Damage Mechanics of Electronics on Metal-Backed Substrates in Harsh Environment

204 IEEE TRANSACTIONS ON COMPONENTS AND PACKAGING TECHNOLOGIES, VOL. 29, NO. 1, MARCH 2006

Damage Mechanics of Electronics on Metal-BackedSubstrates in Harsh Environments

Pradeep Lall, Member, IEEE, Mohd Nokibul Islam, John Evans, Jeffrey C. Suhling, Member, IEEE, and Tushar Shete

Abstract—Increased use of sensors and controls in automotiveapplications has resulted in significant emphasis on the deploymentof electronics directly mounted on the engine and transmission.Increased shock, vibration, and higher temperatures necessitatethe fundamental understanding of damage mechanisms whichwill be active in these environments. Electronics typical of officebenign environments use FR-4 printed circuit boards (PCBs).Automotive applications typically use high glass-transition tem-perature laminates such as FR4-06 glass/epoxy laminate material( g = 164.9 C). In automotive underhood application envi-ronments, metal-backing of PCBs is being targeted for thermaldissipation, mechanical stability, and interconnections reliability.

In this study, the effect of metal-backed boards on the in-terconnect reliability has been evaluated. Previous studies onelectronic reliability for automotive environments have addressedthe damage mechanics of solder joints in plastic ball-grid arrays(BGAs) on nonmetal backed substrates and ceramic BGAs onnonmetal backed substrates. Other failure mechanisms investi-gated include delamination of PCB from metal backing.

The test vehicle is a metal backed FR4-06 laminate. Metalbackings investigated include aluminum and beryllium copper.Three adhesives have been investigated for metal backing in-cluding arlon, pressure sensitive adhesive, and pre-preg. The useof conformal coating for reliability improvement has also beeninvestigated. Component architectures tested include plastic BGAdevices, C2BGA devices, quad flat no-lead (QFN), and discreteresistors. Reliability of the component architectures has beenevaluated for hot air solder level and electroless Ni/Au finishes.

Crack propagation and intermetallic thickness data has been ac-quired as a function of cycle count. Reliability data has been ac-quired on all these architectures. Material constitutive behavior ofarlon and pressure sensitive adhesive has been measured using uni-axial test samples. The measured material constitutive behaviorhas been incorporated into nonlinear finite element simulations.Predictive models have been developed for the dominant failuremechanisms for all the component architectures tested.

Index Terms—Metal-back, reliability, solder joint.

Manuscript received May 1, 2004; revised May 11, 2005. This work was sup-ported by the National Science Foundation Center for Advanced Vehicle Elec-tronics (CAVE) at Auburn University. This work was recommended for pub-lication by Associate Editor P. McCluskey upon evaluation of the reviewers’comments.

P. Lall, J. Evans, and J. C. Suhling are with the Department of Mechanical En-gineering and Center for Advanced Vehicle Electronics (CAVE), Auburn Uni-versity, Auburn, AL 36849 USA (e-mail: [email protected]).

M. N. Islam was with the Department of Mechanical Engineering and Centerfor Advanced Vehicle Electronics (CAVE), Auburn University, Auburn, AL36849 USA. He is now with Amkor Technology, Chandler, AZ 85248 USA.

T. Shete was with the Department of Mechanical Engineering and Center forAdvanced Vehicle Electronics (CAVE), Auburn University, Auburn, AL 36849USA. He is now with Tyco Electronics Power Systems, Mesquite, TX 75149USA.

Digital Object Identifier 10.1109/TCAPT.2006.

I. INTRODUCTION

I N PRIOR generations of automotive electronics based onsurface mount technology and laminated substrates; pack-

aging of microprocessor and application-specific integrated cir-cuit (ASIC) chips was typically performed using quad flat packs(QFPs) and plastic leaded chip carriers (PLCCs). As pin countshave increased past 160, plastic ball-grid array (PBGA) compo-nents have become desirable for use in automotive power train(engine and transmission) controller modules. Such modules areexposed to the underhood environment, and typical reliabilityspecifications require the ability to survive a certain number ofthermal cycles in the range of 40 C to 125 C [7], [13]. Thesesystems must also typically meet automotive vibration require-ments while exceeding ten years and 100 000 miles of operation.

In order to limit the effects of the vehicle environment,electronics modules are often separated from the mechanicalsystems which they control. Locations like vehicle “firewalls”and fender wells offer the ability to sink module-generatedheat while reducing exposure to temperatures created by themechanical systems and allowing some access to airflow avail-able under-the-hood. However, these modules are constantlyunder pressure to improve thermal efficiency while reducingcost. The effect of temperature cycling on the reliability ofmicroelectronic packages is the topic of much research [11].

Increasing automotive engine controller functionality, plussmaller and hotter engine compartments are placing greater de-mands on the thermal design of engine control modules. Thetraditional method of adding the heat sinks with the componentson printed circuit boards (PCBs) is very costly and providessignificant thermal resistance to the surroundings. The thermalrequirements have been met with improved heat-sink design,and thermal attachment materials promoting thermal conduc-tivity between the electronics and the module casing, i.e., tra-ditional modules using double sided reflowed components andthermally conductive pads to thermally connect the substrateand the metal housing. This design increases the material cost ofthe module by adding the thermal enhancing material and cre-ates an added manufacturing process for attaching the thermalpad to the housing. While this design meets the thermal and re-liability requirements for the module, the system design is notoptimal.

Metal-backed substrates are an attractive alternative from thethermal point of view. An alternative to the previous design in-volves structurally attaching the metal directly to the substrateproviding a direct thermal path. This design has the added ad-vantage of allowing the metal to act as a module “cover” andemploys a single-path manufacturing process. Utilizing this de-sign significantly reduces the module system cost, and improves

1521-3331/$20.00 © 2006 IEEE

LALL et al.: DAMAGE MECHANICS OF ELECTRONICS ON METAL-BACKED SUBSTRATES 205

Fig. 1. BGA and chip resistors test vehicle.

product quality by limiting the assembly to single-pass reflowexposure. However, attaching the metal directly to the substrate(FR-4) increases the assembly’s effective coefficient of thermalexpansion (CTE).

The metal backed PCB’s disadvantage of high CTE relative tobare FR4-06 leads to increased susceptibility to solder joint fa-tigue failure due to thermal cycling. The focus of this research isto evaluate the system-level issues related to metal-backed sub-strate interconnections reliability, damage mechanism, providemodeling techniques to further investigate design alternatives,reliability prediction, and to provide recommendations to auto-motive system designers.

There have been several studies on electronic reliabilityfor automotive environments. However, none of the previousstudies address the damage mechanics of electronics on metalbacked substrates. Published data on crack propagation hastargeted ceramic BGAs [2], [3], [5] and plastic BGAs onglass-epoxy laminate [6], [8], [9], [14], [15]. In this work, theeffect of metal-backed boards on the interconnect reliabilitywill be evaluated. The crack propagation data as a function ofthermal cycle has been collected. The cycles-to-crack initiationand the crack propagation rate has been benchmarked withthe previous nonmetal backed board studies. Solder crackpropagation as a function of electrical resistance has also beeninvestigated. The relationship between the normalized resis-tance changes with the crack propagation has been investigated.Other failure mechanisms investigated include delaminationof PCB from metal backing. Nonlinear finite element analysishas been used to develop predictive models, which have beencorrelated with experimental data.

II. TEST VEHICLE

Testvehiclewasdesignedtoevaluatetheperformanceofmetal-backed technology, using components and materials availablefor high volume automotive programs. The test vehicle targetedboth the thermal efficiency and the component reliability of awide variety of substrate, attachment, and encapsulant options.The substrates used were standard hot air solder level (HASL)

Fig. 2. Images of Uncycled 15-mm BGA cross section.

finished high glass-transition temperature, glass-epoxy laminate(FR4-06) attached to 2.54-mm aluminum with a variety of ad-hesives. Crack propagation data was acquired on PCBs attachedwith pressure sensitive adhesive (PSA).

The boards contain six trace layers to simulate the thermalmass of a true production board, though all functional traceswere run on the topmost layer. All pads on the board were nonsolder mask defined (NSMD) and had an HASL finish. The re-sistor and BGA daisy chains were routed to plated-through holesat the edge of the board where soldered wire connections couldbe made for use in resistance monitoring during the thermal cy-cling tests. A photograph of an assembled test board is shownin Fig. 1. The Component test matrix of this project is providedin Table I. A typical cross section of an uncycled 15-mm BGAon metal-backed boards with other adjacent material layers areshown in Fig. 2. Components analyzed include I/O counts in


TABLE ICOMPONENT TEST MATRIX

Fig. 3. Reflow profile for metal-backed board.

the range of 193–388, I/O pitch in the range of 0.8 to 1 mm, andpackage sizes in the range of 15 to 27 mm. Resistors including2512 and 1225 have also been included in the study. The testboard contains two 15-mm BGA, one 27-mm BGA, two 16-mmC2BGA, two 2512-chip resistors, and two 1225-chip resistors(Fig. 1).

Since the PCBs were attached to metal backing prior to com-ponent reflow, the reflow profile was modified to accommodatethe thermal mass of the assembly. Belt speed, peak temperature,ramp time, and cooling rates were optimized to ensure reliablereflow process while avoiding package warpage (Fig. 3).

III. RELIABILITY TESTING

Thermal cycling ( 40 C to 125 C) of the assembled testboards was performed in a Blue-M environmental chamber. Thethermal cycle duration was 90 min, with 20 min at each extreme.Typical thermocouple results from under a component on thetest board are illustrated in Fig. 4.

A total of 39 boards were used for the crack propagationstudy. The boards were placed vertically in the chamber, andthe wiring passed through access ports to the data acquisitionsystem. Monitoring of the various daisy chain networks wasperformed throughout the cycling using a high accuracy digitalmultimeter coupled with a high performance switching systemcontrolled by LabView software. Failure of a daisy-chain net-work was defined as the point when the resistance change be-came 10 or higher.

Component failure in this section is assumed to mean failureof the solder joints that form the electrical connection betweenthe component and the circuit board. The resulting failure datawere statistically analyzed using two parameter Weibull models.The standard parameters in such an approach are the Weibull

Fig. 4. Test Board thermal cycle (�40 C to 125 C).

slope , and the characteristic life , which is the number ofcycles required to fail 63.2% of the samples from a particularcomponent of the test matrix. From these values for a partic-ular component configuration, the cumulative failures (percent)after any number of thermal cycles can be predicted. When com-paring the various component reliabilities in this study, we haveused the value of %, the number of cycles necessary to cause1% of the parts in a sample set to fail.

The primary focus of this research was to understand the re-liability implications of using metal-backed PCBs on damageinitiation and progression in component solder interconnects.The metal-backing is intended to provide good thermal perfor-mance without degrading the thermomechanical reliability ofcomponents. This creates the need for the attachment material tohave properties providing very good thermal conductivity whileproviding mechanical decoupling between the laminate and themetal. In this test vehicle, PSA has been selected for both relia-bility and crack growth measurement studies.


Fig. 5. Effect of aluminum metal backing on reliability of 2512 resistors withArlon and PSA adhesives.

IV. RELIABILITY DATA

Reliability data on 2512 resistors has shown that PSA pro-vides higher thermal fatigue reliability than arlon, althoughlower reliability than bare FR4-06 (Fig. 5). All the configura-tions have very similar -slopes, in the neighborhood of 3–4,indicating similarity in failure mechanism, which is solder jointfatigue in this case. The CTE of metal backplanes (e.g., alu-minum has CTE 23.6 ppm/ C) is greater than that of FR4-06.A coupled assembly consisting of metal-backed laminate willhave a higher effective CTE. This greater thermal expansioncauses additional stress in the solder joints of the componentsand degrades solder joint reliability (Fig. 5). Data is consistentwith expected behavior. Closed-form computations indicatethat introducing metal back-planes will cause serious degrada-tion in the life expectancy of the solder joints associated withthe BGA and resistors components [17].

Effect of various supplemental restraints including, selectiveencapsulation and conformal coats has also been examined. Ad-dition of conformal coating improves the characteristic life to2000 cycles but reduces the shape factor to 3.02 from 3.8 forbare-FR406. Addition of selective encapsulation increases char-acteristic life to neighborhood of 2800 cycles but reduces theshape factor to 2.5. A reduction in the shape factor in both of theabove mentioned cases reduces the time-to-1% failure (Fig. 6).

The Weibull reliability data plot for the metal-backed BGAcomponents is shown in Fig. 7. Data indicates, that the relativereliability of the various components is strongly related to thedie size. The 15-mm BGA and 27-mm BGA exhibited compa-rable reliabilities relative to each other, as shown in the plot,of time-to 1% failure. The 27-mm BGA exhibited a time-to 1%failure of 900 cycles. The C2BGA exhibited a time-to 1% failureof 700 cycles. All the packages have very similar -slopes, in-dicating similarity in failure mechanism, which is solder jointfatigue in this case.

V. CRACK GROWTH MEASUREMENT

A. Experimental Techniques

Samples were cross-sectioned at various level of thermal cy-cling. Several sections were made per sample to study the crack

Fig. 6. Effect of supplemental restraint mechanisms on reliability ofcomponents on metal-backed substrates.

Fig. 7. Weibull plot of 1% failure for BGA component.

initiation and propagation in various rows. The cross sectionswere characterized by scanning electron microscopy (SEM)using a JEOL JSM 840 instrument operated at an acceleratingvoltage of 20 kV. Thus, it was possible to obtain topographiccontrast in secondary electron images (SEIs) of the polishedsamples without resort to etching. Backscattered electron im-ages (BEIs) were also used to produce atomic number contrast(phases with a high average atomic number appear bright). Allsamples were imaged as polished.

B. Crack Propagation Measurement

Extensive failure analysis has been performed on the 15 mm,C2BGA, 27-mm BGA, 2512-, and 1225-chip resistor solderjoints, and only the highlights will be presented in this paper.Multiple test boards were periodically removed from thechamber after every 125 cycles. Solder joint cracks have beenmeasured at each temperature cycling data-interval. An averagevalue of crack length across samples has been calculated foreach set and cycle-count. The samples were polished to matchthe measured crack length with the intended pad-diameter forBGA packages (Fig. 8).

Cross section measurements were validated by prying-offsome of the samples. In BGA components, solder joint crackingoccurred simultaneously at both the top and bottom of the solderjoint (Fig. 9). In almost all cases, the primary crack initiated at


Fig. 8. Image of crack initiation, corner ball, 15-mm BGA.

Fig. 9. Simultaneous crack propagation top and bottom of the solder joint, C2BGA.

Fig. 10. Typical image of crack initiation and propagation of 2512 chip resistorsolder joint.

the package interface and later a secondary crack initiated at theboard interface. Once the primary crack initiated, it progressedthrough the joint-area, and caused complete failure. Crackspropagation path was parallel to pad-to-solder interface in allcases until complete separation (Figs. 9 and 12).

The 2512 resistors exhibited a different crack propagationbehavior compared to BGA packages. Cracks initiated under-neath the part for resistors and then either followed a pathdiagonal to resistor termination or grew vertically parallelto the resistor termination. Typical cross section images ofcrack initiation and propagation in 2512 resistors are shownin Figs. 10 and 11. Similar to the BGA packages, the cracklengths in discrete resistors were also measured versus cyclecount. The crack lengths on both terminations of a resistor

Fig. 11. Cross section of a completely cracked 2512 resistor.

Fig. 12. 27-mm BGA solder joints crack propagation or growth onmetal-backed boards at various levels of thermal cycling (�40 to 125 C).

were added together similar to the primary and secondarycrack propagation in the BGA devices.

The 2512s exhibited faster crack propagation compared to1225s consistent with their lower characteristic life. The 2512resistors also exhibited faster crack propagation than the 15-mmBGA, 16-mm C2BGA, and 27-mm BGA packages. Crackpropagation data from the discrete resistors and BGA packageshave been shown in Figs. 13–15. The crack growth data in thisstudy indicates that characteristic crack growth rate stays fairlyuniform during the thermal cycle tests. Since, the primary and


Fig. 13. Crack propagation data for 15-mm BGA, C2BGA, and 1225 chipresistor.

Fig. 14. Crack propagation data for 15-mm and 27-mm BGA thermal balls.

Fig. 15. Crack propagation data for 2512 and 1225 chip resistor.

secondary crack lengths have been added and plotted togetherwith the number of thermal cycles, therefore the crack lengths arelarger than one pad-diameter. Thermal and perimeter balls crackdata are plotted separately. Crack propagation rate for thermalballs in comparison with perimeter balls was higher for 15-mmBGA packages and lower for 27–mm BGA packages examinedin this study (Figs. 13 and 14). The crack propagation rate fordiscrete resistors is significantly higher than BGA packages(Fig. 15).

The number of crack initiation cycles has been found bysolving the regression fit of the data for a crack length of zero.

Fig. 16. Typical intermetallic formation after 625 cycles (40 C to 125 C).

TABLE IICRACK INITIATION AND CRACK PROPAGATION

RATES ON METAL-BACKED BOARDS

Fig. 17. Comparative crack initiation/unit inelastic strain energy betweenmetal/non-metal boards.

Fig. 18. Comparative crack growth rate/unit inelastic strain energy betweenmetal/non-metal boards.

It is seen that the crack initiation only accounts for a range of3% to 16.5% of the fatigue life (Table II). Components with the


TABLE IIIVALUES OF ANAND CONSTANTS USED FOR SIMULATION

highest inelastic strain energy density showed the largest crackpropagation rate and the lowest time to 1% failure. In this case,the 2512 resistor has been found the highest crack propagationrate (Table II).

VI. INTERMETALLIC GROWTH BEHAVIOR

Intermetallic growth was monitored as a function of cyclecount. Intermetallic growth, which generally starts right afterthe solder reflow, is fairly thin prior to thermal fatigue damage.Studies on intermetallic growth of Cu–Sn compounds [10], [12],[16] show that two types of intermetallic compounds are gener-ated namely Cu Sn on the copper pad side and Cu Sn on theeutectic solder side. The metal backed samples in this study ex-hibited formation of thicker intermetallic layers. The loss of Snresulting from its diffusion with Cu result in further localizedcoarsening of the eutectic solder microstructure where Pb-richregions are observed adjacent to the Cu Sn intermetallic layer(Fig. 16). The growth of intermetallic compounds at the solderball/pad interface in microelectronic joints may weaken the jointconsiderably. Intermetallic compounds on metal-backed boardshave been found to be consistent with those on nonmetal backedboards.

VII. CRACK GROWTH CORRELATIONS

The inelastic strain energy density has been correlated withthe measured crack growth data. The viscoplastic strain energydensity values were volume-averaged over the interface ele-ments. The typical interface layers are two element layers of1.0-mil thickness. Experimental data on crack propagation hasbeen mapped with the inelastic strain energy density of compo-nents on metal-backed boards from nonlinear simulations.

Fig. 17 shows the comparison of cycle-to-crack initiationversus inelastic strain energy. Cycles to crack initiation inBGAs on metal-backed boards have been benchmarked w.r.tcrack propagation data in ceramic BGAs on nonmetal backedboards reported in Darveaux [2], [3], [5] and crack propagationdata in plastic BGAs on nonmetal backed boards reportedin Lall, et al. [8]. Darveaux’s damage relationships [2], [3],[5] were derived on CBGA assemblies, with predominantlySMD pads and 62Sn36Pb2Ag solder. In addition to significantdifferences in the crack propagation paths for the two pad con-structions, SMD pads fail significantly faster than NSMD padsin thermal fatigue. The thermal mismatch on CBGAs is much

TABLE IVROOM TEMPERATURE CTES OF METAL-BACKED ADHESIVES

larger than PBGA assemblies. Crack propagation in CBGAs isoften observed predominantly on the package side as opposedto both package and board side for PBGAs. Lall, et al. [8]damage relationships were developed on plastic BGA packageson high laminates with NSMD pads. Data indicates thatBGAs on metal backed boards exhibit lower cycles to crackinitiation for all solder joint inelastic strain energy densities.

Fig. 18 shows the propagation rate per unit inelastic energyon metal-backed and non-metal-backed boards. Crack propa-gation rate in metal backed boards has been measured to be1.7 higher than BGA packages with same inelastic strain en-ergy densities on non metal-backed. Therefore, components onthe metal-backed boards would have lower reliability than nonmetal-backed.

VIII. MATERIALS MODEL

Linear and nonlinear, elastic, plastic, creep, temperature, andtime dependent and time-independent material properties havebeen incorporated in the finite element models. It is well knownthat solder is above half its melting point at room temperaturewhich is why time-dependent creep phenomena dominate solderjoint fatigue.

The thermal fatigue failure of electronic packages is asso-ciated with combined plastic-deformation and creep of solderjoints [1]. The Anand Viscoplasticity model, a standard ma-terial in ANSYS, which has been used by several researchersto model the constitutive behavior of solder, has been used inthis study. This constitutive law has been used by Darveaux[2]–[5] in development of damage relationships. The constitu-tive behavior model has therefore been kept the same to enablea benchmark of damage relationships. Anand’s model is splitinto a flow equation and three evolution equations that describethe strain hardening or softening of the materials (Table III).


TABLE VFEM RESULTS ON METAL-BACKED BOARDS

Fig. 19. 2512 quarter symmetry 3-D mesh plot on metal-backed board.

Fig. 20. Three-dimensional quarter symmetry finite-element mesh for (a)C2BGA assembly and (b) 15-mm BGA.

Flow Equation

(1)

Evolution Equations

(2)

(3)

(4)

IX. FINITE ELEMENT MODEL

Solder ball elements were meshed in ANSYS withVISCO107 elements, whereas all other package materialswere meshed with SOLID45 elements. Figs. 19 and 20 showthe three-dimensional (3-D) quarter symmetry finite elementmodels for C2BGA and 15-mm BGA, respectively, mounted onmetal backed FR4-06 laminate. ANSYS was used to simulatethe thermal cycle fatigue. The quarter symmetry model wasused with mapped finite element mesh that varied from approx-imately 40 000 nodes and 35 000 elements to 90 000 nodes and81 000 elements depending on the complexity of the geometryand the number of solder balls (Table IV).

In most of the cases the corner solder ball is the critical jointand of highest deformation. Correlation of predicted fatigue lifewith accelerated test data is shown in Table V. The damage pre-diction is based on crack propagation relationships presented inLall, et al. [8] with the modifications to crack initiation cyclesand crack propagation rate presented in the current paper.

X. CONCLUSION

Results demonstrate that the metal backing of organic lami-nates (i.e., FR4-06) plays a critical role in meeting the reliabilitygoals for harsh environments. The addition of metal backingincreases the CTE of the substrate and creates potential relia-bility issues for many electronic packages. Thus, there existsa trade-off between an increased path for thermal conductivityand reduction in the overall reliability on metal backed boards.The crack measurements show that crack propagation rate onmetal backed boards is much faster 1.7 than nonmetalbacked boards. Further, BGAs on metal backed boards exhibita lower number of cycles-to-crack initiation than packages onnonmetal backed boards for any inelastic strain energy density.

REFERENCES

[1] L. Anand, “Constitutive equations for the rate-dependent deformation ofmetals at elevated temperatures,” Trans. ASME J. Eng. Mater. Technol.,vol. 104, no. 1, pp. 12–17, 1982.

[2] R. Darveaux and K. Banerji, “Constitutive relations for tin-based solderjoints,” IEEE Trans. Compon., Packag., Manufact. Technol. A, vol. 15,no. 6, pp. 1013–1024, Dec. 1992.

[3] R. Darveaux, K. Banerji, A. Mawer, and G. Dody, “Reliability ofplastic ball grid array assembly,” in Ball Grid Array Technology, J. Lau,Ed. New York: McGraw-Hill, 1995, pp. 379–442.


[4] R. Darveaux, “Solder joint fatigue life model,” in Proc. TMS’97 Annu.Meeting, Orlando, FL, Feb. 10–13, 1997, pp. 213–218.

[5] , “Effect of simulation methodology on solder joint crack growthcorrelation,” in Proc. 50th Electron. Comp. Technol. Conf. (ECTC’00),May 2000, pp. 1048–1058.

[6] J. L. Evans, R. Newberry, L. Bosley, S. G. McNeal, A. Mawer, R. W.Johnson, and J. C. Suhling, “PBGA reliability for under-the-hood auto-motive applications,” in Proc. InterPACK’97, Kohala, HI, Jun. 15–19,1997, pp. 215–219.

[7] Road Vehicles—Environmental Conditions and Testing For Electricaland Electronic Equipment, Std. ISO 16 750, 2003.

[8] P. Lall, N. Islam, J. Suhling, and R. Darveaux, “Model for BGA andCSP in automotive underhood environments,,” in Proc. Electron. Comp.Technol. Conf. (ECTC’03), New Orleans, LA, May 27–30, 2003, pp.189–196.

[9] A. Mawer, N. Vo, Z. Johnson, and W. Lindsey, “Board-level characteri-zation of 1.0 and 1.27 mm pitch PBGA for automotive under-hood ap-plications,” in Proc. Electron. Comp. Technol. Conf. (ECTC’99), SanDiego, CA, Jun. 1–4, 1999, pp. 118–124.

[10] J. W. Morris Jr., D. Tribula, T. S. E. Summers, and D. Grivas, “The roleof microstrucure in thermal fatigue Pb-Sn solder joint,” in Solder JointReliability: Theory and Applications, D. Lau, H. John, and R. Van Nos-trand, Eds. New York: Van Nostrand, 1991, ch. 7.

[11] J. Pang, H. L. Kwang, H. Tan, S. Xunqing, and Z. P. Wang, “Thermalcycling aging effects on microstrucural and mechanical properties ofa single PBGA solder joint specimen,” IEEE Trans. Compon. Packag.Technol., vol. 24, no. 1, pp. 10–15, Mar. 2001.

[12] R. E. Pratt, E. I. Stormswold, and D. J. Quesnel, “Effect of solid-stateintermetallic growth on the fracture toughness of Cu/63 Sn-37 Pb solderjoints,” IEEE Trans. Compon. Packag., Manufact. Technol. A, vol. 19,no. 1, pp. 134–141, Mar. 1996.

[13] Recommended Environmental Practices for Electronic Equipment De-sign, Std. SAE J1211, Society of Automotive Engineers, Nov. 1978.

[14] A. R. Syed, “Creep crack growth prediction of solder joints during tem-perature cycling- An engineering approach,” Trans. ASME, vol. 117, pp.116–122, Jun. 1995.

[15] , “Thermal fatigue reliability enhancement of plastic ball grid array(PBGA) packages,” in Proc. Electron. Comp. Technol. Conf. (ECTC’96),Orlando, FL, May 28–31, 1996, pp. 1211–1216.

[16] P. T. Vianco, J. J. Stephens, and J. A. Rejent, “Intermetallic compoundlayer development during the solid state thermal aging of 63Sn-37Pbsolder/Au-Pt-Pd thick film couples,” IEEE Trans. Compon. Packag.,Manufact. Technol. A, vol. 20, no. 4, pp. 478–490, Dec. 1997.

[17] B. A. Zahn, “Comprehensive solder fatigue and thermal characterizationof a silicon based multi-chip module package utilizing finite elementanalysis methodologies,” in Proc. 9th Int. ANSYS Conf. Exhibition, Aug.2000, pp. 1–17.

Pradeep Lall (M’00) received the B.E. degree in me-chanical engineering from the Delhi College of Engi-neering, Delhi, India, the M.S. and Ph.D. degrees inmechanical engineering from the University of Mary-land, College Park, and the M.B.A. degree from Kel-logg Graduate School of Management, NorthwesternUniversity, Evanston, IL.

He is currently the Thomas Walter AssociateProfessor with the Department of Mechanical En-gineering and Associate Director of the NationalScience Foundation (NSF) Center for Advanced

Vehicle Electronics (CAVE), Auburn University, Auburn, AL. He has tenyears of industry experience. He was previously with Motorola’s WirelessTechnology Center. He has published extensively in the area of electronicpackaging with emphasis on modeling and predictive techniques. He hasauthored and coauthored several book chapters and is an Associate Editor forthe ASME Journal of Electronic Packaging. He is lead author of Influence ofTemperature on Microelectronic and Device Reliability (Boca Raton, FL: CRC,1997). He holds three U.S. Patents.

Dr. Lall received three-Motorola Outstanding Innovation Awards, five-Mo-torola Engineering Awards, four-Publication Awards, and is a Six-SigmaBlack-Belt in Statistics. He is a member of Beta Gamma Sigma and an Asso-ciate Editor for the IEEE TRANSACTIONS ON COMPONENTS AND PACKAGING

TECHNOLOGIES, and the IEEE TRANSACTIONS ON ELECTRONIC PACKAGING

MANUFACTURING.

Mohd Nokibul Islam received the B.S. degree in me-chanical engineering from Bangladesh University ofEngineering and Technology (BUET), Chittagong, in1996, and the M.S. and Ph.D. degrees in mechanicalengineering from Auburn University, Auburn, AL, in2001 and 2005, respectively.

He is with Amkor Technology, Chandler, AZ. Hewas a Graduate Research Assistant in the Departmentof Mechanical Engineering, Auburn University, from1998 to 2005. Prior to joining Auburn University, hewas a Lecturer in the Department of Mechanical En-

gineering, Bangladesh Institute of Technology, Chittagong, from (1997 to 1998.His current research is in the area of electronics packaging reliability. He haspublished many papers in the area of electronic packaging reliability with em-phasis on modeling and other predictive techniques.

Dr. Islam received the Best Paper Award from IMAPS’02. He is a member ofASME and SEM.

John Evans received the M.S. degree in electricalengineering from Auburn University, Auburn, ALand the Ph.D. degree in engineering managementand manufacturing systems engineering from theUniversity of Alabama, Huntsville.

He is an Associate Professor of industrial andsystems engineering and Associate Director of theNational Science Foundation Center for AdvancedVehicle Electronics (CAVE), Auburn University.Before joining Auburn University in 2001, hespent 17 years at Daimler Chrysler Corporation,

Huntsville, AL. During his tenure at DaimlerChrysler, he was a Design Engi-neer, Lead Engineer, Financial Specialist, Electronics Packaging Supervisor,and Technology Manager. His most recent position was Manager of strategicbusiness and advanced technology for DaimlerChrysler Hunstville, Electronics.He also was an Adjunct Assistant Professor (1991–2001) in the Departmentof Industrial and Systems Engineering and a Lecturer in the AdministrativeScience Department, University of Alabama, Huntsville.

Jeffrey C. Suhling (M’01) received the B.S. degree(with distinction) in applied mathematics, engi-neering, and physics and the M.S. and Ph.D. degreesin engineering mechanics from the University ofWisconsin, Madison, in 1980, 1981, and 1985,respectively.

He joined the Department of Mechanical Engi-neering, Auburn University, Auburn, AL, in 1985,where he currently holds the rank of Quina Dis-tinguished Professor. At Auburn, he serves as theDirector of the National Science Foundation (NSF)

Center for Advanced Vehicle Electronics, which is an Industry UniversityCooperative Research Center (IUCRC) with 12 member companies. He hasauthored over 150 technical publications, and has advised over 50 graduatestudents. His research interests include the application of analytical, numerical,and experimental methods of solid mechanics to problems in electronicpackaging.

Dr. Suhling received the the National Science Foundation Graduate Fellow-ship (1980–1983), the Outstanding Mechanical Engineering Faculty MemberAward in 1990, the College of Engineering Birdsong Superior Teaching Awardin 1994, and the College of Engineering Senior Research Award in 2001. He isa member of Phi Beta Kappa, Phi Kappa Phi, and Sigma Xi.

Tushar Shete received the B.E. degree in mechanical engineering from PuneUniversity, Pune, India, and the M.S. degree in mechanical engineering fromAuburn University, Auburn, AL, in 2004.

He is a Reliability Engineer with Tyco Electronics Power Systems, Mesquite,TX. Prior to that, he was a Research Assistant in the Department of MechanicalEngineering, Auburn University, working in the field of electronic packagingreliability under harsh thermomechanical environment, with emphasis on mod-eling and experimental techniques.

Damage Mechanics of Electronics on Metal-Backed Substrates in Harsh Environment

Documents

Transcript of Damage Mechanics of Electronics on Metal-Backed Substrates in Harsh Environment