A. Faigon

Electrically Erasable Metal–Oxide–Semiconductor Dosimeters

Metal-oxide-semiconductor (MOS) dosimetry including the reset of the sensor device for its reuse (reutilization) is described. The method consists in restoring the shifted threshold voltage after irradiation to a predefined value by the injection of a Fowler-Nordheim tunnel current. The amount of interface states per unit area is initially saturated in order to ensure repeatability. The method was tested on 70 nm pMOSFETs exposed to a 60Co source. After successive irradiations and erasures amounting several tens of kGy[SiO2], the devices exhibit a dispersion smaller than 2% in the responses.

Trapping effects in thin oxynitride layers in metal‐insulator‐semiconductor devices

The trapping characteristics of thin oxynitride films obtained by the oxidation of a thermally nitrided silicon surface were studied under both tunneling and hot electron injection. Comparison with standard oxide layers yields the following differences: No net positive charge generation is observed in the investigated layers, and the rate of surface states generation is about one order of magnitude smaller. The electron trap density is estimated to be ∼6×1017 cm−3 with a capture cross section of ∼10−17 cm2.

Experimental evidence and modeling of two types of electron traps in Al2O3 for nonvolatile memory applications

Al2O3-based dielectrics are currently considered as promising materials to use in nonvolatile memories. The electron trap density in this material is much higher than in conventional SiO2, being their characteristics critical for the application. Conventional capacitance-voltage (C-V) techniques were used to study the main effects of the electron traps on the electrical characteristics of MOS capacitors with atomic layer deposited Al2O3 as insulating layer. More detailed information about the trapping kinetics was obtained through the study of the constant capacitance voltage transient. Two different types of traps were found. One is responsible for the instabilities observed in C-V measurements, the other has characteristic trapping times three orders longer. A physical model is presented to explain the observed trapping kinetics exhibiting good agreement between experiments and simulations. The energy levels of the studied traps were determined at 2.2 and 2.6 eV below the Al2O3 conduction band, with den...

Zero Temperature Coefficient Bias in MOS Devices. Dependence on Interface Traps Density, Application to MOS Dosimetry

In this paper the influence of temperature fluctuations on the response of thick gate oxide metal oxide semiconductor dosimeters is reviewed and the zero temperature coefficient (ZTC) method is evaluated for error compensation. The response of the ZTC current to irradiation is studied showing that the error compensation impoverishes with absorbed dose. Finally, an explanation and analytic expression for the shifts in the ZTC current with irradiation based on the interface traps creation is proposed and verified with experimental data.

Extension of the Measurement Range of MOS Dosimeters Using Radiation Induced Charge Neutralization

This work proposes a new biasing technique to extend the dose measurement range of MOS dosimeters. The technique consists on alternating stages of positive oxide charge buildup with stages of radiation induced charge neutralization, maintaining an uniform sensitivity along the whole measurement. The technique was applied with 70 nm MOS dosimeters, extending the dose measurement range from less than a kilogray to more than 675 kGy without showing wear effects. An initial saturation of interface traps creation ensures repeatability in the responses.

Reducing Measurement Uncertainties Using Bias Cycled Measurement in MOS Dosimetry at Different Temperatures

Temperature dependence of MOS dosimeters response used under the Bias Controlled Cycled Measurement technique is investigated. The use of the biasing technique allows the compensation of temperature-induced changes in the response of the sensors, and reduces at least ten times the dose measurement error caused by undesired threshold voltage shifts.

Field Oxide n-channel MOS Dosimeters Fabricated in CMOS Processes

This paper presents a new technique to build MOS dosimeters using unmodified standard CMOS processes. The devices are n-channel MOS transistors built with the regular Field Oxide as a thick radiation-sensitive gate. The devices were fabricated in two different commercial 0.6 μm CMOS processes, gate oxide thicknesses of ~600 nm and ~400 nm. Responsivities up to 4.4 mV/rad with positive bias, and 1.7 mV/rad with zero gate bias were obtained in the thicker oxides. The effect of charge trapped in the oxide and interface states on the shift in the threshold voltage are analyzed.

Floating Gate PMOS Dosimeters Under Bias Controlled Cycled Measurement

Floating Gate Metal Oxide Semiconductor (FG-MOS) structures, designed and fabricated in a CMOS process, were irradiated under the Bias Controlled Cycled Measurement (BCCM) novel technique conditions. Results presented in this work show the possibility of using such structures with the BCCM technique to measure ionizing radiation absorbed dose over a range of several kGy without significant loss of sensitivity. Transients observed after the bias switch are related to the evolution of the charge distribution between the floating gate and oxide traps near the semiconductor.

Modeling of the I–V characteristics of high-field stressed MOS structures using a Fowler–Nordheim-type tunneling expression

Abstract The gate current–voltage characteristic of a high-field stressed metal-oxide-semiconductor structure with trapped charge within the insulator barrier is consistent with a Fowler–Nordheim-type tunneling expression. Instead of considering a correction for the cathode electric field as usual, we use an effective local electric field that takes into account the distortion of the oxide conduction band profile caused by the trapped charge. An energy level at the injecting interface, introduced as an optimization parameter of the model, controls the tunneling distance used for calculating the effective field. Trap generation in the oxide is induced by high-field constant current stress and subsequent electron trapping at different injection levels is monitored by measuring the associated flat band voltage shift. The model applies for positive gate injection regardless the stress polarity and the involved parameters are obtained by fitting the experimental data without invoking any particular theoretical model for the trapping dynamics. In addition, it is shown how the presented model accounts for consistently both the current–voltage and voltage–current characteristics as a function of the injected charge through the oxide.

Analysis of experimental current oscillations in MOS structures using a semi-empirical tunneling model

Abstract A semi-empirical model has recently been proposed to simulate the tunneling characteristics of very thin or thick thermal oxides used as a gate insulator in metal-oxide-semiconductor structures. Its salient feature is a constant wave number in the insulator gap combined with an energy dependent attenuation factor for the direct tunneling probability. The tunneling analysis was based on the WKB approximation. In order to incorporate the oscillatory behavior of the tunneling current observed experimentally in the Fowler-Nordheim injection regime, which was disregarded before, here we use the transfer matrix method for calculating the transmission probability. The parameter set involved in the model is extracted from the experimental current vs voltage curves and from its logarithmic derivatives. The obtained values are in excellent agreement with those generally accepted for the MOS system. It is shown in this work that the resonant behavior is totally consistent with the principles on which the former model was constructed.