Delayering, also know as deprocessing, are techniques for getting to the various layers in a chip. Typically this is for imaging or microprobing but is also used to alter circuits or to prepare for fault injection. Barring noninvasive methods like CT scan, any delayering procedure will involve removing parts of the chip and will render it permanently non-functional.

Selectively

Main article: Selective removal

Wet etching

Main article: wet

Abrasive

Main article: abrasive

Electrochemical

Main article: electrochemical

Reactive Ion (plasma) Etching (RIE)

Main article: rie

Almost exclusively a professional shop method. Very good results but high barrier to entry ($5000 for a used machine if you're lucky).

Thickness

TODO: get numbers for passivation (SiO2), metal, etc

tox (oxide thickness) scales down with CMOS.

130 nm technology has gate oxides of 20Å or thinner http://en.wikipedia.org/wiki/CMOS

0.5 um copper (http://books.google.com/books?id=MlAlzr_2K8MC&lpg=PA64&ots=sRv-scJ7vW&dq=cmos%20metal%20thickness&pg=PA64#v=onepage&q=cmos%20metal%20thickness&f=false)

Chemical-Mechanical Planarisation (CMP) considerations

It may not be necessary to delayer depending on the type of analysis being done. For example, 7400 and similar chips often have so few layers and the process technology is so large that a top metal photograph can show transistors.

No CMP:

Original caption: “Second metal layer and polysilicon layer can be seen through top metal layer on Microchip PIC16F877 microcontroller. 500× magnification.” Copyright 2005 Sergei P. Skorobogatov, used with permission

CMP:

Original caption: “Nothing can be seen under the top metal layer on Microchip PIC16F877A microcontroller as the layers inside the chip were planarised during fabrication. 500× magnification.” Copyright 2005 Sergei P. Skorobogatov, used with permission

Lift off

Beck defines “lift off” as “total removal of all layers down to the substrate”. It doesn't seem this is a widely used term but is the only term that we can find for this process. Lift off is important for crystallographic etching. One reason for its lack of popularity might be that the semiconductor field usually refers to lift off as a fabrication technique (see for example http://www.springerlink.com/content/j25g014709338534/#section=85698&page=1).

A. Zonenberg etch rate measurement

FIXME: move this somewhere better

Copper etch - 70 ml 3% H2O2 + 10 ml conc. HCl. Heat to 70C, agitate with magnetic stirrer.

Test sample was Altera Stratix IV. Silver or tin plated copper. Plating was gone after <1 min in etchant, metallurgy of the plating was not studied. 20 minutes of etching = 0.13mm removal so 6.5 um/min. This number is likely +/- 30% or so because the temperature of the solution slowly increased over the etch period.

References

  1. “Semi-invasive attacks - A new approach to hardware security analysis.” Sergei P. Skorobogatov.
 
delayer/start.txt · Last modified: 2018/02/28 06:55 by mcmaster
 
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution 4.0 International
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki