February_EDFA_Digital

38 / 58

edfas.org
ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 19 NO. 1
38
screen is painted with black-and-white pixels to highlight
potential circuit issues and their physical location in the
field of view. An example was shown where normal laser
voltage imaging could not be used due to the low duty
cycle. Instead, the LVT technique was used to visually
signal anomalies. Further root-cause analysis required
tracing the circuit with normal laser voltage probing (LVP).
The minimum gate time is 500 ps, but it is dependent on
the type of signal being probed. The hardware upgrade
adds a new box to the system but does not require a new
rack. The highest bandwidth for the current LVT system
is 450 MHz.
The next presentation, given by Dr. Ramya Yeluri of
Intel Corp., was entitled “Use of Second Harmonic and
Thermal Effects of Laser Voltage Probing.” Dr. Yeluri pre-
sented the background for LVP, stating how an incident
beam, through the backside of silicon, is modulated and
realized in the reflected beam. This change is produced
by the difference in reflectivity with respect to time.
Free-carrier density change is the dominant effect, while
electroabsorption at the active junction is a secondary
effect. It is not possible to separate the effect of gate and
drain spatially on current technologies, due to the small,
compact geometries relative to the laser spot size. All peri-
odic signals can be expressed as a sumof sine and cosine
functions. Perfect squarewaves have odd harmonics only,
so second harmonics can detect deviations (such as a
missing bit, badduty cycle, or bad slew rate). The first case
study included a distorted duty-cycle clock. Typically, the
duty cyclewould need to bemeasured by probing at each
location along a suspected path, but the use of frequency
mapping at the second harmonic enabled detection of
duty-cycle distortion without requiring probing for each
node separately. The second case study demonstrated
how thermal effects can be detected with LVP. For this
case, there was an ohmic short to VSS, which caused con-
tention on the output pMOS of an inverter. While probing
the inverter, the observed amplitude was much higher,
due to the thermal effect. These techniques can be used
with the existing infrared tools and provide for efficiency
improvements in debug cases.
All three presentations were well received, with many
follow-up discussions and questions and answers. Each
of the presentations was tied to either innovations or
efficient use of the existing optical probe tools in today’s
IC semiconductor industrial complex. The content showed
that innovations by both original equipment manufactur-
ers and the customer were still very robust and that future
potential ones exist. Approximately 55personnel attended
the Contactless Fault IsolationUser Groupmeeting, which
was sponsored by Checkpoint Technologies.
ISTFA 2016 FIB USER GROUP
Moderators: Steven Herschbein, Globalfoundries, and Michael Wong, Intel Corp.
Focused ion beam (FIB)-driven sample-preparation
techniques, failure analysis methods, chip circuit
modification/editing, and improved tools continue
to be popular topics at ISTFA. This year, the FIB User
Group was held Thursday afternoon, at the very end of
the conference. Despite the late time, the gathering was
well attended, with a peak attendance of approximately
60 people. Many familiar faces from the failure analysis,
design debug, and materials community along with tool
vendors joined us for light refreshments and stimulating
discussion. Six presentations covered a wide range of
topics, including a survey of conventional as well as novel
conductive polymer coatings used in capping and charge
control; two papers covering the advances andmechanics
of slicing, harvesting, and imaging scanning/transmission
electron microscope (S/TEM) lamellae; the feasibility of
producing an FIB-SEMcluster tool for advanced analytics;
and a circuit edit example that used capacitance rather
than resistance for controlled delay. The organizers also
wish to thank Thermo Fisher Scientific for its sponsorship.
The session’s first speaker was Bryan Tracy of EAG
Laboratories, who presented “Capping and Coatings
for FIB XTEM.” Bryan’s background is in SEM and TEM,
and success in these endeavors is highly dependent on
selection of the right protective coating for the applica-
tion, to deal with charge control and milling artifacts. He
listed multiple key requirements for a good cap, which
included the avoidance of nano-waterfall streaks, pro-
tecting the surface from e-beam and ion beam damage,
understanding specific cap chemical avoidance when