The reliability of integrated-circuit wiring depends strongly on the development and relaxation of stresses that promote void and hillock formation. In this paper an analysis based on existing models of creep is presented that predicts the stresses developed in thin blanket films of copper on Si wafers subjected to thermal cycling. The results are portrayed on deformation-mechanism maps that identify the dominant mechanisms expected to operate during thermal cycling. These predictions are compared with temperature-ramped and isothermal stress measurements for a 1 μm-thick sputtered Cu film in the temperature range 25–450 °C. The models successfully predict both the rate of stress relaxation when the film is held at a constant temperature and the stress-temperature hysteresis generated during thermal cycling. For 1 μm-thick Cu films cycled in the temperature range 25–450 °C, the deformation maps indicate that grain-boundary diffusion controls the stress relief at higher temperatures (>300 °C) when only a low stress can be sustained in the films, power-law creep is important at intermediate temperatures and determines the maximum compressive stress, and that if yield by dislocation glide (low-temperature plasticity) occurs, it will do so only at the lowest temperatures (<100 °C). This last mechanism did not appear to be operating in the film studied for this project.
Stresses in thin films can affect phase transformation temperatures. We show, however, that stress relief and grain growth in Co thin films occur at temperatures below the hcp to fcc phase transformation. Temperature ramped (10 °C/min) in situ stress, differential scanning calorimetry, and resistivity were measured on evaporated 56–275 nm Co films, and microstructure was examined with transmission electron microscopy. As-deposited films are under tensile stress, ≊1×109 Pa, which is relieved between 120 and 275 °C. The as-deposited small grains (15–20 nm) undergo secondary grain growth at 230–350 °C, giving a mixed phase bimodal grain size distribution with the smallest grains (both hcp and fcc) 20–25 nm in diameter and the largest grains (〈0001〉 hcp) 150–200 nm in diameter. The hcp grain growth is signalled by an exothermic calorimetric peak at 282 °C with an enthalpy of 0.15±0.04 kJ/mol and a resistivity decrease. The hcp to fcc phase transformation is signaled by an endothermic calorimetric peak with an enthalpy of 0.27 ±0.07 kJ/mol at 380 °C although transmission electron microscopy reveals that the transformation is incomplete. There is no indication of a reverse transformation on cooling. Additional heat treatment to 736 °C, on the thinner 56 nm Co film on SiO2 substrates, results in agglomeration and an accompanying decomposition of SiO2 catalyzed by Co. The results indicate that as-deposited tensile stress does not control the phase transformation temperature in Co thin films.
Scanning ion microscopy (SIM) employing focused ion beam (FIB) imaging was used to study the grain structure of thin copper films as a function of annealing temperature from 20 to 500 °C. Accurate measurement of grain size is obtained for grains as small as 60 nm, allowing the microstructure of copper to be analyzed on small-grained samples which show poor contrast in scanning electron microscopy. Moreover, the short sample preparation time provides an advantage over transmission electron microscopy (TEM). The growth and coalescence of small (<100 nm) grains in the initially bimodal grain size distribution occurs in the temperature range of 250–350 °C in films of 1000 nm thickness. This grain growth takes place concurrently with the relaxation of compressive stress as observed by temperature-ramped stress measurement. Also, temperature-ramped in situ TEM examination confirms that coarsening of small grains is the dominant grain growth mechanism up to 500 °C.
The addition of 0.4–8.6 at. % Co to Cu thin films strongly influences the temperature evolution of microstructure, stress, and resistivity. For concentrations near 1 at. % Co in coevaporated Cu-Co on oxidized Si, normal grain growth begins at about 75 °C, about 50 °C lower than in pure Cu. There is an abrupt decrease in resistivity and stress at a temperature which increases with Co content from 120 °C (0% Co) to 250 °C (8.6 at. % Co), and coincides with precipitation of Co within Cu grains. A dramatic change in texture is observed in both coevaporated and electroplated Cu-Co films upon annealing above 250 °C. As-deposited films have a three-component texture of (111) fiber, (200) fiber, and random but annealed films have a dominant (200) fiber texture. This ‘‘cube’’ texture differs from the dominant (111) texture of annealed pure Cu, and appears to be coupled to an abnormal grain growth process since many grains are observed to be larger than ten times the film thickness. It is proposed that segregation of Co to external surfaces or to Cu grain boundaries may favor this (200) texture by selectively affecting grain-boundary mobility or the surface energy driven grain growth.
This paper investigates a simple noncontact optical thermometry technique based on the laser interferometric measurement of the thermal expansion and refractive index change of a thin transparent substrate or temperature sensor. The technique is shown to be extendible from room temperature to at least 900 degrees C with the proper choice of a thermally stable sensor. Sensor materials investigated included c-axis A1(2)O(3), MgO, MgAl(2)O(4) (spinel), Y(2)O(3)-ZrO(2) (yttria stabilized zirconia), and fused silica. Calibration data were taken at 633 nm by measuring the sensor response to known temperature changes. These data provided (1) the information needed for quantitative thermometry (i.e., the functional relationship between interference fringes and temperature for samples of known thickness) and (2) the thermal coefficient of refractive index for those materials with known thermal expansion coefficients.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.