# Super- and Ultracapacitors: Thousands of Farads Thanks to Double Layer Technology

In radio technology capacitors sometimes only have a few pF. Electrolytic capacitors reach μF, perhaps even mF. But there are capacitors that deliver thousands of Farads. How do they work?

Supercapacitors first showed up in consumer electronic applications as “gold caps” – capacitors that could only withstand a low voltage, but had sufficient capacity to replace backup batteries for RAM memory or real-time clock chips in computers.

In LED rear lights for bicycles, they caused astonishment because they continued to shine for minutes when the bike had to stop at a traffic light, without a battery being installed. Initially, however, the peak power capability of these components was low and the internal resistance (equivalent serial resistance – ESR) relatively high.

The technology has now been further developed and now we have supercapacitors, also called uItracapacitors, in mass production with capacities of several thousand Farads. In terms of storage capacity, they can compete with small accumulators. However, the physics of supercapacitors are different, and thus they behave electrically differently from accumulators.

## Physics determine capacity

First of all, supercapacitors are really capacitors: Their capacitance is determined by two conductive surfaces facing each other. The larger the surface, the smaller their distance and the higher the dielectric constant of the dielectric between them, the higher the capacitance:

$C=&space;\varepsilon&space;\left&space;(&space;\frac{A}{d}&space;\right&space;)$

C = capacitance, A = area, d = distance, ε = dielectric constant.

Consequently, an air or vacuum capacitor has a small capacitance value because d is quite high here and ε and A are low. But it offers a high dielectric strength.

### Film Capacitor

A film capacitor has a significantly higher capacitance, as more surface area and a higher dielectric constant are available and the film allows the distance to be reduced while maintaining a high dielectric strength. Depending on the dielectric constant of the material, ceramic capacitors sometimes offer even higher capacitances with possible restrictions in dielectric strength and capacitance stability.

### Electrolytic Capacitor

Electrolytic capacitors have an even higher capacitance because no mechanically manufactured dielectric is used here, but a thin, chemically produced oxide layer. If the base material is very rough, the area and thus the capacitance increases further. The dielectric strength is lower and the capacitor has a specific polarity – if handled incorrectly (wrong polarity, overvoltage, overcurrent, over temperature) it can fail prematurely. Ultracapacitors are usually of symmetrical design.

In principle, the polarity of the applied voltage should not matter. However, a polarity is defined during production when the capacitor is charged for the first time and the user should stay with this polarity later on. If the polarity of an ultracapacitor is reversed, there is no risk of a total failure with explosion as with an electrolytic capacitor, but a permanently reduced service life and performance is to be expected.

## Helmholtz double layer

Supercapacitors are double layer capacitors whose underlying principle, the Helmholtz double layers, have been known for over 130 years. They are only a few molecular layers wide in the nanometer range, which results in a further capacity increase of up to a factor of 10,000 compared to the electrolytic capacitor.

For the same reason, however, they have a low maximum voltage allowed, which is around 3 V for individual cells in today’s technology. For higher voltages, the cells can be connected in series, as with batteries; for more than two cells that reach 5 to 5.5 V permissible operating voltage, measures for even voltage distribution are usually required.

However, electrochemical processes such as those in batteries and accumulators, in which the electrode material changes structurally and thus wears out, play only a minor role in double layer capacitors. In the designs of supercapacitors commonly used today, they only contribute a few percent of the capacitance. There are supercapacitor technologies in which this proportion predominates or is equivalent (pseudo and hybrid capacitors). They have lower charge/discharge cycle numbers, but still far higher than accumulators. However, these variants are not currently of great importance and are not the subject of this article.

## Physics and Chemistry

However, ion shifts and chemical formations of these ions do play a role in the double layer, which is why supercapacitors are also referred to as electrochemical capacitors. This is the only way to explain the enormously high field strengths of up to 5000 kV/mm in the double layer – a normal dielectric would fail.

Charge and discharge currents of double layer capacitors can be very high, deep discharge is no problem, 100,000 charge and discharge cycles and more are possible with a lifetime of more than 20 years. The capacities are already around 1/10 of those of accumulators. Thus supercapacitors are much more powerful than accumulators in cyclic operation: they can even drive racing cars or public transport vehicles such as electric buses, which can be recharged via contacts at a short stop at the bus stop.

For example, the Fraunhofer Institute for Material and Beam Technology IWS in Dresden has had hybrid buses manufactured that can travel up to 2 km purely electrically to the next charging station after 15 seconds of charging at the bus stop – a diesel engine will only be activated for longer journeys.

## Old technology

The principles of electrolytic capacitors and supercapacitors were discovered at comparable times – in 1875 by Eugène Adrien Ducretet (electrolytic capacitor) and already in 1853 by Hermann von Helmholtz (supercapacitor), with Helmholtz defining the double layer effect in 1879.

However, while the aluminium electrolytic capacitor was used industrially from 1892 and manufactured using the technology known today from 1931 onwards, the supercapacitor was forgotten for many years: It was not until 1957 that the first patents on “capacitors with porous carbon electrodes” were granted. The mechanism of action of the double layer was unknown to the patent applicant.

Figure 1: Structure of the dielectric of an ultracapacitor

In 1962, a Standard Oil canoe drove over a lake in Ohio with electricity from a car battery sized supercapacitor for a demonstration of ten minutes, but Standard Oil saw no market opportunity and sold the patents to NEC. In addition, even the developers of the supercapacitors were initially unaware of the difference in functionality from electrolytic capacitors; Standard Oil had still classified them as electrolytic capacitors.

In 1982, the Pinnacle Research Institute developed capacities over 1000 F with low internal resistances for military purposes as the “PRI Ultracapacitor”. From 1992, Maxwell continued this development and its marketing. By the end of the 1990s, other innovative manufacturers such as Cooper Bussmann/Powerstor (now Eaton) joined the efforts.

## More robust than accumulators

The temperature resistance of supercapacitors is higher than that of accumulators. Also higher powers are possible at lower temperatures. However, certain limit values must not be exceeded, otherwise the electrolyte will evaporate. A typical end of life is defined as a capacitance loss of 20...30% or an increase of the internal resistance to twice the value – a sudden total failure of a supercapacitor is rare with correct treatment.

The large capacity of the supercapacitors is achieved not only by the extremely thin insulation layer but also by the fact that supercapacitors use carbon electrodes. These are very porous and rough – mostly activated carbon is used. More than 3000 square meters can be achieved with one gram of carbon powder. Graphene and carbon nanotubes have already been tested, but are still too expensive for mass products. At least with graphene this is supposed to change in the next years.

Figure 2: Temperature dependence of ultracapacitors

## Not suitable as filter capacitor

Supercapacitors are not filter elements like normal capacitors and electrolytic capacitors – they are primarily energy storage devices. The internal resistance at higher frequencies is unsuitable for filtering, especially with clocked power supplies and converters – even at 10 Hz only a fraction of the capacity of the supercapacitor is effective because the ions on the double layer do not move fast enough.

In addition, the internal resistance is generally higher than with electrolytic capacitors, which is why its use as a filter and smoothing capacitor is unsatisfactory and can lead to overheating and failure.

In contrast, ultracaps can easily bridge a few seconds of power failure in UPSs without the need for constant maintenance and inspection like UPSs with accumulators. They are even well suited as a starting aid in automobiles, as their performance does not radically deteriorate at low temperatures like that of conventional starter batteries. Just the price is not yet competitive for this application.

## Battery and supercapacitor in a team

Supercapacitors can also be used as buffers if batteries are available and necessary for the power supply, but the device to be supplied draws very pulsed current like an optical smoke detector: Here, the ESR of the batteries becomes too high, especially as the discharge progresses; with a supercapacitor solution connected in parallel, the batteries can be used much longer before a current pulse causes undervoltage, and lithium-ion batteries suitable for long-term use can be used instead of the alkaline-manganese cells suitable for higher currents.

However, problems may occur in certain situations: Supercapacitors are not designed for months of operation without recharging because of the leakage current. Anyone who combines a lithium battery designed for 10 years of operation with a supercapacitor to provide higher peak currents could be confronted with an unexpectedly early discharge of the battery.

## Electrical characteristics

Supercapacitors are pure secondary energy storage devices. Although the self-discharge of today’s cells is low, it is not suitable for the independent supply of devices for months. However, self-discharge is at least low enough to bridge days and sometimes weeks.

For safety reasons, the components are also not delivered charged like accumulators and are normally not pluggable - the peak currents in the event of incorrect operation (short circuit) would be very high and could cause serious damage. Unlike batteries or accumulators, supercapacitors do not supply a chemically determined voltage that is constant over a longer period of time and only drops rapidly at the end of discharge, but, like any capacitor, a voltage that sinks linearly with a constant current draw.

The output voltage of a supercapacitor power supply can be kept constant via voltage regulators; however, when the capacitor voltage drops to ½ of the output value, ¾ of the stored charge has already been discharged. It is therefore not worthwhile to discharge even further with wide range transformers, although deep discharge is not a problem for supercapacitors in principle. On the other hand, there is no sudden failure of the energy storage when a final discharge voltage is reached.

## Higher voltage: supercapacitor arrays

The end of the lifetime of a supercapacitor is usually defined with a capacitance loss of 70...80% and/or an increase of the ESR to 200...300%. A circuit that is to provide a certain supply voltage and capacity with a supercapacitor array must be dimensioned accordingly with a reserve: If a discharge is really planned in seconds and not in hours, the internal resistance will cause a voltage drop, which must be compensated by a correspondingly higher charging voltage and thus more supercapacitors connected in series. This in turn reduces the capacitance due to the series connection.

A correct dimensioning of the individual capacitor cells is therefore calculated with these “end-of-life” parameters and not with the parameters of a brand-new supercapacitor. This ensures long-term operation of the circuit within its setpoints.

### The voltage distribution must be taken into account:

- First, capacitance tolerances in serial circuits will lead to a lower voltage drop on the larger ultracapacitor, which means, the smaller capacitor will have a higher voltage drop and thus may be overloaded. Without further precautions, this reduces its lifetime and capacity, which further increases the overload and ultimately causes the array to significantly lose performance.

- Second, the static leakage currents of ultracapacitors are low, but just as different. This can also result in an imbalance in the voltage distribution.

### Passive and Active Balancing Solutions

Passive or active balancing solutions can be used to remedy this situation and to balance the voltage distribution.

Passive solutions are resistors connected in parallel to the ultracapacitors, which must be significantly lower than the leakage resistances of the capacitors. This solution is inexpensive and reliable, but increases the leakage currents. Active solutions only bypass the ultracapacitor when its permissible operating voltage is reached.

Ultracapacitor modules from Eaton and Maxwell already contain a suitably dimensioned balancing solution. The manufacturers also

Figure 3: Capacitance curve of an older ultracapacitor model at 65°C ambient temperature

offers suitable ready-to-install modules for customers who want to assemble their own arrays. The creepage distances relevant for higher array voltages must be taken into account, especially for in-house designs. Modules assembled by the manufacturers are checked to comply with the necessary safety distances.

If a single prematurely aged capacitor in a module is to be replaced, a new ultracapacitor can upset the balance. In such cases it may be advisable to install a pre-aged capacitor.

## Negative influences

A rule of thumb at Eaton is that the service life of a supercapacitor increases by a factor of 2.2 if

- the operating voltage is reduced by 0.2 V

- the ambient temperature drops by 10°C

If particularly long running times are required, operation at reduced voltage and not too high temperatures is helpful.

Of course, the formulas can be adapted as required. If you define the end of the lifetime at 50% capacity and 200% ESR, the circuit will work longer, but needs larger supercapacitors to work as desired at the end of the lifetime.

According to DIN EN 62391-2, the limit values are 70% capacitance and 400% ESR (fourfold ESR), but the latter is problematic in practice, since a quadruple ESR means a load capacity reduced to ¼ or a fourfold heating of the supercapacitor due to the current load during operation. Such large deviations are not tolerable in high-performance applications. For this reason, the DIN limits are rarely used.

## Round or square?

In order to make full use of a given volume, cuboid capacitors appear to be more advantageous at first. However, like with film capacitors round windings are more stable in operation and cheaper to manufacture.With arrays, cooling is also easier here, while cuboid capacitors sitting on top of each other without gaps are difficult to cool. Although square designs are available for special applications, round designs are usually the more sensible choice.

Figure 4: Ultra-capacitor module 48 V / 165 F

## Dramatic total failure rare

Ultracapacitors are relatively temperature-in-dependent, as long as the permissible limits are not exceeded, especially upwards. Their capacitance remains practically constant from -40 to 65°C, only the internal resistance increases with decreasing temperatures, but this is comparatively moderate. Rechargeable and non-rechargeable batteries, on the other hand, become unusable in the cold.

The operating temperature range of Maxwell ultracapacitors is between -40 and 65°C. Above 65°C, as well as with overvoltage, lifetime and characteristic data suffer very quickly; exceeding this temperature should therefore be avoided. High temperature and voltage are particularly critical at the same time – if a very long service life is important, these limit values should not be exhausted for a long time. Eaton allows up to 85°C depending on the electrolyte, but exceeding these limits is also critical.

Humidity is not critical in hermetically sealed cells – especially the larger varieties – as long as no water is deposited on the capacitors and thus causes leakage currents. The ultracapacitors themselves do not react to moisture if they are originally packed or installed. The air pressure is just as uncritical, which is important for use in aviation.

## Store short-circuited

In contrast to batteries, the deep discharge of an ultracapacitor is absolutely uncritical. In fact, the service life is virtually unlimited when stored in a discharged state and transport is possible with discharged and bridged connections to eliminate the risk of discharge with high currents due to short circuits during transport or high, dangerous touch voltages in modules. Modules must be bridged during transport. If the bridging is removed, a maximum residual voltage of approx. 0.2 V per cell can build up after some time.

### Leakage currents in ultracapacitors

By the way, leakage currents in ultracapacitors are for the most part not caused by a defective dielectric. Rather, the huge surface of the carbon electrodes leads to a time constant of about one second for charging and discharging which is unusually high for a capacitor – for an accumulator this would be a very low time constant! It may take up to 72 hours until the capacitor is completely charged or discharged.

### Last 0.5% of the surface

The last 0.5% of the surface is particularly difficult to reach due to the porous material and the geometry. This leads to the fact that it can take hours to days until these parts are charged – and this shows up electrically as a supposed “leak”. Conversely, these 0.5% are “to blame” for the fact that after a complete discharge of an ultracapacitor after some time the already mentioned residual voltage of up to 0.2 V per cell can build up, if the discharge circuit is removed. In contrast to accumulators, no chemical processes are involved here, but regions are noticeable that were not completely discharged before.

## Electrode production

Because this market is so attractive, many suppliers with sometimes mixed quality cavort here. Even a do-it-yourself construction of a supercapacitor from household utensils is described on Youtube (https://youtu.be/gTt_YBzJ_Dk) – but with extremely aggressive pipe cleaner (sodium or potassium hydroxide or potash lye) as electrolyte, with just 1.2 V maximum operating voltage, without overpressure protection and then stored in the bedroom above the bed to supply the alarm clock. Amusing, if it wasn’t almost macabre. But even the industrial ultracapacitors are sometimes overwhelmed by actual applications and fail prematurely, especially in the case of vibrations in vehicles.

### Carbon Electrodes

Qualitative differences in ultracapacitors are often due to the carbon electrodes. Their area determines the capacity, where almost all industrial manufacturers reach comparable values in a given volume. However, there are different manufacturing processes that affect stability. Wet processes, in which the activated carbon is first sprayed in solution and then develops its porous structure during drying, sometimes lead to more unstable electrodes. In mobile applications, especially in the vicinity of vibrating aggregates such as engines or assembly locations in the vehicle shaken by the journey, these can quickly “crumble” and thus fail.

Maxwell’s drying process is intended to be advantageous here for particularly large cells; in addition, the basic structure is particularly robust. Maxwell calls this “Durablue technology”: It is particularly robust against both individual shocks and continuous oscillations and vibrations.

## Ultracapacitors Development

### Supporting batteries and accumulators

So much for the current state of the art, which is far from being the end of development. However, ultracapacitors have now reached technical maturity and are enough tried and tested that they can be used reliably even in demanding, harsh environments, supporting batteries and accumulators or even replacing them completely.

### Lower capacity compared to batteries

The lower capacity compared to batteries is often no problem at all, because the available discharge current determines the application and not the total capacity that can be removed. Cordless tools, for example, are regularly stored and not held continuously in the hand for hours. In these phases they can already be completely recharged, because with ultracapacitors this does not take hours but seconds.

### Not stressed by high discharge currents

In addition, ultracapacitors are not stressed by high discharge currents. They can even be used in applications in which NiCd accumulators, which are now undesirable for environmental reasons, were still being used because NiMh and lithium accumulators deliver lower peak currents.

## Possible applications

A popular application are decentralized power supplies with high pulse load. Ultracapacitors are used for starting aids for diesel engines, whether in trucks, construction machinery or emergency power generators, but also directly in emergency power supplies, which can take over without delay in the event of a power failure and still function safely after decades without battery changes.

This is eminently important even for supposedly simple applications such as the lighting of emergency exits – so far such devices are often supplied by central battery systems in the basement of the building, which fail after a few years in an emergency if they are not constantly monitored and replaced if necessary.

### Peak current interception and energy recovery

Other applications include peak current interception and energy recovery in elevators and hybrid vehicles. In container ports and warehouses, ultracapacitors can supply port cranes, forklifts and other vehicles. Here they are far superior to accumulators because charging times are not measured in hours but in seconds and minutes and high pulse currents of up to thousands of amps are available.

In fact, only the heating of the ultracapacitors limits a possible repetitive pulse load – a short-circuit can become critical for the connected devices and the connecting cables, but not for the capacitors themselves, nor for the deep discharge. There is no risk of fire, as with lithium batteries, due to heat, overload or mechanical damage.

### High currents required at short notice

Also innovative are units that can quickly turn wind turbines out of the wind in gusty conditions or help to open and close doors in aircraft: The high currents required for this at short notice would cause unnecessary weight and costs with conventional wiring without intermediate storage – with ultracapacitors, power supply is much simpler.

In addition, the door can be opened in an emergency without an on-board power supply. In automobiles, they can also support window regulators, door locks, power steering, belt tensioners and other electric motor-driven units.

### Literature

[1] Helmholtz double layer, Wikipedia: Double Layer

[2] Super- and ultracapacitors at HY-LINE Power Components: HY-LINE Supercaps