On Optimization of Some Parameters in Ultrasonic Metal Welding Understanding of the weld forming process leads to a way of optimizing some parameters in ultrasonic welding

B Y U . I . C H A N G A N D J . FRISCH

Nomenclature

A

Tangential displacement of the sonotrodetip vibration 1 0 3 i n .

A op Optimum tangential displacement of the sonotrode tip vibration, 10~3 Radius of the contact area, in. Radius of the inner boundary ot a slip annulus, in. Diameter of a sphere, or diameter of the disk specimen, in. Young's modulus, psi Constant Shear modulus, psi Normal load or w e l d clamping force, lb Electric power input to the transducer, watts Optimum electric power input, w a t t s Normal stress at the contact area, psi Maximum normal stress at the contact area, psi PmM = 3 / N 2 IT a 2 max S Tangential shear force, lb t Ultrasonic pulse time or weld time, sec X s M i n i m u m sonotrode tip displacement required for a fully developed slip annulus, 10 3 i n . X p Maximum allowable displacement for sublayer plastic deformation w h e n an optimum w e l d is produced, 1 0" 3 in. X A small sonotrode tip displacement by w h i c h a slip annulus is produced, 10 3 in. " Poisson's ratio, dimensionless P

ABSTRACT. The fundamental bonding mechanisms of ultrasonic welding are discussed and t w o basic bond forming processes are suggested. A n optimum w e l d condition for the electric power inputs was formulated using elastic and plastic analysis and a semianalytic expression for the optimum electric power input was thus obtained. The validity of the expression w a s checked by a series of experiments w i t h 2024-T351 aluminum and OFHC copper. A n expression for the coefficient of friction for alternating tangential motion was also obtained from this analysis. The strength characteristics of the welds are compared for different test conditions and failure modes. Introduction Ultrasonic welding is a method of joining similar or dissimilar metals by applying high frequency shear vibration and normal pressure to the w e l d interface. The mechanical energy transmitted to the w e l d area produces a sound metallurgical bond between t w o metals. The major advantage of this welding technique over conventional fusion joining processes is low heat input at the weld. Thin

ry Shear yield stress, psi T

y

Shear stress, psi Maximum allowable shear strain for sublayer plastic deformation at the weld, %

p. Coefficient of friction w h e n an oscillating tangential vibration is applied to the contact between an elastic sphere and an elastic flat, dimensionless

24-s | J A N U A R Y

1974

U. I. CHANG is associated with the Welding Development Department, Manufacturing Development Office of the Ford Motor Company. J. FRISCH is Professor of Mechanical Engineering, University of California, Berkeley, California. Paper was presented at the 54th A WS Annual Meeting held in Chicago during Apri/2-6, 1973

foils or wires can be welded to thick members and the unfavorable changes in material properties due to heat at or around the w e l d are less significant. Also, shrinkage and distortion problems are absent from this welding technique. This joining process is utilized in spot-type w e l d ing, ring welding, line welding, and continuous-seam welding (Ref. 1). For spot-type welding, the variables under control of the welding machine operator are tip radius, normal load, electric power input to the transducer and weld time. Proper adjustment of these variables is essential to minimize the required energy and optimize weld quality for any material combination. The vibratory frequency, w h i c h may range from 10,000 to 175,000 Hz, is determined by the welding machine design and is not believed to be critical (Ref. 1) in ultrasonic welding. The basic mechanism by w h i c h ultrasonic welds are produced is believed to be solid-state bonding (Refs. 1 -7). Static normal pressure and oscillating shear stresses at the w e l d interface result in localized slip at the weld interface and plastic deformation in a thin sublayer enveloping the interface. This process breaks up contaminant films and produces an area of metal-to-metal contact. Even though this joining technique is widely used in industry, the basic theory for this welding process is not completely understood because of the complexity involved in the formation of welds. Good welding practice often relies on a trial and error method even though some variables have an experimental equation for guidance. The objective of this study ic to develop clearer understanding of the bonding mechanism and processes, and to formulate an expression of variables for the optimum welding conditions. In order to eliminate the c o m plexity involved in a multi-interface problem found in commercial ultrasonic welding, the sonotrode tip in a conventional welding system was, for experimental purposes, replaced by a disk specimen w i t h a spherical radius as s h o w n in Fig. 1. This disk specimen was ultrasonically welded to a flat block specimen w h i c h was tightly fixed to a massive anvil. Such a configuration enabled direct application of vibratory energy to the weld interface, and simplified the analysis of weld formation.

cepted mechanism is solid state bonding some investigators have suggested that ultrasonic welding is another form of fusion welding activated by the heat generated through friction and plastic deformation, or at least it is a strongly heat-assisted welding process (Refs. 8-11). The localized temperature rise at the weld in ultrasonic welding is due to the combined effect of elastic hysteresis, localized interfacial slip, and plastic deformation. Temperature measurements made w i t h materials covering a wide range of melting temperatures show that the maximum average interface temperature w h e n good welds are produced ranges from 35 to 5 0 % of the absolute melting temperature of the material, suggesting no melting in the weld zone (Ref. 1). These observations strongly support the solid state bonding m e c h anism. Adhesion, one of the solid state bonding mechanisms, requires an intimate contact of the interface. The presence of surface films is detrimental to achieve atomically close contact of t w o metal surfaces. Surface films, especially oxide films, should either be removed or broken in such a way that clean metals be in contact. Bond strength then depends upon the areas w h e r e metal-to-metal contact is achieved. In ultrasonic welding of metals, the breaking of contaminant films for intimate metal-to-metal contact is accomplished by the combined alternating shear stresses at or around the weld, w h i c h result from the normal load and oscillating t a n gential force. The relative tangential displacement between a pair of contacting bodies can cause localized interfacial slip and sublayer plastic deformation around the interface if no gross sliding is assumed. Here " s l i d i n g " refers to the uniform movement or displacement of one contacting surface over another w h i l e " s l i p " is used for localized tangential displacement at the contacting surface. Gross sliding can occur w h e n the relative displacement is large enough or the frictional force is small enough to slide.

M e c h a n i s m of U l t r a s o n i c W e l d i n g

In ultrasonic welding, both localized slip and sublayer plastic deformation are desirable. The interfacial slip breaks up surface films allowing metal-to-metal contact at higher asperities and subsequently a large number of small bonded areas are formed over the entire contacting area (Ref. 4).

Ultrasonic welding of metals consists of interrelated, complex processes such as plastic deformation, work-hardening, breaking of contaminant films, fatigue, crack formation and propagation, fracture, generation of heat by friction and plastic deformation, recrystallization, and interdiffusion. Although the generally ac-

The plastic deformation in the sublayer enveloping the interface can occur w h e n the relative displacement is larger than that necessary to cause slip and the frictional stress is higher than the f l o w stress of sublayer material. If the frictional stress is lower than the f l o w stress, gross sliding w i l l occur. Defining p. as the coefficient of

SONOTRODE TIP

a. COMMERCIAL TYPE WELDING CONFIGURATION


O-

20 40 60 80 ELECTRIC POWER INPUT

20 40 60 80 100 120 ELECTRIC POWER INPUT - P (WATTS) C ) . NORMAL LOAD : 5 LBS.

100 120 P (WATTS)

d ) . NORMAL LOAD : 10 LBS. D - AIR 0.2 SEC. WELD TIME _ g _ AIR 1.0 SEC. WELD TIME

0.2 SEC. WELD TIME 1.0 SEC. WELD TIME

I I I I I I I I I ) 20 40 60 80 100 120 ELECTRIC POWER INPUT P (WATTS)

0 20 40 ELECTRIC POWER INPUT

0

AIR -0.2 SEC. WELD TIME A I R - 1 . 0 S E C . W E L D TIME

Q

i 9 |fl| ? i i i T " j i 3

b ) . NORMAL LOAD : 5 LBS.

e ) . NORMAL LOAD : 20 LBS.

6 do

QG£r•©••

f ) . NORMAL LOAD : 30 LBS.

AIR • 0.2 SEC. WELD TIME (TENSILE) AIR • 1.0 SEC, WELD TIME (TENSILE) AIR -1.0 SEC. WELD TIME (SHEAR) VAC. 1.0 SEC. WELD TIME (TENSILE)

OO-

A I R - 0 . 2 SEC. WELD TIME AIR -1.0 SEC. WELD TIME

I I 40- -

Q a

m u 20 5 S £ 30- -

UJ CC

I I I 20 40 60 80 ELECTRIC POWER INPUT

100 120 P (WATTS)

Fig. 11 — Tensile weld strength T351 aluminum

or shear weld strength

mum weld strengths, plotted as functions of normal load, are s h o w n in Fig. 12. Curves of optimum electric power input versus normal load are s h o w n in Fig. 13 w h i c h shows that the optimum normal load increases w i t h increased electric power input. These power values have been converted to corresponding tip displacements of the disk specimen by using 30-s I J A N U A R Y

1 974

l""l | I I I I I 20 40 60 80 100 120 ELECTRIC POWER INPUT • P (WATTS)

vs. electric

power

input for

2024-

the relationship s h o w n in Fig. 3, and these optimum tip displacements, as functions of normal load, are s h o w n in Fig. 14. Experiments w e r e performed to obtain slip annuli for different normal loads. To produce such slip annuli, electric power inputs ranging from 4 to 1 5 W, w h i c h correspond to tip displacements of 0.57 x 10 to 0.94 x

1 0 3 in., were applied. The b / a ratios of the slip annuli w e r e measured from the broken welds w i t h a microscope and the values of Xs were calculated using Eq. 7. Calculated values are listed in Table 1 and agree w i t h the observed values of X s . Some of the slip annuli are s h o w n in Fig. 15. For comparison, fully developed slip annuli made under different normal load or power input are s h o w n in Fig. 16. Calculations for the plastic displacement X p require a value for the allowable shear strain at the w e l d Y . In general, ductile materials should have higher Y than brittle materials. The analytical prediction of Y is difficult, if possible, because of high strain rates involved in a high f r e quency vibration, temperature rise, fatigue problems, size effect at the weld, and geometric constraint imposed by surrounding material. Therefore, this value has been obtained from one set of data (N = 3 0 lb) and is then used for the rest of the calculations. The computed value of Y was found to be equal to approximately one-half the typical tensile elongation (Ref. 22). One half of the radius of contact area was calculated for different normal loads and then multiplied by a constant value of Y to obtain X p . The optimum tip displacement A o p was obtained by adding X P to Xs as shown in Table 1. The calculated values of Aop's are plotted in Fig. 14 for comparison w i t h observed data. Equation 12 was also used to calculate Aop for OFHC copper welding. Observed optimum power inputs for OFHC copper w i t h 5 lb normal load were 25 W for 1.0 sec w e l d time and 35 W for 0.2 sec w e l d time (Ref. 19). A calculated value of Aop from Eq. 12 gave 1.05 * 10 3 in. for the optimum tip displacement. Again, the value of Y was taken as one-half of the reported elongation value (Ref. 23). Calculations are listed in Table 2.

Table 2 — Calct lated and Observed Values of O p t i m u m Sonotrode Tip Displacement for O F H C Copper Welding in Air

N, lbs 5 20

Power, W

Ampl., 10-3 in.

b/a

Xs c a t , IO"3 in.

2

0.47

0.316

0.52

8

0.73

1.00

0.73

Xpcal., 10-3 in.""

A0p cal., IO"3 in.

0.266

0.53

1.05

0.422

0.844

1.57

a/2 i0-2 in.(a)

"op

t, sec

obs., W

obs., 10 3 in.

0.2 1.0 1.0

35 25 60

1.10 0.97 1.58

"op

(a) For OFHC copper (Ref. 23): E = 1 7 * 10 fi psi.; elongation (in 2 in.) = 45 to 35%. (b) y = 2 0 % w a s used.

EXPERIMENTAL VALUES OF OPTIMUM TIP DISPLACEMENT Q 02 SEC. WELD TIME - O 1JJ SEC.WELD TIME

120- MATERIAL : 2024-T351 ALUMINUM WELD TIME : 0.2 SEC. H 1.0 SEC. O ENVIRONMENT : AIR

110- MATERIAL : 20Z4-T351 ALUMINUM WELD TIME : 0.2 SEC. • 1.0 SEC. O — ENVIRONMENT I AIR

100- 90- -

J

0.5- \ I I I

et-ef I I III

5 6 7 8 910

H

20

40 50

NORMAL LOAD • N (LBS.)

Fig. 12 — Optimum tensile weld strength as a function of normal load for 2024T351 aluminum in air

The values of p. were calculated by using Eq. 13 and are plotted for different applied normal loads in Fig. 17. These values of p. were obtained by substituting X s as listed in Table 1 into Eq. 13, and are found to be exceedingly high compared to the ordinary coefficient of friction for degreased 2024-T351 aluminum. Furthermore, the curve shows that p. is not a constant but a function of normal load N. A n experimental equation of p. as a function of the normal load for 2024-T351 aluminum was obtained from the curve s h o w n in Figure 17. (16)

Another expression of X s for 2 0 2 4 T351 aluminum has been obtained by substituting Eq. 1 6 into Eq. 6. XS = 1 8 . 8 ( N / E 2 d ) 1 / 3

(6-a)

From Eqs. 8, 6-a, and 10 A o p =[18.8