The method of differential-Thompson transformation (DTTR) is applied for the first time to the field of electromagnetic scattering. Complex objects and their computational areas are transformed by DTTR to a regular area. By applying the finite-difference-Time-Domain (FDTD) technique, the fieM distribution is solved conveniently. Numerical a mples show good comparison with exact results. 0 ABSTRACT A highly accurate absorbing boundary operation has been devebped to efficiently and accurately truncate the computational domain when using the finite-di~erence-timeime-domain method. Referred to as the compkmentary operators method (COM), this technique consim of averaging the solutions of two indepndent boundary operators that are complementary to each other. Because of its independence of the wave number k,, the boundary operation can significantly reduce arti$cial reflections arising from obliquely incident traveling waves as well as evanescent waves. 0 1995 John Wley & Sons, lnc.

We report on the resonant, voltage tunable emission of terahertz radiation ͑0.4 -1.0 THz͒ from a gated two-dimensional electron gas in a 60 nm InGaAs high electron mobility transistor. The emission is interpreted as resulting from a current driven plasma instability leading to oscillations in the transistor channel ͑Dyakonov-Shur instability͒.Plasma waves in a gated two-dimensional electron gas have a linear dispersion law, similar to that of sound waves. The transistor channel acts as a resonator cavity for plasma waves that can reach THz frequencies for a sufficiently short ͑nanometer-sized͒ field effect transistor. 1 As was predicted in Ref. 2, when a current flows through a field effect transistor, the steady state can become unstable against the generation of plasma waves ͑Dyakonov-Shur instability͒ leading to the emission of an electromagnetic radiation at plasma wave frequencies. The emission is predicted to have thresholdlike behavior. It is expected to appear abruptly after the device current exceeds a certain threshold value for which the increment of the plasma wave amplitude exceeds losses related to electron collisions with impurities and/or lattice vibrations.The excitation of plasma waves in a field effect transistor channel can be also used for the detection of terahertz radiation. 3 Recent reports demonstrated a resonant 4 detection in GaAs-based high electron mobility transistors ͑HEMTs͒ and in gated double quantum well heterostructures. 5 This is the first report of resonant THz emission by plasma generation. The terahertz emission ͑0.4 -1.0 THz͒ was obtained by using an InGaAs HEMT with a 60-nm-long gate. We show that the results can be interpreted assuming that the emission is caused by the current driven plasma instability leading to terahertz oscillations in the channel through Dyakonov-Shur instability.Lattice-matched InGaAs/AlInAs HEMTs grown by molecular beam epitaxy on an InP substrate were used in this study. The active layers consisted of a 200 nm In 0.52 Al 0.48 As buffer, a 20 nm In 0.53 Ga 0.47 As channel, a 5-nm-thick undoped In 0.52 Al 0.48 As spacer, a silicon planar doping layer of 5ϫ10 12 cm Ϫ2 , a 12-nm-thick In 0.52 Al 0.48 As barrier layer, and, finally, a 10-nm-silicon-doped In 0.53 Ga 0.47 As cap layer. Details of the technological process are given elsewhere. 6 The gate length was 60 nm, and the drain-source separation was 1.3 m. An InP-based HEMT was chosen for its high InGaAs channel mobility and high sheet carrier density.Output and transfer characteristics are shown in Fig. 1. The low field, linear output region is marked by the dotted line. The deviation of the I d (U sd ) curve from linear behavior indicates the beginning of the saturation region. The arrow indicates the emission threshold voltage, U sd ϳ200 mV at I d ϳ4.5 mA. The horizontal dashed line shows the level of the current saturation (I d ϳ4.8 mA). The I d (U sd ) characteristic shows an unstable behavior for U sd higher than 300 mV. This well-known phenomenon is related to a self-excitation a͒ Also at

As Moore's law reaches its end, traditional computing technology based on the Von Neumann architecture is facing fundamental limits. Among them is poor energy efficiency. This situation motivates the investigation of different processing information paradigms, such as the use of spiking neural networks (SNNs), which also introduce cognitive characteristics. As applications at very high scale are addressed, the energy dissipation needs to be minimized. This effort starts from the neuron cell. In this context, this paper presents the design of an original artificial neuron, in standard 65 nm CMOS technology with optimized energy efficiency. The neuron circuit response is designed as an approximation of the Morris-Lecar theoretical model. In order to implement the non-linear gating variables, which control the ionic channel currents, transistors operating in deep subthreshold are employed. Two different circuit variants describing the neuron model equations have been developed. The first one features spike characteristics, which correlate well with a biological neuron model. The second one is a simplification of the first, designed to exhibit higher spiking frequencies, targeting large scale bio-inspired information processing applications. The most important feature of the fabricated circuits is the energy efficiency of a few femtojoules per spike, which improves prior state-of-the-art by two to three orders of magnitude. This performance is achieved by minimizing two key parameters: the supply voltage and the related membrane capacitance. Meanwhile, the obtained standby power at a resting output does not exceed tens of picowatts. The two variants were sized to 200 and 35 μm2 with the latter reaching a spiking output frequency of 26 kHz. This performance level could address various contexts, such as highly integrated neuro-processors for robotics, neuroscience or medical applications.