Thanks to an advantageously low electrical resistivity, Cu has been the metal of choice for interconnects in modern nanoelectronics, covering virtually all layers of the back end of line. However, the resistivity of nanostructured Cu rises dramatically as the dimensions are reduced (<~ 30 nm), as a result of extra carrier scattering at the sidewalls and on (smaller) grain boundaries. At a system level, this unacceptably leads to more latency and a higher power consumption as we move forward into the CMOS roadmap. Subsequently, alternative metals, and most promisingly Ru, are being considered for replacement of Cu. Despite a lower bulk electrical conductivity than Cu, these metals have indeed shown a reduced sensitivity to dimensions such that they might outperform Cu at the smallest dimensions. In this context, characterizing these metals and understanding how their electrical properties are impacted by dimensions, grain size, etc become of the utmost importance. Unfortunately, on materials with dimension-dependent electrical resistivity, it is complex to discriminate between resistivity and geometrical effects and no existing technique can extract them separately.
This PhD project proposes to further develop and combine two in-line techniques, i.e. micro four-point probe (m4PP) and scatterometry (OCD) to shed light on the electrical and optical properties of nanometer-wide metal lines. While m4PP measures the electrical resistance of the line, OCD determines its optical reflectance across a broad wavelength range (~250-950 nm), making both techniques sensitive to electrical resistivity and geometry. By itself, m4PP always needs a separate geometrical measurement with e.g. electron microscopy (EM) to isolate the resistivity information. Conversely, OCD is considered a stand-alone geometrical measurement as all the material-dependent information is determined independently on a large solid pad. However, this assumes that the electrical conductivity of the metal is identical in the pad and in the narrow line, which is known not to be the case at reduced dimensions. In other words, both techniques need each other but they also provide each other with the required information to solve this intricate problem.
This project will focus on metals belonging to the CMOS roadmap, i.e. mostly Pt-group metals such as Ru, Ir, Pt etc. Ultra-shallow (10-100 nm thick) blanket layers will first be considered to understand how size effects the resistivity and the optical constants. The latter models will help better understand the behavior in narrow lines. Lines of different dimensions (width, height, sidewall angle,...) and different cladding materials will then be investigated. Being still at its infancy when it comes to confined and densely packed structures, the m4PP technique will have to be further developed and understood. To deepen our understanding, attempts will also be made to extract carrier mobilities using various m4PP probe configurations and/or dedicated structures. Finally, a thorough investigation of the optimal use of the combined m4PP/OCD information will have to be carried out.
This work will be done in the metrology group of the advanced patterning department of imec disposing of a multitude of patterning and characterization techniques in support of this project. It will also be done in very close collaboration with the process engineers of imec and its industrial partners.
Required background: physics (solid-state, semiconductors), material science, engineering
Type of work: 60% experimental, 40% theoretical
Supervisor: Ingrid De Wolf, ,
Daily advisor: Janusz Bogdanowicz
The reference code for this position is 2020-113. Mention this reference code on your application form.