Voltage Power Spectral Density Research Articles

Electrical, Optical and Thermal Properties of Ge-Si-Sn-O Thin Films.

This work evaluates the electrical, optical and thermal properties of Sn-doped GexSi1-xOy thin films for use as microbolometer sensing materials. The films were prepared using a combination of a radio frequency (RF) magnetron and direct current (DC) sputtering using a Kurt J Leskar Proline PVD-75 series sputtering machine. Thin films were deposited in an O2+Ar environment at a chamber pressure of 4 mTorr. The thicknesses of the thin films were varied between 300 nm-1.2 µm by varying the deposition time. The morphology and microstructure of thin films were investigated by atomic force microscope (AFM) imaging and X-ray diffraction (XRD), while the atomic composition was determined using the energy dispersive spectroscopy (EDS) function of a scanning electron microscope. The thin film with an atomic composition of Ge0.45Si0.05Sn0.15O0.35 was found to be amorphous. We used the Arrhenius relationship to determine the activation energy as well as temperature coefficient of resistance of the thin films, which were found to be 0.2529 eV and -3.26%/K, respectively. The noise voltage power spectral density (PSD) of the film was analyzed using a Primarius-9812DX noise analyzer using frequencies ranging from 2 Hz to 10 kHz. The noise voltage PSD of the film was found to be 1.76 × 10-11 V2/Hz and 2.78 × 10-14 V2/Hz at 2 Hz and 1KHz frequencies, respectively. The optical constants were determined using the ellipsometry reflection data of samples using an RC2 and infrared (IR) VASE Mark-II ellipsometer from J A Woollam. Absorption, transmission and reflection data for a wavelength range of 900 nm-5000 nm were also determined. We also determined the optical constant values such as the real and imaginary parts of refractive index (n and k, respectively) and real and imaginary part of permittivity (ε1 and ε2, respectively) for wavelength ranges between 193 nm to 35 µm. An optical band gap of 1.03 eV was determined from absorption data and using Tauc's equation. In addition, the thermal conductivity of the film was analyzed using a Linseis thin film analyzer employing the 3ω method. The thermal conductivity of a 780 nm thick film was found to be 0.38 Wm-1K-1 at 300 K. From the data, the Ge-Si-Sn-O alloy was found to be a promising material for use as a sensing material for microbolometers.

Generation–recombination noise in uniformly doped semiconductors with contacts of finite surface recombination velocities

The formula for the voltage noise spectrum due to generation–recombination (G–R) rate fluctuations is derived for uniformly doped majority-carrier n-type semiconductor samples with contacts of finite surface recombination velocities. In the derivation of the formula, a new boundary condition with finite surface recombination velocities considered is used instead of the ohmic boundary condition. The G–R voltage noise spectra are calculated to become saturated and equal in the high electric field region irrespective of surface recombination velocities. However, the behaviors of slow surface recombination velocities in the low electric field region differ from those of fast ones. The profiles of the electron density fluctuations of slow ones differ significantly from those of fast ones in the low-field region. In the high field region, a steep rise in the magnitude of the electron density fluctuation occurs near the voltage contact toward which electrons drift. The various location dependencies of the contribution of noise sources to the voltage power spectral density are also derived for low and high field regions as well as low and high surface recombination velocities, and analytical expressions are developed for the voltage power spectral densities and electron density fluctuations.

Solar Orbiter/RPW antenna calibration in the radio domain and its application to type III burst observations

Context.In order to allow for a comparison with the measurements from other antenna systems, the voltage power spectral density measured by the Radio and Plasma waves receiver (RPW) on board Solar Orbiter needs to be converted into physical quantities that depend on the intrinsic properties of the radiation itself (e.g., the brightness of the source).Aims.The main goal of this study is to perform a calibration of the RPW dipole antenna system that allows for the conversion of the voltage power spectral density measured at the receiver’s input into the incoming flux density.Methods.We used space observations from the Thermal Noise Receiver (TNR) and the High Frequency Receiver (HFR) to perform the calibration of the RPW dipole antenna system. Observations of type III bursts by the Wind spacecraft are used to obtain a reference radio flux density for cross-calibrating the RPW dipole antennas. The analysis of a large sample of HFR observations (over about ten months), carried out jointly with an analysis of TNR-HFR data and prior to the antennas’ deployment, allowed us to estimate the reference system noise of the TNR-HFR receivers.Results.We obtained the effective length,leff, of the RPW dipoles and the reference system noise of TNR-HFR in space, where the antennas and pre-amplifiers are embedded in the solar wind plasma. The obtainedleffvalues are in agreement with the simulation and measurements performed on the ground. By investigating the radio flux intensities of 35 type III bursts simultaneously observed by Wind and Solar Orbiter, we found that while the scaling of the decay time as a function of the frequency is the same for the Waves and RPW instruments, their median values are higher for the former. This provides the first observational evidence that Type III radio waves still undergo density scattering, even when they propagate from the source, in a medium with a plasma frequency that is well below their own emission frequency.

Suppressing 1/f Noise in Graphene Transistors By Area Scaling

Flicker noise or 1/f noise refers to processes in which the power spectral density (PSD) is inversely proportional to frequency and is typically the dominant noise source in transistors at low frequency. We report our work on measurements of unprecedentedly low 1/f noise in graphene field effect transistors, which we attribute to large device area. Large area graphene grown by chemical vapor deposition (CVD) has potential for a variety of applications including biomolecular sensors, bolometric photodetectors and ion sensitive field effect transistors (ISFET) [1]. For all these applications, low-frequency 1/f noise is found to be the dominant factor that determines sensor resolution limits. Hence, the absolute value of 1/f noise PSD serves as a crucial performance metric for graphene sensor applications.Previous studies of 1/f noise in graphene devices has been performed using both CVD grown graphene and exfoliated graphene. Balandin et al [2] has studied 1/f noise in exfoliated graphene and how it varies with substrate and charge carrier concentration. Karnatak et al [3] has shown that contact resistance can play a dominant role in 1/f noise. To date, there has been no experimental study of 1/f as the graphene channel is scaled up to mm lengths.We report here our work on 1/f noise measurement of graphene field effect transistors with varying channel and contact geometries. In our experiments, CVD grown graphene on copper was transferred onto fused silica coated with 100 nm of parylene using a standard wet transfer process. Parylene was used as an interface between the graphene and fused silica, which has been shown by Fakih et al [1] to reduce both drift and hysteresis in graphene ISFETs. The transferred graphene was processed with two photolithography steps to fabricate devices with areas ranging from 12µm2 to 36mm2. Electrical contact to the graphene was made directly with Au. The 1/f noise was measured by first applying a dc bias using a low noise lithium-ion battery, and the generated voltage fluctuations were then measured using a low noise voltage pre-amplifier and a 24-bit digitizer.The voltage PSD of 1/f noise in graphene, will typically have the form of SV=Vo 2K/f. Where Vo is the dc voltage applied across the device, f is the frequency, and K is the unitless noise constant. K is experimentally approximated to be (1/N)∑nSVnfn/Vo 2, where SVn is the voltage PSD measured at n different frequencies fn. The work from Balandin et al [2] demonstrates typical K values of 10-8, and work from Karnatak et al [3] shows values as low as 10-9. Our devices have unprecedented low measured K values of 10-14, which we attribute to the large device area of 36 mm2. Our results experimentally demonstrate that the noise parameter can be decreased by orders of magnitude by working with large area devices. Our work suggests that for graphene based sensor applications, large device area is favourable for improved sensor resolution.

Photoresponse Performance Evaluation of ZnO UV Photodetector Based on Noise Analysis

The ZnO film photosensitive chips prepared by sputtering method were post-treated by air-anneal at different temperatures and Ar or air-plasma treatments for different time. Then, photosensitive chips were packaged into UV photodetectors with a standard TO-5 package to meet the sensitivity, selectivity, and stability standards. Noise analysis was used as an efficient tool to evaluate the quality and reliability of ZnO UV photodetectors, and the overall performance evaluation based on 1/f noise physical model is made to build the relationship between photoresponse and surface defect state density. The 1/f noise comes from defect-related carriers’ number fluctuation, which are trap and detrap processes of free electrons from oxygen vacancies at the surface of ZnO. This can be analyzed semi-quantitatively by the surface trap energy density obtained by voltage power spectral density. For the photosensitive chip treated by 300° air-anneal, we can obtain the largest responsivity 6.7 A/W, the smallest noise equivalent power $2.7 \times 10^{-9}$ W, and the largest specific detectivity $9\ast 10^{10}$ Jones at 300 nm under the bias of 5 V. The excellent photoresponse performance is also indicated by large linear dynamic range and fast response and recovery speeds.

Noise Properties of YBCO Nanostructures

Voltage noise measurements on close to optimally doped YBa <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> Cu <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">7-δ</sub> nanostructures have been performed. The measured resistance noise at temperature T = 96 K (above critical temperature T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</sub> = 85 K) shows a quadratic dependence on the bias current, e.g., the voltage power spectral density S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</sub> αV <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> . Moreover, the normalized voltage noise S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</sub> /V <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> is inversely proportional to the device volume. This is a clear indication that the noise is the result of an ensemble of independent resistive fluctuators, evenly distributed within the sample volume. For our structures, we obtain a product S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">V</sub> /V <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> × Vol. = const. ≈6 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-33</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> /Hz resulting in a Hooge's parameter 3.4 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-4</sup> , which is among the lowest reported in literature. At lower temperature, T = 2 K (well below T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</sub> ) the total voltage fluctuations are given by the combined effect of critical current fluctuations and resistance fluctuations. For the critical current noise, we obtain a product S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">I</sub> /I <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</sub> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> × Vol. = const. ≈6 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-32</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> /Hz. The larger value of the relative critical current noise is most probably due to the fact that the critical current is determined by edge effects whereas the resistance is given by the total volume of the device.

Reduction of Low Frequency Noise Impact to Terahertz Detectors in CMOS

Variation of low frequency noise including 1/f noise and Lorentzian spectra is the dominant source of variations for noise equivalent power (NEP) of THz detectors using a forward biased diode connected NMOS transistor. Elimination of the impact of low frequency noise variation to NEP with 1-V amplitude AC biasing at 44MHz is demonstrated in 830-GHz detectors using a diode connected PMOS transistor. It should be possible to apply this technique to eliminate the impact of low frequency noise variations in THz NMOS and Schottky diode detectors. Lastly, the low frequency noise voltage power spectral density of AC biased diode connected PMOS transistors is comparable to that of commercial GaAs detectors.

Noise Reduction of Amorphous SixGeyO1–x–yThin Films for Uncooled Microbolometers by Si3N4Passivation and Annealing in Vacuum

This paper presents the fabrication and reduction of noise voltage power spectral density (PSD) of Si x Ge y O1– x – y uncooled infrared microbolometers with four different compositions. The noise reduction was achieved by passivating Si x Ge y O1– x – y with Si3N4 layers and by annealing the devices in vacuum at 200 °C, 250 °C, or 300 °C with different time intervals from 1 to 5 h. The voltage noise PSD was measured with a bias current between 0.07 and $0.6~\mathrm {\mu }\text{A}$ before and after annealing. The lowest measured noise voltage PSD before annealing was on devices with Si0.053Ge0.875O0.072 and Si0.041Ge0.902O0.057 films. They were $7.42\times 10^{-15}$ and $2.07\times 10^{-14}~\text{V}^{2}$ /Hz at 23 Hz, respectively. The corresponding 1/ ${f}$ -noise coefficients, $\textit{K}_{f}$ , of the devices were $3.65\times 10^{-14}$ and $3.01\times 10^{-14}$ , respectively. The voltage noise PSD was reduced as the annealing time and temperature were increased. The lowest measured noise was $1.96\times 10^{-14}~\text{V}^{2}$ /Hz at the corner frequency, 12 Hz, on a device with Si0.034Ge0.899O0.067 film annealed at 300 °C for 4 h, before annealing the noise was $4.09\times 10^{-13}~\text{V}^{2}$ /Hz at 12 Hz. This corresponds to a factor of 24 reduction of noise. The testing results demonstrate that annealing at higher temperature 300 °C reduced the low-frequency voltage noise PSD more than that of 200 °C and 250 °C temperatures.

Electrical transport and low-frequency noise in chemical vapor deposited single-layer MoS2 devices

We have studied temperature-dependent (77–300 K) electrical characteristics and low-frequency noise (LFN) in chemical vapor deposited (CVD) single-layer molybdenum disulfide (MoS2) based back-gated field-effect transistors (FETs). Electrical characterization and LFN measurements were conducted on MoS2 FETs with Al2O3 top-surface passivation. We also studied the effect of top-surface passivation etching on the electrical characteristics of the device. Significant decrease in channel current and transconductance was observed in these devices after the Al2O3 passivation etching. For passivated devices, the two-terminal resistance variation with temperature showed a good fit to the activation energy model, whereas for the etched devices the trend indicated a hopping transport mechanism. A significant increase in the normalized drain current noise power spectral density (PSD) was observed after the etching of the top passivation layer. The observed channel current noise was explained using a standard unified model incorporating carrier number fluctuation and correlated surface mobility fluctuation mechanisms. Detailed analysis of the gate-referred noise voltage PSD indicated the presence of different trapping states in passivated devices when compared to the etched devices. Etched devices showed weak temperature dependence of the channel current noise, whereas passivated devices exhibited near-linear temperature dependence.

New numerical low frequency noise model for front and buried oxide trap density characterization in FDSOI MOSFETs

In this paper, we present a new numerical model of the inversion charge power spectral density in FDSOI MOSFET devices with ultra thin body. Numerical simulation results are compared to those of the classical formulation and to experimental data. A good agreement of measurements is obtained with the proposed model. The results show that the noise behavior in FDSOI MOSFETs is strongly related to the front and buried oxides defects, even if the channel is located at the front interface. In other words, the classical formulation of the flat-band voltage power spectral density (PSD) overestimate the front oxide trap density and no more holds true in SOI MOSFETs LFN characterization.

Spectroscopy of electronic thermal noise as a direct probe of absolute thermoelectric coefficients

The utilization of thermal fluctuations or Johnson/Nyquist noise as a generalized spectroscopic technique to experimentally measure transport properties is applied to Pt and W metal films. Through cross-correlation and autocorrelation functions obtained from voltage power spectral density measurements, multiple transport coefficients are obtained through the Green–Kubo formalism. Supported rigorously by the underlying fluctuation-dissipation theorem and Green–Kubo transport theory, this novel experimental technique provides a direct measurement of absolute Seebeck and Peltier coefficients in addition to the electrical resistivity, electronic contribution to thermal conductivity, and Lorenz number. This work reports the validation results of the experiment accomplished through the use of materials with thermoelectric properties widely accepted by the thermoelectric community, Pt and W. Further validation of the data was accomplished by comparing the resistivity results to standard collinear four-probe resistivity measurements. Spectroscopic results for resistivity at 300 K resulted in 5.3% and 2.5% agreement with four-probe resistivity measurements for Pt and W, respectively. The Seebeck coefficient measurements at 300 K showed agreement with published values within 3.8% and 7.5% for Pt and W, respectively. The electronic thermal conductivity measured 66% and 75% of the total thermal conductivity for Pt and W, respectively, at 300 K.

Noise reduction of a-Si 1 − xGe xO y microbolometers by forming gas passivation

We report the reduction of noise voltage power spectral density (PSD) of amorphous Si 1 − x Ge x O y microbolometers by forming gas passivation. The microbolometers fabricated from Si 1 − x Ge x O y were passivated in a rapid thermal annealing chamber in the presence of forming gas at 250 °C with different intervals of time starting from 0.5 h to 8 h. The noise voltage PSD was measured at different bias currents before passivation and after different interval of passivation time. It was found that the noise voltage PSD of the bolometers decreased as the passivation time increased. The 1/ f-noise coefficient ( K f ) was decreased from 7.54 × 10 − 7 to 2.21 × 10 − 10 after 8 h of forming gas passivation performed at 250 °C.

An improved physics-based 1/ f noise model for deep sub-micron MOSFETs

1/ f noise has been investigated on the drain to source voltage of deep sub-micron p-channel MOSFETs, with 5 μm channel width and 0.32–1.2 μm varying lengths. The MOSFETs were fabricated using 0.18 μm medium doped drain technology. The measurements were performed in strong inversion and linear operation region at room temperature. Noise parameters for BSIM3 simulation were extracted from the 1/ f noise measurements. The oxide trap densities calculated from these noise parameters, however, showed a significant difference from transistor to transistor on the same die. Moreover, the drain voltage power spectral density exhibited a stronger dependence on channel length than the 1/ L dependence expected from BSIM3 model. A modified BSIM3 model has been developed, based on threshold voltage variation along the channel. The modified model provides a good quantitative agreement with the noise power spectral density as a function of gate voltage. In addition, the modified flicker noise model more accurately predicts the oxide trap density. Carrier scattering coefficient was also obtained from the noise fitting parameters.

Channel length scaling of 1/ f noise in 0.18 μm technology MDD n-MOSFETs

Low-frequency (LF) noise (1 Hz–100 KHz) measurements have been performed on 12 n-channel MOSFETs with 5 μm width and 0.23–10 μm varying lengths. The MOSFETs were fabricated using a 0.18 μm, medium-doped-drain (MDD) technology. Interface state density and oxide trap density were extracted from the sub-threshold slope characteristics of I– V GS measurements, and from 1/ f noise measurements, respectively. The McWhorter theory based on long-channel MOSFETs was found to break down for MOSFETs with L<0.6 μm. The drain voltage power spectral density varied inversely with channel length, S V d ∝L −1 for L≥1 μm as expected from the McWhorter theory, but showed a stronger dependence on channel length for smaller devices. Following the Tsai and Ma formulation, a modified McWhorter model has been used, based on threshold voltage variations across the channel. The new model more accurately predicts the Si–SiO 2 interface trap density and the channel length dependence of 1/ f noise magnitude in sub-micron MDD MOSFETs. Three HSPICE 1/ f noise models were evaluated. Although the highest level model provided a good fit to the data for L≥1 μm, for shorter channel devices all three levels underestimate the noise magnitude.

Active ladder filter noise analysis by means of continuants

The method of calculation by means of continuants of active ladder filters noise voltage power spectral density is proposed in this brief. The connection between noise properties and sensitivity of the filter transfer function is established. >