Detection of electromagnetic waves using MEMS antennas

beigecakeΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 6 μήνες)

53 εμφανίσεις

Detection of electromagnetic waves using MEMS antennas


P.G. Datskos
*,a, b
, N.V. Lavrik
a
, J. Tobin
a

, and L.T. Bowland
a


a
Oak Ridge National Laboratory, Oak Ridge, T
N

37931
-
6054
;


b
University of Tennessee, Knoxville, T
N

37996
-
1200


ABSTRACT



We describe the design, fabrication and characterization of simple micromechanical structures that are capable of
sensing static electric time varying electromagnetic fields. Time v
arying electric field sensing is usually achieved using
an electromagnetic antenna and a receiver. However, these antenna
-
based approaches do not exhibit high sensitivity
over a broad frequency (or wavelength) range. An important aspect of the present wo
rk is that, in contrast to traditional
antennas, the dimensions of these micromechanical oscillators can be much smaller than the wavelength of the
electromagnetic wave. We characterized the fabricated micromechanical oscillators by measuring their respon
ses to
time varying electric and electromagnetic fields.



Keywords:

MEMS, oscillator, electromagnetic field, antenna


1.

INTRODUCTION



In
Wireless transmission of signals by modulating the frequencies of radio waves has been a form of communicating
information and was linked to available technology. Early receivers were only capable of receiving on/off signals such
as Morse code but the invention of vacuum tubes enabled reliable audio communication.

In
the frequency band 3 MHz


300 GHz the sensing

of electromagnetic waves is usually achieved using an antenna. In fact, the first antennas were
built Hertz who placed dipole antennas at the focal point of parabolic reflectors for both transmitting and receiving
.

1


Depending on their applications, antennas generally c
an be omnidirectional or only weakly directional antennas and
directional or beam antennas.
Omnidirectional

antennas receive or radiate more or less in all directions and are
employed when the relative position of the transmitter is unknown or arbitrary

2
. They are also used at lower frequencies
where a directional antenna would be too large, or simply to cut costs in applications where a directional antenna isn't
required. Directional are used to preferentially radiate or receive in a part
icular direction or directional pattern.
Antennas beyond the dipole or vertical designs are usually intended to increase the directivity and the gain of the
antenna. The vast majorities of designs are fed with a balanced line and are based on the dipole a
ntenna with additional
components, which increase its directionality.

R
esonant antenna
s

3

ha
ve a broader frequency response but its
bandwidth
suffer
o
r is substantially reduced when the distance from the transmitter is large
.

More recently
plasma antenna
s have
been investigated where
plasma is used instead of the metal elements of a traditional antenna
.
3
-
5


Antenna d
esigns such as phased arrays
6
,
7

consist of two or more simple antennas, which are connected together
through an electrical network. Antenna arrays also employ basic (omnidirectional or weakly directional) antenna type,
such as dipole, loop or slot antennas.
In such systems, the sensit
ivity is dependent on the design of the antenna and the
receiver electronics. However, these antenna
-
based approaches do not produce systems that exhibit high sensitivity over
a broad frequency range. Furthermore, since the antenna size is related to the

wavelength of the signal, traditional
electric field detectors are relatively large (especially at lower frequencies) so that compact detectors and arrays of
detectors are not possible.




Recently

studies have reported the development of working radio de
vice based on nanosized radio
-
wave receivers
and detectors fabricated from a single carbon nanotube.
8
,
9


In those devices the amplification and demodulation rely on
field
-
emission from carbon nanotubes due to sharp
-
point geometry t
hat concentrates the electric field.
8


Other studies
also demonstrated that
multiwall carbon nanotubes
interact

with light as simple dipole radio

antennas

and
they show
both the
polarization and

the length antenna effect.

The second, the antenna length effect, maximizes

the response when
the antenna length is a proper multiple

of the half
-
wavelength of the radiation.
10

*datskospg
@ornl.gov;

phone 1 865 574
-
6205; fax 1 865 574
-
9407






2.

THEORY

2.1

Physical Model

In our present work we used a model that involves a charged resonator (see Figure 1). Our approach is based on the
interact
ion of electric charges (localized on the cantilever) with external electric fields. For the sake of simplicity we
will assume that the cantilever device is composed of an electrically nonconductive stem and a conductive microsphere.
If a charged microsp
here is attached at the end of a cantilever the amplitude response, Δ
x
(
ω
) as a function of frequency,
ω, for such a cantilevered charge is given by


x
(

)

q

E
m

2


0
2


2




0
/
Q


2










(1)

where
q

is the electric charge present on the sphere,
E

is the external
field, m is the mass of the system and
Q

is the
quality factor of the resonator.

The mechanical response of the device is depicted schematically in Figure
1
, where the
charge,
q
, at the end of the stem follows the local variations in the external electric

field
.






















Figure
1
.
The detector mechanical resonance frequency can be matched to the
E
-
field frequency. Images (A)
through (E) depict an incoming sinusoidal
E
-
field. The charged NMS responds to the
E
-
field by deflecting
sinusoidally, shown for time
t
=
t
0

to
t
=
t
4
. One complete cycle of motion is
shown during

this time interval. Once
the wave passes the detector, the detector stops responding.






Table 1.
Parameter values

used in our present c
alculations

Mass,
m

(
k
g)

Quality
Factor,
Q

Spring constant,
k

(N/m)

Cantilever length

(

m)

Charge
,
q

(C)

B
andwidth,
B

(Hz)

10
-
15

3.5
×
10
4

5.2
×
10
-
5

400

10
-
12

10


2.2

Fundamental limits: thermomechanical noise

In every mechanical system at a finite temperature there is a component of noise, which is described as
thermomechanical noise. The thermomechanical noise for a cantilever at a temperature
T
and bandwidth
B

is given by


x
(

)

q

E
m

2


0
2


2




0
/
Q


2









(
2
)

At frequencies below the resonance frequency (
ω
<<
ω
0
) this noise becomes


x
(

)
T
M
2
1
/
2

4
k
B
T
B
Q
k

0
3

2


0
2


2


0
4
/
Q
2




®




4
k
B
T
B
Q
k

0







(
3
)

In order to obtain analytical expressions for the fundamentally limited detector performance we evaluate the signal
-
to
-
noise ratio, Δ
x
/<
δx
2
>
1/2

(
in the limit
ω
<<
ω
0
)


x

x
T
M
2
1
/
2

q
m
1
4
k
B
T
B
Q
k

0

2


0
2


2


2

0
2
/
Q
2
E







(
4
)

Using the values in
listed in
Table 1 we plotted in Figure 2 the
thermomechanical

noise and respons
e
to a time varying
electric field
and
as a function of frequency
.


Figure
2
.
Cantilever antenna
response
to an
E
-
field and thermochemical noise
as a function of frequency
.





Defining, the Noise Equivalent Electric Field Difference (
NEEFD
) as the electric field for which the signal to noise ratio
equals

unity, we obtain

N
E
E
F
D

m
q
4
k
B
T
B
Q
k

0

2


0
2


2


2

0
2
/
Q
2







(
5
)

The model presented above allows the
evaluat
ion of the
fundamentally limited performance
and the
expected response
for the MEMS antenna
for
various sizes, resonance frequencies,
and
electric charge.


3.

EXPERIMENTAL

3.1

Fabrication

We used low stress silicon
-
rich silicon nitride (SiNx) as a structural layer for cantilever fabrication. 500 nm thick SiNx
layers were deposited on 100 nm Si (100) wafers using low
-
pressure chemical vapor deposition (LPCVD). The
cantilever processing sequence (Figure
3
) relied of a single layer photolithography plasma etch of SiNx and anisotropic
wet etch of Si. In brief, a SiNx layer was patterned using contact photolithography and after defining cantilever
geometries in SF6 pla
sma, the cantilever arrays were undercut by KOH anisotropic etching. In final step, PVD was used
to deposit 60 nm Au metallization with a 10 nm Ti adhesion layer.


Figure
3
. Microfabrication

steps employed to produce the cantilevers used in the present studies. A SiNx layer
was deposited on a Si wafer using LPCVD and was subsequently patterned photo
-
lithographically and the
cantilever arrays were defined by KOH anisotropic etching to undercu
t the cantilever structures. PVD was used to
deposit Ti and Au layer.


3.2

Finite element analysis

We conducted finite element analysis in order to evaluate mechanical resonances and distribution of electrical charge in
the tested system.

Using COMSOL soft
ware package we constructed two models, mechanical and electrostatic, which
provided eigenfrequency analysis and charge distribution, respectively. Mechanical parameters of the gold and silicon
nitride layers used in the model are given in Table 1. The f
irst
vibrational
mode of the cantilever
[Figure 4(a)]
deduced
from the eigenfrequency analysis was 3,450 Hz
.

This value was
consistent with
the
experimentally measured value of
cantilever resonance frequency, 3,550 Hz. The electrostatic model
consisted o
f a cantilever in proximity to counter




electrode and
bias voltage was applied to
the
counter electrode with respect to the
metal (
gold
)
-
coated cantilever
antenna.
This model allowed
us to visualize and quantify
the spatial
distribution of the
electric
cha
rge on the cantilever
tip in the geometry depicted in Figure
4
(b) and the corresponding expanded view in Figure 4(c)
. The total charge was
analyzed as a function of the applied bias and was found to be
10
-
12
C for a voltage bias of 200 V
.


As it can be see in
Figures 4(b) and 4(c) the electric charge is mostly concentrated at the end
(tip) of the cantilever closest to the counter
electrode. Based on this information the model presented earlier with the charged sphere at the end of the canti
lever still
a
pplies

and we can use the expressions developed in section 2 to estimate the response of the MEMS antenna.




Figure
4
.

FEA modeling was used to theoretically evaluate the MEMS antenna used in the present work. The
charge,
q
, in the resonator was a function of the applied bias voltage and distance from the counter electrode
f
(
V
bias
,

r). The calculated eigenfrequency and d
eformation are shown in 5(a). The potential distribution is shown
in 5(b). The calculated charge density reveals that charge tends to concentrate at the tip 5(c).


charge concentrates

at the tip
!


(a)

(b)

(
c
)





3.3

Measurement set
-
up

We characterized individual MEMS antenna structures using the setup dep
icted in Figure 5.
The MEMS antenna was a
bimaterial
SiNx
cantilever with
length
400


m, width
100


m and a thickness of
500 nm
. The bimaterial was gold and
the thickness was
60

nm.

The motion of the cantilever was measured using an optical readout tech
nique
. When the read
laser beam is focused on the cantilever, the reflected beam deflects in accordance to cantilever bending. The amount of
cantilever bending can be measured by directing the reflected spot onto a position
-
sensitive photodetector (PSD).

In our
measurements the displacement of the charged tip the cantilever (antenna) was due to the due to the nitration with the
electric field
.
Geometrical evaluations of our setup along with calibration tests allowed us to convert the PSD output
voltage

into a deflection of the cantileve
r
.

The conversion coefficient found to be 2.5 mV/nm.


The charge on the
cantilever was due to the
bias voltage
that
was applied to a counter electrode with respect to the gold
-
coated cantilever

antenna
. An external EM source was excited with a sinusoidal
input.
This
EM
was used to excite the
MEMS antenna

and the corresponding mechanical response was measured using the optical readout.
























Figure
5
.

The experimental setup used in the present studies. A bias voltage was applied to a counter electrode
with respect to the gold
-
coated cantilevers. An external EM source was excited with a sinusoidal signal, which
was detected with the MEMS antenna.





3.4

Respo
nses of MEMS antenna


We used the experimental setup described in the previous section to measure the response of the cantilever antenna
to
external electric fields. First we
measured the amplitude of oscillation of the cantilever antenna
as a function of

the
distance of the emitter source
under different bias conditions.
In Figure 6(a) we plotted th
is
r
esponse of the MEMS
antenna to incoming
electromagnetic waves
as a function of
the
emitter source distance

and
for applied bias voltages

of
40 V and 400 V
.
The cantilever response is
inversely proportional to the distance from the excitation source. A curve fit
was performed for the experimentally measured responses.

In Figure 6(a) we plotted the fitted curve for the case of 40
V bias. The fitted curve was:
x
(
R
) = 10
-
6

R
-
2
.0681

and the correlation was 0.9805. The 1/
R
2

dependence confirms the
quasi
-

electrostatic regime.


In Figure 6
(b)
we plotted the
amplitude
r
esp
onse of the MEMS antenna to an electric field
E
=10
-
2

V
/m as a function
of frequency. The
measured
thermomechanical noise is also
plotted
as a function of frequency.
This noise was in the
range of sub
-
nm while the signal was in the range of 10’s nm. T
he
resulting
signal to noise ratio was
found to be
> 100
which gives a limit of detection of < 10
-
4

V/m.



Figure
6
.
(a) Response of the MEMS antenna to incoming EM wave as a function of emitter source distance is
shown for dif
ferent applied bias voltages. (b) Response of the MEMS antenna to an electric field
E
=10
-
2

V/m as a
function of mechanical frequency. The thermomechanical noise is also shown as a function of frequency. The
signal to noise ratio was >

100 which gives a
limit of detection of < 10
-
4

V/m.


4.

CONCLUSIONS

The implemented approach combines

a c
antilever based
E
-
field detector
and
integrates
in
a

MEMS

antenna, a high
Q

resonator and a frequency mixer in a highly compact design
. The
figures of merit
we demonstrate
d
include detectability
of
E
-
fields in the sub mV/m range and frequencies
in the kHz range.
The limit of detection was < 10
-
4

V/m and the
frequency
range can be extended in the MHz
using heterodyning
. The i
mplemented devices rely on a straightforward
tec
hnological sequence suitable for scaled up fabrication

and f
urther improved sensitivity can be obtained by integrating
a high charge density electre
te into the resonating element.

Depending on the material the
charge
lifetime can be quite
long. For examp
le in electrets charges do not decay for hundreds of years.







5.

ACKNOWLEDGEMENTS

The work performed

was supported by and the

Laboratory Director’s Research and Development

Program of Oak Ridge
National Laboratory
.
Oak Ridge National Laboratory is operated
for the U.S. Department of Energy by UT
-
Battelle
under Contract No. DE
-
AC05
-
00OR22725.

REFERENCES

[1]

H. Hertz, Annalen der Physik und Chemie
36
, 1 (1889).

[2]

C. A. Balanis, New Jersey: John Wiley & Sons, Inc
8
, 21
-
31 (2005).

[3]

J. Rashed and C. T. Tai, Antennas and Propagation, IEEE Transactions on
39

(9), 1428
-
1430 (1991).

[4]

Z. H. Qian, R. S. Chen, H. W. Yang, K. W. Leung and E. K. N. Yung, Microwave and optical technology letters
47

(2), 147
-
150 (2005).

[5]

J. P. Rayner, A.

P. Whichello and A. D. Cheetham, Plasma Science, IEEE Transactions on
32

(1), 269
-
281 (2004).

[6]

R. C. Hansen,
Phased array antennas
. (Wiley
-
Interscience, 2009).

[7]

R. J. Mailloux,
Phased array antenna handbook
. (Artech House, 2005).

[8]

K. Jensen, J
. Weldon, H. Garcia and A. Zettl, Nano Letters
7

(11), 3508
-
3511 (2007).

[9]

C. Rutherglen and P. Burke, Nano Letters
7

(11), 3296
-
3299 (2007).

[10]

Y. Wang, K. Kempa, B. Kimball, J. Carlson, G. Benham, W. Li, T. Kempa, J. Rybczynski, A. Herczynski and
Z. Ren, Applied physics letters
85

(13), 2607
-
2609 (2004).