Is There a Substitute for Polycarbonate Film Capacitors?

For the PDF version of this article, click here.

Many of today's electronic systems rely on wound polycarbonate film capacitors for operation at temperatures exceeding +105°C. First introduced in the 1960s, the majority of “Mil Spec” capacitors are based on this dielectric ^[2]. However, in the summer of 2000, the sole supplier of raw polycarbonate film announced it would stop producing the film by the close of that year. Of course, this caused considerable consternation in many sectors of the electronics industry, among capacitor users and suppliers alike ^[3]. Since polycarbonate would remain available only for a period of time based on existing film stocks and from the last time run by the film manufacturer, a course of action was needed to replace this dielectric.

There are other commercially available plastic dielectric films that can operate at the same or higher temperatures as polycarbonate. Possible candidates include:

Polyethylene napthalate (PEN)
Polyphenylene sulfide (PPS)
Polyimide (PI)
Polytetrafluoroethylene (PTFE)

Table 1, on page 32, compares the characteristics (based on the manufacturer's data) of these candidates compared to polycarbonate (PC). We refer to PTFE and PI by DuPont's trademark names, “Teflon” and “Kapton,” respectively. For all others, we will use the abbreviations, as listed above.

The high dissipation factor of PEN rules it out in many applications. Teflon has a very low dissipation factor, but is limited by a low dielectric constant. It's also extremely expensive, and available films have inconsistent thickness variations, making it frustrating for the capacitor manufacturer to use. The dielectric constant and dissipation factor characteristics of Kapton make it an interesting candidate, but the material is only available in very thick films, making it impractical for wound capacitors where volumetric efficiency is a premium. Thus, the best candidate to replace PC is PPS.

Polyphenylene sulfide isn't a newly discovered film dielectric. You may be familiar with PPS as the Phillips Petroleum Company's high temperature thermoplastic molding material, trademarked as “Ryton.” PPS dielectric film is the result of a joint development effort by Phillips and Toray Industries Inc. of Japan. Data on film capacitors using this dielectric first appeared in 1987 ^[5].

PPS Characteristics

The PPS polymer is crystalline in nature with a remarkably stable structure, high temperature resistance (UL temperature index of +240°C), inherent flame resistance, and excellent resistance to attack by solvents and moisture. The question is how does the replacement of PC with PPS affect capacitor design and performance?

PPS and PC have about the same dielectric constant, so the size of a PPS replacement capacitor is the same as the PC original. One difficulty in generalizing this statement is that the PPS dielectric films are not always available in the same thickness as those of PC. The breakdown strength of PPS is slightly higher than that of PC — 400V per micron thickness for PPS compared to 300V for PC. This is important when considering replacement designs. If the original PC design was based on l0µm thick film, it could be replaced with a 9µm thick PPS film. This actually decreases the overall voltage stress on the dielectric by 3% as well as gaining a modest reduction in the final capacitor's size.

The stability of capacitance over the operating temperature range of a device is always an important performance attribute. Fig. 1, on page 31, compares the capacitance change with temperature for PC and PPS. Using the initial value as the capacitance at room temperature. Although the two dielectrics exhibit a different behavior over the temperature range, they both have an overall change of capacitance of about 2% from low to high temperature.

From -55°C to +125°C, PPS has a superior, overall capacitance stability with temperature — 80 ppm/°C for PPS compared with 150 ppm/°C for PC. From -55°C to +85°C, PPS is virtually flat at 7 ppm/°C.

Typically, established PC capacitor designs are limited to +125°C operation and many PC specifications, from a variety of sources, stipulate that you must derate or reduce the voltage rating of a PC capacitor at this temperature. In some cases, this voltage-derating requirement is as much as 50%! PPS can be full rated, with no voltage derating up to +150°C.

To prove this, we constructed 25 l0µF PPS capacitors rated at 50Vdc using the same film thickness that would be used if the design were PC based. We life-tested these parts for 2000 hr at +150°C with no voltage derating. The applied test voltage was 70Vdc, or 140% of the rated voltage. Table 2 summarizes the test results.

To evaluate the shelf life of capacitors manufactured with PPS a series of date codes were found in our “Finished Goods Inventory.” The ages of these capacitors ranged from 9 to 2 years. We life-tested samples of these stock units at +125°C without voltage derating. The actual test voltages were extreme, from 240% to 400% of the actual voltage rating. For each date code we tested 25 units of these axial leaded, non-hermetic, tubulars for 500 hr ^[6].

These life tests illustrate that PPS can operate reliably at +125°C, without the need of the voltage derating required by typical PC capacitor designs. PPS can also be used at temperatures as high as +150°C, and capacitors produced with PPS have an exceptional shelf life.

In the course of collecting data to qualify a PPS alternative (MIL-PRF-83421/06) to the long established PC based MIL-PRF-83421/01 style capacitor, almost 24,000,000 unit-test hours have been collected. This clearly demonstrates that PPS is a reliable dielectric.

Another performance attribute of particular concern in ac applications is a dielectric's loss characteristics. Capacitor losses are typically represented by the dissipation factor (DF), in units of percent and equivalent series resistance (ESR), expressed in mΩ. The DF and ESR performance of a capacitor with temperature and frequency are critical criteria when making capacitor selections for ac applications. If these values are too high at the application temperature or at the operating frequency, it will result in reduced capacitor life, or in the extreme case, catastrophic failure.

To compare the DF and ESR for PPS and PC, we constructed capacitors of 2µF, 50Vdc with the same dielectric film thickness, using the same form factor, and identical termination methods. Figs. 2 and Fig. 3, on page 37, compare PPS and PC for DF and ESR, showing that PPS exhibits lower DF and ESR vs. frequency than PC.

Temperature Effects

Of all the comparisons between these two dielectrics, the effect of temperature on DF has caused the most skepticism about the ability of PPS to replace PC. In Fig. 4, on page 38, you can see that PPS experiences a dramatic increase in DF starting at approximately +100°C, while PC remains relatively flat up to +125°C.

If the increasing DF behavior of PPS with temperature is actually a limiting factor, then ac operation at temperatures above +100°C should result in an immediate capacitor failure due to a thermal runaway condition or a very short life observable by degradation in the electrical characteristics. To test this hypothesis, we conducted two experiments. One was to apply ac voltage at various frequencies to a capacitor at +125°C while monitoring the PPS capacitor's heat rise. The other was to run a direct comparison of PC and PPS capacitors during +125°C ac operation for an extended period.

In the first experiment, we constructed an 8µF, 300Vdc PPS capacitor using 9µm thick film, as an 18AWG copper axial leaded, wrap-and-fill tubular. The capacitor size was about 1.25-in. in diameter by 2.25-in. long. We took measurements initially at +25°C of capacitance at 1 kHz, DF at 1 kHz, ESR at 100 kHz, and insulation resistance (IR) at 300Vdc.

We then fitted the capacitors with a thermocouple and placed them inside a pressboard box of about 5 in.×3 in.×2 in. We added another thermocouple inside of this container. The entire setup was then placed inside an air-circulating oven stabilized at +125°C until all three thermocouples (the unit, the box, and the oven) agreed. The purpose of the pressboard box was to prevent convection cooling of the capacitor surface. After stabilization, we applied a sinusoidal ac signal to the unit and monitored the heat rise of the capacitor body until this temperature had remained unchanged for at least three hours. We repeated this four different times, at four different ac signal frequencies. Table 3 lists the results of these tests.

All these tests resulted in the PPS capacitors reaching temperatures in excess of +125°C, actually approaching +140°C. This is well above the established PC maximum of +125°C and the suggested maximum ac operation temperature of +100°C. The tests showed that the PPS can withstand temperatures of +150°C ac operation at temperatures where the DF characteristics are quite high compared to PC.

Out of curiosity, a PC capacitor of similar rating and size was used to replace the PPS unit in one of the 40,000 Hz tests. Shortly after the ac signal was applied, the end of the PC capacitor melted off.

At the conclusion of all testing, the PPS capacitors were checked again for the same electrical characteristics measured prior to any testing. No degradation was noted in capacitance, DF, ESR, or IR.

We performed additional ac testing comparing PC and PPS snubber units constructed with hybrid metalized/extend foil electrodes ^[7]. The applied ac sinusoidal voltage was 247Vac at 5400 Hz, and the test temperature was +125°C. The PC and PPS units were rated at 0.1µF and 1000Vdc. The approximate capacitor size was 0.500-in. diameter by 1.800-in. long and used 18AWG copper wires as terminations. Electrical measurements were made prior to the test and after 1000 hr of operation. Thermocouples were fitted to the body of five capacitors of each dielectric type to monitor heat rise during operation compared to the test chamber temperature.

The results of this ac life test showed that neither the PC nor PPS capacitors were degraded after 1000 hr of ac operation at +125°C. However, the PPS capacitors did run hotter than the PC capacitors by approximately 5°C.

These ac tests clearly demonstrate that PPS, even with its increasing loss characteristics with temperature, can replace PC in ac applications without any dire consequences.

The final comparison between PC and PPS is the insulation resistance (IR) performance with temperature. Insulation resistance is typically expressed in units of MΩ×µF. Fig. 5 is a plot of IR performance from +25°C to + 150°C for PC and PPS. Their performance is essentially the same.

Economics is a major consideration in today's competitive business environment. If the cost is astronomical to implement replacement material or to solve a problem, then you have no solution at all — just interesting information.

PPS can replace PC. It does have advantages over PC for certain applications. While the cost of a capacitor depends on many factors, the cost of PPS dielectric film is comparable to the historical cost of PC dielectric film before it was announced that it would be discontinued. Furthermore, the supply of PPS has never been interrupted — unlike PC — since the material was first made available in film form as a capacitor dielectric some 20 years ago.

Novel Option

Although PPS is a great alternative to PC, there is another option. Preliminary experimental work has shown that by creating a monolithic device using existing commercially available film dielectrics, it's possible to create another material with the capability of +125°C operation with excellent ac capabilities. For purposes of this discussion, this dielectric will be referred to as monolithic polymer film (MPF).

While the evaluation of MPF is in the early stages, some ac and dc test data has been collected. Tables 4 and 5 summarize the test data collected at the time of this writing. Table 4 provides the MPF results at +125°C on dc life test at 140% rated voltage. Table 5 provides MPF results at +125°C ac operation. The MPF capacitors evaluated in both tests were 20µF, 200Vdc rated, 18AWG axial, copper leaded, non-hermetic configurations, 1.3-in. diameter by 2.25-in. long.

Based on these evaluations, you can use MPF at +125°C in ac and dc applications. There are some differences in performance between the MPF compared to both PC and PPS. You can see these differences summarized in Table 6.

Although the current formulation of the MPF capacitor will not enable it to achieve +150°C operational temperature like PPS, it does exhibit lower DF characteristics with temperature and frequency. This makes it an ideal candidate for extreme ac applications. The major difference is the dielectric constant of MPF, which is less than the dielectric constant of PPS and PC (3.0). Therefore, MPF capacitors will be larger, but still capable of +125°C operation. Nevertheless, this makes it a possible PC replacement.

For now, the future of obtaining capacitors capable of operation at +125°C, as well as +150°C, has been secured by the ability of the PPS dielectric to directly replace PC. In almost every respect, PPS compared to PC is nearly identical or superior in performance. In spite of this, we are always searching for the “perfect” dielectric film and those that are capable of operation at temperatures approaching +200°C or higher.

Experimental films with the greatest potential being investigated today are Dow Chemical's polyparaphenylene benzoaxole (PB), Foster-Miller's polyparaphenylene benzobisthiazole (PBT), and 3M Corp.'s proprietary aromatic polyester (FPE) ^[8].

Many technical papers have been written in the past few years comparing potential high temperature polymer films; however, none of these films have become commercially available. They continue to be merely interesting topics for presentations and research subjects for academia. Grzybowski and McCluskey have stated that, “no manufacturer has expressed any real confidence in the development of a series of high temperature wound film components…” ^[9]. Sadly, these potential high temperature dielectrics linger tantalizingly outside the reach of the component manufacturer since no one has committed to producing these materials on a commercial scale.

Acknowledgment

The assistance in this work of a number of individuals at Dearborn Electronics Inc. is gratefully acknowledged. Without their contribution, this work would not have been possible. They are: Donna C. Meister, Douglas H. Smith, James E. Hollborn, Mark W. Rumler, William M. Donelan, Fredrick T. Rotolo, and James P. Sherry.

References

Anderson, Ron, “Select The Right Plastic Film Capacitor for Your Power Application,” Proceedings of IEEE Industrial Applications Society, San Diego, Calif., October 7, 1996.
Adleson, Leonard and Rice, Herbert L., “Polycarbonate-Foil Capacitors,” Sprague Electric Company Technical Paper No. 63-5, originally presented at Electronic Components Conference, Washington, D.C., 1963.
Roos, Gina, “Film Cap Makers Facing Crunch,” Electronic Buyers News, September 9, 2000.
Biancomano, V., “Film Capacitors,” Electronic Design, November 13, 1986.
Mannheim, Daniel, “Polyphenylene Sulfide (A New Film Capacitor),” Proceedings of Capacitor and Resistor Technology Symposium (CARTS), Garden Grove, Calif., 1987.
Dearborn Electronics Inc., 880P Product Bulletin.
Kowalsky, Jerold D. and Rice, Herbert L., “A-C Applications of Plastic Film Capacitors,” Proceedings of Capacitor and Resistor Technology Symposium (CARTS), Phoenix, 1983.
Zogbi, Dennis M., “Extruded and Metallized Films for the Global Capacitor Industry,” Passive Component Industry, May/June 2000.
Grzybowski, Richard R. and McCluskey, F. Patrick, “High Temperature Performance of Polymer Film Capacitors,” Transactions of the 3rd International High Temperature Electronics Conference, Albuquerque, New Mexico, 1996.