Process development and process adaption guidelines for the deposition of thin-walled structures with IN718 using extreme high-speed directed energy deposition (EHLA3D)

The laser based directed energy deposition (DED-LB) process is a flexible technology for additive manufacturing and modification, coating, and repair in the field of, e.g., flight and aviation, tooling, and medical industry.¹ As such, DED-LB is increasingly gaining relevance in the industry, which needs to improve productivity, flexibility, and material, as well as energy efficiency while overall reducing the production cost.

To meet the increasing demand of the industry, Fraunhofer ILT introduced the extreme highspeed-laser material deposition (EHLA), a modified variant of DED-LB. In EHLA, the focus of the powder gas jet is positioned above the melt pool so that the powder particles are melting above the processing surface by absorbing most of the beam power in flight (see Fig. 1).^2,3

Due to this, the dwelling time for melting a powder particle in the melt pool is reduced so that the process feed rate can be increased to up to 200 m/min.^2,3 As a comparison, the process feed rate of a conventional DED-LB process is in the range of 1 m/min.^4,5 Besides the major increase in the productivity in coating application, the EHLA principle enables to deposit coating thicknesses in the range of 50 μm with a high cooling rate, which are comparable to the laser powder bed fusion (LPBF) process.⁶ Paired with a smaller heat input compared to conventional DED-LB, difficult to weld materials, such as Cu and Al,^7,8 and unconventional materials such as metallic glasses and high entropy alloys can be processed by EHLA. The smaller heat input results in a smaller risk to heat induced defects like hot cracks and, in addition to that, the cooling rate can be adjusted according to the material by the variation in the feed rate. Due to these benefits, the EHLA process has been successfully transferred to the industry, especially as an antiwear and corrosion coating technology.^3,9

One limitation of the EHLA process is the capability to process only rotational symmetric components such as cylinders, shafts, and disks. This limitation results from the machine types that are modified lathes, reaching the feed rates of up to 200 m/min by the rotation of the component (see Fig. 2).

To utilize the process benefits of EHLA for freeform coating, repair, and additive manufacturing (EHLA3D), Fraunhofer ILT successfully demonstrated the EHLA3D process on a tripod machine from the ponticon GmbH. This tripod can operate with a lateral feed rate of up to 200 m/min. Especially relevant for additive manufacturing, the EHLA principle enables to combine high productivity and structural resolution due to the resulting thin coating thicknesses (see Fig. 3).¹⁰ 

As a comparison, 200 μm thick thin walls can be achieved by L-PBF with the materials 316L stainless steel and tungsten.^11,12 With a powder mass flow in the range of 4 g/min walls with a 500 μm thickness can be deposited by DED-LB with the materials 316L and IN718.¹³ So far, the additive manufacturing with the EHLA3D process was successfully demonstrated on the ponticon machine with the materials 316L,^10,14 IN718, IN738, and Al-alloys. So far, no research literature exists regarding the EHLA3D deposition of the thin-walled structure.

The principle of the ponticon machine is a stationary processing optics while a tripod kinematic is moving the built platform. This concept was specifically designed for high-speed application. For this study, a conventional machine concept with moving processing optics was used, demonstrating the feasibility of EHLA3D with CNC-type machines. Due to a simpler machine kinematic, this machine concept allows a higher grade of scalability and lower procurement costs as CNC-machines are widely used and accepted in the industry.

All experiments were conducted on a modified 5-axis CNC-machine with a translatory moving processing optics (3-axis, x-y-z axis) and a tilt/turn-table (2-axis, B + C axis). The machine can operate the processing optics with a feed rate of up to 30 m/min. The used processing optics is a Beo D70 by Trumpf GmbH and is combined with a LDF4000-8 beam source from Laserline GmbH with a beam parameter product of BPP = 8 mm mrad. The specified wavelength is λ = 1080 nm, and the focus diameter of the setup is d_f = 400 μm, which results from a 2:1 aspect ratio and a fiber core diameter of d_core = 200 μm. The processing optics is equipped with a motorized collimation for an adjustable processing beam diameter. The powder nozzle is a Highno. 4.0 by HD Sonderoptiken with a stand of distance of 9 mm and a typical powder focus diameter of 0.7 mm at a powder mass flow of 3 g/min¹⁵ (see Fig. 4).

The powder material, used for the experiments, is the commercially available Amdry 1718 from Oerlikon Metco GmbH. According to the manufacturer, this powder alloy is similar to the alloy Inconel 718 and is specified with a powder particle distribution of 15–45 μm. The chemical composition is provided in Table I.

The substrate materials are hot rolled 1.4301 steel plates with a thickness of 10 mm. Before the experiment, the substrate surface is cleaned with ethanol and no further pretreatment is applied.

In a first stage, a process parameter analysis was conducted on single-track weld beads. By the variation in the process parameters, the influence of the parameters can be evaluated on the weld bead geometry and quality. Every single-track weld bead has a length of l_bead = 35 mm. The process parameters used and variated in the single-track study are summarized in Table II. For the experiments, the process feed rate is kept constant at v_f = 30 m/min. As the aim of the study is to deposit filigree structures with a maximum resolution, beam diameters with d_beam ≤ 1 mm are defined for the process development.

The deposited single-track beads are evaluated by metallographic cross sections. For every single track, three samples located at the beginning, middle, and end of the weld bead are used for the metallographic evaluation (see Fig. 5). The following values are measured by microscopy and used for the evaluation:

• - Track geometry:

  •  o Track height

  •  o Track width

  •  o Aspect ratio:

    $Aspectratio= t r a c k h e i g h t t r a c k w i d t h .$

    (1)

• - Dilution zone depth

• - Deposition area

  •  o Powder efficiency η:

    $η= D e p o s i t i o n a r e a ∗ v f ∗ ρ I N 718 m ˙ ,$

    (2)

where ṁ is the powder mass flow.

For evaluation, the mean value out of the three cross sections was used.

After the process parameter study, parameter sets that are fulfilling the following criteria are selected for the deposition of thin walls:

• - Sufficient dilution zone between 20 and 50 μm

• - No defects

•  o Cracks

•  o Pores

•  o Bonding defects

In a second stage, two selected parameter sets are transferred for thin-wall deposition by stacking the single-tracks with an adjustment of the Δz-increment. For all probes, a bidirectional deposition strategy was applied (see Fig. 6).

The correct z-increment of each parameter set is determined iteratively by comparing the programmed set height with the actual deposited built height of the thin wall after each deposition. If the set height of the wall exceeded the height of the actual deposited wall, the z-increment was decreased. The decrease was conducted in 0.01 mm steps until the wall height matched the programmed set height. After the determination of the correct z-increment, the beam power was decreased to operate with the minimum required heat input. Each thin-walled probe has a length of 100 mm and a deposition height of 10 mm. Similar to the single tracks, metallographic cross sections are prepared at the beginning, middle, and end positions of each thin wall for the thickness and defect evaluation (see Fig. 7). The thickness of each cross section is also measured in the bottom, middle, and top positions of each probe. For the metallographic analysis, two thin-walled probes are deposited, resulting in six cross sections for the calculation of the mean thickness.

To evaluate the resulting surface quality of the thin walls, five further probes with each parameter sets were deposited and its surfaces analyzed by with light interferometry (WIM). The used device for the measurement is a Nexview™ NX2 by Zygo (see Fig. 8).

On each probe, the Sa and Sz values are measured out of a 5 × 5 mm² measuring field. A bandpass filter of a spline type with a cutoff wavelength of λ_short = 3.125 μm in the short period and λ_long = 2500 μm in the long period was applied for the data evaluation.

After evaluation, the developed process parameter sets are defined. To evaluate the effect of each process parameter, thin-walled probes are deposited with a deviation of each process parameter, respectively. Depending on the process result, a process parameter guideline for thin wall deposition with EHLA3D is set up.

The aim of the process parameter analysis is to evaluate the effects of each process parameter on the single-track bead properties deposited on a high-speed process. The resulting track geometries at variated beam diameter and powder mass flow are presented in Fig. 9. The beam power was kept constant at P_L = 1100 W except for the parameter set with the smallest beam diameter, for which a beam power of P_L = 500 was used.

Depending on the used parameter, track widths between 270 and 920 μm and track heights between 30 and 150 μm can be deposited. The bead geometries deposited with the parameters d_beam = 0.8 mm and d_beam = 1.0 mm are in a comparable range so that mainly, the set powder mass flow is affecting the bead geometry. An increasing powder mass flow results in a higher track height, while the track width stays nearly constant. As the beam diameter is affecting the melt pool width, a smaller track width was achieved with the beam diameter d_beam = 0.4 mm. The effect of the beam power on the bead geometry is presented in Fig. 10.

According to the parameter variation, the beam power does not greatly affect the resulting bead geometry. The aspect ratios at different values of beam diameter, powder mass flow, and beam power are presented in Fig. 11.

As presented in Fig. 9, the powder mass flow has the biggest effect on the aspect ratio, which varies between 0.11 and 0.22 in this study. In addition, Fig. 11 indicates that the set beam diameter does not greatly affect the resulting aspect ratio. However, as the energy source of the process, the beam power has an influence on the extent of the dilution zone in the substrate material, as presented in Fig. 12.

Within this parameter study, the resulting dilution zone depth was < 60 μm. At a beam diameter of d_beam = 1 mm and powder mass flow ṁ = 15 g/min, the dilution zone depth falls below 20 μm at the tested beam power. For d_beam = 1 mm and ṁ = 20 g/min, the dilution zone is not visible by microscopy and bonding defects begin to form. In EHLA, the powder particles absorb a proportion of the beam energy during the flight toward the processing surface. Due to this, the dilution zone depth decreases with an increasing powder mass flow. The amount of energy absorbed during in flight also affects the dilution zone depth as presented in Fig. 13.

The carrier gas flow has a major effect on the resulting dilution zone as the powder particle velocity is mainly depending on the set gas flow. Higher carrier gas flows result in a higher particle velocity and in a shorter beam absorption time so that less energy is transferred to the processing surface. According to the parameter study, the shielding gas flow did not affect the dilution zone. Furthermore, both process gases (carrier and shielding gas flow) did not affect the resulting bead geometry.

To evaluate the productivity of the process, the powder efficiency of the selected parameter sets is presented in Table III.

The highest and comparable powder efficiencies are achieved with the beam diameter of d_beam= 1 mm and of d_beam= 0.8 mm as the beam diameter is in the same range as the powder focus diameter. As such, a smaller beam diameter like d_beam = 0.4 mm results in lower powder efficiency as only a proportion of the powder is irradiated by the laser beam. Due to this, one parameter set with d_beam= 1 mm is selected for the transfer to the deposition of thin-walled probes. To investigate the process limitation of the EHLA3D process, one parameter set with d_beam= 0.4 mm is also selected for the process transfer to thin walls. The parameter sets defined for the transfer to thin-wall deposition are shown in Table IV.

Thin-walled probes were deposited by using the selected process parameter sets from the single-track evaluation. The selected parameter sets for the thin-wall deposition are summarized in Table IV. For all probes, a bidirectional deposition strategy was applied as a unidirectional path planning results in isotropic material deviations at the starting and ending points of the walls. This effect is presented in Fig. 14.

To deposit thin walls, the correct Δz-adjustment per layer was determined in a first stage. The Δz-increment was initially set as the average height of the single-track, which resulted in the deposition of smaller wall heights compared to the programmed set height. Due to this, the Δz-increment was iteratively decreased in 0.01 mm steps until the actual deposition height matched the set deposition height. As presented in Fig. 15, a too high Δz-increments resulted in the missing material at the middle of the probe. A too low Δz-increment resulted in a feasible deposition of the probes; however, the actual working distance between the deposited wall and the nozzle tip decreased, which wore off the powder nozzle tip.

For the deposition of thin walls, the initial beam power used in the single track experiments needed to be decreased as a too high beam power resulted in local material accumulation distributed over the probe. The material accumulation due to heat accumulation caused by too high beam power is likely. This effect is presented in Fig. 16. The adapted beam power and parameter sets for the thin-wall deposition are presented in Table V.

According to the experiments, the variation in the shielding gas flow Q_S between 8 and 14 l/min did not affect the deposition outcome of thin walls. In contrast to the shielding gas, as presented in Fig. 17, a higher carrier gas flow resulted in material accumulation at each corner of the probes.

After the evaluation of the effects of the process parameter on thin-wall probe deposition, parameter sets with beam diameters of d_Beam=1 mm and d_Beam= 0.4 mm were used to further evaluate the probe quality by metallographic analysis and surface roughness measurement. The comparison of the metallographic cross sections resulting from the parameter sets is presented in Fig. 18.

According to metallographic evaluation, the process parameter sets resulted in defect free deposition of the thin-walled probes. In both parameter sets, the thin wall consists of a solid inner wall with fused powder accumulation attached on the side. Potentially, the powder accumulation can be reduced by further process parameter adaptions or postprocessing steps to improve the surface quality. The average thickness and the surface property evaluation are summarized in Table VI.

The deposited wall thicknesses increased gradually with the deposition height as the heat input accumulates with the height and result in a wider melt pool. At the upper position, a wall thickness of 450 μm can be achieved with the d_beam = 1 mm parameter set and 310 μm with the d_beam = 0.4 mm parameter set. At the lower position, both parameter sets resulted in a range of 200 μm thickness. The resulting surface roughness of both parameter sets is in the comparable range of S_A ≈ 30 μm and S_Z ≈ 140 μm, having a higher deviation with the d_beam = 0.4 mm parameter set. The characteristic of the thin-walls is resulting from the EHLA-principle (see Fig. 1) in which the high cooling rate results in the rapid solidification of the melt pool. Due to this, compared to DED, the lateral extension of the melt pool is decreasing so that single-tracks with smaller widths can be stacked for the deposition of thin-walled structures.

As demonstration, the parameter set with d_beam = 1 mm was used to deposit a rotational symmetric and thin-walled nozzle mock-up. The feed rate of v_feed = 30 m/min was realized by rotating the turn/tilt table. The curvature was deposited by tilting the table with a simultaneous adjustment of the processing optics in the x-direction. The deposition time was t_process = 25 min, and the demonstrator is presented in Fig. 19. The built height was 300 mm with the smallest diameter at 60 mm and the largest diameter at 120 mm. Due to the stability of the EHLA process, no control or parameter adaptions during the process were required.

A process parameter study was conducted to evaluate the effect of the process parameter on the deposition outcome of a single-track bead. Within the range of the tested parameters, the following key results can be summarized from the parameter study:

• The beam diameter defines the track width, while the track height mainly depends on the amount of used powder mass flow. With increasing powder mass flow, the deposited track height is increasing while the width stays nearly constant.

• The beam power does not greatly influence the track geometry but the extent of the resulting dilution zone.

• The carrier gas flow has a major effect on the creation of the dilution zone. The dilution zone depth decreases with increasing carrier gas flow due to an increase in the powder particle velocity. In an EHLA process, this results in a shorter beam absorption time of the particles.

• The process gas flows (shielding gas and carrier gas) have a neglectable influence on the track geometries.

• In this study, the powder efficiency decreases when the used beam diameter is smaller than the powder focus diameter. To investigate the process limitation, the productivity trade-off with a beam diameter of d_beam = 0.4 mm was accepted.

Based on the selected single-track parameters, a process parameter study on the deposition of thin walls was conducted. Derived from the results generated within the range of experiments, the following process parameter adaption guidelines can be defined:

• The correct Δz-increment should be determined by iteratively by comparing the set height with the actual built height. A too high Δz-increment results in the missing material in the middle of the thin-wall, while a too low Δz-increment wears of the powder nozzle tip.

• Material accumulation distributed over the edge of the wall indicates too high beam power. The material accumulation is likely due to the accumulating heat input at a too high beam power. The beam power and, hence, the heat input, should be iteratively decreased to a minimum so that the thin wall deposition is still feasible.

• Material accumulations at the corners of the thin walls result from too high carrier gas flows

• According to the experiments conducted in this work, the effect of the shielding gas on the deposition of thin walls is neglectable.

Parameter sets with a beam diameter d_beam= 1 mm and d_beam= 0.4 mm were developed for the deposition of thin-walled structures. According to the metallographic analysis, the wall thickness increases gradually with the deposition height as the heat input accumulates during the process. With the parameter set d_beam= 1 mm, a range of 450 μm thickness and with d_beam= 0.4 mm, 300 μm thickness are achieved. The surface roughness of S_A ≈ 30 μm and S_Z≈ 140 μm for both parameter sets mainly results from accumulated and fused powder particles attached to the wall. The improvement of the as built surface quality can be further investigated in the following studies.

This work provides an initial assessment of the feasible structural resolution of the EHLA3D process with the material IN718. The results can be used to identify possible use-cases for the technology such as the manufacturing of cooling channels and ribs. In addition to thin-walled structures, the demonstrated thickness of 300 μm provides a first assessment of possible achievable accuracy regarding near-net-shape additive manufacturing.

The authors have no conflicts to disclose.

Min-Uh Ko: Formal analysis (equal); Investigation (equal); Methodology (equal); Supervision (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Zongwei Zhang: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Thomas Schopphoven: Funding acquisition (equal); Writing – review & editing (equal).