Dielectric Response of Various Partially Cured Epoxy ... - IEEE Xplore

4 downloads 61 Views 477KB Size Report
4 Nottingham Trent University, UK. Abstract: The purpose of this work is to demonstrate the post- treatment importance using dielectric measurements on.
Dielectric Response of Various Partially Cured Epoxy Nanocomposites I. Preda 1,2, H. Couderc 2,3, M. Fréchette 2, S. Savoie 2, F. Gao 4, R. Nigmatullin 4, S.Thompson 4, J. Castellon 1 1

Institut d’Électronique du Sud, Université Montpellier 2, France 2 Institut de Recherche d’Hydro-Québec (IREQ), Canada 3 École de Technologie Supérieure, Université du Québec, Canada 4 Nottingham Trent University, UK

curing and of a post-curing treatment upon dielectric results. Also, the influence of nanoparticles on the post heat treatment will be considered.

Abstract: The purpose of this work is to demonstrate the posttreatment importance using dielectric measurements on nanocomposites. During this post-treatment which is the application of heat over time, crosslinking completes and ageing sets in. Both effects usually take place in the same time. Besides the expected differences between the filled and unfilled epoxy samples, the influence of silane coupling agents and the use of ultrasonic waves on the clay was demonstrated. Since the samples used in the experiment were unusually soft, supposedly because of the insufficient initial curing (4h at 100°C), a total of 38 h of heat treatment was applied, after which an improvement of the mechanical rigidity was noticed. This heat treatment was also considered extremely useful in order to remove humidity from the samples volume. Comparing the results obtained for two samples, with and without the 38h of post treatment, the treated sample was found to have a more stable dielectric response at 20°C.

Montmorillonite (MMT) is a crystalline, layered clay mineral with a central alumina octahedral sheet sandwiched between two tetrahedral silica sheets. The electrostatic bonds between the resulting sheets are done by sodium or magnesium ions in the natural state. A partial exchange of the sodium ions with organic ions is carried out in order to make the surface hydrophobic to obtain exfoliated structures [4]. The chosen organoclay – Cloisite 30B (C30B) – is a natural MMT modified with a quaternary ammonium ion (MT2etOH methyl, tallow, bis-2-hydroxyethyl, quaternary ammonium). Clays as MMT are naturally hydrophilic and absorb water from its environment. Considering the exchange of sodium ion with organic ions is partial, an organoclay can still absorb water molecules [5]. The polymer – particle interface is well known as a potential location for water [6]. Thus, nanocomposites, because of a high number of interfaces, are particularly vulnerable to the water uptake and its consequences. A post-heat treatment was proven to have a substantial effect on water removal and consequently on the dielectric properties of epoxy-organoclay nanocomposites [79]. This is due mainly to the fact that it contributes to the end of the curing reaction and to the removal of residual chemical products [4].

I. INTRODUCTION Epoxies are thermoset polymers in which the chains are ended by an epoxy group. A curing agent is used to open this epoxy ring and to interconnect the polymer chains. Usually, epoxy resins are prepared from low molecular weight liquid resins that are transformed into high molecular weight polymers by curing. Since they have low viscosity during shaping, thermosets have favourable processing properties, which, in combination with their generally good performance, had resulted in a large variety of applications in mechanical and dielectric field. As the curing reaction proceeds [1, 2], the resin goes through several stages going from liquid to gelation (when the material will exhibit viscoelastic properties) and ending with its rigidification (when the network is hardened and the expected mechanical rigidity is obtained). Curing reactions are known to be strongly exothermic. Also, knowing that a too long curing process or a too high curing temperature could damage the material, a low temperature is usually used, as the temperature within the material will increase as well. Consequently, curing needs to be completed by a posttreatment at higher post-cure temperature [3].

II.

EXPERIMENTAL

A. Materials and Sample Preparation The epoxy resin used in the study was diglycidyl ether of bisphenol A based resin, commercially named DDR332. The selected hardener was poly(propylene glycol) bis(2aminopropyl ether) (PPG-b-AP). (3-Glycidyloxypropyl) triethoxysilane (GOPTES) was used as a coupling agent. As mentioned before, the used organoclay was Cloisite 30B. To further improve the compatibility between the clay and epoxy, the clays were treated using silane GOPTES. 15 g of clay were dispersed in 600 ml of distilled water by mixing using magnetic stirring for 3 hours. 100 g of 10 wt% solution

In this study, several partially cured organoclay/epoxy nanocomposite samples, with or without additional clay treatment, were used in order to demonstrate the effect of a complete

978-1-4577-0986-9/11/$26.00 ©2011 IEEE

1 660

III. RESULTS

of GOPTES in distilled water were stirred for 1.5 hours to initiate silane hydrolysis at room temperature. The solution was added into the clay suspension and the resulting mixture was stirred for 48 hours at room temperature to obtain surface modified clay. The modified clays were separated from the solution by filtration and finally dried using freeze drying.

Considering the sample’s initial soft aspect, it was presumed that the initial curing – 4 hours at 100ºC – was insufficient and that the crosslinking process was not complete. Several tests were carried out using neat epoxy test samples in order to find the best way to end the crosslinking process and also to remove the contained water. Several heat-treatments were tested and promising results were found when the samples were cured at 100ºC for 38 h. After the performed heat treatment, the samples were in a stable but aged dielectric state.

The stoichiometric ratio of epoxy/NH2 groups suggests a content of 34 g of PPG-b-AP for 100 g of DDR332. In order to produce under-cured samples, all samples were produced using 64 g of PPG-b-AP for 100 g of DDR332. Glass surfaces were coated with a mould release agent. Following this process, the samples were cured at 100-105°C for 4 hours.

The evolution of the imaginary permittivity versus frequency during the curing of the neat epoxy sample is shown in Figure 1. All the samples showed the same behavior. It can be seen that the samples show an important conduction phenomenon at low frequencies and a part of the left side of the α polarization peak at high frequencies. It can also be seen in the insert that as the post heat treatment moves on, the values of the low frequency conductivity are decreasing. As this decrease is slowing down with time, it can be deduced that the material tends to reach its stable dielectric state.

For one particular sample containing Cloisite 30B, the preparation procedure was altered in order to investigate the influence of ultrasonic treatment (sonication). The required amounts of hardener and Cloisite 30B were premixed using a glass rod and sonicated for 1 min at full amplitude using an ultrasonic processor equipped with microtip. The clay content for all the composites was 2 wt%. Except during transportation, the samples were kept in a controlled dry and UV proof environment, at room temperature.

10

eps"

B. Broadband Dielectric Spectroscopy

10

The frequency range that was used for all the experiments was 10-1 to 106 Hz. For all the measured values, a 0.03 mrad dissipation factor resolution is considered. Samples were disks with a thickness of around 1 mm and a radius of around 20 mm. The samples were studied in a parallel plate configuration. Also, 50 µm thick silver foil electrodes were used to ensure a good contact between the sample and the measuring instrument’s brass electrodes. A sinusoidal voltage of 1 Vr m s was applied and, for each point, the final values for the obtained parameters represent the mean of 10 consecutive measurements. All measurements were made under nitrogen with a ±0.01ºC deviation in temperature.

10

4.8

1h 5h 10h 15h 20h 25h 30h 35h 38h

Increasing time

4.6

4.4 -1

10

frequency(Hz)

α peak

Fig.1: Imaginary permittivity versus time during the 38h curing at 100ºC for the neat epoxy sample.

Using BDS measurements, the real and the imaginary part of the complex dielectric permittivity were acquired. The complex dielectric function that connects the two parts is given by the following formula:

with f the frequency of the applied voltage, ε' the real or conservative part and ε'' the imaginary or loss part. For all the available samples, the following measurement protocol was used: an initial characterization was performed at 20°C, followed by 38 hours of post heat treatment at 100°C, as it is explained below, and finally another set of characterization measurements performed at 20°C and at 100°C. Fig. 2: Imaginary permittivity evolution during the 38h curing at 0.1Hz.

2 661

α peak α peak

β peak

Fig. 3: Imaginary permittivity spectra at 20°C before and after the treatment, and spectra at 100°C after the first and the last hour of the treatment. for the neat epoxy sample

Fig. : 4. Real permittivity of the epoxy nanocomposite samples measured at 20°C after 38h at 100°C

The imaginary permittivity at 0.1Hz of all tested samples, with and without nanoclay, during the 38 h curing can be seen in Figure 2. The imaginary permittivity shows for all samples an usual decrease during the heat treatment. The initial exponential considerable decrease suggests that two phenomena are taking place: completion of the crosslinking process and water removal. After approximately 20 h of heat treatment, the decrease becomes linear and slows down, which means that the first phenomenon – the sample’s crosslinking – is complete. The water removal continues for approximately 20 hours. In order to emphasize the effect of the 38 h curing, the imaginary permittivity spectra at 20°C before and after the treatment for the neat epoxy sample are shown in Figure 3. Spectra at 100°C after the first and the last hour of the treatment are also plotted. As shown in the first two figures, during the process, the imaginary permittivity has an important decrease at low frequencies. At 20°C, one decade decrease is observed at low frequencies. So, the low frequency conductivity is reduced, mainly as a consequence of the water removal. The high frequency part, which is the local β relaxation peak, is not affected by the treatment. As in epoxy, the β relaxation is characteristic from crankshaft motions of the hydroxylether groups [10], located along the molecular chains, it can be deduced that the treatment have no influence on the local motions. In the same time, there is a significant change in the imaginary part for low frequencies, which characterizes the conductivity phenomenon – visible on both 20°C and 100°C measurements.

Fig. 5: Imaginary permittivity of the epoxy nanocomposite samples measured at 20°C after 38h at 100°C

At this point, their dielectric responses could be used in the characterization process. The real and imaginary parts of the complex dielectric permittivity, obtained at 20°C after the post-heat treatment process, are shown in Figures 4 and 5. As shown in Figure 4, the real permittivity of the samples has similar values at high frequencies, but for low frequencies a 10% difference can be seen between the epoxy + Cloisite 30B sample and the nanocomposite sample with ultrasound treatment used in preparation. It can be observed that the samples containing organoclay – except the one where sonication was used during preparation – have a lower permittivity compared to the neat-epoxy. Similar results have been found in the literature [11]. The neat epoxy and the silanised sample have similar real permittivity values at 20°C after the heat treatment.

As mentioned before [7], the post-heat treatment is well known to have an influence on the dielectric response stability of epoxy materials. So, once the heat treatment was performed, all the samples were considered to be in a stable but aged dielectric state.

Two different relaxation peaks are shown in Figure 5. The high frequency peak corresponds to β relaxation and represents the dipole movements of lateral chains of the epoxy

3 662

-

polymer. The low frequency part corresponds to α relaxation high frequency wing peak and is associated with the main chain movement before the glass transition occurs (around 40°C for these samples).

the filled samples are less reticulated than the neat epoxy sample because the fillers block the molecular movement of the polymeric chains.

The applied ultrasound treatment does not have a significant influence over the post heat treatment’s length, even though it produces an exfoliated structure.

As shown in Figure 5, the 38 h treatment seems to be most effective on the C30B untreated epoxy sample. The neat epoxy and the silane-sonicated sample have similar values at 100°C after the heat treatment. These results can also be found at 20°C, once the treatment is complete. Epoxy/C30B nanocomposite has the lowest real and imaginary permittivity values at 0.1Hz (4.08 and 0.07, respectively).

ACKNOWLEDGMENT Financial support from the ANASTASIA European Project (Advanced NAno-Structured TApeS for electrotechnical high power Insulating Applications) is acknowledged. Also, the support from the Nanodielectric program at Hydro-Québec’s Research Institute (IREQ) is gratefully acknowledged.

IV. DISCUSSION It was previously shown in [6] that the epoxy matrix in the micro-filled systems absorbs the same amount of water than the pure epoxy, for a given humidity. In the same time, an extra amount of water can be located around the nanoparticles in the interface zone [11-12]. Water could be linked in organoclay in two forms : bounded at the untreated sites at the surface (like in clay) or trapped in the organoclay stacks [5]. An additional clay functionalization such as silane surface treatment is known to make the clay’s surface more hydrophobic [13]. This characteristic is confirmed by the results presented in Figure 2. Before the heat treatment, the neat epoxy sample presents lower dielectric loss than the nanocomposite sample. Firstly, this could indicate that the organoclay samples are less reticulated that the neat epoxy one because the C30B filler could prevent crosslinking. Secondly, as previously discussed, more water is present in the filled samples and during the heat-treatment, the water contained in the clay will migrate from the C30B layers to the epoxy matrix, before being removed from the sample [4]. As expected, the untreated C30B sample has the highest dielectric loss when the post-heat treatment begins. The initial difference between the sonicated and unsonicated samples could be explained by induced the improvement of dispersion, leading to thinner C30B stacks [14].

REFERENCES [1] [2]

[3] [4]

[5]

[6]

[7]

[8]

[9]

V. CONCLUSIONS

[10]

In this work, the importance of post-treatment upon dielectric response in the frequency domain has been shown using epoxy nanocomposite samples. Several post-heat treatments were used in order to complete the insufficient initial curing and also to remove water. At the end of the treatment, the samples were in a stable but aged dielectric state.

[11]

[12]

[13]

The samples react differently at the post-heat treatment because: - the filled and unfilled samples do not contain the same amount of water knowing that the filled samples have a water layer around the Cloisite 30B stacks;

[14]

T. Kaiser, “Highly crosslinked polymers”, Prog. Polym. Sci., vol.14, 1989, pp. 373-450. M.T. Aronhime, John K. Gillham, “Time-temperature-transformation (TTT) cure diagram of thermosetting polymeric systems”, Adv. Polym. Sci., vol. 78, 1986, pp. 83-113. D. Mulligan, “Cure monitoring for composites and adhesives”, Rapra Review Reports, vol. 14, 2003, pp. 3-4. H. Couderc, M.F.Fréchette, S.Savoie, E.David, , “Nanofiller effect during post-heat treatment of micro-loaded epoxy”, Proceedings IEEE Conference on Electrical Insulation and Dielectric Phenomena, 2010, pp. 608-611. W. Liu, S.V. Hoa, M. Pugh, “Water uptake of epoxy-clay nanocomposites: Experiments and model validation”, Compos. Sci. Technol, vol. 68, 2008, pp. 2066-2072. C. Zou, J.C. Fothergill, S.W. Rowe, “The effect of water absorption on the dielectric properties of epoxy nanocomposites”, IEEE Trans. Dielectr. Insul, vol. 15, 2008, pp. 106-117. M.F.Fréchette, E.David, H.D. Martinez, S.Savoie, “Post-Heat treatment effect on the dielectric response of epoxy samples”, Proceedings IEEE Conference on Electrical Insulation and Dielectric Phenomena, 2009, pp. 705-709. M.F Fréchette et al,. “Nanostructured polymer microcomposites: A distinct class of insulating materials”, IEEE Trans. Dielectr. Insul, vol. 15, 2008, pp. 90-105. M.F. Fréchette, H.D. Martinez, S ; Savoie, A.Krivda, L.E Schmidt, D.Zegarac, “Dielectric study with epoxy-based nanostructured microcomposites containing Silica”, Conference Record IEEE International Symposium Electrical Insulation, 2010, pp. 1-4,. J.Y. Lee, Y.W. Song, S.W. Kim, H.K. Lee, “Dielectric and molecular dynamic studies of sub-Tg by thermally stimulated current (TSC) analysis for liquid crystalline epoxy thermosets”, Mat. Chem. Phys, vol. 77, 2002, pp. 455-460. R. Sarathi, R.K. Sahu, P. Rajeshkumar, “Understanding the thermal, mechanical and electrical properties of epoxy nanocomposites”, Mat. Sci. Eng., vol. 445–446, 2007, pp. 567–578. C Zou, J.C. Fothergill, S.W. Rowe, “A «Water Shell» model for the dielectric properties of hydrated silica-filled epoxy nano-composites” Conference Record IEEE International Conference on Solid Dielectrics, 2007, pp. 389-392. X.Huang, Y.Zheng, P.Jiang, Y.Yin, “Influence of nanoparticle surface treatment on the electrical properties of cycloaliphatic epoxy nanocomposites” IEEE Trans. Dielectr. Insul, vol. 17, 2010, pp. 635643,. M.Kurimoto et al, “Permittivity characteristics of epoxy/alumina nanocomposite with high particle dispersibility by combining ultrasonic wave and centrifugal force”, IEEE Trans. Dielectr. Insul, vol. 17, 2010, pp. 1268-1275.

4 663