Development of a Complimentary Suite of Characterization Tools for the Analysis of MOS Device Degradation Using Dynamic Stimuli

Open Access
Author:
Passmore, Lucas Jay
Graduate Program:
Engineering Science and Mechanics
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
May 21, 2009
Committee Members:
  • Osama O Awadelkarim, Dissertation Advisor
  • Osama O Awadelkarim, Committee Chair
  • Joseph Paul Cusumano, Committee Chair
  • Samia A Suliman, Committee Member
  • Jerzy Ruzyllo, Committee Member
Keywords:
  • semiconductor
  • pca
  • health monitoring
  • mosfet
  • reliability
  • mos
  • dynamic
Abstract:
This work is focused on the development of two new techniques to monitor the health condition and reliability of metal oxide semiconductor field effect transistors (MOSFETs). Currently used techniques do not allow for the testing of MOSFETs in their native mode of operation and are often not suitable for measuring production devices and require special test structures. Since MOSFETs are fully dynamical systems, the best course of action for extraction of information about their current health state is in the evaluation of fully dynamical data sets. The first developed technique is an adaptation of the well known charge pumping (CP) technique. We have modified traditional CP, which utilizes substrate current measurement to determine Si-SiO2 interface traps in MOSFETs, to enable its application to MOSFETs with no substrate contacts by using the drain current. This latter type of MOSFETs includes those used in CMOS, thin-film, and nanowire transistors. These transistors are all three-terminal devices and our CP adaptation is, therefore, referred to as three-terminal charge pumping (3TCP). The efficacy of 3TCP was evaluated and it showed excellent agreement with both traditional charge pumping measurements, and transfer characteristic measurements. While 3TCP is a step towards the use of dynamical response to gate stimulus, it still employs quasi-static measurements of the net current flow through the MOSFETs. To transition into a fully dynamical analysis of MOSFET degradation, one must move beyond traditional characterization techniques which are limited to quasi-static approaches. In this work we have developed a technique to characterize the health state of MOSFETs through the extraction of degradation metrics is based upon the transient response of the drain to source current, Ids, to a step input to the gate of the device. This Ids response is analyzed by applying principal component analysis (PCA) to determine eigen-transients of the system. We refer to this technique as degradation assessment via eigen-transient analysis (DAVETA). We observe that DAVETA produces a measure of device degradation that is much more sensitive than quasi-static methods. More importantly, however, DAVETA is able to track the evolution of MOSFET degradation in real time and while the device is in its normal mode of operation, which makes it possible to predict remaining time to failure. We have applied DAVETA to study degradation measured in u-shaped vertical MOSFETs and planar MOSFETs, as well as MOSFET degradation obtained from simulation program with integrated circuit emphasis (SPICE) simulation experiments. We found excellent agreement between DAVETA's results and those of traditional quasi-static experiments. We also show from SPICE simulations that DAVETA is capable of distinguishing between different degradation mechanisms in response to an identical gate stimulus offering the possibility of developing a single tracker to distinguish multiple failure modes.