Study of TiW/Au Thin Films Metallization Stack for High Temperature and Harsh Environment Devices on 6H Silicon Carbide

This is the authors' abstract. We don't add key points for this paper.

In this work, metallizations suitable for SiC devices and stable at high temperatures furthermore working in harsh environment are investigated. The Au/TiW/Ni2Si multilayer has been studied concerning its physical and chemical stability, and its electrical behaviour at several temperatures up to 600°C in air. Depth profiling by XPS and x-ray diffraction were used to study the thermal stability of TiW films as diffusion barriers between a gold overlayer and the Ni2Si/SiC ohmic contact. Transmission line model (TLM) was adopted to measure the contact resistivity and to test the electrical stability upon thermal stresses. Measurements were performed on samples before and after thermal stresses at temperature ranging from 200°C to 600°C in vacuum and O2 environment. We demonstrate that the metallization scheme is reliable up to 450°C for 100 hrs. Introduction In the recent years the request of modern industry for devices and sensors able to operate at high power and high temperatures (T>200°C) is constantly increasing[1]. This target is beyond the physical limits of conventional semiconductors like, silicon and gallium arsenide. Furthermore, the most existing electronic devices are limited to operate at temperatures below 200°C, primarily due to the degradation of intrinsic properties of the associated device materials including metallization and packaging. Silicon carbide is presently the most favourable semiconductor for high power and high temperature applications because of its excellent physical properties[2,3] and its commercial availability in relative large diameter wafers (up to 75 mm) of high electronic quality. However, an unsolved problem is a high temperature stable and reliable metallization scheme for SiC devices or sensors working in harsh environment at temperatures higher than 400°C [3]. Many solutions have been proposed, however, their reliability is not always fully investigated while the role of some metals and the failure mechanism is not always clearly explained. Several metals have been investigated for ohmic contacts on n-type SiC, namely, Ni, Al/Ni/Al, Si /Ni, NiCr, W/Cr/Ni, Ta, TaC, Re, W, WSix, Ti, TiSi2, TiC, Cr, CoSi2, and MoSi2[4,5,6]. As a result Ni based ohmic contacts have been suggested to be superior candidates due to their reproducibility and low specific contact resistance. The excellent electrical integrity of Ni-based ohmic contacts is due to the formation of Ni silicide after high temperature annealing at 950°C[7,8,9]. In particular, it is well known that the Ni2Si/SiC contact has a specific contact resistance that remains substantially unchanged upon annealing in N2 environment as long as one hour at 1000°C. But, frequently, the up-most metallization layer (bond pads) is exposed to noninert environments, i.e. air. So, the used metal should not react with air (mainly with oxygen) and this is not the case of Ni2Si. To shelter the contact metallization from the ambient, Ni2Si should be covered with a conductive metal layer with a high resistance to the reactive ambient. In this work the Au/TiW/ Ni2Si/SiC multilayer is investigated concerning its physical and chemical stability, and its electrical stability at several temperatures up to 600°C, varying the annealing environment (O2 atmospheric pressure and under turbo vacuum P=10 Torr). Materials Science Forum Vols. 457-460 (2004) pp. 873-876 online at http://www.scientific.net © 2004 Trans Tech Publications, Switzerland Licensed to CNR Area della Ricerca di Genova Genova Italy All rights reserved. No part of the contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 150.145.42.7-21/09/04,15:49:17) Journal Title and Volume Number (to be inserted by the publisher) 2 Experimental The deposition of 100 nm thick nickel contacts was performed by e-gun evaporation at a pressure of 1×10 mbar and an average rate of 5 nm/s. A subsequent thermal treatment of the Ni/SiC was performed by rapid thermal annealing in N2 atmosphere at 950 °C for 60 s. To study the Au and its diffusion barrier characteristics, firstly a TiW film, 600 nm thick, was deposited on a 200 nm thick Ni2Si layer by sputtering at room temperature using a W-Ti (10%) target. A subsequent Au layer, 100 nm thick, was sputtered on the film. The multilayers film was then subject to thermal treatments at several temperatures in O2 environment or under a 10 Torr turbo vacuum. X-ray diffraction (XRD) was performed to monitor changes in the crystalline phase and X-ray photoelectron spectroscopy (XPS) was carried out to determine the depth profiles of the present metals and to determine the diffusion processes in the first 100nm, while to see interdiffusion phenomena near the ohmic contact depth profiles were carried out by Secondary Ion Mass Spectroscopy (SIMS). The electrical characterisation was performed by measuring the specific contact resistance ρc with the transmission line method (TLM) In this a thermally grown oxide layer was used as electrical isolation between the transmission line structure, an array of seven equal area pads (180×100 μm) open by standard lithography. The electrical stability of the contact was checked after 3 hours annealing at several temperature up to 600°C. Results and discussion The contact resistivity was measured on samples before (ρco) and after (ρcf) thermal stresses. The measured values at room temperature are 1.8±0.04 ×10 Ω×cm. The ρsn =ρco/(ρcf) ratio is reported in Fig. 1 versus the process temperature. The square filled dots refer to thermal stresses in vacuum while the open circles are the values obtained when processing in oxygen. In the first case no variation in ρsn is present, the contact remains substantially unchanged after a thermal stress up to 3 hours at 600°C, instead, in the second case, ρsn have an abrupt variation at this temperature that we can’t even measure it. However, till 575°C considering a thermal treatment for three hours there aren’t substantially variation in resistivity. Furthermore, long time treatments up to 100hrs in O2 environment were performed. In this case the contact resistivity up to 450°C remains substantially unchanged demonstrating for this system up to this temperature a good thermal stability. 0 100 200 300 400 500 600 0,6 0,8 1,0 1,2 1,4 1,6 990