High Voltage Tantalum Polymer Capacitors

By Jayson Young, Product Manager Tantalum & Aluminum; Jake Qiu, Senior Scientist; Randy Hahn, Senior Director Valve Metal Technology; KEMET Electronics

Solid tantalum (Ta) capacitors that have intrinsically conductive polymers as the cathode material have been widely used in the electronics industry due to their low equivalent series resistance (ESR) and “non-burning/ non-ignition” failure mode. Intrinsically, conductive polymers are electrically conductive at the molecular level. Unlike polymeric materials whose electrical conductivity is imported from the presence of foreign conductive particles (for example, silver paints or adhesives), a single molecule (a polymer chain) of an intrinsically conductive polymer is conductive. Various types of conductive polymers, including polypyrrole, polyaniline, and polyethyldioxythiophene (PEDT), can be applied to dielectric oxides to serve as the primary cathode material for solid electrolytic capacitors. Two methods of coating the dielectric oxide with conductive polymers are widely used in the industry—in-situ polymerization and electrochemical polymerization. In in-situ polymerization, the monomer of an intrinsically conductive polymer is brought into contact with an oxidizer to drive the polymerization reaction. In electrochemical polymerization, the reaction is driven by an externally supplied current. The major drawback of conductive polymer capacitors, regardless of the type or method of application, is their relatively low working voltages compared to their MnO2 counterparts. For example, the maximum voltage rating of existing surface mount polymer tantalum capacitors is 25 V, while that of their MnO2 counterparts is 50 V. When subjected to voltages exceeding 25 V, the existing polymer capacitors have, to varying degrees, reliability issues. This limitation has restricted the use of these devices to relatively low voltage applications (<20 V). New developments in material and process technologies have overcome these limitations leading capacitors, including both polymer- (polyethyldioxythiophene, or PEDT) and MnO2-based capacitors, was measured. BDV is plotted against formation (anodization) voltage in Figure 1. to the introduction of a new line of high voltage tantalum polymer capacitors with excellent long-term reliability. Research and Analysis: The ability to withstand high voltage can be best characterized by the breakdown voltage (BDV) of the capacitors. Higher BDV corresponds with better reliability at a given application voltage. In an attempt to understand the differences in behavior between polymer Ta capacitors and their MnO2 counterparts, the BDV of a wide range of Ta capacitors, including both polymer- (polyethyldioxythiophene, or PEDT) and MnO2-based capacitors, was measured. BDV is plotted against formation (anodization) voltage in Figure 1. [caption id="attachment_504" align="alignright" width="252" caption="Figure 1: BDV Characteristics of Existing Conductive Polymer and"][/caption]
In the low formation voltage region (<30 V), as shown in Figure 1, the BDV of both polymer and MnO2 capacitors are close to the formation voltage. However, as formation voltage increases above 30 V, the BDV curves for MnO2 and conductive polymer capacitors diverge. For MnO2 capacitors, the BDV continues to increase with increasing formation voltage; whereas, the BDV of polymer capacitors levels out. Above a formation voltage of approximately 100 V, the BDV of conductive polymer capacitors is almost unaffected by further increases in formation voltage. This explains why increasing dielectric thickness, the most important and commonly used approach in making high voltage capacitors, is virtually ineffective for making high voltage polymer capacitors beyond about 25 V ratings. Higher voltage-rated capacitors, 35 V for example, would require a BDV of much greater than 60 V to ensure their sound surge performances and long-term reliability.[caption id="attachment_505" align="alignright" width="252" caption="Figure 2: BDV of Polymer Anodes Using Alternative PEDT and PANI"][/caption]

The relatively low BDV of polymer capacitors suggests that the dielectric in polymer capacitors has been degraded. Efforts were made to understand the causes of this problem. Through a series of studies, a breakthrough was found when comparing the BDV of 25 V-rated capacitors using various polymer processes. Anodes using the existing polymerization (PEDT) process were evaluated and compared to anodes with a new/alternative PEDT and/or PANI polymer process. As illustrated in Figure 2, on page 24, a higher BDV was obtained with the new polymer process regardless of the polymer compositions (PEDT or PANI).

In addition to high BDV, the anodes processed with the alternative polymer process also showed excellent “selfhealing” behavior, which is illustrated in Figure 3. As the applied voltage increased, the current passing through the anode decreased, demonstrating the characteristic aging behavior attributed to self-healing in a solid electrolytic capacitor. The current spikes at about 78 V, which indicates a partial breakdown of the dielectric. However, the consequent drop of the current under constant voltage demonstrates the damaged dielectric was gradually “healed,” possibly due to thermal degradation or de-doping of the polymer at the defective dielectric sites.

Figure 3: “Self-Healing” Behavior of New Polymer Process Anode

Developing High Voltage Polymer Capacitors:
As evidenced in this article, the use of this new process presented a solution to the low BDV problem from which the existing polymer technology was suffering. The next technical challenge was to make polymer capacitors with desired high BDV and good reliability while maintaining other important properties such as low ESR and good stability. To achieve this, the new polymer process was placed into a pilot scale development project and designated as the T521 series to differentiate the new process from the more established polymer process. BDV of the T521 Series Figure 4 compares the BDV of anodes coated with MnO2 as the cathode, anodes processed using established polymerization processes, and anodes processed using the new process. The BDV of capacitors processed with the new polymerization process possess a BDV performance level very similar to the MnO2 capacitors.

Figure 4: BDV Characteristics of Existing Conductive Polymer, MnO2, and the New Process Capacitors


Figure 5 is a BDV probability plot of a T521 series 15 μF/35 V capacitor. The median value (50%) is 78 V, far exceeding the BDV of any available polymer tantalum capacitor.
[/caption

Reliability Testing of the T521 Series
In recent years, multiple studies have been conducted to compare and quantify the reliability of polymer tantalum devices (including PEDT) to their MnO2 predecessors. These studies reached similar conclusions in that the polymer cathode systems are, at the least, as reliable as the well-established MnO2 cathode design when used within their recommended operating ranges. In order to qualify the new polymer process, additional reliability studies were conducted. Components were placed on a conventional 85°C life test and charged at rated voltage for 2000 hours.

Figure 6: 85ºC Life Test for New Polymer Process (DC Leakage)

Figure 5: BDV Behavior of a 15 μF, 35 V, 100 mΩ, 7343-19 Case Size Polymer Device


Figure 7: 85ºC Life Test for New Polymer Process (ESR)

Initial measurements were taken immediately following board mounting of the devices. The devices were measured again after 250, 500, 1000, and 2000 hours of testing. The reliability of the polymer tantalum capacitor can best be determined from assessing the DC leakage and ESR performance of the capacitor over the duration of the life test. Following 2000 hours of testing, the dielectric showed no signs of degradation, as evidenced by the reduction in DC leakage shown in Figure 6. In addition, the ESR performance was found to be stable throughout the 2000 hours of testing (Figure 7). This initial study demonstrates the new polymer process has, at a minimum, reliability characteristics similar to that of the currently available low voltage PEDT polymer tantalum products.

Electrical Performance:
The primary goal for developing a polymer formulation for higher voltage polymer tantalum capacitors was to improve the electrical performance. While the benefits of a “non-burning/non-ignition” failure mode and a much improved voltage derating factor resulted from the use of intrinsically conductive polymers as a replacement for MnO2, these polymers were originally developed as cathode systems for solid electrolytic capacitors in the early 1990s in order to reduce ESR and the RC ladder effect. To quantify the benefits this technology would bring to higher voltage applications, a comparison study was conducted using several MnO2 tantalum capacitor technologies commonly used for higher voltage applications such as decoupling 20 to 24 V power input rails.

Component Selection for a 20 to 24 V Power Input Rail:
For this comparison, component selection was based on the highest capacitance value and lowest ESR that has been achieved in a 50 V MnO2 tantalum chip design. Today, the highest CV ratings commonly available in a standard commercial series design offers up to 15 μF of capacitance in a 7343-43 package size with advertised ESR limits of around 700 mΩ. Low ESR designs are also commonly available with ESR offerings as low as 200 mΩ.

In addition, a more costly multi-anode design of the same capacitance value and case size was offered with an ESR limit of 75 mΩ. Due to the polymer device’s ability to perform more reliably closer to its rated voltage (30% recommended voltage derating versus a 50% recommended voltage derating for MnO2), a 35 V-rated polymer component was selected with an ESR limit of 100 mΩ. The advantage of using a lower-rated voltage polymer device yielded the additional benefit of a much smaller package size (low profile 7343-19). Figure 8 summarizes the list of components that were selected for this evaluation.

Figure 8: Component Selection for Performance Comparison
Polymer Ta (T521 Series) 15 μF 35 V 30% 100 mΩ 7.3 x 4.3 x 1.9 (Low Profile)
MnO2 Ta (Commercial) 15 μF 50 V 50% 700 mΩ 7.3 x 4.3 x 4.3
MnO2 Ta (Low ESR) 15 μF 50 V 50% 200 mΩ 7.3 x 4.3 x 4.3
MnO2 Ta Multi-Anode (MAT) 15 μF 50 V 50% 75 mΩ 7.3 x 4.3 x 4.3

ESR Performance:
Initial ESR measurements were taken from 10 KHz to 1 MHz to establish a baseline for ESR comparison of the four component types (Figure 9). The resulting ESR demonstrates the advantages of the polymer cathode when compared to the commercial and low ESR MnO2 devices. In addition, the polymer design was found to have only slightly higher ESR than the much larger and more costly multianode MnO2 design when operating at frequencies below 30 KHz and showed no significant difference in ESR at frequencies above 30 KHz.

Figure 9: ESR vs. Frequency for Tantalum Capacitors


Capacitance vs. Frequency
With ESR behavior documented, the RC ladder effects were then analyzed to determine the effect on capacitance

Figure 10: Capacitance vs. Frequency for Tantalum Capacitors


with frequency. As shown in Figure 10, the commercial and low ESR MnO2 technologies lose 67% and 40% of capacitance, respectively, at around 300 KHz while the MAT device loses only 14% of initial capacitance. The polymer device demonstrates a capacitance response similar to that of the MAT device with only a 13% drop in capacitance at 300 KHz.

dv/dt Response:
Once this initial assessment of component performance was completed, the next objective was to demonstrate how this technology would be of benefit on higher voltage input rails when compared to MnO2 technologies. The improvements in ESR and capacitance roll-off can be viewed in a time domain, as shown in the dv/dt plot in Figure 11. Using the same four tantalum capacitor technologies as above, a dv/dt plot was constructed to demonstrate this element. The dv is expressed as volts per ampere, as the current is another independent variable with this response. The dt is expressed in μS. We can see that there is no discernable difference in dv between the polymer and multianode MnO2 device up to and beyond 100 μS. However, the commercial and low ESR MnO2 technologies demonstrate a

Figure 11: dv/dt Plot of Tantalum Capacitors

more rapid decline in voltage almost immediately. If we look at a time interval of 30 μS, we see the MAT MnO2 and the polymer both experience a voltage drop of around 2.5 V per ampere. However, the commercial MnO2 and low ESR MnO2 experience a dv of 4.3 and 3.2 V per ampere, respectively, within the same time domain. By targeting a specific application need, a piece count assessment can be conducted. For this exercise, a dv/dt of less than 1 V per ampere/30 μS was selected. To maintain this requirement, the piece count of each capacitor technology was increased until the dv/dt was met. As shown in Figure 12, the minimum piece count necessary to maintain this dv/dt was five times commercial MnO2, four times low ESR MnO2, three times multi-anode MnO2, and three times polymer capacitors.

Figure 12: dv/dt Plot of Multiple Piece Count of Tantalum Capacitors (dv=1 V/dt=30 μS)

Cost Factors:
Regardless of the improvements seen in the performance of polymer technology, the bottom line is the total cost for the solution. To conduct a cost analysis, the dv/dt requirements analyzed were used to construct a cost model. In addition, a commercially available high cap MLCC structure was added to account for other volumetrically efficient package designs being considered today. The commercial MLCC was selected as the state of-the-art capability of an X7R dielectric 22 μF/25 V design in a 2220 SMT chip. The total cost solution for each of these component technologies is given in Figure 13. As shown, the polymer solution represents the lowest cost solution between the selected technologies given the lower piece count and smaller case size requirement. From this point, the cost increases significantly for the commercial MnO2 and low ESR MnO2 solutions due to the higher piece count required and larger case size. The high cap ceramic solution maintains the next highest solution cost due to poor dv/dt performance, which resulted in a much higher piece count as well as the higher premium placed on high cap ceramics. Finally, while the MAT MnO2 design maintained the same piece count as the polymer design, an even higher premium is required due to the higher manufacturing cost of MAT products and the larger case size.

Figure 13: Capacitor Cost for dv/dt Solution

References
1. Prymak, J., “Improvements with Polymer Cathodes in
Aluminum and Tantalum Capacitors,” IEEE 2001-APEC
Conference 2001.
2. Reed, E., “Characterization of Tantalum Polymer
Capacitors,” NASA Electronic Parts and Packaging
Program, NEPP Task 1.21.5, Phase 1, FY05.
3. Reed, E., “Characterization of Tantalum Polymer
Capacitors,” NASA Electronic Parts and Packaging Program,
NEPP Task 1.21.5, Phase 2, 2006.
4. Prymak, J., “New Tantalum Capacitors in Power Supply
Applications,”1998 IEEE Industry Applications Society
Annual Meeting, IEEE-IAS, St. Louis, 1998.
5. Prymak, J., “Replacing MnO2 with Conductive Polymer
in Solid Tantalum Capacitors,” CARTS USA 1999
Proceedings, The Components Technology Institute,
Inc., New Orleans, March 1999.
Editor’s Note: This research article was presented at CARTS USA
2008.

Readers Who Read This Story Also Look At These Additional Resources: (1) Conductive Polymer Capacitors: World Markets: Technologies & Opportunities: 2010-2015 ISBN # 1-893-211-88-6 (2010). (2) Capacitors: Costs To Produce: 2010 (ISBN #1-893211-00-2)