Accelerated Ageing of Aluminum ElectrolyticCapacitor

Rashmi ShuklaDepartment of Electrical Engineering

NIT Durgapur, IndiaEmail: rashmishukla1893@gmail.com

Md. Waseem Ahmad, Nikunj Agarwal, Sandeep AnandDepartment of Electrical Engineering

IIT Kanpur, IndiaEmail: mdwaseemahmad@gmail.com,

nikunj.aggarwal007@gmail.com, me.sandeepanand@gmail.com

Abstract—Electrolytic capacitors are extensively usedin filtering and power decoupling in power electronicconverters. Low cost and high energy density of thesecapacitors make them popular. However, these capacitorsdegrade with age due to heating and evaporation ofelectrolyte. This process may take about 2 to 7 yearsdepending on the application, ambient conditions andcapacitor construction. This paper explores a method foraccelerated ageing of capacitor. The method is based onweight loss of a capacitor due to evaporation of electrolyte.Experimentation is done to validate the performance ofthe method. Values of equivalent series resistance (ESR)and capacitance (C) during degradation are measured andreported in this paper.

Index Terms—Aluminum Electrolytic Capacitor; Con-dition monitoring; Diagnosis; Equivalent series resistance(ESR); MPPT; Photovoltaic (PV)

I. INTRODUCTION

Reliability of power electronic converters is a majorissue [1]. Since failure rate of Aluminum Electrolytic Ca-pacitors (AECs) is approximately twice of that of powersemiconductors switches [2], reliability of convertersis severely affected by use of AECs. Despite of thisdrawback, electrolytic capacitors are popularly used inconverters due to their low cost and volume as comparedto film type capacitors [3]. Degradation of electrolyticcapacitor is mainly due to its electrolyte loss with time,which leads to rise in equivalent series resistance (ESR)and fall in capacitance value [4]. ACEs degradation isstrongly affected by its operating conditions, such asvoltage, current, frequency and temperature. These con-ditions lead to internal heating and increase in the coretemperature, which results in increase in the gas pressureinside the capacitor. High pressure causes the electrolyteto evaporate through the end seals. Loss of electrolyteleads to weight loss of the electrolytic capacitor. Thisalso reduced the capacitance and increases the ESR. In

extreme conditions the pressure relief vent may operatedue to excess temperature and generation of gases [5].Degradation of AECs could also lead to failure of thepower electronic converter [6].

Several methods are suggested in literature for accel-erated ageing of electrolytic capacitors [7]–[9]. In [7],capacitor is tested under different ambient temperatureand ageing of capacitor is done by injecting ripplecurrent in capacitor. In this three differnet tests areperformed, one is with lower ambient temperature andlower magnitude of continuous injected ripple current.In other two ambient temperature is high and alsomagnitude of injected current is high but the injectedcurrent is intermittent, not continuous. In [8], capacitoris aged by applying constant dc voltage and intermittenttemperature across it. In [9], thermally overstressedaccelerated ageing of capacitor is done. Here capacitor issubjected to a constant temperature higher than its rated.

In this paper, accelerated ageing of AEC is achievedby puncturing the capacitor from the top. The objectiveto perform accelerated degradation is twofold. First is toassess the long term performance of power electronicsconverter employing AECs, and second to test the effec-tiveness of diagnostic and prognosis technique employedto access the current health and remaining life of AECsin power electronics converter. The puncturing of AECincreases the rate of loss of electrolyte by evaporation,thereby increase the aging effect [10]. Reduced volumeof electrolyte leads to increase in ESR. This process isfurther accelerated at elevated temperature. In this paper,the effect of multiple punctures on the accelerated ageingof AEC is also evaluated, which is not done in [10].

The paper is organized as follows. Section II describesthe construction of AECs and motivation for choosingpuncturing as accelerated degradation technique. In sec-tion III the experimental setup for accelerated testing isdiscussed. Section IV includes the experimental results.

Fig. 1. Detailed Structure of Capacitor

Conclusions are provided in section V.

II. ALUMINIUM ELECTROLYTIC CAPACITOR

In order to analyse the degradation phenomenon, it isimportant to understand the detailed structure of AEC.

Fig. 1 shows the internal structure of capacitor. Elec-trolyte papers are interleaved in between two strips ofaluminum foil (anode and cathode) and are wound togive it a cylindrical shape [11]. The electrolyte generallyused are solute (ammonium borate) in solvent such asethylene glycol. Due to the impregnation of electrolyte,the cathode and anode becomes electrically connected[12]. There is also an aluminum oxide layer on anodefoil, which acts as the dielectric. This thin oxide layerappears due to the oxidation of electrolyte. This oxidelayer provides high dielectric constant and rectifyingproperties to the AECs. The whole assembly is put ina can and sealed using material rubber, rubber backedphenolic, molded phenolic resin or polyphylenesulfide[13].

The simplified model of AEC is represented by a se-ries combination of capacitance and ESR. The ESR of acapacitor represents the resistances offered by aluminumoxide, electrodes, spacer and electrolyte (foil, tabbing,leads, and ohmic contacts) [14]. These two parameternamely capacitance and ESR are indicative of healthof capacitors [15]. The heating produced due to theelectrical or thermal overstress, results in evaporation ofelectrolyte. This increases the ESR which further resultsin more power dissipation. Also capacitance of AECsdecreases, thereby affecting the ripple voltage handlingcapability [16]. Most of the capacitor manufacturersspecify that capacitors are considered unusable whenESR value reaches 2.8 times of the initial ESR valueand capacitance decrease in excess of 20% of the initialvalue [17].

The motivation of paper is to analyse the effect onchange in ESR and capacitance when evaporation rateof electrolyte are accelerated. The increased evaporationrate is achieved by puncturing the capacitors. This ac-celerate the weight loss of capacitor i.e. loss in volumeof electrolyte. ESR and volume of electrolyte are relatedby,

ESRt

ESR0=

(V0

Vt

)2 (1)

where, ESR0 and ESRt are the equivalent seriesresistances (ohm) of the new and aged capacitor, re-spectively. Similarly, V0 and Vt are the initial volumeof electrolyte and volume of electrolyte after an intervalof time t, respectively. The relationship of depletion inthe electrolyte volume and the effective surface area isgiven below [18],

V (t) = V (0) −A0(t)jeot (2)

where, V(0) is initial dispersion volume and V(t) isthe dispersion volume at time t, At is the oxide surfacearea of evaporation and jeo is the evaporation rate.

III. EXPERIMENTAL SETUP

In the suggested method, the capacitor is puncturedand kept at elevated temperature in an oven as shown inFigs. 2 & 3. The heating temperature is suitably chosento avoid spill of the electrolyte from the puncturedpoint. This is necessary because the degraded capacitorsare further needed for use in the power electronicsconverter. The evaporation rate is constant in this caseas after puncturing the system becomes an open system.Capacitors of rated temperature 105◦C, maximum ratedvoltage 35 V and initial ESR of 0.1067 ohm was used.The capacitor under observation was punctured withdiameter of 0.83 mm.Datasheet of these capacitors aregiven in [19].

Fig. 2. Open Vent of Punctured Capacitor

Fig. 3. Experimental Setup

All the data collected was analysed to observe howthe ESR and the capacitance values were changing overthe period of time. As per data sheet of capacitor, itis considered completely degraded when its ESR valueincreased to 200% of its initial value and capacitancevalue has decreased by 20%.

IV. EXPERIMENTAL RESULTS AND ANALYSIS

According to the procedure described in flowchart asshown in Fig. 4, data was collected at a regular intervalof 5-6 hours.

ESR increased with time, following the trend asshown in Fig. 5. This increase is mainly due to lossof electrolyte. During this aging test of capacitor, thevalue of capacitance also decreased rapidly as shownin Fig. 6. This curve matches with the graph given bymanafacturer in the datasheet [19]. This indicates theaccelerated rate of degradation of capacitor. Weight ofthe capacitors was recorded after every regular intervalsof 4-5 hours. Plot of weight loss vs time is shown inFig. 7. Further, the effect of multiple punctures in thecapacitors is observed. Three capacitors are used, with1, 3 and 5 punctures, respectively. The results are showin Fig. 8. It is observed that as the no. of puncturesincreases, ageing is also accelerated with respect to thatof one puncture. But the gain in acceleration of ageingis not significantly high.

Time taken by electrolytic capacitor to degrade nor-mally is much more than the time taken in accelerated

Measure the initial values of capacitor

Heat the capacitor at rated temperature in oven

After 5-6 hours cool the capacitor to room

temperature

Measure the weight of the capacitor

Measure ESR (ohm) and Capacitance (µF)

ESR>200%

of

initial value

?

NO

YES

Testing

Complete

Puncture a hole in the capacitor

Fig. 4. Flow chart explaining procedure of experiment

0

20

40

60

80

100

120

140

160

0 20 60 80

Perc

enta

ge in

crea

se in

ESR

40Time (in hours)

Percentage increase in ESR

Fig. 5. Percentage variation in ESR with time

tests. Lifetime in accelerated tests can be evaluated usingfollowing equations [20],

Lx = L0.KT .KR.KV (3)

where, L0 = rated lifetime, KT = temperature accel-

-14

-12

-10

-8

-6

-4

-2

00 20 40 60 80

Perc

enta

ge d

ecre

ase

in C

apac

itanc

e

Time (in hours)

Percentage decrease in Capacitance

Fig. 6. Percentage variation in capacitance with time

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 5 10 15 20 25 30 35 40

Wei

ght L

oss (

in g

ms)

Time (in hours)

Fig. 7. Weight loss vs. time

eration factor, KR = ripple current acceleration factor,KV = voltage acceleration factor and,

KT = 2

(T0 − Ta

10K

)(4a)

KR = K[1 −

(IaI0

)2]∆T0

10Ki (4b)

KV =[Va

Vr

]−n (4c)

where, T0 and Ta are the maximum and ambienttemperature, respectively. Ia and I0 are the appliedand maximum ripple current, respectively. ∆T0 is theincrease in the core temperature. Va and Vr are the

0

20

40

60

80

100

120

140

0 10 20 40 50 60

Perc

enta

ge in

crea

se in

ESR

30Time (in hours)

1 Puncture 3 Punctures 5 Punctures

Fig. 8. Percentage change in ESR with change in no. of punctures

operating and rated voltage, respectively.In this experiment, degradation of aluminum elec-

trolytic capacitor is accelerated by accelerating weightloss of electrolyte. Reduced electrolyte leads to increasein ESR and hence degradation of capacitor. So lifetimecan be calculated using following equation,

L = L0

[ Wpunctured

Wunpunctured

]x (5)

where, L and L0 is the actual lifetime and ratedlifetime, respectively, x= experimentally determined con-stant. Wpunctured and Wunpunctured is the weight lossof punctured capacitor in time t and weight loss ofunpunctured capacitor in the same time t respectively.This accelerated method takes lesser time as comparedto other methods.

V. CONCLUSION

AEC limits the lifetime of power electronic devices.To estimate the lifetime, a proper diagnostic techniqueis required. This accelerated ageing test is done todevelop a technique for health monitoring of capacitor.The whole process took around 60-70 hours to degrade.In 70 hours, ESR increased by 138% and capacitancedecreased by 12.8%. The degradation in capacitanceis accompanied by weight loss. This weight loss isaccelerated by puncturing. Weight loss of the capacitoris related to weight loss of electrolyte. Also the effect oflarger no. of punctures is marginal.

REFERENCES

[1] S. Yang, A. Bryant, P. Mawby, D. Xiang, L. Ran and P. Tavner,“An Industry-Based Survey of Reliability in Power Electronic

Converters”, IEEE Trans. on Ind. Appl., vol. 47, no. 3, pp.1441-1451, May-June 2011.

[2] Huai Wang, Dao Zhou, Blaabjerg, F.,“A Reliability OrientedDesign Method for Power Electronic Converters”, in AppliedPower Electronics Conference and Exposition (APEC), 2013Twenty-Eighth Annual IEEE , pp.2921-2928, 17-21 March2013.

[3] H.Ma and L.Wang, “Fault diagnosis and Failure Prediction ofAluminium Electrolytic Capacitors in Power Electronic Con-verters”, in Proc. of IEEE 31st Annu. Conf. of Ind. Electron.Soc. (IECON), 6 Nov. 2005, Raleigh, NC.

[4] Chetan S. Kulkarni, Jose R. Celaya, Kai Goebel and GautamBiswas, “Physics Based Electrolytic Capacitor DegradationModels for Prognostic Studies under Thermal Overstress”,Proceedings of First European Conference of the Prognosticsand Health Management Society. 2012.

[5] Gasperi Michael L., “Life Prediction Model for AluminumElectrolytic Capacitors”, Industry Applications Conference,1996. Thirty-First IAS Annual Meeting, IAS’96., ConferenceRecord of the 1996 IEEE, Vol. 3. IEEE, 1996.

[6] Bo Sun, Xuejun Fan, Yuan, C.A., Cheng Qian, Guoqi Zhang,“A degradation Model of Aluminum Electrolytic Capacitors forLED Drivers,” in Thermal, Mechanical and Multi-Physics Sim-ulation and Experiments in Microelectronics and Microsystems(EuroSimE), 2015 16th International Conference on , pp.1-4,19-22 April 2015.

[7] Sankaran, V.A.; Rees, F.L.; Avant, C.S., “Electrolytic capacitorlife testing and prediction,” in Industry Applications Conference,1997. Thirty-Second IAS Annual Meeting, IAS ’97., ConferenceRecord of the 1997 IEEE , vol.2, no., pp.1058-1065 vol.2, 5-9Oct 1997.

[8] Jano, R.; Pitica, D., “Accelerated ageing tests for predictingcapacitor lifetimes,” in Design and Technology in ElectronicPackaging (SIITME), 2011 IEEE 17th International Symposiumfor , vol., no., pp.63-68, 20-23 Oct. 2011.

[9] Celaya, J.R.; Kulkarni, C.; Saha, S.; Biswas, G.; Goebel, K.,“Accelerated aging in electrolytic capacitors for prognostics,”in Reliability and Maintainability Symposium (RAMS), 2012Proceedings - Annual , vol., no., pp.1-6, 23-26 Jan. 2012

[10] Stephani G, Cartmill K., Arnold J., Nicholas C.,Greg C., “ANew Method for testing Electrolytic Capacitors to Compare LifeExpectancy, IAMPS Device Packaging Conference 2013.

[11] United Chemi-Con, Inc., “Undersfanding Aluminum Elec-trolytic. Second Edition”, Richmond, IL, 1995.

[12] “General Description of Aluminum Electrolytic Capacitor”,Nichion Corp.

[13] C. Kulkarni, G. Biswas, and X. Koutsoukos, “A PrognosisCase Study for Electrolytic Capacitor Degradation in Dc-DcConverters, PHM Conference 2009, October 2009.

[14] A. Bengt, Electrolytic Capacitors Theory and Applications,RIFA Electrolytic Capacitors, Sweden, 1995.

[15] Kulkarni C.S., G. Biswas, J. Celaya, and K. Goebel. “PrognosticTechniques for Capacitor Degradation and Health Monitoring”,in 8th International Workshop on Structural Health Monitoring(IWSHM), 2011.

[16] IEC, “60384-4-1 Fixed capacitors for use in electronicequip-ment”, 2007-03.

[17] EIA-463-A, “Fixed Aluminum Electrolytic Capacitors for Al-ternating Current Motor Starting, Heavy Duty (type 1) and forStandard Duty (type 2) (ANSI/EIA-463-a-99), EIA.

[18] Rusdi M., Y. Moroi, H. Nakahara, and O. Shibata, “Evaporationfrom Water Ethylene Glycol Liquid Mixture”, Langmuir, 2005,pp. 7308-7310.

[19] SAMWHA Electolytic Capacitor, “MINIATURE ALUMINUMELECTROLYTIC CAPACITORS”, Datasheet [Online].Available: http://www.samwha.com/electric/product/list pdf2/RD.pdf

[20] Jano, R., Pitica, D., “Accelerated Ageing Tests for PredictingCapacitor Lifetimes,” in Design and Technology in ElectronicPackaging (SIITME), 2011 IEEE 17th International Symposiumfor, pp.63-68, 20-23 Oct. 2011.