The Advanced Manufacturing Office (AMO) in the U.S. Department of Energy is sponsoring research to formulate high-quality metal powders optimized for metal additive manufacturing (AM), including hard-to-build alloys such as high-temperature refractory alloys. AM refractory multi-principal-element alloys (MPEAs), comprising elements with melting points of 1850-2200°C, offer the potential for step-change improvements in extreme high-temperature resistance needed for the newest high-efficiency gas turbines and other industrial applications.

If such AM refractory MPEAs can be made stronger and more elastic, it would usher in an era of U.S. material manufacturing leadership in commercial refractory alloys with superior resistance to extreme heat and wear.


A Tool for Optimizing AM Powder Formulations

A project team led by Ames Laboratory is now developing a computational Tool that will enable U.S. manufacturers to create and customize their own formulations for new, high-performance MPEAs. This three-year project on metal powders for additive manufacturing (P4AM – see sidebar) is accelerating progress with physics-based simulations to estimate and optimize the properties of AM powders in a huge swath of formulations.

Samples produced from the most promising AM formulations then undergo rigorous testing to see how their mechanical and chemical properties compare to those of materials produced through traditional bulk casting and wrought processing. The data from these experiments is fed into powerful high-performance computers that use artificial intelligence and machine learning to optimize the process. The team will then additively “print” those powders into actual components for energy, automobile and aerospace applications.

Upon completion, the Tool will reduce the number of process steps and avoid the inefficiencies of a trial-and-error approach to alloy development. The Tool is expected to cut the time required for an industrial user to develop a new alloy for an application by two-thirds – from more than a year to just a few months. Most importantly, the end products are expected to perform far better than their existing counterparts and provide greater energy efficiency throughout their life cycle.

Now midway through the project, the research team has identified key principles to guide MPEA formulations:

  • Use high-quality, spherical metal-powder particles.
  • Choose a narrow selection of elements and their proportions based on physics-based modeling.
  • Recognize value in unexpected results to uncover promising new directions for research


Start with Highest-Quality Spherical Powder Particles

While computational models typically represent AM powders as perfect spheres, most commercially produced powders fall short of this ideal. As shown in Figure 1, commercial powders can include numerous protrusions (“satellites”) as well as voids or pores. Such irregularities can disrupt particle alignment in the alloy structure, introducing flaws or otherwise adversely affecting the desired alloy properties.


Fig. 1.  An X-ray tomography image of a commercial metal-powder

particle with satellites and pores

(courtesy Argonne National Laboratory)


Ames recently developed a high-pressure gas atomization process to produce high-quality powders for AM that are consistently sized and near perfectly spherical. Numerous images from high-energy X-ray 3D tomography at Argonne National Laboratory’s Advanced Photon Facility showed that the best sample powders from Ames were orders of magnitude less porous than commercial powders (0.01% porosity for the Ames particles compared to 2-4% porosity for commercial benchmark powders).

Because modelers can safely “assume a spherical particle” for powders produced via Ames’ gas atomization, their physics-based models do a better job of predicting alloy properties and designing the AM processes that carefully position the particles into specific crystalline microstructures. More importantly, higher-quality powders translate into additively printed parts that possess fewer defects and greater reliability.


Physics-Based Modeling for MPEAs

The design space for refractory alloys no longer has just two or three dimensions but five or six (see Multi Element Alloy sidebar). The number of refractory MPEA combinations and the computational power needed to model them is exponentially larger than that for traditional alloys. To find a good starting point in this five- or six-dimensional design space, the project team’s initial model used the long-standing “Hume-Rothery” rules of composition.

Theorists at Ames began with refractory molybdenum (Mo) and then identified combinations with tungsten (W), tantalum (Ta), titanium (Ti) and zirconium (Zr) as a promising design space for a refractory MPEA. The team used molecular dynamic simulations to predict the alloys most likely to offer increased stability with greatly enhanced elasticity in comparison to commercial alloys. Researchers narrowed their design for high-strength refractory alloys using density functional theory (DFT) and first-principals theory-based simulations to predict phase stability and mechanical behavior.

Figure 2 shows some of these simulation results as a 2-D cut through the 5-D design space. While near-equiatomic, high-entropy alloys containing Mo are known for high-temperature strength and corrosion resistance, this particular slice of the 5-D chart indicates that the equiatomic alloy (designated with a “1” in Figure 2) would not be the optimum choice because of its high formation energy (about 15 mRy). Based on these predictions, samples of the off-equiatomic MPEA refractory combinations were manufactured that validated their enhanced properties, including high-temperature strength, thermal stability and ductility – along with strong resistance to corrosion, fracture and creep.


Fig. 2.  A two-dimensional (2-D) slice through the five-dimensional design space of an MPEA alloy

consisting of five component elements: Mo, W, Ta, Ti and Zr. The colors are based on DFT results.

The larger plot shows the energy of formation. The inset plot illustrates that the maximum

strength occurs in the same off-equiatomic region of the design space

where the alloy might be most stable.¹


Unexpected Results

Increased use of physics-based models, machine learning and ever-more powerful computers might suggest that the days of serendipitous breakthroughs are behind us, but researchers can still make valuable discoveries by finding opportunities in unexpected circumstances.

Sandia theorists had suggested that the team try a six-element mixture that combines Mo, niobium (Nb) and Ta (based on earlier refractory successes) with Zr, Ti and aluminum (Al). Mo, Nb and Ta are established refractory elements with melting points above 2200°C and high hardness at room temperature. Zr and Ti are refractory elements with somewhat lower melting points (1855°C and 1668°C, respectively). Aluminum, while not a refractory element, is known to reinforce higher-melting-point intermetallic compounds.

To determine the compositional starting place, the Sandia experimentalists requested that Ames use physics-based modeling to identify the six-element compositions where multiple alloy metal phases coexist. The team ordered the blends suggested by Ames, but when the powder arrived, the blend constituents were based on weight percent rather than atomic percent. This mixture represented a major shift in composition due to mixed signals between theorists and engineers (lesson learned).

Although the elements were in entirely different proportions than planned, Sandia researchers decided to move forward with the printing. After annealing, the result was a wonderfully complex (and likely high strength) microstructure (Fig. 3). In preliminary mechanical analyses, the new Sandia MPEA, named “Kustalloy,” shows extreme toughness and possibly the highest strength ever measured in an alloy. Even as microstructural and physical property analyses are underway to confirm these findings, the team’s theorists are expanding their model to explore this novel region of composition space for additional alloys, which might exhibit even better hardness, phase mixture and other properties.


Fig. 3.  This composite image shows the complex microstructure (transmission electron

microscopy at lower left) and chemical mapping (color key at upper right) of six elements in an AM,

multiple-principal element alloy (MPEA). This new “Kustalloy” exhibits potentially record-breaking

hardness at high temperatures

(courtesy Sandia National Laboratories).


Theorists and experimenters agree that the discovery might never have happened using a purely computational approach based on long-standing assumptions like the H-R rules. While those rules have held up well for nearly a century, they may not accommodate more complex MPEAs like Kustalloy.²


P4AM Project Team

Ames Laboratory leads the five-lab team and coordinates all project activities. Each lab brings to the team a range of specialized expertise, sophisticated equipment and capabilities not available to many industrial companies today.

  • Ames Laboratory: Alloy design, selection and characterization; sample and powder production
  • Oak Ridge National Laboratory (ORNL): Ni-based superalloy melting scans and additive manufacturing (AM) builds, verification of AM processing method and parameter selection
  • Sandia National Laboratory (SNL): Alloy design input for multiple-principle element alloys (MPEA), sample melting, AM builds and characterization
  • Kansas City National Security Campus (KCNSC): Aluminum alloy melting scans and AM builds and characterization
  • Argonne National Laboratory (ANL): Characterization of AM powders (internal porosity and other features) using ANL’s world-class Advanced Photon Source


Industrial Partners

Private partners will use their expertise to identify target alloy properties, produce alloy samples via AM, characterize performance and analyze production economics.


Next Steps on Tool Development

Ames’ AM alloy selection Tool will now benefit from data and machine learning from Sandia’s new Kustalloy prints as well as information drawn from less successful alloys tested earlier. Data still to be collected and fed into the computers for a smarter Tool includes: verified alloy strengths, elasticity and temperature resistance, production tests and economic feasibility.

In addition to the Tool itself, improved powder production methods for AM feedstock materials will be customized in coordination with each of the participating powder producers. Industry partners will use these powders to additively manufacture the alloys and characterize their properties to verify and validate the prescribed processes and product quality.

High-temperature materials often determine the overall efficiency of the larger systems in which they serve a function.³ Aerospace, energy, defense, transportation and other industries that require materials with high thermal resistance will benefit significantly from novel alloys that offer superior performance at lower cost. The integrated AM alloy design Tool – from powder to component – will help the private sector move closer to achieving the promise of AM to deliver energy and cost benefits for multiple applications.


Multi-Element Alloys

Alloys comprising three or more elements are called “multi-element” alloys. The figure is a ternary (3-axis) diagram of alloy composition with the axes showing percentage composition of the element (or element pair as in Figure 2) at each apex. The white circles are alloy combinations with >95% of a “principal element” that are the traditional multi-element alloys (MEAs). Computational and experimental metallurgy have advanced to the point where researchers can both model and fabricate alloys comprising three to six elements in large (up to 35%) proportions.

The shift in the 1980s and 1990s from the corners of this ternary diagram to the center was inspired by a physics-based observation that the high configurational entropy of alloys with five or more elements in near-equiatomic combinations might stabilize them in a solid-solution phase. This suggested that such alloys would be able to withstand extremely high temperatures and pressures.

One of the first high-entropy alloys (HEAs) – the exactly equiatomic FeCrMnNiCo, also known as the “Cantor” alloy – has been the subject of considerable research. While the term HEA is sometimes used to describe the full (orange) space of multi-principal element alloys in which no single element is dominant, it more accurately refers to the near-equiatomic alloys near the center of the chart. Here we describe the broader set of alloys with the term “multi-principal element alloys” (MPEAs). Each principal element in an MPEA makes up at least 5% of the atomic weight of the alloy. MPEAs have unique microstructures that promise extraordinary strength, toughness and thermal stability.

Axes are atomic weight percent for elements a, c and e. The blue dashed line running through the HEA is a cut from 100% “e” to 50% “a” and 50% “c.”

For more information: Contact Tina Kaarsberg, Ph.D., technology manager, Advanced Manufacturing Office at the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy, Washington, D.C.; tel: 202-586-5112; e-mail:; AMO web:




  1. P. Singh et al., “Design of high-strength refractory complex solid-solution alloys,” NPJ Computational Materials (2018) 4:16 ; doi:10.1038/s41524-018-0072-0
  2. Zongrui Pei, et al., “Machine-learning informed prediction of high-entropy solid solution formation: Beyond the Hume-Rothery rules,” NPJ| Computational Materials 6, 07 May 2020.
  3. Dr. Wolfram Knabl, Head of Development Core Technologies and Materials, Webinar: Refractory metals and their applications, May 2020. Accessed Oct. 2, 2020, at 50