The shape of things to come

All-fiber kW beam shaping enables high-productivity industrial processing

Dahv Kliner & Alex Kingsbury

Published in PhotonicsViews - Apr. 2, 2024

The build rate, material quality, and spatter production in a laser powder-bed fusion tool depend on the laser beam size and shape (Source: F. Eibl, Aconity3D)

All-fiber beam shaping is revolutionizing laser-based manufacturing. This capability is a commercially accessible reality in cutting, welding, and additive manufacturing tools released by leading integrators worldwide. These advanced tools increase productivity and part quality and introduce entirely new production capabilities, driving the displacement of legacy lasers and non-laser technologies in existing applications and spurring the development of new markets.

Material processing is the largest segment of the laser market, with macro processing (kW-class laser powers) making up over half of this segment through more than 3 billion US dollars of annual laser sales. Macro processing has experienced more than two decades of nearly uninterrupted growth, driven by several laser advances, including increased power, efficiency, and reliability, and a steady decrease in price. One technology has dominated these advances: fiber lasers, which now comprise well over half of the macro processing market. The dominance of fiber lasers results from their significant practical and performance advantages, particularly at high power, because of their elimination of free-space optics. Free-space optics and their associated mounts, alignment fixtures, and cooling infrastructure result in substantial system complexity and cost, increase size and weight (especially problematic for process heads), cause optical loss, suffer from thermal lensing at high power, and introduce instability and failure modes because of their sensitivity to alignment, environmental conditions, and contamination. Fiber lasers minimize or eliminate all of these problems.

In recent years, another trend has emerged that is providing further laser performance advantages in existing applications and, importantly, enabling entirely new capabilities. This trend is beam shaping, specifically the delivery of laser beams with optimized, tunable spatial and divergence profiles.

It has long been recognized that the laser beam size, shape, and divergence determine the materials processing performance by controlling heat input into the workpiece, which in turn determines the feature size, part quality, production rate, process window, and process flexibility (e.g., materials compatibility). These metrics ultimately determine the tool productivity and versatility and thus the economics of the production process.

Most lasers have fixed beam characteristics, and various downstream methods have been developed to optimize and vary the beam shape. Most methods employ manually changeable or motorized free-space optics such as zoom process heads, fiber couplers with variable launch conditions, switchable optics, or deformable mirrors. These approaches have shown the value of being able to optimize and tune the beam characteristics, but they suffer from the known disadvantages of free-space optics.

More recently, a family of lasers has emerged that use all-fiber technology to provide a composite beam shape comprising a central spot surrounded by a ring. In this approach, different fiber lasers are coupled into different guiding zones (cores) in a fiber to provide the composite beam shape. While eliminating free-space optics, this technology is not tunable in the same sense as the previous free-space technologies because only one beam shape is available at full power. The laser power must be reduced in one of the guiding zones to realize other beam shapes. This design has proven advantageous in applications for which the desired beam shape is known and does not need to be significantly varied, such as welding in battery production lines.

Fig. 1 Typical beam settings for three-zone (left, 3 – 5 kW) and two-zone (right, 3 – 20 kW) CFX fiber lasers used in metal cutting and welding. The images show near-field spatial profiles, and the corresponding second-moment (D4σ) beam diameters are given on the y-axes.

Corona fiber lasers

nLight has developed an all-fiber technology, Corona, which allows real-time tuning of the beam shape with no loss of power [1]. Briefly, the output of a fiber laser is coupled (spliced) to a fiber that enables the beam characteristics (position, shape, etc.) within the fiber to be varied by perturbing the fiber, and the perturbed beam is ‘trapped’ in one or more separate guiding zones of a multicore output fiber to generate the desired beam shape.

The number and dimensions of the guiding zones in the Corona feeding fiber can be optimized for different applications. For example, Fig. 1 shows typical beam profiles for three-zone and two-zone feeding fibers in the CFX product family. In both cases, the diameter of the central core is 100 µm, and the outside diameter (OD) of the largest ring is 330 µm. For the three-zone fiber, the OD of the central ring is 215 µm. We have found that supplying predefined beam shapes, as shown in Fig. 1, is preferable to continuous tuning of the shape for process optimization and stability. Industrial laser sare often deployed in electrically noisy environments, in which analog control signals can be unstable on a variety of timescales. By providing defined, digitally selectable beam shapes (known as ‘Index’ settings), the end user is ensured that the beam characteristics will be stable for years. CFX fiber lasers are widely used in metal cutting and welding. The settings shown in the left panel of Fig. 1 are those typically used for cutting. For welding applications, it is advantageous to provide additional beam shapes, as shown in the right panel of Fig. 1.

Fig. 2 shows typical beam profiles for the AFX product family, which employs a single-mode (SM) input laser. In this case, the central core provides an SM beam (mode-field diameter ≈14 µm), and the surrounding annulus has an OD of 40 µm.

Fig. 2 Typical beam settings for AFX fiber lasers used in additive manufacturing and thin-sheet welding. The images show near-field spatial profiles, and the corresponding second-moment (D4σ) beam diameters are given on the y-axis.

Applications

Metal cutting

Metal cutting is the largest industrial market for high-power lasers, with laser revenue of >2 billion US dollars annually. This application was pioneered by CO2 lasers. By 2017, fiber lasers had largely displaced CO2 lasers because of their numerous advantages. In particular, for cutting thin sheet with N2 assist gas, fiber lasers provided significantly higher cutting speeds. CO2 lasers maintained an advantage, however, for one important application: cutting of mild-steel plate (thickness above ~6 mm) with O2 assist gas, where CO2 lasers offered significantly better edge quality than legacy, fixed-beam fiber lasers. As a result, job shops had to either accept poor performance for mild-steel cutting or maintain both CO2 and fiber laser tools.

The initial CFX products addressed this longstanding problem [2]. The smallest (100 µm) index setting provides cutting speed and edge quality similar to conventional (fixed-beam) fiber lasers with standard 100 µm feeding fiber, maintaining the expected benefits of fiber lasers. The larger and ring-shaped CFX beams, not available in legacy fiber lasers, provide a dramatic improvement in O2 cutting performance for mild steel, achieving cutting speed and edge quality that equal CO2 lasers. Furthermore, the intermediate CFX Index settings (Fig. 1) can provide better edge quality for stainless steel and aluminum than that obtained with conventional fiber lasers, even those at higher power. CFX fiber lasers have thus enabled highly versatile metal-cutting tools. End-users can now have a single tool that performs optimum cutting for a range of materials and thicknesses rather than tolerating poor performance for some jobs or supporting multiple tools.

Having addressed the problem of thick mild steel cutting with the large ring-shaped CFX beams, we next considered fiber designs to optimize N2 cutting performance for thin sheet metal (another large application). It is known that smaller beams provide faster N2 cutting speeds, and some tools employ a 50 µm rather than a 100 µm feeding fibers, although the maximum thickness is then further restricted. We redesigned the CFX feeding fiber to provide a 50 µm Index 0 beam for faster N2 cutting while maintaining the 330 µm annular guiding region for O2 cutting of mild steel, in a product known as Mach Ultra.

Fig. 3 Cutting speed for 3 mm stainless steel with N2 assist gas for a conventional fiber laser (100 µm feeding fiber) at 4, 5, 6, and 8 kW (grey bars) and for Mach Ultra (50 µm beam diameter) at 5 kW.

We validated the cutting performance of Mach Ultra for both N2 and O2 cutting [1]. Fig. 3 shows the N2 cutting speed for 3 mm stainless steel for a conventional 100 µm fiber laser at powers of 4 – 8 kW (grey bars). The green bar shows the cutting speed for 5 kW Mach Ultra (Index 0). The Mach Ultra cutting speed at 5 kW is equivalent to that of a conventional fiber laser at 8 kW, illustrating the dramatic benefits of delivering the laser power in an optimized format. The ability to use a lower laser power while maintaining the same cutting speed and edge quality provides a lower cost per manufactured part by reducing both the up-front cost (laser and chiller) and the operating costs (electricity, downtime for cleaning or replacing optics in the cutting head). As expected, Mach Ultra maintains excellent performance for O2 cutting of mild steel using the 330 µm donut beam.

Metal welding

Metal welding is the second-largest industrial market for high-power lasers, with significant growth driven by automotive applications. Lasers provide substantially higher throughput than conventional welding processes and can generate precise, low-heat-input welds, resulting in increased productivity and part quality. The high power and beam quality of fiber lasers enables ‘keyhole welding’ (so called because of the creation of a high-aspect-ratio vapor channel), which provides high speed and minimal distortion or heat-affected zone. Use of a larger spot size results in a shallower conduction weld, which is advantageous for aesthetic reasons and to minimize post-processing steps. Different beam properties at the workpiece are required to transition between these two welding modes and to optimize the weld quality for different materials and joint designs. CFX’s tunable beam shape (Fig. 1) allows the weld depth and width to be tailored, providing the best performance for each weld.

Fig. 4 Powertrain weld tests performed at the same speed, showing significant spatter reduction when using CFX. Spatter particles are visible both during the welding process (top photos) and in photos of the finished parts. (Source: Optoprim / nLight, first published in: Industrial Laser Solutions, p. 9, Sept./Oct. 2019)

In keyhole welding, spatter is a well known problem that can cause contamination of the finished part (requiring post processing and causing latent failures), a reduction in weld quality, and contamination of the tooling (increasing consumables and maintenance). To reduce spatter, the keyhole stability must be improved and the collapse of the vapor channel prevented. The Corona beam shape directs laser power where it is needed to open the keyhole, and the beam tunability provides minimum spatter for a wide range of materials and welding speeds. For example, Fig. 4 compares spatter generation when welding medium-carbon steel parts representative of an automotive powertrain using a standard beam shape and an optimized CFX beam shape.

Welding of thin metal (~50 µm – 2 mm) is becoming increasingly important for e-mobility applications, including the production of batteries and fuel cells, with high volumes demanding high welding productivity and quality. These applications require lower powers and smaller beams than the welding examples above, and SM lasers are usually used.

Fig. 5  Welding of 50 µm stainless steel bipolar plates for fuel cells. Both welds were performed at 1.3 m/s and 1 kW laser power, and the images show 3.4 mm of weld seam. The humping defect is evident in the SM weld (upper) and has been eliminated in the AFX weld (lower). (Source: D. Bartels, BLZ).

A typical fuel cell for an electric vehicle requires >1 km of weld seam in the bipolar plates within the fuel cell. Using an SM laser, the weld speed is limited to <1 m/s by potential humping defects, and increasing the laser power does not allow faster weld speeds. Work at both the Bavarian Laser Center (BLZ) and the Edison Welding Institute (EWI) has shown that AFX can increase the welding speed on stainless steel foil to at least 1.3 m/s without humping and while maintaining a narrow weld seam (Fig. 5). Similarly, AFX welding of high-strength aluminum stabilizes the melt pool depth and width, reduces hot cracking, decreases surface roughness and humping, and increases the process window [3].

EWI researchers showed that AFX can be used to control the weld penetration depth and interface width when welding 250 µm foils of aluminum to copper in a battery cell welding application. They also demonstrated that AFX can significantly reduce the spatter when welding aluminum and copper (key materials employed in battery production), similar to the CFX results shown in Fig. 4.

In the AFX welding examples above, the optimum beam shape depended on the material and the welding speed, highlighting the unique Corona benefit of maintaining full power for all beam shapes, thereby maximizing the versatility of the tool.

Additive manufacturing

In contrast to cutting and welding, additive manufacturing (AM), and specifically laser powder-bed fusion (LPBF), represents a smaller but rapidly growing market. The ability to fabricate parts that are lightweight, high in complexity, and easily customizable is essential for applications such as aerospace, medical, dental, and industrial. Since its inception in 1995, LPBF has emerged as the leading candidate for high-volume production of additively manufactured parts. CO2 lasers were used in early LPBF machines, but fiber lasers were quickly adopted due to their higher power density and reliability.

To date, the accepted standard in LPBF tools has been an SM fiber laser, although this approach has several disadvantages. At power densities that ensure complete melting across the melt pool, the high intensity peak in the center of the Gaussian beam translates to a heat input and corresponding temperature profile that is above the boiling point, causing melt-pool instability. When depositing alloys, the non-uniform temperature profile causes vaporization of the lighter, lower-boiling-point elements, such as magnesium and zinc, causing compositional variance in the fabricated part. Additionally, the melt pool experiences a Marangoni effect with the central downwards force of the Gaussian beam, leading to ejection of material from the melt pool into the gaseous processing environment. This material, observed as soot and spatter, is the leading cause of porosity and other defects in LPBF-fabricated components (similar to the effects discussed in welding above).

Fig. 6 The effect of beam shape on the melt-pool temperature profile and resultant material cross section (aluminum alloy). Tm denotes the melting point of the material.

Modifying the beam shape changes the temperature response within the melt pool and thus the cross section of the solidified LPBF track, as illustrated in Fig. 6. An SM beam shape results in a strongly peaked temperature profile, which can result in material ‘balling’ or ‘humping’ (as well as porosity, soot, spatter, and compositional variability) [4]. The situation is improved using a flat-top beam shape, but the temperature profile still exhibits a central peak. Transferring power from the center of the beam to the periphery allows the temperature profile to be flattened, as shown for the ring beam [5]. The ideal beam shape is material dependent, and the ability of AFX to vary the power distribution between the central SM spot and the ring (Fig. 2) is essential for optimizing the temperature profile and thus the LPBF process and material quality.

A higher melt-pool stability expands the LPBF process window, allowing the laser power (i.e. heat input) to be increased, which in turn provides higher productivity by enabling faster scan speeds and larger layer thicknesses without compromising quality. Using AFX, a material density of >99.5 % has been reported in parts made up to 7× faster than is possible with an SM beam, vastly improving the economics of additively manufactured parts [4, 6].

Fig. 7 Mechanical properties of Inconel 718 fabricated using an LPBF tool equipped with AFX (Source: M. Bechers & F. Eibl, Aconity3D)

In addition to addressing longstanding LPBF limitations, AFX introduces an entirely new capability to AM. Multiple groups have shown that controlling the thermal profile in the melt pool leads directly to control of the microstructure and thus the local material properties, opening the possibility to achieve previously unattainable performance and functionality in LPBF-fabricated parts. For example, Fig. 7 shows the measured ductility and ultimate strength of Inconel 718, fabricated with a standard SM beam and with the AFX ring beam. In addition to being able to tune these important properties locally, AFX allows both to be increased simultaneously, in contrast to bulk materials, in which they are often anticorrelated.

Conclusions

All-fiber beam shaping is revolutionizing laser-based manufacturing in both established and emerging applications. This capability is a commercially accessible reality in CFX- and AFX-based tools released by leading integrators.

References

[1] D.A.V. Kliner et al.: Proc. of SPIE 11981, p. 119810C (2022), and references therein.

[2] D.A.V. Kliner and B. Victor: Industrial Laser Solutions, p. 23 (Sept./Oct. 2018).

[3] F. Nahr et al.: J. Manuf. Mater. Process. 7, p. 93 (2023).

[4] J. Grünewald et al.: Metals 11, 1989 (2021); J. Laser Appl. 35, 042009 (2023).

[5] A. Rudolf: Photonics Views 20(1), p. 28 (2019); https://doi.org/10.1002/ phvs.202300008 [6] T. Lantzsch et al.: Proc. of SPIE 11994, p. 1199405 (2022).

DOI: 10.1002/phvs.202400012