Synaptic behaviors of a single metal–oxide–metal resistive device

The mammalian brain is far superior to today’s electronic circuits in intelligence and efficiency. Its functions are realized by the network of neurons connected via synapses. Much effort has been extended in finding satisfactory electronic neural networks that act like brains, i.e., especially the electronic version of synapse that is capable of the weight control and is independent of the external data storage. We demonstrate experimentally that a single metal–oxide–metal structure successfully stores the biological synaptic weight variations (synaptic plasticity) without any external storage node or circuit. Our device also demonstrates the reliability of plasticity experimentally with the model considering the time dependence of spikes. All these properties are embodied by the change of resistance level corresponding to the history of injected voltage-pulse signals. Moreover, we prove the capability of second-order learning of the multi-resistive device by applying it to the circuit composed of transistors. We anticipate our demonstration will invigorate the study of electronic neural networks using non-volatile multi-resistive device, which is simpler and superior compared to other storage devices.     

Gate-modulated generation–recombination current in n-type metal–oxide–semiconductor field-effect transistor

Gate-modulated generation–recombination（GMGR） current IGMGRinduced by the interface traps in an n-type metal–oxide–semiconductor field-effect transistor（n MOSFET） is investigated. The generation current is found to expand rightwards with increasing the reversed drain PN junction bias, and the recombination current is enhanced as the forward drain bias increases. The variations of IGMGRcurves are ascribed to the changes of the electron density and hole density at the interface, NSand PS, under the different drain bias voltages. Based on an analysis of the physical mechanism, the IGMGR model is set up by introducing two coefficients（m and t）. The coefficients m and t can modulate the curves widths and peak values. The simulated results under reverse mode and forward mode are obviously in agreement with the experimental results. This proves that this model can be applicable for generation current and recombination current and that the theory behind the model is reasonable. The details of the relevant mechanism are given in the paper.     

Thermophoresis/diffusion as a plausible mechanism for unipolar resistive switching in metal–oxide–metal memristors

We show that the SET operation of a unipolar memristor could be explained by thermophoresis, or the Soret effect, which is the diffusion of atoms, ions or vacancies in a steep temperature gradient. This mechanism explains the observed resistance switching via conducting channel formation and dissolution reported for TiO₂ and other metal-oxide-based unipolar resistance switches. Depending on the temperature profile in a device, dilute vacancies can preferentially diffuse radially inward toward higher temperatures caused by the Joule heating of an electronic current to essentially condense and form a conducting channel. The RESET operation occurs via radial diffusion of vacancies away from the channel when the temperature is elevated but the gradient is small.     

Multifunctional silicon-based light emitting device in standard complementary metal–oxide–semiconductor technology

A three-terminal silicon-based light emitting device is proposed and fabricated in standard 0.35 μ m complementary metal--oxide--semiconductor technology. This device is capable of versatile working modes: it can emit visible to near infra-red (NIR) light (the spectrum ranges from 500 nm to 1000 nm) in reverse bias avalanche breakdown mode with working voltage between 8.35 V--12 V and emit NIR light (the spectrum ranges from 900 nm to 1300 nm) in the forward injection mode with working voltage below 2 V. An apparent modulation effect on the light intensity from the polysilicon gate is observed in the forward injection mode. Furthermore, when the gate oxide is broken down, NIR light is emitted from the polysilicon/oxide/silicon structure. Optoelectronic characteristics of the device working in different modes are measured and compared. The mechanisms behind these different emissions are explored.     

Angular optical characteristics of light scattered by double-layer metal–oxide nanoparticles

The interaction of Tb(III)-2-{[(4-methoxy benzoyl) oxy]} methyl benzoic acid binary complex with nucleosides (adenosine, cytidine, guanosine and inosine) was investigated using UV and fluorescence methods. The reaction of Tb-complex with cytidine, guanosine and adenosine is accompanied by shift to longer wavelength in the absorption band, while there is a blue shift in the absorption band with an enhancement in the molar absorptivity upon the reaction with inosine. The fluorescence intensity of Tb(III)-2-{[(4- methoxy benzoyl) oxy]} methyl benzoic acid binary complex at λ = 545 nm (5D4 → 7F5) was decreased with the addition of the nucleoside molecule following the order: cytidine > inosine > guanosine > adenosine.     

Investigation on latch-up susceptibility induced by high-power microwave in complementary metal–oxide–semiconductor inverter

The latch-up effect induced by high-power microwave(HPM) in complementary metal–oxide–semiconductor(CMOS) inverter is investigated in simulation and theory in this paper. The physical mechanisms of excess carrier injection and HPM-induced latch-up are proposed. Analysis on upset characteristic under pulsed wave reveals increasing susceptibility under shorter-width pulsed wave which satisfies experimental data, and the dependence of upset threshold on pulse repetitive frequency(PRF) is believed to be due to the accumulation of excess carriers. Moreover, the trend that HPMinduced latch-up is more likely to happen in shallow-well device is proposed.Finally, the process of self-recovery which is ever-reported in experiment with its correlation with supply voltage and power level is elaborated, and the conclusions are consistent with reported experimental results.     

Effect of NO annealing on charge traps in oxide insulator and transition layer for 4H-SiC metal–oxide–semiconductor devices

The effect of nitric oxide(NO) annealing on charge traps in the oxide insulator and transition layer in n-type4H–Si C metal–oxide–semiconductor(MOS) devices has been investigated using the time-dependent bias stress(TDBS),capacitance–voltage(C–V),and secondary ion mass spectroscopy(SIMS).It is revealed that two main categories of charge traps,near interface oxide traps(Nniot) and oxide traps(Not),have different responses to the TDBS and C–V characteristics in NO-annealed and Ar-annealed samples.The Nniotare mainly responsible for the hysteresis occurring in the bidirectional C–V characteristics,which are very close to the semiconductor interface and can readily exchange charges with the inner semiconductor.However,Not is mainly responsible for the TDBS induced C–V shifts.Electrons tunneling into the Not are hardly released quickly when suffering TDBS,resulting in the problem of the threshold voltage stability.Compared with the Ar-annealed sample,Nniotcan be significantly suppressed by the NO annealing,but there is little improvement of Not.SIMS results demonstrate that the Nniotare distributed within the transition layer,which correlated with the existence of the excess silicon.During the NO annealing process,the excess Si atoms incorporate into nitrogen in the transition layer,allowing better relaxation of the interface strain and effectively reducing the width of the transition layer and the density of Nniot.     

Electro-optic metal–insulator–semiconductor–insulator–metal Mach-Zehnder plasmonic modulator

The performance of a CMOS-compatible electro-optic Mach-Zehnder plasmonic modulator is investigated using electromagnetic and carrier transport simulations. Each arm of the Mach-Zehnder device comprises a metal–insulator–semiconductor–insulator–metal (MISIM) structure on a buried oxide substrate. Quantum mechanical effects at the oxide/semiconductor interfaces were considered in the calculation of electron density profiles across the structure, in order to determine the refractive index distribution and its dependence on applied bias. This information was used in finite element simulations of the electromagnetic modes within the MISIM structure in order to determine the Mach-Zehnder arm lengths required to achieve destructive interference and the corresponding propagation loss incurred by the device. Both inversion and accumulation mode devices were investigated, and the layer thicknesses and height were adjusted to optimise the device performance. A device loss of <8 dB is predicted for a MISIM structure with a 25 nm thick silicon layer, for which the device length is <3 μm, and <5 dB loss is predicted for the limiting case of a 5 nm thick silicon layer in a 1.2 μm long device: in both cases, the maximum operating voltage is 7.5 V.     

Properties of metal–insulator–metal waveguide loop reflector

A new type and easy-to-fabricate metal–insulator–metal(MIM) waveguide reflector based on Sagnac loop is designed and investigated.The transfer matrix theoretical model for the transmission of electric fields in the reflector is established,and the properties of the reflector are studied and analyzed.The simulation results indicate that the reflectivity strongly depends on the coupling splitting ratio determined by the coupling length.Accordingly, different reflectivities can be realized by varying the coupling length.For an optimum coupling length of 750 nm, the 3-dB reflection bandwidth of the MIM waveguide reflector is as wide as 1.5 μm at a wavelength of 1550 nm, and the peak reflectivity and isolation are 78%and 23 dB, respectively.     

Fabrication and characterization of the normally-off N-channel lateral 4H–SiC metal–oxide–semiconductor field-effect transistors

In this paper, the normally-off N-channel lateral 4H–Si C metal–oxide–semiconductor field-effect transistors(MOSFFETs) have been fabricated and characterized. A sandwich-(nitridation–oxidation–nitridation) type process was used to grow the gate dielectric film to obtain high channel mobility. The interface properties of 4H–Si C/SiO_2 were examined by the measurement of HF I–V, G–V, and C–V over a range of frequencies. The ideal C–V curve with little hysteresis and the frequency dispersion were observed. As a result, the interface state density near the conduction band edge of 4H–Si C was reduced to 2 × 10~(11) e V~(-1)·cm~(-2), the breakdown field of the grown oxides was about 9.8 MV/cm, the median peak fieldeffect mobility is about 32.5 cm~2·V~(-1)·s~(-1), and the maximum peak field-effect mobility of 38 cm~2·V~(-1)·s~(-1) was achieved in fabricated lateral 4H–Si C MOSFFETs.     

Reactivation of damaged rare earth luminescence centers in ion-implanted metal–oxide–silicon light emitting devices

Charge trapping and quenching of electroluminescence (EL) in SiO₂ layers implanted by Ge and rare earth (RE) ions during hot electron injection were investigated. In case of the SiO₂:Ge layer the EL quenching is caused by the transformation of the luminescent defects (Ge–Si or Ge–Ge) to optically inactive centers during hot electron excitation, whereas the EL from rare earth centers is quenched due to the electron trapping by RE-centers or their surroundings, but not due to their optical deactivation. Therefore, the flash lamp post-injection annealing releasing trapped electrons reactivates RE centers and increases the operating time of metal–oxide–silicon light emitting devices (MOSLEDs). PACS 72.20.Jv; 73.40.Qv; 73.50.Gr     

Cross-shaped metal–semiconductor–metal plasmonic crystal for terahertz modulator

The transmission and tuning properties of a cross-shaped plasmonic crystal based on periodic metal–semiconductor–metal (MSM) structures have been investigated in the terahertz (THz) regime. According to the mode analysis, we find that the different resonance modes in the plasmonic crystal show the different changes when this device is actively controlled by the carrier injection of the MSM structures. The longitudinal modes disappear, while the horizontal mode moves to a higher frequency. The former leads to an intensity modulation at 0.5 THz and 1.1 THz when the groove depth h = 60 μm, and the later leads to a band blue-shift from 1.325 THz to 1.38 THz. These results will be applied to THz modulation and tunable filtering.     

Plasmonic finite-thickness metal–semiconductor–metal waveguide as ultra-compact modulator

We propose a plasmonic waveguide with semiconductor gain material for optoelectronic integrated circuits. We analyze properties of a finite-thickness metal–semiconductor–metal (F-MSM) waveguide to be utilized as an ultra-compact and fast plasmonic modulator. The InP-based semiconductor core allows electrical control of signal propagation. By pumping the core we can vary the gain level and thus the transmittance of the whole system. The study of the device was made using both analytical approaches for planar two-dimensional case as well as numerical simulations for finite-width waveguides. We analyze the eigenmodes of the F-MSM waveguide, propagation constant, confinement factor, Purcell factor, absorption coefficient, and extinction ratio of the structure. We show that using thin metal layers instead of thick ones we can obtain higher extinction ratio of the device.     

Plasmonic modulator based on gain-assisted metal–semiconductor–metal waveguide

We investigate plasmonic modulators with gain material to be implemented as ultra-compact and ultra-fast active nanodevices in photonic integrated circuits. We analyze metal–semiconductor–metal (MSM) waveguides with InGaAsP-based active material layers as ultra-compact plasmonic modulators. The modulation is performed by changing the gain of the core, that results in different transmittance through the waveguides. A MSM waveguide enables high field localization and therefore high modulation speed. Bulk semiconductor, quantum wells and quantum dots, arranged in either horizontal or vertical layout, are considered as the core of the MSM waveguide. Dependences on the waveguide core size and gain values of various active materials are studied. The designs consider also practical aspects like n- and p-doped layers and barriers in order to obtain close to reality results. The effective propagation constants in the MSM waveguides are calculated numerically. Their changes in the switching process are considered as a figure of merit. We show that a MSM waveguide with electrical current control of the gain incorporates compactness and deep modulation along with having a reasonable level of transmittance.     

Stacked lateral double-diffused metal–oxide–semiconductor field effect transistor with enhanced depletion effect by surface substrate

A stacked lateral double-diffused metal–oxide–semiconductor field-effect transistor(LDMOS) with enhanced depletion effect by surface substrate is proposed(ST-LDMOS), which is compatible with the traditional CMOS processes. The new stacked structure is characterized by double substrates and surface dielectric trenches(SDT). The drift region is separated by the P-buried layer to form two vertically parallel devices. The doping concentration of the drift region is increased benefiting from the enhanced auxiliary depletion effect of the double substrates, leading to a lower specific on-resistance(Ron,sp). Multiple electric field peaks appear at the corners of the SDT, which improves the lateral electric field distribution and the breakdown voltage(BV). Compared to a conventional LDMOS(C-LDMOS), the BV in the ST-LDMOS increases from 259 V to 459 V, an improvement of 77.22%. The Ron,sp decreases from 39.62 m?·cm~2 to 23.24 m?·cm~2 and the Baliga's figure of merit(FOM) of is 9.07 MW/cm~2.     

Self-assembly of metal–organic frameworks and graphene oxide as precursors for lithium-ion battery applications

We fabricated composites of Fe₂O₃/reduced graphene oxide as lithium-ion batteries anode material with controlled structures by employing self-assembly of metal–organic frameworks (MOFs) and polymer-functionalized graphene oxide as precursors. By electrostatic interaction, the negatively charged MOFs, Prussian Blue (PB), are assembled on poly(diallyldimethylammonium chloride) (PDDA)-functionalized graphene oxide (positive charge). Then the PB cubes become FeOOH nanosheets when treated with sodium hydroxide. Upon further annealing, the FeOOH nanosheets transform to Fe₂O₃ nanoparticles while the graphene oxide become reduced graphene oxide simultaneously. It was found that the composites have good performance as anode of lithium-ion battery. This work shows a new way for self-assembling MOFs and 2D materials.     

Resistive switching effects on the spatial distribution of phases in metal–complex oxide interfaces

In order to determine the key parameters that control the resistive switching mechanism in metal–complex oxides interfaces, we have studied the electrical properties of metal/YBa₂Cu₃O₇−δ${\mathrm{O}}_7-\delta$ (YBCO) interfaces using metals with different oxidation energy and work function (Au, Pt, Ag) deposited by sputtering on the surface of a YBCO ceramic sample. By analyzing the IV$\mathit{IV}$ characteristics of the contact interfaces and the temperature dependence of their resistance, we inferred that ion migration may generate or cancel conducting filaments, which modify the resistance near the interface, in accordance with the predictions of a recent model.     

Corrugated-enhanced second harmonic generation in metal–insulator–metal plasmonic waveguides

Nonlinear effects such as second-harmonic generation (SHG) are important for applications such as switching and wavelength conversion. In this study, the generation of second harmonic in metal–insulator–metal (MIM) plasmonic waveguides was investigated for both symmetric and asymmetric structures. Symmetric means that the metals at the top and bottom of the dielectric layer are the same and asymmetric means that the metals at the top and bottom of the dielectric layer are different. Two different structures are considered here as plasmonic waveguide for generation of second harmonic and analyzed using finite-difference time domain method. Besides the structure has grating on both sides for more coupling between photons and plasmons. The wavelength duration of grating per length unit (number of grooves) will be optimized to reach the highest second harmonic generation. To perform this optimization, the wavelength of operation λ = 458 nm is considered. It was shown that field enhancement in symmetric MIM waveguides can result in enhancement of SHG magnitude compared to the literature values and asymmetric device results in more than two orders of magnitude enhancement in SHG compared to symmetric structure. It is also shown that the electric field of second harmonic depends on the thickness of crystal (insulator). So, its thickness is optimized to achieve the highest electric field.