Parametric investigation of latch-up sensitivity in 1.25 μm CMOS technology

The dependence of latch-up sensitivity in 1.25 μm p-well, CMOS devices on layout parameters, backside contacts, and thickness of epitaxial layers has been investigated by analysis and testing using a specially designed test chip. Retrograde wells and small source-to-substrate/well contact spacings play a critical role in suppressing latch-up, especially in non-epitaxial devices, by reducing the gains of vertical transistors. The backside contact directs high currents to the n+ source without developing large IR drops underneath the p+ source, thereby increasing the holding current. The epitaxial layer devices exhibit the holding voltages substantially higher than the ones for the non-epitaxial devices. Analyses and measurements have shown that the vertical transistors are in the-active mode at latch-up in the epitaxial devices in contrast with the saturation mode in the non-epitaxial devices. There exists an optimum epitaxial layer thickness, at which the holding voltage is a maximum. The optimum spacing is primarily determined by the n+ to p+ source spacing.

[1]  D.S. Zicherman,et al.  A highly latchup-immune 1-µm CMOS technology fabricated with 1-MeV ion implantation and self-aligned TiSi2 , 1986, IEEE Transactions on Electron Devices.

[2]  E. Sangiorgi,et al.  Three-dimensional distribution of CMOS latch-up current , 1987, IEEE Electron Device Letters.

[3]  R.A. Martin,et al.  Three-dimensional effects in CMOS latch-up , 1986, 1986 International Electron Devices Meeting.

[4]  A. Ochoa,et al.  A Discussion of the Role of Distributed Effects in Latch-Up , 1981, IEEE Transactions on Nuclear Science.

[5]  J.A. Seitchik,et al.  An analytic model of holding voltage for latch-up in epitaxial CMOS , 1987, IEEE Electron Device Letters.

[6]  A. A. Witteles,et al.  The Dependence of Latch-Up Sensitivity on Layout Features in CMOS Integrated Circuits , 1986, IEEE Transactions on Nuclear Science.