Considering Beyond-CMOS Metrology

Metrology has become one
of the main pillars upon which the semiconductor industry bases its
progress. Uncertainty rarely transmutes science into technological
progress for the simple reason that if something can’t be
measured and quantified, it becomes very difficult to direct or
control it. The ITRS has done much to define what the needs will be
for each node, alerting academia, industry, and others in the
metrology community about what the requirements will be and which
are the necessary technologies to be developed.

Now, as the semiconductor
industry apparently inevitably progresses toward a post-CMOS future
that lies in an increasingly steeper nanoscale regime, metrology
must also gird itself in preparation for the beyond-CMOS world,
which will include new materials being used now and others as yet
unthought of, as well as structures and devices that will require
manufacturing processes as beyond ours as ours are to cutting
rubylith.

At the last SEMICON West,
my attention was caught by a presentation authored by Alain
Diebold, Empire Innovation Professor of Nanoscale Science at
the College of Nanoscale Science and
Engineering at the University at Albany
(New York).The presentation was done at the
ITRS Public review and is the report of the Metrology Technical
Working Group. It dealt with “Beyond-CMOS Metrology,”
and during it, he called attention to the fact that if we are to go into the exotic
jungles of molecular electronics, ferromagnetic logic, and spin
logic devices, we are going to have to be able to image what we are
doing.

In the case of molecular
electronics, we are—and will be—dealing with soft
structures made out of carbon, which can be easily damaged by
whatever type of microscope is used. With inorganic materials, of
which ferromagnetic and spin logic devices would be composed, this
particular problem doesn’t exist, but some form of spin
microscopy will have to be developed. Things might get somewhat
more complicated, in that some of these future devices could
contain graphene, again making them vulnerable to some forms of
metrology. Carbon nanofilms are very difficult to do microscopy on,
yet some kind of high-resolution microscopy will be essential.

Simulations of TEM
imaging stacked layers of graphene are used to understand
optimum experimental conditions.  This figure shows the atomic
configuration of
A A A stacking of graphene layers used for image
simulation. 
Source: Florence
Nelson, College of Nanoscale Science and
Engineering.

In addition to just microscopy needs,
there are measurement requirements unique to nanoscale materials,
such as the need to define the quantum confinement affecting them.
Are you trying to make a device based on excitons? Then you will
have to measure some aspects and qualities of
the excitons. There will also be Berry phase effects, such as those
observed during carrier transport measurements of graphene by
Philip Kim group at Columbia University (New York). In the
area of spin, some breakthrough work has been done by Vince LaBella
also of the College of Nanoscale Science and Engineering and
others, using beam ballistic emission electron microscopy,
to obtain information about this state.

Another factor to consider is
that as we dive deeper into nanoscale dimensions, band structure
can change optical properties. If one is to do an optical
measurement, all of these variables must be understood.

New methodologies, such as a novel use of a single-electron
transistor to map out where electrons and holes are in graphene are
being explored in Amir Yacoby’s group at Harvard University
(Cambridge, Mass.). These form
“puddles,” a number of holes where there are
electrons—carriers puddling together. This is a spectacular example of something
predicted in theory that is being measured using a new
technique.

The materials area offers
innumerable challenges. For example, we will need to find defects
in them. Part of this involves learning exactly what effect
geometry can have on these new materials; such as how the edge of a
graphene structure impacts its band structure. We will have to
discover how to look for contamination in these futuristic
materials, as well as understand a nanoscopic measurement when it
becomes mesoscopic; that is, going from the very localized to the
more wide-ranging. Macroscale devices will require mapping in a
wider area than has been necessary when we measure structures in
very localized places.

These are the things that keep
the metrology community awake at night. We need advances in microscopy and
metrology for things like optical measurements, local spin,
and new things like electron hole puddles.

We already have very
sophisticated systems, like transmission electron
microscopy—aberration-corrected TEMs. However, getting that
same resolution at very low operating voltages to avoid damaging
materials like carbon films is not easy. There will be a growing
need to link the modeling and simulation; everything from a
theoretical calculation of band structure in a nanowire to optical
measurements that will be done on an assembly of nanowires. Both
extension and invention will be needed.

Meanwhile, the ongoing
collaboration between national labs, industry, and academia must
continue, with increased focus on the characteristics and effects
of new materials’ applications to future devices. These
challenges are not going to be met in just a couple of years; they
will require a sustained effort.