Influence of Printing Parameters on Microstructure Influence

Introduction

Metal 3D printing has seen lots of attention in the past few years because of its ability to create near net shape parts but has not gotten to the same level as other forms of 3D printing since the mechanical properties of part fabricated by such a process cannot be predicted due to several defects. This put a huge hindrance in the quality control procedures of 3D printed metals. There are no standard way of testing and certifying these printed metals. A step in the right direction of coming up with a standardized test for 3D metal parts is to understand the effect of the printing parameters on the microstructure and properties of the As-printed part. These standards are required for quality control purposes for improving safety and performance. In order to come up with standards that cover testing methods, material and design of 3D printed part, the system needs to be extensively investigated. There several areas of metal additive manufacturing that requires testing and standardization. The two major components of metal 3D printing that need studying are the materials and processes.

The material feed stock used in these 3D printed machines need to be tested and certified to be used in the appropriate printer. The properties of the materials that are certified and control include the material particle size distribution. Also, the material parking factor is another property that need to be considered when choose material for 3D printing. If a company or individual prints any other material, that has not been certified, there is a potential to void one’s warranty.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

Lastly the printing process needs to be standardized to better compare the performance of different additive processes per intended application. The process parameters can be categorized as laser parameters or build parameters. The laser parameter include power, the scan speed, the type of laser input (continuous or pulsating). Build parameters include the build direction, orientation, layer thickness.

Our study is narrowed to the process parameter influence on the properties of the printed part. The parameters that are studied here include the laser power, the scan speed and the hatching distance. These parameters are the most sensitive parameters that affect the nature of the final part. Also, these are among the parameters that can easily be changed without voiding the warranty of the 3D printing equipment.

In this study, a range of printing parameters is investigated to ascertain their influence on the microstructure. The laser powers used are 50W, 200W, and 250 W with speeds of 250mm/s,750mm/s,1250mm/s, and 1750 mm/s. The hatch spacing is also varied, 80microns,100microns,120microns, and 140micros 140 μm. A very common way of comparing printing parameters is the use of Volumetric Energy Density (V.E.D) which is representative of the laser power per unit volume delivered to the part during the printing process. This quantity combines the varying parameters used in the experiment, namely laser power(P), scan speed(v) and the hatch spacing(h). The energy density is calculated using the formula; Volumetric Energy Density(J/mm3) = /(ℎ∗ ∗ ). Other parameters that are used in place of the Volumetric Energy density are the Global Energy Density and Linear Energy Density.

Method

Metallographic preparation of samples

Twenty-five samples with different printing parameters were fabricated using EOS M280 DMLS machine using AISI 316L stainless steel powders provided by the same company (EOS). The chemical composition of the 316L powder is listed in table 1. The steel powder has a size distribution between 15 μm– 45 μm. The powders provided by EOS was certified by Nadcap to be used in this model of DMLS machine. The various parameters used to fabricate these samples are shown in table 2. There are three different laser powers 150,200 and 250W. Each sample is printed with the with varying hatch spacing (0.08,0.10,0.12 and 0.14 mm) and laser scan speeds (250,750,1250 and 1750 mm/s). The scanning strategy used in this process is the stripe pattern with stripe overlap and width of 0.08 mm and 10mm respectively. Square cuboids, measuring 20 mm x 20 mm x 15 mm, were fabricated in chamber filled with nitrogen gas with chamber temperature of 375 K

A 5 mm thickness of the sample was cut from the top of each cuboid with Wire Electric Discharge Machining. The 5 mm thickness pieces were mounted in phenolic and metallographically prepared by grinding and polishing. The samples were ground with 120,500,800 and 1200 grinding grit SiC papers. Polishing was done with Al₂O₃ suspension to a mirror finish.

Table 1 Printing parameters for AISI 316L Alloy

Determining the relative density

The relative densities of the printed parts are determined by using Archimedes principle according to the “ASTM B962-17 Standard Test Methods for Density of Compacted or Sintered Powder Metallurgy (PM) Products Using Archimedes’ Principle”. The samples are cleaned, first with diluted cleaning soap (Windex) for 30 minutes to remove dirt trapped on the sample surface. Secondly, the sample are rinsed completely with deionized water to get the cleaning soap off the sample surface for 30 min. Lastly, the samples are then cleaned, also for 30min, with ethanol to get rid of any remaining solvent on the sample. All the cleaning processes is done with a Branson Ultrasonic Bath with pressure 68 Pa. The samples are weighed in both air and distilled water. The mass of the samples in air is recorded as A at room temperature with a Mettler Toledo balance. For the mass in water, the pores on the surface of the sample are coated with lacquer to prevent water from entering the sample. Lacquer #8__ was used to seal the surface pore of the printed samples. These samples are made to cure under room temperature for 8 hours. Another layer of the lacquer was applied to make sure that the pores are all successfully covered whiles making sure not to apply to much lacquer. Apply too much of the lacquer will cause a huge increase in mass which might inherently go over the maximum mass (in addition to the mass of the water filled beaker) specified for the balance.  The coated sample is weighed in air and note as B. Then, the coated samples are submerged and weighed to find the mass under water. The densities are determined using this formula;

Part density, ρ_p = _w =the density of water

A graph of relative densities against the energy densities is shown in figure 1 to clearly show the relation between these two quantities.

Figure 1 Density of samples as a function of Energy Density

Determining the porosities

The porosity is determined using the image analysis software, image J. The un-etched samples were examined under light microscopy using the Zeiss M2 Imager Microscope. Micrographs under various magnification were taken by the microscope to studying the porosity level. For each sample, 10 separate images under the same magnification was used for processing. Each of the 10 images are taken through thresholding, were they are binarized, based on the threshold value that clearly distinguishes the gray scale values of the pores from the sample itself. The average porosity is determined from the percentage area fraction from each single micrograph. This average porosity is then used as the porosity of the samples. The porosity levels are plotted against the Volumetric Energy Density of the samples shown in figure 2 below.  The correlation of between the density and porosity is also determined and graphed as represented in figure 3.

Figure 2:: Porosity against the Energy Density of the Samples

Figure 3 Density and Porosity Correlation

Hardness Testing

Hardness measurements were obtained using the Vickers Micro-Hardness Tester. This tester using a pyramid diamond tip indenter to apply a known force to the point of interest on the sample surface. The hardness of the printed samples was tested with a force of 300gf which is equivalent to 2.94N. Ten different indentations were done per sample with dwell time for each indentation is 15 seconds. During the hardness testing, all indents that fell into pores were repeated. Indents that fall within a pore gave a reading that deviated widely from the other measurement values. The hardness values are obtained by measuring the diagonals of each quadrangular pyramid shaped indent and find the corresponding hardness to the measured diagonals. The average hardness values are done determined and the standard deviations used as the error bars for the graph shown in figure 4. The hardness values are plotted against the V.E.D of the samples.

Figure 4: Micro-hardness of DMLS AISI 316L samples

-Less than 1%

-More than 1%

Investigating the surface particle morphology

Samples with less than 1% porosity level were chosen for further microstructural applications. These samples were re-polished after the light microscopy examinations and immediately etched with marble’s reagent. The etchant used is a combination of CuSO₄, HCl and H₂O. Each sample was immersed fully for 25 seconds for the etchant to react fully with the sample. The samples were removed from the etchant and rinsed thoroughly under running water and then neutralized with ethanol. The samples were studied with a scanning electron microscope (SEM) to study morphology of the sample surface. The SEM was used with 9kV of energy, with a beam intensity of 9.0 and at 25 mm working distance. The SEM chamber is kept under vacuum in the presence of nitrogen gas. The entirety of the sample was studied under low magnification. Micro and Nano sub-particles of each samples were examined at high magnifications. The samples were wrapped in carbon tape to reduce the rate of charging on the sample surface. Electron Dispersive Spectroscopy (EDS) was carried out to find out there were any chemical variations among the samples. The microstructures of these eight samples are shown in figures 5-12 below.

Results

Relative density

Densities of the parts from using the ASTM B962-17 Standard Test Methods for Density of Compacted or Sintered Powder Metallurgy (PM) Products Using Archimedes’ Principle, is plotted against V.E.D as shown in figure 1. The density of the part increases as the VED increases but it reaches a region where it plateaus and remains steady till it reaches a point where is begins to drop again. This is corroborated by the results from the porosity level of the samples as a function of VED.

Porosities

To study just the scan speeds and hatching distance effect on the measured physical properties, all the other printing parameters were kept constant. The energy density is the quantity used to represent the amount of laser energy supplied to each sample per cubic millimetres. Generally, the porosity level begins to reduce as the energy density increases because more energy is fed into the material for fusion. Hence as the energy increase more of the particles fuse together create less pores in the part. The study showed that for AISI 316L parts, when the V.E.D is lower than 20 J/mm3, the porosity is approximately 50% and higher. V.E.D within 30-40 J/mm3 gives a porosity level between 1%-20% and having the energy density higher than 40 J/mm3 bring the pore counts further to a minimum value (less than 1%). It is very crucial to not that as the V.E.D goes beyond a certain threshold (~120 J/mm3 for AISI 316L) the pore begin to re-emerge. The pores reappear is attributed to the fact that at that high energy level the material starts to evaporate, and the laser penetrates deeper than is expected which results in keyhole formation (a site with affinity to pore formation)¹. The energy density does not correspond to an exact porosity level but a range. There are samples with the same energy density values which show a slight variation in porosity because the individual parameters are entirely different from each other. Comparing the individual parameters also shed more light on the matter. For samples with just the scanning speed or hatching spacing changing had a linear relationship with the level of porosity. Samples with varying scan speed showed a direct relation to porosity level, as the laser scan speed increases so does the level of porosity in the part, same can be said for increase in hatch spacing and porosity level

 However, the laser power follows the same trend as the V.E.D against porosity for the same reasons. It is evident from the plot shown in figure 1 that the porosity of the samples initially starts to decrease as the energy supplied to the sample increases. However, the porosity reaches a minimum value at 0.017% area fraction and then starts to increase after passing this threshold value. There was an outlier at energy density 55.56 J/mm³ which had a porosity level of 2.4%.

Hardness

The hardness values of the samples do not seem to have a particular trend against the V.E.D, but the values are a little higher than the hardness of wrought 316L steel which is around 195 HV². The higher hardness of the 3D printed parts is a result of the fine microstructures observed on the surface of the part. These fine structures increase the deformation density of the samples and therefore make it harder. The rapid heating and cooling cycles of the printing process is gives rise to these fine substructures.

Surface particle morphology

The first samples with the least amount of V.E. D showed a lack of fusion voids with un-melted powders within these voids. This is mainly due to the low amount of energy delivered to the powders to cause fusion of these particle together. This is generally due to a combination of high scan speed and low laser power³

G3C2

b

a

Bigger pores

Un-melted

powder

Smaller pores

d

c

Micro-crack

Crack

f

e

Columnar Structures

Cellular Structures

Figure 10 sample G3C2 showing non-uniform distribution of bigger pores  amist a uniform distribution of small pores a)  Unmelted powder b)  Crack c) Micro-crack d)  Cellular Structures e)  Columnar Structures f)

G4B2

b

a

Micro-crack

Pores

d

c

Grain with pores around the boundary

Cellular structures

Micro pores on grain boundary

Columnar structures

Melt pool boundary

Grain boundary

Figure 11 Sample G4B2 showing Relatively few amounts of bigger pores a) micro-crack b) grain with pores along the grain boundary c) Columnar structures d) e)

 Sample G3C2 has a hatch spacing of 0.12 mm, power of 200W and speed of 1250mm/s. The fast speed causes the laser to pass over the powders so fast that there is not enough time for the powders to absorb enough energy to fuse together. Also, it is seen that there are cracks in the samples, both macro and micro-cracks are seen as shown in figure 5. As the V. E. D increases the number and size of crack reduces, and the pores also begin to reduce in size and number. However, as the number of pores reduce in the samples we can also see that the micro pores begin to align along grain boundaries. In some cases, there are micro pores aligned within the grains of the sample. In both cases, there is high probability of intergranular and intragranular fractures when the sample is subject to mechanical loading conditions. In all the eight remaining samples, there are fine sub structures seen on the nanometre level on the surface of these samples. These fine substructures are seen to be both fine cellular structure and columnar structures.

G1C2

a

b

Micro-crack

pores

Columnar structures

c

d

Cellular structures

Figure 6 Sample G1C2 showing non-uniform distribution of bigger pores on the sample  a)micro-crack passing through a grain b)Cellular structure c) Columnar structures d)

Discussion

The correlation of Density and porosity is known to be a linear relationship. The graph in figure 3 tells a slightly different story. This is mainly because both methods have challenges with the techniques. For the porosity measurement, Image J was used in the analysis which raises a lot of problems. The first is that this type of analysis is more of a qualitative method since it does not look at the pores in the entirety of the sample but just the surface. The thresholding done in this process fails to capture a lot of the small pores due to the level of the image quality obtained with the microscope. The density measure does not consider the amount of un-melted powders available in the sample that contributes to the mass but not the microstructure of the part. The best way to go about this is to determine the density of the part with a better method like using the gas pycnometer and then using the density to calculate the porosity based on the densification of the part. Another method will be to employ a Nano-CT/Synchrotron CT to analyse the porosity and then calculating the densification from the porosity

G2B1

Alignment of small pores

Crack

Pores

Grain

d

c

Columnar structures

b

a

Figure 8 Sample G2B1 showing bigger pore with non-uniform distribution with few small pores aligned a) cellular structures b)  Columnar structures c)

G4B3

a

b

Pores

Cellular structures

Columnar structures

c

d

Alignment of pores within a grain

Micro-crack on grain boundary

Alignment of pores on the grain boundary

Figure 12 Sample G4B3 showing very few big pores a)  Cellular and columnar structures b)  alignment of pores within a grain and on grain boundary c)  micro-crack formed because of pores alignment on grain boundary d)

G1C3

a

b

Alignment of small pores

Big pores

Small pores

Columnar structures

c

Cellular structures

Figure 7 Sample G1C3 showing equal distribution of small and bigger pores a)  Weak points created by pores aligning b)   Columnar and cellular structures c)

G2B2

a

b

Grain boundary

Pore

Micro pores along grain boundary

c

c

Columnar structures

Cellular Structures

Figure 9 Sample G3C2 showing a single small pore a) Columnar structures b) Cellular structures c)

G1B2

b

a

Alignment of pores

pores

d

c

Cellular structures

Columnar structures

Figure 5 Micrographs of sample G1B2 showing small pores uniformly distributed throughout the surface of the sample a) pores aligned forming weak points on the sample b) cellular substructures c) and columnar structures d)

Future work

• X-ray Diffraction for Identification of phases present in 3D printed 316L samples and also for residual stress analysis
• Differential Scanning Calorimetry to study any phase transformation of 3D printed Stainless Steel Alloy that occurs within 0-600°C
• Steady state deformation behaviour of 3D printed 316L stainless steel Alloy (Tensile Testing)
• High strain rate deformation behaviour of 3D printed 316L Stainless Steel Alloy (Compressive Testing)

References

1.  Peen Rawn. 3D Printing of 316L Stainless Steel and Its Effect on Microstructure and Mechanical Properties. 2017.

2.  Yusuf S, Chen Y, Boardman R, Yang S, Gao N. Investigation on Porosity and Microhardness of 316L Stainless Steel Fabricated by Selective Laser Melting. Metals (Basel). 2017;7(2):64. doi:10.3390/met7020064

3.  Yusuf SM, Gao N. Influence of energy density on metallurgy and properties in metal additive manufacturing. Mater Sci Technol (United Kingdom). 2017;33(11):1269-1289. doi:10.1080/02670836.2017.1289444