Metadata only
Author
Date
2017-07-12Type
- Conference Paper
ETH Bibliography
yes
Altmetrics
Abstract
Over the past decade, the utilization of technology CAD (TCAD) tools has become widespread in industry and academia.The enormous progress in computing hardware and software technology, together with significant advances in physical modelling accuracy, and speed and robustness of numerical algorithms, have made TCAD to a cost-effective albeit powerful technology that complements experimental approaches to wafer processing and metrology. Technology CAD has its origins in the 1960s in the pioneering work of Hermann Gummel and Don Scharfetter at Bell Laboratories. It was Gummel in 1964 who applied von Roosbroeck’s semiconductor transport theory to the one-dimensional (1D) numerical solution of a (then modern) bipolar transistor. In his seminal paper, Gummel also proposed basic concepts for tackling the mathematical and numerical issues that are inherent in semiconductor transport. Together with Scharfetter, Gummel published another ground-breaking paper, inventing the Scharfetter-Gummel Discretization Scheme that set the stage for all following work in numerical device simulation. Only a few years later, in the early to mid 70s, both Slotboom and Prof. Engl’s group at Aachen (among others, this summary is not exhaustive, we could also mention the work by groups at IBM Hitachi or in Vienna) came forward with the first results on 2D modeling of devices. During the 1980s, research activities in numerical mathematics, physical models and software engineering florished, and many important results were obtained that laid the basis for todays TCAD industry. At the same time, universities began to (more or less freely) distribute their results in the form of software packages. Programs such as MINIMOS or PISCES-II were widely used in academia and industry, indirectly setting the stage for the creation of the TCAD industry [1].
Typically, TCAD is used in R&D departments to investigate new technologies. It is the best solution for the early technology investigation because the TCAD results can express the trend of the device behavior under different conditions or criteria [2,3].
As the development evolves, TCAD can be really useful to study the behavior of the transistor in an accurate way when it is coupled to some measurements data, what is called calibration. In addition, if the TCAD model is calibrated, it is possible to predict virtually and accurately the behavior of other devices that share the same process flow steps but with different conditions. This is needed to predict the device behavior in a range that no test devices are produced.
Another area where TCAD is helpful is the technology transfer where TCAD can reduce by 40 % the migration cost according to the international technology roadmap of semiconductor (ITRS). The innovative domain, where TCAD is also helpful and has demonstrated its impressive advantage, is the technology optimization and yield management. By simulation, the TCAD engineer can investigate some geometrical effect, doping variation, and decouple the physical phenomena to improve the transistor behavior, the power consumption, and to optimize any process step by coupling device, process, stress and reliability engineering.
TCAD tools can be found under two categories: commercial tools and academic tools. The commercial tools, what we deal with, are developed to cover the most wide range of devices such as MOS transistors, bipolar transistor, three-five devices, lasers and even they can be coupled to optoelectronics simulations such as sensors. These tools are user-friendliness, flexible and respond to the industry needs. The models are the mainly common in use models that are implemented using the most advance algorithms and solvers.
The strength of these tools is the compatibility and the integration between the different tool levels as their support. The academic tools are more device-type-oriented or application-oriented and generally have the most complicated models that make them highly accurate; however, unpractical.
Commonly, TCAD simulations are two dimensional (2D), where the device simulated corresponds to a cut in the middle of the real device. This has demonstrated its efficiency to investigate wide devices where the two width sides are considerably fare from each other. As the device is shrinking, where the sides are getting closer, their influence is becoming important. Due to the side and corner effects, the need for three-dimensional (3D) simulations is becoming mandatory to simulate accurately, by including all physical effects, the behavior of the device.
Special attention has to be paid to 3D simulation due to the limitation of some models in the TCAD tools especially at the process level. The oxidation is the major issue for 3D simulations from model implementation. However, the totality of the other 2D models is extended for the 3D simulations for process and device simulations. Another challenge for the 3D simulations is the meshing engine and techniques. To simulate correctly a device under 3D, it is recommended to optimize the mesh generation because the run time and the accuracy of the results are closely related to the mesh size and quality. In addition, the 3D process flow generation is also based on emulating the structure generation, this means that the process steps are reproduced by geometrical operation. This can lead to high complexity of the structure generation if the user did not follow a methodology to achieve the structure.
References
[1] W. Fichtner, N. Braga, M. Ciappa, V. Mickevicius, and M. Schenkel, ``Progress in technology CAD for power devices, circuits and systems,'' in Proceedings International Symposium on Power Semiconductor Devices and IC's (ISPSD), pp. 1-9, 2005.
[2] Y. Saad, “TCAD-Based Three-Dimensional Modeling of Nonvolatile Memories”, Hartung-Gorre Verlag
[3] L. Sponton, “From Manufacturing Variability to Process-Aware Circuit Simulation”, Hartung-Gorre Verlag Show more
Publication status
unpublishedPublisher
ETH ZurichEvent
Subject
TCAD SimulationOrganisational unit
03380 - Huang, Qiuting (emeritus) / Huang, Qiuting (emeritus)
Related publications and datasets
Notes
educational materialMore
Show all metadata
ETH Bibliography
yes
Altmetrics