Technical Note: Microstructural Evolution During Inertia Friction Welding of Austenitic Stainless Steels
BY J. C. LIPPOLD AND B. C. ODEGARD
associated microstructural changes which 1.27 cm (0.5 in.) in diameter were cut Introduction occur during the inertia friction welding from 1.27 cm (0.5 in.) bar stock. The Inertia friction welding is a solid-state of 300-series stainless steels. Future publi blanks were machined such that the fay welding process, which utilizes the fric cations are expected to report the effects ing surfaces were perpendicular to the tional heat generated between a rotating of composition, microstructure, and specimen axis with a surface finish of 32 /i and a stationary workpiece to produce a welding variables on the strength and in or better. metallurgical bond. Little or no melting fracture behavior of a variety of inertia The welding conditions are summa occurs along the weld interface and this friction welded austenitic stainless steels. rized in Table 1. Samples were cleaned in makes the process extremely attractive acetone prior to welding in order to where difficult-to-weld or dissimilar met Experimental Procedure remove any residual machining lubri als are to be welded. The likelihood of cants. The specimen pairs were mea solidification discontinuities is minimized A free-machining grade of austenitic sured prior to welding and again follow or completely eliminated. Other advan stainless steel — Type 303S — was selected ing welding in order to determine the tages of inertia friction welding include to study the microstructural response of total axial upset. reduced development and fabrication austenitic stainless steels during inertia Following welding, representative sam time, lower cost, and increased repro friction welding. The selection of this ples were sectioned axially and metallo ducibility. In addition, in similar metal alloy was based on several considera graphicaliy prepared. A mixed acid welds the joint preparation is minimal tions: etchant containing equal parts of nitric, because the process is essentially "self- 1. The as-received material is heavily hydrochloric, and acetic acid was used to cleaning"; the metal flow directed radially banded and contains sulfide stringers reveal microstructural details. outward along the weld interface along these bands. Since the sulfide parti removes any contaminants associated cles are relatively stable to almost the Mechanical Testing with the initial faying surfaces. melting point of the alloy, they form a Despite its advantages, inertia friction "fingerprint" of the metal flow which Smooth and notched tensile samples welding has not been widely accepted as occurs during welding. were machined from the welded blanks. a viable replacement for fusion welding. 2. With the exception of the sulfur The diameter of the smooth bar gage It is instead, viewed as a "last-ditch- content, the composition of Type 303S section was 4.95 mm (0.195 in). Two effort" process which is used only when stainless steel is equivalent to that of Type notch depths were used, including 1.23 conventional methods fail. This may be 304, a commonly used commercial stain mm (0.048 in.) and 0.825 mm (0.032 due in part to the requirement for sym less steel. in.). metry of the rotating member. However, 3. The alloy is normally unweldable This procedure made it possible to many conventionally welded joints are, in using fusion welding techniques. It was of measure weld strength as a function of fact, symmetric and capable of being interest to determine if Type 303S stain radial distance from the center of the inertia friction welded. less steel could be joined by inertia fric welded blanks. The strength may not be Although several review articles have tion welding. uniform, since the surface velocity addressed the process variables and The (wt.-%) chemical composition of decreases to zero at the axis of rotation. associated temperature distribution dur the alloy is 18.25 Cr, 9.03 Ni, 1.70 Mn, The tensile samples were tested at a ing inertia friction welding (Refs. 1-5), the 0.56 Si, 0.04 C, 0.03 N, 0.02 P, and 0.17 S. crosshead velocity of 0.02 mm/s (0.05 metallurgy of the process has received Weld blanks 5.08 cm (2.0 in.) long and ipm). little attention. If inertia friction welding is to gain wider use as a replacement for fusion welding, the metallurgical re Table 1—Inertia Friction Welding Conditions'3' sponse of materials during inertia friction welding must be better understood. 2 2 2 This report focuses on the characteris Spindle moment of inertia: 1.5 lbm-" (6.32 X 10" kg m ) Spindle rotational speed: 6600 RPM tics of metal flow, localized heating, and Axial force: 6000 lbf (26.4 kN) Energy:(b) 81.94 kj/in.2(1.27 J/m2) Average axial upset: 0.142 in. (3.61 mm) / C LIPPOLD and B. C. ODEGARD are with Sandia National Laboratories, Livermore, Cali (a 2 > ET (k|/in. ) • fornia. 4332 (cross sectional area)
WELDING RESEARCH SUPPLEMENT 135-s Results Bond Line Optical Metallography The inertia friction welding process for solid bars results in extensive metal flow radially outward along the weld interface. NERTIA WELD The photomacrograph shown in Fig. 1 reveals the nature of this flow near the Type 303S axis of the Type 303S weld. Note that the flow lines, as delineated by the banding in the base metal, bend sharply in the vicin ity of the weld interface. The rapid change in flow direction suggests that the temperature gradient in this region is relatively steep. As a result the elevated temperature regime in which plasticity occurs is narrow and effectively limits plastic flow to a narrow band in close proximity to the weld interface. Within that region flow occurs radially outward parallel to the weld interface resulting in the realignment of intermetallic stringers perpendicular to the original base metal orientation. At higher magnification (Fig. 2), the sulfide particles appear refined and homogeneous in contrast to the base metal structure. The combination of high temperature and plastic flow causes the elongated stringers (Fig. 2A) to fracture and partially dissolve, thus forming the low aspect ratio particles in the weld region (Fig. 2B). At the weld interface it is 3 likely that the sulfides melt and act as a "lubricant" on the faying surfaces. No metallographic evidence of sulfide melting was observed. However, the small amount of external flash produced at the outside diameter of the weld 0.4 mm suggests that the presence of liquid films during welding reduced the amount of Fig. 1 —Macrostructure along the weld Interface of an inertia friction weld in a solid bar of Type energy dissipated in the form of frictional 303S stainless steel. Note the rapid change in stringer orientation close to the weld interface heating at the mating surfaces. A microhardness traverse across the weld region at a position midway smooth bars. By comparison, reference between the center of the bar and the samples of Type 304L stainless steel have external surface (half-radius position) is exhibited notched-to-smooth strength shown in Fig. 3. The hardness drops ratios of 1.6 and 2.1 for the shallow and steeply in the region of the weld where deep notch, respectively (Ref. 6). metal flow is significant. Near the center These results indicate that there is little of the weld the hardness flattens out, change in weld strength as a function of corresponding to the homogeneous radial distance from the rotational axis. microstructural region shown previously in Fig. 2A. The hardness peak located ® , 40 wri precisely at the weld interface results from localized deformation which occurs during the final stages of the welding process. The nature of this deformation is discussed elsewhere in this report.
Tensile Test Results z Q .. . The results of both smooth bar and ' ;•' a in Table 2. Failure in the smooth gage tensile bars of 303S occurred at the weld 60 100 (B) I 40pm | interface after minimal elongation. Fur DISTANCE (mils) ther, the notched samples exhibited Fig. 3 — Hardness profile parallel to specimen Fig. 2 — Sulfide morphology in: A — base metal; almost no ductility and only a moderate axis midway between the rotational axis of the B — weld region increase in tensile strength above the bar and the outside surface 36-s | JANUARY 1984 Table 2—Tensile Test Results-*)- YS, ksi(MPa) UTS, ksi(MPa) Elong, % UTSN/UTSs Smooth bar 303S 60.0 (414) 95.3 (657) 4.6 - 304L 41.8 (288) 88.3 (609) 50 - Notch /\(c) 303S - 103.4(713) 0.5 1.08 304L — 142.1 (980) 9.5 1.61 Notch 6 From Fig. 1 it can be seen that near the the hardness "trough" is narrowest at the center of the bar the region of metal flow center of the bar and widest near the is confined to a narrow band adjacent to outside diameter. In addition, the hard the weld interface. Proceeding radially ness gradient is the steepest at the center, Fig. 4 —Fracture surface of smooth bar tensile outward from the specimen center the corresponding to the region where the sample: A-low magnification showing spiral region of plastic flow gradually broadens temperature gradient is the greatest and pattern; B — high magnification revealing and eventually merges with the weld the metal flow is essentially zero. The "woody" fracture morphology expulsion. The expulsion represents the softening results from the recovery and metal which has been extruded from the recrystallization that occurs at sub-solidus joint. Since the temperature gradient temperatures. This softening is mitigated Fracture Morphology along the axis varies as a function of to some extent by the work hardening promoted by the localized metal flow. A low magnification SEM fractograph position along the interface (Refs. 6-8) of the smooth bar fracture surface is (that gradient being greater near the The slight rise in hardness located at shown in Fig. 4A. A distinct spiral pattern center than the outside surface), the the weld interface in Fig. 3 results from is evident emanating from the center of width of the region in which metal flow deformation associated with the final rev the bar. At higher magnification (Fig. 4B), occurs will increase proportionally upon olution of the flywheel. The majority of fhe fracture surface appears to exhibit a moving radially outward from the cen metal flow occurs while the parts are "woody" morphology with no evidence ter. spinning against each other; however, as of the discrete sulfide particles observed The hardness drop along the weld the relative velocity drops the parts form metallographicaliy along the weld inter interface, shown in Fig. 3, is also a result a metallurgical bond and the remaining face (Fig. 2). of the temperature excursion in the vicin flywheel energy is dissipated as localized deformation along the weld interface. Closer examination at X6000 reveals ity of the weld interface. The hardness the presence of the sulfide inclusions. The behavior mirrors the onset of metal flow; The spiral pattern on the weld fracture fractograph in Fig. 5A shows that the surface in Fig. 4A provides a "fingerprint" sulfides are present as thin, lenticular of the metal flow which occurs in and particles superimposed on the fracture near the plane of the weld. Since one surface. part is spinning against the mating station In many cases, the particles were frac ary part, metal flow is not simply radially tured (arrows in Fig. 5A), probably during outward; it is also rotational about the the tensile test. The EDS spectrum in Fig. center of the specimen while flowing 5B, collected at a point corresponding to gradually toward the outside diameter. one of these particles, indicates that the The flow pattern in inertia friction welds particles are probably manganese-rich is geometry dependent; the spiral pattern sulfides. (Since the Cr K^ peak overlies the observed in Fig. 4A can be expected when welding solid bars since metal can Mn Ka peak, CrKjs peak is much higher than normal due to Mn enrichment of the not flow toward the center. When annu particle.) lar samples are welded, for example, metal flow will occur towards both the inside and outside diameter and the spiral Discussion pattern would not be formed. The banded structure of Type 303S The severe plastic deformation which free-machining stainless steel provides a occurs along the weld interface also has unique opportunity to study the nature of an effect on the aligned stringers. As the metal flow and microstructural develop metal turns and flows in the plane of the ment during inertia friction welding. By weld interface, intermetallic particles evaluating both metallographic cross sec making up the stringers are swept along tions parallel to the specimen axis and and subjected to high temperatures and fracture surfaces along the weld inter triaxial stresses. As a result of this thermo- Fig. 5—Sulfide particles on fracture surface: face, it is possible to construct a three- A — thin lenticular particles exhibiting second mechanical processing, the particles are dimensional picture of the weld interface ary cracking; B — EDS spectrum from one of broken up and deformed in the weld microstructural evolution. these particles (dark arrows) region. The original sulfide stringers WELDING RESEARCH SUPPLEMENT 137-s shown in Fig. 2B are squeezed flat and Deformation is restricted to a narrow tract Number DE-AC04-76DP00789. form the thin lenticular particles observed band along the weld interface whose on the fracture surface in Fig. 5. Since width is dictated by the local temperature References metal flow and, hence, particle motion gradient in that region. Recovery and 1. Oberle, T. L. 1970. Inertia welding. Metal along the weld interface is perpendicular recrystallization in the weld region Construction 2(5): 193-195. to the specimen axis, the sulfides become resulted in a significant drop in hardness 2. Wang, K. K., and Wen Lin. 1974. Fly aligned in an orientation transverse to the relative to the surrounding base metal. wheel friction welding research. Welding lour tensile axis. The combination of particle Tensile tests of the inertia friction nal 53(6):233-s to 241-s. shape, orientation, and the brittle nature welds in Type 303S stainless steel resulted 3. Anon. 1979. Inertia welding: simple in principle and application. Welding and Metal of the sulfide results in preferential failure in failure in the weld region with little Fabrication 47(8):585-590. at these sites and provides the "finger ductility. At low magnification the frac 4. Wang, K. K. 1975 (April). Friction weld print" of the metal flow pattern observed ture surface exhibited a spiral pattern ing. Welding Research Council bulletin no. in Fig. 4A. Such particles can and in this representing a "fingerprint" of the metal 204. case did influence the strength of the flow along the weld interface. At high 5. Kiwalle,). 1967. Flywheel friction welding weld. magnification, the fracture exhibited a as a design and production tool. ASTME paper Despite the inferior strength and ductil "woody" appearance with thin, lenticular no. AD57-198. ity of the Type 303S stainless steel inertia sulfide particles coating the surface. The 6. Cheng, C. ). 1962. Transient temperature friction weld, it was possible to produce a orientation and shape of the sulfide parti distribution during friction welding of two similar materials in tubular form. Welding lour metallurgical bond which was essentially cles result from the severe thermo- nal 41(12):542-s to 550-s. free of discontinuities. mechanical deformation which occurs 7. Cheng, C |. 1963. Transient temperature during the inertia friction welding pro distribution during friction welding of two Summary cess. dissimilar materials in tubular form. Welding lournal 42(5):233-s to 240-s. A heavily banded Type 303S free- A ckno wledgment 8. Wang, K. K., and Nagappan, P. 1970. machining stainless steel has been used to Transient temperature distribution in inertia reveal the pattern of metal flow during This work was supported by the U.S. welding of steels. Welding journal 49(9):419-s the inertia friction welding of solid bars. Department of Energy, DOE, under Con to 426-s. WRC Bulletin 284 April, 1983 The External Pressure Collapse Tests of Tubes by E. Tschoepe and J. R. Maison An experimental program was performed to confirm or refute the applicability of Figure UG-31 in Section VIII, Division 1 of the ASME Boiler and Pressure Vessel Code to the design of tubes under external pressure. Commercially available tubes were subjected to external pressure until collapse occurred. The data generated indicates the current ASME design rules for tubes under external pressure are suitable for continued application. Publication of this report was sponsored by the Subcommittee on Shells of the Pressure Vessel Research Committee of the Welding Research Council. The price of WRC Bulletin 284 is $12.00 per copy, plus $5.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, Room 1301, 345 East 47th St., New York, NY 10017. WRC Bulletin 283 February, 1983 A Critical Evaluation of Fatigue Crack Growth Measurement Techniques for Elevated Temperature Applications by A. E. Carden The report contains a discussion and evaluation of several crack length measurement techniques at elevated temperature and presents results from the experimental technique developed at the University of Alabama. Publication of this report was sponsored by the Subcommittee on Cyclic and Creep Behavior of Components of the Pressure Vessel Research Committee of the Welding Research Council. The price of WRC Bulletin 283 is $12.00 per copy, plus $5.00 for postage and handling (foreign + $8.00). Orders should be sent with payment to the Welding Research Council, 345 East 47th St., Room 1301, New York, NY 10017. 38-s | JANUARY 1984