Precipitation Reactions in Age-Hardenable Alloys During Laser Additive Manufacturing
- 3.1k Downloads
We describe and study the thermal profiles experienced by various age-hardenable alloys during laser additive manufacturing (LAM), employing two different manufacturing techniques: selective laser melting and laser metal deposition. Using scanning electron microscopy and atom probe tomography, we reveal at which stages during the manufacturing process desired and undesired precipitation reactions can occur in age-hardenable alloys. Using examples from a maraging steel, a nickel-base superalloy and a scandium-containing aluminium alloy, we demonstrate that precipitation can already occur during the production of the powders used as starting material, during the deposition of material (i.e. during solidification and subsequent cooling), during the intrinsic heat treatment effected by LAM (i.e. in the heat affected zones) and, naturally, during an ageing post-heat treatment. These examples demonstrate the importance of understanding and controlling the thermal profile during the entire additive manufacturing cycle of age-hardenable materials including powder synthesis.
Many classes of alloys owe their high strength to the presence of finely dispersed second phase particles (i.e. phases different from the matrix phase). Since they form by precipitation phase transformations, they are called precipitates, and the materials featuring them, precipitation-strengthened alloys,1 Examples are most Al alloys,2, 3, 4 many Ni-based alloys5, 6, 7 and some steels.8, 9, 10, 11 During conventional processing, these materials undergo two subsequent heat treatments. First, in a homogenization treatment in the single-phase region of the phase diagram, i.e. at relatively high temperature, all elements are brought into solid solution. This is followed by a rapid quenching, designed to limit or fully suppress any precipitation during cooling. Precipitates occurring after quenching from the solutionized state are called primary precipitates. Subsequently, the material is annealed at a lower temperature, where the remaining solutes that are now in a supersaturated state are allowed to precipitate. In this age-hardening step, the desired fine dispersion of particles forms. These particles are called secondary precipitates and are typically only up to a few nanometres in size.
During atomization, either if the quenching rate is not high enough to suppress precipitation, or when the time spent in the liquid state is not sufficient to completely dissolve pre-existing, coarse precipitates.
During the actual LAM processing, i.e. during cooling from the liquid state after deposition. Again, this might occur if the cooling rate is not fast enough.
During the intrinsic heat treatment, i.e. during temperature peaks.
During the regular ageing heat treatment applied to the final part.
Of the above list, in traditional processing only possibility number 4 would be considered as desired option for precipitation strengthening, as the particles that result from too slow cooling (in steps 1 and 2) are either too coarse, or not dispersed homogeneously throughout the microstructure (e.g., concentrated at grain boundaries), or both. Secondary precipitation during the intrinsic heat treatment is an interesting possibility for desired precipitation that would allow shortening or even completely avoiding subsequent ageing treatment. Primary precipitates forming during solidification may be beneficial for grain refinement.
The control of precipitation during additive manufacturing is important, because solution heat treatment after LAM, which might be used to re-dissolve unwanted precipitates is not desirable or in some cases even impossible. Solutionizing increases the cost and complexity of the entire manufacturing process and subsequent quenching might induce warpage of the treated part, spoiling the (near-) net-shape nature of the LAM process. In some alloys such as the supersaturated Al-Sc-alloy discussed below, solutionizing cannot be used at all to re-dissolve solutes.
In this paper, examples of all four types of precipitation occurring in LAM-produced alloys are given, after a brief description of the common experimental techniques. The alloys under investigation are a maraging steel, an Al-Sc alloy and a nickel-based superalloy.
Laser Additive Manufacturing
LAM processes and corresponding main process parameters used for making the materials in this study
Volume energy density, E V (J mm−3)
Scan speed, v S, (mm s−1)
Layer height, D L, (µm)
Specimens were deposited using SLM and LMD in the shape of cubes with side lengths of 10–20 mm. They were cut along the build direction (cross-section view) and metallographically prepared. Most of the analyses presented in this work were obtained by using SEMs (scanning electron microscopes; JEOL 6500F and Zeiss Merlin) equipped with EDS (energy dispersive spectroscopy) detectors. For the analysis of fine precipitates, atom-probe tomography (APT) was employed. For this method, samples are lifted from polished cross-sections by a FIB (focused ion beam)-liftout technique (using an FEI Helios NanoLab 600i dual-beam microscope) and also sharpened by FIB milling. The samples are measured in a Cameca LEAP 3000 HR X local-electrode atom probe, using laser pulsing at a repetition rate of 250 kHz, a pulse energy of 0.4 nJ and a base temperature of 60 K (with the exception of the Al-Sc-alloy, for which voltage pulsing at a pulse fraction of 15% is applied at a base temperature of 40 K).
Precipitates in the Raw Powder: A Supersaturated Al-Sc-Alloy
Al–Sc alloys have been shown to possess an outstanding combination of strength, ductility and corrosion resistance.14 A new class of Al-Sc-(Mg-Zr)-alloys called Scalmalloy® has been developed for use in additive manufacturing.3 In this alloy, the Sc content is well above the maximum equilibrium solubility of ~0.3wt.%. Therefore, to obtain a homogeneous supersaturation of Sc in the Al matrix, rapid solidification (and further, rapid cooling) needs to be employed. It has been shown that this way, the entire Sc content can be used in the strengthening precipitation reaction yielding Al3(Sc,Zr) particles.3
We investigated specimens produced by SLM from two different batches of Scalmalloy® powder. The first batch had been atomized by a standard gas atomization technique (ECKA Granules, Fürth, Germany) while the second batch was manufactured using the electrode induction-melting gas atomization (EIGA) technique (TLS Technik, Bitterfeld, Germany). In this crucible-free technique, the bottom part of a rotating bar is melted by induction heating and the resulting flow of liquid metal becomes atomized in a nozzle system using inert gas.
Using the measured volume fraction of particles (~0.3%) and their measured composition, the amount of Sc removed from the matrix by being bound inside these large particles can be estimated: 0.09 wt.%. This small amount of scandium accounts for only a moderate drop in the strength of the material. Fractographic analysis of tensile test specimens made from this material reveal that the large precipitates also do not act as crack initiation sites and hence have only a small deteriorating impact on mechanical strength and ductility.
Precipitation during material deposition: Al-Sc alloy and Ni-base superalloy
Microsegregation occurs during solidification of most alloys (containing elements with partitioning coefficient not equal to one), provided the cooling rates are not high enough to completely trap solute into the growing crystal.15 We observed microsegregation both in a maraging steel (18Ni-300, see next example) and in a Ni-base superalloy (Inconel 738LC®). Evidently, even the relatively fast cooling rates prevailing in the SLM process (estimated by simulations to be 104–106Ks−1, depending on process parameters)16 are not sufficient for (complete) solute trapping. As the remaining liquid phase becomes strongly enriched in solutes during solidification, the formation of intermetallic phases may become energetically favorable.
Due to the presence of the precipitates both inside the grains and on the grain boundaries, the exact mechanism of precipitation is not clear. They could emerge in the solid state during cooling after solidification or they could be the result of re-heating during the deposition of an adjacent track/layer (see next example) and afterwards stimulate nucleation during partial re-melting of the layer as the next one is deposited above.
Precipitation During Intrinsic Heat Treatment: Maraging Steel
A characteristic feature of LAM processes, in particular the SLM process, are the high cooling rates experienced by the processed material owing to the small size of the melt pool and effective heat conduction through the substrate and already deposited material. It can therefore be expected that no precipitation in the bulk material during cooling after solidification takes place. This can indeed be shown by analyzing the material using statistical analysis of APT datasets.
To prove that this clustering (early stage of precipitation) does not occur during cooling after solidification but rather during the intrinsic heat treatment, i.e. re-heating of the material during the deposition of additional layers, we also investigated samples taken from the very top of the specimen. In the uppermost layer, and in particular in the last track of this layer, no intrinsic heat treatment takes place. Indeed, the RDF corresponding to this situation is again nearly equal to one for all pair distances (cf. the black, dotted line in Fig. 5a).
Precipitation During Ageing Heat Treatment: Maraging Steel
Finally, further precipitation reactions occur during post-manufacturing heat treatments. Even though some of the initial solute content may not be available for precipitation any more, as explained in the previous examples, the remainder of solute atoms reacts in the same way as in conventionally produced materials. A potential difference exists in the different defect density of LAM- and conventionally produced material. Conventionally produced material is solution heat treated before precipitation and hence in recrystallized state, which is relatively poor in defects (e.g., dislocations, dislocation cells, and low-angle grain boundaries). Residual stresses generated during the LAM process can be released by plastic deformation, generating a higher density of dislocations. This presence of dislocations could influence the nucleation rate and the spatial distribution of nucleation sites. We observed precipitation in a LMD-produced and post-heat treated (8 h at 480°C, i.e. until peak hardness is reached) maraging steel and compared the morphology, number density and chemical composition of the emerging particles with the ones observed in a conventionally produced and identically heat-treated specimen (see Fig. 5b). Despite the different processing routes, no significant differences between the samples could be observed (note that in other regards, e.g., in the presence of retained and reversed austenite, grain size and morphology, and texture, significant differences between the samples do exist).
Precipitation of Al3(Sc,Zr) during the production of starting material powder in a supersaturated Al-Sc alloy.
Precipitation of carbide, boride and a Zr-rich intermetallic phase during the LAM process, i.e. during solidification in a Ni-based superalloy. Similarly, precipitation of Al3(Sc,Zr), Al6Mn and Mg17Al12 from a supersaturated Al-Sc alloy during cooling after deposition was observed.
Early stages of Ni3Ti precipitation (Ti–Ti clustering) during the intrinsic heat treatment occurring in LMD of a maraging steel.
Precipitation of Ni3Ti, Ni3Mo and Fe7Mo6 during the ageing heat treatment after LAM in a maraging steel.
These findings reveal that LAM can be used to produce supersaturated alloys that show precipitation during age hardening annealing, as intended. In addition, LAM even provides the potential to dispense with the need of a post-heat treatment by exploiting its intrinsic heat treatment. However, care has to be taken in the production of the powder and in the choice of process parameters to make sure that no undesired precipitation reactions occur in the process chain. Hence, a detailed knowledge of the thermal profile experienced by the material before and during LAM, as obtained, e.g., by process simulations, is necessary.
Open access funding provided by Max Planck Society (Max Planck Institute for Iron Research). The authors would like to thank F. Palm, Airbus Group Innovations, for providing the SLM-produced Scalmalloy® specimens, J. van Humbeeck, KU Leuven, for providing the SLM-produced maraging steel specimens and S. Kleber, Böhler Uddeholm GmbH for providing the conventionally produced maraging steel (Böhler V720). The support by M. Kuzmina, P. Kürnsteiner and S. Ocylok with experiments is gratefully acknowledged.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.