The following article is a component of the April 1999 (vol. 51, no. 4) JOM and is presented as JOM-e. Such articles appear exclusively on the web and do not have print equivalents.

Spray Forming: Overview

The CMSF Process: The Spray Forming of Clean Metal

W.T. Carter, Jr., M.G. Benz, A.K. Basu, R.J. Zabala, B.A. Knudsen, R.M. Forbes Jones, H.E. Lippard, and R.L. Kennedy
JOM-e Logo

A clean-metal spray-forming facility has been under development at the General Electric Research and Development Center for the past four years to develop an inexpensive alternative to powder metallurgy for the production of several nickel-based superalloys used in the rotating components of aircraft engines. The approach combines electroslag refining with spray forming to enable the economical production of oxide-free, homogeneous superalloys suitable for fatigue-life-critical applications. In work at the R&D facility, the system has been used to fabricate a cylindrical workpiece called a preform. After spray forming, the preform is thermomechanically processed and forged. Following inspection, the forging is heat-treated to produce the desired service microstructure and machined into the shape of a disk.


Ceramic inclusions play a significant role in the low-cycle fatigue life of components made from superalloys. A ceramic inclusion is brittle compared to the surrounding metal, and it will crack early in the life of the alloy—possibly as early as the first loading cycle. The fractured inclusion then acts as a crack-nucleation site for the surrounding metal, with an initial size being approximately equal to the inclusion size. Because the crack is short for a large portion of the alloy's service, a reduction in oxide size is required to maintain suitable performance at increasing stress levels. This requirement poses a metals-processing challenge that has been aggressively attacked in both the cast and wrought industry as well as the powder-metallurgy (P/M) industry.

In the cast and wrought industry, the required high level of oxide cleanliness is achieved by using a triple-melt procedure.1 Here, vacuum-induction melting (VIM) is used to achieve the appropriate composition, electroslag refining (ESR) is used to remove the oxides, and vacuum-arc remelting (VAR) is used to achieve the desired microstructure and homogeneity.

Many aerospace disks are made from material processed via the triple-melt approach, but today’s trend in advanced military and commercial aircraft-engine turbine disks is toward highly alloyed, segregation-prone superalloys that require fine, homogeneous microstructures. These nickel-based alloys cannot be manufactured via the triple-melt procedure and are instead manufactured using P/M processing. The processing sequence starts with VIM starting material that is melted, atomized, and solidified in a tower to form metal powder. The powder is sieved to remove large particles, blended, canned, degassed, and then extruded to consolidate the powder into a dense solid billet. The billet is subsequently cut into forging blanks, which are forged, sonically inspected, heat-treated, and machined to form disks. P/M processing results in high-quality, high-performance metallic products, but the cost of P/M processing can be two to four times more than conventional cast and wrought metals.

The only oxide-removal step in P/M processing is the sieve step. The sieve opening size is determined by the design engineer using statistical lifing methods. As the atomized metal is passed through the sieve, all oxide particles larger than the sieve opening fail to pass through the sieve and are rejected. Any large metal particles are also rejected. The amount of metal rejected in the sieve may be a high percentage of the metal atomized, resulting in a significant yield loss and increased cost. Much research is devoted to obtaining a high fraction of fine powder, increasing the yield at the sieve stage. After sieving, extreme care must be taken in each subsequent processing step to avoid recontaminating the powder with foreign material.

The primary cost drivers in P/M are the limited yield from raw input material to finished billet and the cost of extrusion. Extrusion is particularly expensive because large extrusion presses are required to reconsolidate the powder and achieve large billet diameter.

Figure 1a Figure 1b
Figure 1. The clean-metal spray forming process: (a) a schematic of the process; (b) spray forming.
Clean-metal spray forming (CMSF), an inexpensive alternative to P/M for several nickel-based superalloys, is being pursued at the General Electric Corporate Research and Development Center. During spray forming,2,3 a stream of liquid metal is gas-atomized to form a spray of fine liquid-metal droplets, which are accelerated from the atomization zone by fast-flowing atomizing jets. The droplets impinge on a collector where they solidify into a coherent, near-fully-dense preform. Continuous movement of the collector and atomizer, along with careful control of the gas-to-metal flow ratio, allows large preforms to be produced. In this process, the liquid metal is cooled so that it loses most (more than 75 percent) of its heat of fusion to the atomizing gas while in flight. The balance is lost to the preform and impinging gases after deposition. The spray-forming process is capable of producing fine-grained homogeneous microstructures with equiaxed grains and more than 98 percent theoretical density for a wide range of alloys.

Spray forming can compete economically with existing cast and wrought technologies if the resulting geometric shape is near-net, resulting in substantially less material loss during subsequent processing as compared to conventional routes. This is not the approach taken in the work reported here, but it is an area of active research by other researchers.4

Spray forming can also compete with powder processing for billet-making because of the reduced number of processing steps and increased yield. In the past, this possibility has not been exploited as a cost-effective alternative for superalloys to be used in critical fatigue-limited applications because of the lack of a primary oxide-removing process. It is a solution to this problem that led to the CMSF process.


General Electric Corporate Research and Development began investigating spray-formed superalloys for rotating components in 1983. A bottom-pouring induction melting system was combined with a spray-forming facility for the production of test specimens. The material properties of spray-formed superalloys were shown to match their P/M counterparts in all cases except low-cycle fatigue. The source of the problem was identified as particles from the ceramic crucible of the induction melting system, which served as crack initiators. Even when very clean, triple-melted feedstock was used, oxide particles from the bottom-pouring induction system eroded from the crucible liner or nozzle and appeared in the spray-formed billet, thus negating any benefit of previously refined metal.5,6 GE aircraft-engine designers concluded that ceramic-free melting is an absolute requirement for spray-formed materials for rotating components and sought to find an appropriate method. CMSF, pioneered by GE and its partners (Allvac, ALD Vacuum Technologies GmbH, and Osprey Metals, Ltd.), is the result.

CMSF has the potential for producing P/M equivalent alloys at low cost. The process, shown in Figure 1, combines two conventional processes—ESR and spray forming—using an all-copper, inductively coupled, water-cooled guide system referred to as the cold-walled induction guide (Figure 2). The ESR process refines incoming material, eliminating ceramic inclusions upon melting of the input ingot. The cold-walled induction guide transfers the refined liquid metal to the spray-forming system and assures that no ceramic inclusions are reintroduced when the liquid is poured from the crucible, as would otherwise occur with conventional ceramic nozzles.

The CMSF approach replaces the mechanical removal of oxides (sieving) performed in P/M with a more reliable thermochemical reaction in the ESR furnace; hence, the yield loss associated with sieving is reduced. The new process requires no ceramic furnace liners or transfer nozzles and, thus, assures a clean superalloy for critical rotating applications.

Figure 2 Figure 3
Figure 2. The input ingot is lowered into the inductively coupled, water-cooled guide system—the cold-walled induction guide. Figure 3. The CMSF processing sequence.

CMSF can replace P/M in the product manufacturing stream (Figure 3). A four-step process is envisioned:

After isothermal forging, the workpiece is ultrasonically inspected, heat treated, and machined in the shape of a disk. The reduced number of processing steps as compared to the P/M processing route, the higher yield, and the use of conventional forging presses in conversion results in a net reduction in cost of the resulting forging billet.


Several technical benefits can be expected from the new process.

In addition to the technical benefits, several economic benefits can be expected.


Oxide Cleanliness

Oxide cleanliness is the primary incentive for developing the CMSF process. For this study, the electron-beam (EB) button test7 was used for evaluation. Five 38 mm X 38 mm X 162 mm blanks from the top half of the preform were machined into EB electrodes. Cleanliness was evaluated by measuring the various parameters of the oxide raft on the surface of the EB-melted button. Using secondary and backscattered electron imaging techniques, oxide particles greater than or equal to 0.5 mil2 were detected, counted, and measured for surface area. The test utilized a semi-automated scanning electron microscope (SEM) equipped with an energy dispersive x-ray (EDX) system and a thin-window, solid-state light element detector. Surface area measurements were performed using an integrated image analysis system. For each individual button, the total oxide area was calculated and divided by the button weight, resulting in the specific oxide area or oxide cleanliness. The qualitative x-ray composition of the five largest oxides in each button was also determined.

Table I. 718 Electron-Beam Button Cleanliness Data
Material Source Total Number of Oxides Total Oxide Area (mil2) Button Weight (g) Specific Oxide Area (mm2/kg) Major EDX Peaks
VIM Bottom 414.9* 610.2 0.439 Al/O
140.0* 611.1 0.148 Al/Mg/O
278.3* 581.1 0.309 Al/Mg/O
Standard Triple Melt 0 0 537.3 0.000
0 0 522.3 0.000
23 8.6 499.6 0.049 Not Measured
VIM/ESR/VAR 2 0.7 561.3 0.001 Not Measured
0 0 527.3 0.000
0 0 553.0 0.000
High-Rate ESR Top 19 20.9 540.4 0.025 Not Measured
14 12.6 511.3 0.016 Not Measured
55 49.9 501.1 0.064 Not Measured
High-Rate ESR Bottom 62 86.2 509.6 0.109 Not Measured
89 89.9 509.3 0.114 Not Measured
63 78.1 503.4 0.1 Not Measured
CMSF Top 0 0 649.5 0
1 0.02 683.6 ~0
0 0 676.7 0
0 0 660.5 0
0 0 654.3 0
*Oxide agglomerate

Table I compares the oxide cleanliness of as-sprayed 718 with the cleanliness of the initial VIM electrode, standard-melt-rate ESR material, and rapid-melt-rate ESR materials. The rapid-melt-rate ESR data were taken from an ingot that was conventionally ESR melted at approximately 14 kg/min., which is the operating set point for the CMSF process. The cleanliness measurements show that the change in melt rate does not influence cleanliness. The data suggest a marked improvement over VIM in oxide cleanliness as a result of CMSF. We do not suggest that CMSF should consistently produce cleanliness levels better than conventional ESR, which is itself outstanding, but anticipate that long-term data will show levels similar to VIM/ESR.

The EB button test specimens for the CMSF data were taken from a preform that was sprayed using nitrogen gas; a nitride raft formed at the top of all buttons. Metallography was performed on cross sections taken from some of the buttons to determine if the observed nitride raft obscured oxides; no hidden oxides were found. EB buttons taken from conventionally melted, nitrogen-atomized preforms show large numbers of oxides and high specific oxide areas despite the presence of the nitride raft.

As-Sprayed Hot Workability

Table II. Results of Rapid-Strain-Rate Hot-Tensile Test As-Sprayed Alloy 718*
Temperature (°C) Ultimate Tensile Strength (MPa) Elongation (%) Reduction in Area (%)
871 392.4 5.8 3.1
461.4 42.2 48.4
927 305.5 69.3 58.8
271.7 10.8 16.0
982 195.8 22.7 17.0
238.6 38.2 42.8
1,038 155.8 86.4 78.4
153.1 95.3 81.7
1,093 113.1 100.1 82.7
111.7 72.9 74.6
*Test bars were cut from the center location and oriented parallel to the preform longitudinal axis. All tests were performed on the as-sprayed material.
The microstructure of as-sprayed 718 is characterized by an equiaxed ASTM 6 grain size with isolated porosity. Following a limited amount of hot deformation, the material exhibited solution-annealed plus aged tensile properties, hardness, and solution-annealed grain size that are acceptable relative to the requirements outlined in AMS 5662.

Rapid-strain-rate elevated-temperature tensile properties were obtained to evaluate the hot ductility characteristics of the material between 871°C and 1,093°C. Both the strength and ductility exhibit significant variability below 982°C (Table II), possibly related to porosity, but appear to be adequate for hot workability.


The proof-of-concept phase for the CMSF program is complete, and ongoing effort is concentrating on economic issues such as spray-forming yield and conversion yield.


This work was performed under funding from the U.S. Defense Advanced Research Projects Agency agreement F33615-96-2-5262.

1. R.L. Kennedy et al., "Superalloys Made by Conventional Vacuum Melting and a Novel Spray Forming Process," 13th International Vacuum Congress, editor/s (City: Publisher, 1995), pages.
2. A.G. Leatham and A. Lawley, "The Osprey Process: Principles and Applications," Int. J. of Powder Met., 29 (4) (1993), pp. 321–329.
3. P.S. Grant, "Spray Forming," Progress in Materials Science, 29 (1995), pp. 497–545.
4. T. Tom and K. Bowen, "SprayCast-X for Aerospace Applications," Third Int. Conf. on Adv. Materials and Processing, editor/s (City: Publisher, 1998), pp. 1681–1686.
5. H.C. Fiedler, T.F. Sawyer, and R.W. Kopp, "Spray Forming—An Evaluation Using IN718," Proc. of the 1986 Vacuum Metallurgy Conference on Specialty Metals Melting and Processing, editor/s (City: Publisher, 1986), pp. 157–165.
6. H.C. Fiedler et al., "The Spray Forming of Superalloys," J. of Metals, 39 (8) (1987), pp. 28–33.
7. J.M. Moyer et al., "Advances in Triple Melting Superalloys 718, 706 and 625," Superalloys 718, 625, 706 and Various Derivatives, ed. E.A. Loria (Warrendale, PA: TMS, 1994), pp. 39–48.


W.T. Carter, Jr., M.G. Benz, A.K. Basu, R.J. Zabala, and B.A. Knudsen are with General Electric Corporate Research and Development.

R.M. Forbes Jones, H.E. Lippard, and R.L. Kennedy are with Allvac, a division of Allegheny Teledyne Industries.

Copyright held by The Minerals, Metals & Materials Society, 1999

Direct questions about this or any other JOM page to jom@tms.org.

Search TMS Document Center Subscriptions Other Hypertext Articles JOM TMS OnLine