Production-Oriented VDDT and IDDQ Device Testing Methods Based on Multiple Power Supply Pad Measurements

点击翻译按钮获取中文摘要

The International Technology Roadmap for Semiconductors identifies a need for innovative testing and diagnostic methods for digital and mixed-signal devices. Traditional testing strategies are becoming less effective for several reasons. As device operational frequencies increase, the cost of functional test equipment, capable of testing devices at their native speed (at-speed), is becoming cost prohibitive. Testing methods which use slower, cheaper test equipment need to be able to detect defects that cause at-speed delay failures. The increasing complexity and diversity of these devices also make it difficult to access internal nodes and achieve good fault coverage and parametric device information. Technology trends, such as increases in device leakage currents, have reduced the effectiveness of alternative tests such as IDDQ. This research is designed to address these short-coming by investigating device transient and novel quiescent signal techniques. In previous work, process-tolerant VDDT and IDDQ methods have been demonstrated to (1) detect resistive shorting and open defects, (2) predict performance in defect-free devices, and (3) predict defect location in defective devices. These methods are based on the cross-correlation of multiple static and transient power supply signals. Cross-correlation is used to calibrate for process and technology-related variations, such as shifts in transistor betas and increases in leakage current. This research focuses on the implementation of these techniques in a production test environment, and on identifying their capabilities and limitations. The scalability of the methods to large commercial devices is investigated with industrial partners. Production test environment issues, such as measurement noise and instrumentation sampling requirements, test generation strategies and specialized hardware are also investigated.