backside

The “professional” mid range solution. Typically some form of Nd:YAG, either the fundamental or some harmonic. Expect minimum of a few thousand USD for a system, probably many more. Probably the most economical would be to piece together a Mitutoyo microscope. Of course, be careful not to look into your home spun laser while firing without good reason.

The “professional” high end solution used for small processes. Equipment is rather expensive. Some estimates seem to put them around an order of magnitude of 500,000 USD.

FIB time can often be rented on an hourly basis from universities and research labs.

azonenberg is beginning to experiment with FIB. Lab notes

Microprobes can cut traces by moving back and forth. Professional units use peizo's (ie ultrasound).

By masking the chip and then etching (ex: with HF) wires can be removed. Masking compounds range from photoresists to nail polish.

Usually done with a fib. Can be done temporarily with microprobes or possibly on a high process chip very carefully with conductive epoxy.

JM's highlights from watching the presentation

exposing wires

• laser
  • expensive
  • more reliable with experience
  • better option for smaller areas
  • better option for planar chips
• scratching
  • good option for larger, non-planar chips
  • lower cost

fuses

• binary search by covering chip
  • try to erase with UV light
• trace VPP
  • charge pump or external
• confirm guess via microprobing

op amp vs ttl amp for gain

• microcircuits damaged or distorted when disturbed
• op amp in volt follower mode
  • “more transparent electrically”
  • look at noise level
    • drops when makes contact
    • not necessarily the right wire though

fib probing

• shoots ions instead of electrons
• filled hole with platinum to make a bond pad

anti-anti-re

• cut charge pumps
• jump security mesh
• go from back side
  • challenge: can destablize chips
  • tends to be used more by FA than RE
• peizo probing
  • they didn't like it
  • I've seen others use it

future/upcoming anti-re

• analog meshes
  • The IBM PCI card thing has one externally but maybe not plausible yet at chip level

No mention of masked etching. Ex: putting PR or nail polish and etching with HF

• Functional IC Analysis: http://nedos.net/host2012.pdf
• Breaking and Entering through the Silicon: http://nedos.net/ccs2013.pdf
• Low-Cost Chip Microprobing
  • http://events.ccc.de/congress/2012/Fahrplan/events/5124.en.html
  • http://media.ccc.de/browse/congress/2012/29c3-5124-en-low_cost_chip_microprobing_h264.html