**2. Experimental**

## *2.1. Materials*

To investigate the influence of the laser process parameters on the properties of the cutting edge and the influence on the electrochemical performance, double-sided coated electrodes with industrially available material components were used. For the anode, the active material SMGA4 (91 wt.%; Hitachi, Japan), with a specific capacity of 360 mAhg−1, was coated on a 10 μm thick copper collector (Sumisho Metallex, Japan). Subsequently, the coating was compacted to a density of 1.5 gcm<sup>−</sup>3, which results in the porosity of 32.25% and a total thickness of 123 μm. On the cathode side, the active material NMC 111 (90 wt.%; BASF, Germany), with a specific capacity of 165 mAhg−1, was coated on 20 μm thick aluminum collector (Hydro Aluminum Rolled Products, Germany) and compacted to a degree of 2.8 gcm<sup>−</sup>3. The described compaction led to porosity of 31.35% and a total electrode thickness of 143 μm. For the anode and cathode, a conductivity additive SFG6L (2 wt.% anode, 2 wt.% cathodes; Imerys, Switzerland), a carbon black C65 (2 wt.% anode, 4 wt.% cathodes; Imerys, Switzerland), and a PVDF binder (5 wt.% anode, 4 wt.% cathodes; Solvay, Italy) were utilized. The cell manufacturing was carried out with a 27 μm thick separator (Separion) and a conventional LiPF6 electrolyte (UBE Industries Ltd., Japan). The conductive salt LiPF6 was solved in a concentration of one mole in a solvent consisting of ethylene carbonate (EC), dimethyl carbonate (DMC), and ethyl methyl carbonate (EMC), with a volumetric ratio of the solvent components of 1:1:1. To suppress the evolution of gas during the first charging, the electrolyte contained 2 wt.% of Vinylene Carbonate (VC). As further additives for reducing hydrogen formation at high voltages, the electrolyte contained 3 wt.% of cyclohexylbenzene (CHB).

#### *2.2. Analysis of the Cutting Edge Characteristics*

The prescriptive characteristics of the electrode cutting edge are shown in Figure 5a by means of a microsection. These characteristics were determined by light microscopy (VHX 2000 light microscope (LM), Keyence, Osaka, Japan) and laser scanning microscope (VK-X Series 3D Laser Scanning Confocal Microscope (LSM), Keyence, Osaka, Japan). Here, the parameters chamfer width and heat-a ffected zone (HAZ) were considered to be the significant influencing factors on the electrochemical performance and, therefore, investigated further. The HAZ is defined as an area where the active material is thermally stressed but not removed. The chamfer width is characterized by active material removal and a melt formation zone.

**Figure 5.** Analysis of the prescriptive cutting edge characteristics: (**a**) Schematic microsection of a laser-generated cutting edge, 1. Chamfer width (chw), 2. The heat-a ffected zone (HAZ), 3. Melt formation, 4. Ablation, 5. Melt superelevation, α Chamfer angle; (**b**) LSM image of a cutting edge; (**c**) Analysis of a cutting edge by LSM data, based on [10].

To measure these characteristics, LM and LSM images were taken of the upper side of the electrodes at the cutting edges. The areas of the heat-a ffected zone and the chamfer width could be clearly separated by a combined measuring method. By means of the LSM topography images (Figure 5b,c), the chamfer width was measured. By subtracting the chamfer width from the entire a ffected area (LM

images), the heat-affected zone could be quantified. For the determination of contaminant products as a result of the laser-material interaction, SEM (FEI Quanta 650, Thermo Fisher, Waltham, MA, USA) and EDX (Oxford X-Max 80 mm2, Oxford Instruments, Abingdon, England) images of the cut electrodes were taken.

#### *2.3. Cell Format and Manufacturing*

To evaluate the influence of the cutting edge on the electrochemical performance, pouch cells with 15 compartments were built (total cell capacity about 9 Ah). The relatively high number of compartments was chosen to emphasize the effect of the cutting edge properties of the electrodes, maximizing the effects of the cutting edge characteristics on the electrochemical performance. The surface of the anode coated with active material was 16,484 mm<sup>2</sup> and, based on the geometry shown in Figure 6a, gave a cutting edge to surface ratio of 0.030 mm<sup>−</sup>1. The cathode is defined by an area of 15,209 mm<sup>2</sup> and a ratio of cutting edge to the surface of 0.031 mm<sup>−</sup>1. For the anode, a circumferential overlap of 2.5 mm resulted from the illustrated geometries for the anode and the cathode. This overlap guaranteed the total stress of the cathode as a reference in these examinations and ensured the correct balancing of the compartment.

By means of a z-folding process (prototype plant, Jonas & Redmann, Berlin, Germany), the singularized electrodes were stacked alternating between a separator to form an electrodeseparator-composite (ECS). Subsequently, the individual collectors of the electrodes were joined to a tab via ultrasonic welding (Ultraweld F20, Branson Ultraschall, Hannover, Germany). For this purpose, the 15 single anode collectors were welded to a nickel tab with an energy of 200 j, and cathodes collectors to an aluminum tab with an energy of 100 j at an oscillating sonotrode amplitude of 30 μm. Subsequently, the ECS was dried under vacuum for 120 ◦C for 16 h. In the following step, the ECS was inserted into the pouch bag and filled under argon atmosphere with electrolyte and sealed. Finally, the filled cells were tempered for 4 h at 60 ◦C to support the complete wetting of the electrodes. The finished assembled cell is shown in Figure 6b [12].

**Figure 6.** Cell format and pouch cell design: (**a**) Electrode format, (**b**) Complete pouch cell.

#### *2.4. Cell Diagnostic*

After assembly and wetting, the manufactured cells were placed in a climate chamber (WKM Inc., Lachendorf, Germany) at 20 ◦C and connected to a battery tester (Series XCTS, Basytec Inc., Asselfingen, Germany) with a fixed torque of 2.54 Nm. Due to the high capacity of the battery cells, the tests required a high safety environment. Therefore, the climate chambers were equipped with a fire extinguishing system (Wagner Group Inc., Langenhagen, Germany). In the event of an accident, the climate chamber is flooded with nitrogen gas. Furthermore, an activated carbon filter (Stöbich technology Inc., Goslar, Germany) will filter the exhaust air in the pipe duct.

In our experiments, the cells were formed in two cycles. They were first charged and discharged at 1/10 C, and in the second formation cycle with 1/2 C. Upper and lower cut-off voltages were 4.2 V and 2.9 V, respectively, for all charge-discharge cycles. To characterize the cells, a capacity test at 1/10 C and a pulse test (1 C for 1 s) to determine the internal resistance were performed. After the formation process, the cells were matured over eight days with a state of charge (SOC) of 50% at 20 C. Then, the aging of the cells began with a C-rate test with different discharge-rates from 1/5 to 2 C, which lasted 20 cycles. Long-term cycling was then started at 1 C for 100 cycles. After this, the cyclization was paused, and the internal resistance was measured in a pulse test before the C-rate test was repeated. These aging investigations were repeated periodically until at least 450 cycles were reached. In this study, 5 cells per laser variation were analyzed, and only the long-term cycling was considered.

#### *2.5. Laser Cutting Plant and Key Parameter*

The laser cutting plant used in this study was an in-house development and construction. Due to its modular structure and the process-immanent advantages of the scanner system, it is suitable for a large number of different electrode formats. The beam source used was a nanosecond pulsed fiber laser with a central emission wavelength of 1059–1065 nm (G4 Pulsed Fiber Laser, SPI Lasers UK Ltd., Southampton, UK). The average power of the fiber laser was 72 W with a peak pulse power of up to 13 kW and an M<sup>2</sup> of <1.6. Guidance and focusing of the laser beam were performed by a 3-axis laser beam deflection system (AXIALSCAN 30/FOCUSSHIFTER, Raylase AG, Wessling, Germany) with a working field of 400 × 400 mm2. The first two dimensions of the scanner were needed to drive the spot over the workpiece to create the cutout. The third dimension was needed to ensure a constant spot size of ~74 μm with a focus depth of 0.6 mm on one level over the entire working field. All laser cuts were made in one pass. The fully automated handling system was carried out by simple roll to roll and pick and place operation (Figure 7a), controlled by an Arduino Mega 2560. Since the cut could only be realized in the focus level of the laser spot, a special negative form for the positioning of the electrode had to be built for each electrode format (Figure 7b). The positioning of the electrode was done by negative pressure on holes surrounding the cutting curve. Even though a cutting on the fly was possible with the used remote scanner system, a static cutting operation was used to guarantee constant cutting speed over the complete cutting length and an easy format change.

**Figure 7.** Laser cutting plant: (**a**) Front view of the laser cutting plant; (**b**) Negative form.

 **E** 

#### **3. Results and Discussion**

**D** 

The presentation of the results has been divided into three sections. In the first section, the results of the influence of the laser parameters on the cutting edge characteristics, chamfer width, and heat-affected zone have been considered and discussed. The investigations focused on the pulse repetition frequency, the cutting speed, and the pulse length, as well as the influence of the number of hits per surface increment, and intensity at a constant energy density. In the second section, the influence of the cutting edge characteristics and corresponding laser process parameters on the electrochemical performance has been examined. Building on these results, the last section would present further investigations of the cutting process and the cutting edge, explaining the electrochemical behavior.

#### *3.1. Influence of the Laser Process Parameters on the Cutting Edge Characteristics*

In a first step, we investigated the influence of the laser parameters on the cutting edge properties. For this purpose, the pulse repetition frequency (PRF) and the cutting speed (Vc) were deliberately varied with constant power and pulse duration. The resulting cut edges were evaluated according to the analysis methods presented. On the basis of the obtained data, models could be developed by means of the analysis program Design Expert 11, which describes the examined parameter space. The experimental design was made with a D-optimal strategy and comprised 13 experiments with five repeated measurements each. The response surface model was adapted to the measured values by fitting it to a quadratic polynomial of the form of Response = Intercept \* + A + B + AB + AB<sup>2</sup> + A2B A<sup>2</sup> B2. The results for the anode (Figure 8) showed that both parameters influenced the formation of the chamfer width and the heat-a ffected zone. In Figure 8a model, it could be seen that the formation of the chamfer width steadily decreased with increasing PRF. This tendency could also be observed with increasing cutting speed. The smallest chamfer width for this model was obtained at the maximum achievable speed of 700 mms<sup>−</sup>1, with a pulse repetition frequency of 490 kHz. Considering the additional laser parameters, this resulted in a number of hits per area increment of 51 and an energy density of 141 jcm−2. The decrease in the chamfer width with increasing speed or decreasing energy density could be explained by the lower energy input per area. The dependence on energy density has already been confirmed by previous publications [24]. The decrease of the chamfer width with increasing pulse repetition frequency could be attributed to di fferent mechanisms, which result from the adjusted mode of the energy input. Due to the constant average power and the constant pulse duration, the pulse peak power had to drop with increasing pulse repetition frequencies to reduce the energy per pulse. As a consequence, the intensity and the energy decreased with increasing pulse repetition frequency, and thus the width of the threshold intensity or the spot diameter, which led to an ablation, becoming smaller due to the Gaussian intensity distribution. Besides, the increased number of hits of 360 (at 490 kHz and 100 mms<sup>−</sup>1) could lead to a more intensive harmonic laser-plasma interaction, which reduced the energy impinging on the target by shielding e ffects and thus reduced the energy density. In addition, the reduced pulsed peak power could lead to a smaller broadening of the plasma formation. Since the plasma was also involved in the removal of material and the thermal load on the surface, the proportion of the total material removal became less at higher frequencies. Literature regarding ns laser-induced breakdown spectroscopy describes threshold pulse fluences for forming a plasma of 1.01 jcm−<sup>2</sup> for aluminum and 1.46 jcm−<sup>2</sup> for copper. As the smallest pulse fluence at 490 kHz for this System was 3.45 jcm−2, the plasma formation, in general, would always occur for the examined parameter space [26].

**Figure 8.** Model for the influence of the pulse repetition frequency and cutting speed on the anode cutting edge: 72 w, 240 ns; (**a**) Influence on the chamfer width (Cubic fitting: R<sup>2</sup> = 0.77, Adjusted R<sup>2</sup> = 0.75, Predicted R**<sup>2</sup>** = 0.73); (**b**) Influence on the heat-affected zone (Cubic fitting: R<sup>2</sup> = 0.94, Adjusted R<sup>2</sup> = 0.93, Predicted R<sup>2</sup> = 0.93).

From the heat-affected zone model (Figure 8b), it could be seen that increasing the PRF could reduce the heat-affected zone. The cause for the dependence of the thermal load on the PRF could be explained by the reasons given above for the dependence of the chamfer width on the PRF. The correlation of the HAZ with the cutting speed showed the opposite behavior to the chamfer width. As the cutting speed increased, the HAZ increased in the range of 70–350 kHz. This could be explained by the fact that with increasing cutting speed, the chamfer width and the kerf were becoming steadily smaller. This means that less material was removed for higher cutting speeds. Due to the Gaussian intensity distribution and the increasing speed, the energy input profile changed to the effect that the material which was no longer ablated underwent such high thermal stress that there was an optical change. This means that the final product cut with high speed contained a larger active material area that is thermally stressed, which would be completely removed at lower speeds. The investigations on the cathode were carried out in smaller parameter space (Figure 9a) with respect to the speed (100–400 mms<sup>−</sup>1) because due to the higher material thickness of the collector already at 455 mms<sup>−</sup>1, the cut-through limit for high PRF was reached. The results for the formation of the chamfer width as a function of the PRF and Vc showed similar tendencies as the anodic model. Both with increasing PRF and with increasing Vc, the chamfer width decreased significantly. The results showed that the smallest chamfer widths could only be achieved through the combination of low energy densities and high PRF. Possible causes for these dependencies could be transferred from the explanations to the anode. The model presented for the development of the heat-affected zone for cathodes (Figure 9b) could be adjusted with an R<sup>2</sup> of 0.93 and allowed a 92% reliable prediction. With increasing PRF, the HAZ decreased significantly until it approached zero at 490 kHz. Here, the influence of the cutting speed played only a minor role. Despite the low significance, an increase in the cutting speed led to a reduction in the HAZ. These results were opposite to the results for the anode. The difference in behavior was explained by two facts. Firstly, the intensity and energy difference between the ablation threshold and the thermal stress threshold were smaller for the cathode active material than for the anode active material. Secondly, the ablation threshold of the cathode active material was higher than that of the anode active material.

**Figure 9.** Model for the influence of the pulse repetition frequency and cutting speed on the cathode cutting edge: 72 w, 240 ns; (**a**) Influence on the chamfer width (Cubic fitting: R<sup>2</sup> = 0.91, Adjusted R<sup>2</sup> = 0.90, Predicted R<sup>2</sup> = 0.88); (**b**) Influence on the heat-affected zone (Cubic fitting: R<sup>2</sup> = 0.94, Adjusted R<sup>2</sup> = 0.93, Predicted R<sup>2</sup> = 0.92).

Further investigations outside of the considered parameter space in the range of very low cutting speeds (50 mms-1) and very high energy densities, respectively, showed that on the cathode and anode cutting edge, either no or very small HAZ could be identified. This was caused by the slope of the intensity distribution and the very high energy input. As a result, the areas that were previously only thermally stressed at lower intensities were subjected to material removal at higher energy densities. Furthermore, the high energy density at 50 mms<sup>−</sup><sup>1</sup> led to an increase in the ablation area, since the ablation thresholds of the collector and active material differ significantly.

With regard to the strong influence of the PRF on the chamfer width and the HAZ, the influence of the pulse duration and the pulse peak power on different PRF was investigated in a further study. For this purpose, cuts were performed at a constant energy density with a variation of the pulse duration and pulse peak power. Pulse duration was kept constant (240 ns) with the effect that the pulse peak power decreased with increasing frequency (70 kHz: 13 kW, 102 kHz: 6 kW, 200 kHz: 2 kW, 291 kHz: 1 kW, 403 kHz: 0.7 kW, 490 kHz: 0.55 kW). The pulse duration was shortened (240–20 ns) to keep the pulse peak power quasi constant (70 kHz: 13 kW, 102/200/291 kHz: 10 kW, 403/490 kHz: 9 kW). The results for the chamfer width and the heat-affected zone derived from these experiments are shown in Figure 10. The PRF variation with constant pulse duration showed the same tendencies as in the previously presented models in Figures 8 and 9.

A reduction of the pulse length with a quasi-constant pulse peak power led to a larger chamfer width and HAZ, both at the anode and at the cathode (Figure 10). On the anode side, the reduction of the pulse length led to a significant enlargement of the heat-affected zone for the PRF 102 and 200 kHz. Due to the higher PRF and the high intensities, the plasma formed was of higher energy, leading to enhanced thermal stress of the electrode surface and, thus, potentially to a higher ablation. The reduction of the HAZ by the increased frequency of the ns laser pulses was thus determined largely by the low pulse energy. Only by a much greater reduction of the pulse length of less than 10 picoseconds, a cold cutting is possible [14].

**Figure 10.** Influence of the pulse duration at different pulse repetition frequencies on the chamfer width and heat-affected zone of the anode (**a**) and the cathode (**b**) cutting edge.

The increased chamfer width for the anode and the cathode at higher pulse peak powers at higher PRF confirmed the assumption that was previously made on the anode and cathode model. In addition to the decrease in pulse peak power, the decrease in pulse energy at higher PRF led to a reduction of the chamfer width.

In the following experiment, the scalability of the cutting speed or the influence of the number of hits at constant energy density was investigated (Figure 11). The results for the anode showed that at constant energy density, the chamfer width increased at a reduced rate. A reason for this was the lower intensity and the correlated intensity distribution, as well as the lower energy of the pulse, which narrowed the profile of the material removal threshold. As a result, the geometric distance between a full cut and the ablation of the active material increased (see Figure 3.). In general, the cutting kerf, as well as the entire area in which material was removed, became smaller as a result. Because of the displacement of the removal thresholds, a larger chamfer width was produced at lower speeds for a specific energy density.

**Figure 11.** Influence of the intensity and the number of hits per area increment on the chamfer width at a constant energy density of 328 jcm−2.

The fluence resulting from the reduction of the average power to 33.3% was still 1.15 jcm−<sup>2</sup> and was thus above the limit for the formation of plasma for aluminum. Despite the high number of hits, it could be assumed that shielding effects were negligible as the intensity was much lower, and the larger plasma formation led to increased removal of the active material. When the average energy was

reduced to one-third of the maximum power, no cut could be made through the copper collector at a frequency of 490 kHz for the energy density being studied. Due to the strong reduction of the pulse energy and peak pulse power, the beam could no longer be coupled because of the low absorption of the copper. When cutting copper, it is first heated by the radiation until it oxidizes [10]. As soon as the material oxidizes, the radiation can be much better coupled into the material, and only then leads to the sufficiently high absorption of the laser radiation for a complete cut.

Cathode investigations showed similar tendencies, which were much less pronounced. Compared to the anode, the cathode could be cut at 33% of the maximum average power, although the cathode's aluminum collector was twice as thick as the anode's copper collector. The results basically showed that the necessary energy density could be used as a second parameter to define a cut-through limit.

#### *3.2. Influence of the Cutting Edge Characteristics and the Process Parameters on the Electrochemical Performance*

Based on the knowledge gained from the experiments, a parameter study was developed to investigate the influence of the presented product and process properties on the electrochemical performance of the electrode, or the cell. In a first study, a cell was built with anodes and cathodes cut using the same laser and system parameters. The parameter configuration for this cell was referred to as V0 and served as a reference system for the variation of the parameter configuration. This parameter configuration was very reliable in terms of possible fluctuations in the focus position and the layer thickness. In a first limitation, it was examined whether the cutting edge on the anode side, or the cathode side, had a greater influence on the electrochemistry and, thus, the performance of the cell. In further experiments, only the cutting edge of the performance controlling electrode was examined. For the first experiments, cells with anodes (V1) and cathodes (V2) with a very large chamfer width were built. To produce this very large chamfer width, the electrodes were cut with a very high energy density and intensity. In the following, the parameters PRF (V3), Vc (V4), and τ (V5) were varied at a constant energy density at the electrode. Based on the parameter configurations shown in Table 1, it was possible to evaluate the previously presented cutting edges characteristics and process characteristics with regard to their influence on the electrochemical performance of the cell.



Cutting speed (Vc), Pulse repetition frequency (PRF), Pulse duration (τ), Energy density (ED), Intensity (*IPeak*), Number of laser pulses per surface increment (*nline*).

The evaluation of the electrochemical performance showed that the cutting edge of the cathode exerted the greater influence. Therefore, the influence of cut edge characteristics and process configurations at the cathode was investigated. In the following, the characteristics chamfer width and heat-affected zone have been shown for the examined parameter configurations V0 to V5.

The comparison of the features in Figure 12 showed that the total affected area consisting of the heat-affected zone and chamfer width was the largest for the reference parameter V0 for both the anode and the cathode. The largest chamfer width with almost identical characteristics showed the parameters V1 for the anode and V2 for the cathode. The further investigations with the parameters V3 to V5 showed, on the cathode side, the smallest influenced area with an average chamfer width of 70 μm to 80 μm.

**Figure 12.** Response to the parameters, shown in Table 1.

The investigation of the influence of the presented cutting edge characteristics on the electrochemical performance was carried out by means of the cyclization routine defined in Section 2.4. The result of the diagnosis of the electrochemical performance is shown in the form of normalized cyclization curves in Figure 13. The cyclization curves showed that in comparison to the reference (V0), the cathode cutting edge (V2) exerted a significantly greater influence on the cycle stability than the anode (V1). In this case, the mean value for all cells, with the anode (V1), lied on the mean value of the reference cells (V0) with a similar standard deviation after 350 cycles. The cells with the cathode (V2) were far above this value and thus had significantly higher cycle stability. All tests to determine the influence of the chamfer width (V3–V5) on the cathode side showed that with low chamfer width, no improvement of the cycle stability could be achieved compared to V2. Since the cells V2 to V5 had higher cycle stability than the reference cells V0, the statement could be made that the heat-affected zone on the cathode side had a greater influence on the electrochemical stability of the cell than the chamfer width. Furthermore, the cycling curves of cells V3 and V5 were nearly identical after 350 cycles. From this, it could be deduced that the pulse duration or pulse peak power as process parameters did not exert a significant influence on the electrochemical performance. The cells with cathodes cut at constant line energy at reduced speed (V4) showed greater cycle stability after 350 cycles than the V3 and V5 cells. Despite the optimized process control, the cycle stability of the cells V4 was below that of V2.

The cells V3 to V5 showed a greater capacity drop than the cells from the series V2, although they had a small chamfer width and no heat-affected zone. This was an indication that in addition to the previously known and analyzed product features, another, previously unrecognized feature, influenced the electrochemical performance.

**Figure 13.** Capacity fading during the long-time cyclization of the investigated electrode/ cell configuration.

#### *3.3. Further Investigations of the Cutting Process and the Electrode Surface to Explain the Electrochemical Behavior*

In addition to the physical characteristics of the cutting edge, it was found that the different parameters led to a different degree of flying sparks, as shown in Figure 14. This could be recorded by imaging techniques. The sparking could lead to contamination of the electrode and thus affect the electrochemical performance. The images showed that with increased pulse repetition frequency (V0 → V3), the distance of the spark flight increased by 2.5 times. This could be explained by the increased number of hits if the cutting speed remained the same. If the number of hits on the active material, as well as on the solid and molten collector, increased, the number and acceleration of the ablation products would increase too. A reduction of the cutting speed with constant line energy (V4) showed a strong reduction of the sparks. The lower level of the spark formation was due to the lower energy and intensity per pulse, as well as the reduced travel speed. The profile of the sparkling flight was very similar to the sparking profile at low pulse repetition frequency and traversing speed (V2). Since variation V2 was cut with a lower pulse repetition frequency than variation V4, this could lead to less contamination of the electrode and thus explain the impairment of the performance of V4 compared to V2.

In addition to the assessment of the flying sparks, the recordings could strengthen the assumptions made in the previous sections concerning the formation of the plasma. A qualitative comparison of the images in Figure 14 showed that V0, V2, and V5 showed a clearly purple-colored plasma formation, whereas, for the other parameters—V3 and V4, the plasma presumably lied below the ablation gas phase, and thus was significantly smaller. This also explained why at shorter pulse lengths or higher pulse peak powers, the material removal and the formation of the HAZ at a constant energy density was greater.

**Figure 14.** Optical process investigations regarding the flying sparks.

The change in spark travel might lead to a change in the degree of contamination of the electrodes and, thus, in addition to the cut edge quality, adversely affect the electrochemical performance. The contaminations of the electrode were analyzed by SEM/EDX images and are shown in Figure 15. Contamination products in the form of metal spatters could be identified on all electrodes tested. The reference cathode had the strongest metal spatter contaminations. The lowest contamination occurred with the parameter variations V2 and V4.

These results showed that fewer sparks formation led to less contamination of the electrode surface in the area of the cutting edge. Based on the results of the electrochemical diagnosis and the analysis of the cutting edges, it could be assumed that the contaminations on the cathode surface exerted a greater influence on the electrochemical performance than the heat-affected zone and the chamfer width.

 

**Figure 15.** *Cont*.

**Figure 15.** SEM/EDX analysis of the cathode surface next to the cutting edge: Contamination of the surface by molten aluminum splashes are marked turquoise: 1. Electrode surface, 2. Cut zone.

## **4. Conclusions**

By means of a pulsed nanosecond fiber laser, electrodes could be cut without significantly a ffecting the cell cycle stability. The influence of the laser parameters or the cutting edge showed only a small influence on the examined variations. Considering a linear behavior, the parameter V1 led to a theoretical number of cycles of 1670 until the cells reached a state of health (SOH) of 80%. The best parameter V2 resulted in a cycle number of 1760. Since the capacity drop tended to decrease, it could be assumed that the cells would also reach higher cycles up to a SOH of 80%.

The examined characteristics—chamfer width and heat-a ffected zone—could be adjusted by means of the pulse duration, the pulse repetition frequency, and the cutting speed. In principle, the results showed that the chamfer width decreased with decreasing energy density or with increasing cutting speed. Furthermore, with a constant pulse duration and energy density, higher pulse repetition frequencies could result in a smaller chamfer width. This was probably due to the fact that both the decrease of the energy and the decrease of the intensity led to a geometric shift of the removal and the thermal load threshold. This shift was probably explained by the Gaussian intensity profile and the corresponding slope of the intensity in relation to the spot size, as well as by the intensity and energy-dependent formation of the plasma. The experiments with di fferent pulse durations with the same pulse energy showed, for di fferent pulse repetition frequencies and constant energy density, that the material removal and the thermal load, among the pure laser-material interaction, depended on the plasma or the intensity. The analysis of the size of the heat-a ffected zone revealed that it was essentially influenced by the pulse repetition frequencies. The cause of this dependency could be explained by the

same assumptions that have been made previously for the formation of the chamfer width. Both the anode and the cathode showed similar tendencies. In general, the anode could be cut much faster than the cathode due to the double material thickness of the collector, so that the parameter space for the model development in terms of speed for the cathode turned out smaller than for the cathode.

Investigations to evaluate the influence of the chamfer width and the heat-a ffected zone on the electrochemical performance of large-sized multicompartment pouch cells showed that the HAZ had a greater influence than the chamfer width. Furthermore, the results showed that contamination products in the form of metal spatter influenced the electrochemical performance more than the width of the chamfer and probably also the HAZ. It could be shown by imaging techniques that with very low speeds and high intensities and pulse energies with a moderate number of hits, metal spatters could be reduced.

Furthermore, it was possible to reduce the melting spatters even for a high number of hits per surface increment with a low traversing speed, intensity, and pulse energy. The results showed, so far, that the formation of the metal spatters was a function of the number of hits per surface increment, the speed of travel, the intensity, and the energy per pulse.
