Antennas & RF Sensors: Changing the Way We Live

(from mobile telephony to electronic textiles and RFIDs)

John L. Volakis
Th Ohi St t U i it The Ohio State University
volakis@ece.osu.edu
Coupled
Uncoupled Uncoupled
Coupled
Uncoupled Uncoupled
L1
C1
L3 LM
C3
CM
(V
1
I
1
)
(V
3
I
3
)
Coupled
Uncoupled Uncoupled
Coupled
Uncoupled Uncoupled
L1
C1
L3 LM
C3
CM
(V
1
I
1
)
(V
3
I
3
)
c
0
L2
C2
0
L3 LM
0
C3
0
CM
LM
(V
2
I
2
) (V
4
I
4
)
Inductive
coupling
Capacitive
coupling
0
L2
C2
0
L3 LM
0
C3
0
CM
LM
(V
2
I
2
) (V
4
I
4
)
Inductive
coupling
Capacitive
coupling
1
Why so much interest in Antennas and RF
Devices?
Miniaturization need for sensors
and mobile systems
Compact communication systems
with higher data rates
More than 50% of a system-on-a-chip
consists of passive RF devices
RF & Mi crowave
Magazi ne 2006
Conformal antennas for commercial and
military applications
Multifunctionality to
decrease complexity & cost
I EEE Spectrum
2
Wireless World
3
4
Enabling RF Technology for Daily Life Convenience
60 trillion wireless sensor till 2010-12, viz. 10,000 for every person in the world
Wearable Antennas and Sensors are being
developed with conductive textiles to:
Monitor Health (blood pressure, fluid levels etc.)
P iti D t ti
Engineered Substrates for:
Miniaturization,
Multifunctionality & Greater Bandwidth
Position Detection
Decrease Weight and Complexity of Existing
Wireless Devices
System design to communicate various
h h
Challenges:
Convenient way of Substrate Patterning & Printing
(Metamaterials) is not yet realizable
Integration of RF circuits is not realistic in existing
RF components to each other
Integration of RF circuits is not realistic in existing
technology (only in low constant materials)
From syringe
Coated matrix
template
Repeat printing process
on top of solid sheets
Health Monitoring
Challenges:
Incorporate various RF technology
(Bluetooth, GPS, Health Monitoring)
into existing devices
- Small RF devices Challenges:
- Tissue Characterization
- Understand effect of RF
signals on human body
- EM Compatibility with
biological structure
Small RF devices
- High gain antennas for far distance
communication (GPS)
5
Sensor Systems/System on a Chip Overview
P D h i l d Purpose: Detect chemical agents and
vapor mixed in the air at low levels.
Deployment: Sensor can be placed in the Deployment: Sensor can be placed in the
air-ducts of air conditioning systems.
Size: 2-3cm per side. Fairly flat
6
RFID Everywhere
Automobile industry
A i ti
Electronic Toll Collection
Access Control
Animal Tracking
Automobile industry
Aviation
Animal Tracking
Inventory Control
Tracking Runners in
Races!
2000 RFID points
Anti-Theft
Pharmaceuticals
ID Cards ID Cards
7
Ohio States RFID Bug Tracking Radar Goes Up to 190ft (63m)
RF Front End
1. 5.9 GHz CW
signal transmitted
2. Signal intercepted by
tag (up to 200 away),
harmonic re-radiated
Xmit: 5.9 GHz
3. 11.8 GHz
signal received,
down-converted
~200
Transmit/Mixing Circuits
Proposed Array:
Rec: 11.8 GHz
4. IF Signal: Indicates
presence of tag
Rec. Circuit
Proposed Array:
-Xmit ~ 14 x 7
-Rec ~ 7 x 3.5
-Gain ~ 20 dBi
Meandered Loop Tag
Tag measures
9.5 mm x 9.5
mm mm.
9.5 mm ~ 0.19
at 5.91 GHz
(Xmit freq.)
8
Polyphase Multi-band Predistortion Linearization to
Suppress Intermodulation Interference
9
RFID Reader System Remains a Challenge
- RFID readers and tagging promise to revolutionize retail, gg g p
supermarket and warehouse shelving systems, including
purchasing process and habits
Fading is a major issue. Spatial and
R
e
R t il Sh l
Two distributed
antennas mounted
vertically behind
t d d t il h lf
Fading is a major issue. Spatial and
polarization diversity of the distributed
antennas overcome fading in both static
and dynamic environments.
e
a
d
e
r
Retail Shelves
RFID Portal
standard retail shelf
Various metallic and
non-metallic RFID
tagged items arranged
on shelves
E
l
o
n
g
a
t
e
d
t
e
n
n
a
Loaded
pallet
d

R
F
I
D

c
o
v
e
r
a
g
r
i
b
u
t
e
d

A
n
t
g
e

a
r
e
a
D
i
s
t
r
Reader
Two distributed reader
antennas on each sidewall
10
Container Product Tracking Systems
For a typical container 10x10x40 with tightly stacked cartons, we need to use 8
distributed antennas to illuminate entire container. Folded tags are also needed to
Container Side View
distributed antennas to illuminate entire container. Folded tags are also needed to
get good performance.
10
40 40
Folded tag deployment
Antenna Deployment (Top View)
6 4 6 4
6 4 6 2
Folded tag deployment
2
6
We place 2 antennas to cover a 10x10x10
volume of the container. For the entire
container, we need 8 distributed antennas
11
Real Time THz Imaging of Excised Tissue
TARGET APPLICATION:
Breast Cancer Detection
1 in 8 women will have
breast cancer in their lifetime
Digital
Processing
THz Imaging System Overview
Ti S l breast cancer in their lifetime
96% curable if caught early
-----------------
American Cancer Society, Inc,
Tissue Sample
Surveillance Research
Current Breast Cancer
THz
Detector Array
Object Lens
THz
Image
Working Principles
Detection Issues:
X-rays not of sufficient
resolution/information; also
not desirable due to inherent
Direct detection of THz radiation using diodes having
superior noise performance
Build compact detector layouts suitable for 2D arrays
not desirable due to inherent
ionization.
80% of biopsies
unnecessary
E i d ti i i i
Explore/optimize array topologies for maximum
number of detectors (equivalently pixels), power
transfer and detection performance
Investigate suitable imaging algorithms to be
12
Excised tissue imaging is
desirable during operation
but very slow
Investigate suitable imaging algorithms to be
incorporated into the imager signal back-end
Frequencies: 100GHz, 500GHz, 800GHz
Body Worn Antenna Coverage with Diversity Module
20dB 20dB 30dB fluctuation reduces to 3dB 30dB fluctuation reduces to 3dB 4dB 4dB
d
B
m
]
Antenna 3 Antenna 4
v
e
d

p
o
w
e
r

[
Antenna 1
m
a
l
i
z
e
d

r
e
c
e
i
v
Ant1- front torso
Ant2- back torso
N
o
r
mAnt3- R. shoulder
Ant4- L. shoulder
Module output
iWAT2009 13
Azimuth angle [deg]
Rotator
Miniature Antennas and Arrays for THz Imaging
Multiband Antennas:
Detection at multiple frequencies
Smaller size antennas to avoid aberrations
1um 3um
94um
1um 3um
94um
Antennas with controlled radiation
patterns
More antenna elements (pixels)
580um 580um
Metamaterial
50um
1
1
1
u
m
18um
11.4
1.8 0.6
3um
1.8um
7
5
u
m
3um
1.8
50um
1
1
1
u
m
18um
11.4
1.8 0.6
3um
1.8um
7
5
u
m
3um
1.8
-15
-10
-5
0
dB(S
11
)
Parasitically
Loaded
280um
420um
280um
420um
Asymmetrically
Fed
Much smaller
Metamaterial
miniature antennas
for small pixel size
300 400 500 600 700 800 900 1000
-30
-25
-20
Frequency(GHz)
Dual band folded slot antenna for 500 &
500GHz
800GHz
Loaded Fed
metamatetial-based
multiple antennas for
pattern control
Dual band folded slot antenna for 500 &
800GHz breast cancer detection
Imaging arrays
with multiple
detector (antenna
Photonic Crystals &
Metamaterials
array) layouts for
pattern diversity
Metamaterials
High gain apertures and miniature
arrays
Negative refraction for sub-
wavelength focusing
14
wavelength focusing
Wireless Landscape
1000
?
5000
2005 Wireless Landscape
1000
??
5000
2005 Wireless Landscape
500
s
)
?
100
Continually Evolving
802.15.3a
UWB
500
s
)
??
100
Continually Evolving
802.15.3a
UWB
Future wireless systems will focus on high
10
a
t
e

(
M
b
p
s
50
802.16
802 15 3
802.11a/g
Fixed
UWB
802.11n
10
a
t
e

(
M
b
p
s
50
802.16
802 15 3
802.11a/g
Fixed
UWB
802.11n
y g
data rates to deliver content as rich as
that in current wired systems
i
s
s
i
o
n

r
a
UMTS / HSDPA /
802.11b
WiFi 1
5
Canopy
WiMax
802.15.3
Bluetooth
Mobile/Nomadic
802.16e / 20
i
s
s
i
o
n

r
a
UMTS / HSDPA /
802.11b
WiFi 1
5
Canopy
WiMax
802.15.3
Bluetooth
Mobile/Nomadic
802.16e / 20
0.1
0.5
T
r
a
n
s
m
Blackberry
GSM/TDMA
GPRS/EDGE
UMTS / HSDPA /
1xEVDO
ZigBee
802.15.4
Bluetooth
802.15.1
0.1
0.5
T
r
a
n
s
m
Blackberry
GSM/TDMA
GPRS/EDGE
UMTS / HSDPA /
1xEVDO
ZigBee
802.15.4
Bluetooth
802.15.1
Range (meters)
10 100 1000 10 000
100 000
y
1
Range (meters)
10 100 1000 10 000
100 000
y
1
15
Antenna/RF Sensor Challenges and Solution Trends
Applications
Solution Trends
pp
Multifunctional Radios
Software Radios/Radars
Solution Trends
Materials
Magnetics films/Multiferroics
Short Range Giga-bit Comm.
& Data transfer
UWB i i f MH t
g
Composites (polymers,
mixed material systems,
multi layered emulated
UWB imaging from MHz to
THz.
Sensing and monitoring
multi-layered, emulated
anisotropy)
High speed interconnects
g g
Identification
Body worn systems
g p
Structurally reinforced smart
skins (carbon nanotubes,
i )
THz imaging
ceramics )
ferroelectric
ferromagnetic
16
Military Challenges Military Challenges
OSD report (Appendix C, p.9): Over the long term
2012-2030, JTRS, including wideband networking 2012 2030, JTRS, including wideband networking
waveforms (WNM), will provide a fully integrated
information system network to include active and
passive information operation management across
the joint and combined environment. The system will
also include a self-establishing and self healing smart also include a self establishing and self healing smart
network, which will automatically manage the RF
domain.
WNM will operate from 225-400MHz, and is
intended to enhance or replace current intended to enhance or replace current
systems such as the Single Channel Ground
and Airborne Radio System (SINGARS, 30-
88MHz), EPLRS (450-470 MHz), and Link 16.
Gi th t t i l l th i Given that typical wavelength size
antennas are /2 long/wide (5m or
~15ft at 30MHz), they must not only be
structurally embedded but also much structurally embedded, but also much
smaller (miniature) than currently
available to accommodate future real
estate.
17
Future Smaller Aircraft Needs
Challenges
S ll i UAV d RF ti ( i ti i i Small size UAVs and RF congestion (communications, imaging
and information gathering on small platforms)
high speed communications: from 15kb/sec to
5Mbits/sec, 3000 times higher speeds g p
30-5000MHz ultrawide band antennas (with wavelength
5m or 15ft) must be reduced in size by a factor of ~5-10
and integrated on small platforms with gains for high
data rates data rates
High gain/narrow-band antennas for radar, imaging,
guidance, satellite communication, and other specific
applications
t Gl b l H k S tC t i till ---current Global Hawk SatCom antenna is still a
(rotating) reflectors,
---future apertures must be conformal and suitable even
for small UAVs
Entire UAV surface may serve as a RF (smart) skin, implying structural
integrity of the RF functional layers
Volumetric exploitation to address low frequency performance is a
major design and computational challenge.
Multiphysics tools must address concurrent structural/fatigue and RF
performance requirement; need for multiphysics design optimization
tools
UAV sections may be removable and re-bonded for multiple missions
18
Traditional antennas will be a thing of the past!!
Artificial or Real Artificial or Real
Materials?
19
Antenna Miniaturization Approaches Antenna Miniaturization Approaches
Increase Antenna Electrical Size
Slow down currents flowing on antenna,
using inductive and capacitive loading :
L
1
eff

2 2
eff
E+ = 0 =
Load antenna near fields
with material
r
,
r
:
eff
= , =
g
C
L C
eff
eff eff
E+ = 0, =
w
eff eff

eff
Increase Radiation, Bandwidth and Efficiency
Compared to un-miniaturized Antennas
c ,
Artificial Transmission Line (ATL)
Material Loading
20
Matching Networks Significantly
Improves Low Frequency Performance
0
10
Final Operating
Frequency 178
MH (60% t t l
-20
-10
5 segments
Possible matching Circuit
from Optimization
A d 80% Mi i i ti
15
MHz (60% total
minimization)
-40
-30
Non Optimized
Case A(Best)
Case A With Matching
2
0
2
from Optimization
Around 80% Minimization
0.2 0.4 0.6 0.8 1
40
-6
-4
-2
-12
-10
-8
Non Optimized
Case A(Best)
Final Operating Frequency
can go down to 175MHz
GAcase13.02d.FlarDipole
Run.2.case.1.iter.1.gen.8.run.44
0.15 0.2 0.25 0.3
-14
Case A With Matching
21
Array plane
Munks Current Sheet Antenna
superstrate
Z
1
+
=jZ
o
tan(kd
o
)
d
Array Geometry
d
o
Ground Plane
Impedance looking into the
ground plane is inductive
Capacitive load
(realized via the
i t di it l interdigital
capacitor fingers)
to cancel
inductive ground g
plane effect
Interdigital capacitive loads; As the frequency increases Interdigital capacitive loads; As the frequency increases,
capacitance increases to cancel inductive load caused by the
ground plane. This results in nearly purely resistive impedance,
giving rise to the large bandwidth
22
2-18GHz Performance
F B d 2 18 GH
30
30
8x8 Active Array
Broadside Gain
30
30
8x8 Active Array
Broadside Gain
Frequency Band: 2-18 GHz
Polarization: Dual Linear
Overall Size: 22 x22
Total Elements: 2664 Dual Pol
Active Elements: 64 Dual Pol
20
25
25
20
i
)
20
25
25
20
i
)
Active Elements: 64 Dual Pol
Aperture Thickness: 0.8
5
10
15
G
a
i
n

(
d
B
i
)
15
10
5
a
i
n

(
d
B
i
5
10
15
G
a
i
n

(
d
B
i
)
15
10
5
a
i
n

(
d
B
i
-5
0
5
G
CSA11 Measured w/ coupling plates
Theoretical 8X8 Unit Cell Gain
5
0
-5
G
a
-5
0
5
G
CSA11 Measured w/ coupling plates
Theoretical 8X8 Unit Cell Gain
5
0
-5
G
a
0 4 W/element
-10
2 4 6 8 10 12 14 16 18
Frequency (GHz)
2 4 6 8 10 12 14 16 18
Frequency (GHz)
-10
-10
2 4 6 8 10 12 14 16 18
Frequency (GHz)
2 4 6 8 10 12 14 16 18
Frequency (GHz)
-10
0.4 W/element
23
Frequency (GHz) Frequency (GHz)
23
Enabling Miniaturization of Broadband Apertures
Just another way of
capacitive coupling
Munks/Harris Corp. Current Sheet Aperture (CSA) Munk s/Harris Corp. Current Sheet Aperture (CSA)
Matching with Negative Elements
0 F
77pH
-1.45pF
-256nH Antenna final size: /12x/18
3.5:1 bandwidth
No use of any material loading
3
Realized Gain before
e
n
n
a
-128pF
-18.5H
0pF
-77pH
Input
-3
0
3
240 to 1050MHz
(4 5:1)
A
n
t
e
p
Values from
optimization
200 300 400 500 600 700 800 900 1000 1100
(4.5:1)
Keep (almost)
th
Decrease the
t i b
-1000
0
500
Reactance
0
3
the same
bandwidth
antenna size by
3
50 200 400 600 800 1000 1200
-2000
75
150
25
50 75 100 125 150 175 200 225 250 275 300
-3
82 to 275MHz
(3.5:1)
50 100 150 200 250 300
-150
-75
0
75
After
25
benefits for magnetic materials are there
0
5
10
-15
-10
-5
t
a
l

G
a
i
n

(
d
B
)
M3 ferrite
-25
-20
-15
R
e
a
l
i
z
e
d

T
o
t
free s t andi ng
over i nfi ni t e ground pl ane, h1 = 3''
M3-ferrite
cylindrical
cavity
0 50 100 150 200 250 300
-40
-35
-30

i nfi ni t e ground pl ane, w1= 3'', w2= 4'', h1 = 3''
i nfi ni t e ground pl ane, w1= 1'', w2= 1'', h1 = 3''
i nfi ni t e ground pl ane, w1= 1'', w2= 1'', h1 = 4''
i nfi ni t e ground pl ane, w1= 1'', w2= 1'', h1 = 5''
0 5 0 50 100 150 200 250 300
F requenc y (M Hz )
0.5
26
Limitation of current magnetic materials & losses
Existing materials turn out to be
very lossy at UHF frequencies, for
Material losses
cause gain
reduction
free
antenna applications
Losses cancel out the
miniaturization obtained by the
free
space
miniaturization obtained by the
ferrites
Material losses need to be lower
than 0 005 in order for the ferrite
tan must be <0.005
Material specifications
than 0.005 in order for the ferrite
loading to be beneficial
Gain reduction
appears as shifting
up in frequency
p

r
=
r
(impedance matching)
losses<0.005
(<500 MHz)
values : 8 14
f : -15dB frequency, losses
f
o
: -15dB frequency, lossless

r
,
r
values : 8 - 14
27
Antenna Miniaturization Techniques
1 Slow down current flow by 1. Slow down current flow by
material loading
Homogeneous material (broadband)
Inhomogeneous, anisotropic material
k
| c
c
=
k
c
v
g
g p
(narrowband)
Magnetic Material (narrowband, bias
required)
2. Control current phase velocity by
e
|
k
c
v
p
=
2. Control current phase velocity by
controlling L and C of equivalent
transmission line model
k
L
L
C C C
e
L L
C C C
v
1
=
eff
L
Z =
28
eff eff
p
C L
v
eff
c
C
Z
Temex Microwave Ferrite Materials Catalog,
http://www.temex.net/temex/catalog/TEM01.pdf
28
Eploiting 3D Anisotropy
3D anisotropy modifies the k-e diagram of the material medium
Concurrent slow wave velocity & impedance matching lead to
small RF devices with high sensitivity and high gain vs bandwidth.
e
| c
c
=
k
c
v
g
k
v
Material DNA
RBE mode
Frozen Modes
DBE mode
e
k
|
k
c
v
p
=
Flat sections, imply slow velocity
DBE mode
MPC mode
k
math concepts to RF designs and actual antenna realizations
Flat sections, imply slow velocity
10
60
90
120

0
5
30
210
240 300
150
330
180 0
DDM
Measurement
HFSS
x
240
270
300
Concept
y
12dB vs. 1.7dB Simulations
Fabrication
Testing
100% aperture
efficiency
29
Realization and Demonstration of Degenerate
BandEdge (DBE) Mode in a Periodic Assembly
3
4
5
6
E
|
m
a
x
x
z
y
PEC
5-folds amplification
12
15

3 2 1 0
0
1
2
Layer #
|
Field profile for reception
y
i
E
=52.5
o
3
6
9
|
H
|
/
|
H
i
n
c
|
Field profile for reception
3 Layers=0.96cm
0.28
0
5 7.5 10 12.5 15 17.5 20
0
3
Frequency(GHz)

New FEM Domain
Decomposition Tools
required (for 16M DOFs):
1 4GB 30GB
Feeding Schemes: Slot on ground plane, dipole in B
1
layers
Easier to implement, external feed/matching
---1.4GB vs 30GB
---CPU: 2.5 hrs
5
10
60
90
120

Simulated (DDM)
Measured (OSU-ESL)
Measured (NASA)
0
30 150
180 0
Directivity:10.15dB, over 100% aperture efficiency
210
240
270
300
330
30
Performance of the Degenerate BandEdge -DRA Antenna
10
4
t
h
Optimal
Curve
Leaky Wave
DRA
10
2
n

x

B
a
n
d
w
i
d
t
y
Region
Region
2
10
0
G
a
i
n
0 5 10 15 20 25 30
10
-2
ka

o
/8.12
o
/8.12

o
/18
45
o
DBE Antenna
Prototype
Hybrid DBE Mode
DRA Antenna 31
Emulation of Anisotropy using Standard Printed Lines
32
Printed Lines with Lumped Elements allow for much Greater
Control of the Dispersion Diagram and Propagation Properties
Forbidden Frequency Band
P


Propagation Band Structure
Flatter higher Q/gain (slow velocity)
Inflection point high Q, greater
bandwidth (slow velocity, excellent
t hi t di l t i i t f
q
u
e
n
c
y
q y
SIP
k=t points








e
R
B
E



e
S
I
P
matching at dielectric interface zero
acceleration)
kd = 0 lower frequency resonance
(small size)
kd t d d DBE/
F
r
e
q
0
2





























kd = t standard DBE/resonance
Realization:
4-port circuit model
0 t 2t
e
0
1



e
0
Propagation Constant (K)
Port 1
Port 3
Port 2
Port 4
C
L
C
C C
Microstrip TRLs
Patent
Pending
Very low resonances achievable
Dashed: Ordinary
materials
p g ( )
L
C C
Series Chip Capacitors Parallel Chip Inductors
Engineered Crystals to Control k-e diagram
Interdigital Capacitors
Shunt Inductors
33
Equivalent Circuit Realization of the DBE Antenna
(TL pair with Controlled Coupling, Patent Disclosure)
Bloch analysis
(with transfer
ABCD matrices)
L1 = L2 = L3 = 1nH,
Capacitively Loaded TLs:
Coupling Capacitor tunes DBE
Coupled
Uncoupled Uncoupled
H
z
)
L1 L2 L3 1nH,
C1 = 10pF, C2 = C3 = 1pF
L1
C1
0
L3 LM
C3
0
CM
LM
(V
1
I
1
)
(V I )
(V
3
I
3
)
(V I )
Uncoupled Phase shift Coupled Coupling capacitor and
r
e
q
u
e
n
c
y

(
G
H
CM = 0.5pF
CM = 1.5pF
CM= 2 5pF
L2
C2
0
L3 LM
C3
0
(V
2
I
2
) (V
4
I
4
)
Inductive
coupling
Capacitive
coupling
Uncoupled Phase shift
difference due to separate
LC loading emulates
anisotropic tensor
Coupled Coupling capacitor and
inductors realize propagation in
rotated anisotropic tensor
K Bloch Wave Number
F
CM = 2.5pF
(DBE)
DBE design Guidelines: K Bloch Wave Number
n
c
y

(
G
H
z
)
CM = 2.5pF
(DBE)
DBE design Guidelines:
Unit cells with capacitively loaded TLs (C > L) are tuned to DBE via coupling
capacitor (CM).
Inductively loaded (thin) TLs (L < C) are tuned to DBE through inductive
coupling (LM).
Pros & Cons of CM tuning: CM can be easily realized with chip capacitors.
K Bloch Wave Number
F
r
e
q
u
e
n
However (C > L) TLs are very thick on low contrast substrates.
Pros & Cons of LM tuning: Thin TLs have L > C. But, realizing LM on printed
circuits is challenging (needs transformers)
Combination of thick and thin TLs for miniaturization & CM tuning
34
DBE Antenna Design Using Lumped Circuit Model
Choose Circuit Layout Suitable for
Circular Cascading
1
Coupling
Coupling capacitor
2
Microstrip and Lumped Element Realization via
Full Wave Simulators
C1 C2
CM
C2 C1
p g
Coupling capacitor
Loading
L
1
L
2
L
2
L
1
CL
Capacitively
Loaded Lines
Substrate: Duroid, =2.2 tan = 0.0009
2 inch x 2inch x 0.125inch
(5.08 cm x 5.08 x 0.3175cm) Same with Ground Plane
capacitor = 1pf
Loaded Lines
G
H
z
)
CM = 0 pF
CM 0 5 F
Coupling
F
r
e
q
u
e
n
c
y

(
CM = 0.5pF
CM = 0.65pF
(DBE)
K Bloch Wave Number
35
Optimal Antenna Size Using DBE Mode
Printed on High c
r
Substrate
0.85 inch (2.16 cm)
50
Coaxial Cable
H = 0.5 inch (1.27 cm)
Measured 3~3.5% Bandwidth
Coaxial Cable
(Capacitively
Coupled)
0.88 inch
(2.24 cm)
Substrate: Alumina, =9.6 tan = 0.0003
2 inch x 2inch ~ (5.08 cm x 5.08 cm)
(Same with Ground Plane)
4.25 dB Gain at 1.445 GHz
CoorsTek AD-995 substrate:
Alumina = 9.7 2inch x 2inch x 50mil
The antenna has dimensions of ~

0
/9.6 x
0
/9.3 x
0
/16 at 1.445 GHz
Realization
of these elements on devices
with self-scanning capabilities
should provide for satellite to
This design seems to be the smallest
antenna anywhere for the given gain and
bandwidth
should provide for satellite to
handheld communication
36
DBE Antenna with Lumped Capacitor Loading
Electric field at DBE
Resonance
Antenna Layout
2
0
-5
0
45 -45
Measured Gain
1inch = 2.54cm
Capacitively
coupled coax
Measured
S
-10
-15dB
90
135
-90
135
20 % shift
B
(
S
1
1
)
e
n
c
y

(
G
H
z
)
CM = 0 pF
CM = 0.5pF
CM = 0.65pF
e
n
c
y

(
G
H
z
)
CM = 0 pF
CM = 0.5pF
CM = 0.65pF
-6
-4
-2
0
2
(
d
B
)

S
11
135 -135
180
20 % shift
0.4%BW
2dB Realized Gain
Explore 20%shift in
d
B
K Bloch Wave Number
F
r
e
q
u
e
p
(DBE)
K Bloch Wave Number
F
r
e
q
u
e
p
(DBE)
2 2. 2 2. 4 2. 6 2. 8 3
-16
-14
-12
-10
-8
Frequency (GHz)
S
1
1


C
M
=0pF
C
M
=0. 5pF
6.3dB Gain (w/o
capacitors)
1.8dB Gain
p
resonance
frequency for
reconfiguration
(time varying
DBE antenna)
CM = 0pF (0.9%
BW)
CM = 0.5pF
(0.4%BW)
CM = 0.65pF
(DBE) (0.25% BW)
Thicker substrate increases the
6.9dB Broadside Directivity
Frequency (GHz)
Radiation efficiency is 35% due to the lossy lumped
elements
37
Thicker substrate increases the
bandwidth:
125mil = 0.25% BW vs. 275mil 1.5%BW
Varactor Loaded Tunable Antennas
Antenna Design Experimental verification
Varactors (cm)
Skyworks smv-1405 varactors:
2.67pF (0V bias) - 0.63pF (30V bias)
Antenna Design Experimental verification
Control signal
(18nH)
Control ground (18nH) RF input
Antenna Footprint (on
r
=2.2 substrate):
2" 2" 0.25"
/4 0 /4 0 /31 @1 490GHz (cm = 0 0pF) Tunable from1 12GHz to 1 31GHz
0
/4.0
0
/4.0
0
/31 @ 1.490GHz (cm = 0.0pF)

0
/4.3
0
/4.3
0
/35 @ 1.366GHz (cm = 0.5pF)

0
/4.8
0
/4.8
0
/39 @ 1.218GHz (cm = 1.0pF)

0
/5.4
0
/5.4
0
/43 @ 1.087GHz (cm = 1.5pF)
Tunable from 1.12GHz to 1.31GHz
Improving antenna performance: Improving antenna performance:
1) Varactors tunable from 0pF to 2pF (1GHz 1.5GHz)
2) Low loss varactors
3) Smaller capacitances for higher frequencies
Higher order K- Diagram Using
Multiple Transmission Lines
Coupled
Uncoupled Uncoupled
(V I )
(V I )
Coupled
Uncoupled Uncoupled
Coupled
Uncoupled Uncoupled
(V I )
(V I )
1
I
1
V
2
I
4
I
4
V
I
11
L
11
C
21
L
21
C
M1
C
M3
C
1
I
1
V
2
I
4
I
4
V
I
11
L
11
C
21
L
21
C
M1
C
M3
C
Realization of Symmetric SIP:
SIP at ~2 9GHz
L 1
C 1
0
L 2
C 2
0
L 3 L M
L 3 L M
C 3
0
C 3
0
C M
LM
(V
1
I
1
)
(V
2
I
2
)
(V
3
I
3
)
(V
4
I
4
)
Inductive
coupling
Capacitive
coupling
L 1
C 1
0
L 2
C 2
0
L 3 L M
L 3 L M
C 3
0
C 3
0
C M
LM
(V
1
I
1
)
(V
2
I
2
)
(V
3
I
3
)
(V
4
I
4
)
Inductive
coupling
Capacitive
coupling
2
V
2
I
3
V
3
I
5
I
5
V
6
I
6
V
12
L
12
C
13
L
13
C
22
L
22
C
23
L
23
C
M2
C
M3
C
2
V
2
I
3
V
3
I
5
I
5
V
6
I
6
V
12
L
12
C
13
L
13
C
22
L
22
C
23
L
23
C
M2
C
M3
C
550mil
SIP at 2.9GHz
e
n
c
y

(
G
H
z
)
3 Uncoupled lines 3 Uncoupled lines 3 Coupling Capacitors
Coupled lines
3 Uncoupled lines 3 Uncoupled lines 3 Coupling Capacitors
Coupled lines
Effect of Tree-way Coupling (CM3 nonzero) on K- Diagram:
Lower order K branch can display a 6
th
order behavior
Coupling
Capacitors
F
r
e
q
u
K- Bloch Wavenumber
2.5
3
10
9
H
z
)
Higher order
K- curve
2.5
3
10
9
H
z
)
Higher order
K- curve
Higher
order K-
Curve
Lower order K- branch can display a 6
th
order behavior.
Symmetric stationary inflection points (similar to SIP in MPCs)
2.7
10
9
z
)
2.7
10
9
z
)
Symmetric Stationary
Inflection Points in the
p
1.5
2
F
r
e
q
u
e
n
c
y

(
G
H
1.5
2
F
r
e
q
u
e
n
c
y

(
G
H
Curve
2.5
2.6
F
r
e
q
u
e
n
c
y

(
G
H
z
2.5
2.6
F
r
e
q
u
e
n
c
y

(
G
H
z
Inflection Points in the
Propagation Spectrum
Achieved without resorting
39
0 t/4 t/2 3t/4 t 5t/4 3t/2 7t/4 2t
1
K Bloch Wavenumber
0 t/4 t/2 3t/4 t 5t/4 3t/2 7t/4 2t
1
K Bloch Wavenumber
t/2 3t/4 t 5t/4 3t/2
2.4
K Bloch Wavenumber
t/2 3t/4 t 5t/4 3t/2
2.4
K Bloch Wavenumber
L11 = L12 = L13 = L21 = L22 = L23 = 1nH, C11 = 10pF, C12 = 5pF, C13 = 1pF,
C21 = C22 = C23 = 1pF, CM1 = 2pF, CM2 = 2pF, CM3 = 2.6pF
to lossy ferromagnetic
layers!
Magnetic Substrates and Realization of
Magnetic Photonic Crystal Modes
N 2 1
Z
01
Z
02
Z
03
Z
04

excitation
N = 64
excitation
DBE vs. MPC Antenna Performance
40
Tuning Non-Reciprocal K-e Diagrams.
Ferrite substrate only under the coupled sections.
H
0
dielectric
ferrite
ProvidesSeveralAdvantages:
Ferritelosseskepttoaminimum(only
smallferritesubstratesections) s a e i e subs a e sec io s)
Highepsilon(c
r
15)ferrite(
r
>1)
addstominiaturization
Radiation mechanism similar to patch
H
0
ferrite
fringing
fields
Radiationmechanismsimilartopatch
antennas(fringingfieldsondielectric
substrate)
Ferrite inclusion allows for design Ferriteinclusionallowsfordesign
flexibilityformatchingandtunability
Standardbandwidthincreasetechniquescanbe
41
q
incorporated
Specific Design with Magnetic Inserts
2
c
r
=2.2,
tano =0.0009
ferrite:c =15,
tano <1.4E4,
AH = 6Oe
Bandgap
Bandgap
.
.
2

0.87
0
.
8

AH 6Oe
H
0
=80kA/m
thickness=100mil
NovelMPCconceptimprovesbeneficialpropertiesof
DBEantennas:
Close to optimal gain and bandwidth Closetooptimalgainandbandwidth
Broadsideradiationpattern(suitableforphasedarrays)
Smallerfootprints(suitablefortightlypackedarrays)
21% footprintreductionwithMPCDesign(overDBE
antenna)
12% 14% B d id h I
42
Substrate: Rogers Duroid c
r
= 2.2, tano = 0.0009
Ferrite inclusions: Calcium Vanadium-doped Garnet (CVG)
(4M
s
= 1000G, H = 6Oe).
12%14% BandwidthImprovement
Forthethinsubstrate(d=100mil),12%improvement
Forthethicksubstrate(d=400mil),14%improvement
Substrate Thickness Effect
MPC, d = 100mil
MPC, d = 200mil MPC, d = 300mil
MPC d = 400mil MPC, d = 400mil
FBW = 0.7635%
Gain = 6 2845dB
FBW = 2.2302%
Gain = 6 5290dB
FBW = 3.8922%
Gain = 6.4109dB
Gain 6.2845dB Gain 6.5290dB
Gain 6.4109dB
FBW = 5.7554%
Gain = 6.3185dB
DBE-2, d = 400mil
FBW = 5.0314%
43
Gain = 6.7626dB
MPC Antenna with Unbiased Ferrite Blocks
6 dB
Realized Gain
Reflection
5.19 dB
6.23 dB
d
~330 MHz shift
2.04 dB
330 MHz shift
2.00 3.00 4.00
Freq [GHz]
2.467
3.465
2 467
3 465
2 799
3 794
44
2.467
3.465
2.799
3.794
Realized Gain with Biasing
5 76 dB 5.76dB
1.55dB
0.03dB
MPC
resonance
at2.35GHz
Linearly polarized
45
Linearlypolarized
Broadsideradiation
Performance Comparisons
MPC
2
2
0.87
DBE
2
2
1 05
Patch
2
2
1 2
2
0.5
0.8
0.87
2
0 5
0.92
1.05
2
1.3
1.2
0 5 0.5 0.5
Allantennasareon2x2finitegroundplanes g p
46
G/Q Footprint Comparison
DBE
4 x 4
PATCH
1.8 x 1.8
PATCH
4 x 4
c
r
= 2.2
P i t d MPC
DBE
PATCH
2 x 2
PATCH
PATCH
2 x 2
DBE
2 x 2
MPC
2 x 2
c
r
= 9.6
Printed MPC
2 x 2
unbiased
2 x 2
DBE
PATCH
1.2 x 1.2
(Elliptical)
OSU-ESL double loop
Max. radius 6cm
DBE
1.8 x 1.8
DBE
1 5 1 5
1.2 x 1.2
PATCH
1.5 x 1.5
PATCH
1.2 x 1.2
(Linear)
1.5 x 1.5
Iizuka-Hall Printed Antenna
10cm x 1cm (no GP)
Iizuka-Hall Wire Antenna
10cm x 1cm (no GP)
47
Future Radios will Likely be on Hybrid
Substrates
Radio
Receiver
Interconnects
Beamformer Design
between Silicon
and Polymer
48
High speed data lanes overlaid on top of standard FR4 board targeting.
A SERDES I/O signaling over a backplane
utilizing OTT interconnects
Issue: Limiting Speeddue to
Interconnects
Interconnected ICs
a b a b
Polymer data lanes
Passives/discrete ICs
M l i
Standard/low speed interconnect
FR-4 board
Multi-processors
multi-chip bus connection.
chip-to-chip link
49
Printing on Polymers is Challenging
Printing on smooth
PDMS surface (no
surface modifications)
Printing on
microte t red PDMS
Cracks formed on
metal surface
microtextured PDMS
(only evaporation)
Printing on Printing on
microtextured PDMS
(evaporation and
electroplating)
Peeling from PDMS
surface
Printing on microtextured
d f difi d and surface modified
composite surface
50
Transfer Molding/Lift-off and Electroplating
Lift-off negative mask in acetone
Transfer Molding
PDMS-Ceramic Composite
Plastic negative mask on PDMS mold
Stamp Mold
Transfer Molding
PDMS-Ceramic Composite
1m
10-200 m PDMS-Ceramic Composite
Thin copper seed layer (~1 um)
PDMS-Ceramic Composite
Reactive Ion Etching
Stamp Mold
PDMS-Ceramic Composite
PDMS-Ceramic Composite
Electroplating PDMS-Ceramic Composite p g
Negative pattern mask
M
+
M
+
M
+
Reference electrode
Flexible, thick conduction lines on composite
Evaporated thin seed layer (cathode)
Polymer composite substrate
Copper electrode (Anode)
Cu
+2
sulfate-sulfuric acid solution
51
Flexible Microstrip Line Via Electrodeposition
Micro textured surface: Micro-textured surface:
Flexible microstrip line:
74 mm
-0.1
0
0.1
Measured insertion losses (S21)

-0.1
-0.05
0
S21

-0.1
-0.05
0
S21

-0.5
-0.4
-0.3
-0.2
d
B
-0.3
-0.25
-0.2
-0.15
d
B
-0.3
-0.25
-0.2
-0.15
d
B
0 9
-0.8
-0.7
-0.6
0 degree
30 degree
90 degree
-30 degree
-90 degree
0 5
-0.45
-0.4
-0.35
Printed TL
simulated PEC TL
0 5
-0.45
-0.4
-0.35
Printed TL
simulated PEC TL
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
-0.9
frequency (GHz)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-0.5
frequency (GHz)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-0.5
frequency (GHz)

Conducting loss: @ 1 GHz
m dB mm
mm
dB
/ 35 . 1 1000
74
1 . 0
=
High flexibility and high conductivity.
52
Carbon Nanotube Printing on Polymer-
Ceramic Composites
Carbon Nanotube/Silver coated textiles for integration into polymer-ceramic Carbon Nanotube/Silver coated textiles for integration into polymer ceramic
composites.
Key features:
Mechanically very flexible & strong Mechanically very flexible & strong
Strong adhesion with polymer-ceramic
composites
Scalable production process
Comparable performance compared to traditional Comparable performance compared to traditional
antennas
Developed E-textile patch gain was within 90% of the ideal PEC
patch gain)--- Measured gain was 6dB at 2GHz (viz. 53
Direct CNT Printing on Polymer
Step 1: Grow Aligned CNTs on
a Glass Substrate
Key Critical Parameters:
CNT Length
CNT Density
CNT Vertical Uniformity
Step 2: Spin Coating
with Polymer Solution
Key Critical Parameters:
Spin Coating Rate
Polymer type (higher shrinkage rate
h d) when cured)
Step 3: Transfer to
Larger Polymer Substrate Larger Polymer Substrate
54
Step 1: Growing Aligned CNTs
This process involves using a quartz tube This process involves using a quartz tube.
1. Prepare the platform for
the CNT (Si, SiO2 or Quartz)
Si, SiO2, Quartz etc.
Key Observations:
Length of CNTs are very critical
to achieving high conductivity.
b i id b
2. Deposit Catalyst (Fe)-Optional
to achieving high conductivity.
We have found out that CNTs
with about 100um length has the
best conductivity characteristics.
Density of CNTs are very critical
to higher conductivity
3. Grow Nanotubes inside Quartz Tube
to higher conductivity
We found that the grow-process is very vital
to achieving denser aligned nanotubes. We
developed our own process to achieve the best
performance
Uniformity of CNTs in vertical
C
2
H
2
CH
Uniformity of CNTs in vertical
direction
We found out that CNTs tend to pull each other
at the tips. Therefore, they have higher
contact area. This leads to higher conductivity.
In other words, we are using aligned CNT
t hi li d ti f hi h
Quartz Tube Furnace 79400 (1000 F)
CH
4
process to achieve non-aligned tips for higher
contact area.
55
Step 2: Spin Coating With Polymer
Take off the silicon
wafer by Hydrofluoric (HF) acid
PDMS
Or peel it off
(Before Curing: Resistance is R0)
(After Curing: Resistance is R0/2) (After Curing: Resistance is R0/2)
Key Observations:
Curing leads to shrinkage in polymer, which in turn
improves conductivity.
Higher shrinkage ratio is critical to achieving higher Higher shrinkage ratio is critical to achieving higher
conductivity. For PDMS, it has about 2% shrinkage. For
instance rubber has 4% which leads to 100% improvement
in conductivity.
56
Step 3: Transfer to Larger Polymer Substrate
57
Carbon Nanotube Printing on Polymers for Layered
Packaged Electronics?
Over 6 dB Gain
10
E-plane pattern
antenna layer:
microstrip patches,
antenna layer:
microstrip patches,
Within 1dB of
the Perfectly
Conducting Patch
5
0
5
dipoles, bowties and etc.
feeding layer:
feeding networks,
impedance tuning stubs
dipoles, bowties and etc.
feeding layer:
feeding networks,
impedance tuning stubs
-15
-10
-5
d
B
circuits layer:
filters, mixers,
LNAs and etc.
circuits layer:
filters, mixers,
LNAs and etc.
-200 -150 -100 -50 0 50 100 150 200
-25
-20
u (degree)
58
Packaging Trends A Bigger Challenge?
*Photo from StratEdge data sheet
Low Cost ICs Still Need Viable
Packaging g g
Likely bigger challenge than the
chipset for MMW/THz
Commercial Packaging extends to Commercial Packaging extends to
MMW frequencies
DC to 50+ GHz
Progress continues up in frequency
Plastic Packages in use to ~30 GHz
59
Multi-physics, multi-domain co-simulation of system-in-package (SiP)
Working with
Prof Jin-Fa Lee Prof. Jin-Fa Lee
60
Many Thanks to my Students & Collaborators at ElectroScience
61