Use of Electrodeposited Silver as an Aid in Diffusion Welding

Ring tests are used to evaluate silver plating procedures, and a two level, four variable factorial experiment is conducted to gain a better understanding of variables influencing the diffusion welding process

BY J. W. DINI, W. K. KELLEY, W. C. COWDEN, AND E. M. LOPEZ

ABSTRACT. Sound, high strength diffu­ into close contact, but without macro­ Experimental Details sion welded joints were obtained with scopic deformation occurring. In some Specimens silver plated stainless , beryllium and cases the weld formation results in the uranium. The procedures used for pre­ complete disappearance of the interface, Cylindrical butt specimens of the type paring the substrates for plating were first and often 100% joint efficiencies can be shown in Fig. 1 were used for this work. characterized by using ring shear tests, achieved. The process is accomplished These were plated with 0.5 or 3.0 mil which provided quantitative information entirely in the solid state at temperatures (0.0005 or 0.003 in.) of silver on one face, on weld adhesion. under 0.7 Tm (where Tm is the and then two specimens were diffusion For the joining studies, a full factorial point in degrees absolute of the lower welded together using varying conditions experiment with two levels and four material). Good descrip­ of time, temperature, and load. Diffusion variables (24) was run with . tions and overviews of the process can welded specimens were then machined Fewer experiments were run with berylli­ be found in the literature (Refs. 1, 2). per Fig. 2 to provide a reduced section in um and uranium. Best joints were Additional layers of material in the the joint area for tensile testing. Materials obtained with welding conditions which form of coatings or foils are often used as used as substrates included Type 304 included 1 hour at a temperature of joining aids for a variety of reasons (Refs. stainless steel, beryllium, and unalloyed 600°C (1112°F), of 30,000 psi 1, 3); these are to: uranium. (207 MPa) and a plating thickness of 1 mil 1. Promote flow. The Type 304 stainless steel was (0.001 in.) of silver. 2. Obtain clean surfaces. annealed and hot-rolled; its strength 3. Promote diffusion. was 43,000 psi (296 MPa) and the tensile Introduction 4. Minimize undesirable intermetallics. strength 110,000 psi (758 MPa). The 5. Temporarily establish eutectic melt­ beryllium had a yield strength of 37,000 The joining of dissimilar has ing to promote diffusion of base metals. psi (255 MPa), and its tensile strength was widespread industrial application, particu­ 6. Minimize Kirkendall porosity. 58,500 psi (403 MPa). The uranium was larly in the aerospace and nuclear indus­ 7. Reduce joining temperature. high purity rolled rod with a yield strength tries. On occasion, metals to be joined 8. Reduce dwell time. of 50,000 psi (344 MPa) and a tensile cannot be fusion welded or brazed 9. Scavenge undesirable elements. strength of 138,000 psi (951 MPa). because of metallurgical incompatibility. The intermediate coatings used for Most of the work was done with In addition, some metals may require a diffusion welding can be applied by stainless steel, and a full factorial experi­ post-joining heat treatment that rules out either or deposi­ ment with two levels and four variables and . In such cases tion. A recent survey summarized the use (24) was completed with this material. an alternative process that can be used is of electroplated coatings as an aid in When beryllium and uranium specimens diffusion welding. diffusion welding (Ref. 4). Other authors were diffusion welded, one-half of each Diffusion welding is a process whereby (Refs. 5 and 6) have described the use of joint was a stainless steel specimen. This a joint is effected by heating two clean vacuum deposited coatings for this type was done to minimize the cost of machin­ smooth surfaces while they are pressed of application. The purpose of the work ing additional specimens, which is an reported in this paper is to show that expensive proposition for these metals. stainless steel, beryllium, and uranium can /. W. DINI, W. K. KELLEY, W. C COWDEN and No precautions were taken to obtain a E. M. LOPEZ are with the Lawrence Livermore be diffusion welded with the aid of elec­ specified surface finish on the faces to be National Laboratory, Livermore, California. troplated silver. joined. Typically, the finish was in the

26-S | JANUARY 1984 Z ui 2 a. O 30°0'±0°4' a-J BOTH ENDS >Ul NO UNDERCUT a OR N I B (.4975) 0 BOTH ENDS • x -/r<17Si.0005f o ' E*3

removed all the plated deposit except for Insulation ring 2

Table 1—Surface Roughness on Faying a. O Surface of Cylindrical Butt Joint Specimens _i >UJ Surface roughness, UJ It in a Material , RMS x Type 304 stainless 9, 10, 12 o QC steel

WELDING RESEARCH SUPPLEMENT I 27-s Table 2—Silver Plating Solution Formulation

Silver cyanide 40 g/L Potassium cyanide 60 g/L Potassium carbonate 15 g/L Free cyanide 41 g/L Temperature 21-27°C (70-81 °F) Current density 54-161 A/m2 (5-15 A/ft2)

composition were kept constant.

Results

Ring Shear Tests SPECIMEN UNDER TEST (CUT AWAY VIEW)

SUBSTRATE- The cleaning/activating procedures that were evaluated are those that have proved successful for preparing stainless steel, beryllium and uranium for plating with other metals such as or . Ring shear test results are included in Table 4. The procedures providing the best results are those that were used for specimens for the diffusion welding studies. Complete details on these proce­ dures are provided in the Appendix. For all three substrates, the best procedures (ALL DIMENSIONS ARE IN mm) 1 .60 resulted either in failure of the silver ns5 deposit or a combination of failure in the Fig. 4 — Ring shear test specimen and die substrate and the silver. With stainless steel, the key step in the in the literature (Ref. 7). A wealth of considerable information in return for a activation procedure is anodic treatment quantitative ring shear data is available for minimum of experimentation (Ref. 8). in sulfuric acid followed by a Wood's plating on stainless steel, beryllium and Stainless steel was chosen for this work nickel strike; this has been well docu­ uranium. However, no data were avail­ since, as mentioned earlier, these speci­ mented in previous literature (Ref. 9). able for silver deposited on these sub­ mens were the least expensive to fabri­ With this technique, an adherent thin strates. Therefore, these tests were run cate. In every experiment, each variable deposit of nickel is applied to the stainless to verify that procedures that have can have one of two states (high +, or steel, and this serves as a base for subse­ worked for other electrodeposited coat­ low —). The four variables included in this quent coating. ings on these substrates worked equally study and the limits chosen were: time — For good adhesion to beryllium, it is well for silver. 1 hour (h) and 4 h; temperature — 300 absolutely essential to apply an adherent and 600°C (572 and 1112°F); welding immersion zinc deposit (this is called a Silver Plating Solution pressure-10,000 and 30,000 psi (68.9 "zincate treatment") before - and 206.8 MPa); and plating thickness- 1 positing the primary of interest Plating was done in a 10 liter (10.6 mil and 6 mil (0.001 and 0.006 in.). (Ref. 10). Our data show that zincate quarts) silver cyanide plating solution of Plating variables were not included as treatment time is critical, with the stron­ the formulation shown in Table 2. Agita­ part of this investigation. Operating con­ gest bonds obtained with a 60 second(s) tion was kept constant with an electric ditions such as plating current density, immersion. These good results with stirrer. No proprietary additive was used, temperature, solution flow, and solution beryllium are of particular significance. since compounds of this type which are used to brighten deposits and refine the grain structure also increase the impurity content of the coatings. Table 3 shows Table 3—Impurity Content of Silver Deposits**' that a deposit produced in a solution with an additive had nearly twice as much Concentration, ppm carbon as a deposit produced in an Element No additive Proprietary additive'*5' additive-free solution. Copper 8 10 Lead 1 2 Factorial Experiment with Stainless Steel 2 2 4 4 Nickel 3 A two level, four variable factorial (2 ) Iron 10 8 experiment was run with stainless steel 30 23 specimens. This was done to get a better Nitrogen 4 4 understanding of the key variables which 11 5 influence the joining process. Moreover, Carbon 60 110 9 this type of design strategy has been Sulfur 9 shown to be a powerful for under­ 2 2 '"'Deposits were plated at 108 A/m (10 A/ft ), standing plating processes and provides (b>Lea-Ronal bright silver process "K". Lea-Ronal. Inc.. Jamaica, New '

28-s | JANUARY 1984 strength v. time, pressure, or thickness, Table 4—Ring Shear Data for Silver Electroplated Stainless Steel, Beryllium and Uranium included in Fig. 5 help to visually explain the analysis. It is to be noted that: Ring shear 1. Temperature was significant when strength, psi(a) Material Activation procedure diffusion welding pressure was 10,000 psi Type 304 stainless steel Clean, anodic sulfuric treatment. Wood's nickel 27100 (68.9 MPa); it was nonsignificant when strike, silver strike, silver plate the pressure was 30,000 psi (206.8 Beryllium Clean, 30 s zincate, silver strike, silver plate 18200 MPa)-Fig. 5A. Beryllium Clean, 60 s zincate, silver strike, silver plate 24200 2. Temperature was significant for 6 6061 aluminum Clean, 30 s zincate, silver strike, silver plate 21000 mil (0.006 in.) thick silver coating, but it 18400 6061 aluminum Clean, 60 s zincate, silver strike, silver plate was nonsignificant for 1 mil (0.001 in.) Uranium Clean, ferric chloride etch, silver plate 1000 thick coatings-Fig. 5B. Uranium Clean, ferric chloride etch, nickel plate (0.25 18300 mil), silver plate 3. Pressure was significant at 300°C (572 °F) but was nonsignificant at 600° C "psi X 0.00689 = MPa (1112°F)-Fig. 5C. 4. Pressure was significant for 6 mil (0.006 in.) thick silver coatings but was This is because, traditionally, a copper less steel specimens. Time was a nonsig­ nonsignificant for 1 mil (0.001 in.) thick strike has been used after the zincate nificant variable. On the other hand, coatings —Fig. 5D. step and this is followed by the deposit of thickness, pressure and temperature 5. Thickness was significant at 10,000 interest. Since we wanted to avoid any were significant as were the interactions psi (68.9 MPa) but was nonsignificant at extraneous metals to minimize harmful of temperature with pressure and tem­ 30,000 psi (206.8 MPa)-Fig. 5E. intermetallics during solid state welding, perature with thickness. Plots of joint 6. Thickness was significant at 300°C no copper strike was used in any of this work. The ring shear data in Table 4 verify that substitution of a silver strike Table 5-—Factorial Experiment for Stainless Steel'a for the copper strike did not degrade adhesion. X, x2 x3 Xa Failure load, Uranium is one of the most difficult Trial (Time] (Temp.) (Load) (Thickness) psi metals to plate upon adherently, because I - - - - 44,500 its surface has a tendency to oxidize and 2 + - - — 42,400 become passive. True chemical bonding 3 - + - - 56,800 is not obtained between electrodeposits 4 + + - — 55,000 and uranium (Ref. 11). Therefore, the 5 - - + - 61,900 uranium surface must be roughened by 6 + - + - 66,300 either chemical, electrolytic or mechani­ 7 - + + - 48,000 cal means to provide some degree of 8 + + + - 57,300 mechanical adhesion. The most success­ 9 - - - + 11,800 and 3,000 10 18,100 and 4,000 ful techniques involve chemical or elec­ + - - + 11 - + - + 53,900 and 48,900 trolytic treatment in acid solutions con­ 12 + + - + 50,900 taining chloride , followed by removal 13 - - + + 48,700 of chloride reaction products in nitric acid 14 + - + + 53,200 before plating. The process we prefer for 15 - + + + 46,500 unalloyed uranium consists of etching in a 16 + + + + 49,100 solution containing 1400 grams/liter fer­ '•'The our variables had high (+) and low (—) values. including 1 a nd 4 hours (time), 300 and 600°C (temperature), 10.000 and ric chloride (Ref. 12). 30.000 psi load) and I and 6 mi (thickness). Poor bond strengths were obtained when silver was plated over uranium samples that had been etched, and this is Table 6—Analysis of Variance (ANOVA) Data for Diffusion Welded Silver Plated Stainless consistent with past experience with cop­ Steel Specimens per and electroless nickel deposits (Ref. i (b1 ) 13). Good bond strengths are obtainable Source D.F.('' Sum sq. Mean sq. F Ratio, Fn P (F > F0)W when etched uranium is plated with nick­ TIME (TI) I 2,772.4 el. For this reason, a nickel deposit was TEMP (TE) I 43,642.7 8.69 .06 used as an intermediary coating before PRESS (P) 85,816.6 17.09 .025 silver plating, and this provided good THICKNESS (TH) 89,549.0 17.84 .024 adherence, e.g., a bond strength of TIXTE 5.0 18,300 psi (126.2 MPa). TIXP 3,202.8 TEXP 131,242.8 26.14 .015 TIXTH 1.4 Diffusion Welding—Stainless Steel TEXTH 39,364.2 7.84 .068 The factorial experiment conducted PXTH 12,136.6 TIXTEXP 323.6 with silver-plated stainless steel speci­ TIXTEXTH 872.8 mens is detailed in Table 5 along with the TIXPXTH 1,285.0 results. Good bonds were obtained using TEXPXTH 12,369.1 a variety of conditions with the mean TIXTEXPXTH 38.1 strength being 47,800 psi (329.6 MPa). RESIDUAL 7 1 15,062.5 5,020.8 Table 6 summarizes the analysis of vari­ ance (ANOVA) data from the 24 factorial (a) D.F. —degrees of freedom. (b) Sum sq — sum of squares. experiment with diffusion welded stain­ (c) P (F > F0)-The probability that F is greater than F0.

WELDING RESEARCH SUPPLEMENT 129-s P=30.TH=I 1- 60 TE=300.P=30 60 60 60 56 56 TE=600.TH=I 56 56 z TE=6C0.P=30 TE=300.TH=I 52 i 52 TE=600.P=I0 « 52 i 52 = 30.TH = 6 2Ul TE=600,TH=6 £ 48 £ 48 48 £ 48 -I0.TH= I a. 44 O 44 _i "5, 5, 44 44 ^ UJ . 40 | 40 40 S 40 U>l ™ 36 " 36 36 " 36 Q — 32 o 32 1 3Z I 32 X 28 28 28 28 TE=300.TH=6 24 24 24 24 u TE=300.P=I0

60 . TH»I 60 60 P=30 56 s

but was nonsignificant at 600°C as-plated specimens show how the -Fig. 5F. deposit follows the machining marks of 7. Time was nonsignificant— Fig. 5G. the stainless steel. Cells are built up that Scanning electron photomicrographs are similar to the ductile dimples found on of some fractured joints are shown in ruptured surfaces. Figs. 6 and 7. With strengths greater than 40,000 psi (275.8 MPa), the ruptured surfaces showed well developed ductile Beryllium dimples. In these cases silver-to-silver joint separation occurred by void coales­ Good welds were obtained with beryl­ cence—Fig. 6. With poor welds, e.g., Fig. 6 — Good ductile rupture on welded stain­ lium joined to stainless steel using a vari­ less steel specimen. Diffusion welding condi­ specimens at low temperature (300°C or ety of conditions —Table 7. The mean tions included 4 h at 600°C (1112°F) with 572°F) and low pressure (10,000 psi or weld strength was 34,300 psi (236.5 MPa) 30,000 psi (206.8 MPa) load. Silver plating 68.9 MPa), incomplete contact is clearly excluding two sets of specimens that thickness was I mil (0.001 in.) and joint visible — Figure 7. The ridges exhibited exhibited very low strengths — 3200 and strength was 57,300 psi (395 MPa). X500 joining, but the valleys were not welded. 2500 (22 and 17 MPa) and 14,900 and (reduced 49% on reproduction) In addition, other areas did not make 2200 psi (103 and 15.2 MPa) due most contact during welding. likely to the low temperature used for Figures 8 (0.5 mil or 0.0005 in. silver) joining these specimens. and 9 (3.0 mil or 0.003 in. silver) of Failure typically occurred in the berylli-

Fig. 7— Welded stainless steel specimen after testing. Conditions included I h at 300°C (572°F) and 10,000psi (68.9 MPa). Light areas were in contact and welding occurred; dark areas were not in contact during welding, joint Fig. 8 — 3.0 mil (0.003 in.) Ag as-deposited on Fig. 9 — 0.5 mil (0.0005 in.) Ag as-deposited on strength was 11,800 psi (81 MPa). X50 (re­ stainless steel. X200 (reduced 50", on repro­ stainless steel. X200 (reduced 50% on repro­ duced 49",, on reproduction) duction) duction)

30-s | JANUARY 1984 Table 7—Diffusion Welding Results for Beryllium Joined to Stainless Steel Ul Time Temperature Bonding pressure Silver plating thickness Failure load, psi 5 /- = 1 hr\ /- = 300°C\ /- = 10,000 psi\ = 1 mil\ oQ. V+ = 4 br) V+ - 600°C^ \+ = 30,000 psi/ (7 = 6 mi7 — — - Not done > + — - — Not done Ui + - - 34,900 a + + — — 31,400 - + — 3,200 and 2,500 o + - + — 14,900 and 2,200 um, and it is important to point out that The data of Table 7 were averaged for 3. Thickness and load were relatively X no special precautions were taken during each variable and plotted along with data insignificant for both beryllium and urani­ o tensile testing to assure proper alignment obtained for similar conditions with stain­ um but were significant for stainless steel.

52 > 56 48 56 56 a i 52 i 44 52 =5 52 Q. CL o Q. OC 5 48 ^ .40 u> 48 c5 48 < | 44 1=36 zi 44 ^44 UJ c Ui | 40 £ 32 40 a, 40 UJ OC 5 36 "' ?R CD 36 " 36 cn 1 32 1 24 32 ~ 32 "3 ~3 28 20 c 28 28 "3 24 16 24 24 o0. 20 12 20 20 —I Ul 10 4 Hr 300°C 600°C Kpsi 30 Kpsi I Mil 6 Mil > Time T empera ure Pressure Thickness Ul Fig. 10 — Influence of welding variables on joint strength for uranium oa

52 44 •— 52 52 ce a. Cl ,*: 48 40 -X. 48 i 48 a. m 44 36 44 t a 44 (0~) 40 32 4) 40 a. | 40 O 36 28 (0 36 | 36 o 42 24 'o iz —i " 32 28 20 —>2 8 | 28 24 16 24 X 24 o 20 12 20 OC 20 < I Hr 4 Hr 300°C 600°C Mil 6 Mil 10 Kpsi 30 Kpsi Ul Ui Time Temperature Thickness Pressure UJ OC Fig. 11 — Influence of welding variables on joint strength for beryllium

WELDING RESEARCH SUPPLEMENT j 31-s t- 52 56 48 56 56 zUJ i 52 i 44 52 / i 52 0s- Q- Q. ex. a 48 5 40 48 / £ 48 a.(/> / o_l £44 = 36 44 / f.44 Ul C c > S 40 • 32 o» 40 0. 40 c / UQl 4) " 36 " 28 36 / " 36 X 1 32 1 24 to 32 1 32 —3 o(K 28 20 O 28 28

With beryllium, strong silver-silver joints could not be examined, since the specimens broke entirely in the beryllium. Two specimens welded at low tempera­ ture - 300°C(572°F) for 1 and 4 h welding time at 30,000 psi (206.8 MPa)-did not show the good silver-silver joint typical of that seen on stainless steel. Figure 13 shows a region of beryllium fracture and silver- silver rupture typical of these specimens. Use of high temperature (600°C, i.e., 1112°F) during welding resulted in com­ plete failure in the beryllium.

Uranium

Regardless of the pretreatment pro­ cess, good joints were obtained with uranium under a variety of welding con­ ditions—Table 8. Thickness, load, and time were insignificant, but temperature was very important as it was with berylli­ A- • um and uranium. This is shown in Fig. 10 and Table 8. The weakest joints were always those that included a low welding :0KV X 5 1 0 18@U 095 47036 BSE temperature (300°C or 572°F) as part of the joining process. Fig. 13 —Beryllium specimen showing fracture in beryllium (left) and silver (right). Welding conditions were 1 h, 300°C (572"F), 30,000 psi(206.8 MPa), and 6 mil (0.006 in.) of silver. The joint In general, the diffusion welds were failed at 23,000 psi (158.6 MPa) stronger if no etch was used, or if the

Table 8—Diffusion Welding Results for Uranium Joined to Stainless Steel

Time Temperature Bonding pressure Silver plating thickness Failure load, psi /- = 300CX /- = 10,001,000 psi\\ /- = 1 mil\ Pretreatment (--ii) \+ = 600C/ \+ = 30,001,000 psipsij/ \+ = 6 mil/ + + 12,700 FeCI3 etch + 36,300 Plus 0.5 mil Electroplated Ni then - + + - 40,000 electroplated Ag + + + - 36,300 FeCI3 Etch + + - + 44,500 plus 0.25 mil Electroplated Ni then - + + + 35,400 electroplated Ag + + + + 47,300 FeCI3 + - + + 23,600 Electroplated Ag + + - + 26,400 + + + + 28,200 No FeCI3 etch + - + + 39,100 Electroplated Ag + + - + 33,600 — + + + 35,900

32-s | JANUARY 1984 etch was used and followed by a thin thickness, pressure and temperature 12. Waldrop, F. B., and Bezik, M. |, 1971. coating of nickel. The fact that good were significant as were the interactions U.S. patent 3,573,120. welds were obtained when no etch was of temperature with pressure and tem­ 13. Dini, J. W., and lohnson, H. R. 1981 used is of particular interest and is an perature with thickness. (Oct.) Plating and Surface Finishing 68:64. 14. Dini, |. W., and lohnson, H. R. 1982 example of a situation where the pres­ Although full factorial experiments (Sept. 2). Plating on and zirconium. sure and heat used for diffusion welding were not run with beryllium and uranium, UCRL 87111. overcame the deficiencies encountered enough data were obtained to reveal with less than optimum plating adhesion. that time was insignificant for these Similar improvements in joint strength materials. On the other hand, tempera­ have been obtained with diffusion ture was very significant, and thickness welded titanium and zirconium substrates and load were relatively insignificant. For Appendix even though less than optimum plating joining of these materials the use of 1 h at A. Procedure for Cleaning, Activating and procedures had been used to provide 600°C (1112°F) and 30,000 psi (206.8 Plating Silver on Type 304 Stainless Steel the bonding layer (Ref. 14). Although MPa) is suggested. The optimum plating good joints were obtained with no etch thickness would be 1 mil (0.001 in.) of 1. Degrease. prior to plating, it is recommended that silver. 2. Electroclean in Oakite 195 at 70°C, uranium parts for solid state welding be (158°F) cathodic 1V2 min; then anodic etched prior to plating. This is suggested Acknowledgments 30 s. because without the etch there is no 3. Water rinse. The work described in this paper was adhesion between the coating and urani­ 4. Immerse 10% (wgt) sulfuric acid for sponsored by the um, and deposits could easily be dam­ 3 min. Department at LLNL, and we are thankful aged prior to welding. 5. Anodic activate in 70% (wgt) sulfu­ for the support. We would like to ric acid for 3 min at 538 A/m2 (50 acknowledge W. W. Feng, who heads A/ft2). Recommendations the M.E. Research Program, for help and 6. Water rinse. assistance throughout the program, 7. Immerse in 2% (vol) hydrochloric For optimum diffusion welding results W. D. Ludemann for many technical acid for 30 s. it is recommended that stainless steel, discussions and for providing scanning 8. Wood's nickel strike for 5 min at beryllium and uranium be cleaned, acti­ electron microscopy support, and R. W. 538 A/m2 (50 A/ft2). Composition of vated and plated with silver using the Mensing for providing the analysis of the Wood's nickel strike: nickel chloride — processes described in the Appendix. For factorial experiment with stainless steel. 240 g/L; hydrochloric acid—125 ml/L; the diffusion welding operation, the use The work was performed under the Temperature- 25°C (77°F). of high temperature, high pressure, and auspices of the U.S. Department of 9. Water rinse. low thickness-e.g., 600°C (1112°F), Energy by Lawrence Livermore National 10. Immerse in potassium cyanide, 10 30,000 psi (206.8 MPa) and 1 mil (0.001 Laboratory under Contract W-7405-ENC- in.) —is suggested. If the parts are of a g/l for 1 min. 48. 2 complex configuration, a thicker silver 11. Silver strike for 1 min. at 108 A/m 2 plating can be used without noticeably (10 A/ft ). Composition of silver strike: degrading weld strength. Since time was References silver cyanide —7.4 g/L; potassium cya­ shown to be insignificant for all three nide—75 g/L; temperature —25°C 1. Meiners, K. E. 1973. Materials and pro­ materials, 1 h would be quite adequate to (77°C); —stainless steel. cesses for the 70's: cost effectiveness and facilitate joining. 12. Silver plate to final thickness at 54 reliability. Proceedings 5th national SAMPE 2 2 technical conference. Kiamesha Lake, New A/m (5 A/ft ). Composition of silver York. plating solution: silver cyanide —40 g/L; Summary 2. Olson, D. L, and Liby, A. L. 1979. Berylli­ potassium cyanide —60 g/L; potassium carbonate —45 g/L; temperature —25°C Stainless steel, beryllium and uranium um science and , eds. D. R. Floyd and J. N. Lowe, vol. 2, pp. 275-96. were diffusion welded with the aid of (77°F); agitation — stirring rod. 3. Cine, C. L. 1966. An analytical and electroplated silver. Sound, strong joints experimental study of . were obtained with V2 in. (12.7 mm) B. Procedure for Cleaning, Activating and Welding lournal 45(11):481-s to 489-s. Plating Silver on Beryllium diameter cylindrical butt specimens under 4. Dini, ). W. 1982 Use of electrodeposition a variety of diffusion welding conditions. to provide coatings for solid state bonding. 1. Degrease. The procedures used for preparing the We/ding /ournal 61(11):33-39. 2. Scrub with pumice. substrates for plating were first character­ 5. Naimon, E. R., Doyle, ). H., Rice, C. R., 3. Water rinse. ized by using ring shear tests which pro­ Vigil, D., and Walmsley, D. R. 1981. Diffusion 4. Etch in 20 parts water, 20 parts nitric vided quantitative information on adhe­ welding of aluminum to stainless steel. Weld­ acid, 1 part hydrofluoric acid for 5 min, ing journal 60(11):17-20. sion. 23°C (73°F). 6. Knowles, |. L, and Hazlett, T. H. 1970. For the diffusion welding studies, a full High-strength, low-temperature bonding of 5. Rinse. factorial experiment with two levels and 6. Zincate in solution containing sulfu­ 4 beryllium and other metals. Welding journal four variables (2 ) was run with stainless 49(7):301-s to 310-s. ric acid 25 ml/L, potassium fluoride 15 steel. Fewer experiments were done with 7. Dini, |. W„ and lohnson, H. R. 1978. g/L, zinc 30 g/L, 1 min, 27°C beryllium and uranium. Adhesion measurement of thin films, thick (81 °F). The four variables included in the study films and bulk coatings, ed. K. L. Mittal, pp. 7. Water rinse. were: time—1 and 4h; temperature — 305-26. Philadelphia: American Society for 8. Strip zincate in 50% nitric acid. Testing Materials, STP 640. 300 and 600°C (572 and 1112°F); bond­ 9. Water rinse. 8. Dini, |. W., and lohnson, H. R. 1981 ing pressure-10,000 and 30,000 psi 10. Repeat step 6. (Feb.). Plating and Surface Finishing 68:52. (68.9 and 206.8 MPa); and silver plating 9. Wood, D. 1938. Metal Industry 36:330. 11. Water rinse. thickness-1 and 6 mil (0.001 and 0.006 10. Dini, |. W., and lohnson, H. R. 1976 12. Silver strike —see stainless steel in.). With stainless steel the mean joint (lune). Plating and Surface Finishing 63:41. procedure (appendix A) for details. strength was 47,800 psi (329.6 MPa). 11. Dini, ). W., and Musket, R. C. 1973. 13. Silver plate —see stainless steel Time was a nonsignificant variable, while Plating 60:811. procedure (appendix A) for details.

WELDING RESEARCH SUPPLEMENT 133-s C. Procedure for Cleaning, Activating and 9. Water rinse. 55°C(131°F). Plating Silver on Unalloyed Uranium 10. Immerse in nitric acid (50% vol.) 17. Water rinse. for 3 min. 18. Immerse in sulfuric acid (10% wgt) 1. Degrease. 11. Water rinse. for 1 min. 2. Immerse in nitric acid (50% vol.) for 12. Etch in 1400 g/L ferric chloride, 19. Water rinse. 3 min. 32°C (90°F), 5 min. 20. Immerse in potassium cyanide, 10 3. Water rinse. 13. Water rinse. g/L for 1 min. 4. Scrub with pumice. 14. Immerse in nitric acid (50% vol.) 21. Silver strike —see stainless steel 5. Water rinse. for 3 min. procedures (appendix A) for details. 6. Immerse in nitric acid (50% vol.) for 3 min. 15. Water rinse. 22. Silver plate —see stainless steel 16. Plate with —0.25 mil in nickel sul- procedure (appendix A) for details. 7. Water rinse. famate solution, 215 A/m2 (20 A/ft2), 8. Scrub with pumice.

WRC Bulletin 280 August, 1982 The Varestraint Test by C. D. Lundin, A. C. Lungenfelter, G. E. Grotke, G. G. Lessmann, and S. J. Matthews

The Varestraint Test, or one of its various modifications, is the most utilized test for evaluation of hot cracking sensitivity. This monograph presents the experience of several researchers in their use of the Varestraint Test. It is not intended to be a standardization document, but a utilization guide.

Publication of this report was sponsored by the Subcommittee on Heat Resistant Alloys of the High Alloys Committee of the Welding Research Council.

The price of WRC Bulletin 280 is $10.00 per copy, plus $3.00 for postage and handling (foreign— $5.00). Orders should be sent with payment to the Welding Research Council, 345 East 47th St., New York, NY 10017.

WRC Bulletin 285 July, 1983 Indices and Flexibility Factors for Concentric Reducers by E. C. Rodabaugh and S. E. Moore

This report was developed as part of the ORNL Piping Program funded by the U. S. Atomic Energy Commission. The recommended stress indices in the report were incorporated into the ASME and Code, Section III, in 1977. Finite Element Analysis of Eccentric Reducers and Comparisons with Concentric Reducers by R. R. Avent, M. H. Sadd, and E. C. Rodabaugh

This report was developed to provide stress indices for eccentric reducers, and includes recommen­ dations for relatively minor work changes in the NB-3680 portion of the ASME Code which would extend the coverage to include eccentric reducers.

Publication of these reports was sponsored by the Sub-Committee on Piping Pumps and of the Pressure Vessel Research Committee of the Welding Research Council.

The price of WRC Bulletin 285 is $14.25 per copy plus $5.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, Rm. 1301, 345 E. 47th St., New York, NY 10017.

34-s | JANUARY 1984