Temperature dependence of resistance in reactively sputtered RuO2 thin films

Abstract

Ruthenium dioxide (RuO2) is a stable conductive and semi-transparent oxide. It has been investigated in the past for several applications: (i) good diffusion barrier properties in integrated circuit processing [1], (ii) improvement of reliability when used as electrodes for ferroelectric capacitors [2], and (iii) thin film resistors capable of achieving zero temperature coefficient [3, 4]. The property of nearly zero temperature coefficient of resistance is very important in circuits requiring a high degree of accuracy. Previously there are several reports on the characteristics of the microstructure and resistivity of sputtered RuO2 under different deposition conditions such as oxygen flow rate and annealing temperature [3–6]. But there are still uncertainties concerning the physical mechanisms responsible for the zero temperature coefficient, and the type of charge carriers. In this letter, we will provide a more comprehensive description of the electrical characteristics of reactively sputtered RuO2 film, including some Hall effect measurements which have not been reported before. N-type silicon wafers with (100) orientation were chosen as the substrates. Silicon dioxide (SiO2) was grown by thermal oxidation to a thickness of 500 nm. RuO2 thin films were reactively sputtered by r.f. magnetron sputtering of a ruthenium target in a mixture of Ar and O2. The relative partial pressure of oxygen to the total gas pressure, p(O2)/p(O2 + Ar), was set to be 45%. Prior to sputtering, the system was pumped down to a base pressure of less than 3 × 10−7 Torr. During sputtering, the input power was about 5 W/cm2, and the total pressure was 0.6 Pa. The substrate temperature during deposition was varied between 50 ◦C and 150 ◦C. The samples were then annealed in a nitrogen ambient at different temperatures from 200 to 500 ◦C for 30 min. Rutherford backscattering (RBS) verified that the atomic ratio of oxygen to ruthenium atoms in the deposited films is 2. The microstructures of the films prepared under different conditions were examined by X-ray diffraction (XRD). The resistivity of the films were measured in Van der Paul structure on a die size of 1 cm2 for temperatures ranging from 30 K to 400 K. Hall measurements on the films were also performed at 30 K with the same sample structure. The temperature dependence of the resistivity of the RuO2 films deposited at 100 ◦C substrate temperature and annealed at different temperatures is shown in Fig. 1. The behavior for samples deposited at other substrate temperatures 50 ◦C and 150 ◦C are similar (not shown). In Fig. 1 the resistance values at the low temperatures around 30 K are almost independent of temperature for all annealing conditions. But at temperatures above 100 K, the resistance shows definite trends with temperature variation. The temperature coefficient is negative for annealing temperatures of 200 ◦C and 300 ◦C, but is positive for annealing temperatures of 400 ◦C and 500 ◦C. To correlate with the microstructure, the X-ray diffraction patterns (XRD) are shown in Fig. 2 for the same sample. It shows that the film is crystallized with a sharp peak at 28 ◦ (corresponding to