Increasing productivity in laser-based powder bed fusion of plastics using defocused Gaussian beams

The production of large, intricate plastic components necessitates high-volume manufacturing to justify the significant investment in mold tooling. Laser-based powder bed fusion of plastics (PBF-LB/P) is an additive manufacturing process that presents a compelling alternative, enabling cost-competitive small-series production of such components for medical devices, motor vehicles and aerospace applications. However, additive manufacturing aims to produce components with high dimensional accuracy, particularly thin-walled or lattice structures. Therefore, industrial PBF-LB/P machines are equipped with CO₂ lasers, which operate at a wavelength of 10.6 µm. At this wavelength, semi-crystalline thermoplastics show high absorbance (Laumer et al., 2016), resulting in an optical penetration depth of the laser radiation in the range of typical layer thicknesses between 100 and 300 µm (Yamauchi et al., 2021). However, the wavelength also determines the laser beam radius w(z) as described in equation (1). The beam radius is defined as the radial distance at which the light intensity decreases to 1/e² compared to the axial value. The waist beam radius w₀ in the focal point (z = 0) along the optical axis can be calculated from the wavelength λ, the beam quality factor M², the focal length f, and the raw beam diameter d_r, as shown in equation (2):

Following the laser beam as it leaves the scan system, the rays converge at the focal point (Figure 1), where the beam radius is at a minimum and the beam intensity is at a maximum. In the standard PBF-LB/P process, the focal point coincides approximately with the powder bed surface. The component is then manufactured using parallel scan lines with a hatch distance h. In addition to the parallel scan vectors, a contour line is commonly scanned to improve the surface quality of the component. This contour line has a contour offset from the original component contour and a hatch offset between itself and the parallel hatch lines.

Depending on the design of the PBF-LB/P system and the hardware used, the beam waist diameter d₀ = 2w₀ of many industrial systems is approximately 400 µm. Such systems reliably manufacture geometrical features such as walls and pins down to 500 µm (Allison et al., 2019). The market launch of systems such as the EOS Formiga P110 FDR further exemplifies the trend for high-dimensional accuracy. This machine is equipped with a Co. laser that operates at a wavelength of around 5 µm, enabling a minimum wall thickness of 220 µm (EOS GmbH, 2024a). Nonetheless, this focus on precision hinders overall productivity since smaller beam diameters require more scan lines to melt the cross-section of a component in one layer. The resulting increase in laser exposure time affects the total processing time, which comprises the time needed for preheating the powder to processing temperature, the actual printing time and the time for cooling the part and part cake to ambient temperature. Unless other system components enable preheating and cooling outside the process chamber, only the actual printing time, which consists of the individual processing steps laser exposure (including material consolidation) and recoating, can be minimized in state-of-the-art PBF-LB/P systems. Minimizing the printing time can improve productivity, which can also be affected by other variables, such as the usable build volume and the packing density of the print jobs.

Thus, current research is focused on accelerating the individual processing steps, laser exposure and powder recoating. However, coating speeds are limited since increased coating speeds decrease the packing density (Sillani et al., 2021). Such changes in the bulk properties alter the powder bulk’s optical penetration depth and conduction properties (Drummer et al., 2015), which, in turn, negatively affect the mechanical properties (Niino and Sato, 2009). Therefore, coating speeds must be balanced between short coating times and part quality. In conclusion, optimizing laser exposure as the remaining processing step can achieve significant productivity gains. For example, Sillani et al. (2022) reported that laser exposure on industrial machines capable of manufacturing large components, such as the EOS P500 (build volume: 500 × 330 × 400 mm³), can account for up to 80% of the printing time, depending on the recoating speeds.

To improve the build rate, a fundamental understanding of the limitations of conventional process parameters regarding the interaction between the laser beam and the plastic powder is required. Using the scan speed v, the hatch distance h and the layer height s, the build rate b can be calculated:

The build rate b in mm³/s is used as a quantity to measure the productivity of the process concerning the volume rate of plastic processed during laser exposure. This definition of the build rate b is equal to the denominator of the volumetric energy density E_v, which is frequently used in industry and research to describe the energy input from laser radiation into the plastic material:

In the context of PBF-LB/P, the laser power P_L, scan speed v and hatch distance h are of particular significance, as they represent the most relevant laser and scan parameters for the industry standard scan pattern, as shown in Figure 1. The layer height s accounts for the volumetric energy absorption of lasers in plastic materials. However, only a layer height of approximately half the depth of a single melt track ensures sufficient layer adhesion to achieve high mechanical properties (Li et al., 2020). Furthermore, an optimal balance between laser power and scan speed is paramount for successful processing since these parameters primarily influence a single melt track’s width and depth (Yuan et al., 2020).

By adjusting the scan speed and laser power at constant energy levels, the interaction time between the plastic powder and the laser beam changes, influencing the melting behavior. At high scan speeds and laser powers, the width of single melt tracks corresponds to the laser beam diameter, as shorter interaction times may hinder heat conduction (Drummer et al., 2014). Also, the depth of single melt tracks decreases, corresponding to the powder’s optical penetration depth, as the heat conduction in the z-direction decreases with minimized interaction times (Drummer et al., 2014). This can ultimately lead to compromised layer adhesion (Riedlbauer et al., 2014), negatively impacting mechanical properties and part density (Drexler et al., 2015).

For space-filling strategies, the upper limit for the hatch distance is the width of a single melt track (Li et al., 2021). The width of a single melt track increases with increasing energy input (Li et al., 2021). Thus, the required overlap (see Figure 1) between two melted scan tracks changes when the laser power or scan speed varies. However, the melt width is approximately limited to the laser beam diameter when increasing scan speed and laser power (Drummer et al., 2014). The hatch distance, therefore, does not provide sufficient potential for improving the build rate if a focused beam is used.

Furthermore, process parameters are not transferable between materials as variations in thermal, optical and rheological properties influence the transient behavior of the plastic melt temperature and, thus, the resulting mechanical properties (Bierwisch et al., 2021). In conclusion, increasing the build rate by increasing laser power, scan speed and hatch distance utilizing a focused laser beam is already limited by a material-dependent process window that yields sufficient laser-powder interaction.

Since optimizing the conventional laser and scan parameters has been extensively researched, further aspects must be considered to maximize the build rate. For example, PBF-LB/P systems can integrate additional optical elements, including a third optical axis alongside the standard 2D galvanometer scanners. Used conventionally, this enables dynamic control of the laser focus, maintaining a consistent laser beam diameter across the entire scanning field regardless of beam deflection by the scanning mirrors. The design is particularly advantageous for high stitching quality when using multiple lasers. Furthermore, these focusing units are less expensive than objectives like F-theta lenses, which are traditionally employed for flat field correction to ensure uniform laser beam diameters on the build surface.

While such 3D scanning modules are typically used for precise laser beam positioning, they can also intentionally increase the beam diameter through beam defocusing by shifting the focal point below or above (as illustrated in Figure 2) the powder surface. Defocusing results in a broader but flatter intensity distribution at the powder surface. However, it cannot be assumed that the distribution is still Gaussian for an actual optical setup. Nevertheless, Meyer and Witt (2020) demonstrated that defocusing a CO₂ laser leads to wider melt tracks, enabling hatch distances up to 750 µm without compromising layer height. This was achieved by using a laser power of up to 130 W to maintain a consistent specific energy input in the increased laser spot area. Yamauchi et al. (2025) investigated volumetric parts using a laser power of up to 82 W with defocused laser beams up to 1,000 µm in diameter. The authors visually observed that smoke formation occurs for defocused laser beams at higher laser power to scan speed ratios compared to focused laser beams, underlining the higher usable laser powers for defocused laser beams. Yamauchi et al. (2025) examined three parameter combinations with defocused laser beams for density measurements while keeping the energy density constant. The authors found that the part densities remained constant between the parameter combinations. However, it remains unclear whether a change in energy density leads to changes in part properties, such as density, and how defocused laser beams affect surface-near porosity.

In the PBF of metals, current research is focused on using alternative beam shapes, such as ring profiles, to enhance productivity and optimize process dynamics (Holla et al., 2023). While this approach may be beneficial for extending hatch distances in PBF-LB/P, such optical elements are currently not commercially available for CO₂ lasers. Conversely, Nussbaum et al. (2021) examined using high-intensity infrared projectors in PBF to melt large areas simultaneously with exposure times of around 8 s. Although the produced plastic parts exhibit high mechanical properties, the projector system design is limited to a 2.1 × 1.6 cm² area. However, it is essential to note that industrial infrared projector systems capable of processing significantly larger plastic components are now commercially available (Visitech, 2024). Nevertheless, additive manufacturing systems equipped with projector systems could be a suitable solution to increase productivity. Similarly, EOS announced a technology using an array of up to one million diode lasers operating simultaneously to reduce the printing time independent of the build job’s packing density and the part’s geometry (EOS GmbH, 2024b). However, the economic viability of these systems can only be determined after commercial launch.

In conclusion, current approaches to enhance productivity in PBF-LB/P often necessitate significant upfront investments in new equipment or hardware components. In contrast, this work explores a more cost-effective strategy: using scan strategies with defocused Gaussian beams to improve build rates and, thus, productivity. The widespread availability of a third optical axis in PBF-LB/P systems, or the potential for retrofitting, justifies this approach. Furthermore, implementing the laser beam diameter as an additional processing parameter requires mainly software updates.

Consequently, this research examines the impact of defocused Gaussian beams on part properties and achievable build rates. Volumetric samples are produced with the focused laser beam and two different beam magnifications for the defocused laser beam. The laser power, scan speed and hatch distance are selected to examine a wide range of volumetric energy densities. The Archimedean density of the samples is then determined to investigate whether samples with sufficient density can be obtained using defocused beams despite the increased build rates. In addition, it is investigated whether the volumetric energy density, as with the focused laser beam, can indicate the density of the samples independent of the beam diameter. Cross-sectional images of selected samples are taken to describe the porosity of the sample and the effect of defocused laser beams on the near-surface areas. Finally, the build rate improvement is determined and related to the density of the obtained samples.

Polyamide 12 (PA12) powder (PA2200, EOS GmbH, Krailling, Germany) is used for all experiments. The mixing ratio of virgin and used powder is 50:50 Wt.%, which is a common ratio in the state-of-the-art. Bulk density measurements (DIN EN ISO 60) show a value of 0.451 ± 0.005 g/cm³, indicating the comparability of the powder batches. In addition, tapped density is measured as 0.526 ± 0.006 g/cm³, resulting in a Hausner ratio of 1.17 ± 0.02. Therefore, the PA12 powder can be characterized as free-flowing (Geldart et al., 1984).

A self-built PBF-LB/P research system is used for the experiments, as depicted in Figure 3. The powder reservoir, the build platform (build volume 150 × 106 × 200 mm³) and the overflow container are arranged side by side in the process chamber of the PBF-LB/P system. The fresh powder is applied using a knife-shaped recoater. Heating cartridges are installed on both sides of the build platform to control the temperature of the powder bulk. Two parallel infrared (IR) twin-tube heaters (Heraeus Noblelight GmbH, Kleinostheim, Germany) allow independent temperature control of the powder surface by a pyrometer (CT LT22, Optris GmbH, Berlin, Germany).

To enable the defocusing of the laser beam, the optical components are mounted on an optical bench that can be moved vertically (z-direction). Two linear drives connected to the printer’s frame realize the movement. Self-locking screw drives enable the z-positioning of the optical bench, which is attached to the carriages of the linear axes (positioning accuracy: ± 0.1 mm). The Gaussian laser beam is generated by a 120 W CO₂ laser (Iradion i120F, Iradion Laser GmbH, Krailling, Germany) with a wavelength λ of 10.6 µm and a beam quality factor M² of 1.2. A galvanometer scanner (intelliSCAN III 20, SCANLAB GmbH, Puchheim, Germany), combined with a f-theta lens with an effective focal length of 300 mm deflects the laser beam to melt the plastic powder.

The required z-position of the optical bench for the defocused laser beam can be calculated using equations (2) and (3). Based on the optical setup and the laser’s average beam diameter, the raw beam diameter d_r is set to 15 mm. According to equation (2), this results in a theoretically focused beam waist diameter of d₀ = 2w₀ = 324 µm (beam magnification of ×1.0). For simplicity, this work assumes a normal distribution of intensity for defocused beam diameters. Future research is needed to describe the actual intensity distribution of defocused beams. The beam radius can then be calculated as a function w(z) of the vertical offset z. For the later examined defocused beam diameters d_2.5 = 2.5d₀ = 810 µm (beam magnification of ×2.5) and d_3.5 = 3.5d₀ = 1134 µm (beam magnification of ×3.5), this results in a relative offset of z_2.5 = 14.8 mm and z_3.5 = 21.7 mm, respectively. Table 1 summarizes the investigated beam diameters. The calculated beam diameter range in Table 1 corresponds to a positioning accuracy of 100 µm, which can be set repeatably on the z-axis screw drives. The optical assembly is moved upwards so that the laser always operates in the diverging range of the laser beam.

Defocusing the laser beam by moving the optical assembly results in a varying distance between the scan system and the powder surface. The machine control must consider this, as process parameters such as scan speed and laser beam diameter are based on a fixed working distance. A correction factor f_corr is introduced as the quotient of the focused working distance (equal to the focal length f of the f-theta lens) and the sum of the focal length and the applied offset z_offset:

This correction factor is applied to the scan speed and the scan pattern in the x- and y-dimension, ensuring that the part size is constant independent of the beam magnification.

Cubes (edge length 15 mm, n = 4) are produced on the PBF-LB/P research system to examine the density of the obtained samples. The aim is to understand how the defocused laser beam, as a novel process parameter, interacts with established parameters such as laser power, hatch distance and scan speed for volumetric samples. All samples are printed with a constant layer height of 0.1 mm and a recoating speed of 100 mm/s. The temperature of the infrared heaters T_IR and the heating cartridges T_C were set to T_IR = 168°C and T_C = 160°C.

The interaction between the laser beam and the powder must be considered when selecting the print parameters. If the energy input is too low, the plastic powder will not melt sufficiently to provide adequate adhesion between adjacent layers. The plastic will degrade near the surface if the energy input is too high. Therefore, upper and lower limits are introduced for the energy input within which the plastic powder is processed. This work uses the limits suggested by Chatham et al. (2021), who have reported a plateau of mechanical properties of PA12 between 0.20 and 0.40 J/mm³ for focused laser beams. To take account of a possible extended usable energy density range when using defocused laser beams, the lower energy density limit is set to 0.15 J/mm³ in this work. This measure assumes that the increased beam diameter in the defocused regime results in a more frequent overlap if the hatch is not increased by an equivalent extent and, e.g. the scan speed is increased instead. Owing to the more frequent overlap, less energy may be required to produce samples of the same density. The upper limit of the volumetric energy density is not raised as smoke development in the PBF-LB/P research system, as reported by Yamauchi et al. (2025), owing to a too-high energy density has to be prevented.

Table 2 lists the variations in laser power, scan speed and hatch distance used to produce cubic samples with beam magnifications of ×1.0, ×2.5 and ×3.5. As the laser beam diameter increases, the intensity decreases for a given power, meaning a higher laser power is required for defocused beams to compensate for the larger exposed area. However, as observed in preliminary experiments, high laser powers are impractical for focused beams, as the peak intensity becomes excessive, potentially leading to material degradation (Peyre et al., 2015). In contrast, low laser powers in defocused beams result in insufficient melting. The hatch distance is also constrained by the laser spot size, with distances ≥ 0.5 mm only feasible for defocused beams. Scan speeds were selected to maintain the volumetric energy density E_v, within the range of 0.15 ≤ E_v ≤ 0.40 J/mm³. Industry-standard scan speeds of around 4000 mm/s were used for focused beams, while defocused beams enable speeds up to 10,000 mm/s. The increased scan speed with defocused beams is possible due to the approximately constant interaction time between the laser and the powder because of the increased defocused beam diameter. In addition, the build rate b in Table 2 expresses the plastic volume (mm³) processed per second during laser exposure. Thus, eight focused and 19 defocused parameter sets were printed with four samples each.

For the sample parameters described in the previous section, process parameters with significantly higher build rates than the state-of-the-art are investigated by defocusing the laser beam. Owing to the comparatively high scanning speeds (see Table 2), the synchronization of laser power and scanner movement plays a crucial role. The switch-on behavior of the laser can be characterized by the laser rise time. The laser rise time is defined as the time it takes the laser to rise from 20 to 80% power. For the laser of the PBF-LB/P research system, the rise time is 45 µs according to the calibration sheet of the laser. For the maximum scanning speed of 10,000 mm/s used in the experiments, this means that due to the laser rise time alone, 450 µm of the scan vectors will not be exposed to the nominal power. This is still a minimum estimate since, contrary to the definition of the laser rise time, the laser must cover a larger power difference from 0 to up to almost 100%, and at the same time, the radius of the defocused laser beam (around 0.6 mm for ×3.5 beam magnification) also adds to the scan vector length that receives less energy because of the beam’s Gaussian intensity distribution. To compensate for this effect, a pre-contour line without hatch offset is used for all samples. The contour line is printed with the same parameters as the hatch lines. For the hatch lines, a 90° alternating pattern is used.

The ability to produce dense components is a prerequisite for using the defocused laser beam in the PBF-LB/P process. Therefore, density measurements based on the Archimedean principle are performed according to DIN EN ISO 1183–1 (immersion method). The sample’s edge length of 15 mm ensures that it meets the recommended minimum mass of 1 g as specified in the standard. All samples are weighed in air and ethanol (ρ_eth = 0.79 g/cm³) on a laboratory balance in a controlled environment with a defined room temperature. Samples are immersed in ethanol for 30 s before the mass is recorded. This procedure allows for repeatability but also overestimates the density of samples with open porosity in surface-near areas because of ethanol infiltration.

Microscopic cross-sectional images are made for selected samples to complement and further investigate the results of the Archimedean density measurements. Images of the polished sample cross-sections were taken using a light microscope (Leica DM6 M, Leica Microsystems GmbH, Germany) with an attached CMOS camera (Leica DMC, 4500, Leica Microsystems GmbH, Germany).

The results of the density measurements for the 27 printed parameter combinations are shown in Figure 4. The highest density attainable with the PBF-LB/P research system and a focused laser beam at 30 W is 0.99 g/cm³ with a resulting volumetric energy density of 0.40 J/mm³. This is in the range of PA12 components produced with industry-standard parameter sets on industrial machines for which 2%–3% residual porosities are commonly reported in the literature (Dewulf et al., 2016). Therefore, 0.99 g/cm³ is defined as the minimum acceptable density for parameter combinations utilizing defocused laser beams (gray area in Figure 4).

As a general trend, the relationship between volumetric energy density and Archimedean density is consistent regardless of beam magnification. While there is a rapid increase in density at low energy densities, the rise in density flattens out at higher energy densities. An independent influence of laser power or scan speed on the component density cannot be identified. Both ×2.5 and ×3.5 beam magnifications reach the minimum acceptable density at 0.3 J/mm³, while the focused beam reaches it at a higher energy density of 0.34 J/mm³. In addition, the maximum achievable density at the same energy density of 0.4 J/mm³ is higher for the defocused samples than for the focused samples. This indicates that the defocused laser beam requires less energy input to achieve a specific part density.

The logistic fits ρ_fit in Figure 4 underline the higher achievable part densities at the same volumetric energy densities when using defocused laser beams, especially in the low energy density range. The coefficients of the fits and the weighted R² values are listed in Table 3. As there is a large variation in the measurement uncertainty between data points, incorporating weights into the R² value accounts for the varying reliability of the Archimedean density measurements. Future studies using infrared thermography are needed to evaluate the individual influence of laser power, scan speed, hatch distance and beam magnification on the laser-powder interaction and resulting part properties, such as density, in more detail.

The influence of the beam magnification on the porosity is illustrated by the cross-sectional images in Figure 5 for the two parameter combinations produced for beam magnifications ×1.0 and ×3.5. The visually observable density decreases in the cross-section for the sample produced with ×3.5 beam magnification and 0.24 J/mm³ (60 W, 10,000 mm/s). This is in contrast to the mean values of the Archimedean density. As a difference between visual appearance and Archimedean density is only observed for samples produced with low volumetric energy density, it can be assumed that the fluid infiltrates the surface-near region. The cross-sectional image in Figure 5 highlights that the sample with a beam magnification of ×3.5 at 0.24 J/mm³ is characterized by a poor bond between the layers, resulting in large open pores connected to the sample surface, corresponding to an overestimation of the Archimedean densities. In contrast, the displayed parameter combination for 0.34 J/mm³ (60 W, 7,000 mm/s) shows a good agreement between Archimedean density and the visual appearance of pores. This is attributed to the reduced edge porosity, reducing infiltration effects during Archimedean density measurements. Therefore, the trends observed in Figure 4 are still valid since the increased ratio of open pores is limited to low energy input, and no sample produced with low energy input reaches values close to the target density of 0.99 g/cm³.

Figure 5 further highlights that the porous region at the part’s side surface is wider for a scan speed of 10,000 mm/s compared to a scan speed of 7,000 mm/s. This effect is because of the increased influence of the laser rise time and the reduced line energy input into the contour line at 10,000 mm/s. The contour exposure cannot compensate for the increased porosity in surface-near regions. This highlights the need to further optimize contour exposure, especially at high scan speeds.

Samples manufactured with an energy density < 0.30 J/mm³ tend to show insufficient adhesion between layers, as seen for 0.24 J/mm³ in Figure 5. The density increases with increasing E_v, reaching a plateau between 0.30 and 0.40 J/mm³ for both focused and defocused laser beams. Even though there are slight variations in density between the different beam magnifications, the volumetric energy density seems reasonable for an initial parameter guess for defocused laser beams. It can, therefore, be assumed that suitable parameter combinations for defocused beams can also be derived for other plastic powder materials based on established volumetric energy density ranges used for focused laser beams.

Greiner (2023) reported that maintaining energy density by lowering scan speed and laser power results in increased melt temperatures and a low number of large pores for focused laser beams. Similarly, Drummer et al. (2014) showed that combining high laser powers with high scan speeds at equivalent energy density results in high heating rates that favor heat accumulation at the melt surface and reduce melt depth, ultimately increasing porosity. These effects are not observed for defocused laser beams, even if scan speeds of 10,000 mm/s or laser powers of 120 W are employed. Therefore, it can be assumed that dense samples can be manufactured at even higher scan speeds and laser powers, resulting in a further improved build rate when using defocused laser beams.

The improved particle fusion is likely attributed to an extension of the period the plastic material is maintained at higher temperatures, as suggested by the findings of Hofmann et al. (2024). Therefore, future studies should evaluate if using even higher laser power and scan speeds combined with high-frequency overlapping of scan tracks can further reduce porosity if defocused laser beams are employed in PBF-LB/P.

Nevertheless, the highest density of 1.01 g/cm³ is achieved if a laser power of 120 W is combined with a comparatively low scan speed of 4,000 mm/s using a defocused laser beam. The cross-sectional image of the sample with the highest density in Figure 6 shows nearly no pores in the bottom half. In contrast, the final layers, especially the top 2 mm of the sample, show similar porosity as focused samples. A potential explanation is provided by Soldner et al. (2021), who used process simulation of PBF-LB/P to demonstrate that partially crystallized structures are remelted during the layer-by-layer manufacturing process if focused laser beams are used, enabling extended time for particle fusion in subsurface layers. Consequently, the influence on the specific combination of laser and scan parameters is likely more relevant for defocused laser beams as the laser-powder interaction time can vary widely. However, samples produced with defocused laser beams not only exceed the industry standard density of 0.99 g/cm³, but the combination of scan speed and hatch distance significantly influences the time needed for laser exposure, which will be discussed next.

Section 2.3 introduced the build rate as the product of hatch distance, scan speed and layer height to estimate the volume processed by the laser per time. Figure 7 illustrates the achieved Archimedean densities as a function of build rates for beam magnifications ×1.0, ×2.5 and ×3.5. The choice of the reference build rate (dashed line in Figure 7) is based on box plots in a review by Kusoglu et al. (2021), in which the 100 most cited peer-reviewed articles in the field of PBF of plastics were statistically evaluated concerning, e.g. the material used, scan speeds and hatch distances. Taking the 75^th percentile (75% of the reported values are below this value) of the reported scan speed v₇₅ = 3,500 mm/s, hatch distance h₇₅ = 0.25 mm and layer height s₇₅ = 0.1 mm for PA12 as a reference, this results in a reference build rate of 87.5 mm³/s, which is nearly equal to the obtained build rate for focused Gaussian beams with P_L = 30 W in this publication.

Per each laser power, the density of the samples decreases as the build rate increases. This can be explained by the increase in build rate at constant laser power being accompanied by a decrease in energy density. Furthermore, higher laser powers allow for increased build rates. Considering the highest density parameter combination at P_L = 30 W as a reference for focused laser beams, a build rate increase from 80 to 350 mm³/s can be observed while maintaining a density of at least 0.99 g/cm³. For 60 and 90 W, a reduced slope is present at a low build rate. This indicates the existence of a maximum achievable density per applied laser power. A reduced slope cannot be observed for the defocused samples printed at the highest investigated laser power of 120 W. Therefore, it may be possible to achieve even denser parts by further reducing the build rate at 120 W.

Looking more closely at the effect of beam diameter magnification on the achievable densities, higher densities can be realized for 90 and 120 W at comparable build rates for the ×2.5 magnification, except for the highest investigated build rate of 525 mm³/s, where the densities are nearly identical. This can be attributed to the effect that the increased cross-sectional area of the laser beam at ×3.5 magnification increases surface-near porosity (see Figure 5), as the defocused laser beam has an enlarged area of lower energy intensities in its cross-section.

Using the 75th percentile as a reference, defocused laser beams can increase the build rate by a factor of 4. However, Figure 7 shows that the associated volumetric energy densities limit the achievable build rates at acceptable densities. Therefore, an increased laser power greater than 120 W and higher scan speeds could allow an even higher build rate.

The precision of a CO₂ laser in the PBF-LB/P process enables the production of dense components with fine features down to 0.5 mm. However, while beneficial for small-scale precision, a focused beam requires numerous scan lines for larger components while limiting the available range of process parameters such as laser power, scan speed and hatch distance. To address this, the present study explores increasing build rates by defocusing the laser beam, allowing the use of larger beam diameters, larger hatch distances and higher scan speeds while still achieving industry-standard component density. The laser was defocused from a calculated diameter in the build plane of 324 µm to 1,134 µm, with laser power reaching up to 120 W to maintain energy intensity across the enlarged spot area. Cubes were manufactured on a research PBF-LB/P system using energy densities of 0.15–0.40 J/mm³, typically applied to focused beams, but with defocused beams using hatch distances up to 0.75 mm and scan speeds up to 10,000 mm/s. The resulting Archimedean density measurements and cross-sectional images provide the following key findings:

• volumetric energy density values commonly used for focused laser beams can be effectively applied to defocused laser beam scan strategies;

• defocused laser beams increase build rates by a factor of 4 while still maintaining industry-standard part density (∼ 0.99 g/cm³) due to the capability of using scan speeds up to 10,000 mm/s;

• the density of PA12 components can be increased to as high as 1.01 g/cm³ with defocused laser beams, even at build rates higher than the research standard (build rate of 87.5 mm³/s); and

• with defocused laser beams, it is possible to achieve part densities in line with the industry standard and beyond, even with a lower energy density than focused laser beams.

Furthermore, this work demonstrated the significant impact of laser beam intensity distribution on forming porous edge areas, mainly when defocused beams were used. Additionally, the importance of the laser rise time in CO₂ lasers will become more pronounced at scan speeds above 10 m/s. Therefore, future studies should focus on path planning, particularly optimizing contour exposure strategies, to enhance structural integrity, dimensional accuracy and surface quality for defocused laser beams.

The findings underscore the necessity for advanced scanning strategies in PBF-LB/P. While this study did not incorporate such technology, future research could implement movable lenses in the optical path of the CO₂ laser to dynamically adjust the laser beam diameter rather than relying on vertical movement of the optical assembly. This would eliminate the need for component size corrections through software scaling factors, simplifying the build process preparation.

In addition, future investigations must explore the influence of laser and scan parameters for defocused laser beams on the laser–powder interaction and the resulting part properties. Infrared thermography will be used to identify links between the selected parameters, the temperature–time behavior of the plastic and the final part properties.

Finally, it is important to note that the hardware limitations of the PBF-LB/P research system in this study were insufficient to determine the upper performance limits of defocused laser beams. Higher productivity gains may be achievable with faster scanning systems and other beam sources.

The authors thank Optoprim Germany GmbH and Iradion Laser Holding GmbH for providing the CO₂ laser used in the experiments. Moreover, the authors thank SCANLAB GmbH for providing the galvanometer scanner.

Open Access funding enabled and organized by Projekt DEAL.

Data will be made available on request.

Declaration of generative AI and AI-assisted technologies in the writing process: While preparing this work, the authors used “DeepL Write” and “Grammarly” to improve the manuscript’s language and readability. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the publication's content.