Technical Papers

THE LITTLEST BIG CAPACITOR-
AN EVANS HYBRID

David A. Evans
Evans Company
72 Boyd Avenue
East Providence, RI 02914

 

ABSTRACT

High-energy-density capacitors with a tantalum metal anode, Ta2O5 dielectric, aqueous electrolyte, and RuO2 cathode have been described.1 Known as Evans Hybrid capacitors, these devices have the high energy density of electrochemical capacitors with the improved a.c. behavior of electrolytic capacitors. Single-cell working voltages up to 215 volts have been demonstrated. Capacitors of this type, with multiple elements arranged in parallel within a single case, and improved packaging techniques have now been assembled. Performance data on a 50 V, 18 mF hybrid capacitor are presented.

 

INTRODUCTION

Electrolytic capacitors enjoy wide popularity because of their high working voltage, good a.c. performance, and low cost. Electrochemical capacitors (ECs) are attractive because they offer very high energy density, high power capability, and potential low cost. Now there is a capacitor offering the advantages of both technologies.

The Evans Hybrid capacitor (US patent 5,369,547) combines the best features of both these capacitor types. In the Evans Hybrid, the cathode in an electrolytic capacitor is replaced with a large capacitance value EC cathode. Because the EC cathode requires little volume, available space can be used to increase the size of the anode. The resulting capacitor has several times the energy density of the original. Now this technology has attracted interest in applications commonly featuring tantalum or aluminum electrolytic capacitors. Applications include pulse power, filtering, and communications. Capacitances up to 1 F and working voltages up to 450 V are often specified.

As the performance of electrical-particularly electronic-equipment improves rapidly, the performance of capacitors must also improve. Capacitor technology paces some applications. For a capacitor, improvement usually means better electrical performance and a higher energy density. The Evans Hybrid offers both-energy density comparable to ECs with a.c. performance like electrolytic capacitors.

The objective of this work was to develop working prototype hybrid capacitors which could match the electrical performance of aluminum electrolytic capacitors while offering substantial space and weight efficiencies.

 

DISCUSSION

The electrolytic capacitor is really two capacitors connected in series by an electrolyte. The total capacitance, CT, is described as follows

1/CT = 1/CA + 1/CC

where CA and CC are the anode and cathode capacitances. From this, one can see that CT will be less than either CA or CC. The electrodes of an electrolytic capacitor are electrochemically oxidized, producing a dielectric oxide coating. The thickness of this oxide can be adjusted by the manufacturer to give any desired cell working voltage.

ECs are also comprised of a pair of capacitors, which function as electrodes, series connected by an electrolyte. What sets ECs apart from electrolytic capacitors is the way charge is stored, and the absence in ECs of a dielectric. ECs rely on charge storage in the electric-double-layer. One type of EC, called a pseudocapacitor, also stores charge in the electrochemically reversible change in oxidation state of its electrode material. RuO2, the pseudocapacitor material used in this work, has a very high specific capacitance, on the order of 100 F/g of electrode. Because ECs lack a dielectric, however, the EC cell voltage is limited to the breakdown potential of the electrolyte. For aqueous electrolytes, this is about 1 V/cell. To built higher voltage devices, multiple cells must be assembled in series, increasing part count and production cost. Questions of voltage balance are often answered by reducing cell voltage, decreasing energy density and raising reliability concerns for high-voltage ECs. 2

Figure 1. Prototype hybrid capacitor schematic view.

The Evans Hybrid capacitor exploits the inherent high energy density of pseudocapacitor electrodes while maintaining the high cell voltage, higher rate capability, and simple design of an electrolytic capacitor. Figure 1 is a schematic view of a two-anode-pellet version of the design adopted for this work. Although these capacitors had the familiar stacked arrangement of ECs, all were single cells. Capacitors using from 1 to 4 anode pellets have been constructed in similar fashion.

The capacitors had a 0.5 mm thick polypropylene case which was heat-sealed upon assembly. Tab ends from the cathode foils and wire lead(s) from the anode pellet(s) exited and electrolyte was initially added through holes in the cover. These holes were then sealed with a polymer sealant/adhesive which also supported the nickel tab terminals. The anode pellet(s) were pressed from a commercial capacitor grade tantalum powder. These were vacuum sintered according to established practice, and had a density of about 5 g/cm3. The Ta2O5 dielectric was electrochemically formed on the anode(s) in a phosphoric acid solution. The cathode electrodes were single or double sided RuO2 on 0.05 mm tantalum foil, and had a capacitance of about 140 mF per coated side. The electrodes were spaced by 0.13 mm thick non-woven separators. Sulfuric acid solution was used for the electrolyte.

Figure 2 is a photograph of a one-anode-pellet hybrid capacitor according to the description. Although the form of these capacitors was circular, almost any shape could be made. Figure 3 is a photograph of a disassembled two-anode-pellet capacitor. The construction is simple and compact, and the modular design permits multiples of similar electrode elements to be stacked in parallel, increasing capacity or reducing equivalent series resistance (ESR). This allows a large number of device ratings to be built using a small set of common parts.

Figure 4. Volume fraction for hybrid capacitor components

Figure 5. Weight fraction for hybrid capacitor components

Figure 4 is a pie-chart showing how the volume of a three-anode-element hybrid capacitor is occupied. The relatively large electrolyte volume includes electrolyte filling the space within the anode. Relative weights of components for the same capacitor are shown in Figure 5. The weight fraction of active components-electrodes and electrolyte, is 90% for the device shown. This is the highest packaging efficiency yet achieved for this type of capacitor, and contributes to its high energy density. Included in the segment marked 'other' are the separators and terminals.

ESR is critical to capacitor performance. The ESR of an EC is primarily related to the cell geometric area, and the electrode thickness. Hybrid capacitors should display similar properties. Several hybrid capacitors of different voltage and capacitance ratings were made to investigate the effect of changing anode thickness and dielectric thickness on device ESR and a.c. performance. Table 1 lists the ESR for three of the devices. The 12 g anodes were twice the thickness of the 6 g anodes. From the table, one can see that increasing the number of anodes in parallel, or decreasing the anode thickness indeed lowers the device ESR, as was expected. In this way, the ESR can be designed, independent of voltage or capacitance rating.

Table 1. Hybrid capacitor characteristics

 item  capacitance
(120 Hz)
 voltage  ESR
(1 kHz)
 weight  dimensions
dia x height
 anode
pellet weight
 # of
anodes
 1  18 mF  50 V  20 mW  56 g  36 mm x 12 mm  12 g  3
 2  18.1mF 28 V   10.2 W  30 g  36 mm x 8 mm  6 g  3
 3  8 mF  32 V  <50 mW  20 g  36 mm x 7 mm  12 g  1


One goal of this work was to match the electrical performance of an aluminum electrolytic capacitor. Comparisons were made to a Nippon Chemicon series 36DA, 18 mF, 50 V high-capacitance, low-ESR aluminum electrolytic capacitor. A distinguishing characteristic of ECs is their highly non-ideal impedance behavior. The main cause of this is their use of porous electrodes. The result is an apparent sharp drop in capacitance with increasing dV/dt. This behavior is easily observed using a technique known as complex impedance analysis. Resistance and reactance data of a capacitor over a range of a.c. frequencies are gathered and compared. Where the data given above are "snapshots" of capacitor performance at a fixed frequency, impedance analysis gives a picture of performance over a frequency spectrum. From this, the performance of a capacitor in any application can be accurately predicted, and a comparison between various capacitor technologies can be made.

Figure 6. Nyquist plot comparing the a.c impedances of a 18 mF, 50 V hybrid capacitor and a 1 F, 1 V RuO2 pseudocapacitor.

Impedance data can be presented in a number of ways useful to the engineer. One, known as the Nyquist plot, giving reactance vs. resistance as a function of frequency, can be used to investigate the behavior of porous electrodes. The Nyquist plot of an ideal capacitor is a vertical line intersecting the resistance (R) axis at the ESR value. Capacitors with porous electrodes display a characteristic curve. Figure 6 is a Nyquist plot comparing the a.c impedances of a 18 mF, 50 V hybrid capacitor and a 1 F, 1 V RuO2 pseudocapacitor. For comparison, 56 of the 1 F, 1 V ECs connected in series would have a rating of 18 mF, 50 V, the same capacitance and voltage rating as the hybrid capacitor, but the ESR would be many times higher. The curve for the EC shows the usual porous electrode behavior, but the hybrid capacitor curve is nearly vertical, indicating very ideal behavior. In Figure 7, the Nyquist plot for the hybrid capacitor is displayed with the plot from the 18 mF, 50 V aluminum electrolytic capacitor. While the hybrid capacitor and the electrolytic capacitor differ slightly in ESR, the overall shape of the curves is very similar. Further comparisons are made in the following figures.

Figure 7. Nyquist plots of an 18mF, 50 V hybrid and an 18 mF, 50 V aluminum electrolytic capacitor.

Figure 8 shows the resistance and reactance vs. frequency for the hybrid and aluminum capacitors. Over the frequency range of 0.01 Hz to 100 kHz, these capacitors exhibit nearly identical impedance behavior. The point of lowest reactance occurs at the frequency of self-resonance. Figure 9 compares the phase angle vs. frequency for the hybrid and aluminum capacitors. This is the difference in phase between the peaks in the voltage and the current waveforms in an a.c. circuit. An ideal capacitor has a phase angle of -90°, meaning the voltage lags the current by _ of a wavelength. A resistor has a 0° phase angle, and an inductor has a +90° phase angle. For efficient capacitor use, the a.c. frequency should remain below the point on the plot where the phase angle is -45°, about 1 kHz for the capacitors shown.

Figure 8. Resistance and reactance vs. frequency for an 18 mF, 50 V hybrid and an 18 mF, 50 V aluminum electrolytic capacitor.

Figure 9. Phase angle vs. frequency for the hybrid and aluminum capacitors


The plot shown in Figure 10 is useful in comparing variation in capacitance over frequency. The model here is a series RC circuit, and

C = 1 / (2 p f Z'')

where C is capacitance, f is frequency, and Z'' is the magnitude of the reactance. This plot gives capacitance vs. frequency for the 18 mF, 50 V hybrid and aluminum electrolytic capacitors. The two curves have a very similar shape, with the aluminum electrolytic capacitor having slightly lower capacitance. The apparent spike in capacitance value for both devices at a frequencies above about 2.5 kHz is not real. As Z" ® 0 in the preceeding equation, C ® ¥, in other words, C spikes. Since this happens outside the frequency range of efficient operation anyway, it is of little consequence.

Using these methods, the most efficient capacitor for a particular application can be readily selected. In addition, changes in capacitor performance over time are easy to detect.

Figure 10. Capacitance vs. frequency for the hybrid and aluminum electrolytic capacitors.

Other aspects of the 18 mF, 50 V hybrid and aluminum electrolytic capacitors are compared in Table 2. From the relation

E = 0.5C V2,

where E is energy, C is capacitance, and V is voltage, the energy stored by each capacitor is 22.5 joules. From Table 2, the hybrid capacitor has a gravimetric energy density 2.5 times higher, and a volumetric energy density 11 times higher than the comparable aluminum electrolytic capacitor!

 

CONCLUSION

An Evans hybrid capacitor ESR and capacitance can be designed for any desired value using a small set of common parts. The electrical performance of the hybrid is very similar to the performance of an aluminum electrolytic capacitor of the same capacitance and voltage rating. The savings of volume and weight for the hybrid are considerable. The energy density of the 50 V hybrid closely approaches that of a 50 V EC, but with far superior a.c. performance. The Evans hybrid capacitor is a very attractive alternative to aluminum electrolytic capacitors in applications where energy density is important.

 

Table 2. Weight, volume, and energy density of the aluminum electrolytic and hybrid capacitors.

 capacitor weight (g)   dimensions
dia x height (mm)
 volume
(cm3)
gravimetric energy
density (J/g)
 volumetric energy
density (J/cm3)
18 mF, 50 V
hybrid
 56  36 x 12  12.2  .40  1.84
 18 mF, 50 V
aluminum
electrolytic
 141  51 x 67  136.9  .16  .164

 


REFERENCES

1. D. A. Evans, "High Energy Density Electrolytic-Electrochemical Hybrid Capacitor", Proc.14th Capacitor and Resistor Technology Symposium, 3/94.

2. J. R. Miller, "Electrochemical Capacitor Voltage Balance: Cell Uniformity Requirements for High-Voltage Devices", Proc. 36th International Power Sources Symposium, pp. 15-18, 6/94.

 

 

 

Back to Technical Papers Pag

 

 

© Copyright 2007 Evans Capacitor Company. All rights reserved.