EE World Online organized and published “virtual roundtable” bringing together three experts in supercapacitor technology to share their experience and practical insights into supercapacitors: Chad Hall (CH), Co-Founder / Sr. Vice President Sales & Operations, with Ioxus; Eric DeRose (ED), Global Product Manager – SuperCapacitors, with AVX Corp.; and Jason Lee (JL), Global Product Manager for Supercapacitors, with Eaton Corp.
What determines the ESR of supercapacitors?
The primary effects for ESR are cell design and materials. Most of the ESR in supercapacitors is determined by how the manufacturer makes connections from the active electrode to the terminals that the customer sees. The next biggest effect are materials – electrolyte and electrode materials and how the electrode is processed.
CH: ESR in the supercapacitor cell comes from two things; chemistry and cell design. The chemistry used, depending on the carbons, conductive carbons, binders, separator, and electrolyte all lead to an ESR level. Thickness of the electrode will also affect the ESR. The cell design in how the electrode foils are attached to, and how much of the foils are attached to, to terminals, drive the cell ESR numbers as well. The lowest ESR is achieved by using a properly designed chemistry, with good connections to the terminals. Often, however, you need to look at how the ESR increases over time, rather than the initial rating. The increase in ESR is typically used to determine the system performance at end of life, and just because a cell as low ESR out of the box, does not mean it will stay low. Cycling, time spent at temperature and voltage, all play a role in the increase in ESR. Some manufactures have improved this ESR gain significantly by using chemistry. Lastly, since almost no ultracapacitor is used alone, the module construction and design drive a significant part of the ESR seen at a system level. Some companies screw/bolt bus bars to the cell terminals (higher ER), some weld bus bars to the cell terminals better), and one welds the cells directly together, end-to-end (best) to reduce ESR.
Is there a trade-off between supercapacitor ESR and other performance specifications?
ED: Achieving lower ESR comes at the expense of DCL (leakage current) specification, and vice versa. They are inversely proportional.
JL: Typically, higher capacitance means lower ESR. There is also the effect of temperature on ESR. The lower the temperature, the higher the ESR, especially as you get below freezing. However, ESR does not come down much as temperature is increased from room temperature.
CH: There is a definite trade-off between ESR and energy density of a supercapacitor. To me, an ultracapacitor should have lower ESR / high power – that is what it is chemically designed to be. You can increase the energy, but it causes increased ESR / lower power, and often lower cycle life. The proper system should have a high-power capacitor with a high energy battery, as that is what both chemistries were made to be.
What factors impact supercapacitor ESR in real-world applications – operating temperature? Discharge rate? If the ESR varies, is it a linear or non-linear characteristic?
CH: Temperature will affect ESR. Low temperatures (from approximately -20C to -40C) will see the ESR increase approximately 2x. The ESR stays fairly flat on the upper-temperature end. Over time, high temperatures will lead to degradation of the cell, generating gas and causing a reduction of capacitance and an increase in ESR. The variance ESR effect for temperature will show a fairly flat line from +65C to approximately -20C, where a noticeable increase occurs due to the conductivity of the electrolyte at lower temperatures. Self-heating will typically correct this within a few cycles.
ED: Since rise of ESR is considered a failure mechanism of supercapacitors, failure rate of supercapacitors, in general, are non-linear. Operating temperature and applied voltage have the biggest impacts on aging in the application. Current profile could come into play as it generates internal heat buildup depending on the frequency and depth of discharge.
JL: Temperature and voltage are the primary effects on ESR over time. The higher the temperature and higher operating voltage at the cell level will cause the ESR to increase faster. Temperature can also affect the ESR as the electrolyte mobility decreases in cold temperatures. Additionally, ESR variation over the operating life depends heavily on the manufacturer. Do they do burn-in prior to shipping, how pure are the materials, how clean is the manufacturing process.
How much does supercapacitor ESR vary between various device models? and from manufacturer to manufacturer?
ED: Supercapacitor ESR will certainly vary between models, series, and from manufacturer-to-manufacturer. In general, ESR decreases as capacitance increases. Furthermore, AVX manufacturers special low ESR designed supercapacitors (SCC LE Series) that feature low ESR characteristics compared to our own standard series (SCC Series) or in the industry. These can be critical in an application or for solutions that require multiple cells in series as ESR is additive. However, the sacrifice made to achieve that low ESR performance comes at the expense of DCL as they are typically inversely proportional. What I think would be certainly expected – you will definitely see ESR specifications vary throughout the industry. Design engineers should also be mindful of the specific wording used in supplier datasheets that could cause added confusion when comparing two products or consider what is truly a side-by-side comparison.
CH: ESR varies tremendously between cell types and constructions and from manufacturer to manufacturer. The lower-cost products often have a higher ESR to start with, and it increases significantly in a short period. Mid-range products will start with a low ESR, but that ESR will also increase quickly and faster as it ages. The best products on the market start with very low ESR and that ESR rise over time stays low, across a wide temperature range and throughout its life in cycling applications.
JL: On datasheet specifications, the ESR does not appear to vary a lot between manufacturers. There is some, but the actual as shipped ESR varies quite a bit. In addition, the ESR change as the part changes varies widely between manufacturers. As noted above, ESR is largely affected by the interconnects. Thus, the device model and form factor have a large effect on ESR. So, coin cells are orders of magnitude higher than cylindrical cells which are orders of magnitude higher than large axial cells. Typically, higher capacitance also relates to lower ESR as the electrode area is higher.
What are the non-ideal characteristics of actual supercapacitors compared with theoretical devices and what are the sources of those non-ideal characteristics? How can those non-ideal aspects be minimized?
JL: Lifetime effects are the first. A lot of testing and characterization is modeled. However, manufacturing variation can affect the lifetime, so parts do not perform to the model. This is typically overcome by building in the margin to the design, either lower operating voltage per cell or designing in higher capacitance.
CH: The non-ideal characteristics of supercapacitors (a.k.a. ultracapacitors or EDLCs) is the lower voltage limit than other electronics and the lower energy density than batteries. The chemistry (electrolyte) chosen typically dictates a number of the characteristics of a cell for the maximum voltage, which in turn drives the energy density. The formula of E=1/2CV2 (energy = one-half of the capacitance X the voltage squared) shows a direct correlation of energy to voltage. The material chosen (electrolytes are typically acetonitrile based or propylene carbonate-based), including the conductive salts, the carbons (typically organic-based from carbonized coconut shells for example), and the separator (cellulosic of a PTFE for example) are used. The carbon is activated via steam (lower contaminates, but less activation which leads to lower capacitance and lower power) or using a chemical process such as KOH (higher capacitance and power). All of these material choices lead to the ability to function over the life differently, and there are capacitors that use lower-cost materials that result in shorter life (higher capacitance loss or faster resistance gain), or products that take advantage of highly engineered materials to provide the highest power and energy over a longer life. The lower-quality materials result in more chemical side reactions, that drive leakage current, and poor performance. The life of an ultracapacitor is typically described at end-of-life (EOL) by reaching a 20% capacitance loss or a 200% increase in resistance (ESR or equivalent series resistance).
Conventional capacitors are offered in standard sizes with standardized electrical ratings. Are there industry-standard sizes (physical sizes or electrical capacities) for supercapacitors?
ED: In general, the same thing applies for standard-sized supercapacitors. For example, a 10x30mm can supercapacitor is generally 10 Farads across the industry. You may come across certain suppliers offer it as an 11F or 12F with different capacitance tolerances possibly, and this same trend spans in other can sizes as well, but you can partially chalk that up as marketing strategy. What truly differs are the other electrical parameters such as DCL (leakage current) or ESR (equivalent series resistance) that directly impact performance in use of the application. Those are absolutely not standard across the industry based on size or capacitance.
JL: Many of the sizes mimic electrolytic capacitors for supercapacitors used in electronics applications. Up to 600F, they come in cans leveraged from electrolytic sizes. There are also slight changes to this for customer-specific applications. As you get to larger capacitances, the “standards” have been set by the early suppliers. Today, we have 60mm diameter cells with varying lengths to meet capacities from 650F to 3400F. Coin cell sizes are based on button cell batteries.
CH: The industry that produces supercapacitors (a.k.a. ultracapacitors) has developed standard sizes over time, with some standardization on termination. There are two types of cells produced; laminated pouch cells (where the cells’ electrodes are stacked or wound in a flat manner) or cylindrical cells. Of the larger cells used in most power industrial applications, the standard ratings for voltage are 2.7V or 2.85V. Standard capacitance sizes are 100F, 350F, 600F, 1200F, 2000F, and 3000F. Termination ranges from flat tabs on pouch cells to solderable terminals on the 100F-600F, and screw or weldable terminals on the 600F – 3000F cells. There is an IEC specification for testing cells and UL-810A for safety testing ultracapacitors. More industry standards should be developed for testing products to make it easier for customers to choose the right quality product.
Are supercapacitors most-often used alone or in combination with other energy storage devices such as batteries? Or conventional capacitors?
CH: In most cases, supercapacitors are used without another energy storage product, but often there are advanced power electronics involved such as DC/DC converters, which may drive the need for filtering capacitors to reduce high frequencies across the supercapacitors. Due to the nature of ultracapacitors, being very high power (10-40kw/kg), long cycle life (1,000,000 charge/discharge cycles), the wide temperature range (-40C to +65C or even +85C), the low resistance (.02mOhms), or the high round trip efficiency (typically 95-99%), ultracapacitors should be paired with batteries (high energy, low power, low cycle life, narrower temperature window). Capacitors will take or give almost any current you want to give/take, and not care. A single 3000F cell is typically capable of providing 3,000A for 1 second. And it can do so with very little heat generation. This makes supercapacitors a wonderful product to use for large current demands (UPS, vehicle acceleration, cranes, automated guided vehicles, starting engines, wind turbine pitch control, etc.). The fast charge acceptance of a supercapacitor allows for much higher brake energy regeneration than advanced batteries, and they will happily perform this cycle a million times. Allowing the battery to handle the lower power, steady-state, energy demands, and the capacitor to handle the peak loads of acceleration or load shifts, makes an extremely efficient system design. Often, supercapacitors are looked at as too expensive, but that is usually because the designer is not sizing the system properly. Working with your ultracapacitor supplier in the early design phase often dramatically lowers system costs.
JL: In terms of number of cells, they are still used most often alone. However, there are many applications where they are combined with batteries such as water meters and electric busses. They are not typically combined with conventional capacitors.
ED: Supercapacitors are often used in conjunction with primary or secondary batteries. They are ideal for peak power assist applications where the supercapacitor provides the necessary current pulse(s) that would otherwise drain the battery, as well as power hold up applications. In general, they are a value add in extending lifetime of the battery and application. There are instances in which supercapacitors can completely replace a battery but those are usually much larger designs such as industrial applications.
What is the most misunderstood aspect of supercapacitor technology/operation? How does that translate into challenges for design engineers using supercapacitors?
JL: Cost remains the most misunderstood aspect of supercapacitors. We still get many inquires which ask what the cost per watt-hour is. It’s really understanding the advantages and limitations so that they are applied to the right application.
CH: The most misunderstood aspect of supercapacitors is often the system sizing. This is a challenge because designers will assume the cost is too high, and look past supercaps. If the power electronics are allowed to use a wider voltage window (full rated V to ½ rated V) and allow the full energy level of the supercaps to be used, it helps reduce the capacitor costs. Also part of the cost is the ultracapacitor module design and how that fits into the system architecture. There are modules made specifically for standard racks, which allow for low-cost system building blocks. The designers should work with their ultracapacitor provided to fill out application worksheets and have discussions about the system needs. Lifetimes can be achieved for 20+ years of maintenance-free operation, but perhaps the system only needs to work for 7 years? If that is the case, the capacitor manufacturer should be able to use their life modeling, based on the system sizing inputs, and reduce size by increasing the derated of the volts per cell and achieve the proper life at a reasonable cost.
What is usually the biggest challenge engineers face when first using supercapacitors?
ED: I will take this question and the one that precedes it as one and the same as I answer here. The biggest challenge or misunderstanding in my opinion would be properly sizing a supercapacitor solution for end of life due to its failure mechanisms. Supercapacitors are much more complex than conventional board-level capacitors, so in turn understanding how failure is influenced by voltage and temperature is never a “one size fits all” equation. All we can do is rely on decades of industry experience and internal captured test data to better characterize expected lifetimes. Taking a different view on this as well, specifically as it relates to design engineers first using supercapacitors, certainly be conscious of the amount of energy and current potential they are using or “playing around with.” Especially when designing for higher voltage applications where multiple cells are necessary to attain that voltage level with supercapacitors, the current potential can be dangerous or even dead.
JL: How to size supercapacitors and accounting for the voltage drop. Many engineers are used to working with batteries or a more constant voltage source. Understanding how the voltage drops as it powers the load and the effect of the current on this.
CH: Typically one of the biggest challenges engineers face when first using supercapacitors is to not fully understand their system needs. By really understanding the energy and power needs, the system design can be altered to ensure they are using the right size products. Supercapacitors can be used to deliver very high currents for a short period (starting an engine or lifting a container), or they can be used to deliver a small amount of current for a longer period, acting as more of a battery (powering LEDs for example). Knowing how much power is usable, and looking at the energy needs will allow for flexibility in choosing chemistries of energy storage.
What should designers do to maximize supercapacitor performance and lifetimes?
JL: Lower the operating voltage per cell is the main “knob” designers have in order to maximize lifetime. The typical method is to put more cells in series, but this increases the ESR of the system. This can be overcome by adding capacitance as this typically correlates with lower ESR.
ED: Supercapacitor lifetime is directly tied to applied voltage and temperature. Make sure to adhere to recommended voltage deratings for high-temperature performance or long lifetime expectancy, and when in doubt seek advice from the manufacturer for their guidance & resources.
CH: Two things really affect the life of an ultracapacitor; heat and voltage. Proper module design will reduce the thermal and electrical effects, by employing thermally conductive materials to pull heat from the cell, or balancing and limiting the voltage to the cells. High rate duty cycling will often drive the internal temperature higher than ambient, and this temperature rise needs to be considered when sizing the system. Voltage is also typically de-rated on a per-cell basis to achieve a very long life. Lower temperatures (down to -40C) do not affect life but can cause an increase of up to 2x for the ESR. These should be considered when looking at life.
The original article was posted by EE World Online:
