Production and Evaluation of Meltblown Poly(phenylene sulfide)
Roy M. Broughton, Jr.1,
Chris Eash2 and Larry C. Wadsworth3
1
2
University of
3
University of
Abstract
Manufacturers are currently supplying poly(phenylene
sulfide), PPS, for use in continuous melt production of
nonwoven fabrics by spunbond and meltblowing (MB) processes. Aside from a preliminary study made by these
authors in a paper presented on MB PPS at the INDA Filtration Conference in
November 2006, little evaluation of the MB PPS products has been reported. A significant thermal shrinkage problem with
MB PPS was noted and procedures were suggested to minimize shrinkage, mainly by
slowing the PPS fiber cooling rate to allow polymer crystallization to occur
during the melt blowing process. Among
the procedures studied, elevating the temperature of the MB room and increasing
the weight of the web were found to be the more effective in minimizing thermal
shrinkage of
Introduction
Poly(phenylene sulfide) fiber has been available for more than 30 years (1, 2). It has been the fiber/fabric of choice for moderately high temperature filtration in chemically corrosive environments (3). It is reported to be nonflammable (OI = 44), stable for use at 200 °C and resistant to organic solvents, acids and bases in the range pH 2 – 12, and oxidizing agents as well as 10% HCl (4).
Nonwovens are the structures of choice for filtration applications because of their improved filtration capacity and their depth filtration capability which retards plugging of the filter. The inclusion of a filtration layer having fine fibers is well understood to produce more effective removal of small particulate and meltblowing (MB) is generally recognized as an effective way of producing fibers finer than those available from typical melt spinning or spunbond (SB) processes. The natural progression of the logic above is the attempt to produce MB PPS nonwovens. At least one supplier suggests and offers PPS for MB applications (5, 6).
Although the polymer manufacturer has
publicized MB and offers a MB grade, little data has been published on the
properties of the MB PPS fabrics. In a
preliminary study, the current authors found that the MB PPS fabric exhibited
thermal shrinkage that would render its use unacceptable in high-temperature
filtration applications (7). This high
degree of thermal shrinkage is typical of fiber melt extrusion where the
polymer has a Tg substantially above room temperature and therefore passes through
its optimum crystallization temperature rapidly during quenching, as reported
by Spruiell and Janke (8). The as-spun
fibers of PPS (like PET) should be expected to cool before crystallization and
exhibit substantial shrinkage as they are subsequently heated to their
crystallization temperature. The
expected results were found for MB PPS by Japanese researchers who suggested a
calendaring step to remedy the problem (9).
The dimensional instability of PPS fiber during heating has also been
recognized in patented polymer additives introduced to reduce the shrinkage
during calendaring (10). This current
work attempts to address the concern about dimensional instability of MB PPS
fabric during heating, by slowing the MB PPS fiber cooling rate through manipulation
of the MB process parameters.
Experimental
PPS polymer (Fortron 0203HS, from Ticona) was dried overnight in a Conair dessicant dryer at 105 °C and a -40 °C dewpoint. The resin was extruded on a 6 inch wide melt-blowing line at the University of Tennessee (TANDEC). The MB die geometry and extrusion profile are given in Table 1. A nosepiece with a 60 degree included angle was utilized with spinneret diameters of 0.018 in. (0.46 mm), an orifice L/D of 10/1 and with 20 holes/in. There were 121 holes over the 6 inch die. The air knife gap was 0.080 in. (2.0 mm) and the die tip setback was 0.080 in. (2.0 mm). Different methods for decreasing MB fiber cooling rates were studied, and fabric properties were evaluated, including filtration efficiency (FE) to 0.1 μm NaCl, DSC and thermal shrinkage at 125 °C for five minutes. FE and ∆P to an aerosol of 0.067 µm NaCl particles at a flow rate of 32 l/min (5.3 cm/s face velocity) were determined with the TSI Model 8130 Filtration Tester. Thermal shrinkage was determined by placing a 4 x 6 in. specimen without restraining the edges onto a metal pan and then putting it into a forced draft oven at 125 °C for 5 minutes. After the specimen was removed from the oven it was measured again in both machine (MD) and cross machine (CD) directions, and percentage shrinkage based on the original dimensions was calculated.
Average fiber diameters of the PPS web samples were determined from scanning electron microscopy. Ten fiber width measurements were made from two web specimens (1 x 2 inch dimensions) taken diagonally from the two edges (2-3 inches from actual edge of web) and from the middle of the web for a total of 20 diameter measurements per each sample. An image analysis program from the Scion Company was used. Fiber diameters were converted from pixels to micrometers using the factor, 3.99/1 pixels/µm. Selected samples were evaluated by Differential Scanning Calorimetry (DSC).
The following process parameters were employed to prevent rapid cooling of the PPS web below the glass transition temperature at around 89°C. The spinneret diameter of 0.018 inches (0.46 mm), compared to the more typical 0.015 in. (0.38 mm) allowed larger diameter strands with a higher volume to surface area ratio in order to delay fiber cooling. The 0.080” (2.0 mm) air gap and setback is believed to have reduced the rate of adiabatic expansion (11) and thereby reduced the air cooling rate of the MB quench air when compared to a more traditional 0.030”(0.76 mm) air gap and setback used for polypropylene MB production.. In order to reduce the fiber cooling rate due to the mixing of the MB process air free jet with cooler ambient air, the ambient air was allowed to heat up to 115°F (46°C) by closing all doors to the MB room and turning off all fresh air ventilation. The temperature of the MB collector belt was monitored and the PPS MB web was allowed to reside on the collector for different lengths of time to determine if more time resulted in less thermal shrinkage by slowing down the time required for the web to cool to the Tg of PPS.
Table 1. MB die geometry, extrusion profile and air temperature parameters
MeltBlown
Line Width |
6 |
Inch |
Hole
Size |
0.018 |
Inch |
Number
of Holes |
121 |
Holes |
Air
Gap |
0.08 |
Inch |
Setback |
0.08 |
Inch |
Air
Temperature |
310 |
°C |
Extruder
Zone 1 Temp |
290 |
°C |
Extruder
Zone 2 Temp |
300 |
°C |
Extruder
Zone 3 Temp |
310 |
°C |
Die
Temp |
310 |
°C |
Polymer
Throughput |
0.44 |
ghm |
Melt
Pressure |
5.5 |
Bar |
Air
Plenum Pressure |
1.03 |
Bar |
As shown in Table 1, the MB die and hot attenuation air temperatures were both set at 310 °C . The polymer throughput remained a constant 0.44 g/hole/min with the resulting melt pressures of 5.5 bar. The process parameters downstream of the MB die are given in Table 2. The MB process air pressure for sample 1 was 1.38 bar (20 psi), but was reduced to 1.03 bar (15 psi) due to fiber breakage with sample 1. The die-to-collector distance was maintained at a constant 280 mm (11 in.) throughout the trial. Two belt speeds were used in order to provide low and high basis weight samples for comparison. The web was removed from the collector belt at two locations corresponding to 1.1 meters and 1.8 meters after deposition to provide a short and long residence time on the collector belt. The ambient MB room temperature was allowed to rise to 43 °C (109 °F) by closing the doors and turning off the ventilation and exhaust fans. With samples 21 and 22, the doors were opened and the ventilation was turned on to allow the MB room to cool to 15 °C (outside temperature was approximately 10 °C).
Results and Discussion
A typical DSC trace
for MB PPS (heated at 20 oC per minute) is shown in Figure 1. This figure provides an adequate explanation of
the shrinkage problem: the MB PPS is not
stabilized by crystallinity, and when crystallization occurs, it is preceded or
accompanied by significant shrinkage. The DSC shows a Tg of about 85 oC,
a crystallization exotherm at about 120o C, and a melting endotherm
at about 280 oC.
Recrystallization on cooling occurred at about 198o C and
there was not further crystallization during a second scan up to the melting
point. No consistent effect was observed when
comparing shrinkage with the specific heat of crystallization or specific heat
of fusion.
Figure 1. DSC trace for Sample 2 - heated and cooled
The data for all samples appear in Tables 2 and 3, and show substantial thermal shrinkage variability over the run. With the MB room at the high ambient temperature of 43 °C (Samples 1-20), thermal shrinkage varied from an acceptable low value of 3 – 5 % to a quite unacceptable high of 30%; whereas, thermal shrinkage of the webs produced at a room temperature of 15 °C (Samples 21 and 22) had thermal shrinkage values ranging from 34 to 45%. By far the most noticeable correlation in these data is the dramatic increase in thermal shrinkage when the ambient room conditions were changed from 43 to 15o C. Before the collection of sample 21 the doors were opened and the temperature dropped to a mild 15 °C (59 °F). Samples 21 and 22 exhibit thermal shrinkage values of over 30% in the MD and CD, with a notable shrinkage of 45.1% resulting in the CD of sample 22. This operation was not repeated with more samples as the room would have required several hours to heat to 43 °C. We can say that ambient air temperature appeared to have the greatest impact of any process variable resistance to thermal shrinkage of the MB PPS web - as shown by samples 21 and 22.
When this dramatic effect is removed from the data and one looks at what other processing variable affect thermal shrinkage, there is some supporting evidence for the idea that thermal history during extrusion is important. There is a trend (with substantial scatter) indicating that increased residence time on the belt accompanies an increased basis weight and a reduced shrinkage (Figures 2 and 3). This is consistent with and reinforces the idea that slowing the rate of cooling can provide increased fabric crystallinity and, therefore, increased thermal dimensional stability. It is not known to what extent these effects would provide a sufficient thermal history to allow a more moderate ambient collection temperature.
Table 2. Process parameters downstream of the MB die
Belt
Speed (m/min) |
Distance
on Belt (m) |
Time
on Belt (s) |
Weight
(g/m2) |
Thickness
(Mils) |
MB
Room Temp (°C) |
|
|
|
|
|
|
|
|
1 |
12.0 |
1.1 |
5.5 |
21.5 |
10.2 |
43 |
2 |
12.0 |
1.1 |
5.5 |
19.1 |
11.4 |
43 |
3 |
6.9 |
1.8 |
15.7 |
47.0 |
10.7 |
43 |
4 |
6.8 |
1.1 |
9.8 |
36.3 |
10.9 |
43 |
5 |
4.7 |
1.8 |
22.9 |
54.9 |
12 |
43 |
6 |
4.7 |
1.1 |
14.0 |
57.0 |
19.8 |
43 |
7 |
13.8 |
1.8 |
7.8 |
20.2 |
8.9 |
43 |
8 |
13.8 |
1.1 |
4.8 |
17.7 |
9.5 |
43 |
9 |
4.6 |
1.8 |
23.3 |
53.6 |
15.2 |
43 |
10 |
4.6 |
1.1 |
14.3 |
55.1 |
11.9 |
43 |
11 |
7.0 |
1.8 |
15.5 |
33.1 |
10.1 |
43 |
12 |
7.0 |
1.1 |
9.4 |
37.1 |
10.1 |
43 |
13 |
14.0 |
1.8 |
7.7 |
16.5 |
9.1 |
43 |
14 |
14.0 |
1.1 |
4.7 |
17.0 |
7.4 |
43 |
15 |
7.0 |
1.8 |
15.3 |
40.8 |
9.6 |
43 |
16 |
7.1 |
1.1 |
9.3 |
33.4 |
8.6 |
43 |
17 |
4.6 |
1.8 |
23.7 |
57.9 |
12.8 |
43 |
18 |
4.5 |
1.1 |
14.5 |
46.3 |
10.4 |
43 |
19 |
13.8 |
1.8 |
7.8 |
21.2 |
10.1 |
43 |
20 |
13.8 |
1.1 |
4.8 |
18.9 |
8.0 |
43 |
21 |
|
1.8 |
7.5 |
16.8 |
9.9 |
15 |
22 |
13.8 |
1.1 |
7.5 |
14.2 |
15 |
15 |
Note: Die-to collector distance was 280 mm (11 in.) for all samples
Collector belt temperature was 85 °C for all samples
Table 3. MB PPS fiber diameter, filtration efficiency to 0.067 μm NaCl at a face velocity of 5.3 cm/s, and thermal shrinkage
Sample No. |
Fiber Diameter (μm) |
FE (%) |
Thermal Shrinkage @ 125o C (%) |
|
|
|
|
MD |
CD |
|
|
|
|
|
1 |
|
|
16.8 |
11.5 |
2 |
2.1 |
16.0 |
8.8 |
4.4 |
3 |
2.4 |
16.8 |
5.3 |
2.7 |
4 |
3.3 |
18.4 |
4.4 |
4.4 |
5 |
|
|
3.5 |
5.3 |
6 |
|
|
5.3 |
2.7 |
7 |
|
|
5.3 |
2.7 |
8 |
|
|
15.9 |
13.3 |
9 |
3.2 |
20.8 |
4.4 |
3.5 |
10 |
2.8 |
17.3 |
4.4 |
3.5 |
11 |
3.7 |
15.3 |
8.8 |
6.2 |
12 |
3.2 |
12.0 |
23 |
5.3 |
13 |
3.5 |
8.2 |
15.9 |
7.1 |
14 |
3.2 |
5.9 |
30.1 |
7.1 |
15 |
4.2 |
10.8 |
11.5 |
6.2 |
16 |
4.4 |
12.7 |
8.0 |
2.7 |
17 |
4.3 |
20.9 |
4.4 |
3.5 |
18 |
3.6 |
24.3 |
4.4 |
2.7 |
19 |
3.8 |
6.9 |
7.1 |
8.8 |
20 |
|
|
8.8 |
9.7 |
21 |
4.3 |
5.7 |
33.6 |
40.7 |
22 |
|
|
39.8 |
45.1 |
Figure 2. Effect on shrinkage of residence time on belt
Figure 3. Effect of
fabric weight on shrinkage
Figure 4 shows the expected result that increasing the fabric weight increases the filtration efficiency and since increased weight accompanies decreased shrinkage, it follows in Figure 5 that there is an inverse relationship between shrinkage and filtration efficiency.
Figure 4. Filtration
efficiency versus weight of MB PPS webs.
Figure 5. Filtration efficiency versus fabric sample shrinkage
Conclusions:
Shrinkage of MB PPS fabrics is substantial and too large for high temperature applications unless precautions are taken to minimize the effect.
Shrinkage seems particularly sensitive to the rate of cooling which can be controlled by maintaining a hot ambient temperature surrounding the take up.
A slower take up (resulting in a heavier fabric as well as
a longer time on the collection belt) also reduces the thermal shrinkage.
These techniques allow the production of a MB PPS with
sufficient dimensional stability to produce useful high temperature filtration
products.
Acknowledgements
The donation of the PPS polymer by Ticona is appreciated. The authors are also grateful to TANDEC for donating the line time on the 6-inch MB line and to Shawn Stahl, Constance Eastman, Munya Mundava, Hasan Kocer, and Fatma Kilinc-Balci for assisting in the trials and for performing the testing.
References
1. Robert Campbell, “P-phenylene sulfide polymers” U. S. Patent 3919177, November 11, 1975.
2. Robert Campbell and Harrold Yelton, “P-phenylene sulfide polymers, U. S. Patent 4024118, May 17, 1977.
3. R. P Donovan, pp
74 in Fabric Filtration for Combustion
Sources, Marcel Dekker,
4. http://www.celanese.com/index/mr_index/mr_news/mr_news_business/mr_news_
business-fullpage.htm?id=21569
5. http://www.celanese.com/index/mr_index/mr_news/mr_news_business/mr_news_
business-fullpage.htm?id=16863
6. http://www.ticona.com/redesign/markets/technical_fibers___fabrics
7. Larry C. Wadsworth, Roy M. Broughton, Jr., Christopher A. Eash and Peter Tsai
“Advanced
High Temperature Melt Blown Polyphenylene Sulfide Filter Media” Proceedings
on CD ROM, Filtration 2006 Conference,
8. J.E. Spruiell,
Chris J. Janke. “A review of the measurement and development of crystallinity
and it's relation to properties in neat poly(phenylene sulfide) and its fiber reinforced
composites”. US Department of Energy,
9. Isoda, Hideo,
Hayashi Klyohide, and Katsumasa Yamamoto (Toyobo Co., Ltd,
10. Auerbach, Andrew; Brueck, Martin, and Srinivasan, Ramesh, “Methods of making spunbonded fabrics from blends of polyarylene sulfide and a crystallinity enhancer”, U. S. Patent Application 20050269011, December 8, 2005
11. Ting Chen, Xinhou Wang, Xiubao Huang. “Modeling
the Air-Jet Flow Field of a Dual
Slot Die in the Melt Blowing Nonwoven Process”. Textile Research Journal 2004; 74 1018.