May_EDFA_Digital

42 / 58

edfas.org
ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 19 NO. 2
42
After graduating from Purdue University, I started
out at Delco Electronics in 1986 and really enjoyed
the work plus the lasting friendships that formed.
I helped solve a pinhole oxide problem that had plagued
the fab for years by using a hole in aluminum foil on an
old Bausch & Lomb illuminator, forming a small spot of
light. Scanning the light manually across the read-only
memory row line, I identified the locus of the failure
by monitoring the change in leakage. Corrective action
ensued based on scanning electron microscopy (SEM)
images of the pinhole. Approximately one year into the
job, the economy went sour for automotive electronics. I
got a job interview with AMI Semiconductor in Pocatello,
Idaho, and went west.
The AMI fab had issues with oxide leakage, and this
is when I first conceived passive voltage contrast, which
is still widely used today. The gate oxide (GOX) test struc-
tures frequently failed at probe, and the root cause was
identified as damaged GOX without an understanding of
when this occurred in theprocess. Passive voltage contrast
allows differentiationof conductors from insulators, using
the electronbeamat various tilt angles. It was determined
that the GOX damage occurred before the source/drain
implant. This meant the poly plasma etch process was to
blame, and successful corrective action ensued. Passive
voltage contrast became an in-line screen at AMI. The
work was published at the EOS/ESD Symposium and
ISTFA 1990.
PATENTS: US5892539, US5764409,
US6112004
I left AMI and went to Fremont, Calif., to work for
Waferscale Integration, Inc. (WSI). I managed the failure
analysis lab and had minimal equipment to work with in
the beginning: a wire-caged lab equipped with an SEM,
analog oscilloscope, optical microscope, and an old
probe station. Over the years, with the support of a great
manager, I was able to build the lab, adding a confocal
microscope, memory tester, probe station with laser
INVENTOR'S CORNER
cutter, and so on. We needed photon emission capabil-
ity, so I looked at what was available. After seeing how
poorly they were designed and how hard they were to
use, I concluded that I could build my own. A frame-edit
videocassette recorder and a Sony extended infrared
charge-coupled device sensor ended up smoking the
doors off the current “scientific-grade” and Gen III inten-
sified systems at a fraction of the cost. I contracted with
Alpha Innotech (a bio-tech company) to add the FA1000
photon emission and vibration couplers to allowportable
test head interfaces to their product offering. This system
sold very well for them.
PATENTS: US7872485, US7323888
After seven years at WSI, I resigned and have been
pursuing my own ventures since 2000.
I realized a need in the industry for localizing fault
locations. I conceived the idea of rastering a laser over the
diewhile looping test functionally. Coordinating the laser
position to the pass/fail status along with synchronized
Idd mapping could provide this fault isolation. Ultimately,
this concept became the stimulus-induced fault test (SIFT)
tool. I published this work in 2002 and had up to one year
after publication to patent it. I filed the provisional patent
in August 2003.
PATENTS: US8797052, US9411002
Building on the SIFT work, I then developed time-
of-flight SIFT (TOFSIFT). Realizing that the industry was
moving to stacked dice or situations with no direct access
to the active layers, I began to experiment with pulsed
heat from a CO2 laser to measure the propagation time
and therefore the distance to a known substrate diode.
The target areamust be able towithstand the laser power.
This technique uses a continuous-wave thermal laser to
heat the device under test locally at known coordinates
and cycle the beam as a function of the propagation
time it takes the thermal wave to reach the defect. By
trilateration at different stimulus points, the defect can
THE PASSION OF INVENTING
Jim Colvin, CEO, FA Instruments, Inc.