An Experimental Study on the Performance of Two Temperature Sensors Based on 4H SiC Diodes Procedia Engineering 168 ( 2016 ) 729 – 732 Available online at www sciencedirect com 1877 7058 © 2016 The Au[.]
Available online at www.sciencedirect.com ScienceDirect Procedia Engineering 168 (2016) 729 – 732 30th Eurosensors Conference, EUROSENSORS 2016 An experimental study on the performance of two temperature sensors based on 4H-SiC diodes S Raoa,*, G Pangalloa, F.G Della Cortea a Università degli Studi “Mediterranea”, Dipartimento di Ingegneria dell’Informazione, delle Infrastrutture e dell’Energia Sostenibile (DIIES), Via Graziella Feo di Vito, 89122 Reggio Calabria, Italy Abstract The performance of two temperature sensors based on 4H-SiC diodes are investigated Both devices show a good linear dependence on temperature of the difference between the forward bias voltages measured on two identical diodes (Schottky or p-i-n) yet biased at different constant currents The Schottky diodes-based sensor shows a high sensitivity (S=5.11mV/K) whereas the p-i-n structure has a highly linear output proportional to the absolute temperature © 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license © 2016 The Authors Published by Elsevier Ltd (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference Keywords: Terms—Schottky diodes, p-i-n diodes, power semiconductor devices, sensors silicon carbide, temperature sensors, wide band gap semiconductors; Introduction In the last decade, Schottky, p-i-n and p-n diodes have been explored for the realization of temperature sensors [14] The favorable chemical and physical properties, e.g high thermal conductivity (3-5 W/cm°C) and high critical electric field for breakdown setup (Ec = 2-5 MV/cm), make 4H-silicon carbide (SiC) a suitable material for high power and high temperature applications [5,6] Recently, diode-temperature sensors based on 4H-SiC were designed and fabricated to work in hostile and harsh environments [7-11] However, the accuracy of such single-diode sensors is affected by the non-linear behavior with temperature of the saturation current, IS, in particular when the bias current intensity, ID, is comparable with IS [1,12] * Corresponding author Tel.: +39 09651693274; E-mail address: sandro.rao@unirc.it 1877-7058 © 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference doi:10.1016/j.proeng.2016.11.262 730 S Rao et al / Procedia Engineering 168 (2016) 729 – 732 In this work, we present two proportional-to-absolute-temperature (PTAT) sensors, overcoming therefore the nonlinear effects of IS, based on 4H-SiC Schottky and p-i-n diodes, respectively The sensors have been realized using two identical diodes integrated on the same chip and biased with different currents kept constant over the considered temperature range Sensors structure In our setup, two p-i-n diodes, D1 and D2, with almost identical ID-VD characteristics, were driven by two external and independent current sources, providing constant ID1 and ID2 currents [Fig 1(a)] over the whole temperature range The two currents are described by the analytical expression: I D1, = I S 1, e · § qV D , − R S , I D , ¨ −1 ¸ ¸ ¨ η , kT ¹ © (1) where IS1,2, RS1,2, VD1,2 and Ș1,2 are the saturation current, the series resistance, the diode voltage drop and the ideality factor for D1 and D2, respectively If the two diodes show the same ideality factor (Ș1=Ș2=Ș), the difference between the voltage drops across the two diodes (VD2 - VD1) can be written from (1) as: Δ V D = V D − V D1 = kT η ln (r ) + R S (I D − I D ) q (2) where r=ID2/ID1 is the bias current ratio Eq (2) indicates that, for a fixed ID2/ID1 ratio, the sensor output, ǻVD, is linearly proportional to T if Ș is highly stable with temperature and the contribution of RSÂ(ID2-ID1) can be considered negligible Fig Electrical circuit of the PTAT sensor (a) Schematic cross section of the 4H-SiC integrated Schottky (b) and p-i-n diode (c) The 4H-SiC diodes were fabricated by CNR-Institute for Microelectronics and Microsystems, unit of Bologna (I) Standard photolithography and wet chemical etching were used to pattern, 150ì150 àm2 Ti/Al Schottky contacts, spaced each other by about 155 µm [Fig 1(b)] The circular p+-type anode regions were obtained simultaneously by ion implantation of Al through a SiO2 mask designed to pattern the vertical p-i-n diodes with p+-type area of 7.54×10−4 cm2 The distance between the two p-i-n diodes is ~190 µm [Fig 1(c)] The microchip contains several diodes with the common cathode consisting of a commercial n+-4H-SiC substrate [13] The chip was packaged and the Ti/Al metal contacts were bonded using thin Al wires, 50 µm in diameter, to a custom printed circuit board (PCB) to allow an electrical connection to the measurement set-up In Fig (b) and (c), the schematic cross sections of the PTAT sensors, each consisting of two geometrically similar integrated diodes biased by different currents kept constant over the considered temperature range, T=293-573 K, are shown S Rao et al / Procedia Engineering 168 (2016) 729 – 732 Experimental results and discussion The devices have been tested in a thermostatic oven (Galli G210F030P) setting the reference temperature through its internal PID digital microcontroller Two calibrated and certified resistance temperature detectors (RTDs) based on platinum wire (PT100), with an accuracy of ±0.3 K, were placed in contact with the PCB, through a thermal conductive paint, very close to the device under test in order to monitor, during measurements, the temperature set points A comparison of the measured ǻVD, in a range from (down to) T=293 K up to (from) 573 K for several values of ID1 and current ratio, r, is reported in Fig together with the best-linear fitting of the experimental data The plots show both the high linear dependence of ǻVD on T and the corresponding sensitivities calculated from the slope of the ǻVD-T characteristics To evaluate numerically how well the linear model truly fits the experimental data, the coefficient of determination (R2) [14] was calculated For Schottky diodes-based PTAT sensors all of the characteristics show a good degree of linearity (R2 > 0.998) for ID1 ranging from 0.5 mA up to 1.3 mA and for different current ratios, r As reported in Fig 2(a), for ID1=1.3 mA and r=4.9, the calculated sensitivity is S=2.85 mV/K, increasing for higher r and lower bias currents ID1 For r=18.4 and ID1=0.5 mA we get a high value of S=5.11 mV/K Here the sensor shows its maximum linearity, R2=0.9995, corresponding to a rmse, with respect to the best-linear model, of ±0.6 K Whereas, the p-i-n diodes-based PTAT sensor, Fig.2 (b), shows a good degree of linearity (R2 > 0.999) for ID1 ranging from 180 ȝA up to 3.3 mA, with 1.1