Friction Welding of Incompatible Materials

The feasibility of using a metal interlayer to friction weld certain similar and dissimilar metals was established

BY F. SASSANI AND J. R. NEELAM

ABSTRACT. A modified method for fric­ Although friction welding is inherently ferred to as the "three-element meth­ tion welding of incompatible materials a versatile process, it does not result in a od") in joining incompatible materials. was investigated. Friction welding of joint of acceptable quality in some cases, This paper considers the outcome of this brass to copper, bronze to steel and such as the welding of titanium to nickel method that has been used to overcome titanium to nickel with different interlayer where no weld is formed, and in silver some of the situations which result in a materials was performed, and varying alloys to themselves where a brittle weld brittle weld or no weld at all. degrees of success and mechanical joint is formed. Such combinations, in this strengths were observed. paper, have been termed "incompati­ Three-Element Friction Welding Metallurgical analyses and observa­ ble." While some other methods arrive at tions of the extent of metallic bonding a good metallic bond between these The method involves the use of a third and diffusion showed that incompatible materials, the possibility of using friction element, the "interlayer" or the "inter­ materials can be friction welded with an welding has advantages ranging from mediate material," which is placed as a interlayer and mechanically improved simplicity and efficiency to quality and buffer in a small premachined recess be­ joints obtained. productivity. tween the two incompatible base materi­ The objective of the research dis­ als being welded —Fig. 1. The element welds with both of the base materials introduction cussed here is to examine the potential of and, by remaining as a thin metallic bond­ a modified friction welding process (re- Friction welding has become a widely ing layer between them, forms a joint. used process for joining both similar and From tabulated weldability data compiled dissimilar materials in such industries as by The Welding Institute, England (Ref. 3), Table 1 was setup. For given combina­ the automotive industry. In the majority KEY WORDS of cases, there are no problems relating tions of incompatible materials, this table to different melting points, cast structure Friction Welding immediately provides some possible in­ at the interface, inclusions and microseg­ Incompatible Metals terlayers. Several other materials can also regation. One-hundred-percent efficient Brass to Copper be thought of as interlayers, but many welds are common (Ref. 1), and with Bronze to Steel can be ruled out as being ineligible due to some dissimilar materials, the strength Titanium to Nickel lack of certain physical properties and in­ obtained in the joint can be better than Interlayer Materials ability to create conditions conducive to the strength of the weaker material (Ref. Dissimilar Materials bonding. An ideal intermediate material 2). For these reasons and for those of Three-Element Weld should generally: Copper Interlayer manufacturing economics, friction weld­ A) Be able to create conditions ame­ Aluminum Interlayer ing is a preferred process when size and nable for metallic bonding to geometry of the workpieces permit. occur.

Incompatible materials Table 1—Incompatible Material Combinations, Type of Weld and Possible Interlayers

Type of Possible Material Combination Weld Formed Interlayer

Aluminum alloys/magnesium alloys No weld Aluminum Brass/copper No weld Aluminum Bronze/plain carbon steel No weld Aluminum, bronze(s Bronze/steel alloy No weld Aluminum Recess nterlayer Magnesium alloys/magnesium alloys No weld Aluminum Magnesium alloys/stainless steel No weld Aluminum Fig. 7 —Schematic of component arrangement Nickel/titanium No weld Aluminum in three-element friction welding Niobium/stainless steel No weld Aluminum Niobium/zirconium alloys No weld Aluminum Silver/titanium No weld Copper Plain carbon steel/titanium No weld Aluminum, copper F. SASSANI is an Assistant Professor and /. R. Plain carbon steel/tungsten carbide, cemented Brittle weld Aluminum NEELAM is a former Graduate Research Assis­ Stainless steel/titanium Brittle weld Aluminum tant, Department of Mechanical Engineering, Stainless steel/zirconium alloy Brittle weld Aluminum University of British Columbia, Vancouver, B.C., Canada. (a) The tests conducted resulted in brittle welds.

264-s I NOVEMBER 1988 Table 2—Composition of Materials Used in Table 3—Experimental Conditions and Results for Ti-Ni Combination with Cu Interlayer the Experiments Speed: 1120 rpm Material Composition (%) Friction Forge Friction Time Forge Time Pressure Pressure Ultimate Tensile Strength Aluminum 0.6 Si, 1 Mg, 0.25 Cu, (s) (s) X1000psi(MPa) X1000 psi (MPa) X1000 psi (MPa) (6061) 0.25 Cr, Al (balance) 7 7 5.0 (34.47) 10.0 (68.95) Failed on dropping Brass 65 Cu, 35 Zn (yellow 5 7 5.0 (34.47) 10.0 (68.95) Failed by slight bending brass) 5 7 6.25 (43.1) 10.0 (68.95) 9.38 (64.70) Copper 99.9 Cu (commercially 5 7 6.25 (43.1) 12.5 (86.2) 19.65 (135.48) pure) 5 14 6.25 (43.1) 12.5 (86.2) 4.13 (28.53) Naval 60 Cu, 0.75 Sn, 5 7 6.25 (43.1) 10.0 (68.95) 2.78 (19.16) brass 29.25 Zn 3 7 6.25 (43.1) 10.0 (68.95) 4.45 (30.69) Nickel 99.9 N (commercially 3 7 6.25 (43.1) 10.0 (68.95) 11.07 (76.32) pure) 5 7 6.25(43.1) 10.0 (68.95) 6.85 (47.29) Phosphor 90 Cu, 10 Sn bronze Steel 0.25 C, 0.45 Mn, (1025) 0.04 P (max), Table 4—Experimental Conditions and Results for AI-AI Combinations 0.04 S (max) Titanium 99.9 Ti (commercially pure) Speed: 1120 rpm, No Interlayer Friction Time Forge Time Friction Pressure Forge Pressure Ultimate Tensile Strength (s) (s) X1000 psi (MPa) X1000 psi (MPa) X1000 psi (MPa) 3.0 10.0 6.25(43.1) 7.5 (51.71) 26.41 (182.14) B) Be compatible with both of the 3.5 10.0 6.25 (43.1) 10.0 (68.95) 25.79 (177.82) base materials. 3.5 10.0 6.25 (43.1) 17.5 (120.66) 28.47 (196.30) C) Have the passivity to the formation 3.5 10.0 6.25 (43.1) 17.5 (120.66) 27.13 (187.06) of brittle intermetallics at the tem­ peratures encountered. Speed: 1120 rpm, Al Interlayer 4.0 10.0 D) Flow at the same temperature and 6.25 (43.1) 10.0 (68.95) 26.41 (182.14) pressure ranges as those of the 3.5 10.0 6.25 (43.1) 12.5 (86.2) 25.79 (177.82) 3.0 10.0 6.25 (43.1) 12.5 (86.2) 28.47 (196.30) base materials. 3.0 10.0 6.25 (43.1) 3.75 (25.85) 8.41 (57.99) With regard to item A, some materials, such as brass, flake away in the form of chips and thus adhesion, alloying or con­ tinuous plastic deformation does not take overview of the interaction between dif­ pursued further with incompatible mate­ place. In such cases, conditions for metal­ ferent material combinations and set the rials, experiments were conducted on lic bonding do not occur. groundwork for main experiments. For other materials that were generally con­ example, because of similar melting sidered to be suitable for the friction points and the base materials' close welding process. The first set of experi­ Qualitative Considerations resemblance, copper was expected to ments involved joining pairs of aluminum These four properties set the guide­ form a good metallic bond with its own alloy (6061) specimens with an interlayer lines for the selection of a material as an alloy brass, but the experiments proved of the same alloy. The properties of the interlayer. Materials such as lead or alumi­ otherwise. Another example was steel joints obtained were similar in all respects num are immediately excluded when (melting point 1450°C/2642°F) when to those of the conventional two-piece tungsten is welded, for they cannot cre­ used as an interlayer between nickel (mp joints. Table 4 shows the results for both ate conditions to promote metallic bond­ 1455°C/2651°F) and titanium (mp cases. Even after the flash was turned ing, since they flow at much lower stress­ 1820°C/3308°F). The conditions re­ down, there were no indications of their es and temperatures than tungsten does. quired for metallic bonding did not mate­ being three-element specimens. When In other cases, however, the reasons for rialize and it was titanium, with a higher subjected to a tensile test, the specimens elimination are not so evident. For exam­ melting point, that was completely failed across the interface after necking in ple, bronze may have been considered extruded away because of its low hot a ductile mode of fracture. In several as a suitable intermediate material strength. cases, such as shown in Fig. 2, the failure between copper and brass (which do not For the same material combination, took place outside the weld plane. In the weld) as it has a melting point close to the copper, with a lower melting point of next set of preliminary experiments, an melting point of the two. However, the 1083°C (1981 °F), was used as the inter­ intermediate material dissimilar to one of results of the preliminary experiments layer instead of steel. While copper the base materials was used. In two indicated that this reasoning was not formed a good weld with titanium by the different sets of experiments, 1025 plain entirely adequate. conventional friction welding process, it carbon steel was friction welded to 6061 did not form a satisfactory weld with it in aluminum with aluminum and steel as the three-element method. Experimental interlayers, respectively. The aluminum Preliminary Experiments conditions and results of the tensile tests interlayer produced a joint that failed adjacent to the steel interface, but not in For all the experiments conducted in are given in Table 3. The inconsistency of the plane of the weld. The failure was this work, the specimens were 0.5 in. the ultimate tensile strength is apparent, similar to that of a two-piece joint. The (12.7 mm) in diameter. The interlayers particularly with multiple experiments steel interlayer, however, did not were 0.04 in. (1 mm) thick and 0.4 in. (10 under identical conditions. Under the ten­ produce a good metallic bond for it did mm) in diameter. The compositions of sile test, the specimens always failed or not adequately interact at the steel-to- the materials and the interlayers used are separated at the copper/titanium inter­ steel interface to create the required given in Table 2. The purpose of the face. conditions for bonding. The third trial preliminary experiments was to obtain an Before the three-element method was

WELDING RESEARCH SUPPLEMENT 1265-s 350 o- 50 F o o o Without Interlayer : 45 - With Copper Interlayer UJ 300 cc

ct 40

250 IS, 35 z

200 5 30 ' 2 t^)_ 8 10 12 14 FORGE PRESSURES I000PSI) I ' 40 60 80 100 FORGE PRESSURE (MPa) Fig. 2-A three-element friction welded specimen (all aluminum) failed Fig. 3 — Variation of UTS versus forge pressure for bronze/steel in tension in the base material. The weld plane is at the middle of the combination welded at 3600-psi (25-MPa) frictioning pressure and 3-s machined recess frictioning time

used a totally dissimilar material as the Brass/Copper these copper patches had been through interlayer. In this case, nickel was used a ductile mode of fracture, as numerous Three materials were thought of as between two steel rods. During the small dimples could be seen. The rest of interlayers for this combination: alumi­ experiments, most of the nickel was usu­ the area had failed largely by interface num, phosphor bronze and naval brass. ally expelled. These specimens, under the separation. However, a 1000X magnifi­ The experiments with the first two were cation of this area showing some patches tensile test, failed along the weld plane unsuccessful. Although aluminum welds of localized ductility indicated that a fair after necking at a strength of almost well with copper (Ref. 9) in conventional amount of metallic bonding had also equal to that of a pure steel specimen. friction welding, it did not work as an taken place. General observation was Tables 3 and 4 show representative interlayer (Ref. 10). The remaining experi­ that in these areas the base copper was results from preliminary experiments. ments were carried out with naval brass. the weakest part. The naval brass interlayer between cop­ Main Experiments per and brass resulted in joints of a tensile Some problems, such as inconsistency strength as high as 30,000 psi (207 MPa). of the weld results, were encountered After the preliminary experiments, the In the tensile tests, the failures took place with this combination of materials, as is feasibility of the method was established, mostly at the brass/naval brass or the evident from Table 5. One factor was and further experiments were carried out naval brass/copper interface, and oc­ thought to be the randomly varying rela­ with incompatible material combinations. curred indiscriminately at both interfaces. tive motions between the three ele­ Of the several combinations of materials Table 5 shows some results under differ­ ments, which resulted in an uneven rate that do not weld, three combinations ent experimental conditions. A closer of heat generation. Due to this uneven were selected as base materials for these look at the failed surfaces showed that rate of heat input, the amount of melt-off experiments: brass/copper, bronze/steel the failure had taken place largely by for each cycle varied so that workable and titanium/nickel. interface separation. In a particular case conditions for welding this combination From Table 1, aluminum was consid­ which failed at the copper/naval brass of materials could not be established with ered a suitable element for use as the interface, some copper patches were certainty. intermediate material since it forms a exposed on the fracture surface especial­ good weld with each of the six materials. ly towards the periphery. Examination Bronze/Steel In addition to aluminum, several other under the scanning electron microscope Welding experiments were performed materials that were not evident from (SEM) showed that the failure across Table 1 were also used, and results of the with 1025 steel and bronze specimens. experiments are discussed later. While it is accepted that the tensile test does not evaluate a specimen's impact Table 5—Experimental Conditions and Results for Brass-Copper Combination strength or fatigue properties (Refs. 4, 5), it was used as the sole quantitative crite­ Speed: 1120 rpm: Naval Brass Interlayer rion test for all the friction welded speci­ Friction Forge mens, though occasional qualitative Friction Time Forge Time Pressure Pressure Ultimate Tensile Strength bending tests were also performed. (s) (s) X1000 psi (MPa) X1000 psi (MPa) X1000 psi (MPa) More advanced in-process and nonde­ 7 10 6.25 (43.1) 10.0 (68.95) 22.70(156.51) structive testing of friction welded joints 7 10 6.25 (43.1) 10.0 (68.95) 15.10 (104.11) has been under investigation (Refs. 6-8) 7.5 10 6.25(43.1) 10.0 (68.95) 10.69 (73.70) and will become more common in the 8 10 6.25 (43.1) 10.0 (68.95) 10.70 (73.77) coming years. In all the experiments, the 8 10 6.25 (43.1) 11.25 (77.57) 8.26 (56.95) process parameters were manipulated to 8 10 5.0 (34.47) 11.25 (77.57) 19.43 (133.90) improve (not necessarily maximize) the 9.5 10 5.0 (34.47) 11.25 (77.57) 13.58 (93.63) tensile strength. 10.5 10 5.0 (34.47) 11.25 (77.57) 30.29 (208.85)

266-s | NOVEMBER 1988 The welds obtained were brittle and the experiments was titanium and nickel. weak in tension and bending. An interlay­ An attempt to weld these two materials Table 6—Experimental Conditions for Results Shown in Fig. 3 er of copper was used, and the improve­ directly resulted in a joint of a very low ment in the tensile strength was substan­ strength. Even though titanium has a tial. The rates of heat generation and much higher melting point than nickel, it Without With Interlayer Interlayer melt-off were more uniform with this starts to flow at a much lower tempera­ material combination than they were ture. The result is a complete extrusion of Friction time (s) 4 5 with brass and copper. The variation in titanium without any reasonable joint Friction-pressure 7.5 (51.7) 8.75 (60.3) the melt-off was less than 0.0625 in. (1.6 being formed. X1000 psi (MPa) mm). The metallic bond strength also Welding was attempted between the Forge time (s) 10 Speed, rpm 1120 responded to the change in one of the two with several interlayer materials. process parameters, the forge pressure. Although steel appeared to be a reason­ The tensile strength increased with an able candidate due to its strength, melting initial increase in the forge pressure and point, etc., it did not form a good metallic then decreased, as shown in Fig. 3. The bond with titanium. Titanium also flows at Table 7—Experimental Conditions for Results Shown in Fig. 6 strength of the joints made without an much lower pressures than steel does. interlayer was less affected by forge Copper by itself welds well with titanium, pressure. The experimental conditions but the joints obtained between titanium Friction time (s) 15 are given in Table 6. and nickel with the copper interlayer Friction-pressure 7.5 (51.7) The tensile test specimens with copper failed at the copper/titanium interface in X1000 psi (MPa) interlayer separated at the interface with­ the tensile tests. This combination has not Forge time (s) 15 out necking and the maximum ultimate yet been fully investigated. However, Speed, rpm 1120 tensile strength obtained was 47,770 psi failure at the copper/titanium interface is (330 MPa), about a 40% improvement thought to have been caused by the over the strength obtained without an presence of brittle intermetallic com­ increase in the frictioning time, the interlayer. The failures that appeared to pounds. The best results obtained for this strength peaks at a frictioning time of 2 s, have taken place in the copper interlayer incompatible nickel and titanium combi­ followed by a sudden drop, and then left a layer of copper on both surfaces. A nation were with the use of aluminum as there is a gradual increase in strength low magnification of the steel specimens the interlayer. Table 7 shows the experi­ rising to give another peak of the UTS at showed a general tendency for interface mental conditions when the aluminum 15 s, the maximum frictioning time avail­ separation, especially at the center of the interlayer was used. The most influential able with the present setup of the specimen where the welding is usually parameters in this case were the forge machine used. pressure and the frictioning time. Figure 6 worst —Fig. 4. A 200X magnification (Fig. The reason for this variation in strength shows that when the forge pressure was 5) shows some areas with original is due to a number of factors: increased from 5000 psi (34 MPa) to machining marks present on the 1) Mode of bonding 16,000 psi (110 MPa) the ultimate tensile unbonded steel surface and adjacent a) Mechanical interlocking strength increased from about 20,100 psi dimpled areas. The dimpled areas are b) Metallic bonding (139 MPa) to 41,800 psi (288 MPa). The macroscopic evidence of fracture in the 2) Thickness of the interlayer (not copper metallic bonding layer. The experimental conditions and the plot between the ultimate tensile strength for investigated) bronze counterpart surface showed 3) Physical properties of interlayer more uniform bonding all over the sur­ tests of different frictioning times are shown in Table 8 and Fig. 7, respectively. 4) Extent of the formation of brittle face, and ductile fracture had taken place intermetallic phases. both in the copper and in the bronze. For each set of experimental conditions, three tests were performed. Every data The mode of bonding is purely point on Fig. 7 indicates the maximum, mechanical in the initial stages as the asperities become interlocked and some Titanium/Nickel minimum and the average value of the ultimate tensile strength. With the of the harder material plows into the The third combination of materials for softer. This was noticed by the fact that

Fig. 4 —Failed steel surface from bronze/copper/steel combination Fig. 5-200X of outlined area in Fig. 4 showing flow lines on the steel showing mottled copper over some areas (40X) surface and fine dimples on the copper layer

WELDINC RESEARCH SUPPLEMENT | 267-s 250 | 350

K 300 o HJ 200 tu 250 cc I- M 200 LU _l 150 w 150 z LU *- 100 LU I- 8 10 12 14 16 2 50 FORGE PRESSURES IOOO PSI) L I I I 0 ± 0 >-»- 3 6 9 12 40 60 80 100 120 FORGE PRESSURE (MPa) FRICTIONING TIME (sec.) Fig. 6 — Variation of UTS versus forge pressure for titanium/nickel Fig. 7 — The effects of frictioning time for a titanium-aluminum-nickel combination with an aluminum interlayer. Frictioning pressure was 4000 joint welded at 4000-psi (27.5-MPa) frictioning pressure and 16,000-psi psi (27.5 MPa) with frictioning time of 15 s (110-MPa) forging pressure when the specimens welded at around 2 being welded, a possibility exists for the which adversely affects fatigue and frac­ s of frictioning time were subjected to formation of intermetallic compounds ture toughness. the tensile test, clickings were heard as (Ref. 4). The intermetallics in general are the separation of the interlockings took brittle and are weak in tension, fatigue Extended Analysis place. Titanium, like aluminum, has a very and bending properties. The formations stable oxide coating which generally pre­ and growth of these phases can be In the titanium-aluminum-nickel joint, vents further oxidation and protects it controlled by varying the friction welding there is a possibility of diffusion of alumi­ from the environment. With longer fric­ process parameters, such as decreasing num into nickel or titanium, titanium into tioning time, it is thought that more oxide the speed and increasing the axial pres­ aluminum or nickel into aluminum. is displaced, thereby exposing unoxidized sure during the frictioning period, etc. An electron microprobe analyser was titanium and aluminum and encouraging The rate of heat input is higher at lower used on longitudinally cut and polished metallic bonding to occur. speeds over a certain minimum limit. welded specimens to determine the inci­ When subjected to tensile tests, most Similarly, by having large loads, the heat dence and, if any, the extent of diffusion. of the failures occurred by interface sep­ input rate can be increased to reduce the Analysis was done on three types of aration either at the nickel or the titanium process time, but both of these have joints that arise from this combination junctions with the aluminum interlayer. limitations. Low speeds generate high (Fig. 7): Type I joints, seemingly good Some traces of aluminum always adhered torques, which can lead to work-holding joints resulting at low frictioning time of to the base material's surface where the problems. Too high friction pressure about 2 s; Type II joints, bad joints interface separation had taken place. As results in high rates of melt-off, an unnec­ occurring at times greater than 2 s, with the frictioning time is quite substantial (15 essary extrusion of metal that can cause these particular joints obtained at around s) for the 0.5-in. diameter specimens localized, unrecrystallized plastic damage, 6 s; and Type III joints, good joints obtained at about 15 s of frictioning time. Fig. 8 —A nickel scan for a Type II joint The concentration profile at the alumi­ over a spotted area num-nickel and aluminum-titanium inter­ suspected to be face gave a sharp transition without any nickel rich. The dark indication of movement of aluminum into area on the top-right either of the two base materials. This was portion is nickel the case with all the joints, Types I, II and (1200X) III. As expected, no diffusion of either nickel or titanium into aluminum was observed for Type I joints. But in Type II joints there was some evidence of move­ ment of nickel into aluminum, while titani­ um showed no diffusion at all. An

Table 8—Experimental Conditions for Results Shown in Fig. 7

Friction-pressure 7.5 (51.7) X1000 psi (MPa) Forge time (s) 15 Forge pressure 12.5 (86.1) X1000 psi (MPa) Speed, rpm 1120

268-s [ NOVEMBER 1988 Fig. 9 — Nickel scan for a Type 111 joint at a region close to the center atFig. 10 — Nickel scan for a Type 111 joint at a region close to the periphery. 200X. Nickel is seen as the dark area Numerous dark regions indicate extensive movement of nickel into aluminum at long frictioning times absorbed electron image of this joint is to the periphery, the spots were more enhancement using a computer program shown in Fig. 8, the image is magnified widely dispersed and were globular. The MAGIC (microprobe analyzer general 1200 times. Figure 8 also shows the absorbed electron image, with the con­ intensity corrections). Since only two ele­ microprobe scan of the Ni peak. A high centration profile as the scan passes over ments were being analyzed and were Ni response is apparent in the dark upper some of these spots, is given in Fig. 10, predominant, the concentration of alumi­ corner, decreasing rapidly to zero at the which shows numerous small peaks all of num was determined by difference. nickel-aluminum interface. Localized re­ them having some nickel. A light-shaded region was also found gions of high Ni are apparent within the A quantitative analysis of the electron along the nickel-aluminum interface, but aluminum layer, probably indicative of microprobe results provided approxi­ this region could not be analyzed due to the formation of a nickel-aluminum inter­ mate stoichiometric compositions of the its thinness (approximately 1 to 2 microns metallic compound. All the small dark nickel-rich region within the aluminum thick) and the difficulty in obtaining areas are nickel-rich regions though the layer. For the sake of simplicity, the count counts accurately because of substantial occurrence of spots of the size showing taken was assumed to be of 100% alumi­ difference in the atomic weights of the the peak were not as common as the num, even though the aluminum used elements being analyzed. smaller spots. had alloying elements (1% silicon, 0.6% The Type III, aluminum-nickel junction magnesium). The nickel and aluminum Conclusions also showed evidence of nickel diffusion intensities were taken at different posi­ into the aluminum layer. The shape and tions within the aluminum layer under a A method of using an intermediate the location of this transfer varied along steady operating condition of the equip­ layer of a same or different composition the radius. At locations nearer to the ment. The results are shown in Table 9. in the conventional friction welding pro­ center (Fig. 9), the nickel-rich regions The concentration percentages were cess was investigated for both compati­ were of plate-like structure and were obtained after atomic number correction, ble and incompatible materials. Com­ close to the interface. At locations nearer absorption correction and fluorescent pared to the two-element method, the

Table 9--Quantitative Analysis of the Nickel and o a Polished Type III Specimen at Various Locations

Spot Weight Atomic Unknown Standard Unknown Standard Number Element Percent Percent K-Ratio Intensity Intensity Background Background Iterations

1 Ni (Ka) 73.33 55.82 0.7203 56829 78896 109 189 2 Al(a) 26.67 44.18 2 Ni (Ka) 57.88 38.71 0.5269 44411 78896 109 189 2 Al(a) 42.12 61.29 3 Ni (Ka) 78.48 62.63 0.7735 61026 78896 109 189 2 Al(a) 21.52 37.37 4 Ni (Ka) 69.77 51.47 0.6838 53937 78896 109 189 2 Al(a) 30.23 48.53 5 Ni (Ka) 57.20 38.05 0.5560 43870 78896 109 189 2 Al(a) 42.80 61.95 6 Ni (Ka) 39.97 23.43 0.3842 39314 78896 109 189 2 Al(a) 60.03 76.5

(a) Determined by difference.

WELDING RESEARCH SUPPLEMENT 269-s three-element method resulted in 100% gram. Further analysis is required to introduction to the process. Exploiting friction efficient welds (in terms of tensile determine the effect of these phases on welding in production, ed. E. D. Nicholas, pp. strength) when welding aluminum to alu­ the strength of the joint. Future studies 2-9. The Welding Institute, England. minum with a similar interlayer, and over are directed toward determining the 4. Wallach, E. R. 1978. Intermetallics in sol­ 95% when welding steel to steel with a functional relationships between material id-phase welds. Proc. 4th Intl. Conf. on Advancesin Welding, pp. 11-21. The Welding nickel interlayer. properties, process parameters and the Institute, England. Attempts to weld a number of incom­ strength of the welded joints. 5. lessop, |. |. 1978. Friction welding dissim­ patible material combinations using sev­ This study established the feasibility ilar metals. Proc. 4th Intl. Conf. on Advances in eral intermediate materials achieved dif­ and practicality of the three-element fric­ Welding, pp. 23-36. The Welding Institute, ferent degrees of strength. A 40% tion welding of incompatible materials. England. improvement in strength was observed 6. Wang, K. K., and Ahmed, S. 1976. Ultra­ for joints obtained between bronze and Acknowledgments sonic detection of weld strength for dissimilar- steel using a copper interlayer. The metal friction welds. Proc. 4th North American This work was supported by the Natu­ strength of the joint obtained between Metalworking Research Conf, ed. T. Altan, ral Sciences and Engineering Research the incompatible combination of titanium pp. 384-389, SME, Detroit, Mich. Council of Canada. Assistance of Profes­ and nickel with an aluminum interlayer 7. Wang, K. K., Reif, C. R„ and Oh, S. K. sor B. Hawbolt of the Department of 1982. In-process quality detection of friction was stronger by 13% than the joint Metals and Materials Engineering, U.B.C., welds using acoustic emission techniques. obtained between the interlayer material and permission to use the SEM facility are Welding journal 61(7):312-s to 316-s. to itself. The variation of the strength of appreciated. 8. Oh, S. K., Hasui, A., Kunio, T., and Wang, this joint with friction time was attributed K. K. 1982. Effects of initial energy on acoustic to different metallic bonding mecha­ emission relating to weld strength in friction References nisms. welding. Trans. Japanese Welding Society In the titanium-aluminum-nickel combi­ 1. Hazlett, T. N. 1962. Properties of friction 13(2):15-26. nation, a microprobe analysis showed welded plain carbon and low alloy steels. 9. Nicholas, E. D. 1975. Friction welding of movement of nickel into aluminum when Welding Journal 4X2): 49-s to 52-s. copper to aluminum. Metal Construction 7(3):135-141. the frictioning time was long. The nickel- 2. Jennings, P. 1970. Some properties of dissimilar metal joints made by friction weld­ 10. Neelam, J. R. 1984. A study of interlayer rich areas in the aluminum layer, when ing. Proc. of The Conference on Advances in assisted friction welding of incompatible analyzed quantitatively, showed the Welding Processes, pp. 147-1152, The Weld­ materials. M.A. Sc. Thesis. University of British presence of several phases, most compo­ ing Institute, England. Columbia, Canada. sitions being in agreement with phases 3. Nicholas, E. D. 1979 (with revised 11. Hansen, M. 1958. Constitution of Binary shown on the aluminum-nickel phase dia­ amendment chart 1983). Friction welding: an Alloys. McGraw-Hill.

WRC Bulletin 329 December 1987 Accuracy of Stress Intensification Factors for Branch Connections By E. C. Rodabaugh

This report presents a detailed examination of the stress intensification factor (SIF) formulations for perpendicular branch connections that are specified in American standard codes for use in the design of industrial and nuclear Class 2 and 3 piping systems. Publication of this report was sponsored by the Subcommittee on Piping, Pumps and Valves of the Pressure Vessel Research Committee of the Welding Research Council. The price of WRC Bulletin 329 is $20.00 per copy, plus $5.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, Suite 1301, 345 E. 47th St., New York, NY 10017.

WRC Bulletin 333 May 1988 Bibliography on Fatigue of Weldments and Literature Review on Fatigue Crack Initiation from Weld Discontinuities By C. D. Lundin The bibliography together with a review of the present state of assessment of the factors which affect fatigue crack initiation make up this document. The bibliography was compiled through the efforts of many students at The University of Tennessee utilizing the previously available bibliographies and computer searches.

Publication of this report was sponsored by the Subcommittee on Failure Modes in Pressure Vessel Components of the Materials and Fabrication Division of the Pressure Vessel Research Committee of the Welding Research Council. The price of WRC Bulletin 333 is $20.00 per copy, plus $5.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, Suite 1301, 345 E. 47th St., New York, NY 10017.

270-s | NOVEMBER 1988