<<

Weldability of and Its Alloys

Tungsten and its alloys can be successfully joined by gas tungsten-arc , gas tungsten-arc braze welding, electron beam welding and by chemical vapor deposition

BY N. C. COLE, R. G. GILLILAND AND G. M. SLAUGHTER

ABSTRACT. The of tungsten eliminated all but a small amount of but it should be noted that the CVD and a number of its alloys consolidated porosity and also eliminated the prob­ material contained more than the nor­ by arc , powder , or lems associated with the high tempera­ mal amounts of fluorine. chemical-vapor deposition (CVD) tech­ tures necessary for welding (such as Various sizes and shapes of tung­ niques was evaluated. Most of the mate­ large grains in the weld and heat-af­ rials used were nominally 0.060 in. thick fected zones). sten and tungsten alloys were joined sheet. The joining processes employed for comparison. Most of them were were (1) gas tungsten-, (2) Introduction products although gas tungsten-arc braze welding, (3) elec­ some arc-cast materials were also tron beam welding and (4) joining by Tungsten and tungsten- alloys welded. Specific configurations were CVD. are being considered for a number of used to determine the feasibility of Tungsten was successfully welded by advanced nuclear and space applica­ building structures and components. all of these methods but the soundness tions including thermionic conversion All materials were received in a fully of the welds was greatly influenced by devices, reentry vehicles, high temper­ cold worked condition with the excep­ the types of base and filler ature fuel elements and other reactor tion of the CVD tungsten, which was (i.e. powder or arc-cast products). For components. Advantages of these ma­ example, welds in arc-cast material were received as-deposited. Because of the terials are their combinations of very increased brittleness of recrystallized comparatively free of porosity whereas high melting temperatures, good and large-grained tungsten the materi­ welds in powder metallurgy products strengths at elevated temperatures, al was welded in the worked condition were usually porous, particularly along high thermal and electrical conductivi­ to minimize grain growth in the heat- the fusion line. For gas tungsten-arc ties and adequate resistance to corro­ (GTA) welds in V]0 in. unalloyed tung­ affected zone. Because of the sten sheet, a minimum preheat of 150° sion in certain environments. Since brittleness limits their fabricability, high cost of the material and the C (which was found to be the ductile- relatively small amounts available, we to-brittle transition temperature of the the usefulness of these materials in structural components under rigorous designed test specimens that used the base ) produced welds free of minimum amount of material consist­ cracks. As base metals, tungsten- service conditions depends greatly alloys were weldable without preheat, upon the development of welding ent with obtaining the desired in­ but porosity was also a problem with procedures to provide joints that are formation. tungsten powder products. Pre­ comparable in properties to the base heating appeared not to affect weld po­ metal. Therefore, the objectives of Procedure rosity which was primarily a function these studies were to (1) determine Since the ductile-to-brittle transition of the type of . the mechanical properties of joints temperature (DBTT) of tungsten is The ductile-to-brittle transition tem­ produced by different joining methods above room temperature, special care peratures (DBTT) for gas tungsten-arc in several types of unalloyed and al­ must be used in handling and machin­ welds in different types of powder metal­ 1 lurgy tungsten were 325 to 475° C, as loyed tungsten; (2) evaluate the ing to avoid cracking . compared to 150° C for the base metal effects of various modifications in heat causes edge cracking and we have and that of 425° C for electron beam- treatments and joining technique; and found that grinding and electrodis- welded arc-cast tungsten. (3) demonstrate the feasibility of charge leave heat checks Braze welding of tungsten with dis­ fabricating test components suitable on the surface. Unless they are re­ similar filler metals apparently did not for specific applications. moved by lapping, these cracks may produce better joint properties than did propagate during welding and subse­ other joining methods. We used Nb, Ta, quent use. W-26% Re, Mo and Re as filler metals Materials Tungsten, like all , in the braze welds. The Nb and Mo 1 caused severe cracking. Unalloyed tungsten in /36 in. must be welded in a very pure atmos­ Joining by CVD at 510 to 560° C thick sheets was the material of most phere of either (gas tung­ interest. The unalloyed tungsten in this sten-arc process) or (electron study was produced by powder metal­ beam process)2 to avoid contamina­ N. C. COLE and G. M. SLAUGHTER are with Metals and Div., Oak Ridge lurgy, arc casting and chemical-vapor tion of the weld by interstitials. Since National Laboratory, Oak Ridge, Tenn.: deposition techniques. Table 1 shows tungsten has the highest R. G. GILLILAND is with the University of . Milwaukee. the impurity levels of the powder met­ of all metals (3410° C), welding allurgy, CVD and arc-cast tungsten equipment must be capable of with­ Paper presented at the American Welding products as received. Most fall within Society 49th Annual Meeting, , standing the high service tempera­ April 1-5. 1968. the ranges nominally found in tungsten tures.

WELDING RESEARCH SUPPLEMENT | 419-s cu 1 u CD i to § 0> — t- 43 l_ 43 H 3 o 3 P o z ~J Srg "a 2 £ 8 3 a> 12 , c oi 3 il o -a 43 " J3 a a C3 c> CD 00 S ~3~ 3 00 3 43 O 0) c "" o £ g > 33 _ U c*-l CA 43 1 ca _f CD rt c ed C » o CC *-» CD "3 00 CD «u , 2 "3 S.S C O iS 3 (_ rt "3 £*£ "3 S 3>'E -* CD 05 0 O 3 . ""«§ P 3 «-> ^5 td I- « w 43 IS o .3 O •— -CO a 1)43 C o S •a c 43 O c CO 43 c c O > >* oo o a, P. s -a TD .33 3J 3 5 a. d 5, „, 2 O 43 C^H (3 LT) CD Ol .3 3 43 I § Qu. CA •s " »- ^E O oCD ^H 3 •5 rt 3 w •a S c C P CA o-£ CD TJ CD 3 E t" i £43 -c H O U •- t> r 00 • • _ CD CD •g.8 - c l-l 43 73 < " 4C0D CED t- cfl c -.23 _g o •3 -3 o 'g -S _g 2 on - a s c 3 w o o c JS 3 "C "30 td 2 CD rt £ 2 ~«H 2 c >-o £ 0 -n u 2 0 <1> — CD 03 rt O — c > c a 3 IS 60 3 U ,. 0) c C CD p ' C •3 g 3 CD CD r3 a 3 C 2 0 2 « p ".Ho c c Pi CD .St £&S J gx E.2 ° £ 2 S'S 3 ca 23 2? a-w-3 o i o ^_cd , "03 CD

CA CA w "3 >>" 00 T3 o o CD 43 43 O V- c C CD CD O jo t- o 'J rt cs o C 43 2 CD > UD CM CD 5 "J 00'S CD 2 O 3 .5 43 -r- 3 3j 2 V O 00 «> a o ^ 43 3 g 33 CA CA CA O x 5 S -a 00"" •3 C CD O 0 E .E ° CD 5 43 1 O O m c O C rt 00 __ H CA 8 3 d CD Poo 3 "3 3 u CD • - 53 c u rt CD CD CD O rt CA D, ci to , *'& 3 CD P Q C 0 - & o 3 CA a 1 CM m e 00 >-. 'w CA '5 43 o 2 rt 00 a > rt CD CO '—1 00 UD CD CD c c w 2 CD 3 33 >, S* ft g o ?- O clE'^rt "3 '^ CD .£ 43 CD 43 o CA CD •2 S^ CA CD CA C ob 3 CA " O CD it"^ E S CD S CD E 10 S3 R fa CD CO _, 00 ^ .3 43 X O 03— o -" •" 43 g Pi c2 e- 33 43 — CD J3 OB" 0 H rt 43 CA 1 0 C 1_ rt 00 C 43 OO s ° eo •a H CD >, rt .3 00 CA CD t2p 3 , ti S c CD CD CD .2 c cu C T3 CD C •3 ^ -S, CD CD CD o 033 c u CD '.a 2 43 .13 •S O S3 ta & 5 143 43 4S to t- •X5 o3 o, 43 " 43 c • & " ^ cw 3 ^24= ^ 1 r».*H s >, C0 CD O p ,+-•*-» C4H o *« « a -a T3 1-1 S3 *~^ 2 2P P, CD CA 0L 0 p • rt T3 CD 43 o S3 LA 3 43 2 ^3 43 - CD c u OH to •C3D 3w oO C cw CD C " O •. 3 - .3 o > a Cum .a 3 CD O i- • b -a > o a "3 £ o i-H 3 CD CD O 4* 033 b a O «J 3 .2 33 'p rt o tS 3. £ CA O CD •K 14 CQ T3 ., .2 C .3 D.T3 t. tH i-t CD Pi N CO T3 -0 3 CA Q. > 2 O c "9 o c o E S O CD CD E a rt 331 CD &. 43 £3 3a 3x3i •£ & rt CD O.T3 3T o S 5 CNJ CM CD og

CD UD UD CD CD CM O O O ,-H C 43 1 1 c O 00 2 CD 2 to CU CA CD ;| 3 H CD K. E .3! 3 P, CA R c- 7 J3 II O 6 O '0 4= u 3 CD —'<3 6 Pi g E 42 P. CO. ^ <3 rt CD o_ pq I "- p .2 43 X3 «£ •° 43 o a. 10 Fs 2 CA CD . o 00 ^r ^T CM ^- _ '5 rt "" CA 3 < . 3 CA T3 -a >* «> » & cr ca CA CM 2 LO CD . CD 2 o o _ ^2C •o rt tR £ «3 £ > > CD O 0 3 F . "ra CD g is CD 3 cp CNJ OJ qCM r^H ^ E CA -o 2 4^ I u p > 3 0) O C £ § *os 3 •3 CD H £ ° 2 > V V 43 CD 4^li5 3 o ,9 ^ 3 -"3 2 * " Oi .s'r 3 £ *j o 3 l/> m E •6 ^343 N .2,i5^ 3 Sh a: CM 1 ^E^ 3 I 3 4> ra D.-3 "rt £ CD H w ro M a a a o __ 43 i"S N r- —1 H 2 S c-43 ^ PH CO o 2 x £ I •o a-i r*- oo r-^. E43 4?33 £ co £ Pi o .5 2 3 0) CM •s a. id 3 43 P. 43 43 '^ >1 o O 3 £ o

CA CA T3 U cu P3 U " T3 U rt J, •C 2 3 i 43 3 cu CD r. O <£ to fl s.a S £ .8 u « ^ rt 'CA > E oi cu cu CA _& C3 GO _ • — " ° ^ P 1- .2 3 43 T33 0JS CD 2 c •a c O CU <0 6 CD CD s 2 9 E c 43 "rt -3 43 u P i« cw O 3 cH D 3 CD cd — CD '-" 0 — 3 43 P Q M "a fi E N S rt •^3 rt p < Pi CD u & y u o 3 o °Jcl CU w „ *^ C°D Cp 3 10 * E 43 ^H "S 2? ai •3 C < S 00 ±3 " _s 2 o cu 2 S Si 43 o 2 E c 2 "d IU - H B > a> S ' _ 00 D. 43 CU - ~ c_ c a 5 CD OS 0 T3 ~ 43 5 43 >, C 3 tu • •- CD 43 3 % t! 8 HIS — fd ^j cc 3 P.T3 rt ,, CA CD .3 rt 3) O •3 43 D.2 CD O h o - 43 p ^ O CD U 1- CO 2 o < 3 u f £»-*• -p o is .y B O 7j Ei. ft •" s * O cu 3 60 j= ° D.43 i? 5 2 u rt >^ co 00 £ „ 2a -23 1' 2 E |r-C3 O < CD 2 CD '3 O 43 cA S Pi 0. cc B -o .2, 3 420-S i SEPTEMBER 1971 [" HEATER

.-THERMOCOUPLE

CARTRIDGE HEATER

Fig. 1—Automatic welding apparatus. A (left)—apparatus incontrolled atmosphere chamber (arrow points to preheating fixture; B (right)—schematic of preheating fixture

Results for Unalloyed Tungsten grains. The columnar grains have gas specimen. To ensure a weld free of bubbles at grain boundaries caused by cracks, preheating at least to the General Weldability 8 fluorine impurities . Consequently, if DBTT of the base metal is recom­ Gas Tungsten-A rc Welding — In the fine grain substrate surface is re­ mended. Electron beam welds in pow­ gas tungsten-arc welding of V16 in. moved before welding, the weldment der metallurgy products also have the thick unalloyed sheet, the work must does not contain a metallographically weld porosity mentioned previously. be substantially preheated to prevent detectable heat-affected zone. Of Gas Tungsten-Arc Braze Welding— brittle failure under stress induced by course, in worked CVD material (such In an effort to establish whether braze . Figure 2 shows a typi­ as extruded or drawn tubing) the welding could be used to advantage, cal fracture produced by welding with­ heat-affected zone of the weld has we experimented with the gas tungsten- out proper preheating. The large grain the normal recrystallized grain struc­ arc process for making braze welds size and shape of the weld and heat- ture. on powder metallurgy tungsten sheet. affected zone are evident in the frac­ Cracks were found in the columnar The braze welds were made by pre- ture. Investigation of preheating tem­ grain boundaries in the HAZ of sever­ placing the along the butt peratures from room temperature to al welds in CVD tungsten. This crack­ joint before welding. Braze welds were 540° C showed that preheating to a ing, shown in Fig. 5, was caused by produced with unalloyed Nb, Ta, Mo, minimum of 150° C was necessary for rapid formation and growth of bub­ Re, and W—26% Re as filler metals. consistent production of one-pass butt bles in the grain boundaries at high- As expected, there was porosity at the welds that were free of cracks. This temperatures". At the high tempera­ fusion line in metallographic sections temperature corresponds to the DBTT tures involved in welding, the bubbles of all joints (Fig. 6) since the base of the base metal. Preheating to high­ were able to consume much of the metals were powder metallurgy prod­ er temperatures did not appear to be grain boundary area; this, combined ucts. Welds made with and necessary in these tests but material with the stress produced during cool­ filler metals cracked. with a higher DBTT, or configurations ing, pulled the grain boundaries apart The of welds and braze that involve more severe stress con­ to form a crack. A study of bubble welds were compared by means of a centrations or more massive parts, formation in tungsten and other metal study of -on-plate welds made may require preheating to higher tem­ deposits during heat treatment shows with unalloyed tungsten and W—26% peratures. that bubbles occur in metals deposited Re as filler metals. The gas tungsten- The quality of a weldment depends below 0.3 Tm (the homologous melt­ arc welds and braze welds were made greatly upon the procedures used in ing temperature). This observation manually on unalloyed tungsten pow­ fabricating the base metals. Autog­ suggests that gas bubbles form by der metallurgy products (the low enous welds in arc-cast tungsten are coalescence of entrapped vacancies porosity, proprietary (GE-15) grade essentially free from porosity, Fig. and gases during annealing. In the and a typical commercial grade). 3A, but welds in powder metallurgy case of CVD tungsten, the gas is Welds and braze welds in each mate­ tungsten are characterized by gross probably fluorine or a fluoride com­ rial were aged at 900, 1200, 1600 porosity, Fig. 3(b), particularly along pound10. and 2000° C for 1, 10, 100 and 1000 the fusion line. The amount of this Electron Beam Welding — Unal­ hr. The specimens were examined porosity, Fig. 3B, particularly along loyed tungsten was electron beam metallographically, and trav­ 3C, in welds made in a proprietary, welded with and without preheating. erses were taken across the weld, heat- low porosity product (GE-15 pro­ The need for preheat varied with the affected zone, and base metal both duced by Co., Cleve­ thickness, length and shape of the as-welded and after heat treatment. land). Gas tungsten-arc welds in CVD Table 2—Typical Analyses of Interstitials in Una lloyed Tungsten After Welding tungsten have unusual heat-affected zones due to the grain structure of the Interstitial content, ppm base metal7. Figure 4 shows the face Weld metal Base metal — and corresponding cross section of Type of tungsten F N2 O, C F N2 0, C such a gas tungsten-arc butt weld. Typical powder metallurgy 14.7 4 56 <20 2.9 5 41 <20 Note that the fine grains at the sub­ Proprietary powder metal­ strate surface have grown due to the lurgy, low porosity 7.9 1 120 <20 <2 1 140 <20 heat of welding. Also evident is the Chemically vapor deposited 9.3 1 72 <20 10.5-23 1 52 <20 Arc cast 1 1 10 14 1 1 7 14 lack of growth of the large columnar

WELDING RESEARCH SUPPLEMENT | 421-s ..- 7\77 - ;:,.„,... .1, -.-',.' ::*;.. .

•_,

WmmMtZjwms,''.

Fig. 2—Gas tungsten-arc weld in unalloyed tungsten. A (top)—fracture which occurred while welding without a preheat; B (bottom)—end view of the fracture. Note the contrasting intergranular failure of weld and heat-affected zone versus the cleavage type failure of the base metal

Since the materials used in this study in hardness at the fusion line. With were powder metallurgy products, increasing aging temperature, the varying amounts of porosity were hardness of the braze weld decreased present in the weld and braze weld until, after 100 hr at 1600° C, the deposits. Again, the joints made with hardness was the same as that of the typical powder metallurgy tungsten unalloyed tungsten base metal. This base metal had more porosity than trend of decreasing hardness with in­ those made with the low porosity, creasing temperature held true for all proprietary tungsten. The braze welds aging times. Increasing time at a con­ made with W—26% Re filler metal stant temperature also caused a simi­ had less porosity than the welds made lar decrease in hardness, as shown for with the unalloyed tungsten filler metal. an aging temperature of 1200° C in No effect of time or temperature Fig. 7B. was discerned on the hardness of the welds made with unalloyed tungsten Joining by Chemical Vapor Deposi­ as filler metal. As welded, the hard­ tion—Joining of tungsten by CVD ness measurements of the weld and techniques was investigated as a meth­ base metals were essentially constant od for producing welds in various and did not change after aging. specimen designs. By use of appropri­ However, the braze welds made with ate fixtures and masks to limit deposi­ W—26% Re filler metal were consid­ tion to the desired areas, CVD and erably harder as produced than the powder metallurgy tungsten sheets base metal (Fig. 7). Probably the were joined and end closures on tub­ higher hardness of the W-Re braze ing were produced. Deposition into a weld deposit was due to solution bevel with an included angle of about hardening and/or the presence of o- 90 deg produced cracking, Fig. 8A, finely distributed in the so­ at the intersections of columnar grains lidified structure. The tungsten- growing from one face of the bevel rhenium phase diagram11 shows that and the substrate (which was etched localized areas of high rhenium con­ away). However, high integrity joints tent could occur during rapid cooling without cracking or gross buildup of and result in the formation of the impurities were obtained, Fig. 8B, hard, brittle o- phase in the highly when the joint configuration was segregated substructure. Possibly the o- changed by grinding the face of the phase was finely dispersed in the base metal to a radius of V2 in. grains or grain boundaries, although tangent to the root of the weld. none was large enough to be identified Fig. 3—Centogenous gas tungsten-arc To demonstrate a typical applica­ by either metallographic examination welds in tungsten. A (top)—arc-cast tion of this process in fabrication of or X-ray diffraction. tungsten, B (middle)—typical powder metallurgy tungsten, C (bottom)—propri­ fuel elements, a few end closures were Hardness is plotted as a function of etary, low porosity powder metallurgy made in tungsten tubes. These joints distance from the braze-weld center tungsten. Etchant: 25% H20», 25% were leak-tight when tested with a line for different aging temperatures NH..OH, 50% H.0 (reduced 50% in re­ mass spectrometer leak detec­ in Fig. 7A. Note the abrupt change production. tor.

422-s I SEPTEMBER 1971 :•'••

... -A-:•:-;..7 : ~, Fig. 4—Gas tungsten-arc butt weld in chemically vapor deposited tungsten; top view and cross section. Etchant: H20a NHXIH, H20

V

J

Fig. 5—Cracking and bubble formation in heat-affected zone of weld in chemically vapor-deposited tungsten. Same general area is shown at increasing magnifications. Etchant: H202, NH,OH, H20

Mechanical Properties Bend Tests of Fusion Welds— Ductile-to-brittle transition curves were determined for various joints in unalloyed tungsten. The curves in Fig. 9 shows that the DBTT of two powder metallurgy base metals was about 150° C. Typically, the DBTT (the lowest temperature at which a 90 to 105 deg bend could be made) of both materials increased greatly after weld­ ing. The transition temperatures in­ creased about 175° C to a value of 325° C for typical powder metallurgy tungsten and increased about 235° C to a value of 385° C for the low porosity, proprietary material. The difference in the DBTTs of welded and unwelded material was attributed to the large grain size and possible redistribution of impurities of the welds and heat-affected zones. The test results show that the DBTT of typical powder metallurgy tungsten welds was lower than that of the pro­ prietary material, even though the lat­ ter had less porosity. The higher DBTT of the weld in the low porosity tung­ sten may have been due to its slightly larger grain size, Fig. 3A and 3C. The results of investigations to de­ termine DBTTs for a number of joints in unalloyed tungsten are sum­ marized in Table 3. The bend tests were quite sensitive to changes in testing procedure. Root bends ap­ peared to be more ductile than face bends. A properly selected stress Fig. 6—Braze welds on powder metallurgy tungsten sheet. The various fille after welding appeared to lower the metals used are columbium, , W—26% Re, molybdenum, and rhe DBTT substantially. The CVD tung- nium. As-polished

WELDING RESEARCH SUPPLEMENT | 423-s TYPICAL POWDER METALLURGY o BASE METAL -4 WELDS PROPRIETARY LOW POROSITY POWDER METALLURGY • BASE METAL * WELDS

80 BASE METAL r n ' 60 • rn E&/ 1 WF I ns 20 i / 1 J JZ

200 300 400 TEMPERATURE (°C1 . '. TEMPERATURE- 200* C Fig. 9—Ductile-to-brittle transition ^-AS W LMo curves for powder metallurgy tungsten ""^r^fe0" base metal and welds contrast to unalloyed tungsten, W— 7 wm 26% Re tubing and sheet were autog- ! Fig. 8—Joints made by chemical vapor enously welded without the need for deposition process between powder preheat. Powder metallurgy W—26%

—— WELO MET metallurgy tungsten sheets. A (top)— Re, like the unalloyed powder metal­ crack in joint at intersections of col­ lurgy tungsten, also exhibited weld umnar grains. Bevel was 90 deg included porosity, Fig. 10A. The absence of DISTANCE FROM WELD CENTE! angle. B (bottom)—joint free of cracks. porosity in arc cast W—26% Re, Fig. Fig. 7—Effect of aging on hardness of Bevel was ground to a radius of y2 in. tangent to the root of the weld. Etchant: 10B, again illustrates the influence powder metallurgy tungsten welded of the process history of the base with W—26<% Re filler metal. A (top)— H202, NH,OH, H20 (reduced 57% in re­ effect of varying temperature at con­ production) metal upon weldability. stant 100-hr aging time. B (bottom)— Tungsten-rhenium-molybdenum ma­ effect of varying time at constant 1200° Bend Tests of Braze Welds—Gas terial can also be welded without C aging temperature. pyramid tungsten-arc braze welds made with preheating. However, a high tempera­ hardness tested at 1-kg load Nb, Ta, Mo, Re, and W—26% Re as ture stress relief near the recrystalliza­ filler metals were also bend tested and tion temperature is needed before sten had, as welded, the highest DBTT the results are summarized in Table 4. welding. Without a sufficient stress (560° C); yet when it was given a The most (a 90 deg bend relief, severe centerline cracking may 1 hr stress relief of 1000° C after angle at 525° C) was obtained with a be encountered at the centerline, as welding, its DBTT dropped to 350° rhenium braze weld. (This amount of shown in Fig. 11. A stress relief of C. Stress relief of arc welded powder bending is probably borderline since 1300° C for 1 hr eliminated the prob­ metallurgy tungsten for 1 hr at 1800° the specimen was cracked when re­ lem. Powder products again exhibited C reduced the DBTT of this material moved from the testing rig.) large amounts of porosity. by about 100° C from the value de­ Although the results of this cursory Electron Beam Welding—Electron termined for it as-welded. A stress study indicate that a dissimilar filler beam welds in tungsten alloys are relief of 1 hr at 1000° C on a joint metal may produce joints with illustrated in Fig. 12. End caps were made by CVD methods produced the mechanical properties inferior to welded to several test capsules with a lowest DBTT (200° C). It should be those of homogeneous welds in tung­ defocused electron beam. Both the noted that, while this transition tem­ sten, some of these filler metals may cap and the capsule shown in Fig. perature was considerably lower be useful in practice. 12A were made of powder metallur­ than any other transition temperature gy W—25% Re. Note the gross poros­ determined in this study, the improve­ Results for Tungsten Alloys ity at the fusion lines. The specimen ment was probably influenced by the shown in Fig. 12B had an arc cast lower strain rate (0.1 vs 0.5 ipm) General Weldability cap and a powder metallurgy capsule. used in tests on CVD joints. Gas Tungsten-Arc Welding—In No porosity was present near the fu-

Table 3—Ductile-to-Brittle Transition Temperatures of Joints in Una lloyed Tungsten

Surface Approximate Joining technique Type of tungsten Type bend in tension Joint condition DBTT" (°C) Gas tungsten-arc Powder metallurgy Longitudinal Faceb As welded 450 Longitudinal Root As welded 325 Transverse Face As welded 475 Longitudinal Face Stress relieved, 1800°C <350 Stress relieved, 2800°C >500 Powder metallurgy Longitudinal Root As welded 385 (low porosity) CVD Longitudinal Face As welded 560 CVD Longitudinal Face Stress relieved, 1000°C 350 Electron-beam Arc cast Longitudinal Face As welded 425 CVD Powder metallurgy Longitudinal Face Stress relieved, 1000°C 235b CVD Longitudinal Face Stress relieved, 1000°C 200" a Ductile-to-brittle transition temperature, defined as lowest temperature at which specimen bent fully (90 to 105 deg without cracking. b Bend test strain rate changed to 0.1 ipm (all others tested at 0.5 i pm).

424-s | SEPTEMBER 1971 Table 4—Results of Longitudinal Bend Tests at 525° C of Gas Tungsten-Arc Braze Welds in Powder Metallurgy Tungsten Bend angle at which first crack Filler metal occurred (deg) Visual observations of tested specimens Niobium 14 Fractured across entire width Tantalum 53 Cracks in weld, heat-affected zone, and base metal W-26% Re 17 Cracks in weld and heat-affected zone Molybdenum 44 Fractured across entire width Rhenium 90 Very slight crack in weld

assurance of making the joint without sion line with the arc cast cap. Figure Fig. 10—Autogenous gas tungsten-arc 12(c) shows both an arc cast cap and melting through the thin tube. welds in W—26% Re. A (top)—weld in capsule of W—5% Mo, which con­ Metallographic examination of powder metallurgy base metal. B (bot­ tains no porosity at either interface. these prototype welds revealed that tom)—weld in arc-cast base metal. This type of evidence is found fre­ they had complete penetration, no Etchant: H202, NH.OH (reduced 62% in quently with both unalloyed tungsten cracks and only a small amount of ) and tungsten alloys. fine porosity. Figure 13B shows the difference in grain growth exhibited Conclusions Component Fabrication by the two tungsten tubes. In the Tungsten and many of its alloys can Another part of our program con­ bottom tube, the grains grew very be successfully joined by welding, cerns determining the feasibility of little as compared to the rapid grain braze welding, and chemical vapor fabricating test components from growth in the vertical tube. Since the deposition, provided certain techniques tungsten and tungsten alloys. A porosity shown was along the inter­ are used. Special machining proc­ demonstration assembly simulating a face between the braze weld deposit esses must be employed, the material loop was successfully tung­ and the fine grained tungsten tube, a must be handled with care, and equip­ sten arc, braze welded, Fig. 13A. higher concentration of impurities ment must be capable of producing Nondestructive inspection revealed may be present in the grain bound­ and handling the extreme heat needed the welds to be helium leak-tight and aries of the fine grained tube, which for welding tungsten. For single-pass crack-free. It was constructed from may have slowed the grain growth. welds in unalloyed tungsten, the 0.275-in. OD by 0.035-in. wall CVD Upon closer examination the o- workpiece must be heated to at least tungsten tubing. It was welded manu­ phase of the tungsten-rhenium system the DBTT of the base metal before ally in a chamber with a controlled was discovered near the edge of the welding to avoid transverse cracking. atmosphere of very pure . In an braze weld, Fig. 13C. It was posi­ Neither W—26% Re nor W—25% effort to reduce the required heat tively identified by microprobe analy­ Re—30% Mo required this preheat input, W—26% Re was chosen as the sis and by the use of special etching because the transition temperatures of filler metal. techniques (0.5 N NaOH solution pref­ the base metals are below room tem­ erentially attacks o- phase in tungsten- perature. However, a preheat may be In general, problems increase in desirable for large or complex struc­ making manual welds as the tempera­ rhenium). Sigma phase, since it is hard and brittle, is an undesired microcon- tures, which may require multipass ture increases (glove deterioration, welds. inconvenience). Also, in build­ stituent. Microsegregation during sol­ ing complex components, the lower idification of the weld metal probably Possibly because of solid solution heat required for braze welding is a produced areas high enough in rheni­ hardening and the presence of o- definite advantage in decreasing weld um content to form the o- phase. phase, W—26% Re weld metal is stresses and reducing the size of the heat-affected zone. In the case of a small thin tube welded to a large component, a filler metal with a lower melting point increases the ease and

W-25% Re W-25%Re W-3%Mo Fig. 12—Electron beam welds in tungsten alloy capsules. Porosity is located only Fig. 11—Effect of stress relief on welds at the interface of weld metal and powder metallurgy product. A (left)—powder in W—25% Re—30% Mo. A (top)—stress metallurgy tungsten cap and tube, B (center)—arc-cast tungsten cap and powder relieved 1 hr at 950° C. B (bottom)— metallurgy tungsten tube, and C (right)—arc-cast tungsten cap and tube (reduced stress relieved 1 hr at 1300° C 25% in reproduction)

WELDING RESEARCH SUPPLEMENT | 425-s Fig. 13—Simulated corrosion loop braze welded with W—26% Re filler metal. A (left)—completed loop, B (middle)— cross section of one of the braze-welds; C (right)—higher magnification (X100) of the root of the weld. Note the a phase along the edge of the root. Etchant: 25% NH4OH, 25% H202, and 50% H»0. harder than unalloyed tungsten (480 However, more development must be Suppl., 528-s to 542-s (1969). 3. "Vapors Create Tungsten Joints," vs 400 dph). Aging for increasing done before this process can be ap­ Age 19(21). p. 76-77 (May 1963). times and temperatures up to 1600° C plied. 4. Heestand, R. L., Federer, J. I., and Leitten, C. F., Jr., Preparation and Evalu­ and 1000 hr decreases the hardness of ation of Vapor Deposited Tungsten, the W—26% Re weld metal to that of A cknowledgements ORNL-3662 (August 1964). unalloyed tungsten. 5. Schaffhauser, A. C., "Low-Tempera­ The authors gratefully thank J. D. ture Ductility and Strength of Thermo- Powder products, whether unal­ Hudson for preparing, welding and chemically Deposited Tungsten and Effects loyed tungsten or tungsten alloys, have of Heat Treatment," pp. 261-276, Summary testing specimens, G. E. Moore of the of the 11th Refractory Composites Work­ porosity in the weld zone, particularly Welding and Facility, Plant ing Meeting, AFML-TR-66-179 (July along the fusion line. The amount of 1966). and Equipment Division, for manually 6. Evaluation Test Methods for Refrac­ porosity depends on the process his­ welding the bead-on-plate specimens. tory Metal Sheet Materials, Materials Ad­ tory of the base metal as well as of its visory Board Refractory Metal Sheet Roll­ We thank R. L. Heestand (now at ing Panel, MAB176-M (Sept. 6, 1961). Re­ composition. BMI), R. G. Donnelly, A. C. vised. As expected, the DBTT of all 7. Lundin, C. D., and Farrell, K., "Dis­ Schaffhauser, R. E. McDonald, W. C. tribution and Effects of Gas Porosity in grades of tungsten was greatly in­ Robinson (now with Union , Welds in CVD Tungsten," WELDING JOURNAL, creased by welding. Stress relief be­ Greenville, S.C.) and J. I. Federer for 49(10). Research Suppl., 461-s to 464-s fore welding reduced the cracking sus­ (1970). their invaluable assistance in material 8. Schaffhauser, A. C., and Heestand, ceptibility of the welds and heat treat­ procurement and consultation. The R. L., "Effect of Fluorine Impurities on ments after welding appeared to im­ Grain Stability of Thermochemically De­ work of the following groups of the posited Tungsten." pp. 204-211, 1966 IEEE prove the ductility of the welds. The Metals and Ceramics Division is also Conference Record of the Thermionic Con­ DBTT for unalloyed tungsten welds appreciated: the Mechanical Proper­ version Specialist Conference, Nov. 3 and ranged from 325 to 560° C, depend­ If, 1966, Houston, Texas, Institute of Elec­ ties Group for testing the bend speci­ trical and Electronics Engineers, New ing on the type of base metal and mens, the Metallography Group for York, testing conditions. Use of a dissimilar 9. Farrell. K., Houston, J. T., and preparing the photomicrographs and filler metal (braze weld technique) Chumley, J. W., "Hot Cracking in Fusion metallographic samples, and the Re­ Welds in Tungsten," WELDING JOURNAL, did not improve the properties of the 49(3), Research Suppl., 132-s to 137-s ports Office for preparing the manu­ resulting joint but appeared to cause (1970). script. 10. Farrell, K., Federer, J. I., Schaff­ further embrittlement. hauser, A. C, and Robinson, W. C., Jr., "Gas Bubble Formation in Metal Depos­ Chemical vapor deposition is a References its," pp. 263-267, Chemical Vapor Deposi­ feasible and promising process for tion 2nd Intern. Conf., ed. by J. M. Blo- 1. Barth, V. D.. Physical and Mechani­ cher, Jr., and J. C. Withers, The Electro­ joining tungsten. Of all the joints cal Properties of Tungsten-Base Alloys, chemical Society, New York, 1970. studied, those made by CVD methods DMIC-127, pp. 6-10 (March 1960). 11. English, J. J., Binary and Ternary followed by a stress relief of 1000° C 2. Lessman, G. G, and , R. E., Phase Diagrams of Columbium, Molybde­ "The Weldability of Tungsten Base Al­ num, Tantalum and Tungsten, DMIC-152, had the lowest DBTT (200° C). loys," WELDING JOURNAL, 48(12), Research p. 92 (April 28, 1961).

NEW WELDING RESEARCH COUNCIL BULLETINS

WRC BULLETIN 160: "High-Frequency Resistance Welding" by D. C. Martin WRC BULLETIN 161: "The Fabrication and Welding of High-Strength Line- " by H. Thomasson The price of either WRC Bulletin 160 or 161 is $1.50 per copy. Orders for single copies should be sent to the American Welding Society, 345 East 47th St., New York, N.Y. 10017. Orders for bulk lots, 10 or more copies, should be sent to the Welding Research Council, 345 East 47th St., New York, N.Y. 10017.

426-s I SEPTEMBER 1971