**1. Introduction**

Due to continuing human-induced CO2 emissions, the global warming of the earth and associated negative consequences of extreme weather are steadily increasing [1]. The impact of climate change has increased awareness of the population in industrialized countries of environmentally friendly or carbon-neutral behavior. These interests are driving the development of new, greener, and more efficient technologies for the major CO2 producing sectors, which are electrical energy production and individual mobility. To reduce CO2 emissions in the individual mobility sector, electric mobility is considered a key technology [2,3]. In order to maintain the positive CO2 balance of electric mobility, the use of electrical energy from renewable energy sources is required, but also the production of cells must be made more e fficient and, therefore, more ecological. In particular, a more e fficient large-scale cell production can be achieved by reducing rejects by optimizing existing or developing new technologies, as material costs dominate over operating and investment costs [4].

Lithium-ion pouch cells are considered to be the most e ffective electrochemical technology. Due to their advantages regarding the high volumetric utilization of the installation space of the battery pack, they are especially well suited for automotive batteries. Because of the possibility to customize the cell geometry, the pouch cell leads to less dead volume compared to conventional cylindrical 18650 cells [5]. Besides, the high volumetric energy densities on battery level pouch cells have high volumetric energy of 466 WhL−<sup>1</sup> and a specific energy density of 241 Whkg−<sup>1</sup> on cell level (Cathode: LG Chem. NCM 111, Anode: LG Chem. Graphite) [6]. Even though cylindrical 18650 cells have a volumetric energy density, which is about 20% higher than those of pouch cells [5], the homogeneous mechanical behavior during charging and discharging of pouch cells lead to longer cycle lifetime, which is to be considered a long term advantage over the higher energy density [7]. For pouch cells, in general, it is fundamental to cut the endlessly coated electrodes and make them suitable for the following stacking process. It is imperative to look more closely at the cutting process itself since each cut could lead to manufacturing errors in terms of contamination and cutting edge quality [8]. Due to a large number of processes, it is necessary to optimize every step to improve ecological productivity. Figure 1 shows the evolution of the accumulated production rejects of a conventional cell production line schematically.

**Figure 1.** Influence of the reject rate of every single process on the total reject rate of a cell production line.

As shown in Figure 1, each process can contribute to the overall e fficiency of the production line. Due to the numerous processes of battery production, even small reject rates can lead to a high overall reject rate and, therefore, to low utilization of raw materials. Already at a reject rate of only 1% per process, the accumulated rejects at the end of the line reach almost 14%. Therefore, each process should be at least in the range of a 4σ reject rate, respectively, with a reject rate of under 0.09% with regard to ecological production. Investigating and improving the singulation process can contribute to accomplishing these high standards.

Besides the established shearing or die-cutting (DIN 8588) [9], laser cutting is becoming increasingly common in cell production lines due to its process-immanent advantages over the contact-based singulation method [10]. Especially for very thick and fragile electrodes, or pure lithium metal anodes, laser cutting is no longer an alternative but state of the art technology due to the lack of contact and the associated lack of mechanical stress [11]. In the following section, we have given a summary of the numerous studies in this field and laser ablation generally.

#### *State of the Art Laser Cutting of Electrodes*

The interactions between the material and the laser-generated photons during the cutting process is very dynamic and very complex due to a large number of possible and partly mutually dependent influencing factors. The key factors of a laser cutting plant are wavelength (λ), average power (*Pavg*), spot size *dspot*, laser profile and Rayleigh length in focus, cutting speed (*vc*), the number of passes, and cutting angle. In the case of pulsed laser beam sources, the additional factors are pulse peak power *Ppeak*, pulse energy (*EP*), pulse repetition frequency (*PRF*), and pulse shape, as well as pulse duration (τ). The relevant material properties of the electrode are the composition of the coating (anode/cathode), collector material, coating thickness, collector thickness, absorption coefficient, and degree of compaction of the electrode (Figure 2).

**Figure 2.** Relevant primary process parameters and material properties of a continuous wave (cw) and pulsed laser cutting process for electrodes [10,12].

For a simplified description of the influence of the primary process parameters in Figure 2, the secondary parameters energy density (ED) [13], intensity (*I*), pulse fluence (*Hp*), and number of laser pulses per surface increment (*nline*) [13] for pulsed beam sources can be used to explain the ablation or cutting behavior. Considering a certain wavelength, average power, and spot size, these parameters can be adjusted by *vc*, *PRF*, and τ. The secondary parameters are defined by the following equations:

$$ED = \frac{P\_{avg}}{v\_c \, d\_{spot}} \left[\frac{j}{cm^2}\right] \tag{1}$$

$$I = \frac{4\ P\_{\text{avg}}}{\pi \, d\_{\text{spot}}^2} \text{ for cw and } I\_P = \frac{4\ P\_{\text{Peak}}}{\pi \, d\_{\text{spot}}^2} \text{ for pulsed laser with } P\_{\text{Peak}} \approx \frac{P\_{\text{avg}}}{PRF} \left[\frac{W}{cm^2}\right] \tag{2}$$

$$H\_p = \frac{E\_P}{\pi} \left[ \frac{j}{cm^2} \right] \tag{3}$$

$$m\_{\rm lin\varepsilon} = \frac{\text{PRF}\left(d\_{\rm spot} + \text{ }\pi\text{ }\upsilon\_{\rm c}\right)}{\upsilon\_{\rm c}}\tag{4}$$

The energy density describes the energy input per surface increment (cutting length times const. spot size) on the material to be processed, depending on the cutting speed, photonic power, and spot size. This parameter can be used to describe the scalability of the cutting speed for a defined laser/scanner system and material or to define the necessary energy input for a quality cut and the cut-through limit. Since the energy density represents the average power input independent of the peak power, the intensity is necessary to further describe the cutting process with a pulsed laser beam

source. In addition, the pulse fluence and the number of pulses, which hit per surface increment, must be specified. Based on the energy density, the intensity, the pulse fluence, and the number of hits per area increment, the amount of material removal and the material removal behavior can be derived. Figure 3 schematically shows the number of hits per surface increment as a function of the *vc*, *PRF*, τ, and the *dspot*, as well as the effects on the electrode cutting edge characteristics. The ablation thresholds *a* and *b*, as well as the heat-affected zone *c* shown in Figure 3, can be derived from the mentioned secondary parameters.

**Figure 3.** Laser scanning microscope image of a laser-cut electrode (**left**); Ablation thresholds depending on the number of hits, intensity, pulse fluence, energy density, as well as the laser- and corresponding plasma-intensity-profile (**right**): (a) cut-through threshold, (b) ablation threshold, and (c) thermic influencing threshold [10].

As the type of energy input can lead to different removal mechanisms, a comparison of continuous wave (cw) and pulsed systems based on the energy density is only conditionally possible. Here, a distinction can be made between thermal (cw/pulsed) and athermal (pulsed) dominant removal processes. The thermal ablation generally proceeds in three successive phases, and, in the case of a continuous cut, also in parallel phases. In the first phase, the photons are absorbed by the surface and penetrate a near-surface area. In the following second phase, the temperature increases at the surface as a result of the absorption, and in deeper zones by thermal conductive effects. The rising temperature leads to a transformation of the state of matter from solid to liquid and liquid to gas, or directly from solid to gas. In the third phase, the penetration depth increases and thus the melting or evaporation zone, wherein the material is expelled in liquid or gaseous form from the kerf. The material ejection can be further distinguished into fusion, sublimation, and photochemical cutting/ablation. The athermal removal process is characterized by the fact that the duration of the photonic energy input is too short for initiating heat conduction. This removal process can be achieved with pulse durations of less than 10 ps. Furthermore, due to the short exposure time of pulses in the ps and fs range, the spatial extent of the resulting plasma is negligibly small [14]. Considering the high intensities, athermal processes can be further distinguished between sublimation cutting and photochemical ablation.

Figure 4 shows the processes that occur after the first impact of the photons on the processed material. On average, the photons are absorbed in 10 fs, where the photonic energy is converted into thermal energy within 100 fs by electron-electron relaxation of the electrode systems of the covalent bonds. This is followed by an electron-phonon relaxation after about 1 to 10 ps, which leads to a heat transfer of the electrons into the lattice structure. Parallel to this process, the ablation begins, and after about 100 ps, the phonon-phonon relaxation leads to heat conduction [15].


**Figure 4.** Time-scaled processes during a laser pulse material interaction, based on [16,17].

By means of the beam source used in this study, it is possible to generate pulses in the ns range. Due to the relatively long pulse duration of 240 ns, we can only realize a thermally affected cutting process (thermal removal process) with this system [14]. As a result of the composite structure and the associated varying material properties over the thickness of the electrode, it can be assumed that several removal mechanisms take place simultaneously and/or serially during cutting. The cutting of the porous coating will be characterized by a photochemical and sublimation proportion, whereas the cutting of the metallic collector will be characterized by sublimation and a fusion proportion. In addition to the direct interaction between laser and material, a plasma which correlates to the intensity distribution of the laser will also interact with the electrode. This can lead to material removal or thermal loading of the active material, in addition to thermal conduction effects.

The laser-induced plasma is caused by the high intensities of the individual pulses, which leads to strong oscillations of the free electrons, enabling them to knock bounded electrons out of neutrally charged atoms. The avalanching increase of free and strong oscillating electrons leads to a large number of free electrons and positively charged species. The resulting high-energy plasma absorbs the photons of the laser radiation by the inverse Bremsstrahlung (IB) and the photoionization (PI). In this case, IB is considered to be the main absorption mechanism. The photons in this mechanism are absorbed by free electrons as they collide with neutral or ionized species. This leads to an increase in the energy of the electrons and thus to an increase in the degree of ionization and the temperature of the plasma. If the temperature and density of the plasma rise above a certain level, the plasma can shield the area from being cut from the laser radiation. This effect is referred to as the plasma shielding effect [18].

Investigations on cw and pulsed laser beam cutting of electrodes have already been carried out in previous studies. The results showed that the use of single-mode cw fiber lasers made it possible to achieve very high, industry-relevant cutting speeds due to the high average power and the achievable low spot sizes. Studies showed that it was possible to cut an anode (120 μm) with 11,666 mms<sup>−</sup><sup>1</sup> and a cathode (130 μm) with 10,000 mms<sup>−</sup><sup>1</sup> with a single-mode cw fiber laser at a wavelength in the infrared range (1070 nm), an average power of 5000 W, and a spot size of 25 μm [19]. This resulted in an energy density of 2000 jcm−<sup>2</sup> for the cut-through limit of the cathode and an energy density of 1714 jcm−<sup>2</sup> for the anode. The lower required energy density for the anode was probably due to the low collector thickness and the lower degree of compaction. In another study, an anode (50 μm) with a collector thickness of 30 μm was cut with a cutting speed of 2000 mms<sup>−</sup><sup>1</sup> using a single-mode cw Ytterbium fiber laser (1070 nm, 250 W, and 23 μm spot size). The much lower required energy density of 543 jcm−<sup>2</sup> could be explained by the lower material thickness and by the higher intensity [20]. Based on these results, further investigations using the same system showed that compacted (56 μm) and non-compacted anodes (70 μm) with a collector thickness of 20 μm could be cut at a speed of 5000 mms<sup>−</sup><sup>1</sup> [21]. The low energy density of 217 jcm−<sup>2</sup> required for the singulation suggested that the collector thickness was

the dominating speed-influencing parameter of the electrode. Considering the strong influence of the collector thickness, as well as the intensity of the focused laser spot, it was plausible that an energy density of 818 jcm−<sup>2</sup> at 5000 mms<sup>−</sup><sup>1</sup> led to a cut-through of an anode (100 μm) with a current collector thickness of 10 μm by using a single-mode cw fiber laser (1070 nm, 450 W, 11 μm spot size) [22]. In principle, the investigations with cw laser systems showed that the achievable chamfer width (total ablated area) was less than 50 μm and increased with increasing energy densities [21].

The trend of these dependencies was also found in investigations with pulsed laser beam sources. In addition to this, the investigations showed that with pulsed beam sources in the nanosecond range, higher pulse repetition frequencies enabled higher cutting speeds and led to a smaller chamfer width [19]. By using an ns pulsed fiber laser (1070 nm, 100 W, 50 μm spot size, 500 kHz, and 30 ns), it was possible for Kronthaler et al. [23] to cut an anode (114 μm) with a collector thickness of 10 μm at a speed of 1200 mms<sup>−</sup>1. Using the same system parameter, a slightly higher cut-through limit of 1250 mms<sup>−</sup><sup>1</sup> could be achieved for a 124 μm thick cathode with a collector thickness of 20 μm [23]. From the given parameters, an energy density of 160 jcm−<sup>2</sup> resulted, for the cut-through limit, in an intensity per pulse of 3.4 × 10<sup>8</sup> Wcm−<sup>2</sup> and a hit number of 20. Lutey et al. showed in their research that the number of hits per area increment caused an increase in the plasma shielding effect. With higher numbers of hits, a higher average power was needed to realize a cut. At 125 hits per area increment (500 kHz), an energy density of 660 jcm−<sup>2</sup> was needed, whereas, with a hit number of 5 (20 kHz), only an energy density of 352 jcm−<sup>2</sup> sufficed [24].

Further results on laser cutting of electrodes showed that low energy densities and intensities were necessary for singulation for a smaller wavelength. At a wavelength of 1064 nm, an energy density of 448 jcm−<sup>2</sup> and an intensity of 28.5 × 10<sup>8</sup> Wcm−<sup>2</sup> were necessary [24], whereas, at 355 nm, an energy density of 340 jcm−<sup>2</sup> and an intensity of 1.7 × 10<sup>8</sup> Wcm−<sup>2</sup> were sufficient to cut a 120 μm thick anode [19]. These results could not be confirmed by studies with a laser in the green electromagnetic spectrum (532 nm, 1 ns, 6 W). For the singulation of an anode (130 μm), with the same collector thickness, an energy density of 560 jcm−<sup>2</sup> and an intensity of 50.8 × 10<sup>8</sup> Wcm−<sup>2</sup> were needed. This could be attributed to the fact that the number of hits per surface increment of 165 increased the plasma shielding effect, and thus reduced the maximum cutting speed [25].

In order to realize the most dynamic and fast cutting processes possible in a cutting plant, the beam guidance on the workpiece is conventionally carried out by means of a remote scanner system. As a result of the low mass and the associated low inertia of the deflection mirrors, very high speeds and repetition accuracies can be achieved in the horizontal plane. Due to the varying distance of the electrode to the scanner system, a focus adjustment in the vertical direction is necessary. This adjustment can be produced by means of additional lenses (focus-shifter) or by a static f-theta objective. Considering the static beam refocusing, the advantage of an f-theta objective over a focus shifter is the higher repetition accuracy and wear-freedom. However, the dynamic focus by means of a focus shifter allows us to adjust the working field and to customize the focus in certain areas.
