S' S' Babu

1
Use of APFIM and the importance of thermodynamic
and kinetic modeling techniques for studying
non-equilibrium partitioning non-equilibrium
partitioning in Nickel-base superalloy welds

• S. S. Babu
• Oak Ridge National Laboratory
• Oak Ridge, TN 37831
• Microscopy and Microanalysis for phase
  transformation studies, May 30 to June 1, 1999,
  Fukuoka, Japan.

Research Sponsored by Basic Energy Sciences
Division of the Department of Energy
2
Why weld? Welding plays an important role in
economical reuse and reclamation of used and
failed Ni-base superalloy blades.

• Ongoing work has shown that it is possible to
  weld single-crystal Ni-base superalloys.
• However, non-equilibrium microstructure
  development during solidification and solid-state
  decomposition may be expected.
• Weld microstructure will control the weld
  properties and microstructural stability during
  service.

3
Experimental

• Alloys (2 - 3 mm single crystal thick sheets)
• PWA-1480 Ni - 5.0 Al - 10.0 Cr - 1.5 Ti -
  5.0 Co - 12.0 Ta - 4.0 W (wt.)
• CMSX-4 Ni - 5.5 Al - 6.5 Cr - 1.1 Ti - 7.0
  Co - 7.0 Ta - 6.5 W (wt.)
• (small amounts of Mo and Re)
• CM247DS Ni - 5.5 Al - 8.0 Cr - 0.8 Ti -
  9.0 Co - 3.2 Ta - 9.5 W (wt.)
• Welding procedure
• Electron Beam Welding (EB) (Slow weld cooling)
• accelerating voltage 100-125 kV, beam current
  7.5-10 mA, and welding speed 4.2 x 10-3 m s- 1
  (Preheat 500 C)
• Pulsed Laser Beam(PL) (Rapid weld cooling)
• average power 90-240 W, pulse rate 40 s-1 and
  welding speed 2.1 ??10-3 (low) and 12.7 ?? 10-3
  ms-1 (high)
• Optical, Transmission electron and Atom-Probe
  Field-Ion microscopes were used for
  characterization.

4
TEM revealed ? precipitates inside ? dendrite.

• Size of ? precipitates varied from 0.05 to 0.5
  µm.
• Average volume percentage of ? was 75.
• Eutectic ? ? was observed along the
  inter-dendritic boundary.

5
FIM micrograph showing ?, ?, and ?-? interfaces.
PWA-1480

• No indication of solute segregation at ?-?
  interfaces.
• How do the elements partition between ? and ??

6
APFIM concentration profile from PWA-1480 welds
shows no segregation at ?-? interface.

• Not much difference in the composition of ? and
  ?phases was observed and were similar to the
  base metal values 3.

7
No eutectic constituents at the solidification
cell boundaries were observed in the laser welds.
low-speed laser welds

• Fine L12 - ordered ? phases (75) were
  observed.
• Similar microstructure was observed in high-speed
  laser welds

8
Shape of concentration profiles indicated a
redistribution of alloying elements between ? and
? phases by diffusion.

• Arrows indicate the direction of diffusion flux
  in between g and g phases. However,
  concentration gradients within ? phase are not
  large. No solute segregation was observed at
  interfaces.

9
Comparison of calculated lattice misfit at ?-?
interface in laser welds and in single-crystal
base metal.

• Lattice misfit 100 ? (a? - a?)/ (a?
  a?)/2
• a???a? are the lattice parameter of ? and ?
  phases.
• This was calculated with the measured
  concentration profile.
• Lattice misfit in laser welds is positive
  compared to that of negative value in the
  single-crystal base metal.
• will influence the microstructural evolution
  during post weld heat treatment and service

10
CM247DS samples were solutionized at 1573 K for
5 min and were continuously cooled at different
rates.

• Microstructures were characterized with
  transmission electron microscopy and were
  compared with base metal microstructure.

11
Number density of ? precipitates in CM247DS
increased with an increase in cooling rate during
continuous cooling.
12
How can we use the above result to design welding
processes?The above data need to be used to
extend thermodynamic and kinetic calculations to
rapidly cooling weld metals.
13
Ni-Al-Cr quasi-binary phase diagram suggests a
eutectic reaction during weld solidification.
This is supported by experimental observations.

• Eutectic decomposition was observed in PWA-1480
  EB welds.
• This phase diagram was calculated with ThermoCalc
  software.

14
Diffusion controlled growth of primary g phase
into liquid was calculated for Ni - 20 at. Al -
5 at.Cr alloy.
100 C/s
10 C/s

• During faster weld cooling g phase forms with
  less elemental partitioning into liquid.
  Therefore, eutectic formation can be suppressed.
  This was in agreement with CMSX-4 laser weld
  microstructure development.

15
Solid-state diffusion controlled growth
calculations predicted concentration gradients in
the g phase when cooled at 75 C/s compared to 1
and 10 C/s.
Al
Cr

• The concentration profiles are due to incomplete
  redistribution of alloying elements.
• These calculations were performed by using
  DicTra.
• However, further work is necessary to describe
  the nucleation kinetics of g during continuous
  cooling.

16
Summary and Conclusions

• Weld thermal cycles affect the microstructure
  development significantly.
• Number density of g precipitates increases with
  cooling rate
• Composition of g precipitates different from
  base metal
• Lattice misfit of g precipitates with g
  different from base metal
• It is possible to use the above changes so that
  one can modify the welding or post-weld heat
  treatment procedures.
• However, we need to use the microanalysis results
  to improve the thermodynamic and kinetic
  calculations. This is an important need so that
  one can extend these calculations to rapid
  materials processing.

17
Summary of Results

• In PWA-1480 electron beam welds,
• Eutectic ? phase was observed along
  inter-dendritic boundaries. Coarse ? (gt 100 nm)
  precipitates were found in the ? matrix.
  Elemental partitioning between ? and ? phase was
  similar to that in PWA-1480 base metal
• In CMSX-4 laser welds,
• Negligible eutectic constituents were observed.
  Fine ? (10-50 nm) precipitates were found in
  the ? matrix. APFIM analysis showed large
  concentration gradients within the ? phase
  surrounding the ? precipitate
• Transformation studies in CM247DS
• Number density of ? precipitates increased with
  an increase in cooling rate.
• Thermodynamic and kinetic calculations agreed
  with experimental weld-microstructure development.

18
Extensive concentration variations were observed
inside the ? phase in CMSX-4 welds.

• This observation necessitates further
  investigation.

19
No evidence of solute segregation at ??/ ?
interface in CMSX-4 welds.
CMSX-4

• Other FIM micrographs indicated fine distribution
  of ? and ? phases.