Solidification and Liquation Cracking

Overview

Welding

Solidification and Liquation Cracking Issues in Welding Sindo Kou

Solidification cracking can occur in the fusion zone during the solidification of the weld metal. Liquation cracking, on the other hand, can occur in the partially melted zone during the solidification of the liquated material. These two types of cracking are reviewed in this article, including the factors that affect cracking and the remedies. INTRODUCTION Solidification cracking can occur in the fusion zone and liquation cracking can occur in the partially melted zone during welding. During the welding of an alloy, the weld pool is surrounded by semisolid materials (Figure 1).1 For convenience, circular-patch welding is used as an example here. The workpiece shown in Figure 1 consists of an inner piece in the form of a circular patch and an outer piece, which is held tightly to the fixture with bolts (not shown) that go through the holes. The pool boundary is the liquidus temperature of the weld metal during welding. Behind the weld pool is the solidifying material called the mushy zone, which is a semisolid consisting of dendrites (S) and the interdendritic liquid (L). Behind the mushy zone is the completely solidified material called the fusion zone or the weld metal. In front of and on the sides of the weld pool is a semisolid consisting of partially melted base-metal grains (S) and the intergranular liquid (L). Here, the temperature is above the eutectic temperature of the base metal (or the solidus temperature if the workpiece is solutionized to dissolve eutectic completely before welding). Behind the partially melted grains on the pool sides is the completely solidified material called the partially melted zone (PMZ). 2003 June • JOM

Figure 2a and b shows examples of solidification cracking and liquation cracking in circular-patch welds of aluminum alloys. As shown, solidification cracking occurs more or less along the centerline of the weld, while liquation cracking takes place along the outer edge of the weld. These two types of cracking will be discussed in this article.

between grains.8 When localized tensile stresses/strains in the material exceed its resistance to cracking, solidification cracking occurs. ASSESSING SOLIDIFICATIONCRACKING SUSCEPTIBILITY Several tests have been used to evaluate susceptibility to solidification cracking,9,10 including the Houldcroft

SOLIDIFICATION CRACKING Solidification cracking, as shown in Figure 3a, is intergranular. The fracture surface often reveals the dendritic morphology of the solidifying weld metal.1 These two characteristics suggest that solidification cracking occurs during the terminal stage of solidification, where dendrites have grown nearly fully into grains that are separated from one another by a small amount of liquid in the form of grain-boundary films. At this time, the weld metal can be rather weak and thus susceptible to cracking in the presence of tensile stresses/strains. Tensile stresses/strains can be induced in the weld metal if it cannot contract freely during cooling, for instance, in a highly restrained workpiece such as in circular-patch welding (Figure 1). The weld metal tends to contract during cooling because of solidification shrinkage and thermal contraction. Various theories of solidification cracking have been proposed.2–7 According to Borland,8 there are essentially three solidification cracking theories—the shrinkage-brittleness theory, the strain theory, and the generalized theory that includes the relevant ideas from the first two theories. According to the generalized theory, cracking can take place in a material in which continuous liquid films separate grains or in which some solid-solid bridges exist

Figure 1. Semisolid materials, namely, partially melted base-metal grains and the mushy zone, around the weld pool of an alloy during welding (for instance, circular-patch welding).

37

The crack length serves as the index of cracking sensitivity.

solidification cracking

FACTORS AFFECTING SOLIDIFICATION CRACKING

a

b 5 mm Figure 2. Macrographs showing cracking in aluminum circular-patch welds: (a) solidification cracking; (b) liquation cracking.71

test,11–14 Varestraint test,15–18 circularpatch test,19–21 and Sigamajig test.22 The Varestraint test and the circular-patch test are used often and described briefly here. In the Varestraint test, developed by Savage and Lundin,15 an augmented tensile strain is applied to the test specimen by bending it to a controlled radius at an appropriate moment during welding. Both the amount of the applied strain and the crack length (either the total length of all cracks or the maximum crack length) serve as the index of cracking sensitivity. The specimen can also be bent transverse to the welding direction. This transverse Varestraint test16 may promote cracking inside the weld metal more than outside the weld metal. In the circular-patch test19–21 (shown in Figure 1), where the outer piece is restrained by bolts, the weld tends to contract upon cooling, inducing tension at its outer edge and compression at its inner edge. The tensile stresses and strains at the outer edge are in the direction transverse to the weld. This is why liquation cracking occurs along the outer edge of the weld (Figure 2b). 38

Several factors can affect the solidification cracking susceptibility of weld metals, including1 the solidification temperature range, the amount and distribution of liquid at the terminal stage of solidification, the primary solidification phase, the surface tension of the grain-boundary liquid, the grain structure, the ductility of the solidifying weld metal, and the tendency of weldmetal contraction and the degree of restraint. All these factors are directly or indirectly affected by the weld-metal composition. The first two factors are affected by microsegregation during solidification. Microsegregation in turn can be affected by the cooling rate during solidification. In some cases, the primary solidification phase can also be affected by the cooling rate (for instance, austenitic stainless steels). Generally speaking, the wider the solidification temperature range, the larger the area that is weak and susceptible to weld-solidification cracking. DuPont et al.23 showed that in niobiumbearing superalloys, the maximum crack length in Varestraint testing increases with increasing solidification temperature range. S and P can widen the solidification temperature range of steels, stainless steels, and nickel-base superalloys significantly, thus increasing their cracking susceptibility.24–27 In pure aluminum there is no soluterich liquid of a low freezing point at grain boundaries to cause solidification cracking. In highly alloyed aluminum, on the other hand, there is abundant solute-rich liquid of a low freezing point between grains to backfill and “heal” any incipient cracks that form.3 Somewhere in between these two cases there exists a composition where just enough solute-rich liquid can form thin continuous grain-boundary films to cause the maximum sensitivity to cracking. Figure 4a through d shows the effect of composition on the solidificationcracking sensitivity of several binary aluminum alloys.28–32 Figure 4e shows the crack sensitivity in pulsed-laser welding of Al-Cu alloys.33 Each curve shows a maximum crack sensitivity.

Dupont et al.23 observed in some niobium-bearing superalloys that the amount of the terminal liquid undergoing the L→γ + Laves reaction is small and remains isolated, thus not really contributing to solidification cracking. This type of morphology, unlike the continuous terminal liquid undergoing the same reaction in the more cracksusceptible alloys, should be more resistant to crack propagation throughout the mushy zone. For austenitic stainless steels, the susceptibility to solidification cracking is much lower when the primary solidification phase is δ-ferrite rather than austenite. 34–36 As the ratio of the chromium equivalent to nickel equivalent increases, the primary solidification phase changes from austenite to δ-ferrite and cracking is reduced. As shown by Takalo et al.,35 this change occurs at about Creq/Nieq = 1.5, where Creq = Cr + 1.37Mo + 1.5Si + 2Nb + 3Ti and Nieq = Ni + 0.3Mn + 22C + 14.2N + Cu. However, this value can increase with increasing cooling rate during welding because of undercooling, for instance, to Creq/Nieq = 1.6–1.7 in pulsed-laser welding.36 The effect of the surface tension of the grain-boundary liquid is such that the smaller the dihedral angle, the more the grain-boundary liquid tends to form continuous films on grains.37 Nakata and Matsuda16 showed that solidification cracking in several aluminum alloys increases with deceasing dihedral angle of the grain-boundary liquid. For aluminum alloys, welds with finer equiaxed grains were found less susceptible to solidification cracking.16,38–40 Fine equiaxed grains can deform to accommodate contraction strains more easily; that is, they are more ductile. Liquid feeding and the healing of incipient cracks can also be more effective in fine-grained material. In addition, the grain-boundary area is much greater in fine-grained material and, therefore, segregates of low melting points are less concentrated at the grain boundary. Nakata and Matsuda16 proposed that the ductility of the solidifying weld metal has a strong effect on solidification cracking. They used the transverse Varestraint test to determine ductility of the solidifying weld metal as a function JOM • June 2003

b 0.5 µm Figure 3. Micrographs showing cracking in aluminum welds: (a) solidification cracking; (b) liquation cracking.

of temperature. The resultant ductility vs. temperature curve is called the ductility curve, with the liquidus temperature TL as the upper bound and the temperature at the tip of the longest crack as the lower bound. The slope of a line passing through TL and tangent to the ductility curve is called the critical strain rate for temperature drop (CST), that is, the critical rate at which the strain varies with temperature drop. According to Nakata and Matsuda,16 CST correlates well with the cracking susceptibility of the weld metal. Both the tendency of weld-metal contraction and the degree of restraint are factors that can affect solidification cracking.1 As already mentioned, the weld metal tends to contract due to solidification shrinkage and thermal contraction. Austenitic stainless steels have relatively high thermal expansion coefficients (as compared to mild steels) and, therefore, are often prone to solidification cracking. The solidification shrinkage of aluminum is as high as 6.6%.5 The thermal expansion coefficient of aluminum is roughly twice that of iron-based alloys. As a result, many aluminum alloys are susceptible to solidification cracking. As mentioned previously, tensile stress/strains are induced in the weld metal when the weld metal cannot contract freely during welding. 2003 June • JOM

Relative Crack Sensitivity

a

Solidification cracking can be reduced by adjusting the weld-metal composition, the grain structure, and the welding condition.1 When welding aluminum alloys, it is desirable to have a weld-metal composition that is away from the peak of the crack sensitivity curve (shown in Figure 4). To reach the desired weldmetal composition, a filler metal of a proper composition can be used and the welding parameters can be selected to achieve the desired dilution ratio (the fraction of the base metal in the weld metal). The maps of crack sensitivity vs. composition for ternary aluminum alloys are also useful.31,41 When welding austenitic stainless steels, as mentioned previously, it is desirable to have a weld-metal composition such that the weld-ferrite content is at a level of 5–10% in order to avoid solidification cracking. Quantitative relationships between the weld-ferrite content and the weld-metal composition in austenitic stainless steels have been determined by Schaeffler,42 DeLong,43 Kotecki,44,45 Balmforth et al.,46 and Vitek et al.47,48 With the help of such quantitative information, proper filler metals and dilution ratios can be selected to insure enough weld-ferrite content. When welding carbon and low-alloy steels, it is desirable to have a weldmetal composition such that the weldmetal Mn/S ratio is sufficiently high to avoid solidification cracking. Increasing the Mn/S ratio promotes the formation of MnS rather than FeS. This is because the high melting point and the globular morphology of MnS tend to make sulfur less detrimental. At relatively low carbon levels the solidification-cracking tendency can be reduced by increasing the Mn/S ratio.49 However, at higher carbon levels (i.e., 0.2–0.3% C), increasing the Mn/S ratio is no longer effective. 50 In such cases lowering the weld-metal carbon content, if permissible, is more effective. The grains of the weld metal can be refined by inoculation and arc oscillation to reduce solidification cracking. Grain refining by inoculation has been achieved in steels and stainless steels

with titanium and aluminum nitride,51,52 and in aluminum alloys with titanium and zirconium.53–56 Dudas and Collins32 produced grain refining and eliminated solidification cracking in an Al-Zn-Mg filler weld by adding small amounts of zirconium to the filler metal. Solidification cracking has been reduced by the following means: Garland57 grain refined welds of aluminummagnesium alloys by mechanically vibrating the arc during welding; Kou and Le58 grain refined welds of 5052 aluminum by magnetic-arc oscillation (20 Hz); David and Liu59 refined iridium welds by magnetic-arc oscillation; and Tseng and Savage60 refined the solidification structure within the grains of HY-80 steel, though not the grains themselves. Kou and Le58,61,62 used low-frequency (1 Hz) magnetic-arc oscillation during gas-tungsten-arc welding of aluminum alloys and forced the weld-metal columnar grains to change their growth direction periodically. Solidification cracking was reduced effectively because a crack had to change its direction periodically in order to propagate down

Composition of Weld, % Alloying Element Total Crack Length (mm)

REMEDIES FOR SOLIDIFICATION CRACKING

Copper Content (wt%)

Figure 4. Effect of composition on crack sensitivity of some aluminum alloys. (a–d) from Reference 32; (e) from Reference 33. (Reprinted from Kou.1)

39

Figure 5. Mechanism of liquation cracking in the partially melted zone of a full-penetration aluminum weld.71

the weld along grain boundaries. In addition, favorable welding conditions can be used to reduce solidification cracking. For instance, reducing restraint helps reduce solidification cracking. Making the weld surface slightly convex can help prevent surface shrinkage cracks that tend to occur in welds with a concave surface.63 This is because the surface is less in tension when it is slightly convex. Reducing the weld depth-to-width ratio can help prevent centerline cracking that tends to occur in welds with a steep angle of abutment between columnar grains growing from opposite sides of the weld pool.63 This type of cracking is often observed in deep and narrow welds produced by electron beam or submerged arc welding. LIQUATION CRACKING Liquation cracking, as shown in Figure 3b, is intergranular just like solidification cracking. However, unlike solidification cracking, the fracture surface does not reveal any dendritic morphology. The presence of a liquid phase at the intergranular fracture 40

surface is sometimes evident,64 but not always.65,66 Liquation cracking occurs in the PMZ of a weld, which is immediately outside the fusion zone, because of heating above the eutectic temperature (or the solidus temperature if the workpiece is completely solutionized before welding) in the PMZ.1 Liquation can occur along the grain boundary as well as in the grain interior. Grainboundary liquation makes the PMZ susceptible to liquation cracking. Liquation cracking1,70 has also been called edge-of-weld cracking,32 basemetal cracking,67 hot cracking,68 and heat-affected zone cracking.69 The name liquation cracking is used here. Huang and Kou71 showed in Figure 5 the development of liquation cracking in the PMZ of a full-penetration aluminum weld. Since the PMZ is weakened by grain-boundary liquation, it can crack if the solidifying and contracting weld metal is stronger than it. Huang and Kou72 used curves of temperature (T) vs. solid fraction (fS) to help understand liquation cracking. Since the strength of a semisolid is

likely to increase primarily with the solid fraction, they proposed that if the weld-metal fS exceeds the PMZ fS during PMZ terminal solidification, the solidifying and contracting weld metal can become stronger than the PMZ it pulls, and liquation cracking is likely to occur if tensile stresses/strains and liquation are both significant in the PMZ. For example, Huang and Kou72 made full-penetration, circular-patch welds in 6061 aluminum with filler metals 5356 (Al-5Mg) and 4043 (Al-5Si). At the dilution ratio of about 65%, severe liquation cracking occurred in the weld made with 5356 but no cracking was observed in the weld made with 4043. Figure 6 shows the T-fS curves for both the weld metal and the PMZ at the fusion boundary, calculated using Pandat73 based on the multicomponent Scheil equation. The curves show that, at about 65% dilution, 5356 made the weld metal fS greater than the PMZ fS during PMZ terminal solidification while 4043 made the weld metal fS less than the PMZ fS throughout PMZ solidification. The susceptibility of the PMZ to hot cracking can be evaluated using several different methods, such as the Varestraint test,74–76 circular-patch test,65,71,72 and hot-ductility test.66 FACTORS AFFECTING LIQUATION CRACKING The liquation cracking susceptibility of the PMZ can be affected by the extent of liquation, the grain structure, the hot ductility, and weld-metal contraction and the degree of restraint.1 The greater the extent of liquation, the more likely that liquation cracking can occur. The extent of liquation in the PMZ depends on the tendency of the material to liquate and the level of heat input. The tendency of an aluminum alloy to liquate increases with increasing freezing temperature range and fraction of liquid during freezing. For instance, alloy 7075 liquates much more severely than alloy 6061 because the former has a wider freezing temperature range and a greater fraction of liquid during freezing.77 The tendency to liquate can also increase with increasing grainboundary segregation in the workpiece material. As-cast materials, for instance, JOM • June 2003

tend to have more grain-boundary segregation. The higher the heat input, the greater the liquation and the more likely liquation cracking can occur. The heat input depends on the welding process and parameters and the workpiece thickness. Miyazaki et al.76 observed in 6061 aluminum welds made with filler 5356 that liquation cracking increased with increasing heat input per unit length of weld. A base metal with coarser grains is more susceptible to liquation cracking. The coarser the grain, the less ductile the PMZ becomes. Furthermore, the coarser the grains, the smaller the grain-boundary area and hence, the more concentrated the impurities or low-melting segregates are at the grain boundary. Increasing liquation cracking with increasing grain size has been observed in aluminum alloys76,78 and Inconel 718.79,80 Hot ductility can also play a role in liquation cracking. Several different criteria have been used for interpreting hot-ductility curves. 81 It has been suggested that an alloy that recovers its ductility easily during cooling tends to be more crack resistant.66 Both the tendency of weld-metal contraction and the degree of restraint can affect the susceptibility to liquation cracking.1 The more the weld metal contracts, the more likely the PMZ will crack. The more severely the workpiece is restrained, the more likely liquation cracking will occur.

Katoh et al.,74 Kerr et al.,75 and Miyazaki et al.76 studied liquation cracking in 6000-series alloys including 6061. Longitudinal liquation cracking occurred when alloy 6061 was welded with a 5356 filler but not with a 4043 filler. As shown previously, Huang and Kou72 showed that T-fS curves can be used to help select the weld-metal composition to avoid liquation cracking. The sensitivity of the PMZ to liquation cracking can be reduced by decreasing the degree of restraint and, hence, the level of tensile stresses. The size of the PMZ and, hence, the extent of PMZ liquation can be reduced by reducing the heat input. In order to minimize liquation cracking, the heat input can be kept low by using multi-pass welding or low-heat-input welding processes (such as electron beam and gas-tungsten-arc welding) when possible. Liquation cracking can be reduced by selecting the proper base metal for welding if it is feasible. The basemetal composition, grain structure, and microsegregation can affect the susceptibility of the PMZ to liquation cracking significantly. Using base metals that have low sulfur and phosphorus can help reduce the freezing temperature range and, hence, liquation cracking. Using fine-grained materials helps reduce liquation cracking because of better ductility and lower concentrations

of liquation-inducing segregates. As-cast materials are particularly susceptible to liquation cracking because of the presence of low-melting segregates along grain boundaries. Upon heating during welding, severe grain-boundary liquation can occur in the PMZ, making it highly susceptible to liquation cracking. Examples include as-cast 304 stainless steel82 and as-cast corrosion-resistant austenitic stainless steel.83 By heat treating to homogenize the castings before welding, liquation cracking can be reduced. CONCLUSION Experimental data of crack sensitivity vs. composition are useful for selecting the proper weld-metal composition to avoid solidification cracking. To avoid liquation cracking at the same time, however, the weld-metal composition should be such that the weld metal solidifies after the partially melted zone does. Curves of temperature vs. solid fraction of the weld metal and the base metal (same as the partially melted zone) can help select the proper weld-metal composition. ACKNOWLEDGEMENTS This work was supported by the National Science Foundation under Grant No. DMR-0098776. The authors are grateful to Bruce Albrecht and

2003 June • JOM

(

Liquation cracking can be reduced by controlling the weld-metal composition, the welding condition, and the base metal. The weld-metal composition can be adjusted by selecting a proper filler metal and dilution ratio. Metzger67 observed liquation cracking in 6061 aluminum welds made with Al-Mg fillers at high dilution ratios but not in welds made with Al-Si fillers at any dilution ratios. Gittos et al.65 observed liquation cracking in alloy 6082 welded with NG61 (equivalent to 5356) filler at high dilution ratios (about 80%) but not in welds made with NG21 (equivalent to 4043) filler at any dilution ratios.

)

REMEDIES FOR LIQUATION CRACKING

( )

Figure 6. Temperature vs. solid fraction for 6061 base metal (PMZ) and weld metal of 6061 welds made with filler metals 5356 and 4043 at about 65% dilution. Curves calculated using Pandat 73 of CompuTherm LLC.72

41

Todd Holverson of Miller Electric Manufacturing Company, Appleton, Wisconsin for donating the welding equipment (including Invision 456P power source and XR-M wire feeder and gun) and for their technical assistance. The authors thank Y.A. Chang, a professor at the University of WisconsinMadison, for letting them use the database required for calculating the T-fS curves. References 1. S. Kou, Welding Metallurgy, 2nd edition (New York: John Wiley, 2003), pp. 263–339. 2. G.J. Davies and J.G. Garland, Int. Metal Rev., 20 (1975), p. 83. 3. D.C.G. Lees, J. Inst. Metals, 72 (1946), p. 343. 4. A.R.E. Singer and P.H. Jennings, J. Inst. Metals, 73 (1947), p. 273. 5. M.C. Flemings, Solidification Processing (New York: McGraw-Hill, 1974). 6. H.F. Bishop et al., Trans. AFS, 68 (1960), p. 518. 7. J.C. Borland, Br. Weld. J., 7 (1960), p. 508. 8. J.C. Borland, Welding and Metal Fabrication, 47 (January/February 1979), pp. 19–29. 9. R.D. Stout, Weldability of Steels, 3rd edition, ed. R.D. Stout and W.D. Doty (New York: Welding Research Council, 1978), p. 252. 10. R.W. Messler, Jr., Principles of Welding, Processes, Physics, Chemistry and Metallurgy (New York: Wiley, 1999), pp. 557–589. 11. P.T. Houldcroft, Br. Weld. J., 2 (1955), p. 471. 12. J.A. Liptax and F.R. Baysinger, Weld. J., 47 (1968), p. 173s. 13. J.G. Garland and G.J. Davies, Metals Const. Br. Weld. J., 16 (December 1969), p. 565. 14. J.H. Rogerson, B. Cotterell, and J.C. Borland, Weld. J., 42 (1963), p. 264s. 15. W.F. Savage and C.D. Lundin, Weld. J., 44 (1965), p. 433s. 16. K. Nakata and F. Matsuda, Trans. JWRI, 24 (1995), p. 83. 17. K.F. Krysiak et al., Welding, Brazing, and Soldering, Volume 6 (Materials Park, OH: ASM International, 1993), p. 443. 18. J.N. DuPont, J. R. Michael, and B.D. Newbury, Weld. J., 78 (1999), p. 408s. 19. M.J. Cieslak, Weld. J., 66 (1987), p. 57s. 20. S.A. David and J.J. Woodhouse, Weld. J., 66 (1987), p. 129s. 21. T.W. Nelson et al., Weld. J., 76 (1997), p. 110s. 22. G.M. Goodwin, Weld. J., 66 (1987), p. 33s. 23. J.N. DuPont, C.V. Robino, and A.R. Marder, Sci. Technol. Weld. Join., 4 (1999), p. 1. 24. G.R. Pease, Weld. J., 36 (1957), p. 330s. 25. D.A. Canonico et al., Welding Research Council Symposium on Effects of Minor Elements on the Weldability of High-Nickel Alloys (New York: Welding Research Council, 1969), p. 68.

42

26. W.F. Savage, E.F. Nippes, and G.M. Goodwin, Weld. J., 56 (1977), p. 245s. 27. D.H. Kah and D.W. Dickinson, Weld. J., 60 (1981), p. 135s. 28. A.R.E. Singer and P.H. Jennings, J. Inst. Metals, 73 (1947), p. 197. 29. W.I. Pumphrey and J.V. Lyons, J. Inst. Metals, 74 (1948), p. 439. 30. J.D. Dowd, Weld. J., 31 (1952), p. 448s. 31. P.H. Jennings, A.R.E. Singer, and W.L. Pumphrey, J. Inst. Metals, 74 (1948), p. 227. 32. J.H. Dudas and F.R. Collins, Weld. J., 45 (1966), p. 241s. 33. E.J. Michaud, H.W. Kerr, and D.C. Weckman, Trends in Welding Research, ed. H.B. Smartt, J.A. Johnson, and S.A. David (Materials Park, OH: ASM International, 1995), p. 154. 34. A. Gueussier and R. Castro, Rev. Metall., 57 (1960), p. 117. 35. T. Takalo, N. Suutala, and T. Moisio, Metall. Trans., 10A (1979), p. 1173. 36. T.J. Lienert, Trends in Welding Research, ed. J.M. Vitek et al. (Materials Park, OH: ASM International, 1998), p. 726. 37. C.R. Smith, Trans. AIME, 175 (1948), p. 15. 38. F. Matsuda et al., Trans JWRI, 12 (1983), p. 93. 39. F. Matsuda et al., Trans JWRI, 13 (1984), p. 57. 40. M.J. Dvornak, R.H. Frost, and D.L. Olson, Weld. J., 68 (1989), p. 327s. 41. W.L. Pumphrey and D.C. Moore, J. Inst. Metals, 73 (1948), p. 425. 42. A.L. Schaeffler, Metal. Prog., 56 (1949), p. 680. 43. W.T. Delong, Weld. J., 53 (1974), p. 273s. 44. D.J. Kotecki, Weld. J., 78 (1999), p. 180s. 45. D.J. Kotecki, Weld. J., 79 (2000), p. 346s. 46. M.C. Balmforth and J.C. Lippold, Weld. J., 79 (2000), p. 339s. 47. J.M. Vitek, Y.S. Iskander, and E.M. Oblow, Weld. J., 79 (2000), pp. 33s and 41s. 48. J.M. Vitek, S.A. David, and C.R. Hinman, Weld. J., 82 (2003), p. 43s. 49. R.B. Smith, Welding, Brazing, and Soldering, Volume 6 (Materials Park, OH: ASM International, 1993), p. 642. 50. J.C. Borland, Br. Weld. J., 8 (1961), p. 526. 51. G.N. Heintze and R. McPherson, Weld. J., 65 (1986), p. 71s. 52. W.A. Petersen, Weld. J., 53 (1973), p. 74s. 53. B.P. Pearce and H.W. Kerr, Metall. Trans., 12B (1981), p. 479. 54. F. Matsuda et al., Trans. JWRI, 12 (1983), p. 93. 55. H. Yunjia et al., Weld. J., 68 (1983), p. 280s. 56. S. Sundaresan et al., Science and Technology of Welding and Joining, 5 (2000), p. 257. 57. J.G. Garland, Metal Const. Br. Weld. J., 21 (1974), p. 121. 58. S. Kou and Y. Le, Weld. J., 64 (1985), p. 51. 59. S.A. David and C.T. Liu, Grain Refinement in Castings and Welds, ed. G.J. Abbaschian and S.A. David (Warrendale, PA: The Metallurgical Society of AIME, 1983), p. 249. 60. C. Tseng and W.F. Savage, Weld. J., 50 (1971), p. 777. 61. S. Kou and Y. Le, Metall. Trans., 16A (1985), p. 1887.

62. S. Kou and Y. Le, Metall. Trans., 16A (1985), p. 1345. 63. O.W. Blodgett, Weld. Innovation Q., 2 (3) (1985), p. 4. 64. R.G. Thompson, ASM Handbook, volume 6 (Materials Park, OH: ASM International, 1993), p. 566. 65. N.F. Gittos and M.H. Scott, Weld. J., 60 (1981), p. 95s. 66. M.J. Cieslak, ASM Handbook, Welding, Brazing and Soldering, vol. 6 (Materials Park, OH: ASM International, 1993), p. 88. 67. G.E. Metzger, Weld. J., 46 (10) (1967), pp. 457s–469s. 68. J.E. Steenbergen and H.R. Thornton, Weld. J., 49 (2) (1970), pp. 61s–68s. 69. N.F. Gittos and M.H. Scott, Weld. J., 60 (6) (1981), pp. 95s–103s. 70. T. Ma and G. Den Ouden, Int. J. for the Joining of Materials (Denmark), 11 (3) (1999), pp. 61–67. 71. C. Huang and S. Kou, “Liquation Cracking in Full-Penetration Aluminum Welds: Binary Al-Cu Welds,” submitted to Welding Journal. 72. C. Huang and S. Kou, “Liquation Cracking in Full-Penetration Aluminum Welds: Al-Mg vs. Al-Si Fillers for Welding Al-Mg-Si Alloys,” submitted to Welding Journal. 73. Pandat, computer software developed by CompuTherm LLC, Madison, WI. 74. M. Katoh and H.W. Kerr, Weld. J., 66 (1987), p. 360s. 75. H.W. Kerr and M. Katoh, Weld. J., 66 (1987), p. 251s. 76. M. Miyazaki et al., Weld. J., 69 (1990), p. 362s. 77. C. Huang, G. Cao, and S. Kou, “Liquation Cracking in Partial-Penetration Aluminum Welds: Assessing Tendencies to Liquate, Crack and Backfill,” submitted to Science and Technology of Welding and Joining. 78. C. Huang, S. Kou, and J.R. Purins, Proceedings of Merton C. Flemings symposium on Solidification and Materials Processing, ed. R. Abbaschian, H. Brod, and A. Mortensen (Warrendale, PA: TMS, 2001), p. 229. 79. R.G. Thompson et al. Weld. J., 64 (1985), p. 91s. 80. H. Guo, M.C. Chaturvedi, and N.L. Richards, Science and Technology of Welding and Joining, 4 (1999), p. 257. 81. W. Yeniscavich, Methods of High-Alloy Weldability Evaluation, (New York: Welding Research Council, 1970), p. 1. 82. W.R. Apblett and W.S. Pellini, Weld. J., 33 (1954), p. 83s. 83. M.J. Cieslak, ASM Handbook, Welding, Brazing and Soldering, vol. 6 (Materials Park, OH: ASM International, 1993), p. 495. Sindo Kou is professor and chair of the Department of Materials Science and Engineering at the University of Wisconsin. For more information, contact Sindo Kou, University of Wisconsin, Department of Materials Science and Engineering, 1509 University Avenue, Madison, Wisconsin 53706; (608) 262-0576; fax (608) 262-8648; e-mail [email protected].

JOM • June 2003