May_EDFA_Digital

55 / 58

55
ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 18 NO. 2 A good example of a patentable idea that resulted from
observation happened to a colleague andme some years
ago. The structure being analyzed was testing gate oxide
breakdowns. Several of the failed samples were depro-
cessed to expose the poly layer, where silicon-selective
etching with choline hydroxide was used to uncover the
gate oxide. The etch removed the poly layer and, because
there was a hole in the gate oxide, it also etched a small
amount of the silicon substrate at the breakdown point.
The location, size, and pattern of the breakdown can nor-
mally be seen through lowSEMvoltage settings. On these
particular samples, a large quantity of particles propa-
gated out of the holes after etching. At first we assumed
there was contamination in the oxide and these particles
were the root cause of the problem. Energy-dispersive
spectroscopy revealed that the particles were comprised
ofmetallic elements—badnews for the gate oxideprocess.
However, the particles were odd-looking, as if they were
grownone on topof the other, and theywere large enough
that they should have been seen optically at the poly
stage. So, what was their origin?
A close examination of the sample-preparation
process found that metal tweezers were used while
etching the poly, and they were the source of the par-
ticles. That finding began an investigation into why the
particles were attracted to the holes in the gate oxide.
It was discovered that when the semiconductor device
was optically illuminated, a p/n junction was formed as a
result of photovoltaic phenomena. The metal ions from
the etch solution, where the metallic tweezers had been
used, migrated to the oxide hole area due to the voltage
pull created at the p/n junction, leading to the buildup of
ions. What a great way to tell if there’s metal contamina-
tion in a solution, but who besides failure analysts has
access to an SEM to look at the sample after exposure?
We needed a device andmethod that was easy to use and
practical for trapping metal ions and detecting the metal
ion contamination in a solution.
A very simple circuit designwas drawn upwith electri-
cal leads connected to two small exposed p/n junctions in
close proximity. After the circuit is exposed to the solution,
the resistance can be measured between the junctions.
Depending on the space between the two junctions and
the amount of time exposed to the solution, the quantity
of metal ion can be determined. Being failure analysts,
we didn’t have the resources to create a wafer with these
exposed p/n junctions, and therefore, we didn’t have an
operational device, but that didn’t stop us from filing for
a patent. The theory behind the phenomenon was solid,
and the findings were reproducible. Illustrations of our
circuit were drawn up, a brief description of the process
andwhat itwouldaccomplishwaswritten, andafilingwas
submitted.[4] It was literally that easy. Of course, from that
point the lawyer or patent attorney took over and added
any other variations that could be incorporated into the
invention, as well as the appropriate wording. Before the
patent was finally awarded, it took several years to go
through the entire process of searches to ensure that the
idea did not previously exist. It was worth the wait.
Now, you’re probably thinking, “Wait a minute. This
patent doesn’t have anything to dowith the foundry busi-
ness.” That is true; we most likely will never manufacture
this invention, but that doesn’t mean the patent has no
value for the company. Patents can be licensed or sold to
other companies. Through that possibility, we may end
up making those circuits for another company, thereby
providing profitability for them and for us. So, all those
ideas for improving the analytical side for lab efficiency
or capability may still be worth the pursuit.
REFERENCES
1. N. Wingfield: “Jury Awards $1 Billion to Apple in Samsung Patent
Case,” The New York Times online ed., Aug. 24, 2012,
2. American Intellectual Property Law Association: “Summary of the
America Invents Act,”
3. G. Quinn: “A Brave New Patent World—First to File Becomes Law,”
IP Watchdog, March 13, 2013,
4. R.-L. Chiu and J. Higgins: Solid State Sensor for Metal Ion
Detection and Trapping in Solution, U.S. Patent 9,000,783, Apr.
7, 2015,
ABOUT THE AUTHOR
JasonHiggins graduated fromITT
Technical Institute with a Bachelor’s
degree in automation robotics. He
joined TSMC WaferTech in 1997 as
a reliability FA engineer and lab
supervisor. Jason specializes in fault
isolation and physical failure analysis
on low-yield andfield-return samples
from a variety of product functions and technologies. He
has been an Associate Editor of EDFAmagazine since 2008.