Multifunctional polyetherimide nanocomposite foam

Abstract

Polyetherimide/multi-wall carbon nanotube nanocomposite foams were fabricated and their electrical and mechanical properties characterized. The composites with 0.5–3 wt% of carbon nanotube loadings were fabricated using a solution-based method. Different foaming conditions were studied along with the residual solvent effects. It was found that the residual solvent significantly affected the mechanical properties of the nanocompsites. Electrical conductivity of the composite foams with the same wt% loading of carbon nanotubes was found to be dependent on the relative density of the foams. Foams with a higher relative density had a higher electrical conductivity. The fabricated polyetherimide nanocomposite foam is thermally stable, with a glass transition temperature close to that of the neat polyetherimide. Despite the volume expansion due to foaming, the polyetherimide nanocomposite foams achieved an electrical conductivity of 10⁻⁷ S/cm at 2 wt% carbon nanotube loading and 45% relative density. This makes them a multifunctional material for lightweight and electrostatic dissipative applications.

Keywords

Carbon nanotubes electrical conductivity thermomechanical properties dynamic mechanical thermal analysis nanocomposite foam

Introduction

Polymer nanocomposite foams have attracted much attention due to their improved mechanical, thermal and electrical properties. Early research on polymer nanocomposite foams has focused on improving foam morphology and the mechanical properties. Nanoparticles such as nanoclay, carbon nanofiber and carbon nanotubes (CNTs) were used with a wide range of polymer matrices to increase the nucleation density, thereby reducing the cell size, reinforce the polymer foam and impart other desirable properties such as fire retardancy.¹ Polymer nanocomposite foams with improved electrical conductivity have recently been developed for lightweight electrostatic dissipative (ESD), electromagnetic interference (EMI) shielding and lightning-strike protection applications.²^–⁴ It is well known that a small amount of highly conductive nanoparticles, when well dispersed, could form a conducting network in the polymer matrix, thereby significantly improving the electrical conductivity of the otherwise insulating material. Making these polymer nanocomposites into foams would lead to the formation of a new class of multifunctional materials that are both lightweight and electrically conductive.

Traditionally, ESD and EMI prevention have relied on metal-based materials, either thin sheet metals or metal-coated plastics. These materials suffer from disadvantages of being heavy, prone to corrosion and difficult to process. Conductive polymer nanocomposite foams are preferred compared with metal-based materials, especially for lightweight applications such as those found in aircrafts, spacecrafts and automobiles. The porous structure of these foams also provides an added benefit for EMI shielding due to enhanced magnetic energy absorption by wave scattering.³

Most of the existing polymer nanocomposite foams have low temperature thermoplastic matrices with a glass transition temperature (T_g) ranging between −65 and 105°C and a melting temperature ranging from 65 to 240°C, including polyvinylidene fluoride, poly(methyl methacrylate), polystyrene and low-density polymer ethylene.³^–¹¹ The pore sizes of these foams are from sub-micron to a few hundred micrometers and the porosity ranges between 15% and 85%. Thermosetting foams usually have a larger pore size, sometimes on the scale of millimeters.^12,13 The electrical conductivity varies dramatically among different studies, from 0.1 S/cm to 6 × 10 ⁻ ⁸ S/cm depending on the type and loading of the nanoparticles. In general, a higher loading of the conducting phase contributes to a higher conductivity, and graphene-based materials are more conductive than CNT-based ones.^3,6 Although their electrical conductivities are high enough for ESD or EMI shielding applications, the matrix materials all have a low service temperature and may not be used for high performance applications, such as aircraft interior.

In this study, we fabricate polyetherimide (PEI)/multi-wall carbon nanotube (MWCNT) nanocomposite foams and characterize their electrical and mechanical properties. PEI is a high performance polymer with a T_g of 217°C and a melting temperature of 340°C. Using PEI as the matrix material, the resultant nanocomposite foams could be used in a more stringent environment where high service temperature is required. In addition, PEI performs well with low smoke and low toxicity while it burns. The fabricated composite foams could be used for aerospace and automotive components to satisfy the lightweight, ESD prevention as well as fire safety requirements.

Experimental

Materials

PEI powder (ULTEM 1010 P) was provided by SABIC Innovative Plastics Co. PEI is a semi-transparent thermoplastic material with high strength and stiffness. The molecular structure of PEI is shown in Figure 1. The density and mechanical and thermal properties are shown in Table 1. Functionalized MWCNTs with the carboxyl group (–COOH) were purchased from Cheap Tubes, Inc. Functionalized CNTs were used for better dispersion.¹⁴ The purity of the tubes was greater than 95%; the inner diameter was 3–5 nm, outer diameter was 8–15 nm and length was 10–50 µm. The COOH group content was 2.5%. Dichloromethane (DCM) from Fisher Scientific was used as the solvent to dissolve PEI and suspend MWCNTs. For foaming, industrial grade CO₂ was used as the blowing agent in a solid-state foaming process.

Figure 1.

The molecular structure of PEI.²⁴

Table 1.

Properties of PEI (ULTEM 1010P).

Density (g/ml)	1.27
Tensile modulus (GPa)	3.6
Tensile stress at yield (MPa)	110
Tensile strain at yield (%)	7
Glass transition temperature (°C)	217
Service temperature (°C)	170
Processing temperature (°C)	350
Thermal conductivity (W/m °C)	0.22

PEI: polyetherimide.

Sample preparation

PEI/MWCNT nanocomposites of 0.5–3 wt% with 0.5 wt% loading intervals were prepared using the solvent casting method. PEI powder and MWCNTs were separately added to DCM of equal amounts. Probe-type sonication (VC 750, Sonics and Materials Inc, USA) was applied at room temperature for 1 h with a power level of 150 W in order to obtain a homogeneous DCM/CNT suspension. The PEI solution was then added to the DCM/MWCNT suspension. The mixture was stirred with a magnetic stirrer for 30 min at room temperature. The solution was probe sonicated at 150 W for 1 h. The resultant PEI/MWCNT/DCM solution was cast in a Teflon-coated metal mold to go through solvent extraction steps. The thickness of the obtained nanocomposite samples after solvent extraction ranged from 300 to 400 µm.

Solvent extraction

Two steps of solvent extraction were performed and the weight of samples was recorded after each step. The cast samples were first placed in a vacuum oven (MTI, USA) at 100°C for 3 days. They were then saturated in supercritical CO₂ at 8 MPa and 35°C for another 3 days. In the latter step, the supercritical CO₂ introduced a plasticizing effect in PEI such that the residual DCM molecules could diffuse out easily.¹⁵^–¹⁷ CO₂-saturated samples were subsequently placed in the vacuum oven for 1 day at 100°C to extract the remaining CO₂ molecules.

Foaming of PEI/MWCNT nanocomposites

The PEI/MWCNT nanocomposites were foamed using a solid-state foaming method. Solid-state foaming is a two-stage process. The sample was first saturated in high-pressure CO₂ using a syringe pump (Teledyne ISCO 260D, USA) for 24 h and then foamed using hot platens. A detailed description of this procedure can be found in Zhou et al.¹⁸ Two groups of experiments were conducted, yielding foamed samples with different porosities. The experimental conditions are shown in Table 2. At least two samples were foamed under each condition. The sample size before foaming was approximately 30 × 30 mm and the thickness was between 300 and 400 µm.

Table 2.

Foaming conditions.

Conditions	Group 1	Group 2
Saturation pressure (MPa)	8	8
Saturation time (h)	24	24
Saturation temperature (°C)	35	20
Gas concentration (%)	∼11	∼13
Desorption time (min)	15	3
Foaming temperature (°C)	175	195
Foaming time (s)	20	1

Material characterization

Direct current (DC) electrical conductivities (both volume and surface) were measured with a Multimeter (HHM93, Omega, USA) for low resistivity and a Mega-Ohmmeter (Super Mega-Ohmmeter, Hioki 8220, Japan) for high resistivity with a set of high-resolution electrodes (SME 8311, Hioki, Japan). Densities of the samples were obtained using the following equation,

Density = \frac{Wa}{Wa - Ww} (0.9975) {g / cm}^{3}

(1)

where w_a is the sample weight in air and w_w is that in water. The ratio between the foamed and unfoamed sample densities is defined as the relative density of the foam. Microstructures of the samples were examined with a Quanta 650 FEG scanning electron microscope (SEM, FEI Company, USA). The samples were freeze-fractured in liquid nitrogen and the exposed foam surface was sputter coated with gold and palladium (Au/Pd) for imaging. Mechanical properties were tested with a TA Q800 dynamic mechanical analyzer (DMA) (TA Instrument, USA) with a heating rate of 3°C/min and a constant frequency of 1 Hz.

Results and discussion

Solvent extraction

The amount of solvent extraction by vacuum oven for 5 days was about 7% of the total sample weight for all the samples prepared. However, bubble formation was still observed when the temperature was raised to the T_g of pristine PEI. This suggested that there was still a considerable amount of residual solvent left in the sample. To further remove the residual solvent, 3 days of supercritical CO₂ solvent extraction was performed after 3 days of vacuum oven extraction. The sample weight reduction results are shown in Figure 2. It is seen that the average total weight reduction after both the vacuum oven and supercritical CO₂ solvent extraction was 13%, showing that the supercritical CO₂ solvent extraction was able to reduce the residual solvent by at least another 6% after the vacuum oven treatment.

Figure 2.

Solvent extraction results at various MWCNT loading levels (3 days of vacuum oven and 3 days of supercritical CO₂ solvent extraction).

The effects of residual solvent on the mechanical property of the fabricated nanocomposite samples were examined with DMA tests. The results are shown in Figure 3. For samples only treated with vacuum oven solvent extraction, the T_g was found to be around 170°C, much lower than that of pristine PEI, which is 217°C. The T_g dropped by almost 50°C, indicating the significant negative effect of residual DCM. With the supercritical CO₂ solvent extraction, the T_g of PEI returned to 215°C. From the DMA data, it is also seen that the existence of residual solvent lowered the storage modulus by about 20%. Kumar et al.¹⁹ fabricated PEI/CNT nanocomposites using the solvent casting method with 4 h vacuum extraction of the residual solvent at 60°C. For neat PEI samples that were fabricated, the average storage modulus was found to be 1.5 GPa at 50°C and the T_g was 170°C. These values were significantly lower compared with those found in this study.

Figure 3.

DMA results from neat PEI showing the residual solvent effect. Multiple tests were conducted, figure showing typical results.

Volume DC conductivities of the fabricated nanocomposites samples were measured and the results are shown in Figure 4. It is seen that regardless of the solvent extraction method, a percolation threshold existed between 0.5 and 1 wt% of CNT loading. The amount of residual solvent had little effect on the electrical conductivity of the fabricated samples.

Figure 4.

Volume DC electrical conductivity of nanocomposites with various MWCNT loadings and residual solvent levels.

Residual solvent is undesirable in nanocomposites since it has a plasticizing effect on mechanical strength.^19,20 In our study it was found that the residual solvent has another undesirable effect, i.e. it could act as a blowing agent, making the foaming process hard to control. Therefore, the residual solvent needed to be reduced to a minimum level after the solvent casting process, even though it does not affect the electrical conductivity of the nanocomposite samples.

Foam morphology

SEM images of unfoamed and foamed samples with two different foaming conditions at various MWCNT loading levels are shown in Figure 5. From unfoamed nanocomposite samples, it is clear that MWCNTs were well dispersed throughout the polymer matrix. For the morphology of foamed samples, different cell sizes were found with different foaming conditions, especially for neat PEI samples. Relative densities of nanocomposite foams were about 45% for Group 1 samples and 75% for Group 2 samples. The fact that Group 2 samples had a smaller cell size is due to a number of reasons, including a higher gas concentration, a higher foaming temperature and a shorter foaming time. Group 2 samples were saturated at room temperature, where the solubility of CO₂ is higher than that at an elevated temperature. The gas concentration of Group 2 samples was thus higher than that of Group 1 samples after full saturation. A higher gas concentration and foaming temperature will yield a higher nucleation density. Therefore, more and smaller bubbles were created in Group 2 samples. Since the foaming time of Group 1 samples was much longer than that of Group 2 samples, bubbles grew bigger and led to a lower relative density for Group 1 samples. Within Group 2 samples, the foaming characteristics of neat PEI (0 wt%) and PEI nanocomposites were significantly different. Neat PEI had a large amount of small pores, which is consistent with our previous results,²¹ while the composites had much larger pores. It is believed that the difference is caused by the residual solvent that was left in the composite samples even after the rigorous solvent extraction process. DCM at the foaming temperature is a blowing agent and may have contributed to the generation of larger pores.

Figure 5.

Unfoamed and foamed samples with various MWCNT levels. Scale bars: 2 µm

Electrical properties of PEI nanocomposites and foams

Volume electrical conductivities of the PEI/MWCNT composites and their foams are shown in Figure 6. Dramatic increase in electrical conductivity was found for both the composites and foams between 0.5 and 1 wt% of MWCNT loading. To determine the percolation threshold (φ_c) for both unfoamed and foamed samples, the following power law relationship is used.²²

σ \propto (φ - φ_{c})^{t}

(1)

where σ is the electrical conductivity, φ is the MWCNT weight fraction and t is the critical exponent depending on the system dimensionality. The critical exponent t takes a value of 1.33 for a two-dimensional system and 2 for a three-dimensional system. In this research t = 2 was used. Log–log plots of conductivity versus reduced mass fraction for unfoamed and foamed samples (Group 1) were generated, as shown in Figure 7(a) and (b). Good linear fits were found with the log–log data. The percolation threshold φ_c for unfoamed samples was determined to be 0.6 wt% and that for Group 1 foamed samples was 0.9 wt%. Group 2 samples were not analyzed due to the lack of data points. In general, foaming shifted the percolation threshold slightly higher in terms of the wt% loading. On the other hand, electrical conductivities of foamed samples were relatively lower than those of unfoamed samples. The difference could be as high as 100 times. This was because of the fact that some of the conducting paths in the composites were broken due to the volumetric expansion caused by foaming. For Group 2 samples having a higher relative density, the conductivity reduction was less than 10 S/cm compared with the unfoamed ones. The volumetric expansion in this case was not as high, since the relative density is 75% and the majority of conducting network were retained in the polymer matrix. Overall, the volume electrical conductivity of both groups of foamed samples with more than 2 wt% MWCNT loading was high enough for the electrostatic charge dissipation application, according to the American National Standard Institute (10 ⁻ ¹¹S/cm required).²³

Figure 6.

Volume DC electrical conductivity of nanocomposites and foams with various MWCNT loading levels.

Figure 7.

(a) A log–log plot of conductivity versus reduced mass fraction for unfoamed samples. (b) A log–log plot of conductivity versus reduced mass fraction for Group 1 samples.

The results of surface electrical conductivity measurements are shown in Figures 8 and 9. Similar to the volume electrical conductivity, foaming did not shift the percolation threshold significantly; however, foamed samples had a lower electrical conductivity compared with the unfoamed ones. The electrical conductivity depended on the relative density of the foamed samples. Group 1 samples (45% relative density) had a surface electrical conductivity 100 times lower than that of unfoamed ones. Group 2 samples (75% relative density) had it 10 times lower. The reduced electrical conductivity was also attributed to the surface expansion caused by foaming. Therefore, a lower degree of foaming maintained a higher surface conductivity and vice versa. A notable characteristic in this experiment is that the electrical conductivity of the bottom surface for both composites and foams was higher than that of the top surface by a factor of 10. This is believed to be due to the gravitational effect of MWCNTs in the DCM/PEI/MWCNTs fabrication process. More MWCNTs fell at the bottom of the samples, which allowed more chances to form a conducting network at the bottom surface. The surface electrical conductivity values of both groups of foams were in the range of the electrostatic charge dissipation applications when the MWCNT loadings were greater than the 1.5 wt% percolation threshold.

Figure 8.

Surface DC electrical conductivity of nanocomposites and Group 1 foams with various MWCNT loading levels.

Figure 9.

Surface DC electrical conductivity of nanocomposites and Group 2 foams with various MWCNT loading levels.

Mechanical properties of composites and foams

The DMA results of both unfoamed and foamed samples are shown in Figure 10(a) to (c). Storage moduli at 50°C were normalized against the relative density of the samples, where the storage modulus was divided by the relative density of the sample. The storage moduli of unfoamed nanocomposites varied from 2.3 to 2.5 GPa regardless of the MWCNTs loading, which indicated that the addition of MWCNTs did not significantly improve the mechanical properties of PEI with the loadings tested in this study. However, the storage modulus of neat PEI obtained in this research was comparable or higher than those reported in other studies. For example, Liu et al.²⁴ reported a storage modulus of 2 GPa at 50°C using a casting and imidization process. The storage modulus of neat PEI reported by Kumar et al.¹⁹ was 1.5 Gpa. The normalized storage modulus of the foamed samples averaged 1.5 and 1.8 GPa for Group 1 and Group 2, respectively, and both were lower than that of unfoamed samples. However, comparing samples with similar densities, the MWCNTs reinforced foams on average showed a higher storage modulus than did neat PEI foams. This effect is more pronounced in Group 2 samples, which had a higher relative density (75%).

Figure 10.

(a) DMA results of unfoamed PEI/MWCNT composites (only showing neat, 1.5 wt% and 3 wt% samples). (b) Normalized storage and loss moduli of PEI/MWCNT composite foams (Group 1). (c) Normalized storage and loss moduli of PEI/MWCNT composite foams (Group 2).

In general, T_g of all the samples was found to be around 220°C, which is almost identical as the reference value of neat PEI (217°C). This shows that compounding and foaming of PEI in this study did not change the thermal stability of the neat PEI matrix. Therefore, combining the reduced density and improved electrical conductivity, the PEI nanocomposite foams fabricated in this study could be used for low-density electrostatic dissipation applications in high temperature and high mechanical stress conditions.

Conclusions

PEI/MWCNT nanocomposite foams were fabricated and their electrical and mechanical properties characterized. Residual solvent was a concern affecting the mechanical property and the foaming characteristics of the nanocomposites. The percolation threshold of PEI/MWCNT composites was found to be between 0.5 and 1 wt% for the volume electrical conductivity. Foaming did not shift the percolation threshold but reduced the electrical conductivity. The electrical conductivity of the composite foams depended on the relative density. Foams with a higher relative density had a higher electrical conductivity. The fabricated PEI nanocomposite foam was thermally stable, with a T_g that was almost the same as that of the original PEI. The storage moduli of MWCNT-reinforced foams on average were higher than that of the neat PEI foams with a similar relative density. With the high performance polymer matrix, the nanocomposite foams fabricated in this study could be used for low-density electrostatic dissipation applications in high temperature and high mechanical stress situations.

Footnotes

Funding

This work was supported by the National Science Foundation through Grant CMMI-0728287/CMMI-1062226.

Acknowledgements

The authors are grateful to Dr Edward Kung from SABIC Innovative Plastics and Dr Samra S Sangari from Boeing for helpful discussions.

References

Lee

Zeng

Cao

. Polymer nanocomposite foams. Compos Sci Technol 2005; 65: 2344–2363.

Huang

. Electromagnetic interference (EMI) shielding of single-walled carbon nanotube epoxy composites. Nano Lett 2006; 6: 1141–1145.

Zhang

Yan

Zheng

. Tough graphene− polymer microcellular foams for electromagnetic interference shielding. Appl Mater Interfaces 2011; 3: 918–924.

Thomassin

Pagnoulle

Bednarz

. Foams of polycaprolactone/MWNT nanocomposites for efficient EMI reduction. J Mater Chem 2008; 18: 792–796.

Huynen

Quiévy

Bailly

. Multifunctional hybrids for electromagnetic absorption. Acta Mater 2011; 59: 3255–3266.

Eswaraiah

Sankaranarayanan

Ramaprabhu

. Functionalized graphene–PVDF foam composites for EMI shielding. Macromol Mater Eng 2011; 296: 894–898.

Chen

Ozisik

Schadler

. The influence of carbon nanotube aspect ratio on the foam morphology of MWNT/PMMA nanocomposite foams. Polymer 2010; 51: 2368–2375.

Zeng

Hossieny

Zhang

. Synthesis and processing of PMMA carbon nanotube nanocomposite foams. Polymer 2010; 51: 655–664.

Rizvi

Kim

Naguib

. Synthesis and characterization of novel low density polyethylene–multiwall carbon nanotube porous composites. Smart Mater Struct 2009; 18: 104002–104002.

10.

Yang

Gupta

Dudley

. Novel carbon nanotube- polystyrene foam composites for electromagnetic interference shielding. Nano Lett 2005; 5: 2131–2134.

11.

Yang

Gupta

Dudley

. Conductive carbon nanofiber–polymer foam structures. Adv Mater 2005; 17: 1999–2003.

12.

Shi

. Ultralight conductive carbon nanotube–polymer composite. Small 2007; 3: 408–411.

13.

Fletcher

Gupta

Dudley

. Elastomer foam nanocomposites for electromagnetic dissipation and shielding applications. Compos Sci Technol 2010; 70: 953–958.

14.

Sahoo

Rana

Cho

. Polymer nanocomposites based on functionalized carbon nanotubes. Prog Polym Sci 2010; 35: 837–867.

15.

Sasaki

Takishima

Masuoka

. Supercritical carbon dioxide extraction of benzene in poly (vinyl acetate) and polystyrene (Part 2). Sekiyu Gakkaishi 1990; 33: 304–310.

16.

Koegler

Patrick

Cima

. Carbon dioxide extraction of residual chloroform from biodegradable polymers. J Biomed Mater Res 2002; 63: 567–576.

17.

Wang

Kramer

Sachse

. Effects of high pressure CO₂ on the glass transition temperature and mechanical properties of polystyrene. J Polym Sci: Polym Phys Edn 1982; 20: 1371–1384.

18.

Zhou

Wang

. Fabrication of functionally graded porous polymer via supercritical CO₂ foaming. Compos Part B: Eng 2011; 42: 318–325.

19.

Kumar

Caceres

. Dramatic property enhancement in polyetherimide using low-cost commercially functionalized multi-walled carbon nanotubes via a facile solution processing method. Nanotechnology 2009; 20: 465708–465708.

20.

Fried

. Polymer science and technology, Upper Saddle River, NJ: Prentice Hall PTR, 1995.

21.

Zhou

Vaccaro

Sundarram

. Fabrication and characterization of polyetherimide nanofoams using supercritical CO₂. J Cell Plast 2012; 48: 239–255.

22.

Bauhofer

Kovacs

. A review and analysis of electrical percolation in carbon nanotube polymer composites. Compos Sci Technol 2009; 69: 1486–1498.

23.

American National Standard Institute (ANSI/ESD) S541-2008. 2008.

24.

Liu

Tong

Zhang

. Preparation and characterization of carbon nanotube/polyetherimide nanocomposite films. Compos Sci Technol 2007; 67: 406–412.