February_EDFA_Digital

S-2 / 62

ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 18 NO. 1
S - 2
that accuracy and speed in diagnosing issues is essential to
semiconductor manufacturer success or failure.
The last portion of the panel was dedicated to an open
conversation between the attendees and the panelists.
It was discussed that, to achieve success, everyone must
own the issue across organizations. It was mentioned that
the FA engineer should always be involved in any problem-
solving session. One attendee also commented that
spending money for tools is the easy part; finding FA engi-
neers with the appropriate skills is harder. This comment
led to several exchanges on what defines an FA engineer,
and most attendees must have recognized themselves in
the different descriptions given. The panel concluded on
a topic proposition for the ISTFA 2016 Panel Discussion:
“The Next-Generation FA Engineer.” See you next year in
Fort Worth, Texas!
ISTFA 2015 CONTACTLESS FAULT ISOLATION USER GROUP
Moderators: Patrick Pardy and Dan Bockelman, Intel Corporation
Optical probing techniques continue to be critical
in the failure analysis (FA), fault isolation (FI), and
product development space. As the industrymoves
into the 10/14/16/22 nm geometries, many wonder if the
current techniques will generate the results needed to
improve yield, debug, and characterization and to move
products tomarket. Or, instead, are newtechnologies badly
needed to sustain today’s pace of innovation in the process
space? To that end, the topics this yearwere specific toward
new techniques and technologies. An inflection point in
the wavelengths of current infrared (IR) tools to potential
shorter wavelengths solutions was discussed. New studies
and some initial toolset work to include visible laser
probing was completed, as well as an alternative 1154 nm
wavelength for current IR toolsets. A full presentation on
debug data techniques, for the foundry- and design-only-
based scenario, was also given at this year’s conference.
Further discussion followed on a number of design-for-
test (DFT) and scan failures, and the physical toolsets/
techniques used to help solve silicon issues were demon-
strated. Future recommendations onDFT, test looping, and
test complexities moving forward were discussed as well.
Finally, a new technique used to debug/characterize silicon
designs in a motherboard/platform setup was discussed.
For a new integration of an infrared emission microscopy
(IREM) systemand amotherboard, themeans todebug and
characterize new products and features was completed.
Dr. Christian Boit (Berlin University of Technology,
Berlin, Germany) gave the first presentation, “Contactless
Visible Light Probing for Nanoscale ICs through 10 µm
Bulk Si.” This presentation focused on imaging and
optical techniques with visible laser light. Using confocal
microscopy to obtain better degrees of freedom in probing,
the near-infrared plus silicon solid immersion lens (SIL)
combination is only good for >20-nm-node technologies.
However, moving forward, a minimum 2 × improvement
will be needed to keep up with the aggressive shrink for
future process nodes. Because the numerical aperture for
IR-based optical tools is already at the limit (3.5 for silicon),
one of the only other adjustments is lambda (wavelength)
for resolution improvements. Using shorter-wavelength
lasers has advantages, as described above, but very big
challenges as well. One of these challenges is that silicon
is very absorptive at the shorter wavelengths, especially
those in the visible range. For his work at 650 nm, Dr. Boit
described the reasonable absorption depth (AD) for the
650 nm laser to be ~3.5 µm remaining silicon thickness
(rst). This rst depth provided the optimal working dis-
tance for the focal plane using the confocal system. The
advantage here is that only the focal plane is transferred
and the back surface reflection is suppressed, which may
increase resolution by 1 to 1.5 ×. For 10 µm of rst, the AD
is 6× that of 3.5 µm rst, but using a confocal system still
provides a good image. Dr. Boit used a Zeiss laser scanning
microscope air gap (non-SIL) system for his experiments.
He measured improvements up to 1.68 ×, using an older
process technology. Dr. Boit stated that the bandgap for
the SIL must be >1.9 eV and that GaP, transparent to 550
nm, was identified as a good candidate material. He also
stated that the technology will not be useable with GaP
above 2.25 eV; instead, other materials (i.e., SiC) will need
to be explored. However, SiC has a worse index of refrac-
tion and themaximumsilicon thickness decreases further,
so there are challenges with respect to future materials as