Pulsladung von Bleibatterien
Pulse charging of batteries - controlled deposition of metal?
René Groiß, Harry Döring, Jürgen Garche, Energy Storage and Energy Conversion Division of Zentrum für Sonnenenergie und Wasserstoffforschung, Helmholtzstr.8, 89081 Ulm,
At the Zentrum für Sonnenenergie und Wasserstoffforschung pulse charging is examined. This review, which is focused on lead-acid batteries, will show the problems arising during fast charging and will give suggestions how they could be solved by pulse charging procedures.
First measurements show that at least the oxygen evolution problem which is also a heat problem is reduced through pulse charging.
An important requirement for the increase in the number of electric vehicles is the improvement of the batteries. Only if the driving range, life time and price become comparable to internal combustion engine vehicles battery powered vehicles can be sold.
Fast charging will improve both the driving range of an electric vehicle and the life time of its battery. To reach recharge times comparable to the refilling of an internal combustion engine vehicle advanced charge regimes have to be used. The work at the Zentrum für Sonnenenergie und Wasserstoffforschung (ZSW) shall examine whether the advantages of pulse charge regimes will justify the increased expenditure of the charge equipment they require. The aim is to deliver a pulse charge procedure to recharge electric vehicle lead-acid batteries with advanced tubular plates within a few minutes without loosing or even increasing cycle life.
As most of the EVs employ lead-acid batteries, the fast charging work at the ZSW is focused on these types of batteries.
The fast charging aims of the ALABC (Advanced Lead-Acid Battery Consortium) for 1998 are: 50% recharge within 3min, 80% recharge within 10min and 100% recharge within 30min
Basic reactions of lead-acid batteries:
The main charging reactions of the lead-acid batteries are:
at the positive electrode:
at the negative electrode:
In flooded batteries the main side reactions are oxygen evolution at the positive electrode
and hydrogen evolution at the negative electrode.
In valve regulated lead-acid batteries, as practically all lead-acid batteries for EV applications are, oxygen produced in reaction ( 3) diffuses in gas channels through the immobilized electrolyte to the negative electrode where it is consumed electrochemically
With additional charge reaction ( 2) the sum is zero. The so called oxygen cycle (reactions ( 3) and ( 5)or ( 6) ) prevents valve regulated lead-acid batteries from excessive water loss. Reaction ( 5) or ( 6) increases the potential of the negative electrode which hinders the hydrogen evolution reaction ( 4). Reaction ( 5) and ( 6) are mainly limited by the diffusion of the oxygen to the negative electrode [ 3]. The efficiency of the oxygen cycle is high enough to prevent water loss being one of the main life limiting criteria in valve regulated lead-acid batteries. But as the oxygen cycle produces heat which has to be handled and which decreases the charge efficiency, oxygen evolution is the main problem in fast charging lead-acid batteries [ 21] [ 22].
An important side reaction in lead-acid batteries is the formation of a passivation layer between the grid and the active mass, mainly of the positive electrode. This layer usually consists of PbO [ 10]. While Pb and PbO2 are good electronic conductors, the conductivity of PbO is quite low which makes the passivation layer disrupt the connection between the grid and the active mass [ 14]. Additionally PbO is a semiconductor [ 6]. Because of different stoichiometry and impurities the PbO layers can be p-type or n-type doped. The resulting p-n junctions can also hinder the current flow.
Especially at low sulfuric acid concentrations (deep discharge) a open circuit corrosion reaction occurs:
Anodic corrosion is an important side reaction, too:
Note, the lead in this case is coming from the grid and not from the active mass [ 5]. So anodic corrosion not only decreases positive grid conductivity but also consumes water. This water loss can be the limiting factor in cycle life (number of charge and discharge cycles) of lead-acid batteries.
Another life limiting factor of lead-acid batteries is the increase in particle size. This can be
The charge rate at the valve regulated lead-acid battery is mainly limited by exhaustive oxygen production[ 4]. This produces heat by the oxygen cycle and can damage the battery by a mechanical shedding of the positive active mass by oxygen bubbles [ 12] [ 20]. Therefore the aim is to chose the maximum charge rate without too much oxygen evolution.
Most of the commercial pulse chargers use current pulses to gain additional information, namely about the internal resistance of the battery and the cables . With this information, the constant current charging can be prolonged without too much oxygen production. This has been recognized very early [ 8] and can be applied to all types of batteries which are charged by a constant voltage regime, for example lead-acid and lithium systems.
fig. 1: Scheme of current and voltage vs. time during pulse charging
At the time of charge current interruptions an immediate drop of the electrode potentials can be observed, followed by a slower decrease. This immediate drop is caused by the ohmic part of the internal resistance, while the slower decrease is a result of the overpotentials. As the ohmic part does not influence electrode reactions it can be added to the battery control voltage without harming the battery.
fig. 2: Resistance compensated charging of a lead-acid battery cell
This is demonstrated in fig. 2. The battery voltage exceeds the control voltage. The resistance compensated battery voltage is kept constant after the control voltage is reached. If charging goes further, the two voltages come closer as the IR drop decreases because of the decreasing current. By the resistance free charging method, which is also used by commercial battery chargers like the Norvik Minitcharger®, the constant current phase could be prolonged and the current in the constant voltage phase could be increased, compared to a standard CC-CV-charging (constant current, followed by constant voltage). Therefore the battery can be recharged faster.
fig. 3: Dependency of the oxygen production current on lead dioxide in 5M sulfuric acid
The second reason for employing pulse charge algorithms is to reduce gassing. As can be seen from fig. 3, the oxygen evolution current has a Tafel dependency from the potential and is clearly influenced by temperature [ 15]. Therefore it can be concluded that the oxygen evolution reaction is kinetically controlled, while the charge reaction, especially at high charge currents, is diffusion controlled. So, if we assume that the two reactions do not influence each other, the oxygen evolution itself is not much limited by diffusion. As the charge reaction especially at higher current densities will be limited by dissolution, diffusion and recrystallization processes, the balance of the two reactions is shifted to the oxygen evolution reaction after the current has been switched on and the charge reaction limiting processes become dominating [ 13]. So, if the charge current is switched off before the charge reaction starves and if the relaxation time between the charge pulses is long enough, the oxygen evolution reaction will be reduced.
A third matter for choosing a pulsed charge current is the structure of the active masses. One of the great failure modes of a lead-acid battery, as mentioned, is the enlargement of the active mass particles. In electroplating it is a well known fact that higher currents cause finer crystal structures, even if the average current is the same [ 18]. Additionally, by the locally higher overpotentials caused by a pulsed charge current, large lead sulfate crystals can be charged which are lost for conventional charge methods.
A pulse charge technique will also have an influence on the structures of the corrosion and passivation layers especially of the positive plate. But so far it is not known whether this is a good or a bad influence.
An effect which is today also rarely examined is the effect of discharge pulses between the charge pulses.
Some papers [ 9] [ 17] assume that by pulse charging the diffusion of the charge products into the bulk of the electrolyte will be enhanced. It is supposed that the higher acid concentration produced during a charge pulse will cause a higher diffusion rate. But looking at second Fick’s law, this is wrong, regarding the mean over time. The second Fick’s law can be written as
So, if we split the concentration into two terms, we can write
what can be rewritten to
As no mixed derivations occur, the diffusion of c1 and c2 are independent from each other, even though both concentrations refer to the same ions in the same solution. Looking at the first Fick’s law, we can do the same:
We can split the flux j into two parts
This enables us to rewrite a periodic flux by:
t’ being the period of j(t). So, j2 will be the medium flux which is related to the electrical current i2 by the Faraday constant.
The equations for the periodic part and the medium flux can be solved independently from each other. Afterwards, the resulting concentrations have to be added.
This resulting concentration will be the concentration caused by the constant part of the charge current (which is the medium charge current) plus a concentration profile caused by the periodic part of the charge current, which is zero in mean over time. Therefore the ion transport by diffusion will not be enhanced by a pulsed charge current. For the periodic part of the current, the changes in acid concentration will only reach into a limited penetration depth, which is given by
See fig. 4 for the resulting concentration profile.
fig. 4: Calculated concentration profiles for a sinusoidal current at different times
We suppose that the charge pulses usually employed for pulse charging do not even reach the bulk electrolyte. For positive lead-acid electrodes the pure double layer capacity of a metallic surface is about 10µF/cm2 [ 11] [ 24], for a lead dioxide surface in sulfuric acid at least 10mF/cm2 [ 23]. As the BET surface is approximately 5m²/g active mass [ 19] and the specific capacity is about 0.05Ah/g the double layer capacity of a metallic surface will be about 10F/Ah. So, for a charge current of 20C (which is twenty times the amperehour capacity in amperes, e. g. 54A for a 2.7Ah battery) a change in potential of maximum 2V/s would be expected. As the observed changes in positive electrode potential are smaller, this double layer capacity is the bottom limit. So the charge pulses stuck within the double layer and do not reach the bulk of the electrolyte.
Compared to electroplating, the charging of a lead-acid battery is quite different, even if at the negative electrode metallic lead is produced.
In electroplating, the whole lead is in solution before it is deposited to the surface. In battery charging, the lead is located in the discharged active mass and the solubility inside the electrolyte is very bad because of the sulfate ion. So, the lead has to be dissoluted immediately before it can be precipitated.
Another difference is the active surface area which is much larger in battery charging (because of the porous structure of the active masses) than in electroplating. Therefore shedding of the precipitated material because of a bad connection to the bulk material is not a problem in battery charging. But, as electroplating is usually done once but battery charging and discharging has to be repeated very often, keeping the high surface area is a problem in battery charging, especially at the negative electrode. Therefore expanders are used which keep the high surface during the whole cycle life [ 1].
But the most important thing is that in battery charging attention has to be paid on both electrodes, whilst in electroplating the counter electrode is usually not of interest.
At the ZSW the influence of charge and discharge pulses on oxygen evolution and active mass morphology shall be examined. Therefore we want to learn about the balances between the reactions occurring during charging of a lead-acid battery.
fig. 5: Scheme of the pulse charge equipment
For the following experiments a pulse charge equipment has been built which is able to do nearly all pulse charge procedures. It consists of a charge control computer (586 PC), which has the user interface and controls charging at a high level. The pulse control computer is a fast microcontroller which controls the pulse charging at the low level and is also able to sample the current and all potentials. The connection to the PC is made by an ISA slot. The data acquisition unit, also connected via ISA slot, gives an additional independent possibility to sample up to 32 channels with a sample rate of maximum 200kHz. The mass spectrometer, controlled by the PC via RS232, is used to analyze the gas inside the battery.
Every charge subcycle (for example including a 100ms charge pulse and a 100ms rest period) can be splitted into up to 50 parts. Within each part, all potentials (positive, negative or whole battery), resistance compensated or not, duration and current can be controlled. Every parameter can be controlled in dependence of every other parameter or measured value, for example potential in dependence of temperature. Switching to the next step (charge, discharge, rest) can be controlled by all measured or set values.
To make single electrode measurements, a method has been developed to attach a standard Hg/Hg2SO4 reference electrode by a diaphragm in the cell wall to a valve regulated lead-acid battery while keeping the battery gas tight. With this method the reference potential has been stable for half a year now and the batteries have kept tight.
To analyze the gas which is produced while charging, a quadrupole mass spectrometer system (Balzers 420 series) is connected to the battery. By a special inlet system (fast vent) and the secondary electron multiplier of high sensitivity the amount of gas which is consumed by the mass spectrometer is kept low. Additionally gas pressure and temperature inside the battery are measured.
Since now, small commercial valve regulated lead-acid battery single cells (Sonnenschein dryfit A506) have been used to do all the measurements. The small size of the cells (4.2Ah nominal capacity at ten hour discharge, 2.7Ah at one hour discharge) is necessary because of the limited power of the pulse charge equipment (100A, 10V). The cells are thermally isolated to come closer to the thermal situation of a larger battery and to make coulometric heat measurements analyzing the temperature course.
After a charge cycle, the battery rests for 10 minutes, is discharged with a constant current of 2.7A and rests for another 10 minutes. If the battery temperature is higher than 28°C, the equipment will wait in open circuit until the battery has reached 28°C.
fig. 6: Example of pulse charging (Sonnenschein dryfit A506, 10% state of charge)
Within a charge subcycle the changes in battery voltage are caused mainly by the negative electrode, as can be seen in fig. 6.
It is well known that the positive electrode potential is the potential which varies with time. But usually the measured time scale with standard equipment is too long to measure the fast changes of the negative electrode potential. However, with our fast equipment the slow change of positive electrode potential cannot be seen in the measured region. Instantaneous ohmic changes could be observed on both electrodes.
fig. 7: Oxygen production current fraction of battery current as a function of state of charge
Nevertheless, oxygen evolution is influenced by a pulse charging algorithm, although no changes in positive electrode potential except for the ohmic drop can be observed during charge pulses and rest periods in between, which can be seen in fig. 7. The fraction of current leading to gas evolution is calculated by heat measurements: The heat, which has not been produced by the charge reaction or the ohmic drop must be produced by the oxygen cycle. Because of the low current at the end of charging, the fraction could not be calculated for higher states of charge. It can be clearly seen that the increase of the oxygen evolution current fraction is about 5% SOC (state of charge) later for the pulsed current charge. The other parameters, medium charge current (C5, which is one fifth of the nominal ampere-hour capacity in amperes) and control voltage (2.4V resistance compensated) have been the same.
fig. 8: Charge factor for constant current and pulse charging (dryfit A506, 10C)
In fig. 8, the charge procedure has been changed between constant current and pulse charging after six consecutive charge and discharge cycles. For pulse charging, the charge factor (ratio between charged and discharged capacity between a consecutive charge and discharge cycle) is lower by about 4% in relation to the constant current charge. Another interesting fact is the "relaxation effect" of the charge factor. The reason of this effect is not well understood so far. Probably this could be an effect of the changes in internal resistance by the different charge regimes. Note, the charging in fig. 8 is switched off after a fixed time. So, if the internal resistance is higher after repeated constant current charging, the time for the constant (medium) current phase of the following pulse charge procedure would be reduced as the potential increase is faster and therefore the time for overcharging is reduced. The charge factor goes down.
The reduced oxygen production can also be observed in the single electrode potentials during charging:
Note, the battery voltage as the sum of positive and negative electrode potential first increases during the constant current phase (which is very short here because of the high charge current) and is decreased afterwards because of the rising temperature. Without this temperature compensation of the battery voltage the so-called thermal runaway could occur: The rising temperature decreases the oxygen evolution voltage which increases the oxygen production. By the oxygen cycle, the temperature is further increased, which increases oxygen production, the temperature will rise without any control.
fig. 9: Electrode potentials vs. SOC during charging for different charge regimes.
It is evident that the charge potential of the positive electrode could be decreased by a pulse charge regime, especially at low SOC. By employing discharge pulses between the charge pulses the potential shift is about twice the shift observed for single charge pulses. This shift can be explained by the reduced oxygen production at the positive electrode. As there is less oxygen diffusing to the negative electrode, the potential of the negative electrode becomes more negative, consumption of oxygen leads to a more positive potential. This is the effect which also appears if a valve regulated lead-acid battery with oxygen cycle is compared to a flooded battery: The negative electrode potential of the latter is more negative, the positive electrode potential is shifted in the negative direction. The impact of this shift in potential to the cycle life of the negative electrode is not known so far.
Pressure measurements and measurements by the mass spectrometer have shown that even for very fast charging no gas is exhausting through the valve out of the battery. Inside the battery there is very less oxygen found, all the other gas inside the battery is hydrogen. The partial pressure during rest is approximately 10hPa. During charging, the partial pressure increases to about 20hPa.. The increase in pressure during charging can be explained only by the increase of the temperature (ideal gas law).
By a pulse charge procedures it is possible to recharge lead-acid batteries within a time comparable to the refilling time of an internal combustion engine vehicle, even with a simple resistance free charging voltage equipment like the commercially available minitchargerÒ . With a improved pulse charge equipment it is possible to further influence the balance between the charge and the oxygen production reaction and shift it towards the charge reaction, especially if discharge pulses between the charge pulses are used.
By pulse charging the battery voltage is shifted towards the negative electrode, an effect which is enlarged by discharge pulses. The oxygen production will start later for pulse charging than for constant current charging.
But so far the quantitative and even the qualitative interactions between the charge and the oxygen production reaction are not well understood. The simplest model is to describe the Tafel-like oxygen production and the diffusion controlled charge reaction at the positive electrode for their own, but even the hindering of the oxygen production will not be described by this simple model. Therefore new models to describe the pulse charging of a lead-acid battery have to be found and, as the potential during a pulse seems to be not suitable, possibilities to check these models have to be thought up.
[ 1]: G. I. Aidman: J. Power Sources 59 (1996) 25
[ 2]: ALABC Research Plan 1997-1999, ALABC, P.O. Box 12036, Research Triangle Park, NC 27709 USA
[ 3]: S. Boduardo, M. Maja, N. Penazzi: J. Power Sources 55 (1995) 183
[ 4]: T. G. Chang, E. M. Valeriote, D. M. Jochim: J. Power Sources 48 (1994) 163
[ 5]: B. Culpin, D. A. J. Rand: J. Power Sources 36 (1991) 415
[ 6]: H. Döring, J. Garche, H. Dietz, K. Wiesener: J. Power Sources 30 (1990) 41
[ 7]: K. W. Klontz, N. H. Kutkut: Design considerations of extremely fast charging systems for electric vehicles, Electric Vehicle Symposium 13 (1996) 751
[ 8]: K. V. Kordesch: J. Electrochem. Soc. 119 (1972) 1053
[ 9]: L. T. Lam, H. Ozgun, L. M. D. Cranswick, D. A. J. Rand: J. Power Sources 42 (1993) 55
[ 10]: P. Mattesco, N. Bui, P. Simon, L. Albert: J. Electrochem. Soc. 144 (1997) 443
[ 11]: M. Metikoë -Hukoviº , R. Babiº , S. Briniº : J. Power Sources 64 (1997) 13
[ 12]: J. K. Nor, D. R. Smith: Very fast battery charging and battery energy management, Electric Vehicle Symposium 12 (1995) 117
[ 13]: D. K. Nowak: J. Power Sources 40 (10992) 105
[ 14]: D. Pavlov: J. Power Sources 53 (1995) 9
[ 15]: D. Pavlov: J. Power Sources 64 (1997) 131
[ 16]: D. Pavlov, E. Bashtavelova, D. Simonnson, P. Ekdunge: J. Power Sources 30 (1990) 77
[ 17]: J. Podrazhansky: Method and apparatus for charging, thawing and formatting a battery, United states patent 5307000 (1994)
[ 18]: J. Puippe: Pulse-Plating: Elektrolytische Metallabscheidung mit Pulsstrom, Eugen G. Leuze Verlag, Saulgau (1990)
[ 19]: P. Reinhardt: Untersuchungen zur Struktur und den elektrochemischen Eigenschaften von Bleidioxid in positiven Elektroden des Bleistarterakkumulators, Dissertation, Uni Dresden, 1974
[ 20]: E. M. Valeriote, T. G. Chang, D. M. Jochim: Comparative fast charging of lead/acid batteries, 26th Intl. Symp. on Automotive Technology and Automation, Aachen, Germany (1993)
[ 21]: E. M. Valeriote, D. M. Jochim: J. Power Sources 40 (1992) 93
[ 22]: E. M. Valeriote, J. K. Nor, V. A. Ettel: Very fast charging of lead/acid batteries, 5th ILZRO Int. Lead/Acid Seminar Vienna (1991)
[ 23]: M. P. Vinod, K. Vijayamohanan: J. Power Sources 50 (1994) 67
[ 24]: G. L. Wei, J. R. Wang: J. Power Sources 52 (1994) 193
Die Rechte dieses Artikels liegen beim Autor. Die Wiedergabe, auch auszugsweise, ist nur mit Genehmigung des Autors gestattet. Richten Sie Ihre Anfrage mit Angabe von Ort und Zweck der Wiedergabe an: .
The rights of this article belong to the author. Reproduction, also partial reproduction, needs the permission of the author. Send your request to: