Skeleton Technologies White Paper
Skeleton Technologies White Paper
Skeleton Technologies White Paper
equivalent
series resistance
Equivalent series resistance is the underdog of ultracapacitor specifications. Limelight is often
spoken for by Farads and operating voltage and the resulting energy density. However, when
deployed for high pulse power loads – the strongpoint of ultracapacitors – ESR has the single
strongest impact towards system efficiency and working temperature. These, in turn, determine
the percentage of energy that can be retrieved from the ultracapacitor in “real life” and the
operating temperature, which is the sole strongest predictor of ultracapacitor lifetime.
The Ragone plot shown below depict the dependence of the usable energy content at different
power levels. A 10 second discharge equates to a discharge power of 1.5 kW per kg of weight
(considering the ultracapacitor weight of approx. 0.5kg, the capacitor is discharge at approx 0.75
kW). At this point the ultracapacitor has energy content of 4 Wh/kg.
At five times higher power demand – the ultracapacitor is discharged within 2 seconds – heat
losses are such that power drops from ideal 7.5 kW to 5 kW and energy content is down to 2.7 Wh/
kg of usable energy. The area between the two lines equals the heat and contact resistance losses.
Please see technical annex for the mathematical representation of this and other ESR related ultra-
capacitor performance implications.
The Ragone plot below shows the difference between similarly sized capacitors with a 3-fold differ-
ence in internal resistance. At modest power levels - 10 second discharge - the product with higher
ESR levels manages 3.3 Wh/kg (-30%) while Skeleton product manages over 4 Wh/kg (-11.5%).
In conclusion: the competing product tested is essentially unusable at power levels over 1 kW/
kg and Skeleton product’s efficiency still outperforms the competing product at five times higher
power demands.
Contact and total resistance - manufacturer quoted ESR values are not the whole story
As a de-facto standard manufacturers quote ESR values obtained in a 10 ms voltage drop test.
These values are relevant only in the framework of application times below 1 second. For real-life
comparisons a five-second voltage drop test or a 0.1 Hz test with alternating current yield an ESR
value more relevant for most real-life applications.
Contact resistance is the resistance seen in the current path, including capacitor terminals, the
aluminium foil current collector, the microscopic boundary gap between electrode carbon layer and
the foil current collector and, finally, the carbon structure of the electrodes. The contact resistance
makes up approximately 50% of the total ESR. Total resistance value adds “pore resistance”, the
effect of which gains in amplitude in the timeframe between 0.5 and 5 second application time,
with the ESR elevation gains dropping off after that. The source of pore resistance is the result of
atomic forces between electrolyte ions and the carbon structure of the electrodes, which hold them
in place (labelled “ensnarement” in the scientific literature).
A comparison of two products with closely matched weight. Measurements performed with Electrochemical
Impedance Spectrometry, from 100 Hz to 0.1 Hz. Contact resistance is measured at 10ms(100 Hz) and total
resistance at 10 seconds (0.1 Hz). The plot is adapted from the Nyquist plot by substituting the derivative
Y-axis with reciprocal of AC frequency (1/v) in seconds.
Lithium Ion capacitors are often praised for their energy density, but are in practice let down
by approximately 10-fold lower life-expectancy and lower power density. Hybrid capacitors use
one “traditional” ultracapacitor electrode on one side and a lithium-ion battery electrode on the
other. This combination of the two technologies creates a compromise between a battery and a
supercapacitor.
The ultracapacitor rule of thumb - by which total resistance is approximately double that of contact
resistance – does not hold in a hybrid system. As seen in the test example below, LiCap exhibits a
4-fold increase in resistance values already at 1sec. The internal test results with a commercially
available 7500F hybrid capacitor show how 10 ms resistance value of 0.34 mOhm is elevated to a
1s resistance value seen in testing was 1.36 mOhm.
Internal resistance has an inverse relationship to temperature – higher temperatures lead to lower
resistance. The temperature effect can be mainly attributed to the electrolyte, more specifically, the
viscosity of the electrolyte solvent, in this case acetonitrile (AN). Acetonitrile freezes below -40OC
and boils at 82OC with the viscosity highest at the lowest temperature.
Case Study
The task is to design a system with the ability to deliver 275 kW of power for 10 seconds
at higher than 95% efficiency. Nominal voltage range is approx. 850 – 425 V, dictated
by minimum voltage requirements are loosely based on a pair of Yasa 750 motors
(http://www.yasamotors.com/wp-content/uploads/2015/09/YASA-750-Product-Sheet.pdf).
No converters are to be used.
The number of cells connected in series is an easy requirement to calculate: with a nominal poten-
tial of 2.85 V at least 299 cells are needed. The number of parallel strings is an interplay of heat-
loss vs cell capacity. Among Skeleton products the two capacitor models closest to the design
requirements are 300 pcs of 3200 F cells connected in a single string or 600 pcs of 1800 F cells
connected in two parallel strings.
The 1800 F system with 600 cells has superior characteristics: nearly 3% higher round-trip
efficiency and also more capacity, which combined add around 18% to the duration of the 275 kW
output.
While a better system, the one with1800 F cells is over-engineered for the purpose as additional
efficiency and capacity beyond 95% and 10 seconds of acceleration assistance are not seen as
design drivers. Added weigh, cost and complexity more than outweigh the marginal benefits to be
had from the overengineered system.
A simulation of the 300 cell system (2 charge and discharge cycles at constant power) is shown
below to convey a more concise overview of where the ESR related losses occur.
If the ESR of the capacitors doubled (both contact and total resistance), the heat losses would be
higher by 2.8% points and the maximum duration of the acceleration assistance would drop to 9.75
seconds. Below we are comparing the discharge voltage curves of three systems: ideal, nominal
and 2x ESR.
Starting at the same level of charge (voltage at 855 V) the ideal system would supply 10.63 seconds
of 275 kW output while the 1xESR and 2xESR systems deliver 10.16 seconds and 9.75 seconds
respectively. Changes in the slope of the voltage curve for the ideal system stem purely from the
squared relationship between voltage and energy content while real life systems are affected also
by ESR and therefore see a faster drop-off of the voltage curve.
To summarize: The comparative efficiency of each system can be derived from the root mean square
(RMS) of the area below the voltage curve. The ESR losses can be observed as 1) the initial voltage
drop, 2) increasing angle of descent of the voltage curve at lower voltages.
TECHNICAL ANNEX
ESR definition
The equivalent series resistance ESR, also known as the internal resistance, is a physical value
representing the Ohmic losses in a simplified equivalent diagram comprising a resistor and an ideal
capacitor in a serial circuit (Fig. 1).
Fig. 1
Series equivalent diagram
of an ultracapacitor
The ESR is caused by non-ideal properties of the ultracapacitor materials, i.e. pore and contact
resistances, ionic inertia, electrolyte conductivity, impurities etc. Ultracapacitor manufacturers use
different methods to express the ESR values, Skeleton Technologies has chosen for engineering
calculations the 5 s current cutoff method as it represents most accurately Ohmic losses and their
effects.
Let us use symbols U, I, Ri for the voltage, current and ESR, respectively.
1. According to the Ohm’s law, the current flowing through a resistance causes a voltage drop
∆U=I • Ri
2. While an ultracapacitor is charged or discharged, some of the energy is lost with the dissipation
power ∆P=I² • Ri i.e. current squared.
3. The dissipated power causes the ultracapacitors to warm up with the final temperature rise of
∆Tf=∆P • Rth = I² • Ri • Rth , where Rth is the thermal resistance.
The ESR becomes evident in multiple values describing the ultracapacitor’s performance:
1. matched impedance power
2. usable power
3. power at 95 % efficiency
SCA3200 Generic
Rated voltage 2.85 V 2.85 V
Rated capacitance 3200 V 3400 V
ESR 0.13 mΩ 0.22 mΩ
Thermal resistance, case-ambient 3.2 °C/W 3.2 °C/W
Based on the equation we can plot constant current application times for
both (Fig. 2). As seen, the SCA3200 outperforms a generic 3400 F ultracapacitor at currents
exceeding 840 A, where lower ESR compensates higher capacitance.
Fig. 2
Constant current application times for SCA3200 and generic 3400 F ultracapacitors
Thermal benefits
From the equation ∆Tf=I2 • Ri • Rth it can be seen, that the final temperature rise at constant
current operation is proportional to the ESR. In practice it means that less cooling efforts are
needed or the ultracapacitors can be deployed in warmer environments. Moreover, keeping the
temperature under control helps to avoid premature ageing. An example of constant current
operation temperature rise is given in Fig. 3 below.
Fig. 3
Temperature rise at constant current operation, caused by ESR
Conclusions