Isothermal calorimeter for measuring heat generation rate in individual supercapacitor electrodes

Abstract

Heat generation in electric double layer capacitors (EDLCs) may lead to temperature rise and reduce their lifetime and performance. This study aims to measure the heat generation rate in individual carbon electrode of EDLCs under various charging conditions. First, the design, fabrication, and validation of an isothermal calorimeter are presented. The calorimeter consisted of two thermoelectric heat flux sensors connected to a data acquisition system, two identical and cold plates fed with a circulating coolant, and an electrochemical test section connected to a potentiostat/galvanostat system. The EDLC cells consisted of two identical activated carbon electrodes and a separator immersed in an electrolyte. Measurements were performed on three cells under galvanostatic cycling for different current density, electrolyte composition, and polarity. The measured time-averaged irreversible heat generation rate was in excellent agreement with predictions for Joule heating. The reversible heat generation rate in the positive electrode was exothermic during charging and endothermic during discharging. By contrast, the negative electrode featured both exothermic and endothermic heat generation during both charging and discharging. The results of this study can be used to validate existing thermal models, to develop thermal management strategies, and to gain insight into physicochemical phenomena taking place during operation.

Highlights

  • An isothermal calorimeter was designed, manufactured, and carefully validated
  • The device can measure heat generation rate at each electrode of supercapacitors
  • Its capabilities were illustrated with EDLC electrodes and various electrolytes
  • Irreversible heat generation rate was due to Joule heating
  • Reversible heat generation rate was significantly lower at the negative electrodes

Materials and Methods

1. Isothermal calorimeter

An isothermal calorimeter was designed, fabricated, and validated to measure instantaneous heat generation rate in electrical energy storage systems. The experimental apparatus consisted of (i) two thermoelectric heat flux sensors (HFS) connected to (ii) a data acquisition (DAQ) system (34972A LXI, Keysight Technology), (iii) two identical instrumented cold plates fed with a circulating coolant (Dynalene HC-50, Dynalene Inc.) from (iv) a temperature-controlled chiller (Polystat, Cole-Parmer), (v) two flow meters (FLR-1012, Omega), and (vi) an electrochemical test section containing a two-electrode cell immersed in an electrolyte and connected to (vii) a potentiostat/galvanostat (SP 150, Bio-Logic Science Instruments). A vertical clamp was used to hold the electrochemical test section and the cold plates together and to ensure good thermal contacts among them. Finally, the entire calorimeter and the cold plates were wrapped in 13 mm thick thermal insulation (Ceramic fiber, Morgan Thermal Ceramics), with thermal conductivity of 0.07 W/m.K to minimize heat losses to the surrounding.

2. EDLC devices

The EDLC devices tested consisted of two identical activated carbon electrodes separated by a 350 μm glass fiber separator (GF85 filter, Advantec MFS Inc.). Different electrolytes and associated potential windows were tested to assess the effect of ions size and valency and their asymmetry on the performance and thermal behavior of EDLC devices, as summarized in Table 1. Devices 1 and 3 used organic electrolytes made of 1M of lithium hexafluorophosphate in ethylene carbonate:dimethyl carbonate (EC:DMC) with 1:1 weight ratio and 1M of tetrabutylammonium tetrafluoroborate (TBATFB) in acetonitrile solvent, respectively. By contrast, Device 2 used an aqueous electrolyte made of 1M of citric acid in deionized (DI) water. Citric acid was chosen because it does not corrode the stainless steel current collector, unlike most aqueous electrolytes. Finally, each EDLC device was assembled, installed, and sealed in the electrochemical test section inside the glove box to avoid any contact with air. Table 1: Electrolyte composition and galvanostatic operating conditions for the three carbon-based EDLC cells studied.

3. Analysis

Q̇i(t) (in mW) in electrode "i" in contact with heat flux sensor "j" can be expressed as,

where Ai is the footprint area of the electrode. Here, the subscript "i" refers to the positive "+" or negative "-" electrode. The time-averaged heat generation rate at electrode "i" subjected to a galvanostatic cycle of period tcd was estimated by integrating the instantaneous heat generation rate Q̇i(t) over one period, i.e.,

where n is the cycle number, large enough to have reached oscillatory steady state. In addition, the instantaneous reversible heat generation rate Q̇rev,i(t) at each electrode can be evaluated by subtracting the time-averaged heat generation rate from the instantaneous heat generation rate Q̇i(t), i.e.,

In order to effectively compare the reversible heat generation rate at each electrode, the instantaneous reversible heat generation rate Q̇rev,i(t) was averaged over a galvanostatic charging step of duration tc to yield,

Note that, by definition, time-averaging of the reversible heat generation rate Q̇rev,i(t) at electrode "i" over an entire galvanostatic cycle of period tcd yields . Finally, the total instantaneous, time-averaged, reversible, and time-averaged of reversible heat generation rates in the entire cell can be expressed as ,

.

Results and Discussion

and were positive and proportional to the square of the current I2 in both positive and negative electrodes with a coefficient of proportionality corresponding to their respective resistances R+ and R-R+ and R-. These results confirm that the two electrodes constituting the device were nearly identical. Figures 1(d) to 1(f) also indicate that the measured total irreversible heat generation rate was in excellent agreement with predictions for the heat generation rate due to Joule heating in the entire device given by,

.

Figure 1: Heat generation rates Q̇+(t) at the positive electrode (blue), Q̇-(t) at the negative electrode (red), and Q̇T(t) in the entire cell (black) as functions of the dimensionless time for current I=6 mA for (a) Device 1, (b) Device 2, and (c) Device 3 for five galvanostatic cycles. Time-averaged heat generation rates under galvanostatic cycling as functions of I2 for current I ranging between 2 and 6 mA for (d) Device 1, (e) Device 2, and (f) Device 3 (Table 1).

Figure 2 indicates that the reversible heat generation rate Q̇rev,i(t) was significantly different at the positive and negative electrodes and was independent of cell polarity. At the positive electrode, Q̇rev,+(t) was systematically exothermic during charging and endothermic during discharging. By contrast, the reversible heat generation rate Q̇rev,-(t) at the negative electrode was both exothermic and endothermic during either charging or discharging. In addition, at the positive electrode was proportional to the current I while at the negative electrode was systematically lower than at the positive electrode and independent of the current I.

Figure 2: Reversible heat generation rates (a) in the entire cell, (b) at the positive electrode, and (c) at the negative electrode as functions of the dimensionless time for two galvanostatic cycles under current I=6 mA for Devices 1, 2, and 3. Time-averaged reversible heat generation rates during a charging step (d) in the entire cell, (e) at the positive electrode, and (f) at the negative electrode as functions of current I ranging between 2 and 6 mA for Devices 1, 2, and 3.

Conclusion

The present study designed, assembled, and carefully validated an isothermal calorimeter to investigate the temporal evolution of the heat generation rate in EDLC devices. This calorimeter was able to measure separately the instantaneous heat generation rates at each electrode of a two-electrode device with resolution as low as 10 μW and uncertainty of 3%. Heat generation measurements were demonstrated on three EDLC devices consisting of two identical activated carbon electrodes and different organic and aqueous electrolytes under galvanostatic cycling. First, the three devices were characterized using (i) cyclic voltammetry to obtain the gravimetric capacitance and (ii) galvanostatic cycling under constant current I to obtain the total internal resistance. Second, the measured time-averaged irreversible heat generation rates at each electrode were similar and proportional to I2. The total irreversible heat generation rates measured in the entire EDLC cell were in excellent agreement with predictions for Joule heating. Third, the reversible heat generation rate Q̇rev,i(t) was significantly different at the positive and negative electrodes and was independent of cell polarity. At the positive electrode, Q̇rev,+(t) was systematically exothermic during charging and endothermic during discharging. By contrast, the reversible heat generation rate Q̇rev,-(t) at the negative electrode was both exothermic and endothermic during either charging or discharging. In addition, at the positive electrode was proportional to the current I while at the negative electrode was systematically lower than at the positive electrode and independent of the current I. The difference in thermal behavior at the positive and negative electrodes may be due to parasitic reversible redox reactions, solvation/desolvation, and/or differences in ion size and transport properties in the electrolytes. Unfortunately, understanding the causes of such differences falls outside the scope of the present study but will be the subject of future research. Finally, the present results can be used to developed thermal management strategies for EDLCs. The isothermal calorimeter assembled could be used for other electrical energy storage systems and to gain insight into physicochemical phenomena taking place during charging and discharging.

Publication

O. Munteshari, J. Lau, A. Krishnan, B. Dunn, and L. Pilon, 2018. Isothermal Calorimeter for Measurements of Time-Dependent Heat Generation Rate in Individual Supercapacitor Electrodes, Journal of Power Sources, Vol. 374, pp. 257-268. doi:10.1016/j.jpowsour.2017.11.012 pdf

Related publications

A. Likitchatchawankum, R.H. DeBlock, G. Whang, M. Frajnkovič, O. Munteshari, B.S. Dunn, L. Pilon, 2021. Heat Generation in Electric Double Layer Capacitors with Neat and Diluted Ionic Liquid Electrolytes Under Large Potential Window Between 5 and 80 ºC, Journal of Power Sources, Vol. 488, 229368. doi:10.1016/j.jpowsour.2020.229368 pdf

A. Likitchatchawankum, G. Whang, J. Lau, O. Munteshari, B. S. Dunn, and L. Pilon, 2020. Effect of Temperature on Irreversible and Reversible Heat Generation Rates in Ionic Liquid-Based Electric Double Layer Capacitors, Electrochimica Acta, Vol. 338, 135802. doi:10.1016/j.electacta.2020.135802 pdf

O. Munteshari, A. Borenstein, R. H. DeBlock, J. Lau, Y. Zhou, A. Likitchatchawankun, R. Kaner, B.S. Dunn, and L. Pilon, 2020. In Operando Calorimetric Measurements for Activated Carbon Electrodes in Ionic Liquid Electrolytes Under Large Potential Windows, ChemSusChem, Vol. 13, pp. 1-15. doi:10.1002/cssc.201903011 pdf

O. Munteshari, Y. Zhou, B.-G. Mei, and L. Pilon, 2019. Theoretical Validation of the Step Potential Electrochemical Spectroscopy (SPECS) Technique and Multiple Potential Step Chronoamperometry (MUSCA) Methods, Electrochimica Acta, Vol. 321, 134648. doi:10.1016/j.electacta.2019.134648 pdf

O. Munteshari, J. Lau, A. Likitchatchawankum, B.-A. Mei, C.S. Choi, D. Butts, B.S. Dunn, L. Pilon, 2019. Thermal signature of ion intercalation and surface redox reactions mechanisms in model pseudocapacitive electrodes, Electrochimica Acta, Vol. 307, pp. 512-524. doi:10.1016/j.electacta.2019.03.185 pdf

A. Likitchatchawankum, A. Kundu, O. Munteshari, T.S. Fisher, and L. Pilon, 2019. Heat Generation in All-Solid-State Supercapacitors with Graphene Electrodes and Gel Electrolytes, Electrochimica Acta, Vol. 303, pp. 341-353. doi:10.1016/j.electacta.2019.02.031 pdf

O. Munteshari, J. Lau, D. Ashby, B. Dunn, and L. Pilon, 2018. Effects of Constituent Materials on Heat Generation in Individual EDLC Electrodes, Journal of the Electrochemical Society, Vol. 165, No. 7, pp. A1547-A1557. doi:10.1149/2.0771807jes pdf

O. Munteshari, J. Lau, A. Krishnan, B. Dunn, and L. Pilon, 2018. Isothermal Calorimeter for Measurements of Time-Dependent Heat Generation Rate in Individual Supercapacitor Electrodes, Journal of Power Sources, Vol. 374, pp. 257-268. doi:10.1016/j.jpowsour.2017.11.012 pdf