- Email: [email protected]
Forward and reverse differential-pulse effects applied in the formation of a solid electrolyte interface to enhance the performance of lithium batteries
Electrochimica Acta 147 (2014) 582–588
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Forward and reverse differential-pulse effects applied in the formation of a solid electrolyte interface to enhance the performance of lithium batteries Fu-Ming Wang a,b, *, Jung-Chi Wang a , John Rick b a b
Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei, Taiwan Sustainable Energy Center, National Taiwan University of Science and Technology, Taipei, Taiwan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 July 2014 Received in revised form 30 September 2014 Accepted 1 October 2014 Available online 7 October 2014
In Li-ion batteries, the solid electrolyte interface (SEI) plays a crucial role in transferring Li ions into active materials through an electrochemical driving force. SEI is a composite layer containing of inorganic and organic components, which are fabricated by the salt degradation products and partial or complete reduction products of the solvent of the electrolyte at the battery's initial charge-discharge cycle. The chemical properties of SEI and the electrochemical driving force must be mutually optimized so as to strengthen its integrity, while minimizing irreversible SEI formation; thereby suppressing the decomposition at high temperatures. In this study, we investigated a new method of creating the SEI, i.e. the forward and reverse differential-pulse (FRDP) method, which balances the reaction kinetics of SEI formation. Furthermore, the use of the FRDP method also creates a SEI with a modified kinetic reaction route that affects battery performance. Here, we present data from the first charge-discharge and cycle performance at a high rate and a high temperature, obtained using scanning electron microscopy, electrochemical impedance spectroscopy, X-ray photoelectron spectroscopy, and Li+-diffusion kinetics analysis. Our findings indicate that the use of the FRDP method for generating the SEI results in a 58% reduction in the SEI's ionic diffusion activation energy and a 4.5% increase in battery capacity at room temperature, while increasing battery performance at 60 C stability compared to batteries in which the SEI is formed using the constant-current method. ã 2014 Elsevier Ltd. All rights reserved.
Keyword: Solid electrolyte interface Pulse Lithium ion battery Forward and reverse Kinetic Dynamic
1. Introduction Li-ion batteries are widely used as a source of rechargeable power for portable electronic devices. In addition to their potential use in various current and future devices, Li-ion batteries have potential to be used in high-power systems such as electric vehicles (EVs) and hybrid EVs. The high reliability and uniform standard of the EV battery design is postulated to be comparable to that of batteries manufactured for current 3C (Communications, Computers and Consumer Electronics) applications. To obtain high power, the rapid ionic transfer that results from the electrochemical reaction, which occurs at the electrode/electrolyte interface, must be accelerated. In lithium batteries, this electrode/electrolyte interface is named the solid-electrolyte interface (SEI) or the passivation layer, and it is used for de-solvating Li ions that are solvated, due to the carbonates' high
* Corresponding author at: IB 606, 43 Keelung Road, Section 4, Taipei 106, Taiwan, R.O.C. Tel.: +886 2 27303755; fax: +886 2 27376922.; e-mail: [email protected]. edu.tw http://dx.doi.org/10.1016/j.electacta.2014.10.004 0013-4686/ ã 2014 Elsevier Ltd. All rights reserved.
dielectric constant. During the initial charging of these batteries, Li ions are extracted from the cathode and intercalated into the anode through a non-aqueous electrolyte, such as ethylene carbonate (EC), propylene carbonate (PC), or ethyl methyl carbonate (EMC). When these carbonates contain a lithium salt (LiPF6), a complex SEI layer including organic and inorganic compounds can be readily generated on the anode's surface through a reduction reaction [1,2]. Typically, when the SEI's organic compounds contain (CH2OCO2Li)2, ROCO2Li and RCO2Li, theyenhance the battery life and the charge-transfer rate because they provide ion-hopping sites on the highly functional groups (-C=O and -C-O-) of the SEI [1,3]. However, ensuring the formation of the organic compounds is challenging because the electrochemical reaction of the cyclic carbonates undergoes substantial mutations. Zhang et al., demonstrated that a twoelectron pathway is used for the electrochemical decomposition of EC and a one-electron pathway is used for the decomposition of PC because the electron-donating group, -CH3, augments the electron density within the cyclic structure [1] and thereby restricts the flow of electrons and generates an organic compound [1,4]. Research conducted using electrochemical impedance spectroscopy (EIS) has
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
583
(a) Bulk Electrolyte
Li+
Bulk Electrolyte
SEI MCMB CC
e- (CC)
Cu
Cu
(b) Bulk Electrolyte
Li+
Bulk Electrolyte
SEI MCMB FRDP
e- (FP)
Cu
Cu
e- (RP) Fig. 1. The solid electrolyte interface formation concepts of lithium ion battery applied by (a) CC and (b) FRDP charging protocols.
revealed that the SEI is formed in two early stages [5]: in the first stage, a porous, highly resistive, and dimensionally unstable SEI layer grows before Li ions intercalate into graphite, whereas the second stage occurs concurrently with the intercalation of Li ions and generates a highly elaborate and conductive organic SEI layer. Several reports have indicated that diverse approaches can control the formation of an effective SEI in Li-ion batteries. Electrolyte additives, which are low-molecular-weight monomers that exhibit high reduction potentials, are used for reinforcing SEI formation and preventing the polarization of the electrochemical reaction. Chemical compounds based on vinylene carbonates [6,7], sulfones [8], maleimides [9–13], and phosphates [14,15] were previously developed for generating an SEI that prevented the exfoliation effects on carbon when PC was used as the solvent: using this process, a 3D electrochemically active surface area was formed and frame-retardant applications were improved. However, large-scale solubility and electrochemical stability of the additives in electrolytes cannot be readily maintained. Recently, considerable attention has been devoted to modifying the surface of electrodes or particles by using so-called artificial SEI-formation methods such as atomic layer deposition/ molecular layer deposition (ALD/MLD) [16,17], solid state synthesis [18], and radio frequency sputtering coating [19]. Miller et al., used coarse-grained lattice models for studying the reductive reaction of Li+ and showed that pulse charging can be used to suppress dendrite formation on the anode's surface when the polarization is high: the analysis also revealed that dendrite formation is linked to the competition between the timescales of Li+ diffusion and the reduction of SEI formation. Miller et al., suggested the use of short pulse durations shifts this competition
in favor of Li+ diffusion and thus lowers the propensity for dendrite formation [20]. Previous results [4,20] have indicated that DP charging affects the formation and the composition of the SEI layer in distinct and dynamic ways; however, the dynamic electrochemical equilibrium of both Li+ diffusion and the electron transfer must be re-examined because Li+ diffusion is extremely slow compared with electron conduction. We conducted this study to determine whether balanced SEI formation on the anode surface can be achieved by further refining the DP protocol. In this study, we investigated the use of a new forward and reverse differential-pulse (FRDP) method developed for SEI formation. Fig. 1 shows the SEI formation concepts of the charging protocols of CC and FRDP. Owing to Li+ diffusion in the bulk electrolyte being much slower than electron transfer in the current collector, we anticipate that the diffusivity of Li+ can be enhanced by the reverse pulse (RP) effect. Fig. 1a shows that the slow movement of lithium ions achieves a non-uniform SEI formation from electron transfer in the CC protocol; in contrast Fig. 1b shows that the RP approach causes the uniform movement of lithium ions resulting in a homogeneous SEI layer. We examined the reaction kinetics, the charge–discharge characteristics, cycle performance, and the electrochemical properties of Li-ion batteries in which an SEI film was grown. Xray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM) were used to examine the composition and the morphology of SEI films prepared using FRDP and CC protocols. Our results can help guide the development of a new concept that can be applied in the preparation of elaborate SEI layers in commercial Li-ion batteries.
584
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
TC
PF
TD
PR TC
Fig. 2. A typical current variation curve of pulse charging including the charging time (TC), the magnitude of the forward pulse (PF), the pulse duration time (TD) and the magnitude of the reverse pulse (PR).
2. Experimental The Li-ion batteries used in this study were two-electrode-type 503,759 C cells, which were provided by SYNergy ScienTech Corp., Taiwan (ALB, aluminum-plastic laminated film featuring exterior dimensions of 5.0 37 59 mm); the battery capacity was 1000 mAh. The cathode consisted of 91 wt% LiCoO2 as the active material, 6 wt% polyvinylidene difluoride (PVDF) as the binder, 2wt % graphite (KS-6), and 1 wt% of vapor grown carbon fiber (VGCF) as the conductive additive. The graphite anode was composed of 93 wt% mesocarbon microbeads (MCMB-2528, Osaka Gas), 4 wt% PVDF as binder, and 3 wt% KS-6 as the conductive additive. The electrolyte used was 1 M lithium hexafluorophosphate (LiPF6) mixed in EC/EMC (1:2 in volume ratio) and 2 wt% vinylene carbonate. The experimental batteries were charged as follows: the batteries were CC charged at 0.1C for 10 min and were then subjected to DP charging at a forward pulse (FP) magnitude of 0.2C and RP magnitudes of 0 and 0.002C, with a 1.6-ms duration and a charging time of 0.4 ms up to 4.2 V. The charged voltage was then maintained at 4.2 V until the current was below 0.01C. The batteries were then discharged at a 0.2C constant-current rate to 2.75 V. Fig. 2 shows characteristics typical of DP charging; the four parameters shown are the magnitude of the FP (PF), the charging time (TC), the pulse duration time (TD), and the magnitude of the RP (PR). In this study, we developed two new FRDP methods that avoided ionic transfer polarization affecting the reaction kinetics during SEI formation, and thereby shortened the charging period. Based on previous results, we set the duration to appropriately 1.6 ms and the charging time to 0.4 ms [4]. In the case of the control batteries, the charging was performed as follows: the batteries were CC charged at 0.1C for 10 min and then subjected to CC charging at 0.2C up to 4.2 V, the charged voltage was maintained at 4.2 V until the current was below 0.01C The batteries were then discharged at a 0.2C constant-current rate to 2.75 V. All the charging programs used and the code names of batteries are listed in Table 1. The morphologies of the MCMB electrodes dissembled from the 503,759C cells were examined by performing SEM on a LEO1530 microscope (JSM-6480 JEOL) at an accelerating voltage of 15 keV. The MCMB surface composition was determined using XPS (PHI, 1600S) to evaluate the SEI characteristics. To prepare samples
for SEM and XPS analyses, in a dry room, the MCMB electrodes were washed using dimethyl carbonate and then dried under vacuum for 24 h before examination. To avoid exposure to moisture, all samples were prepared in an Ar-filled glove box. When electrochemical impedance spectroscopy (EIS) was performed, a Biologic (VMP3) potentiostat/galvanostat was used in the frequency range of 10MHz to 0.01Hz at an AC amplitude of 5 mV at 25 C. To measure cycling ability, the batteries that were charged using the FRDP and CC charging protocols were further CC charged at 1C and discharged at 1C for 50 cycles at 60 C: the measurements were obtained using a Maccor Battery Tester, Series 4000. 3. Results and discussion 3.1. The 1st charge-discharge characteristics observed using four methods of SEI formation In Figures 2 and 3, all the charge curves of FRDP protocol exhibited low-voltage plateaus and were unaffected by RP charging, indicating that FP charging efficiently nullified the activation and concentration polarization that occurs while the lithium ions are being de-intercalated from the cathode. Fig. 3 also shows that the average working discharge voltages of the batteries used for testing the FRDP1 and FRDP2 protocols were 3.798 and 3.796V, respectively, which are almost equal, but are considerably higher than the voltage measured for the battery used for testing the CC method (3.751V). The discharge capacities of batteries using these four methods are summarized in Table 2. The results in Fig. 3 and Table 2 clearly reveal that the RP charging affected the discharge capacity of the FRDP batteries, which indicates that RP charging effectively decreased the irreversible electrochemical polarization on the anode's surface, where Li+ diffusion is in equilibrium with electron transfer and thus helped obtain a battery capacity higher than that of the CC battery. These results indicated that the FRDP protocols can be used to construct an effective SEI. 3.2. Analysis of EIS characteristics To reveal the differences between the FRDP and CC protocols, we performed impedance analysis to identify the effects of using these methods for forming the SEI. Fig. 4 shows the EIS results
Table 1 The program of CC and FRDP charging protocols for the SEI formation Cells (number)
Charging protocol procedure
CC FRDP1 FRDP2
0.1 C CC charging for 10 min ! 0.2 C CC charging to 4.2 V ! CV charging to 4.2 V 0.1 C CC charging for 10 min ! 0.2 C forward and 0 C reverse pulses (TP/0.4 mS, TR/1.6 mS) charging to 4.2 V ! CV charging to 4.2 V 0.1 C CC charging for 10 min ! 0.2 C forward and 0.002 C reverse pulses (TP/0.4 mS, TR/1.6 mS) charging to 4.2 V ! CV charging to 4.2 V
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
585
4.2
E vs (Li/Li+)/V
4.0
3.8
3.6
3.4
3.2
CC FRDP1 FRDP2
3.0 0
200
Q/mAh
400
600
800
1000
Q/mAh Fig. 3. The 1st charge and discharge curves of the LiCoO2/ MCMB (503759 type) full cells on (a) CC, (b) FRDP1 and (c) FRDP2 formation protocols at room temperature.
obtained after SEI formation. The equivalent-circuit model shown in Fig. 4 represents the battery's internal construction, which is illustrated as five components: the series resistance (R1), which represents the wire and electrolyte resistances; the first resistance (R2)-capacitance element, which represents the impedance of the bulk SEI; the second resistance (R3)-capacitance circuit, which generates the interface on the anode surface; the Warburg element (W1), which reflects the ionic-diffusion resistance of the electrode materials; and the interface capacitance (C1) of the battery [9]. Because the same electrolyte and wire were used in all batteries, the R1 values in Fig. 4 are all approximately the same. However, the R2 value of the FRDP2 battery (18 mV) was lower than the values of the FRDP1 (21 mV) and CC (21 mV) batteries. This result indicated that when the FRDP2 protocol was used, a unique SEI was formed on the anode's surface as a consequence of the shortest RP process being used, which implies that the compounds that compose this SEI are highly diffusive and that the SEI provides a suitable infrastructure for Li+ diffusion. The R3 value of the FRDP2 battery shows that the charge-transfer resistance (12 mV) on the anode's surface of this battery was also lower than those of all other batteries, which indicates that the use of the FRDP2 protocol adequately reduces the electrolyte in a lowpolarization electrochemical reaction. Thus, when the FRDP2 protocol was used, the SEI was formed on the anode's surface, which indicates that a connection between the SEI and the anode's surface was established effectively. The value of W1 indicates that the use of the FRDP2 protocol not only affected SEI formation on the anode's surface, but also improved the diffusion impedance of electrode materials; this is because when this method was used, the diffusion of Li+ ions and the transfer of electrons occurred concurrently in an equilibrium state and balanced the reaction kinetics. The discharge capacity of the batteries shown in Fig. 3 and Table 2 agrees with the results shown Table 2 The discharge capacity of the 1st cycle of the batteries at room temperature. Formation Protocol
CC
FRDP1
FRDP2
Discharge Capacity/mAh
1020.7
1052.4
1064.8
in Fig. 4, and these results together suggest that the RP process of battery formation can be used in commercial battery manufacturing. 3.3. The morphological effects of FRDP on MCMB electrodes Fig. 5 shows SEM images of morphologies of MCMB surfaces that were generated using the four SEI formation protocols; the images were obtained after completing the first cycle at room temperature. These four photographs show that the morphologies of the MCMB samples are nearly identical. However, in the SEM image of Fig. 5a, the MCMB particles appear to be distributed unevenly and the SEI on the MCMB surface is inhomogeneous (Fig. 1a), which implies that the SEI formed using the CC protocol was extremely small. By contrast, in Figs. 5b and 5c, the SEI can be observed clearly and is seen to uniformly cover the MCMB particles, which indicates that the use of the FRDP protocols triggered the electrochemical reduction of the electrolyte and generated the SEI. However, the SEIs formed after using the two FRDP methods covered the MCMB surface to various extents. When the RP effect was absent (FRDP1 battery; Fig. 5b), VGCF and KS6 dispersed onto the parts of the MCMB surface that were covered by the SEI. When the FRDP2 protocol was used and the RP effect was present (Fig. 1b), the SEI formed was uniformly distributed and exhibited large-scale cladding on the MCMB particles (Fig. 5c). Single MCMB particles were connected by the SEI, indicating that Li+ diffusion is enhanced and proceeds readily when the use of a suitable RP effectively reduces the polarization of SEI formation by affecting the reaction dynamics. 3.4. XPS analysis of the SEI Fig. 6a shows the spectra of the C1s on the MCMB surface of the batteries in which the four SEI formation protocols were used; the spectra were collected after completing the first cycle at room temperature. The peak detected in these spectra at approximately 289 to 291 eV was assigned to the carbonate group, whereas the broad peak at approximately 283 to 286 eV was assigned to a group of superimposed peaks, including the C-C (graphite) peak at
586
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
0.16 0.14 R1 )
0.12
!Z"/
0.10
R2 ( )
R3 ( )
W1 ( )
C1 (Coul)
CC
0.073
0.021
0.018
0.033
175
FRDP1
0.073
0.021
0.018
0.024
215
FRDP2
0.072
0.018
0.012
0.021
300
0.08 0.06 0.04 0.02 0.00 0.00
CC FRDP1 FRDP2 0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Z'/! Fig. 4. Impedance Spectra comparisons of the LiCoO2/ MCMB (503759 type) full cell after formation (Inset is the equivalent circuit used to fit experimental curves). Schematic representation of the item of equivalent circuit with four formation protocols has been arranged inside the figure.
284.2 eV, the ROCO2Li peak at 289-to-291 eV, the Li2CO3 peak at 290 eV, the C-O-C peak at 286.2 eV, and the C-H peak at 285.5-to286 eV [3,13,21,22]. The 284-to-285-eV region in Fig. 6a represents the sp2 of the highly conductive (-C-C-) carbon-carbon structure of the MCMB and the conductive carbons of VGCF and KS-6. The intensity of the peaks in the 284-to-285-eV region was the highest when the
FRDP2 protocol was used, which indicates that the MCMB electrode was highly conductive because a thin SEI was formed; this agrees with the result shown in Fig. 5. Furthermore, the intensity of the peaks of FRDP2 protocol in the 289.5-to-290.5-eV region was high and did not shift compared to the CC protocol, indicating that the FRDP2 method reinforces the kinetic route that forwards a two-electron pathway reaction of organic SEI formation
Fig. 5. The SEM micrograph of MCMB electrode after battery formation with the formation protocols of (a) CC, (b) FRDP1 and (d) FRDP2 methods.
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
low-polarization reduction, yielding highly diffusive ionic products (e.g., lithium ethylene dicarbonate) [23]. The use of the FRDP2 protocol resulted in the abundant formation of SEI reduction products in the 289.5-to-290.5-eV region, which indicates that a weak RP effect shown in Fig. 1b generates an unhindered electron flow during SEI formation. A comprehensive survey of the results presented thus far indicates that the SEI formed using the FRDP2 protocol was thin, but it contained high levels of organic compounds that support ionic diffusivity. Furthermore, as shown in Fig. 6a, a clear shoulder in the 285.5to-286.6-eV region was detected only in the case of the FRDP1 battery. Previous studies [1,3] have suggested two reasons for this. Low-molecular-weight electrolyte additives such as maleimides [9–13], vinylene carbonate (VC) [6,7], or sulfones [8] can be electrochemically polymerized; thereby causing C-H bonding of the polymer on the anode surface. In this study, the VC was used to the additive in the electrolyte. In addition, Novak et al. [3] suggests that this shoulder peak corresponds to the ringopening products generated from EC; these products, which can be expressed as the polymer of -(CH2CH2O)n- (polyethylene oxide, PEO). According to above literature, the PEO provides extremely low ionic conductivity and limits the ionic diffusion at the interface. Fig. 6b depicts O1s spectra of the MCMB surface, wherein a multi-component lode with an oxygen atom at 529-to-536 eV is deconvoluted and may include peaks to C&9552;O of ROCO2Li at 532.5 eV, C-OH at 533 eV, Li2CO3 at 532 eV, LiOH at 531.9 eV and CO at 531 eV [3,13]. According to Fig. 6b, the use of the FRDP2 protocol results in agreement with the Fig. 6a that the peak (532-to-532.5 eV) intensity was typically higher than that when other protocols were used. This result demonstrates a highly ionic conductive compound was formed.
10000
(a)
Intensity
8000
CC FRDP1 FRDP2
6000
4000
2000
294
292
290
288
286
284
282
Binding Energy/eV
25000
(b) CC FRDP1 FRDP2
Intensity
20000
15000
10000
5000
538
536
534
532
530
528
587
526
3.5. Cycle-life characteristics at high temperatures
Binding Energy/eV Fig. 6. The (a) C1s and (b) O1s XPS spectrum of the MCMB electrode surface after battery formation with the four formation protocols.
[1]. Our results showed that when the FRDP2 protocol was used, the peak intensity was typically higher than that when other protocols were used. A previous study [1] indicated that when EC was used as the electrolyte solvent, an SEI was readily formed during the electrochemical reduction reaction. A stable twoelectron pathway for the decomposition of EC is favored, due to the
Fig. 7 displays measurements of the cycle life of the batteries in which the SEI was formed using the FRDP and CC protocols at room temperature. At the end of seven cycles, all the batteries exhibited maintainable capacity retention at 60 C. However, after 50 cycles, among the batteries tested, the FRDP2 battery exhibited the highest stability and the lowest cycle fading. The results shown in Fig. 7 agree with the results of XPS analysis shown in Fig. 6 and indicate that the thin SEI that was formed using the FRDP2 protocol effectively connected particles and thus enabled Li+ diffusion, and
1.00
CC FRDP1 FRDP2
-1.5
-1.8
-1
Log (Rct /
Q/Ah !
-1
)
0.99
0.98
-2.1
CC FRDP1 FRDP2
0.97
0
10
-2.4 20
30
40
50
Cycle number Fig. 7. The cycle performance of LiCoO2/ MCMB (503759 type) full cell with the four formation protocols under 1C charge and 1C discharge rate at 60 C measurement.
3.2
3.4
3.6 -1
3.8
-1
1000T /K
Fig. 8. The Arrhenius plot of the reciprocal resistances corresponding to chargetransfer components on MCMB surface formed by three formation protocols.
588
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
Table 3 The Activation Energy derived for the Rct on MCMB surface formed by four formation protocols. Formation Protocol
CC
FRDP1
FRDP2
Discharge Capacity/mAh
29.48
25.84
17.33
further that the RP used for forming the SEI must be precisely estimated to eliminate the polarization of the electrochemical reaction on the anode surface. In the FRDP1 battery, a slow (0.2C) FP charging was used for SEI formation, with the results indicating that PEO cannot significantly support ionic diffusion on an interface, where the impedance of the interface between the SEI and the anode's surface is high, compared with interfaces that are formed more rapidly (0.5C), i.e. by forward charging [4]. These results indicate that, previously unappreciated FP and RP charging, effectively counter forward charging and disrupt the reactiondynamics and the equilibrium of Li+/electron transfer. 3.6. Charge-transfer activation energies A previous study [24], suggests that the two measurement EIS semicircles, shown in Fig. 4, can be attributed to the diffusion of Li+ through the SEI layer at medium frequencies and to the chargetransfer reaction on the electrode's surface at low frequencies. The charge-transfer component is closely associated with the chemistry of the SEI. A simple, thermally activated process is used for forming the charge-transfer component Rct: 1 ¼ A0 eEa =RT Rct where A0,R, and Ea denote a pre-exponential constant, the standard gas constant, and the activation energy, respectively. The Arrhenius equation can be used to extract the activation energies of Li+ transport during these processes by plotting log(R1) versus reciprocal temperature (1/T plots). Thus, we plotted the reciprocal Rct temperature dependences to derive the activation energies required for Li+ transport across the charge-transfer component (Fig. 8). The slope obtained in the case of each formation protocol was converted to activation energies using the following equation [24]: Ea = -19.144 slope (kJ mol1) Table 3 lists the activation energies of the charge-transfer components that were formed using the four SEI-formation protocols. The results clearly indicate that when we used the FRDP2 protocol, we obtained the smallest activation energy (17.33 kJ mol1) for the bulk SEI/anode surface interface; indicating that the electrochemical reaction proceeded efficiently and exhibited the least amount of polarization. This result agrees with the results of our XPS, EIS, SEM, and cycle-life measurements. 4. Conclusions In this study, the optimized FRDP protocol (FRDP2) is used to balance the dynamic of electron transfer and the ionic diffusion at the initial electrochemical reaction of SEI formation in lithium batteries. However, an appropriate pulse magnitude must be optimized, which is the key that reinforces the kinetic route of SEI formation, promoting the unhindered electron flow to the EC reaction and VC reduction. The performance of two-electrode commercial-type batteries was investigated with the results for the battery made using the FRDP3 SEI-formation protocol demonstrating that the use of RP can lead to an outstanding battery performance with respect to high temperature cycling ability. Compared with the CC method of SEI formation, this FRDP technique should provide greater in the future EV market.
Acknowledgement The author is grateful for the financial support from the Ministry of Science and Technology (MOST) of Taiwan, R.O.C, under grant numbers 102-2221-E-011-016-MY3 and 103-ET-E-011-003ET. Reference [1] S.S. Zhang, A review on electrolyte additives for lithium-ion batteries, J. Power Sources 162 (2006) 1379. [2] A.M. Haregewoin, E.G. Leggesse, J.C. Jiang, F.M. Wang, B.J. Hwang, S.D. Lin, Comparative study on the solid electrolyte interface formation by the reduction of alkyl carbonates in lithium ion battery, Electrochim. Acta 136 (2014) 274. [3] P. Verma, P. Maire, P. Novak, A review of the features and analyses of the solid electrolyte interphase in Li-ion batteries, Electrochim, Acta 55 (2010) 6332. [4] F.M. Wang, H.Y. Wang, M.H. Yu, Y.J. Hsiao, Y. Tsai, Differential pulse effects of solid electrolyte interface formation for improving performance on highpower lithium ion battery, J. Power Sources 196 (2011) 10395. [5] S.S. Zhang, K. Xu, T.R. Jow, EIS study on the formation of solid electrolyte interface in Li-ion battery, Electrochim. Acta 51 (2006) 1636. [6] A.V. Cresce, S.M. Russell, D.R. Baker, K.J. Gaskell, K. Xu, In Situ and Quantitative Characterization of Solid Electrolyte Interphases, Nano Lett. 14 (2014) 1405. [7] I.A. Shkrob, Y. Zhu, T.W. Marin, D. Abraham, Reduction of Carbonate Electrolytes and the Formation of Solid- Electrolyte Interface (SEI) in Lithium-Ion Batteries. 2. Radiolytically Induced Polymerization of Ethylene Carbonate, J. Phys. Chem. C 117 (2013) 19270. [8] R. Wagner, S. Brox, J. Kasnatscheew, D.R. Gallus, M. Amereller, I. Cekic-Laskovic, M. Winter, Vinyl sulfones as SEI-forming additives in propylene carbonate based electrolytes for lithium-ion batteries, Electrochem. Commun. 40 (2014) 80. [9] F.M. Wang, M.H. Yu, C.S. Cheng, S.A. Pradanawati, S.C. Lo, J. Rick, Phenylenedimaleimide positional isomers used as lithium ion battery electrolyte additives: Relating physical and electrochemical characterization to battery performance, J. Power Sources 231 (2013) 18. [10] C.S. Cheng, W.R. Liu, F.M. Wang, A novel ionic host solid electrolyte interface formation on reduced graphene oxide of lithium ion battery, Electrochim. Acta 106 (2013) 425. [11] C.S. Cheng, F.M. Wang, J. Rick, Aqueous Additive for Lithium Ion Batteries: Promotes Novel Solid Electrolyte Interface (SEI) Layer with Overall Cost Reduction, Int. J. Electrochem. Sci. 7 (2012) 8676. [12] C.S. Cheng, F.M. Wang, J. Rick, Breakthrough additive technology for improving the performance of high-power lithium ion batteries, MRS Commun. 2 (2012) 5. [13] F.M. Wang, H.M. Cheng, H.C. Wu, S.Y. Chu, C.S. Cheng, C.R. Yang, Novel SEI formation of maleimide-based additives and its improvement of capability and cyclicability in lithium ion batteries, Electrochim. Acta 54 (2009) 3344. [14] M. S. Ding K. Xu, S. Zhang, T. R. Jow J. L. Allen, Evaluation of Fluorinated Alkyl Phosphates as Flame Retardants in Electrolytes for Li-Ion Batteries I. Physical and Electrochemical Properties, J. Electrochem. Soc. 150 (2003) A161. [15] K. Xu, S. Zhang, J.L. Allen, T.R. Jow, Evaluation of Fluorinated Alkyl Phosphates as Flame Retardants in Electrolytes for Li-Ion Batteries I. Performance in Cell, J. Electrochem. Soc. 150 (2003) A170. [16] H.Y. Wang, F.M. Wang, Electrochemical Investigation of an Artificial Solid Electrolyte Interface for Improving the Cycle-ability of Lithium Ion Batteries Using an Atomic Layer Deposition on a Graphite Electrode, J. Power Sources 233 (2013) 1. [17] A.J. Loebl, C.J. Oldham, C.K. Devine, G.N. Parsons, P.S. Solid electrolyte interphase on lithium-ion carbon nanofiber electrodes by atomic and molecular layer deposition Fedkiw, J. Electrochem. Soc. 160 (2013) A1971. [18] J. Chong, S. Xun, X. Song, G. Liu, V.S. Battaglia, Surface stabilized LiNi0.5Mn1.5O4 cathode materials with high-rate capability and long cycle life for lithium ion batteries, Nano Energy 2 (2013) 283. [19] J.J. Kim, T. Yoon, K.J. Lee, S.M. Oh, Y.E. Sung, The effects of radio frequency sputtering of TiO2 on Li[Li0.07Ni0.38Co0.15Mn0.4]O2 Cathode for Lithium Ion Batteries, J. Nanosci. Nanotechnol. 13 (2013) 7924. [20] M.Z. Mayers, J.W. Kaminski, T.F. Miller, Suppression of Dendrite Formation via Pulse Charging in Rechargeable Lithium Metal Batteries, J. Phys. Chem. C 116 (2012) 26214. [21] F.M. Wang, M.H. Yu, Y.J. Hsiao, Y. Tsai, B.J. Hwang, Y.Y. Wang, C.C. Wan, Aging Effects to Solid Electrolyte Interface (SEI) Membrane Formation and the Performance Analysis of Lithium Ion Batteries, Int. J. Electrochem. Sci. 6 (2011) 1014. [22] S.A. Pradanawati, F.M. Wang, J. Rick, In Situ formation of pentafluorophosphate benzimidazole anion stabilizes high-temperature performance of lithium-ion batteries, Electrochim. Acta 135 (2014) 388. [23] M. Nie, D. Chalasani, D.P. Abraham, Y. Chen, A. Bose, B.L. Lucht, Lithium Ion Battery Graphite Solid Electrolyte Interphase Revealed by Microscopy and Spectroscopy, J. Phys. Chem. C 117 (2013) 1257. [24] K. Xu, Y. Lam, S.S. Zhang, T.R. Jow, T.B. Curtis, J. Phys. Chem. C 111 (2007) 7411.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Forward and reverse differential-pulse effects applied in the formation of a solid electrolyte interface to enhance the performance of lithium batteries Fu-Ming Wang a,b, *, Jung-Chi Wang a , John Rick b a b
Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei, Taiwan Sustainable Energy Center, National Taiwan University of Science and Technology, Taipei, Taiwan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 July 2014 Received in revised form 30 September 2014 Accepted 1 October 2014 Available online 7 October 2014
In Li-ion batteries, the solid electrolyte interface (SEI) plays a crucial role in transferring Li ions into active materials through an electrochemical driving force. SEI is a composite layer containing of inorganic and organic components, which are fabricated by the salt degradation products and partial or complete reduction products of the solvent of the electrolyte at the battery's initial charge-discharge cycle. The chemical properties of SEI and the electrochemical driving force must be mutually optimized so as to strengthen its integrity, while minimizing irreversible SEI formation; thereby suppressing the decomposition at high temperatures. In this study, we investigated a new method of creating the SEI, i.e. the forward and reverse differential-pulse (FRDP) method, which balances the reaction kinetics of SEI formation. Furthermore, the use of the FRDP method also creates a SEI with a modified kinetic reaction route that affects battery performance. Here, we present data from the first charge-discharge and cycle performance at a high rate and a high temperature, obtained using scanning electron microscopy, electrochemical impedance spectroscopy, X-ray photoelectron spectroscopy, and Li+-diffusion kinetics analysis. Our findings indicate that the use of the FRDP method for generating the SEI results in a 58% reduction in the SEI's ionic diffusion activation energy and a 4.5% increase in battery capacity at room temperature, while increasing battery performance at 60 C stability compared to batteries in which the SEI is formed using the constant-current method. ã 2014 Elsevier Ltd. All rights reserved.
Keyword: Solid electrolyte interface Pulse Lithium ion battery Forward and reverse Kinetic Dynamic
1. Introduction Li-ion batteries are widely used as a source of rechargeable power for portable electronic devices. In addition to their potential use in various current and future devices, Li-ion batteries have potential to be used in high-power systems such as electric vehicles (EVs) and hybrid EVs. The high reliability and uniform standard of the EV battery design is postulated to be comparable to that of batteries manufactured for current 3C (Communications, Computers and Consumer Electronics) applications. To obtain high power, the rapid ionic transfer that results from the electrochemical reaction, which occurs at the electrode/electrolyte interface, must be accelerated. In lithium batteries, this electrode/electrolyte interface is named the solid-electrolyte interface (SEI) or the passivation layer, and it is used for de-solvating Li ions that are solvated, due to the carbonates' high
* Corresponding author at: IB 606, 43 Keelung Road, Section 4, Taipei 106, Taiwan, R.O.C. Tel.: +886 2 27303755; fax: +886 2 27376922.; e-mail: [email protected]. edu.tw http://dx.doi.org/10.1016/j.electacta.2014.10.004 0013-4686/ ã 2014 Elsevier Ltd. All rights reserved.
dielectric constant. During the initial charging of these batteries, Li ions are extracted from the cathode and intercalated into the anode through a non-aqueous electrolyte, such as ethylene carbonate (EC), propylene carbonate (PC), or ethyl methyl carbonate (EMC). When these carbonates contain a lithium salt (LiPF6), a complex SEI layer including organic and inorganic compounds can be readily generated on the anode's surface through a reduction reaction [1,2]. Typically, when the SEI's organic compounds contain (CH2OCO2Li)2, ROCO2Li and RCO2Li, theyenhance the battery life and the charge-transfer rate because they provide ion-hopping sites on the highly functional groups (-C=O and -C-O-) of the SEI [1,3]. However, ensuring the formation of the organic compounds is challenging because the electrochemical reaction of the cyclic carbonates undergoes substantial mutations. Zhang et al., demonstrated that a twoelectron pathway is used for the electrochemical decomposition of EC and a one-electron pathway is used for the decomposition of PC because the electron-donating group, -CH3, augments the electron density within the cyclic structure [1] and thereby restricts the flow of electrons and generates an organic compound [1,4]. Research conducted using electrochemical impedance spectroscopy (EIS) has
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
583
(a) Bulk Electrolyte
Li+
Bulk Electrolyte
SEI MCMB CC
e- (CC)
Cu
Cu
(b) Bulk Electrolyte
Li+
Bulk Electrolyte
SEI MCMB FRDP
e- (FP)
Cu
Cu
e- (RP) Fig. 1. The solid electrolyte interface formation concepts of lithium ion battery applied by (a) CC and (b) FRDP charging protocols.
revealed that the SEI is formed in two early stages [5]: in the first stage, a porous, highly resistive, and dimensionally unstable SEI layer grows before Li ions intercalate into graphite, whereas the second stage occurs concurrently with the intercalation of Li ions and generates a highly elaborate and conductive organic SEI layer. Several reports have indicated that diverse approaches can control the formation of an effective SEI in Li-ion batteries. Electrolyte additives, which are low-molecular-weight monomers that exhibit high reduction potentials, are used for reinforcing SEI formation and preventing the polarization of the electrochemical reaction. Chemical compounds based on vinylene carbonates [6,7], sulfones [8], maleimides [9–13], and phosphates [14,15] were previously developed for generating an SEI that prevented the exfoliation effects on carbon when PC was used as the solvent: using this process, a 3D electrochemically active surface area was formed and frame-retardant applications were improved. However, large-scale solubility and electrochemical stability of the additives in electrolytes cannot be readily maintained. Recently, considerable attention has been devoted to modifying the surface of electrodes or particles by using so-called artificial SEI-formation methods such as atomic layer deposition/ molecular layer deposition (ALD/MLD) [16,17], solid state synthesis [18], and radio frequency sputtering coating [19]. Miller et al., used coarse-grained lattice models for studying the reductive reaction of Li+ and showed that pulse charging can be used to suppress dendrite formation on the anode's surface when the polarization is high: the analysis also revealed that dendrite formation is linked to the competition between the timescales of Li+ diffusion and the reduction of SEI formation. Miller et al., suggested the use of short pulse durations shifts this competition
in favor of Li+ diffusion and thus lowers the propensity for dendrite formation [20]. Previous results [4,20] have indicated that DP charging affects the formation and the composition of the SEI layer in distinct and dynamic ways; however, the dynamic electrochemical equilibrium of both Li+ diffusion and the electron transfer must be re-examined because Li+ diffusion is extremely slow compared with electron conduction. We conducted this study to determine whether balanced SEI formation on the anode surface can be achieved by further refining the DP protocol. In this study, we investigated the use of a new forward and reverse differential-pulse (FRDP) method developed for SEI formation. Fig. 1 shows the SEI formation concepts of the charging protocols of CC and FRDP. Owing to Li+ diffusion in the bulk electrolyte being much slower than electron transfer in the current collector, we anticipate that the diffusivity of Li+ can be enhanced by the reverse pulse (RP) effect. Fig. 1a shows that the slow movement of lithium ions achieves a non-uniform SEI formation from electron transfer in the CC protocol; in contrast Fig. 1b shows that the RP approach causes the uniform movement of lithium ions resulting in a homogeneous SEI layer. We examined the reaction kinetics, the charge–discharge characteristics, cycle performance, and the electrochemical properties of Li-ion batteries in which an SEI film was grown. Xray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM) were used to examine the composition and the morphology of SEI films prepared using FRDP and CC protocols. Our results can help guide the development of a new concept that can be applied in the preparation of elaborate SEI layers in commercial Li-ion batteries.
584
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
TC
PF
TD
PR TC
Fig. 2. A typical current variation curve of pulse charging including the charging time (TC), the magnitude of the forward pulse (PF), the pulse duration time (TD) and the magnitude of the reverse pulse (PR).
2. Experimental The Li-ion batteries used in this study were two-electrode-type 503,759 C cells, which were provided by SYNergy ScienTech Corp., Taiwan (ALB, aluminum-plastic laminated film featuring exterior dimensions of 5.0 37 59 mm); the battery capacity was 1000 mAh. The cathode consisted of 91 wt% LiCoO2 as the active material, 6 wt% polyvinylidene difluoride (PVDF) as the binder, 2wt % graphite (KS-6), and 1 wt% of vapor grown carbon fiber (VGCF) as the conductive additive. The graphite anode was composed of 93 wt% mesocarbon microbeads (MCMB-2528, Osaka Gas), 4 wt% PVDF as binder, and 3 wt% KS-6 as the conductive additive. The electrolyte used was 1 M lithium hexafluorophosphate (LiPF6) mixed in EC/EMC (1:2 in volume ratio) and 2 wt% vinylene carbonate. The experimental batteries were charged as follows: the batteries were CC charged at 0.1C for 10 min and were then subjected to DP charging at a forward pulse (FP) magnitude of 0.2C and RP magnitudes of 0 and 0.002C, with a 1.6-ms duration and a charging time of 0.4 ms up to 4.2 V. The charged voltage was then maintained at 4.2 V until the current was below 0.01C. The batteries were then discharged at a 0.2C constant-current rate to 2.75 V. Fig. 2 shows characteristics typical of DP charging; the four parameters shown are the magnitude of the FP (PF), the charging time (TC), the pulse duration time (TD), and the magnitude of the RP (PR). In this study, we developed two new FRDP methods that avoided ionic transfer polarization affecting the reaction kinetics during SEI formation, and thereby shortened the charging period. Based on previous results, we set the duration to appropriately 1.6 ms and the charging time to 0.4 ms [4]. In the case of the control batteries, the charging was performed as follows: the batteries were CC charged at 0.1C for 10 min and then subjected to CC charging at 0.2C up to 4.2 V, the charged voltage was maintained at 4.2 V until the current was below 0.01C The batteries were then discharged at a 0.2C constant-current rate to 2.75 V. All the charging programs used and the code names of batteries are listed in Table 1. The morphologies of the MCMB electrodes dissembled from the 503,759C cells were examined by performing SEM on a LEO1530 microscope (JSM-6480 JEOL) at an accelerating voltage of 15 keV. The MCMB surface composition was determined using XPS (PHI, 1600S) to evaluate the SEI characteristics. To prepare samples
for SEM and XPS analyses, in a dry room, the MCMB electrodes were washed using dimethyl carbonate and then dried under vacuum for 24 h before examination. To avoid exposure to moisture, all samples were prepared in an Ar-filled glove box. When electrochemical impedance spectroscopy (EIS) was performed, a Biologic (VMP3) potentiostat/galvanostat was used in the frequency range of 10MHz to 0.01Hz at an AC amplitude of 5 mV at 25 C. To measure cycling ability, the batteries that were charged using the FRDP and CC charging protocols were further CC charged at 1C and discharged at 1C for 50 cycles at 60 C: the measurements were obtained using a Maccor Battery Tester, Series 4000. 3. Results and discussion 3.1. The 1st charge-discharge characteristics observed using four methods of SEI formation In Figures 2 and 3, all the charge curves of FRDP protocol exhibited low-voltage plateaus and were unaffected by RP charging, indicating that FP charging efficiently nullified the activation and concentration polarization that occurs while the lithium ions are being de-intercalated from the cathode. Fig. 3 also shows that the average working discharge voltages of the batteries used for testing the FRDP1 and FRDP2 protocols were 3.798 and 3.796V, respectively, which are almost equal, but are considerably higher than the voltage measured for the battery used for testing the CC method (3.751V). The discharge capacities of batteries using these four methods are summarized in Table 2. The results in Fig. 3 and Table 2 clearly reveal that the RP charging affected the discharge capacity of the FRDP batteries, which indicates that RP charging effectively decreased the irreversible electrochemical polarization on the anode's surface, where Li+ diffusion is in equilibrium with electron transfer and thus helped obtain a battery capacity higher than that of the CC battery. These results indicated that the FRDP protocols can be used to construct an effective SEI. 3.2. Analysis of EIS characteristics To reveal the differences between the FRDP and CC protocols, we performed impedance analysis to identify the effects of using these methods for forming the SEI. Fig. 4 shows the EIS results
Table 1 The program of CC and FRDP charging protocols for the SEI formation Cells (number)
Charging protocol procedure
CC FRDP1 FRDP2
0.1 C CC charging for 10 min ! 0.2 C CC charging to 4.2 V ! CV charging to 4.2 V 0.1 C CC charging for 10 min ! 0.2 C forward and 0 C reverse pulses (TP/0.4 mS, TR/1.6 mS) charging to 4.2 V ! CV charging to 4.2 V 0.1 C CC charging for 10 min ! 0.2 C forward and 0.002 C reverse pulses (TP/0.4 mS, TR/1.6 mS) charging to 4.2 V ! CV charging to 4.2 V
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
585
4.2
E vs (Li/Li+)/V
4.0
3.8
3.6
3.4
3.2
CC FRDP1 FRDP2
3.0 0
200
Q/mAh
400
600
800
1000
Q/mAh Fig. 3. The 1st charge and discharge curves of the LiCoO2/ MCMB (503759 type) full cells on (a) CC, (b) FRDP1 and (c) FRDP2 formation protocols at room temperature.
obtained after SEI formation. The equivalent-circuit model shown in Fig. 4 represents the battery's internal construction, which is illustrated as five components: the series resistance (R1), which represents the wire and electrolyte resistances; the first resistance (R2)-capacitance element, which represents the impedance of the bulk SEI; the second resistance (R3)-capacitance circuit, which generates the interface on the anode surface; the Warburg element (W1), which reflects the ionic-diffusion resistance of the electrode materials; and the interface capacitance (C1) of the battery [9]. Because the same electrolyte and wire were used in all batteries, the R1 values in Fig. 4 are all approximately the same. However, the R2 value of the FRDP2 battery (18 mV) was lower than the values of the FRDP1 (21 mV) and CC (21 mV) batteries. This result indicated that when the FRDP2 protocol was used, a unique SEI was formed on the anode's surface as a consequence of the shortest RP process being used, which implies that the compounds that compose this SEI are highly diffusive and that the SEI provides a suitable infrastructure for Li+ diffusion. The R3 value of the FRDP2 battery shows that the charge-transfer resistance (12 mV) on the anode's surface of this battery was also lower than those of all other batteries, which indicates that the use of the FRDP2 protocol adequately reduces the electrolyte in a lowpolarization electrochemical reaction. Thus, when the FRDP2 protocol was used, the SEI was formed on the anode's surface, which indicates that a connection between the SEI and the anode's surface was established effectively. The value of W1 indicates that the use of the FRDP2 protocol not only affected SEI formation on the anode's surface, but also improved the diffusion impedance of electrode materials; this is because when this method was used, the diffusion of Li+ ions and the transfer of electrons occurred concurrently in an equilibrium state and balanced the reaction kinetics. The discharge capacity of the batteries shown in Fig. 3 and Table 2 agrees with the results shown Table 2 The discharge capacity of the 1st cycle of the batteries at room temperature. Formation Protocol
CC
FRDP1
FRDP2
Discharge Capacity/mAh
1020.7
1052.4
1064.8
in Fig. 4, and these results together suggest that the RP process of battery formation can be used in commercial battery manufacturing. 3.3. The morphological effects of FRDP on MCMB electrodes Fig. 5 shows SEM images of morphologies of MCMB surfaces that were generated using the four SEI formation protocols; the images were obtained after completing the first cycle at room temperature. These four photographs show that the morphologies of the MCMB samples are nearly identical. However, in the SEM image of Fig. 5a, the MCMB particles appear to be distributed unevenly and the SEI on the MCMB surface is inhomogeneous (Fig. 1a), which implies that the SEI formed using the CC protocol was extremely small. By contrast, in Figs. 5b and 5c, the SEI can be observed clearly and is seen to uniformly cover the MCMB particles, which indicates that the use of the FRDP protocols triggered the electrochemical reduction of the electrolyte and generated the SEI. However, the SEIs formed after using the two FRDP methods covered the MCMB surface to various extents. When the RP effect was absent (FRDP1 battery; Fig. 5b), VGCF and KS6 dispersed onto the parts of the MCMB surface that were covered by the SEI. When the FRDP2 protocol was used and the RP effect was present (Fig. 1b), the SEI formed was uniformly distributed and exhibited large-scale cladding on the MCMB particles (Fig. 5c). Single MCMB particles were connected by the SEI, indicating that Li+ diffusion is enhanced and proceeds readily when the use of a suitable RP effectively reduces the polarization of SEI formation by affecting the reaction dynamics. 3.4. XPS analysis of the SEI Fig. 6a shows the spectra of the C1s on the MCMB surface of the batteries in which the four SEI formation protocols were used; the spectra were collected after completing the first cycle at room temperature. The peak detected in these spectra at approximately 289 to 291 eV was assigned to the carbonate group, whereas the broad peak at approximately 283 to 286 eV was assigned to a group of superimposed peaks, including the C-C (graphite) peak at
586
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
0.16 0.14 R1 )
0.12
!Z"/
0.10
R2 ( )
R3 ( )
W1 ( )
C1 (Coul)
CC
0.073
0.021
0.018
0.033
175
FRDP1
0.073
0.021
0.018
0.024
215
FRDP2
0.072
0.018
0.012
0.021
300
0.08 0.06 0.04 0.02 0.00 0.00
CC FRDP1 FRDP2 0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Z'/! Fig. 4. Impedance Spectra comparisons of the LiCoO2/ MCMB (503759 type) full cell after formation (Inset is the equivalent circuit used to fit experimental curves). Schematic representation of the item of equivalent circuit with four formation protocols has been arranged inside the figure.
284.2 eV, the ROCO2Li peak at 289-to-291 eV, the Li2CO3 peak at 290 eV, the C-O-C peak at 286.2 eV, and the C-H peak at 285.5-to286 eV [3,13,21,22]. The 284-to-285-eV region in Fig. 6a represents the sp2 of the highly conductive (-C-C-) carbon-carbon structure of the MCMB and the conductive carbons of VGCF and KS-6. The intensity of the peaks in the 284-to-285-eV region was the highest when the
FRDP2 protocol was used, which indicates that the MCMB electrode was highly conductive because a thin SEI was formed; this agrees with the result shown in Fig. 5. Furthermore, the intensity of the peaks of FRDP2 protocol in the 289.5-to-290.5-eV region was high and did not shift compared to the CC protocol, indicating that the FRDP2 method reinforces the kinetic route that forwards a two-electron pathway reaction of organic SEI formation
Fig. 5. The SEM micrograph of MCMB electrode after battery formation with the formation protocols of (a) CC, (b) FRDP1 and (d) FRDP2 methods.
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
low-polarization reduction, yielding highly diffusive ionic products (e.g., lithium ethylene dicarbonate) [23]. The use of the FRDP2 protocol resulted in the abundant formation of SEI reduction products in the 289.5-to-290.5-eV region, which indicates that a weak RP effect shown in Fig. 1b generates an unhindered electron flow during SEI formation. A comprehensive survey of the results presented thus far indicates that the SEI formed using the FRDP2 protocol was thin, but it contained high levels of organic compounds that support ionic diffusivity. Furthermore, as shown in Fig. 6a, a clear shoulder in the 285.5to-286.6-eV region was detected only in the case of the FRDP1 battery. Previous studies [1,3] have suggested two reasons for this. Low-molecular-weight electrolyte additives such as maleimides [9–13], vinylene carbonate (VC) [6,7], or sulfones [8] can be electrochemically polymerized; thereby causing C-H bonding of the polymer on the anode surface. In this study, the VC was used to the additive in the electrolyte. In addition, Novak et al. [3] suggests that this shoulder peak corresponds to the ringopening products generated from EC; these products, which can be expressed as the polymer of -(CH2CH2O)n- (polyethylene oxide, PEO). According to above literature, the PEO provides extremely low ionic conductivity and limits the ionic diffusion at the interface. Fig. 6b depicts O1s spectra of the MCMB surface, wherein a multi-component lode with an oxygen atom at 529-to-536 eV is deconvoluted and may include peaks to C&9552;O of ROCO2Li at 532.5 eV, C-OH at 533 eV, Li2CO3 at 532 eV, LiOH at 531.9 eV and CO at 531 eV [3,13]. According to Fig. 6b, the use of the FRDP2 protocol results in agreement with the Fig. 6a that the peak (532-to-532.5 eV) intensity was typically higher than that when other protocols were used. This result demonstrates a highly ionic conductive compound was formed.
10000
(a)
Intensity
8000
CC FRDP1 FRDP2
6000
4000
2000
294
292
290
288
286
284
282
Binding Energy/eV
25000
(b) CC FRDP1 FRDP2
Intensity
20000
15000
10000
5000
538
536
534
532
530
528
587
526
3.5. Cycle-life characteristics at high temperatures
Binding Energy/eV Fig. 6. The (a) C1s and (b) O1s XPS spectrum of the MCMB electrode surface after battery formation with the four formation protocols.
[1]. Our results showed that when the FRDP2 protocol was used, the peak intensity was typically higher than that when other protocols were used. A previous study [1] indicated that when EC was used as the electrolyte solvent, an SEI was readily formed during the electrochemical reduction reaction. A stable twoelectron pathway for the decomposition of EC is favored, due to the
Fig. 7 displays measurements of the cycle life of the batteries in which the SEI was formed using the FRDP and CC protocols at room temperature. At the end of seven cycles, all the batteries exhibited maintainable capacity retention at 60 C. However, after 50 cycles, among the batteries tested, the FRDP2 battery exhibited the highest stability and the lowest cycle fading. The results shown in Fig. 7 agree with the results of XPS analysis shown in Fig. 6 and indicate that the thin SEI that was formed using the FRDP2 protocol effectively connected particles and thus enabled Li+ diffusion, and
1.00
CC FRDP1 FRDP2
-1.5
-1.8
-1
Log (Rct /
Q/Ah !
-1
)
0.99
0.98
-2.1
CC FRDP1 FRDP2
0.97
0
10
-2.4 20
30
40
50
Cycle number Fig. 7. The cycle performance of LiCoO2/ MCMB (503759 type) full cell with the four formation protocols under 1C charge and 1C discharge rate at 60 C measurement.
3.2
3.4
3.6 -1
3.8
-1
1000T /K
Fig. 8. The Arrhenius plot of the reciprocal resistances corresponding to chargetransfer components on MCMB surface formed by three formation protocols.
588
F.-M. Wang et al. / Electrochimica Acta 147 (2014) 582–588
Table 3 The Activation Energy derived for the Rct on MCMB surface formed by four formation protocols. Formation Protocol
CC
FRDP1
FRDP2
Discharge Capacity/mAh
29.48
25.84
17.33
further that the RP used for forming the SEI must be precisely estimated to eliminate the polarization of the electrochemical reaction on the anode surface. In the FRDP1 battery, a slow (0.2C) FP charging was used for SEI formation, with the results indicating that PEO cannot significantly support ionic diffusion on an interface, where the impedance of the interface between the SEI and the anode's surface is high, compared with interfaces that are formed more rapidly (0.5C), i.e. by forward charging [4]. These results indicate that, previously unappreciated FP and RP charging, effectively counter forward charging and disrupt the reactiondynamics and the equilibrium of Li+/electron transfer. 3.6. Charge-transfer activation energies A previous study [24], suggests that the two measurement EIS semicircles, shown in Fig. 4, can be attributed to the diffusion of Li+ through the SEI layer at medium frequencies and to the chargetransfer reaction on the electrode's surface at low frequencies. The charge-transfer component is closely associated with the chemistry of the SEI. A simple, thermally activated process is used for forming the charge-transfer component Rct: 1 ¼ A0 eEa =RT Rct where A0,R, and Ea denote a pre-exponential constant, the standard gas constant, and the activation energy, respectively. The Arrhenius equation can be used to extract the activation energies of Li+ transport during these processes by plotting log(R1) versus reciprocal temperature (1/T plots). Thus, we plotted the reciprocal Rct temperature dependences to derive the activation energies required for Li+ transport across the charge-transfer component (Fig. 8). The slope obtained in the case of each formation protocol was converted to activation energies using the following equation [24]: Ea = -19.144 slope (kJ mol1) Table 3 lists the activation energies of the charge-transfer components that were formed using the four SEI-formation protocols. The results clearly indicate that when we used the FRDP2 protocol, we obtained the smallest activation energy (17.33 kJ mol1) for the bulk SEI/anode surface interface; indicating that the electrochemical reaction proceeded efficiently and exhibited the least amount of polarization. This result agrees with the results of our XPS, EIS, SEM, and cycle-life measurements. 4. Conclusions In this study, the optimized FRDP protocol (FRDP2) is used to balance the dynamic of electron transfer and the ionic diffusion at the initial electrochemical reaction of SEI formation in lithium batteries. However, an appropriate pulse magnitude must be optimized, which is the key that reinforces the kinetic route of SEI formation, promoting the unhindered electron flow to the EC reaction and VC reduction. The performance of two-electrode commercial-type batteries was investigated with the results for the battery made using the FRDP3 SEI-formation protocol demonstrating that the use of RP can lead to an outstanding battery performance with respect to high temperature cycling ability. Compared with the CC method of SEI formation, this FRDP technique should provide greater in the future EV market.
Acknowledgement The author is grateful for the financial support from the Ministry of Science and Technology (MOST) of Taiwan, R.O.C, under grant numbers 102-2221-E-011-016-MY3 and 103-ET-E-011-003ET. Reference [1] S.S. Zhang, A review on electrolyte additives for lithium-ion batteries, J. Power Sources 162 (2006) 1379. [2] A.M. Haregewoin, E.G. Leggesse, J.C. Jiang, F.M. Wang, B.J. Hwang, S.D. Lin, Comparative study on the solid electrolyte interface formation by the reduction of alkyl carbonates in lithium ion battery, Electrochim. Acta 136 (2014) 274. [3] P. Verma, P. Maire, P. Novak, A review of the features and analyses of the solid electrolyte interphase in Li-ion batteries, Electrochim, Acta 55 (2010) 6332. [4] F.M. Wang, H.Y. Wang, M.H. Yu, Y.J. Hsiao, Y. Tsai, Differential pulse effects of solid electrolyte interface formation for improving performance on highpower lithium ion battery, J. Power Sources 196 (2011) 10395. [5] S.S. Zhang, K. Xu, T.R. Jow, EIS study on the formation of solid electrolyte interface in Li-ion battery, Electrochim. Acta 51 (2006) 1636. [6] A.V. Cresce, S.M. Russell, D.R. Baker, K.J. Gaskell, K. Xu, In Situ and Quantitative Characterization of Solid Electrolyte Interphases, Nano Lett. 14 (2014) 1405. [7] I.A. Shkrob, Y. Zhu, T.W. Marin, D. Abraham, Reduction of Carbonate Electrolytes and the Formation of Solid- Electrolyte Interface (SEI) in Lithium-Ion Batteries. 2. Radiolytically Induced Polymerization of Ethylene Carbonate, J. Phys. Chem. C 117 (2013) 19270. [8] R. Wagner, S. Brox, J. Kasnatscheew, D.R. Gallus, M. Amereller, I. Cekic-Laskovic, M. Winter, Vinyl sulfones as SEI-forming additives in propylene carbonate based electrolytes for lithium-ion batteries, Electrochem. Commun. 40 (2014) 80. [9] F.M. Wang, M.H. Yu, C.S. Cheng, S.A. Pradanawati, S.C. Lo, J. Rick, Phenylenedimaleimide positional isomers used as lithium ion battery electrolyte additives: Relating physical and electrochemical characterization to battery performance, J. Power Sources 231 (2013) 18. [10] C.S. Cheng, W.R. Liu, F.M. Wang, A novel ionic host solid electrolyte interface formation on reduced graphene oxide of lithium ion battery, Electrochim. Acta 106 (2013) 425. [11] C.S. Cheng, F.M. Wang, J. Rick, Aqueous Additive for Lithium Ion Batteries: Promotes Novel Solid Electrolyte Interface (SEI) Layer with Overall Cost Reduction, Int. J. Electrochem. Sci. 7 (2012) 8676. [12] C.S. Cheng, F.M. Wang, J. Rick, Breakthrough additive technology for improving the performance of high-power lithium ion batteries, MRS Commun. 2 (2012) 5. [13] F.M. Wang, H.M. Cheng, H.C. Wu, S.Y. Chu, C.S. Cheng, C.R. Yang, Novel SEI formation of maleimide-based additives and its improvement of capability and cyclicability in lithium ion batteries, Electrochim. Acta 54 (2009) 3344. [14] M. S. Ding K. Xu, S. Zhang, T. R. Jow J. L. Allen, Evaluation of Fluorinated Alkyl Phosphates as Flame Retardants in Electrolytes for Li-Ion Batteries I. Physical and Electrochemical Properties, J. Electrochem. Soc. 150 (2003) A161. [15] K. Xu, S. Zhang, J.L. Allen, T.R. Jow, Evaluation of Fluorinated Alkyl Phosphates as Flame Retardants in Electrolytes for Li-Ion Batteries I. Performance in Cell, J. Electrochem. Soc. 150 (2003) A170. [16] H.Y. Wang, F.M. Wang, Electrochemical Investigation of an Artificial Solid Electrolyte Interface for Improving the Cycle-ability of Lithium Ion Batteries Using an Atomic Layer Deposition on a Graphite Electrode, J. Power Sources 233 (2013) 1. [17] A.J. Loebl, C.J. Oldham, C.K. Devine, G.N. Parsons, P.S. Solid electrolyte interphase on lithium-ion carbon nanofiber electrodes by atomic and molecular layer deposition Fedkiw, J. Electrochem. Soc. 160 (2013) A1971. [18] J. Chong, S. Xun, X. Song, G. Liu, V.S. Battaglia, Surface stabilized LiNi0.5Mn1.5O4 cathode materials with high-rate capability and long cycle life for lithium ion batteries, Nano Energy 2 (2013) 283. [19] J.J. Kim, T. Yoon, K.J. Lee, S.M. Oh, Y.E. Sung, The effects of radio frequency sputtering of TiO2 on Li[Li0.07Ni0.38Co0.15Mn0.4]O2 Cathode for Lithium Ion Batteries, J. Nanosci. Nanotechnol. 13 (2013) 7924. [20] M.Z. Mayers, J.W. Kaminski, T.F. Miller, Suppression of Dendrite Formation via Pulse Charging in Rechargeable Lithium Metal Batteries, J. Phys. Chem. C 116 (2012) 26214. [21] F.M. Wang, M.H. Yu, Y.J. Hsiao, Y. Tsai, B.J. Hwang, Y.Y. Wang, C.C. Wan, Aging Effects to Solid Electrolyte Interface (SEI) Membrane Formation and the Performance Analysis of Lithium Ion Batteries, Int. J. Electrochem. Sci. 6 (2011) 1014. [22] S.A. Pradanawati, F.M. Wang, J. Rick, In Situ formation of pentafluorophosphate benzimidazole anion stabilizes high-temperature performance of lithium-ion batteries, Electrochim. Acta 135 (2014) 388. [23] M. Nie, D. Chalasani, D.P. Abraham, Y. Chen, A. Bose, B.L. Lucht, Lithium Ion Battery Graphite Solid Electrolyte Interphase Revealed by Microscopy and Spectroscopy, J. Phys. Chem. C 117 (2013) 1257. [24] K. Xu, Y. Lam, S.S. Zhang, T.R. Jow, T.B. Curtis, J. Phys. Chem. C 111 (2007) 7411.