Presented at the COMSOL Conference 2010 Boston

Effective Medium Theory of Nanodielectrics

for Embedded Energy Storage Capacitors

- Rajesh Bikky1, Nacer Badi*1, Abdelhak

Bensaoula2
1 – Center for Advanced Materials, UH; 2 – Department of Physics,
UH
* – email address: nbadi@uh.edu
 Contents
 Introduction

 Background theory

 Simulation

 Results

 Conclusion

 Future work

 Preview of empirical results

 Introduction

 Capacitors are passive devices that store

electrical charge because of charge separation
between plates.

 Storing energy is a key issue in all electrical

equipment.

 Passive components occupy about 70% of

space on PCBs[1].

 Need for higher storage in compact space is

increasing by the day.
1. R. K. Ulrich and L. W. Schaper, “Integrated Passive Component Technology,” IEEE press, Wiley-
Interscience, 2003.
 Introduction
 Embedded Capacitors replace discrete capacitors
in modern technologies like system in package
(SIP) and system on chip (SOC)[2].

 Advantages - Miniaturization, increase component

density and higher electrical performance [3].

 Applications of Embedded Capacitors include:

2. J. Lu, “High Dielectric Constant Polymer Nanocomposites for Embedded Capacitor Applications,” Georgia
Institute of Technology, 2008.
3. S. K. Saha, “Nanodielectrics with giant permittivity”, Bull. Mater. Sci., Vol. 31, No. 3, June 2008, pp. 473–477.
 Introduction

 Traditional ceramic capacitors cannot be used as

embedded capacitors because of poor volumetric
efficiency.

 Ceramic based dielectric material makes it hard to

manufacture (reproducibility, stability and cost) and
shows limited flexibility due to high ceramic
loading.

 Polymer based dielectrics can be potential

embedded capacitors because of their advantages
and compatibility with PCB materials.
 Background Theory:
POLYMERS
 High processability, mechanical flexibility, electrical
breakdown strength

 Low dielectric constant is a serious drawback for polymers

to be used as embedded capacitors.

 This can be overcome by adding „fillers‟ to polymer matrix.

 Composites cannot be made with ceramic fillers of smaller

size as their permittivity decreases with decrease in size[4].

 High loading of ceramic fillers in polymer defeat the purpose

of composites.
4. J. Lu and C. P.Wong, “Recent advances in high-k nancomposite materials for embedded capacitor
application,” IEEE Trans. Dielectr. Electr. Insul.,vol. 15, no. 2, pp. 1322–1328, Oct. 2008.
 Background Theory: METAL
NANOFILLERS
 Metal nanoparticles are promising filler materials as low
loading will suffice.

 They have interesting electrical, magnetic and physical

properties[5].

 Metal-Polymer percolative composites are identified as

strong candidates to realize high K[6].

 Nanoparticles based dielectrics have higher

capacitance density because of small thickness of the
5. L.film.
Nicolais and G. Carotenuto, Metal-Polymer Nanocomposites. John Wiley & Sons Inc.: Hoboken, NJ,
2005
6 . Yang Shen et al, “High Dielectric Performance of Polymer Composite Films Induced by a Percolating
Interparticle Barrier Layer”, Adv. Mater. 2007, 19, 1418–1422.
 Background Theory:
CAPACITOR
 Amount of charge stored in a capacitor is denoted
by its capacitance.
C – Capacitance
K 0 A K – Dielectric Constant
C ε0 – Permittivity of free
d space
A – Area of plates
d – Distance between
 Greater the value of K, greater the C value.
plates

 Relationship between polarization, electric field

and capacitance
P   0  e Eeffective P – Dielectric Polarization
where χe – Susceptibility
e  K 1 E – Electric field between
plates
 Background Theory: COMPOSITE
CAPACITOR

V
E
d
V = voltage applied to the E Epolarizatio
plates
Eeffectiven = E -
Epolarization
As Eeffective decreases, P and K increases
 Background Theory: EFFECTIVE MEDIUM
THEORY (EMT)
 Effective dielectric constant of medium is calculated
with EMTs[9]. Few popular EMTs are
EMT Formula εeff - effective permittivity
Model of the medium
 h
Maxwell – 1 2 f ( i )
 i  2 h εi - dielectric function of
Garnett  eff   h[ ]
i  h inclusion (filler)
1 f ( )
 i  2 h
Symmetric εh – dielectric constant of
1
Bruggeman
 eff  [3 f ( i   h )  2 h   i  host polymer
4
EMT (1  3 f ) 2  i2  2(2  9 f  9 f 2 ) i h  (3 f  2) 2  h2 ]
f – fraction of inclusion

Asymmetric  i   eff  eff 1A A=0.5 for disks and 0.33

 (1  f )( )
Bruggeman i  h h for spheres.
9. William M. Merrill, et al, “Effective Medium Theories for Artificial Materials Composed of Multiple Sizes of
Spherical Inclusions in a Host Continuum”, IEEE Transactions on antennas and propagation, vol. 47, no. 1,
January 1999.
 Background Theory: DRUDE
THEORY
 Dielectric constant of size dependent Au
nanoparticles[8] is
 i   freeelectrons   boundelectrons
2 2
pf pf
i  1 
 2  i  2   2  i
f 0 b
ω - frequency

ωpf - plasma frequency = 1.3*1016 Hz

ϓf - size dependent damping factor = 2.65*1014 Hz

ϓb – bound electron damping term = 2.4*1014 Hz

ω0 - bound electron constant term = 7*1015 Hz

8. A. E. Neeves and M.N.Birnboim, “Composite Structure for the enhancement of nonlinear-optical
susceptibility”, J.Opt Soc. Am. B. Vol 6, No. 4, April 1989.
 Background Theory: PERCOLATION
THEORY
 Huge changes occur in physical and electrical
properties for a critical loading of nanoparticles,
known as percolation threshold, fc[10].

 At loading above fc , dielectric loses its insulating

properties and becomes conductive.

 Power law describes properties of system at fc

K
 ( f  f c )s S is a constant of
Kh value 1

10. C. W. Nan, Y. Shen and Jing Ma, “Physical properties of composites near percolation”, Annu. Rev. Mater.
Res. 2010. 40:3.1–3.21
 Simulation: SETUP

 A 2-step procedure is employed to calculate

effective permittivity of the medium.

 First step is to generate an arrangement of

nanoparticles in PVP matrix.

 Disks and spheres are randomly arranged for 2D

and 3D modeling respectively.

 Number of fillers is chosen according to desired

loading of nanoparticles.
 Simulation: SETTINGS
 AC/DC module – In plane electric currents model is
used.

 In sub domain settings, appropriate materials are

selected from material library. Conductivities and
relative permitivities of PVP and Au are applied to the
geometry.

 Using boundary conditions, one face is set as input

voltage while its opposite is ground. Other faces are
set to periodic condition.

 Drude, EMT and Percolation theory expressions are

 Simulation: SOLVERS and POST-
PROCESSING
 Parametric solver sweeps frequencies from 1kHZ
to 1peta Hz at constant loading.

 Parametric solver is also used at constant

frequency when loading is varied from 0 to 1.

 Post-processing is used to create slice plots for

3D models and surface plots for 2D models.

 Global expression plots and data points values

are acquired from post-processing.
 Results: 2D MODELING

Additional
polarization is
created at surface
of fillers in direction
of electric field.
With increase in net
polarization,
effective dielectric
constant also
increases as
Electric polarization patterns in 2D
governed by
equations nanodielectric with f = 0.134
 Results: 2D MODELING

f = 0.38 f = 0.78

Effective dielectric constant increases with increase

in loading of nanoparticles.
 Results: 3D MODELING

3D and 2D models
follow same
pattern of
polarization
increase when
there is an
increase in loading
of nanoparticles.

Electric polarization patterns in

3D nanodielectric with
 Results: 3D MODELING

f=0.078 f=0.155

Effective dielectric constant increases with increase in

loading of nanoparticles.
 Results: K CALCULATION USING
EMTS
 Effective dielectric constant of the medium is
calculated using above mentioned EMT
models.

• All
EMTs predictions are
close to each other at low
loading.

• EMTs fail at high loading

values because they
Real part of dielectric ignore the inter-particle
function of 2D composite effects which are
calculated using various predominant at those
EMTs values.
 Results: K CALCULATION USING
PERCOLATION THEORY for 3D MODEL
 Effective dielectric constant of the medium at fc is
calculated for 3D model.

Dielectric
constant
increases
gradually with
increase in
loading. There is
a sudden and
huge increase in
K value near
percolation
 Results: K CALCULATION USING
PERCOLATION THEORY for 2D MODEL
 Effective dielectric constant of the medium at fc is
calculated for 2D model.

An increase of
magnitude 200
times is
observed at fc. K
value of polymer
increased from 7
at f=0 to 1400 at
fc.
 Conclusions

 Electric field and polarization patterns of 2D and 3D

nanodielectrics are observed.

 At low loading, EMTs and Percolation theory

predictions are close and both theories predict
gradual increase in dielectric constant.

 EMTs fail at high loading but percolation theory

takes in to account the metal-insulator nature of the
composite and predicts huge increase in value of K
at percolation threshold.

 At percolation threshold, K is determined as 1400,

where as for a bare polymer this value is just 7.
 Future Work

• Thiswork will be extended other inexpensive

metals such as Ag.

• Studieswill be continued with particle of more

complex structures, such as core-shell type.

• Comparison between analytical and empirical

values will be carried out.
 Preview of Empirical Results
• Nanoparticles are synthesis in our labs. Studies on PVP
films were conducted.
• Parallel plate nanodielectrics are fabricated with different
loadings of nanoparticles.
Loading of Loading of Breakdown Dielectric
nanoparticles nanoparticles field constant
(by weight) (by volume) (V/µm) (K)

0 0 130 7
0.88 0.056 80 8.1
2.2 0.14 81 6.6
3.4 0.217 75 4.5
10 0.064 33 20

I-V Characteristics of nanodielectrics.

Data used to plot the graph is
tabulated.
 Preview of Empirical Results
• SEM structural analysis, I-V characterization and
capacitance measurements are carried out for fabricated
devices.
Loading of Loading of Breakdown Dielectric
nanoparticles nanoparticles field constant
(by weight) (by volume) (V/µm) (K)

0 0 130 7
0.88 0.056 80 8.1
2.2 0.14 81 6.6
3.4 0.217 75 4.5
10 0.64 33 20

Breakdown field of nanodielectrics Vs

Loading of nanoparticles. Data used to
plot the graph is tabulated.

With increase in loading of nanoparticles, breakdown field

decreases and dielectric constant increases.