Megumi Akai-Kasaya | Osaka University (original) (raw)
Papers by Megumi Akai-Kasaya
IEICE Transactions on Information and Systems
Edge computing, which has been gaining attention in recent years, has many advantages, such as re... more Edge computing, which has been gaining attention in recent years, has many advantages, such as reducing the load on the cloud, not being affected by the communication environment, and providing excellent security. Therefore, many researchers have attempted to implement neural networks, which are representative of machine learning in edge computing. Neural networks can be divided into inference and learning parts; however, there has been little research on implementing the learning component in edge computing in contrast to the inference part. This is because learning requires more memory and computation than inference, easily exceeding the limit of resources available for edge computing. To overcome this problem, this research focuses on the optimizer, which is the heart of learning. In this paper, we introduce our new optimizer, hardware-oriented logarithmic momentum estimation (Holmes), which incorporates new perspectives not found in existing optimizers in terms of characteristics and strengths of hardware. The performance of Holmes was evaluated by comparing it with other optimizers with respect to learning progress and convergence speed. Important aspects of hardware implementation, such as memory and operation requirements are also discussed. The results show that Holmes is a good match for edge computing with relatively low resource requirements and fast learning convergence. Holmes will help create an era in which advanced machine learning can be realized on edge computing.
![Research paper thumbnail of Adsorption and Light Emission of a Racemic Mixture of [7]thiaheterohelicene-2,13-carboxaldehyde on Au(111), Cu(001), and NiAl(110) Surfaces Investigated Using a Scanning Tunneling Microscope](https://attachments.academia-assets.com/112987780/thumbnails/1.jpg)
](The Journal of Physical Chemistry C, 2021
The coverage-dependent adsorption of a racemic mixture of [7]thiaheterohelicene-2,13-carboxaldehy... more The coverage-dependent adsorption of a racemic mixture of [7]thiaheterohelicene-2,13-carboxaldehyde on Au(111), Cu(001), and NiAl(110) surfaces was investigated using a scanning tunneling microscope (STM). At a low coverage, the adsorption process for helicene molecules on Au(111) was strongly affected by surface reconstruction. At monolayer saturation coverage, the dominant molecular structure observed on Au(111) characteristically had separated self-assembled twin rows aligned in ⟨112̅ ⟩ directions. The helicene molecules within these separated twin rows are preferentialy arranged in a zigzag pattern with alternating (M)-and (P)-enantiomers. With increasing molecular coverage, the molecular structural transition from self-assembled twin rows to self-assembled single rows was observed. STM-induced light emission (STM-LE) investigation of helicene molecules on Au(111) showed the suppression of plasmon light emission over the molecules. The adsorption of helicene molecules on Cu(001) and NiAl(110) was quite different from that observed on Au(111). Neither the formation of self-assembled twin rows nor the molecular arrangement into different domains was observed. The formation of molecular clusters on Cu(001) and NiAl(110) was observed. STM-LE investigation of helicene molecules adsorbed on Cu(001) showed the suppression of plasmon light emission over these molecules. In contrast to metallic Au(111) and Cu(001) substrates, STM-LE investigations revealed the enhancement of light emission above the molecular clusters formed on metallic NiAl(110), suggesting plasmon-enhanced molecular light emission.
SPIE Proceedings, 2013
ABSTRACT Some of Morpho butterfly species have a mysterious physical coloration. Their blue color... more ABSTRACT Some of Morpho butterfly species have a mysterious physical coloration. Their blue color has both high reflectivity (>60%) and a single color in too wide angular range (> ± 40° from the normal), which are contradicting with each other from viewpoint of the optical interference. A key to the mechanism of the specific Morpho-color was suggested to be the nano-randomness in arrangement of the nanostructures on its scale, which prevents the rainbow interference. However, concrete optical roles of the nano-randomness remained still unclear. Using finite-difference time-domain (FDTD) analysis, we have recently investigated the optical role of different kinds of randomness in the nanostructure on the Morpho butterfly’s scale. The results revealed clearly several independent roles of different kinds of randomness. On the other hand, by inproving the accuracy of simulation, we have found new aspects on the analysis, especially for the number of random components (nano-trees). These new aspects will give important hint and caution to futher simulation on the optical properties of this specific colorations that have wide potential applications. The direction obtained by the numerical simulations to analyze optically complex random structures will serve not only to understand the scientific principles, but also to design the optical properties of artificial materials.
Surface Science, 2005
A significant improvement of the conductivity of polydiacetylene thin films by iodine doping was ... more A significant improvement of the conductivity of polydiacetylene thin films by iodine doping was observed using independently driven double-tip scanning tunneling microscopy. One-dimensional conduction along the polydiacetylene backbone was obtained and the conductivity of the iodine-doped polydiacetylene thin film was estimated to be (3 ± 0.3) • 10 À3 S/cm, which is three orders of magnitude higher than that of the nondoped polydiacetylene thin film and five orders of magnitude higher than that reported previously. The results of visible light absorption spectroscopy and atomic force microscopy showed that distinct changes in the electronic properties of the polydiacetylene backbone in the absence of a close-packed arrangement of polydiacetylene molecules.
Surface and Interface Analysis, 2008
We have measured a nanoscale fluorescence image of copper phthalocyanine (CuPc)/Au nanostructures... more We have measured a nanoscale fluorescence image of copper phthalocyanine (CuPc)/Au nanostructures on an indium tin oxide (ITO) substrate using a scanning tunneling microscope (STM). The quantum efficiency (QE) of the intrinsic fluorescence of CuPc is very low; however, the plasmon enhancement effect on an Au surface increases fluorescence efficiency. To obtain the nanoscale fluorescence image, in this study, we used up‐conversion fluorescence through a triplet‐triplet annihilation (TTA) process. Tunnel electrons induced fluorescence and tip‐induced plasmon (TIP) emission simultaneously; however, the up‐conversion fluorescence from CuPc could be clearly separated from TIP emission by spectroscopic filtering. The TIP showed different properties between the ITO substrate and the Au surface, and the CuPc fluorescence was enhanced only on the TIP of Au nanostructures. This study would lead to the application of a new fluorescence imaging method based on plasmon‐assisted STM‐induced light...
Japanese Journal of Applied Physics, 2006
We extend our previous formulation of low-energy QCD in terms of an effective lagrangean containi... more We extend our previous formulation of low-energy QCD in terms of an effective lagrangean containing operators of dimensionality d ≤ 6 constructed with pseudoscalars and quark fields, describing physics below the scale of chiral symmetry breaking. We include in this paper the vector and axial-vector channels. We follow closely the Extended Chiral Quark Model approach and consistently work in the large-N c and leading log approximation and take into account the constraints from chiral symmetry and chiral symmetry restoration. The optimal fit of all parameters gives further support to a heavy scalar meson with a mass ∼ 1 GeV and a value of the axial pion-quark coupling constant g A 0.55 to 0.66, depending on some assumptions concerning the Weinberg sum rules.
Surface Science, 1998
Semi-empirical molecular orbital calculations reveal the local surface density of states for the ... more Semi-empirical molecular orbital calculations reveal the local surface density of states for the adsorbed molecules on the Si surface. The organic molecules adenine, thymine, cytosine, and pentacene, which are adsorbed on Si(100)2×1 surfaces have been imaged by scanning tunneling microscopy (STM). The molecular images obtained by STM exhibit distinct shapes corresponding to the expected shapes for adsorption configurations. The energy level diagrams of the molecular orbitals (MOs) of the Si cluster on which the molecules are adsorbed are shown. The calculated MOs for adenine and thymine are in good agreement with the molecular images observed in STM. The bias dependence image of adsorbed cytosine is also explained by the calculated MOs of the molecule.
Physical Review Letters, 2003
A molecular wire candidate, the polydiacetylene chain, fabricated in a substantial support layer ... more A molecular wire candidate, the polydiacetylene chain, fabricated in a substantial support layer of monomers self-assembled on a highly ordered pyrolytic graphite surface, was studied using scanning tunneling microscopy and spectroscopy. The density of states of individual polymers and constituent monomers were observed on the same surface, and then compared with the calculated results. The spectrum delineating the density of states of the polydiacetylene wire clearly reveals the theoretically predicted-band and band edge singularities of the one-dimensional polymer.
Applied Physics Letters, 2014
Currently, single walled carbon nanotubes (SWNTs) field effect transistor (FET) devices can be fa... more Currently, single walled carbon nanotubes (SWNTs) field effect transistor (FET) devices can be fabricated by either chemical vapor deposition (CVD) or solution casting. Comparing to the CVD-based
Applied Physics Letters, 2014
We fabricated single-walled carbon nanotube (SWNT) field-effect transistor (FET) devices on flatt... more We fabricated single-walled carbon nanotube (SWNT) field-effect transistor (FET) devices on flattened electrodes, in which there are no height difference between metal electrodes and the substrate. SWNT-FET fabricated using bottom contact technique have some advantages, such that the SWNTs are free from electron irradiation, have direct contact with the desired metal electrodes, and can be functionalized before or after deposition. However, the SWNTs can be bent at the contact point with the metal electrodes leading to a different electrical characteristic of the devices. The number of SWNT direct junctions in short channel length devices is drastically increased by the use of flattened electrodes due to strong attractive interaction between SWNT and the substrate. The flattened electrodes show a better balance between their hole and electron mobility compared to that of the non-flattened electrodes, that is, ambipolar FET characteristic. It is considered that bending of the SWNTs in the non-flattened electrode devices results in a higher Schottky barrier for the electrons.
Physical Review Letters, 2015
Electronic transport was investigated in poly(3-hexylthiophene-2,5-diyl) monolayers. At low tempe... more Electronic transport was investigated in poly(3-hexylthiophene-2,5-diyl) monolayers. At low temperatures , nonlinear behavior was observed in the current-voltage characteristics, and a nonzero threshold voltage appeared that increased with decreasing temperature. The current-voltage characteristics could be best fitted using a power law. These results suggest that the nonlinear conductivity can be explained using a Coulomb blockade (CB) mechanism. A model is proposed in which an isotropic extended charge state exists, as predicted by quantum calculations, and percolative charge transport occurs within an array of small conductive islands. Using quantitatively evaluated capacitance values for the islands, this model was found to be capable of explaining the observed experimental data. It is, therefore, suggested that percolative charge transport based on the CB effect is a significant factor giving rise to nonlinear conductivity in organic materials. The interface between an organic semiconductor and a dielectric layer plays a critical role in carrier transport in organic field-effect transistors (OFETs), because the intrinsic transport characteristics are governed by only a few molecular layers at the interface. Good organic conductors often have a low-dimensional configuration, e.g., quasi-one-dimensional (1D) structures or two-dimensional (2D) layers. Nevertheless, a large number of fundamental questions remain to be answered regarding the charge-transport mechanism, particularly in low-dimensional structures. One such question concerns the nonlinear behavior that is often observed in the current-voltage (I-V) characteristics of organic conductors. Even for materials that exhibit good linear I-V characteristics near room temperature (RT), nonlinearity can occur as the temperature is reduced, reflecting a decrease in conductivity. This effect has been interpreted using a variety of mechanisms-such as charge hopping, trapping, tunneling, and emission-either within the organic material or at the interfaces. Explanations have been proposed involving classical physical models previously developed for inorganic materials; what these explanations have in common is that the expression for the current contains an exponential term involving the electric field (E) and the temperature (T). However, the observed nonlinearity cannot be fully explained using these conventional models or combinations of them. Recently, it has been reported that the I-V characteristics obey a power-law relationship in low-dimensional organic materials such as polymer nanofibers [1], nanotubes [2], and polymer films [3,4], as is the case for carbon nanotubes and inorganic quantum wires. The observed power-law relationship for polymer materials has been put forward as evidence for tunneling into a 1D Luttinger liquid because of the quasi-1D structure of these materials [2,3]; however, power-law behavior was also observed for a three-dimensional (3D) polymer film [4]. The origin of such power-law behavior in organic materials is still under debate [5]. On the other hand, in inorganic granular materials, the power-law dependence of the I-V characteristics has commonly been attributed to dissipative tunneling processes, such as that associated with a Coulomb blockade (CB) [6-8]. The CB effect has been confirmed during charge transport through a single molecule spanning adjacent electrodes [9,10], although it has rarely been suggested as the origin of nonlinear conduction in larger condensed organic conductor systems [4,11,12]. CB transport occurs in systems consisting of an array of small conductive islands connected by narrow junctions, provided the tunneling resistance between neighboring sites is significantly larger than the quantum resistance (≫h=e 2), the capacitance associated with each island is sufficiently small, and the energy corresponding to an additional electron charge at each site is large compared to k B T. Here, h is Planck's constant, e is the charge of an electron, and k B is Boltzmann's constant. Since the nature of the individual sites (i.e., whether they are metallic, superconducting, or semiconducting) is irrelevant [13], there is no reason why the CB effect should not emerge in organic materials that consist of small conducting segments. In the present Letter, an investigation into charge transport is carried out through a 2D conjugated polymer PRL 115, 196801 (2015) P H Y S I C A L
Nanoscale, 2017
Detection and use of physical noise fluctuations in a signal provides significant advantages in t... more Detection and use of physical noise fluctuations in a signal provides significant advantages in the development of bio-and neuro-sensing and functional mimicking devices. Low-dimensional carbon nano-materials are a good candidate for use in noise generation due to the high surface sensitivity of these materials, which may themselves serve as the main building blocks of these devices. Here, we demonstrate that the addition of a molecule with high redox activity to a carbon nanotube (CNT) field-effect transistor provides tunable current fluctuation noise. A unique charge-trap state in the vicinity of the CNT surface due to the presence of the single molecule is the origin of the noise, which generates a prominent and unique slow discrete random telegraph signal in the device current. The power spectral density reveals the peculiar frequency limit of the fluctuation for different types of molecules depending on their redox activity and adsorption configuration. These results indicate that the detected noise will provide new opportunities to obtain electronic information for a single molecule combined with a nanotube surface, and that controllability of the noise may contribute to the expansion of noise utilization in future bio-inspired devices.
APPLIED PHYSICS LETTERS, 2017
Stochastic resonance (SR) is an intrinsic noise usage system for small-signal sensing found in va... more Stochastic resonance (SR) is an intrinsic noise usage system for small-signal sensing found in various living creatures. The noise-enhanced signal transmission and detection system, which is probabilistic but consumes low power, has not been used in modern electronics. We demonstrated SR in a summing network based on a single-walled carbon nanotube (SWNT) device that detects small subthreshold signals with very low current flow. The nonlinear current-voltage characteristics of this SWNT device, which incorporated Cr electrodes, were used as the threshold level of signal detection. The adsorption of redox-active polyoxometalate molecules on SWNTs generated additional noise, which was utilized as a self-noise source. To form a summing network SR device, a large number of SWNTs were aligned parallel to each other between the electrodes, which increased the signal detection ability. The functional capabilities of the present small-size summing network SR device, which rely on dense nanomaterials and exploit intrinsic spontaneous noise at room temperature, offer a glimpse of future bio-inspired electronic devices.
Nature communication, 2018
In contrast to AI hardware, neuromorphic hardware is based on neuroscience, wherein constructing ... more In contrast to AI hardware, neuromorphic hardware is based on neuroscience, wherein constructing both spiking neurons and their dense and complex networks is essential to obtain intelligent abilities. However, the integration density of present neuromorphic devices is much less than that of human brains. In this report, we present molecular neuromorphic devices, composed of a dynamic and extremely dense network of single-walled carbon nanotubes (SWNTs) complexed with polyoxometalate (POM). We show experimentally that the SWNT/POM network generates spontaneous spikes and noise. We propose electron-cascading models of the network consisting of heterogeneous molecular junctions that yields results in good agreement with the experimental results. Rudimentary learning ability of the network is illustrated by introducing reservoir computing, which utilises spiking dynamics and a certain degree of network complexity. These results indicate the possibility that complex functional networks can be constructed using molecular devices, and contribute to the development of neuromorphic devices.
IEICE Transactions on Information and Systems
Edge computing, which has been gaining attention in recent years, has many advantages, such as re... more Edge computing, which has been gaining attention in recent years, has many advantages, such as reducing the load on the cloud, not being affected by the communication environment, and providing excellent security. Therefore, many researchers have attempted to implement neural networks, which are representative of machine learning in edge computing. Neural networks can be divided into inference and learning parts; however, there has been little research on implementing the learning component in edge computing in contrast to the inference part. This is because learning requires more memory and computation than inference, easily exceeding the limit of resources available for edge computing. To overcome this problem, this research focuses on the optimizer, which is the heart of learning. In this paper, we introduce our new optimizer, hardware-oriented logarithmic momentum estimation (Holmes), which incorporates new perspectives not found in existing optimizers in terms of characteristics and strengths of hardware. The performance of Holmes was evaluated by comparing it with other optimizers with respect to learning progress and convergence speed. Important aspects of hardware implementation, such as memory and operation requirements are also discussed. The results show that Holmes is a good match for edge computing with relatively low resource requirements and fast learning convergence. Holmes will help create an era in which advanced machine learning can be realized on edge computing.
The Journal of Physical Chemistry C, 2021
The coverage-dependent adsorption of a racemic mixture of [7]thiaheterohelicene-2,13-carboxaldehy... more The coverage-dependent adsorption of a racemic mixture of [7]thiaheterohelicene-2,13-carboxaldehyde on Au(111), Cu(001), and NiAl(110) surfaces was investigated using a scanning tunneling microscope (STM). At a low coverage, the adsorption process for helicene molecules on Au(111) was strongly affected by surface reconstruction. At monolayer saturation coverage, the dominant molecular structure observed on Au(111) characteristically had separated self-assembled twin rows aligned in ⟨112̅ ⟩ directions. The helicene molecules within these separated twin rows are preferentialy arranged in a zigzag pattern with alternating (M)-and (P)-enantiomers. With increasing molecular coverage, the molecular structural transition from self-assembled twin rows to self-assembled single rows was observed. STM-induced light emission (STM-LE) investigation of helicene molecules on Au(111) showed the suppression of plasmon light emission over the molecules. The adsorption of helicene molecules on Cu(001) and NiAl(110) was quite different from that observed on Au(111). Neither the formation of self-assembled twin rows nor the molecular arrangement into different domains was observed. The formation of molecular clusters on Cu(001) and NiAl(110) was observed. STM-LE investigation of helicene molecules adsorbed on Cu(001) showed the suppression of plasmon light emission over these molecules. In contrast to metallic Au(111) and Cu(001) substrates, STM-LE investigations revealed the enhancement of light emission above the molecular clusters formed on metallic NiAl(110), suggesting plasmon-enhanced molecular light emission.
SPIE Proceedings, 2013
ABSTRACT Some of Morpho butterfly species have a mysterious physical coloration. Their blue color... more ABSTRACT Some of Morpho butterfly species have a mysterious physical coloration. Their blue color has both high reflectivity (>60%) and a single color in too wide angular range (> ± 40° from the normal), which are contradicting with each other from viewpoint of the optical interference. A key to the mechanism of the specific Morpho-color was suggested to be the nano-randomness in arrangement of the nanostructures on its scale, which prevents the rainbow interference. However, concrete optical roles of the nano-randomness remained still unclear. Using finite-difference time-domain (FDTD) analysis, we have recently investigated the optical role of different kinds of randomness in the nanostructure on the Morpho butterfly’s scale. The results revealed clearly several independent roles of different kinds of randomness. On the other hand, by inproving the accuracy of simulation, we have found new aspects on the analysis, especially for the number of random components (nano-trees). These new aspects will give important hint and caution to futher simulation on the optical properties of this specific colorations that have wide potential applications. The direction obtained by the numerical simulations to analyze optically complex random structures will serve not only to understand the scientific principles, but also to design the optical properties of artificial materials.
Surface Science, 2005
A significant improvement of the conductivity of polydiacetylene thin films by iodine doping was ... more A significant improvement of the conductivity of polydiacetylene thin films by iodine doping was observed using independently driven double-tip scanning tunneling microscopy. One-dimensional conduction along the polydiacetylene backbone was obtained and the conductivity of the iodine-doped polydiacetylene thin film was estimated to be (3 ± 0.3) • 10 À3 S/cm, which is three orders of magnitude higher than that of the nondoped polydiacetylene thin film and five orders of magnitude higher than that reported previously. The results of visible light absorption spectroscopy and atomic force microscopy showed that distinct changes in the electronic properties of the polydiacetylene backbone in the absence of a close-packed arrangement of polydiacetylene molecules.
Surface and Interface Analysis, 2008
We have measured a nanoscale fluorescence image of copper phthalocyanine (CuPc)/Au nanostructures... more We have measured a nanoscale fluorescence image of copper phthalocyanine (CuPc)/Au nanostructures on an indium tin oxide (ITO) substrate using a scanning tunneling microscope (STM). The quantum efficiency (QE) of the intrinsic fluorescence of CuPc is very low; however, the plasmon enhancement effect on an Au surface increases fluorescence efficiency. To obtain the nanoscale fluorescence image, in this study, we used up‐conversion fluorescence through a triplet‐triplet annihilation (TTA) process. Tunnel electrons induced fluorescence and tip‐induced plasmon (TIP) emission simultaneously; however, the up‐conversion fluorescence from CuPc could be clearly separated from TIP emission by spectroscopic filtering. The TIP showed different properties between the ITO substrate and the Au surface, and the CuPc fluorescence was enhanced only on the TIP of Au nanostructures. This study would lead to the application of a new fluorescence imaging method based on plasmon‐assisted STM‐induced light...
Japanese Journal of Applied Physics, 2006
We extend our previous formulation of low-energy QCD in terms of an effective lagrangean containi... more We extend our previous formulation of low-energy QCD in terms of an effective lagrangean containing operators of dimensionality d ≤ 6 constructed with pseudoscalars and quark fields, describing physics below the scale of chiral symmetry breaking. We include in this paper the vector and axial-vector channels. We follow closely the Extended Chiral Quark Model approach and consistently work in the large-N c and leading log approximation and take into account the constraints from chiral symmetry and chiral symmetry restoration. The optimal fit of all parameters gives further support to a heavy scalar meson with a mass ∼ 1 GeV and a value of the axial pion-quark coupling constant g A 0.55 to 0.66, depending on some assumptions concerning the Weinberg sum rules.
Surface Science, 1998
Semi-empirical molecular orbital calculations reveal the local surface density of states for the ... more Semi-empirical molecular orbital calculations reveal the local surface density of states for the adsorbed molecules on the Si surface. The organic molecules adenine, thymine, cytosine, and pentacene, which are adsorbed on Si(100)2×1 surfaces have been imaged by scanning tunneling microscopy (STM). The molecular images obtained by STM exhibit distinct shapes corresponding to the expected shapes for adsorption configurations. The energy level diagrams of the molecular orbitals (MOs) of the Si cluster on which the molecules are adsorbed are shown. The calculated MOs for adenine and thymine are in good agreement with the molecular images observed in STM. The bias dependence image of adsorbed cytosine is also explained by the calculated MOs of the molecule.
Physical Review Letters, 2003
A molecular wire candidate, the polydiacetylene chain, fabricated in a substantial support layer ... more A molecular wire candidate, the polydiacetylene chain, fabricated in a substantial support layer of monomers self-assembled on a highly ordered pyrolytic graphite surface, was studied using scanning tunneling microscopy and spectroscopy. The density of states of individual polymers and constituent monomers were observed on the same surface, and then compared with the calculated results. The spectrum delineating the density of states of the polydiacetylene wire clearly reveals the theoretically predicted-band and band edge singularities of the one-dimensional polymer.
Applied Physics Letters, 2014
Currently, single walled carbon nanotubes (SWNTs) field effect transistor (FET) devices can be fa... more Currently, single walled carbon nanotubes (SWNTs) field effect transistor (FET) devices can be fabricated by either chemical vapor deposition (CVD) or solution casting. Comparing to the CVD-based
Applied Physics Letters, 2014
We fabricated single-walled carbon nanotube (SWNT) field-effect transistor (FET) devices on flatt... more We fabricated single-walled carbon nanotube (SWNT) field-effect transistor (FET) devices on flattened electrodes, in which there are no height difference between metal electrodes and the substrate. SWNT-FET fabricated using bottom contact technique have some advantages, such that the SWNTs are free from electron irradiation, have direct contact with the desired metal electrodes, and can be functionalized before or after deposition. However, the SWNTs can be bent at the contact point with the metal electrodes leading to a different electrical characteristic of the devices. The number of SWNT direct junctions in short channel length devices is drastically increased by the use of flattened electrodes due to strong attractive interaction between SWNT and the substrate. The flattened electrodes show a better balance between their hole and electron mobility compared to that of the non-flattened electrodes, that is, ambipolar FET characteristic. It is considered that bending of the SWNTs in the non-flattened electrode devices results in a higher Schottky barrier for the electrons.
Physical Review Letters, 2015
Electronic transport was investigated in poly(3-hexylthiophene-2,5-diyl) monolayers. At low tempe... more Electronic transport was investigated in poly(3-hexylthiophene-2,5-diyl) monolayers. At low temperatures , nonlinear behavior was observed in the current-voltage characteristics, and a nonzero threshold voltage appeared that increased with decreasing temperature. The current-voltage characteristics could be best fitted using a power law. These results suggest that the nonlinear conductivity can be explained using a Coulomb blockade (CB) mechanism. A model is proposed in which an isotropic extended charge state exists, as predicted by quantum calculations, and percolative charge transport occurs within an array of small conductive islands. Using quantitatively evaluated capacitance values for the islands, this model was found to be capable of explaining the observed experimental data. It is, therefore, suggested that percolative charge transport based on the CB effect is a significant factor giving rise to nonlinear conductivity in organic materials. The interface between an organic semiconductor and a dielectric layer plays a critical role in carrier transport in organic field-effect transistors (OFETs), because the intrinsic transport characteristics are governed by only a few molecular layers at the interface. Good organic conductors often have a low-dimensional configuration, e.g., quasi-one-dimensional (1D) structures or two-dimensional (2D) layers. Nevertheless, a large number of fundamental questions remain to be answered regarding the charge-transport mechanism, particularly in low-dimensional structures. One such question concerns the nonlinear behavior that is often observed in the current-voltage (I-V) characteristics of organic conductors. Even for materials that exhibit good linear I-V characteristics near room temperature (RT), nonlinearity can occur as the temperature is reduced, reflecting a decrease in conductivity. This effect has been interpreted using a variety of mechanisms-such as charge hopping, trapping, tunneling, and emission-either within the organic material or at the interfaces. Explanations have been proposed involving classical physical models previously developed for inorganic materials; what these explanations have in common is that the expression for the current contains an exponential term involving the electric field (E) and the temperature (T). However, the observed nonlinearity cannot be fully explained using these conventional models or combinations of them. Recently, it has been reported that the I-V characteristics obey a power-law relationship in low-dimensional organic materials such as polymer nanofibers [1], nanotubes [2], and polymer films [3,4], as is the case for carbon nanotubes and inorganic quantum wires. The observed power-law relationship for polymer materials has been put forward as evidence for tunneling into a 1D Luttinger liquid because of the quasi-1D structure of these materials [2,3]; however, power-law behavior was also observed for a three-dimensional (3D) polymer film [4]. The origin of such power-law behavior in organic materials is still under debate [5]. On the other hand, in inorganic granular materials, the power-law dependence of the I-V characteristics has commonly been attributed to dissipative tunneling processes, such as that associated with a Coulomb blockade (CB) [6-8]. The CB effect has been confirmed during charge transport through a single molecule spanning adjacent electrodes [9,10], although it has rarely been suggested as the origin of nonlinear conduction in larger condensed organic conductor systems [4,11,12]. CB transport occurs in systems consisting of an array of small conductive islands connected by narrow junctions, provided the tunneling resistance between neighboring sites is significantly larger than the quantum resistance (≫h=e 2), the capacitance associated with each island is sufficiently small, and the energy corresponding to an additional electron charge at each site is large compared to k B T. Here, h is Planck's constant, e is the charge of an electron, and k B is Boltzmann's constant. Since the nature of the individual sites (i.e., whether they are metallic, superconducting, or semiconducting) is irrelevant [13], there is no reason why the CB effect should not emerge in organic materials that consist of small conducting segments. In the present Letter, an investigation into charge transport is carried out through a 2D conjugated polymer PRL 115, 196801 (2015) P H Y S I C A L
Nanoscale, 2017
Detection and use of physical noise fluctuations in a signal provides significant advantages in t... more Detection and use of physical noise fluctuations in a signal provides significant advantages in the development of bio-and neuro-sensing and functional mimicking devices. Low-dimensional carbon nano-materials are a good candidate for use in noise generation due to the high surface sensitivity of these materials, which may themselves serve as the main building blocks of these devices. Here, we demonstrate that the addition of a molecule with high redox activity to a carbon nanotube (CNT) field-effect transistor provides tunable current fluctuation noise. A unique charge-trap state in the vicinity of the CNT surface due to the presence of the single molecule is the origin of the noise, which generates a prominent and unique slow discrete random telegraph signal in the device current. The power spectral density reveals the peculiar frequency limit of the fluctuation for different types of molecules depending on their redox activity and adsorption configuration. These results indicate that the detected noise will provide new opportunities to obtain electronic information for a single molecule combined with a nanotube surface, and that controllability of the noise may contribute to the expansion of noise utilization in future bio-inspired devices.
APPLIED PHYSICS LETTERS, 2017
Stochastic resonance (SR) is an intrinsic noise usage system for small-signal sensing found in va... more Stochastic resonance (SR) is an intrinsic noise usage system for small-signal sensing found in various living creatures. The noise-enhanced signal transmission and detection system, which is probabilistic but consumes low power, has not been used in modern electronics. We demonstrated SR in a summing network based on a single-walled carbon nanotube (SWNT) device that detects small subthreshold signals with very low current flow. The nonlinear current-voltage characteristics of this SWNT device, which incorporated Cr electrodes, were used as the threshold level of signal detection. The adsorption of redox-active polyoxometalate molecules on SWNTs generated additional noise, which was utilized as a self-noise source. To form a summing network SR device, a large number of SWNTs were aligned parallel to each other between the electrodes, which increased the signal detection ability. The functional capabilities of the present small-size summing network SR device, which rely on dense nanomaterials and exploit intrinsic spontaneous noise at room temperature, offer a glimpse of future bio-inspired electronic devices.
Nature communication, 2018
In contrast to AI hardware, neuromorphic hardware is based on neuroscience, wherein constructing ... more In contrast to AI hardware, neuromorphic hardware is based on neuroscience, wherein constructing both spiking neurons and their dense and complex networks is essential to obtain intelligent abilities. However, the integration density of present neuromorphic devices is much less than that of human brains. In this report, we present molecular neuromorphic devices, composed of a dynamic and extremely dense network of single-walled carbon nanotubes (SWNTs) complexed with polyoxometalate (POM). We show experimentally that the SWNT/POM network generates spontaneous spikes and noise. We propose electron-cascading models of the network consisting of heterogeneous molecular junctions that yields results in good agreement with the experimental results. Rudimentary learning ability of the network is illustrated by introducing reservoir computing, which utilises spiking dynamics and a certain degree of network complexity. These results indicate the possibility that complex functional networks can be constructed using molecular devices, and contribute to the development of neuromorphic devices.