Abstract—Silicide formation as a result of the reaction of metals with silicon is a widely studied topic in semiconductor industry since silicides form an essential part of modern day Integrated Circuits (ICs). In most situations the fundamental kinetics of silicide formation are analyzed using elaborate techniques such as X-ray Diffraction (XRD) and Rutherford Backscattering (RBS). From those data the resulting silicide thickness is then derived. We propose a simple and elegant alternative in the form of in situ four point probe resistance measurements. Subsequently, we calculate the resulting silicide thickness using a simple 3-layer (metal – metal silicide – silicon) resistive model. The devices under test consisted of Pd sputtered onto bare Si wafers. Pd was chosen as it is well described in literature and reacts already at low (< 150° C) temperatures. We looked at two different temperature regimes to validate our model: the high temperature (150 – 225° C) and the low temperature (50 – 150° C) regime. The results obtained via this simple measurement technique showed that silicide formation of Pd contacts on Si occurs already at 57° C. Furthermore, below approximately 100° C our results indicated that apart from Pd 2 Si formation also mixed layers of Pd in Si and vice versa exist whereas at higher temperatures these mixed layers are absent or not significantly present. For the high temperature data we found a good agreement between our results and literature. We therefore consider this measurement technique as an elegant, universally applicable method for monitoring silicide formation of metal-silicon contacts. Index Terms— Modeling, silicides, semiconductor devices, interconnect. I. INTRODUCTION ilicides form a basic part of integrated circuits in which they can be used as Ohmic contacts, Schottky barrier contacts, local interconnects, diffusion barriers, and gate electrodes. Knowledge of the temperature dependent growth kinetics of silicides is very important for proper device manufacturing. Many silicides grow according to diffusion based reactions [1], which can be characterized via an activation energy E A . For the derivation of E A one needs to Manuscript received October 20, 2009. This work was supported by the Netherlands Technology Foundation (STW). E. J. Faber, R. A. M. Wolters, B. Rajasekaran, C. Salm, J. Schmitz are with the Semiconductor Components Group, MESA+ Institute for Nanotechnology, University of Twente, P.O. Box 217, 7500AE Enschede, The Netherlands (phone: +31-534892669; fax: +31-534891034; e-mail: e.j.faber@ewi.utwente.nl). R.A.M. Wolters is with NXP Semiconductors Research, Eindhoven, The Netherlands. know the formed silicide thickness d as a function of time and temperature. In literature d is often derived via complex, physical characterization techniques like X-Ray Diffraction (XRD) or Rutherford BackScattering (RBS). We propose a simple and elegant alternative for monitoring layer growth via in situ four point resistance measurements. For this we studied the silicide formation of Pd 2 Si via Pd layers on crystalline Si. Palladium is of interest as a high work function (5.12eV [2]) metal to form good contacts to p-type silicon. Pd on Si, however, is known to form Pd 2 Si at temperatures even below 200° C [3,4]. Pd 2 Si Silicide formation at such low temperatures should be taken into account when devices are processed under conditions with unknown temperature, such as sputtering of Pd onto Si surfaces. For the purpose of this study, however, this well known behavior is exploited in evaluating our measurement technique. We looked at both high temperature (> 150° C) kinetics of Pd with Si and low temperatures (< 150° C) kinetics. The high temperature measurements are also well described in literature using analysis via XRD and RBS techniques and make a comparison with our proposed method possible. The resistance measurements are complemented with XRD and RBS. II. THEORY The reaction of metallic Pd layers with bulk Si has been extensively studied. Below approximately 800° C only one stable phase is formed, namely Pd 2 Si. The formation of this silicide follows diffusion limited kinetics [1]. The formed layer thickness of the silicide d Pd2Si grows according to the Arrhenius equation: kT E Si Pd A te D d − = 0 2 2 (1) where D 0 [m 2 s -1 ] is the pre-exponential term, t [s] is the time, E A [eV] is the activation energy, k [eV·K -1 ] is Boltzmann constant and T [K] the temperature. Values for the activation energy vary between 0.9 – 1.5 eV [4,5]. Most studies in literature calculated the Pd 2 Si thickness via backscattering techniques [5-7] or X-ray Diffraction [3,4]. In this study we determined the Pd 2 Si thickness via in-situ resistance measurements. The basis for the calculation of d Pd2Si is a three layer resistive model of the samples in the form of a stack of Monitoring silicide formation via in situ resistance measurements Erik J. Faber, Rob A. M. Wolters, Bijoy Rajasekharan, Cora Salm, Jurriaan Schmitz S -67-