In
,a

‘11:

. a.

1.11).

raft-uh
3:,

L..-

. r T
‘ 1 . . ~FM. .5.
_ . 5m. . . u , .
.n: .7. . ‘

 

A8 30.5. ..| \l

    

I
3'

THEOtS

RAH

llllHlllllllllllllllllllllllﬂljljllll

31293 01771

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

An Investigation of Aging Kinetics and
Creep Deformation in Composite Solders

with In—Situ Particualte Reinforcements
presented by

Sunglak Choi

has been accepted towards fulfillment
of the requirements for

Master's Materials Science

degree in

 

 

ﬂwmsa

I
Major professor

0-7539 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earIier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

use W14

AN INVESTIGATION OF AGING KINETICS AND
CREEP DEF ORMATION IN COMPOSITE SOLDERS
WITH IN-SITU PARTICULATE REINFORCEMENTS

By

Sunglak Choi

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Materials Science and Mechanics

1997

ABSTRACT

AN INVESTIGATION OF AGING KINETICS AND
CREEP DEFORMATION IN COMPOSITE SOLDERS
WITH IN-SITU PARTICULATE REINFORCEMENTS

By

Sunglak Choi

The motivation for this study is to improve the reliability of lead-free solder joints in
electrical systems while removing lead. Solder microstructures under service conditions
at high homologous temperatures tend to be highly unstable, resulting in failure of solder
joints. This is caused by enhanced microstructural coarsening resulting from thermally-
induced or mechanical stress. To improve the reliability of solder joints as well as to
eliminate lead, lead-free in-situ composite solders based on eutectic Sn/Ag
(96.58n/3.5Ag), which is a potential candidate for replacing leaded solders, were
developed. A similar composite approach also was used with the eutectic Sn/Pb matrix
for comparison. The creep behavior of unaged and aged composite solder joints was
improved by a factor of 100-1000 as a consequence of the composite phase. Although the
creep resistance of aged joints were degraded, composite solders still exhibited better

creep behavior as compared to the corresponding non-composite ones.

To My Parents and Wife

Eunsang Choi, Oksun Lee, and Junghee Sung

iii

ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to my academic advisor and co-
advisor, Professor Thomas R. Bielerand Professor K.N. Subramanian, for thier guidance
and encouragement. I also wish to thank the members of my committee, Professor Dahsin
Liu and Professor Patrick Kwon, for their valuable suggestions and comments on this

work.

Particularly, I would like to express my appreciation to Composite Materials and
Structures Center for the ﬁnancial support to make this research possible. I would also
like to thank the technican in Physics Department, Reza Loloee, for his contribution in
making specimen, and Alan Gibson and Jeff Mcdougall for their helps as friends and co-

workers.

I would like to express my gratitude to my parents and wife, for their never-ending

love and suport during my graduate study.

iv

Table of Contents

CHAPTER 1 : INTRODUCTION .................................................................................... 1
CHAPTER 2 : REVIEW OF RELATED LITERATURE ............................................. 4
2.1 GENERAL CHARACTERISTICS OF SOLDERING ............................................ 4
2.1.1 Soldering ......................................................................................................... 4

2.1.2 Solder .............................................................................................................. 4

2.1.3 Flux ................................................................................................................. 5

2.1.4 Substrate .......................................................................................................... 7

2.1.5 Solder Joint ..................................................................................................... 8

2.1.6 Service Conditions of Solder Joints .............................................................. 11

2.2 SOLDER MATERIALS ........................................................................................ 14
2.2.1 Tin/Lead Solder ............................................................................................ 14

2.2.2 Tin/Silver Solder ........................................................................................... 21

2.2.3 Composite Solder .......................................................................................... 26

2.3 CREEP BEHAVIOR OF SOLDERS .................................................................... 32
2.3.1 The Description of Creep Deformation ........................................................ 32

2.3.2 Creep Mechanisms ........................................................................................ 34

2.3.3 Deformation behavior of Solder Joint ........................................................... 36

2.3.4 Creep Deformation of Sn/Pb Solder ............................................................. 37

2.3.5 Creep Deformation of Sn/Ag Solder ............................................................. 42

2.4 FATIGUE BEHAVIOR OF SOLDERS ................................................................ 44
2.4.1 Fatigue Deformation of Sn/Pb Solder ........................................................... 44

2.4.2 Fatigue Deformation of Sn/Ag Solder .......................................................... 48

2.5 STRATEGY FOR DEVELOPING AN IN-SIT U COMPOSITE SOLDER .......... 49
CHAPTER 3 : EXPERIMENTAL PROCEDURE ....................................................... 51
3.1 MATERIALS ......................................................................................................... 51
3.1.1 Solder Material .............................................................................................. 51

3.1.2 Substrate Material ......................................................................................... 52

3.1.3 Flux ............................................................................................................... 52

V

3.2 THE PREPARATION OF SOLDER JOINTS ...................................................... 55

3.2.1 The Pre-Treatment for Manufacturing a Solder Joint ................................... 55

3.2.2 Soldering Process .......................................................................................... 55

3.3 AGING OF THE SOLDER JOINTS ..................................................................... 59

3.4 CREEP TESTING OF COMPOSITE SOLDER JOINTS ..................................... 62
CHAPTER 4 : RESULTS ............................................................................................... 65
4.1 MICROSTRUCTURAL OBSERVATION OF BULK MATERIALS ................. 65
4.1.1 Eutectic Sn/Pb (63Sn/37Pb) and Sn/Ag (96.58n/3.5Ag) Solders ................. 65

4.1.2 Slow Cooled Eutectic Sn/Pb and Sn/Ag Composite Solders ........................ 70

4.1.3 Remelted Composite Solders ........................................................................ 74

4.2 MICROSTRUCTURES OF COMPOSITE SOLDER JOINTS ............................ 76

4.3 AGING OF COMPOSITE SOLDER JOINTS ...................................................... 81
4.4 CREEP DEFORMATION OF COMPOSITE SOLDER JOINTS ........................ 88
4.4.1 Scratch Technique for Shear Strain Measurement ........................................ 88

4.4.2 Eutectic Sn/Ag Composite Solder Joints ...................................................... 91

4.4.3 Eutectic Sn/Pb Composite Solder Joints ....................................................... 95

4.4.4 The Fracture Surfaces of Composite Solder Joints ....................................... 97
CHAPTER 5 : ANALYSIS AND DISCUSSION ........................................................ 102
5.1 AGING OF COMPOSITE SOLDER JOINTS .................................................... 102

5.2 CREEP DEFORMATION OF COMPOSITE SOLDER JOINTS ...................... 112
5.2.1 Eutectic Sn/Ag and Sn/Pb Composite Solder Joints ................................... 112

5.2.2 The Fracture Surfaces of Composite Solder Joints ..................................... 117
CHAPTER 7 : CONCLUSIONS .................................................................................. 118
BIBLIOGRAPHY .......................................................................................................... 120

vi

List of Tables

Table l Creep Deformation Parameters for Eutectic Sn/Ag Composite Solder Joints ...... 94
Table 2 Creep Deformation Parameters for Eutectic Sn/Pb Composite Solder Joints ...... 96
Table 3 Relative Diffusivities of elements in Solders ..................................................... 105

Table 4 Aging Parameters For Eutectic Sn/Pb Composite Solder Joints
Used in First and Second Aging ......................................................................... 109

vii

List of Figures

Figure 1 Schematic Diagram of Typical Joint Configuration
(a) Die or Substrate Bonding (b) Solder Bump

(c) Leadless Chip Carrier (LLCC) (d) Leaded Chip Carrier attachment ............. 10
Figure 2 Sample Geometries for the Testing of Solder Joint

(a) Single Shear Lap Joint, (b) Double Shear Lap Joint, (c) Butt Joint ............... 10
Figure 3 Schematic Diagram of Thermally Induced Strain in Solder Joint ....................... 13
Figure 4 The Service Condition for Solders ...................................................................... 13
Figure 5 Sn/Pb Phase Diagram .......................................................................................... 16
Figure 6 Cu/Sn Phase Diagram .......................................................................................... 19
Figure 7 Sn/Ag Phase Diagram .......................................................................................... 22
Figure 8 Schematic Diagram of Induction or Radio Frequency (RF) Furnace .................. 53
Figure 9 The Dimensions of The Cu Substrate .................................................................. 54

Figure 10 Schematic Diagram of
(a) The Solder Joint Fixture Assembly and
(b) A Fabricated Single Shear Lap Joint ............................................................ 57

Figure 11 The Temperature versus Time Proﬁle for Manufacturing of
(a) Eutectic Sn/Pb Composite Solder Joints and

(b) Eutectic Sn/Ag Composite Solder Joints ...................................................... 58

Figure 12 The Schematic Diagram of the Temperature-Time History
in Thermal Cycling ........................................................................................... 60
Figure 13 The Creep Fixture .............................................................................................. 63
Figure 14 Optical Microstructure of Eutectic Sn/Pb Solder Ingot ..................................... 68
Figure 15 Optical Microstructure of Eutectic Sn/Ag Solder Ingot .................................... 68

Figure 16 Microstructural Variation of Eutectic Sn/Pb Solder
Due to Variation of Cooling Rate During Splat Cooling .................................. 69

Figure 17 Microstructural Variation of Eutectic Sn/Ag Solder
Due to Variation of Cooling Rate during Splat Cooling ................................... 69

viii

Figure 18 Microstructure of Slow Cooled Eutectic Sn/Pb Composite Solder
(a) Can, Intermetallic n—phase Morphology (b) Eutectic Sn/Pb Matrix ........ 72

Figure 19 Microstructure of Slow Cooled Eutectic Sn/Ag Composite Solder
(a) Cu,Sn, Intermetallic n-Phase Morphology (b) Eutectic Sn/Ag Matrix ....... 73

Figure 20 The Modiﬁed Morphology of Cu(,Sns Intermetallic n-Phase due to Splat
Cooling in (a) Eutectic Sn/Pb composite solder and
(b) Eutectic Sn/Ag Composite Solder ................................................................ 75

Figure 21 SEM Microstructure of Eutectic Sn/Pb Composite Solder Joint
Containing Cu6Sn, Intermetallic n-Phase Particles ........................................... 77

Figure 22 SEM Microstructure of Eutectic Sn/Ag Composite Solder Joint
Containing Cu,,Sn5 Intermetallic n-Phase Particles ............................................ 77

Figure 23 The Temperature versus Time Proﬁle for Manufacturing and
Corresponding Microstructure of Eutectic Sn/Pb Composite Solder Joints
Used in (a) First Aging and (b) Second Aging .................................................. 82

Figure 24 The Microstructure of Eutectic Sn/Pb Composite Solders
after Static Aging at (a) 50, (b) 80, (c) 100, (d) 120°C for 100 hours ............... 83

Figure 25 The Microstructure of Eutectic Sn/Pb Composite Solder Joints
after 20 Aging Cyclies (20 Hours at Aging Temperature)
(a) 20 Aging Cyclies between RT and 60°C (20 Hours at 60°C)
(b) 20 Aging Cyclies between RT and 120°C (20 Hours at 120°C) ................... 84

Figure 26 The Microstructural Change of Eutectic Sn/Ag Composite Solders
after Static Aging at (a) 60, (b) 100, (c) 120, (d) 150°C for 100 hours ............. 85

Figure 27 The Microstructure of Eutectic Sn/Ag Composite Solder Joints
after 20 Aging Cyclies (20 Hours at Aging Temperature)
(a) 20 Aging Cyclies between RT and 100°C (20 Hours at 100°C)
(b) 20 Aging Cyclies between RT and 150°C (20 Hours at 150°C) ................... 86

Figure 28 Two Scratches Across The Eutectic Sn/Ag Composite Solder Joint
(a) before loading (b) after creep deformation at 21.7 MPa for 374 ksec
(c) shear strain distribution across the solder joint ............................................. 90

Figure 29 Creep Curve of Average Shear Strain versus Time
for Unaged Eutectic Sn/Ag Composite Solder Joints ........................................ 92

Figure 30 Creep Curve of Average Shear Strain versus Time
Aged Eutectic Sn/Ag Composite Solder Joints ................................................. 94

ix

Figure 31 Creep Curve of Average Shear Strain versus Time
for Aged Eutectic Sn/Pb Composite Solder Joints ............................................ 96

Figure 32 The Fracture Surface Showing Pores in
(a) Eutectic Sn/Ag Composite Solder Joints Aged at 100°C
(b) Eutectic Sn/Pb Composite Solder Joints Aged at 100°C .............................. 98

Figure 33 SEM Micrograph showing
(a) The Smeared Matrix in Eutectic Sn/Ag Composite Solder Joint
and EDAX Result
(b) The Cu6Sn5 Intermetallic Particles in Fracture Surface and EDAX Result ...99

Figure 34 SEM Micrograph Showing Side View of Fracture Surface of

Eutectic Sn/Ag Composite Solder Joint Aged at 60°C For 100 Hours ............ 101
Figure 35 Arrhenius Plot for Aging of Eutectic Sn/Pb Composite Solder ...................... 104
Figure 36 Arrhenius Plot for Aging of Eutectic Sn/Pb Composite Solder ...................... 108
Figure 37 Arrhenius Plot for Eutectic Sn/Ag Composite Solder ..................................... 111

Figure 38 The Normalized Steady-State Behavior
Eutectic Sn/Ag Composite Solder joints ......................................................... 116

Figure 39 The Normalized Steady-State Behavior
Eutectic Sn/Pb Composite Solder joints .......................................................... 116

CHAPTER 1

INTRODUCTION

Soldering is one of the oldest and most widely used metallurgical joining techniques.
Soldering is technically deﬁned as joining two base materials through the use of a third

“ﬁller” metal with a melting temperature below 425°C (800°F) [1, 2].

The earliest recorded evidence of soldering was found in Mesopotamia in 4000 BC.
when the gold-based hard solders were developed. Use of these solders was also found in
Egypt (3600 B.C.), Ur (3400 B.C.), Greece (2600 B.C.), and other Mediterranean regions.
The ﬁrst mention of soldering is 3000 years ago in the Bible (Isaiah, 41:7) [3]. The ﬁrst
use of soft Sn/Pb solders historically appears to be about 1900 BC. It was the Celts and
the Gauls in Northern and Central Europe that developed soft soldering. It was noted by
the ﬁrst Romans to enter the region that the Celtic pieces with high quality were
constructed with tin-lead soﬁ soldering. The Romans borrowed the expertise of the Celts
on soft soldering technique and used tin/lead solders extensively in their construction of
aqueducts. The soft soldering was continuously used since then as a joining method in
fabrication of many crafts, plumbing, water container, and larger structures (e.g.,
automobile radiators). It was in the early 20th century that soldering technology entered
the electronic industry. In those early applications, the soldering technology served a
reliable means to provide electrical continuity between components. The electronic

revolution has promoted ﬁrrther miniaturization and increased performance of devices

with the invention of silicon-chip technology. Soft soldering has been adapted to

microchip packaging and higher level system assembly [2].

The most common solder materials are tin/lead solders. These solders have been used
to join metals for at least a millennium, and they have been the most outstanding material
for the interconnection and packaging of modern electronic components and devices
during the past several decades. The prevalent use of tin/lead solders is due mainly to

their low cost and unique combination of material properties [4].

Lead-bearing solders are categorized as either tin-based tin/lead solders or lead-based
lead/tin solders. Tin-based tin/lead solders comprise the largest group of alloys used in
electronic industry. The typical tin-based tin/lead solders are eutectic tin/lead
(63Sn/37Pb) and near eutectic tin/lead (6OSn/40Pb) that are used most extensively in
circuit-board assembly because their melting temperatures are low enough that processing
can be compatible with printed wiring board materials and electronic devices. Lead-based
lead/tin solders are not used in printed wiring board assembly because they show poor
ﬂow and their higher melting temperatures are incompatible with printed wiring board
materials. Within that group, 95Pb/5 Sn and 90Pb/ 10Sn are used in the controlled collapse

chip component (C‘) attachment of silicon chips to package boards [2].

However, in the light of environmental and health concerns, there have been
increasing research efforts during the past few years to develop lead-free solders to
replace the Sn/Pb alloys [5]. Lead and lead-bearing compounds are among the most toxic

substances, and the environmental and health issues concerning the toxicity of Pb

threaten to ban the use of lead-bearing solders in electronic applications. These concerns
have inspired a great deal of research efforts in the development of suitable Pb-free solder

alloys.

The potential candidates of lead-free solders which have been investigated are
eutectic Sn/Bi (4ZSn/58Bi), eutectic Sn/Ag (96.58n/3.5Ag), eutectic Sn/Zn (9ISn/9Zn),
and eutectic Sn/In (48.3Sn/51.7In). Even though some lead-free solders have been in use
in special applications for several years (e.g., eutectic Sn/Bi (428n/58Bi) for step
soldering which illustrates the use of progressively lower melting point solder alloys in
subsequent assembly levels to preserve the integrity of the previously solidiﬁed joints [6],
and eutectic Au/Sn (80Au/ZOSn) for creep-resistant soldering in optical-ﬁber alignment

and packaging [6] ), there is no obvious substitute for lead/tin solder alloys at present.

Miniaturization in the microelectronics industry and the widespread use of surface
mount technology (SMT) which was introduced in the 19608 have demanded higher
performance on solders. Solders have to provide structural function by mechanically
supporting the electronic devices in addition to playing a role of electronic contacts
between printed circuit substrates and surface mount devices [7]. The mechanical
properties of solders have become more important because of the small-size of solder
joints in high density surface mount device (SMD) packages. Thus the development of
alternative lead-free solder with superior mechanical properties to Sn/Pb solder to
improve reliability of solder joint while removing lead in electronic packaging appears to
be desirable. This thesis describes the development of in-situ composite solders and

preliminary evaluation of improvements of properties.

CHAPTER 2

REVIEW OF RELATED LITERATURE

2.1 GENERAL CHARACTERISTICS OF SOLDERING
2.1.1 Soldering

Soldering is deﬁned as “a joining process by which two substrates are joined using a
ﬁller metal (solder) which forms strong bonds with the ﬂux on solidiﬁcation [1]. The
substrate materials remain solid during the joining process and the solder is distributed
between the properly ﬁtted surfaces of the joint by capillary action. The fundamental
feature of the soldered joints is that a metallurgical bond is produced at the
solder/substrate interface. The soldering temperatures in processing typically range from
30 to 50°C (55 to 90°F) above the liquidus temperature to ensure appropriate heating of

the substrate and sufﬁcient ﬂow of the solder [8].

2.1.2 Solder

Solder is usually deﬁned as fusible alloys with a liquidus temperature below 425°C
(800°F) [9]. Solders are typically binary or ternary alloys. Most solder alloys are eutectic

or near eutectic to take advantage of a low melting point.

The melting temperature of solders is a critical characteristic in practical joining
processes because it determines the minimum processing temperature and the maximum
operating temperature of the system. The mechanical and physical properties of solder

such as strength, creep-fatigue resistance, thermal expansion coefﬁcient or conductivity,

and corrosion are also of importance because the solder provides a means of fastening
two components together. When a molten solder is used in processing, its surface tension

is another important factor in determining its wettability with the substrates [10].

The solder can be typically applied in two different ways, (i) coating the surface, and
(ii) an application of discrete amount. The common types of discrete solder placement are
solder paste and preforms. The solder paste refers to a homogeneous mixture of
prealloyed solder power and ﬂux in a composition that is a stable suspension and suitable
for dispensing. The solder preform is deﬁned as a shaped form of solder that has a
determined amount of metal with or without a ﬂux core of ﬁxed size. These special forms
of solder are used when proper amounts of the solder is necessary at individual joints to

form a satisfactory joint [11].

2.1.3 Flux

Fluxing is used to produce a reliable bond between solder and substrate by improving
the wetting properties of the solder/substrate system. Flux has signiﬁcant inﬂuence on the
wetting behavior of solder with substrates that determines the solderability of the solder,
so the selection of ﬂux and the development of new ﬂuxes for a speciﬁc soldering

application are important [12].

The major function of flux is to remove the thin tarnish layers on the surface of the
substrates during the initial steps of the soldering process, allowing the molten solder to

spread and react with the substrate. The additional roles of the ﬂux are that it lowers the

surface tension of the solder, so that it more readily ﬁlls gaps and holes by capillary

action and protects the metal surface from reoxidation during the heating stages.

The principal components of the ﬂux are active chemical components such as halides
for oxide removal, wetting agents to enhance surface spreading, and a vehicle to dilute
and mix the cleaning compound and wetting agents together. The vehicles are usually
water, isopropyl alcohol, glycerin, glycol (for liquid ﬂuxes), and petroleum jelly (for ﬂux

pastes or cream).

Fluxes are categorized by their cleaning agent as following : rosin-base ﬂuxes,
organic-acid ﬂuxes ( also called intermediate or water-soluble ﬂuxes), and inorganic-acid
ﬂuxes. The most common ﬂuxes are rosin-based in the electronic applications. The rosin-

based ﬂuxes are divided into three classes according to their activity [13] :

0 RA - rosin base, fully activated ﬂuxes with a strong ionic activator such as halide,

o RMA - rosin base, mildly activated ﬂuxes containing lower concentrations of halides
or activated with organic acids (halide free), and

o R - rosin base ﬂuxes without any activators, relying on the very weak organic acids in

natural rosin for the ﬂuxing action.

2.1.4 Substrate

Substrate is a base material on which electronic devices are soldered using a ﬂux. The
surface layer of a substrate consists of the so called solderable layer to which the molten
solder metallurgically reacts and the additional coating or protective layer to preserve
solderable layers from excessive oxidation on the solderable surface during atmospheric
storage, prevent dissolution of the base metal, provide a good wettable surface for

soldering [14].

The solderable surface can be either the substrate itself, usually copper, or coatings of
elemental metals such as nickel or copper deposited by electroplating, electroless plating,
evaporation, pretinning (solder dip coating), sputtering, chemical vapor deposition
(CVD). The protective layer is an electroplated ﬁlm that is entirely absorbed into the
solder during soldering process. Protective layers are precious metals such as gold [15].

Less protective layers are electroplated tin and electroplated tin-lead solder.

Intermetallic compounds normally form at the interface between the solder and the
substrate. Intermetallic reactions typically proceed much more slowly on Ni than on Cu
substrates [16]. Sn/Pb solder system forms Cu/Sn intennetallic compounds (Cu6Sn5 and
Cu3Sn) on Cu substrate [17]. Other solder systems, such as Sn/Ag, Sn/Cu, and Sn/Bi, all
form the same intennetallic compounds as Sn/Pb solder on both Cu [18, 19] and Ni
substrates (Ni38n4, Ni3Sn2 intennetallic compounds) [20]. Eutectic Sn/In solder is
different from other Sn-based solder since ternary compounds are formed on Cu

substrate, which are Cuzln3Sn on the solder side and Cu2(Sn, In) on the Cu side. On Ni

substrate a ternary In/Sn/N i intennetallic forms whose stoichiometry has not been

determined [16, 21].

The dissolution of the substrate into the solder when molten solder is in contact with
substrate is another concern in electronic applications because the dissolved metal may
adversely affect the solder microstructure and material properties [22]. Such dissolution
of the substrate can (a) alter the alloy composition which may raise or lower its melting
temperature, (b) alter the microstructure [16] and/or (c) cause dissolved metal to come out

of solution in the form of undesired precipitates during the cooling process [15].

The dissolution of the substrate should be avoided to prevent degradation of solder
joints. Foster [15] found that presence of gold in solder matrix forms gold/tin
intennetallic compound which deteriorates the mechanical properties of solder alloy and
recommended most of gold should be removed from the joint area. The removal of gold
could be achieved by abrasion or by dissolving the gold in molten solder. Goldstein and
Morris [16] reported that the Cu dissolution in the solder joint could be avoided by
additional coating of nickel on Cu substrate since Ni diffuses much more slowly than Cu

into the solder joint.

2.1.5 Solder Joint

Solder joints consist of three main components : the electronic device, solder, and the
substrate. A variety of solder joints geometries in electronic applications have been
developed. The selection of the connection type is determined by manufacturing,

electronic and service considerations. Schematic examples of typical joint conﬁguration

in electronic products are shown in Figure l [23, 24], and sample geometries that are used

for the testing of solder joints are illustrated in Figure 2 [25, 26].

The two major classes of soldering technologies are through-hole technology and
surface-mount technology. Through-hole technology refers to the soldering of the leaded
device packages through a hole on substrate. The leads are inserted into the holes in the
printed wire board (PWB) and are soldered in place. In the surface-mounted technology
smaller components are soldered directly to the surface of the printed circuit board
(PCB). Surface-mounted technology made it possible for components to be mounted on

both sides of the board.

10

 

 

 

 

 

 

 

(a) (b) (C)

 
 
 

 

 

 

Gullwing

 

 

 

((1)

Figure 1 Schematic Diagram of Typical Joint Conﬁguration
(a) Die or Substrate Bonding (b) Solder Bump
(c) Leadless Chip Carrier (LLCC) (d) Leaded Chip Carrier attachment

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) (b) (C)

Figure 2 Sample Geometries for the Testing of Solder Joint
(a) Single Shear Lap Joint, (b) Double Shear Lap Joint, (c) Butt Joint

11
2.1.6 Service Conditions of Solder Joints

Soldering is by far the most economic and popular method to interconnect
components to printed circuit boards or chips to substrate. With the advent of surface
mounting technologies in 19603, the miniaturization and functional integration of
electronic modules have been pursued much more. As a consequence, solder joints
became structural components by mechanically supporting the electronic devices [7], in
addition to providing electrical connection between surface mounted devices and
substrates. Consequently solder joints in electronic applications experience more severe

conditions that critically affect the deformation behavior and reliability of solder joints.

The principal sources of deformation in service conditions are thermal ﬂuctuations
and mechanical vibrations. The thermal ﬂuctuations stem from the heat generated within
the devices themselves and temperature changes from the external operating
circumstances [1]. When an electronic module is subject to thermal ﬂuctuation in service,
thermal strains which are cyclic in nature are induced in the joint due to the difference in
the coefﬁcients of thermal expansion produced by the various materials used. Since the
solder is usually much softer than the other joint components, most of deformation

generated is concentrated within the thin layer of solder joint.

The thermally-induced strain can be depicted using the schematic diagram in Figure 3
[24, 27]. Assumptions which should be applied in a simple model are that there are

neither temperature gradients nor transients, and that the solder is fully pliant, or relaxes

12

fully. If the temperature changes from To to T, this will then cause a thermal shear strain 7

distributed uniformly through the solder joint.

(1;..13
2

..T)

 

a

7 _ H
y is the thermally induced shear strain, AOL (= or, — a2) is the difference in expansion
coefﬁcients between printed circuit board and component, AT is the amount of

temperature change, L is the distance between the pinning points, and H is the height of

the joint.

Solders have inherently low melting temperatures which are critical for solder
processing and operation in electronic applications. These low melting temperatures of
solders give a disadvantage in service conditions, or even at room temperature which
corresponds to a high homologous temperature. The homologous temperature of a
material refers to the ratio of a given temperature to the melting temperature in degrees K
[28]. The conditions for solders are compared to high temperature structural alloys in
Figure 4 [29]. These service conditions for solder consequently mean that time dependent
effects such as creep, stress relaxation, surface and microstructural changes should be

considered in the analysis of deformation behavior, in addition to cyclic phenomena.

13

Surface Mount Component

 

 

 

 

 

 

 

 

 

 

 

 

 

a, > or,
Solder‘Joint
2 PCB 2
i '1 Li Tn >To
i i 01» Tu i i
I i i I I i
8 ' I 07, TI E I I 2
l I I 1

Figure 3 Schematic Diagram of Thermally Induced Strain in Solder Joint [24, 27]

Homologous Temperature, TH(=T/Tm)

 

 

 

 

1.0
0'9“ Eutectic Sn/Pb Solder
0.81— : 3“?“ “MAL
' Eutectic .
95Pb/5Sn Au/Si Allo N' Alloy C0 0y
i :1 v 1
0.6-
Stainless Steel
0.5— f
0.4— I '
Silver'Solder
0.3—-
J 1 1 1 1 1 1 1 )1 1 1
300 400 500 600 700 800 900 1000 “1700 1800
Melting Point of Alloy, Tm(K)

Figure 4 The Service Condition for Solders [29]

14
2.2 SOLDER MATERIALS
2.2.1 Tin/Lead Solder

Lead-bearing solders consist of the largest group of alloys used in electronic products.
Within that group, the tin/lead eutectic (63 Sn/37Pb) and near eutectic (60Sn/40Pb)
solders are used most extensively in circuit-board assembly. The lead-rich lead/tin solders
(SSn/95Pb, IOSn/90Pb) are commonly used at relatively high temperature applications
(100°C to 250°C) such as for joining sections of automotive radiators. The Sn/Pb solders
with high melting temperature are also used in controlled-collapse chip connection (C4)

technology to attach integrated-circuit chips to packages for use in high speed computers

[30].

The tin/lead solder has been widely used in electronic applications because of the
several advantages of the tin/lead solders. The eutectic tin/lead(63Sn/37Pb) and near
eutectic (60Sn/40Pb) solders have sufﬁciently low melting temperature making them
compatible with printed wired board materials and electronic devices during processing.
Lead is relatively inexpensive and there are abundant resources of lead located in the
United States [2]. Lead as an impurity in the tin/lead solders also prevents the white ([3)
tin to gray (or) tin allotropic transition when cooling through 13°C which is referred to as
“tin pest” because of the loss of structural integrity by tin in cold service [31]. Lead also
lowers the surface tension in tin-lead solders relative to pure tin [32, 33], thereby
improving the solderability of circuit boards and devices by facilitating the wetting of the

solder. The wettibility of 6OSn/40Pb solder appears to be better than any other solder

15

alloys in terms of lower contact angle (9C), faster wetting rate, and shorter time required

to form joints [34].

According to the phase diagram of tin/lead [3 5], the binary system of tin/lead shown
in Figure 5 is a simple eutectic which has lead-rich or—Pb phase and tin-rich B—Sn phase.
The tin-rich B—Sn phase has negligible equilibrium solubility for lead at room
temperature. The lead-rich a—Pb phase can contain 1~2 wt.% tin at room temperature and
the amount of Sn soluable in the lead-rich or—Pb phase increases considerably as
temperatures are elevated. The solubility of Pb in B—Sn phase also increases appreciably

with temperature.

Upon solidiﬁcation, a two-phase microstructure is developed, the morphology of
which is dependent upon the cooling rate. Slow cooling rate produces a lamellar
microstructure of alternate layers of lead-rich and tin-rich phases. Volumes of lamellae of
similar orientation are known as colonies (or grains). With increasing cooling rate, the
colony size and interlamellar spacing decrease and with a high cooling rate, a ﬁne

spheroidal structure is obtained.

350

16

Atomic Percent Sn

0 10 20 30 4O 50 60 70

l l l

80

I

100

 

300 -

N N
O LII
O C
1 1

Temperature °C
3
O

100%

50-

 

0

. 3-7.502'C

 

 

18.3

61.9

97.8

(13311)

 

 

0
Pb

I

10

Figure 5 Sn/Pb Phase Diagram [35]

I

20

I

30

40 50 60
Weight Percent Sn

I

70

80

90

100
Sn

17

The microstructure of Pb/ Sn alloys has great inﬂuence on deformation behavior and
mechanical properties. The microstructure of Pb/Sn solder alloys formed by fast cooling
rate exhibit superplastic behavior in creep tests as a consequency of the ﬁne globular
microstructure, grain boundary sliding at low stress and low strain rate and reduced
strength, increased ductility, and a signiﬁcantly increased the fatigue life by delaying
crack initiation and growth [36, 37]. These desirable effects on mechanical behavior and
properties result from ﬁne and equiaxed microstructure and uniform distribution of

eutectic phases which can be obtained from the fast cooling rate on solidiﬁcation.

The microstructure of tin/lead alloys has the inherent microstructural instability which
arises from high solid solubilities of the constituent elements within the two phases
combined with service at high homologous temperature. Due to the high homologous
temperature (T/Tm = 0.64 at 20°C) of tin-lead alloy even at the room temperatures, it is
subjected to the precipitation of tin colonies within the lead-rich phase and overall
coarsening of the lead-rich and tin-rich phases [38]. The microstructural changes due to
precipitation and coarsening result in degradation of mechanical properties such as

strength, ductility [39, 40], and microhardness [30].

The microstructural instability of eutectic and near eutectic Sn/Pb solder is also
observed under thermal fatigue (thermal cycling), isothermal fatigue (mechanical
cycling), and creep in shear at high temperature which occur in solder joints [41, 42]. The
combination of shear strain and high temperature in these deformations creates regions of
concentrated plastic deformation, which result in an inhomogeneously coarsened band

similar to the recrystallization and grain growth, within the solder joints. This

18

heterogeneously coarsened band which is the weakest region becomes a source of crack

initiation and propagation in solder joints.

Although the majority of the investigations have revealed that failures in the joints
occur through the solder via a heterogeneous coarsening of the microstructure, some of
the studies performed had failures occurred through the interface with intermetallic layers
present in solder joints after exposure to high homologous temperature during
deformation [43, 44]. It was also found that the joint strength was decreased as the
thickness of the interfacial intermetallic layers increased [45]. Because the intermetallic
layers tend to be brittle and the strain is induced in the joint due to the differing
coefﬁcients of thermal expansion between the solder and substrate, the reliability of a

solder joint can be degraded by the presence of interfacial intermetallics.

For the Cu substrate with Sn/Pb solder, Sn reacts with Cu forming Cu,,Sn5 n-phase
which has a hexagonal crystal structure [46] adjacent to solder, and Cu3Sn e-phase which
has a pseudo-hexagonal crystal structure [47] next to the Cu substrate as shown in Figure
6 [48]. Cu68n5 n-phase is formed during solidiﬁcation but Cu3Sn e-phase is found to form
during aging after solidiﬁcation. The interfacial intermetallic layers grow faster when the
solder is molten and will also continue to grow after the joint has solidiﬁed. The Cu3Sn 8-
phase was considered to grow at the expense of Cu,,Sn5 n-phase [49, 50]. The mean total
thickness of the interfacial intennetallic layers was observed to increase linearly with the
square root of aging time and the growth was faster for higher annealing temperature,

suggesting that the growth of interfacial intennetallic layers is controlled by the volume

diffusion

l9

. Atomic PercentTin
0 10 20 30 40 50 60 70 80 90 100
12m L I l l l l l l l

 

 

 

 

 

 

 

 

 

 

 

 

 

0102030405060708090100
Cu WeightPercentTin Sn

Figure 6 Cu/Sn Phase Diagram [48]

20

mechanism and the rate is controlled by the interdiffusion of active species, i.e. Cu and

Sn [50].

The growth of the interfacial intennetallic layers appeared to alter the fracture
behavior, wherein at as-soldered condition the joint failed through the solder but after
aging, the joint showed brittle fracture through the interfacial intermetallic layers with
increased thickness of the intermetallic layer. The failure behavior and strength of solder
joint was very sensitive to the thickness of Cu68n5 interfacial layer [17, 51]. The growth
of the interfacial intermetallic layers was observed to be suppressed in the presence of Ni

in the Sn-Pb solder by increasing the activation energy for the solid state growth [50, 52].

The Cu-Sn intennetallic compound was also found in the Sn—Pb solder matrix in the
joint [53]. Cu,,Sn5 n-phase precipitates in rod or whisker form were observed in the Sn
grains of solder joints. Cu diffused through the interfacial Cu68n, n-phase layer and
dissolved into the solder during reﬂow of the solder joint. The Cu,,Sn5 n-phase
precipitated from the supersaturated solid solder with Cu which was produced by further
cooling because of the decreasing solubility of Cu in the solid solder with decrease in
temperature. The bulk solder with Cu68n, n-phase showed a smaller elongation at failure

and higher work hardening rate than the pure solders [54].

2 1
2.2.2 Tin/Silver Solder

Tin/silver solder, especially eutectic tin/silver (96.SSn/3.5Ag), is being considered as
a potential lead-free candidate because tin/silver solder is basically free of toxic lead and
it is better for high temperature applications due to its high melting temperature, and
silver is available in sufﬁcient quantities to be considered as minor alloying element [55,

56].

Like eutectic lead/tin solder (63Sn/37Pb), eutectic tin/silver solder (96.SSn/3.5Ag)
consists of two phases which are essentially pure tin-rich base-centered tetragonal (bct) [3-
phase with less than 0.04% Ag in volume fraction, and Ag3Sn intennetallic orthorhombic
short-range ordered s-phase with less than 5% in volume fraction [56, 57]. The Sn-Ag
phase diagram shown in Figure 7 [57] indicates virtually no solid solubility of Ag in Sn-
rich phase even at temperatures near the melting point. Eventually all the Ag in the

96.5 Sn/3.5Ag resides in the Ag3Sn intennetallic compound.

The microstructure of 96.58n/3.5Ag might be altered depending on the soldering
process, process parameters, thermal history after soldering, alloying elements, and
deformation. In laser reﬂow soldering, the size and morphology of Ag3Sn intennetallic
compound was ﬁner and the interfacial layer between solder and substrate was much
thinner in as-solidiﬁed condition of the joint as a result of more rapid cooling rate
(104°C/sec). Ag3Sn within the Sn-rich matrix, and Cu-Sn intennetallic compounds at

interfacial layer were observed to coarsen with aging after soldering, becoming more

Temperature °C

22

Atomic Percent Sn

 

 

 

 

 

 

 

 

0 10 20 30 40 50 50 70 30 90 100
1°00 961.93'C
8001
12.5 21.0
60 ~
0 724'
(Ag) 24.58 27480‘0 52.0
400- 231.9681‘C
‘p——8 \
22m
200- , . . 96.5
" 1255 i (BSn)—*
0 N \“T """ 13.0 """ «(aSn)
010 20 304050 60 7o 80 90100
Ag Weight Percent Sn Sn

Figure 7 Sn/Ag Phase Diagram [57]

23

signiﬁcant at higher temperature. Even though different microstructures are produced
according to different soldering processes, they tend to be altered in high temperature

environments [56].

The melting temperature (221°C) of 96.SSn/3.5Ag is higher than that of 63Sn/37Pb
(183°C) upon which current manufacturing technologies are based. 96.SSn/3.5Ag might
be suitable for more demanding high temperature service environments such as for under-
the-hood service in automobiles and avionics systems. The most desirable Pb-free solder
would possess almost the same melting characteristics as Sn/Pb solder. The incorporation
of a solute addition into the matrix was considered to be the best way to depress the
melting temperature of the 96.SSn/3.5Ag without the degrading the microstructure and
mechanical properties. The solute additions should be such that the solute resides solely

in solution with the matrix phase and not cause adverse microstructural segregation [4].

The depression of melting temperature of the 96.SSn/3.5Ag was reported with
improved microstructure and mechanical properties by addition of 1% Zn and 1%
Zn/0.5% Cu to the 96.5Sn/3.5Ag. Addition of 1% Zn suppressed the melting temperature
of the 96.SSn/3.5Ag to 217°C and reﬁned the microstructure, especially making rod shape
of Ag3Sn more spherical but the ductility decreased. 0.5% Cu addition helps to recover

ductility by decreasing of the effective grain size [5 8].

96.SSn/3.5Ag solder, like 63Sn/37Pb, forms the two intennetallic compounds
(Cu68n5, Cu3Sn) with Cu substrate. Interfacial reactions in solder and substrate systems

are of particular importance to the manufacturability and reliability of solder joints.

24

Excessive growth of interfacial intennetallic compounds during and aﬁer soldering can
be detrimental to reliability of solder joints due to their consumption of protective or
solderable ﬁnishes and the brittle nature of the intennetallic compounds. The growth of
intermetallic compounds is substantially controlled by two processes which are
diffusional process, the atomic transport of Sn and Cu through the intermetallic layer, and

the reaction process at the solder/intennetallic and the intennetallic/Cu interfaces.

With molten Sn/Ag solder, the growth rate followed the relationship x = ktm’ for the
total interfacial layer and x = kt"2 for Cu3Sn layer in a previous study [59]. However, a
differetn study showed a different result of parabolic relation (x =kt1’2) for total interfacial
layer [56, 60]. ‘x’ is the thickness of the alloy layer, ‘k’ is a temperature dependent

reaction rate constant, and ‘t’ is the time of exposure.

The interfacial layer growth may have a direct relationship with the fracture
mechanism. During aging at high temperature, and tensile and shear loading, cracks were
found within the interfacial layers or interface of the intermetallic layer and solder, and
appeared to propagate leading to the failure of the entire joint after the growth of
intennetallic compounds at the interfacial region [56, 59]. Ag was not observed in the

interfacial layer, implying that Ag was not involved in the interfacial reaction [59, 61].

The gold coating on solderable layer of electronic parts is usually used to protect the
solderable layer during the storage for a long time because gold is resistant to oxidation

and shows good wettability. However, the dissolution of substrate into the molten solder

25

may affect the mechanical properties of solder by forming intennetallic compound within

the solder matrix.

The 96.SSn/3.5Ag was more tolerant for dissolved gold than 63Sn/37Pb. The tensile
fracture elongation was found to be less sensitive for the 96.SSn/3.5Ag with dissolved
gold than 63Sn/37Pb. The tensile fracture of 96.SSn/3.5Ag still exhibited ductile fracture
with 5 mass% gold. These results were attributed to much smaller size of the AuSn, gold
intermetallics formed in the 96.SSn/3.5Ag solder joints compared to the 63Sn/37Pb joints
[62]. The Cu dissolution forming Cu,Sn5 intennetallic compound into the 96.SSn/3.5Ag
solder matrix of the joint was observed in the laser reﬂow soldering in which the
96.SSn/3.5Ag was superheated typically from 400°C to 600°C followed by the rapid
cooling (104°C/sec). Cu dissolution was much more dependent on the temperature
reached than the time at temperature. Such Cu dissolution is expected to increase strength

of the joints [56].

Adequate solderability is essential for any practical soldering application, even if the
melting temperatures are satisfactory. Solderability is controlled by the wettability of the
solder which is another signiﬁcant property of solders in determining a reliable solder
joint and is deﬁned as the ability for a liquid solder alloy to spread on a solid surface.
Wettability can be inﬂuenced by several factors such as solder composition, ﬂux,
temperature, surrounding atmosphere, etc. The kinetics of the wetting behavior are
dominated by the physical and chemical processes at the solder-substrate, solder-ﬂux, and

ﬂux-substrate interfaces.

26

The wettability of molten 96.SSn/3.5Ag appeared to be poorer than molten
63 Sn/37Pb but still acceptable for practical soldering applications. It was attributed to the
high values of 7”, which describes the surface tension between solder and ﬂux, caused by
higher surface tension of silver itself. The wettability of the 96.SSn/3.5Ag can be
improved by using ﬂux which decreases its surface tension [34, 63]. In reﬂow
experiments, the 96.SSn/3.5Ag exhibited better wetting characteristics than the
63Sn/37Pb, displaying less sensitivity to the change of reﬂow atmosphere between
nitrogen and air [64]. The minor addition of alloying elements which are Sb, Bi, Cu, and

In did not produce substantial improvement on wettability [65].

2.2.3 Composite Solder

With the advent of surface mount technology and the pursuit of miniaturization in the
electronic industry, the solder was demanded to be an electrical connection between
electronic devices and printed circuit substrates, as well as a mechanical support the
devices as a structural means. As a result, the reliability of solder joints has grown in

importance in electronic packaging technology.

The service temperature for solders are generally high relative to their absolute
melting point and solder microstructures tend to be highly unstable under typical
operating conditions. Solder materials also are readily subject to thermal stresses induced
due to the thermal expansion mismatch between components and substrates during

thermal cycling. Consequently, alternative solder materials possessing stable

27

microstructure and improved mechanical strength to ensure reliability of solder joint in

electronic applications need to be developed.

Approaches to strengthen the conventional solder materials by introducing ﬁne
second phase particles to solder to create a composite have been under investigation [66,
67, 47, 70 - 74]. Such particles may be introduced either by precipitation from
supersaturated solid or liquid solution or by external addition of foreign particles as
dispersoids. The precipitation strengthening may not be suitable method because the
precipitates are generally prone to coarsening unless one or more of the constituent
elements in the precipitate phase are chosen to have low solubility and diffusivity in the
solder matrix. On the other hand, foreign dispersoid particles introduced within solder
alloys would not coarsen easily since the elements or compound involved can be chosen

to have low solubilities and diffusivities or no reactivity with the matrix solder alloy [68].

Several approaches has been adapted using Sn/Pb solder to develop advanced
alternative solders, called composite solder. Among them, three principal approaches

have been pursued.

1. Blending conventional solder paste and powdered metals or intermetallics, and

subsequently reﬂowing the mixture [66, 69, 70, 71]

2. Mechanical alloying of conventional solder particles and high strength intennetallic

powders [67, 72]

28

3. Incorporation of ﬁne dispersoids by rapid solidiﬁcation processing of induction
melting and gas atomization using solder alloy with small addition of third element [47,

73, 74]

The ﬁrst approach has a limitation with the size of second phase reinforcements in the
solder matrix, which is determined by the size of the reinforcement powders that are
typically more than 10 pm. The approach of mechanical alloying produces signiﬁcant

oxygen contamination and may not be a suitable process [47, 73].

Several types of particles were used to produce composite solder with stable
microstructure and improved mechanical properties. The representative particles used are
the powder of intennetallic compound (Cu,Sn,, Cu3Sn), elemental metals such as Cu, Ni,
Ag, Au to introduce intennetallic compound with Sn in the solder matrix, aluminium

oxide [75], and carbon ﬁber.

The second phase dispersoids should be small enough (typically less than 500nm) and
closely spaced (generally < 1 pm) to yield signiﬁcant strengthening without considerable
loss in ductility. Incorporation of dispersoids by rapid solidiﬁcation processing produced
in-situ composite solders with large volume fraction of small, incoherent dispersoids in
solder matrix and improved strength, increased fatigue lives and creep resistance [47, 73,
74]. Fine dispersoids in the solder matrix was believed to stabilize dislocation
substructures and microstructures by pinning dislocations and grain boundaries, and act

as effective deformation barriers at elevated temperatures [74].

29

It is also important to uniformly distribute small dispersoid particles in sufﬁcient
quantity and maintain this dispersion after the alloy is melted and reﬂowed to affect
mechanical properties of the solid solder. A magnetic ﬁeld was applied to molten solders
containing ﬁne Fe particles to overcome the gravity-induced particles segregation
problem and to achieve a microstructure with a unifome distributed three-dimensional

network of the dispersoid particles [68].

The dispersoids should be well wetted by the solder or should possess a similar
speciﬁc density to solder, otherwise they will tend to sink or ﬂoat during the composite
preparation or during soldering processes. When carbon ﬁber was used as a dispersoid to
composite solder, carbon ﬁber was coated with copper to improve wetting with solder
matrix. Composite solder containing carbon ﬁber coated with copper exhibited better

mechanical properties than that containing uncoated carbon ﬁber [76].

The Cu,,Sn5 intennetallic compound has been adapted in the previous studies as a
second phase that provides several beneﬁts to the composite solder. The Cu,,Sn5
intermetallic compound is readily wet by solders with the use of ﬂux [77]. The solubility
of Cu,,SnS is very low in Sn/Pb solder at typical reﬂow temperatures as well as practically
zero in solid Sn/Pb. The density of Cu6Sn5 which is 8.23 g/cm3 at room temperature is
quite close to that of liquid eutectic Sn/Pb which is 8.2 g/cm3 at the eutectic temperature.

These similar densities are expected to promote a uniform distribution of particles [69].

The dispersoids could make composite solder more resistant to creep fatigue

deformations by acting as heterogeneous nucleation sites to produce a very ﬁne, initial

3O

microstructure of solder matrix upon reﬂow, and by serving as obstacles to grain growth,
crack growth, and microstructural coarsening to retain the stable microstructure of solder

even after thermal exposure [67 69, 77].

There is a need for a solder with a low coefﬁcient of thermal expansion, since the
stresses are thermally induced in solder joints due to the difference in thermal expansion
coefﬁcient between substrate and solder. Carbon ﬁber with nearly zero thermal expansion
coefficient as a dispersoid was attempted to adjust thermal expansion coefﬁcient of a
solder. Composite solder with the amount of carbon ﬁber which makes thermal expansion

coefﬁcient of solder to be similar to that of substrate exhibited prolonged fatigue lives

[76].

The dispersoids added to the Sn/Pb solder strongly affected the growth of the
intermetallic layers at the solder/Cu substrate interface. Cu containing particles such as
Cu68n5, Cu3Sn, and elemental Cu increased the thickness of the Cu—rich interfacial layer
(Cu3Sn) by reducing the amount of Sn available for interfacial reaction. The dispersoids
acted as a sink for Sn in the solder matrix. The dispersoids without copper, such as Ag,
Au, Ni enhanced the formation of Sn-rich interfacial layer (Cu68n5). One possibility was
that these elements diffuse to the intennetallic/substrate and act as a diffusion barrier
impeding the diffusion of Cu out of the substrate. This reduced the amount of Cu
available to form Cu3Sn layer with Sn and the Cu6Sn5 increased at the expense of the

Cu3Sn [71].

31

The strengthening effect of the second phase in composite solders could be offset
because of the formation of microvoids in solder [65, 70]. The strength of composite
solder appeared not to correlate with the volume fraction of dispersoids but to correlate
with the concentration of voids in the microstructure of the composite solder. These voids
acted as stress concentration centers and decreased the strength of the composite solder.
The formation of microvoids appeared to be very sensitive to processing method of
composite solder. The origin of voids is not known. The conceivable causes are ﬂux or
gas bubbles trapped by the higher viscosity of the composite solder melt during

solidiﬁcation.

A discussion about microvoids formation was given by Clough et. al [70]. The
upward velocity, ‘v’, of a ﬂux bubble in the melt can be described by the Stoke’s Law

given below

g(Ap)d2
18m

where g = 9.80 m/s2 is the acceleration due to gravity, Ap = p5 - pb is the difference
between the density of the solder (p5) and the bubble (pb), ‘d’ is the bubble diameter, and

n is the dynamic viscosity of the melt. It was found experimentally that the viscosity of
the solder melt increases with the additions of increasing amounts of particles to the melt.
As a result, the velocity of ﬂux bubbles decreases as the additions of the particles
increase. Thus the presence of large amounts of reinforcements tends to cause large

volume fractions of residual porosity to be trapped in the solder [68].

32

2.3 CREEP BEHAVIOR OF SOLDERS
2.3.1 The Description of Creep Deformation

Creep is a time-dependent deformation of a material when it is subjected to a constant
load at constant temperature. Creep is important at temperatures greater than 0.5Tm where
Tm is the melting point of a material in degrees absolute. Creep strain is measured as a
function of time and the slope of the creep strain-time curve is referred as a creep strain-
rate. Creep deformation is categorized into three stages : (1) primary or transient, (11)
secondary, constant rate, or quasi-viscous, and (III) tertiary. When a constant load is
applied to a specimen, the creep-rate decreases with time after an initial rapid elongation
of the specimen, then reaches a steady-state in which the creep-rate does not change with
time, and ﬁnally the creep-rate increases rapidly with time until fracture occurs. The

creep-rate in steady-state deformation is known as the minimum creep-rate.

The creep resistance of a material is oﬁen explained using the minimum creep-rate in
steady-state in creep deformation. The creep deformation in secondary or steady-state
stage is commonly described using a generalized, dimensionless rate equation called the

Mukherjee-Bird-Dom equation,

 

é...kT _ A(g)"(_b_)”
DEb ' E d
where 5'5. is the steady state strain rate, k is Boltzmann’s constant, T is the absolute

temperature, 0' is the applied stress, E is the elastic modulus of the material, b is the

magnitude of Burgers vector, d is the grain size of the material, A is a material constant, n

33

is stress exponent dependent on deformation mechanism, and p is a constant related to
grain size. The secondary stage of creep is a period of nearly constant creep-rate which
results from a balance between the competing processes of strain hardening and recovery.
the recovery mechanisms that are most important at high temperatures involve diffusion
of atoms through the crystals or along the grain boundary. In the above equation D is the
diffusion coefﬁcient for the diffusion mechanism that controls the recovery rate and is

expressed such as

D 2 Do exp(-k%)

where Do is the pre-exponential constant and Q is the activation energy for the rate-
limiting diffusion process. The temperature inﬂuence on the creep-rate is dominated by
the exponential temperature dependence of the diffusivity, so as an approximation, the

Mukherjee-Bird-Dom equation is often used

where A' is a material-speciﬁc constant.

34
2.3.2 Creep Mechanisms

Creep deformation in high homologous temperature can occur by various mechanisms
and each mechanism may have its own unique effect on the microstructure and may cause
its own distinctive form of damage. Thus, it is important to be able to predict what
mechanism(s) will be prevalent under a given set of stress, temperature, and
microstructure conditions. Deformation mechanisms map is a graphical description of
creep mechanisms in the stress versus temperature space that provide a convenient way to
predict what mechanism(s) will be predominant under a given set of deformation

conditions.

The deformation mechanisms for creep can be divided into three groups, diffusional
creep, dislocation creep, and grain boundary sliding. The dominant creep mechanism at
low stresses and high temperatures is diffusion of vacancies in response to the applied
stress. There are two types of diffusional creep. Nabarro-Herring creep represents the
deformation by diffusion of vacancies through the lattice and Coble creep describes the
diffusion of vacancies along grain boundaries which is a rapid path at lower temperatures.
The stress exponent, n, is one for both mechanisms and the activation energy is higher for
lattice diffusion than that for boundary diffusion. Coble creep is more favored for ﬁner

grain sizes compared to Nabarro-Herring creep.

Dislocation climb is a rate-controlling process of creep at moderate to high stress.
Climb is a process in which the dislocation can change the glide plane by emission or

absorption of vacancies. This diffusion controlled process limits the strain-rate by

35

allowing dislocations to bypass or penetrate obstacles such as dislocation tangles, cell
walls, precipitates, and dispersoids only as fast as climb occurs. There are a variety of
stress exponents ranging from 4 to 9 at higher stresses which imply many deformation

and recovery processes involving climb.

In solid solution alloys, dislocation glide in response to shear stress can also be
retarded. At moderate to high stress, the glide of dislocations can be the rate-limiting
mechanism because the dislocation velocity is limited by the solute atom atmospheres
that must move with the dislocation. Dislocation processes become more important with
increasing grain size, increasing stress, and/or decreasing temperature. For this

mechanism, the stress exponent, n, is about three.

At very high stress, the strain-rate increases dramatically with a small increase of
stress. This behavior is often referred to as power law breakdown, where the inﬂuence of

diffusion based processes become less important.

Grain boundary sliding (GBS) is also known as a fundamental mechanism for creep
deformation of superplastic materials. Superplastic deformation accompanies very large
strains before fracture and exhibits a stress exponent, n, between one and three and an
activation energy close to the value for grain boundary diffusion. GBS has a strong
dependence on the grain or phase size. Superplastic deformation occurs in a material with
stable, ﬁne, and equiaxed grains. The distinctive characteristic of superplastic materials is
high strain-rate sensitivity, m (m = l/n), which is about 0.3 to 0.9. The grain boundary

sliding in superplastic deformation requires some accommodation mechanisms such as

36

diffusion, boundary migration, or intragranular slip to maintain material continuity during

the GBS process.

All creep processes exhibit tertiary creep after the minimum creep rate is reached, and
this is related to crack initiation and propagation. Grain boundary sliding without
accommodation process required to maintain material continuity during the sliding
process is known to be responsible for this stage by nucleating intergranular cavity on

grain boundaries or interphase boundaries.

2.3.3 Deformation behavior of Solder Joint

Solder joints in electronic modules undergo temperature cycling which is caused by
power on/off cycles, daily temperature variations, and seasonal changes. Solder joints are
subject to a cyclic loading condition imposed by thermal stresses induced by the
difference between the thermal expansions of electronic components. Since the thermal
cycling ordinarily exhibits the slow strain-rate, solder joints basically undergo low
frequency thermal fatigue. The service temperatures of solder joints corresponding to
high homologous (T 2 0.5Tm) temperature and relatively long hold times are often
encountered during service conditions. Thus time-dependent, thermally-activated
deformation or creep becomes signiﬁcant at high homologous temperature greater than
0.5. Consequently, damages or failures of solder joints are caused by creep-fatigue
interaction which typically describes the cyclic application of a load at temperatures

where time dependent, thermally-activated processes are signiﬁcant.

37

The creep-fatigue Wage in solder joints has been identiﬁed in the analysis of
fracture surface of solder joints which were thermally cycled, where fatigue striations
were observed side by side with creep voids and extensive intergranular cracking [78].
The eutectic Sn/Pb solder subjected to isothermal cyclic loading at room temperature
showed that deformation occurred primarily by creep process and was very sensitive to
hold time or frequency resulting in increasing creep rate and decreasing cycles to failure
with increasing hold time [79, 80]. Creep was also observed to be the predominant mode
of deformation for isothermal fatigue in which good correlation was found between the

isothermal fatigue life of solder joints and the amount of creep strain per cycle [81].

2.3.4 Creep Deformation of Sn/Pb Solder

The creep behavior of mechanically worked and recrystallized eutectic Sn/Pb solder
which has various grain sizes ranging from a few micron to several microns is known to
exhibit sigmoidal relationship between strain-rate and stress. This relationship divides the
variation of the steady-state strain-rate with stress into three distinct regions: region I at
low stresses in which stress exponent, n, is equal to three, region 11 at intermediate
stresses where it is close to two, and region III at high stresses where it is equal to or

larger than six.

In region III with highest strain-rate, dislocation climb controlled creep was found to
be dominant deformation mode. Region III exhibited a stress exponent larger than six and

activation energy similar to that of lattice diffusion in pure lead or tin. There was no

38

dependence on the phase size in region III, so the exponent on grain size, p, was equal to

zero [82].

Region II at intermediate strain-rate (~10'5-10’2 see") was found to be dependent on
superplastic deformation according to the results of many researchers. Eutectic lead/tin
solder showed exceptional ductility up to 4000% when pulled in tension at a strain rate in
this region [83]. Region II was revealed to have a stress exponent close to two and
activation energy similar to that for grain boundary diffusion and to be sensitive to grain
size. Two different activation energies were observed depending on temperature which
are 44.7 i 1.1 kJ/mol below 0.89 Tm and 81.1 i 3.9 kcal/mol above 0.89 Tm. Since the
effective diffusivities were dominated by the grain boundary diffusivities even at a
homologous temperature of 0.97, this change in activation energy with temperature
seemed to suggest a change in creep mechanism, not a change in diffusion path from
grain boundaries to the lattice [84]. The other study indicated that dominant deformation
mode in region 11 was grain boundary sliding with accommodation process of diffusional

creep to maintain grain coherency [85].

The existence of region I in creep behavior of eutectic Sn/Pb is controversial. Region
I at very low strain-rates (_<.10'5 sec") showed the stress exponent is equal to three, the
grain size exponent is 2.3, and the activation energy of ~84 kJ/mol which is similar to
that for lattice self diffusion [86]. Region I was speculated to result from diffusional creep
but it is incompatible with experimental results because diffusional creep requires a stress
exponent of unity at low stresses. It was concluded that Region I may result from a

distinct deformation processes operating in sequence with region 11. However, region I

39

was not observed in the experiments by Murty [87] and Lam et. al. [88]. The stress
exponent was observed to remain constant down to the lowest strain-rate (~10‘7 sec"),
indicating that the superplastic deformation is still operative with n s 2 and with a

activation energy equal to that for grain boundary diffusion.

Some attempts have been made to construct a deformation mechanism map in the
space of normalized grain size versus normalized shear stress [89] and normalized strain-
rate versus normalized stress [90] for the eutectic Sn/Pb alloy. A signiﬁcant feature of the
experimental results for creep behavior of eutectic Sn/Pb is that a higher activation
energy occurs in region I than in region 11. It was anticipated that two additional
processes, Nabarro-Herring and Coble diffusional creep, may also contribute to the

deformation of these materials under some conditions [89].

Sn/Pb solder in an as-solidiﬁed joint condition was also shown to exhibit superplastic
deformation in creep testing. Mei et. al. identiﬁed two regions in creep testing of as-
solidiﬁed Sn/Pb solder joints [37]. At low stress and strain-rate a stress exponent value
close to two and activation energy for grain boundary diffusion was measured, implying
superplastic deformation whose mechanism is grain boundary sliding. At higher stresses
and strain-rates a stress exponent of seven and activation energy for self-diffusion

indicated that dislocation climbing was the operating mechanism.

The superplastic behavior of the as-solidiﬁed Sn/Pb solder joint was attributed to the
initial microstructure of the joint which is ﬁne and equiaxed [36, 91]. When eutectic

Sn/Pb is solidiﬁed in lamellar eutectic microstructure which results from the slow cooling

40

rate, the dominant mechanism of creep at normal strain-rate is a bulk mechanism which
has a stress exponent in the range of four to seven and activation energy close to that for
bulk diffusion. The quenched joint with initially more globular, ﬁne-grained
microstructure showed the transition from region 111 (dislocation climb process) to region
II (superplastic deformation process) which was accompanied by a change in activation
energy, while slow-cooled joints showed no transition to a superplastic deformation
process [92]. The steady-state creep-rate was microstructure sensitive at low stress and
strain rate in the grain boundary creep region, while at high stress and strain-rate the

creep-rate was microstructure insensitive in the matrix creep region.

A different creep behavior was revealed for lead-rich Sn/Pb solder alloy. Lead-rich
Sn/Pb bulk solder alloy with 3 to 9% Sn and grain size ranging from 0.01 to 2.5mm was
tested under creep conditions at high temperatures. Four distinct regions in normalized

stress versus strain-rate creep curve were observed [93, 94].

Region I in the lowest stress regime showed viscous creep with the strain exponent of
1.05i0.15. Diffusional creep deformations such as Nabarro-Herring and Coble creep
were considered to be responsible for this region. However, since an activation energy
close to that of lattice self-diffusion in pure lead and the grain size of the materials is too
large, Coble creep was not regarded as dominant deformation mode for region I. On the
other hand, the material constant, i.e. prefactor A, in region I which is 9.97x10‘lo was
4000 times greater than the predicted value for Nabarro-Herring creep and the grain size
was still too large to expect Nabarro-Herring creep to be operative. Thus the grain size

independent viscous creep which was known to be the dislocation based Harper-Dom

41

creep was considered to be responsible at these large grain size, but the underlying

mechanism is still speculative.

Regions II and III were observed to operate with n values of 4.6 and 3.2 respectively.
As stress increases to region 11, stress exponent increased to 4.6 and the climb of the edge
dislocations (climb I) appeared to be rate-controlling step for this region. Region III with
the stress exponent close to 3 appeared with further increasing stress and indicated creep
due to viscous glide of edge dislocations controlled by the locking of the gliding
dislocations by solute atoms with the activation energy to be identiﬁed with that for the

diffusion of the solute atoms in the lead matrix.

As stress increases to the other extreme, region IV, stress exponent increased to 4.74
and dislocation climb (climb 11) became the dominant deformation mechanism again
where the dislocations were able to get free from the solute atmosphere locking. The
experimental data at the highest applied stress exhibited a higher strain-rate than that

predicted from the power law with n = 4.7, which is indicative of power law break-down.

The characteristics of creep behavior were inﬂuenced by the intentionally added
second phase. The Ag3Sn particles which were introduced by adding Ag element in
eutectic Sn/Pb, which do not normally cavitate during superplastic deformation, caused
cavitation at many of the interfaces between Ag3Sn and matrix [95]. Addition of
elemental gold appeared to suppress the grain boundary sliding mechanism of creep in
ﬁne-grained eutectic Sn/Pb [96]. No transition from bulk creep to grain boundary sliding

creep was observed in Au-containing solder joints and the stress exponent between 4 and

42

6 remained down to the lowest strain-rates of 10‘7sec". The integrity of the materials
during creep deformation requires that the grain boundary sliding should be accompanied
by some bulk deformation so that the grains do not separate from one another. It was
suggested that Au may inhibit the bulk deformation processes that must accompany grain

boundary sliding.

2.3.5 Creep Deformation of Sn/Ag Solder

Eutectic Sn/Ag solder exhibited different creep behavior from Sn/Pb solder. Primary
creep was found to be signiﬁcant, contributing to large portion of the total creep
deformation [97]. For eutectic Sn/Ag solder tested at 298 K to 453 K (0.6Tm to 0.92Tm)
with stresses varying from about 1 MPa to 30 MPa, the primary creep followed a stress-

dependent relationship.

Steady-state creep for most of eutectic Sn/Ag bulk solder and joint was observed to
increase with temperature and stress. The steady-state region showed a stress exponent of
n z 10 and an activation energy of 50.2 kJ/mol for bulk and joint [97], z 6.7 to 8.4 and
62.7 kJ/mol for bulk solder [98], and z 5.5 and 38 kJ/mol for a ball grid array
conﬁguration [99]. Eutectic Sn/Ag solder was identiﬁed to be controlled by dislocation

climb.

The exponent of 10 is a much larger value than that generally noted in creep due to
dislocation climb where usually the n value varies from about 4 to 7. The data might be
analyzed in terms of a exponental stress dependence in tertiary creep but such dependence

is expected to occur at high stresses rather than low stresses used in tests [97]. Thus the

43

power-law creep was expected to be applicable. The activation energy of 12 kcal/mol was
about one half of that for self diffusion and was identiﬁable with either grain boundary or
dislocation-pipe diffusion. In a relation with activation energy, which is 46 kJ/mol for
pure tin at low temperature, a relatively large value for the stress exponent along with the
observed activation energy indicated the low-temperature climb of edge dislocations to be

the rate-controlling mechanism.

In stress-rupture tests, eutectic Sn/Ag solder exhibited temperature dependence on
stress to give the same life time [100, 101]. The stress-rupture properties of the eutectic
Sn/Ag soldered ring-and-plug joint decreased with increasing temperature in the range of
100—150°C resulting in a lower stress for a given life time at high temperature. The stress-

rupture properties were superior to those of the eutectic Sn/Pb solder.

44

2.4 FATIGUE BEHAVIOR OF SOLDERS

2.4.1 Fatigue Deformation of Sn/Pb Solder

Isothermal Fatigue (Mechanical Cycling)

In low cycle fatigue, the number of cycles to failure (Nf) is usually less than 10‘, as

described by Cofﬁn-Manson relation [102],
Nf” A}? = C

where Nfis the number of cycles to failure, A}? is the cyclic plastic strain range, ,6 is the
Cofﬁn-Manson exponent, and C is a constant. This relation is plotted on Ayp versus Nfas

a single straight line on log-log scales with a slope of -,6. Increasing the strain range

during cycling leads to a decrease in the number of cycles to failure.

6OSn/40Pb solder joint followed the Coffm—Manson relation when it was tested in
shear in the 1% to 10% high plastic strain range at 35 and 150°C at a frequency of 0.3 Hz
[103]. Eutectic Sn/Pb solder also followed the Cofﬁn-Manson relation above 0.6 % total
strain range (total plastic + total elastic) at 25°C [104]. However, low tin lead-based
solder (96.5Pb/3.SSn) tested in tension-tension at low plastic strain range below 0.6% and
at temperatures from 5 to 100°C ﬁt much better to two straight lines with a breakpoint at
about 0.3% plastic strain rather than a single straight line in the plot on Ayp versus Nf.
The two lines were shown to have high value of slope (,6) at low strain ranges below
0.3% and low value of slope (,6) at high strain ranges above 0.3%. The break in the lines

in the plot of the Cofﬁn-Manson relation was attributed to the change in the fracture

45

mode from intergranular at low strain ranges to mixed transgranular-intergranular at high

strain ranges [105].

The inﬂuence of frequency on isothermal fatigue life is signiﬁcant. The number of
cycles to failure for 6OSn/40Pb solder tested in shear at 35 and 150°C was reduced
substantially when the cycling frequency was decreased below approximately 3x104Hz at
35°C and 3x10'3Hz at 150°C [103]. At frequencies above the 3x104Hz at 35°C and 3x10’
3Hz at 150°C, the decrease in the number of cycles to failure with increasing frequency
was less signiﬁcant. For 96.5Pb/3.SSn solder [105], practically no change in the number
of cycles to failure was observed on varying frequency in the comparatively high
frequency region. A reduction in the number of cycles to failure with decreasing
frequency below 10'2Hz was found at 25, 50, 80°C. The frequency effect on the fatigue
life of 96.5Pb/3.5Sn solder appeared to be a function of the strain range : the higher the

strain range, the larger the effect of frequency on fatigue life.

Hold time at maximum strain exhibited a dramatic effect on the number of cycles to
failure of 96.5Pb/3.SSn and eutectic Sn/Pb solder [104, 105]. Hold time of a few minutes
at maximum strain reduced the number of cycles to failure by almost an order of
magnitude. Further increase in hold time eventually led to a constant value of the number
of cycles to failure. On the other hand, the fatigue life of 60Sn/40Pb solder in tests
employing equal hold time at positive and negative shear strains was observed to be
unaffected by such combination of hold times [106]. However, for 96.5Pb/3.SSn solder,

equal hold time at both tensile and compressive strains was found to be damaging by

46

reducing the number of cycles to failure as compared to tensile or compressive hold times

separately [105].

Thermal Fatigue (Thermarl Cycling)

In thermal cycling, solder joints undergo thermomechanical strains which eventually
lead to joint failure. This strain is produced by the combination of temperature
ﬂuctuations experienced by the package and the differing thermal expansion coefﬁcients
between materials used in the package. The source of temperature ﬂuctuations that result

in thermal fatigue can be either environmental or due to on-off power cycling.

The failure of near eutectic Sn/Pb (6OSn/40Pb) solder joint under thermal cycling was
preceded by the formation of a locally coarsened region along the direction of maximum
shear within the joint. The coarsened region was the weakest point of the joint and failure

occurred exclusively through this coarsened microstructure [41].

The localized coarsening was due to high local shear strain imposed during thermal
cycling on the solder joint by the thermal expansion differences between joined materials
at high temperature. The combination of shear strain and high temperature created regions
of concentrated plastic deformation within the solder joint. The accumulated deformation
along with the elevated temperature caused local microstructural changes. The localized
coarsening was a thermally-assisted stress-driven phenomenon and it appeared that

recrystallization and subsequent grain growth are the causes of the observed coarsened

band [42].

47

An increase in the amount of shear strain imposed on 6OSn/40Pb solder joint during
thermal cycling was signiﬁcant, resulting in more rapid localized coarsening and more
rapid failures. In contrast to isothermal fatigue behavior, decreasing the strain-rates
drastically increased the number of cycles to failure. This was attributed to the delay or
elimination of the formation of coarsened region in isothermal tests, which resulted from
dynamic recovery during thermal cycling. At the high temperature part of the thermal
cycle, and the slower strain-rate, the dislocation substructure recovered faster than it work
hardened which tended to minimize the subsequent recrystallization and growth of the

solder microstructure [107].

It was observed that extensive stress relaxation occurred during the elevated
temperature hold time, and very little stress remained after one minute at that
temperature. The stress at low temperature portion had little or no relaxation over a
period of hours. A drop in load with increasing number of thermal cycles was a
complicated function of microstructural coarsening and the formation of cracks that

decrease the area carrying the load [107].

48
2.4.2 Fatigue Deformation of Sn/Ag Solder

Isothermal Fatigge (Mechanical Cycling]

In the isothermal fatigue conducted at room temperature under constant total shear
strain control of Ay, = 0.04 to 0.79, the near eutectic Sn/Ag solder (96Sn/4Ag in wt %)
joint showed far superior fatigue resistance compared to eutectic Sn/Pb solder joint at the
high strain regime (A7, 2 0.2), but was far inferior to eutectic Sn/Pb solder joint at the
lower strain range (A7, < 0.2) [108]. It was suggested that the superior resistance of
eutectic Sn/Ag solder joint at the high strain regime must be due to a higher resistance to
crack initiation rather than to crack propagation, since this solder was observed to have

higher crack growth rate.

Previous studies [100, 101] by other investigators under cyclic tension/compression
with a mean stress of zero and a 0.2 mm/min strain rate using ring in plug joint have
shown that the isothermal fatigue resistance of eutectic Sn/Ag solder was better than

Sn/Pb solder at 20 and 100°C.

Thermal Fatigl_re (Thermal Cycling)

Near eutectic Sn/Ag solder (9SSn/5Ag in wt %) joint showed a longer fatigue life
than Sn/Pb solder in the thermal cycling from —30 to +100°C and 10 cycles/day.
9SSn/5Ag solder joint exhibited fatigue failure through the intermetallic layer between
the solder and Cu substrate that coarsened during thermal cycling. It was believed that the
fatigue failure for 95Sn/5Ag solder joint was postponed because the microstructural

coarsening that led to earlier failure in Sn/Pb solder did not occur in the 9SSn/5Ag solder

49

[109]. Another study on thermal fatigue of Sn/Ag solder also revealed that eutectic Sn/Ag
and 95 Sn/5Ag solder joints were more fatigue resistant than eutectic Sn/Pb solder joint in

the thermal cycling from —54 to +75°C [10].

2.5 STRATEGY FOR DEVELOPING AN IN-SI T U COMPOSITE SOLDER

The reliability concerns of solder joints in addition to the environmental concerns for
toxic materials place more demands for high performance of solder joints in electronic
applications. With these concerns for solder joints in electronic packaging, lead-free

composite solders incorporating particulate reinforcement have been considered.

This study will aim at the development of lead-free, in-situ composite solders to
improve reliability of solder joints in electrical systems while removing lead. The lead-
free, in-situ composite solder will be based on eutectic Sn/Ag (96.SSn/3.5Ag) solder and
developed by corporating ﬁne Cu6Sn, intennetallic particles as a composite phase to
make composite solder. A similar approach is applied to eutectic Sn/Pb (63Sn/37Pb) for

comparison.

Eutectic Sn/Ag (96.SSn/3.5Ag) solder has been considered as a potential candidate
for replacing leaded solders because it is free of toxic lead, better for high temperature
applications due to its high melting temperature, and silver is available in sufﬁcient
quantities to be considered as minor alloying element. Cu6Sn5 intennetallic compound is
known to wet readily by solders with the use of ﬂux. It is therefore expected that the
particulate will be well bonded to the matrix of the resulting composite solders. Phase

diagrams show no intermediate phases between Cu.,Sn5 and Sn, or between Cu and Ag.

50

Based on the literature review, the presence of the composite phase will affect
microstructural evolution which causes continuous change in mechanical properties.

Aging of composite solder joints will be performed to evaluate microstructural evolution.

The service temperatures for solder joints corresponding to high homologous
temperature (T > 0.5 Tm) and relatively long hold times are often encountered during
service conditions, indicating that time-dependent, thermally-activated deformation or
creep becomes important. Creep deformation has been observed to occur in isothermal
cyclic loading, and the amount of creep was very sensitive to hold time or frequency. The
effect of particulate reinforcements on the creep behavior of the in-situ composite solder

will be evaluated.

The mechanical properties of solder joints are known to be sensitive to the
microstructure of solder materials as reviewed in literature survey. Thus it is considered
that the microstructural evolution due to aging will affect the creep properties of solder
materials. The creep testing of unaged and aged composite solder joints will be done to

investigate the effect of microstructural evolution on creep properties.

CHAPTER 3

EXPERIMENTAL PROCEDURE

3.1 MATERIALS
3.1.1 Solder Material

The solder materials used in this study are eutectic Sn/Ag composite solder and
eutectic Sn/Pb composite solder. Composite solders were prepared by induction melting
of constituent materials. The base or matrix material of the composite solder was eutectic
Sn/Ag solder (96.SSn/3.5Ag) and eutectic Sn/Pb solder (63Sn/37Pb) which were
provided as ingots with dimensions of 2 x 36.5 x 1.5 cm3 by Ford Scientiﬁc Laboratory.
Cu,,Sn5 intennetallic n-phase was introduced into eutectic Sn/Ag and Sn/Pb solders as
second phase to make composite solders by adding the determined amount of pure Cu
and Sn elements to produce 20 vol % Cu,,Sn5 intennetallic n-phase while retaining the

eutectic composition of the matrix.

The fabrication of the composite solder was performed in an Induction or Radio
Frequency (RF) fumace shown schematically in Figure 8. The inside of the quartz tube
had an inert atmosphere of argon(Ar) and helium (He) gases to prevent the oxidation
during the melting of the component materials. Eutectic Sn/Ag solder or eutectic Sn/Pb
solder with additional Cu and Sn elements were placed in the graphite cup inside the
quartz tube and the temperature was raised by induction heating to a temperature above

the melting temperature of pure Cu, which has the highest melting point, 1085°C, among

51

52

the constituent materials. The temperature of the melt was measured using an optical

pyrometer.

After maintaining the temperature above the melting point of pure Cu for about 20
minutes to ensure complete melting of all constituent materials, induction furnace power
was turned off to solidify the melt. It was observed by visual observation to take about 10
min for the melt to solidify into a disk-shaped solid, and it took almost 2 hours to reach
room temperature. The composite solder ingot was cut using a diamond saw, polished,
and etched by a 3.3% nital (a mixture of HNO3 and methanol) to analyze the

microstructure with an optical microscope.

3.1.2 Substrate Material

The substrate material used to make solder joints using composite solder was pure
copper. Cu dogbone halves which had the dimensions shown in Figure 9 were machined
using EDM, to obtain a reduced width of 1 mm from a copper sheet, to simulate the

actual solder joint geometry in the electronic applications.

3.1.3 Flux

The ﬂux used in this study was a liquid ﬂux called 200-L Flux, provided by Ford
Scientiﬁc Laboratory. The 200-L ﬂux contained 25 ~ 35 wt % zinc chloride, 2 ~ 7 wt %
ammonium chloride, 1 ~ 5 wt % hydrochloric acid, and water as a balence. This ﬂux was

used to ensure reliable bonding of solder joints.

53

 

Mechanical Pump

Figure 8 Schematic Diagram of Induction or Radio Frequency (RF) Furnace

54

 

 

Figure 9 The Dimensions of The Cu Substrate

55
3.2 THE PREPARATION OF SOLDER JOINTS
3.2.1 The Pre-Treatment for Manufacturing a Solder Joint

The composite solders of about 20 grams produced by induction melting were
remelted using a small ceramic crucible with a size of 2.7 cm diameter and 3.0 cm height
in a furnace at a temperature above the melting point of Cu3Sn intennetallic e-phase,
676°C. They were poured vertically onto a 5 mm thick Cu plate to produce a circular splat
of about 5 cm diameter to obtain very ﬁne Cu,Sn5 intennetallic particles in the composite
solder due to a fast cooling rate. The splat cooled composite solder was rolled resulting in
about 100 pm thick sheet of composite solder which was cut into preforms of about 1.0

mmz.

Copper dogbone halves were painted with solder mask to prevent the spread of solder
beyond the 1 mm2 soldered area at the end of each half. The area soldered was pre-
cleaned with a 50 % aqueous HNO3 just prior to soldering to remove the thin tarnish

layers.

3.2.2 Soldering Process

Copper dogbone halves were placed in an aluminum ﬁxture designed to hold the Cu
dogbones in their proper alignment. Copper dogbone halves were ﬂuxed with 200-L
liquid acid ﬂux to remove a thin tarnish layer on Cu dogbone halves and to improve
wettability between solder preforms and copper dogbone halves. Eutectic Sn/Ag

composite solder and eutectic Sn/Pb composite solder preforms, 1.0 mm2 x 100 pm thick,

56

were sandwiched between the two Cu dogbone halves. The assembly is shown in

Figure 10.

The entire aluminum ﬁxture was placed on a preheated (about 500°C) hot plate and
allowed to reach a temperature of 270°C for the eutectic Sn/Ag composite solder, and
250°C for the eutectic Sn/Pb composite solder before being removed to the non-metallic
laboratory bench to cool. The Cu6Sn, intennetallic n-phase remained in the solid state
and only the eutectic Sn/Ag and eutectic Sn/Pb matrix were reﬂowed and solidiﬁed
during soldering process. Ten solder joints were manufactured in the aluminum ﬁxture at

a time following the heating/cooling history shown in Figure 11.

One side of the single shear lap joint was polished using 2000 grit sand paper and
0.05 pm SiO2 colloidal suspension. The surface of eutectic Sn/Ag composite solder joint
was etched with a solution of 50 ml H20, 50ml acetic acid, and 1 drop of H202 for
approximately 4 seconds, the specimen was washed with methanol prior to observing the
microstructure in Hitachi S2500 scanning electron microscope. It was not necessary to
etch the eutectic Sn/Pb composite solder joints because each phase in eutectic Sn/Pb
composite solder was clearly observed in the scanning electron microscope due to the

large difference of atomic numbers between Sn and Pb.

57

Composite Solder (1 mm!)

  
  
   
 
 
 
  

Solder Mask ——> .

200-L Flux

ThrmCeo-ouple 6mm- “ ‘ " ' 3.5mm

 

(b)

Figure 10 Schematic Diagram of
(a) The Solder Joint Fixture Assembly and
(b) A Fabricated Single Shear Lap Joint

58

 

 
    
 

 

 

 

 

 

 

  

 

 

 

300 ._ t I I I 1— v I I I I v I r r 1 I v Y t
250 7 ------------------------------- ~ ---------------- . -------------------------------- j
A t 2
CU " . .1
a zoo :- ---------------- 1 ---------------- i --------------- -:
g *- E1113 'c (Mr; or Eutectic sn/Pb Solder.) :
a - E E E
3 150 _ ---------------------------------------------------------------------------- "“
a - -1
a . .
0 .
E- 100 _ ------------------------------- : --------------------------------------------- —
50 ---------------- 1 -------------------------------- 1
1i ‘
0 P A L r l 1 1 r I j i 1 I 1 r r l 4 1 r
0 2 4 6 8 10
Time(minutes)
(a)
300 I I I IIIIII ! I I I I I I I I I I I I I rI I I I I 4
C i 1 1
250 L ----------------------- --------------------------- 1f -------------------------- J
G b 22I'C(M.P. of Eutectic :SnlAg Solder) ‘
x .............’ ---------------------------- 'y -------------------------- —d
I-
:1
H
a
I-
0 .
=- a
E 5
9 t
5" s s
1
0 ......... 1 r- ......... «
0 5 10 15
Time(minutes)

(b)

Figure 11 The Temperature versus Time Proﬁle for Manufacturing of
(a) Eutectic Sn/Pb Composite Solder Joints and
(b) Eutectic Sn/Ag Composite Solder Joints

59

3.3 AGING OF THE SOLDER JOINTS
Aging of solder joints was conducted by placing the solder joints on a piece of
aluminum which was on a hot plate, that had a known and steady temperature at

particular locations.

The eutectic Sn/Pb composite solder joints were statically aged for 100 hours at four
different temperatures (120, 100, 80, 50°C), and cyclically aged at three different
temperatures (120, 100, 60°C). The eutectic Sn/Ag composite solder joints were aged for
100 hours at four different temperatures (150, 120, 100, 60°C), and cyclically aged at
three different temperature (150, 120, 100°C). Aging by thermal cycling between the
three predetermined aging temperatures and room temperature was performed by turning
on and off hot plate 40 times. The time at the aging temperature and room temperature
was one hour respectively, resulting in two hours per cycle as shown in Figure 12. The
thermal cycling provided 20 hours at elevated temperature in a thermal cycling process.
There was an overnight hold time at room temperature for three times during the

experiment of thermal cycling.

For eutectic Sn/Pb composite solder joints, aging by thermal cycling was performed
again to inspect the repeatability of the result from ﬁrst aging by thermal cycling. Aging
by thermal cycling was accomplished using an appliance timer that turned the hot plate

on and off every hours.

60

T

T
one hour one hour one hour one hour

 

 

7"

'1‘

 

 

one cycle

Figure 12 The Schematic Diagram of the Temperature-Time History

in Thermal Cycling

61

After aging, the eutectic Sn/Ag composite solder joints were repolished slightly and
re-etched. Such a procedure was essential to eliminate original etching grooves and to
remove the inﬂuence of surface conditions such as oxidation, so that a direct comparison
of the microstructural changes due to aging could be made. It was not necessary to re-etch
the eutectic Sn/Pb composite solder joints because each phase was clearly observed in the
scanning electron microscope. Secondary electron images of the microstructure of the
solder joints were recorded using an Hitachi S2500 scanning electron microscope prior to

and aﬁer aging.

Micrographs of eutectic Sn/Ag composite solder joints were scanned with an AGFA
model scanner. The scanned images were then analyzed using LECO image analysis
software to determine Ag3Sn particle size and growth as a result of aging. The software
measured the area of each Ag3Sn particle assuming that the Ag3Sn particles to be circular.
An average particle area was calculated from which an average particle radius was
determined. By using the average particle size before and after aging, and by knowing
aging time and temperature, the aging kinetics were determined using a simple kinetic

model.

For eutectic Sn/Pb composite solder joints, a number of lines were drawn in different
directions on a micrograph of an eutectic Sn/Pb composite solder joint. The number of
intercepts between the line and interface boundary were counted. The length of the line
was divided by the number of intercepts to obtain a length/intercept. Several
length/intercepts were obtained for each micrograph from which an average

length/intercept was calculated. The length/intercept was used to determine coarsening

62

effects as a result of aging, with and without including the intercept length of the Cu6Sn5

phase.

3.4 CREEP TESTING OF COMPOSITE SOLDER JOINTS

Creep tests were performed with the eutectic Sn/Ag and eutectic Sn/Pb composite
solder joints. Both unaged and aged specimens were deformed with dead loading at room
temperature as illustrated in Figure 13. Scratches were made by grinding solder joint on a
rough cloth across the polished side of the solder joints and the amount of creep
deformation was measured using the location and shape of scratches present across the
joints. The specimens were removed periodically ﬁ'om the creep ﬁxture shown in Figure
13 to photograph the scratches with optical microscope and then reloaded in the creep

ﬁxture.

The geometry of the solder joints was such that the deformation was in simple shear
mode at constant load, which also corresponded to a constant shear stress. In creep tests
for eutectic Sn/Ag composite solder joints, a constant load was used for each unaged
specimen. For aged specimens a constant load of 2.25 lbs (10 N) for eutectic Sn/Pb
composite solder joints and 3.5 lbs (15.6 N) for eutectic Sn/Ag composite solder joints
was applied ﬁrst. After the strain-rate reached a steady-state value, the load was increased
to 2.5 lbs (11.1 N) for eutectic Sn/Pb composite solder joints and 4 lbs (17.8 N) for
eutectic Sn/Ag composite solder joints. Since the mechanical properties of solder and
solder joint are microstructurally sensitive as reviewed in the previous chapter, this

method was used to reduce the number of test joints and to facilitate the evaluation of the

63

 

Figure 13 The Creep Fixture

64

stress dependence of strain-rate of solder joints with same microstructure by retaining

nearly the same microstructure during creep testing.

All micrographs containing scratches were scanned with an AGFA model scanner and
the scanned images were converted to PICT ﬁles aﬁer adjusting the contrast to provide
optimal clarity of the scratch, using Adobe Photo software. Datathief software was used
to digitize the scratch data and these data were analyzed in spreadsheet programs to
determine the shear strain as a function of time as follows : X axis was put in the
direction of the thickness of solder joint and Y axis in the normal direction of the
thickness of solder joint. The software measured X and Y value for 15 positions on a
scratch. The slopes (dY/dX) between each position on a scratch were calculated from
which an average slope was determined. Since the scratch line before creep testing was
oblique with respect to the direction of solder joint thickness, an average slope of scratch
before creep testing was subtracted by an average slope of the deformed scratch after a

given time to obtain a shear strain value at a given time.

The examination of the fracture surfaces was performed using a Hitachi $2500
scanning electron microscope after creep rupture of the solder joint. These micrographs
indicated that there were a number of large pores in the solder joints, and these regions
could not carry the load. The area of pores larger than 50 pm in diameter were subtracted
from the soldered area to obtain the load carrying area used to determine the average
stress operating in the solder joint. The solder joints were found to have a porosity up to

20 % of the solder joint area.

CHAPTER 4

RESULTS

4.1 MICROSTRUCTURAL OBSERVATION OF BULK MATERIALS
4.1.1 Eutectic Sn/Pb (63Sn/37Pb) and Sn/Ag (96.SSn/3.5Ag) Solders

The microstructure of eutectic Sn/Pb solder shown in Figure 14 consisted of two
phases of lead-rich a-Pb phase and tin-rich B-Sn phase which are the same microstructure
identiﬁed in the literature review. The dark regions represents the lead-rich or-Pb phase
and the bright regions represents the tin-rich B-Sn phase. This microstructure appeared
intermediate one between globule and lamellae microstructure since the colony

boundaries which deﬁne regions having similar orientation are not clear in all areas.

The microstructure of cast eutectic Sn/Ag (96.SSn/3.5Ag) solder shown in Figure 15
were characterized by a colonized eutectic microstructure consisting of tin-rich primary
dendrites (appearing white in optical micrography) which are surrounded by regions of a
true eutectic phase of tin-rich B-phase with dispersed rod-shpaed Ag3Sn intennetallic a-

phase.

The existence of tin-rich dendrites in the microstructure of the eutectic Sn/Ag solder
was also observed in another study [5 8] where these dendrites appeared to be quite stable
and were present after months of room temperature aging. The presence of this tin rich

dendrites in cast eutectic Sn/Ag solder is indicative of the nonuniform cooling rate during

65

66

solidiﬁcation [58]. Another study [98] observed only the true eutectic phase for slow-

cooled eutectic Sn/Ag cast bulk where the tin-rich B-phase dendrites were not observed.

In the true eutectic phase Ag3Sn intennetallic s-phase appeared to have an acicular
shape and areas having similar orientation of Ag3Sn Intermetallic e-phase are visible. In
eutectic Sn/Ag alloy, the volume fraction of Ag3Sn intermetallic a-phase is very small

(i.e., less than 5 % in volume) [56].

To evaluate the effect of cooling rate on microstructure, eutectic Sn/Ag and Sn/Pb
solder droplets were splat-cooled on COpper and microstructurally examined as shown in
ﬁgure 16 and 17. The left sides of the micrographs shown in Figure 16 and 17 exhibits a
globule microstructure for eutectic Sn/Pb solder and a ﬁner grained microstructure for
eutectic Sn/Ag solder caused by fast cooling rate when the solder was in contact with the
copper plate resulting in rapid heat withdrawal and solidiﬁcation. The microstructure of
eutectic Sn/Ag solder on this side is characterized by the eutectic phase being evenly
distributed. In contrast, the right sides of the micrographs shown in Figure 16 and 17
showed a slower cooling rate near the air-cooled side than the left sides of the
micrographs. Eutectic Sn/Pb solder shows the lamellar eutectic microstructure on this
side. The microstructure of eutectic Sn/Ag solder in this region is characterized by a
colonized eutectic consisting of Sn-rich primary dendrites ranging in size from 5 to 25

pm which are surrounded by the true eutectic phase.

As shown in the Figure 16 and 17, it was evident that the microstructures of eutectic

Sn/Pb and Sn/Ag solders were dependent upon the cooling rate. The variation in cooling

67

rate in the thickness of the splat from severe to moderate from the Cu surface to the outer

surface of the solder caused the inhomogeneous phase distribution.

v- v .. 4'

 

Figure 15 Optical Microstructure of Eutectic Sn/Ag Solder Ingot

C

‘i‘? ‘r

- "it .‘ as
has??? Iggy w
,.. “3‘? ‘.« 1' “4""
$9”) .- ,,

{1‘17 '.
. _%“}‘ ﬁgvgk‘i” ‘

    
  

 

<— Faster Cooling Rate Slower Cooling Rate —->

Figure 16 Microstructural Variation of Eutectic Sn/Pb Solder
Due to Variation of Cooling Rate During Splat Cooling

 

— Faster Cooling Rate Slower Cooling Rate —>

Figure 17 Microstructural Variation of Eutectic Sn/Ag Solder
Due to Variation of Cooling Rate during Splat Cooling

70
4.1.2 Slow Cooled Eutectic Sn/Pb and Sn/Ag Composite Solders

Eutectic Sn/Pb composite solder produced by induction melting and slow cooling is
shown in Figure 18. Eutectic Sn/Pb composite solder contained Cu,,Sns intennetallic 11—
phase of long, rod-like shape in the eutectic Sn/Pb matrix. The eutectic Sn/Pb matrix
clearly showed lamellar microstructure resulting from the slow cooling rate during

solidiﬁcation where colonies having similar orientation are easily visible.

Eutectic Sn/Ag composite solder shown in Figure 19 exhibited Cu.,Sn5 intermetallic
n—phase surrounded by colonized eutectic Sn/Ag microstructure. Cu,,Sn5 intennetallic 11—
phase had a rod shape shorter than that in the eutectic Sn/Pb composite solder. Eutectic
Sn/Ag matrix in composite solder retained nearly the same microstructure as eutectic
Sn/Ag solder. Eutectic Sn/Ag matrix showed the tin-rich B-phase dendrites and the true
eutectic phase of tin-rich B-phase with dispersed Ag3Sn intermetallic e-phase. In the
eutectic Sn/Ag matrix the primary and secondary dendrites of tin-rich B-phase without
Ag3Sn intennetallic s-phase are easily visible and the true eutectic phase comprised of

Ag3Sn intennetallic e-phase of acicular shape and tin-rich B-phase.

The addition of Cu and Sn elements to introduce 20 vol% Cu,Sn, intermetallic 11-
phase exhibited no effects on the microstructure of eutectic Sn/Pb and Sn/Ag matrix,
since the microstructure of eutectic Sn/Pb and Sn/Ag matrix of composite solders was the
same as that without composite phase. A previous study showed that adding 1% Zn to
eutectic Sn/Ag suppressed the formation of the tin-rich B-phase dendrites and introduced

a ﬁner and more uniform eutectic microstructure throughout the alloy [4]. In this study

71

Cu and Sn elements added to eutectic Sn/Ag and Sn/Pb solders appeared to be consumed
to form 20 vol% Cu68n5 intennetallic n-phase without affecting the microstructure of

eutectic Sn/Pb and Sn/Ag microstructure.

72

 

 
     

 

' ’ I

v' j

\‘-4" ‘4» f ‘5 I
.§~'b . _ ”' w "\:,~
17¢;‘~$r=*\\~

        
 

,_ .1' 4.1:, '1'

Hi I}. “if

,Nfé t {g
' wﬁ'ﬁ
‘9; I ’

 

   

' 1

Figure 18 Microstructure of Slow Cooled Eutectic Sn/Pb Composite Solder
(a) Can, Intermetallic n—phase Morphology (b) Eutectic Sn/Pb Matrix

73

 

Figure 19 Microstructure of Slow Cooled Eutectic Sn/Ag Composite Solder
(a) Cu6Sn, Intermetallic n—Phase Morphology (b) Eutectic Sn/Ag Matrix

74

4.1.3 Remelted Composite Solders

The eutectic Sn/Pb and Sn/Ag composite solders were remelted in a fumace and splat
cooled on a Cu plate. Figure 20 shows the morphology of Cu68n5 intermetallic n-phase in
eutectic Sn/Pb and eutectic Sn/Ag composite solders as a result of fast cooling rate by
quenching on a Cu plate. The Cu6Sn5 intermetallic n-phase was changed by such a
treatment to a very ﬁne particle shape with the size less than 2 pm distributed
homogeneously throughout the eutectic Sn/Pb and eutectic Sn/Ag matrix. It should be
noted that the size of Cu-rich Cu-Sn dispersoids was about 0.1~1.0 um [47] and 0.2~0.5
um [74] in in-situ Sn/Pb composite solder produced by rapidly solidifying Sn/Pb/Cu
ternary alloys containing 2.5, 5, and 7.5 wt % Cu by induction melting/gas atomization

(IMGA).

The solder joints were fabricated with preforms of the remelted eutectic Sn/Pb and
Sn/Ag composite solder cut out of rolled sheet at the reﬂow temperature below the
melting temperature of Cu,.,Sn5 intennetallic n-phase, retaining this morphology of Cu6Sn5

intermetallic n-phase.

75

 

(b)

Figure 20 The Modiﬁed Morphology of Cu,Sn, Intermetallic n-Phase due to Splat
Cooling in (a) Eutectic Sn/Pb composite solder and
(b) Eutectic Sn/Ag Composite Solder

76

4.2 MICROSTRUCTURES OF COMPOSITE SOLDER JOINTS

In order to obtain solder joints of the scale of those found in actual electronic solder
joints with representative, air-cooled microstructures, small single shear lap joints were
manufactured using the copper dogbones halves and the heating history described in the

experimental procedure.

Figure 21 illustrates the typical, as-formed microstructure of eutectic Sn/Pb composite
solder joint. The microstructure is characterized by Cu,Sn5 intennetallic particles evenly
distributed in the eutectic Sn/Pb matrix without agglomeration of the particles. The Sn/Pb
matrix surrounding Cu6Sn5 intennetallic particles showed a typical globular eutectic
microstructure which consisted of lead-rich or-phase of bright areas and tin-rich B-phase

of dark areas.

The eutectic Sn/Ag composite solder joint shown in Figure 22 exhibits typical,
colonized microstructure of eutectic Sn/Ag matrix containing Cu,,Sn5 Intermetallic
particles. The eutectic Sn/Ag matrix in the joint revealed nearly same microstructure like
bulk eutectic Sn/Ag solder which consisted of tin-rich dendrites and regions of a true
eutectic phase comprised of tin-rich B-phase and Ag3Sn e-phase. Ag3Sn e-phase showed

equiaxed shape rather than acicular shape in bulk solder.

 

Figure 21 SEM Microstructure of Eutectic Sn/Pb Composite Solder Joint
Containing CuGSn, Intermetallic n-Phase Particles

 

Figure 22 SEM Microstructure of Eutectic Sn/Ag Composite Solder Joint
Containing Cu,,Sn5 Intermetallic n-Phase Particles

78

The Cu,,Sn5 intennetallic particles had undergone a change of size from less than 2
pm to 3-8 pm upon reﬂow of eutectic Sn/Pb and Sn/Ag composite solders but retained
nearly the same particle shape. Cu,,Sn5 intermetallic compound was also found as an
interfacial layer between the Cu substrate and the eutectic Sn/Pb and Sn/Ag composite
solders. Cu,,Sn5 interfacial layer formed between the eutectic Sn/Pb and Sn/Ag composite
solder and Cu substrate appeared to be a continuous interfacial layer. However, closer
observation reveals channels in the Cu,,Sn5 interfacial layer. These channels have been

proposed to provide an open diffusion path for Cu during solidiﬁcation [110].

The density of Cu,,Sn5 intennetallic compound, 8.23g/cm3, at room temperature is
reasonably close to that of liquid eutectic Sn/Pb solder, 8.2 g/cm3, at the eutectic
temperature [69]. These similar densities of Cu6Sn, intermetallic compound and liquid
Sn/Pb solder indicates that the buoyancy of Cu6Sn5 intermetallic particles during reﬂow
should be small, and this will promote a uniform distribution of particles without ﬂoating

of Cu6Sn5 intennetallic particles in eutectic Sn/Pb matrix during reﬂow.

The coarsening of the Cu6Sn5 intermetallic particles from less than 2 pm to 3 ~ 8 pm
upon reﬂow of eutectic Sn/Pb composite solder could be expected during reﬂow and
solidiﬁcation due to the fast rate of diffusion of Cu in Sn shown in Table 3. The previous
studies [47] revealed that Cu-Sn intennetallic particles experienced drastic changes in
morphology from particle shape to plate shape as well as in size from 0.2-1 pm to greater
than 1 um upon reﬂow of in-situ Sn/Pb bulk composite solder. In the composite solders

produced by reﬂowing mixtures of Cut,Sn5 particles and eutectic Sn/Pb solder paste, most

79

of Cu,,Sn5 intermetallic particles appeared to be in the form of sintered clusters of small

particles ranging from about twenty to sixty microns [69].

A previous study [111] showed that during reﬂow of dispersion-strengthened Sn/Pb
solder joint, the dispersion-strengthened solder revealed ﬁner microstructure than that of
the conventional solder upon solidiﬁcation suggesting that the dispersoids acted as
heterogeneous nucleation sites resulting in ﬁne microstructure. It is unclear in this study
if the Cu,,Sns intennetallic particles similarly provided heterogeneous sites for nucleation
of Sn/Pb matrix, since the microstructure of eutectic Sn/Pb matrix is also subject to

change depending on the cooling rate upon solidiﬁcation.

However, it appeared that the Cu,Sn, intennetallic particles affected the growth of
colony in Sn/Pb matrix which is volumes of lamellae of similar orientation. The size of
colony was checked by counting the number of intercept of colony boundary and a linear
line. The size of colony in Sn/Pb composite solder joint was smaller than Sn/Pb solder
joint without composite phase. There was areas with no speciﬁc orientation like globular
microstructure in Sn/Pb composite solder joint, indicating that the Cu,,Sn5 intermetallic

particles caused the matrix not to form the colony by limiting the growth of colony.

A previous study [56] on the eutectic Sn/Ag solder joint produced by laser soldering
showed that Cu6Sn, intermetallic n-phase was formed in the form of dendrites by
dissolution of Cu substrate during reﬂow and dispersed throughout the eutectic Sn/Ag
matrix. Laser soldering superheated eutectic Sn/Ag solder to high temperatures, typically

from 400°C to 600°C, and this superheating was considered to result in more rapid

80

dissolution of Cu from substrate and great amounts of Cu,.,Sn5 intermetallic in the bulk
solder. However, Cu6Sn5 intennetallic n-phase was not formed in infrared (IR) soldering
which heats the solder to only 240 to 260°C. It was concluded that Cu dissolution is much

more dependent on the temperature reached than the time at the temperature.

In this study the temperature reached during soldering is similar to that in infrared(IR)
soldering. It could be thought that all of the Cu,,Sns intermetallic n-phase resulted from
preformed Cu68n, intermetallic n-phase by intentionally adding Cu and Sn elements, not

Cu dissolution from substrate upon reﬂow and solidiﬁcation.

The previous study [56] also found that in laser soldering the Ag3Sn intermetallic a-
phase was much ﬁner and appeared only at magniﬁcations over 4000X. The small size of
this phase was attributed to the rapid cooling rate, typically 104°C/sec, and such rapid
cooling promoted nucleation but severely limited the growth. In this study, the cooling
rate was about 1°C/sec as shown in Figure 11, which appeared to be fast enough to result
in Ag3Sn intermetallic c-phase of ﬁne particle shape with size less than 0.5 pm. However,
this result is controversial because the rod-like Ag3Sn intermetallic e-phase was formed in

infrared (IR) soldering in which the cooling rate is 10°C/sec.

81

4.3 AGING OF COMPOSITE SOLDER JOINTS

Figure 23 shows the monitored heating/cooling history when the eutectic Sn/Pb
composite solder joints used in ﬁrst and second aging were manufactured. These two
heating/cooling histories show that the eutectic Sn/Pb solder joints used in ﬁrst and
second aging were fabricated in a similar condition resulting in similar microstructure

shown in Figure 23.

The microstructure of eutectic Sn/Pb composite solder joints after static aging are
shown in Figure 24. By comparison with Figure 23, the eutectic Sn/Pb composite solder
joints revealed dramatic coarsening in size and shape of tin-rich or-phase and lead-rich [3-
phase when aged at 50, 80, 100, and 120°C for 100 hours. All of the globular
microstructure observed in the as-solidiﬁed microstructure disappeared for all of the aged
joints. It was apparent that increasing the aging temperature caused more coarsening in

both phases.

Figure 25 shows the microstructure of eutectic Sn/Pb composite solder joints after 20
cyclic aging between room temperature and aging temperature (20 hours at aging
temperature). By comparison with Figure 24, the 100 hours static aged solder joints
showed more coarsening in both phases than after 20 hours during cyclic aging between
room temperature and aging temperature. The time at aging temperatrue for static aging

was 5 times longer, accounting for the greater amount of coarsening.

Temperatu re('C)

u.-
o
9

Temperature (‘C)

300

250

§

I-
U!
0

 

82

UI
O

 

o 2 4 6 8 lo
Time(minutes)

 

o 2 4 6 8 10
Time (minutes)

 

(b)

Figure 23 The Temperature versus Time Proﬁle for Manufacturing and
Corresponding Microstructure of Eutectic Sn/Pb Composite Solder Joints
Used in (a) First Aging and (b) Second Aging

83

 

Figure 24 The Microstructure of Eutectic Sn/Pb Composite Solders

aﬁer Static Aging at (a) 50, (b) 80, (c) 100, (d) 120°C for 100 hours

s4

 

(a) (b)

Figure 25 The Microstructure of Eutectic Sn/Pb Composite Solder Joints
after 20 Aging Cyclies (20 Hours at Aging Temperature)
(a) 20 Aging Cyclies between RT and 60°C (20 Hours at 60°C)
(b) 20 Aging Cyclies between RT and 120°C (20 Hours at 120°C)

85

 

(C) (d)

Figure 26 The Microstructural Change of Eutectic Sn/Ag Composite Solders
after Static Aging at (a) 60, (b) 100, (c) 120, ((1) 150°C for 100 hours

86

 

(a) (b)

Figure 27 The Microstructure of Eutectic Sn/Ag Composite Solder Joints
after 20 Aging Cyclies (20 Hours at Aging Temperature)
(a) 20 Aging Cyclies between RT and 100°C (20 Hours at 100°C)
(b) 20 Aging Cyclies between RT and 150°C (20 Hours at 150°C)

87

The microstructures of eutectic Sn/Ag composite solder joints obtained after static
aging are shown in Figure 26. The eutectic Sn/Ag composite solder joints exhibited
signiﬁcant coarsening in the Ag3Sn c-phase particle size when aged at 60, 100, 120, and
150°C for 100 hours. Similar results were observed in a previous study on the coarsening
of Ag3Sn e-phase particle [56]. Large acicular Ag3Sn c-phase were sometimes observed

as shown in Figure 26 (b) and (c).

It was evident that increasing the aging temperature had a dramatic effect on Ag3Sn a-
phase particle size. By comparison with Figure 22, the joint aged at 60°C retained most of
the colonied characteristics observed in the as-solidiﬁed microstructure. The tin-rich
dendrites were apparent in matrix, but the surrounding Ag3Sn e-phase particles coarsened.
The specimen aged at 100°C retained few colonied characteristics of as-solidiﬁed
microstructure. After aging at 120 and 150°C, all of the colonied characteristics observed
in the as-solidiﬁed microstructure have disappeared as a result of extensive Ag3Sn 8-

phase particle coarsening.

Figure 27 shows the microstructure of eutectic Sn/Ag composite solder joints after 20
cyclic aging between room temperature and aging temperature (20 hours at aging
temperature). By comparison with Figure 26, cyclic aged solder joints for 20 hours
showed similar coarsening of Ag3Sn particle to static aged solder joints for 100 hours.
Since the 20 hours aging time for cyclic aging was shorter than 100 hours for static aging,
a similar coarsening in static and cyclic aging indicated that the rate of coarsening is

faster in cyclic aging than in static aging.

88

In contrast, Cu,Sn, intermetallic particles for eutectic Sn/Pb and Sn/Ag composite
solder joints showed no obvious coarsening during the times and temperatures used in

this study. The Cu6Sn5 intermetallic particles were stable at elevated temperature up to

150°C for 100 hours.

4.4 CREEP DEFORMATION OF COMPOSITE SOLDER JOINTS
4.4.1 Scratch Technique for Shear Strain Measurement

Scratches made across the joint area were used to determine the strain distribution and
average shear strain across the eutectic Sn/Ag and Sn/Pb composite solder joint. Figure
28 shows scratches used to determine the average shear strain and strain distribution of an
eutectic Sn/Ag composite solder joint before loading and after creep deformation at
21 .7MPa for 374 ksec. The scratch line was nearly linear in the as-made condition before

loading and after some time the scratch deviated from its original conﬁguration.

The deviation of the scratch line from the original position was not homogeneous
across the joint, indicating that the strain distribution was inhomogeneous across the joint
as shown in Figure 28 (c). Since the deviation of scratch from its original position was
dominated by deformation in the eutectic Sn/Ag matrix, the Cu,SnS intermetallic particles
in which a less deviation of the scratch was observed caused this inhomogeneity in shear
strain across the joint. This effect is most apparent in the drop in strain on the right of the
upper scratch. In addition, the deviation of scratch from its original position was not
observed in Cu substrate and Cu,Sn, n-phase interfacial layer between eutectic Sn/Ag

composite solder and Cu substrate. The shear strain was computed by measuring the

89

displacement of the scratch, Ay, using 7 = Ay/Ax, where Ax refers to the distance from the
interface where the displacement was measured. The average shear strain was calculated
from the inhomogeneous shear strain distribution and was used to make plots of shear

strain versus time.

90

 

(a) (b)

0.1

0.08

P
c
as

Shear Strain (y)
P
i

0.02

 

24 48 72 96 120
Width Dimension of Solder Joint (um)

(C)

Figure 28 Two Scratches Across The Eutectic Sn/Ag Composite Solder Joint
(a) before loading (b) after creep deformation at 21.7 MPa for 374 ksec
(c) shear strain distribution across the solder joint

91

4.4.2 Eutectic Sn/Ag Composite Solder Joints

Figure 29 shows the creep curve of average shear strain versus time for unaged
eutectic Sn/Ag composite solder joints which were crept with 21.7 MPa and 25.8 MPa.
Immediately after loading, the unaged eutectic Sn/Ag composite solder joints exhibited a
brief period of primary creep in which the strain-rate decreased to its minimum, or steady
state value. The strain-rate of the steady state region was higher in high stress condition
than low stress condition. The rupture time differed signiﬁcantly for the two different
stress conditions, but the strain at which the rupture process started was near 0.2 for both

conditions.

92

 

' T I T 7 I l

; . . 1
Unaged i fracture fracture |
Specimens 7) 0'8 P195 '

I

l

1

‘ a-é 21.7 MPa
l . + 25.8 MPa

0 4
l

0.3

 

Average Shear Strain

\ 0.2e-7ls

l 1 A l

1000 1500 2000
Time (ksec)

 

 

 

 

 

Figure 29 Creep Curve of Average Shear Strain versus Time
for Unaged Eutectic Sn/Ag Composite Solder Joints

93

Figure 30 shows the creep curve of average shear strain versus time for eutectic
Sn/Ag composite solder joints which were aged at 150, 120, 100, 60°C for 100 hours. The
creep experiments were performed under dead loading condition of 3.5 lb (about 19 MPa
for solder joint aged at 120, 100, and 60°C, and 16.3 MPa for solder joints aged at 150°C)
and 4 lb (about 18.6, 21.1, 22 MPa for solder joints aged at 150°C, 100°C, 60°C) as shown
in Table l. The same loading condition produced slightly different stress conditions for
each solder joint due to slightly different load carrying areas as determined from the

fracture surface.

At the ﬁrst stress condition of about 19 MPa, the solder joints aged at 100 and 60°C
showed similar strain-rate (1.8e-7/s) at steady-state condition, while the solder joint aged
at 120°C exhibited faster strain-rate (7.2e-7/s). At the second stress condition of about 21
MPa for the joint aged at 100°C and about 22 MPa for the joint aged at 60°C, the solder
joint aged at 100°C revealed faster strain-rate (l6e-7/s) than that aged at 60°C (2.9e-7/s).
It was evident that the solder joints aged at higher temperature showed higher steady-state
strain-rate. The solder joint aged at 150°C showed much lower strain and strain-rate
(0.2e-7/s for the ﬁrst stress condition and 2.5e-7/s for second stress condition) compared
to the other solder joint aged at lower temperatures because it had a lower applied stress

than the other solder joints.

94

 

0.4 ...t...,...!4.
'agedat150°C i

"_ aged at 120°C , ]
Fiji’iiiaged at 100°C ..... ...... . ................ _

0.3 t
r - -v- Riy=l6e-7/s .
_ agedat60°C 9 §.211MPa _

...................

 
  
 
 

‘- 19.2 MPai .VH= 2. 9e-7/s
§ 22. 0 MPa-

Average Strain

    

'18.5MPa

 

Q... a‘ .
y- = 0. 2e—7/s

  

 

0 200 400 600 800
Time(ksec)

Figure 30 Creep Curve of Average Shear Strain versus Time
for Aged Eutectic Sn/Ag Composite Solder Joints

Table 1 Creep Deformation Parameters for Eutectic Sn/Ag Composite Solder Joints

 

 

 

 

 

 

 

 

 

Unaged Eutectic Aged Eutectic Sn/Ag Composite Solder Joints
Sn/Ag Composite
Solder Joints 150°C 120°C 100°C 60°C
Thickness of 120 57 70 100 80 67
Joints (urn)
Soldered 8.2 9.3 10.1 8.4 9.0 10.4
Area (105um2)
Pore Area 1.0 0.67 0.53 0.16 0.5 2.3
(105ml)
Porosity (%) 12.5 7.2 5.2 2.0 5.6 22.3
Applied 3.5 5 3.5 4 3.5 3.5 4 3.5 4
Loads (lbs)
Applied 21.7 25.8 16.3 18.6 19 18.5 21.1 19.2 22
Stress (MPa)

 

 

 

 

 

 

 

 

 

 

 

95
4.4.3 Eutectic Sanb Composite Solder Joints

The average shear strain versus time for eutectic Sn/Pb composite solder joints which
were aged at 120, 100, 80, 50°C for 100 hours is shown in Figure 31. The creep
experiment of the eutectic Sn/Pb composite solder joints were performed under dead
loading condition of 2.25 lb (about 8.9, 10.2, 10.5, 8.8 MPa for solder joint aged at 120,
100, 80, and 60°C) and 2.5 lb (about 11.7, 9.8 MPa for solder joints aged at 80°C, 50°C)
as shown in Table 2. The same loading condition produced slightly different stress
condition for each solder joint due to variations in the soldered area and pore area present

in solder joints.

As with eutectic Sn/Ag composite solder joints, a higher aging temperature in caused
a higher steady-state strain-rate in creep deformation, even though the applied stress for
each solder joint was different due to the porosity. The solder joints aged at 80°C and
60°C also showed a similar trend at the second stress condition. It was evident that aging
at the higher temperature produced higher steady-state strain-rates. The solder joint aged
at 50°C exhibited a large soldered area and no apparent porosity, resulting in a low

applied stress.

96

Average Strain

._ .
.9 "

. o-

9 MPa

 

0 500 1000 1500 2000 2500 3000 3500
Time (ksec)

Figure 31 Creep Curve of Average Shear Strain versus Time
for Aged Eutectic Sn/Pb Composite Solder Joints

Table 2 Creep Deformation Parameters for Eutectic Sn/Pb Composite Solder Joints

 

97
4.4.4 The Fracture Surfaces of Composite Solder Joints

Representative fracture surfaces of crept eutectic Sn/Ag and Sn/Pb composite solder
joints are shown in Figure 32. The fracture surface showed that the solder joints
contained big and small pores up to 20 % of the soldered area which did not carry the
load during the creep deformation. The porosity present in each solder joint was
subtracted from the soldered area to calculate the applied stress to solder joints. The
soldered area and porosity were different for each solder joint, as shown in Tables 1

and 2.

The fracture process consisted of smearing and tearing actions. The eutectic Sn/Ag
and Sn/Pb matrix were identiﬁed to be smeared through the whole fracture surface during
fracture process. Figure 33 (a) shows a smeared area on the fracture surface of an eutectic
Sn/Ag composite solder joints showing a high peak intensity of Sn-rich B-phase as
identiﬁed by EDAX. There is a small Ag peak to the left of the Sn peak. The small Ag3Sn
particles which could not be visually identiﬁed in the fracture surface due to the their
small size were considered to be distributed over the whole area of fracture surface.
Cu6Sn5 intennetallic n-phase particles were pulled out of their original position, and
distributed over the fracture surface by smearing action of two faces of fractured joint.

They were identiﬁed as Cu containing particles as shown in the EDAX proﬁle in Figure

33 (b).

98

...r—v'

200 um

200 um

 

(b)

Figure 32 The Fracture Surface Showing Pores in
(a) Eutectic Sn/Ag Composite Solder Joints Aged at 100°C
(b) Eutectic Sn/Pb Composite Solder Joints Aged at 100°C

99

      

0 - 20 no
005 Preset: 100: lenlnsl Us
63: 39k Dad

 

213:

 

FW: 0 - 20 not!
100; Presets 100: Rminingl Os
1 1275 Dead

(b)

Figure 33 SEM Micrograph showing
(a) The Smeared Matrix in Eutectic Sn/Ag
Composite Solder Joint and EDAX Result
(b) The Cu6Sn, Intermetallic Particles
in Fracture Surface and EDAX Result

100

The composite solder joints showed two kinds of fracture paths. Most fractures
showed the fracture through bulk composite solder and sometimes fracture near the
interface between composite solder and interfacial layer was identiﬁed in the side view of

fractured solder joints such as the joint in Figure 34.

101

Interfacial Fracture Through
Fracture Bulk Solder

200 um

 

Figure 34 SEM Micrograph Showing Side View of Fracture Surface of
Eutectic Sn/Ag Composite Solder Joint Aged at 60°C For 100 Hours

CHAPTER 5

ANALYSIS AND DISCUSSION

5.1 AGING OF COMPOSITE SOLDER JOINTS
The analysis of aging effects was performed using length/intercept measurement
between Pb-Sn interfaces for eutectic Sn/Pb composite solder joints, and measurement of

the particle size of the Ag3Sn s-phase for eutectic Sn/Ag composite solder joint.

The aging kinetics (activation energy for diffusion) was determined from the known

aging time and temperature, using the equation [112]
r3 — r03 = ktexp(—Q / RT)

where r0 is the length/intercept for eutectic Sn/Pb and the Ag3Sn particle radius for
eutectic Sn/Ag before aging, r is the length/intercept for eutectic Sn/Pb and the Ag3Sn
particle radius for eutectic Sn/Ag after aging, t is the aging time (100 hours for static
aging and 20 hours for cyclic aging), k is a proportionality constant, Q is the activation
energy for the aging process, T is the temperatrue in degrees Kelvin, and R is the gas

constant (8.314 J/mol °K).

The above equation was modiﬁed as follows

3 3
I (r -r0)/ = _ Q 1 k
°g{ 1‘ 2.3031iT+ 0g

 

102

103
and the activation energy for the aging process was determined from the slope,

(-Q/2.303R), ofthe log{(r3 — r03) / t} versus 1/T plot.

Figure 35 shows the Arrhenius plot of the change of size with inverse temperature for
Sn/Pb composite solder joints. The slope provides an activation energy of 60 kJ/mol for
cyclic aging, and 60 kJ/mol at high temperatures and 21 kJ/mol at low temperatures for
static aging. A similar activation energy at high temperatures was found for static and
cyclic aging conditions, indicating that the same mechanism was active for both aging
conditions. However, a change in mechanism is apparent at lower temperatures for static
aging, where the rate of coarsening appears to become faster than observations in cyclic

aging, with an activation energy of 21 kJ/mol.

An activation energy of 60 kJ/mol for static aging at high temperatures and cyclic
aging corresponds to the approximately half of activation energy for lattice diffusion of
Sn and Pb shown in Table 3. Previous studies [113] showed that diffusion in the lattice as
well as in the grain boundaries has been found to be appreciable at the temperatures
which the solders may experience during operation. It also implies that an effective
diffusion coefﬁcient involving lattice and grain boundary diffusion should be used for
explanation of diffusion process. Since the activation energy for grain boundary diffusion
has been observed to be approximately the half of activation energy for lattice diffusion,
the observed activation energy of 60 kJ/mol implies that grain boundary or interfacial

diffusion was dominant in the aging process.

104

 

 
  
   
 

 

 

 

 

 

0-5 T 1 I r 1
; : i Sn/Pb;
5 . é Com osite Solder
0f ........... g.-Cyclr.cAgrng..-..--..--..---_--.-..-..;.P ............ ............ _
A
In
a 2 = =
.. -0-5 r """"""" g ' wa'l‘ernp'eraturewr
E ‘ § StaticAging
:5. . .
v gQ=21kJImoI
g _1 _ ...................................... x ....... = .............. =I ...........
m .
g .
s... .
mi. 3 i i 5 i
:1 -1.5 — ------------- -------------- -------------- ;~ ------ \--—-a ------------
60 g s s s
2 . ~ ; I i
_2 -..rftStatisAgins ............. ............ _
'l— Cyclic Aging g g
; g 2 2 1.
-2.5 l l l l L
0.0025 0.0027 0.0029 0.0031
1rr (K'l)

Figure 35 Arrhenius Plot for Aging of Eutectic Sn/Pb Composite Solder
(The error bars were determined from the average value of ro and standard deviation of r
values evaluated from 7 measurements of the length/intercept for each aging condition)

Table 3 Relative Diffusivities of elements in Solders [114]

105

 

 

 

 

 

 

 

 

 

 

 

 

Diffusivity Activation (Dt)"2 pm for 100 hours
Do (cmz/sec) Q (kJ/mol) 25°C 75°C 150°C
Pb in Pb 0.995 107.4 0.002 0.052 1.399
Pb in Sn 4.0 99.6 0.012 0.201 8.58.5
Sn in Sn 10.7 105.1 0.012 0.254 6.364
i c axis
Sn in Sn 7.7 107.2 0.007 0.150 4.005
n c axis
Ag in Sn 0.18 77 0.454 4.24 44.8
J. c axis
Ag in Sn 0.0071 55.1 7.5 37.0 200.3
M c axis
Cu in Sn 0.0024 33.1 369 964 2658
J. c axis
Cu in Sn 2 x 10'0 at 25°C 848528
[I c axis

 

 

 

 

 

 

106

However, an activation energy of 21 kJ/mol at lower temperatures for static aging and
this transition to lower activation energy was not understood at this time. Thus to evaluate
the repeatability of this result, aging by thermal cycling was performed again using an
appliance timer that turned the hot plate on or off each hour without an overnight hold

time.

Figure 36 describes observed aging kinetics from the second aging experiment. A
similar activation energy of 25.5 kJ/mol for static and 20.3 lemol for the cyclic aging
condition was also observed. However, a transition to higher activation energy was not

observed at higher temperatures.

Several issues that could affect coarsening kinetics were investigated. Coarsening
effects as a result of aging was evaluated without including the intercept length resulting
from the area of the Cu,,Sn5 composite phase, and similar activation energies were
obtained. Since the Cu6Sn, composite phases in each solder joint occupied almost equal
intercept lengths correlated with the 20 % volume fraction, it was concluded that the
Cu6Sn5 composite phase had the same effect for all solder joints. Also, use of rJ-ro2 was

considered instead of r3-ro3, and this resulted in only a small increase in activation energy.

The differences in the initial microstructures used in ﬁrst and second aging
experiments due to differences in the heating/cooling history were compared. The time
span between the beginning of wetting and the beginning of solidiﬁcation for these two
heating/cooling history shown in Figure 23 was also similar resulting in similar

microstructure. The initial lengths/intercept, r0, shown in Table 4 for the eutectic Sn/Pb

107

solder joints used in ﬁrst and second aging were also similar. Consequently the aging
result was not affected by the initial microstructure of eutectic Sn/Pb solder joints used in

the ﬁrst and second aging experiments.

The effect of specimen geometry was considered such as thickness of the solder joint.
The thickness of solder joints was also checked because the thickness of solder joints
affects to the thermally induced shear strain between substrate and solder due to the
difference of thermal expansion coefficient which can inﬂuence the aging effect.
However, in this study the thickness of eutectic Sn/Pb solder joints used in ﬁrst and

second aging cannot be correlated to higher or lower values in the aging kinetic plots.

In the several regards discussed above, the repeatibility of the results from aging

experiments is doubtful and low activation energies are not understood at this moment.

108

 

       
   

 

 

 

 

 

0 l l l I
I.- ' 3 A '
.\ _g=20.3 kJ/mol g
A T“ -~; -\
E -1- -\
E ; :
3' St t.~ A . 9 5 1
p. _ ,. .,.,alc..,glng ..........
-1.5 : . _
,5; Q = 25.5 kJ/mol ; ,\
e . 1 .
s -2—
g“ t
-2.5
in
_3 1 1 1 1 1
0.0025 0.0026 0.0027 0.0028 0.0029 0.003 0.0031
1/1‘ (K!)

Figure 36 Arrhenius Plot for Aging of Eutectic Sn/Pb Composite Solder
(The error bars were determined from the average value of ro and standard deviation of r
values evaluated from 7 measurements of the length/intercept for each aging condition)

109

Table 4 Aging Parameters For Eutectic Sn/Pb Composite Solder Joints
Used in First and Second Aging

(r0 is the initial length/intercept, r is the length/intercept,
and t is the thickness of solder joints)

 

 

 

 

 

 

 

 

 

 

 

 

 

Static Aging (100 hours) Cyclic Aging (20hours)
l 20°C 1 00°C 80°C 50°C 120°C 100°C 60°C
ro(um) 1.26 1.17 1.12 1.43 1.19 1.25 1.15
r (mm) 3.87 2.70 2.33 2.09 2.44 1.91 1.26
First Standard 0.26 0.16 0.27 0.13 0.13 0.18 0.16
Aging Deviation
Experi of ro
-ment Standard 0.89 0.41 0.29 0.41 0.14 0.34 0.24
Deviation
of r
t(p.m) 116 70 103 70 116 121 80
r0 (pm) 1.07 1.23 1.2 1.13 1.07 0.99 1.13
r (pm) 2.44 2.28 1.99 1.54 1.87 1.71 1.47
Second Standard 0.24 0.28 0.25 0.25 0.09 0.28 0.21
Aging Deviation
Experi of ro
-ment Standard 0.44 0.36 0.40 0.26 0.26 0.36 0.33
Deviation
of r
t (um) 98 76 117 77 98 105 92

 

 

 

 

 

 

 

 

 

 

110

Figure 37 shows the Arrhenius plot of the size change with inverse temperature for
the Sn/Ag composite solder. A similar activation energy of about 37 kJ/mol was found
for static and cyclic aging conditions for all temperatures used in this study. The rate of

coarsening was faster in cyclic aging as compared to static condition.

The activation energy of about 37 kJ/mol corresponds to about a half of activation
energy 77 kJ/mol _L c axis (55 kJ/mol II c axis) for Ag in Sn as shown in Table 3. This
suggests that grain boundary or interfacial difﬁrsion was prevalent in aging process. Since
the diffusion distances for Ag in Sn shown in Table 3 are quite large, even at ambient
temperatures, they are consistent with extensive Ag3Sn coarsening. A previous study
[115] on coarsening effects of eutectic Sn/Ag solder joints without composite phase
revealed an activation energy of 48 kJ/mol, which corresponds favorably to the activation
energy 55 lemol || c axis for Ag in Sn. The rate of coarsening in the non-composite
Sn/Ag solder joints was much lower, by a factor of about 5 to 100. The activation energy
of 37 kJ/mol for Sn/Ag composite solder joints was lower than observed for non-

composite Sn/Ag solder joints.

111

 

-3.5 T I I I I I I I TTI I I I T T T T I I I I I I l I I I I ‘ I I I I I I I T 1'
t z I I I : r
t .

1_J L l

_ .\f § -§- Cyclic Aging . 3
_.--_.....f.-.\.-.\..3....I.--.i....Q.=..§3.7.5lenlole...3 .......... g ........ _

I
A

1
I . l
- o 1
b , t '
I I t
I I I t
- l l
. . . I
. l l
.— ................................................. . ............................ —
l I
- I I . l
I n . t
. t . 1
I I l
. l -
l t I
. I t -
- . i
t .

O

      

0

log {(r3-r 3)/t} (um, hrs)

-
l
t
r
v
u

............................................................................
t

—.— 5 Static Aging .
Q = 36.8kJ/mole

I

5"

U!
l

 

 

 

0.0023 0.0025 0.0027 0.0029 0.0031

l/T (°K")

Figure 37 Arrhenius Plot for Eutectic Sn/Ag Composite Solder

 

112
5.2 CREEP DEFORMATION OF COMPOSITE SOLDER JOINTS
5.2.1 Eutectic Sn/Ag and Sn/Pb Composite Solder Joints

The unaged eutectic Sn/Ag composite solder joints revealed that 20 % shear strain
could be accommodated in the composite solder joint before tertiary creep commences as
shown in Figure 29. This is similar to the amount of creep strain that non-composite
Sn/Pb solders can sustain, but lower than eutectic Sn/Ag solder can sustain in double
shear lap chip carrier conﬁguration [99]. The single shear lap joint used in this study
appeared to cause more severe damage during creep deformation than double shear lap

joint.

The eutectic Sn/Ag solder joint aged at 150°C showed much lower strain and strain-
rate compared to the other solder joint aged at lower temperatures. The fracture surface of
the solder joint aged at 150°C revealed that this solder joint had larger soldered area and
less pore area than other solder joints. Consequently the solder joint aged at 150°C had
larger load-carrying area which is estimated by subtracting the pore area from the
soldered area and lower actual applied stress (16.3 MP3 for 3.5 lb and 18.6 MPa for 4 lb)
than the other solder joints. The low applied stress due to larger soldered area and lower

porosity resulted in a low strain and strain rate.

The creep resistance of a material is often explained using the minimum creep-rate in
steady-state in creep deformation. The minimum creep-rates in steady-state of eutectic
Sn/Ag and Sn/Pb composite solder joints shown in Figure 29, 30, and 31 are compared

with eutectic Sn/Ag and Sn/Pb solder joints without the composite phase in Figure 38 and

113

39 in which the normalized strain-rate was plotted versus normalized shear stress. The
normalization parameters were the same as used by R. Darveaux and K. Baneiji [99]. The
solid line curves [99] are for eutectic Sn/Ag and Sn/Pb solder joints without composite

phase, which was obtained by incorporating the temperature dependent shear modulus.

It is apparent in Figure 38 that all of the aged eutectic Sn/Ag solder joints showed
higher strain-rates than the unaged eutectic Sn/Ag solder joints indicating that the solder
joints in the unaged condition are more creep resistant. The eutectic Sn/Ag composite
solder joints showed approximately two to three orders of magnitudes lower strain-rates
than non-composite eutectic Sn/Ag solder joints. The eutectic Sn/Pb composite solder
joints exhibited approximately one to two orders of magnitudes lower strain-rates than
non-composite eutectic Sn/Pb solder joints. The degree by which the strain-rates were
decreased depended on the amount of aging and indicated that eutectic Sn/Ag and Sn/Pb
composite solder joints containing Cu,,,Sn5 intermetallic particles had higher creep

resistance.

All of the aged eutectic Sn/Ag composite solder joints exhibited a higher strain-rate
than the unaged eutectic Sn/Ag composite solder joints. The aging of the solder joints
resulted in coarsening of small Ag3Sn e-phase particles present in the matrix of the solder.
The size of Ag3Sn e-phase particles increased and the number of Ag3Sn e-phase particles
decreased due to coarsening. A previous study [97] regarded eutectic Sn/Ag solder as a
dispersion-hardened alloy because in eutectic Sn/Ag solder, most of the Ag resides in
Ag3Sn intermetallics particles that are ﬁnely distributed in the B-Sn matrix. It was also

considered that the ﬁnely distributed Ag3Sn intermetallics particles act like dispersed

114

particles that can block dislocation glide and climb. In this study the effects of aging
appeared to be consistent with softening resulting from overaging in precipitate
strengthened alloys. The higher strain-rates in aged solder joints indicate that coarsened
Ag3Sn s-phase particles in the aged condition were less effective in blocking dislocation

motion than in the unaged condition.

Since all of the aged data appear to fall on a similar line nominally parallel to the
unaged data, it appears that a small amount of aging results in enough coarsening to make
the contribution of Ag3Sn particles to the strength, and the Cu,,Sn5 particles then dominate

the creep resistance.

The eutectic Sn/Ag and Sn/Pb composite solder joints showed much higher creep
resistance than the eutectic Sn/Ag and Sn/Pb solder joints without composite phase. This
result indicated that creep resistance without sacriﬁcing ductility is possible in composite
solders. The increase in creep resistance by the incorporation of hard particles was
observed in previous studies [70] in which eutectic Sn/Pb solder exhibited an increasing
trend in the creep resistance with increasing volume fraction of Cu.,Sn5 intermetallic

particles.

The exact mechanism by which the creep strength of eutectic Sn/Ag and Sn/Pb
composite solder joints was increased is uncertain at this time. However, it was apparent
that the Cu,Sn5 intennetallic particles impeded plastic ﬂow of the eutectic Sn/Ag and
Sn/Pb matrix. During the creep deformation the scratches present in Cu,,Sn5 intennetallic

particles showed no deviation from its original position and such deviation was observed

115

only in eutectic Sn/Ag and Sn/Pb matrix. The Cu68n5 intennetallic particles appeared to
be 3 effective barrier to plastic ﬂow of eutectic Sn/Ag and Sn/Pb matrix during creep
deformation resulting in lower strain-rates in eutectic Sn/Ag and Sn/Pb composite joints

than eutectic Sn/Ag and Sn/Pb solder joints without composite phase.

Y°(T/G)eXP(Q/kT)

Y(T/G)6XP(Q/kT)

116

 

—
G
u

u
an

H
9.

in
C
(a

 

1i. Darveaux & K'Bhnerji

1 ﬁ' 1 I I

—4
—1

3...- I. ' ‘ ‘ ‘s‘blaéi-‘jbiiits aged-if 150°C """"""""""""""""""""" 1;
g-----'--~-solder~joints-aged int-12:09 --------------------------------------- '2'
gang... solder iointsagedatjﬂilffc ..................................... ;
: solder joints aged at 60§°C
VH9 ' "iiiia'gEél some? joints """ """""""""""""""""""""""" E

 

 

Figure 38 The Normalized Steady-State Behavior

2 'R. ba'rv'eiii's. x. ﬁane'rn' ' ‘ f 7 j
10‘ 3 9"‘Eb’l’ci'é'r"j'éiiii'iiééii‘ii'i‘i’d‘é """"""""""""""""""""""" 3
:5- 0 "Sblde‘r'joint‘a‘ged'at100°C”“”“""“’ """""""""""""""" 1
103 Em‘m-solder-jointsafgedat-809G--~-----~--E --------------------------- ;
if....-solderjointsafgcd.at.5!l.°.C..-----.---..1 ............................ E
10' F ............................. .............................. ............................ 3;
:5---_Q-?-52.8-kJ/mol.g ............................. ............................. g

of Eutectic Sn/Ag Composite Solder joints

 

 

 

 

 

 

t/G

Figure 39 The Normalized Steady-State Behavior
of Eutectic Sn/Pb Composite Solder joints

 

117

5.2.2 The Fracture Surfaces of Composite Solder Joints

The formation of pores in eutectic Sn/Pb solder joints containing Cu68n5 intermetallic
particles and Au element was reported in the previous studies [70, 96] in which the
amount of porosity tended to increase with the volume fraction of Cu,Sn, and Au. It was
observed [70] that for Sn/Pb composite solder the inverse relationship appeared to exist
between vol % porosity and strength, i.e., lower volume fractions of porosity correspond
to higher amounts of strenghtening, and vice versa. Large volume fractions of porosity in
the matrix can inhibit or prevent composite strengthening by reducing effective joint
cross sectional area. Pores within the bulk of the joint can act as crack nucleation sites,
just as the external surface does. However, in this study composite solder joints did not

show any signiﬁcant relation between porosity and creep strength.

Previous studies [70, 96] discussed a possible origin for the pores present in the joints
is through ﬂux or gas bubbles trapped by the viscosity of the melt. It was also considered
that porosity will increase with increasing amounts of reinforcement because the viscosity
of the melt increases with increasing amounts of reinforcement, resulting in the decrease
of the velocity of ﬂux or gas bubbles motion and large volume fractions of residual
porosity. The origin of porosity present in composite solder joints and the effect of

viscosity on porosity remain as an open question to evaluate.

CHAPTER 7

CONCLUSIONS

1. The lead-free Sn/Pb and Sn/Ag composite solders containing very ﬁne Cu68n5
intennetallic particles of about less than 2 um in size were successfully developed by a
splat quenching method. The small single shear lap joints of the scale of those found in
actual electronic solder joints with lead-free composite solder were manufactured using

the copper dogbone halves.

2. The aging of composite solder joints caused microstructural coarsening of the
Sn/Pb and Sn/Ag matrix. However, The Cu,,Sn5 intennetallic particles were resistant to
microstructural coarsening and stable within the regime of time and temperature used in
this study. Aging kinetics for eutectic Sn/Ag composite solder joints were faster in cyclic
aging conditions, and the activation energy of the coarsening process is comparable with
grain boundary diffusion of Ag atoms. The measurement of aging kinetics for eutectic

Sn/Pb composite solder joints did not provide any conclusive result.

3. The composite solder joints exhibited a lower creep strain-rate than solder joints
without the composite phase, indicating that the composite solder is more creep resistant.
The Cu68n5 intermetallic particles appeared to cause a lower creep strain-rate in the

composite solder.

4. The fracture surface of composite solder joints showed that composite solder joints

contained porosity up to 20 % which can inhibit or prevent composite strengthening by

118

119

reducing effective joint cross section. The composite solder joints showed two kinds of
fracture paths which are through bulk composite solder and near the interface between

composite solder and interfacial layer.

BIBLIOGRAPHY

10.

11.

12.

BIBLIOGRAPHY

W.J. Plumbridge, “Review Solders in Electronics”, J. Mater. Sci., 1996, Vol. 31, pp.
2501-2514

P.T. Vianco and DR. Frear, “Issues in the Replacement of Lead-Bearing Solders”,
JOM., 1993, Vol. 45, No. 7, pp. 14-19

M.M. Schwarz, “Fundamentals of Soldering”, ASM Handbook, 1990, Vol. 6, pp.
126-137

M. McCormack and S. Jin, “The Design and Properties of New, Pb—Free Solder
Alloys”, Proc. IEEE/CPMT Int '1 Electronics Manufacturing Technology Symp.,
1994, pp. 7-14

D. Shangguan and A. Achari, “Evaluation of Lead-Free Eutectic Sn-Ag Solder For
Automotive Electronic Packaging Applications”, ibid., pp. 25-37

J. Seyyedi, “Thermal Fatigue Behaviour of Low Melting Point Solder Joints”,
Solder. Surf Mt. Tech, 1993, No. 13, pp. 26-32

J .H. Lau and D.W. Rice, “Solder Joints Fatigue in Surface Mount Technology: State
of the Art”, Solid State T ech., 1985, pp. 91-104

P.T. Vianco, “Soldering in Electronic Applications”, ASM. Handbook, 1990, Vol. 6,
pp. 985-1000

H.H. Manco, Solders & Soldering, 3’ded., McGraw-Hill, New York, 1992, pp. 11

J. Glazer, “Metallurgy of Low Temperature Pb-Free Solders for Electronic
Assembly”, International Materials Reviews, 1995, Vol. 40, No. 2, pp. 65-93

H.H. Manko, Solders and Soldering, 3'ded., McGraw-Hill, 1992, pp. 407-439

L.J. Matienzo and RR. Schaffer, “Wetting Behavior of Eutectic Tin/Lead Solder and
Fluxes on Copper Surface”, J. Mater. Sci., 1991, Vol. 26, pp. 787-791

120

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

121

AD. Romig, Jr., Y.A. Chang, J.J. Stephens, D.R. Frear, V. Marcotte, and C. Lea, in
Solder Mechanics - A State of the Art Assessment, ed. D.R. Frear, W.B. Jones, and
KR. Kinsman, The Minerals, Metals and Materials Society, Warrendale, PA, 1991,

pp. 33

SK. Kang, and AK. Sarkhel, “Lead (Pb)-Free Solders for Electronic Packaging”, J.
Electron. Mater., 1994, Vol. 23, No. 8, pp. 701-707

F .G. Foster, “Embrittlement of Solder by Gold from Plated Surface”, AST M STP
319,1962, pp. 13-19

J.L. Freer Goldstein and J.W. Morris, Jr., “The Effect of Substrate on The
Microstructure and Creep of Eutectic In-Sn”, Metal]. T rans., 1994, Vol. 25A, pp.
2715-2722

D.R. Frear and RT. Vianco, “Intermetallic Growth and Mechanical Behavior of Low
and High Melting Temperature Solder Alloys”, ibid., pp. 1509-1523

L. Zakraysek, “Intermetallic Growth in Tin-Rich Solders”, Weld. Res. Supp., 1972,
Vol. 51, pp. 5368-541S

Z. Mei and J.W. Morris, Jr., “Characterization of Eutectic Sn—Bi Solder Joints’, .1.
Electron. Mater., 1992, Vol. 21, No. 6, pp. 599-607

S.K. Kang and V. Ramachandran, “Growth Kinetics of Intermetallic Phases at the
Liquid Sn and Solid Ni Interface”, Scripta Metall., 1980, Vol. 14, pp. 421-424

J.W. Morris, Jr., J .L. Freer Goldstein, and Z. Mei, “Microstructure and Mechanical
Properties of Sn-In and Sn-Bi Solder”, JOM., 1993, Vol. 45, No. 7, pp. 25-27

W.G. Bader, “Dissolution of Au, Ag, Pd, Pt, Cu, and Ni, in a Molten Tin-Lead
Solder”, Weld. Res. Supp., 1969, Vol. 48, pp. 5518-5578

K.R. Kinsman, in Solder Mechanics - A State of the Art Assessment, ed. D.R. Frear,
W.B. Jones, and KR. Kinsman, The Minerals, Metals and Materials Society,
Warrendale, PA, 1991 , pp. xx

W.L. Winterbottom, “Converting to Lead-Free Solders : An Automotive Industry
Perspective”, JOM., 1993, Vol. 45, No. 7, pp. 20-24

C]. Thwaites and W.B. Hampshire, “Mechanical Strength of Selected Soldered
Joints and Bulk Solder Alloys”, Weld Res. Supp., 1976, pp. 3238-3293

P.T. Vianco, “General Soldering”, ASM Handbook, 1990, Vol. 6, pp. 964-983

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

122

RM. Hall, in Solder Joint Reliability - Theory and Applications, ed. J .H. Lau, Van
Nostrand Reinhold, New York, 1991, pp. 307-309

G.E. Dieter, Mechanical Metallurgy, 3'ded., McGraw-Hill, pp. 301

R. Satoh, K. Arakawa, M. Harada, and K. Matsui, “Thermal Fatigue Life of Pb-Sn
Alloy Interconnections”, IEEE Trans. Comp. Hybrids Manuf Tech, 1991, Vol.
CHMT-l4, No. 1, pp. 224-232

D.R. Frear, J .B. Posthill, and J.W. Morris, Jr., “The Microstructural Details of B-Sn
Precipitation in a 5Sn-95Pn Solder Alloy”, Metal]. T rans., Vol. 20A, 1989, pp. 1325-
1333

H.H. Manco, Solders & Soldering, 3'ded, McGraw-Hill, New York, 1992, pp. 63

MA. Carroll and ME. Warwick, “Surface tension of some Sn-Pb Alloys : Part 1
Effect of Bi, Sb, P, Ag, and Cu on 6OSn-40Pb Solder” Mater. Sci. Tech, 1987, Vol.
3, pp. 1040-1045

T. IIDA and Roderick.I.L. Guthrie, The Physical Properties of Liquid Metals,
Oxford University Press, New York, 1988, pp. 134

RT. Vianco, F.M. Hosking, and J.A. Rejent, “Wettability Anslysis of Tin-Based,
Lead Free Solders”, Proc. Conf NEPCON. West, Anaheim, CA, 1992, pp. 1730-
1738

Binary Alloy Phase Diagrams, 2“ded., ed. T.B. Massalski, ASM International, Metal
Park, Ohio, 1991, pp. 3014

Z. Mei, J.W. Morris, Jr., M.C. Shine, and T.S.E. Summers, “Effects of Cooling Rate
on Mechanical Properties of Near-Eutectic Tin-Lead Solder Joints”, J. Electron.
Mater., 1991, Vol. 20, No. 8, pp. 599-608

Z. Mei, D. Grivas, M.C. Shine, and J.W. Morris, Jr., “Superplastic Creep of Eutectic
Tin-Lead Solder Joints”, J. Electron. Mater., 1990, Vol. 19, No. 11, pp. 1273-1280

R.J. Klen Wassink, Soldering in Electronics, 2“"ed, Electrochemical Publications
Limited, Scotland, 1989, pp. 146-147

B.T. Lampe, “Room Temperature Aging Properties of Some Solder Alloys”, Weld.
Res. Supp., 1976, pp. 3303-3403

J.G. Duh, K.C. Liu, “Microstructural Evolution in Sn/Pb Solder and Pd/Ag Thick
Film Conductor Metallization”, Proc. Conf IEEE ECTC., 1991, pp. 653-657

41

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

123

D. Frear, D. Grivas , and J.W. Morris, Jr., “A Microstructural Study the Thermal
Fatigue Failures of 6OSn-40Pb Solder Joints”, J. Electron. Mater., 1988, Vol. 17, No.
2, pp. 171-180

D. Tribula, D. Grivas, D.R. Frear, and J.W. Morris. Jr., “Microstructural
Observations of Thermomechanically Deformed Solder Joints”, Weld Res. Supp.,
Vol. 68, No. 10, pp. 4043-4093

L.Quan, D. Frear, D. Grivas, and J.W. Morris, Jr., “Tensile Behavior of Pb-Sn
Solder/Cu Joints”, J. Electron. Mater., 1987, Vol. 16, No. 3, pp. 203-208

H.N. Keller, “Temperature Cycling of HIC Thin-Film Solder Connections”, IEEE
Trans. Comp. Hybrids Manuf Tech, 1981, Vol. CHMT-4, No. 1, pp. 132-139

W.J. Tomlinson and NJ. Bryan, “The Strength of Brass/Sn-Pb-Sb Solder Joints
Containing O to 10% Sb”, J. Mater. Sci., 1986, Vol. 21, pp. 103-109

A. Gangulee, G.C. Das, and MB. Bever, “An X-Ray Diffraction and Calorimetric
Investigation of the Compound Cu63n5”, Metal]. Trans., 1973, Vol. 4, pp. 2063-2066

S.M.L. Sastry, T.C. Peng, and RI. Lederich, K.L. Jerina, and CG. Kuo,
“Microstructures and Mechanical Properties of In-situ Composite Solders”, Proc.
Conf NEPCON. West, Anaheim, CA, 1992, pp. 1266-1275

Binary Alloy Phase Diagrams, 2“ded., ed. T.B. Massalski, ASM International, Metal
Park, Ohio, 1991, pp. 1481

L. Zakraysek, “Intermetallic Growth in Tin-Rich Solders’, Weld. Res. Supp., 1972,
pp. 5363-5413

Alex C.K. So and Y.C. Chan, “Aging Studies of Cu-Sn Intermetallic Compounds in
Annealed Surface Mount Solder Joints”, Proc. Conf IEEE ECTC., 1996, pp. 1164-
1171

J .O.G. Parent, D.D.L. Chung, and I.M. Bernstein, “Effects of Intermetallic Formation
at the Interface Between Copper and Lead-Tin Solder”, J. Mater. Sci., 1988, Vol. 23,
pp. 2564-2572

R. F. Pinizzotto, E.G. Jacobs, Y. Wu, J .A. Sees, L.A. Foster, and C. Pouraghabagher,
“Formation and Growth of Intermetallics in Composite Solder” J. Electron. Mater.,
1993, Vol. 22, No. 7, pp. 355-359

L.E. Felton, K. Rajan, R]. F icalora, and P. Singh, “n—Cu68n5 Precipitates in Cu/Pb-
Sn Solder Joints”, Scripta Metal], 1991, Vol. 25, pp. 2329-2333

54

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

124

D. F rear, D. Grivas, J .W. Morris, Jr., “The Effect of CuGSns Whisker Precipitates in
Bulk 6OSn-40Pb Solder”, J. Electron. Mater., 1987, Vol. 16, No. 3, pp. 181-186

J .H. Vincent and and G. Humpston, “Lead-Free Solders for Electronic Assembly”,
Geneal Electronic Company (GEC) J. Res., 1994, Vol. 11, No. 2, pp. 76-89

W. Yang and R.W. Messler, Jr., and LE. Felton, “Microstructure Evolution of
Eutectic Sn-Ag Solder Joints”, J. Electron. Mater., 1994, Vol. 23, No. 8, pp. 765-772

Binary Alloy Phase Diagrams, 2"ded., ed. T.B. Massalski, ASM International, Metal
Park, Ohio, 1991, pp. 96

M. McCormack and S. Jin, “Progress in the Design of New Lead-Free Solder
Alloys”, JOM., 1993, Vol. 45, No. 7, pp. 36-40

J. London and D.W. Ashall’ “Compound Growth and Fracture at Copper/Tin-Silver
Solder Interfaces”, Brazing & Soldering, 1986, No. 11, pp. 49-55

E.K. Ohriner, “Intermetallic Formation in Soldered Copper-Based Alloys at 150°C to
250°C”, Weld. Res. Supp., 1987, Vol. 66, No. 7, pp. 1913-2023

P.T. Vianco, K.L. Erickson, and PL. Hopkins, “Solid State Intermetallic Compound
Growth between Copper and High Temperature, Tin-Rich Solders-Part
1:Experimental Analysis”, J. Electron. Mater., 1994, Vol. 23, No. 8, pp. 721-727

M. Harada and R. Satoh, “Mechanical Characteristics of 96.SSn/3.5Ag Solder in
Microbonding”, IEEE Trans. Comp. Hybrids Manuf T ech., 1991, Vol. 13, No. 4, pp.
736-742

I. Artaki, A.M. Jackson, and RT. Vianco, “Evaluation of Lead-Free Solder Joints in
Electronic Assemblies”, J. Electron. Mater., 1994, Vol. 23, No. 8, pp. 757-764

C. Melton, “Reﬂow Soldering Evaluation of Lead Free Solder Alloys”, Proc. Conf
IEEE ECTC., 1993, pp. 1008-1011

M.E. Loomans, S. Vayman, G. Ghosh, and ME. Fine, “Investigation of Multi-
Component Lead-Free Solders”, J. Electron. Mater., 1994, Vol. 23, No. 8, pp. 741-
746

J.L. Marshall, J. Sees, and J. Calderon, “Microcharacterization of Composite
Solders”, Proc. Conf. NEPCON. West Anaheim, CA, 1992, pp. 1278-1283

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

125

H.S. Betrabet, S.M. McGee, and J.K. McKinlay, “Processing Dispersion-
Strengthened Sn-Pb Solders to Achieve Microstructural Reinforcement and
Stability”, Script Metal], 1991, Vol. 25, pp. 2323-2328

M. McCormack, S. Jin, and G.W. Kammlott, “Enhanced Solder Alloy Performance
by Magnetic Dispersions”, IEEE Trans. Comp. Hybrids Manuf Tech-Part A, 1994,
Vol. 17, No. 3, pp. 452-457

J.A. Wasynczuk and GK. Lucey, “Shear Creep of Cu,Sn5 Eutectic Composites”,
Proc. Conf NEPCON. West, Anaheim, CA, 1992, pp. 1245-1255

R.B. Clough, R. Patel, J.S. Hwang, and G. Lucey, “Preparation and Properties of
Reﬂowed Paste and Bulk Composite Solder”, ibid., pp. 1256-1265

R.F. Pinizzotto, Y. Wu, E.G. Jacobs, and LA. foster, “Microstructural development

in composite Solders Caused by Long Time, High Temperature Annealing”, ibid., pp
1284-1298

H.S. Betrabet and SM. McGee, “Towards Increased Fatigue Resistance in Sn-Pb
Solders by Dispersion Strengthening”, ibid., pp 1276-1277

C.G. Kuo, S.M.L. Sastry, and KL. Jerina, “Fatigue Deformation of In-Situ
Composite Solders”, 15’ Int ’1. Conf Microstructures and Mechanical Properties of
Aging Materials, ed. P.K. Liaw, R. Viswanathm, K.L. Murty, E.P. Simonen, and
DR. Frear, The Minerals Metals & Materials Society, TMS, Warrendale, PA, 1993,
pp. 417-423

C.G. Kuo, S.M.L. Sastry, and KL. Jerina, “Tensile and Creep Properties of In-Situ
Composite Solders”, ibid., pp. 409-415

S. Jin, “Solder Materials Issues in High Density Interconnection and Packaging”,
F ina] Program ASM-T MS Materials Week ’96, ASM International and The Minerals,
Metals & Materials Society, Cincinnati, Ohio, 1996, pp. 116

CT. Ho and D.D.L. Chung, “Carbon Fiber Reinforced Tin-Lead Alloy as a Low
Thermal Expansion Solder Preform”, J. Mater. Res., 1990, Vol. 5, No. 6, pp. 1266-
1270

J .L. Marshall, J. Calderon, J. Sees, G. Lucey, and J.S. Hwang, “Composite Solder”,
IEEE Trans. Comp. Hybrids Manuf Tech, 1991, Vol. 14, No. 4, pp. 698-702

A.I. Attarwala, J.K. Tien, G.Y. Masada, G. Dody, “Conformation of Creep and
Fatigue Damage in Pb/Sn Solder Joints”, J. Electron. Packag, 1992, Vol. 114, pp.
109-111

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

126

RC. Weinbel, J .K. Tien, R.A. Pollak, and SK. Kang, “Creep-Fatigue Interaction in
Eutectic Lead-Tin Solder Alloy”, J. Mater. Sci., 1987, Vol. 22, pp. 3901-3906

J .K. Tien, B.C. Hendrix, and AI. Attarwala, “Creep-Fatigue Interaction in Solders”,
IEEE Trans. Comp. Hybrids Manuf Tech, 1989, Vol. 12, No. 4, pp. 502-505

M.C. Shine, and LR. Fox, “Fatigue of Solder Joints in Surface Mount Devices”
ASTMSTP 942, 1988, pp. 588-610

D. Grivas, K.L. Murty, and J .W. Morris, Jr., “Deformation of Pb-Sn Eutectic Alloys
at Relatively High Strain Rates”, Acta Metal]., 1979, Vol. 27, No. 5, pp. 731-737

M.M.I. Ahmed and T.G. Langdon, “Exceptional Ductility in the Superplastic Pb-62
Pct Sn Eutectic”, Metal]. T rans., 1977, Vol. 8A, No. 11, pp. 1832-1833

B.P. Kashyap and GS. Murty, “Experimental Constitutive Relations for High
Temperature Deformation of a Pb-Sn Eutectic Alloy”, Mater. Sci. Eng., 1981, Vol.
50, No. 2, pp. 205-213

S.T. Lam, A. Arieli, and AK. Mukherjee, “Superplastic Behavior of Pb-Sn Eutectic
Alloy”, Mater. Sci. Eng., 1979, Vol. 40, No. 1, pp. 73-79

F.A. Mohamed and T.G. Langdon, “Creep Behavior in the Superplastic Pb-62% Sn
Eutectic”, Phil. Mag., 1975, Vol. 32, No. 4, pp. 697-709

G.S. Murty, “Stress Relaxation in Superplastic Materials”, J. Mater. Sci., 1973, Vol.
8, No.4, pp. 611-614

S.T. Lam, A. Arieli, and AK. Mukheijee, “Superplastic Behavior of Pb-Sn Eutectic
Alloy”, Mater. Sci. Eng., 1979, Vol. 40, No. 1, pp. 73-79

F.A. Mohamed and T.G. Langdon, “Deformation Mechanism Maps for Superplastic
Materials”, Scripta Metal]., 1976, Vol. 10, No. 8, pp. 759-762

E. Pink, A. Marquez, and A. Grinberg, “A Note on Rate Sensitivities of Flow
Stresses of Eutectic and Eutectoid Alloys”, Scripta Metal]., 1981, Vol. 15, pp. 191-
194

Z. Mei, J.W. Morris, Jr., M.C. Shine, “Superplastic Creep of Eutectic Tin-Lead
Solder Joints”, J. Electron. Packag., 1991, Vol. 113, pp. 109-114

D. Tribula, J.W. Morris, Jr., “Creep in Shear of Experimental Solder Joints”, J.
Electron. Packag., 1990, Vol. 112, pp. 87-93

127

93. K.L. Murty and I. Turlik, “Deformation Mechanisms in PbSn Alloys : Application to
Solder Reliability in Electronic Packaging”, Proc. ASME & JSME Joint Conf,
Advances in Electronic Packaging, San Jose, CA, 1992, pp. 309-317

94. R. Darveaux, K.L. Murty, and I. Turlik, “Predictive Thermal and Mechanica,
Modeling of a Developmental MCM”, JOM., 1992, Vol. 44, No. 7, pp. 36—41

95. CW. Humphries and N. Ridley, “Effect of Relatively Hard Particles on Cavitation of
Microduplex Pb-Sn Eutectic during Superplastic Flow” J. Mater. Sci., 1977, Vol. 12,
pp. 851-853

96. PA. Kramer, J. Glazer, and J.W. Morris, Jr., “The Effect of Low Gold
Concentrations on the Creep of Eutectic Tin-Lead Joints”, Metal]. T rans., 1994, Vol.
25A, pp. 1249-1257

97. H. Yang, P. Deane, P. Magill, and KL. Murty, “Creep Deformation of 96.SSn-3.5Ag
Solder Joints in a Flip Chip Package”, Proc. Conf IEEE ECTC., 1996, pp. 1136-
1142

98. C.H. Raeder, G.D. Schmeelk, D. Mitlin, T. Barbieri, and W. Yang, “Isothermal
Creep of Eutectic SnBi and SnAg Solder and Solder Joints”, Proc. IEEE/CPMT Int ’1
Electronics Manufacturing Technology Symp., 1994, pp. 1-6

99. R. Darveaux and K. Baneiji, “Constitutive Relations for Tin-Based-Solder Joints”,
Proc. Conf IEEE ECTC., 1992, pp. 538-551

100.C.J. Thwaites, “Some Metallurgical Studies Related to the Surface Mounting of
Electronic Components”, Circuit World, 1984, Vol. 11, No. 1, pp. 8-12

101.M.E. Warwick, “The Wetting and Mechanical Properties of Lead-Free Capillary
Plumbing Solders”, Brazing & Soldering, 1985, No. 8, pp. 20-26

102.8. Vaynman, M.E. Fine, and DA. Jeannotte, in Solder mechanics, Solder Mechanics
- A State of the Art Assessment, ed. D.R. Frear, W.B. Jones, and KR. Kinsman, The
Minerals, Metals and Materials Society, Warrendale, PA, 1991, pp. 159

103.H.D. Solomon, “Low-Frequency, High-Temperature Low Cycle Fatigue of 6OSn-
40Pb Solder”, ASTM STP 942, 1988, pp. 342-370

104.E.C. Cutiongco, S. Vaynman, M.E. Fine, and DA. Jeannotte, “Isothermal Fatigue of
63Sn-37Pb Solder”, J. Electron. Packag., 1990, Vol. 112, pp. 110-114

1058. Vaynman, M.E. Fine, and DA. Jeannotte, “Isothermal Fatigue of Low Tin Lead
Based Solder” Metall. Trans., 1988, Vol. 19A, pp. 1051-1059

128

106.H.D. Solomon, “The Inﬂuence of Hold Time and Fatigue Cycle Wave Shape on The
Low Cycle Fatigue of 60/40 Solder”, Proc. Conf IEEE ECC., 1988, pp. 7-12

107.D.R. F rear, “Microstructural Evolution During Thermomechanical Fatigue of 628n-
36Pb-2Ag and 6OSn-40Pb Solder Joints”, Proc. Conf IEEE ECTC., 1990, pp. 518-
524

108.Z. Guo, A.F. Sprecher, H. Conrad, and M. Kim, “Monotonic Properties and Low-
Cycle Fatigue of Several Soft Solder Alloy Systems”, Proc. Conf. Materials

Developments in Microelectronic Packaging .' Performance and Reliability, ed. P.
Singh, ASM International Material Park, Ohio, 1991, pp. 155-162

109.J.L. Marshall, “Characterization of Solder Fatigue in Electronic Packaging”, Brazing
& Soldering, 1988, No. 15, pp. 4-9

110.H.D. Blair, T.Y. Pan, R.P. Cooper, J.M. Nicholson, R.H. Poulson, S.C. Stine, K.
Ezis, and D. Mitlin, “Copper Dissolution in Molten Solders, Proc. Conf NEPCON.
West, Anaheim, CA, 1994, pp. 543-555

111.H.S. Betrabet, J.K. McKinlay, and SM. McGee, “Dynamic Observation of Sn-Pb
Solder Reﬂow in a Hot-Stage Environmental Scanning Electron Microscope”, J.
Mater. Sci., 1992, Vol. 27, pp. 4009-4015

112.D.A. Porter and KB. Easterling, Phase Transformations in Metals and Alloys, 2"d
ed., Chapman & Hall, London, 1981, pp. 315

113.D. Gupta and J .M. Oberschmidt, “Difﬁlsion Processes in Lead Based Solders Used
in Microelectronic Industry”, Design and Reliability of Solders and Solder
Interconections, ed. R.K. Mahidhara, D.R. Frear, S.M.L. Sastry, K.L. Murty, P.K.
Liaw, and W. Winterbottom, The Minerals, Metals &Materials Society, 1997, pp.
59-64

114.Smithell’s Metals Reference Book, 7th ed., ed. E.A. Brandes and GB. Brook,
Butterworth/Heinemann, Oxpord, 1992

115.A.W. Gibson, S.L. Choi, K.N. Subramanian, and TR. Bieler, “Issues Regarding
Microstructural Coarsening Due to Aging of Eutectic Tin-Silver Solder”, Design and
Reliability of Solders and Solder Interconections, ed. R.K. Mahidhara, D.R. Frear,
S.M.L. Sastry, K.L. Murty, P.K. Liaw, and W. Winterbottom, The Minerals, Metals
&Materials Society, 1997, pp. 97-103

nrcaronN srnr: UNIV. LIBRARIES
1|HIIWII"IIIIIIIIWIIIWIN11111111leWWI
31293017710900