Transcript Surface characterization of UV irradiated nanocrystalline

Electrical conductivity phenomena in
an epoxy resin-carbon-based
materials composite
www.polito.it/micronanotech
www.polito.it/carbongroup
M. Castellino*, A. Chiolerio**, M. Rovere*, M.I. Shahzad*, P. Jagdale* & A. Tagliaferro*
*Applied Science & Technology Department - Polytecnich of Turin , **IIT – Torino - Center for Space Human Robotics
10129 Turin, Italy
[email protected]
Introduction
Aim of this work
The effective utilization of Carbon Based Materials (CBMs) in composite applications
depends strongly on their ability to be dispersed individually and homogeneously
within a matrix. To maximize the advantage of CBMs as effective reinforcement for
high strength polymer composites, they should not form aggregates and must be well
dispersed to enhance the interfacial interaction within the matrix. Our protocol for
solution processing method includes the dispersion of CBMs in a liquid medium by
vigorous stirring and sonication, mixing the CBMs dispersion solvents in a polymer
solution and controlled evaporation of the solvent.
A thermoset commercial epoxy resin, used in the automotive field, has been chosen for this
study together with 16 different kinds of Carbon based materials: 13 different commercial
Carbon NanoTubes (CNTs), including Single and Multi-Walled CNTs both as grown and
functionalized, carbon beads and powders.
Different weight % (1 and 3 wt.-%) concentrations of CBMs in polymer resin were tried to
study the electrical behaviors of the polymer Nano-Composites (NCs). Therefore the best
composite has been chosen in order to study its conductivity behavior much more in details
(from 1 to 5 wt.-%).
Samples
Thermoset epoxy resin
+
N
Type
Diameter
(nm)
Length
(μm)
1
Multi wall
30-50
10-20
Purity
(weight
%)
> 95
2
Multi wall
<8
10-30
> 95
3
Short thin MW
9.5
1.5
> 95
4
Single wall
1-2
5-30
> 90
5
Single wall
1-2
0.5-2
> 90
6
Single wall
2
several
> 70
7
-COOH Functional SW
2
several
> 70
8
-COOH Functional MW
9.5
1.5
> 95
9
CNT powder
-
-
-
10
Carbon balls
1000
-
-
11
Carbon mix
-
-
-
12
Multi wall annealed
<8
10-30
> 95
13
Graphitized Multi wall
20-30
10-30
>99.5
14
Multi wall
18-35
>10
97
15
Multi wall
25-45
>10
98.5
16
Multi wall
6-10
>10
>90
Resin + 1 wt.-% of MWCNTs type 13
Electrical Properties
Electrical measurements were performed
using the so called “Two Point Probe (TPP)
method” (Schroder, 1990) with a Keithley-238
High Current Source Measure Unit, used as
high voltage source and nano-amperometer.
NCs samples showed three different electrical
behaviors: noisy, linear and non-linear
responses, which depend on dispersoids
amount and characteristics.
N
CBMs
(wt-%)
Resistance
(Ohm)
Comments
1
1
50 ∙ 109
noisy signal
3
240 ∙ 106
not linear
1
240 ∙ 109
noisy signal
3
165 ∙ 109
noisy signal
1
120 ∙ 103
not linear
3
90
linear
1
47 ∙ 109
noisy signal
3
17 ∙ 106
almost linear
1
500 ∙ 109
noisy signal
3
470 ∙ 106
not linear
1
160 ∙ 109
noisy signal
3
3 ∙ 106
almost linear
1
120 ∙ 109
noisy signal
3
11 ∙ 106
almost linear
1
20 ∙ 103
not linear
3
80
linear
2
3
20
Resin + 3 wt.-% of type 9
4
Voltage (V)
10
5
0
-10
6
Noisy signal
-20
-10
-2,0x10
20
-10
0,0
-1,0x10
-10
Current (A)
7
-10
1,0x10
8
Resin + 1 wt.-% of type 3
2- Current density distribution in
a composite volume
Percolation theory
Fluctuation-induced tunnelling
σ  (p-pc)t
ln σ  -Ap-1/3
D . Stauffer, et al. “Introduction to percolation theory.” Taylor and Francis, London, 1994
9
1
200 ∙
109
noisy signal
B . Kilbride, et al., JAP 92, p. 4024, 2002
0
-1
10
3
0
-1
noisy signal
-3
10
-10
1
3
11
-20
Not linear
-4
-4
-1,8x10
-1,2x10
-5
-6,0x10
0,0
-5
-4
6,0x10
1,2x10
3
12
13
5
14
0
15
-5
Linear
-1,0x10
-2
-5,0x10
0,0
Current (A)
-2
5,0x10
170 ∙ 109
120 ∙ 109
noisy signal
noisy signal
noisy signal
1,8x10
Resin + 3 wt.-% of type 8
-1
190 ∙
109
-1
1,0x10
16
1
500 ∙ 109
3
109
120 ∙
-2
noisy signal
-4
Current (A)
-10
1
160 ∙ 109
noisy signal
noisy signal
1
120 ∙ 109
noisy signal
3
170 ∙ 106
noisy & not linear
1
500 ∙ 109
noisy signal
3
27 ∙ 106
almost linear
1
100 ∙ 109
noisy signal
3
2 ∙ 109
almost linear
1
60 ∙
109
noisy signal
3
8 ∙ 106
almost linear
-5
10
-7
10
-9
10
Exp
Fit
-11
10
pc = 0.36 v.%
t = 1.8
-13
10
-15
10
0,0
0,3
0,6
0,9
Volume (%)
1,2
log (Conductivity)
10
10
100 ∙ 109
Conductivity (S/cm)
Voltage (V)
1- Mesh distribution on the control
volume of the FEM simulation
2,0x10
10
Voltage (V)
Finite Element Method (FEM) simulation was performed, using the commercial code Comsol Multiphysics™, of a
composite material slab characterized by different resistivities, having the same dimensions of real samples, with the aim
of evaluating the volume interested by the higher fraction of current density and estimating the penetration depth of DC
currents into the sample thickness.
An example of the simulation control volume is given in Figure 1, where the tetrahedral mesh of Lagrangian cubic
elements is shown. In Figure 2, the current density is distributed almost in the whole sample, with the exception of the
portions close to the electrodes, where the effective path avoids the sample bottom and edges.
Based on these simulations, the effective electrical path was estimated to be: 3 mm thick (same thickness of the sample),
3 cm width (same width of the sample) and 1 cm long (sample length reduced by the electrode size and dead ends).
-3
-4
-5
Exp
Fit
-6
-7
A = 1.44
-8
3,5
4,0
4,5
5,0
5,5
6,0
6,5
7,0
(Volume fraction)^-1/3
New model in progress...
3D statistical resistor network taking into account:
• tunneling effect between neighboring CNTs;
• CNTs dimensions, structure and orientation.
Conclusions
A detailed electrical characterization, made making use of sophisticated Finite Element Method (FEM) simulations and a careful realization of a measurement setup, allowed to collect
confident estimates for the resistances for each of the samples above described. Several conduction behaviors have been found: from highly conductive NCs, which showed linear Ohmic
curve (i.e. samples 3 and 8), to non-linear diode-like trend up till completely insulating one (R > 109 W).
We have applied physical models such as the percolation theory and the fluctuation-mediated tunneling theory. Parameters extracted from the model fitting allowed us to conclude that
the lowest percolation threshold may be found for our resin. Nevertheless a new conductivity model is needed, which has to take into account for CNTs dimensions and spatial
distribution inside the polymer matrix. Some of the results here reported have been already published in: A. Chiolerio et al (2011) Electrical properties of CNT-based Polymeric Matrix
NanoComposites. In: Yellampalli Siva (ed). Carbon Nanotubes-Polymer Nanocomposites, p. 215-230.