Status of Lead-Free 2015: A Perspective · 2015-04-17 · 4/14/2015 3 © Indium Corporation Intro:...

4/14/2015

1

© Indium Corporation

Status of Lead-Free 2015: A Perspective

• Ron Lasky, PhD, PE• Senior Technologist Indium Corp• Instructional Professor Dartmouth

• [email protected]


Dr. Ron Lasky: World’s Strongest Senior Archer?

• Native of Binghamton• Graduate of BCC, Cornell,

BU, Cornell– PhD in materials science

• NYS Professional Engineer• More than 30 years in

electronic and optoelectronic packaging at IBM, Universal Instruments, Cookson Electronics

• Author of 6 books• Currently a Senior

Technologist for Indium and an Instructional Professor at Dartmouth College

2

4/14/2015

2


Thayer School of Engineering at Dartmouth

Sylvanus Thayer Dartmouth 1807, USMA ’08Father of West PointFounder of Engineering at DartmouthFather of Technology in US

Dartmouth: The Smallest and Best Ivy

Slide #3


Agenda

• Introduction• History• Current Alloys

– Use by Industry– Reliability– Process Issues– New Failure Modes

• Future Alloys• Tin Whiskers• Conclusions• Appendices

4/14/2015

3


Intro: RoHS: True of False

1. You are safer because of RoHS

Slide #5


Slide #6

False

• But they are!

4/14/2015

4


Slide #7


RoHS was established to…

• The purpose of this Directive is to approximate the laws of the Member States on the restrictions of the use of hazardous substances in electrical and electronic equipment and to contribute to the protection of human health and the environmentally sound recovery and disposal of waste electrical and electronic equipment

Slide #8

4/14/2015

5


Slide #9 Slide #9


True or False #2

• Lead‐Free has no process or performance benefits.

Slide #10

4/14/2015

6


False: LF: The Good News:Coalescent Performance

Comparison

Slide #11

Pb Before Reflow

Pb Paste Before Reflow

Pb paste printed onto Cu coupon

Pb free Paste Before Reflow

Pb free paste printed onto Cu coupon

Pb Paste After Reflow

Pb paste fused onto Cu coupon Pb free paste fused onto Cu coupon

Pb free Paste After Reflow


The Green Generation

Slide #12Global Foundation, Business Week 02/04/08

4/14/2015

7


Lead-Free Reliability Thought

• “Everyone knows there is no data on LF reliability right?”

• Wrong

– >$7,000 B of products made, some since 2001

• No major reliability issues

– >$100sM of R&D

• Long term (>8 years) data still sketchy

13


The Electronics Industry

About 25% Pb‐Free Exempt in 2014

4/14/2015

8


The Electronics Industry

Courtesy Prismark


Perspectives

• An imaginary question for Apple CEO Tim Cook– How difficult was the transition to Pb‐Free for Apple?– Pauses and looks a little confused: “Now I remember, it wouldn’t make the top 50 challenges Apple has faced since 2003.”

• Real question to Thayer School IT staff in 2011– You purchase $millions in electronics each year, how has the reliability been since the implementation of Pb‐free in 2006?

– Pauses and looks confused, “What is Pb‐Free?

• But implementing Pb‐Free and RoHS likely cost many $10B

4/14/2015

9


Lead-Free Early History

• As early as the 1990s NCMS looks at Lead‐free

• Tin‐bismuth solders rejected

– Low temp ternary alloy formation with lead contamination

– Fillet lifting in wave soldering

– Brittle

– Poor performance above 100oC

• The industry settles on SAC


Slide #18

But is Near Eutectic Best? B. Boettinger, K. W. Moon of NIST performed studies to determine the true

Sn-Ag-Cu eutectic.

Alloys in shaded area have freezing range <10°C.

0 0.5 1 1.5 2 2.5

wt% Cu

0

1

2

3

4

5

6

7

8

wt%

Ag

250 C

230 C

270 C

220 C

230 C 270

C

290

C

250

C

230

C

310

C

Estimation of Ternary Liquidus Surface, 10/23/99 Based on Marquette saturation data, with NWU and NIST thermal analysis.

Ag3Sn

Cu6Sn5

Sn

NIST experimental work showed that the eutectic composition is approximately Sn3.5Ag 0.9Cu. (+/- 0.1%) (In agreement with Loomis and Fine)

4/14/2015

10


SAC Tombstoning

Slide #19

• Still not as low as SnPb due to greater temperature gradient


Slide #20

• Reduced chip size causes greater vulnerability toward chip disturbance at reflow soldering (Tombstoning, billboarding, wicking)

• Increase surface tension may reduce wetting speed, but cause more defects due to higher horizontal pulling force

• Need alloy with a slower wetting speed at melting temp via other approaches, such as a pasty alloy with high mass fraction of solid at onset of melting.

y = 0.082e-5.0766x

R2 = 0.8882

0%

2%

4%

6%

8%

0% 20% 40% 60% 80%

Mass Fraction of Solid (%)

To

mb

sto

nin

g R

ate

(%). 3.5-1

3.8-0.7Sn63

3-0.5

2.5-0.82-0.5

(Indium)

Mass Fraction of Solid Estimation

Symmetry assumed for eutectic

0%

1%

2%

3%

4%

5%

6%

7%

Sn2Ag0

.5Cu

Sn2.5

Ag0.8

Cu

Sn3Ag0

.5Cu

Sn3.5

Ag1Cu

Sn3.8

Ag0.7

Cu

63Sn3

7Pb

To

mb

sto

nin

g R

ate

(%)

.

(Indium)

0.56N/m

0.51N/m

Wetting Speed/Tombstoning Challenge

4/14/2015

11


Thermal Cycling and Low Ag

Slide #21


Slide #22

Less Ag: Smaller Ag3Sn Platelets

93.6Sn4.7Ag1.7Cu 95.5Sn3.8Ag0.7Cu 5000X 5000X

1500X2000X

Source: T.Y. Lee, W.J. Choi, K.N. Tu, J.W. Jang, S.M. Kuo, J.K. Lin, D.R. Frear, K.Zeng, and J.K. Kivilahti, “Morpology, kinetics and thermodynamics of solid state aging of eutectic SnPb and Pb-Free solders (Sn3.5Ag, Sn3.8Ag0.7Cu, and Sn0.7Cu) on Cu”, J. Mater. Res. 17(2) (2002).

4/14/2015

12


Slide #23

Failures AlongAg3Sn Intermetallic


Lead-Free History Since 1999

• Circa 1999: Near Eutectic SAC387 is initial LF Alloy of Choice– Eutectic is alloy of choice as it is replacing eutectic SnPb

– 100s millions of mobile phones assembled with this alloy to date

• SAC305 becomes IPC’s SVPC Alloy of Choice Circa 2006– Exhibits less tombstoning than SAC387

– Uses less silver

• SAC105 has acceptance in 2007‐2008 to now– Less silver

– Performs better in drop shock than SAC305, but worse in Thermal Cycle (TC)

– Disadvantage: Tm is 225C vs SAC305 219C

– Only saves a penny per gram of solder paste cost with $17/oz Ag

• Sn/Cu and Sn/Cu+Ni are lower cost options for wave– Tm is 227CSlide #24

4/14/2015

13


SAC305 IPC SPVC

• SMT Only

• Copper Dissolution (wave)

• Solder Pot Attack (wave)

• Poor Hole Fill (wave)

• High Ag Cost (Solder Bar)

• Drop Shock Testing

Slide #25

Images courtesy of ASKBobWillis.com


Slide #26

CuSn Needles in Solder Joints

4/14/2015

14


Cu Pad Dissolution From Pb-Free

Dennis Barbini, “Cost effective wave soldering”, Vitronics, Oct. 2003

Ref: Angela Grusd, “Connecting to Lead-Free Solders”, Circuit Assembly, p.32-38, August, 1999.


Alloy Selection for Wave

• Alloys currently under consideration:

– SAC305

– 99.3Sn/0.7Cu

– Sn/Cu + 0.1 Ni

• 99.3Sn/0.7Cu is preferred for price reasons over SAC

• Sn $18.10/kg• Zn $2.13/kg• Cu $5.94/kg• Bi $16.50/kg• In $650.00/kg• Ag $550/kg• Sb $8.15/kg• Ni $12.70/kg• Pb $1.87/kg• So…………

• 99.3Sn/0.7Cu = $18.01/Kg• SAC305 = $34.00/Kg

SAC105 =$23.27/kg

4/14/2015

15


Reliability

Equal or better?Imagine You as an Engineer in 1910

with the task of finding a perfect solder in all respects


Older Work:NEMI Test Plan

Ref: Bradley; Summary of Pb‐Free Solder Reliability;Motorola QuickStart Seminar‐; Ft. Lauderdale, FL; February 2005

Slide #30

4/14/2015

16


Summary of NEMI Results

Slide #31



169 CSP Results


Slide #32

4/14/2015

17


Henshall:Low Silver Investigation

• Compared SAC305, low silver and Pb BGA solder balls

• SAC305 paste

• OSP, NSMD pads

• Tg=170oC PWB

• 1.27 to 0.5 mm pitch

• 0.32‐0.62 mm ball dia

• Pack Size: 8‐45 mm

• 0‐100 and ‐40‐125 CTC

Slide #33

SMTAI 2010


Henshall’s Weibull Plot

1000.000 100001.000

5.000

10.000

50.000

90.000

99.000

Cycles to Failure

Percent Failed

0 to 100 C1.0 mm pitchAll Data

Sn‐PbSAC305

SAC205

SAC105

SACX

•Most data sets “clean”• Common anomaly was“late” failures

• Investigating cause

34

‐40‐125C results were mixed!

Dopants they studied do not affect TC

4/14/2015

18


Coyle Etal

• 0‐100C

• Dwell from 10, 30, 60 min

Slide #35

© Indium Corporation36

Coyle’s ATC Weibull Plot for10 Minute Dwell Time

4/14/2015

19

© Indium CorporationSlide #37

“Test vehicles assembled with lead‐free materials (notably tin‐silver‐copper) exhibited lower reliability under some test conditions.”

http://www.teerm.nasa.gov/LeadFreeSolderTestingForHighReliability_Proj1.html

© Indium CorporationSlide #38

4/14/2015

20


Some NASA Data

Slide #39


Slide #40

Take out SnPb/LF mix‐LF Better or MB: 28‐LF Much Better: 23‐SnPb Better or MB: 17‐SnPb Much Better: 5‐SnPb Much Much: 9‐Draw: 1

4/14/2015

21


Slide #41

What about Sn/Pb Paste; SAC Ball?


225 C Peak Profile

0

50

100

150

200

250

0 1 2 3 4 5

217 C

183 C

4/14/2015

22


43

SnPb and SAC Reflow Well

Oxidation Resistance Testing– 90s Time above liquidus (TAL)

– 25oC above liquidus

– Sn/Pb and SAC alloy at 208C (11C below SAC Tm)

– SAC sphere

Poor Oxidation Barrier Good Oxidation Barrier


Mixed Alloy System

• A quaternary alloy

• Uncontrolled

• Hard to model

• Poorest choice for long term reliability–Especially mission critical

• Use only if supported by careful process and reliability evaluation

4/14/2015

23


Process Issues

• Higher Processing Temperature

– Component Concern

– Pad Cratering

• “New” Failure Modes: Discussed in Appendix II

– Graping

– Head‐in‐Pillow

– Increased Voiding

• Halogen Free: Discussed in Appendix II


Process Issues Continued

• View of an Enlightened Solder Paste Supplier: You should be able to assemble Lead‐Free:

– Reflow in air

– Halogen free or not

– With minimal graping, HIP, voiding

– Lead and pad finishes of your choice

4/14/2015

24


Slide #47

Poor Coalescent Good CoalescentPoor coalescent is attributed to powder oxidation during reflow process in air atmosphere Courtesy Vahid Goudarzi Motorola

Solder Paste Makes a Difference.

But SAC solder does not flow as well as SnPb solder


Questions1. How has SAC fared?

– $Trillions in products

– No major reliability concerns

2. Industry consensus on military/aero, medical?– Mil/Aero should stick with PbSn for at least five more years of reliability

data

• Refinish and reball

– Medical (ex body) should go with lead‐free

3. Will SnPb be with us for the foreseeable future?– Absolutely

4. What is the impact of PB‐free on component providers?– Matte Sn with nickel flash for tin whisker mitigation

– Component temperature concern?

5. Are exemptees converting?– We are at a point where exemptees are only mil/aero, so probably not

4/14/2015

25


Questions (con’t)

7. Is there comfort with a hybrid SnPb Lead‐free process

– If there is, there should not be

– A quaternary alloy of unknown composition exists. Yikes!

8. Is the industry waiting for a 2nd generation alloy – They exist, are good, but few are buying

– SAC305 works and a requal is too expensive

9. What is the impact of drop off in thermal cycling after isothermal aging?

– Are aging temperatures realistic?

– 125C is 398/505= 0.79

– Like aging steel at 1200C, which is orange hot

– $7 Trillion since 2001 with no major reliability issues


Auburn Isothermal Again Study

Isothermal Aging Effects onthe Harsh Environment Performance of Lead‐FreeSolder Joints, Zhang etal, SMTAI2012, Orlando, FL

4/14/2015

26


A Heretical Thought

• Are our TC Tests too Rigorous? <100C, CTE =2.6 ppm/C

<60C, CTE =50

<50C, CTE =60


Future Alloy: Dopants: Foundation of Future Improvements?

• Ge

• Co

• Ni

• P

• Ti

• Mn

• Bi

• Etc………..

Slide #52

4/14/2015

27


Slide #53

SnCu (Ni, Ge, Co, etc)

• Nihon Superior’s Sn100C (Sn 0.7Cu .05Ni .006Ge) 227C

• Ag Free allowing a big cost reduction

• Ni and Co – reduces solder pot and copper erosion, provides nucleation sites

• Ge – Dross inhibitor


The Effect of Dopants: Ni

Slide #54

4/14/2015

28


Ge

Slide #55


New Alloys:Work Performed by Lee etal

on SACm

• SACm– 98.5Sn<1Ag0.5Cu0.05Mn

• Two Tests:– JEDEC Drop Test (JDT)

• 0‐250 hours for aging at 150C or 250 TCT

– Thermal Cycle (‐40 to 125C)• 0‐250 hours aging at 150C

– Lee etal at http://www.indium.com/technical‐documents/whitepaper/achieving‐high‐reliability‐lowcost‐leadfree‐sac‐solder‐joints‐via‐mn‐doping

Slide #56

4/14/2015

29


Slide #57

Older Work:Doped SAC105 Alloys


Drop Test

Slide #58

4/14/2015

30


JEDEC Drop Test: SnPB, SAC105, SAC305, SACm

1000100101

99

90807060504030

20

10

5

32

1

Data

Per

cent 3.57261 139.748 0.964 10 5

1.33995 120.958 0.952 10 51.05990 58.854 0.925 10 52.16891 179.272 0.974 11 4

Shape Scale Corr F CTable of Statistics

SnPbSAC105SAC305SACM

Variable

Weibull plot for JDT performance for TFBGA (NiAu) devices with various sphere alloys assembled on OSP treated PCBs. The assemblies were pretreated with 250 cycles of TCT prior to JDT.


SnPb, SAC105 SAC305 and SACm:TCT Data

Slide #60

4/14/2015

31


Aging and IMC Thickness

Slide #61


Is LF a Reliability “Time Bomb?”

Slide #62

TRT= 430oK = 0.85

TMT Sn= 505oK

4/14/2015

32


Conclusions

• For 0‐100C type TC SAC305>SAC105>SnPb

• DS: SAC105 ≈ SnPb > SAC305

• For ‐40‐125C TC the results are mixed

– LF reliability for this condition is still “unproven”

• Adding dopants can improve results in DS and TC

– SACM has better drop shock performance than SAC105 and much better drop shock performance than SAC305

– SACM matches SAC305 in thermal cycle performance (‐40‐125C) and are better than SnPb Re characteristic life

• Unaged SnPb is better than all SACs Re first failure

– Hence, SACM is a better choice than SAC105 or SAC305

– The mechanism for the improved performance is attributed to a stabilized microstructure with a uniform distribution of IMC particles

• Aging has a dramatic effect on both drop shock and thermal cycle life

Slide #63


The Future of Pb-Free Alloys2010 Tim Jensen Prediction

What do they really want?

1. SAC105: Lower alloy cost

2. SAC105+: Lower alloy cost, lower failure rates

3. Bi/Sn: Lower alloy cost, lower PCB/component costs, lower processing costs (Stage 2.5)

4. Real Target: 180‐200°C, low material costs

4/14/2015

33


True or False #3

• There are now more tin whisker fails than hard drive fails.

65


False

• No significant number of TW fails reported.

– NASA has tracked hundreds

– http://nepp.nasa.gov/Whisker/

• > 50M hard drive fails/year

– Google Study: 8% of HDs fail each year

66

4/14/2015

34


What Are Tin Whiskers?• Very thin (1‐10 short (< 1

mm) tin filaments

• They are created by compressive stresses in tin. Typically found in bright platings

• Much, but not all, is understood about their formation and mitigation– http://nepp.nasa.gov/Whisker/

• Other metals produce whiskers: Zinc, silver, etc

67


68

4/14/2015

35


69


BUT ARE TW FAILS COMMON TODAY?

70

4/14/2015

36


Not Much Newer than 2006

71

http://nepp.nasa.gov/whisker/failures/index.htm


What’s the Concern?

• They can cause shorts

– But often fuse, creating intermittent problems

• Their incubation time is unknown

– Can be minutes to decades

• Mitigation is possible, but not elimination

• “Absence of evidence is not Evidence of Absence” Henning Leidecker, NASA

– But is perhaps evidence of a low RPN

72

4/14/2015

37


Component Finish Options

• Component or Connector Lead Finish is Biggest TW Risk

• Sn is easiest and most obvious choice– Sn Whiskers still a major concern

• TI uses Ni/Pd/Au

• Others recommend: Sn/Bi and Sn/Cu

• Suppliers recommend Sn/Ag/Cu for BGA

73


• At the interface between two materials, a joint will form when each of the materials diffuse into one another to create a strong bond. Intermetallic stress occurs when one material, in this case copper, diffuses more quickly into the other, namely tin, creating a bulk diffusion in one direction.

Sn Whiskers: Intermetallic Formation

74

4/14/2015

38



Cu

Sn

CuSnx

As the new intermetallic grows mostly into the space previously occupied by the tin, stresses are relieved by the formation of tin whiskers.

75


76

• Such diffusion can be dramatically reduced with a layer of Ni or Ag (etc)

• Ni Layer can also help prevent surface discoloration


Cu

Sn

SnOx

Ni, Ag, etc

SnNix

4/14/2015

39


Whisker Mitigation

Plating layer (Sn99.3Cu0.7)

Nickel

Substrate (Cu)

With Ni LayerNo Whisker Growth

Without Ni LayerWhiskers Observed in a Few Days


Lead Does Have Dramatic Positive Benefit, but so would 2% Bismuth

78

4/14/2015

40


79

Whisker Mitigation Techniques

Due to Compressive Stress

•Augmented by:

•Cu/Sn Intermetallic Formation at Grain Boundary

•Small Grain Size

•Thin Solder Layer

•Clean Solder Surface

•Alleviated by:

•Ni Plating on Cu

•Large Grain Size

•Thick Solder Layer

•Alloying with Bi


What about Coatings

• Work at CALCE suggests parylene coatings are promising

• For failure, two surfaces must be penetrated

80

4/14/2015

41


The Whisker IndexXu, Zhang, Fan, Abys etal

• Whisker Index=S*N*d*L*f(L)

• S= area, N=number of whiskers, d=whisker diameter, L=whisker length, f(L) relates to whisker length

• WI is reduced from 100,000 to 10 by changing from bright to satin bright tin plating alone

• Ni flash reduces WI to near 0

81


82


• The Whisker Index is a number that tells you the amount of whiskers in a unit area

• Whiskers are almost entirely removed by adding a layer of Ni

From Dr. Chonglun Fan, Cookson Electronics

4/14/2015

42


FMEA: Failure Modes and Effects Analysis

• Probability of TW– With Mitigation: Small, but not zero

• Severity if TW exists– Low for consumer product

– Potentially High for Mission Critical device

• Detection– High Risk: hard to detect In most cases TW RPN

• RPN=P*S and D is not high– Low for consumer product

– May be unacceptable for Mission Critical Product

• Majority of TW fails had no mitigation

83


What Should Your Strategy Be?

• Non Mission Critical Products– Avoid tin where possible ‐Work with Suppliers

– BOM Scrub ‐Parylene Coating

– Matte Sn ‐ Ni Flash

– 2% Bi ‐ Anneal to reduce stresses

• Mission Critical Products– Above plus

– CALCE TW Risk Calculation• http://www.calce.umd.edu/whats_new/2005/whiskerrisksoftware.htm

– FMEA

– Hire expert?

84

4/14/2015

43


For a 40 Year Mission Critical Product, RPN may be high. Consider PERM’s 3 Bears

85


Conclusions

• With mitigation, TWs can be reduced to an acceptable risk for consumer products

• Mission Critical Product will require special attention

86

4/14/2015

44


Acknowledgments

• This overview relies on the work of:– Greg Hensall Etal

• Low Silver BGA Sphere Metallurgy Project,

SMTAI 2010, Orlando, FL

– Richard Coyle, etal

• The Effect of Silver Content on the Solder Joint Reliability of a Pb‐free PBGA Package, SMTAI 2010, Orlando, FL

– Ning‐Cheng Lee etal

• ACHIEVING HIGH RELIABILITY LOW COST LEAD‐FREE SAC SOLDERJOINTS VIA MN OR CE DOPING, ECTC 2009

• And Older Work

– NASA’s Work

http://www.teerm.nasa.gov/LeadFreeSolderTestingForHighReliability_Proj1.html

Slide #87


APPENDIX I: TIN PEST

88

4/14/2015

45


Slide #89

Tin Pest in SnCu0.5

Transformation of Beta‐Tin into Alpha‐Tin in Sn‐0.5Cu at T <10oC from Y. Karlya, C. Gagg, and W.J. Plumbridge, “Tin pest in lead‐free solders”, Soldering and Surface Mount Technology, 13/1 [2000] 39‐40.


Slide #90

What is Tin Pest?

• Normal or “white” (beta) tin – 7.31 g/cm3

– Stable tetragonal structure above 13.2oC

• “Grey” or (alpha) tin– 5.77 g/cm3

– Stable cubic structure below 13.2oC

• The transition is slow (18 months or so), but cumulative– Expansion would destroy solder joints

• Maximum rate is at about ‐30oC• Why? Grey tin has lower entropy than white

– Lower temps likely drive the change

4/14/2015

46


Slide #91

Tin Pest is Not New News

• Napoleon’s Buttons

• Observed in church organ pipes

– Attributed to Satan => Tin Plague

• German

– Plague = Pest


Slide #92

The Mechanism

• Occurs where expansion is easiest– Surfaces, corners

– Tensile stress acerbates

• Retarded by– Compressive stress

– Alloying with soluble metals: e.g. Bi, Sb, Pb

• Metals that form intermetallics do not suppress: e.g. Ag, Cu

4/14/2015

47


Slide #93

Tin Pest Retardant

Alloying Metal

Tin Pest Retardant

% Concentration for effective Inhibition

Bi Strong 0.3 Sb Strong 0.5 Pb Strong 5.0 Cu None-

Weak ? >> 5.0

Ag Weak-Mod

? > 5.0


Slide #94

Concern for Pb‐Free

• SAC will likely be susceptible to tin pest

• Some preliminary work agrees

• Contaminants may help– Organic

– Inorganic

4/14/2015

48


Slide #95

The Need

• More work to understand Tin Pest in field conditions

• Support the work of Plumbridge etal and Havia etal

• Addition of Bi in SAC

– Patent issues?


Slide #96

Tin Pest References

[1] Homer, C. E., Watkins, H. C. Transformation of Tin at Low Temperatures, The Metal Industry, 1942, LX, 22, London.

[2] Y. Karlya, C. Gagg, and W.J. Plumbridge, “Tin pest in lead‐free solders”, Soldering and Surface Mount Technology, 13/1 [2000] 39‐40.

[3] Information from Tin Technology literature archive courtesy of Kay Nimmo ([email protected])

[4] [email protected]

[5] [email protected]

Appendix I: Tin Pest

4/14/2015

49


Appendix IILead‐Free Failure Modes:

Graping, head‐in‐Pillow, Voiding (with Halogen‐Free)


Slide #98

Graping

4/14/2015

50


Slide #99

Graping• Miniaturization of electronics has

given us 01005 passives and 0.4 to 0.3 mm pitch packages

• The resulting smaller solder paste deposits = surface powder oxidation

• Higher process temperatures (i.e. 227C vs 217C) may exacerbate graping (and tombstoning)

• Dopants such as P and Ge may help (control oxidation)


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20 25 30 35 40 45 50 55

SA/V

ol (

1/m

ils)

Side of Square Aperture (mils)

Surface Area/Volume (4 mil stencil)

100

4/14/2015

51


Small Deposit Coalescence Challenge


Graping vs. Aperture Size

4/14/2015

52


Smaller Apertures Require Finer Paste Particle Sizes

103

Powder

Size

Diameter Range

TYPE microns

3 25 45

4 20 38

5 15 25

6 5 15


Particle Size Effect on Reflow Result

Typical results: Type 3 (left) vs. Type 6 (right) using the sameno-clean flux chemistry and reflow profile (RTP).

4/14/2015

53


Ramp to Peak Profile

105


Soak Profile

106

4/14/2015

54


Profile Optimization


Ramp-to-Peak (RTP) vs. Soak

108

Typical results: RTP profile (left), soak profile (right)using the same Type 6 powder size and flux chemistry (no-clean),

4/14/2015

55


Round Aperture vs. Square Aperture

109

Typical results: Circular aperture/pad (left), square aperture/pad(right) using the same Type 3 powder size, area ratio, flux chemistry(no-clean), and reflow profile (RTP).


Why the Difference?

• AR for square:

– D2/4Dt= D/4t

• But area of circle = D2/4 = 0.785 D2

• D2 for a square

– >25% more volume

• Plus experiments show better transfer efficiency from square apertures

– >40% more volume

110

•AR > 0.66 for good printing•Some modern pastes can print <0.50

•Should be verified experimentally

4/14/2015

56


Water-Soluble vs. No-Clean

111

Typical results: Water-soluble (left) vs. no-clean (right)using the same Type 6 powder size and reflow profile (RTP).


Solder Mask Defined is Better

112

Typical results: Non-solder mask defined pad (left),solder mask defined pad (right) using the same Type 6powder size, flux chemistry (no-clean), and reflow profile (RTP).

4/14/2015

57


Paste A S O

The pads in red circle showed graping symptom

Graping Paste “A” vs. Others


• What is head‐in‐pillow?– A defect in which both the solder

paste and the BGA ball reflow but they do not coalesce together.

• How does it happen?1. Component warps during preheat

and soak of profile2. Paste and ball separate prior to

melting3. Paste and ball coalesce separately4. Oxide layer forms on surface of

molten solder5. Component warps back during

cool down but it is already solidified or oxide layer is too thick for paste and ball to coalesce together.

Key Solder Paste Attributes:

•Flux Activity and Oxidation Barrier

•Small Deposit Coalesence (anti‐graping)

•Slump Resistance/Over Print Capable

•Print Performance/No Tailing

The Head-in-Pillow(HIP) Challenge

4/14/2015

58


How HIP Happens


• Type 1: Excessive BGA Oxide or Contamination– The oxide film on solder ball is too thick for solder paste to break through to coalesce.

• Type 2: Excessive Warpage– Solder solidified already when the BGA ball is brought back in contact with solder dome on pad

• Type 3: Warpage + Oxidation– Solder still in molten state when solder ball was brought in contact with solder dome on pad. However, the oxide film prevent the coalescence of two solder bodies.

Types of HIP

4/14/2015

59


Areas of Investigation:Components

• Component Warpage – ball loses contact with the paste

• Coplanarity• Sphere Oxidation

– Uncontrolled manufacturing– Fifo– Packaging– Storage

• Silver Segregation– Cases seen as high as 36%

silver content at the surface.– Silver tailing

• Cleaning process after ball‐attach– Contaminants from solution– Incomplete cleaning– Hydroxide formation


Areas of Investigation:Printing Process (Secondary)

• Setup– Poor registration

– Improper board setup

– Aperture area ratio

– Stable board support and clamping system (vacuum)

– Poor stencil design

• Materials– Transfer efficiency

– Slump resistance

4/14/2015

60


Areas of Investigation: Reflow Overview

• Profile– Ramp rate and soak time

– Peak temperature

– Environment (air vs nitrogen)

• Flux exhaustion– Critical variable: total heat input

– Sublimation/Vaporization

– Small deposits

• Flux activity– Arrhenius equation applies –

higher temps, more active

• pH (acidity) and Acid Value– Acid/Base systems make this an

incomplete/poor predictor

0

0.2

0.4

0.6

0.8

1

0.00001 0.0001 0.001 0.01 0.1 1

P (Oxygen Patrial Pressure)

S (

So

lder

ing

Per

form

ance

)

Air

K=0.5

K=1

K=2

K=5

K=25

S = 1/ (1 + K x P)

K = B0 / B

where

B0 is oxidation barrier

capability of a typical RMA flux, B is that of a given flux

P = oxygen partial pressure

y = -0.0707Ln(x) + 0.8191R2 = 0.9892

0.5

0.6

0.7

0.8

0.9

1

0.1 1 10 100

Flux Quantity (mg)

Flu

x F

ract

ion

Bu

rn-O

ff

TGA Study

Heating profile


Areas of Investigation: Reflow - Preheat

• Fast ramp rate increases slumping

• Slow ramp rate minimizes spacing between paste and solder ball

• Added Considerations:– Metal %– Powder size– Solvent– Print height

4/14/2015

61


Areas of Investigation: Reflow - Soak

• Oxidation occurs the entire time that the board is in the preheat and soak stages of the profile

• Oxidation barrier of solder paste critical

• Other considerations:– Soak profile helps voiding

– High air flow rates


Areas of Investigation: Reflow - Peak

• Peak temperature affects the amount of warping

• TAL impacts the amount of oxides that form on molten solder surface

• Other considerations:– Intermetallic formation

– Low Ag BGA’s

4/14/2015

62


Areas of Investigation: Reflow - Cool Down

Source: Kinyanjui R; Effect of Ball Size on Undercooling of SAC Solder Joints; SMTAI 2005

• Undercooling a positive attribute for HIP

• Closer to pure Sn the alloy is, the more undercooling

• Smaller balls, more undercooling


How Oxidation Barrier Helps HIP

4/14/2015

63


Voiding• Where is voiding an issue?

– BGA’s and CSP’s with microvia‐in‐pad

– LGAs and QFNs with large ground pads

• How does HF Matter?– Major cause of voiding is oxide

removal by activator– HF activators aren’t as effective

• More total activator• More likely to have oxides remaining

at liquidus

– High surface tension of Pb‐Free solders


Slide #126

0

10

20

30

40

97.5

Sn2Ag0

.5Cu

96.5

Sn2.5

Ag0.8

Cu

96.5

Sn3Ag0

.5Cu

95.5

Sn3.5

Ag1Cu

95.5

Sn3.8

Ag0.7

Cu

63Sn3

7Pb

Vo

idin

g (

area

%)

Ag voiding

SAC Sn63

Alloy Effect on Voiding

4/14/2015

64


Slide #127

Impact of Soaking on Voids


Slide #128

Flux Outgassing Rate vs. Heat Input

4/14/2015

65


Slide #129

Outgassing Rate for Two Different Fluxes


Green Evolution: Current Need vs. Long Term Effect

• In 1948, Paul Hermann Müllerwas awarded the Nobel Prize for developing DDT that virtually eradicated malaria and yellow fever from most developed countries.

• In 1962, Rachel Carson’s Silent Spring which began a widespread ban on DDT use.

From Nobel Prize

to

EPA Ban

Slide #130

4/14/2015

66


Greenpeace Greener Guide

Slide #131


Halogens and Green Movement

Courtesy of http://www.elementsdatabase.com

4/14/2015

67


Halogens in Electronics

• PCB Laminate Materials– HF alternatives are more expensive to manufacture and may be more sensitive in high reliability applications

• Components– HF components are being developed, but can be challenging in complex IC’s

• PVC (Primarily in Cables)– HF alternatives are often more brittle and more costly

• Soldering Materials– HF materials could suffer in soldering performance (HIP, graping, long profiles)

Slide #133


Appendix III:Motorola Pb‐Free Cellphone

Assembly

Courtesy: Vahid Goudarzi, Motorola ([email protected])

4/14/2015

68


Slide #135

Motorola’s Process Criteria

• Paste must have good response to pause, tack, slump and other printing metrics

• The process/paste must show good coalescence and solder joint quality in a broad reflow process window

• The reliability of the finished product must be as good or better than the standard Pb solder

• The process must be simple and robust so that it can be transferred to other locations world wide


Paste Evaluation/Manufacturing Process Development

• Screen Printing Evaluation

• Reflow Profile Development

• Tackiness Measurement

• Surface Insulation Resistance (SIR) Evaluation

Slide #136

4/14/2015

69


Slide #137

Lead‐free solder paste suppliers & materials

A3B3

A1B1

C1

A2B2

C2

Phase # 1 Phase # 2 Phase # 3

ABC

Paste Suppliers

Flux Vehicles

Paste evaluation & selection strategy:Work with 8 preferred paste suppliers to develop a lead-free solder paste that meets Motorola’s manufacturing quality & product reliability requirements

These studies were completed using Sn/Ag/Cu, Entek finish boards, & air atmosphere

Pb‐free Solder Paste Evaluation

The Finalists:


Optimum Print Speed, Squeegee Pressure, & Snap Off was set per paste supplier recommendation and validated

Stencil Printing Evaluation

• Objective: To ensure Pb‐free paste performs consistently as a function of time

• Variables:

– Abandon time @ t=0, t=1, & t=4 hours

– Solder paste (A1, B1, C1, A2, B2, C2, A3, B3)

• Output:

– Volume measurement using laser system

– Visually inspect for smearing and selected apertures for clogs.

4/14/2015

70


Selected Inspection SitesBased on gage R&R results

RSC

CSR

ABCDEF

1 2 3 4 5 6 7 8 9 1011 12 13 14 15 16 171 819 202122 23 24

1 2 3 4 5 6 7 8 9 101112 13 14 15 16 171 81 9 20 21 22 23 24

OSP Finish Test Vehicle for Paste Evaluation


Slide #140

Pb‐free Solder pastes performed well @ abandon time = 0h

Solder Paste Volumetric Measurement or 12 mils SMD pads @ t=0 h

200

300

400

Paste C1 Paste B3 Control All Pairs

Tukey‐Kramer

0.05

Volume

Paste C2 Paste A1Paste B1

4/14/2015

71


Slide #141

Paste C2 failed @ abandon time=1h

100

200

300

400

Paste B1Paste B3 Control All Pairs

Tukey‐Kramer0.05

Volume

Paste C1 Paste C2 Paste A1

Solder Paste Volumetric Measurement for 12 mils SMD pads @ t=1 h


Slide #142

Reflow Profile Development

• Objective: To determine reflow process window & identify a Pb‐ free paste which requires MINIMUM peak temp.

• Variables:– Peak temperature– Time above liquidus– Solder Paste (A1, B1, C1, A2, B2, C2, A3, B3)

• Output:– Coalescent performance– Solder joint quality

4/14/2015

72


Slide #143

Pb-free 2X3 Full Factorial Reflow DOE• Minimize peak temp. to reduce thermal stress on Components• Interaction between peak temp. & time above liquidus

60Sec. 70Sec. 80Sec.

229C

237C

245C

Time Above Liquidus

Peak Te

mperatu

re

Peak temp

Time above

liquidus

Ramp rate

Lead-free reflow profile

P1

P2

P3

P7P4

P5

P6

P8

P9

Selected paste MUST perform equally well @ P1 through P9 in air atmosphere

Reflow Profile Matrix


Slide #144

Inspection criterion:

Coalescent performance @ P1,P2, P3, P4, P5, P6, P7, P8 &P9

Poor Coalescent Good CoalescentPoor coalescent is attributed to powder oxidation during reflow

process in air atmosphere

Reflow Profile Development Cont.

4/14/2015

73


Slide #145

Poor Solder Joint

Insufficient toe fillet

Good Solder Joint

Inspection criterion:

Wetting performance @ P1,P2, P3, P4, P5, P6, P7, P8 &P9

Insufficient toe fillet results in field reliability issues

Reflow Profile Development Cont.


Flux Tackiness Measurement

• Objective: To ensure flux provides sufficient tackiness to hold components in place during manufacturing processes

• Variables:– Paste life @ t=0;t=1h t=2h; t=4h; t=8h – Pb‐free solder pastes

• Output:– IPC‐TM‐ 650 Test Procedure: Measure the force required to Separate a 5mm diameter probe from paste

– Shake Test ‐Automated vision inspection after placement

4/14/2015

74


Slide #147

IPC650 Tack Test ‐ Tack test evaluation result

Tack Test

0

0.5

1

1.5

2

2.5

0 1 2 4 8

Time (Hrs)

Tac

k (g

/mm

2) Indium 51A

Indium 92J

Indium 232-99-2

Indium SMQ 230

Tack Test

0

0.5

1

1.5

2

2.5

0 1 2 4 8

Time (Hrs)

Tack

(g

/mm

2)

Control #1

Control #2

B1

B3

Flux Tackiness Measurement Cont.


Slide #148

1) Populated PCBs after 0, 4, and 8 hours

2) Image components to determine X, Y, and Theta offsets.

3) Place PCBs on XY table of Chip Shooter & shake PCBs for 120 Sec.

4) Image components to determine X, Y, and Theta offsets

5) Determine delta for before & after shake process

0-1

23

0

-3

12

1

1

-1

21

0

10

-1

1

-4 -2 0 2 4 6

caps

switch

tantalum

caps

3.2x1.6

resistor

X-Offset Y-Offset Theta

Component placement offset after 120 second of shaking by chip shooter

Flux Tackiness Measurement Cont.

To ensure flux provides sufficient tackiness to hold components in place during assembly process

4/14/2015

75


Slide #149

Surface Insulation Resistance

Supplier B Surface Insulation Resistance Test

24 72 120 168 216 264 312 360 408 456

Hours

Ohm

s Channel 1

Channel 11

Channel 21

Channel 153

Channel 163107

109

Motorola SIR Test Boards = B25 Test Board + Solder mask

SIR requirements is minimum of 10 8 Ohms


Printing PasteVisual Inspection

Volumetric data

ReflowP1,P5,P9

P1,P2,P3,P4,P5,P6,P7,P8,&P9

TackinessInstron IPC650

Shake Test

QualitySolder Joint

ALT

SIRJ-STD B25

Motorola

NT

NT

NT NT

NT NT

NTNT

NT

NT

NT

NT

NT

NT

NT

NT NT

NT

NT

NT

NT

NT

NT NT

Passed Failed NT Not tested

NT

NT

NT

NT

NT

NT

NT

NT

Paste B3 met all requirements

Paste Final Evaluation Results

PasteA1

PasteC1

Phase 1PasteA2

PasteC2

Phase 2PasteB3

PasteA3

Phase 3

PasteB1

PasteB2

4/14/2015

76


Slide #151

Solder Coalescent Comparison @ P1, P5, & P9

B3

@

P1

B3

@

P1

B3

@

P5

B3

@

P5

B3

@

P9

B3

@

P1

B3

@

P1

B3

@

P1

B3

@

P1

B3

@

P1

B3

@

P5

B3

@

P9P9

B3

@

P5

Good Coalescent

Supplier B paste

A1

@

P 1

A1

@

P5

A1

P9

@

Poor Coalescent

Supplier A Paste

Paste A1 does not fully coalesce and result in grainy joint due to powder oxidation in air atmosphere

Paste Final Evaluation Results Cont.


Slide #152

No significant difference in solder joint fillet @ P1, P5, & P9 usingB3 solder paste

Lead‐free @ P1(229;60)

Lead‐free@ P9(245;80) Leaded @ 210 C

Lead‐free @ P5(237;70)

Paste Final Evaluation Results Cont. Solder Joint Evaluation @ P1, P5, & P9

4/14/2015

77


Slide #153

B3@P1

B3@P5

B3@P9

Sn/Pb

Solder Joint Evaluation @ P1, P5, & P9

No significant difference in solder joint fillet @ P1, P5, & P9 usingB3 solder paste

Paste A1

Insufficient toe fillet

Paste Final Evaluation Results Cont.


Slide #154

Intermetallic formationPaste Final Evaluation Results Cont.

P1 P5

P9

0.0023 mm 0.0025 mm

0.0025 mm

4/14/2015

78


Slide #155

Recommened 235C profile

0

50

100

150

200

250

300

0 1 2 3 4 5 6 7 8

Time/min

Tem

p/C

Ramp 0.7 deg/sec

Peak 235CTime above 217C: 70s

Peak Temp. = 235 C +/‐ 5C; Time Above Liquidus = 70Sec +/ 10Sec

Recommended Profile for B3


Slide #156

Product Level & Solder Joint Reliability Evaluation

• Drop Test

• Shear Test

• Liquid-to-Liquid Thermal Shock

• ALT for different Products

Reliability Evaluation

Pb free solder joints MUST perform equal or better than leaded solder joints

4/14/2015

79


Slide #157

1) Dropping products from 5 feet 2) Vert. & horiz.vibration for 2 hrs3) Thermal shock for 48 hrs4) Repeated step 1 thru. 3 X times 5) Measure % joint cracks on shields

Drop Test

Shield solder joint cracking is significantly reduced using B3

Pb-free radios

0

5

10

15

20

25

% of crack

0

5

10

15

20

25

Leaded radios

% of crack

Drop test vehicle

Reliability Evaluation Cont.


Slide #158

0.5mm CSP

CSP

SOIC

0.5mm Conn.

BGA

0.5mm QFP

0.75mm CSP

0.8mm CSP

20X40 Cap.

DIME

0.5 mm CSP

• 6X6 mm Package size • 0.5 mm pitch partial array • 0.3 mm solder balls size

Solder Joint Reliability Evaluation Test Vehicle


4/14/2015

80


Slide #159

•0

•1

•2

•3

•4

•5

•6

•7

•8

•9

•10

•Ceramic

•Inductors

•Tantalum

•Capacitors

•Small

•Capacitors

•Ferrite Bead •Mid-size

•capacitors

•Sh

ea

r a

t fa

ilure

(kg

)•[

afte

r th

erm

al s

hock

]

•SnPb

•SnAgCu

Shear Test

No significant difference in shear force after LLTS.



Slide #160

Variables:

‐ Solder Paste (Paste B3 & Pb Paste)

‐ Component Type ( 0402, 0603, 0805, BGAs, CSPs, VCO,Transformer)

Output:

‐ Electrical test at every 75 cycles for 450 cycles‐ Red dye analysis at 150, 300, and 450 cycles

Liquid‐to‐Liquid thermal shock evaluation (‐55 C to +125 C)


4/14/2015

81


Slide #161

Failed Joint

Passed Joint

Failed Joint

Red dye evaluation result

No significant difference in cracked area in leaded and Pb‐free joints

Cra

ck a

rea

, %

-25

0

25

50

75

100

125

Pb-free Sn-Pb

Solder

All Pairs

Tukey-Kramer

0.05

Joint crack data for different components

Liquid‐to‐Liquid thermal shock results after 450 cycles



Slide #162

Products Built with Pb‐free Paste

* i1000 iDEN

* Concorde iDEN

* i1000 Charger

* i700 iDEN

* i85 iDEN

Products built with Pb‐free solder paste and passed ALT

4/14/2015

82


Slide #163

Appendix IV: Dopants: Fragility

• 01005 passives and 0.3mm CSP

• Very small solder joints

• Grain Structure and/or Intermetallic could traverse entire joint

• Ti, Ni and Co – reduce undercooling and refine grain size


Slide #164

• Miniaturized devices more portable, thus more risk to be dropped.

• Small joints more vulnerable to shock.• Joints with low fragility desired

– Lower hardness (such as low Ag SAC)– Dopant which reduce IMC thickness, scallop

size, or fragility, such as Mn, Ti, Y, Bi, Ce, Ni, Co, Pt

– Form more ductile solder via high Cu– Dopants which reduce Kirdendall void

formation, such as Ni, In, high Cu– Reduce spalling, such as high Cu– Fine grains to nullify effect of anisotropic

orientation of Sn crystal

(Kao)

(Indium)

Drop Test Performance (Mean value)

0

10

20

30

40

Sn1

.1A

g0.4

5Cu0

.1G

e

Sn

1.1A

g0.

47C

u0.

06N

i

Sn1

.07A

g0.4

7Cu

0.08

5Mn

Sn

1.1A

g0.

64C

u0.

13M

n

Sn

1.13

Ag0

.6C

u0.

16M

n

Sn

1.1A

g0.

45C

u0.

25M

n

Sn1

.07A

g0.

58C

u0.

037C

e

Sn

1.09

Ag

0.47

Cu0

.12C

e

Sn1

.05A

g0.

56C

u0.

3Bi

Sn1

.16A

g0.5

Cu0

.08Y

Sn1

.0A

g0.4

9Cu0

.17Y

Sn1

.05A

g0.7

3Cu0

.067

Ti

Sn1

.0A

g0.4

6Cu0

.3B

i0.1

Mn

Sn1

.05A

g0.4

6Cu0

.6B

i0.0

67M

n

Sn1

.19A

g0.

49C

u0.

4Bi0

.06Y

Sn1

.15A

g0.

46C

u0.

8Bi0

.08Y

Sn

1.05

Ag0

.64C

u0.2

Mn

0.02

Ce

SA

C30

5

SA

c38

7

SA

C10

5

Sn6

3

No

. of

Dro

ps

to

Fa

ilure

As-reflowed

After aging

Fragility Challenge

4/14/2015

83


Slide #165

Kirkendall Voids

Cu

Sn

CuSnx

As the new intermetallic grows mostly into the space previously occupied by the tin, stresses are relieved by the formation of tin whiskers.

Kirkendall voids form in copper

Cu


Slide #166

Fragility Challenge‐ Anisotropic Crystal Orientation Effect

• Small joint is more vulnerable to shock, due to – Reduced compliance due to

increasing IMC contribution– The joint may contain only one

grain. The anisotropic nature of grain orientation can cause uneven stress distribution from joints to joints.

– When C‐axis is parallel to pad surface (orange color), crack more prone to happen.

• Desire– Dopants such as Ti which suppress

undercooling hence suppress instantaneous large crystal formation

– Dopants such as Ni, Co which refine the grain size.

(Bieler et al)

C-axis parallel to pad more prone to crack, due to large CTE mismatch & high modulus

Single grain

4/14/2015

84


Slide #167

Copper Tin Intermetallic Thickness

• Cu6Sn5 and Cu3Sn

• Necessary Evil

• Thick Intermetallic Layer = Joint Cracking

• Mn, Ti, et al suppress intermetallic formation

Status of Lead-Free 2015: A Perspective · 2015-04-17 · 4/14/2015 3 © Indium Corporation Intro:...

Documents

Transcript of Status of Lead-Free 2015: A Perspective · 2015-04-17 · 4/14/2015 3 © Indium Corporation Intro:...