external source. Asymmetric injection/extraction barriers are created in LEDs by using metals with different work functions for the anode and cathode. Electron (hole) injection from the anode (cathode) is enhanced while extraction at the cathode (anode) is inhibited to ensure recombination within the active layer. In contrast, LECs typically use the same metal for both the cathode and anode. Asymmetric injection/extraction barriers are instead created by introducing mobile ions into the OSC layer, which can be rearranged by an applied field. When a voltage is applied between the electrodes of an LEC, anions accumulate at the cathode and cations at the anode as depicted in Figure 1a. These accumulations of ions are also known as electronic double layers (EDLs). The anionic EDL improves hole injection but inhibits electron injection at the cathode, whereas the cationic EDL does the opposite at the anode. Two models are commonly used to describe steady-state transport in LECs: the so-called electrodynamic (ED)17−20 and 1084
DOI: 10.1021/acsami.8b15042 ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
Research Article
ACS Applied Materials & Interfaces ECD models.17,21−26 The ED model describes devices which begin with non-Ohmic contacts. When a bias is applied to the device during normal operation, the majority of the voltage drops at the electrode interfaces causing the formation of EDLs which screen the applied field. As a result, negligible fields exist in the bulk of the channel. EDL formation at the contacts improves injection of carriers as described above. By comparison, the ECD model characterizes devices which begin with Ohmic or close to Ohmic contacts. In this case, the applied voltage drops linearly across the channel. Ions respond to the uniform electric field throughout the bulk of the semiconductor layer. Consequently, ion rearrangement in the bulk leads to three distinct regions. One region with high cation density, one with high anion density, and an “intrinsic” region sandwiched between the two which has a low density of both kinds of ions. Such devices are often called dynamic pnjunctions. The p- and n-like regions correspond to areas of high anion and cation densities, respectively. These space charges form a nonuniform electric field within the channel somewhere in the bulk of the device, away from the electrode interfaces and result in the asymmetric electrical conduction of the channel. This is exactly what is observed in PCBM ratchets as shown in Figure 2b (bottom) and matches well with SKPFM data on devices reported as exemplifying the ECD mode. Further evidence for the existence of a pn-like junction in our devices can be seen in the voltage dependence of the total device resistance. From the potential profile in Figure 2b (bottom), one can calculate the differential resistance profile in the channel defined as r (x ) =
1 d V (x ) I dx
pushes the junction further into reverse bias (9 V) or further into forward bias (−9 V). Working Mechanism of the n-Type PCBM Ratchet. In the previous section, we have shown the presence of a pn-like junction in the channel of PCBM ionic−organic ratchets. This results in a slightly different working mechanism in comparison to that of the previously reported p-type ionic−organic ratchets where the asymmetric conductivity is realized due to an injection barrier at a channel/electrode interface. The anionreach (p-like) part of the channel is adjacent to the source electrode and the cation-reach (n-like) one is adjacent to the drain electrode. This explains the rectification polarity: a negative potential applied to the drain electrode and, thus, to the n-like part of the pn-junction lowers down its potential barrier and reduces resistance (forward bias). The location of the pn-junction is shifted toward the drain electrode most likely due to different mobilities of the cations and anions of the organic salt (see Figure 1b). As shown above, there are no injection barriers for electrons at the source and drain electrodes. Therefore, when a positive potential is applied to the gate, electrons are injected from both source and drain electrodes (as opposite to the p-type ratchets). However, as the pn-junction is shifted closer to the drain electrode, a major part of the channel is charged by electrons injected from the source electrode as the internal electric field in the pn-junction suppresses the transport of electrons from the cation-rich part (adjacent to the drain electrode) toward the anion-rich part (see Figure 3a).
(2)
V(x) in eq 2 is the potential at a particular location in the channel, relative to the source electrode, and I is the measured current. The differential resistance profile is plotted in Figure 2c for two bias amplitudes. Figure 2c shows more clearly how the junction position lies in the channel away from the interfaces. The differential resistance peak position corresponds to the center of the junction where the resistance gradient is greatest. Integration of a differential resistance function gives the total resistance of the device at a specific applied bias. For +5 and −5 V, the total device resistances are 11.56 and 7.27 MΩ, respectively. For +9 and −9 V, the total device resistances are 12.34 and 5.57 MΩ, respectively. For either the 5 or 9 V amplitudes, the total resistance is smaller for the negative drain bias, indicating a larger current flow from source to drain. The resistance of a pn-like junction can be modified by an externally applied bias. A forward bias (negative potential applied to the cation-reach region adjacent to the drain electrode) reduces the junction resistance, whereas a reverse bias (positive potential to the cation-reach region adjacent to the drain electrode) increases the junction resistance. This finding well correlates with the rectification polarity of the stressed PCBM ratchet (black dashed line in Figure 1c). The asymmetry in the measured currents and extracted resistances for either ±5 or ±9 V are therefore attributed to the forward and reverse biasing of the pnrectifying junction. Moreover, in going from a 5 to a 9 V amplitude, the difference in total resistance between the positive and negative bias gets larger. This is expected in the presence of a pn-junction where increasing the applied voltage
Figure 3. Schematic representation of the working mechanism of ntype PCBM ionic−organic ratchets.
If the gate potential changes to a negative one, the injected electrons are extracted by the source and drain electrodes with the same probability since the pn-junction does not prevent the transport of electrons from the anion-reach to the cationreach part of the channel (Figure 3b). Thus, the drain electrode collects more electrons than it injects. As a result, within one period, a directed movement of electrons from the source to drain is realized. Therefore, the resultant dc electrical current flows from the drain to source (as shown in Figure 1a) and the I−V characteristic of the operating n-type PCBM ratchet is located in the second quadrant as opposite to p-type ratchets. 1085
DOI: 10.1021/acsami.8b15042 ACS Appl. Mater. Interfaces 2019, 11, 1081−1087
ACS Applied Materials & Interfaces
■
Frequency Dependence of Short-Circuit Current. We now revisit the frequency dependence of Isc. We showed that the slope of Isc versus frequency on a log−log plot is only 0.5 (see Figure 1d) in our devices and deviates from the chargepump model (eq 1) and p-type devices which show a slope of 1. According to eq 1, this difference may stem from the channel-gate capacitance and/or η acquiring a frequency dependence in our PCBM ratchets. If the channel-gate capacitance could be accurately and independently measured as a function of frequency, η could be obtained from eq 1. Impedance spectroscopy was employed to determine the total device capacitance as a function of frequency (see Figure S5a). The results and analysis of the impedance study are presented in detail in Section II of the Supporting Information. To obtain only the channel-gate capacitance, parasitic contributions to the total device capacitance must be subtracted using an equivalent capacitance model. This requires accurate determination of the individual source and drain capacitances, which could not be accomplished with confidence in this study because of the measurement error being comparable to the data values. The junction which also has its own capacitance, as well as its position within the channel, further complicates the calculation of η. Although these issues prevent determination of the absolute value of η, the total device capacitance is expected to carry the same frequency dependence as the channel-gate capacitance. The results suggest that the frequency-dependent capacitance on its own does not account for the reduction of slope of Isc in Figure 1d to a value of 0.5 (see Figure S5b). The remaining frequency dependence of Isc in PCBM ratchets must therefore be associated with a frequency-dependent η. The frequency dependence of both the channel-gate capacitance and η lead to Isc being reduced at higher frequencies. The power density of this device is 0.2083 mW/mm2. The power density depends on many factors, most importantly, the frequency dependence of η. This is described in detail in the Supporting Information. Therefore, understanding the origin of this frequency-dependent η is an important topic to be revisited in our future work.
Research Article
EXPERIMENTAL SECTION
Sb-doped silicon wafers (0.01−0.03 Ω cm) with 200 nm of thermally grown SiO2 were purchased from WRS materials. [6,6]-Phenyl C61butyric acid methyl ester (PCBM) was purchased from Solenne BV. BCB or CYCLOTENE 3022-46 (BCB) was purchased from DOW Chemicals and was diluted from the received solution to 1 wt % in toluene. The organic salt TBABr was purchased from Acros Organics. Stock solutions of pristine PCBM and TBABr were separately made in chloroform and stored in a nitrogen glovebox (
