JEC COMPOSITES MAGAZINE - Issue #85 - November/December 2013 - 45

cure monitoring

Monitoring the dielectric cure
of an epoxy-graphite prepreg
Lambient Technologies tested samples of an
epoxy-graphite prepreg using dielectric cure
monitoring to study the material's cure over the
desired temperature profile. This report presents
and discusses Log(ion viscosity) and the slope
of Log(ion viscosity) curves, which indicate the
progress of the cure. The data clearly show cure
of the prepreg despite the presence of conductive
graphite fibres, and the extraction of critical points
identifies characteristic features of the cure such
as minimum ion viscosity, maximum slope of
Log(ion viscosity) and the time to a chosen end of
cure.

By
Huan Lee, Principal Electronics Engineer,
Lambient Technologies
Steve Pomeroy, Principal Software Engineer,
Lambient Technologies

he prepreg was placed on top of a Mini-Varicon sensor
with a paper filter in between. The filter allows epoxy to
flow to the sensor while preventing graphite fibres from
contacting the elecDefinitions
trodes and creating
Loss factor (unitless) is defined as:
a short circuit. The
ε" = σ'/ ω
(Equation 1)
prepreg was then
Relative conductivity (unitless) is defined as: cured using the folσ' = σ/ε0
(Equation 2)
lowing temperature
where:
profile:
σ = conductivity (ohm-1 - cm-1)
- Ramp at 5 °F/min
ε0 = 8.86 x 10-14 F/cm
from room temperaω = 2pf (radians/sec)
ture to 260 ±5 °F
f = frequency of excitation signal
- Hold at 260 °F for
(Hz)
90 minutes +10/-0
Ion viscosity (ohm-cm), also known as
minutes
resistivity, is defined as:
The first prepreg
IV = ρ = 1/σ
(Equation3)
sample was moni-

tored at 1 Hz, 10 Hz, 100 Hz, 1.0 KHz, 10 KHz and 100 KHz at
periodic intervals for the duration of its cure. Lambient Technology's CureView software acquired and stored the data, and
performed post-analysis and presentation of the data.
CureView can extract critical points only for data obtained at
a single frequency, so multiple frequencies had to be used for
the first sample to determine an optimum single frequency for
the second sample. The second sample was monitored at 10 Hz
only.

Results of multi-frequency cure monitoring
Figure 1 shows raw data from the cure of the first sample.
Log(loss factor) curves are plotted for 1 Hz, 10 Hz, 100 Hz, 1.0
KHz, 10 KHz and 100 KHz, with 1 Hz being the uppermost
curve and 100 KHz the lowermost curve.
The heat source for these tests did not have optimum PID
parameters for temperature control, which resulted in visible
temperature fluctuations. In turn, these temperature fluctuations caused small variations in loss factor, but did not affect the
overall quality of the data or the insights which they provide.
The loss factor data are clean with almost no noise, except at
100 Hz and 1 KHz after about 65 minutes. Poor grounding in
the building and high electrical noise in the test area caused this
noise, which would not appear under typical conditions.
The data for 1 Hz and 10 Hz show the effects of electrode
polarization, which occurs when materials are highly conductive. Electrode polarization causes an artificial dip in loss factor
during the early parts of cure, as shown in Figure 1. If not properly understood, these artifacts are confounding and can cause
misinterpretation of the cure process. Fortunately, the effect
of electrode polarization can be removed mathematically as
described in detail in Lambient Technologies Application Note
9: Electrode polarization and boundary layer effects.
After removal of electrode polarization effects, the loss factor
curves ideally will resemble the data presented in Figure 2,
taken from the cure of a typical epoxy, which has also been
corrected for electrode polarization. All loss factor peaks
occur at the same time and represent the mechanical viscosity
minimum. If a frequency-independent (or DC) conductivity
dominates the loss factor, then by equation 1, the loss factor is
inversely proportional to frequency, i.e. the loss factor decreases
by a factor of 10 (1.0 on a Log10 scale) as frequency increases

No85 November - December 2013

/ jec composites magazine 45

JEC COMPOSITES MAGAZINE - Issue #85 - November/December 2013

Table of Contents for the Digital Edition of JEC COMPOSITES MAGAZINE - Issue #85 - November/December 2013