Beyond MLCCs: The Rise of the Silicon Capacitor

Whether used for energy storage, power decoupling or tuning and filtering, capacitors are critical components in every electronic design. Nowadays, multi-layer ceramic capacitors (MLCCs) have become ubiquitous, being deployed in everything from smart phones to electronic content-laden vehicles, As a result, the market for these miniature components is predicted to reach five trillion devices by 2025.

However, they are reaching the end of the road as a suitable solution to many designers’ needs.

Silicon capacitors are one way that engineers can address the latest design problems in terms of performance, size, stability and susceptibility to threats such as vibration, temperature, and electrical noise. Empower Semiconductor’s E-CAP technology is an example of how capacitors are keeping pace with advances in other component types.

 

Increasing Application Challenges for MLCCs

MLCCs have been great servants and workhorses for the electronics industry and their existence has been, and indeed will continue to be, important for many applications. But it is a fact that the direction of travel of some designs is increasingly exposing their limitations.

As data rates increase so do power requirements, and the need to address these in space envelopes that are typically getting smaller even as end applications incorporate more processing power and functionalities raises the specter of problems associated with noise and thermals.

Look inside many pieces of modern equipment and you will see a lot of board space occupied by MLCCs. The number of these tiny off-the-shelf devices is dictated by the high-power, data hungry nature of the applications. This is often further exacerbated by the requirement to not only extend the frequency bandwidth response of the design, but also guarantee the effective capacitance value once all deratings due to voltage, operating temperature and aging are accounted for. To mitigate these requirements, more and more devices are traditionally added to the circuitry. Ultimately the combination of these factors means the long-serving MLCC is often no longer an elegant, viable or sensible solution. Other specifications that become more of an issue as end application power and frequency increase, are Equivalent Series Resistance (ESR) and Equivalent Series Inductance (ESL). In simple designs, the low values of these parasitics that move the capacitor value away from being ‘pure’ – i.e. what it says on the can / package – have little or no impact on performance and reliability. However, as power and frequency are increased, efficiency can be impacted and unexpected performance and results seen. As a result, designers are actively seeking alternatives, one of which is the silicon capacitor.

 

Introducing High-Performance Silicon Capacitors

As power goes up and available space reduces, the spotlight falls on power density.

Silicon capacitors offer significantly increased power density, as illustrated by Empower’s E-Cap technology, which typically offers five times the capacitance density possible with MLCCs. E-CAP integrates what would have previously required multiple discrete components onto a single die in a single package. This dramatically reduces component count and can give a valuable five to seven times reduction in board real estate required for a given capacitive solution. BoM and total cost of ownership (due to single rather than multiple component insertions) are also cut back.

Interconnects made on the die within a package as opposed to on the PCB mean that the reliability and ruggedness of this approach are also better.

 

Figure 1. E-CAP solutions offer 5x capacitance density versus traditional MLCCs. Image used courtesy of Bodo’s Power Systems [PDF]

 

In terms of both electrical and physical specifications, no two applications are the same, so the ability to source integrated solutions that are customized for specific needs is appealing to design engineers. From an electrical perspective, E-CAP usefully allows matched capacitance values ranging from 75 picofarads (pF) to 5 microfarads (µF) (@2V) to be integrated into an array.

Passives have often been ‘afterthoughts’ to the key processing technology and, as a result, the shape of the available board space once major components have been placed may not be regular nor the enclosure height very generous. This means that the flexibility to offer an integrated package that matches the application requirements is vital. E-CAP solutions can take account of the X and Y dimensions on the designer’s wish list and with an overall package height down to just 50µm, headspace is less likely to be an issue. Additionally, packaging interconnect options based on bumps, pads or pillars allow designers to choose the best solution based on system constraints.

Integrated and versatile form factors also facilitate innovative approaches to locating capacitors - for example, ‘in-package’ and ‘under-package’ SoCs and in PCB filtering and bypass applications. As mentioned, with power and data intensity increasing in a smaller space envelope, performance issues can emerge, many of which are due to derating. In this area, silicon capacitors are less needy and cause minimal design headaches for engineers used to working with MLCCs.

For example, temperature de-rating for MLCCs can be in excess of 10% up to 85°C, while de-rating for the effects of aging can be in the order of 5 to 10% over 10,000 hours of operation. In a discrete MLCC solution, this can necessitate additional components with the associated cost and layout issues that presents. In contrast, E-CAPs are barely impacted by temperature and aging and so do not require over-design to mitigate effects. In addition, they require no AC or DC bias de-rating.

 

Figure 2. Small size and electrical and mechanical robustness open up alternative mounting options for integrated silicon capacitor solutions. Image used courtesy of Bodo’s Power Systems [PDF]

 

While ESR and ESL parasitic characteristics become a potential issue with MLCCs as power and frequency increase, silicon capacitor solutions exhibit these at much lower levels removing uncertainty around performance. In addition, although Empower’s E-CAPs have lower nominal capacitance, their superior frequency response and ESL over MLCCs results in lower impedance at high frequencies.

 

Custom Parts With Low Risk

The desire for shortened design cycles to get compelling new products to market quickly leads to the perennial choice of ‘make versus buy’ (custom versus off-the-shelf) solutions. MLCCs clearly fall into the latter category but thanks to an optimized design flow, custom silicon capacitor solutions can still help designers deliver within tight timescales.

Empower, for example, based on learnings from the development and launch of its Integrated Voltage Regulator (IVR) technology, has developed a process to smoothly take customer requirements for E-CAP solutions from concept to production.

 

Table 1. Compared to standard MLCCs, E-CAP technology almost entirely removes the need to consider de-rating.

Parameter	Standard MLCC	E-CAP
Temperature de-rating	-11% up to 85°C	Negligible-～0.3% (measured in ppm/K) - equivalent to COG
DC bias de-rating	44% @ 3V	None
Aging	～5-10%/10k hrs	<0.001%/10k hrs
ESL	>100pH (100nF)	<10pH (100nF)

 

At the beginning of the process this means a clear definition of technical requirements and application constraints, which then leads to a custom design proposal that can quickly move to design and the fabrication of engineering samples. The predictable performance of E-CAP due to minimal parasitics and the robust nature of an integrated package containing multiple capacitive elements, removes the risk from the process, resulting in parts that are tightly aligned to application needs.

 

This article originally appeared in Bodo’s Power Systems [PDF] magazine.