,’

UCRL-JC-128691PREPRINT

Characterization of Electron Beam MeltedjUranium - 6% Niobium Ingots

R. H. McKoon

This paper was prepared for submittal to theElectron Beam Melting and Refining State of the Art 1997 Conference

Reno, NVOctober 5-7,1997

October 31,1997

,\

DISCLAIMER

This document was prepared as an account of work sponsored by an agency ofthe United States Government. Neither the United States Government nor theUniversity of California nor any of their employees, makes any warranty, expressor implied, or assumes any legal liability or responsibility for the accuracy,completeness, or usefulness of any information, apparatus, product, or processdisclosed, or represents that its use would not infringe privately owned rights.Reference herein to any specific commercial product, process, or service by tradename, trademark, manufacturer, or otherwise, does not necessarily constitute orimply its endorsement, recommendation, or favoring by the United StatesGovernment or the University of California. The views and opinions of authorsexpressed herein do not necessarily state or reflect those of the United StatesGovernment or the University of California, and shall not be used for advertisingor product endorsement purposes.

CHARACTERIZATION OF ELECTRON BEAM MELTEDURANIUM - 65%0NIOBIUM INGOTS

Robert H. McKoonLawrence Livermore National Laboratory

Livermore, California 94551

A study was undertaken at Lawrence Livermore NationalLaboratory to characterize uranium, 6’% niobium ingotsproduced via electron beam melting, hearth refining andcontinuous casting and to compare this material withconventional VIM/skull melt /VAR material. Samples of boththe ingot and feed material were analyzed for niobium, tracemetallic elements, carbon, oxygen and nitrogen. Ingot sampleswere also inspected metallographically and via microprobeanalysis.

Introduction

Since the mid 1940’s, uranium metal has been fabricated into a widevariety of parts at the Department of Energy’s Y-12 complex in Oak Ridge,Tennesse. In the case of the uranium-6% niobium alloy, the reactivity of thismaterial in the liquid state results in large amounts of non-recyclable scrap andwaste being generated during production. Approximately half of this waste is adirect result of producing the alloy and the remaining half results fromthermomechanical component fabrication processes. The low yield of thecurrent multi-step flow sheet creates substantial environmental, safety andhealth liabilities and high cost in the production of these parts.

By use of Electron Beam Cold Hearth Refining (EBCHR), about 80’?ZOofthe uranium waste and scrap generated under current manufacturing methodscan be eliminated. A single EBCHR melt can replace the current three meltingstep process currently in use and as shown in Figure 1. In addition, the abilityto recycle materials stockpiled from current manufacturing processes couldreduce the requirement for virgin feedstock to zero for the foreseeable futureby working down existing stores of scrap. As an added benefit, it is estimatedthat this will also reduce radiation exposure levels to plant workers by about30Y0.

*

Current Status

In 1993, Lawrence Livermore National Laborato~(LLNL) retrofitted anexisting, on-site, uranium-qualified vacuum processing system as an EBCHRfurnace. In 1994, funding was obtained to demonstrate the capability ofproducing 5.5-inch-diameter U-6 Nb ingots meeting Y-12 specifications and todesign and construct a scrap feeder capable of recycling chopped Y-12 plate scrap[1]. A modeling effort was also initiated to better understand the relationshipsbetween input process parameters and final ingot structure.

This facility is complete and has produced U-6 Nb from both virgin andscrap feedstock which meet Y-12 specifications. In 1996, the furnace wasmodified to allow the production of 8.25-inch diameter ingots, which arenecessary to meet the requirements of the current Y-12 production flowsheet.Figure 2 shows typical ingots in both 5.5” and 8.25” diameters.

Current work centers around characterizing the EBCHR produced U-6Nb and comparing it to conventional Vacuum Induction Melted/SkullMelted/Vacuum Arc Remelted (VIM/SM/VAR) material produced at the Y-12plant. Previous work by Y-12 established the characterization protocol for U-Nb ingots during each stage of the VIM, SM and VAR processing [2] andvariations in Nb composition top to bottom and with radial location withinthe ingot [3].

Procedures

A full sized 8.25” diameter ingot, 24” long and weighing 315 kg,produced under typical operating conditions was selected for characterization.The raw materials for producing this ingot was 100% scrap plate. Electronbeam power and power distribution over the melt area, the hearth and cruciblewere held constant during the run and a constant casting rate of 370 kg/hr wasachieved. Table 1 shows typical casting rates for VA.R systems at the Y-12 plantwhich have roughly twice the capacity of the present LLNL 250 kW system.

Diameter - in. Length - in. Melt Rate - kg/hrk AVAR 8.25 30-40 630-900EB 8.25 24 300-500

Table 1. Comparison of VAR and EB melting parameters

The characterization procedures used closely followed the evaluationsfirst devised by Y-12 [2] for standard VIM/SM/VAR product. A process flowdiagram for these procedures is shown in Figure 3. The characterization effortsvrere organized into sectioning, wet chemical analysis, interstitial analysis,microprobe analysis and metallography according to the following procedures.

1.0

2.0

1.1

1.2

1.3

1.4

Top d Bottom of Ingot for Chem”cal AnalvsisThe ;Onp2“ portion and the bottom ~“ portion were cut from the ingot.Center and edge samples were machined from the bottom of the toppiece and the top of the bottom piece.

Half I ~ot for XRF Macro AnalvsiSThe ;enmaining 20” long ingot was cut in half lengthwise to generatetwo hemi-cylinders. A - 1“ thick slab (S.25” x -20”) was cut from theback hemi-cylinder, and the front (larger width) face of the slab waspolished to a 16y RMS finish. The polished slab surface was air etchedand photographed.

90 t Slice SamDles for Chemical AnalvsisSlices from the front hemi- cylinder were cut at measured distancesalternating from 1.1” and 3.0” along its length according to Figure 4.The top sides of the 1.1” thick slices were machined to a 120g inchfinish. 0.125” wide by 0.015” deep did channels were then machinedfrom the top faces of the samples as shown in Figure 5, at 0.5” intervalsalong the ingot radius.

Carbon /Oxv~en and Metallom a~hic Sam~l~Four cubes a~proximately l-ficfi by l-inc~ were cut from each l-inchthick slice as shown in Figure 6.

Chemical Analysis

2.1

2.2

. .loblum and T ace Metalhcs

Machine turnin~s from the top and bottom pieces, and from ingotslices obtained in section 1.3 were analyzed for niobium and tracemetallics via Inductively Coupled Plasma - Emission Spectroscopy(ICP-ES).

G bon. o ygen nd NitrogenC, ~ and ; anal~ses of samples generated in section 1.4 wereperformed using standard LECO combustion equipment. Cut andpolished samples were acid etched in lN HN03 for 1 minute, thenrinsed in de-ionized water, dried and immediately stored in acetoneuntil ready for analysis.

3.0 Metallo~rar)hv and MicromobeThe 1“ &b>s generated i; section 1.4 were examined for grain size,niobium dendritic microstructure, and carbide and oxide size anddistribution using the following procedure:

3.1

3.2

Results

etallou vhvSamples ;ere polished using established procedures for uraniumalloys. Carbide and oxide inclusions, grain size and microstructurewere observed and recorded in digital format.

micro-segregationabove were examined in an electron microprobe forof niobium.

Full ingot slice

A photograph of the ingot slice showing typical U-Nb banding is shownin Figure 7. Banding is a result of the solidification dynamics for this particularalloy system and indicate areas of high and low niobium within the ingot on amacro scale. The liquid pool at steady state, as judged from the band profiles isabout 1.3” deep compared with a pool depth of 4“[4] for typical VAR product.EB processing gives the ability to independantly vary rate and power input topromote shallow or deep pools as desired for a particular metal. To promotethe desired shallow solidification front, power was input to the ingot as a ringshaped beam directed to the periphery of the ingot only. As a result, bandingappears less pronounced in the center of the ingot compared to the edgeregions. It should be remembered, however, that this ingot was cast at roughly2/3 of the nominal VAR rate so that a somewhat shallower pool would beexpected. Note also the absence of porosity throughout the ingot. Electronbeam hot topping was used to insure a completely sound ingot.

Ingot top and bottom

Analytical results for the feed material, and the ingot top and bottom(center and edge) for niobium and trace metallic elements are shown below inTable 2. Niobium is within the specified range of 5.2-6.5 percent by weight inthe areas of merit for the Y-12 specification: Top center and edge, and bottomcenter.

Previous Y-12 investigations (5) on purification of uranium by electron beamrefining have shown substantial reductions in high vapor pressure metallicimpurities after a single electron beam melt. Table 2 shows similar results forthis current investigation. Chromium and manganese levels, the two highvapor pressure constituents measured, are preferentially lowered during

electron beam processing. The more reactive metals (Me, Ni, Ti, and Zr) alsoappear to be somewhat lower in the ingot than in the feed although it is notclear if this is due to their higher vapor pressures with respect to uranium, orto their reactivity with carbon in the melt to form insoluble metal carbides.

Overall, tramp impurities are lowered by EBCHR, resulting in a modestimprovement in material purity. Whether or not this leads to improvementsin the final part remains to be seen. The niobium appears to be slightlyenriched compared to the feed, however as will be shown, this is at odds withresults of niobium analysis for the individual ingot slices.

Nb Al Cr Mn Mo Ni Ti ZrSpec. 5.2-6.5* 75 75 75 75 500

Feed 6.09 0 23 14 16 20 61 31

Bottom-edge 6.87 0 7 2 2 18 73 18

Bottom-center 6.28 0 7 4 6 14 61 16

Top-edge 6.31 0 15 8 18 15 56 22

Top-center 6.18 0 26 4 8 20 66 22

* The Y-12 specification (Document number 00-M-200) calls for niobiumanalysis of the top center and edge and the bottom center only.

Table 2. Ingot top and bottom analysis

Niobium distribution

Niobium distribution as a function of both radial position and top tobottom location within the ingot are shown in Figure 8 for both the tvpicalVIM/SM/VAR material and ;he EBCHR materia~ From these curv& it canbe seen that in both cases, niobium tends to be higher in the center of the ingotnear the bottom of the ingot whereas toward the top of the ingot, the situationreverses and niobium is lower in the center. Various theories have been putforth in the literature for the solidification patterns seen in this alloy, some ofwhich imply that judicious programming of the electron beam on the ingotduring casting may lead to more optimal ingot structures. For this particularset of casting conditions, however, EBCHR appears to produce a productcomparable to the standard Y-12 material.

Carbon, Oxygen and Nitrogen

Plots of carbon, oxygen and nitrogen as a function of radial position andlongitudinal location within the ingot are shown in Figure 9. From this data,it appears that carbon may be reduced slightly during melting whereas oxygenand nitrogen show slight increases. These small changes, however, areprobably not statistically significant, especially since free energy calculations foroxygen and uranium oxide reacting in the liquid state show that this is not afavored reaction.

Metallography

Metallographically, the material appears similar to the baseline Y-12 U-6Nb. Grain size, center to edge, and distribution and size of inclusions appeartypical of the VIM, SM., VAR material. The material appears free of majorinclusions and voids. Figure 10 is a typical metallographic photo showing thatcarbides are randomly distributed and all below 20 microns in size.

Microprobe analysis

A plot of an electron microprobe scan of both the VIM/ SM/VAR andthe EBCHR material is shown in Figure 11. As expected, since theinterdendritic coring giving rise to these micro-variations are a function of theparticular alloy system rather than processing conditions, the EBCHR materiallooks the same as standard VAR material.

Discussion

This EBCHR ingot produced from 100°/0recycled scrap plate appears to becomparable to the standard Y-12 product in all areas studied. Gross segregation(top-bottom) is comparable. Macro-segregation (banding) of the EBCHRmaterial appears to be slightly less distinct than in the standard Y-12 material.Micro-segregation (dendritic coring) is same as for the Y-12 material. Carbon,oxygen and nitrogen appear to be unchanged by EB processing, whereas in theY-12 material, carbon is increased by about 50 ppm with each melt. The slightdifferences seen between Y-12 reported results and LLNL reported resultsprobably have more to do with differences in analytical equipment andtechniques between the two institutions than in actual variations in ingotchemistries.

This study has shown that a single EBCHR melt can produce uranium6°/0niobium ingot material meeting current specifications and which appearscomparable to the baseline, triple melted material. The technology is nowready to be transferred to the Y-12 production facility where furthercharacterization and statistical comparisons will need to be made prior to finalcertification of the process. Components are at present being fabricated fromEBCHR ingot material, and will be evaluated over the course of the comingyear.

REFERENCES

1. R. H. McKoon, “Progress toward uranium scrap recycling via EBCHR,” UCRL-JC-118173, Lawrence Livermore National Laboratory, October 1994.

2. L. R. Chapman, “Metallurgical characterization of vacuum-induction-melted,skull-arc-melted and vacuum-arc-remelted uranium-niobium alloy castings,”Y-12 Development Division Technical Progress Report, Y-2414-3, September1989.

3. Y-12 Development Division Technical Progress Report Y-2024-3, March 1977,pg. 16.

4. Y-12 Technical report Y-2302-3, August 1982, pg. 100.

5. G. L. Powell, A. L. Williamson, J. J. Dillon, J. B. Soward, “Electron-BeamMelting of Mulberry and Uranium,” Y-12 report Y/AJ-169, November 19, 1976.

*This work was performed underby Lawrence Li vermore National

the auspices of the U.S. Department of EnergyLaboratory under contract No. W-7405-Eng-48.

UOnmwNb pool

f

Skull: sdidllhdu.sNb9Ml

(( I/’=0--a—0---0\O

cut Arcmolt Sswingot8intoforgingbinds

Figure 1: Current VIM/SM/VAR production route for U-6Nb

Figure 2: 5.5” and 8.25” diameter U-6Nb ingots

●

cast ingot Saw cut off top Saw cut ingot Saw cut hami- Machinachain. cut4 Cut $amplaa for4 & bottom 2- - tongitudinaltyto v cytindmto make w samptesfrom * mataltographu + interstitial

hami-cylindar 10 Sticas fiva 1“Stiiee aampfas analysisA

I I I I

21E3Machine off Cut W“ thickcenter& edge stabfromsamples hemi-cytindaf

Analyze turnings Machine onefor Nb face of slab to

60URMSfinish

II

B+=-l

*ES!EIrlRecordmicrostructure

1 1

ImMiompiobaanatyah0?samples

Figure 3: Ingot characterization flow sheet

TOPBO170M

T“ I I I I I 1I

I III II II II

I I. .

IT I AT ;T; BT ;T; CT ;T; DT ;T; ET(8.25) ,

II II II III II II II III II II II III II II II III I 1 1 1 I

II I I

L.20-_JFigure4: Ingot slicing plan

1i

o.+’~ ~

31I .13x8pl

-J 0.50

.50 x 7 pl

Figure5: Radial sample locations

(Ml) I (M2) ‘ (N13) ‘ (M4)’

1

‘ -+0 l+--

Figure 6: Metallographic

(typical)

sample locations

1.0

7

Figure7 Ingot slice showing typical U-6~bandtig

5 .b

t

hdiil Paifk (mm)

a) Typical VIM/SM/VAR material

I I I I I I I I75 so 2s c Z 50 7s

6.4F_ I

6.2 -;

s‘------559 mm

# 5.6— 479 mm

----375 mm

’165 mm

5

4.8 J.,

.110 .90 .70 -50 -30 -10 10 30 50 70 90 110

Radial Position - mm

b) EBCHR melted ingot

Figure 8: Edge to center niobium distribution in 210 mm diameter ingots atgiven distances above bottom of ingot

0070

60Eg 50

e 402& 30u

20

10

0

-——. .—— .. ——. —-. . .1

I

. -------- ---r ‘~:..

---=-S.=* ,---- .““. .#q-l----

---- 0 -./

---- /-- -“ I ~ Starting

II

‘-,--MI

II

.-. ‘--M2! M3I

1 ~M4

5T.— —

a) Carbon80

70

60Eg 50

g- 40al2300

20

10

0

4T 3T 2T IT

Sample Position

/0

./-

~ Starting

‘+-Ml-.. ---M2.. _.. - M3

~ M4

5T 4T 3T 2T lT

b) OxygenSample Position

60

50

EQ 40n.

5 30m

g 20

10

0

I >. I

~ Starting

–+--Ml..- ---M2

M3

~ M4

I[

5T 4T 3T 2T IT

c) Nitrogen Sample Positon

Figure 9: Interstitial element levels in EBCHR ingot

Figure 10 Typical carbide distribution in EBCHR produced U-6Nb

Ret Y-12 Cw. m Re$m, Y.2702Jan. 1978. Pg. 8

Figure 11: Macrosegregation in U-6Nb ingots

Subject Index

1. Uranium - all pages

2. Uranium-Niobium alloy - all pages

3. Electron beam melting - all pages

Technical Inform

ation Departm

ent • Lawrence Liverm

ore National Laboratory

University of C

alifornia • Livermore, C

alifornia 94551