Metallurgical Examination of Powder

Metallurgy Uranium Alloy Welds

G . M . A L E X A N D E R - M O R R I S O N , A. G. DOBBINS,
R. K. HOLBERT, and M. W. DOUGHTY

Inertia welding provided a successful technique for joining full density, powder metallurgy
uranium-6 wt pct niobium alloy. Initial joining attempts concentrated on the electron
beam method, but this method failed to produce a sound weld. The electron beam welds
and the inertia welds were evaluated by radiography and metallography. Electron beam
welds were attempted on powder metallurgy plates which contained various levels of
oxygen and nitrogen. All welds were porous. Sixteen inertia welds were made and all
welds were radiographically sound. The tensile properties of the joints were found.to be
equivalent to the p/m base metal properties.

INTRODUCTION EXPERIMENTAL TECHNIQUE

The application of powder metallurgy techniques to the Materials
uranium-6 wt pct niobium alloy (U-6Nb) is being pursued
for two reasons. First, the arc-melt, wrought U-6Nb alloy is The chemistries and mechanical properties of the wrought
chemically inhomogeneous with respect to niobium, where- and p/m base metals are listed in Tables I and II. The metals
as the powder metallurgy U-6Nb material has no macro- were sampled in the fully heat treated condition: a vacuum
segregation and minimal microsegregation. Second, the heat treatment at 800 ~ for one hour and water quench. The
wrought processing techniques for U-6Nb do not fabricate resulting structures are shown in Figure 1. The wrought
near-net-shape parts, while the powder metallurgy process U-6Nb structure is uniform and characterized as duplex. The
will aim all processes at producing near-net-shape parts. inclusions, primarily uranium and niobium carbides, are
Development efforts are underway to produce a variety of located within the grains. The powder metallurgy material is
near-net-shape U-6Nb powder metallurgy parts. also characterized by a duplex structure, but the grain
The near-net-shape part applications will require that boundaries are jagged and irregular. The p/m material con-
the powder metallurgy (p/m) material be joinable. This tains a high density of fine inclusions (oxides, nitrides,
project was initiated to determine a joining technique and carbides) which are uniformly dispersed throughout
for two material combinations: p/m-to-p/m and p/m-to- the material.
wrought. To be considered joinable (more specifically weld-
able), the technique must produce a radiographically sound Welding Methods
joint and the joint must be equivalent in strength and Past work with the U-6Nb alloy ~ and with powder metal-
ductility to the base metal. lurgy superalloys 2 has demonstrated that two welding
methods would be applicable to p/m U-6Nb alloy: electron
G.M. ALEXANDER-MORRISON, A, G. DOBBINS, R.K. beam and inertia welding.
HOLBERT, and M.W. DOUGHTY are with Martin Marietta Electron Beam Welding. Electron beam (EB) welding is
Energy Systems, Oat: Ridge Y-12 Plant, Oak Ridge, TN 37830.
This paper was presented at the ASM Materials Week Sym- the standard joining process for arc-melt, wrought U-6Nb
posium on Metallography, 14-15 Oct. 1985, Toronto, ON, Canada. alloy parts. Variations on the standard EB procedure were

70 VOL.8, NO. 1, JUNE 1986 9 AMERICANSOCIETYFOR METALS J. MATERIALSFOR ENERGYSYSTEMS

 Table I. Chemistry of Base Metals (Weight Percent) Table III. Electron Beam Weld Parameters

Composition Wrought P/M, EB* P/M, Inertia Voltage 120 kV

Amperage 8 or 4 milliamps for a full or
Niobium 5.89 5.85 5.87 partial weld penetration
Carbon 0.0065 0.0080 0.0072 Travel 12.7 mm per second (30 ipm)
Oxygen 0.0030 0.110 to 0.022 0.0095 Beam sharp and with a linear path
Nitrogen 0.0040 0.060 to 0.014 0.0160
Hydrogen 0.0001 0.0002 0.0001
Uranium balance balance balance
prepared. The four plates were butt-welded together and
*EB--electron beam welding material
two single pass welds made on the composite sample. The
welds in each material were examined by radiography and
metallography.
Inertia Welding. For this study, three material combina-
Table II. Mechanical Properties of O-6Nb Base Metals tions, using solid 15.9 m m (0.625 inch) diameter cylinders,
were attempted. The c o m b i n a t i o n s were wrought-to-
Powder wrought, p/m-to-wrought, and p/m-to-p/m. Two prelim-
Property Wrought Metallurgy
inary welds were made to determine the most favorable
Tensile strength, MPa (ksi) 827 (120) 844 (122)
welding parameters and to check the amount of upset gener-
Yield strength, MPa (ksi) 141 (21) 159 (23)
Elongation percent 31 28 ated during welding. Using the selected inertia parameters
Reduction-in-area, percent 41 36 (listed in Table IV), sixteen welds were made. The process-
Modulus, • 10 3 ksi 10 10.1 ing for each weld is listed in Table V.

tested on the p/m material; a plate, 2.79 mm (0.110 inch) Table IV. Inertia Weld Parameters
thick, of p/m U-6Nb was tested with full and partial pene-
tration welds to determine the most favorable welding Flywheel speed, initial 2500 rpm
parameters. Two parameter combinations, summarized in Flywheel speed at final load 100 rpm
Initial load 2169 kg (985 lbf)
Table III, were selected for further testing. Because
Final load 4338 kg (1970 lbf)
the original p/m plate contained 200 wppm oxygen and
130 wppm nitrogen, a composition variable was included in
the tests. Three p/m pressings, each with a different oxygen
and nitrogen level, were machined into plates. A fourth
Table V. Inertia Weld Combinations
plate, of wrought U-6Nb with a low impurity level, was also
Material Post-Weld Tensile
Weld Combination* Heat Treatment Specimen
1 W/W no no
2 W/W yes yes
3 W/W yes yes
4 P/W yes yes
5 P/W yes yes
6 P/W yes yes
7 P/W no no
8 P/P yes yes
9 P/W no yes
10 P/P no yes
11 P/P no yes
12 P/W no yes
13 P/P no no
14 P/P no yes
15 W/W no yes
16 W/W no yes

Fig. 1--Structure of U-6Nb base metals: (a) wrought, (b) powder metal- *W--wrought, P--powder metallurgy
lurgy (original magnification 100•

J. MATERIALSFOR ENERGY SYSTEMS VOL. 8, NO. 1, JUNE 1986 71

 All p/m materials were welded in the fully heat treated
condition. Samples 1, 7, and 13 were mounted for metal-
lographic examination after welding. Samples 2 through 6
and weld 8 were heat treated after welding. All samples
except for 1, 7, and 13 were tensile tested and then
mounted for metallographic examination. Immediately after
welding, all welds were radiographed.

Metallographic Preparation
Specimens were prepared for polishing by rough grinding on
SiC 240, 320, 400, and 600 grit papers with water. Be-
tween grit sizes, the samples were ultrasonically cleaned to
remove all large abrasive particles. Polishing was accom-
plished on vibratory polishers using a 0.3/z aluminum ox-
ide powder in water on a chemotextile* polishing cloth to

*Buehler LTD-Texmet

remove the paper scratches. Final polishing was on a syn-

thetic chemically resistant cloth** and used 0.05 p. alumi-

**Buehler LTD-Chemomet

num oxide powder in a 5 pct chromium trioxide, demin-

eralized water solution. An electrolytic etchant (10 pct ox-
alic acid) was used with a stainless steel cathode, specimen
anode, and 3 to 5 v open circuit for 5 to 10 seconds.

RESULTS AND DISCUSSION

Electron Beam
Figure 2 displays the full and partial penetration welds for
each of the powder metallurgy plates and for the wrought
U-6Nb plate. All of the p/m welds are porous, while the
wrought welds displayed no porosity. The best p/m weld is
the partial penetration weld in p/m plate 3 (Figure 2(c)).
Fig. 2--Electron beam welds (originally 8x): (a) p/m plate 1, oxygen
The radiograph of the composite plate confirmed that this 1048 ppm, nitrogen 591 ppm; (b) p/m plate 2, oxygen 736 ppm, nitro-
weld had the least number of pores per linear inch of bead: gen 168 ppm; (c) p/m plate 3, oxygen 222 ppm, nitrogen 136 ppm;
(d) wrought plate, oxygen 34 ppm, nitrogen 45 ppm. (p/m Plate 1 sample
89 pores with a diameter less than 1.3 mm. This plate also was reversed in the mount.)
contained the lowest amount of oxygen and nitrogen:
222 ppm oxygen and 136 ppm nitrogen.
photograph of Weld 1: a wrought-to-wrought weld sample.
Inertia
The joint is characterized by three regions: the fusion zone
All welds were radiographically sound. The results of the (FZ); the twisted zone (TZ); and the heat affected zone
tensile tests are listed in Table VI. The results were catego- (HAZ). Similar regions have been noted in welds of p/m
rized by material combination and heat treatment. All superalloys) After sample preparation, the FZ of Weld 1
strengths were equivalent to the p/m base metal strengths. (Figure 4) etched very dark and could not be resolved at low
The elongations and reduction-in-area values were lower magnifications. With a very light etch and at high mag-
than desired (a side effect of the increased strength). For nifications (1000 to 2000 • the joint of Weld 1 is found to
the nonpost-weld heat treated welds, the strengths and consist of very fine grains (Figure 5). The inclusions are
reduction-in-area values were the highest, but their elon- aligned parallel to the fusion line. Adjacent to the fusion
gation values were the lowest. zone is a region which appears twisted and distorted (refer
Prior to examining the p/m weld structures, a review of to Figure 4). This distortion is a characteristic of inertia
the wrought-to-wrought weld was required. Figure 3 is a welds. The material farther away from the FZ and neigh-
72 VOL. 8, NO. 1, JUNE 1986 J. MATERIALS FOR ENERGY SYSTEMS
 Table VI. Tensile Results for InerUa Welds

Property**
Material* UTS YS El RA N***
W/W HT 800 (116) 138 (20) 27 31 2
W/W nHT 869 (126) 186 (27) 21 33 2
P/W HT 820 (119) 145 (21) 26 34 3
P/W nHT 834 (121) 186 (27) 26 40 2
P/P HT 841 (122) 138 (20) 24 21 1
P/P nHT 869 (126) 186 (27) 21 29 3

*Materials: W--wrought; P--powder metallurgy; HT--given post-

weld heat treatment; nHT--not given post-weld heat treatment.
**Property abbreviations: UTS - - ultimate tensile strength MPa (ksi);
Y S - yield strength, MPa (ksi); E l - elongation in 2.54 cm gage, percent;
R A - - reduction-in-area, percent.
***N--number of 0.63 cm diameter tensile specimens.
Fig. 3--Weld 1: wrought-to-wrought.

boring the twisted zone is the heat affected zone. The struc- Weld 7 is a p/m-to-wrought weld sample. A series of
ture is similar to the base metal after a short solution anneal. photographs, both bright field and polarized light, helps to
Figure 6 displays the structure of Weld 2: the tensile com- demonstrate the different inertia welding responses of p/m
parison of wrought-to-wrought Weld 1. Both the bright field and wrought materials. Weld 7 is shown in Figure 7. The
and polarized light views reveal the prior fusion line. The fusion zone also consists of very fine grains (Figure 8). At
polarized light view, Figure 6(b), displays the desirable a lower magnification, the deformation pattern of the joint
grain size change for a heat treated inertia weld. The fusion is similar to the wrought-to-wrought welds (Figure 9): a
line is barely noticeable except for a slight refinement in the twisted zone followed by the heat affected zone. The depth
grain size. of the twisted zone is equal for both the wrought and the

Fig. 4--Structure of wrought-to-wrought Weld 1 (bright field, original magnification 100•

J. MATERIALS FOR ENERGY SYSTEMS VOL. 8, NO. 1, JUNE 1986 73

Fig. 5--Weld 1 joint consists of very fine grains and inclusions parallel to FZ. (a) Bright field, 1000x; (b) bright field, 2000•

Fig. 6--Structure of wrought-to-wroughtWeld #2 after heat treatment and tensile testing: (a) bright field; (b) polarized light (original magnification50x).

p/m. Figures 10 through 13 display the differences between

the wrought and p/m materials' respective twisted zones and
heat affected zones. The wrought material contains much
larger grains than the p/m. Figures 10 and 11 were taken at
a 45 deg rotation to the fusion line. At this angle the polar-
ization is maximum and the twins are visible in the large
grains. The HAZ for each material resembles the respective
base metal.
The tensile comparison to Weld 7 is the p/m-to-wrought
sample, Weld 6. After welding, the joint was heat treated,
tested, and mounted for metallography (Figure 14). The
most noticeable difference between the two materials is their
size and distribution of pits. The pits are the remnants of
Fig. 7--Weld 7: p/m-to-wrought (p/m material is on the right). inclusions in the materials. During tensile testing, the car-

74 VOL.8, NO. 1, JUNE 1986 J. MATERIALS FOR ENERGY SYSTEMS

 Fig. 8 - - V i e w of Weld 7 fusion zone (original magnification 1000•

bide and nitride inclusions fracture. These inclusions are

then removed during the polishing and leave behind the pits.
The wrought material in Figure 14(a) contains fewer but
larger pits than the p/m material. Notice that the pits provide
a large notch when the inclusion is located on the sur-
face. Concentrating on the fusion line, Figure 14(a) shows

Fig. 10--Twisted zone in the wrought side of Weld 7: (a) bright field;
Fig. 9 - - W e l d 7: fusion and twisted zones (p/m material is on the right). (b) polarized light (original magnification 250x).

J. MATERIALS FOR ENERGY SYSTEMS VOL. 8, NO. 1, JUNE 1986 75

Fig. 11--Twisted zone in the p/m side of Weld 7: (a) bright field; (b) polarized light (original magnification 250x).

Fig. 12--Heat affected zone adjacent to the twisted zone in the wrought side of Weld 7: (a) bright field; (b) polarized light (original magnification 250x).

76 VOL. 8, NO. 1, JUNE 1986 J. MATERIALS FOR ENERGY SYSTEMS

Fig. 13--Heat affected zone adjacent to the twisted zone in the p/m side of Weld 7: (a) bright field; (b) polarized light (original magnification 250•

several vertical wavy lines where the fusion occurred. In and 7, etched dark and was unresolved at low magni-
Figure 14(b), the joint is indistinguishable from the base fications. The upset on this weld is uniform and is consid-
metal. This tensile specimen failure initiated on the wrought ered an acceptable amount of flash. Weld 11 is another
side of the joint and ran at a 45 deg angle across the weld p/m-to-p/m weld with no post-weld heat treatment; how-
and into the p/m material. This ductile fracture and the high ever, this weld was mechanically tested. Figure 16 displays
elongation (26 pct) categorized this weld as successful. the joint and the fracture profile. As expected, the joint is
Weld 13 (Figure 15) is a p/m-to-p/m weld which did not visible as a dark line. The fracture occurred in the p/m base
receive a post-weld heat treatment. The joint, like Welds 1 metal at a significant distance from the joint. The fracture

Fig. 14--Structure of Weld 6 joint after heat treatment and tensile testing: (a) bright field; (b) polarized light (original magnification 50 • ).

J. MATERIALS FOR ENERGY SYSTEMS VOL. 8, NO. 1, JUNE 1986 77

 bright field and polarized light (Figure 17). There was no
evidence of a twisted zone or a heat affected zone in this
weld. One possible cause for the difference in this weld and
the other non-heat treated welds is that the joint under-
went a significant strain. The width of the martensitic bands
(variants) increases with strain and makes the substructure
more visible. 4 Further evaluations of the other p/m-to-p/m
welds will reveal whether this substructure is unique to this
one weld or not.

SUMMARY
Electron beam welding was attempted with three p/m plates
containing various levels of oxygen and nitrogen. All p/m
Fig. 15--Weld 13: p/m-to-p/re. welds were porous. The least porosity occurred in a partial
penetration weld in p/m plate 3. This plate contained the
lowest levels of oxygen and nitrogen. Sixteen inertia welds
were made which combined both wrought and p/m U-6Nb
alloy cylinders. The most successful p/m weld was the com-
bination of p/m to wrought. The metallography of the
unheat-treated welds revealed that three distinct regions are
present: fusion zone (a fine-grained region), a twisted zone
(consisting of distorted grains with some twinning), and the
heat affected zone (minimal grain growth noted). The heat
treatment given to the welds successfully blends the prior
fusion zone into the base metals. The properties of all the
p/m-to-wrought and the p/m-to-p/m welds were similar to
the p/m base metal properties. The inertia welding method
is considered a potentially successful welding method for all
Fig. 16--Weld 11: p/m-to-p/m tensile; no post-weld heat treatment. p/m U-6Nb alloy; whereas, the electron beam method will
require p/m material with much lower levels of oxygen and
displayed ductile characteristics, and the elongation was nitrogen in order to be successful.
acceptable but lower than the nominal p/m base metal value:
24 pct v s 28 pct.
CONCLUSIONS
Next to the fusion zone, a highly twinned region was
found. Upon closer inspection, the region had a well- (1) Electron beam welding will not be a successful join-
defined martensitic substructure which is visible in both ing technique for p/m U-6Nb unless the oxygen and nitro-

Fig. 17--Martensitic substructure in Weld 11: (a) bright field; (b) polarized light (original magnification 250•

78 VOL. 8, NO. 1, JUNE 1986 J. MATERIALS FOR ENERGY SYSTEMS

gen levels can be greatly decreased below the current levels REFERENCES
of 400 ppm total.
1. E W. Turner and L. D. Johnson: "Joining of Uranium Alloys", Physical
(2) Inertia welding is a successful joining technique for Metallurgy of Uranium Alloys, Army Materials Technology Conference
p/m U-6Nb alloy and does not require that the oxygen and Series, Brook Hill Publishing Co., 1976, pp. 145-85.
2. P. Adam and H. Wilhelm: "Welding of PM Superalloys", High Tem-
nitrogen levels be extremely low. perature Alloys for Gas Turbines, Dreidel Publishing Company,
(3) The most successful inertia weld is the p/m-to- Dordrecht, 1982, pp. 909-30.
3. P. Adam: "Material Response in Inertia Welding of Superalloys", Mate-
wrought joint. rials Technology and Testing, 1982, vol. 13, pp. 258-62.
(4) Additional evaluations of the p/m-to-p/m joints will 4. V.C. Hemperly: "An Examination of Temperature Induced Shape
Memory of Uranium-5.3-6.9 Weight Percent Niobium Alloys", Y-
be required before their utilization in part designs. 2059, Oak Ridge Y-12 Plant, P.O. Box Y, Oak Ridge, TN 37830,
Sept. 1976.
ACKNOWLEDGMENT
Prepared for the United States Department of Energy by
Martin Marietta Energy Systems, Inc., under Contract DE-
AC05-84OR21400.

J. MATERIALS FOR ENERGY SYSTEMS VOL. 8, NO. 1, JUNE 1986 79