Arrêt du plomb dans les équipements électroniques  2

Alliages de Substitution pour les Productions Electroniques: 2

Alliages et température de fusion : (Lingots, barres, baguettes, grenailles, Préformes fils) 2

1°) Alliages binaires: 2

2°) Alliages tertiaires: 2

Fils avec âme décapante  2

Crèmes à braser (pâtes & encres) 2

Glossaire des Acronymes   3

LEAD FREE SOLDERS IN ELECTRONICS   3

ABSTRACT  3

INTRODUCTION   3

• Health Concerns  4

• Greater Thermal Stress of Components  4

• Use of Temperature Sensitive Components and  4

Substrates  4

LEAD FREE ALTERNATIVE SOLDER ALLOYS   4

1. Sn/Ag  4

2. Sn/Ag/Cu  5

3. Sn/Cu  5

4. Sn/Ag/Cu/Sb  5

6. Sn/Ag/Bi/Cu  6

7. Sn/Bi 6

8. Sn/Sb  6

9. In/Sn  7

10. Sn/Zn  7

11. Au/Sn ( 7

DISSOLUTION KINETICS OF COPPER   7

THE EFFECT OF ISOTHERMAL AGING   8

Experimental Procedure  8

Results and Discussion  9

CREEP DEFORMATION   9

CONCLUDING REMARKS   12

ACKNOWLEDGEMENTS   12

REFERENCES   12

Integrity of Solder Joints from Lead-free Solder Paste   12

ABSTRACT  12

INTRODUCTION   13

• Health Concerns  13

• Greater Thermal Stress of Components  13

LEAD FREE ALTERNATIVE SOLDER ALLOYS   14

Sn/Ag  14

96.5Sn/3.5Ag 221°C  14

Sn/Ag/Cu  14

95.5Sn/4.0Ag/0.5Cu 217-219°C  14

Sn/Cu  14

99.3Sn/0.7Cu 227 °C  14

Sn/Ag/Cu/Sb  14

96.2Sn/2.5Ag/0.8Cu/0.5Sb (known as Castin) 217-220 °C  14

Sn/Ag/Bi 14

91.8Sn/3.4Ag/4.8Bi 200-216°C  14

Sn/Ag/Bi/Cu  14

90Sn/2.0Ag/7.5Bi/0.5Cu (138) 198-212°C  14

Sn/Bi 15

42Sn/58Bi 138°C  15

Sn/Sb  15

95Sn/5Sb 232-240°C  15

In/Sn  15

52In/48Sn 118°C  15

Sn/Zn  15

91Sn/9Zn 199°C  15

Au/Sn  15

80Sn/20Au 280°C  15

DISSOLUTION KINETICS OF COPPER   15

THE EFFECT OF ISOTHERMAL AGING   16

Experimental Procedure  16

Results and Discussion  16

CREEP DEFORMATION   16

Valuable Lessons Learned  18

CONCLUDING REMARKS   19

ACKNOWLEDGEMENTS   19

REFERENCES   19

Lead-free alloys   19

1. Introduction  19

2. Alternative materials and solder 21

3. Oxidation tendency of different 23

4. Dissolution of copper 23

5. Low-melting binary and ternary  24

6. Intermetallic compounds  24

7. Alternative surface metallizations  25

8. Alternative PCB materials  26

9. Concluding remarks  27

 

Alliages de Substitution pour les Productions Electroniques:

Alliages et température de fusion : (Lingots, barres, baguettes, grenailles, Préformes fils)

Ref.	composition	Liquidu	Solidus	densté
§   S 100	Sn100	233	E	7,80

1°) Alliages binaires:

§	Sn90	Ag10	300	227	7,32
§   SA 2.5	Sn97,5	Ag2,5	226	221	7,34
§   SA 3	Sn97	Ag3	232	E	7,28
§   SA 3.5	Sn96,5	Ag3,5	221	E	7,36
§   SA 5	Sn95	Ag5	240	232	7,25
§   SA 5	Sn95	Ag5	222	183	7,42
§   SB 40	Sn60	Bi40	170	138	8,12
§   SB 5	Sn95	Sb5	248	179	8,43
§   SC 0,7	Sn99,3	Cu0,7	227	E
§   SC 3	Sn97	Cu3	320	227
§   SI 42	Sn58	In42	145	118	7,30
§   SI 48	Sn52	In48	131	118	7,30
§   SS 2	Sn98	Sb2	240	221	7,39
§   SS 3	Sn97	Sb3	240	235	7,28
§   SZ 9	Sn91	Zn9	199	E	7,27

2°) Alliages tertiaires:

§   SAB	Sn	Ag	Bi
§   SAC 305	Sn96,5	Ag3,0	Cu0,5	217	218	7,39
	Sn95,75	Ag3.5	Cu0,75	217		7.38
§   SAC 385	Sn95,7	Ag 3,8	Cu0,5
§   SAC 387	Sn95.5	Ag3.8	Cu0.7	217	E
§   SAC 408	Sn95,2	Ag4	Cu0,8
§   SAS 25.10	Sn65	Ag25	Sb10	238	232	7,26

2°) Alliages quaternaires:

§   SABC	Sn	Ag	Bi	Cu
§   SAC 0307 Co	Sn98,07	Ag0,3	Cu0,7	Co 0,03	217	227	7,32

 

 

Ces changements peuvent imposer des modifications importantes au niveau des process, des équipements, et des composants.

En contrepartie, les alliages "Sans Plomb" présentent une meilleure résistance au fluage et des caractéristiques de tenue mécanique plus élevées que l'alliage Sn63Pb37.

Fils avec âme décapante

Désignation de l'alliage		Zone de fusion		Poids.
Alliages sans plomb RoHS		Solide	Liquide	spéc	Domaines d'utilisation
	Sn 97 Ag2 Cu Sb	216 °C	221 °C	7.3	Soudure fine sur C.I. - Soudage de composants pour constructions radio, télévision,  retouches, trous métallisés. Soudure pour professionnel, câblage conventionnel.  - Appareil de mesure.  - Circuits imprimés cuivre nu.  - Améliore la résistance mécanique. Soudage sur circuits dorés, céramiques argentées et toutes bases argentées.  - Soudure utilisée pour des ensembles travaillant à des températures élevées (lampes, moteurs,…) Soudure sur céramique argentée, circuits imprimés, basse température, points juxtaposés.  Soudure spéciale pour grande résistance mécanique.  - Soudures sans plomb pour contact alimentaire.
	Sn 96 Ag	221 °C	eutectique	7.3
	Sn 100	232 °C	Fusion	7.3
	Sn 97 Cu Sn 99 Cu	227 °C 227 °C	320 °C eutectique	7.3 7.3

Crèmes à braser (pâtes & encres)

 

Tableau des Crèmes

 

Glossaire des Acronymes

Le monde du "sans plomb" regorge d’acronymes, d’abréviations et de termes qui ne vous sont peut-être pas familiers. Grâce à cette page, nous ferons de notre mieux pour lever le voile sur vos doutes et vos interrogations compte-tenu de l’évolution de la législation RoHS.

 

LEAD FREE SOLDERS IN ELECTRONICS

Angela Grusd

Heraeus Inc.

West Conshohocken, PA

ABSTRACT

Lead-bearing solders are used extensively in the electronics industry. In recent years, efforts to develop alternatives to lead-based solders have increased significantly. As researchers began to focus on Pb-free solders they recognised their potential in high temperature applications such as automotive where Sn/Pb solders do not meet the demanding requirements. In particular, the Sn-Ag-X lead free solders offer superior creep resistance at room temperature and 100°C as compared to Sn-40Pb. Results of this work will be presented as well as factors to consider when developing

and implementing lead free alloys such as manufacturability, availability, and cost. One of the most promising replacement alloys is Sn/4Ag/0.5Cu. This alloy will be discussed in detail.

INTRODUCTION

Lead free solders are currently in production in some facilities and some of the “green” companies have proposed timelines for their implementation in the next year. The development of lead free lternatives was initially driven by impending U.S. legislation and Environmental Protection Agency regulations restricting lead usage in the electronics industry due to the toxicity of lead. More recently, the European Commission has proposed legislation aimed at abolishing lead in End of Life Vehicles by 2002. The European Commission has made a second proposal against lead in End of Life Electrical and Electronic Waste, however no firm date has been set.

Japan has similar proposals surrounding lead usage. In December 1997, the Japan Environmental Agency proposed legislation on the disposal of lead scrap. Leadcontaining scrap, such as electronics, must be disposed of in sealed landfill sites to prevent the leaching of lead. A second Japanese proposal was initiated by the Ministry of International Trade and Industry (MITI) Japan

Automobile Industries Association. This called for a 50% voluntary reduction of lead in vehicles (excluding batteries) by 2001 and to one-third by 2003. Several major Japanese Consumer Electronic Manufacturers have publicly announced accelerated plans to eliminate lead solder completely by 2001. Lead free alternatives are being considered for several reasons:

• Health Concerns

It is widely known that lead is related to certain health risks. If lead particles are inhaled or ingested, they accumulate in the human body causing damage to the blood and central nervous systems. Lead poisoning can be detected by blood analysis and the limits are defined by national governments.

The standard upper value of untainted human beings should not exceed 130 µg/l. The upper blood lead limit of people who are exposed to lead at work is, according to the German Biological Place of Work Tolerance values (TRGS 903, June 1994), 300 µg/l blood for women and 700 µg/l for men. In the USA, these critical values have been lowered from 250 to 100 µg/l.

As a result, some industries have already eliminated lead and have found suitable alternatives, for example plumbing solders, tinned cans, lead-free gasoline for vehicles, and leadfree cut crystal glass. The majority of lead consumption is automobile batteries and ammunition. The lead consumption of the electronics industry is relatively small and, according to different sources, lies between 0.5 - 7%.

When choosing alternative metals, consideration must also be given to their health risks as well. Recent studies in the USA and in Europe came to the following conclusions concerning

the toxicology of lead and some alternative metals:[1]

• Cd is extremely toxic and should not be used (high risk). Many companies such as Ford Motor Company have banned Cd-containing materials.

• Pb was also identified as highly toxic (high risk - it is considered harmful to the reproductive system).

• Sb is very toxic and should not be considered a major alloying element (medium risk - in Europe this material is considered potentially carcinogenic).

• Ag and Cu are used in lead-free alloys in small concentrations - in Europe these materials are seen as low risk.

• Sn and Zn are essential elements in the human diet, yet may be toxic if exposures are sufficiently high (low risk).

• Bi is a relatively benign metal with a history of medicinal uses (low risk).

• Greater Thermal Stress of Components

In the automotive industry, more and more circuits are being

placed in the engine compartment in order to reduce the

quantity of cables and, therefore, reduce cost. These under

the hood conditions often see temperatures in excess of

150°C. A leading automobile manufacturer has even

measured temperatures as high as 170°C on a hybrid circuit.

The high thermal stresses imposed on the solder joints at

these temperatures has led automotive manufacturers to

research Pb-free alternatives with high thermal fatigue

resistance, because they observed that Sn-37Pb has poor

thermal fatigue properties even at 125°C. Higher

temperatures dramatically reduce the strength of the

solder joint during thermal cycling, due to greater plastic

deformation of the solder as well as diffusion,

recrystallization and grain growth inside the solder.

For conventional alloys such as Sn62/Pb36/Ag2 (melting

point 179°C) and Sn63/Pb37 (melting point 183°C),

there is a major concern for the mechanical and

microstructural stability and, therefore, the reliability of

the solder joint at an operating temperature of 150°C,

because it is approaching the melting point of the alloy.

• Use of Temperature Sensitive Components and

Substrates

Some industries have driven down costs by replacing

higher cost plastic components with less expensive

plastic ones. These components, however, cannot

withstand the standard reflow temperatures of 210-

230°C. Therefore low melting temperature solder alloys

are used in this case. This is especially apparent with

consumer electronics, as the operation temperatures are

from 0°C to +60°C. This lower temperature range

corresponds to much less thermal stress on the solder

joints as compared to those temperatures found in underthe-

hood applications which typically reach 150°C or

greater. An example of a solder for these lower

temperature applications is Sn/Bi eutectic.

LEAD FREE ALTERNATIVE SOLDER ALLOYS

1. Sn/Ag (96.5Sn/3.5Ag:221°C)

This alloy exhibits adequate wetting behavior and

strength and is used in electronics as well as plumbing.

Several sources have also reported good thermal fatigue

properties as compared to Sn/Pb. Thermal fatigue

damage in solders is accelerated at elevated temperatures.

In the Sn/Pb system, the relatively high solid solubilities

of Pb in Sn and vice versa, especially at elevated

temperatures, lead to microstructural instability due to

coarsening mechanisms. These regions of

inhomogeneous microstructural coarsening are known to

be crack initiation sites. It is well-documented that these

types of microstructures in Sn/Pb alloys fail by the

formation of a coarsened band in which a fatigue crack

grows. By comparison, the Sn/Ag system, has limited

solid solubility of Ag in Sn, making it more resistant to

coarsening. As a result, Sn/3.5Ag eutectic forms a more

stable, uniform microstructure that is more reliable.

Although the Sn/3.5Ag alloy itself exhibits good

microstructural stability, when soldered to copper base

metal, the combination of a higher Sn content (96.5Sn

compared to 63Sn) and higher reflow temperature

environments accelerate the diffusion rates for copper base

metal in Sn. As its corresponding composition is reached, the

brittle Cu6Sn5 intermetallic compound is nucleated and begins

to grow. To slow the diffusion rate and thereby decrease the

growth kinetics, alternative surface finishes such as

immersion gold (Au over Ni over Cu) may be used. The 2 µm

Ni in the immersion gold coating serves as an effective

diffusion barrier, limiting the Cu from diffusing into the

solder and forming the brittle Cu6Sn5 intermetallic compound.

Other surface finishes such as immersion silver (Ag over Cu)

and immersion palladium (Pd over Cu) do not contain a Ni

barrier layer. Their effect on the growth kinetics of the

intermetallic compound layers is under investigation.

2. Sn/Ag/Cu

1) 95.5Sn/4.0Ag/0.5Cu 217-219°C

2) 95.5Sn/3.8Ag/0.7Cu 217-219°C

3) 95.0Sn/4.0Ag/1.0Cu 217-219°C

4) 93.6Sn/4.7Ag/1.7Cu 216-218°C

Because the mechanical stability of the joint is degraded when

the melting point is approached, elevated temperature cycling

produces more damage for Sn/Pb solder (m.p. 183°C) as

compared to higher melting point solders. The melting

temperatures of Sn/Ag/Cu solders make them ideal in high

operating temperatures up to 175°C. As for wetting,

Sn/Ag/Cu solders do not wet Cu as well as Sn/Pb using

commercial fluxes. However, good fillet formation can be

easily achieved provided the fluxes are suitable for higher

temperature use. Soldering in nitrogen atmosphere also

improves wettability using no-clean fluxes. The copper

dissolution test provides a relative measurement of the

solder’s tendency to dissolve Cu from the base metal and

form the Cu6Sn5 intermetallic compound. For alloys 1-3, the

rate of copper dissolution is slower than the Sn/Ag alloy yet

faster than the Sn/Pb eutectic. For alloy 4, the high level of

Cu in the alloy prevented the dissolution of the copper wire

(See Dissolution section).

3. Sn/Cu (99.3Sn/0.7Cu:227 °C)

This alloy might be also suitable for high temperature

applications required by the automotive industry. It is a

candidate especially for companies looking for lead and silver

free alloys. Preliminary testing conducted on this alloy has

shown a significant improvement in creep/fatigue data over

standard Sn-Pb alloys.

4. Sn/Ag/Cu/Sb (96.2Sn/2.5Ag/0.8Cu/0.5Sb

(known as Castin):217-220 °C)

This alloy has similar mechanical properties to the Sn/Ag/Cu

alloy.

5. Sn/Ag/Bi (91.8Sn/3.4Ag/4.8Bi:200-216°C)

In general, bismuth is added to Sn/Ag/X solder alloys in

order to depress the melting point. Another benefit of Bi

addition is greater joint strength as indicated by ring and plug

testing. This particular alloy was developed by Sandia

National Labs. Sandia’s internal studies have found no

electrical failures on surface mount devices following

10,000 thermal cycles using 68 I/O PLCCs, 24 I/O

SOICs, and 1206 chip capacitors on standard FR-4

PCBs. The boards were cycled 0 to 100°C at a ramp rate

of 10°C/ minute. No cracks or deformation were observed

on boards cross-sectioned after 5000 thermal cycles.

Cross-sectional data on 10,000 cycles is being collected.

These results are in good agreement with data collected

by the National Center for Manufacturing Sciences

(NCMS) Lead Free Solder Project, which reported very

good thermal fatigue resistance on OSP printed circuit

boards (Organic Solderability Preservative that protects

copper pads and through-holes). The NCMS High

Temperature Fatigue Resistance Project is currently

evaluating this solder at temperatures up to 160 and

175°C. In combination with Pb from the PCB or

component metallisation, a Sn/Bi/Pb ternary compound

is formed with a melting point of only 96°C. As the trend

toward eliminating lead continues, this alloy may become

more attractive.

6. Sn/Ag/Bi/Cu (90Sn/2.0Ag/7.5Bi/0.5Cu

(138):198-212°C)

Although the addition of Bi to the Sn/Ag/X system

imparts greater strength and improved wetting, too much

bismuth (greater than 5%) leads to the presence of a

small DSC peak near 138°C, corresponding to the binary

Sn/Bi eutectic at 138°C or the ternary Sn/Ag/Bi eutectic

at 136.5°C. For this alloy with 7.5 weight percent Bi, this

corresponds to approximately 1% of the total melting.

This small amount of eutectic melting has an uncertain

effect on joint reliability as the temperature approaches

138°C. This combined with the aforementioned concern

of forming a SnPbBi compound at 96°C, makes this alloy

an unlikely candidate for a Pb-free solder.

7. Sn/Bi (42Sn/58Bi:138°C)

The low melting point of this alloy makes it suitable for

soldering temperature-sensitive components and

substrates. If these contain Pb, the SnPbBi ternary

eutectic compound may form at 96°C, which in turn

adversely affects the thermal fatigue properties. The

NCMS Lead Free Solder Project recently reported the

results of thermal cycle testing at 0/ 100°C and -55/

+125°C for over 5000 cycles on OSP boards. The result

was that the Sn/Bi outperformed the Sn/Pb at both

temperature excursions. It was thought that the closeness

of 125°C to the binary Sn/Bi eutectic at 138°C would

cause this alloy to be a poor performer. Two possible

explanations for this unexpected result were presented.

The Sn/Bi alloy may be annealing at 125°C, relaxing the

stresses produced during thermal cycling. A second

explanation was the alloy may be undergoing

recrystallization.

8. Sn/Sb (95Sn/5Sb:232-240°C)

The 95Sn-5Sb solder is a solid solution of antimony in a tin

matrix. The relatively high melting point of this alloy makes

it suitable for high temperature applications. The antimony

imparts strength and hardness. In comparing the yield

strengths of several solder alloys, the strength of 95Sn/5Sb

was 37.2 N/mm2 and was nearly the same as 96.5Sn/3.5Ag

(37.7 N/mm2 ).[2] The high strength of this alloy causes the

lowest energy crack path to be at the solder/intermetallic

interface in the case of thinner intermetallics. As the

intermetallic thickens, the crack path is through the

intermetallic layer. Formation of the intermetallic compound

SbSn is possible at these levels of Sb. This phase has a cubic

structure with a high hardness. The wetting behavior was

measured on a wetting balance in air using a standard RMA

flux. The wetting force at 2 seconds for 95Sn/5Sb on a Cu

wire is significantly less than Sn/37Pb and Sn/3.5Ag. In

addition to marginal wetting performance, the toxicity of Sb

has also raised concerns. As with bismuth, antimony is also a

by-product in the production of lead.

9. In/Sn (52In/48Sn:118°C)

The melting point of this alloy makes it suitable to low

temperature applications. With regard to indium, it displays

good oxidation resistance, but is susceptible to corrosion in a

humid environment. It is also a very soft metal and has a

tendency to cold weld. In addition, the 52In/48Sn alloy

displays rather poor high temperature fatigue behavior, due to

its low melting point. The high indium content limits the

widespread use of this alloy due to cost and availability

constraints.

10. Sn/Zn (91Sn/9Zn:199°C)

The presence of zinc in solder alloys leads to oxidation and

corrosion. Samples of bulk alloys that were steam aged for 8

hours exhibited severe corrosion as evidenced by a purplish

color. In powder form, it reacts rapidly with acids and alkalis

and forms a gas. Zinc-containing solder alloys have been

known to react with the flux medium in as little as a day,

resulting in a paste that is “hard as a rock”. Thus, its

compatibility with fluxes and its storage stability is

questionable. The reflowed solder joints do not wet as well as

other lead-free alloys. When wave soldered, this alloy tends

to produce excessive dross. Therefore, manufacturability of

this alloy and zinc-bearing alloys in general is a concern.

11. Au/Sn (80Sn/20Au:280°C)

Au/Sn eutectic solder is a very strong, rigid solder due to the

formation of brittle intermetallic compounds. Problems of

cracked dies have been seen using Au/Sn eutectic solders in

die attach applications. It is not known if the cracks occur

from processing or during thermal cycling. The high cost of

this alloy restricts its use in many applications where cost is a

factor.

DISSOLUTION KINETICS OF COPPER

In the electronics industry, copper is commonly used as a

basis material for

• conductor traces and solder pads on the PCB

• lead frames of SO, QFP, PLCC, and other

components.

Alloys with a high tin content and a higher melting point

have a greater tendency to dissolve copper. If a larger

quantity of copper is dissolved from the base metal into

the solder material, there is excessive formation of the

Cu6Sn5 intermetallic phase. Solder joint reliability can

be adversely affected by the brittle nature of this

intermetallic compound, in particular the mechanical

properties of the solder joint, especially under high

impact conditions.[3]

The extent of copper dissolution in various alloys can be

evaluated by means of a simple test. A 50 gram weight

was attached to a 125 µm diameter copper wire. A small

quantity of A611 liquid flux was brushed on the copper

wire. Then the alloy was placed on the tip of a soldering

iron (pre-heated to 280°C), and the soldering iron was

positioned so it made contact with the copper wire. The time

it took to break the copper wire (i.e. until the copper dissolved

in the solder) was measured and recorded as the dissolution

time.

The following alloy compositions were tested:

• 60Sn/40Pb,

• 95Sn/4Ag/1Cu,

• Castin 96.2Sn/2.5Ag/0.5Sb/0.8Cu,

• 95.5Sn/4Ag/0.5Cu,

• 88Sn/3Ag/8.5Bi/0.5Cu,

• 88.42Sn/3.07Ag/8.51Bi,

• 99.3Sn/0.7Cu,

• 96.5Sn/3.5Ag.

60Sn/40Pb had the slowest rate of dissolution of the copper

wire as expected due to its lower Sn content. For the high tin

solders, the graph shows that the addition of 0.5% copper to

the solder alloy can decrease the dissolution rate dramatically.

In the case of the Sn/Ag/Bi alloy, the effect of adding 0.5%Cu

was to increase the dissolution time from 1.5 minutes for

Sn/Ag/Bi to 3 minutes for Sn/Ag/Bi/0.5Cu.

8.76

4.1

3.42 3

1.8 1.8 1.56 1.25

60Sn/40Pb

Sn-Ag-1Cu

Castin

Sn-Ag-Bi-Cu

Sn-Ag-0.5Cu

Sn-Cu

Sn-Ag-Bi

Sn-Ag

0

2

4

6

8

10

Dissolution Time (min)

Figure 1. Dissolution Kinetics of Copper in Several Solder Alloys.

THE EFFECT OF ISOTHERMAL AGING

It is important to study intermetallic growth formation

because in solder joints with coarsened Cu-Sn

intermetallics, fracture is brittle and occurs through the

intermetallic layer. An aging study was performed on

96.5Sn/ 3.5Ag and 95.5Sn/ 4Ag/ 0.5Cu solder alloys on

copper substrates. The intermetallic layer growth

characteristics of the two alloys were compared in order

to determine the effect of copper addition to Sn-Ag based

alloys.

Experimental Procedure

Two solder pastes were made, 96.5Sn/ 3.5Ag and

95.5Sn/ 4Ag/ 0.5Cu. The pastes were made with -325/

+500 mesh electronic grade (Type 3) powder. A 1000

gram batch of each paste was mixed in a small productionscale

Ross mixer at 89.5% metal loading/ 10.5% flux by

weight. Heraeus V365 no-clean/ halide-free flux was used.

The test pieces were 2” x 2” copper coupons cut from 0.021

inch thick, commercial grade alloy 110 copper foil. They

were then pressed flat and cleaned in acetone. The solder

paste was screen printed through an 8 mil thick, stainless

steel, laser cut stencil on a DEK 247 printer with all printing

parameters kept constant. Therefore, the solder volume is

presumably constant and was not considered a factor in this

study. Six coupons of each alloy were printed. The printing

characteristics of both pastes were very good.

The test coupons were reflowed at Heraeus in a nitrogen

convection reflow oven using a standard profile for the

pastes. The test pieces were then placed in a Lindberg/

Blue M air convection oven held at 150°C. The samples

were aged for periods of 2, 4, 11, 20, and 41 days (984

hours). Following aging, each sample was sectioned across 3

joints for metallographic examination. For each sample, the

average thickness of the resulting interfacial compound was

reported.

Table I. Intermetallic Thicknesses for Solder Alloys Aged at 150°C.

Time (hours) Intermetallic Thickness (µm)* (Cu3Sn + Cu6Sn5)

96.5Sn/ 3.5Ag 95.5Sn/ 4Ag/ 0.5Cu

0 0.25 + 2 0.25 + 2

48 0.5 + 3.25 0.5 + 2.5

96 0.75 + 2.5 0.75 + 2.5

264 1 + 2.5 1 + 3

480 1.5 + 3 1.5 + 4.5

984 2.5 + 4 2.5 + **

*The standard deviation for the measurements is on the order of 0.5 µm.

**A nonuniform morphology of the Cu6Sn5 layer precluded a characteristic thickness measurement.

Results and Discussion

It is widely known that copper is soluble in molten Sn-

Ag-X solders. The dissolution of copper results in the

formation of å-phase Cu3Sn and ç-phase Cu6Sn5. Due to

the concentration gradient, the Cu-rich Cu3Sn phase

forms adjacent to the copper substrate. Cu3Sn has a more

planar structure. The more Sn-rich Cu6Sn5 phase forms

adjacent to the Sn-based solder and has a scallop-edge

appearance. The reason why Cu6Sn5 has a scallop-edge

appearance may be due to the fact that Cu6Sn5 dissolves

faster along the grain boundary. Between the Cu6Sn5

grains, there are molten solder channels extending all the

way to the Cu3Sn/Cu interface. Since the Cu3Sn

compound layer is so thin, these channels serve as fast

diffusion and dissolution paths of Cu in the solder to feed

the interfacial reaction.[4] This interfacial layer grows

during solid-state aging as the tin and copper diffuse to

the interface and react.

The growth kinetics of the intermetallic compounds was

found to be similar as expected due to the similar Sn

contents and reflow temperatures of the two alloys. The

microstructural features of the Sn-Ag-X alloys are also

similar. The matrix is polygranular Sn with a grain size

in the as-solidified condition of approximately 1 µm.[5]

Five phases can be seen in the SEM micrographs given

in Figures 6 to 17: Sn, Ag3Sn, Cu6Sn5, Cu3Sn, and Cu.

CREEP DEFORMATION

Recent work indicates that similar failure mechanisms are

involved in thermal fatigue in shear and unidirectional creep

in shear. Also, since the temperatures during thermal fatigue

represent high solder homologous temperatures, creep

deformation is involved. Creep deformation is the timedependent

plastic flow of a material under constant load at

elevated temperature. As the homologous temperature (the

ratio of the test temperature to the melting temperature on an

absolute temperature scale) increases, the ease with which

plastic flow occurs also increases. Creep is significant at a

homologous temperature greater than 0.5. Therefore, creep

deformation occurs in solders even at room temperature.

Every high temperature excursion results in a straining of the

solder joint as the constraining materials expand different

amounts. By understanding the mechanisms that lead to

fatigue failures, researchers can use the appropriate

metallurgical strategy to slow down or stop these mechanisms

and thus develop an improved, more fatigue resistant solder

alloy.[6]

The elevated operating temperature and operative strain rates

imply that creep is the major deformation mode during low

cycle fatigue. Also, the observation that solder joint fatigue

failures and creep failures appear the result of similar

metallurgical mechanisms indicates that both techniques can

be used to study the fatigue failure mechanism and relative

solder alloy fatigue resistance. As such, it becomes important

to understand how the solder microstructure accommodates

the applied strain. For the Sn-Ag-X solders, the strain

accommodation occurs through the tin matrix at individual

Sn-Sn grain boundaries.[5]

Creep testing was performed on samples of the same

dimension and preparation method as that used for

standard tensile testing of solder alloys. All creep testing

was performed at International Tin Research Institute

(ITRI). To generate the creep-rupture data, the solder

alloys were cast into dumbbell-shaped test samples

having 20 mm gauge length and 2 mm diameter. They

were cast at a temperature of 50 degrees above the

liquidus into a heated steel mold. The mold was then

water cooled. Samples were then subjected to the

standard aging procedure of 24 hours at 125°C, in

addition to at least 24 hours at room temperature for the

benefit of stabilizing the microstructure as much as

possible.

Samples were held isothermally at both room

temperature and 100°C. A weight was hung from the

sample during the test representing an applied stress, and

the time to rupture was recorded. Samples and test

method conformed to the British Standard BS3500: part

3: 1969 “Method for Creep and Rupture Testing of

Metals.” Time to rupture was determined by measuring

electrical resistance across the sample; after fracture the

resistance became infinite and timing stopped. All

samples were tested in duplicate. Data was collected for

3 different loads (4, 8, and 16 MPa) at both temperatures.

Results for time to failure at 100, 500, and 1000 hours were

recorded. Loads were applied which were expected to give

lifetimes in the region of those times but extrapolation was

carried out to estimate values for the times required. The

results at 25°C are shown in Figure 2. To interpret the data,

compare the times to rupture for a similar applied stress on

the two alloys. For example, for an applied stress of 4 MPa,

Sn/40Pb failed after 265 hours, whereas the

95.5Sn/4Ag/0.5Cu alloy took 3000 hours for failure to occur.

Figure 3 presents the data collected at ITRI for several

candidate lead free alloys compared to Sn/40Pb. The

95.5Sn/4Ag/0.5Cu alloy performs the best at room

temperature compared to Sn/3.5Ag eutectic, Sn/0.7Cu

eutectic, and Sn/40Pb. As expected, the Sn/Ag/Cu and

Sn/Ag alloys behave similarly due to their similar

microstructural development. The graphic representation of

the Sn/Cu data greatly differs from that of the other three,

perhaps indicative of a different failure mechanism. The

results of creep testing at 100°C are presented in Figure 4. At

100°C, the Sn/Ag and Sn/Ag/Cu curves appear switched from

the 25°C results with the best performer now being the Sn/Ag

eutectic alloy. Figure 5 shows the creep-rupture data for the

Heraeus Sn/4Ag/0.5Cu alloy tested at both room temperature

and 100°C. As expected, higher temperatures allow materials

to creep at a faster rate, thereby reducing the time to failure.

Figure 3. Creep-Rupture Data for Several Candidate Lead Free Alloys Compared to Sn-40Pb at 25°C.

Figure 4. Creep-Rupture Data for Several Candidate Lead Free Alloys Compared to Sn-40Pb at 100°C.

 

Figure 5. Creep-Rupture Data for the Heraeus Sn/4Ag/0.5Cu Alloy Tested at 25°C and 100°C.

CONCLUDING REMARKS

Recent work with candidate lead free alloys indicate a

significant improvement in reliability over Sn/Pb.

Figures 2-4 clearly demonstrate a superior creep

resistance over Sn/Pb for all lead free alloys tested

including Sn/Ag eutectic, Sn/Cu eutectic, and

Sn/4Ag/0.5Cu at both room temperature and 100°C.

Although the Sn/Cu eutectic outperformed Sn/40Pb, it

did not perform as well as the Sn-Ag-X alloys.

An aging study of both the 95.5Sn/4Ag/0.5Cu and

96.5Sn/3.5Ag solder alloys was performed in order to

evaluate the growth kinetics of the intermetallic layers

following extended heat treatment. An understanding of

the microstructural evolution that occurs at the

solder/copper interface at elevated temperatures is helpful

to understand the failure mechanisms that dominate at

elevated temperatures. Creep occurs when materials

under constant stress, below the tensile stress, slowly

deform and finally fracture. The creep rate is dependent

on alloy composition and microstructure and is strongly

temperature dependent. Because Sn/Ag and Sn/Ag/Cu

have similar microstructures, they behave similarly

during isothermal aging and creep testing.

ACKNOWLEDGEMENTS

The author is grateful to acknowledge the support of the

International Tin Research Institute for mechanical

testing of the alloys. The metallographic preparation and

scanning electron microscopy of the samples used to study

aging was performed by F.W. Gayle and L. Smith at the

National Institute of Standards and Technology and is greatly

appreciated. Also, thanks to Dr. M.R. Notis of Lehigh

University and A. Z. Miric of Heraeus Germany for their

useful discussions.

REFERENCES

1. National Center for Manufacturing Sciences, Lead-Free

Solder Project Final Report, August 1997.

2. P.T. Vianco, K.L. Erickson, and P.L. Hopkins, “Solid

State Intermetallic Compound Growth Between Copper

and High Temperature, Tin-Rich Solders-Part I:

Experimental Analysis,” Sandia National Labs (Contract

Number DE-AC04-94AL85000), 1994.

3. G. Humpston and D.M. Jacobson, Principles of Soldering

and Brazing, ASM International, 1993.

4. H.K. Kim and K.N. Tu, “Kinetic Analysis of the

Soldering Reaction Between Eutectic SnPb Alloy and Cu

Accompanied by Ripening,” Physical Review B, Vol. 53,

No. 23, June 15, 1996, p. 16028.

5. D.R. Frear, “The Mechanical Behavior of Interconnect

Materials for Electronic Packaging,” J.Metals, (May

1996), pp. 49-53.

6. J.W. Morris, Jr. and D. Tribula, “Creep in Shear of

Experimental Solder Joints,” Journal of Electronic

Packaging, June 1990, Vol. 112, pp. 87-93.

 

 

 

Integrity of Solder Joints from Lead-free Solder Paste

Angela Grusd

Heraeus Inc.

West Conshohocken, PA

ABSTRACT

This past year the Lead Free Movement has taken center stage in the electronics assembly industry. In

December 1997, the Japan Environmental Agency proposed legislation on the disposal of lead scrap. Leadcontaining

scrap, such as electronics, must be disposed of in sealed landfill sites to prevent the leaching of lead.

A second Japanese proposal was initiated by the Ministry of International Trade and Industry (MITI) and the

Japan Automobile Industries Association. This called for a 50% voluntary reduction of lead in vehicles

(excluding batteries) by 2001 and a 66.67% reduction by 2003. The committee of JEIDA (Japan Electronic

Industry Development Association) and the lead-free soldering committee of JIEP (Japan Institute of Electronics

Packaging) outlined the lead-free road map for soldering on January 30, 1998. Following the road map, several

major Japanese Consumer Electronic Manufacturers initiated their own road map and publicly announced

accelerated plans to eliminate lead solder completely by 2001.

Matsushita (Panasonic):

It is reported that by 2001, all lead will be eliminated from major electronics products. Since October 1, 1998 a

compact mini-disc player has been available in Japan with a label that reads “Produced for the Environment.” It

contains a Sn/Ag/Bi/Cu lead free solder. The product is about half the size of a compact disc player, kind of a

portable walkman. In March of 1999, this product will be launched in Europe. They report that installation of

lead free solder can be achieved such that no problems regarding material properties, installaiton quality, and

reliability arise in actual production and implementation is being proceeded to meet their 2001 target.

Following this product, Matsushita will introduce lead free audio stereos, car stereos, and televisions into Japan

at the beginning of 1999 and in Europe in the summer of 1999. Japanese sites in Europe are already being

affected and advised to change to lead free by Japan. This is expected to have a great impact on the electronics

industry in the United States as well.

Sony:

By 2001, all lead will be eliminated except for high density electronics packaging.

Toshiba:

By 2000, lead will be eliminated from mobile phones.

Hitachi:

By 1999, lead usage reduced to half of that in 1997. By 2001, all products will be lead-free.

In April,1998 the Japanese started a project similar to the National Center for Manufacturing Sciences Pb-free

Project called the NEDO project. The project goal is to make a data base on lead-free soldering, to select leadfree

solders, and to establish the process for lead-free solders. The total budget is 350,000,000 yen/two years.

Project members are from major electronics companies, device companies, solder companies, and several

universities.

Another important alloy is the Sn/Ag/Cu system. Sn/Ag/Cu is also on the list of the JEIDA lead free roadmap.

In addition, the European Brite-Euram Consortia recommended Sn/Ag/Cu as the general purpose solder. The

Sn-Ag-X lead free solders offer superior creep resistance at room temperature and 100°C as compared to Sn-

40Pb. Results of this work will be presented as well as factors to consider when developing and implementing

lead free alloys such as manufacturability, availability, and cost. One of the most promising replacement alloys

is Sn/4Ag/0.5Cu. This alloy will be discussed in detail.

INTRODUCTION

Lead free alternatives are being considered for several reasons:

• Health Concerns

It is widely known that lead is related to certain health risks. If lead particles are inhaled or ingested, they

accumulate in the human body causing damage to the blood and central nervous systems. Lead poisoning can be

detected by blood analysis and the limits are defined by national governments. The standard upper value of

untainted human beings should not exceed 130 µg/l. The upper blood lead limit of people who are exposed to

lead at work is, according to the German Biological Place of Work Tolerance values (TRGS 903, June 1994),

300 µg/l blood for women and 700 µg/l for men. In the USA, these critical values have been lowered from 250

to 100 µg/l.

As a result, some industries have already eliminated lead and have found suitable alternatives, for example

plumbing solders, tinned cans, lead-free gasoline for vehicles, and lead-free cut crystal glass. The majority of

lead consumption is automobile batteries and ammunition. The lead consumption of the electronics industry is

relatively small and, according to different sources, lies between 0.5 - 7%.

When choosing alternative metals, consideration must also be given to their health risks as well. Recent studies

in the USA and in Europe came to the following conclusions concerning the toxicology of lead and some

alternative metals:[1]

• Cd is extremely toxic and should not be used (high risk). Many companies such as Ford Motor Company

have banned Cd-containing materials.

• Pb was also identified as highly toxic (high risk - it is considered harmful to the reproductive system).

• Sb is very toxic and should not be considered a major alloying element (medium risk - in Europe this

material is considered potentially carcinogenic).

• Ag and Cu are used in lead-free alloys in small concentrations - in Europe these materials are seen as low

risk.

• Sn and Zn are essential elements in the human diet, yet may be toxic if exposures are sufficiently high (low

risk).

• Bi is a relatively benign metal with a history of medicinal uses (low risk).

• Greater Thermal Stress of Components

In the automotive industry, more and more circuits are being placed in the engine compartment in order to

reduce the quantity of cables and, therefore, reduce cost. These under the hood conditions often see

temperatures in excess of 150°C. A leading automobile manufacturer has even measured temperatures as high

as 170°C on a hybrid circuit. The high thermal stresses imposed on the solder joints at these temperatures has

led automotive manufacturers to research Pb-free alternatives with high thermal fatigue resistance, because they

observed that Sn-37Pb has poor thermal fatigue properties even at 125°C. Higher temperatures dramatically

reduce the strength of the solder joint during thermal cycling, due to greater plastic deformation of the solder as

well as diffusion, recrystallization and grain growth inside the solder.

For conventional alloys such as Sn62/Pb36/Ag2 (melting point 179°C) and Sn63/Pb37 (melting point 183°C),

there is a major concern for the mechanical and microstructural stability and, therefore, the reliability of the

solder joint at an operating temperature of 150°C, because it is approaching the melting point of the alloy.

LEAD FREE ALTERNATIVE SOLDER ALLOYS

Sn/Ag

96.5Sn/3.5Ag 221°C

This alloy exhibits adequate wetting behavior and strength and is used in electronics as well as plumbing.

Several sources have also reported good thermal fatigue properties as compared to Sn/Pb. Thermal fatigue

damage in solders is accelerated at elevated temperatures. In the Sn/Pb system, the relatively high solid

solubilities of Pb in Sn and vice versa, especially at elevated temperatures, lead to microstructural instability due

to coarsening mechanisms. These regions of inhomogeneous microstructural coarsening are known to be crack

initiation sites. It is well-documented that these types of microstructures in Sn/Pb alloys fail by the formation of

a coarsened band in which a fatigue crack grows. By comparison, the Sn/Ag system, has limited solid solubility

of Ag in Sn, making it more resistant to coarsening. As a result, Sn/3.5Ag eutectic forms a more stable, uniform

microstructure that is more reliable.

Although the Sn/3.5Ag alloy itself exhibits good microstructural stability, when soldered to copper base metal,

the combination of a higher Sn content (96.5Sn compared to 63Sn) and higher reflow temperature environments

accelerate the diffusion rates for copper base metal in Sn. As its corresponding composition is reached, the

brittle Cu6Sn5 intermetallic compound is nucleated and begins to grow. To slow the diffusion rate and thereby

decrease the growth kinetics, alternative surface finishes such as immersion gold (Au over Ni over Cu) may be

used. The 2 µm Ni in the immersion gold coating serves as an effective diffusion barrier, limiting the Cu from

diffusing into the solder and forming the brittle Cu6Sn5 intermetallic compound. Other surface finishes such as

immersion silver (Ag over Cu) and immersion palladium (Pd over Cu) do not contain a Ni barrier layer. Their

effect on the growth kinetics of the intermetallic compound layers is under investigation.

Sn/Ag/Cu

95.5Sn/4.0Ag/0.5Cu 217-219°C

Because the mechanical stability of the joint is degraded when the melting point is approached, elevated

temperature cycling produces more damage for Sn/Pb solder (m.p. 183°C) as compared to higher melting point

solders. The melting temperatures of Sn/Ag/Cu solders make them ideal in high operating temperatures up to

175°C. As for wetting, Sn/Ag/Cu solders do not wet Cu as well as Sn/Pb using commercial fluxes. However,

good fillet formation can be easily achieved provided the fluxes are suitable for higher temperature use.

Soldering in nitrogen atmosphere also improves wettability using no-clean fluxes. The copper dissolution test

provides a relative measurement of the solder’s tendency to dissolve Cu from the base metal and form the

Cu6Sn5 intermetallic compound.

Sn/Cu

 99.3Sn/0.7Cu 227 °C

This alloy might be also suitable for high temperature applications required by the automotive industry. It is a

candidate especially for companies looking for lead and silver free alloys. Preliminary testing conducted on this

alloy has shown a significant improvement in creep/fatigue data over standard Sn-Pb alloys. However, the Sn-

Ag-X alloys perform better in creep testing.

Sn/Ag/Cu/Sb

96.2Sn/2.5Ag/0.8Cu/0.5Sb (known as Castin) 217-220 °C

This alloy has similar mechanical properties and reliability characteristics to the Sn/Ag/Cu alloy. However,

there is some concern regarding the toxicity of Sb.

Sn/Ag/Bi

91.8Sn/3.4Ag/4.8Bi 200-216°C

In general, bismuth is added to Sn/Ag/X solder alloys in order to depress the melting point. Another benefit of

Bi addition is greater joint strength as indicated by ring and plug testing. This particular alloy was developed by

Sandia National Labs. Sandia’s internal studies have found no electrical failures on surface mount devices

following 10,000 thermal cycles using 68 I/O PLCCs, 24 I/O SOICs, and 1206 chip capacitors on standard FR-4

PCBs. The boards were cycled 0 to 100°C at a ramp rate of 10°C/ minute. No cracks or deformation were

observed on boards cross-sectioned after 5000 thermal cycles. Cross-sectional data on 10,000 cycles is being

collected. These results are in good agreement with data collected by the National Center for Manufacturing

Sciences (NCMS) Lead Free Solder Project, which reported very good thermal fatigue resistance on OSP

printed circuit boards (Organic Solderability Preservative that protects copper pads and through-holes). The

NCMS High Temperature Fatigue Resistance Project is currently evaluating this solder at temperatures up to

160 and 175°C. In combination with Pb from the PCB or component metallization, a Sn/Bi/Pb ternary

compound is formed with a melting point of only 96°C. As the trend toward eliminating lead continues, this

alloy may become more attractive.

Sn/Ag/Bi/Cu

90Sn/2.0Ag/7.5Bi/0.5Cu (138) 198-212°C

Although the addition of Bi to the Sn/Ag/X system imparts greater strength and improved wetting, too much

bismuth (greater than 5%) leads to the presence of a small DSC peak near 138°C, corresponding to the binary

Sn/Bi eutectic at 138°C or the ternary Sn/Ag/Bi eutectic at 136.5°C. For this alloy with 7.5 weight percent Bi,

this corresponds to approximately 1% of the total melting. This small amount of eutectic melting has an

uncertain effect on joint reliability as the temperature approaches 138°C. This combined with the

aforementioned concern of forming a SnPbBi compound at 96°C, makes this alloy an unlikely candidate for a

Pb-free solder.

Sn/Bi

42Sn/58Bi 138°C

The low melting point of this alloy makes it suitable for soldering temperature-sensitive components and

substrates. If these contain Pb, the SnPbBi ternary eutectic compound may form at 96°C, which in turn adversely

affects the thermal fatigue properties. The NCMS Lead Free Solder Project recently reported the results of

thermal cycle testing at 0/ 100°C and -55/ +125°C for over 5000 cycles on OSP boards. The result was that the

Sn/Bi outperformed the Sn/Pb at both temperature excursions. It was thought that the closeness of 125°C to the

binary Sn/Bi eutectic at 138°C would cause this alloy to be a poor performer. Two possible explanations for this

unexpected result were presented. The Sn/Bi alloy may be annealing at 125°C, relaxing the stresses produced

during thermal cycling. A second explanation was the alloy may be undergoing recrystallization. Also, no filletlifting

was observed due to the eutectic nature of the alloy.

Sn/Sb

95Sn/5Sb 232-240°C

The 95Sn-5Sb solder is a solid solution of antimony in a tin matrix. The relatively high melting point of this

alloy makes it suitable for high temperature applications. The antimony imparts strength and hardness. In

comparing the yield strengths of several solder alloys, the strength of 95Sn/5Sb was 37.2 N/mm2 and was nearly

the same as 96.5Sn/3.5Ag (37.7 N/mm2 ).[2] The high strength of this alloy causes the lowest energy crack path

to be at the solder/intermetallic interface in the case of thinner intermetallics. As the intermetallic thickens, the

crack path is through the intermetallic layer. Formation of the intermetallic compound SbSn is possible at these

levels of Sb. This phase has a cubic structure with a high hardness. The wetting behavior was measured on a

wetting balance in air using a standard RMA flux. The wetting force at 2 seconds for 95Sn/5Sb on a Cu wire is

significantly less than Sn/37Pb and Sn/3.5Ag. In addition to marginal wetting performance, the toxicity of Sb

has also raised concerns. As with bismuth, antimony is also a by-product in the production of lead.

In/Sn

52In/48Sn 118°C

The melting point of this alloy makes it suitable to low temperature applications. With regard to indium, it

displays good oxidation resistance, but is susceptible to corrosion in a humid environment. It is also a very soft

metal and has a tendency to cold weld. In addition, the 52In/48Sn alloy displays rather poor high temperature

fatigue behavior, due to its low melting point. The high indium content limits the widespread use of this alloy

due to cost and availability constraints.

Sn/Zn

91Sn/9Zn 199°C

The presence of zinc in solder alloys leads to oxidation and corrosion. Samples of bulk alloys that were steam

aged for 8 hours exhibited severe corrosion as evidenced by a purplish color. In powder form, it reacts rapidly

with acids and alkalis and forms a gas. Zinc-containing solder alloys have been known to react with the flux

medium in as little as a day, resulting in a paste that is “hard as a rock”. Thus, its compatibility with fluxes and

its storage stability is questionable. The reflowed solder joints do not wet as well as other lead-free alloys.

When wave soldered, this alloy tends to produce excessive dross. Therefore, manufacturability of this alloy and

zinc-bearing alloys in general is a concern.

Au/Sn

80Sn/20Au 280°C

Au/Sn eutectic solder is a very strong, rigid solder due to the formation of brittle intermetallic compounds.

Problems of cracked dies have been seen using Au/Sn eutectic solders in die attach applications. It is not known

if the cracks occur from processing or during thermal cycling. The high cost of this alloy restricts its use in

many applications where cost is a factor.

DISSOLUTION KINETICS OF COPPER

In the electronics industry, copper is commonly used as a basis material for

• conductor traces and solder pads on the PCB

• lead frames of SO, QFP, PLCC, and other components.

Alloys with a high tin content and a higher melting point have a greater tendency to dissolve copper. If a larger

quantity of copper is dissolved from the base metal into the solder material, there is excessive formation of the

Cu6Sn5 intermetallic phase. Solder joint reliability can be adversely affected by the brittle nature of this

intermetallic compound, in particular the mechanical properties of the solder joint, especially under high impact

conditions.[3]

The extent of copper dissolution in various alloys can be evaluated by means of a simple test. A 50 gram weight

was attached to a 125 µm diameter copper wire. A small quantity of A611 liquid flux was brushed on the copper

wire. Then the alloy was placed on the tip of a soldering iron (pre-heated to 280°C), and the soldering iron was

positioned so it made contact with the copper wire. The time it took to break the copper wire (i.e. until the

copper dissolved in the solder) was measured and recorded as the dissolution time.

The following alloy compositions were tested: 60Sn/40Pb, 95Sn/4Ag/1Cu, Castin 96.2Sn/2.5Ag/0.5Sb/0.8Cu,

95.5Sn/4Ag/0.5Cu, 88Sn/3Ag/8.5Bi/0.5Cu, 88.42Sn/3.07Ag/8.51Bi, 99.3Sn/0.7Cu, and 96.5Sn/3.5Ag.

60Sn/40Pb had the slowest rate of dissolution of the copper wire as expected due to its lower Sn content. For

the high tin solders, the graph shows that the addition of 0.5% copper to the solder alloy can decrease the

dissolution rate dramatically. In the case of the Sn/Ag/Bi alloy, the effect of adding 0.5%Cu was to increase the

dissolution time from 1.5 minutes for Sn/Ag/Bi to 3 minutes for Sn/Ag/Bi/0.5Cu.

Figure 1. Dissolution Kinetics of Copper in Several Solder Alloys.

THE EFFECT OF ISOTHERMAL AGING

It is important to study intermetallic growth formation because in solder joints with coarsened Cu-Sn

intermetallics, fracture is brittle and occurs through the intermetallic layer. An aging study was performed on

96.5Sn/ 3.5Ag and 95.5Sn/ 4Ag/ 0.5Cu solder alloys on copper substrates. The intermetallic layer growth

characteristics of the two alloys were compared in order to determine the effect of copper addition to Sn-Ag

based alloys.

Experimental Procedure

Two solder pastes were made, 96.5Sn/ 3.5Ag and 95.5Sn/ 4Ag/ 0.5Cu. The pastes were made with -325/ +500

mesh electronic grade (Type 3) powder. A 1000 gram batch of each paste was mixed in a small productionscale

Ross mixer at 89.5% metal loading/ 10.5% flux by weight. Heraeus V365 no-clean/ halide-free flux was

used.

The test pieces were 2” x 2” copper coupons cut from 0.021 inch thick, commercial grade alloy 110 copper foil.

They were then pressed flat and cleaned in acetone. The solder paste was screen printed through an 8 mil thick,

stainless steel, laser cut stencil on a DEK 247 printer with all printing parameters kept constant. Therefore, the

solder volume is presumably constant and was not considered a factor in this study. Six coupons of each alloy

were printed. The printing characteristics of both pastes were very good.

The test coupons were reflowed at Heraeus in a nitrogen convection reflow oven using a standard profile for the

pastes. The test pieces were then placed in a Lindberg/ Blue M air convection oven held at 150°C. The samples

were aged for periods of 2, 4, 11, 20, and 41 days (984 hours). Following aging, each sample was sectioned

across 3 joints for metallographic examination. For each sample, the average thickness of the resulting

interfacial compound was reported.

Table I. Intermetallic Thicknesses for Solder Alloys Aged at 150°C.

Time (hours) Intermetallic Thickness (µm)* (Cu3Sn + Cu6Sn5)

96.5Sn/ 3.5Ag 95.5Sn/ 4Ag/ 0.5Cu

0 0.25 + 2 0.25 + 2

48 0.5 + 3.25 0.5 + 2.5

96 0.75 + 2.5 0.75 + 2.5

264 1 + 2.5 1 + 3

480 1.5 + 3 1.5 + 4.5

984 2.5 + 4 2.5 + **

*The standard deviation for the measurements is on the order of 0.5 µm.

**A nonuniform morphology of the Cu6Sn5 layer precluded a characteristic thickness measurement.

Results and Discussion

It is widely known that copper is soluble in molten Sn-Ag-X solders. The dissolution of copper results in the

formation of å-phase Cu3Sn and ç-phase Cu6Sn5. Due to the concentration gradient, the Cu-rich Cu3Sn phase

forms adjacent to the copper substrate. Cu3Sn has a more planar structure. The more Sn-rich Cu6Sn5 phase

forms adjacent to the Sn-based solder and has a scallop-edge appearance. The reason why Cu6Sn5 has a scallopedge

appearance may be due to the fact that Cu6Sn5 dissolves faster along the grain boundary. Between the

Cu6Sn5 grains, there are molten solder channels extending all the way to the Cu3Sn/Cu interface. Since the

Cu3Sn compound layer is so thin, these channels serve as fast diffusion and dissolution paths of Cu in the solder

to feed the interfacial reaction.[4] This interfacial layer grows during solid-state aging as the tin and copper

diffuse to the interface and react.

The growth kinetics of the intermetallic compounds was found to be similar as expected due to the similar Sn

contents and reflow temperatures of the two alloys. The microstructural features of the Sn-Ag-X alloys are also

similar. The matrix is polygranular Sn with a grain size in the as-solidified condition of approximately 1 µm.[5]

Five phases were identified in the SEM micrographs: Sn, Ag3Sn, Cu6Sn5, Cu3Sn, and Cu.

CREEP DEFORMATION

Recent work indicates that similar failure mechanisms are involved in thermal fatigue in shear and unidirectional

creep in shear. Also, since the temperatures during thermal fatigue represent high solder homologous

temperatures, creep deformation is involved. Creep deformation is the time-dependent plastic flow of a material

under constant load at elevated temperature. As the homologous temperature (the ratio of the test temperature to

the melting temperature on an absolute temperature scale) increases, the ease with which plastic flow occurs also

increases. Creep is significant at a homologous temperature greater than 0.5. Therefore, creep deformation

occurs in solders even at room temperature. Every high temperature excursion results in a straining of the solder

joint as the constraining materials expand different amounts. By understanding the mechanisms that lead to

fatigue failures, researchers can use the appropriate metallurgical strategy to slow down or stop these

mechanisms and thus develop an improved, more fatigue resistant solder alloy.[6]

The elevated operating temperature and operative strain rates imply that creep is the major deformation mode

during low cycle fatigue. Also, the observation that solder joint fatigue failures and creep failures appear the

result of similar metallurgical mechanisms indicates that both techniques can be used to study the fatigue failure

mechanism and relative solder alloy fatigue resistance. As such, it becomes important to understand how the

solder microstructure accommodates the applied strain. For the Sn-Ag-X solders, the strain accommodation

occurs through the tin matrix at individual Sn-Sn grain boundaries.[5]

Creep testing was performed on samples of the same dimension and preparation method as that used for

standard tensile testing of solder alloys. All creep testing was performed at International Tin Research Institute

(ITRI). To generate the creep-rupture data, the solder alloys were cast into dumbbell-shaped test samples

having 20 mm gauge length and 2 mm diameter. They were cast at a temperature of 50 degrees above the

liquidus into a heated steel mold. The mold was then water cooled. Samples were then subjected to the

standard aging procedure of 24 hours at 125°C, in addition to at least 24 hours at room temperature for the

benefit of stabilizing the microstructure as much as possible.

Samples were held isothermally at both room temperature and 100°C. A weight was hung from the sample

during the test representing an applied stress, and the time to rupture was recorded. Samples and test method

conformed to the British Standard BS3500: part 3: 1969 “Method for Creep and Rupture Testing of Metals.”

Time to rupture was determined by measuring electrical resistance across the sample; after fracture the

resistance became infinite and timing stopped. All samples were tested in duplicate. Data was collected for 3

different loads (4, 8, and 16 MPa) at both temperatures. Results for time to failure at 100, 500, and 1000 hours

were recorded. Loads were applied which were expected to give lifetimes in the region of those times but

extrapolation was carried out to estimate values for the times required. The results at 25°C are shown in Figure

2. To interpret the data, compare the times to rupture for a similar applied stress on the two alloys. For

example, for an applied stress of 4 MPa, Sn/40Pb failed after 265 hours, whereas the 95.5Sn/4Ag/0.5Cu alloy

took 3000 hours for failure to occur.

Figure 3 presents the data collected at ITRI for several candidate lead free alloys compared to Sn/40Pb. The

95.5Sn/4Ag/0.5Cu alloy performs the best at room temperature compared to Sn/3.5Ag eutectic, Sn/0.7Cu

eutectic, and Sn/40Pb. As expected, the Sn/Ag/Cu and Sn/Ag alloys behave similarly due to their similar

microstructural development. The graphic representation of the Sn/Cu data greatly differs from that of the other

three, perhaps indicative of a different failure mechanism. The results of creep testing at 100°C are presented in

Figure 4. At 100°C, the Sn/Ag and Sn/Ag/Cu curves appear switched from the 25°C results with the best

performer now being the Sn/Ag eutectic alloy. Figure 5 shows the creep-rupture data for the Heraeus

Sn/4Ag/0.5Cu alloy tested at both room temperature and 100°C. As expected, higher temperatures allow

materials to creep at a faster rate, thereby reducing the time to failure.

Figure 2. Creep-Rupture Data for Heraeus Sn-4Ag-0.5Cu and Sn-40Pb at 25°C.

Figure 3. Creep-Rupture Data for Several Candidate Lead Free Alloys Compared to Sn-40Pb at 25°C.

Figure 4. Creep-Rupture Data for Several Candidate Lead Free Alloys Compared to Sn-40Pb at 100°C.

Figure 5. Creep-Rupture Data for the Heraeus Sn/4Ag/0.5Cu Alloy Tested at 25°C and 100°C.

Valuable Lessons Learned

This first lesson appears too easy and is often overlooked. When testing Bi-containing lead free alloys, many

companies have had to scrap many hours worth of otherwise valuable data because they used the standard Sn/Pb

component finish or PWB board metallization. As reported earlier, in combination with Pb from the PWB or

component metallization, a Sn/Bi/Pb ternary compound is formed with a melting point of only 96°C. These

joints would not be able to undergo thermal cycling and are technically unsalvageable.

This second valuable lesson is not so obvious and was discovered by the NCMS High Temperature Fatigue

Resistant Solder Project Team. Components used for the PWB and hybrid assemblies for the

Thermomechanical Fatigue Test build had leads or castellations, in the case of the LCCCs, pretinned with pure

tin. The large volume of Sn on the leads and castellations resulted in a contamination of the final solder joints.

The problem was initially identified in the case of the high lead alloys because they melted at the eutectic

temperature of 183°C and had the eutectic Sn/Pb morphology. The problem was less severe in the Sn-based

lead free alloys but still a major problem. With the LCC 44 component, a 50% dilution of the alloys was

reported. The Sn contamination for the PLCCs was far less as expected, since the “tinning” volume is much

less, with only a 15 -20 micron thick coating on the leads. A 10% dilution was reported for the PLCC

components. The selected alloys were far enough from their targeted compositions that the NCMS Team chose

to repeat the entire assembly. In the future work of the project, components will not be supplied with pure tin or

tin-lead solder. The preferred alternative is to prealloy the components with the test alloys to avoid

contamination.

CONCLUDING REMARKS

Recent work with candidate lead free alloys indicate a significant improvement in reliability over Sn/Pb.

Figures 2-4 clearly demonstrate a superior creep resistance over Sn/Pb for all lead free alloys tested including

Sn/Ag eutectic, Sn/Cu eutectic, and Sn/4Ag/0.5Cu at both room temperature and 100°C. Although the Sn/Cu

eutectic outperformed Sn/40Pb, it did not perform as well as the Sn-Ag-X alloys.

An aging study of both the 95.5Sn/4Ag/0.5Cu and 96.5Sn/3.5Ag solder alloys was performed in order to

evaluate the growth kinetics of the intermetallic layers following extended heat treatment. An understanding of

the microstructural evolution that occurs at the solder/copper interface at elevated temperatures is helpful to

understand the failure mechanisms that dominate at elevated temperatures. Creep occurs when materials under

constant stress, below the tensile stress, slowly deform and finally fracture. The creep rate is dependent on alloy

composition and microstructure and is strongly temperature dependent. Because Sn/Ag and Sn/Ag/Cu have

similar microstructures, they behave similarly during isothermal aging and creep testing.

ACKNOWLEDGEMENTS

The author is grateful to acknowledge the support of the International Tin Research Institute for mechanical

testing of the alloys. The metallographic preparation and scanning electron microscopy of the samples used to

study aging was performed by F.W. Gayle and L. Smith at the National Institute of Standards and Technology

and is greatly appreciated. Also, thanks to Dr. M.R. Notis of Lehigh University and A. Z. Miric of Heraeus

Germany for their useful discussions.

REFERENCES

1. National Center for Manufacturing Sciences, Lead-Free Solder Project Final Report, August 1997.

2. P.T. Vianco, K.L. Erickson, and P.L. Hopkins, “Solid State Intermetallic Compound Growth Between

Copper and High Temperature, Tin-Rich Solders-Part I: Experimental Analysis,” Sandia National Labs

(Contract Number DE-AC04-94AL85000), 1994.

3. G. Humpston and D.M. Jacobson, Principles of Soldering and Brazing, ASM International, 1993.

4. H.K. Kim and K.N. Tu, “Kinetic Analysis of the Soldering Reaction Between Eutectic SnPb Alloy and Cu

Accompanied by Ripening,” Physical Review B, Vol. 53, No. 23, June 15, 1996, p. 16028.

5. D.R. Frear, “The Mechanical Behavior of Interconnect Materials for Electronic Packaging,” J.Metals, (May

1996), pp. 49-53.

6. J.W. Morris, Jr. and D. Tribula, “Creep in Shear of Experimental Solder Joints,” Journal of Electronic

Packaging, June 1990, Vol. 112, pp. 87-93.

 

Previously published in German in

Productronic, 11/97.

©Productronic, 1997

Anton Zoran Miric

W.C. Heraeus GmbH, Hanau, Germany

Angela Grusd

Heraeus Cermalloy Division, West Conshohocken, Pennsylvania, USA

[ 19 ]

 

 

 

 

Lead-free alloys

In recent years, efforts to develop alternatives to leadbased solders have increased dramatically. These efforts began as a response to potential legislation and regulations restricting lead usage in the electronics industry. Lead is extremely toxic when inhaled or ingested. As researchers began to focus on Pb-free solders, they recognized their value in high temperature applications (e.g. automotive manufacturing) where Sn/Pb solders do not meet the requirements.

There are many factors to consider when developing lead-free alloys: manufacturability, availability, reliability, cost and environmental safety. Of these, the most challenging and time consuming is the reliability of alternative solders. The lead-free alloys available cannot be used as a drop-in replacement for the SnPb or SnPbAg. The introduction of lead-free solder alloys may mean having to use alternative component and PCB metallizations, PCB materials, solder fluxes, etc.

1. Introduction

Lead is used in the electronics industry as a part of the solder material alloy. For some time, now, alternative materials have been called for, for several reasons.

Greater thermal stress of components

In the automotive industry, more and more circuits are being placed in the engine compartment. This can reduce the quantity of cables, the vehicle weight and bring a cost benefit. Under such conditions the thermal stress is caused by temperatures in excess of 150°C. A leading automobile manufacturer has measured temperatures of 170°C on a hybrid circuit and therefore requires good thermal fatigue properties at temperatures up to 175°C. Higher temperatures dramatically reduce the firmness of the solder joint during thermal cycling, due to:

• greater plastic deformation of the solder;

• recrystallisation and grain growth inside the solder.

At such high temperatures, the mechanical properties of the conventional alloys Sn62/Pb36/Ag2 (melting point 179°C) and Sn63/Pb37 (melting point 183°C) deteriorate. The

service temperature is c. 97 per cent of the melting temperature and there is a major concern for the mechanical and microstructural stability and the reliability of the solder joint.

Minimisation of health risks

The use of lead alloys is related to certain health risks, e.g. in the USA, the use of lead-containing wood varnishes for outdoor applications has caused soil pollution. Some industries already do without lead: tinned foods, water mains, lead-free petrol for vehicles, lead-free colours. Lead-free cut crystal glass was presented at the Ambiente ’97 in Frankfurt.

If chronically exposed to lead, an accumulation of lead in the human body is possible, which may cause damage to the blood and central nervous systems. Lead poisoning can be detected by blood analysis and the limits are defined by national governments. The standard upper value of untainted human beings should not exceed 130µg/l. The upper lead limit in the blood of people who are exposed to lead at work is, according to the German Biological Place of Work Tolerance values (TRGS 903, June 1994), 300µg/l blood for women and 700µg/l for men. In the USA, this critical value has been lowered from 250 to 100µg/l.

In the manufacturing environment, lead constitutes only a minor direct danger. The only risk is caused by breathing in floating, dried powder particles, or through the occurrence of gross mistakes in meeting necessary safety measures.

A greater danger lies in the contamination of ground water and soil by electronic scrap (disposal of electronic assemblies and solder dross).

Several proposals have been made to limit, tax or ban the use of lead in electronics. The US Congress has received several proposals. One contained a proposal for a general tax on lead production. The European Parliament received a proposal aimed at abolishing lead and mercury (batteries excluded), cadmium, etc., in all vehicles, licensed as of 1 January 2002.

Nevertheless, one has to remember that the major part of lead consumption is for automobile batteries and ammunition.

The lead consumption of the electronics industry is relatively small and, according to different sources, lies between approximately 7 per cent and 0.5 per cent.

When choosing alternative metals, consideration must also be given to their health risks as well. Recent studies in the USA and in Europe came to the following conclusions concerning the toxicology of lead and some alternative metals:

• Cd is extremely toxic and should not be used (high risk).

• Pb was also identified as highly toxic (high risk – in Europe it is considered harmful to the reproductive system).

• Sb is very toxic and should not be considered a major alloying element (medium risk – in Europe this material is considered potentially carcinogenic).

• Ag and Cu are used in lead-free alloys in small concentrations – in Europe these materials are seen as low risk.

• Sn and Zn are essential elements in the human diet, yet may be toxic if exposures are sufficiently high (low risk).

• Bi is a relatively benign metal with a history of medicinal uses (low risk).

Use of temperature-sensitive components and substrates

Many industries strive to reduce costs. Components and substrates are used that cannot withstand the general common reflow temperatures of 210 to 230°C. The low melting temperature solder alloys are used in this case.

This is especially apparent with consumer electronics, as the operation temperatures are from Tmin 0°C to Tmax +60°C. This is a very low stress compared to that found in automobiles (inside the vehicle Tmin –55°C to Tmax +95°C, engine compartment Tmin –55°C to Tmax +125 to 175°C).

2. Alternative materials and solder

alloys

Some alternative basis metals and their alloys are shown in Table I.

Sn/Ag: 96.5Sn/3.5Ag; 221°C

This alloy exhibits adequate wetting behaviour and strength and is used in electronics as well as soldering water mains.

Several sources have also reported good thermal fatigue properties as compared to Sn/Pb. Thermal fatigue damage in solders is accelerated at elevated temperatures. In the Pb- Sn system, the relatively high solid solubilities of Pb in Sn and vice versa, especially at elevated temperatures, lead to microstructural instability due to coarsening mechanisms.

These regions of inhomogeneous microstructural coarsening are known to be crack initiation sites. It is well-documented that these types of microstructures in Pb-Sn alloys fail by the formation of a coarsened band in which a fatigue crack grows. By comparison, the Sn-Ag system has limited solid solubility of Ag in Sn, making it more resistant to coarsening. As a result, Sn-3.5Ag forms a more stable, uniform microstructure that is more reliable.

Although the Sn-3.5Ag alloy itself exhibits good microstructural stability, when soldered to copper base metal, the combination of a higher Sn content (96.5Sn compared to 63Sn) and higher reflow temperature environments accelerates the diffusion rates for copper base metal in Sn. As its corresponding composition is reached, the brittle Cu6Sn5 intermetallic compound is nucleated and Soldering & Surface Mount

Technology

Alternative basis metals to lead and their alloys Worldwide production/ Melting Priceb capacity/capacity reserve Typical Price of alloya Melting Element point (°C) Toxica (US$/kg) (thousand tons)c composition (US$/kg) point (°C) Remarks Suitable Sn 232 No 5.3 160/281/81 Pb 327 Highly 0.6 Very high 62Sn/36Pb/2Ag 6.4 179 Well-established alloys; low price Yes 63Sn/37Pb 3.6 183 Ag 960 No 142.5 13/15/1.5 96.5Sn/3.5Ag 10.1 221 Good thermal fatigue; fast Cu dissolution Yes Cu 1,083 No 2.4 8.000/10.200/2.200 95.5Sn/4Ag/0.5Cu 10.8 216-219 Preliminary tests show good thermal fatigue Yes 95.5Sn/3.8Ag/0.7Cu 10.5 217-219 properties at high temperatures; favourably 95Sn/4Ag/1Cu 10.8 216-219 priced 99.3Sn/0.7Cu 5.3 227 Bi 271 No 8.2 4/8/4 58Bi/42Sn 7.0 138 Suitable for low temperature applications Yes 90Sn/2Ag//7.5Bi/0.5Cu 8.2 198-212 Creation of low melt. Sn/Bi phase at 138°C 91.8 Sn/3.4Ag/4.8Bi 10.0 200-216 Fatigue at high temperature under evaluation Sb 630 Very 2.0 78.000/122.000/44.000 95SN/5Sb 5.1 232-240 High melting point; toxicity concern Yes 65SN/25Ag/10Sb 39.3 230-235 Creation of needle-shaped Ag3Sn phases 96.7Sn/2Ag/0.8Cu/0.5Sb 8.0 217-220 Similar properties as Sn/Ag/Cu In 157 Very 250 0.1/0.2/0.1 52In/48Sn 132.5 118 Expensive; limited availability; indium tends to No 97In/3Ag 246.8 143 corrode in combination with humidity and is 77.2 Sn/20In/2.8Ag 58.1 189 very soft 86.4Sn/11In/2Ag/0.6Sb 34.9 Creation of low-melt. In/Sn phase at 118°C Zn 419 No 1.5 6.900/7.600/700 91Sn/9Zn 5.0 199 Problem of oxidation. Strong dross formation No Au 1,063 No 10,509 Low 80Au/20Sn 8,403.3 280 Much too expensive No Cd 320 Extremely 1.2 Medium 67Sn/33Cd 3.9 170 Toxicity No Notes: aInformation relates to pure metals only bBudget prices valid in July 1997 cCurrent lead consumption worldwide = approximately 60,000 tons per annum (source: US Bureau of Mines)

 

10/1 [1998] 19–25 MCB University Press [ISSN 0954-0911] [ 20 ]

Anton Zoran Miric and Angela Grusd Lead-free alloys Soldering & Surface Mount Technology 10/1 [1998] 19–25 Table I [ 21 ]

Anton Zoran Miric and Angela Grusd Lead-free alloys Soldering & Surface Mount Technology 10/1 [1998] 19–25 begins to grow. To slow the diffusion rate and thereby decrease the growth kinetics, alternative surface finishes such as immersion gold (Au over Ni over Cu) may be used.

The Ni in the immersion gold coating serves as a diffusion barrier, limiting the Cu from diffusing into the solder and forming the brittle Cu6Sn5 intermetallic compound. Other

surface finishes such as immersion silver (Ag over Cu) and immersion palladium (Pd over Cu) do not contain a Ni barrier layer. Their effect on the growth kinetics of the intermetallic compound layers is under investigation.

Sn/Ag/Cu

1 95.5Sn/4.0Ag/0.5Cu; 216-219°C

2 95.5Sn/3.8Ag/0.7Cu; 217-219°C

3 95.0Sn/4.0Ag/1.0Cu; 216-219°C

4 93.6Sn/4.7Ag/1.7Cu; 216-218°C

Because the mechanical stability of the joint is degraded when the melting point is approached, elevated temperature cycling produces more damage for Sn/Pb solder (melting

point 183°C) as compared to higher melting point solders.

The melting temperatures of Sn-Ag-Cu solders make them ideal in high operating temperatures up to 175°C. As for wetting, Sn-Ag-Cu solders do not wet Cu as well as Sn/Pb

using commercial fluxes. However, improved wetting is possible if fluxes are designed specifically for higher temperature use. Soldering in nitrogen atmosphere also improves wettability (see Plates 1 and 2). The copper dissolution test provides a relative measurement of the solder’s tendency to dissolve Cu from the base metal and form the Cu6Sn5 intermetallic compound. For alloys 1-3, the rate of copper dissolution is slower than the Sn/Ag alloy yet faster than the Sn/Pb eutectic. For alloy 4, the high level of Cu in the alloy prevented the dissolution of the copper wire.

 

Sn/Cu: 99.3Sn/0.7Cu; 227°C

This alloy might be also suitable for high temperature applications required by the automotive industry. It is a candidate especially for companies looking for lead and silver free alloys. Preliminary testing conducted on this alloy has shown good fatigue performance.

Sn/Ag/Cu/Sb: 96.7Sn/2Ag/0.8Cu/0.5Sb

(known as Castin-Alloy); 217-220°C

This alloy has similar properties to the Sn/Ag/Cu alloy.

Sn/Ag/Bi: 91.8Sn/3.4Ag/4.8Bi; 200-216°C

In general, bismuth is added to Sn-Ag-X solder alloys in order to depress the melting point. Another benefit of Bi addition is greater joint strength as indicated by ring and plug testing. This particular alloy was developed by Sandia National Labs. Sandia’s internal studies have found no electrical failures on surface mount devices following 10,000 thermal cycles using 68 I/O PLCCs, 24 I/O SOICs, and 1,206 chip capacitors on standard FR-4 PCBs. The boards were cycled 0 to 100°C at a ramp rate of 10°C/ minute. No cracks or deformation were observed on boards cross-sectioned after 5,000 thermal cycles. Cross-sectional data on 10,000 cycles boards are being collected. These results are in good agreement with data collected by the NCMS Lead Free Solder Project, which reported very good thermal fatigue resistance on OSP printed circuit boards (Organic Solderability Preservative that protects copper pads and through-holes). The NCMS High Temperature Fatigue Resistance Project is currently evaluating this solder at temperatures up to 160 and 175°C.

In combination with Pb from the PCB or component metallization, a Bi/Pb compound is formed with a melting point of only 97°C. As the trend toward eliminating lead continues, this alloy may become more attractive.

Sn/Ag/Bi/Cu: 90Sn/2.0Ag/7.5Bi/0.5Cu;

(138) 198-212°C

Although the addition of Bi to the Sn-Ag-X system imparts greater strength and improved wetting, too much bismuth (greater than 5 per cent) leads to the presence of a small DSC peak near 138°C, corresponding to the binary Sn/Bi eutectic at 138°C or the ternary Sn/Ag/Bi eutectic at 136.5°C. For this alloy with 7.5 weight per cent Bi, this corresponds to approximately 1 per cent of the total melting. This small amount of eutectic melting has an uncertain effect on joint reliability as the temperature approaches 138°C. This combined with the aforementioned concern of forming a BiPb compound at 97°C, makes this alloy an unlikely candidate for a Pb-free solder.

Sn/Bi: 42Sn/58Bi; 138°C

The low melting point of this alloy makes it suitable for soldering temperature-sensitive components and substrates. If these contain Pb, the BiPb compound may form at 97°C, Plate 2

Solder joints with a lead-free Sn/Ag/Cu alloy – reflow in nitrogen Plate 1

Solder joints with a lead-free Sn/Ag/Cu alloy – reflow in normal temperature [ 22 ]

Anton Zoran Miric and Angela Grusd Lead-free alloys Soldering & Surface Mount Technology 10/1 [1998] 19–25 which in turn adversely affects the thermal fatigue properties.

The NCMS Lead Free Solder Project recently reported the results of thermal cycle testing at 0/100°C and –55/+125°C for over 5,000 cycles on OSP boards. The result was that the Sn/Bi outperformed the Sn/Pb at both temperature excursions. It was thought that the closeness of 125°C to the binary Sn/Bi eutectic at 138°C would cause this alloy to be a poor performer. Two possible explanations for this unexpected result were presented. The Sn/Bi alloy may be annealing at 125°C, relaxing the stresses produced during thermal cycling. A second explanation was the alloy may be undergoing recrystallization.

From a raw material perspective, bismuth is a by-product of lead mining. If the output of lead decreases, less Bi will be obtained. This may cause a rise in Bi costs. If this alloy were used, known resources would last only 16 years as its worldwide consumption would double. There are other disadvantages to this high bismuth-containing alloy. At this high bismuth content oxidation occurs very rapidly when the alloy is exposed to air. This requires the use of adequately activated fluxes. Owing to its high Bi content, this alloy displays little elasticity.

Sn/Sb: 95Sn/5Sb; 232-240°C

The relatively high melting point of this alloy makes it

suitable for high temperature applications. The antimony

has the effect of imparting strength and hardness to the

alloy. Formation of the intermetallic compound SbSn is

possible at these levels of Sb. This phase has a cubic structure

with a high hardness. The wetting behaviour was

measured on a wetting balance in air using a standard RMA

flux. The wetting force at two seconds for 95Sn/5Sb on a Cu

wire is significantly less than Sn/37Pb and Sn/3.5Ag. In

addition to marginal wetting performance, the toxicity of Sb

has also raised concerns. One source reported that with

more than 4 per cent antimony in the alloy the tensile

strength is reduced and the joints have a variable and occasionally

low fatigue strength. As with Bi, antimony is also a

by-product in the production of lead.

Sn/Ag/Sb: 65Sn/25Ag/10Sb (known as

Motorola J Alloy); 230-235°C

Motorola J alloy is a relatively high temperature alloy that

displays good creep resistance. Most notably, this alloy is

currently used as a replacement for Au-Si die attach material.

One company that uses Motorola J has expressed concern

that the alloy may be too strong and sometimes breaks

the die. This behaviour can be explained in terms of the

alloy microstructure and resulting properties. With this

relatively high amount of antimony, a large amount of the

hard SbSn phase is likely present. These SbSn cubes can act

as crack initiation sites and eventually lead to failures.

Therefore, the high strength which the high antimony

content imparts may prove to be too stiff for microelectronics

applications where compliance to shear stresses is a

requirement.

The large amount of Sb may also be responsible for the

very poor wetting behaviour of Motorola J alloy. Wetting

balance measurements taken in air using a standard RMA

flux show that the force at two seconds is considerably less

than Sn/37Pb, Sn/3.5Ag, and even 95Sn/5Sb, which is

known to be marginal/difficult in circuit assembly. The

rapid oxidation of antimony is likely a factor. Another

source reported that the poor wetting behaviour was due to a

Ag content of >4 per cent, due to a decreased fluidity. This

observation can be attributed to the formation of a substantial

amount of the needle-like Ag3Sn phase which is solid

until 480°C. A large amount of solid Ag3Sn particles could

inhibit solder wetting and spreading. It is likely that both the

high Sb and Ag levels that are present in this alloy contribute

to its poor wetting behaviour. The needle-like Ag3Sn

phase may also act as a crack nucleation site, affecting

fatigue behaviour.

From an electronics manufacturing standpoint, this alloy

is too strong, wets poorly, and is too expensive with 25 per

cent Ag.

In/Sn: 52In/48Sn; 118°C

The melting point of this alloy makes it suitable to low

temperature applications.

Indium alloys are more compatible with gold than tin,

and the dissolution therein of gold is considerably slower.

With regard to indium, it displays good oxidation resistance,

but is susceptible to corrosion in a humid environment. It is

also a very soft metal and has a tendency to cold weld. In

addition, the 52In/48Sn alloy displays rather poor high

temperature fatigue behaviour, due to its low melting point.

The high indium content limits the widespread use of this

alloy due to cost and availability constraints.

Sn/In/Ag/Sb: 86.4Sn/11In/2Ag/0.6Sb

Sn/In/Ag: 77.2Sn/20In/2.8Ag; 189°C

A low-melting, binary SnIn phase, with a melting point of

118°C, may occur with these alloys. They have most of the

disadvantages as mentioned previously for the binary In/Sn

alloy, e.g. poor thermal fatigue.

Sn/Zn: 91Sn/9Zn; 199°C

The presence of zinc leads to oxidation and corrosion. It

also reacts with acids and alkalis. Thus, its compatibility

with flux and its storage stability is critical. When wave

soldered, this alloy tends to produce a lot of dross.

Au/Sn: 80Sn/20Au; 280°C

The application of this alloy is restricted, due to the very

high price and the limited availability of gold.

3. Oxidation tendency of different

molten alloys in air and dissolution of

initial oxides in nitrogen atmosphere

A paper presented at Nepcon 1997 showed that each alloy

has a different tendency to form oxides, when exposed to air

at high temperatures (see Table II).

On the other hand, most of the oxides are soluble in a

liquid solder, when they are heated in a nitrogen atmosphere

above their melting points. The solubility of the oxides in

molten solder increases with increasing temperature (see

Table III).

4. Dissolution of copper

In the electronics industry, copper is commonly used as a

basis material for

• conductor tracks and solder pads on the PCB;

• lead frames of SO, QFP, PLCC, and other components.

Some alloys have a greater tendency to dissolve copper

than others. Alloys with a high tin content and with a

higher melting point are especially critical.

If a larger quantity of copper is dissolved in the solder

material, there is excessive formation of Cu6Sn5 intermetallic

phases. These phases are very brittle and

adversely effect the mechanical properties of the solder

joint.

The extent of copper dissolution of various alloys can

be discovered by means of a simple test: a hanging weight

of 50g is attached to a copper wire of diameter 125µm. A

small quantity of flux is applied to one point on the copper

wire. Then, the trial alloy is held on to this point with a

soldering iron, which has been pre-heated to 280°C. The

time it takes to break the copper wire (i.e. until the copper

is dissolved inside the solder) is measured (see Figure 1).

For surface metallizations like Ni-Au or Pd, this is a less

critical problem. However, it has been reported that, if the

[ 23 ]

Anton Zoran Miric and

Angela Grusd

Lead-free alloys

Soldering & Surface Mount

Technology

10/1 [1998] 19–25

alloys have a high tin content like 95Sn/5Sb or

96.5Sn/3.5Ag, a Ni protective coating of 2µm is not

enough to avoid copper dissolution and the consequent

growth of brittle intermetallic phases.

The mixing-in of copper with the solder alloy reduces

the tendency of the solder to dissolve copper out of the

metallization.

5. Low-melting binary and ternary

phases

If lead-free solder materials are being applied, the PCB

and SMD metallizations should also be lead-free. In

combination with lead, some lead-free alloys may create

low-melting binary and ternary phases/compounds (see

Table IV).

Such low-melting phases have a bad influence on the

reliability of the solder pad. This especially concerns

thermal fatigue at higher temperatures.

One might assume that SnPb is not dissolved at solder

temperatures below the melting point of 183°C. However,

in practice, it has been shown that SnPb can actually be

dissolved in SnBi at temperatures below 183°C. Gold, too,

goes into solution in tin, when exposed to the usual reflow

temperatures of approximately 220°C, despite its melting

point of 1,063°C.

Low-melting phases may also occur with ternary or

quaternary systems. This is not always necessarily caused

by the presence of lead. Such low-melting phases also

appear in combination with indium and bismuth (see

Table V).

6. Intermetallic compounds

The long-term integrity of a solder joint very much depends

on the intermetallic compounds that are formed during the

soldering process. When soldering to copper or nickel a

formation of intermetallic compounds is necessary to

achieve a good joint and their presence in the solder joint

significates good wetting. However if this layer becomes

too thick the base metal or finish may be consumed by the

solder and this leads to dewetting and poor joint reliability –

more over some of the intermetallics are very brittle which

decreases fatigue strength especially if this layer is thick.

Intermetallic compounds continue to grow very slowly also

at room temperature, but under this condition they will not

normally become thick enough to detrimentally influence

the reliability of the joint. However, at elevated temperatures,

especially when the time above the melting point

during reflow is very long, the growth of an intermetallic

layer will be very excessive and will very negatively influence

the reliability of the solder joint.

The melting points of intermetallic compounds (Cu3Sn,

Cu6Sn5, Ag3Sn, Ni3Sn4, AuSn, AuSn2, Ag4Sn4 etc.) gives

an indication about the zone growth potential at room, or at

working temperature (see Table VI). Rule of thumb: the

lower the melting point, the larger the potential for growth.

Table II

Oxide thickness: initial and after oxidizing the solder preform in air at a temperature that was

140°C above the melting point of the alloy

Oxidation Oxide thickness (angstroms) Dominant

Alloy temperature (°C) Initial After 10 min. After 50 min. oxide type

Sn99.3/Cu0.7 367 20 50 50 Sn-oxide

Sn96.5/Ag3.5 361 30 50 50 Sn-oxide

Sn63/Pb37 323 30 50 500 Sn-oxide

Bi58/Sn42 278 350 800 Sn-oxide

Sn95/Sb5 380 20 875 1,425 Sn-oxide

Sn91/Zn9 339 70 200 325 Zn-oxide

52In/48Sn 257 20 175 600 In-oxide

Note:

Before oxidizing the preform in air, the initial oxides were removed by heating the preform in nitrogen to 500°C and

then holding it for ten minutes in a flow of hydrogen (hydrogen reduces oxides); afterwards, the preform was cooled

in nitrogen to a temperature that was 140°C above the solder’s melting point. Then, a nitrogen flow was switched to

air flow, to start oxidation. Finally, the preform was cooled to room temperature in nitrogen, and the oxide thickness

was measured using auger electron spectroscopy

Table III

Temperature at which initial oxides dissolve and at which the solder starts to spread

Temperature for dissolving Temperature in N2 at which solder

Melting point initial oxides in N2 preform starts to spread (ppm)b

Alloy (°C) 10ppma (°C) 10 100 1,000 10,000

Sn99.3/Cu0.7 227 »245 230 234 245 No spread

Sn96.5/Ag3.5 221 »240 230 238 240 No spread

Sn63/Pb37 183 »260 205 207 270 No spread

Sn95/Sb5 238 »250 246 255 258 No spread

Sn91/Zn9 199 > 500 No spread

52In/48Sn 118 »210 200 No spread

Notes:

aA fluxless solder preform was heated on a glass substrate in nitrogen (10ppm oxygen residue) until the temperature

was reached at which the molten solder changed from its initial shape (flat) to a nearly spherical shape (an

oxide-free surface forms a nearly spherical shape on glass, driven by liquid surface tension).

bA fluxless preform was heated on copper (coated with Ni/Au) in nitrogen (10 to 10,000ppm oxygen residue), until

the temperature was reached at which the molten solder started to spread

Figure 1

Copper disolution of various alloys

10

8

6

4

2

0

Sn/Pb

Sn/Ag/Cu/Sb

Sn/Ag/Bi/Cu

Sn/Ag/Bi

Sn/Ag/Cu

Sn/Ag

Dissolution Time in Minutes

8.76

3.42 3

1.8 1.56 1.25

Table IV

Influence of Pb on some lead-free binary systems: the

creation of low melting phases

Lowest melting point Lowest melting point

System in the binary system (°C) in combination with lead (°C)

Sn-Bi 138 97

Sn-Zn 199 183

Sn-Ag 221 179

Sn-Sb 232 183

Table V

Creation of low melting phases in ternary and quaternary

systems

Usual melting point Lowest melting point

System of the system (°C) Phase/(°C) Phase /(°C)

90Sn/7.5Bi/2Ag/0.5Cu 212 Sn-Bi/138 Bi-Pb/97

77.2Sn/20In/2.8Ag 189 Sn-In/118 Sn-Pb-Ag/179

86.4Sn/11In/2Ag/0.6Sb 221 Sn-In/118 Sn-Pb-Ag/179

65Sn/25Ag/10Sb 233 Sn-Pb-Ag/179

[ 24 ]

Anton Zoran Miric and

Angela Grusd

Lead-free alloys

Soldering & Surface Mount

Technology

10/1 [1998] 19–25

The ductility and also the firmness of the solder pad are

negatively influenced by excessive growth of the intermetallic

zones.

The very low melting temperature of SnAu intermetallic

compounds is consistent with fast gold dissolution into Sncontaining

alloys.

7. Alternative surface metallizations

(PCB and component)

Today, the eutectic SnPb alloy is most commonly applied as

an end metallization of PCBs and components (HAL = hot

air levelling). It does not make much sense, however, to

replace lead as a solder material, and to continue using lead

as a coating. Furthermore, a combination of lead-free and

lead-containing alloys may result in a deterioration in the

mechanical properties of a solder connection (see also

section 5).The pre-tinned pads also have spherical and

irregularly thick solder tips (typical film thickness is 25µm

in the middle of the pads, and 1µm at their edges). They are

not suitable for ultra-fine pitch components. There is no firm

sealing between the stencil and the PCB/pads, and the

solder paste gets into spaces between the pads. During the

placing process, component movement may occur.

Below are listed alternatives which are already well

established in electronics manufacturing.

Ni-Au

Typical film thickness: 3-5µm Ni and 0.15-0.25µm Au. Au

avoids oxidation of Ni, and is quickly dissolved at solder

temperature (dissolution speed 4µm/s Au in SnPb solder at

250°C):

• Even and coplanar surface, well suited to ultra-fine pitch

applications.

• Expensive.

• Au is a good protection against oxidation, but only if it is

applied properly: not too thin and not too porous (otherwise,

Ni migrates to the surface and oxidizes), but not too

thick either (otherwise, brittle AuSn4 intermetallic phases

are formed if the gold content is >3wt%); not too high a

phosphor content of nickel (<9 per cent) etc.

• Various chemicals used during the deposition of Ni and

Au may be dangerous to the environment (cyanide

bath/potassium gold cyanide). Nickel itself is also seen as

a material risky to the environment.

Cu with organic passivation (OSP = organic

solderability preservative)

Typical film thickness of the organic passivation is 0.1-

0.5µm:

• Even and coplanar surface, well suited to ultra fine pitch

applications.

• Favourably priced.

• Good protection against oxidation at normal soldering

temperatures, and if soldered under nitrogen. There is an

increased danger of oxidation. If soldered in a normal

atmosphere, especially with double-sided reflow or at

higher soldering temperatures – there is a higher risk of

oxidation and solderability issues during the application

of mildly activated fluxes.

Possible alternatives are:

• Thin silver coatings (0.07-0.1µm), which should avoid

problems that usually occur in combination with silver,

e.g. electromigration and dendrite growth.

• Thin palladium coatings (palladium 94-97 per cent,

phosphor 3-6 per cent – film thickness approximately

0.2µm). An electroless deposition on copper or on nickel

is possible. The dissolution speed of Pd in tin is much

lower than that of Au. Pd is cheaper than gold, but,

nevertheless, relatively expensive. Palladium is a catalytic

material that tends to react with organic molecules

in the atmosphere – a non-solderable, organic layer is

formed after a longer storage time.

Not only the solder material itself, but also the metallization,

has a great influence on the thermal fatigue of a solder

pad. Results with the Sn96.5/Ag3.5 solder material are

much better in combination with a Pd metallization than

with a Cu metallization (dissolution of copper and extensive

creation of brittle Cu6Sn5 phases – see also section 4). At

higher temperatures, the same alloy displays bad cycle

stability, when combined with the BiSn coating (lowmelting

Bi-Sn phases, 138°C).

Also wettability test results with the SnAgCu and SnAg

alloys were better on Ni-Au than on protected Cu pads.

8. Alternative PCB materials

Standard PCB materials (glass/epoxy FR4) can be heated up

to between 260 and 280°C. According to MIL-P-13949 G,

the maximum permissible heat is specified for 10s +1/ –0s,

in a solder bath at 287 ± 6°C. DIN IEC 249 specifies 10s for

FR2 and FR3, and 20s for FR4 at 260 +5/–0°C. The glass

transition temperature for standard FR4 is 130-145°C. At

temperatures above 280-300°C, the process of decomposition

of the polymerized resin begins.

The temperature resistance of the FR4 material is sufficient

for most lead-free alloys, but with some alloys that

have higher melting points, the reflow temperature of 260-

280°C may be exceeded. In this case, alternative PCB

materials must be used, e.g.:

• FR5, glass transition temperature of 180°C.

• Glass/BT epoxy, glass transition temperature of 180°C.

• Glass/polyimide, glass transition temperature of 250°C.

The use of PCB materials with a higher thermal resistance

results in much higher costs. The estimated relative costs

are: FR4 = 100; FR5 = 150; glass/BT epoxy = 250; glass /

polyimide = 350-550.

9. Concluding remarks

None of the alternative alloys discussed can be offered as a

suitable drop-in replacement for the SnPb eutectic or neareutectic

alloys. The introduction of alternative alloys

requires compromises (solder temperature, solder flux,

wettability, costs, different component and PCB metallizations,

etc.). One of the most important rules of thumb is: a

lead-free alloy should not be contaminated with any lead.

Despite the higher melting points (e.g. 95.5Sn/4Ag/0.5Cu

– 217°C) of some alloys, the solder temperature does not

necessarily have to be significantly higher than the general

reflow temperature in IR ovens. Recently-made reflow

ovens, having a sensible combination of IR rays and forced

convection, allow for a minimization of the temperature

differences between the smaller chip components and the

larger PLCCs and QFPs.

Due to their limited availability, Bi and In cannot completely

replace SnPb. In any case, Sn will remain as one of

the basis metals. Cd is out of question, because it is toxic.

Sb and Cu are somewhat toxic, but less toxic than Pb and

Cd. They should only comprise a minor part of the alloy.

The higher costs of a new alloy do not necessarily restrict

widespread application. Usually, the placement costs for a

typical populated board comprise in total <10 per cent

Table VI

The lowest melting point of various intermetallic compounds

Element Sn (°C) Bi (°C) Sb (°C) In (°C)

Cu 415 –a 568 687

Ag 480 –a 558 305

Ni 795 469 620 449

Au 252 373 460 544

Zn –a –a 455 –a

Note:

aIntermetallic compounds do not occur

[ 25 ]

Anton Zoran Miric and

Angela Grusd

Lead-free alloys

Soldering & Surface Mount

Technology

10/1 [1998] 19–25

(approximately 15 per cent of this percentage is material

costs) of the total costs of that board – the components

themselves represent the major cost.

Various institutes, like ITRA (International Tin Research

Association) and NCMS (National Centre for Manufacturing

Sciences) extensively test different alloys. The results of

these tests should provide the industry with extensive

information on candidate lead-free solder alloys.

In order to apply a new lead-free alloy in the assembly

operation, it must be available as solder paste, solder bar, or

solder wire. It must be possible to do the rework with the

solder wire.

Some renowned manufacturers like Boeing announced

the introduction, in the near future, of lead-free alloys into

their production.

Mechanical properties, the total costs of the finished

board, environmental aspects, and the recycling possibilities

with alternative alloys, all very much influence the introduction

of new, lead-free alloys. Environmental legislation may

also play a significant role in this regard.

Many manufacturers will compromise the shininess of

the joint and/or can accept less wettability, and even a

slightly higher melting point. The electronics industry will

not, however, convert to lead-free solders unless the reliability

characteristics are at least as good as Sn-Pb eutectic.

Recent work with candidate lead-free alloys (e.g. Sn/Ag or

Sn/Ag/Cu) indicate that they do exhibit improved reliability

over SnPb.

Further reading

DeSantis, C., Felty, J. and Tsung-Yu Pan (1997), “Toxicology &

economics/availability of potential lead-free solder alloying

elements”, TMS, Vol. 2, Orlando, FL.

Dong, C.C., Schwarz, A. and Roth, D.V. (1997), “Effects of atmosphere

composition on soldering performance of lead-free

alternatives”, Nepcon 1997.

Draft Proposal for a European Parliament and Council Directive on

End of Life Vehicles (1996), 31 July.

Engelmaier, W. (1994), “Versuchsbeschleunigung für Surface-

Mount Lötverbindungen”, VTE, No. 1.

Hampshire, W.B. (1993), “The search for lead-free solders”, Soldering

& Surface Mount Technology, No. 14.

Hampshire, W.B. (1996), “Some problems in switching to lead-free

solders”, Nepcon 1996.

Hannemann, M., Nowottnick, M., Bell, H. and Woock, A. (1995),

“Vergleichende Untersuchungen von Anschlußflächen für die

Flachbaugruppenfertigung”, VTE, No. 1.

James Maguire Boeing Defense & Space Group (1996), “No lead

soldering overview”, Green Manufacturing Symposium, USA,

Vol. 6.

Kang, S.K., Rai, R.S. and Purushothaman, S. (1996), “Interfacial

reactions during soldering with lead-tin eutectic and lead-free,

tin-rich solders”, Journal of Electronic Materials, No. 7.

Klein-Wassink, R.J. (1989), Soldering in Electronics, Electrochemical

Publications.

Klein-Wassink, R.J. and Verguld, M.M.F. (1995), Manufacturing

Techniques for Surface Mounted Assemblies, Electrochemical

Publications.

McCormack and Jin, S. (1994), “New, lead-free solders”, Journal of

Electronic Materials, Vol. 23 No. 7.

Melton, C. (1993), Effect of Solder Reflow Process Variables on the

Solder Wettability of Lead Free Solder Alloys, Brasage.

Melton, C. and Fuerhaupter, H. (1996), “Lead-free tin surface finish

for PCB assembly”, Nepcon 1996.

Neue Lote für den automatisierten Lötprozeß (1996), VTE, Nos 1

and 2.

Soldering to Gold Coatings, Publication No. 736, International Tin

Research Association.

Vianco, P.T. and Frear, D. (1993), “Issues in the replacement of leadbearing

solders”, The Journal of the Minerals, Metals &

Materials Society, July.

Weiss, D.G. (1991), “Hohen Ansprüchen genügen”, Productronic.