Dielectric materials play a vital role in the performance of microelectronic devices, because they have the potential to electrically separate conductive components from one another in microcircuits. A material’s dielectric constant, κ, is defined by the ratio of the capacitance of two conductors, divided by a dielectric to the capacitance of the same conductors isolated by vacuum. A circuit’s maximum operating frequency can be limited by the capacitance between conductors. The capacitance is also found to increase in inverse proportion to the distance of separation between the conductors. It is necessary to separate the components of the device, using a material with a dielectric constant as low as possible in order to reduce the size of a microelectronic device and also increase its working frequency. This process is carried out using a specific group of materials known as ultra low-κ (ULK) dielectrics.
Low-κ materials are obtained by a vital transaction between the electrical and mechanical properties. The integration of nanometer-scale pores to decrease κ results in reduced stiffness, strength, and adhesion of the deposited films.
It is important to control the mechanical properties of ULK films during a semiconductor production process in order to ensure that the device survives and yields a reliable, dependable finished product. It is extremely important to incorporate newly developed higher porosity and lower κ materials for every single change performed in the semiconductor node. Currently, the mechanical reliability monitoring of ULK films is gaining immense importance to rapidly identify process difference and maintain high device yields. Nanoscratch and Nanoindentation testing provide a way to determine the critical scratch force (adhesion), modulus (stiffness), and hardness (strength) of ULK films.
Nanoindentation testing of ULK films is executed by pushing a diamond pyramidal probe within the film to a given force, retaining the force for a few seconds, and then withdrawing the probe. Displacement and force are continuously determined throughout the testing process, and based on this, both the modulus and hardness of the material are estimated.
Figure 1 displays an example of a force-displacement curve obtained from a test performed on a ULK film with 200 nm thickness. A cube-corner geometry was used due to its acute geometry, which permits it to pass through a thin, moderately dense skin layer on the ULK film’s surface, probing the properties of the interior portion of the film. The measurement is influenced by the properties of the silicon substrate in situations when the indent is too deep. The finest measurements of the film’s properties are obtained only when the indent depth is more than what was needed to pass through the surface layer.
Figure 1. Representative force-displacement curve from an indent on a ULK film.
Extremely high localized characterization of the mechanical properties of the material is obtained because of the small scale of the nanoindentation process. However, when nanoindentation tests are executed in arrays, they can also be utilized to develop maps of mechanical properties of surface areas that are larger. Bruker’s Hysitron Nanomechanical Metrology Tool (NMT) series includes instruments that are custom designed for nanomechanical testing in process control applications. The translation stage of the ATI 8800 (Figure 2) is capable of handling silicon wafers up to 300 mm diameter, comprising of an adequate range to reach any area on the wafer.
Figure 2. Hysitron ATI 8800, fully automated nanomechanical metrology tool for 24/7 wafer process monitoring.
The ATI 8800 is capable of being used to carry out a number of tests in a pre-determined area of a wafer to check for mapping mechanical properties over a specific area or inter-wafer process variability. In this situation, several 1884 nanoindentation tests were executed to find out the homogeneity of the mechanical properties of a ULK film across the whole surface of a 300 mm wafer (Figure 3). The mechanical property maps generated by the array of tests highlights that modulus and hardness differed by approximately 10 to 15% throughout the ULK film’s surface.
Figure 3. Results from 1884 nanoindentation tests on a 200 nm ULK film, showing a property variability of 10-15% due to non-uniform processing conditions.
The ULK film’s adhesion to the underlying substrate was measured by Nanoscratch tests. In each scratch test, the probe was laterally shifted along the wafer’s plane for a distance of 10 µm while simultaneously increasing the normal force from 1 to 1500 µN. A diamond 90° conospherical probe with a 1 µm radius of curvature was utilized for the tests.
Increase in the normal force in a scratch test results in the probe sinking deeper into the material, increasing the lateral force and introducing increasing stress on the substrate/film interface. When a specific kind of stress is applied, the film delaminates from the substrate, and this delamination event is shown in the data as an unexpected decrease in lateral force together with an increase in normal displacement. The normal force at delamination is documented as a significant normal force and is utilized as a measure of the interfacial failure load or interfacial delamination load of the film (Figure 4).
Figure 4. Representative data from a nanoscratch test showing how the critical load was determined.
The in-situ SPM imaging capability of the instrument was used to obtain a topographical image of scratches at varied points in the test to assure that the initial critical event matches to film delamination, while the extremely larger event that follows is the result of film spallation. Hence, it is important that the instrument utilized for executing such tests is able to precisely identify the relatively slight onset of delamination as opposed to the even more evident film spalling event.
A set of 1884 nanoscratch tests was executed on the wafer to map the interfacial adhesion of the ULK film over the surface.
Figure 5. Results of 1884 scratch tests on a 200 nm ULK film, showing adhesion variability due to non-uniform processing conditions.
Figure 5 displays a considerable difference in film adhesion from one side of the wafer to the other, and over most of the wafer, the adhesion was evidently lower within ~20 mm of the wafer edge.
This information has been sourced, reviewed and adapted from materials provided by Bruker Nano Surfaces.
For more information on this source, please visit Bruker Nano Surfaces.