Wan Fazlida Hanim Abdullah

Design of single neuron on FPGA

This paper presents a digital design of neuron architecture on field-programmable gate array (FPGA). The objective of this project is to translate data from electrochemical sensor signals and process the data with neuron structure on digital hardware. The hardware realization of neural network requires investigation of many design issues relating to signal interfacing and design of a single neuron. Analysis focuses on effect of digital design decisions such as module architecture towards data accuracy and delay. The work touches on analogue to digital interfacing, data structure and digital module design that includes adder, multiplier and multiplier accumulator (MAC). A major component of the algorithm is the design of the activation function. The chosen activation function is the hyperbolic tangent which is approximated by Taylor Series expansion. The neuron is evaluated on an Altera DE2-70 FPGA. The performances are evaluated in terms of functionality, usage of resources and timing analysis. For the data structure, it was demonstrated that increasing the fractional bits will increases the precision. The neuron functionality was demonstrated on digital platform. It was found that less delay were produce by using Carry Look Ahead design compared to Ripple Carry Adder by 25% in the MAC performance.

Chemical field-effect transistor with constant-voltage constant-current drain-source readout circuit

Response of Chemical Field-Effect Transistor (CHEMFET) electrochemical sensors are taken from the output of a readout interface circuit that maintains constant drain-source voltage and current levels. We employ the readout circuit for the purpose of supervised learning training data collection. Sample solutions are prepared by keeping the main ion concentration constant while the activity of an interfering ion varied based on the fixed interference method. Results show that the voltage response demonstrates linear relationship to the ion concentration within the detection limit. However, noise in the form of abrupt and random changes in DC levels reduces correlation and increases mean square error between similarly repeated measurements. We find that referencing the voltage response to the sensor response in DIW prior to measurement greatly improves the repeatability. The process of approximating ionic concentration level is achieved up to 80% recognition by feeding the readout circuit output to a neural network post-processing stage.

TiO 2 -based extended gate FET pH-sensor: Effect of annealing temperature on its sensitivity, hysteresis and stability

In this study, titanium dioxide (TiO₂) thin films have been investigated as a sensing membrane of the extended gate field effect transistor (EGFET) for pH detection application. The sol-gel TiO₂ has been prepared and spin coated onto the indium tin oxide (ITO) coated glass as a substrate. Then the TiO₂/ITO test structures thin films were annealed for 15 min at different temperatures; 300 °C and 400 °C under ambient atmosphere. The pH sensing characterizations of TiO₂ thin films were measured by Semiconductor Parametric Device Analyzer in different pH buffer solutions of pH 4, 7, 10 and 12. The sensitivity of TiO₂ thin film annealed at 400 °C exhibited a higher sensitivity that is 51 mV/pH compared to the thin film annealed at 300 °C gave slightly lower sensitivity of 49 mV/pH. The hysteresis and drift effect for TiO₂ thin films also being investigated in this study. TiO₂ thin films annealed at 400 °C obtain better hysteresis and drift value compared to the TiO₂ thin films annealed at 300 °C.

Effect of film thickness on the memristive behavior of spin coated titanium dioxide thin films

The paper presents the memristive behavior of spin-coated titania thin films. The spin speed during the deposition was varied and their effect on the film thickness and thus, to the memristive behavior was studied. Sample deposited at 5000 rpm produced thinner film thickness which result in better switching behavior. The sample showed larger OFF to ON resistance ratio of 5. All samples were measured 3 times for each positive and negative bias. It was found that the memristive behavior was repeatable for 2nd and 3rd measurement. The crystallinity of the films was characterized using Raman spectroscopy. In our work, it was observed that the film thickness mainly affects the switching behavior of memristive device instead of the crystallinity of the films.

Effect of Post-Deposition Annealing Process on the Resistive Switching Behaviour of TiO2 Thin Films by Sol-Gel Method

The paper presents the memristive behaviour of spin-coated titania thin films on ITO substrate. The sample was annealed in air ambient at different annealing temperature and duration of 250 and 450 °C for 20 and 60 min. The effect of post-annealing process to the physical thickness and crystallinity of the films towards switching behaviour was studied. It was found that the thickness and crystallinity of the films increases as the post annealing process increases. Sample annealed at 250 °C for 20 min with thinner film showed better switching behaviour even though the sample is still in amorphous form. Thus, in our work, we believed that the crystallinity of the films does not affect the switching behaviour of the sample. The reliability of device performance was studied by repeating the measurement for three times.

Sputtered titanium dioxide thin film for Extended-Gate FET sensor application

This paper presents an investigation on sputtered titanium dioxide (TiO2) thin film for application in the Extended-Gate Field Effect Transistor (EGFET) sensor application for pH detection. The TiO2 thin film was fabricated on conductive indium-tin oxide (ITO) covered glass by using RF Sputtering method. An EGFET proof-of-concept setup was constructed using a commercial FET as the transducer and the TiO2/ITO structure was used for the extended gate. The sensor measurement was taken using semiconductor device parametric analyzer, constant-current constant-voltage biasing interfacing circuit and data logger to obtain the sensitivity and the characteristics for output voltage with respect to time. TiO2 thin film on ITO glass as the sensing membrane was compared with a bare ITO glass substrate for the extended gate. The TiO2/ITO glass exhibited the sensitivity of 42.1 mV/pH and good linearity of 0.9997 in the sensing range pH4, pH7 and pH10 which was better than the bare ITO glass sensitivity of 31.3 mV/pH and the linearity of 0.9868.

Chemical Field Effect Transistor Response with Post Processing Supervised Neural Network

This work presents the classification of potassium ion concentration in the presence of interfering ammonium ions from Chemical Field-Effect Transistor (CHEMFET) sensors involving neural network post-processing stage. Data collection for the purpose of supervised learning training data is obtained from sample solutions prepared by keeping the main ion concentration constant while the activity of the interfering ions based on the fixed interference method. The measurement setup includes a readout interface circuit that ensures constant-current constant-voltage across the drain-source for isothermal point operation. The training algorithm is back-propagation with generalized delta rule on a multilayer feed-forward network. Activation function based on the MOSFET drain current equation in the linear region is attempted in the hidden layer. Using function fitting approach, the network aims to find the potassium ion concentration despite the presence of interfering ion, without having to estimate device and chemically related parameters that would otherwise require further experiments.

pH sensing characteristics of silicon nitride thin film and silicon nitride-based ISFET sensor

The pH sensing characterizations of silicon nitride (Si₃N₄) thin film, and Si₃N₄-based ISFET sensor were discussed in this paper. The Si₃N₄ thin film was fabricated with to form silicon/silicon dioxide/silicon nitride (Si/SiO₂/Si₃N₄) structure while Si₃N₄-based ISFET was fabricated and packaged as a pH sensor. Both thin film and ISFET packaged device have low pressure chemical vapor deposited silicon nitride as sensing layer. Source measure unit (SMU), the semiconductor device analyzer model Agilent B1500A and a custom made test jig was used to investigate the electrical properties for both Si₃N₄ thin film and Si₃N₄-based ISFET sensor. The measurements were carrying out in three different buffer solutions with various pH levels of 4, 7 and 10 at room temperature. The pH sensitivity of thin film and ISFET packaged device with silicon nitride were 66.9 mV/pH and 53.6 mV/pH respectively. Both have shown good linearity in chosen pH range.

Fluorescence characteristic of ruthenium nanoparticles as a dissolved oxygen sensing material in gas and aqueos phase

The fabrication and characterization of a fiber optic dissolved oxygen sensor based on the dynamic quenching of fluorescence from a ruthenium complex in gas and aqueous phase is presented. The unclad portion of the multimode optical fiber is coated with an oxygen sensitive ruthenium complex, which is immobilized by the sol-gel route. The intensity of the fluorescence emission has been found to decrease as the oxygen content increases. The dynamic range and the temporal response of the sensor are investigated for three different environment, gas, deionized water and sea water. The performance of the device is evaluated in terms of the results obtained during actual measurements.

Memristor based delay element using current starved inverter

This paper will first review on some applications of newly found passive element, the memristor. Utilizing the beneficial characteristic of memristor where it can remember its last state, more and more improvements on today electronic designs has been proposed. However, it is crucial to observe the behavior of memristor model before applying into circuits, especially when the memristor is coupled with other devices. In this paper, LTspice memristor model is used to simulate memristor behavior and applied to the basic delay element circuit. The circuit used a tristate inverter as the delay element. It controls the current flowing to the parasitic capacitor, thus controlling the delay. The compatibility of memristor with the delay element is also in consideration to ensure the functionality of the circuits. At the end, a basic delay element using inverter and memristor is presented. This paper is divided into 4 sections, including the introduction where few examples of memristor applications are explained. It follows by next section where the inverter delay characteristic is narrated. Section 3 is about a mathematical model of memristor that been used to provide a specific memristor resistance in order to get certain delay value during simulation. Using LT spice, a memristor based delay circuit design is then proposed and the delay is observed by circuit simulation. In conclusion, the calculated R and delay value is then compared to the simulation result in order to verify circuit functionality.