Metal AM at Euro PM2015: Superalloys, powder atomisation and advances in inkjet and LMD processes

Heat treatment and mechanical testing of IN-939 superalloy

The session on superalloys was opened with a contribution from Hakan Brodin (Siemens Industrial Turbomachinery AB, Sweden) on the heat treatment and mechanical testing of a selective laser melted superalloy, IN-939.

Alloy IN-939 is a corrosion resistant nickel-based superalloy, where components are typically manufactured by precision casting and currently mainly used as blade and vane material in industrial gas turbines. Using standard manufacturing routes, the alloy is never used in the as-manufactured condition because of the heavy segregation that would appear after casting. Therefore, a solution heat treatment is always necessary, followed by an aging procedure to achieve the appropriate strength and ductility levels. The nominal material composition for castings is given in Table 1.

The reported study focussed on mechanical testing experiments after heat treatment of IN-939 manufactured by Selective Laser Melting AM technology. A typical as-manufactured IN-939 microstructure is shown in Fig. 1. In this figure, the build direction was perpendicular to the image plane, i.e. out of the plane. The powder used was argon gas atomised and consolidation was carried out using an EOS M270 DM machine, with 195 W laser power and 20 µm layer thickness.

Anisotropy can be a problem in AM materials and, therefore, two test specimen orientations were evaluated, 0° (parallel to the build platform / perpendicular to the build direction) and 90° (parallel to the build direction). The initial heat treatment used during the material development of IN-939 was a two stage heat treatment. Due to the very low tensile and creep ductility achieved, the heat treatment was later modified to improve ductility at the expense of reduced tensile strength. The proposed heat treatment route was then a four stage heat treatment. Heat treatment procedures are shown in Table 2. In the current work, attention was on the heat treatments according to procedures 1 and 2. For products with high performance demands, a HIP procedure was proposed prior to step 1.

Tensile testing at ambient and elevated temperature was performed on a) as-manufactured, b) HIPed and heat treated material. The measured values of yield stress, ultimate tensile stress, elongation at fracture and energy at fracture for as-manufactured IN-939 are shown in Fig. 2. These results display anisotropy and a material with high elongation at fracture compared to the typical values of 2-8% for heat treated precision cast IN-939. The main reason for the good elongation and ductility is the absence of heat treatment, where a ɣ-ɣI microstructure has not developed due to the fast cooling during SLM processing.

In order to allow a comparison of different heat treatments, Fig. 3 shows yield stress and ultimate tensile stress for the 0° direction after different heat-treatments, compared to precision casting data. The initial (as-manufactured) strength of an SLM material is as strong as the corresponding cast material after a solution heat treatment and aging. By using a proper solution heat treatment and aging procedure, the tensile strength of SLM processed material will be superior to a casting.

Stiffness in the different test directions 0° and 90° versus material condition is shown in Fig. 4. Reference in-house data for cast material is also included in this figure. It is clear that hot isostatic pressing is beneficial for material behaviour. Even though the material is already homogeneous before HIP, containing low porosity and being crack free, the results indicate a significant change in stiffness. Creep testing has been performed for a solution strengthened superalloy as rupture tests in the 0° and 90° directions (Figs. 5 and 6). Creep tests have been reported as stress as a function of the Larson-Miller Parameter (LMP). The Larson-Miller parameter is defined as:

LMP = T.[log(t) + C]. 10^-2

where T = temperature [K],t = time 
[h] and C is the Larson-Miller constant, typically set at 20.

The 0° direction is known to be inferior with regard to creep performance (rupture time, elongation at rupture) than the 90° direction. The performance of the 90° direction in the as-manufactured condition can be compared to cast material.

The tensile testing of SLM-processed IN-939 has indicated that the material is anisotropic with regard to strength in the as-manufactured condition. This can be seen in the tensile test results (both yield stress and ultimate tensile stress). Elongation and energy at fracture do not seem to be as significantly influenced by the anisotropy. Material stiffness is lower in the as-manufactured condition compared to what can be expected in a material manufactured by precision casting. A HIP process seems to be able to improve the poor stiffness observed in the as-manufactured material.

The anisotropic behaviour of IN-939 can also be observed in the creep behaviour of the alloy. In the as-manufactured condition, the material creep behaviour is significantly inferior in the 0° direction compared to a casting. On the other hand, in the 90° direction, the creep resistance is on a par with precision cast material. A heat treatment can improve the creep properties in the 0° direction with the consequence that the better performing (90°) direction reduces the resistance against creep damage.

Results in the reported study indicate that a further refinement of the heat treatment could have the potential to remove the creep life anisotropy. Further work is needed to optimise the heat treatment in this respect.

Improving quality and production capacity for Ni based powders

The next paper in this session was from P Vikner, R Giraud and C Mayer (Erasteel, France) and M Sarasola (Metallied Powder Solutions, Spain) and presented the process improvements made by Erasteel to produce quality Ni- and Co-based superalloy powders for AM in terms of interstitial cleanliness, particle size and morphology.

In order to improve the nitrogen and oxygen content in the powders, it important to exercise good control over the vacuum induction melting and the atomisation processes and to limit the contact between air and hot powder. It is also important to prevent moisture and foreign particles from entering the powder. Moisture would have a negative impact on flowability and foreign particles can degrade the performance of the final product. In the production line at Metallied, the powder is constantly under a protective atmosphere with a view to avoiding any pick-up of oxygen, nitrogen, moisture or foreign particles. In order to further improve the cleanliness, all powder handling is done in a new dedicated facility that is separated from the upstream production. An inert gas blender and inert gas packing ensure a clean homogenous powder.

By addressing the sources of nitrogen pick-up at all the stages of powder fabrication, the nitrogen level in the final alloys has been reduced, as shown in Fig. 7 for Ni718 for laser melting powder bed applications, requiring powder particle sizes between 10 and 53 μm. This improvement in Ni718 superalloys will decrease the risk of TiN stringers formation in the powder (or subsequently in the built part). Such stringers, due to their low ductility, would become initiation sites for cracks. With a similar objective of reducing the risk of the formation of detrimental oxides, the oxygen level in Ni718 can also be kept well below the maximum level as shown in Fig. 7 for fine powders between 10 and 53 μm.

In order to exercise control over powder size and morphology, it is necessary to adjust atomisation parameters. It is also desirable to be able to monitor atomisation parameters to ensure that the process is under control. In water and gas atomisation, most of the relevant process parameters can be covered by the “melt-to-gas flow rates ratio”. Increasing this ratio leads to an increase of the median value of the particle size distribution (d50) as shown in the left hand side of Fig. 8 for tin and copper powder.

In the Erasteel process, this ratio is monitored through the in-line measurement of powder temperature at the bottom of the atomiser. For a given gas nozzle configuration, the metal feed rate was varied, leading to a variation in the melt-to-gas flow rates ratio. If the gas feed is kept constant, the particle size will vary as a function of gas temperature. On the right side of Fig. 8, a plot of the d50 against the relative gas temperature for two Ni based alloys, Ni738 and SYP3 (Astroloy), confirms this trend.

For Additive Manufacturing applications, the size and size distribution of the powder are critical, but it is also necessary to control the shape of the powder. The powder must be able to flow well in the Additive Manufacturing devices, to spread well in the case of powder bed applications and to limit the risks of induced porosity. It is therefore important to be able to produce rather regular spherical powder and to limit the presence of satellite particles.

By tuning the gas to melt flow ratio, it is possible to adapt the morphology of powders to targeted Additive Manufacturing applications, as shown in Fig. 9. With a high gas to melt flow ratio, the morphology of particles smaller than 45 μm with spherical, regular particles is adapted to laser melting powder bed technology, but the higher fraction of powders, between 45 and 106 μm, shows very irregular particles. It was thus necessary to tune the gas to melt ratio to improve the morphology of particles between 45 and 106 μm. This regularity of morphology is mandatory for deposition or cladding type technologies (targeted size of particles between 45 and 125 μm depending on the type of machine), with flowability being a determining factor.

In this first stage of process optimisation, Erasteel has focussed effort on the process-related issues of reduced interstitial levels and control over the size distribution and morphology of the powders. This effort is being continued with the view of reaching a complete understanding.

Heat treatment optimisation of Hastelloy X superalloy

The final paper reviewed in this session came from Giulio Marchese, Sara Biamino, Matteo Pavese, Daniele Urgues, Mariangela Lombardi and Paolo Fino (Politecnico di Torino, Italy) and Giangranco Vallillo (GE AVIO s.r.l., Italy) and addressed the optimisation of heat treatment of Hastelloy X superalloy, processed by Direct Metal Laser Sintering (or Selective Laser Melting).

Hastelloy X is a widely used superalloy. It is a nickel-chromium-iron-molybdenum alloy with outstanding high temperature strength and oxidation resistance. Given these properties, it is ideal for gas turbine engines, aircraft, industrial furnaces and chemical processing applications, such as pyrolysis tubes and muffles. The microstructure and hardness of Hastelloy X, processed by DMLS, are very different from those achieved in traditional manufacturing and, therefore, the authors considered it to be necessary to specifically study the response to heat treatment.

The reported work has studied the effect of heat treatment on the microstructure and hardness of a Hastelloy X alloy (composition as in Table 3 and particle size distribution between 15 µm and 53 µm) processed by DMLS, using a EOSINT M280 machine, by GE AVIO s.r.l. According to the literature, performing heat treatment can increase grain size and make it possible to control M23C6 carbide precipitation along grain boundaries, which is thought to increase creep resistance. For these reasons, it was considered crucial to study the solution treatment before the investigation of aging treatment, in order to optimise heat treatment parameters.

The as-built samples, provided by AVIO, had a low level of total porosity of about 0.4% ± 0.1%, as measured by optical microscopy. Metallographic assessment of these samples showed that the rapid and localised melting and cooling in the DMLS process created a very fine microstructure, as illustrated in Fig. 10. Along the plane yz it is possible to observe the shape of melt pools with their contours.

Solution treatments were then carried out at 1175°C and 1066°C. The influences of the various solution treatments on grain size and hardness are summarised in Table 4. For solution treatments at 1175°C, a reduction of hardness and grain growth can be observed (Table 4). Also, the grain growth reaches a limit after a solution time of one hour. The grains do not grow further with increasing treatment time and the hardness remains constant at its lowest value. By contrast, the heat treatment at 1066°C was insufficient to ensure an effective solution response. In fact, the microstructure remained similar to that of the as-built samples (shown in Fig. 10), as confirmed by the small reduction in hardness observed. Therefore, the chosen treatment for the further aging trials was a solution treatment at 1175°C for 60 minutes.

After this solution treatment, the microstructure is composed of equiaxed grains of gamma phase, twins and dark points that could indicate carbides precipitates in the grains and on the grain boundaries. XRD analysis indicates the presence of gamma phase only with a 3.60 Å lattice parameter, while the carbides have not been identified probably because the levels present are below the sensitivity threshold of the instrument. Even if they are present in small quantities, however, FESEM analysis shows some possible residual carbides in the gamma phase matrix, with an extremely fine size lower than 1 μm (Fig. 11). These microstructures are ideal as a starting point for the aging treatments, since no coarse or large carbides are observed. So, appropriate aging treatments can be determined with the aim of controlling the type and shape of carbide precipitates.

Different aging treatments were performed at 745°C and 788°C for 3 h and 6 h. After the aging treatments, the grain size remains the same, while the hardness increases due to precipitation of carbides, as reported in Table 5. Precipitation of the carbides occurs predominantly along the grain boundaries, but they can also be found within the grains, as shown in Fig. 12. The hardness of the Hastelloy X is seen to increase substantially after an aging treatment of 3 h at 745°C or at 788°C, due to a very homogeneous precipitation of carbides. The hardness obtained is similar to that cited in the literature or in technical sheets for commercial alloys.

An important objective of further work by this group will be the determination of the precipitated carbide type, both by XRD and TEM. This knowledge will allow the tailoring of the heat treatment to obtain an appropriate precipitation of M23C6. The differences between the traditional aging treatment and the one optimised for the DMLS-produced Hastelloy X, and the consequent effect on the mechanical properties at both room and high temperature, will allow DMLS to become a versatile and efficient technique to produce components with this superalloy.

Precision ink jet printing on a powder bed

The first contribution to the Special Processes and Materials session came from Robert Frykholm, Bo-Goran Andersson and Ralf Carlstrom (Höganäs AB) and reported on progress with the company’s Digital Metal^® technology.

This technology is based on precision ink-jet printing of an organic binder on a powder bed, followed by a separate sintering treatment to obtain final strength of the component. This approach offers a number of benefits over the powder bed AM technologies, which involve melting of powders. The whole volume of the build box can be maximised as the components can be packed tightly because no account of thermal conductivity needs to be taken. Also, there is no need, in principle, for building support structures during printing because the components are supported by the powder bed in the build box. As printing of the metal powders occurs at room temperature followed by a separate sintering, there is no heat involved during building and printing can be performed without protective gas. Since no melting takes place during building, green components can be produced with very high detail levels and tolerances. Also, since forming and heat treatment are separated, this allows for a wide materials selection range. The complete Digital Metal process is shown in Fig. 13.

With standard processing of 316L stainless steel in the Digital Metal process, a density level of approximately 96% of theoretical can be achieved. If higher density is desired it is possible to apply Hot Isostatic Pressing (HIP). With this technique, almost full density can be reached. However, there is also an alternative route to obtaining high density and that is by reducing the printed volume of the component. The principle in this case is to print only the surface layer of the component and to allow the bulk of the component to be dense packed loose powder contained by the printed shell. This technique is referred to as shell-printing. Even though the powder in the bulk is loose, the powder particles are close packed and not able to move. The principle of shell-printing is shown in Fig. 14.

As well as the positive effect on density, there are additional benefits with this technique. Since the total printed volume is reduced, both time and the amount of organic additive can be reduced. The fact that the amount of organic additive can be reduced is important when processing elements sensitive to carbon, e.g. Ti-alloys. This makes debinding of components easier and faster, both because the total amount of C needed to be removed is reduced, but also because all C is present in the component surface at the start of the process and does not need to migrate from the bulk.

The reported experimental study was performed with 316L stainless steel powder. Conventional printing and shell-printing were both applied. Sintering of the steel was performed in vacuum or hydrogen at 1360-1380°C for 2 h. Hot Isostatic Pressing was performed at 1150°C for 1.5 h with a pressure of 1000 bar. To investigate the mechanical performance of 316L, tensile test bars were produced. To date, the stainless steel 316L has proved to be the workhorse material in the Digital Metal process. It can be processed to close to full density with correct sintering settings. The material is sintered at high temperature, close to the solidus temperature. When sintered in vacuum, a partial pressure of Ar is applied. This is done to reduce the loss of Cr from the surface, which otherwise might occur due to the vapour pressure of Cr. Loss of Cr has to be prevented in order to maintain the corrosion resistance of the steel.

Using the standard process, 316L can be printed and sintered to around 7.7 g/cm3 density (96-97% of theoretical density) (Table 6). To reach higher densities, the sintering temperature could be increased, but this is normally not done since the higher temperature would have a negative effect on achievable tolerances. If higher densities are required, HIP can be performed. A prerequisite for this treatment is that the density of the sintered specimen must be high enough not to have a structure with interconnected porosity. With closed porosity, pressure will be applied uniformly on the outer component surfaces and therefore induce densification. For a standard Digital Metal processed 316L component, the porosity is low enough to allow effective HIP.

To obtain high densification directly in the sintering stage, the shell-printing technique can be applied. The result of applying this process can be seen in Fig. 15. In this case, a cylinder with height 24 mm and diameter 24 mm was produced by printing only a 1 mm thick surface shell. The image shows pore and microstructure after sintering, with the surface shell to the left. As can be seen, the addition of organic constituent has an effect on densification. The surface zone has porosity on the same level as for a standard printed specimen, while the bulk has significantly reduced porosity. This is visualised more effectively in Fig. 16, which displays the structure of the bulk. Only small and evenly distributed pores are present and the densification is almost on the same level as for HIP.

Data from the mechanical testing are given in Table 6 together with MPIF 35 standard data for MIM components. It can be seen that the material is well above the minimum levels and that performance, in fact, matches typical values. There is no large difference between standard and shell printed specimens, although shell-printing offers somewhat higher strength and hardness.

Carbon contents after sintering are shown in Table 7. This table displays a clear effect of the reduced amount of organic constituent in the shell-printing process, with reduction of final C-level from 0.013 wt% down to 0.003 wt%. Comparing with the MPIF 35 standard, it can also be seen that both materials are within the limit for C-content.

Alternative approaches to densifying printed green parts

The next paper remained with ink-jet printing technology and was presented by Juan Isaza, Claus Aumund-Kopp, Sandra Wieland, Frank Petzoldt, Mathis Bauschulte and Dirk Godlinski (Fraunhofer IFAM, Bremen, Germany). The authors highlighted two alternative approaches to densifying the printed green parts; super-solidus liquid phase sintering to high density or partial sintering without shrinkage followed by infiltration.

The paper reported on case studies aimed at developing innovative applications based on each of these approaches. The first case study addressed the potential for producing complex tools (with conformal cooling channels) by liquid phase sintering of printed parts made from either X190CrVMo20 (M390) or HS 6-5-2 (M2) tool steel. X190VrVMo20’s eutectic reaction requires extremely careful control of sintering temperature. At sintering temperatures above 1265°C, the surface of the samples becomes smoother, but the shape begins to deteriorate, because of the high amount of liquid phase. At temperatures below 1260°C, the density and the surface quality decrease because nearly no liquid phase is present. At temperatures around 1263°C, there seems to be an optimum balance between shape retention, surface quality and density.

With additional dwelling steps at different temperatures below the maximum sintering temperature and low heating ramps to ensure shape retention, 99% of theoretical density can be achieved. Another approach to improving the density is to lower the layer thickness of the printing process, reaching slightly higher sintered densities of an additional 0.5%. The hardness after sintering and cooling down is typically between 50 and 55 HRC and can be further improved by heat treatment.

The high speed steel HS 6-5-2 shows two equilibria between melt and solid in its phase diagram. Sintering temperatures in the lower range between 1230°C and 1290°C do not lead to full density in reasonable sintering times. In contrast, a sintering programme in the range of the second equilibrium with a long dwelling time of 90 min at 1320°C followed by a short dwelling time of 15 min at 1330°C leads to a density of 98 to 99% of theoretical density combined with good shape retention and a smooth surface. Lowering the layer thickness to 150 μm compared with 177 μm leads to the same slightly higher sintered densities of an additional 0.5% as observed with the X190CrVMo20 powder.

The surface quality of the parts, which is determined by powder particle size and the surface tension of the liquid phase in sintering, makes secondary finishing operations, at least on functional surfaces, necessary. Nevertheless, large volume machining can at least be avoided and more complex parts, for example with internal opened cavities, can be produced. Hence, applying this AM technology to tool steels could lead to cost savings and shorter production times for tools.

The innovative application, targeted by the infiltration approach, was the manufacture of MRI-compatible medical implants. The aim of the second study was, therefore, to reduce the magnetic susceptibility of titanium while retaining its outstanding material properties. For this, a titanium-silver material was manufactured and tested for MRI compatibility. By combining the paramagnetic titanium (matrix) with the strong diamagnetic silver, the formation of artefacts in MRI should be minimised. Porous samples of alloyed titanium powder (Ti-6Al-4V) were produced by ink-jet printing and were then sintered and infiltrated with silver.

Ti-6Al-4V rods were prepared by means of inkjet-based 3D printing and were subsequently debound and sintered. Titanium gas atomised powder with a median particle size D50 of 38 μm was used. This material had a bulk density of 2.44 g/cm³ and a tap density of 2.71 g/cm³. The binder used during the process is sold under the trade name Ex1-Lab-Binder-02. Due to the shrinkage behaviour during sintering, two different geometries were designed for printing: Ø 4 x 50 mm³ for density levels of 75% and 67% and Ø 4.3 x 56 mm³ for a density level of 79%. Table 8 shows the debinding and sintering basic data for the different samples.

In total, six samples were infiltrated (two samples for each density level) at 980°C with two different soaking times (30 min and 20 min) in the furnace. Silver powder (99.9 %) with a particle size distribution of -10+20 μm was used for the infiltration. Due to the variations in shrinkage behaviour in the set of samples, they were manually post-processed after infiltration to achieve a uniform geometry for subsequent characterisation. The microstructural analyses on all of the samples showed that, after the infiltration, no porosity was left. In the image taken by optical microscope at low magnification (Fig. 17, left), the layered structure originating from the printing process is clearly visible. The higher magnification in REM (Fig. 17, right) reveals that, during the infiltration process, diffusion also takes place and additional phases are formed from the original Ti-6Al-4V and silver. According to XRD assessments, this is mainly TiAg. The amount and distribution of the phases present varies depending on the density of the titanium and the soaking time.

The susceptibility of the samples was determined after post-processing. For each sample, three values were measured and then averaged. The resulting mass susceptibility was calculated from the predetermined density for the samples and the measured volume susceptibility. By combining the original titanium with the diamagnetic silver, it was possible to reduce the magnetic susceptibility of the material. A silver content of 33 vol. % (53 wt.%) leads to a reduction of volume susceptibility of approximately 38%. At a lower silver percentage 25 vol.% (45 wt.%), the reduction of volume susceptibility was up to 29%. The correlation between the titanium content and susceptibility can be seen in Fig. 18 (left). It was also noted that the susceptibility is higher with shorter soaking times.

The electrical conductivity of the samples was determined by four-point measurement. With increasing silver content, it is possible to trace a significant increase in the conductivity compared to solid material (Grade 5) (Fig. 18 right).

To determine the artefact behaviour of the samples in the MRI, scans of the samples were made in a water phantom. The samples were then aligned perpendicular to the static field of the tomograph. To visualise the distortion field caused by the samples, a TRUFI sequence (True Fast Imaging With steady precession) was used.

All samples show strong field distortions during measurement. In order to distinguish between the samples, the diameter of the artefact rings perpendicular to the field direction was measured. The evaluation of all image files was carried out using image processing software. In order to consider the geometry of the samples, the distance of the artefact boundary was calculated to the sample contour being perpendicular to the field direction. The diameter of the distortion is highest in the solid material samples (see Fig 19, upper left for Grade 5 and Fig. 19, upper right for Grade 4) and decreases in the infiltrated samples with increasing silver content from right to left.

It has been demonstrated that, by combining Ti-6Al-4V, with its excellent mechanical properties, with diamagnetic silver, the magnetic susceptibility can be reduced as well as the artefact distortion, opening a new dimension in the production of special medical devices made from titanium. Although images free from artefacts have not been achieved with the presented combination, the significance of the artefact reduction will need to be studied for specific applications.

Process development of gears manufactured by Laser Metal Deposition

Finally, Marleen Rombouts and Gert Maes (Vito, Belgium), Freddy Varspringer (VCST, Belgium) and Scott Wilson (Oerlikon Metco, Switzerland) turned the attention to the Laser Metal Deposition (LMD) process in considering this process’s use for fast production of prototype automotive gears. The drawback of the currently applied production method for such parts is the long delivery time (minimum 6-8 weeks) of the hard metal tooling required to hob the shape of the teeth. Typical failure modes for gears are tooth root fatigue fracture and pitting wear on the tooth flanks (surface contact fatigue).

Laser metal deposition has been performed at Vito on a CNC machine with rotational axis, 1kW fibre laser, processing optics, powder feeder and coaxial cladding nozzle. Two different parameter sets have been used for production; only set 2 resulted in nearly 100% dense teeth, while the material deposited using set 1 contained a significant amount of porosity (the details of these parameter sets were not, however, divulged in the paper). The individual teeth have been built up by laser metal deposition on a S355 steel shaft with diameter of about 60 mm. The dimensions of the teeth are in the range of 5-10 mm high, 2-10 mm wide and 20 mm along the cylindrical axis.

As feedstock, different powders from Oerlikon Metco, including low alloy steel (designated as ‘soft’) and X42Cr13 steel (designated as ‘hard’), have been processed. Gears combining both materials have also been manufactured (‘hard-soft’). The LMD gears have all been tested in the as-built condition.

Using parameter set 1, lack-of fusion defects could be observed on the cross sections, while parameter set 2 resulted in almost 100% dense deposits (Fig. 20). The deposits at optimum processing conditions were characterised by good bonding, low dilution and were almost defect-free (occasionally a small spherical pore). The roughness of the side surface was significantly higher for the low alloy steel deposit.
The average hardness of the low alloy steel after LMD was around 280 HV, while the X42Cr13 (‘hard’) steel resulted in a deposit with average hardness of 550-600 HV. As a reference, typical hardness values for conventionally manufactured 16MnCr5 gears are 350 HV in the core and 700 HV in the carburised surface layer.

The results for tooth bending fatigue of the LMD gears are presented in Fig. 21. For parameter set 1, a higher bending strength during cyclic loading was obtained for the ‘hard’ steel than for the ‘soft’ low alloy steel. A slight improvement in bending strength was obtained for the gear consisting of the ‘soft’ steel in the core region and the ‘hard’ steel at the edges compared to the fully ‘soft’ steel gear. A significant improvement in bending strength was obtained by altering the processing conditions from parameter set 1 to parameter set 2, as a result of the absence of pores in the latter case. At optimum processing conditions, a fatigue bending strength of 800-850 MPa has been reached with the gear consisting completely of ‘hard’ steel. As a reference, case hardened 16MnCr5 steel has a bending stress of around 860 MPa.

Gears, produced using parameter sets 1 and 2 and the hard steel powder, have been subjected to surface contact fatigue tests (see Fig. 22). All tests were stopped before failure occurred and before the normal ending of the test, in order to enable an intermediate evaluation of the gears. These tests indicate an allowable contact stress of around 1100 MPa. As a reference, an allowable contact of 1500 MPa can be expected from case hardened 16MnCr5 steel. The authors reached the final conclusions that the fatigue strengths, under bending and contact stress, already achieved have satisfied the requirement for prototype gears, but that further analysis of the dynamic mechanical behaviour of gears, built using other powders and with additional heat treatments, is on-going.

Author

David Whittaker
DW Associates
231 Coalway Road, Wolverhampton
WV3 7NG, United Kingdom
Tel: +44 (0)1902 338498
Email: [email protected]

References

[1] H Brodin, Heat Treatment and Mechanical Testing of a Selective Melted Superalloy, as presented at Euro PM2015, Reims, France, October 4-7 2015.

[2] P Vikner, R Giraud, C Mayer and M Sarasola, Towards Improved Quality and Production Capacity for Ni based Powders for Additive Manufacturing, , as presented at Euro PM2015, Reims, France, October 4-7 2015.

[3] G Marchese, S Biamino, M Pavese, D Urgues, M Lombardi, P Fino and G Vallillo, Heat Treatment Optimization of Hastelloy X Superalloy Produced by Direct Metal Laser Sintering, , as presented at Euro PM2015, Reims, France, October 4-7 2015.

[4] R Frykholm, B-G Andersson and R Carlstrom, Progress of precision ink jet printing on a powder bed, , as presented at Euro PM2015, Reims, France, October 4-7 2015.

[5] J Isaza, C Aumund-Kopp, S Wieland, F Petzoldt, M Bauschulte and D Godlinski, New Materials and Applications by 3D-Printing for Innovative Approaches, , as presented at Euro PM2015, Reims, France, October 4-7 2015.

[6] M Rombouts, G Maes, F Varspringer and S Wilson, Process Development and Material Properties of Gears Manufactured by Laser Metal Deposition, as presented at Euro PM2015, Reims, France, October 4-7 2015.

World PM2016

In 2016, the Euro PM conference becomes the PM2016 Powder Metallurgy World Congress and Exhibition, taking place in Hamburg, Germany, October 9-13, 2016. The event will include a number of technical sessions focussed exclusively on metal AM. The PM world congress takes place every two years and rotates between Europe, Asia and North America. To find out more visit www.worldpm2016.com