Journal of Porous Materials 11: 173–181, 2004

c 2004 Kluwer Academic Publishers. Manufactured in The Netherlands.

A New Process to Make a Porous PTFE Structure from Aqueous PTFE

Dispersion with the Help of Hydrogel

TAKASHI KUROSE
Department of Polymer Science and Engineering, Yamagata University,
Yonezawa 992-8510, Japan

TATSUHIRO TAKAHASHI∗
Venture Business Laboratory, Yamagata University, Yonezawa 992-8510, Japan
tatsuhiro@ckpss.yz.yamagata-u.ac.jp

KIYOHITO KOYAMA
Department of Polymer Science and Engineering, Yamagata University,
Yonezawa 992-8510, Japan

Received October 22, 2003; Revised March 10, 2004

Abstract. A novel method to make a porous material having relatively large cell diameter (200–300 µm), which
consisted of mainly poly(tetrafluoroethylene) (PTFE), was developed from aqueous PTFE dispersion by using the
characteristics of hydrogel with the addition of carbon nanofiber (CNF). The porous material was produced as
follows: firstly, an aqueous agar gel containing PTFE and CNF was prepared; secondly, the gel was freeze-dried;
thirdly, the dried gel was heat-treated at 400◦ C where the agar was almost decomposed and PTFE became molten.
The porous material showed electric conductivity (about 50
), high porosity (about 96 vol%), and relatively
uniform cell structures without shrinkage during freeze drying and heat treatment. While the method without CNF
resulted in large shrinkage during heat treatment, meaning that CNF prevented the shrinkage. It was explained by the
idea that the existence of rigid CNF, which was dispersed in the cell wall, prohibited the shrinkage of PTFE during
heat treatment. It was unexpectedly found by SEM analysis that the porous materials had another macro-porous
structure inside the cell wall, suggesting that the developed materials had a double porous structure.

Keywords: porous materials, poly(tetrafluoroethylene), agar, carbon nanofiber, freeze drying

1. Introduction
Several reports have been published about the structural
It has been known that a porous structure is formed from observations and the properties of mechanical strength
hydrogel such as agar, alginate, gelatine, and poly(vinyl of the dried gels for agar [1–4], alginate [2, 5], gelatine
alcohol) gel by freeze drying. Due to biodegradable [6], and poly(vinyl alcohol) [7, 8].
feature and high conformity to a living body, porous Poly(tetrafluoroethylene) (PTFE) is known to have
materials made of these hydrogels have been widely unique chemical, thermal resistance, and surface prop-
adopted in food, pharmaceutical, and medical fields as a erties [9, 10]. Porous materials consisted of PTFE will
carrier of ingredients in foods or an absorbent of liquid. have the properties that are light weight, heat and chem-
ical resistance, which are useful as a shock absorber, a
∗ To whom correspondence should be addressed. filter, and a holder of parts used in severe conditions and
174 Kurose, Takahashi and Koyama

environment. It’s difficult to make a porous structure agar was a few wt% and decomposed products during
by a typical polymer melt foaming process with the heat treatment under air were only water and carbon
help of blowing agents, because the intractability of dioxide, which were harmless to environment and
common PTFE originates in the ultrahigh molecu- human body. In this paper, we discussed the properties
lar weight (estimated to be >107 g/mol) and associ- of the porous materials containing PTFE and CNF
ated extraordinary high melt viscosity (>1011 Pa · s) in terms of process, internal structures, porosity, and
[9]. electric property.
To overcome this difficulty, porous materials
of PTFE, i.e., expanded poly(tetrafluoroethylene)
(ePTFE), have been prepared from the PTFE powder 2. Experimental
with a lubricant by extrusion, rolling, and stretching


R 2.1. Materials
(produced as GORE-TEX (pores size: 0.2–5 µm))
[9]. Many macro-pores (0.2–5 µm in diameter) were
formed through stretching the PTFE sheets. However, The commercial food-grade agar (KANTEN


R
the shape of ePTFE has been limited to the shape of KUKKU , Ina Food Industry Co., Ltd.) was used
membrane. Another process to produce porous mate- to make gel. Poly(tetrafluoroethylene) (PTFE) was
rials of PTFE is the method that PTFE particles were used in the form of stabilized dispersions in water. The
welded together under pressure, which had difficulty dispersion contained 60wt% polymer (MDF PTFE
to make high porosity structure. 30-J, Dupont-mitsui fluorochemical Co., Ltd.) (Aver-
Nanocarbon materials such as fullerens, carbon nan- aged particle size; 0.2–0.25 µm). Carbon nanofiber
otubes, and carbon nanofibers, which have high elec- (CNF) (VGCF
R
, Vapor Grown Carbon Fiber, Showa
tric conductivity, heat conductivity, and stiffness, have Denko K.K.) was used (Averaged diameter; 150 nm
been receiving considerable attentions [11–13]. Re- from a catalog, averaged length; 3.9 µm, measured
search and development using these nanocarbon ma- by our method [13]). A surfactant (Sodium Dodecyl-
terials have been concentrated on various shapes of benzenesulfonate, Tokyo Kasei Kogyo Co., Ltd.) was
products compounded with polymer. However, it has used to disperse CNF into the purified water that were
been difficult to make polymeric porous materials con- de-ionized and filtered out by a membrane.
taining high concentration of CNF by a conventional
polymer foaming process with using blowing agents, 2.2. Method of Sample Preparation
because the addition of small amount of CNF caused
extremely high viscosity [14]. The aim of using CNF Our proposed method to make porous materials has 3
in this study was to reduce the volume shrinkage of a steps as follows: (1) preparations of wet gel containing
porous PTFE structure during heat treatment, in addi- PTFE, additionally with CNF; (2) drying the wet gel;
tion to functionalize porous materials with high stiff- (3) heat treatment of the dried gel. Figure 1 shows the
ness and electric conductivity. The electrical conductiv- flow chart of the detailed sample preparation.
ity of CNF inhibits the electron charge of PTFE porous
materials, which is necessary considering the appli- 2.2.1. Preparation of the Wet Gel. A wet gel was
cations of IC foam and the PTFE shock absorbers for prepared from mixing two kinds of solutions. The first
electric parts. The stiffness of CNF enables the strength solution consisted of agar and purified water, which
of PTFE shock absorbers to control. was heated around 95–100◦ C for 5 minutes to dissolve
The features of the new process and the produced agar into purified water. The second solution consisted
porous materials in this study were as follows: (a) the of purified water and PTFE dispersion or CNF, addi-
porous material had high porosity (96 vol%) and large tionally with a surfactant, which were mixed by a stir-
pore size (100–300 µm); (b) many kind of materials, rer. After keeping the temperature for both solutions
such as, synthetic polymer, metal, and inorganic com- between 50 and 60◦ C, two solutions were mixed to-
pound, were available for the process to make porous gether before gelation occurred. The reason, why two
materials; (c) since hydorogel transforms to hydrocol- solutions were prepared separately, was to avoid gen-
loid reversibly, many kind of polymer processing such eration of the bubbles, which was caused by boiling
as extrusion and injection molding were available for of water during mixing. The mixture was pored into a
making various shapes; (d) the amount of addition for plastic mold (10 × 10 × 1 cm (height)), and the gelation
 A New Process to Make a Porous PTFE Structure 175

Table 1. Sample compositions used in this work. steel net and heated under air at 400◦ C for 60 minutes.
Sample composition (g) The real temperature close to samples was detected by
inserting a thermocouple in the furnace. Weight and
Purified water Agar CNF PTFE Surfactant size of samples before and after heat treatment were
A 100 1.5 – – 0.45
measured. An apparent density and a porosity of the
porous materials were also estimated.
B 95 1.5 – 5 –
C 125 1.5 4.2 – 0.45
D 95 1.5 3 5 – 2.3. Characterization

TGA (thermogravimetric analysis) measurements of

agar, PTFE, and CNF were conducted from room tem-
perature up to 500◦ C at the heating rate of 10◦ C/min
under air atmosphere (TGA 2950 thermogravimetric
analyzer, TA instruments Co., Ltd.). PTFE and agar
were dried in the vacuum oven before the measure-
ment. PTFE was dried from aqueous PTFE dispersion.
DSC (differential scanning calorimeter) measure-
ments of PTFE were performed from room temper-
ature up to 380◦ C at the heating rate of 10◦ C/min and
kept at 380◦ C for 5 min, cooled to 150◦ C at the cooling
rate of 5◦ C/min under nitrogen atmosphere (DSC 2920,
TA instruments Co., Ltd.). Since we feared that a lit-
tle bit of decomposed gas during measurement at high
temperature might damage the DSC apparatus, DSC
measurements were performed under nitrogen atmo-
sphere. Aqueous PTFE dispersion was dried in the vac-
uum oven to eliminate water completely before DSC
measurement.
Figure 1. Flow chart of sample preparation to make the wet gel. A cross-section of dried and heat-treated gel was ob-
served by using SEM (scanning electron microscopy)
took place by cooling at room temperature. Table 1 (JSM-5310, JEOL Co., Ltd.). Samples, either dried or
lists the sample compositions to prepare the wet gel. heat-treated, were cut with a cutter for SEM analysis
after immersed for 5 minutes in liquid nitrogen to elim-
2.2.2. Drying of the Wet Gel. The wet gel was cut into inated ductile fracture. A membrane of gold was coated
several pieces (2 × 8 × 1 cm (height)) and these pieces with about 200 Å thickness using an ion spattering ap-
were dried in a vacuum drying oven at room temper- paratus to eliminate charging in electron beam. SEM
ature. Samples were set onto a plastic net instead of a photographs were taken with a Polaroid camera.
metal plate in the vacuum oven to dry homogeneously. Electric property was evaluated by measuring the
Simultaneously two samples were dried at maximum electric resistance with a digital multimeter (CDM-
in each drying process to keep vacuum pressure sub- 27D, Custom Co., Ltd.). The size of prod to touch the
stantially low. The vacuum treatment was conducted sample was 2 mm in diameter and 20 mm in length.
by using an oil rotary vacuum pump whose pumping An interval between prods was 10 mm during mea-
speed was 120 L/min. The temperatures at two points, suring the electric resistance. The electric resistance
inner oven and gel surface, were monitored by a ther- measurement was conducted for both sides at several
mocouple. The pressure was also measured by a Pirani points. Since there were not large differences of the
vacuum gauge. electric resistance among all positions, the values were
averaged in this study.
2.2.3. Heat Treatment of the Dried Gel. The dried gel Specific surface area of porous materials was mea-
was heated in the Muffle furnace (FO310, Yamato Sci- sured with specific surface area analyzer (NOVA-1200,
entific Co., Ltd.). Samples were set onto the stainless Yuasa Ionics Co., Ltd.). Sample was crushed into
176 Kurose, Takahashi and Koyama

1–2 mm and filled into glass cell. Specific surface area PTFE was observed at 348◦ C. The endothermic heat of
of porous materials was measured using the single point PTFE was 62 J/g. The crystallinity of PTFE was 67%,
BET method. based on the enthalpy at melting transition of PTFE
was 93 J/g. A crystallization temperature was observed
around 318◦ C. Considering the heat treatment temper-
3. Results and Discussion ature of samples at 400◦ C, CNF was not changed from
TGA measurement during the process, but agar lost
3.1. Thermal Properties of Materials about 30% of the weight. While, PTFE became molten
by the heat treatment, since melting point of PTFE was
Figure 2 shows results of TGA for agar, CNF, and around 330–350◦ C from DSC measurements.
PTFE. A remarkable thermal decomposition was ob-
served for agar around 260◦ C, and nearly 70 wt% was
lost up to 400◦ C. On the other hand, CNF gave no 3.2. Drying Behavior of the Wet Gel
weight loss up to 500◦ C. In the case of PTFE, the first
weight loss occurred around 200◦ C, which was caused Figure 4 shows real temperature profiles at two posi-
by vaporization of surfactant, since PTFE dispersion tions, inner drying oven and gel surface, as a function of
contained several wt% of the surfactant [9]. A trace time after starting vacuum drying. The former hardly
amount of weight loss was observed around 500◦ C changed during drying. On the other hand, the latter
by decomposition of PTFE. Figure 3 shows melt- abruptly decreased blow 0 ◦ C with a little bit instabil-
ing curves and crystallization behavior correspond- ity at the initial 4 minutes after the vacuum treatment
ing to PTFE. The melting peak of the first run for started, and the temperature increased gradually. The
temperature at the gel surface reached inner oven tem-
perature after 800 minutes, and the both kept constant
at room temperature. Taking these results into consider-
ation, gel surface became frozen blow 0◦ C during the
initial 400 minutes in Fig. 4 even though inner oven
temperature was 25◦ C. This abrupt temperature de-
crease of the gel surface was caused by the evaporation
of water, which absorbed the heat from the gel surface.
The pressure in the drying oven was always 2–4 Torr
during vacuum drying. Since the ice sublimate below
the pressure of 4 Torr at 25◦ C, the wet gel was dried in
the state of freeze drying. It is importantly noted that
all the samples hardly showed volume contraction in
Figure 2. TGA curve of a agar, CNF, and PTFE under air at the this drying process.
heating rate of 10◦ C/min.

Figure 3. DSC melting curve and crystallization behavior of PTFE Figure 4. Temperature behavior as a function of time for inner
under heating up rate of 10◦ C/min and cooling rate of 5◦ C/min. drying oven and gel surface during vacuum drying.
 A New Process to Make a Porous PTFE Structure 177

Figure 5. SEM photograph of the Sample A after drying treatment.

Figure 6. SEM photograph of the Sample B after drying treatment.

3.3. SEM Observations of the Dried Samples mained [1, 7]. The porous structure observed in Fig. 5
would be formed by the same mechanism.
Figure 5 shows the SEM photographs at the cross sec- Figure 6 provides the SEM photographs at the cross
tion of the dried gel for Sample A (see Table 1). The section of the dried gel for Sample B (see Table 1). The
sizes of cells were fairly uniform having about 200– cell structure of Sample B was similar to that of Sample
300 µm in diameter. The porosity of the sample was A in spite of including PTFE particles. PTFE particles
98 vol% based on the calculation from density. Similar size (about 0.2 µm) was substantially smaller than the
structures have been reported in several papers about thickness of cell walls, resulting in that PTFE particles
hydrogel after freeze drying [1, 3, 6]. The formation were included within the cell wall without affecting the
of porous structure from the wet gel was explained by cell structure.
the fact that a dissolved polymer chain was segregated Figure 7 gives the SEM photographs at the cross
from grown ice crystals during freeze drying process, section of the dried gel for Sample C (see Table 1).
resulting in the formation of the network structures for The thickness of the cell wall was clearly thicker,
polymer. After ice was sublimated by the process of compared with those of Sample A (Fig. 5) and B
freeze drying, the porous structure of polymer was re- (Fig. 6). This was caused by the fact that the cell wall of
178 Kurose, Takahashi and Koyama

Figure 7. SEM photograph of the Sample C after drying treatment.

Figure 8. SEM photograph of the Sample D after drying treatment.

Sample C included CNF. It should be noted that cell Figure 8 shows the SEM photographs at the cross
structure became open-cell structure clearly by the ad- section of the dried gel for Sample D (see Table 1). The
dition of CNF. To get deeper insight of the different dried gel of Sample D had disordered cell structures,
cell structures of Sample C, how CNF was dispersed which consisted of agar, CNF, and PTFE. Though the
in the aqueous PTFE dispersion was examined through cell structure of Samples A and B were fairly uniform,
optical micrograph observation (400 magnitude). CNF Samples C and D including CNF showed disordered
existed partially as a dispersed fiber but mostly as an cell structures. Since CNF in the water tended to cohere
aggregated particle structure (diameter; 0.5–10 µm) in spite of an addition of surfactant, the wet gel contain-
even after substantial stirring. CNF had the aggregated ing CNF had aggregate structures of CNF in the wet gel
particle structure as received [13]. Therefore, CNF observed form optical micrograph. Therefore, the ag-
was not completely dispersed as an individual fiber. gregate structures of CNF let the formation of network
Figure 7 (right) detected a particle-like form aggregated structure heterogeneity during the freeze drying,
particles on the cell wall. These aggregated structure resulting in the disordered cell structure of Samples C
induced thick wall cells. and D.
 A New Process to Make a Porous PTFE Structure 179

3.4. SEM Observations of the Heat Treatment the weight of the initial material before thermal treat-
Samples ment) of Samples B and D were 0.71 and 0.81,
respectively.
A heat treatment was conducted under air atmosphere Figure 9 gives the SEM photographs of the cross
at 400◦ C for 60 minutes where PTFE particles be- section for the heat-treated Sample B (see Table 1).
came molten due to higher temperature than the melt- The cell wall of Sample B became obviously thicker
ing point of PTFE (Fig. 3), and agar was almost de- and the cell size was reduced extremely during heat
composed from the result of TGA (Fig. 2). After the treatment. However, the cell structure seemed to be
heat treatment of Samples A and C, both Samples kept. These SEM photographs allowed us to say
were so fragile as not to be handled properly. While that PTFE particles in the cell wall were welded and
Samples B and D, had enough elasticity to be han- agglomerated in order to decrease the surface area of
dled. Though Sample B contracted extremely by heat- PTFE particle during heat treatment, which resulted
treatment (shrink: 92 vol%), Sample D hardly showed in large volume shrinkage.
the volume change (shrink: 2 vol%). The total yield Figure 10 provides the SEM photographs of
of products (the weight of final product divided by the cross section of the heat-treated Sample D

Figure 9. SEM photograph of the Sample B after heat treatment.

Figure 10. SEM photograph of the Sample D after heat treatment.

180 Kurose, Takahashi and Koyama

Table 2. Sample properties of porous materials produced in this study.

Porosity Bulk densty Electric resistivity Shrink after Specific surface

Sample Compositions Treatment (vol%) (g/cm3 ) (
) treatment (vol%) area (m2 /g)

A Agar Dried 98 0.02 >106 0 –

Heated Deconposed (no shape)
B Agar/PTFE Dried 97 0.05 >106 0 –
Heated 73 0.57 >106 92 1
C Agar/CNF Dried 94 0.06 130 0 –
Heated Broken (too weak to be handled)
D Agar/CNF/PTFE Dried 96 0.08 600 0 –
Heated 96 0.07 50 2 7

(see Table 1). The thickness of the cell wall and the lists the sample properties. All the samples showed high
cell sizes of Sample D seemed to be similar to that of porosity, even sample B (73 vol%) showing the large
Sample D before heat treatment (Fig. 8). Many par- shrinkage. Samples C and D containing CNF had elec-
ticles on the cell surface appeared by heat treatment. tric conductively. Since Sample D hardly showed vol-
The particles on the cell surface would be agglomer- ume shrinkage and had CNF on the cell surface, specific
ated particles of CNF, which did not get loose during a surface area of Sample D was 7 m2 /g, which was larger
mixing process of the solutions. than 1 m2 /g for sample B after heat treatment.
Figures 11(a) and (b) show the SEM photographs of The compositional difference between Fig. 11(a)
a cell surface structure for Sample D before (Fig. 11(a)) and (b) was the agar which existed in Fig. 11(a) but
and after (Fig. 11(b)) heat treatment, respectively. It not in Fig. 11(b) from TGA. However, the macro-
was found that many pores of around 1 µm in diameter pores in the cell wall were generated in a large volume
were observed in Fig. 11(b), which was not observed quantity (Fig. 11(b)), which could not be explained
in Fig. 11(a). It was quite unexpected that the macro- only by the decomposition and the disappearance of
porous structure on the cell surface was observed, agar. The careful observation revealed that CNFs were
which seemed to be formed by connected CNF with the partially covered and bridged by PTFE. The observa-
help of molten PTFE. The macro-pores in cell wall and tion led to the interpretation that the heat treatment
CNF network had an interpenetrating structure. Table 2 made PTFE molten and bridged among CNFs beside

Figure 11. SEM photograph of the cell surface for the Sample D after heat treatment.
 A New Process to Make a Porous PTFE Structure 181

the disappearance of agar. This idea was supported by References

the result that Sample C (No PTFE) became so fragile
after heat treatment. The control of shrinkage for the 1. H.M. Tong, I. Noda, and C.C. Gryte, Colloid and Polymer Sci-
Sample D during heat treatment seemed to be origi- ence 262, 589 (1984).
2. A. Nussinovitch, R.V. Silvestre, and M. Peleg, Biotechnology
nated by the dispersion of rigid CNF in the cell wall. Progress 9, 101 (1993).
3. A. Nussinovitch, Z. Gershon, and L. Peleg, Food Hydrocolloids
4. Conclusions 12, 105 (1998).
4. A.Z. Nussinovitch, M.G. Corradini, M.D. Normand, and M.
Peleg, Journal of Texture Studies 31, 205 (2000).
The porous materials consisted of PTFE and CNF, were 5. R. Reifen, M. Edris, and A. Nussinovitch, Food Hydrocolloids
successfully produced by utilizing the characteristics of 12, 111 (1998).
agar gel. The porous materials in our developed method 6. H.W. Kang, Y. Tabata, and Y. Ikada, Biomaterials 20, 1339
had high porosity (about 96 vol%) and relatively uni- (1999).
7. F. Yokoyama, I. Masada, K. Shimamura, T. Ikawa, and K.
form cell structures (200–300 µm). The porous ma- Monobe, Colloid and Polymer Science 264, 595 (1986).
terial made of only PTFE resulted in an extreme large 8. P.J. Willcox, J. Douglas, W. Howie, K.S. Rohr, D.A. Hoagland,
shrinkage during heat treatment (shrink: 92 vol%). The S.P. Gido, S. Pudjianto, L.W. Kleiner, and S. Venkatraman, Jour-
sample containing both PTFE and CNF showed almost nal of Polymer Science: Part B: Polymer Physics 37, 3438
no shrinkage (shrink: 2 vol%). The reduction of the (1999).
9. T. Satokawa, Fluoro Resin Handbook (Nikkan Kogyo Shinbun,
volume shrinkage was explained by the rigid CNF dis- Tokyo, 1990), p. 2, 34, 125, 126.
persed in the cell wall. The sample containing CNF 10. J.G. Drobny, Technology of Fluoropolymers (CRC press LLC,
showed electric conductivity (about 50
). It was un- Boca Raton, 2000), p. 1.
expectedly found that another macro-porous structure 11. J. Liu, A.G. Rinzler, H. Dai, J.H. Hafner, R.K. Bradley, P.J. Boul,
was formed in the cell wall. A. Lu, T. Iverson, K. Shelimov, C.B. Hussman, F.R. Macias, Y.S.
Shon, T.R. Lee, D.T. Colbert, and R.E. Smalley, Science 280,
1253 (1998).
Acknowledgments 12. K. Lozano, J.B. Rios, and E.V. Barrera, Journal of Applied Poly-
mer Science 80, 1162 (2001).
13. E. Satou, T. Takahashi, T. Natsume, and K. Koyama, TANSO
This research was supported by the CLUSTER 209, 159 (2003).
(Nagano Ueda area) of the Ministry of Education, Cul- 14. P. Poetschke, T.D. Fornes, and D.R. Paul, Polymer 43, 3247
ture, Sports, Science and Technology, Japan. (2002).