New Method for Evaluating Strength and Ductility of Brazed Joints Advantages of a new universal type of braze specimen are simple design for easy production and good reproducibility. A brazing fixture is not required and the brazing gap is easy to maintain

BY E. KLAUSN ITZER

Introduction (c) hardness tests; (d) analyses of review is given in Reference 1. The High-temperature brazing has structure; (e) microanalyses; (f) cor­ shapes of specimens have the follow­ found special application in nuclear rosion tests; (g) irradiation tests and ing disadvantages from the point of reactors and best known is the fabri­ post-irradiation tests as per (a) to view of the present objective: (a) cation of brazed spacers for fuel ele­ (d). Brazed specimens are made individu­ ments. Web cross joints of about One and the same shape of speci­ ally. This involves dissimilar brazing 0.016 inch (0.4 mm) stainless steel or men were used in all tests to keep the conditions and high production costs; nickel alloy sheet are usually made. marginal conditions resulting from the (b) The gap is difficult to maintain. Figure 1 as an example shows a manufacture of the brazed specimens This drawback makes itself felt partic­ spacer of Inconel 718 material brazed constant. ularly in high-temperature brazing fol­ with AWS class BNi-7 brazing filler lowed by homogenizing; (c) Speci­ Universally Applied mens are very large as a rule and are metal for use with fuel elements in Brazed Specimen pressurized water reactors. therefore inappropriate for irradiation Initially the stress conditions to tests; (d) A brazing fixture is required The possibility of using stainless in some cases which presents addition­ steel spacers, as shown in Fig. 2 which the brazed joints are exposed permit only the use of high- al difficulties when high brazing tem­ brazed with AWS BNi-5, was to be peratures are involved. These consid­ studied under a sodium-cooled fast temperature brazing filler metals, such as AWS BNi-5. All the following ex­ erations resulted in the design of the breeder research scheme. The joints brazed specimen shown in Fig. 3, are subject to a more severe environ­ planations and discussions should therefore be viewed from this aspect. which largely avoids the above- ment than that which is found in mentioned drawbacks. pressurized-water reactors due to the Brazing was performed without flux higher operating temperature, the under a vacuum. A sleeve is manufactured from the highly corrosive medium and the high Numerous test methods and shapes base metal and a pin is fitted into it. dose of neutrons. of specimens are known. A very good The pin is reduced in diameter over Before the investigations on the brazing filler metal were started, the question of a universal type of braze Forced fit- 1 1 specimen was considered. The results * ' * I'S * I 1 * " ' * , * * 7y\\ •-n "... of these considerations are described - - : _ * •• " J-' - - - - , *:j * ,; •-::, - u in the following paper. * • * •-. • Test Procedures m ., . ~ . *- - - ill To enable brazed joints to be tested li. . j . - i »**<. 1 - -. for their useability tinder the an­ i S i 111 81 1 1 » 1 I i ticipated service conditions—tem­ Fig. 1—Spacer made of Inconel 71* 8 is perature: 1200°F (650°C); serv­ brazed with AWS BNi-7 Venting holes--^ Pockets for brazing metal ice medium: sodium; "fluence": p IO23 n/cm2 (E > 0.5 MeV)—the < following tests were carried out: (a) shearing tests at 68°F (20°C), 750°F Brazing gap---^^ / (400°C), 1200°F (650°C); (b) de­ formation tests at 68°F (20°C), 750°F (400°C), 1200°F (650°C);

E. KLAUSNITZER is with Siemens Akti- Forced fit- engesellschaft, Erlanger, Postfach, W. Z Germany. Paper presented at the Second Interna­ tional AWS-WRC Brazing Conference held in San Francisco, Calif, during April 27- Fig. 2—Spacer made of stainless steel 29, 1971. is brazed with AWS BNi-5 Fig. 3—Brazed test specimen

454-s I OCTOBER 1971 Fig. 4—Typical appearance of a brazed test specimen (stain­ Fig. 5—Eccentric position of the pin (stainless steel brazed less steel brazed with AWS BNi-5) with CPNM 2) the greater part of its length so that Pockets for brazing metal the desired gap is obtained. The sleeve has several holes in opposite places, Punch- enabling the brazing gap to be vented and the brazing filler metal to be applied. Since the brazed specimen can be turned on a lathe, its manufac­ Sleeve ture is accurate, cheap and easily reproducible. The specimen in this form is still not appropriate for the tests enumer­ Holder ated above. It has to be divided into 0.04-0.08 inch (1-2 mm) thick slices. These slices enabled all tests to be performed. Brazed Test Specimen Sleeve and pin are degreased Specimen (methanol in an ultrasonic bath) and then fitted together. The holes lying on one line parallel to the axis of the sleeve are then filled with brazing filler Brazing gap metal. The number and size of the holes should be chosen so as to ensure the brazing gap being completely filled with brazing filler metal. A vacuum induction furnace for indirect heating by means of a susceptor has proved Fig. 7—Schematic of shearing tool suitable for the brazing operation. Brazing was carried out under the (1.2 mm) thick can be obtained from Venting slot' conditions prescribed by the manufac­ the specimen in Fig. 3. turer of the brazing filler metal or About 40 test specimens have been according to our own experience made with different brazing filler met­ (heating rate, brazing temperature, als and no rejects were found. holding time and cooling rate). The first examination after brazing is a Basic Studies of Slices visual test. If the specimens are metal­ A few basic studies were carried lically bright and the braze metal is out to check whether the brazed visible both in the holes and at the specimen had the desired properties. ends of the sleeve and pin, then the It was found that the gap usually is Fig. 6—Improved design of brazed test specimens have been properly brazed. constant over the entire length of the specimen The specimen is then sliced, making specimen. Figure 4 shows a typical occasionally encountered, as Fig. 5 sure that the cuts are taken exactly at brazed joint made with AWS BNi-5. shows. right angles. About 20 slices 0.05 inch An eccentric position of the pin was This inaccuracy presumably is due

WELDING RESEARCH SUPPLEMENT] 455-s Punch

Specimen

-fJ8,05 Fig. 10—Brazed test specimen after shearing test with re- Fig. 8—Details of shearing tool mains of braze metal on the sheared faces

to the small pin diameter of 0.16 then screwed on to prevent the slice TaB = shearing strength in ps inches (4 mm) and to the horizontal from bending upward during the test. (or kgf/mm2) position in which brazing was done. The lower end of the punch has a Utilizing larger specimens, however, centering pilot to fit the hole in the Pmax, = maximum load exerted by was impossible within this research slice. In diffusion brazing, the gap the testing machine in program since the space for the usually is very small and it is of pounds force. planned irradiation tests was limited. particular importance therefore that F0 = cross section under test in A larger specimen will be used in the both the brazed specimens and the inches or (mm2). future to take account of the experi­ shearing tool should be made with ence gained. This is shown in Fig. 6. close tolerances. Details are given in Numerous metallographic examina­ The new position of the brazing filler Fig. 8. tions made on tested slices revealed that shearing always takes place in the metal pockets and venting slots makes The shearing tool is shown in Fig. brazed joint. Figure 10 shows a typi­ a vertical brazing position possible. A 9. It is made of Nimonic 90 and has cal case where residues of braze metal greater number of slices is also proven itself well in about 100 tests at are distributed at random over the achieved in this way. 68°F (20°C), 750°F (400°C) and sheared faces. 1200°F (650°C.) For the test, the Determining Shear Strength shearing tool with the slice inserted is Determining Ductility placed in a testing machine and the A shearing tool (Fig. 7) is used for load is raised until the specimen As stated initially, the brazed joints determining the shear strength of the breaks. The shearing strength is calcu­ considered are subjected to a high slices. This tool comprises a die, a lated from the following formula: degree of "fluence" so that embrittle­ holder and a punch. The slice is ac­ ment is liable to occur. It is therefore commodated in a depression of the die most important that a coefficient where it is centered. The holder is Tali F„ denoting the ductility should be estab-

+*?

I 19 20 21 22 23 24 25 26 27 28 29

Fig. 9—Photo of shearing tool made of Nimonic 90

456-s | OCTOBER 1971 lished. No acceptable suggestions were the deformation. Any other quantita­ the brazed joints. The hardness test found in the technical literature. To tive criteria are impossible to give results (Vickers test using a load of avoid any misunderstanding it should since, as the degree of deformation 0.050 kgf) are also shown. be pointed out that the proposition rises, increasingly disturbing factors All tests were carried out on slices described in the following section is due to friction tend to develop. As is 0.05 inches (1.2 mm) thick obtained not fully satisfactory either but it ap­ evident from Fig. 12, friction causes from the brazed specimen in Fig. 3. pears to give reasonable results. uneven deformation. But it is also The testing rate was constant 0.02 The test method for determining seen how closely the slice adapts itself in/min (0.5 mm/min) in all tests. ductility is sketched in Fig. 11 and to the spherical punch. Figure 13 The shear and ductility test results works as follows: The slice is placed shows testing tools with different diam­ are shown in Tables 3 and 4 respec­ on a soft metallic base (lead or alumi­ eters. They are made of Nimonic 90 tively. For the sake of comparison, num) which has to be replaced for and were employed in more than 100 the base metal was also shear tested each new test. It is then forced into tests at 68°F (20°C), 750°F after it had undergone two different the base by means of a spherical (400°C) and 1200°F (650°C). Fur­ heat treatments. These slices, like the punch in a testing machine. The slice ther tests should reveal whether lubri­ brazed ones, were 0.32 inches (8 mm) adapts itself closely to the spherical cation, the use of other materials for in diameter and 0.05 inches (1.2 mm) punch. The merit of this method is the base, or employment of cylindri­ thick and were taken from the same that the brazed joint is evenly de­ cal punches can improve the testing heat as the sleeves of the brazed formed irrespective of the braze effi­ method. specimens. ciency (ratio of strength of braze metal to base metal). The degree of de­ Test Results Discussion formation depends on the diameter of An extensive test series was carried It is not the objective of this paper the spherical punch and the thickness out on different high-temperature braz­ to discuss the quality of the brazed of the slice. The objective of the test is ing filler metals deposited on stainless joints but rather the effectiveness of to define the degree of deformation at steel. The brazing filler and base met­ the test methods. Nevertheless it which the first incipient crack occurs als used are tabulated in Table 1. The should be mentioned that the micro­ in the brazed joint. The test is there­ brazing conditions are stated in Table graphs of Fig. 14 display excellent fore started with a large punch, the 2. brazed joints with, in some cases, diameter of which is then reduced in Figure 14 shows typical micro­ complete diffusion of the brazing filler steps until an incipient crack is no­ graphs displaying the characteristics of metal (BNi-5) into the base metal. ticed. In a test series using the same slice thickness, the diameter of the spheri­ cal punch can be used as a criterion of

Spherical punch

Specimen

Soft metallic base e.g. aluminium) Fig. 11—Ductility testing tool

Fig. 12—Appearance of specimens after ductility test. Diameter of punch is 40 Fig. 13—Ductility testing tools mm (top), 16 mm (center) and 8 mm (bottom)

WELDING RESEARCH SUPPLEMENT I 457-s 1 ',-/.'' fr -: 195' 1 183 183 206 171 • • . . • 218 180 234 175 .J7^ - ' • • - 249 •'.-"" - •.* ''''• ' ^r\^ *321 197 • til 177 * •,''." 208 .1:78 186 174 182 183 177

^::,i^;..i:.:.:j,:B^: y~- •'• 166 •••>::.:. - 171 171 •* 171. 174 • • 180 173 •' • . ' 180 178

f A'- f. » 177 242 . 499 -183 'j 166' 169 - 175 163 . 189 174 174 172 ~; * , ^"Vi.---x '^^ ££tl.'.

180 183 • :•> -' 193 192 i" 204 199 190 18,5 • • 184 201 J90 ijs.•* » 24*9 299

i.. - 192 202 202 202 kv.. . - v.. .' • i 202 186- 202 195 193

Fig. 14—Typical micrographs show structure and hardness of brazed joints. Top (left) BNi-6, (right) BNi-5; middle (left) Nicobraz 60, (right) CM60; bottom (left) CMNP, (right) CPNM2

458-s | OCTOBER 1971 Shearing Strength Although the slices were taken at random from five brazed specimens, Table 1—Chemical Composition of Brazed Joints there is only a small amount of scatter Brazing in the various test results. - Corr position, % filler metals* Ni Si Fe Mo P Cr M • Pd Cu The shear strength measured on all specimens in which complete dif­ BNi-6 Remainder — — — 11 1 — — — fusion was observed is on the same BNi-5 Remainder 10 — — — 19 — — — Nicrobraz 60 Remainder 8 17 order of magnitude as that of the — — — — — — CM 60 67 10 3 — — 20 — — — base metal. Just a slight reduction CMNP 50.1 11 30 5.4 3.5 — — — — was observed at high temperatures. CPNM 2 15 10 20 55 An exception to this is brazing filler metal CPNM2 at 1200°F (650°C). * By AWS classification where applicable. The sharp drop in shear strength is Note: Base metal is X8 CrNiMo V Nb 16 13, W.Nr. 1-4988. Composition: 0.05% C, 0.5% Si, 1.24% Mn, 16.8% Cr, 13.15% Ni, 0.79% V, 1.44% Mo, 0.67% Ta + Nb, remainder Fe. accounted for by the different chemi­ cal compositions of this filler metal, the latter moreover being a nondiffusa- ble type of brazing filler metal. The high shear strength and the Table 2—Brazing Conditions small spread of the test results are Temperature surprising. One might conclude that Brazing Brazing Vacuum, Temperature, maintenance there is an inherent error in the test filler metal gap,/im torr °F/°C time, minutes method. What evidence can be pro­ BNi-6 15 5 X IO"6 1920 (1050) 120 duced to prove the effectiveness of the BNi-5 15 5 X 10"' 2190 (1200) 60 method described? Nicrobraz 60 15 5 X 10-6 2050 (1120) 90 6 a) The high shear strength is CM60 15 5 X 10" 2150 (1175) 60 CMNP 15 5 X IO"6 2100 (1150) 60 plausible since optimum brazing 6 conditions existed. CPNIV12 15 5 X 10~ 2190 (1200) 5 b) Numerous metallographic examinations were undertaken to prove that the rupture under shear stress actually is in the brazed joint.

Table 3—Shear Test Results RT 750° F (400° C) 1200° F (650° C)

Brazing filler metal Individual results Mean value Individual results Mean value Individual results Mean value

67,570(47.5) 59,710(42.0) 55,140(38.7) AWS BNi-6 68,140(47.9) 68,710(48.3) 59,140(41.6) 61,000(42.9) 53,710 (37.7) 54,000(38.1) 70,570 (49.6) 64,280 (45.2) 53,850(37.8)

59,000(41.5) 54,280 (38.2) 50,350 (35.4) AWS BNi-5 68,280(48.0) 65,140 (45.8) 55,500 (39.0) 56,150(39.5) 49,780(35.0) 50,070(35.2) 68,420 (48.1) 58,570 (41.2) 50,070(35.2)

67,710(47.6) 50,500 (35.5) 48,070 (33.8) Nicrobraze 60 68,430 (48.1) 65,850 (46.3) 54,600 (38.4) 54,020 (38.0) 47,950(33.7) 48,070(33.8) 61,180 (43.0) 57,070 (40.1) 48,070(33.8)

67,000 (47.1) 57,000 (40.0) 43,500(30.6) CM60 71,280(50.1) 70,000(49.2) 55,280 (38.8) 55,900 (39.3) 49,780(35.0) 38,520(34.1) 71,100(50.0) 55,700(39.2) 52,080 (36.6)

72,700(51.1) 60,930(42.8) 47,300(33.3) CMNP 75,540(53.1) 73,250(51.5) 61,900(43.5) 61,300(43.1) 53,350(37.5) 51,020(35.8) 71,570 (50.4) 61,300(43.1) 51,020 (35.8)

66,850(47.0) 56,850(39.9) 35,000 (24.6) CPNM2 67,000 (47.1) 68,560(48.2) 56,930(40.0) 57,070 (40.1) 33,290(23.4) 34,710(24.4) 71,450 (50.3) 57,710(40.5) 35,570(24.9)

Base metal

W. Nr. 1.4988 75,540 (53.2) 55,780(39.2) 50,930(35.8) 2 hr—1920° F (1050° C) 77,000(54.1) 75,850 (53.3) 55,350 (38.9) 55,640 (39.1) 52,210 (36.7) 52,500(36.2) 74,700 (52.5) 55,780 (39.2) 51,640 (36.3)

W. Nr. 1.4988 71,100 (50.0) 56,850(39.9) 51,640 (36.3) 1 hr—2190° F (1200° C) 71,100 (50.0) 71,100(50.0) 55,640(39.1) 56,350(39.6) 52,780(37.1) 52,500(36.9) 71,280(50.1) 56,710 (39.8) 53,340(37.5)

WELDING RESEARCH SUPPLEMENT | 459-s Ductility The spread of the test results was Table 4—Ductility Test Results larger in the ductility tests than in the shear tests. Moreover, big differences Diameter of punch in inches (mm) at instant of incipient cracking were encountered between the various Brazing metal RT 750° F (400° C) 1200° F (650° C) types of brazing filler metal. However, BNi-6 0.63 (16) 1.57(40) 0.47 (12) in principle the results agree with 0.95 (24) 1.57 (40) 0.39 (10) those of the structural examinations 1.57 (40) 0.95 (24) 0.39(10) 1.57 (40) 0.95 (14) and hardness tests in Fig. 14 (i.e., 1.57(40) 0.95(24) incomplete diffusion and great hard­ BNi-5 0.32 (<8) 0.32 (<8) 0.32 (<8) ness involve low ductility). In several 0.39 (10) 0.32 (<8) 0.32 «8) cases no incipient crack could be pro­ 0.32 (<8) 0.32 (<8) 0.32 (8) duced even when the smallest punch 0.47 (12) 0.32 (<8) diameter available was used. There is 0.47 (12) 0.32 (<8) no point in providing even smaller Nicrobraz 60 (120) 0.95(24) 0.39 (10) diameters since the increasing friction 1.57 (40) 0.47 (12) 0.63 (16) 4.72 (120) 0.95 (24) 0.32 (8) would lead to uneven reduction of the 4.72 (120) 0.47 (12) wall thicknesses (see also, Fig- 12). 4.72 (120) 0.3 (<8) Conclusions CM60 0.32 (<8) 0.32 (<8) 0.32 (<8) 0.32 (<8) 0.32 (<8) 0.32 (<8) The experience gathered with these 0.32 (<8) 0.32(8) 0.32 (8) brazed specimens and the new testing 0.32 (<8) 0.32 (<8) methods for evaluating the shear 0.32 (<8) 0.32 (8) 0.32 (<8) 0.32 «8) strength and ductility of brazed joints CMNP 0.32 (<8) are predominantly good. 0.39 (10) 0.32 (<8) 0.32 (8) 0.63 (16) 0.32 (<8) 0.32 (8) The great advantages of the brazed 0.47(12) 0.32 «8) specimen are easy and cheap produc­ 0.63 (16) 0.32 «8) tion and good reproducibility as well CPNM2 0.32 (<8) 0.32 (8) 0.32 (8) as good maintenance of the brazing 0.47 (12) 0.32 (<8) 0.47 (12) gap. It should further be emphasized 0.32 «8) 0.32 (<8) 0.32 (8) that a brazing fixture is not required. 0.32 (<8) 0.47 (12) Slices cut from the brazed specimen are advantageous in that they were brazed under exactly the same condi­ tions. Any undesirable marginal con­ ditions are therefore eliminated from mens. These specimens were in­ c) The shear strength measured the subsequent examinations. The in the base metal is the actual spected subsequently and it was slices are well suited for many testing value of this material (check found that lack of fusion existed, methods. made by applying the formula r e) The results obtained with aB = 0.8 X SB). brazing filler metal CPNM 2 at References d) Markedly lower values were 1200°F (650°C) prove that low 1. DMIC Report 244, August 1968, weid- measured on some of the speci­ shear strength may occur as well. merit Evaluation Methods.

Order now-slip cases for your Welding Journals

• Each case holds 12 issues (yearly volume of the Journal). Made from finest quality binders board- covered with washable Stands upright. Journal issues slip in and out easily. simulated leather.

• Black sides, with back in Decorator's red. Title and AWS symbol • Available from Welding Journal at $3.50 each. (Price outside imprinted in 23K gold. Gold foil provided to enable user to in­ USA or its possessions—$4.50 each. Add 6% sales tax on sert year and volume number within seconds. New York City orders. Allow 3 to 4 weeks for delivery.)

460-s | OCTOBER 1971