Modeling Of Molded Electronic Package Warpage .

2018CONFERENCEMELAKAMALAYSIAI ET M38thInternational Electronics ManufacturingTechnology (IEMT)Conference2018Modeling of Molded Electronic Package WarpageCharacteristic with Cure Induced Shrinkage andViscoelasticity PropertiesWei Keat LohIntel Technology Sdn BhdMalaysiawei.keat.loh@intel.comChih Chung HsuCoreTech System (Moldex3D)Taiwanjimhsu@moldex3d.com2018-09-05Ron W. KultermanFlex Ltd, Austin Tx.USAron.kulterman@flex.comHaley FuiNEMIShanghaihaley.fu@inemi.orgrestricted

iNEMI: Warpage Characteristics of Organic Packages,Phase 4Phase 4 (2016-2017)Phase 3 (2015-2016)Phase 2 (2013-2014)Phase 1 (2011-13)iNEMI TechnologyRoadmap 2013 Identify packagewarpage as keychallenge Literature survey Establish metrologycorrelation betweensites and increaseawareness Reaching out toindustry forcomponent donation (Forming & Storming) Establish currenttechnology packagedynamic warpage(POP, PBGA, FCBGA) Measurementprotocol (effect of “AsIs”, “Bake” andMoisture ExposureTime (MET)) andsample size needed. (Norming)2018CONFERENCEMELAKAMALAYSIAI ET M Continue establishingtechnology packagedynamic warpage (allkind) Dynamic warpagemeasurementmetrology assessment. (Performing) Continue establishingtechnology packagedynamic warpage ofnew packagetechnology (all kinds) Leverage previouseffort to study theimpact of LowTemperature Solderon dynamic warpagerequirement Collaborate withsimulation software toderive modelingapproach in betterpredict the dynamicwarpage

Project Scopes in Phase 4 SOWCharacterization of latest packaging technology Characterize emerging electronic packaging technology dynamic warpage behavior to develop a betterunderstanding of the current development of package construction and material development as listed in Table I. Silicon Interposer and Embedded Silicon Bridge with different package stiffeners and constructions forHeterogeneous Packaging Solution Next generation of POP packages that leverages wafer level process for package construction which includePanel and Wafer level molding. These also include the use of new fiber reinforced mold material for warpagecontrol. System In Package/Multi Chip Package (BGA) with different configuration and layout. Embedded Package (embedded silicon, actives and passives)Assess the impact of Low Temperature Solder (LTS) on package dynamic warpagerequirement To assess the impact of lower temperature solder on package warpage for those packages collected in Phase 2 and3. To establish the risk level based on package technology with respect to warpage only.Assessment the use FEA in optimizing package warpage Establish modeling optimization approach and tools requirement to enable higher accuracy prediction technique Compare modeling with experiment data and identify potential gaps for further development in FEA technique32018CONFERENCEMELAKAMALAYSIAI ET M

ContentBackground and motivationModeling approach and governing equationsSensitivity StudiesSummaryDisclaimer: The work here covers some feasibility study whichcollaborated with iNEMI and we are not endorsing any software inparticular.2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M4

Multi-physics in Assembly ProcessIncoming MaterialsElectronic Package Assembly Package Design –manufacturing variation Substrate warpagevariation Die Warpage Raw materials (differentmaterial characteristics)Chip Attach Mass Reflow TCB NonConductivePaste or FilmEncapsulation &Molding Underfill Injection orCompressionMolding MoldedUnderfillSMT Assembly Process Steps Temperature exposure Time effect (Visco andcreep) Mechanical loadingLid Attach (ifapplicable) Mechanicalloading Adhesive IAI ET M ReflowTemperature/TransientThermal exposure anddistribution Mechanical interaction Board level SJR YieldBall Attach Mechanicalloading ReflowtemperatureTest and FinalInspection Mechanicalloading Coplanarity/Warpage

Current Modeling Approach is itAdequate?Create FEAGeometry,Meshing and applyconstraintboundariesconditionand includeincomingvariationsApply linear andvisco-elasticitymaterialconstitutiveproperties andinclude differentprocessing steps ofhow materials areadded to theconstruction of thepackageCompute themodel from theinitial stress freeconditionHow to accountfor the individualresidual stresses ofthe componentsduring assemblyPost processing toextract predictedpackage warpagefor risk assessmentwith account forprocess, materialand incomingvariation.Significant effort needed to account for multi-physics interaction inpackaging assembly for warpage predictionMany commercial available modeling tools are still working to refinethe modeling needs.2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M6

Moldex3DFlow viscosityCuring kineticsPVTCViscoelasticityFillingStage 1: Compute for1. In mold filling andcuring stage2. Induce residual stressand shrinkage basedon degree of cure andpressureCuringStage 2: Compute for1. Post mold cure stage2. Subsequent Thermalexposure stagePost mold cure deformationCoupled mold flow, curing kinetics, visco-elasticity and shrinkageand conjugate heat transfer in one environment2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M7

Material Characterization in Moldex3DLaboratoryH5DRMCR 502TMA 4000Rotationandoscillationtests forviscoelasticpropertiesCoefficient ofthermalexpansionDSC 8500DMA csDynamicsmechanismanalysis forrelaxationmodulus2018CONFERENCEMELAKAMALAYSIAI ET MThermalconductivityPT-6800Specificvolume withcuring(PVTC)

Test Model for DemonstrationMold flow directionPackage size: 12x12mmPackage thickness: 0.61mmMold thickness: 0.45mmMaterialsSiliconModulus (MPa)172800CTE 1706.3Mold2016-09-08 SIAI ET MPVTC9

Mold Flow Governing EquationsMass balance ( u ) 0 tDensity, ρMomentum balanceRunnerGateTransfer potEMC Du τ p gDtCavityHeat capacity, CpEnergy balance T u T k T : S t Cp CAE procedureFlow/Cure behaviorThermoset conversion balanced n k1 k 2 m (1 )dt(Relationshipbetween τ and uor viscosity, η)Thermal conductivity, kConversion rate, αTemperature-pressure-cure dependent volume, V(T,Post mold cure analysisRelaxation Modulus, G2018CONFERENCEMELAKAMALAYSIAI ET MP, C)

Curing Kinetics Mold MaterialPropertiesThe combined model to investigate the curing kinetics of the given moldbecause of its ability to accurately predict the experimental data. Thecombined model can be expressed as follows:100.0d n k1 k 2 m (1 )dt E1 k1 A1 exp RT )80.0Conversion(%)(90.0T 200oC70.0T 175oCT 150oC60.050.040.030.0 E k 2 A2 exp 2 RT 20.010.00.00100200300400Time (s)2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M11

Structural Stress Governing EquationsStress balance σ(t ) 0,1 C( PVTC ), ( U UT )2Compute for thermal and cure induced strains can be expressedin the following PVTC equation for epoxy: PVTC VS ( P,V , T , C )2018CONFERENCEMELAKAMALAYSIAI ET M

PVTC Mold Material PropertiesThe two domain modified Tait model isused to formulate the specified volume ofresin as below:Molding pressure 0.1MPa111 (1 ) V VuncuredVcuredMolding pressure 10.0MPaMolding pressure 0.1MPaMolding pressure 10.0MPap Vcured / uncured V0 1 C ln 1 B b1S b2 S T , if T TtV0 b1L b2 LT , if T TtUncured stateCured state b3S exp( b4 S T ) , if T TtB b3L exp( b4 LT ), if T TtTt b5 b6 PT T b5 , C 0.08942016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M13

Viscoelasticity Governing Equations C is a tensor and function of relaxationmodulus E(t, T, ) curve:C Cijkl E (t )( ik jl il jk ) E(t ) ij kl(1 )(1 2 )2(1 )Viscoelasticity of Mold Materials3.50E 11 Relaxation stiffness in VE analysis:3.00E 11–log aT E(t) (g/cms2)E (T , C, t ) E (Tref , Cref , t a )where the temperature and curing shift factor2.50E 112.00E 11 C1 (T Tref )Tg 130 oCTg 150 oCTg 170 oCC2 T Tref1.50E 11log aC a1C n a2C ( n 1) . a01.00E 111.0E-20 1.0E-17 1.0E-14 1.0E-11 1.0E-08 1.0E-05 1.0E-02 1.0E 01 1.0E 04 1.0E 07 1.0E 10 1.0E 13Time (s)2018CONFERENCEMELAKAMALAYSIAI ET M

Loading Boundaries ConditionInjectionmolding tocooling2016-09-08 restrictedPost MoldCure2018CONFERENCEMELAKAMALAYSIAI ET MSubsequentSMT reflowsimulation15

Mold Flow SimulationThe mold flow front encapsulates the cavity uniformly andhas the potential to trap voids underneath the chip in C4region.2018CONFERENCEMELAKAMALAYSIAI ET M

Effect of Detail vs ARPAGE (μM)Homogenous modelingDetail ModelSubstrate detail-4-6-8-10-12HomogeneousDetail Model – copper and dielectricpatterns are taken into accountHomogeneous – the effective mechanicalproperties are considered2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIATIME, T (S)The impact of the detail model for thiscase is 3% (not significant) due to smalland thin designI ET M17

Effect of PVTC Properties0500010000150002000025000300000-5WARPAGE (μM)-10-15-200% mold shrinkage1% mold shrinkage1.5% mold shrinkage-25-30-35PVTC properties captures thevolumetric change as a function oftemperature for uncured and curedmold2016-09-08 restrictedTIME, T (S)The impact of mold shrinkage issignificant. Typical FEA doesn’taccount for chemical shrinkage2018CONFERENCEMELAKAMALAYSIAI ET M18

Dynamic Warpage of Package30020TEMPERATURE, T ( C)250200Evironment Temperature, T ( C)10Package Temperature, T ( C)5Warpage0-5150-10-15100WARPAGE 00TIME, T (S)During the reflow cycle, the package warpage changes as a function oftemperature. The package started off concave (-ve) at room temperature, thenmorph to convex at elevated temperature2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M19

Effect of Tg as a function of curingdegree2000180-2Tg (Glass transition temperature)-4140-6120-8100-10Tg ( )160DiBenedetto (Tg1 Tg 0 )Tg ( ) Tg 0 1 (1 ) 800%10%20%30%40%50%60%70%80%90%log acac (Curing shift factor)Vogel Hac exp R 11 T ( ) T g1 g-12100%Curing degree (%)“DiBenedetto equation ” is usually adopted for modeling the Tg as a function ofconversion rate of reactionthe curing shift factor, ac, utilizes Vogel model2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M20

Effect of Tg0500010000150002000025000300000-5WARPAGE (μM)The evolution of Tg as a function ofcuring degree1% mold shrinkage without Tg effect-101% mold shrinkage with Tg effect-15-200.49Sensor(130)-25Cured(130), 0.1MPa0.49Uncured(130), 0.1MPa-30Sensor(170 degree C)Cured(170), 0.1MPa-400Specific volume (cm3/g)-35TIME, T (S)500010000150002000025000300000-10Tg 130Uncured(170), 0.1MPa0.48 V1 300.48 V1 700.470.47Tg 150WARPAGE (μM)-20Tg 1700.4625-3075100125150175200Temperature ( C)-40Lower Tg gives rise to higherwarpage magnitude-50-60-7050TIME, T (S)2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M21

Summary Capturing the packaging assembly process involved requires significant effort inrefining the assumptions, the numerical approach, and the analytical approach.The sensitivity model studied here only covered the molding process by capturingthe mold flow and structural analysis based on fluid dynamics, PVTC, viscoelasticitywith corresponding analytical equations and non-linear properties.The in-mold cure shrinkage which was represented by PVTC data, played asignificant role in determining the final package warpage.A higher cure shrinkage can yield a higher package warpage.The impact of mold cure conversion rate on Tg and PVTC can change thecharacteristic of package warpage from uncured to fully cured mold.The modeling demonstrated that a higher cured Tg mold gives rise to a lowerpackage warpage.The additional reflow cycles included in the model increases the package warpageas the result of changes in the viscoelasticity and stress relaxation.The next logical step is to evaluate other modeling approaches and validation thatcan mimic the actual assembly process as close as possible to enhance theprediction capability.2016-09-08 restricted2018CONFERENCEMELAKAMALAYSIAI ET M22

Package Technology Dynamic WarpageMeasurementGenerous Donation of Samples from IndustryNew variance - MCPFocusing on new and advance package technologies232018CONFERENCEMELAKAMALAYSIAI ET MPictures are mainlyfor illustration only

System In Package/Multi Chip Package (BGA) with different configuration and layout. Embedded Package (embedded silicon, actives and passives) Characterization of latest packaging technology To assess the impact of lower temperature solder on package