Identifying Ultrasonic Testing Based Nondestructive Qualification Parameters for Laser DED Processed IN718

Abstract

This article explores the use of ultrasonic nondestructive evaluation for qualification of laser-DED IN718 samples. The main goal of this article is to identify potential ultrasonic parameters which have highest sensitivity to microstructral changes that result from fabrication of DED samples. The ultrasonic qualification parameters were extracted from ultrasonic testing including velocity and attenuation measurement, and C-Scan imaging. These measurements were further used to extract parameters that quantify the anisotropy, microstructural heterogeneity, and grain scattering. Two laser-DED IN718 samples fabricated with slightly different processing parameters were evaluated to observe the influence of the laser power and scan speed on the qualification parameters. The identified qualification parameters were compared for these two samples, along with a hot-rolled sample that was also used as reference. The results suggest that the anisotropy, attenuation, and heterogeneity were highest in the DED samples compared to the reference sample. The identified qualification parameters seem to capture these changes, suggesting they could be potentially used for qualification of AM parts.

1. Introduction

Laser-processed additive manufacturing (laser-AM) comprises several technologies that allow the fabrication of three-dimensional structures from a CAD model in a layer-by-layer fashion, in contrast to conventional manufacturing where subtraction of material is predominant. Laser-AM has shown progress in the fabrication of complex metallic materials such as nickel-based superalloys [1,2,3,4,5,6,7,8,9], which are within the most important engineering materials due to their superior mechanical properties and behavior under high-temperature conditions, characteristics that make them ideal for application in the aerospace or power generation industry [1,2,3,4]. The eventual adoption of these laser-AM parts on a large scale appears promising. However, one of the principal challenges that AM faces is the ability to produce complex components that can respond with high reliability in critical services. For these reasons, qualification, certification, and product acceptance are important affairs in AM industry that are not yet in their final stage. The Federal Aviation Administration (FAA) was one of the first institutions in approving the use of an AM part for an engine of commercial jets [10,11,12]. It was this event that originated the first standard approved by ASTM and ISO [10]. More recently, there are over a dozen standards that address several aspects of AM, such as terminology, design, materials and processes, and test methods [10,11,12,13,14,15,16,17,18,19,20]. However, there is still progress in developing and expanding the scope of new standards from different agencies around the globe [10]. Part certification and qualification appears to be more challenging since several factors could affect the reproducibility of products, such as feedstock material, the chosen laser-AM technology and post-processing of the as-fabricated product.

Laser Directed Energy Deposition (laser-DED) is among the laser-AM technologies, and it consists of several processing parameters including laser power, scan speed, hatch spacing, layer thickness, particle shape/size distribution and other build conditions. A wrong set of these parameters can result in parts with internal defects such as porosity or lack of fusion [15,16,17,18,19,20]. This becomes particularly important in the fabrication of nickel-based superalloys, where the multiple melting and solidification during fabrication, inherent of any laser-AM technology, result in complex polycrystalline microstructures composed of grains of a wide distribution of sizes, shapes, and orientations. They are considered statistically isotropic when the grains are randomly oriented; however, this rarely happens, especially with parts fabricated with laser-AM technologies where the cooling rates have a direct effect on grain morphology. The as-fabricated parts will exhibit texture or a preferred orientation of grains that will result in anisotropic properties of the part with the build direction being different than the transverse directions. This and other mechanical behavior such as strength, fatigue, hardness, or residual stress have been studied using destructive methods, where the influence of the processing parameters on these mechanical properties can be observed and attributed for example to the presence of porosity, lack of fusion, flaws, etc. [1,2,3,4,5,6,7,8,9,21,22]. At the same time, ASTM and ISO are advancing in their effort to establish a unified standard for several types of AM technologies, including laser-DED and more specifically, with nickel-based superalloys as described by the standards F3055-14a and F3056-14 for Inconel 718 and Inconel 625, respectively. However, this progress does not close the gap in the case of materials with complex microstructure that can present anisotropy, microstructural heterogeneity or even the presence of cluster of grains having different sizes.

Qualification is defined as the evidence that a product design meets requirements in the presence of operational environments, and it can refer to a process, a machine, or a part [23]. A part can be qualified for a specific application by measuring physical properties with quantifiable parameters, for instance by measuring the modulus of metallic parts with destructive (tensile test) or non-destructive methods (ultrasonic characterization). These qualification procedures are compiled within standards and must comply with requirements specific to a certifying authority such as the American Society for Testing and Materials (ASTM), the International Organization for Standardization (ISO), the Federal Aviation Administration (FAA), the American Society of Mechanical Engineers (ASME) or the Aerospace Material Specifications (AMS). Among them, ASTM and ISO are leading the development of standards for the AM industry with the creation of the ASTM Committee F42 in 2009 and the ISO Technical Committee 261 in 2011 [24]. Regarding nickel-based superalloys, the standards F3055-14a and F3056-14a referring to Inconel 718 and Inconel 625 fabricated with powder bed fusion (PBF) technology, respectively, are intended for purchasers or producers and provide specifications for the acceptance of parts under the complying of minimum tensile properties, these standards do not mention specifications for non-destructive testing. The standard F3187-16 is a guide for DED of metals that defines the requirements for this technology application such as space, set-up, process documentation or process monitoring technologies. In its section describing post processing it is mentioned that the ability to perform ultrasonic non-destructive testing can be affected by the unique microstructure of the part developed during deposition. The standard F3122-14 specifies the evaluation of mechanical properties for AM metallic materials, it describes the destructive methods to quantify deformation (tensile test, uniaxial compression, bearing, bending, hardness, etc.) and fatigue (axial force fatigue test, fracture toughness, crack growth) properties, the standard does not mention non-destructive testing. The application of this standard can be affected by the AM material, anisotropy, porosity, testing environment, etc. More recently, and after years of preparation, the standard E3166-20 was published, this standard serves as a guide for nondestructive evaluation (NDE) of after build AM metallic parts for aerospace applications. It discusses NDE methods applicable to aerospace parts fabricated via PBF and DED and includes nickel-based superalloys, among others. It also discusses the typical defects present in AM parts.

However, this standard does not provide a novel approach regarding ultrasonic testing, and it is limited to the methods described in the standards from the ASTM subcommittee E06.07 on ultrasonic methods. Furthermore, several other works have also explored the NDE of laser-based additive techniques [25,26,27]. However, these have not explore complex microstructures like IN718. There is still a gap in the state-of-the-art of ultrasonic NDE qualification procedures for laser-processed nickel-based superalloys that can consider the challenges posed by the material microstructural features.

This article explores different ultrasonic parameters that could potentially be used as nondestructive qualification parameters for laser-DED IN718 parts. These parameters quantify the anisotropy, microstructural heterogeneity, average cluster size and the grain or cluster size contribution to ultrasonic attenuation. The ultrasonic velocity, attenuation and C-scans of the samples are used as the basis for the derivation of the qualification parameters. Two laser-DED IN718 samples fabricated with slightly different processing parameters were evaluated to observe the influence of the laser power and scan speed on the qualification parameters. The identified qualification parameters were compared for these two samples. Additionally, a hot-rolled sample was also used as reference.

2. Methods and Materials

2.1. Materials

IN718 cubes were fabricated using a Laser Engineered Net Shaping (LENS) rapid Laser Directed Energy Deposition machine at Optomec Inc. (Albuquerque, NM, USA). IN718 is a multiphase nickel superalloy with composed of 50–55 wt% of Ni, 17–21 wt% of Cr, 4.8–5.5 wt% of Nb, 2.8–3 wt% of Mo, 0.65–1.15 wt% of Ti, 1 wt% of Co and small amount of Al and Fe [3]. To test the qualification parameters, we used two different sets of processing parameters as shown in

Table 1. Process parameters used in the present study.

Parameter	PP1	PP2
Oxygen and moisture	<20 ppm	<20 ppm
Laser power	900 W	1000 W
Speed	20 mm/s	15 mm/s
Powder feed rate	12.3 g/min	12.3 g/min
Powder particle size	16–44 μm	44–16 μm

. The samples were built as 25 mm cubes on a 12 mm thick stainless steel 304 base plate and were further cut from the base plates and machined to a final dimension of 25 mm (L) × 25 mm (W) × 19 mm (H) samples, the latter being the build direction. The final surface finish after machining both samples was equivalent to a 400-grit surface finish. As can be observed from the processing parameters (PP), the laser power and speed were changed. The objective is to determine how these affect the non-destructive properties of the samples and identify parameters that can be quantified from the NDE measurements. From a destructive analysis, a detailed study has been carried out earlier [1] and will be used in Section 5.

2.2. Ultrasonic Testing

Details of the ultrasonic testing including transducers, frequency, methodology, etc. are given in the Supplementary Materials. As the methodology is similar to previous work, we have added this to the Supplementary Materials, instead of repeating the information. This work measures the ultrasonic velocity, attenuation measurements and C-Scan imaging, which were further used to extract parameters that can be used for qualification.

3. Proposed Qualification Parameters

3.1. Anisotropy Index

Previous work showed that as-fabricated IN718 samples demonstrate considerable acoustic anisotropy compared to wrought samples. There could be several sources of acoustic anisotropy including crystalline anisotropy, multiphase contribution, texture, residual stress etc. The wrought condition is expected is be isotropic, i.e., no anisotropy, and the influence of PP on these characteristics must be explored. From an acoustics perspective, these are coupled effects, which makes it challenging to disentangle them. Therefore, we need a coupled parameter which can account for all these effects using a single parameter, which can be potentially used as a metric/parameter for nondestructive qualification. More generally, it can be used to characterize any metallic sample having any crystal symmetry where their ultrasonic velocities at their principal directions can be measured. This parameter is based on the anisotropy factor previously proposed by Ledbetter and Migliori [28], where their proposed anisotropy index is the squared ratio of the maximum and minimum shear velocities. However, we observed changes in both longitudinal and shear velocities along the different directions, therefore, we redefine this as:

$$A_{LS}={\displaystyle \frac{max\left({v_{L_i}}^2\right)}{min\left({v_{L_i}}^2\right)}}\times {\displaystyle \frac{max\left({v_{S_{ij}}}^2\right)}{min\left({v_{S_{ij}}}^2\right)}}$$

(1)

where, $v_{L_i}$ and $v_{S_{ij}}$ represents the longitudinal and shear wave velocity along the different propagation and polarization directions. Compared to the anisotropy index defined earlier, we also notice that the variation between the longitudinal and shear velocities is not the same. Therefore, it becomes important to include both velocities to define a single parameter. Physically, this parameter will capture the maximum and minimum variation in the velocities induced by anisotropy. The advantage of our proposed parameter is that it is derived from the ultrasonic velocities propagating along the principal directions, thus, we get one parameter that simplifies the qualification process. This becomes particularly important in the case of as-fabricated laser-processed IN718 samples where the nature of the fabrication process is highly influenced by the processing parameters and the material, and results in samples with complex microstructures and different degrees of anisotropy [3]. This anisotropy qualifying parameter can easily be tested. When there is no anisotropy, i.e., $max\left(v_{L_i}\right)=min\left(v_{L_i}\right);max\left(v_{S_{ij}}\right)=min\left(v_{S_{ij}}\right)$ Therefore, in the absence of anisotropy, the ratio $A_{LS}\approx 1$. This will be evaluated for all the samples including a conventional hot-rolled IN718 sample.

3.2. Heterogeneity Parameter

The back-wall amplitude C-Scans will provide information that can be related to the microstructural heterogeneity of the as-fabricated laser-DED IN718 samples. Microstructural heterogeneity is defined as spatial variation of microstructural features such as grain morphology, i.e., shape and size, distribution of phases and other features. While this is expected to be small for conventional microstructures, we can observe a large spatial variation for the IN718 DED structures. The back-wall amplitude is affected by the ultrasonic attenuation of the wave energy as it travels through the sample. The major contribution to the ultrasonic attenuation is grain scattering; thus, if there is microstructural heterogeneity, then a spatial variation of the back-wall amplitude will be observed. To extract parameters which are physically meaningful and represent the microstructures, it is important to utilize the heterogeneity of the ultrasonic response. Incidentally, heterogeneity is also the reason conventional NDE qualification methods fail for DED structures. In this work, we propose to utilize this heterogeneity to qualify the structure. To extract heterogeneity parameters from the C-Scan data, we follow the flow chart shown in Figure 1. The C-Scan images are normalized to the data within the image can be quantified. The color C-Scans are converted into grayscale images for easier processing. Next, the histogram of the image was obtained, and converted to dB scale. Finally, a Gaussian fit was performed on the histogram:

$$G\left(dB\right)=A{e}^{-{((dB-\mu )/\sigma )}^2}$$

(2)

where, $\sigma$ is the standard deviation and $\mu$ is the mean of the distribution. From a given amplitude C-Scan, these two parameters can be extracted that can be utilized to quantify the microstructural heterogeneity and the average wave energy loss. $\sigma$ provides a metric to quantify the distribution of back-wall amplitudes and $\mu$ measures the average energy loss in the total area represented by the C-Scan.

3.3. Attenuation Coefficient

As the wave propagates through a sample, the ultrasonic energy is absorbed and scattered. The attenuation coefficient is a measure of the both these phenomenon (from Supplementary Materials). The details of measuring the attenuation parameter is given in the Supplementary Materials including details of diffraction correction (from Supplementary Materials). A general model to describe the ultrasonic attenuation coefficient is given by [29,30,31]:

$${\alpha }_{model}=A\times f^4+B\times f^2$$

(3)

where A and B are fitting coefficients to this model and serve to scale the contributions towards the fourth and second power of the frequency. In general, the major contribution to the attenuation coefficient is grain scattering, and both parameters A and B quantify different grain scattering mechanisms, i.e., Rayleigh and stochastic scattering. The parameter A will contain contributions to grain scattering from small grains ($\lambda >>D$) and B will contain contributions to grain scattering from large cluster of grains ($\lambda \approx D$). The as-fabricated IN718 laser-DED samples exhibited large cluster of small dendritic grains, ranging from 100–500 μm, making the two parameters A and B suitable to quantify their grain scattering contributions. Given that these microstructural features influence greatly the mechanical properties of the sample, we propose the use of the A and B attenuation fitting parameters as metrics to qualify laser-processed IN718 samples. The experimental setup used to measure the attenuation coefficient is described in the Supplementary Materials.

4. Results

4.1. Anisotropy Index

The degree of anisotropy was first characterized using the parameter $A_{LS}$ defined in Equation (1). The measurements were carried out on both PP1 and PP2 samples along with the wrought sample. Figure 2 shows the $A_{LS}$ values of the three samples. The hot-rolled sample $A_{LS}$ is 1.006 which is closer to the theoretical isotropic value $A_{LS}=1$, which was expected. PP1 sample was measured along three different directions, which also provides a range of values shown in the shaded region. It is apparent that the PP1 values are greater than 1. However, the PP2 values are significantly higher than 1, which suggests that the anisotropy is higher in PP2 samples.

4.2. Heterogeneity Parameter

The standard deviation ($\sigma$) and mean ($\mu$) heterogeneity parameters were extracted from C-Scans of the different samples using the procedure described in Figure 1. The results are shown in Figure 3. In the case of the PP1 sample, amplitude C-Scans for the three principal directions of propagation were analyzed as shown in Figure 3a–c. This produces a range of values for both $\sigma , \mu$ for the PP1 sample. The C-Scan of the build plane of PP2 sample exhibited more variation, having an area with significantly lower back-wall amplitudes as shown in Figure 3e. Therefore, we separated the C-Scan image form PP2 into two regions: a “good area” where the scatter was lower as shown in Figure 3d and the “full area” which is the complete amplitude C-Scan as shown in Figure 3e. Lastly, the C-Scan of the hot-rolled sample was analyzed as shown in Figure 3f. We can observe that it shows considerably lower heterogeneity compared to the DED samples. Figure 3 shows the six back-wall amplitude C-Scans that were analyzed.

The $\sigma$ and $\mu$ parameters that were extracted form the C-Scan processing are shown in Figure 4 were extracted using Equation (3). The $\sigma$ of the PP1 sample ranges from 2.89 dB to 7.72 dB as shown in Figure 4, with the highest value corresponding to its build direction. The $\sigma$ of PP2 “good area” falls within this range and it is closer to the PP1. The PP2 “full area” $\sigma$ however falls outside of the range. This is evidence of the influence that different processing parameters have on the microstructural heterogeneity of IN718 laser-DED samples, in this case by presenting an area with larger variation. The hot-rolled $\sigma$ is small even compared to the lower limit of the PP1. This is expected from a conventionally processed sample where microstructural heterogeneity does not exists or is insignificant.

Moving to the mean parameter, $\mu$ of PP2 “good area” falls within the PP1 window that ranges from −7.38 dB to −10.13 dB. Contrary to the $\sigma$ parameter, the PP2 “good area” $\mu$ value is closer to the lower limit of the PP1 range. This suggests that the PP2 “good area” exhibits a lower average energy loss than the PP1. However, the PP2 “full area” $\mu$ value is outside the PP1 $\mu$ range, which once again suggests that that these two processing parameters result in diffneret microstructural conditions. The hot-rolled $\mu$ value is the lowest of the group, hwich is once again expected.

4.3. Attenuation Coefficient

The experimentally measured attenuation coefficient data was fit with Equation (3) in the 2 to 7 MHz, given that a 5 MHz transducer was used to conduct the experiments. The attenuation of the hot-rolled sample was very small, therefore we used a 30 MHz transducer. Therefore, the frequency range from 10 to 40 MHz was used to fit the attenuation model. Figure 5 shows the attenuation coefficient curves and their corresponding fitted attenuation models for PP1, PP2 and hot-rolled samples. The figure also shows “peaks” or “humps” in the data, which will be discussed later. In general, there is some variability in the 3 different directions, however, there is also some differences between PP1 and PP2.

The fitting parameters in Equation (3) for the different samples are shown in Figure 6. Parameter A, which controls the quartic distribution is shown in Figure 6a. It is apparent that the PP1 values have a slightly broad range and that the PP2 “good area” falls within this range. A of the “full area” is also smaller compared to PP1. Both these follows the same trend of the other parameters including anisotropy and heterogeneity. The parameter B also follows a similar trend, where the PP2 “good area” is smaller than the PP1 value and PP2 “full area” is larger than PP1. These are once again consistent with the other parameters. The $A,B$ of the hot rolled sample could not be plotted in the same figures due to the mismatch of the values. For hot rolled, $A=1.26\times {10}^{-7}, B=8.95\times {10}^{-5}$, which are well outside the range of the DED values. This suggests that the hot rolled sample has a higher contribution from A, and lower contribution from B compared to the DED sample.

5. Discussion

The proposed ultrasonic NDE qualification parameters serve to quantify different aspects of laser-processed IN718 that most of the standard qualification procedures may not be able to address for such complex materials. Their significance in qualification and evaluating the microstructure is given below.

The $A_{LS}$ parameter is very similar to the anisotropic index, which has been used with a great deal of success to quantify the anisotropy. The $A_{LS}$ results suggest that hot-rolled sample agrees well with the isotropic value of ≈1, and the PP1 sample has a small overall variation of $3.00\%(1.112\pm 0.014)$. The fact that this average lies away from 1 signifies the anisotropy of the PP1 structure. More importantly, PP2 seems to have a higher anisotropy factor, which is farther away from 1. This suggests that the $A_{L}S$ parameter that we defined is sensitive to changes in the microstructure resulting from different processing parameters. Unlike the other parameters, $A_{LS}$ was not carried out in “good” and “full area” due the physical limitation of the transducer and sample size. For accurate measurements we had to use a larger diameter transducer which averages the response over a larger area. However, smaller diameter shear and longitudinal transducers can be used in the future to alleviate these concerns.

It can be argued that the microstructural heterogeneity of this complex IN718 DED material is the main reason characterization with most qualification standards becomes impossible. By extracting two parameters, $\sigma , \mu$, from the ultrasonic C-Scans which contain volumetric information of the sample, the heterogeneity can be quantified. Furthermore, this is the first instance of quantifying heterogeneity from C-Scan in this manner, and it is our belief that its adoption as a qualifying parameter for complex materials such as laser-processed nickel-based superalloys will be of significant importance in order to understand how different fabrication processing parameters influence microstructural heterogeneity. Its disadvantage is that since its calculation is only based on amplitude C-Scans, it does not provide information regarding the physical and chemical nature of the heterogeneity.

The A and B parameter from the attenuation coefficient model in Equation (3) quantify the different grain scattering mechanisms [32,33] contributions from the grains and its application for the DED IN718 samples showed interesting results. The A (quartic) parameter quantifies grain scattering in the Rayleigh regime, which in the case of the tested DED samples occurs from the cellular and dendritic columnar grains of average grain diameter of approximately 10 microns, complying with the condition $\lambda \gg D$. Therefore, if the microstructure has small grains, the A contribution is expected to be higher, which is the case for PP1 and PP2 “good area”. Similarly, the hot rolled sample has a much higher contribution from A, which is in good agreement with this. The B parameter quantifies the stochastic scattering contribution to ultrasonic attenuation with the condition that $\lambda \simeq D$. The larger B parameter of PP1 compared to PP2 “good area” suggests that the PP1 sample has a greater number of clusters that comply with $\lambda \simeq D$. Similarly, the B of the hot rolled sample was significantly smaller than both PP1 and PP2, which once again suggests that the contribution from large grains (in relation to wavelength) is minimal in the conventional sample. This is consistent with grain size shown earlier [34] of ≈10 μm, which is significantly smaller compared to ultrasonic wavelength (200 μm at 30 MHz). The challenge with consistently using this parameter is inability to fit the model equation with a good $R^2$ value. As the “peaks” in the $\alpha$ curves become sharper and larger, the goodness of fit will decrease, which will make it challenging to get consistent fit parameters for comparison.

From our previous work [34], the most significant characteristic that make the laser-DED IN718 samples unique is the presence of large cluster of dendritic grains. These clusters tend to produce backscatter that is significantly larger than conventional microstructures. The attenuation and backscatter results in the previous work showed the presence of “peaks” or “humps” in the frequency domain. There could be several reasons for these including signal processing methodology as reported earlier [34]. However, it is worth noting that this is a result of the large backscatter where the amplitude of the backscattered signals were somewhat equivalent to the amplitude of the backwall echo. Therefore, grain scatter is a significant contribution which can be directly related to the peaks or humps. From Figure 5, it is apparent that these peaks occur only in the DED sample, and specifically only in the cases where there is high scattering. For example, PP1 D1–D3 show similar behavior of the peaks. However, PP2 has a “good area” and seems to have a peak near 2 MHz and at 11 MHz which is outside of the probe bandwidth. PP2 “full area” has peaks at several frequencies. This suggests that the presence of these scattering peaks themselves could be used as indicators. It is also interesting to note that the peak width or quality factor is not the same between PP1 and PP2. This could also be another parameter to track. However, it is more important to first understand the reason for the existence of these peaks. Currently we do not understand their origins and speculate that they maybe a result of signal processing or from scattering transition [34]. In a different work by the authors that explored different weight percentages of IN718 and SS420, the peak size, location and number of peaks changed with increasing weight percentage of IN718 [35]. The change in weight percentage results in formation of intermetallics and different phases, which changes the attenuation response [35]. These results seem to suggest that the peak behavior is unique to DED samples compared to conventional samples. This suggest they can be used to qualification as well.

The objective of this article is to explore the different parameters that can be extracted from ultrasonic testing that could be utilized towards qualification. These parameters have to be rigorously tested using a large set of processing parameters and heat treatments and the results must be validated using destructive analysis. Broad definition of microstructural heterogeneity, grain size etc. for DED microstructures are presently missing the literature. Therefore, these need to be clearly defined from a destructive analysis perspective for this work to be completely validated. To limit the size of the article and maintain coherency, these explorations and validations using destructive analysis will be presented elsewhere. Furthermore, the NDE parameters need to be correlated to traditional qualification parameters such as hardness, tensile, fatigue and creep properties [24] carefully. This will help standardizing the NDE parameters and eventually help in identifying accuracy, cost-effectiveness and operational complexity of using these parameters vs. the traditional parameters. The focus of this article is to extract parameters from ultrasonic measurements that could then be used for qualification in the future. The next step is to carry out correlations in a large number of samples and establish these parameters.

6. Conclusions

This article explores the use of ultrasonic testing to qualify laser DED components, specifically focusing on IN718 cubes. The objective was to test the sensitivity of UT to changes that result from variation in processing parameters. Two processing parameters; speed and power were varied, the samples were tested with UT. Several parameters were extracted from the UT including anisotropy, microstructural heterogeneity, and grain scattering. The UT parameters were observed to show some degree of sensitivity between the samples. Compared with the reference sample (hot-rolled), it was observed that the UT parameters showed very good sensitivity. These suggest that the UT parameters could be used as qualification parameters, specifically during process qualification to detect defects, microstructural heterogeneity or post fabrication during heat treatment phase. We believe that these qualification parameters will contribute to the effort to understand the relationship between processing parameters and structural integrity. A thorough validation and evaluation of these parameters by comparing a large set of processing parameters will be carried out in the future.