Surface induced latchup in VLSI CMOS circuits

Takacs, D.; Werner, Christoph; Harter, J.; Schwabe, Ulrich

doi:10.1109/t-ed.1984.21515

Cited by 29 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For future VLSI/ULSI integrated circuits thin, lightly doped epitaxial layers grown on highly conducting substrates will be required to minimize CMOS latch-up and improve radiation hardness. Recent investigations have shown that the closer the highly conductive substrate is placed to the active device region (meaning of course the thinner the epitaxial layer) the more resistant to latch-up the devices fabricated on this structure are (1). Technologies such as plasma-assisted epitaxy (2) and low temperature, low pressure epitaxy (3) are being aggressively pursued to grow these thin epitaxial layers by minimizing the problems of gas phase autodoping and solid-state outdiffusion.…”

mentioning

confidence: 99%

Investigation of Various Substrate Dopants and Epitaxial Growth Techniques for Producing Sharp Transition Epitaxial Wafers

Schmidt

1987

J. Electrochem. Soc.

View full text Add to dashboard Cite

Arsenic, antimony, phosphorus, boron, and gallium doped silicon substrates were used for the fabrication of n/n + or p/p+ epitaxial wafers to study autodoping effects in a horizontal induction heated atmospheric pressure epitaxial reactor. Using the "high-low" technique which uses a high temperature prebake (without HC1 etching) followed by a relatively low temperature and slow growth rate deposition, sharp n/n + transition on arsenic material was achieved. This method appears to produce an epitaxial layer of more uniform resistivity out to the beginning of the transition region for all of the n-type substrates used. This is indicative of reduced autodoping effects. Results show that when the epitaxy employs the high-low technique at a temperature of 1050~ heavily doped arsenic substrates offer unique transition and defect density advantages over either antimony or phosphorus material. This should enable a reduction in n/n + epitaxial layer thickness. The high-low technique did not reduce autodoping for the p-type substrates. The gallium substrates evaluated during these experiments exhibited unexpected behavior in light of the fact that gallium has a diffusion constant nearly identical to that of boron at 1100~ However, the gallium wafers showed little difference in transition between any of the deposition conditions tried. Extreme gas phase autodoping is believed to be the cause for these gallium doped substrate results.For future VLSI/ULSI integrated circuits thin, lightly doped epitaxial layers grown on highly conducting substrates will be required to minimize CMOS latch-up and improve radiation hardness. Recent investigations have shown that the closer the highly conductive substrate is placed to the active device region (meaning of course the thinner the epitaxial layer) the more resistant to latch-up the devices fabricated on this structure are (1). Technologies such as plasma-assisted epitaxy (2) and low temperature, low pressure epitaxy (3) are being aggressively pursued to grow these thin epitaxial layers by minimizing the problems of gas phase autodoping and solid-state outdiffusion. However, specialized equipment is needed to implement these advanced technologies. It is desirable to find a method of improving transition sharpness in conventional atmospheric pressure equipment without modification.The high-low technique was developed by G. R. Srinivasan at IBM (4-6), and is used in bipolar integrated circuit manufacture to reduce autodoping and pattern shift in buried collector structures (which are typically arsenic doped). It was felt that the advantages demonstrated for the HLT in bipolar epitaxy could be transferable to arsenic epitaxial wafer n/n + structures. These results indicate that the benefit of reduced autodoping demonstrated for the HLT in bipolar VLSI does carry over and enables the production of very sharp epitaxy/substrate transitions.This current work concerns itself with conventional atmospheric pressure silicon epitaxy and attempts to determine which substrate dopant and ep...

show abstract

mentioning

confidence: 99%