Ceramic capacitors are rapidly increasing in capacitance and volumetric efficiency (CV/cc) due to higher dielectric constants and smaller dielectric thickness as well as higher layer counts. To compete with ceramic capacitors and meet demands for miniaturization, tantalum (Ta) capacitors also need to increase their volumetric efficiency.
Traditionally, the only way to increase CV/cc in Ta capacitors was to reduce particle size in the Ta powder, thereby increasing the surface area of the anode. This leaves aside packaging efficiency as a common issue for all types of capacitors.
Although Ta powder manufacturers continue to increase powder CV, the application of newly developed, high-CV powder is limited to low-working-voltage (WV) capacitors with very thin dielectrics. Higher-voltage capacitors cannot use high-CV powder because they require a thicker dielectric than that in the low-voltage parts.
The thicker dielectric grows through the “necks” between the powder particles and clogs fine pores between particles, reducing anode surface area and, thereby, CV. This means that for mid-voltage and high-voltage capacitors, which constitute the bulk of Ta capacitors, the applicable powders have been in use for some time.
High reliability and stable leakage current are critical for demanding applications. Issues such as difficult or impossible accessibility for repair, the high cost of equipment, and the potential for personal injury require reliability of the highest level. Additionally, stable dc-leakage (DCL) characteristics are necessary to ensure that designers can develop power supplies that serve their target purposes.
Recently developed techniques do improve the use of Ta in anodes for the creation of additional capacitance, providing higher CV/cc for Ta capacitors while simultaneously improving their stability and reliability. To meet the growing demands of critical applications, newly developed processes show great potential.
ANODE CV VS. POWDER SINTERING CONDITIONS
Press density (d) and sintering temperature (Ts) of Ta powder are the two major parameters that influence utilization of Ta in Ta anodes. Figure 1 shows CV/cc as a function of d and Ts in anodes sintered with 23k CV/g Ta powder. This data shows that it’s possible to increase CV/cc at high press density and low sintering temperature.
But with conventional sintering in a vacuum, low sintering temperature doesn’t provide sufficient bonding between the powder particles and between the particles and lead wire, affecting the mechanical and electrical properties of sintered anodes. This is due in part to oxygen, which dissolves in Ta particles from natural surface oxide during sintering in a vacuum and acts as a sintering inhibitor.
SINTERING IN REDUCING ATMOSPHERE
Sintering in a reducing atmosphere such as magnesium (Mg) vapor results in the removal of oxygen from Ta particles. This intensifies the diffusion of Ta atoms, allowing the growth of “necks” between powder particles at lower temperatures than when sintering in a vacuum.
Our alternative process allows low-temperature sintering in a deoxidizing atmosphere, the initial results of which appear in Figure 2 for 50k CV/g Ta powder. The figure 2 shows CV/cc, oxygen content, and delta volume in Ta anodes sintered in a vacuum (Sintering) and sintered with deoxidizing (D-sintering).
D-sintering increases CV/cc by approximately 35% and radically reduces oxygen content in sintered anodes versus sintering in a vacuum. Increases in CV/cc with D-sintering incur anode expansion, while sintering in a vacuum results in anode shrinkage (Fig. 2b). This difference in volume change between regular sintering in a vacuum and D-sintering is the result of a change in the dominant sintering mechanism.
With low-temperature D-sintering, the dominant sintering mechanism is surface diffusion of Ta atoms. This results in open pores and an expansion of the anode volume, providing the highest possible volumetric efficiency. The reduction in oxygen content also improves dc-leakage current behavior.
Continue on Page 2
LOW AND STABLE DCL CURRENT
The typical way to reduce and stabilize DCL in Ta capacitors is to reduce the electrical field in the dielectric, either by increasing the form factor, which is a ratio between formation voltage and working voltage, or by derating, which is a ratio between the application voltage and the component’s working voltage. In both cases, DCL recedes exponentially. The rate of degradation processes that promote a DCL increase also retreats exponentially.
The problem with this approach is that the volumetric efficiency of Ta capacitors is inversely proportional to both form factor and derating. This is why the real challenge for Ta capacitors is the combination of high stability and reliability with the highest possible volumetric efficiency.
Low oxygen content in the bulk of D-sintered anodes allows suppression of crystallization of the amorphous dielectric, which is the major degradation mechanism in mid- and high-workingvoltage Ta capacitors. To enhance stabilization of the amorphous dielectric, this sintering technique can pair with other anode and dielectric manufacturing techniques that provide chemical purity and structural uniformity to Ta anodes and oxide dielectric.
When implementing this combination of techniques, the dielectric remains amorphous during testing and field application. As a result, DCL stays low for a long period even in conditions that are harsher than normal application conditions.
Figure 3 illustrates the DCL distribution in 22-µF/20-V, D-case Ta capacitors (a) as well as after 2000 hours of life test at 1.32 working voltage (WV) and 85°C (b). Capacitors under test consist of either those employing the crystallization-preventing package of techniques (test group) or conventional technology (control group).
Initial DCL was well below the limit and practically identical for both test and control parts. Yet only parts with the test technology retained low DCL during the life test, while the DCL of the control parts increased by more than an order of magnitude.
An important feature, the capacitors shown in Figure 3 employ 50k CV/g Ta powder, the highest CV/g powder in the industry that is applicable for 20-WV capacitors. The use of higher-CV/g powder provides higher volumetric efficiency to Ta capacitors. However, it also promotes crystallization of the dielectric and, thereby, DCL increases during testing and in the field.
Nevertheless, even with this powder, capacitors employing the package of crystallization-preventing techniques demonstrate low and stable DCL during long accelerated testing. These techniques provide a powerful tool for improving the performance of hightemperature, extended-range, and wet Ta capacitors.
A NEW SCREENING TECHNIQUE
Use of the highest-purity raw materials, advanced technology, and the most sophisticated machinery cannot guarantee that all capacitors will have flawless dielectrics. Some may have defects in their dielectric due to occasional contamination or damage resulting from machinery malfunction or human factors.
Hidden defects in the dielectric that do not heal by the endof- line aging and go undetected by final testing can progressively worsen during field operation and cause capacitor failure. That’s why accelerated aging, surge testing, reflow testing, and other procedures are part of the manufacturing process of Ta capacitors.
One problem with these techniques is they cannot guarantee exclusion of all the non-reliable parts. When intensified, they also can deteriorate the performance and reliability of the general population of the capacitors as a result of testing.
To address this problem, we have a special technique that allows screening of non-reliable capacitors with hidden defects in the dielectric without any damage to the general population of capacitors. Screening entails the simulation of breakdown voltage (BDV) without actually damaging parts.
Continue on Page 3
Low BDV indicates defects in the dielectric and, therefore, a high probability of failure in the field. High BDV indicates flawless dielectric. Figures 4 and Figure 5 show the distribution of screening voltage within a lot of 100-µF/16-V, X-case solid Ta capacitors and the resulting data, respectively. Screening voltage correlates with actual BDV in individual capacitors.
As per the example, about 95% of the distribution lies in the narrow range of voltages while 5% of the distribution spreads out toward low voltages. DCL readings in all parts were much lower than the DCL limit for this rating. According to data in Figure 5, there’s no effect on DCL readings during the screening, confirming the non-destructive nature of the screening process.
Additionally, Figure 6 demonstrates the results after accelerated life testing on screened capacitors versus non-screened capacitors. We observe failures at power-on in non-screened capacitors at an early stage during this accelerated test. In contrast, screened capacitors do not show any early failures, and their time-to-failure distribution is uniform.
Crystallization-preventing techniques provide low and stable DCL as well as higher volumetric efficiency for Ta capacitors with mid-range and high working voltages. Teaming with the unique breakdown simulation screening process, the capacitors demonstrate exceptional reliability. In addition, these anode improvement and testing techniques apply to both surface-mount tantalum and wet tantalum capacitors to provide high reliability and stable leakage current characteristics.