- Email: [email protected]
Performance tuning of lithium ion battery cells with area-oversized graphite based negative electrodes
Journal of Power Sources 396 (2018) 519–526
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Performance tuning of lithium ion battery cells with area-oversized graphite based negative electrodes
T
Tim Daggera,b, Johannes Kasnatscheewc,∗, Britta Vortmann-Westhovena, Timo Schwietersa,b, Sascha Nowaka, Martin Wintera,b,c, Falko M. Schappachera,∗∗ a
MEET Battery Research Center, Westfälische Wilhelms-Universität Münster, Corrensstr. 46, D-48149, Münster, Germany Institute of Physical Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstr. 28/30, D-48149, Münster, Germany c Helmholtz Institute Münster, IEK-12, Research Center Jülich GmbH, Corrensstr. 46, D-48149, Münster, Germany b
H I GH L IG H T S
of negative electrodes increases specific capacity losses. • Area-oversizing loss in overhang area (negative electrode) due to trapped active lithium. • Capacity trapped, but thermodynamically active lithium can be reactivated. • Kinetically in charged state irreversibly increases losses of active lithium. • Duration • Cycle life can be electrochemically improved by periodic discharge modifications.
A R T I C LE I N FO
A B S T R A C T
Keywords: Variation of area ratio of negative and positive electrode Electrode alignment vs. performance Reversible vs. irreversible specific capacity losses Strategy for cycle life extension
The accuracy for positional alignment of the positive electrode vs. the negative electrode is of great importance for the quality of assembly of lithium ion cells. Area-oversized negative electrodes increase the tolerance for electrode alignment. In this study, the impact of area-oversizing of the negative electrode on the specific capacity losses during charge/discharge cycling is systematically investigated by using electrochemical and analytical methodologies. It is shown, that with a higher degree of area-oversizing more active lithium is kinetically trapped in the outer negative electrode areas (“overhang”), causing performance-deteriorating losses in usable specific capacity. Nevertheless, most of this “lost” specific capacity is of reversible nature as the trapped active lithium can be electrochemically recovered, which is analytically proven by inductively coupled plasma-optical emission spectrometry (ICP-OES) and laser ablation-inductive coupled plasma-mass spectrometry (LA-ICP-MS). Given this relation, a periodic application of a short constant voltage step after discharge results in a significant performance increase. In contrast, holding the cell in the charged state is detrimental for cells with area oversized negative electrodes as the amount of reversible and irreversible trapped active lithium increases. Based on the obtained insights, the influence of variations of the electrochemical conditions on charge/discharge cycling performance is discussed.
1. Introduction The steadily increasing demand for alternative energy sources and electro-mobility necessitates adequate electrochemical energy storage devices [1–7]. Among different systems, batteries, particularly the lithium ion batteries (LIBs) are the state-of-the-art (SOTA) benchmark technology [8–10]. In view of commercial claims, issues with respect to specific energy, energy efficiency, cycle life, safety and even recycling still require further research and development efforts [11–13], which
∗
remains challenging given the complexity of a LIB cell during operation [14–17]. The charge/discharge cycling performance of a LIB cell sensitively depends on a wide range of parameters regarding not only the active and inactive materials (including their composition and balancing aspects) [18–25], but also the design of cell geometry [26,27,45]. An example for the influence of the latter is a cell with geometrically oversized area of the negative electrode relative to the area of the positive electrode. This design has practical benefits as it minimizes the
Corresponding author. Corresponding author. E-mail addresses: [email protected] (J. Kasnatscheew), [email protected] (F.M. Schappacher).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.06.043 Received 13 March 2018; Received in revised form 4 June 2018; Accepted 10 June 2018 0378-7753/ © 2018 Elsevier B.V. All rights reserved.
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
risk of lithium plating at the edges of graphite electrode sheets during charge [28]. However, a deteriorating relation between the performance and the oversized area of the negative electrode is reported by Son et al. [29]. It is claimed that the outer areas (overhang) of the negative electrode additionally utilize active lithium for the formation of the solid electrolyte interphase (SEI) [30–32], resulting in increased irreversible specific capacity losses. However, these conclusions made on the interpretations of Coulombic efficiencies (CEs) did not consider possible kinetic origins, which can have a significant impact on the observed phenomena, as well [33]. In fact, recent reports point to an even reversible character of the apparently “lost” specific capacity by revealing changes in the lithium amount in the negative electrode overhang depending on the state of charge and discharge [34–37]. In this work, the reversible and irreversible specific capacity losses are thoroughly investigated by applying electrochemical and analytical measures for varying degrees of the overhang. Given the practical importance of the overhang usage [27,38,39], the goal of this work is not only to clarify but also to circumvent the origin of the accompanied losses in specific capacity.
Table 1 Charge/discharge cycling procedure includes five constant current (cc) formation cycles, a self-discharge experiment and subsequent constant current/constant voltage (cc/cv) charge/cc discharge cycling. The criterion of 0.05C for cv steps implies a cut-off when the detected current during the cv step drops below the threshold value of 0.05C. Cycle
Mode
Charge
Discharge
1 2–5 6 7 8–19 20 21- …
Formation Formation Cycling Cycling + SD Cycling Cycling + dcv Cycling
cc: 0.05C cc: 0.1C cc/cv: 1C/0.05C cc/cv: 1C/0.05C 120 h rest cc/cv: 1C/0.05C cc/cv: 1C/0.05C cc/cv: 1C/0.05C
cc: 0.1C cc: 0.1C cc: 1C cc: 1C cc: 1C cc/cv: 1C/0.05C cc: 1C
cc: constant current. cv: constant voltage. 1C: 1C corresponds to charge or discharge rate of 142 mAh∙g−1. SD: A rest step of 120 h is applied after charge. dcv: constant voltage step during discharge.
discharge mode. The discharge of the 20th cycle comprised a low rate (0.05C) constant voltage step during discharge (dcv). LIB pouch cells were cycled between 4.2 V and 3.0 V. Four different charge/discharge cycling procedures were used to investigate long term cycling according Table 2. The standard procedure contains 5 formation cycles and 494 cycles of cc/cv charge with subsequent cc discharge at 2C. After the 20th, 120th, 220th, 320th, and 420th charge and discharge was a rest step of 50 min, respectively. The second procedure includes a SD experiment (120 rest step after conventional cc/cv charge) after the 7th, 107th, 207th, 307th and 407th charge. The third procedure includes a dcv step to the standard procedure after the 20th, 120th, 220th, 320th, and 420th discharge. The fourth procedure includes both, SD and dcv experiments. For analytical investigations, cells with 78% oversizing of the negative electrode were investigated (Fig. 1 (c)). Two cells were cycled with the procedure depicted in Table 1 till the 6th, 7th or 20th cycle, respectively. In addition to the discharged cells, two uncycled cells were transferred into the glove box (water and oxygen content of less than 0.2 ppm), disassembled and electrodes were washed with DMC. For Inductively coupled plasma-optical emission spectrometry (ICPOES) measurements, six 4 mm Ø discs were punched out of the negative electrode overhang for each cell. ICP-OES measurements were performed using a Spectro ARCOS ICP-OES (Spectro Analytical Instruments) instrument with axial plasma viewing. A standard Fassel type torch (Spectro Analytical Instruments) was employed. For sample introduction, the system's peristaltic pump with a cross flow nebulizer and a double-pass spray chamber (Scott type) was used. Element emission was detected at different individual emission lines simultaneously. Operating conditions for ICP-OES measurements are shown in Table 3 [42]. This method was used to determine the lithium content of the discharged electrode according the literature [42]. For Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), 20 mm × 20 mm rectangles were punched out of a corner of the 40 mm × 40 mm negative electrodes. The 7Li signal (cps = counts per second) was recorded using 184 transient signals per
2. Experimental Single side coated electrodes for the LIB pouch cells were in house made, based on mesocarbon microbeads (MCMB, Osaka) and LiNi1/ 3Mn1/3Co1/3O2 (NMC111, Cellcore) active materials for negative and positive electrodes, respectively. The electrodes were locally separated by an electrolyte soaked Freudenberg 2190 separator. The electrolyte was based on 1 M LiPF6 in ethylene carbonate (EC):dimethyl carbonate (DMC) 1:1 wt% (LP30, BASF). The electrolyte amount, which can impact the performance upon variation, was excessed in all cells [40]. The LIB pouch cells were assembled in the dry room (dew point of at least −60 °C). The determined practical delithiation capacitiy of the MCMB graphite based negative electrodes was 296 ± 10 mAh∙g−1 and of the NMC111 based positive electrodes was 142 ± 2 mAh∙g−1, when cycled at 0.1 C in three-electrode lithium half cells in the range of 0.02 V and 1.5 V vs. Li/Li+ and 4.2 V and 2.5 V vs. Li/Li+, respectively. Providing an active material mass loading of 9.2 mg∙cm−2 and 14.2 mg∙cm−2, respectively, the negative electrode was oversized in capacity in a ratio of 1.3:1. Different negative to positive electrode area ratios were studied. Whereas the positive electrode had the area size of 30 mm × 30 mm, the negative electrode area sizes were varied from 30 mm × 30 mm–33 mm × 33 mm, 40 mm × 40 mm, and 55 mm × 55 mm, respectively (Fig. 1(a)–(d)). Similar mass loading of the overhang (as apparently “inactive” area) was chosen. (With increasing overhang, the amount of “inactive” materials increases, increasing the overall mass and volume of the cell, finally decreasing energy density/specific energy) [41]. The full cells were cycled according to the electrochemical conditions in Table 1. Five constant current (cc) formation cycles (as default cycles for reasons of SEI optimization) were followed by one constant current/constant voltage (cc/cv) charge, cc discharge acclimation cycle at 1C/0.05C (6th cycle). The 7th cycle included a self-discharge (SD) experiment (i.e., a rest step of 120 h) with subsequent cc discharge at 1C. After the SD experiment, cells were cycled in cc/cv charge, cc
Fig. 1. Cells with varying oversized negative electrode area: (a) 0%, (b) 21%, (c) 78% and (d) 236%. 520
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
[43]. Scans were performed using a laser energy of 2 J∙cm−1, 110 μm circular spots, a shot frequency of 10 Hz and a lateral scanning speed of 110 μm∙s−1. Operating conditions for ICP-MS measurements are shown in Table 4.
Table 2 Charge/discharge cycling conditions for four procedures, a standard procedure (without SD and dcv experiments), a procedure with SD experiments, a procedure with dcv experiments, and a procedure including both, SD and dcv experiments. The LIB pouch cells were cycled within the voltage range of 4.2–3.0 V. Cycle
Mode
Charge
Discharge
1 2–5 6 7
Formation Formation Cycling Cycling (+SD)
cc: cc: cc: cc:
8–19 20
Cycling Cycling (+dcv)
21–106 107
Cycling Cycling (+SD)
108–119 120
Cycling Cycling (+dcv)
121–206 207
Cycling Cycling (+SD)
208–219 220
Cycling Cycling (+dcv)
221–306 307
Cycling Cycling (+SD)
308–319 320
Cycling Cycling (+dcv)
321–406 407
Cycling Cycling (+SD)
408–419 420
Cycling Cycling (+dcv)
421–499
Cycling
cc: 0.05C cc: 0.1C cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C
3. Results and discussion 3.1. Effect of area-oversizing of the negative electrode on specific capacities
0.1C 0.1C 2C 2C
Different degrees of area-oversizing of negative electrodes are realized to investigate the influence of oversize on the charge/discharge cycling performance. Fig. 2 depicts conventional charge/discharge cycling data during the initial five formation cycles at lower rates (0.1C) and a subsequent cycle at an elevated rate (1C). The first cycle Coulombic efficiency (FCE) declines with increasing negative electrode area. Cells with no area-oversizing of the negative electrode (0%) reveal best FCEs of 86.10 ± 0.03% (Fig. 2(a)). An oversizing of the negative electrode of 21%, 78% and 236% (Fig. 2(a)–(d) results in a steady decrease of the FCE with values of 84.22 ± 0.02%, 77.00 ± 0.01% and 59.6 ± 0.7%, respectively. Following this trend, discharge capacities, CEs and capacity retentions of the subsequent formation cycles also become worse with increasing the area-oversizing of negative electrode. This relation of poorer performance with increased area-oversizing of the negative electrode is well in accordance with the findings of Son et al. [29]. However, the application of a SD experiment after the 7th charge (high duration in charged state) reveals controversy findings.
cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C
3.2. The application of a self-discharge experiment for cells with areaoversized negative electrodes As depicted in Fig. 3, the capacity loss during the SD experiment (7th cycle) increases with the degree of the area-oversize of the negative electrode. Whereas cells with zero area-oversize exhibit a capacity loss of only 3.2 ± 0.2 mAh∙g−1, cells with an negative electrode area-oversizing of 21% show a capacity loss of already 10.6 ± 0.1 mAh∙g−1 (Fig. 3(a) and (b)). The capacity losses further increase for area-oversized negative electrodes of 78% and 236% to 19.2 ± 0.1 mAh∙g−1 and 41.8 ± 0.5 mAh∙g−1, respectively (Fig. 3(c) and (d)). Immediately after the SD experiment (7th cycle), a salient continuous recovery of the capacity during upcoming cycles is observable for cells with area-oversized negative electrode. Within this recovery, the discharge capacities exceed the charge capacities (implying Coulombic efficiencies > 100%), pointing to a kinetic impairment in the previous cycle that is resolved this cycle, which is in accordance with recent reports [35–37]. Cells with an oversized negative electrode of 21% show a capacity recovery of 3.15 ± 0.07 mAh∙g−1, whereas cells with 78% and 236% negative electrode area-oversizing recovered capacities of 2.8 ± 0.4 mAh∙g−1 and 4.9 ± 0.5 mAh∙g−1 from the 7th to the 19th discharge. To maximize the recovery of – as stated above – only kinetically “lost” specific capacity, a dcv step is applied (20th cycle, Fig. 2) to enable a complete discharge by excluding kinetic limitations [44]. The amount of specific capacity recovery after the dcv step correlates to the area-oversizing degree of the negative electrode. It can be concluded that the area-oversizing leads to additional specific capacity losses, as the active lithium is preferably kinetically trapped in the overhang.
cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C
cc: constant current. cv: constant voltage. 1C: 1C corresponds to charge or discharge rate of 142 mAh∙g−1. SD: A rest step of 120 h is applied after charge. dcv: constant voltage step during discharge. Table 3 ICP-OES system parameters. ICP-OES system parameter
Operating conditions
Radio frequency power Total Argon amount Sample carrier gas Outer plasma gas Torch injector tube (inside diameter)
1400 W 13.65 L∙min−1 0.80 L∙min−1 12.0 L∙min−1 1.8 mm Ø
Table 4 ICP-MS system parameters. ICP-MS system parameter
Operating conditions
Radio frequency power Radio frequency matching Argon carrier gas flow Integration time per mass Sampling depth
1550 W 1.30 V 950 mL∙min−1 0.03 s 6.0 mm
3.3. Mechanism of the capacity recovery after self-discharge experiments The observed relation between area-oversizing of the negative electrode, specific capacity losses and the recovery behavior of the specific capacity points to a mechanism, which is schematically depicted in Fig. 4. It can be deduced, that the intentionally inactive overhang of the negative electrode (black frame in Fig. 4(a) is, by no
sample. A 193 nm ArF Excimer Laser (Analyte Excite Excimer LASystem, Teledyne Cetac) coupled to an ICP-MS System (7700 Series, Agilent Technologies) with helium carrier gas (99.999%, Westfalen AG) with a flow of 900 mL∙min−1 as described in literature were utilized 521
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 2. Charge/discharge cycling performance of LIB cells with different area-oversizing of the negative electrode during initial five cycles at a lower rate (0.1C) and one subsequent cycle at a higher rate (1C). Area-oversizing of the negative electrode amounts to (a) 0%, (b) 21%, (c) 78% and (d) 236%.
(Fig. 4(c)), the graphite based negative electrode gets delithiated preferably in the area where positive electrode and negative electrode are overlapping (where the electrical field is stronger) resulting in different lithiation degrees (SOC) within the negative electrode, that is lower in the overlap area and higher in the overhang area. The lithium ions in the overhang (where the effective electric field is weaker) are hindered
means, inactive, but rather active involving lithium during charge/ discharge cycling. In the SD experiment (7th cycle), which is an extreme condition (long duration in the fully charged state) this gets more evident. Here, in the charged state lasting for 120 h, the lithium ions can equally distribute within the entire negative electrode area including the overhang (Fig. 4(b)). During the following discharge step
Fig. 3. Self-discharge experiments during the 7th charge followed by charge/discharge cycling at 1C. Area-oversizing of the negative electrode amounts to (a) 0%, (b) 21%, (c) 78% and (d) 236%. 522
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 4. (a) Scheme of the area-oversized negative electrode (black) in comparison to the positive electrode (grey) and (b) the lithium distribution of the negative electrode at the end of a charge cv step in the 7th cycle, that is after the 120 h rest step; (c) discharged state after the SD experiment in the 7th cycle and (d) after discharged state, in the upcoming cycles.
in a kinetic manner, resulting in a slowed down mitigation of the ions from the outer electrode area towards the inner area, finally causing incomplete discharge when reaching the discharge cut-off condition, which results in higher specific capacity loss. In the following cycles, the trapped lithium in the overhang further equilibrates over the electrode area, thus also including the movement towards the inner active area, where the lithium gets active again (Fig. 4(d)). Owing to higher movement towards the inner area from the overhang compared to the opposite, a steady increase in the discharge capacity can be observed (Fig. 3). The application of the dcv step (20th cycle, Fig. 3) obviously facilitates the reactivation of the kinetically trapped active lithium.
Table 5 Lithium content of the overhang (78%) of the negative electrode in an uncycled state, before the SD experiment (6th discharge), after the SD experiment (7th discharge) and after the dcv step (20th discharge). Negative electrode overhang
Li-content/mg
uncycled after the 6th discharge after the 7th discharge after the 20th discharge
0.81 0.98 0.74
experiment); (c) of a discharged electrode after the 7th cycle (after the SD experiment) and (d) of the discharged electrode after the 20th cycle (after dcv). All samples were washed with DMC to minimize the amount of lithium residues (e.g. from the electrolyte) on the electrode, which is shown for the uncycled cell (Fig. 5(a)). As reported in literature, in MCMB/NMC full cells, the active lithium remains in the MCMB anode even in the discharged state, particularly at the beginning of charge/ discharge cycling [18]. This is also observable in the discharged state after the 6th cycle, where the negative electrode provides a uniform lithium distribution in discharge state (Fig. 5(b)), including the overhang. Thus, already at the beginning of charge/discharge cycling the overhang includes lithium, which is not practically utilized. It can be deduced that by increasing the negative electrode overhang, more lithium is “trapped” in the overhang and the practically usable specific capacity gets diminished (in line with the results in Fig. 2). In the discharged state, after the SD experiment (7th cycle), the overhang has a higher lithium concentration compared to the active (overlapping) area of the electrodes, supporting the validity of the aforementioned mechanism and the ICP-OES results (Fig. 4 and Table 5). Trapped lithium
3.4. Specific capacity recovery confirmed by ICP-OES and LA-ICP-MS analyses To support the above data interpretation pointing at kinetically trapped lithium in the overhang of the negative electrode, particularly after the SD experiments [36,37], ICP-OES measurements of the overhang are carried out to determine the lithium contents. Samples are taken from the overhang (for analytical investigation an overhang of 78% is used) of the negative electrodes in the uncycled state and after the 6th cycle, the 7th cycle and the 20th cycle in the discharged state. The alternating lithium contents presented in Table 5 support the aforementioned mechanism (Fig. 4). The lithium content in the overhang increases after SD application (0.81 mg–0.98 mg) and decreases after the dcv step (0.98 mg–0.74 mg), supporting the reversible activity of the overhang. For reasons of verification, LA-ICP-MS measurements are performed, providing the additional benefit to visualize the entire electrode (overlap and overhang area). The lower left corner of the electrodes showing the lithium concentration distribution is shown in Fig. 5: (a) of an uncycled electrode; (b) of a discharged electrode after the 6th cycle (before the SD 523
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 5. Lithium distribution in the lower left corner of the negative electrode in (a) an uncycled state and in the discharged state after (b) the 6th cycle (before SD experiment), (c) the 7th cycle (after SD experiment) and (d) the 20th cycle (after dcv experiment). The black line highlights the boundary of the overhang and the overlap area.
resistances remain unchanged (Fig. 6(b)), the beneficial effect of the repeated application of dcv steps can be reasonably attributed to an increased amount of used active lithium (= increasing specific capacity). On the basis of these insights the dcv step likely counteracts the accumulation of kinetically trapped lithium in the overhang. Application of a periodic SD step however deteriorates the performance of the charge/discharge cycling as seen by the lower specific capacity (Fig. 6(a), green) and increased resistance (Fig. 6(b), green). Given the shown relation of the SD step and overhang of the negative electrode, accumulation of trapped active lithium in the overhang can be assumed (decreasing specific capacity), and moreover its partly irreversible depletion for the SEI formation (increasing internal resistance). Nevertheless, the combination of both, dcv steps and SD steps is obviously even worse for long-term charge/discharge cycling compared to application of solely SD steps, as seen by the observed SOH values of 82.91% and 85.38%, respectively. While the SD step leads to an accumulation of active lithium in the overhang and the dcv step to its removal, this test procedure likely leads to a high flow in/out the overhang area. In consequence this facilitates SEI formation of the overhang which finally involves irreversible active lithium. In summary, long durations in the charged state (as it is the case during SD applications) should be avoided with respect to achieve good performance during long term charge/discharge cycling, when the negative electrode is area-oversized (i.e., in typically balanced cells). The responsible accumulation of active lithium during charge/discharge cycling in the overhang is schematically shown in Fig. 7 for both test procedures including SD steps. In case long durations in charged state can be avoided, the performance can be significantly maximized when periodic dcv steps (controlled over-discharge) [23] are applied. In this case, the partly accumulated active lithium in the overhang is mostly of reversible nature and can be re-activated as schematically illustrated in Fig. 7.
in the overhang is not fully usable resulting in decreased subsequent specific capacities. Furthermore, due to the voltage drop, that evolves during the SD experiment, the cell reaches its discharge cut-off earlier (see SI 1), i.e., at the expense of the discharge capacity. This implies that the electrode is less delithiated after the 7th cycle (SD experiment) compared to the cell discharged without previous SD experiment in the 6th cycle. According to the aforementioned mechanism, the application of the dcv step (20th cycle) leads to an almost complete removal of the lithium in the overhang. Therefore, the electrode shown in Fig. 5(d) shows the least lithium in the remaining in the overhang. Based on the results one could conclude that a dcv step increases the electrochemical performance of the cell, as trapped lithium can be easily reactivated. (Additional insight: when conventional cycling conditions are used, the outer area remains predominately inactive, hence the mass balancing between the active areas of cathode and anode is unaffected by additional mass of outer area.) For better verification, long-term effects of the dcv and SD step on the performance are investigated in the following. 3.5. Long-term effects of repetitive self-discharge (SD) and discharge constant voltage (dcv) steps As displayed in Fig. 6(a) LIB cells that are cycled at conventional electrochemical conditions (without SD and dcv steps; according to Table 2), reveal a characteristic decay in specific capacity for the given full cell system and conditions. After 499 cycles, the cells exhibit a specific discharge capacity of 112.3 ± 0.3 mAh∙g−1, which corresponds to a state of health (SOH) of 86.39% (compared to the 6th cycle with 130.0 ± 0.1 mAh∙g−1). After 499 cycles, the cells cycled according to the standard procedure, provided a raise in resistance of about 1.5 Ω as depicted in Fig. 6(b). In case the dcv step is applied every 100 cycles (Fig. 6(a), orange), the charge/discharge cycling performance is significantly improved up to a SOH of 90.22% after 499 cycles. Given the fact that the internal 524
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 6. (a) Charge/discharge cycling results of cells cycled at different procedures (Table 2) having 21% overhang. A standard procedure (black) including only cc/cv charge and cc discharge (black). A standard procedure plus a dcv step after 20th, 120th, 220th, 320th and 420th discharge (orange). A standard procedure plus SD step in cycles 7, 107, 207, 307 and 407 (green). A standard procedure including both, dcv and SD steps (blue). (b) Relative comparison of the growth in internal resistance during charge/discharge cycling. All graphs show average values of two cells and are normalized to their 6th cycle. (Unedited internal resistances with standard deviations are depicted in Figure SI 2.). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
4. Conclusion
and laser ablation-inductive coupled plasma-mass spectrometry (LAICP-MS) confirmed not only its activity, but also its partly reversible nature. Relevant insights into long-term charge/discharge cycling performance could be deduced. Already a periodic application of a short dcv step after several cycles significantly improved the performance by reactivating the amount of trapped active lithium in the overhang area. As revealed by self-discharge (SD) experiments, long durations of a LIB cell in the charged state should be avoided, as they accelerate electrolyte decomposition, which enhanced irreversibly utilized active lithium and increased cell resistance. In addition, for such anode to cathode area ratios, the enlarged amount of trapped active lithium in the overhang area increased its irreversibility (e.g. by immobilization in SEI film formation) and in turn additionally led to increased internal resistances.
In this work, the influence of an area-oversized negative electrode on the charge/discharge cycling performance was thoroughly investigated. Such cell construction is beneficial as it assured a complete overlap of the electrodes. However, the apparently inactive (but in practice active) overhang of the negative electrode significantly influenced the charge/discharge cycling performance. A systematic increase of the overhang area resulted in increased specific capacity losses, particularly for a longer duration in charged state. However, when enhancing the discharge kinetics by application of a discharge constant voltage (dcv) step, a major part of the apparently irreversible specific capacity losses could be recovered, pointing to their mostly reversible nature. The monitoring of the lithium content of the overhang area by inductively coupled plasma-optical emission spectrometry (ICP-OES)
Fig. 7. Schematic illustration of the quantitative distribution of lithium in the area oversized negative electrodes during charge/discharge cycling for the four different test procedures depicted in Fig. 6(a). The cycling procedure with the absence of SD and dcv steps (Standard) leads to a continuous accumulation of lithium in the overhang, which are mostly of reversible nature, thus can be reactivated by using the procedure with repetitive dcv steps (+dcv). In contrast, application of the procedure with SD steps (+SD) leads to a severe accumulation of lithium in the overhang having an increased irreversible amount, thus cannot be re-activated by using the procedure including dcv steps (+SD & dcv). 525
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Acknowledgement [20]
The authors would like to express their thanks to MEET Battery Research Center at the University Muenster. Daniel Brüggemann and Marco Fritzen are acknowledged for cell assembly works. Pascal Noll and Nils Wallus (MEET Battery Research Center) are acknowledged for providing balanced electrodes.
[21] [22] [23] [24]
Appendix A. Supplementary data
[25]
Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.jpowsour.2018.06.043.
[26] [27]
References [28] [29]
[1] R. Wagner, N. Preschitschek, S. Passerini, J. Leker, M. Winter, J. Appl. Electrochem. (2013) 1–16. [2] O. Gröger, H.A. Gasteiger, J.-P. Suchsland, J. Electrochem. Soc. 162 (2015) A2605–A2622. [3] R.W. Schmitz, P. Murmann, R. Schmitz, R. Müller, L. Krämer, J. Kasnatscheew, P. Isken, P. Niehoff, S. Nowak, G.-V. Röschenthaler, Prog. Solid State Chem. 42 (2014) 65–84. [4] D. Andre, S.-J. Kim, P. Lamp, S.F. Lux, F. Maglia, O. Paschos, B. Stiaszny, J. Mater. Chem. 3 (2015) 6709–6732. [5] K.G. Gallagher, S. Goebel, T. Greszler, M. Mathias, W. Oelerich, D. Eroglu, V. Srinivasan, Energy Environ. Sci. 7 (2014) 1555–1563. [6] T. Placke, R. Kloepsch, S. Dühnen, M. Winter, J. Solid State Electrochem. (2017) 1–26. [7] P. Meister, O. Fromm, S. Rothermel, J. Kasnatscheew, M. Winter, T. Placke, Electrochim. Acta 228 (2017) 18–27. [8] M. Winter, R.J. Brodd, Chem. Rev. 104 (2004) 4245–4269. [9] N. Nitta, F. Wu, J.T. Lee, G. Yushin, Mater. Today 18 (2015) 252–264. [10] E.J. Berg, C. Villevieille, D. Streich, S. Trabesinger, P. Novák, J. Electrochem. Soc. 162 (2015) A2468–A2475. [11] J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, S. Nowak, I. Cekic Laskovic, M. Winter, J. Power Sources 359 (2017) 458–467. [12] P. Meister, H. Jia, J. Li, R. Kloepsch, M. Winter, T. Placke, Chem. Mater. 28 (2016) 7203–7217. [13] S. Rothermel, M. Evertz, J. Kasnatscheew, X. Qi, M. Grützke, M. Winter, S. Nowak, Chemsuschem 9 (2016) 3473–3484. [14] M. Evertz, F. Horsthemke, J. Kasnatscheew, M. Börner, M. Winter, S. Nowak, J. Power Sources 329 (2016) 364–371. [15] J.B. Goodenough, K.-S. Park, J. Am. Chem. Soc. 135 (2013) 1167–1176. [16] W. Li, B. Song, A. Manthiram, Chem. Soc. Rev. 46 (2017) 3006–3059. [17] J. Kasnatscheew, B. Streipert, S. Röser, R. Wagner, I. Cekic Laskovic, M. Winter, Phys. Chem. Chem. Phys. 19 (2017) 16078–16086. [18] J. Kasnatscheew, T. Placke, B. Streipert, S. Rothermel, R. Wagner, P. Meister, I. Cekic-Laskovic, M. Winter, J. Electrochem. Soc. 164 (2017) A2479–A2486. [19] J. Kasnatscheew, U. Rodehorst, B. Streipert, S. Wiemers-Meyer, R. Jakelski,
[30] [31] [32] [33]
[34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45]
526
R. Wagner, I. Cekic-Laskovic, M. Winter, J. Electrochem. Soc. 163 (2016) A2943–A2950. J. Kasnatscheew, M. Evertz, R. Kloepsch, B. Streipert, R. Wagner, I. Cekic Laskovic, M. Winter, Energy Technol. 5 (2017) 1670–1679. J. Kasnatscheew, R.W. Schmitz, R. Wagner, M. Winter, R. Schmitz, J. Electrochem. Soc. 160 (2013) A1369–A1374. R. Wagner, S. Brox, J. Kasnatscheew, D.R. Gallus, M. Amereller, I. Celdc-Laskovic, M. Winter, Electrochem. Commun. 40 (2014) 80–83. J. Kasnatscheew, M. Börner, B. Streipert, P. Meister, R. Wagner, I. Cekic Laskovic, M. Winter, J. Power Sources 362 (2017) 278–282. T. Dagger, M. Grützke, M. Reichert, J. Haetge, S. Nowak, M. Winter, F.M. Schappacher, J. Power Sources 372 (2017) 276–285. B. Streipert, S. Roser, J. Kasnatscheew, P. Janssen, X. Cao, R. Wagner, I. CekicLaskovic, M. Winter, J. Electrochem. Soc. 164 (2017) A1474–A1479. P.J. Osswald, S.V. Erhard, J. Wilhelm, H.E. Hoster, A. Jossen, J. Electrochem. Soc. 162 (2015) A2099–A2105. B.R. Long, S.G. Rinaldo, K.G. Gallagher, D.W. Dees, S.E. Trask, B.J. Polzin, A.N. Jansen, D.P. Abraham, I. Bloom, J. Bareño, J. Electrochem. Soc. 163 (2016) A2999–A3009. M. Tang, P. Albertus, J. Newman, J. Electrochem. Soc. 156 (2009) A390–A399. B. Son, M.-H. Ryou, J. Choi, S.-H. Kim, J.M. Ko, Y.M. Lee, J. Power Sources 243 (2013) 641–647. C. Korepp, W. Kern, E. Lanzer, P. Raimann, J. Besenhard, M. Yang, K.-C. Möller, D.T. Shieh, M. Winter, J. Power Sources 174 (2007) 637–642. M. Winter, Z. Phys. Chem. Int. J. Res. Phys. Chem. Chem. Phys. 223 (2009) 1395–1406. G.H. Wrodnigg, J.O. Besenhard, M. Winter, J. Electrochem. Soc. 146 (1999) 470–472. J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, R. Klöpsch, B. Vortmann, H. Hahn, S. Nowak, M. Amereller, A.-C. Gentschev, P. Lamp, M. Winter, Phys. Chem. Chem. Phys. 18 (2016) 3956–3965. M. Lewerenz, A. Warnecke, D.U. Sauer, J. Power Sources 354 (2017) 157–166. M. Lewerenz, J. Münnix, J. Schmalstieg, S. Käbitz, M. Knips, D.U. Sauer, J. Power Sources 345 (2017) 254–263. B. Gyenes, D.A. Stevens, V.L. Chevrier, J.R. Dahn, J. Electrochem. Soc. 162 (2015) A278–A283. J. Wilhelm, S. Seidlmayer, P. Keil, J. Schuster, A. Kriele, R. Gilles, A. Jossen, J. Power Sources 365 (2017) 327–338. E. Schuster, C. Ziebert, A. Melcher, M. Rohde, H.J. Seifert, J. Power Sources 286 (2015) 580–589. J. Kurfer, M. Westermeier, C. Tammer, G. Reinhart, CIRP Ann. - Manuf. Technol. 61 (2012) 1–4. S.J. An, J. Li, D. Mohanty, C. Daniel, B.J. Polzin, J.R. Croy, S.E. Trask, D.L. Wood, J. Electrochem. Soc. 164 (2017) A1195–A1202. J. Kasnatscheew, R. Wagner, M. Winter, I. Cekic-Laskovic, Top. Curr. Chem. 376 (2018) 16. B. Vortmann-Westhoven, M. Winter, S. Nowak, J. Power Sources 346 (2017) 63–70. T. Schwieters, M. Evertz, A. Fengler, M. Börner, T. Dagger, Y. Stenzel, P. Harte, M. Winter, S. Nowak, J. Power Sources 380 (2018) 194–201. J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, S. Nowak, I. Cekic-Laskovic, M. Winter, J. Phys. Chem. C 121 (2017) 1521–1529. S.J. An, J. Li, C. Daniel, S. Kalnaus, D.L. Wood, J. Electrochem. Soc. 164 (2017) A1755–A1764.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Performance tuning of lithium ion battery cells with area-oversized graphite based negative electrodes
T
Tim Daggera,b, Johannes Kasnatscheewc,∗, Britta Vortmann-Westhovena, Timo Schwietersa,b, Sascha Nowaka, Martin Wintera,b,c, Falko M. Schappachera,∗∗ a
MEET Battery Research Center, Westfälische Wilhelms-Universität Münster, Corrensstr. 46, D-48149, Münster, Germany Institute of Physical Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstr. 28/30, D-48149, Münster, Germany c Helmholtz Institute Münster, IEK-12, Research Center Jülich GmbH, Corrensstr. 46, D-48149, Münster, Germany b
H I GH L IG H T S
of negative electrodes increases specific capacity losses. • Area-oversizing loss in overhang area (negative electrode) due to trapped active lithium. • Capacity trapped, but thermodynamically active lithium can be reactivated. • Kinetically in charged state irreversibly increases losses of active lithium. • Duration • Cycle life can be electrochemically improved by periodic discharge modifications.
A R T I C LE I N FO
A B S T R A C T
Keywords: Variation of area ratio of negative and positive electrode Electrode alignment vs. performance Reversible vs. irreversible specific capacity losses Strategy for cycle life extension
The accuracy for positional alignment of the positive electrode vs. the negative electrode is of great importance for the quality of assembly of lithium ion cells. Area-oversized negative electrodes increase the tolerance for electrode alignment. In this study, the impact of area-oversizing of the negative electrode on the specific capacity losses during charge/discharge cycling is systematically investigated by using electrochemical and analytical methodologies. It is shown, that with a higher degree of area-oversizing more active lithium is kinetically trapped in the outer negative electrode areas (“overhang”), causing performance-deteriorating losses in usable specific capacity. Nevertheless, most of this “lost” specific capacity is of reversible nature as the trapped active lithium can be electrochemically recovered, which is analytically proven by inductively coupled plasma-optical emission spectrometry (ICP-OES) and laser ablation-inductive coupled plasma-mass spectrometry (LA-ICP-MS). Given this relation, a periodic application of a short constant voltage step after discharge results in a significant performance increase. In contrast, holding the cell in the charged state is detrimental for cells with area oversized negative electrodes as the amount of reversible and irreversible trapped active lithium increases. Based on the obtained insights, the influence of variations of the electrochemical conditions on charge/discharge cycling performance is discussed.
1. Introduction The steadily increasing demand for alternative energy sources and electro-mobility necessitates adequate electrochemical energy storage devices [1–7]. Among different systems, batteries, particularly the lithium ion batteries (LIBs) are the state-of-the-art (SOTA) benchmark technology [8–10]. In view of commercial claims, issues with respect to specific energy, energy efficiency, cycle life, safety and even recycling still require further research and development efforts [11–13], which
∗
remains challenging given the complexity of a LIB cell during operation [14–17]. The charge/discharge cycling performance of a LIB cell sensitively depends on a wide range of parameters regarding not only the active and inactive materials (including their composition and balancing aspects) [18–25], but also the design of cell geometry [26,27,45]. An example for the influence of the latter is a cell with geometrically oversized area of the negative electrode relative to the area of the positive electrode. This design has practical benefits as it minimizes the
Corresponding author. Corresponding author. E-mail addresses: [email protected] (J. Kasnatscheew), [email protected] (F.M. Schappacher).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.06.043 Received 13 March 2018; Received in revised form 4 June 2018; Accepted 10 June 2018 0378-7753/ © 2018 Elsevier B.V. All rights reserved.
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
risk of lithium plating at the edges of graphite electrode sheets during charge [28]. However, a deteriorating relation between the performance and the oversized area of the negative electrode is reported by Son et al. [29]. It is claimed that the outer areas (overhang) of the negative electrode additionally utilize active lithium for the formation of the solid electrolyte interphase (SEI) [30–32], resulting in increased irreversible specific capacity losses. However, these conclusions made on the interpretations of Coulombic efficiencies (CEs) did not consider possible kinetic origins, which can have a significant impact on the observed phenomena, as well [33]. In fact, recent reports point to an even reversible character of the apparently “lost” specific capacity by revealing changes in the lithium amount in the negative electrode overhang depending on the state of charge and discharge [34–37]. In this work, the reversible and irreversible specific capacity losses are thoroughly investigated by applying electrochemical and analytical measures for varying degrees of the overhang. Given the practical importance of the overhang usage [27,38,39], the goal of this work is not only to clarify but also to circumvent the origin of the accompanied losses in specific capacity.
Table 1 Charge/discharge cycling procedure includes five constant current (cc) formation cycles, a self-discharge experiment and subsequent constant current/constant voltage (cc/cv) charge/cc discharge cycling. The criterion of 0.05C for cv steps implies a cut-off when the detected current during the cv step drops below the threshold value of 0.05C. Cycle
Mode
Charge
Discharge
1 2–5 6 7 8–19 20 21- …
Formation Formation Cycling Cycling + SD Cycling Cycling + dcv Cycling
cc: 0.05C cc: 0.1C cc/cv: 1C/0.05C cc/cv: 1C/0.05C 120 h rest cc/cv: 1C/0.05C cc/cv: 1C/0.05C cc/cv: 1C/0.05C
cc: 0.1C cc: 0.1C cc: 1C cc: 1C cc: 1C cc/cv: 1C/0.05C cc: 1C
cc: constant current. cv: constant voltage. 1C: 1C corresponds to charge or discharge rate of 142 mAh∙g−1. SD: A rest step of 120 h is applied after charge. dcv: constant voltage step during discharge.
discharge mode. The discharge of the 20th cycle comprised a low rate (0.05C) constant voltage step during discharge (dcv). LIB pouch cells were cycled between 4.2 V and 3.0 V. Four different charge/discharge cycling procedures were used to investigate long term cycling according Table 2. The standard procedure contains 5 formation cycles and 494 cycles of cc/cv charge with subsequent cc discharge at 2C. After the 20th, 120th, 220th, 320th, and 420th charge and discharge was a rest step of 50 min, respectively. The second procedure includes a SD experiment (120 rest step after conventional cc/cv charge) after the 7th, 107th, 207th, 307th and 407th charge. The third procedure includes a dcv step to the standard procedure after the 20th, 120th, 220th, 320th, and 420th discharge. The fourth procedure includes both, SD and dcv experiments. For analytical investigations, cells with 78% oversizing of the negative electrode were investigated (Fig. 1 (c)). Two cells were cycled with the procedure depicted in Table 1 till the 6th, 7th or 20th cycle, respectively. In addition to the discharged cells, two uncycled cells were transferred into the glove box (water and oxygen content of less than 0.2 ppm), disassembled and electrodes were washed with DMC. For Inductively coupled plasma-optical emission spectrometry (ICPOES) measurements, six 4 mm Ø discs were punched out of the negative electrode overhang for each cell. ICP-OES measurements were performed using a Spectro ARCOS ICP-OES (Spectro Analytical Instruments) instrument with axial plasma viewing. A standard Fassel type torch (Spectro Analytical Instruments) was employed. For sample introduction, the system's peristaltic pump with a cross flow nebulizer and a double-pass spray chamber (Scott type) was used. Element emission was detected at different individual emission lines simultaneously. Operating conditions for ICP-OES measurements are shown in Table 3 [42]. This method was used to determine the lithium content of the discharged electrode according the literature [42]. For Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), 20 mm × 20 mm rectangles were punched out of a corner of the 40 mm × 40 mm negative electrodes. The 7Li signal (cps = counts per second) was recorded using 184 transient signals per
2. Experimental Single side coated electrodes for the LIB pouch cells were in house made, based on mesocarbon microbeads (MCMB, Osaka) and LiNi1/ 3Mn1/3Co1/3O2 (NMC111, Cellcore) active materials for negative and positive electrodes, respectively. The electrodes were locally separated by an electrolyte soaked Freudenberg 2190 separator. The electrolyte was based on 1 M LiPF6 in ethylene carbonate (EC):dimethyl carbonate (DMC) 1:1 wt% (LP30, BASF). The electrolyte amount, which can impact the performance upon variation, was excessed in all cells [40]. The LIB pouch cells were assembled in the dry room (dew point of at least −60 °C). The determined practical delithiation capacitiy of the MCMB graphite based negative electrodes was 296 ± 10 mAh∙g−1 and of the NMC111 based positive electrodes was 142 ± 2 mAh∙g−1, when cycled at 0.1 C in three-electrode lithium half cells in the range of 0.02 V and 1.5 V vs. Li/Li+ and 4.2 V and 2.5 V vs. Li/Li+, respectively. Providing an active material mass loading of 9.2 mg∙cm−2 and 14.2 mg∙cm−2, respectively, the negative electrode was oversized in capacity in a ratio of 1.3:1. Different negative to positive electrode area ratios were studied. Whereas the positive electrode had the area size of 30 mm × 30 mm, the negative electrode area sizes were varied from 30 mm × 30 mm–33 mm × 33 mm, 40 mm × 40 mm, and 55 mm × 55 mm, respectively (Fig. 1(a)–(d)). Similar mass loading of the overhang (as apparently “inactive” area) was chosen. (With increasing overhang, the amount of “inactive” materials increases, increasing the overall mass and volume of the cell, finally decreasing energy density/specific energy) [41]. The full cells were cycled according to the electrochemical conditions in Table 1. Five constant current (cc) formation cycles (as default cycles for reasons of SEI optimization) were followed by one constant current/constant voltage (cc/cv) charge, cc discharge acclimation cycle at 1C/0.05C (6th cycle). The 7th cycle included a self-discharge (SD) experiment (i.e., a rest step of 120 h) with subsequent cc discharge at 1C. After the SD experiment, cells were cycled in cc/cv charge, cc
Fig. 1. Cells with varying oversized negative electrode area: (a) 0%, (b) 21%, (c) 78% and (d) 236%. 520
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
[43]. Scans were performed using a laser energy of 2 J∙cm−1, 110 μm circular spots, a shot frequency of 10 Hz and a lateral scanning speed of 110 μm∙s−1. Operating conditions for ICP-MS measurements are shown in Table 4.
Table 2 Charge/discharge cycling conditions for four procedures, a standard procedure (without SD and dcv experiments), a procedure with SD experiments, a procedure with dcv experiments, and a procedure including both, SD and dcv experiments. The LIB pouch cells were cycled within the voltage range of 4.2–3.0 V. Cycle
Mode
Charge
Discharge
1 2–5 6 7
Formation Formation Cycling Cycling (+SD)
cc: cc: cc: cc:
8–19 20
Cycling Cycling (+dcv)
21–106 107
Cycling Cycling (+SD)
108–119 120
Cycling Cycling (+dcv)
121–206 207
Cycling Cycling (+SD)
208–219 220
Cycling Cycling (+dcv)
221–306 307
Cycling Cycling (+SD)
308–319 320
Cycling Cycling (+dcv)
321–406 407
Cycling Cycling (+SD)
408–419 420
Cycling Cycling (+dcv)
421–499
Cycling
cc: 0.05C cc: 0.1C cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C cc/cv: 2C/0.05C (120 h rest) cc/cv: 2C/0.05C cc/cv: 2C/0.05C 50 min rest cc/cv: 2C/0.05C
3. Results and discussion 3.1. Effect of area-oversizing of the negative electrode on specific capacities
0.1C 0.1C 2C 2C
Different degrees of area-oversizing of negative electrodes are realized to investigate the influence of oversize on the charge/discharge cycling performance. Fig. 2 depicts conventional charge/discharge cycling data during the initial five formation cycles at lower rates (0.1C) and a subsequent cycle at an elevated rate (1C). The first cycle Coulombic efficiency (FCE) declines with increasing negative electrode area. Cells with no area-oversizing of the negative electrode (0%) reveal best FCEs of 86.10 ± 0.03% (Fig. 2(a)). An oversizing of the negative electrode of 21%, 78% and 236% (Fig. 2(a)–(d) results in a steady decrease of the FCE with values of 84.22 ± 0.02%, 77.00 ± 0.01% and 59.6 ± 0.7%, respectively. Following this trend, discharge capacities, CEs and capacity retentions of the subsequent formation cycles also become worse with increasing the area-oversizing of negative electrode. This relation of poorer performance with increased area-oversizing of the negative electrode is well in accordance with the findings of Son et al. [29]. However, the application of a SD experiment after the 7th charge (high duration in charged state) reveals controversy findings.
cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C cc: 2C
3.2. The application of a self-discharge experiment for cells with areaoversized negative electrodes As depicted in Fig. 3, the capacity loss during the SD experiment (7th cycle) increases with the degree of the area-oversize of the negative electrode. Whereas cells with zero area-oversize exhibit a capacity loss of only 3.2 ± 0.2 mAh∙g−1, cells with an negative electrode area-oversizing of 21% show a capacity loss of already 10.6 ± 0.1 mAh∙g−1 (Fig. 3(a) and (b)). The capacity losses further increase for area-oversized negative electrodes of 78% and 236% to 19.2 ± 0.1 mAh∙g−1 and 41.8 ± 0.5 mAh∙g−1, respectively (Fig. 3(c) and (d)). Immediately after the SD experiment (7th cycle), a salient continuous recovery of the capacity during upcoming cycles is observable for cells with area-oversized negative electrode. Within this recovery, the discharge capacities exceed the charge capacities (implying Coulombic efficiencies > 100%), pointing to a kinetic impairment in the previous cycle that is resolved this cycle, which is in accordance with recent reports [35–37]. Cells with an oversized negative electrode of 21% show a capacity recovery of 3.15 ± 0.07 mAh∙g−1, whereas cells with 78% and 236% negative electrode area-oversizing recovered capacities of 2.8 ± 0.4 mAh∙g−1 and 4.9 ± 0.5 mAh∙g−1 from the 7th to the 19th discharge. To maximize the recovery of – as stated above – only kinetically “lost” specific capacity, a dcv step is applied (20th cycle, Fig. 2) to enable a complete discharge by excluding kinetic limitations [44]. The amount of specific capacity recovery after the dcv step correlates to the area-oversizing degree of the negative electrode. It can be concluded that the area-oversizing leads to additional specific capacity losses, as the active lithium is preferably kinetically trapped in the overhang.
cc: 2C cc (/cv): 2C (/0.05C) 50 min rest cc: 2C
cc: constant current. cv: constant voltage. 1C: 1C corresponds to charge or discharge rate of 142 mAh∙g−1. SD: A rest step of 120 h is applied after charge. dcv: constant voltage step during discharge. Table 3 ICP-OES system parameters. ICP-OES system parameter
Operating conditions
Radio frequency power Total Argon amount Sample carrier gas Outer plasma gas Torch injector tube (inside diameter)
1400 W 13.65 L∙min−1 0.80 L∙min−1 12.0 L∙min−1 1.8 mm Ø
Table 4 ICP-MS system parameters. ICP-MS system parameter
Operating conditions
Radio frequency power Radio frequency matching Argon carrier gas flow Integration time per mass Sampling depth
1550 W 1.30 V 950 mL∙min−1 0.03 s 6.0 mm
3.3. Mechanism of the capacity recovery after self-discharge experiments The observed relation between area-oversizing of the negative electrode, specific capacity losses and the recovery behavior of the specific capacity points to a mechanism, which is schematically depicted in Fig. 4. It can be deduced, that the intentionally inactive overhang of the negative electrode (black frame in Fig. 4(a) is, by no
sample. A 193 nm ArF Excimer Laser (Analyte Excite Excimer LASystem, Teledyne Cetac) coupled to an ICP-MS System (7700 Series, Agilent Technologies) with helium carrier gas (99.999%, Westfalen AG) with a flow of 900 mL∙min−1 as described in literature were utilized 521
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 2. Charge/discharge cycling performance of LIB cells with different area-oversizing of the negative electrode during initial five cycles at a lower rate (0.1C) and one subsequent cycle at a higher rate (1C). Area-oversizing of the negative electrode amounts to (a) 0%, (b) 21%, (c) 78% and (d) 236%.
(Fig. 4(c)), the graphite based negative electrode gets delithiated preferably in the area where positive electrode and negative electrode are overlapping (where the electrical field is stronger) resulting in different lithiation degrees (SOC) within the negative electrode, that is lower in the overlap area and higher in the overhang area. The lithium ions in the overhang (where the effective electric field is weaker) are hindered
means, inactive, but rather active involving lithium during charge/ discharge cycling. In the SD experiment (7th cycle), which is an extreme condition (long duration in the fully charged state) this gets more evident. Here, in the charged state lasting for 120 h, the lithium ions can equally distribute within the entire negative electrode area including the overhang (Fig. 4(b)). During the following discharge step
Fig. 3. Self-discharge experiments during the 7th charge followed by charge/discharge cycling at 1C. Area-oversizing of the negative electrode amounts to (a) 0%, (b) 21%, (c) 78% and (d) 236%. 522
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 4. (a) Scheme of the area-oversized negative electrode (black) in comparison to the positive electrode (grey) and (b) the lithium distribution of the negative electrode at the end of a charge cv step in the 7th cycle, that is after the 120 h rest step; (c) discharged state after the SD experiment in the 7th cycle and (d) after discharged state, in the upcoming cycles.
in a kinetic manner, resulting in a slowed down mitigation of the ions from the outer electrode area towards the inner area, finally causing incomplete discharge when reaching the discharge cut-off condition, which results in higher specific capacity loss. In the following cycles, the trapped lithium in the overhang further equilibrates over the electrode area, thus also including the movement towards the inner active area, where the lithium gets active again (Fig. 4(d)). Owing to higher movement towards the inner area from the overhang compared to the opposite, a steady increase in the discharge capacity can be observed (Fig. 3). The application of the dcv step (20th cycle, Fig. 3) obviously facilitates the reactivation of the kinetically trapped active lithium.
Table 5 Lithium content of the overhang (78%) of the negative electrode in an uncycled state, before the SD experiment (6th discharge), after the SD experiment (7th discharge) and after the dcv step (20th discharge). Negative electrode overhang
Li-content/mg
uncycled after the 6th discharge after the 7th discharge after the 20th discharge
0.81 0.98 0.74
experiment); (c) of a discharged electrode after the 7th cycle (after the SD experiment) and (d) of the discharged electrode after the 20th cycle (after dcv). All samples were washed with DMC to minimize the amount of lithium residues (e.g. from the electrolyte) on the electrode, which is shown for the uncycled cell (Fig. 5(a)). As reported in literature, in MCMB/NMC full cells, the active lithium remains in the MCMB anode even in the discharged state, particularly at the beginning of charge/ discharge cycling [18]. This is also observable in the discharged state after the 6th cycle, where the negative electrode provides a uniform lithium distribution in discharge state (Fig. 5(b)), including the overhang. Thus, already at the beginning of charge/discharge cycling the overhang includes lithium, which is not practically utilized. It can be deduced that by increasing the negative electrode overhang, more lithium is “trapped” in the overhang and the practically usable specific capacity gets diminished (in line with the results in Fig. 2). In the discharged state, after the SD experiment (7th cycle), the overhang has a higher lithium concentration compared to the active (overlapping) area of the electrodes, supporting the validity of the aforementioned mechanism and the ICP-OES results (Fig. 4 and Table 5). Trapped lithium
3.4. Specific capacity recovery confirmed by ICP-OES and LA-ICP-MS analyses To support the above data interpretation pointing at kinetically trapped lithium in the overhang of the negative electrode, particularly after the SD experiments [36,37], ICP-OES measurements of the overhang are carried out to determine the lithium contents. Samples are taken from the overhang (for analytical investigation an overhang of 78% is used) of the negative electrodes in the uncycled state and after the 6th cycle, the 7th cycle and the 20th cycle in the discharged state. The alternating lithium contents presented in Table 5 support the aforementioned mechanism (Fig. 4). The lithium content in the overhang increases after SD application (0.81 mg–0.98 mg) and decreases after the dcv step (0.98 mg–0.74 mg), supporting the reversible activity of the overhang. For reasons of verification, LA-ICP-MS measurements are performed, providing the additional benefit to visualize the entire electrode (overlap and overhang area). The lower left corner of the electrodes showing the lithium concentration distribution is shown in Fig. 5: (a) of an uncycled electrode; (b) of a discharged electrode after the 6th cycle (before the SD 523
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 5. Lithium distribution in the lower left corner of the negative electrode in (a) an uncycled state and in the discharged state after (b) the 6th cycle (before SD experiment), (c) the 7th cycle (after SD experiment) and (d) the 20th cycle (after dcv experiment). The black line highlights the boundary of the overhang and the overlap area.
resistances remain unchanged (Fig. 6(b)), the beneficial effect of the repeated application of dcv steps can be reasonably attributed to an increased amount of used active lithium (= increasing specific capacity). On the basis of these insights the dcv step likely counteracts the accumulation of kinetically trapped lithium in the overhang. Application of a periodic SD step however deteriorates the performance of the charge/discharge cycling as seen by the lower specific capacity (Fig. 6(a), green) and increased resistance (Fig. 6(b), green). Given the shown relation of the SD step and overhang of the negative electrode, accumulation of trapped active lithium in the overhang can be assumed (decreasing specific capacity), and moreover its partly irreversible depletion for the SEI formation (increasing internal resistance). Nevertheless, the combination of both, dcv steps and SD steps is obviously even worse for long-term charge/discharge cycling compared to application of solely SD steps, as seen by the observed SOH values of 82.91% and 85.38%, respectively. While the SD step leads to an accumulation of active lithium in the overhang and the dcv step to its removal, this test procedure likely leads to a high flow in/out the overhang area. In consequence this facilitates SEI formation of the overhang which finally involves irreversible active lithium. In summary, long durations in the charged state (as it is the case during SD applications) should be avoided with respect to achieve good performance during long term charge/discharge cycling, when the negative electrode is area-oversized (i.e., in typically balanced cells). The responsible accumulation of active lithium during charge/discharge cycling in the overhang is schematically shown in Fig. 7 for both test procedures including SD steps. In case long durations in charged state can be avoided, the performance can be significantly maximized when periodic dcv steps (controlled over-discharge) [23] are applied. In this case, the partly accumulated active lithium in the overhang is mostly of reversible nature and can be re-activated as schematically illustrated in Fig. 7.
in the overhang is not fully usable resulting in decreased subsequent specific capacities. Furthermore, due to the voltage drop, that evolves during the SD experiment, the cell reaches its discharge cut-off earlier (see SI 1), i.e., at the expense of the discharge capacity. This implies that the electrode is less delithiated after the 7th cycle (SD experiment) compared to the cell discharged without previous SD experiment in the 6th cycle. According to the aforementioned mechanism, the application of the dcv step (20th cycle) leads to an almost complete removal of the lithium in the overhang. Therefore, the electrode shown in Fig. 5(d) shows the least lithium in the remaining in the overhang. Based on the results one could conclude that a dcv step increases the electrochemical performance of the cell, as trapped lithium can be easily reactivated. (Additional insight: when conventional cycling conditions are used, the outer area remains predominately inactive, hence the mass balancing between the active areas of cathode and anode is unaffected by additional mass of outer area.) For better verification, long-term effects of the dcv and SD step on the performance are investigated in the following. 3.5. Long-term effects of repetitive self-discharge (SD) and discharge constant voltage (dcv) steps As displayed in Fig. 6(a) LIB cells that are cycled at conventional electrochemical conditions (without SD and dcv steps; according to Table 2), reveal a characteristic decay in specific capacity for the given full cell system and conditions. After 499 cycles, the cells exhibit a specific discharge capacity of 112.3 ± 0.3 mAh∙g−1, which corresponds to a state of health (SOH) of 86.39% (compared to the 6th cycle with 130.0 ± 0.1 mAh∙g−1). After 499 cycles, the cells cycled according to the standard procedure, provided a raise in resistance of about 1.5 Ω as depicted in Fig. 6(b). In case the dcv step is applied every 100 cycles (Fig. 6(a), orange), the charge/discharge cycling performance is significantly improved up to a SOH of 90.22% after 499 cycles. Given the fact that the internal 524
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Fig. 6. (a) Charge/discharge cycling results of cells cycled at different procedures (Table 2) having 21% overhang. A standard procedure (black) including only cc/cv charge and cc discharge (black). A standard procedure plus a dcv step after 20th, 120th, 220th, 320th and 420th discharge (orange). A standard procedure plus SD step in cycles 7, 107, 207, 307 and 407 (green). A standard procedure including both, dcv and SD steps (blue). (b) Relative comparison of the growth in internal resistance during charge/discharge cycling. All graphs show average values of two cells and are normalized to their 6th cycle. (Unedited internal resistances with standard deviations are depicted in Figure SI 2.). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
4. Conclusion
and laser ablation-inductive coupled plasma-mass spectrometry (LAICP-MS) confirmed not only its activity, but also its partly reversible nature. Relevant insights into long-term charge/discharge cycling performance could be deduced. Already a periodic application of a short dcv step after several cycles significantly improved the performance by reactivating the amount of trapped active lithium in the overhang area. As revealed by self-discharge (SD) experiments, long durations of a LIB cell in the charged state should be avoided, as they accelerate electrolyte decomposition, which enhanced irreversibly utilized active lithium and increased cell resistance. In addition, for such anode to cathode area ratios, the enlarged amount of trapped active lithium in the overhang area increased its irreversibility (e.g. by immobilization in SEI film formation) and in turn additionally led to increased internal resistances.
In this work, the influence of an area-oversized negative electrode on the charge/discharge cycling performance was thoroughly investigated. Such cell construction is beneficial as it assured a complete overlap of the electrodes. However, the apparently inactive (but in practice active) overhang of the negative electrode significantly influenced the charge/discharge cycling performance. A systematic increase of the overhang area resulted in increased specific capacity losses, particularly for a longer duration in charged state. However, when enhancing the discharge kinetics by application of a discharge constant voltage (dcv) step, a major part of the apparently irreversible specific capacity losses could be recovered, pointing to their mostly reversible nature. The monitoring of the lithium content of the overhang area by inductively coupled plasma-optical emission spectrometry (ICP-OES)
Fig. 7. Schematic illustration of the quantitative distribution of lithium in the area oversized negative electrodes during charge/discharge cycling for the four different test procedures depicted in Fig. 6(a). The cycling procedure with the absence of SD and dcv steps (Standard) leads to a continuous accumulation of lithium in the overhang, which are mostly of reversible nature, thus can be reactivated by using the procedure with repetitive dcv steps (+dcv). In contrast, application of the procedure with SD steps (+SD) leads to a severe accumulation of lithium in the overhang having an increased irreversible amount, thus cannot be re-activated by using the procedure including dcv steps (+SD & dcv). 525
Journal of Power Sources 396 (2018) 519–526
T. Dagger et al.
Acknowledgement [20]
The authors would like to express their thanks to MEET Battery Research Center at the University Muenster. Daniel Brüggemann and Marco Fritzen are acknowledged for cell assembly works. Pascal Noll and Nils Wallus (MEET Battery Research Center) are acknowledged for providing balanced electrodes.
[21] [22] [23] [24]
Appendix A. Supplementary data
[25]
Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.jpowsour.2018.06.043.
[26] [27]
References [28] [29]
[1] R. Wagner, N. Preschitschek, S. Passerini, J. Leker, M. Winter, J. Appl. Electrochem. (2013) 1–16. [2] O. Gröger, H.A. Gasteiger, J.-P. Suchsland, J. Electrochem. Soc. 162 (2015) A2605–A2622. [3] R.W. Schmitz, P. Murmann, R. Schmitz, R. Müller, L. Krämer, J. Kasnatscheew, P. Isken, P. Niehoff, S. Nowak, G.-V. Röschenthaler, Prog. Solid State Chem. 42 (2014) 65–84. [4] D. Andre, S.-J. Kim, P. Lamp, S.F. Lux, F. Maglia, O. Paschos, B. Stiaszny, J. Mater. Chem. 3 (2015) 6709–6732. [5] K.G. Gallagher, S. Goebel, T. Greszler, M. Mathias, W. Oelerich, D. Eroglu, V. Srinivasan, Energy Environ. Sci. 7 (2014) 1555–1563. [6] T. Placke, R. Kloepsch, S. Dühnen, M. Winter, J. Solid State Electrochem. (2017) 1–26. [7] P. Meister, O. Fromm, S. Rothermel, J. Kasnatscheew, M. Winter, T. Placke, Electrochim. Acta 228 (2017) 18–27. [8] M. Winter, R.J. Brodd, Chem. Rev. 104 (2004) 4245–4269. [9] N. Nitta, F. Wu, J.T. Lee, G. Yushin, Mater. Today 18 (2015) 252–264. [10] E.J. Berg, C. Villevieille, D. Streich, S. Trabesinger, P. Novák, J. Electrochem. Soc. 162 (2015) A2468–A2475. [11] J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, S. Nowak, I. Cekic Laskovic, M. Winter, J. Power Sources 359 (2017) 458–467. [12] P. Meister, H. Jia, J. Li, R. Kloepsch, M. Winter, T. Placke, Chem. Mater. 28 (2016) 7203–7217. [13] S. Rothermel, M. Evertz, J. Kasnatscheew, X. Qi, M. Grützke, M. Winter, S. Nowak, Chemsuschem 9 (2016) 3473–3484. [14] M. Evertz, F. Horsthemke, J. Kasnatscheew, M. Börner, M. Winter, S. Nowak, J. Power Sources 329 (2016) 364–371. [15] J.B. Goodenough, K.-S. Park, J. Am. Chem. Soc. 135 (2013) 1167–1176. [16] W. Li, B. Song, A. Manthiram, Chem. Soc. Rev. 46 (2017) 3006–3059. [17] J. Kasnatscheew, B. Streipert, S. Röser, R. Wagner, I. Cekic Laskovic, M. Winter, Phys. Chem. Chem. Phys. 19 (2017) 16078–16086. [18] J. Kasnatscheew, T. Placke, B. Streipert, S. Rothermel, R. Wagner, P. Meister, I. Cekic-Laskovic, M. Winter, J. Electrochem. Soc. 164 (2017) A2479–A2486. [19] J. Kasnatscheew, U. Rodehorst, B. Streipert, S. Wiemers-Meyer, R. Jakelski,
[30] [31] [32] [33]
[34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45]
526
R. Wagner, I. Cekic-Laskovic, M. Winter, J. Electrochem. Soc. 163 (2016) A2943–A2950. J. Kasnatscheew, M. Evertz, R. Kloepsch, B. Streipert, R. Wagner, I. Cekic Laskovic, M. Winter, Energy Technol. 5 (2017) 1670–1679. J. Kasnatscheew, R.W. Schmitz, R. Wagner, M. Winter, R. Schmitz, J. Electrochem. Soc. 160 (2013) A1369–A1374. R. Wagner, S. Brox, J. Kasnatscheew, D.R. Gallus, M. Amereller, I. Celdc-Laskovic, M. Winter, Electrochem. Commun. 40 (2014) 80–83. J. Kasnatscheew, M. Börner, B. Streipert, P. Meister, R. Wagner, I. Cekic Laskovic, M. Winter, J. Power Sources 362 (2017) 278–282. T. Dagger, M. Grützke, M. Reichert, J. Haetge, S. Nowak, M. Winter, F.M. Schappacher, J. Power Sources 372 (2017) 276–285. B. Streipert, S. Roser, J. Kasnatscheew, P. Janssen, X. Cao, R. Wagner, I. CekicLaskovic, M. Winter, J. Electrochem. Soc. 164 (2017) A1474–A1479. P.J. Osswald, S.V. Erhard, J. Wilhelm, H.E. Hoster, A. Jossen, J. Electrochem. Soc. 162 (2015) A2099–A2105. B.R. Long, S.G. Rinaldo, K.G. Gallagher, D.W. Dees, S.E. Trask, B.J. Polzin, A.N. Jansen, D.P. Abraham, I. Bloom, J. Bareño, J. Electrochem. Soc. 163 (2016) A2999–A3009. M. Tang, P. Albertus, J. Newman, J. Electrochem. Soc. 156 (2009) A390–A399. B. Son, M.-H. Ryou, J. Choi, S.-H. Kim, J.M. Ko, Y.M. Lee, J. Power Sources 243 (2013) 641–647. C. Korepp, W. Kern, E. Lanzer, P. Raimann, J. Besenhard, M. Yang, K.-C. Möller, D.T. Shieh, M. Winter, J. Power Sources 174 (2007) 637–642. M. Winter, Z. Phys. Chem. Int. J. Res. Phys. Chem. Chem. Phys. 223 (2009) 1395–1406. G.H. Wrodnigg, J.O. Besenhard, M. Winter, J. Electrochem. Soc. 146 (1999) 470–472. J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, R. Klöpsch, B. Vortmann, H. Hahn, S. Nowak, M. Amereller, A.-C. Gentschev, P. Lamp, M. Winter, Phys. Chem. Chem. Phys. 18 (2016) 3956–3965. M. Lewerenz, A. Warnecke, D.U. Sauer, J. Power Sources 354 (2017) 157–166. M. Lewerenz, J. Münnix, J. Schmalstieg, S. Käbitz, M. Knips, D.U. Sauer, J. Power Sources 345 (2017) 254–263. B. Gyenes, D.A. Stevens, V.L. Chevrier, J.R. Dahn, J. Electrochem. Soc. 162 (2015) A278–A283. J. Wilhelm, S. Seidlmayer, P. Keil, J. Schuster, A. Kriele, R. Gilles, A. Jossen, J. Power Sources 365 (2017) 327–338. E. Schuster, C. Ziebert, A. Melcher, M. Rohde, H.J. Seifert, J. Power Sources 286 (2015) 580–589. J. Kurfer, M. Westermeier, C. Tammer, G. Reinhart, CIRP Ann. - Manuf. Technol. 61 (2012) 1–4. S.J. An, J. Li, D. Mohanty, C. Daniel, B.J. Polzin, J.R. Croy, S.E. Trask, D.L. Wood, J. Electrochem. Soc. 164 (2017) A1195–A1202. J. Kasnatscheew, R. Wagner, M. Winter, I. Cekic-Laskovic, Top. Curr. Chem. 376 (2018) 16. B. Vortmann-Westhoven, M. Winter, S. Nowak, J. Power Sources 346 (2017) 63–70. T. Schwieters, M. Evertz, A. Fengler, M. Börner, T. Dagger, Y. Stenzel, P. Harte, M. Winter, S. Nowak, J. Power Sources 380 (2018) 194–201. J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, S. Nowak, I. Cekic-Laskovic, M. Winter, J. Phys. Chem. C 121 (2017) 1521–1529. S.J. An, J. Li, C. Daniel, S. Kalnaus, D.L. Wood, J. Electrochem. Soc. 164 (2017) A1755–A1764.