Metal Science and Heat Treatment

Vol. 41. Nos. 9

10. 1999

UDC 621.792

E F F E C T OF S T R U C T U R A L I N H O M O G E N E I T Y OF M A T E R I A L ON T H E S T R E N G T H OF S O L D E R E D J O I N T S V. N. Semenov Translated from Metallovedenie i Termicheskaya Obrabotka Metallov, No. 10, pp. 13 - 16, October, 1999.

INTRODUCT|ON

The greater thickness o f the joint in the place where the shells break off is an indication that the clearance between them was not been fully exhausted in soldering o f the structure (in places with an elevated rigidity such as the transition boundary between the cylinder and the cone and the conical part itself). The presence o f the band in the steel implies the tbrmation o f a layer differing in structure from the base metal. This can be caused by the diffusion exchange o f chemical elements between the steel and the solder, because an excess amount o f liquid phase appears in places with an increased clearance (Fig. 4a ); hydrogen diffusing from the bulk o f the metal can accumulate in this zone too. The steel can be charged with hydrogen both in the process o f preparing the parts for depositing galvanic coats and in soldering due to the dissociation o f the moisture present in the volume of the furnace. In contrast to oxygen, hydrogen and iron form interstitial solutions [1]. Consequently, it can be expected that they may accumulate in this zone. The blue-black color on the surfhce o f the steel shell after withdrawing the soldered structure from the soldering furnace is a sign o f the presence of moisture in the furnace.

In tests of soldered bimetallic structures by hydraulic loads, their fracture occurs under a load below the one required by the specifications. The metal breaks off over the steel-bronze interface in the cylinder-cone transition zone (Fig. 1). The soldered structure in question consists of shells: the external shell is made o f high-strength steel o f the VNS type, and the internal shell is made o f a copper-base alloy BrKh0.8. The shells are soldered together with the help of a copper-silver solder at 970~ with a vacuum o f 265 mPa between the shells and an inert gas on the outside. A metaIlographic analysis of microsections cut from a zone close to the defective region has shown that the soldered joint in this place has a width of up to 100 p.m. In regions with a strongly dense joint (on the cylindrical part), the width of the juncture does not exceed 2 tam. The metal breaks off (Fig. 2) on the interface o f the joint and the base metal (steel). In the diffusion zone of the steel we find a band (layer) 1 0 - 15 p.m thick that differs in color from the base metal (Fig. 3).

METHODS OF RESEARCH In order to study the causes o f fracture and create a soldering technology providing the requisite strength of the

Fig. 1. Break-off over a soldered joint of steel with chromium bronze. • 1.

Fig. 2. Boundary zone in break-off over a soldered junction of steel and chromium bronze. • 250. 434 0026-0673/99/0910-0434$22.00 ~ 2000 Kluwer Academic/Plenum Publishers

Effect of Structural lnhomogeneity of Material on the Strength of Soldered Joints

435

Fig. 3. Inhomogeneity (band) in the diffusion zone in steel. • 250.

joint, we performed mechanical tests of soldered joints and metallographic, x-ray spectral, diffraction local, mass spectroscopic, and phase analyses. The mechanical tests were performed on soldered copper-steel disk specimens with a specified clearance between them and without a clearance in the soldering process. The geometry of the surface of the specimens corresponded to the geometry of the soldered shells. The clearance between the soldered disks was 0.05 and 0.1 mm wide. The solder was deposited onto the soldered surfaces by a galvanic method, i.e., a copper coating with a thickness h = 12 !am onto steel and a copper coating with h = 20 ~tm plus a silver coating with h = 5 lam onto the copper alloy. The formation of solder melt occurred due to contact-reaction melting of the coatings beginning at 779~ (the temperature of the copper-silver eutectic). The form and thickness of the coatings and the soldering regimes corresponded to the conditions of preparation of the parts for soldering and the soldering process itself. The metallographic analysis of the soldered joint and the steel layer on the boundary with the soldered joint was performed on microsections with the help of an MIM-8 microscope. The microspectral analysis was performed using Comebax and Joly microanalyzers. The concentration of chemical elements in the steel on the boundary with the soldered joint and in the joint itself was determined by comparing the intensities of the x-ray radiation from the studied specimens with the intensity of standard specimens. Carbides were segregated from the steel in a methanol electrolyte with an additive of hydrochloric and citric acids. The specimens were prepared preliminarily by air hardening from 1000~ and tempering at 610~ The x-ray diffraction analysis was performed in a Debye chamber. The distribution of hydrogen in steel specimens was studied using a laser mass-spectrometric installation based on a GOR-100M laser and an omegatron meter of partial pressure. The microspecimens were fabricated using a ruby laser operating in a mode of free generation. The laser beam was focused on the surface of the specimen, and the

Fig. 4. Microstructure of a soldered joint of steel and chromium bronze with a copper-silver solder (• 400): a) soldering with a clearance; b ) without a clearance.

adjustment was controlled using a microscope under a 200-fold magnification. RESULTS In can be seen from Fig. 5 that the strength of the joint diminishes considerably with the growth of its width due to widening of the clearance between the disks in soldering. The metallographic analysis has shown that in joints obtained by soldering with clearances the metal breaks off over the solder/base metal interface. The specimens fracture like a soldered structure (see Fig. 2). In the diffusion zone of steel soldered with a clearance, we observe bands (layers) (Fig. 4a ), whereas in specimens soldered without a clearance it is absent (Fig. 4b ). The microscopic x-ray spectral analysis has shown that the band formed in the steel differs in chemical composition from the base metal, it contains 29% Cr, 68% Fe, and 1% Ni (the base metal contains 14% Cr, 78% Fe, and 5% Ni). A qualitative analysis reflects this especially vividly. We observe a eutectic consisting of copper and silver in the soldered joint (Fig. 4a ); the base element in it is nickel and partially chromium. The iron that serves the steel base does not dissolve in the joint.

436

V.N. Semenov

~f, MPa 240 I

120-~

O-

80

0

(1.05

0.113

S, 1111TI

Fig. 5. Strength of a steel chromium bronze silver soldered joint as a fi.mction of the width of the clearance.

As a result of a layer-by-layer phase analysis and diffractometry, we established that the action of the solder on the surface layer increases the content of the carbide phase M_,3C~, (Cr~3C6 ) in it to 0.58% (in the base metal its concentration does not exceed 0.5%). A Layer containing phases such as Cu6Sn, NiSn, FeSn, CuO, and Cun 9~S has been observed on the steel/molten solder interlace. The lattice parameter of the metal of the layer a = 0.293 nm, and the thickness is 10 - 15 gin. Under this layer the steel has a lattice parameter a = 0.287 nm. The data of the mass-spectroscopic analysis have shown that the mass fraction of hydrogen in the bulk of the steel after depositing a galvanic coating onto the specimens is (I.I - 2.6) x 10 4 %. After soldering hydrogen is redistributed, namely, its content in the bulk amounts to ( 1 . 3 - 1.7) x 10 -4%, and on the interface with the molten s o t d e r i t i s ( 1 0 - 1 1 ) • 10 4%[2]. We can conclude that soldering of structures with an incompletely filled clearance between the shells is accompanied by complex physicochemical processes on the steel/melt interface, which leads to the formation of a layer in the surface zone of the steel, which differs in chemical composition from the base metal, and the appearance of chemical compounds on its surface. DISCUSSION OF RESULTS Let us consider the reasons for the appearance of a layer in the surthce zone of the steel and the effect of the layer and the chemical compounds on the strength of the soldered joint. A layer-by-layer phase analysis gives us grounds to state that the layer in question is an a-phase, i.e., ferrite. This is confirmed by the value of the lattice parameter of the metal of the layer a = 0.287 nm (a-iron has a = 0.28606 nm [2]) and by the fact that its magnetic properties are higher than those of the base metal.

These results are confirmed by the data of [3], where it is shown that steels with an elevated concentration of chromium and a diminished concentration of nickel and iron (as is observed in the surface layer) become ferritic in the entire temperature range in accordance with the Fe - Cr diagram. The appearance of the co-phase in the steel is due to the fact that the clearance preserved between the shells during the soldering plays the role of a capillary in which the excess molten solder accumulates. When the structure is heated to a temperature of 779~ a liquid phase appears. Since copper (a solder component) and nickel (a chemical element of the steel) form unlimited solid solutions [1], nickel diffuses into the melt. The chemical elements in the surface layer of the steel are redistributed, the content of iron and nickel in it/hlls to 68 and 1%, respectively, and the chromium concentration increases to 29%. The reduction in the iron content and the increase in the chromium content should be considered by proceeding from the tbllowing prerequisites. When the structure is heated, starting from the moment of the appearance of liquid solder, the diffusion of nickel into the melt is accompanied by a y---> a transformation in the surface layer of the steel. Another accompanying process is the segregation of a Cr23Cr carbide phase from the solid solution of the matrix (at t = 800~ its content in the steel is 1.21%). It can be assumed that the carbide concentration in the co-phase should be higher than in the y-phase. This assumption is based on two facts. Firstly, by the data of the phase analysis the surface layer of the steel after soldering the structure (in the zone of interaction with the melt) bears more ?'-phase than the bulk of the metal. Secondly, the conditions for the formation of Cr23C~, in the surface layer are more favorable than in the bulk of the steel because the solubility of carbon in the a-phase is much lower than in the y-phase [4]. Another factor promoting the appearance of an elevated amount of carbides in the a-phase seems to be the stresses, the level of which in the surface layer should be higher than in the bulk of the metal. It is known from [2, 5] that the stresses intensify these processes. At the same time, the segregation of carbides Cr23C6 from the solid solution depletes the a-phase of chromium and carbon. This assumption is based on the fact that the concentration of the segregated carbides in the a-phase is higher than in the y-phase. Since the system should be in thermodynamic equilibrium, the depleted a-phase takes these elements from the y-phase. The diffusion of chromium and carbon from the y-phase to the a-phase leads to the formation of vacancies in the y-phase. Their concentration in the y-phase also increases due to plastic deformation of the bulk layers of the metal, because at a high temperature it is not suppressed much. The presence of excess vacancies in the y-phase on the boundary with the a-phase leads to diffusion of iron atoms into the y-phase, which diminishes the iron concentration in

Effect of Structural lnhomogeneity of Material on the Strength of Soldered Joints

the surface layer o f the steel. The assumption on vacancy transfer of iron and chromium atoms agrees with the data of [4, 5]. However, when the soldering is performed at a temperature exceeding 800~ the carbide phase dissolves, and at t = 970~ its concentration does not exceed 0.17%. Theretore, in the range 800~ < t < 970~ the st-phase is considerably enriched with chromium and carbon. The appearance of the stress gradient in the st-phase is due to two factors. 1. When the structure is heated, complex transtbrmations occur in the surface layer of the steel at the boundary with the melt, namely, first an st - o y transformation and, starting at t = 779~ (the initial temperature of solder melting), a Y--~ c~ transtbrmation connected with the dissolution of nickel in the molten solder. It seems that the observed changes lead to phase hardening, 2. In the zone of contact with the melt, we observe diffusion o f chemical elements of the solder into the steel, which causes the distortion of the parameter o f the st-phase lattice. The effect o f st-iron and the chemical compounds on the strength, of the soldered joint is visually illustrated by the data o f Fig. 5. The role of the st-phase can be explained by the tact that copper, like silver (solder elements), does not form compounds with iron [1, 3], which are the base of the a-phase. In addition, silver is insoluble in solid iron, and the solubility of copper is about 1% [3]. With the second element o f the st-phase, i.e., chromium, the solder components do not form compounds below the soldering temperature either [6, 7]. The action o f the chemical compounds should be considered in light o f the fact that they are low-melting and, consequently, their strength is not high as compared with the metal obtained by diffusion intergrowth through copper-silver in the process o f soldering the structure in the absence of a clearance between the connected shells. It should also be noted that the decrease in the strength o f the soldered joint seems to be affected somewhat by the carbide phase too, because it is a brittle compound and its concentration in the st-phase is higher than in the y-phase. Thus, it can be concluded that all these factors complement each other and thus affect the strength o f bonding between the atoms on the interface o f the st-phase and the crystallized melt, which finally leads to fracture of the soldered structure at a lower test load than is specified. We established that in soldered structures without a clearance between the contacting surfaces a ferrite layer is not formed in soldering. The strength of such a joint is much higher than in the presence o f such a layer. This fact shows that in order to provide the requisite strength o f the joint we should prevent the appearance o f a ferrite layer in the steel shell when the structure is soldered. This can be done only if the clearance is removed in heating o f the shells. In this connection, we considered the creep of the shell material under the action o f excess pressure that occurs when

437

the structure is heated with the aim of creating conditions for removing the clearance. The pressure P was created inside the volume of the furnace by pressing the shells to each other. It is detennined by the fonnula p = P.~ +

[(a 2 - a l )(ts - t o ) - ~ c 2 - ~ r R 1+

-

-~;a]E262

-

Ej81

where Pc~ is the minimum contact pressure between the shells that provides an appropriate formation of the soldered joint, t is the soldering temperature, t 0 is the room temperature, eel and ~_, are the strains due the creep of the materials of the external and internal shells under the load, s~ = U/R is the proportion of the initial assembly clearance between the shells to the radius o f the shell, E I and E. are the moduli of elasticity of the materials of the external and internal shells, and 61 and 6, are the thicknesses o f the walls of the external and internal shells, respectively. The use of pressure in the stage o f heating the structure for soldering allows us to provide quality formation of the soldered joint and its satisfactory strength. With the introduction o f a hold and a designed pressure we observed no cases o f early failure of the structures. CONCLUSIONS 1. The inhomogeneity of the chemical composition that appears in steel on the boundary with the soldered joint markedly worsens the strength of the soldered joint. 2. In order to prevent early fracture o f the soldered s~ructure, we developed theoretical and experimental substantiation for designing the soldering process, which consists of allowance for the special features o f the design o f the structure and the mechanical, thermophysical, and rheological properties of the materials and the solder. REFERENCES 1. M. Hansen and K. Anderco, Structures ofBinal 3' Alloys [in Russian], Vol. 2, Gos. Nauch.-Tekh. Izd. Lit. po Chem. Tsvet. Metallurg., Moscow (1962). 2. V.N. Semenov, Liquid-Metal Embrittlement c?fHigh-Strength Alloys in Their Interaction with Copper-Silver Soldet:~. and a Soldering Technology Author's Abstract of Candidate's Thesis [in

Russian], Moscow (1987). 3. A. E Gulyaev, Metal Science [in Russian], Metallurgiya, Moscow (1977). 4. M. A. Krishtal, D([fusion Mechanisms in h'on Alloys [in Russian], Metallurgiya, Moscow (1972). 5. Ya. I. Frenkel', An bltroduction to the Theot3, of Metals [in Russian], Goz. Izd. Fiz.-Mat. Lit., Moscow (1958). 6. W. R. Hibbard, et al., Trans. ,,tIME, 175, 283 - 294 (1948). 7. A. T. Grigor'ev at al., Vestn. Mosk. Univ.. SeJ: Fiz.-Mat. Estestv. Nauk, No. 9, 77 81 (1954).