Carol A. Handwerker Purdue University
Katsuaki Suganuma Osaka University
Heidi L. Reynolds Sun Microsystems, Inc.
Jasbir Bath Solectron Corporation
E1: Global RoHS and Environmental Regulations
Wednesday PM, April 11, 2007
Room 3005 (Moscone West)
9:30 AM - **E1.1
Environmental Regulations and Materials Technology in the Electronics Industry
Robert Pfahl 1 Show Abstract
1 , International Electronics Manufacturing Inititiative-iNEMI, Herndon, Virginia, United States
This presentation will review current activities in the Electronics industry to address the European Union’s “restriction of the use of certain hazardous substances in electrical and electronic equipment” or RoHS Directive. It will discuss the work that remains to be done over the next years to reduce the ongoing costs of maintaining RoHS compliance. It will then focus on other pending and implemented legislation from around the world, dealing with the design, use, and recycling of materials in electronic products. It will conclude by focusing on the future: What’s ahead, and the alternative directions that industry could take. The thrust of the` argument is that industry should take a proactive approach, work with stakeholders, and direct our activities where there is technical/ecological evidence we could and should be doing a better job to protect the environment. We should involve stakeholders in the process of evaluating alternative technologies to determine trade-offs between product functionality, environmental impact, reliability, safety, and cost. We will give an example of the approach we recommend for working with stakeholders (evaluation of alternatives to brominated flame retardants) and discuss areas we believe warrant industry’s attention.
10:00 AM - **E1.2
Management Methods for Controlling Pollution by Electronic Information Products - China RoHS.
David Towne 1 Show Abstract
1 , Sun Microsystems, Inc., Menlo Park, California, United States
Legislation for the Restriction of Hazardous Substances (RoHS) for the European Union (EU) has been implemented since July 1, 2006. The EU RoHS regulates the use of lead, mercury, hexavalent chromium, PBBs and PBDEs in electronic assemblies. The target date for RoHS legislation in China is March 1, 2007. This talk will highlight the similarities and differences between the EU RoHS and the China RoHS. Challenges with respect to implementation and compliance testing will be discussed. Industry consortia involvement in the development of the "Management Methods for Controlling Pollution by Electronic Information Products" or China RoHS will also be discussed.
10:30 AM - **E1.3
The Greening of Printed Electronics: Novel Materials Selection for Low-Cost Electronics in a RoHS World
Paul Brazis 1 , Jie Zhang 1 Show Abstract
1 , Motorola, Inc., Schaumburg, Illinois, United States
In recent years the international community has become increasingly aware of the environmental impact of manufactured products throughout their entire service life. As a result, regulatory measures such as the European Union Restriction of Hazardous Substances (RoHS) have been enacted and strictly enforced, limiting the selection of substances that may pose harm to consumers and/or their environment after end-of-life. Consequences of non-compliance include large financial penalties and potential ban from doing business within a particular state, nation, or trade bloc. In traditional microelectronics manufacturing, these regulations have forced the need for significant modification of established product design and manufacturing processes, in order to identify and change materials selection away from well-understood systems (most notably, lead-based solder) towards solutions that may impose tradeoffs in reliability or cost.These new environmental regulations need not be simply viewed as obstacles that need to be endured. Rather, these restrictions may also be leveraged as a significant opportunity for directing research towards new approaches that were not previously considered. As an example, this presentation will demonstrate how environmental regulations can steer a program towards solutions that result in better-to-predict health and safety impact while delivering potentially superior performance compared to incumbent methods. This approach will be framed within a description of manufacturable printed electronics work being developed at Motorola. As printed electronics are intended primarily for high-volume, low-cost, and single-use applications, the advantage of achieving improved performance, non-toxicity, and compliance of environmental regulation simultaneously strengthens the market viability for the technology.
11:30 AM - **E1.4
Quantifying Brominated Flame Retardant Content and Emissions in Polymers
Inez Hua 1 2 , Yin Ming Kuo 1 , Brianna Dorie 1 Show Abstract
1 School of Civil Engineering, Purdue University, West Lafayette, Indiana, United States, 2 Division of Environmental and Ecological Engienering, Purdue University, West Lafayette, Indiana, United States
Quantifying Brominated Flame Retardant Content and Emissions from PolymersInez Hua, Yin-Ming Kuo, and Brianna DorieBrominated flame retardants (BFRs) fulfill an important safety function by enabling the manufacturers of electronics, polymers, and other materials to economically meet stringent fire resistance requirements. However, more information is needed about the emissions, fate, and environmental effects of BFRs in the environment. Although the fate and toxicology of some BFRs has been the subject of increasing scientific interest, there is comparatively little information quantifying the potential sources or life-cycle flows of BFRs through the economy. A primary objective of our work is to conduct an inventory analysis on major product categories that incorporate “deca-BDE”, a specific formulation of the PBDEs and tetrabromobisphenol-A (TBBPA). Specific applications for the formulations include the use of TBBPA in epoxy resins and in acrylonitrile butadiene styrene (ABS). Epoxy resins are used for laminating printed circuit boards and printed wire boards; it is estimated that 95% of printed circuit boards contain TBBPA as a flame retardant. Furthermore, a variety of electrical and electronic equipment components are fabricated from ABS, including computer housings, televisions, faxes, and printers. One of the more significant applications of deca-BDE is as an additive to high impact polystyrene (HIP), which is also a material in electrical and electronic equipment. Use, maintenance, re-use and/or recycling and disposal of these products will strongly influence the rates and amounts of flame retardant compounds that are released into the environment. We present results of polymer processing and extraction procedures, during which deca-BDE and TBBPA in polymers are quantified by gas chromatography equipped with various detectors, including an electron capture detector and a mass spectrometer (GC/ECD and GC/MS). These measurements are necessary for estimating historical and current mass flows of deca-BDE and TBBPA. The deca-BDE and TBBPA content varies widely, depending on polymer or product type, as well as age. Also reported are BFR emissions from polymers under simulated use conditions (temperature, humidity).
12:00 PM - E1.5
Bromine Flame Retardant Interactions and Degradation with Gold-Aluminum Wire Bonds
John Osenbach 1 , Joze Antol 1 , Ron Weachock 1 , Carl Peridier 1 , Matt Stahley Stahley 1 , Frank Baiocchi Baiocchi 1 , John DeLucca DeLucca 1 , Dan Gerlach 1 Show Abstract
1 , Agere systems, Allentwon , Pennsylvania, United States
It is well established that bromine additives in the form of flame retardants can be detrimental to the stability of Au-Al wire bonds. In particular bromine has been found to accelerate the degradation of Au-Al wire bonds at high temperatures via corrosion. A number of possible reaction sequences have been proposed to explain the corrosion mechanism. Many of these models require, either explicitly or implicitly, the existence of a film of moisture, which in the presence of Br acts as the electrolyte to promote electrochemical corrosion. Unfortunately, the temperature where many of the studies have been done is in excess of 150C. At temperatures in excess of 150C, surface moisture is expected to be present in only a chemisorbed state. Given that there are strong interaction forces between the chemisorbed moisture and the surface, it is possible that the moisture-electrolyte requirement in these models is not realized. In this paper we present the results of our studies on the degradation of Au-Al wire bonds at temperatures between 150C and 200C. Both mechanical and microstructural data that was collected and analyzed will be presented. Data was taken on devices molded in three different commercially available mold compounds, two Br-containing compounds produced by two different manufacturers and one non-Br-containing compound. In addition, the affect that probing the bond pads prior to wire bonding has on wire bond degradation was also studied. The data indicate that the activation energy for degradation of the wire bonds in Br containing mold compound is substantially higher than that for degradation in Br-free compounds (approximately 2eV and 1eV, respectively). The ramifications of this activation energy difference as it pertains to wire bonding stability will be discussed. Furthermore, the data clearly show that the degradation rate is reduced as the degree of probe damage is increased. Finally, a model based on the formation of thermodymically stable phases in this system is proposed to explain the fundamental mechanisms that lead to Br-induced corrosion of Au-Al wire bonds. This model does not require an electrolyte in the form of adsorbed moisture.
12:15 PM - E1.6
Zn-Sn and Zn-In High Temperature Lead-Free Solders
Katsuaki Suganuma 1 , Jae-Ean Lee 1 , Keun-Soo Kim 1 Show Abstract
1 ISIR, Osaka Univ., Ibaraki, Osaka Japan
Pb-Sn alloys, containing over 85 mass%Pb, have been commonly used in die attach bonding as high temperature solders. The high Pb content in these solders invariably makes them subject to RoHS restrictions; therefore the development of lead-free solders is a target technology. Several high temperature lead-free solder candidates exist, including Sn-Sb, Au-Si, Au-Sn, and Bi-Ag alloys. Besides the good aspects of Sn-Sb, Au-Si, Au-Sn, and Bi-Ag alloys as high temperature lead-free solder, it has some drawbacks such as toxicity of Sb, brittle natures of Bi and Au-Sn intermetallic compounds, and high cost of Au. For these reasons, there is a need to develop new types of affordable and reliable high temperature lead-free solder materials. The potential of newly-designed Zn-Sn and Zn-In alloys as high temperature lead-free solders was evaluated, with particular focus on the fundamental properties and interface stability. The basic properties and the current understandings on these alloys are briefly summarized.
12:30 PM - E1.7
e-Ni/Au UBM & AuSn Solder Bumping Technology for High Temperature LED Applications
Andrew Strandjord 1 Show Abstract
1 , FlipChip International, Phoenix, Arizona, United States
E2: Tin Whisker Nucleation and Growth
Wednesday PM, April 11, 2007
Room 3005 (Moscone West)
2:30 PM - **E2.1
Sn whisker: Understanding and Prevention?
John Osenbach 1 Show Abstract
1 , Agere systems, Allentwon , Pennsylvania, United States
Although solid state whisker growth has been know for more than 60 years, a basic understanding upon which a quantitative model that describes whisker nucleation and growth remains elusive. As such, mathematical expressions that a-prior can be used to predict if and/or when solid state whisker growth in a particular whisker prone film will occur does not yet exit. This is somewhat troubling since the European Restriction of Hazardous Substances (RoHS) legislative requires the elimination of Pb from electronic devices. The most non-disruptive, economical, and widely used way of doing so is to replace the Sn-Pb finish with a pure Sn finish. However, Sn is one of the materials that is prone to solid state whisker growth. In addition to the lack of a quantitative predictive model for Sn-whisker growth, the literature on this topic contains a significant number of seemingly conflicting and sometimes contradictory results that have lead various authors to diametrically opposed conclusions. In spite of the limited data set, the apparent conflicting data, and lack of agreement on the detailed mechanisms that are at the root of solid state Sn-whisker formation and growth, there are a number of variables that are generally believed to influence and/or cause whisker growth including: 1) the intrinsic stress in the as plated film; 2) extrinsic stress arising from chemical reactions between the film the ambient and the substrate alloy; 3) purity of the Sn plate; 4) grain size and grain crystallographic orientation of the Sn plate; 5) lead frame surface finish; 6) pre-plate chemical treatment of the lead; 7) the process used to form the lead, 8) the bend radii of the lead; 9) the Sn-plating thickness; 10) the post plate thermal processing; and 11) the storage and operating ambient. These variables have been used as a starting point for the development of standardized whisker testing procedures, to improve plating chemistries, and to develop Sn-whisker mitigations techniques. Unfortunately, the lack of a fundamental understanding as well as the lack of a predictive model means that passing a standardized test procedure does not insure whisker free Sn. In this talk I will review the current understanding of Sn-whisker nucleation and growth. .Particular attention will be given to some of the more controversial topics and the effectiveness of the current mitigation strategies.
3:00 PM - **E2.2
An Electrical Characterization of Tin Whiskers
Robert Hilty 1 , Ned Corman 1 Show Abstract
1 Technology, Tyco Electronics, Harrisburg, Pennsylvania, United States
Tin whiskers are one of the greatest reliability concerns with lead free electronics. The blueprint for failure is simple – a metallic filament of tin bridges between two conductors and leads to an electrical short. While this makes intuitive sense, there are limits to when this failure mechanism will be active. This research examines the role of tin oxides and surface contaminants on preventing electrical conduction between tin whiskers and an adjacent contact. Contact resistance measurements were made to tin whiskers under various loading schema to evaluate the conditions under which whiskers would lead to electrical conduction. Electrical conductivity to the whiskers requires breakdown of inherent oxide layers present on tin; this can be achieved mechanically, thermally or by way of electrical breakdown. Electrical breakdown voltages of 3 volts or more were required to achieve continuity.
4:00 PM - **E2.3
Sn Whisker Formation: Stress Generation at the Sn-Cu Interface.
Gery Stafford 1 , Carlos Beauchamp 1 , Ugo Bertocci 1 , Espen Sandnes 1 , William Boettinger 1 Show Abstract
1 Metallurgy Division, NIST, Gaithersburg, Maryland, United States
Sn is widely used as a coating in the electronics industry because it provides excellent solderability, ductility, electrical conductivity and corrosion resistance. However, Sn whiskers have been observed to grow spontaneously from Sn electrodeposits and are known to cause short circuits in fine pitch pre-tinned electrical components. Compressive residual stress is thought to play a role in whisker formation on the surface of electrodeposited Sn. However, questions remain as to the dominant source of stress and the precise growth mechanism. One of the most commonly discussed sources of compressive stress in Sn electrodeposits is intermetallic compound (IMC) formation, specifically Cu6Sn5, due to the reaction of substrate Cu with Sn. The volumetric strain for this process is the difference between the molar volume of the Cu6Sn5 and the molar volume of an unalloyed mixture of pure Cu and pure Sn of the same average composition as the intermetallic. The IMC takes up 5.6 % less volume than a rule of mixture combination of the Sn and Cu reactants from which the IMC forms. A tensile stress develops in the IMC layer which in turn induces a compressive stress into the adjacent Sn layer. The mechanism by which the compressive stress is generated in the Sn layer is poorly understood. Pure Sn, Sn-Cu and Sn-Pb alloy electrodeposits vary in their ability to form hillocks and whiskers, although all show similar IMC reaction zones at the deposit- substrate interface. This raises the question whether or not other factors, in addition to interfacial IMC formation, might be responsible for whisker formation, especially since electroplating itself often produces significant residual stress in deposits. This paper presents new data on the kinetics and stresses associated with the formation of Cu6Sn5 intermetallic at the Cu-Sn interface obtained by monitoring the curvature of a Sn-coated Cu cantilever electrode while in solution and under potential control. The force exerted onto the cantilever, as the result of Cu6Sn5 formation, was measured as a function of time and after electrochemically dissolving the un-reacted Sn. This allowed us to separate the IMC stress and the residual growth stress inherent to electrodeposition. We separately examined the process using an electrochemical quartz crystal nanobalance (EQNB) in order to determine the IMC thickness, which in turn allowed us to calculate the average biaxial stress in the IMC film. Mechanisms by which this IMC stress can impart compressive stress into an adjacent Sn layer are discussed.
4:30 PM - **E2.4
Interaction of Stress, Intermetallic Formation and Defect Kinetics in Sn Whisker Formation.
Eric Chason 1 , Lucine Reinbold 1 , Nitin Jadhav 1 , Vernorris Kelly 1 , Jae Wook Shin 1 , Eric Buchovecky 1 , Ramanarayan Hariharaputran 1 , Sharvan Kumar 1 Show Abstract
1 Division of Engineering, Brown University, Providence, Rhode Island, United States
Sn whiskers were discovered over 50 years ago, but many aspects of how, when and why they form are not understood. A deeper understanding would improve our ability to develop whisker mitigation strategies and assess their reliability. We present the results of a series of studies designed to explore the fundamental processes that control whisker formation in Sn layers on Cu. In particular, we focus on the connection between film stress (believed to be the driving force for whisker growth), the formation of the intermetallic (IMC) phase and the growth kinetics of whiskers. We use real-time measurements to quantify the simultaneous evolution of stress, IMC volume and whisker density combined with TEM, SEM and XRD studies of the microstructural evolution. From this work, a picture is emerging of how the IMC particles grow, how stress is generated and relaxed in the Sn matrix and how whisker nucleation is connected to film stress. We find that particles initially nucleate and grow preferentially where grain boundaries in the Sn film intersect the Cu film, with a secondary preference for nucleation at Cu grain boundaries. The growth kinetics are explained by a model that considers the effect of multiple diffusion paths that change as the IMC grows and blocks them. Stress that is generated around the growing IMC particles is mediated by dislocation glide (as seen in TEM) and diffusional creep processes. This distributes the large strain due to the volume mismatch between the phases over a much larger region than would occur by purely elastic stress fields. The presence of the surface oxide is thought to play a critical role in the stress evolution by preventing relaxation at the surface, either by preventing diffusion of point defects or by causing dislocation pileup, or both. Comparison of Sn films with Pb-Sn films illustrates how the presence of Pb modifies the stress and IMC kinetics and suppresses whisker evolution.
5:00 PM - E2.5
Distribution and evolution of stress in Sn layers on Cu
Nitin Jadhav 1 , Lucine Reinbold 1 , Vernorris Kelly 1 , Jae Wook Shin 1 , Anthony Johnson 1 , Michael Task 1 , Sharvan Kumar 1 , Eric Chason 1 Show Abstract
1 , Brown University, Providence, Rhode Island, United States
Stress is generally believed to be the driving force for Sn whisker formation, so a better understanding of how stress evolves and how it is distributed in the Sn layer would be useful for relating stress evolution to whisker growth kinetics. We have used a real-time wafer curvature technique to monitor the stress in bilayers of Sn and Cu deposited on glass substrates by electrodepositon and vapor deposition. The curvature technique measures the integrated value of the stress through the entire multilayer composite, so a single curvature measurement cannot distinguish between stress in the Sn layer, Cu layer or intermetallic layer that forms between them. To separate the stresses in the different layers, we measure the change in curvature that occurs when the Sn layer is removed by selectively etching only that layer. Assuming that the stress in the underlying layers doesn’t change when the Sn is removed, the curvature change can be attributed to the stress in the Sn layer. To understand the distribution of stress in the Sn layer, we compare stress measurements of films structures grown with different thicknesses. For example, we measure the curvature change in several different thicknesses of Sn (600nm, 1200 nm Sn and 2400 nm Sn) deposited on a 600 nm Cu layer. By comparing the change in curvature when the each Sn layer is etched, we can roughly determine how the stress is distributed through the layer, i.e, is the stress uniform through the layer or confined near the Sn/Cu interface. By monitoring a number of samples grown at the same time, we can determine the kinetics of stress evolution. Measurements of the change in weight of the samples are used to confirm that the IMC growth kinetics don’t depend on the Sn layer thickness measured and to relate the stress to the intermetallic volume.
5:15 PM - E2.6
Characterization and Modeling of Intermetallic Growth Kinetics in Sn Layers on Cu.
Eric Buchovecky 1 , Lucine Reinbold 1 , Ramanarayan Hariharaputran 1 , Eric Chason 1 , Sharvan Kumar 1 Show Abstract
1 Division of Engineering, Brown University, Providence, Rhode Island, United States
The formation of whiskers from a Sn film deposited on Cu is intimately linked with the nucleation and growth of the intermetallic phase (IMC) within the Sn. Halting or slowing the growth of IMC could provide a valuable strategy for mitigating whisker formation in Pb-free plating finishes and solders. The measured IMC growth kinetics do not conform to a simple model, suggesting that multiple diffusion mechanisms are interacting to control the growth. To characterize the IMC growth kinetics, we have measured the mass of Cu6Sn5 IMC formed at room temperature as a function of time for both PVD and electroplated Sn films on Cu. We find that the IMC growth curves undergo an abrupt change in slope (more than 20-fold reduction in growth rate) during the period from approximately 4 to 8 days after initial film deposition. SEM observation of the Sn/Cu interface during this period (after chemically stripping the overlying Sn) shows discontinuous IMC grains which occlude only a fraction of the interface area. IMC appears to nucleate and grow preferentially at the intersections of Sn grain boundaries with the Sn/Cu interface. This suggests that the turn-over of the growth curve may be related to sealing off of high-diffusivity pathways for Cu influx by the growth of IMC in the Sn grain boundaries.To better understand the origin of the sudden change in growth rate, we present numerical models in which the nucleation and growth of IMC is restricted to the Sn/Cu interface, and the supply of Cu is controlled by the available area of Sn/Cu interface. We examine the effects of enhanced rates of IMC nucleation, IMC growth, and Cu influx in the Sn grain boundary regions. Comparison of computed growth curves and morphologies with experimental observations allows us to evaluate different mechanisms potentially responsible for the observed decrease in IMC growth rate.
E3: Poster Session: Pb-Free and RoHS-Compliant Materials and Processes
Thursday AM, April 12, 2007
Salon Level (Marriott)
9:00 PM - E3.1
Dissolution Kinetics of Nickel in Lead-free Sn-Bi-In-Zn-Sb Soldering Alloys
Katayun Barmak 1 , David Berry 1 , Vira Khoruzha 2 , Kostyantyn Meleshevich 2 , Vasyl Dybkov 2 Show Abstract
1 Department of Materials Science and Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States, 2 Department of Physical Chemistry of Inorganic Materials, Institute for Problems of Materials Science, Kyiv Ukraine
Because of the toxicity of lead, conventional Sn-Pb solders are gradually being replaced with Sn-base soldering alloys containing additions of other metals (Ag, Bi, Ga, In, Zn, Sb, etc.). To avoid the excessive loss of metals in contact with the soldering alloys during soldering, data on their dissolution rates in molten solders are needed. In this work, we report on the dissolution kinetics of nickel in liquid 87.5% Sn-7.5% Bi-3% In-1% Zn-1% Sb and 80% Sn-15% Bi-3% In-1% Zn-1% Sb soldering alloys. The dissolution rates are obtained by the rotating-disc technique in the temperature range of 250-450 °C. The dissolution process of any metal in the liquid phase is described by the Nernst-Shchukarev equationc = cs[1 – exp(–kSt/v)],where c is the instantaneous concentration of the dissolved metal in the bulk of the liquid phase at a time t, cs is the solubility (or saturation concentration) at a given temperature, k is the dissolution rate constant, S is the surface area of the solid in contact with the liquid, and v is the volume of the liquid phase. Hence, the dissolution process can be fully characterized by two quantities, namely, the solubility (saturation concentration), cs, and the dissolution rate constant, k. The temperature dependence of the nickel solubility in the alloys was found to obey the following equations of the Arrhenius type: cs = 4.94 × 102 exp (–39500/RT) mass% for the former alloy and cs = 4.19 ×102 exp (–40200/RT) mass% for the latter, where R is the gas constant, J mol-1 K-1, and T is the absolute temperature, K. The dissolution rate constants, k, are rather close for both alloys and vary at an angular rotational speed of 24.0 rad s-1 from about 1.1 × 10-5 m s-1 at 250 °C to 6.2 × 10-5 m s-1 at 450 °C. The dependence of the dissolution rate constant, k, upon the square root of the angular rotational speed, ω1/2, is linear, indicative of the diffusion-controlled dissolution process.Using these data, the thickness, xd, of the dissolved part of the nickel layer during the soldering procedure can readily be estimated from a simple equation of the form: xd = cskt/ρNi,where ρNi is the density of nickel, kg m-3, and cs must be expressed in kg m-3.
9:00 PM - E3.10
Electrodeposition Process and Solder Characteristics of the Pb-free Composite Solder Reinforced with Carbon Nanotubes
Kwang-Yong Lee 1 , Eun-Kyoung Choi 1 , Tae-Sung Oh 1 Show Abstract
1 Materials Science and Engineering, Hongik University, Seoul Korea (the Republic of)
While Sn-Pb solders have been extensively used as interconnect materials in electronic industry, there are international legislative actions to limit the use of Pb in electronics. An attractive approach to improve the reliability of a Pb-free solder joint is to use a composite solder by adding reinforcements to a Pb-free solder alloy. In the present study, a new lead-free composite solder, where carbon nanotubes (CNTs) were uniformly dispersed as reinforcements in the Sn matrix, was developed using co-electrodeposition process of the Sn and the CNTs. The Sn-CNT composite solder bath was prepared by mechanical mixing of CNTs, surfactant and Sn solution. The electrodeposition mechanism of the Sn-CNT composite solder could be suggested as a repeating process of Sn layer formation, CNT adsorption, and entrapment of CNTs into the Sn matrix. With increasing the CNT concentration in the bath up to 10 g/l, the CNT content in the Sn-CNT composite solder reached a maximum and remained constant even with further increasing the CNT concentration in a bath. With increasing the CNT concentration in the bath from 0 g/l to 10 g/l, the shear energy increased linearly more than 50% from 1.3×10-5 J to 2.1×10-5 J and the shear displacement was also improved for about 50% from 59um to 89um, indicating that the mechanical reliability of a solder joint can be substantially improved by using a composite solder reinforced with CNTs.
9:00 PM - E3.2
Growth Kinetics of the Intermetallic Layer at the Interface of Solid Metals with Liquid Soldering Alloys
Vasyl Dybkov 1 Show Abstract
1 Department of Physical Chemistry of Inorganic Materials, Institute for Problems of Materials Science, Kyiv Ukraine
The formation of a brittle intermetallic layer weakens the tranzition zone between a solid metal and a soldering alloy. It is therefore essential in practice either to reduce its thickness to a permissible value or to avoid its occurrence at the solid-liquid interface at all. The net rate of the ApBq layer formation between a solid A and a liquid B is equal to the difference of the rates of its growth and dissolution (for more detail, see V.I. Dybkov, Reaction diffusion and solid state chemical kinetics, The IPMS Publications, Kyiv, 2002, Ch. 5; free access at http://users.i.com.ua/~dybkov/V). The dissolution rate is(dx/dt)dissolution = bt = b0 exp(–at), (1)where b0 = csk/ρintφ, a= kS/v, cs is the solubility (or saturation concentration) of a solid metal A in a liquid B at a given temperature, k is the dissolution rate constant, ρint is the density of the intermetallic compound ApBq, φ is the content of A in ApBq, S is the surface area of the solid in contact with the liquid, and v is the volume of the liquid phase. If both components A and B contribute to the layer-growth process, the rate of its growth at the A-ApBq and ApBq-B interfaces is (dx/dt)growth = k0B1[1 + (k0B1x/k1B1)] + k0A2[1 + (k0A2x/k1A2)], (2)where all k0 are chemical constants and all k1 are diffusional constants.Hence, the rate of layer formation is described by the equationdx/dt = k0B1[1 + (k0B1x/k1B1)] + k0A2[1 + (k0A2x/k1A2)] - b0 exp(–at).(3)From this equation, a condition for the absence of the ApBq layer at the interface between A and B can readily be derived. Indeed, the highest value of its first two terms is equal to the sum k0B1 + k0A2. Therefore, atk0B1 + k0A2 < b (4)the ApBq layer cannot occur between A and B since dx/dt < 0. It means the rate of layer growth due to chemical reactions at the A-ApBq and ApBq-B interfaces is less than the rate of its dissolution at ApBq-B interface. Since the dissolution rate gradually decreases with passing time from b0 to 0, the ApBq layer occurs between A and B after some delay.Under conditions of simultaneous dissolution in the liquid phase at a constant rate b, the thickness of the ApBq layer asymptotically tends to its highest value xmax = k1(k0 – b)/ k0b.(5)In this equation, the sum of growth constants was replaced by one constant for simplicity. If the growth rate of the ApBq layer is limited by the rate of diffusion of components A and B across its bulk, so that k0 >> k1/x, while the rate of chemical transformations (partial chemical reactions) far exceeds the dissolution rate (k0 >> b), equation (5) takes a simpler form:xmax = k1/b).(6)This equation allows to evaluate easily the thickness of the ApBq intermetallic layer formed between a solid metal A and a liquid solder B under conditions of simultaneous dissolution in the liquid phase. Examples of calculations are presented.
9:00 PM - E3.3
Fabrication and Characterization of 0.95(Na0.5K0.5)NbO3-0.05LiTaO3 Piezoelectric Ceramics by Pressureless Sintering.
Min-Soo Kim 1 , Sin-Woong Kim 2 , Soon-Jong Jeong 1 , Bok-ki Min 1 , Jae-Sung Song 1 Show Abstract
1 Advanced Materials & Application Research Laboratory, Korea Electrotechnology Research Institute, Changwon, Kyungnam, Korea (the Republic of), 2 Department of Ceramic Science and Engineering, Changwon University, Changwon, Kyungnam, Korea (the Republic of)
Lead oxide-based ferroelectric materials are the most widely used for piezoelectric transducers due to their excellent piezoelectric properties. Considering lead toxicity, there is great interest in developing lead-free piezoelectric materials, which are biocompatible and environmentally friendlier. Recently alkali oxide materials, such as sodium - potassium niobate have drawn much attention due to their ultrasonic applicability and are also considered as promising candidates for a piezoelectric lead-free system. However, it is difficult to sinter such NKN-based materials via conventional sintering process. In this reason, many researchers have investigated hot press or spark-plasma sintering of NKN-based ceramics. From the industrial point of view, pressureless sintering is required for mass production. Therefore, in this study, dense 0.95(Na0.5K0.5)NbO3-0.05LiTaO3 for lead-free piezoelectric ceramics were developed by conventional sintering process. Sintering temperature was lowered by additions as a sintering aid. The electromechanical coupling factor and piezoelectric constant of NKN-5LT ceramics were improved. These excellent piezoelectric and electromechanical properties indicate that this system is potentially good candidate for lead-free material for a wide range of electro-mechanical transducer applications.
9:00 PM - E3.4
Pulsed Laser Deposition of ferroelectric NaNbO3 Thin Films on MgO substrate.
Shinya Oda 1 , Takehisa Saito 1 , Takahiro Wada 1 , Hideaki Adachi 2 Show Abstract
1 Materials Chemistry, Ryukoku University, Otsu, Shiga, Japan, 2 Advanced Technology Research Laboratories, Matsushita Electric Ind. Co., Ltd., Kyoto Japan
NaNbO3 (NN) has been attracting attention as lead-free piezoelectric materials. NN is an antiferroelectric material with high Curie temperature of 365oC. However, high-density NN ceramics showed a P-E hysteresis loops characteristic of ferroelectrics and also showed piezoelectricity . Recently, we have successfully deposited good quality NN films on (100) SrTiO3 (STO) substrate with perovskite structure by Pulsed Laser Deposition (PLD) . In this study, we tried to fabricate epitaxial NN films on (100) MgO substrate with rock-salt structure using similar PLD technique. We think that MgO substrate is better than STO for the deposition of NN films, because MgO crystal is more convenient than STO crystal and easier to be processed. First we deposited the NN films directly on the MgO substrate. However, we could not obtain good quality NN film by the deposition directly on the MgO substrate. Then, we I searched the buffer layer for the deposition of NN film on MgO. We have successfully fabricated good quality NN films on the MgO substrate by the using of K-Ta-O (KTO) buffer layer.
The KTO buffer layer with a thickness of 2 nm was deposited on (100) MgO substrate. Then, SrRuO3 (SRO) lower electrode layer with a thickness of 150 nm was deposited on KTO/MgO substrate. The NN film with a thickness of about 3 μm was deposited at substrate temperature of 750oC. All the layers were deposited by PLD using high-density ceramic targets.
The crystallographic analysis of the films was performed by the conventional θ/2θ scan mode and rocking curve measurements. Surface of the NN film was observed by atomic force microscope (AFM) and surface and cross-section were observed by scanning electron microscopy with ultimate analysis (SEM-EDX). Dielectric properties, such as relative dielectric constant, εr, and loss tangent, tanδ, were measured using an inductance-capacitance-resistance (LCR) meter. The P-E hysteresis loop was observed using a ferroelectric tester, TF2000FE.
X-ray diffraction showed that the NN films were epitaxially grown on a SRO/KTO/(100)MgO substrate by PLD. The full width at half maximum of the rocking curve was as small as 0.14°. The AFM showed that the step of an atomic order could be observed on the film surface and it has confirmed that the surface of the NN film was very flat. The relative dielectric constant, εr, and dielectric loss, tanδ of the film were 350 and 0.05 at 1 kHz, respectively. The P-E hysteresis loop of the NN film was characteristic of ferroelectric behavior with Pr=8.4 μC/cm2 and Ec=64 kV/cm. These values are a little larger that those of the film deposited on STO substrate. The characteristics were discussed with a comparison of the NN film epitaxially grown on (100) STO substrate.
 T. Wada et al., Jpn. J. Appl. Phys.42, 6110-6114 (2003).
 T. Saito et al., Jpn. J. Appl. Phys. 44, 6969-6972 (2005).
9:00 PM - E3.5
First-Principles Calculation of Lead-Free Piezoelectric (K0.5Bi0.5)TiO3
Tomoyuki Koyama 1 , Akio Shigemi 1 , Takahiro Wada 1 Show Abstract
1 Materials Chemistry, Ryukoku University, Otsu, Shiga, Japan
Recently, lead-free piezoelectric materials have come under the spotlight from the viewpoint of environmental protection. (K0.5Bi0.5)TiO3 (KBT) has been studied as a candidate for a lead-free piezoelectric material . KBT is a ferroelectric material with a high Curie temperature of 377°C and has some crystal phases with different crystal system in the wide range of temperatures. At room temperature, KBT with a tetragonal (I4mm) symmetry exhibits ferroelectric behavior and transforms to a paraelectric phase with cubic symmetry at 377°C. In this study, in order to clarify thermodynamic relation of the various KBT phases, we calculated enthalpies of formation for some KBT phases, cubic (Fm3m), tetragonal (I4mm) and virtual rhombohedral (R3m) phases. In addition, we calculated electronic structures of cubic KBT and some compounds such as BaTiO3 and (K0.5La0.5)TiO3 for reference.
The present calculation was carried out by the similar method reported in our recently published paper . All calculations were carried out within the local density approximation (LDA) of density functional theory, using a plane-wave pseudopotential method (CASTEP code). A k-point mesh generated by the Monkhorst-Pack scheme was employed for numerical integrations over the Brillouin zone. A k-point mesh of 6×6×6 was generated for cubic or virtual rhombohedral structures and that of 6×6×7 was used for tetragonal structure.
Theoretically calculated lattice parameters for the cubic and tetragonal phases of KBT were obtained at 0 K, and in contrast, the experimental values were obtained at higher temperatures. However, our theoretically determined lattice parameters successfully reproduced their experimental values within an error of ±3.5%.
The calculated enthalpy of formation for KBT decreases with decreasing crystal symmetry. The lowest theoretical enthalpy of formation, –1508.65 kJ/mol, was obtained for the tetragonal phase. This result corresponds to the experimental result that the tetragonal phase is stable at low temperatures. Our present calculation showed that the enthalpy of formation of the rhombohedral phase is higher than that of the tetragonal phase.
We examined the electronic structure of cubic KBT. The total and local densities of states (DOS) are calculated. The chemical bonding between Ti 3d and O 2p orbitals, and Bi 6p and O 2p orbitals are covalent while the bonding between K 4s and O 2p orbitals is ionic. The valence band mainly consists of the bonding orbital of O 2p, Ti 3d and Bi 6p and the top of the valence band consists of an O 2p component. On the other hand, the conduction band mainly consists of antibonding orbitals of Ti 3d, Bi 6p and O 2p, and the bottom of the conduction band consists of a nonbonding orbital that arises from Ti 3d component.
 T. Wada, A. Fukui and Y. Matsuo, Jpn. J. Appl. Phys. 41, 7025-7028 (2002).
 A. Shigemi and T. Wada, Jpn. J. Appl. Phys. 44, 8048-8054 (2005).
9:00 PM - E3.6
The Study of the Organic Residue Influence on the Crystal Structure of Ge Nanoparticles
Suk Jun Kim 1 , Eric Stach 1 , Carol Handwerker 1 , Ling-Shao Chang 2 , Alexander Wei 2 Show Abstract
1 Materials Engineering, Purdue University, West Lafayette, Indiana, United States, 2 Chemistry, Purdue University, West Lafayette, Indiana, United States
Printed electronics has drawn the interest of many researchers due to several advantages – high production rates accompanied by the ability to switch to low volume as well as high volume designs, low capital cost, and potentially more environmentally friendly processes and materials – than lithography. As a consequence, various candidate materials for all of the components are being studied and applied in printed electronic platforms. Of the materials needed, semiconductor materials that can be formed into devices at low temperatures are essential for the needed electronic components, such as field effect transistors. Germanium nanoparticles for this application are prime candidates for this application and are the subject of the coalescing study reported. The Ge nanoparticles synthesized at room temperature in dimethoxyethane by reduction of GeCl4 with Na(naphthalide) and subsequent reaction with butyl Grignard had an average particle size of approximately 3.8nm with the nanoparticles remaining surrounded by a significant amount of organic residue. All particles were ST12, a metastable tetragonal phase of bulk Ge. In the in-situ TEM experiments, nanoparticle coalescence was observed in the temperature range of 200°C to 400°C. The various coalescence temperatures can be assumed that organic capping layer prevents nanoparticles from coalescing, so that some particles surrounded by relatively a small amount organic residue coalesced at lower temperature. Although the particle size increased to over 70nm during coalescing, the particles remained ST12, and a significant number of nanoparticles formed in the organic residue upon heating to 300°C. Based on TEM results, it can be postulated that the sintering and coalescing led to the broad exothermic peak (150~300°C) and the organic residue transformation and weight loss induced the sharp endothermic peak (300~400°C).
9:00 PM - E3.7
Electrochemical Migration Mechanism of SnPb and Pb-free Solder alloys for PCBs.
YoungRan Yoo 1 , JaYoung Jung 1 , ShinBok Lee 2 , YoungBae Park 1 , YoungChang Joo 2 , YoungSik Kim 1 Show Abstract
1 School of Advanced Materials Engineering, Andong National University, Andong, Gyeongbuk, Korea (the Republic of), 2 School of Materials Science and Engineering, Seoul National University, Seoul Korea (the Republic of)
9:00 PM - E3.8
Dielectric Properties of Barium (Strontium) Titanate/ Polybenzoxazine Composites with 0-3 Connectivity for Electrical Applications
Gasidit Panomsuwan 1 , Hathaikarn Manuspiya 1 , Hatsuo Ishida 2 Show Abstract
1 Petroleum and petrochemical, Chulalongkorn university, Bangkok Thailand, 2 Macromolecular Science and Engineering, Case Western Reserve University, Cleveland, Ohio, United States
Polymer-ceramic composites are good candidate materials for electrical applications, especially embedded capacitors, because they combine good processability of polymer and high dielectric performance of ceramic resulting in low processing temperature and cost. In this study, barium titanate (BaTiO3) and barium strontium titanate (Ba1-xSrxTiO3) powders prepared from sol-gel process were incorporated into polybenzoxazine matrix to form 0-3 connectivity. Polybenzoxazine is used as the polymer matrix because it offers significant properties over conventional thermosetting materials such as light weight, no acid or alkaline as catalyst, and zero volumetric shrinkage. The thermal and dielectric properties of composites were investigated using thermalgravimetric analysis (TGA) and impedance/gain-phase analyzer, respectively. The distribution of ceramic fillers in polybenzoxazine matrix was observed with scanning electron microscope (SEM). From TGA results, it was found that polybenzoxazine starts to decompose at temperature around 250C°. The decomposition temperatures of the composites increase with increase of the ceramic contents. The dielectric properties of these composites, including dielectric constant and dielectric loss, were studied as a function of frequency range of 1 kHz-10 MHz and temperature. It was found that the dielectric constant of composites were nearly constant with changing frequency, indicating low frequency dependence. By adding BaTiO3 fillers, the dielectric constants of the composites significantly increase with increasing BaTiO3 contents. Dielectric loss values of the composites were less than 0.016. Furthermore, the dielectric properties of barium strontium titanate (Ba1-xSrxTiO3) with strontium amount of 0.3/polybenzoxazine composites were also investigated and compared with composites of BaTiO3. The dielectric constants of these composites were fitted with theoretical models for 0-3 connectivity (Yamada’s model).
9:00 PM - E3.9
Microstructures and Dielectric Properties ofCu-deficient and excess CaCu3Ti4O12 Polycrystalline Ceramics
Sung Yun Lee 1 , Duk Keun Yoo 1 , Sang Im Yoo 1 Show Abstract
1 School of Materials Science and Engineering, Seoul National University, Seoul Korea (the Republic of)
Carol A. Handwerker Purdue University
Katsuaki Suganuma Osaka University
Heidi L. Reynolds Sun Microsystems, Inc.
Jasbir Bath Solectron Corporation
E4: Pb-Free Solder: Materials and Properties
Thursday AM, April 12, 2007
Room 3005 (Moscone West)
9:30 AM - E4.1
Applications of Nanoindentation in the Evaluation of Mechanical Properties and Reliability of Sn-Ag-Cu Solders
Babak Arfaei 1 , Junghyun Cho 1 Show Abstract
1 Mechanical Engineering Department, State University of New York at Binghamton, Binghamton, New York, United States
The demand for environmental initiatives within the electronics industry, such as the replacement of lead-tin solders, has increased in the last decade. Despite the fact that the removal of lead from solder has been accelerated by legislation in different parts of the world (especially in the case of the RoHS/WEEE directives in the EU), the lack of reliable information concerning the mechanical properties of Pb-free solder alloys and the intermetallic compounds (IMCs) resulting from the reflow process is still a major concern. This paper reports an investigation done on the Sn-Ag-Cu alloys, which are currently considered to be the best replacement for the Sn-Pb solder. In particular, nanoindentation is used to assess mechanical performance of individual constituents of Pb-free alloys (i.e., Sn-based matrix and various types of IMCs). One purpose of this research is to correlate the nanoindentation behavior to macroscopic solder performance in order to determine the relevance of nanoindentation work in evaluating the solder joint reliability. Further, the effects of microstructure developments on mechanical behavior of Pb-free solders are explored with various Cu concentrations. In order to identify the controlling deformation mechanisms at elevated temperatures, nanoindentation creep tests are employed and measure the stress dependence and activation energy.
9:45 AM - **E4.2
Stress Relaxation of SAC During the Dwell Period of a Thermal Cycle.
Milos Dusek 1 , Davide Maio 1 , Christopher Hunt 1 Show Abstract
1 DEPC, NPL, Teddington, Middlesex, United Kingdom
In accelerated thermal cycling electronics assemblies are subjected to thermo-mechanical low cycle fatigue. The main driver for fatigue is shear strain, which is the resultant of a cyclic temperature profile and different coefficients of thermal expansion (CTE) between components and substrates. The level of stain depends on the level of CTE mismatch in involved materials and causes localised stresses in solder joints. The degradation mechanism resulting from the applied cyclic strain is the development of fatigue cracks. It is these cracks in solder joints that eventually causing electronic circuitry to fail. The natural behaviour of solder joints under a strain is plastic deformation, which relieves concentrated stresses. Reduced stress therefore depresses plastic deformation and minimise fatigue damage. Ideally to gauge stress relaxation, this measurement would be made after an instantaneous application of strain. This is of course not possible, and in the case of thermal cycling the stress relaxation measurement is made at the end of the temperature ramp. Such measurements of stress relaxation expedite the prediction of the required time period to efficiently damage the solder joint in a full cycle. In thermal cycling this necessary time period is called dwell time. It has been established that for SnPb solder that 10 minutes (Ref: 1) is a sufficient dwell time. Currently the dwell time for lead-free solders is predicted to be longer (Ref: 2). The main argument used for this prediction is slower creep rate measured of lead-free solder compared to that of SnPb solder, under the same conditions. The fact is that creep rates are measured at constant load (engineering stress), which is not a case during a thermal cycling dwell where load is naturally relieved. Hence the dwell times necessary for lead-free solder need to be assessed at various temperatures and put into context with prediction models. The FEA modelling has for long time been the only method predicting stress in solder joints, however over simplification of the microstructure and its time dependent behaviour, coupled with dependencies on the model itself, such as mesh size, mean there are limitations to this approach. A measurement machine developed for strain controlled fatigue testing at various temperatures and monitoring true stress permits the validation of the FEA models.(1) IPC standard 9701(2) IPC standard 9701A
10:15 AM - E4.3
Thermal and TEM analysis of Sn Nano-Particles for Use in Nanosolder Applications
Kevin Grossklaus 1 , Eric Stach 1 , Carol Handwerker 1 Show Abstract
1 Materials Science and Engineering, Purdue University, West Lafayette, Indiana, United States
A major obstacle facing the full implementation of Sn-Ag-Cu lead-free solders for many microelectronic applications is the higher melting point of that system compared to the previously used Sn-Pb solders. As an approach to overcoming this obstacle, metallic nanoparticles are being studied as part of an effort to develop a lower melting point solder. By using small particle effects it is hoped that the melting point of some lead-free solder alloys can be lowered close to the melting point of the Sn-Pb eutectic solders. As part of the process of developing a “nanosolder,” the melting and solidification behavior of metallic tin nanoparticles has been characterized by differential scanning calorimetry (DSC) and transmission electron microscopy (TEM). The results of that experimentation will be discussed with respect to modern theory on small particle melting behavior. This is done with the goal of correlating the melting behavior of the nanoparticles as part of a group studied by DSC with the behavior of individual nanoparticles observed by TEM. The impact of particle size and surface condition on the melting behavior of metallic nanoparticles will be discussed with regards to the overall goal of developing a widely applicable and commercially viable nanosolder.
10:30 AM - E4.4
Intermetallic Compounds and Voids Formation in Pb-free Solders on Different Metallizations.
Zhiheng Huang 1 , Paul Conway 1 , Luke Taplin 1 Show Abstract
1 Wolfson School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough United Kingdom
Intermetallic compounds (IMCs) such as Ag3Sn, Cu6Sn5 precipitate within the bulk of Sn-Ag-Cu (SAC) ternary Pb-free solders during solidification after reflow processes. The size, morphology and distribution of those precipitates may vary as a function of the compositions of the solders, the soldering processes, and the sizes and geometries of the solder bumps. In addition, interfacial IMC layers, e.g. Cu6Sn5 or (Cu,X)6Sn5, in which Cu atoms are replaced by X atoms from metallizations, form due to the interdiffusion between the Pb-free solder and the metallizations during soldering or the subsequent aging processes. In some circumstances, a plate-like phase Ag3Sn, or a less frequently observed plate-like phase Cu6Sn5, can grow as a significant fraction of the sizes of the solder bumps from the solder/substrate interface or the edge of the solder bumps.Voids are more frequently observed in Pb-free soldering than in the conventional SnPb soldering. Both materials and processing parameters could contribute to the formation of the voids in Pb-free solders. The quantity, size and distribution of the voids may influence the reliability of the solder joints.As the size of the solder joints in consumer-oriented portable electronic products continues to shrink to meet the ever-increasing consumer needs, the dimensions of the microstructural features such as the IMCs and the voids are becoming comparable to those of the solder joints. As such, the IMCs and the voids are becoming increasingly important as factors that could potentially influence the reliability of the solder joints and thus the whole electronic device, module or system. This paper presents the experimentally characterised features of Pb-free solder joints, namely the quantity, size/thickness, morphology, distribution and orientation, of the IMCs and the voids present in several Pb-free solder/metallization systems. In particular, different microstructural features of the IMCs in different sizes of Pb-free solder bumps are highlighted. The implications of the IMCs and voids on the reliability of the future micro solder joints are also discussed.
11:15 AM - E4.5
Effects of Stress and Strain on Void Growth in Cu3Sn.
Mao Gao 1 , Liang Yin 2 , Pericles Kondos 2 , Peter Borgesen 2 , Eric Cotts 1 Show Abstract
1 Physics and Materials Science, Binghamton University - SUNY, Binghamton, New York, United States, 2 , Unovis Solutions, Kirkwood, New York, United States
The microelectronics packaging industry relies heavily on soldering to electroless or electroplated Ni or Cu. Soldering results in the formation of an intermetallic bond between Sn and either Ni or Cu. Usually these bonds are stronger than the solder, so that loading leads to plastic flow of the latter long before the intermetallic structures fail. However, occasionally this is not the case. A host of different mechanisms may cause bonds to Ni pads to either be weak right after reflow or weaken in subsequent aging. Bonds to Cu pads are usually strong right after reflow, but sometimes they degrade drastically over time. Since nobody currently knows how to control most of these problems they are all subjects of extensive research. Soldering to copper leads first to the dissolution of some Cu and then the nucleation and growth of a nodular Cu6Sn5 intermetallic layer on the pad. This phase is not in equilibrium with Cu, but kinetics limits the initial thickness of a Cu3Sn layer formed between them. A typical reflow may lead to the formation of a 0.5-2µm thick layer of Cu6Sn5 with a 40-50nm layer of Cu3Sn underneath. Both layers grow with time after assembly, and very often a few voids are formed and grow within the Cu3Sn. Their density and sizes are usually small and they are commonly overlooked.On relatively rare occasions the voiding may, however, be strong enough to have catastrophic consequences and tens of millions of dollars have been lost by various companies because of it. In one of the worst cases, a large number of solder joints failed in gentle handling of printed circuit boards after storage for one year at room temperature. Extensive research has led to a thorough documentation of this phenomenon. Notably, the behavior is completely determined by some property of the Cu. Most plated Cu is not nearly as bad as the example above, but in our assessment 5-15% of generally available Cu is poor enough to be a concern for some applications. Unfortunately, high end suppliers using the best available plating bath monitoring tools have not always been able to ensure a consistent performance. In fact, indications are that a single batch of product in between many may suddenly be bad for no discernable reason. We continue to develop faster and simpler tests for ongoing screening of incoming product, but there is a clear need for a more fundamental resolution. The identification of Cu samples of unusually reproducible behavior has allowed for systematic studies of the underlying phenomenon. While it may be reasonable to conclude that such a phenomenon is associated with ‘Kirkendall voiding’, we have argued elsewhere that a consideration of the diffusive fluxes does not support such a mechanism. In the present work we shall show how the voiding does not correlate with the initial Cu purity, grain size, or strain state. Experiments did, however, show an effect of external stresses during aging.
11:30 AM - **E4.6
Correlation of Microstructure and Heterogeneous Failure in Pb Free Solder Joints.
Thomas Bieler 2 , Peter Borgesen 3 , Yan Xing 1 , Lawrence Lehman 1 , Eric Cotts 1 Show Abstract
2 Materials Science and Chemical Engineering, Michigan State University, Lansing, Michigan, United States, 3 , Unovis Solutions, Kirkwood, New York, United States, 1 Physics and Materials Science, Binghamton University - SUNY, Binghamton, New York, United States
There is a clear and present need for a systematic, mechanistic understanding of the thermomechanical behavior of realistic lead free solder joints (for example, the effects of temperature and loading history as well as composition on solder joint integrity). Most Pb free solders have Sn as a main constituent, with various other elements (e.g. Cu and Ag) added to reduce the solder melting point and to enhance wettability. The absence of Pb in these solder alloys causes solder joints to undercool much more (30 to 100C) than SnPb eutectic, causing primary solidification of large intermetallics. The rapid solidification of the Sn phase results in only one to three dominant Sn crystal orientations in a solder joint. The degree of variations in solidification temperature of Sn in Pb free solder joints is also large (10-20C). This results in significant variation in the microstructure of Pb free alloys. The impact of microstructural variability in Pb free solder joints on their resulting properties must be characterized and analyzed before it can be understood or controlled. Thus the variability of ball grid array (BGA), chip scale packaging (CSP), and flip chip components with SnAgCu solder joints were examined as a function of long term thermal cycling. Assemblies were removed for cross sectioning and microstructural characterization at various stages of cycling. The variation in the overall thermomechanical strains with solder joint location across an area array allowed detailed study of a range of temperature-damage combinations. Damage accumulation was examined using optical and electron microscopy techniques. Polarized light microscopy provided delineation of Sn grains, while electron microscopy with EDS provided compositional analysis, and orientation imaging microscopy allowed crystal orientations to be identified. These tools permit a careful examination of 1) formation of intermetallic compounds at solder/metallization interfaces, 2) the size and orientation of Sn grains in the SnAgCu solder, 3) the morphology and number of intermetallic precipitates in the SnAgCu solder matrix, and 4) crack propagation in the solder joint. A clear dependence of the thermomechanical response of solder joints on Sn grain orientation was observed. Solder balls with Sn grains of particular orientation (a-axis perpendicular to the substrate) were observed to fail before neighboring balls with different orientations. Such failures were analyzed in terms of variations in the coefficient of thermal expansion of Sn with crystallographic direction ( along the a-axis it is half the value along the c-axis); joints observed to be damaged had maximum mismatch in the coefficient of thermal expansion between solder and substrate at the joint interface, as well as tensile stress modes during the hot part of the cycle. Localized recrystallization was observed in regions of maximum strain regions caused by differential expansion conditions.
12:00 PM - E4.7
Development of Simple Electrolytes for the Electrodeposition of Pb-free, Sn-based Alloy Solder Films.
Chunfen Han 1 , Qi Liu 1 , Douglas Ivey 1 Show Abstract
1 Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta, Canada
The health hazards posed by Pb in conventional Pb-Sn solders have prompted countries, such as Japan and those from the European Union, to pass legislation prohibiting or restricting the use of these types of solders. One result of this legislation has been the development of a large number of Pb-free alternative solders. Most of these solders are Sn-based; e.g., eutectic Sn-0.7wt%Cu and eutectic and near-eutectic SAC (Sn-Ag-Cu) alloys. The alloy solder films can be electrodeposited from electrolytes containing several additives. The additives help to improve solution stability and deposit morphology. However, the additives often make the electrodeposition process difficult to control and can pose disposal issues.Tin-based, slightly acidic electrolytes, which contain only one additive other than the metal salts, have been developed in our laboratory. These electrolytes (Sn and Sn-Cu) are stable for more than 30 days and can be used at room temperature. Solution compositions and electroplating conditions have been optimized through polarization measurements. Pure Sn, as well as Sn-Cu eutectic and near-eutectic alloy films, with smooth morphology can be deposited, using either direct or pulsed current, at rates exceeding 20 microns/hour.
12:15 PM - E4.8
Electromigration Study of SnAgBi and SnAgBiIn Systems Pb-free Solders.
Albert T Wu 1 , Ming-Shun Chen 1 , Wen-Lin Shih 1 Show Abstract
1 Materials and Mineral Resources Engineering, National Taipei University of Technology, Taipei City Taiwan
The searching of Pb-free solder is an ongoing issues for industrial and academic society. Plenty of Sn based Pb-free systems have been proposed and were being studied intensively. However, none of the Pb-free solder system has satisfied results for completely replacement of eutectic SnPb solder. Among all the choices, SnAgBi and SnAgBiIn solders are two good candidates for their low melting point and good wettability comparing to SnAgCu, which was suggested by NEMI for Pb-free replacement. However, the study on these two solder systems is quite rare. We propose a systematic study of electromigration on SnAgBi and SnAgBiIn Pb-free solder. By changing the composition, the current density, we would like to gather a set of data which could be used for comparison with the well known electromigration behavior of eutectic SnPb and SnAgCu. Hopefully the knowledge could provide better understanding and useful information for the Pb-free community.
12:30 PM - E4.9
Lead-Free Solder Alloys With Small Amounts of Ni, Au and Cu
Chi-man Wu 1 , Ning Zhao 2 1 , Daquan Yu 2 , Lai Wang 2 , Joseph Lai 1 Show Abstract
1 Physics and Materials Science, City University of Hong Kong, Hong Kong, Hong Kong SAR, China, 2 School of Materials Science and Engineering, Dalian University of Technology, Dalian, LiaoNing, China
The interfacial reaction between the molten solder and the surface finishing of the substrate material affects the integrity and reliability of solder joints. In many interconnection processes using solder balls, under bump metallization (UBM) is used. The layer structure in the UBM provides good adhesion to the solder, and facilitates a diffusion barrier to inhibit the reaction between the bump material and the solder. Ni, Cu and Au are commonly used in many UBM systems. It was found that when there was an increase in reflow time and temperature during soldering, Ni, Au and Cu of the UBM could diffuse into and react with molten solder. The result is that, especially at location near the UBM, the composition and the microstructure of the original solder alloy may change. Also, in wave soldering, the solder bath tends to pick up Cu from the copper leads or substrates being soldered. For both examples involving the UBM reaction and the wave soldering bath, the final chemical composition of the original solder alloy could have changed into one that might contain higher concentrations of Cu, Ni and Au. In this respect, it is necessary to study the property and performance of the solders that have new compositions after soldering.The above-mentioned issue can be studied with a basic understanding of lead-free solders with the addition of small amounts of Cu, Ni and Au. The alloys resulting from the additions of Cu, Ni and Au are called multi-component lead-free solder alloys. So Sn-2Cu-0.5Ni, Sn-2Cu-0.5Ni-0.5Au, Sn-3.5Ag-0.5Ni, Sn-3.5Ag-1Cu-0.5Ni and Sn-3.5Ag-2Cu-0.5Ni were made for this study.These alloys were first characterized using differential scanning calorimetry (DSC). X-ray diffraction (XRD) analysis was then carried out. Tensile tests were also carried out on the solder alloys. The solder materials at various conditions were examined with a scanning electron microscope for their microstructure.It was concluded that1. The melting temperatures of all multi-component lead-free solders were very close to those of the original alloys. An exception was found in the Sn-2Cu-0.5Ni-0.5Au alloy such that an endothermal peak was observed on the DSC curves, when compared with that of Sn-2Cu. 2. With the help of XRD, it was found that when the solder composition was Sn-2Cu-0.5Ni, the IMCs formed were Cu6Sn5 and (Cu,Ni)6Sn5. In the case of Sn-2Cu-0.5Ni-0.5Au, besides (Cu,Ni)6Sn5, (Cu,Au)6Sn5 and (Cu,Ni,Au)6Sn5 were also observed. The IMCs formed in Sn-3.5Ag-0.5Ni were Ag3Sn and Ni3Sn4. In both Sn-3.5Ag-1Cu-0.5Ni and Sn-3.5Ag-2Cu-0.5Ni, Ag3Sn and (Cu,Ni)6Sn5 were detected.3. The tensile test results showed that the multi-component lead-free solders have preferable or at least equivalent tensile properties compared with that of the original ones.4. All of the multi-component lead-free solders exhibited ductile rupture at break.
E5: Next Generation RoHS-Compliant Materials and Processes
Thursday PM, April 12, 2007
Room 3005 (Moscone West)
2:30 PM - E5.1
Characterisation of Ferroelectric Domains in KNN
Ralf-Peter Herber 1 , Susanne Wagner 1 , Michael Hoffmann 1 , Gerold Schneider 1 Show Abstract
1 Institute of Advanced Ceramics, Hamburg University of Technology, Hamburg, Hamburg, Germany
The search for lead-free ferroelectrics has been intensified over the last years in order to replace lead zirconate titanate (PZT) ceramics as high-performance piezoelectric materials because of their content of toxic lead. The most promising lead-free candidate seems to be the potassium sodium niobate (KNN) family, which was shown by Saito et al. in 2004*. Since then an increasing number of publications concerning KNN have been published considering processing and electrical properties. In our work we present not only the ferroelectric domain structure of KNN for the first time, but also the first analysis of the domain structure of a ferroelectric material with a morphotropic composition. This is crucial for the understanding of the nanoscale properties and for further improving of the overall piezoelectric properties of this new material.(*Saito, Y. et al. Lead-free piezoceramics. Nature 432, 84-87 (2004))
2:45 PM - E5.2
Deposition and Dielectric Response of CaCu3Ti4O12 Thin and Thick Films for High Energy Density Dielectric Applications.
Jennifer Sigman 1 , Paul Clem 1 , David Williams 1 , Michael Niehaus 1 , Diana Moore 1 , Bruce Tuttle 1 Show Abstract
1 , Sandia National Laboratories, Albuquerque, New Mexico, United States
The current trend of size reduction in capacitor discharge units (CDU) for power supplies and decoupling capacitors for microelectronic circuits will require higher dielectric constant materials that are compatible with standard semiconductor processing techniques. Towards this end, most researchers have focused on ferroelectric materials such as Pb(Zr,Ti)O3 because these materials have shown some of the highest known dielectric constants. However, recently, it has been shown that CaCu3Ti4O12 (CCTO) can possess dielectric constants >10,000 that are relatively temperature insensitive in the range of 100-600K. We, therefore, report on our current efforts with depositing CCTO thin films and thick films by chemical solution deposition (CSD) with carboxylic acid precursors and aerosol spray deposition (ASD). Both deposition techniques allow the quick of deposition quality films with low toxicity. In the CCTO thin films by chemical solution deposition, dielectric constants >2000 and capacitance densities of 4-5 uF/cm2 have been measured. ASD techniques have been used to successfully deposited multilayer thick film stacks with dielectric constants of 2000, loss tangents <3%, and capacitance densities of 0.1uF/cm2. Both approaches demonstrate the successful deposition of a new high energy density dielectric material that is not only Pb-free but also non-ferroelectric. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
3:00 PM - E5.3
BaxSr1-xTiO3 Single Crystal Fibers Grown by LHPG Technique and their Dielectric Properties.
Hathaikarn Manuspiya 1 , Ruyan Guo 2 , Amar Bhalla 2 Show Abstract
1 petroleum and petrochemical college, Chulalongkorn university, Bangkok Thailand, 2 Materials Research Institute, Pennsylvania State University, University Park, Pennsylvania, United States
Single crystal fibers of BaxSr1-xTiO3 (x = 0.05, 0.1, 0.2, 0.35, 0.5) were attempted to grown using the laser heated pedestal growth (LHPG) technique. Crystals were examined by Laue back reflection pattern indicating the high quality crystallinity and <1 1 0> growth direction. The perovskite structure of the crystals was identified by the X-ray diffraction technique. Temperature and frequency dependent dielectric properties were analyzed. The induced dielectric peaks in SrTiO3 were observed and shifted to high temperature as increasing Ba concentration.
3:15 PM - E5.4
Structural and Diffusivity Study of Ba0.9Sr0.1ZrxTi1-xO3 Ceramics Prepared by Sol Gel Method.
Manoj Kumar 1 , Asish Garg 1 Show Abstract
1 Department of physics , IIT Delhi, India, New delhi India
Thursday, April 12New Presenter - E5.4 @ 2:15 pmStructural and Diffusivity Study of Ba0.9Sr0.1ZrxTi1-xO3 Ceramics Prepared by Sol Gel Method. Asish Garg
4:00 PM - E5.5
Development of Organic Stabilizers for Being Used in an Electroless Nickel Plating Bath.
Ke Wang 1 , Liang Hong 1 2 , Zhaolin Liu 2 Show Abstract
1 Chemical & Biomolecular Engineering, National University of Singapore, Singapore Singapore, 2 , National University of Singapore, Singapore Singapore
4:15 PM - E5.6
Lead-free Zero Stress Optic Glass.
Josef Zwanziger 1 2 Show Abstract
1 Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada, 2 Institute for Research in Materials, Dalhousie University, Halifax, Nova Scotia, Canada
Glass is optically isotropic, but when anisotropic stress is applied this symmetry is broken and birefringence is induced. For large scale imaging applications this property is detrimental, and in this case zero stress optic glasses are desirable. Such glasses are deformed by stress but nevertheless show no optical birefringence. It is well-known in the field that lead silicate glass with about 50 mol percent lead oxide content has this property, and this formulation is commercially successful. In order to develop materials compliant with RoHS regulations, we have succeeded in formulating a model that explains both why lead confers zero stress optic response and predicts other glass families with equivalent stress optic response. In the talk I will outline the details of the model and show experimental results on a variety of new lead free zero stress optic glasses.