IEEE Solid-State Circuits Magazine - Spring 2015 - 71

Semiconductor's Approach in mmWave

Propagation Loss (dB)

lagging behind by about two years.
For example, 45-nm technology was
adopted two years later for mobile
devices than for PCs. But for 14 nm, we
commercialized 14-nm FinFET APs just
three months later than for PCs. It will
be no surprise to see 10-nm technology
adopted in mobile applications first.
However, to continue providing
value to consumers, every technology
has to push its own inherent limitations, be it analog or digital, micronsize or nanosize. We are pushing the
limits of silicon processing technology. And we are implementing novel
3D structures, such as vertical NAND
and TSV device stacking.

80
40

Signal Attenuation of mmWave
Theoretical

Requires Massive
Parallel Process

20 dB 30 GHz
3 GHz
30 GHz

0
1
2
3
8 X 8 Array Antenna
Distance Between Tx and Rx (m)

Modem

Modem Technologies for High-Speed Data Communication
S333 LTE-A Modem

100 Mb/s
32 nm

450 Mb/s
28 nm

600 Mb/s
14 nm

1 Gb/s
10 nm

mmWave for 5G
At the same time, we are venturing into never explored areas where
silicon technology could enable new
technologies and applications, such
as mmWave. I would like to elaborate
a little bit more on mmWave for 5G,
which is the radio frequency band
between 30 GHz and 300 GHz. Unlike
the current cellular frequency, which
is depleting fast, mmWave is a virgin
territory. The disadvantage is that,
theoretically, signal attenuation is
more severe at the higher frequency
spectrum. In fact, we can see almost
20 dB more loss at 30 GHz as compared to that of 3 GHz. By employing directive 8 x 8 antenna arrays at
30 GHz, we can reduce the losses to
below the theoretical limit to be even
better than 3 GHz. This innovation in
analog will be well supported by silicon technology. In fact, such a large
array requires massive parallel processing capabilities that should fit in
a smartphone, which the 10-nm technology would be ready to provide in
time for 5G deployment.
There are some technologies that
will not benefit from silicon technology
as they will soon meet their end, such
as the long standing HDD technology. I
would not be surprised to see that SDD
will almost completely replace HDDs
within the next ten years.

Security at the Chip Level
The dark side of the data-driven
world and of having so much data on

Directive 8 x 8 antenna arrays at 30-GHz array require massive parallel processing capabilities that should fit in a smartphone.

Security in Silicon
Secure Processing

Secure Storage

Secure
Processor
PUF
"Unclonable"

Secure
Memory
Secure
USIM
Secure
NFC

Secure
Communication

Samsun
g

NFC

Secure
Solid-State Drive

Secure
Sensor
PUF: Physically Unclonable Function

To protect every aspect of the data path, such as creation, processing, transmission, and
storing, each small chip will have to be equipped with the right security solution.

everything is that private and confidential information is exposed to
the risks of unauthorized access.
In most cases, consumers may not
even realize that there is a leakage of
information. The ultimate goal of security is to allow users to fully enjoy
the data service without any threat
of data leakage. We need to protect
every aspect of the data path, such

as creation, processing, transmission, and storing. Each small chip
will have to be equipped with the
right security solution. We have been
able to include enhanced security in
our chips through our physically unclonable function (PUF) technology.
PUF makes use of the fact that
each Si chip has a distinctive process
signature, which makes it uniquely

IEEE SOLID-STATE CIRCUITS MAGAZINE

s p r i n g 2 0 15

Table of Contents for the Digital Edition of IEEE Solid-State Circuits Magazine - Spring 2015