Steep Increase in Substrate Current in Metal–Oxide–Semiconductor Field-Effect Transistor with Multiple-Gate Configuration

Abstract

A steep increase in substrate current has been observed in a metal–oxide–semiconductor field-effect transistor with a multiple-gate configuration. Regarding gate voltage dependence, the substrate current plotted on a logarithmic scale exhibited a sharp rise with a slope of 6 mV/decade, which is 20 times steeper than that simultaneously measured for the drain current. Since the slope is even 10 times steeper than the ideal subthreshold swing of 60 mV/decade, the upsurge has been discussed using a hypothetical model in which the impact ionization rate is increased by excitation of acoustic standing waves within the device.

Relation between sharp resonance and bias condition in metal–oxide–semiconductor field-effect transistor with multiple gate configuration

“…The author has proposed a new type of oscillator in his previous study, 6) in which a device based on metal-oxidesemiconductor field-effect transistor (MOSFET) with a multiple-gate configuration was fabricated using a typical CMOS process, and a sharp resonance with a quality factor 7,8) of Q > 1500 was preliminarily observed at 300 MHz in the frequency dependence of the impedance measured for the backgate terminal of the device. 6) In another study, 9) commonly detected from the backgate terminal, a steep increase in the substrate current I sub was observed in the gate voltage (V gs ) dependence measured for a device fabricated with a similar device configuration using the same CMOS process. The upsurge in I sub was observed in a limited narrow range of V gs between the threshold voltage V th and the saturation voltage V sat = V ds + V th , where V ds is the drain bias voltage.…”

“…Since the drain current I d simultaneously measured with I sub did not show any remarkable anomalies under a specific bias condition, it was suggested that the upsurge in I sub is caused by an increase in the impact ionization rate ¡, not in I d . 9) From the experimental results, it is reasonable to expect that a sharp resonance would also be observed only under a specific range of bias condition. 10) To confirm this expectation, the relationship between the bias condition and the sharp resonance was investigated in this study.…”

“…Similar discontinuous jumps were also reported in previous papers. 9,11) I sub increased from a background level of 2.6 © 10 ¹10 A to a maximum of 5 © 10 ¹9 A at V gs = 0.93 V, as indicated by point C. At V gs = 0.8 V as indicated by point B, log I sub increased most steeply with a slope of 131 mV/decade, which is steeper than that observed in the V gs -log I d curve (286 mV/decade). Since the slope of I sub being steeper than that of I d is unusual and inconsistent with the conventional I sub theory [14][15][16] widely used in MOSFET models for circuit simulators, 17) the impact ionization rate ¡ should be anomalously increased at point B.…”

“…Since the slope of I sub being steeper than that of I d is unusual and inconsistent with the conventional I sub theory [14][15][16] widely used in MOSFET models for circuit simulators, 17) the impact ionization rate ¡ should be anomalously increased at point B. 9,11) Figure 3 shows frequency dependences of R and X measured from 0.02 to 1.60 GHz in 0.002 GHz steps without the application of bias voltages (V g = V d = 0 V). Both R and X were in approximate agreement with an equivalent series CR model with C = 15 © 10 ¹12 F and R = 47 ³.…”

“…In the case of MOSFET devices, the number of created holes can be electrically monitored from the backgate terminal. 9) Since impedance was measured with a constant signal level (5 mV), the steep change in the impedance shown in Fig. 4(b) is caused by a steep change in electric current.…”