AbstractA porous expanded polytetrafluoroethylene (ePTFE) mem-

brane was prepared from emulsion polymerized PTFE finepowders by a series of mechanical operations, which includ-ed extrusion, rolling, stretching and heat setting. Very smallholes in the PTFE sheet were observed by SEM analyses afterextrusion and rolling, and which were elongated andenlarged by longitudinal stretching. Fibrils between the slitsare observed by SEM analyses. The second stretching opera-tion, transverse stretching, provided a lattice-like porousstructure. After heat setting, an island-like structure wasformed, which is composed of billions of tiny inter-connectedcontinuous fibrils and nodes. The porous structure was stud-ied through SEM and the COULTER POROMETER tester.Results show that the mean pore size and porosity increasewith an increase in longitudinal stretching ratio, transversestretching ratio, and heat setting temperature. The mean poresize decreases and the porosity increases with an increase intransverse stretching rate. Remarkably, increasing transversestretching rate increases porosity while the mean pore sizedecreases slightly, resulting in a membrane with a more uni-form cell and denser cell structure. DSC, WAXD analysis and

mechanical testing show that mechanical processing and heatsetting decrease both the crystallinity and crystallite size.

KeywordsExpanded PTFE membrane, biaxial stretching, porous

structure, crystalline structure

1. IntroductionMembrane technology has been significantly developed in

the last three to four decades and has been extensivelyapplied in human life, industrial fields and scientific research.Expanded polytetrafluorethylene (ePTFE) membranes can beproduced by a series of mechanical operations: extrusion,rolling and stretching [1,2,3]. The biaxially stretched ePTFEmembrane was produced by the first stretching (longitudinalstretching) operation parallel to the rolling direction and sec-ond stretching (transverse stretching) perpendicular to thefirst stretching.

It is well known that PTFE has a high melting point, ischemically inert and strongly hydrophobic. The ePTFE mem-brane has high strength and a smooth surface and each squarecentimeter contains billions of continuous superfine fibrilsinterconnected with each other. These properties make ePTFEmembranes ideally suited for a variety of applications andindustrial processes, especially for harsh chemical environ-ments and high temperatures. Furthermore, ePTFE has lowchemical extractability and has excellent biological compati-bility. In China, during May and June, 2003, this ePTFE lami-nated with fabric was worn by the medical staff in severalhospitals without another single infection of medical person-nel with the Severe Acute Respiratory Syndrome (SARS)virus[4,5].

Studies on Porous andMorphological Structures of Expanded PTFE Membrane Through Biaxial Stretching TechniqueBy Xinmin Haoa,b, Jianchunn Zhangb, Yuhai Guoc,,Huapeng Zhangc

ORIGINAL PAPER/PEER-REVIEWED

31 INJ Summer 2005

a. Donghua University, Shanghai, 200051, People’s Republicof China)b. The Quartermaster Research Institute of the GeneralLogistic Department of CPLA, Beijing, 100088, People’sRepublic of China)c. Shanghai New & Special Textile Material Research Center,Shanghai, 200082, People’s Republic of China)

Researchers have developed biaxial stretching technologyto produce expanded PTFE membranes [6,7,8,9,10,11,12];however, few documents have been reported extensively onthe influence of porous structure formation through biaxialstretching operations and the morphological structurechanges during the processing. In this paper, the influence ofthe biaxial stretching process on the structure and morpholo-gy of expanded ePTFE membrane is investigated.

2 Experimental2.1 Sample Preparation

The starting PTFE powder was produced by emulsionpolymerization (CD123. Asahi Fluoropolymers Co.) with anumber average molecular weight of 5 X 106 was mixed withnaphtha as a lubricant (20%) and then extruded into a rod of13mm in diameter at 180°C by use of an extruder. Thereafter,this rod was rolled between two metal rollers to form a sheetwith a thickness of 120µm. The rolled sheet was longitudinal-ly stretched between two pairs of rollers in an oven at 200°C,while the naphtha was being evaporated to produce the PTFEbase sheet. Then the longitudinally stretched-PTFE base sheetwas stretched along the transverse direction at 140°C andfinally subjected to heat setting at about 280°C for several sec-onds. Heat setting of the stretched membrane is the lastprocess necessary to stabilize the porous structure and ensuregood dimensional and mechanical properties of the finalePTFE membrane. The process machinery is shown in theforegoing paper[4].

2.2 Measurements and CharacterizationObservation of surface morphological structure

The ePTFE membranes were sputtering coated with Au forthe SEM observation using Amary-1845FE.

Analysis of pore diameter and porosityMeasurement of the pore diameter was performed with a

COLULTER POROMETER after the sample ePTFE membranehad been soaked in POROFIL liquid for several minutes.

DSC AnalysisA Perkin-Elmer DSC-7 calorimeter was used to analyze the

thermal behavior of differently processed ePTFE membranes,with the heating rate of 15∞C/min up to 400∞C. About 8mgPTFE membrane sample was used for each DSC measure-ment. The crystallinity of the membrane sample was calculat-ed with equation (1):

Xc = (∆H/∆HO) X 100% (1)Where: ∆H is the melting enthalpy of the samples, and ∆HO

is the melting enthalpy of ePTFE with 100% crystallinity,which is assumed as 69J/g.

WAXD AnalysisThe intensity profiles of the ePTFE membrane samples were

measured with Ni-filtered CuKa irradiation ( angstrom) usinga Rigaku D/Max-B WAXD diffractometer with a tube operat-ing voltage of 40kV. In order to eliminate the effect of the ori-

entation induced by stretching on crystallinity measurement,all samples were cut into powder and then pressed into thesample holder during WAXD testing.

The equatorial WAXD patterns of different ePTFE mem-brane samples were obtained, and the crystallinity and crys-tallite size of different samples were calculated by equation (2)and Scherrer’s equation (3) respectively after peak fitting,where the peak around 2q=18∞ is attributed to the crystallinediffraction of (100) planes and the weak peak around 2q=16∞is attributed to the amorphous contribution.

(2)

Where IC is the area of the resolved crystalline peak, and Ia

is the area of the resolved amorphous peak, K is 0.66.

(3)

Where D(hkl) is the crystallite size normal to the crystallineplane (hkl), here k is assumed to 0.89, λ is the CuKα radiationwavelength of 1.5418 angstrom, β is the height at half width ofthe resolved crystalline peak in radian, θ is the Bragg diffrac-tion angle in degree.

According to Bragg reflection equation (4), is in reverse pro-portion to , which represents the chain packing density of thecrystalline domains and can be calculated from the diffractionpatterns of the membrane samples.

2dsinθ = nλ , (4)

where d is the space between the diffracted crystallineplanes, θ is the Bragg angle, and λ is the X-ray radiation wave-length 1.5418 Å, n is the diffraction order.

Mechanical TestingThe tensile tests of the membrane samples were conducted

using an INSTRON 1122 with a sample size 20mm wide and780mm long and conformed to Chinese standard GB1040-79.The loading rate was 100mm/min, and every sample wasloaded to tensile break.

3. Results and Discussion3.1 Effects of Processing Parameters on the Porous Structures of ePFTE membrane3.1.1 Effect of the longitudinal stretching

The results of the SEM observations after different longitu-dinal stretching are given in Fig. 1 and Fig. 2. The rolling andstretching direction in these two figures are the horizontaldirection.

From Figure 1, it can be seen that some holes on the PTFEbase sheet are formed as a result of extrusion and rolling andthey are elongated by the first longitudinal-stretching step.

Samples in Figure 2 were all processed under the same lon-gitudinal-stretching conditions with 5X, 6X and 8X longitudi-nal-stretching steps, at 4.8m/min stretching rate, and no heat

32 INJ Summer 2005

C = %100Ic

Ic + KIa

D(hkl) =kλ

βcosθ

hao layout 7/5/05 5:55 PM Page 2

setting. Figure 2shows that the effect of ratio of longitudinalstretching on the surface structure of the membrane. Theeffects on porosity and mean pore size are shown in Table 1,showing that both properties increase with increasing longitu-dinal stretching.

3.1.2 Effect of the transverse stretchingThe results of the SEM observations after different trans-

verse stretching degrees are shown in Figure 3. The longitudi-nal stretching ratio was maintained at 2X and the transverse

33 INJ Summer 2005

Figure 1SEM MICROGRAPHS OF THE PTFE ROLLED SHEET AND THE LONGITUDINALLY-STRETCHED

PTFE BASE SHEET

Figure 2EFFECT OF RATIO OF LONGITUDINAL STRETCHING ON THE SURFACE STRUCTURE OF THE

MEMBRANE

(1) PTFE rolled sheet

(1) 5X longitudinal stretching (2) 6X longitudinal stretching

(3) 8X longitudinal stretching

(2) 1X longitudinal-stretching PTFE sheet

hao layout 7/5/05 5:55 PM Page 3

stretching rate was also fixed and the ratio of transversestretching was varied during the sample preparation from2.1X to 8.5X. The transverse stretching rate was fixed at4.8m/min and no heat treatment was employed.

According to the data in Table 2, porosity and pore diameterare increased along with the increase of the ratio of transversewidening. Pore diameter and porosity are controlled in accor-dance with the final application of the product.

3.1.3 Effect of the transverse stretching rateThe effect of the transverse stretching rate on the pore diam-

eter and porosity of the ePTFE membrane is given in Figure 4and Table 3, where the other processing parameters are all thesame except the transverse stretching ratio. None of the mem-branes depicted in Figure 5 were heat-set and they wereprocessed with 2X longitudinal stretching and 4.7X transversestretching.

The data in Table 3 show that the higher the transversestretching rate, the higher the porosity, further, the averagepore diameter declines with the increase of the stretching rate.Literature data reveals that the temperature during transversestretching has little effect on membrane stretchability. This isbecause the activation energy required for the steady growthof fibrils is 11.3 kJ/mol [8,9], which makes it is easy to form

pores in membrane by stretchingmembrane.

3.1.4 Effect of the heat settingtemperatureHeat setting is the last process inthe production of microporousePTFE membrane, which stabi-lizes the structure and enhancesthe mechanical properties of themembrane.

In order to counteract the ther-mal shrinkage and fix themicrostructure of the membrane,the membrane was constrained

during the heat setting. The tem-perature and time of heat set-ting have great influence onthe porous characteristics, thedimensional stability andmechanical properties of thefinal membrane. Figure 5 dis-plays the SEM photographsof the membrane samples andthe effect of heat setting onthe pore characteristics areshown in Table 4. Heat settingtemperatures ranged from245°C to 300°C and all the

34 INJ Summer 2005

Figure 3EFFECT OF THE RATIO OF TRANSVERSE STRETCHING ON THE

STRUCTURE OF THE MEMBRANE

Table 1 EFFECT OF LONGITUDINAL STRETCHING

RATIO ON POROSITYStretching ratio 5 6 8Porosity, % 76.4 87.6 89.1Mean pore diameter, ¶Ãm 0.35 0.47 0.51

Table 2EFFECT OF TRANSVERSE STRETCHINGRATIO ON POROSITY AND THE PORE

DIAMETERStretching ratio 2.1 2.9 4.7 8.5Porosity, % 56.5 58.4 60.4 78.0Mean pore diameter, µm 0.08 0.09 0.12 0.24

Table 3EFFECT OF TRANSVERSE STRETCHING

RATE ON POROSITY AND PORE DIAMETER

Stretching rate, m/min 4.8 6 8Porosity, % 60.4 64.2 70.8Mean pore diameter, µm 0.12 0.11 0.09

1) 2.1X transverse stretching (2) 2.9X transverse stretching

(3) 4.7X transverse stretching (4) 8.5X transverse stretching

hao layout 7/5/05 5:55 PM Page 4

samples were subject to 2X longitudinal stretching, 8X trans-verse stretching, and the transverse stretching rate was set as9m/min.

From Table 4, we can see that the membrane pore diameterincreases with the increase of heat setting temperature. Thehigher of the heat setting temperature, the larger of the node

area, and the space betweenthe nodes also increases. Itmay be that the fibrils con-nected by the nodes willbreak down at higher temper-ature. The pores are formedamong fibrils which intercon-nect the nodes, and the porediameter is dependent on thespace among fibrils.

3.2 Effects of ProcessingParameters on the CrystallineStructure and MechanicalProperty of ePFTE membrane3.2.1 DSC Analysis

The DSC traces and corre-sponding calculated results ofdifferent membrane samplesare given in Figure 6 and Table5.

From Figure 6 and Table 5, itcan be seen that the crys-tallinity and the meltingfusion of the membranedecreases with stretching andheat setting temperature. Thehigher the heat setting tem-perature, the lower theenthalpy and consequentlythe lower the crystallinity.From Figure 6, it can be seenthat the melting of the ePTFEmembrane occurs at a rangeof temperatures, which sug-gests the non-perfectionand/or non-uniformity ofthe crystalline structure ofthe membrane with less per-fect and/or small crystallinedomains melting first, andthe melting range is widenedby stretching and heat set-ting, and after stretching andheat setting the onset andmaximum melting tempera-ture of ePTFE membrane islowered. With 300°C heattreatment, the maximummelting temperature is low-

ered to about 327°C, which is the well documented meltingtemperature of PTFE recrystallized from the melt. The ePTFEmembrane is turned from opaque to transparent when theheat setting temperature is above 300°C, which suggests theeffect of decrease of crystallinity with the increase of heat set-ting temperature.

35 INJ Summer 2005

Figure 4EFFECT OF TRANSVERSE STRETCHING RATE ON THE SURFACE

STRUCTURE OF THE MEMBRANE

(1) 4.8m/min (2) 6m/min

(3) 8m/min

Table 4EFFECT OF HEAT SETTING TEMPERATURE ON POROSITY

Sample No. Heat setting Mean pore Minimum pore Maximum poretemperature diameter, µm diameter, µm diameter, µm

1 245 0.382 0.319 0.4242 280 0.589 0.469 0.6463 300 0.685 0.618 0.835

Table 5CALCULATED RESULTS FROM DSC ANALYSIS

Sample Enthalpy (∆H),J/g Crystallinity XC ,% Tmax,OCSheet 59.636 86.4 342.7Base sheet 57.549 83.4 342.7Non-heat set 54.786 79.4 342.1245° heat set 46.165 66.9 343.3(336.0)280° heat set 32.852 47.6 335(343.7)300° heat set 16.059 23.3 327

hao layout 7/5/05 5:55 PM Page 5

A discernable peak shoulder at lower temperature exists forall samples before heat setting, and it becomes more and morepronounced after heat setting and especially at 300°C heatsetting. Only one lower temperature peak exists with peaktemperature of about 327°C. Several documents have beenpublished dealing with the melting and crystallization behav-

ior of dispersion PTFE poly-mer. Suwa et al [13] attrib-uted the higher temperaturepeak of DSC analysis to thelinear part of the folded rib-bon based on the microstruc-ture model first put forwardby Rahl et al [14] and con-firmed by Chanzy et. al [15]through high-resolution elec-tron microscopy. Further, thelower temperature peak wasattributed to the foldedregion of the folded ribbon.As far as the chain extensionin the crystalline lamellae orribbon is concerned, someauthors suggested the chainsare folded with the chainaxis perpendicular to thelength of the crystallinebands like conventionalpolyethylene after recrystal-lization from the melt, andsome authors like Suwa andRahl at al concluded that

chains are extended with thechain axis along the crys-talline ribbons like polyethyl-ene crystallized under highpressure for virgin dispersion

polymer particles.Based on the above mentioned observation and discus-

sion, it seems that with heat setting above 150°C the virginPTFE polymer becomes more and more chain folded, but thechain extension and folded degree is still open to verificationby other characterization means which is underway in ourfurther study.

3.2.2 WAXD AnalysisThe WAXD profiles and corresponding calculated results

are given in Figure 7 and Table 6.Table 6 shows that the crystallinity and crystallite size of the

membrane is decreased with the stretching and heat settingprocesses, which suggests stretching and heat setting underconstraint leads to the deterioration of crystalline domains ofPTFE membrane, especially the crystalline domains perpen-dicular to the chain axis direction. After longitudinal stretch-ing, the chain packing normal to the chain axis directionbecomes smaller, and with biaxial stretching, the chain pack-ing becomes a little larger, but after heat setting at 245°C and280°C, the chain packing normal to the (100) planes restoresand with 300°C heat setting the chain packing again becomeslarger with recrystallization of small or defective crystallitesunder higher temperature.

36 INJ Summer 2005

Figure 5EFFECT OF HEAT SETTING TEMPERATURE ON THE POROUS STRUC-

TURE OF THE MEMBRANE

Figure 6DSC CURVES OF DIFFERENT PTFE

MEMBRANE SAMPLES

hao layout 7/5/05 5:55 PM Page 6

3.2.3 Mechanical PropertyHeat setting serves as not only a dimensional stabilization

and fixation means, but also a mechanical enhancement. Withheat setting, the tensile breaking extension decreases, the ten-sile modulus increases and the tensile strength also increasesa little, which is illustrated in Figure 8.

4. ConclusionsEmulsion polymerized PTFE particleswith a number average molecular weightof 5X106 were used as the starting mater-ial to make ePTFE membranes with theoperations of extrusion, rolling, longitu-dinal stretching, biaxial stretching andheat setting.

Small holes in the ePTFE sheet areformed after extrusion and rolling, and

are elongated into slits which become larger with increasedlongitudinal stretching. More fibrils are also observed by SEMas the fracture size increases. The second stretching opera-tion, transverse stretching, provides a lattice-like porousstructure. After heat setting, the island-like structure isformed; it is composed of billions of tiny inter-connected con-tinuous fibrils and the nodes.

The mean pore size and porosity increase with an increasein longitudinal stretching ratio, transverse stretching ratio,and setting temperature. The mean pore size decreases, andporosity increases with an increase in transverse stretchingrate. The crystallinity and crystallite size of the membranedecreases with stretching and heat setting, and after heat set-ting, the crystallinity clearly decreases. Also after heat setting,the membrane can no longer be stretched, and the dimension-al stability and mechanical properties are enhanced. Heat set-ting may change the extended chains in the lamellae into fold-ed chains, which will be the subject of future work.

Through controlling the stretching direction, stretchingratio, stretching rate and heat setting temperatures, ePTFEmembranes with different porous characteristics, morphologyand mechanical properties can be engineered for differentproduct end-uses.

Acknowledgements: The financial support of this project from the Beijing

Natural Science Foundation of China with the Program grantnumber of 2052016 is thankfully acknowledged.

We are very grateful to Prof. Larry C. Wadsworth of theMaterials Science & Engineering Department, University ofTennessee for helpful critical suggestions.

Refernces1. S.Oga. Japan Patent 42-13560, 1967.2. R.W.Gore. U.S. Patent 3664906, 1972.3. R.W.Gore. U.S. Patent 3953556, 1976.4. Huang Jizhi, Zhang Jianchun, et al. European Polymer

Journal 2004; 40: 667-671.5. Hao Xinmin, Zhang Jianchun, et al. European Polymer

Journal 2004; 40: 673-678.6. Zhang Jianchun, Huang Jizhi et al. In: Processes tech-

nique of waterproof and permeable textile, China TextilePress; 2003. p.271.

7. Guo Yuhai. A Study on Waterproof & MoisturePermeable Textile Laminated with Microporous ePTFEMembrane. Master Thesis, 1997. p32.

37 INJ Spring 2005

Figure 7WAXD ANALYSIS RESULTS OF DIFFERENT

EPTFE MEMBRANE SAMPLES

Figure 8TENSILE MECHANICAL PROPERTIES OF THE

HEAT SETTED MEMBRANE SAMPLES

Table 6CALCULATED RESULTS OF THE WAXD ANALYSIS

Sample Crystallite size, Å Crystallinity, % 2θ,O d,ÅSheet 113.2 97.2 18.0 4.93Base sheet 93.7 94.2 18.5 4.80Non-heat set 84.9 94.9 17.9 4.96245° heat set 89.3 88.6 18.5 4.80280° heat set 79.9 86.0 18.5 4.80300° heat set 65.3 80.7 18.0 4.93

hao layout 7/5/05 5:55 PM Page 7

8. Taketo Kitamura, Ken-ichi Kurumada, et al. Polym. Eng.Sci. 1999; 39: 2256.

9. Taketo Kitamura, Shohei Okabe, et al. Polym. Eng. Sci.2000; 40: 809.

10 Ken-ichi Kurumada, Taketo Kitamura, et al. J. Mem. Sci.1998; 149: 51-57.

11. Zhang Zhiliang, Chen Shanmei. Chinese Journal ofChemical Physics 2003; 16: 151-155.

12. Chen Shanmei, Li Aoqi. Chinese Membrane Science andTechnology 2003; 23: 19-21.

13. Takeshi Suwa, Tadao Seguchi, et al. J. Polym. Sci., Polym.Phys. Ed. 1975; 13: 2183-2194.

14. F. J. Rhal, M. A. Evanco, et al. J. Polym. Sci. Pt. A-2 1972;10: 1337-1349.

15. H.. D. Chanzy and P. Smith. J. Polym. Sci., Polym. Lett.Ed. 1986; 24: 557-564. — INJ

38 INJ Summer 2005

hao layout 7/5/05 5:55 PM Page 8