JED - April 2016 - 26

The Journal of Electronic Defense | April 2016


circuits that benefit from the tight integration of the NEXT III-V technologies
and silicon CMOS.
One example of how the process
works can be seen in the design and realization of a novel arbitrary waveform
generator that incorporates new heterogeneous circuit technologies. Circuits
include an analog-to-digital converter
and digital-to-analog converter (DAC)
that incorporates InP advancements
from DARPA's Terahertz Electronics program. Although InP doesn't offer the
power density of GaN, it can provide major advantages in high-frequency performance, making it another valuable
tool in a designer's toolbox.
The heterogeneous devices enable
higher switching frequencies and higher
broadband dynamic range on the output,
particularly for the DAC. "This demonstration is also interesting," says Green,
"because it shows that, with the DAHI
program, we can either take older nodes
and make them more capable, meaning we
don't have to always have cutting-edge
silicon CMOS to get cutting-edge capability, but that we can also take the leading
edge of silicon CMOS and make it better
with the integration. This is part of the
reason for DAHI, to demonstrate how it
can enable us to use commercial technologies and make them better to provide differential capabilities for the DOD."
Green describes DAHI as a broad program, aimed at "setting up an ecosystem,
not just a technology demonstrator."
Currently, Northrop Grumman (Redondo
Beach, CA) is providing a foundry "access point" that leverages both its own
work in InP technology, as well as sourcing silicon CMOS from Global Foundries
(East Fishkill, NY), a GaN variant from
HRL Laboratories (Malibu, CA), and
high-Q passive filters from Nuvotronics
(Durham, NC) which incorporate their
PolyStrata® fabrication technique for
creating 3-D waveguide structures that
are optimized for high frequency.
The materials are received at the
Northrop foundry where they are processed for the integration and coupling
to the III-V technologies that have been
made in parallel. Explains Green, "It's a
chiplet style integration that has proven to be valuable from a yield perspective because it enables you to decouple

the yield of any single technology to
some extent."
Northrop Grumman provides an integrated foundry design kit that allows
designers to design into these different technologies seamlessly. They then
perform the mask aggregation and the
backend integration work to deliver
chips to the various performers. In addition to their own work, Northrop is
working with multiple primes on the
program with different design teams
working to realize circuits using the
process including BAE Systems, Rockwell Collins and Teledyne.

"To give a physical
sense of the scaling,
the devices made on
the NEXT program
would fit inside the
gate of the devices
that were made on the
earlier [Wide Bandgap
- Dr. Daniel Green
Green says the program is now midstream, just yielding its second batch of
wafers. "We're seeing good yield so far
on limited devices, and we're now getting into more complex circuits." Green
expects to see further positive results
later this year, "showing that we can not
only get better performance in proof-ofdemonstration circuits, but also start to
yield useful circuits that will start to
flow into different systems."

A common challenge across all of
DARPA's efforts aimed at pushing the
limits of GaN and other materials is
that, along with higher power density,
comes a greater thermal management
problem that must also be addressed.
DARPA's Intrachip/Interchip Enhanced
Cooling (ICECool) program is centered
on attacking the problem at its source
by integrating cooling techniques into

the chip itself. The approach involves
creating microchannels to flow cooling fluid directly under the junctions
of devices. As described by Green, "It's
an aggressive approach, and you have to
create this plumbing on the underside
of the die, but there have been excellent results showing that you can dramatically reduce the thermal resistance
of the chips, buying you both improved
reliability and performance."
In other initiatives, DARPA's Semiconductor Technology Advanced Research Network (STARnet) program,
launched in 2013, is a joint effort with
the semiconductor community sponsoring fundamental research at universities. Currently, the effort includes
approximately 170 researchers organized
into six different research centers. Each
of these research centers focuses on a
specific challenge, such as nonconventional materials and devices incorporating nanostructures with quantum-level
properties, to enable analog, logic and
memory devices for beyond-binary computation. "There are some interesting
avenues emerging from the STARnet
program," says Green, "as well as some
of our own 6.1-oriented research [Basic
Research] efforts that we are considering as a path forward going from here."
In addition to providing revolutionary advances in materials technology,
DARPA must also address the practical
aspects of the challenge, such as making
the technology affordable, useable and
accessible. The Common Heterogeneous
Integration and IP reuse (CHIPS) effort is
directed toward developing strategies to
accomplish this. Says Green, "The idea is
understanding how to take advantage of
a more modular design flow, building on
the theme of leveraging all the technology capabilities, but enabling them to be
used more broadly across multiple applications so you don't have to embark on a
new development cycle every time."
Ultimately, Green says much of their
recent work has been aimed at unlocking the latent capability of GaN. "It still
has a lot of room to run, and there is a
lot of unlocked potential there." Looking forward, he observes that "there are
also other materials that are coming to
the fore that will present opportunities
in the not too distant future." a


JED - April 2016

Table of Contents for the Digital Edition of JED - April 2016

The View From Here
Conferences Calendar
Courses Calendar
From the President
The Monitor
World Report
DARPA Profile: Advanced Materials
Mission Profile: Non-Kinetic Thinking Creates New Possibilities for Air and Missile Defense
Threat Monitor
EW 101
AOC News
Index of Advertisers
JED Quick Look
JED - April 2016 - cover1
JED - April 2016 - cover2
JED - April 2016 - 3
JED - April 2016 - 4
JED - April 2016 - 5
JED - April 2016 - The View From Here
JED - April 2016 - 7
JED - April 2016 - Conferences Calendar
JED - April 2016 - 9
JED - April 2016 - Courses Calendar
JED - April 2016 - 11
JED - April 2016 - From the President
JED - April 2016 - 13
JED - April 2016 - 14
JED - April 2016 - The Monitor
JED - April 2016 - 16
JED - April 2016 - 17
JED - April 2016 - 18
JED - April 2016 - 19
JED - April 2016 - 20
JED - April 2016 - 21
JED - April 2016 - 22
JED - April 2016 - World Report
JED - April 2016 - DARPA Profile: Advanced Materials
JED - April 2016 - 25
JED - April 2016 - 26
JED - April 2016 - 27
JED - April 2016 - Mission Profile: Non-Kinetic Thinking Creates New Possibilities for Air and Missile Defense
JED - April 2016 - 29
JED - April 2016 - 30
JED - April 2016 - 31
JED - April 2016 - 32
JED - April 2016 - 33
JED - April 2016 - 34
JED - April 2016 - Threat Monitor
JED - April 2016 - 36
JED - April 2016 - EW 101
JED - April 2016 - 38
JED - April 2016 - AOC News
JED - April 2016 - 40
JED - April 2016 - Index of Advertisers
JED - April 2016 - JED Quick Look
JED - April 2016 - cover3
JED - April 2016 - cover4