Journal of Power Sources160(2006)
1302–1307
Overcharge investigation of lithium-ion polymer batteries
Yuqun Zeng,Kai Wu,Deyu Wang,Zhaoxiang Wang∗,Liquan Chen
Laboratory for Solid State Ionics,Institute of Physics,Chinese Academy of Sciences,Beijing100080,China
Received21December2005;received in revised form6February2006;accepted6February2006
Available online29March2006
Abstract
The overcharge performances of lithium-ion polymer batteries(LIPB)have been studied by monitoring their temperature variation and analyzing the generated heat during overcharge.The critical concentration of lithium in the cathode material is determined for the thermal runaway of the battery.Solutions against the thermal runaway are proposed based on these results.
©2006Elsevier B.V.All rights reserved.
Keywords:Lithium-ion polymer battery;Overcharge;Cathode material
1.Introduction
Rechargeable lithium-ion batteries(LIB)have been popu-larly accepted in a wide variety of applications including mobile phones,blue teeth,notebook computers,(hybrid)electric vehi-cles,etc.due to their high energy density,long cycle life and other unique properties.The lately developed lithium-ion poly-mer batteries(LIPB),because of their higher energy density and safety than the traditional lithium-ion batteries with liquid elec-trolyte,are expected to share more of the battery market.In spite of the great success in development and market,battery safety remains the main concern of the consumers and the fabrica-tors.This concern becomes more severe when the battery works at high temperatures,high-rate charge and discharge,extreme overcharge and other abusive operations[1].The more energy is stored,the more hazardous will be the energy storage system potentially.
Overcharge performance is an important feature of a bat-tery.Some thermal analyses of LIB materials have been carried out in order to understand the overcharge mechanism[2–4]. Some authors used accelerated rate calorimetry(ARC)while others applied differential scanning calorimetry(DSC),trying to learn more about the reasons for the thermal runaway[5–7]. Tobishima and Yamaki[1,8]reported the overcharge reaction while Leising ducted systematic studies on the over-∗Corresponding author.Tel.:+861082649050;fax:+861082649046.
E-mail address:wangzx@aphy.iphy.ac(Z.Wang).charge of LIB cells[9,10].Most of these studies are emphasizing the importance of the overcharge performances of the battery materials.
Systematic investigations are rare on the overcharge perfor-mances and solutions to the overcharge-induced failure for the newly born lithium-ion polymer batteries.This paper will inves-tigate the thermal performances of Bellcore-type lithium-ion polymer batteries during overcharge and propose some solu-tions to the thermal runaway from the point of views of battery designing and operation.
2.Experimental
The stacked lithium-ion polymer batteries with a capac-ity of650mAh were products of Dongguan Amperex Elec-tronics Technology Co.,Ltd.(ATL).Cathodefilms were pre-pared by mixing LiCoO2and carbon black in a solution of poly(vinyl difluoride)-hexafluoropropylene(PVDF-HFP) binder and dibutyl phthalate
dissolved in ace-tone.The mixture was cast on a to prepare the cathode film.Then the cathodefilm and Al stacked together by hot lamination to prepare the cathode electrode.The Mylarfilm was removed me
anwhile.Analogously were the anodes made with carbonaceous microsphere(CMS)as the active material but hot laminated onto a piece of Cu grid.PVDF-HFP treated porous polypropylenefilm(25␮m thick)was used as the sep-arator.One mole per liter of LiPF6dissolved in EC:DEC:EMC (1:1:1,v/v/v)was used as the electrolyte.Another hot lamina-tion completed the construction of a bi-cell.Seven bi-cells were
0378-7753/$–see front matter©2006Elsevier B.V.All rights reserved. doi:10.1016/j.jpowsour.2006.02.009
Y.Zeng et al./Journal of Power Sources 160(2006)1302–13071303
hermetically sealed in a pocket to form a 650mAh LIPB pack-age.
K-type
thermocouples were built inside the LIPB or sand-wiched between the bi-cells to monitor their internal t
empera-tures.The down-leads of the thermocouple were sealed on one side of the pocket before the electrolyte was filled in.Another thermocouple was fixed at the center of the upper surface of the cell to monitor its external temperature.The testing was conducted on Arbin BT-2000battery tester at 23±3◦C.All batteries were at first charged to 4.2V and then discharged to 3.0V before the overcharge test was carried out.
For the convenience of discussion,the nominal cell balance (CB)of the capacity of the anode to the cathode is defined as CB =
CMS weight at unit area ×300LiCoO 2weight at unit area ×137
The 300and 137mAh g −1are the available capacities of CMS and LiCoO 2,respectively.The CB for all the cells is 1.05unless specified in this paper.
In order to understand the overcharge mechanism,both the charger’s voltage (U )and battery’s voltage (E )were recorded.The U was measured after the battery was charged for 2min,while the E was obtained after the battery stood by for 10min,following the 2-min galvanostatic charge.3.Results and discussion
The theoretical specific capacity of LiCoO 2is 274mAh g −1.The coulombic efficiency in the first cycle o
f an LIPB is ca.92%.This means that the capacity loss of LiCoO 2is roughly 11mAh g −1considering that the actually available (reversible)capacity of LiCoO 2is 137mAh g −1when the bat-tery is charged to 4.2V .Therefore,the actual value of x in Li x CoO 2at full discharge (3.0V)after the initial charge is ((137×0.92+137)/274)=0.96.As the coulombic efficiencies for most commercial cells and laboratory test cells are over 99%in the second and subsequent cycles,it is reasonable to suppose that the coulombic efficiency was 100%if the battery is cycled between 3.0and 4.2V .Then it is calculated that the theoretical value of x at full charge (4.2V)is (x =0.96−0.5=)0.46.Now we further presume that the amount of Li ions extracted from the cathode is proportional to the galvanostatic overcharge time before x =0in Li x CoO 2.Considering the polarization effect of the battery and the decomposition of the electrolyte,this might be not reasonable.However,such a presumption will help esti-mate the Li concentration in the overcharged cathode materials and obtain some interesting information as shown in the follow-ing.
Fig.1presents the voltage and current profiles of a 650mAh LIPB with the 2C/12V overcharge mode.“2C/12V”means that the battery is galvanostatically charged to 4.2V in half an hour.Then it is further charged at the same current den-sity for some predetermined time (4h in total,for example).If the voltage of the cell reaches 12V within this period,the cell will automatically turn to be potentiostatic
ally charged.It is seen that the voltage of the cell increases with time until the highest value (5.1V)is reached at t =57.5min.Then it drops
Fig.1.Plots of the voltage and current of a 650mAh LIPB with the 2C/12V overcharge mode.
sharply to 3.9V .After that,the voltage becomes stable at around 4.0V .
Fig.2shows the internal and external temperature profiles of this battery in the above overcharge test.The variation of its inter-nal temperature is quite similar to that of its external temperature.Both temperatures are just a little over the room temperature until about t =52.0min.Then both of them increase sharply with time.The internal temperature reaches its maximum,167◦C,at t =65.5min while the external temperature reaches its summit,112◦C,at t =67min.The time interval between these two peak values is clearly due to the heat conduction of the battery mate-rials.As the battery is further charged,both the internal and external temperatures drop quickly and then keep unchanged at around 105and 80◦C,respectively.
The same type of cells are charged with 1C/12V and 3C/10V overcharge modes,respectively,to find out the influence of charge rate on their overcharge performances.For the conve-nience of comparison,the voltage and the internal temperature of the battery have been shown in Figs.3and 4,re
spectively.The internal temperature rises obviously after ca.31min and reaches the maximum at t =37min in the 1C/12V overcharge
Fig.2.The time-dependent internal and external temperature plots of the same 650mAh LIPB (as in Fig.1)with the 2C/12V overcharge mode.
1304Y.Zeng et al./Journal of Power Sources 160(2006)1302–1307
Fig.3.Comparison of voltage and internal temperature of a 650mAh LIPB in the overcharge mode of 1C/12V (current =650mA).
test.However,the same type of battery fails to pass the 3C/10V test.Figs.2–4show clearly that the maximum internal tempera-ture increases with increasing charge rate,implying that charge rate is an important factor responsible for the thermal runaway of an overcharged battery.
Cell balance (CB)has been considered another factor that might influence the safety of an overcharged battery because the extra lithium extracted from the cathode material might be plated on the surface of the anode if the CB value is not suffi-ciently high.Therefore,650mAh LIPBs with CB as high as 2are assembled and charged with 1C/12V ,2C/12V and 3C/10V overcharge modes,respectively,in order to evaluate the impact of the CB on their overcharge performances.In comparison with batteries with CB =1.05,these batteries demonstrate no obvious differences in their overcharge performances:their temperatures rise sharply after ca.x =0.16and reach their summits short after x =0.Meanwhile the voltages reach the peak value at ca.x =0.In addition,3C/10V overcharge also leads to fire short after x =0.The temperature is about 190◦C before fire.This indicates that increasing the CB value cannot prevent the thermal runaway of the battery in a rather wide range above its normal value.That
Fig.4.Variation of voltage and internal temperature of a 650mAh LIPB with charge time at 3C rate (3C/10V/4h overcharge mode).
Table 1
The temperature,charge time,calculated Li concentration x in Li x CoO 2and battery voltage at some critical points during overcharge Rate
Items
Onset
temperature Maximum voltage Maximum internal temperature 1C
Temperature (◦C)28.541.8111.5Charge time (min)
95.6106.7127.9
Calculated x in Li x CoO 20.160.07a
V oltage (V)
4.799
5.136  4.4142C
Temperature (◦C)32.746.8169.6Charge time (min)
47.953.563.9
Calculated x in Li x CoO 20.160.07a
V oltage (V)
4.856
5.086  3.8563C
Temperature (◦C)39.943.4797.9Charge time (min)
31.932.737.6Calculated x in Li x CoO 20.160.140.02
V oltage (V)
5.000
5.042
b
The coulombic efficiency of the battery is supposed to be 100%before x =0in Li x CoO 2in the overcharge test.
a The calculated x value is negative at this point.
b The battery catches fire at this point and the corresponding voltage is not available.
is,lithium deposition on the anode surface does not occur or is not the reason for the thermal runaway of the battery with CB =1.05.
LIPB cells with smaller (200mAh)and larger (2000mAh)capacities were also fabricated with the same materials and over-charged with the same modes.The overcharge performances of both cells are similar to and consistent with that of the above batteries.However,the 2000mAh battery caught fire even at 1C/12V overcharge while the 200mAh cell passed the 3C/10V overcharge test.
Summarizing the relationship between the calculated Li con-centration in Li x CoO 2and the internal temperature of the bat-tery,it is interesting to find that the temperature of the bat-tery does not obviously increase before ca.x =0.16and the maximum temperature arrives later than that of the voltag
e.Table 1lists the temperature,charge time,calculated Li con-centration in Li x CoO 2and the battery voltage at some criti-cal points during overcharge.It shows that the temperature at which the temperature begins to rise sharply (onset tempera-ture)and the maximum temperature of the battery all depend strongly on and increase with the charge rate.However,the influence of the charge rate is slight on the maximum voltage that the battery can reach.In addition,the voltages at which the sharp temperature rise begins increases with the charge rate,probably because of the polarization effect of the battery materials.
It is interesting that the onset temperature of the cells with different overcharge modes is at x =0.16although the calculated Li concentration at the maximum voltage and temperature of the cell departs from the reasonable value,independent of the charge rate.This implies that the calculated x value indeed reflects the actual x value in Li x CoO 2before any thermal runaway takes place.Therefore,x =0.16is an important reference point for solutions to the thermal runaway of the battery.
Y.Zeng et al./Journal of Power Sources160(2006)1302–13071305 The above results may be summarized as follows:
(i)x=0.16in Li x CoO2is the critical point for the beginning
of sharp rise of temperature and voltage during galvanos-
tatic overcharge.Both the temperature and the voltage reach
their maximum at ca.x=0.
(ii)Increasing the CB value cannot prevent the thermal run-
away in a wide range above the normal value.
(iii)Charge rate is an important factor that influences the over-
charge performance of the cell,especially those with high
capacities.
(iv)Battery capacity,or more exactly,the ratio of surface area to
the battery capacity is another important factor that impacts
the safety of the battery.
Clearly the quick temperature rise and thefiring of the cell are
all related to the heat generated during1C/12V overcharge and
the heat generation process is related to the exothermic reactions
in the overcharged cell[1–11].Therefore,the abovefindings
actually point to one point:the rate of heat generation(related
to charge rate and exothermic reaction rate)and the rate of heat
dissipation(related to the size and shape as well as the materials)
of the cell are responsible for the battery safety.This means that
the heat generation and heat dissipation is a pair of discrepancy.
If the heat generation and dissipation are in balance,the tem-
perature of the cell will be under control.This occurs only in
low-capacity cells or cells operated with low current density.If
this balance is broken but the rate of heat generation is not very
high,the temperature of the cell will rise and a new balance is
built at a higher level.This is the case for most practical cells.
However,if the balance is broken and the temperature overrides
a critical point before the new balance is built,the cell will catch
fire or be exploded.In this case,the cell fails to pass the over-
charge test.
In order to understand the above thermal performances of the
cell,analysis to the heat generation during the charge process is
necessary.The heat balance of any thermodynamic system can
be written as
i ρi C p
i
∂T
∂t
=∇2(kT)+
i
Q g
i
(1)
At the boundary of a cell
∂T ∂n
surface
=−
h
k
(T surface−T room)(2)
whereρrepresents the mass density of a component in the sys-tem,T for the equilibrium temperature of the system,t for time, C p for the specific heat capacity of a component,k for the heat conduction coefficient,Q for the heat generation rate and h for thefilm coefficient between the cell pocket and air,n for the normal direction of the battery surface.
Dahn and co-workers[11]used the dimensionless Biot num-ber(defined as Bi=h(V/A)/k)as a criteria of heatflow rate within a body when compared to heatflow at the surface(V for vol-ume and A for surface area of the cell)and to evaluate if their cylindrical(18650model)battery can be regarded as a lumped m
ass with uniform temperature.It has been popularly accepted that the temperature of a system can be regarded uniform[12] if its Bi≤0.1.For our650mAh LIPBs,h=13.5W m−2K−1, k=20.06W m−1K−1,V=4.52×10−6m3,A=4.04×10−3m2. Therefore,
the Biot number is calculated to be0.0754.This means that a lumped mass approach is a good approximation for the solid LIPB.That is,our LIPB can be simplified as a lumped mass with homogeneous temperature during cycling. Based on this,Eq.(1)can be simplified as
mC p
d T
d t
=−hA(T−T room)+Q in(3)
where C p=1280J kg−1K−1[13]and Q in is the input electric power.
The heat generation rate Q g can be classified into two parts during charging,due to physical processes(Q p)and due to chemical reactions(Q c,including entropy variation and reac-tion between the active material and the electrolyte,etc.).That is,
Q g=Q p+Q c(4) According to Eq.(3),the generated heat can be calculated if the temperature T is known.Q p can be written as
Q p=I(U−E)(5) Thus Q c can be deduced by Q c=Q−Q p.The generated heat will lead to temperature rise.Clearly temperature rise is easier to probe than the heat generation.Therefore,temperature varia-tion rather than the generated heat will be used in the following discussion.Supposing that all the input energy is converted to heat,Eq.(3)can be used to estimate how much the tempera-ture will rise.Fig.5compares the actual(measured)and virtual (calculated)temperature curves of the cell overcharged with the 2C/12V mode.
Fig.5.Comparison of the virtual(calculated)and actual(measured)tempera-tures of the battery overcharged with the2C/12V mode.The virtual temperature is calculated by supposing that all the input electric energy is converted to heat (Eq.(3)).The measured temperature represents the actual internal temperature of the battery during overcharge.
1306Y.Zeng et al./Journal of Power Sources 160(2006)1302–1307
The electric energy input into a battery will be converted into chemical energy and heat.For an LIPB in charge,the following reactions are expected.On the anode 6C +y Li ++y e −=Li y C 6(6)
On the cathode
LiCoO 2→Li 1−x CoO 2+x Li ++x e −
(7)During the charge process,the chemical energy is stored in the anode and cathode.Take the 2C/12V overcharge mode for example,the charge process may be divided into four periods (Fig.5):
(i)Between A and B,the cell temperature rises very little.This is the normal charge process for most batteries.Most of the input energy is converted to chemical energy and stored in the cell between x =0.96and 0.16.
(ii)Between B (x =0.16)and C (x =0),the calculated tem-perature rises at first and then turns to drop but the cell temperature keeps rising sharply.In this period,most of the input energy is converted to heat.
(iii)Between C and D (after x =0),the output heat becomes
powerbi官方电脑版下载more than the input electric energy.That means that an exothermic chemical reaction occurs in this period.By sub-tracting the total output heat from the total input energy,it is seen that the released chemical energy is only a small portion of the total input energy.
(iv)Far beyond x =0(after D),the calculated temperature is
roughly equal to the actual temperature of the cell,indicat-ing that our supposition for Fig.5(most of the input energy is converted to heat)is reasonable.Fig.6presents the contributions of Q g ,Q p and Q c to the temperature rise of the battery during overcharge.It is seen that the contributions of both the physical and chemical heats before t =45min are very little.The battery temperature rises very little in this period.Between t =45and 75min,the chemical
heat
Fig.6.Contributions of the chemical (Q c )and physical (Q p )heats to the tem-perature rise of the battery with 2C/12V mode.
is responsible for the sharp increase and the maximum of the battery temperature.After the summit of the battery temperature the chemical heat contribution decreases quickly to about zero while the physical heat begins to take its part at this point.The physical heat works to keep the battery temperature at a higher level of balance after t =75min.
The cells that passed the 3C/12V/4h overcharge test were disassembled after the overcharge test.Inspection of each part of the battery shows that
(i)Very little gas is produced,most of which is CO 2.Therefore,electrolyte decomposition is not the main reason for the fire of the cell.
(ii)No obvious changes are observed on the Al and Cu current
collectors.
(iii)The anode material is golden to reddish when the cell is dis-assembled in the glove box.It becomes white after exposed to air for a few seconds.The white species is LiOH due to the violent reaction of Li
C 6and water (moisture)in air.Violent reaction takes place and a lot of gas is given off when the anode is put into water.Therefore the material on the anode after overcharge must be LiC 6and deposited Li.These indicate that the chemical energy stored in the anode during charge is not released or that the anode does not contribute to the above temperature increase of the battery.(iv)The cathode material becomes gray after overcharge.It was
reported that the end member of LiCoO 2at deep charge is highly oxidative CoO 2[14].Therefore,the cathode material will oxidize the electrolyte and get itself reduced.During this process,CoO 2is reduced to CoO x and much of the chemical energy stored in the cathode is released:
CoO 2+electrolyte →CoO x +O 2+gas +heat
(8)
The gas in Eq.(8)includes CO 2and other oxidation products of the electrolyte.The reaction can take place whether or not the cell can pass the overcharge test.The difference between them is the reaction rate and the rate of heat dissipation.For an LIPB that can pass the overcharge test,its chemical energy stored in the cathode must be released at a sufficiently slow speed and dissipated at a rather high speed by which the internal tempera-ture of the cell is below some critical value.In this process,the ene
rgy stored in the anode is remained.
For a cell that cannot pass the overcharge test,however,its cathode releases its chemical energy at a rather high speed in comparison to the heat dissipation rate.In this way,the temper-ature of the cell rises sharply and reaches some critical value,inducing reactions in the anode.Then the anode further releases much of its chemical energy,triggering thermal runaway.It is known that metallic lithium melts at 180◦C and becomes very active at and above this temperature.Therefore,it seems that 170◦C is the highest temperature that the cell can tolerate without catching fire during the overcharge test.With this,it is proposed that the critical temperature for the thermal runaway is ca.180◦C.
In order to show the importance of the temperature control or the rate of heat generation/dissipation for the safety of the