Journal of Minerals & Materials Characterization & Engineering, Vol. 8, No.1, pp 15-24, 2009
jmmce.org Printed in the USA. All rights reserved
15
Nanoporous Structure and Enhanced Thermal Stability of a
Polyimide/Single-Walled Carbon Nanotube Composite
Gerard T. Caneba
Department of Chemi c a l E ng i n e e r in g a n d Center for Environmentally Benign
Functional Material s , Mi ch ig an T e chnological Unive rsi ty, Houghton, MI 49931
Email: caneba@mtu.edu
ABSTRACT
Polymer composites based on polyimides and single-walled carbon nanotubes are of interest in
space and aerospace industries, due to inherent thermal stabilities as thermal/electrical
conductivities of the components. Unfortunately, it is very difficult to disperse the carbon
nanotubes (CNTs) into the polyimide matrix. This work attempts this by predispersion of the
CNTs in the precursor reactive materials, and then polymerization is carried out with the CNTs
in place. When this is done, some unusual behavior has been observed, such as the formation of
nanoporous composite structure and enhanced thermal stability of the products.
Keywords: Polyimide, Single-walled Carbon Nanotubes, HiPco, nanoporous structure, thermally
stable polymer composite
1. INTRODUCTION
High performance polymers, such as polyimides, epoxies, poly(tertrafluoroethylene),
polycarbonate, etc. are being used in aerospace and space systems. Carbon nanotubes are of
interest as lightweight materials with enhanced mechanical, thermal, and electrical properties.
For example, their continuous fibers have been shown to be stronger per weight compared to
stainless steel (up to 100 times stronger) and Kevlar (14 times stronger ). Also, they can posses
16 G.T. Caneba Vol.8, No.1
electrical and thermal conductivities bet te r t h a n Copper. Within a polymeric matrix, thermal
diffusivities can be at least 30 times that of th e n eat polymer. It is also reasonable to expect that
layers of carbon nanotubes can act as excellent diffusion barrier materials, based on tight packing
of fibers in the nanoscale. It is therefore the aim of this proposed program to formulate
composites of various carbon nanotubes with high performance polymers that will possess
enhanced mechanical, electrical and th e rmal conductivity proper t ie s . At the same time these
composites could possess special surface characteristics, such as radiation protection and
monitoring barriers, mass diffusion barriers, tribo-m e cha nica l p rope rti es, etc .
In this work, we introduce a procedure to produce polyimide/single walled carbon nanotube
composite as nanoporous materials, which exhibits enhanced thermal stability properties. The
polyimide we produced came from the condensation polymerization reaction of dianhydrides and
a diamine. An intermediate product, poly(amic acid), was formed. Then, imidization was done
by water removal from the intermediate.
2. EXPERIMENTAL
The dianhydride (Pyromellitic dianhydride, PMDA) and diamine (3,3’-Oxydianiline, ODA) were
dried under vacuum overnight at 110oC. Then, 50g of a 0.1wt% solution of HiPco in NMP was
prepared by adding 0.05g of HiPco to 50g of NMP in a capped 70ml test tube. It was best to use
conical caps for these test tubes because the NMP tends to dissolve the glue that hold in the foam
seal on the non-conical caps. Once the 0.1wt% HiPco/NMP solution was made up it needs to be
sonicated. Before s on i c at i on , th e caps on the tubes were tightly sealed, since the vibrations from
sonication can cause a loose cap to come off. The HiPco/NMP solution was sonicated until
nanotubes appeared to be uniformly distributed throughout the NMP. After sonication was
complete, the dried PMDA was added to the HiPco/NMP solution and the sonication step was
repeated. The amount of PMDA resulted in 1 wt % HiPco in the polyimide. While sonication
was occurring the ODA was weighed out in a 1:1 mole ratio with the PMDA in the HiPco/NMP
solution. The ODA was placed in a 125ml Erlenmeyer flask with a magnetic stirrer. Th e
PMDA/HiPco/NMP solution was added into the 125 ml flask with the ODA, and mixed
overnight to produce the poly(amic acid). Imidization was done the following day by adding in
pyridine and acetic anhydride, each at 2:1 mole ratio with the ODA or PMDA. Products were
obtained after 24 hrs as gel materials, from which solvents and unreacted material were
repeatedly extracted using 190-proof ethanol. Solids were air-dried at room temperature for at
least 24 hrs and then vacuum-dried at ~110oC overnight.
3. RESULTS AND DISCUSSION
Three types of dried bulk products were obtained: (a) HiPco-PI with 0.5 wt % HiPco; (b) HiPco-
PI with 1 wt % HiPco; and, (c) Pure PI. The HiPco-PI products are black porous materials
Vol.8, No.1 Nanoporous Structure and Enhanc ed Thermal Stability 17
(Figure 1) and the pure PI product is a brown dense solid (Figure 2). Both HiPco-PI products
have open-cell structures with about 20-50 nm uniform cells (Figure 3).
Figure 1. Picture of porous solid materials from PI/HiPco that were obtained from a
stoichiometric reaction of 3,3’-ODA and PMDA.
Figure 2. Picture of pure PI solid in film form, which indicates a dense structure.
18 G.T. Caneba Vol.8, No.1
Figure 3. SEM of open cell structure of bulk PI product with 0.5 wt % HiPco, showing 20-50
nm cells. The PI was obtained from a stoichiometric reaction of 3,3’-ODA and PMDA.
Densities seem to be quite low, and these solids exhibited microcracked domains of 100-500
microns (Figures 4 and 5).
Figure 4. SEM of fractured surfaces of bulk PI with 0.5 wt % HiPco, showing 100-200 micron
microcracked domains. The PI was obtained from a stoichiometric reaction of 3,3’-ODA and
PMDA.
Vol.8, No.1 Nanoporous Structure and Enhanc ed Thermal Stability 19
Figure 5. SEM of fractured surfaces of bulk PI with 1 wt % HiPco, showing 100-500 micron
microcracked domains. Also shown on the surfaces is a root st ructure made up of HiPco
bundles. The PI was obtained from a stoichiometric re ac tio n of 3,3’-ODA and PMDA.
This means that 100-500 micron powd ers can easily be obtained from both HiPco-PI nanoporous
materials. This is confirmed after the HiPco-PI porous solid was placed in m-Cresol, which
resulted in the breaking up of the solid into 100-500 micon particles (Figure 6). For the 0.5 wt %
HiPco material, the HiPco bundles are not visible even under 50,000X magnification on an SEM.
For the 1 wt % HiPco sample, root structures are evident (Figure 5), especi ally at a magnification
of 2,000X (Figure 7).
Figure 6. PI/HiPco 100-500 micron particles after adding m-cresol into porous solid in Figure 1
and drying the resulting suspension.
20 G.T. Caneba Vol.8, No.1
Figure 7. SEM of a fractured surface of bulk PI with 1 wt % HiPco, showing a root structure
made up of HiPco bundles at a magnification of 1000X. The PI was obtained from a
stoichiometric reaction of 3,3’-ODA and PMDA.
Protruding structure s from these root patterns are “thorns“ that are about 600 nm long and 100
nm wide (Figure 8); their surfaces seem to indicate aligned nanotube bundles (Figure 9).
Figure 8. SEM of a fractured surface of bulk PI with 1 wt % HiPco, showing “thorns”
emanating from the root structure in Figures 4 and 5. The PI was obtained from a stoichiometric
reaction of 3,3’-ODA and PMDA.
Vol.8, No.1 Nanoporous Structure and Enhanc ed Thermal Stability 21
Figure 9. SEM of a “thorn” from a fractured surface of bulk PI with 1 wt % HiPco. The PI was
obtained from a stoichiometric reaction of 3,3’-ODA and PMDA.
The base of a protrusion also shows nanotube bundles that serve as nanoanchors. Why are the
HiPco/PI materials nanoporous while the pure PI is a dens e material? It is believed that the
HiPco structure served as a rheology modifier for the reactive system, as also evidenced by
observation of incr e as ed v is co si t ies. The gel of the PI with small molecule components could
not turn into dense solids while drying because of the presence of the nanotube bundles and
fibers.
Thermal properties of the nanoporous HiPco-PI composites are quite interesting. TGA analysis
showed the HiPco burn-off temperature of about 500oC in a ir (Figures 10 and 11). PI burn-off
occurred at 500oC and at 615oC and 800oC for 0.5 and 1 wt % HiPco, respectively. In nitrogen
gas, complete gasification of PI was not observed up to 1000oC (Figures 12 and 13). About 50%
of material still remained. Also, in air, the composite with 1 wt % HiPco is more thermally
stable than the one with only 0.5 wt % HiPco. DSC analysis of the 1 wt % HiPco composite
indicates no distinct second-order transition until HiPco burn-off. This means that the polyimide
is unprocessable and cannot be remelted. It should be noted that typical onset of decomposition
temperatures for polyimides based on polypyromellitides are obtained based on th e diamines
used.1 If the diamine is based on a hexamethylene group, the o nset of decompositi o n occurs at
285°C in air and 370°C in Nitrogen gas. For a diamine based on a 1,3-benzene group, the onset
of decomposition occurs at 318°C in air and 530°C in Nitrogen gas. Nanofoams made from
polyimides based on an oxydianiline-pyromellitic dianhydride through grafting with labile
polypropylene glycol oligomers have also been found to decompose at around 600°C.2 These
data indicate that our polyimide-carbon nanotube composites seem to exhibit self-stabilizing
thermal behavior, because as shown in Figures 12 and 13, decomposition does not occur in
Nitrogen even above 1,000°C.
22 G.T. Caneba Vol.8, No.1
It is evident that a number of unusual structural and thermal behavior has been observed during
the formation of PI/HiPco composites from the in situ CNT procedure used in this work. More
experimental efforts will be needed in order to explain these results.
Figure 10. DSC-TGA results of 0.5 wt % HiPco in PI in air.
Figure 11. DSC-TGA results of 1 wt % HiPco in PI in air.
Vol.8, No.1 Nanoporous Structure and Enhanc ed Thermal Stability 23
Figure 12. DSC-TGA comparison results of 0.5 wt % HiPco in PI in air and under nitrogen gas
atmosphere.
Figure 13. DSC-TGA comparison results of 1 wt % HiPco in PI in air and under nitrogen gas
atmosphere.
24 G.T. Caneba Vol.8, No.1
4. CONCLUSION
The physical predispersion of HiPco type single-walled carbon nanotubes in precursor fluids of
polyimide formation not only allowed the dispersion of the CNT into the polymer, but also
resulted in nanoporous composite structure and thermally more stable materials.
Acknowledgements
The author acknowledges the Carbon Nanotubes Group of the Structural Engineering Division of
NASA-Johnson Space Center for various contributions to this research effort, including the
supply of HiPco, SEM work, and some financial ass i s t an c e . Ac k n o w le d g e ment is also provided
to Dr. Patricia Heiden of the Department of Chemistry for pointers on polyimide synthesis.
Finally, the author ackno wledges the Michigan Technological University Center for
Environmentally Benign Functional Materials (CEBFM) for pro viding the setting for manuscript
preparation and continuing efforts from this work.
References
1. Verbicky, Jr., J.R., “Polyimides”, in: Concise Encyclopedia of Polymer Science and
Engineering, J.I. Kroschwitz (Ed.), John Wiley and Sons, New York, 1990, pp.826-828.
2. Han, S.H., Do, J.-S., Abdul Kader, M., Lee, J.-H., Lee, M.-H., and Nah, C., “Preparation
and Characterization of a Polyimide Nanofoam through grafting of labile Poly(propylene
glycol) Oligomers”, Polymers for Advanced Technologies, 15(7), 370-376 (2004).