Establishing laser cutting of components for sulfide-based solid-state batteries

Lithium-ion batteries (LIBs) have been established as the predominant electrochemical energy storage technology in portable electronic devices and battery electric vehicles.¹ The success of this technology was enabled not only through the advances in energy and power densities but also through the advancements in its production, leading to lower costs and high production capacities. As the conventional LIB technology has matured and the materials’ potentials are utilized almost to their theoretical limits, research on alternative technologies for further advancements is increasing.² Solid-state batteries (SSBs) are widely regarded as the next-generation battery technology.³ The use of a solid-state electrolyte instead of the currently used liquid electrolyte promises higher power and energy densities by enabling the use of novel anode active materials such as lithium metal or silicon.³ 

While the advantages of SSB technology have been proven at laboratory scale with small capacity batteries, the large-scale production of SSBs has been only scarcely researched. To enable the large-scale production of SSBs, a “drop-in” technology that is compatible with currently existing and commissioned LIB production facilities is crucial.⁴ One of the most promising technologies is sulfide-based SSBs.⁵ Sulfide-based SSBs can be manufactured similarly to LIBs, making them a compelling candidate from an industrial standpoint. However, closer investigations are needed at each processing step.⁴ One of these under-investigated processing steps is the singling step, the cutting of components from a sheet or coil by mechanical cutting or laser cutting.

The typical components of a sulfide-based SSB with their respective thicknesses are shown in Fig. 1. The specific requirements for the laser cutting of these components, considering the environment, materials, and the product itself, are detailed in the following.

Both the composite cathode and the solid electrolyte (SE) separator contain sulfide SE, which is known to be highly reactive in contact with water. Upon water or moisture contact toxic hydrogen-sulfide gas (H₂S) is formed and the ionic conductivity is decreased.⁶ In order to prevent the described degradation, the here presented study on the laser processing of sulfide-based materials took place within an inert moisture free argon atmosphere.

An industrial laser cutting process could be implemented within a dry room, similar to those currently used for LIB production. As a typical sulfide-based SE will still degrade, however slowly, in a dry room atmosphere, the exposure time of the material and the dew point of the environment must be controlled and tracked. By adhering to these constraints, the material's degradation can be kept within acceptable limits set by the manufacturer.⁷ 

Not only the reactivity of the sulfide-containing components can pose a challenge for the cutting process, sulfide-based composite cathodes contain multiple different materials; an active material, a conductive additive, the SE, and a binder. While the active material, the conductive additive, and the binder in sulfide-based SSBs are similar to the currently used materials for conventional cathodes in LIBs, the SE, which constitutes about 30 wt. %, is anticipated to significantly alter the laser absorption of the cathode and therefore its behavior during laser processing. Multimaterial systems, such as the composite cathode, pose a challenge for laser processing due to the possibly widely different ablation characteristics during the cutting process.

The laser cutting of the sulfide-based SE separator is anticipated to be less challenging compared to the composite cathode, as the separator is mainly made up out of one material, the SE with typically over 95 wt. %, the rest being a binder. However, given the unknown laser ablation behavior of the SE, the feasibility of laser cutting SE separators needs to be investigated.

Sulfide-based SSBs are candidates for the utilization of lithium metal anodes⁸ or the promising concept of silicon anodes.⁹ Both the laser cutting of lithium metal¹⁰ and of silicon anodes¹¹ have been successfully demonstrated in the past, which is why this publication does not explore it further.

As their name suggests, SSBs do not contain a liquid electrolyte, unlike their conventional LIB predecessors. Ion movement between components in SSBs relies on solid-to-solid contact.¹² Therefore, to achieve low ionic resistances between components and to ensure high charge and discharge performance, it is essential to have homogeneous surfaces allowing a large contact area.

Apart from homogeneous surfaces, the solid-to-solid contact in SSBs is typically enhanced through operating the batteries under high pressures in the range of megapascals, allowing the materials to deform and increase their contact area.¹³ The formation of elevated contours or large particles, such as debris on a component's surface, during the cutting process of SSBs can result in the penetration of the SE separator due to the high operation pressures. This, in turn, can lead to the failure of the battery.

Mechanical punching, especially as used in laboratories, tends to produce burrs, motivating the use of potentially burr-free laser cutting for both industrial and research applications.

As the production routes for solid-state batteries are not yet defined and will likely change depending on the desired battery characteristics,⁴ the studies presented in this work have been conducted mainly on uncompressed cathodes. Uncompressed components generally are thicker and have a higher porosity compared to the compressed components in the final battery. Both of these properties increase the difficulty of laser cutting.¹⁴ The feasibility of cutting compressed components was also investigated within this publication. This ensures the transferability and validity of the process studies across different SSB production chains, where singling might occur either before or after compression.⁴ 

The investigated composite cathodes consisted of the active material LiNi_0.6Mn_0.2Co_0.2O₂ (NMC 622), the binder polyiso­butylene (PIB), the conductive additive carbon black (C65), and the solid electrolyte lithium phosphorus sulfur chloride (LPSCl). The respective shares of the materials were approximately 66 wt. % NMC, 3 wt. % PIB, 3 wt. % C65, and 28 wt. % LPSCl. The composite cathodes were coated on both sides of an aluminum current collector using a solvent-based approach. Compressed composite cathodes were obtained by calendering with a pilot-scale calender in an argon atmosphere (as described in Ref. 4).

Due to the high cost of the sulfide solid electrolyte, the bottom coating (cf. layer 3 in Fig. 1) was substituted with a typical LIB cathode coating without a solid electrolyte. This was done to mimic a cathode as it would be implemented in an industrialized multilayer SSB. To ensure the validity of the bottom layer substitution, the validation of the best parameter set was conducted using composite cathodes fabricated with two identical sulfide-containing layers.

The separator was manufactured following a solvent-based route and contained approximately 92 wt. % LPSCl SE and 8 wt. % PIB binder. Laser cutting of uncompressed separators was not investigated, as the components were too fragile to be removed from the coating substrate (siliconized polyester foil) before compressing. Compressed SE separators were obtained by calendering with a pilot-scale calender in an argon atmosphere (as described in Ref. 4) and were removed from the substrate foil before laser processing.

The thicknesses of the investigated components are given in Fig. 1. Further details on the materials cannot be given due to confidentiality agreements.

Process studies were conducted by cutting line geometries defined in the laser processing software (RAYGUIDE, RAYLASE, Germany). Parallel lines 8 mm in length with a 1 mm spacing were cut [cf. Fig. 2(a)]. When multiple passes were investigated, the laser cut started from the same spot on each pass. Validation studies were conducted with the final component’s geometry [cf. Fig. 2(b)]. The laser beam was focused on the samples surface for each cut. The samples were centered in the scanner's working area.

Sample manufacturing and processing were conducted in gloveboxes with an inert argon atmosphere. Oxygen and water contents were monitored and were continuously below 2 ppm.

As ultrashort pulsed laser sources allow for cleaner laser cuts compared to constant wave lasers,¹⁵ a picosecond pulsed ytterbium fiber laser (YLPP-25-3-50-R, IPG Photonics, USA) with a wavelength of 1030 nm delivering an average power of 50 W was used for the experiments. The laser beam source was located outside of the glovebox, and the beam was guided into the box by an optical system. Inside the glovebox, a 2D-scanning optic (Superscan IV, Raylase, Germany) was used to deflect the laser beam. The experiments were performed on the same system as described in detail in Ref. 16. The key performance parameters of the laser setup are listed in Table I.

To assess the quality of the performed cuts, both a laser scanning microscope (LSM) (VK 9710, Keyence, Germany) and a scanning electron microscope (SEM) (IT-200, JEOL, Germany) were used for qualitative as well as quantitative assessments. Qualitative properties investigated included the melt elevation at the cut edge, the heat-affected zone, the amount of weld spatter, and the continuity of the achieved cuts (cf. Fig. 3).

The quality assessments were conducted based on representative sections on the surface of the sample oriented toward the laser source, as the bottom surface typically exhibits fewer or no defects. The respective cuts were inspected in their entirety at low SEM magnification (50×), and a representative section was selected for defect categorization at higher SEM magnification (200×). Quantitative measurements included the ablation depth, in case the laser did not separate the sample completely.

Analysis with the LSM and SEM was conducted in a dry room with a dew point below −40 °C. The samples were transferred to the dry room in gas tight containers and only exposed to the dry atmosphere for the duration of the measurements (<2 h) to keep possible degradation to a minimum.

Preliminary studies were conducted at the different pulse repetition rates (PRRs) between 201 and 1838 kHz (cf. Table I). The studies indicated that low PRRs were unsuitable for the cutting task, as they did not deliver sufficient energy to the material at reasonable (v_s≥ 0.1 m s^–1) scanning speeds. To investigate the influence of the PRR on the quality of the cuts, the two highest PRR levels, 1003, and 1838 kHz, were selected for further studies.

To identify suitable discrete power levels, single passes were made at the two specified PRRs (1003 and 1838 kHz) at different scanning speeds (0.1 m s^–1 ≤ v_s ≤ 2.0 m s^–1) and three different peak fluences (3.12, 4.16, and 5.20 J cm⁻²), the resulting ablation depth was analyzed with an LSM.

Process window studies were conducted to assess the feasibility of laser cutting composite cathodes. In these studies, the scanning speed, the peak fluence, the number of laser passes, and the PRR were systematically varied to gain insight into their influence on the cut.

First, the individual cuts were classified according to the layers of the composite cathode (cf. Fig. 1) that were cut. The assessment was performed based on the visual inspection of SEM images.

Achieving separation of the desired geometry from the sheet or coil is required, but a high edge quality is essential as well, as previously described. To investigate the edge quality, the degree of created debris around the cut was categorized based on visual inspection of SEM images. If no debris originating from the laser cut could be seen in the SEM image, the cut was categorized as no debris. If sporadic debris was found, the category little debris was assigned, if a large amount of debris was observed, the label significant debris was assigned.

In addition to the cut completeness and debris categorization, the presence of cracks in the vicinity of the cut edge was analyzed based on visual inspection of SEM images. If no cracks on the surface of the sample were seen, the parameter set was categorized as no cracks. The differentiation between the categories small and significant cracks was done based on the width, length, and number of the observed cracks.

To estimate the best overall parameter set within the conducted experiments, a quality index (QI) was defined. The defined quality levels were assigned penalty values that increase with defect severity (cf. Table II). The individual values were multiplied by a weighting factor of 1 for both debris (QI_debris) and crack defects (QI_cracks), and a factor of 10 for the cut completeness (QI_cut) to reflect the importance of complete separation. The specific value of 10 allows for the easy differentiation of the various defects; overall QI values below 10 indicate full cuts with varying degrees of defects while parameter sets with QI values of 11 or 12 resulted in a cut with few defects but incomplete separation. These parameter sets may, however, be suitable for components of lesser thickness, enabling high cut quality at high processing speeds.

The best identified parameter set for cutting composite cathodes was applied to create the final cathode geometry for a lab-scale pouch cell. Additionally, a cathode of identical geometry was produced using mechanical stamping. Both compressed and uncompressed composite cathodes were cut. The utilized mechanical stamps were punched against a flat polyethylene board. This method is typically used for pouch cell production in LIB research; however, it is not representative of an optimized mechanical process as described in Ref. 17.

The studies presented in this publication focused primarily on the laser cutting of the composite cathodes. However, the gained understanding of the laser behavior of the sulfide-based composite cathodes was used to demonstrate the feasibility of laser cutting sulfide-based SE separators.

The studies indicated that even at the lowest scanning speeds, a moderate peak fluence of 3.12 J cm⁻² resulted in ablation depths lower than the first material layer (≈140 μm, cf. Fig. 4). Therefore, lower power levels were not considered. Full separation could only be achieved at low speeds (v_s ≤ 0.3 m s^–1), high fluences (F₀ ≥ 4.16 J cm⁻²), and at the higher PRR of 1838 kHz.

Figure 5 illustrates the categorization of the investigated parameter sets of the individual cuts based on the cut completeness as a process map. The data show that at a lower PRR of 1003 kHz, no complete cuts were achieved with a single pass. At a higher PRR of 1838 kHz, complete cuts were achieved at high fluences (F₀ ≥ 4.16 J cm⁻²) and low scanning speeds (0.1 m s^–1 ≤ v_s ≤ 0.2 m s^–1).

Full cuts at high scanning speeds could only be achieved by employing multiple laser passes at the maximum available fluence (F₀=5.20 J cm⁻²). Five laser passes (n = 5) enabled cutting speeds of up to 0.5 m s^–1 at a PRR of 1838 kHz, while speeds of up to 1.6 m s^–1 were achieved at the same PRR with 10 passes (n = 10).

At low speeds (v_s ≤ 1 m s^–1), debris could be found for every power level, PRR, and number of passes (cf. Fig. 6). Higher scanning speeds led to a smaller amount of defects.

The lower investigated PRR of 1003 kHz led to a significantly lower amount of debris across all parameter sets. An increase in debris was observed with an increasing fluence.

The process map on cracking of the samples indicates the significant increase in cracks with the increase in laser passes (cf. Fig. 7). Both PRR and peak fluence had a lesser impact on crack formation. However, slower speeds and higher fluences also contributed to cracking.

The cracking is, therefore, believed to result from the heat impact of the laser beam on the material and especially the repeated temperature changes the material undergoes during multiple laser passes.

The weighted quality index values were summed to establish a baseline for the evaluation of the overall cut quality (cf. Fig. 8). Parameter sets that did not achieve a full separation of at least the first material layer (cf. Fig. 1) were not considered. These boundary conditions combined with a quality gate of QI ≤ 2 led to the identification of a narrow process window with high quality cuts (cf. Fig. 8). A manual comparison based on SEM analysis between these parameter sets was conducted, and the parameter set F₀ = 3.12 J cm⁻², PRR = 1838 kHz; n = 5; v_s= 0.3 m s^–1 identified as leading to the best overall cut quality (cf. Fig. 9). This parameter set was, therefore, used for the validation studies.

The laser-cut and mechanically stamped composite cathodes were compared using SEM images (cf. Fig. 9). Both the compressed and uncompressed laser-cut components exhibited an elevated cut edge on the top surface, while the bottom surface remained completely flat [cf. Figs. 9(a) and 9(c)]. The cross section of the cut shows a small amount of melted material [cf. Figs. 9(a) and 9(c)].

The surface of the laser-cut uncompressed cathode exhibited cracks, as visible in Fig. 9(b), with a length of approximately 170 μm and debris in the vicinity (<100 μm) of the cut edge. No cracking was observed in the compressed cathode [cf. Fig. 9(f)]. In contrast, the uncompressed mechanically stamped composite cathode showed significant cracks parallel to the cut edge [cf. Fig. 9(d)].

The mechanical force of the stamping tool deformed the edge macroscopically, leading to the visibility of parts of the layer 1 surface in the cross sections [cf. Figs. 9(c) and 9(g)].

Delamination can be seen in both mechanically cut specimens; however, it is significantly more pronounced in the uncompressed cathode [cf. Fig. 9(c)]. The laser-cut components show no delamination defects.

Compressed SE separator samples were successfully cut with a parameter set of F₀ = 3.12 J cm⁻², v_s = 1.0 m s^–1, PRR = 1003 kHz, and n = 5. However, the cut edge shows both cracking and melt elevation (cf. Fig. 10). As the separator of a battery typically has a larger area than the electrodes, near edge defects are not expected to influence the electrochemical performance of the separator or the battery.

The studies demonstrated that the laser cutting process is suitable for the cutting of sulfide-based composite cathodes and solid-electrolyte separators. Comparing the laser-cut electrodes to mechanically cut electrodes showed that the laser cuts produced a significantly better edge quality for the SSB application, as they exhibited no macroscopic defects such as delamination or deformation. The comparison was performed using mechanical stamping without a counter punch, limiting the finding’s applicability to research applications.

While mechanical stamping with a negative counter punch would lead to better cut edges, such tools are prohibitively expensive during the research and upscaling of a novel battery design, where electrode geometries often need to be adjusted. Mechanical punching with negative counter punches also produces a burr and surface deformation, although less pronounced.¹⁷ 

The advantages of laser cutting SSB components are, therefore, clearly visible for research applications. A comparison between an optimized mechanical stamping system and an optimized laser system is necessary to better evaluate their potential for industrial applications.

Multiple possibilities can be explored to further improve the cut quality and processing speed of the within this publication presented laser cuts. It was demonstrated that high fluences at low speeds are required to achieve complete cuts at one pass. Higher speeds, however, are necessary to enable a higher cut edge quality, which presents a conflict of goals. As a possible solution, the number of laser passes was increased, which enabled cutting at higher scanning speeds; however, it decreased the cut quality further. Increasing the peak fluence again decreased the cut quality.

Increasing the pulse repetition rate positively influenced both the edge quality and the completeness of cuts. Therefore, initial efforts to improve cut quality should focus on using a laser system reaching even higher pulse repetition rates.

Using a laser source with a shorter pulse width could enhance cut quality further by allowing higher peak power, thereby enabling ablation with a reduced heat impact.¹⁸ In combination with a higher pulse repetition rate, complete cuts with single passes at high scanning speeds could be possible.

To further improve the cutting of thick composite cathodes, an optical setup with a higher Rayleigh length should be considered. If multiple laser passes are required, a 3D scanning optic could allow for a refocusing between passes.

To ensure that the laser-material interaction has no negative impact on the material’s properties during battery operation, electrochemical testing of laser-cut components should be considered.

The authors thank the Bavarian Ministry of Economic Affairs, Regional Development and Energy for funding the research project “Industrialisierbarkeit von Festkörperelektrolytzellen.” They also thank the German Federal Ministry for Economic Affairs and Energy and the Bavarian Ministry of Economic Affairs, Regional Development and Energy for funding the work in the BMW project “IPCEI EuBatIn” (No. 16BZF205). The authors thank BMW for their cooperation. A great thanks goes to Elena Jaimez Farnham, Alessandro Sommer, and Sophie Grabmann. Open Access funding enabled and organized by Projekt DEAL.

The authors have no conflicts to disclose.

Lovis Wach: Conceptualization (lead); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (lead); Project administration (equal); Validation (equal); Visualization (lead); Writing – original draft (lead); Writing – review & editing (equal). Yokubjon Khaydarov: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (supporting); Validation (equal); Visualization (equal); Writing – review & editing (supporting). Pawel Garkusha: Formal analysis (supporting); Writing – review & editing (supporting). Rüdiger Daub: Funding acquisition (lead); Project administration (lead); Resources (lead); Supervision (lead); Writing – review & editing (equal).