INEMI Cu Wire Bonding Project - ICEP2013
iNEMI Cu Wire Bonding Project Update on Technology InvestigationReliability Test with Concerns1Masahiro Tsuriya1 , Andy Tseng2, JoonSu Kim3,m.tsuriya@inemi.org, International Electronics Manufacturing Initiative (iNEMI)2Andy.Tseng@aseus.com, ASE (U.S.) Inc.3JoonSu.Kim@amkor.co.kr, Amkor Technology Korea, Inc.ABSTRACTThe adoption of copper wire bonding is picking up pace across semiconductor product families and systemapplications. iNEMI launched a project in 2010 to survey the industry and determined key concerns of the membercompanies. Based on the outcome of the industry-wide survey, the reliability project has been launched to assessthe effects of leadframe and organic substrate packaging with multiple Cu wire types and process conditions. In thistalk we will provide an overview of the technical project with bondability testing result and some preliminary reliabilitytesting results are discussed. Two bonding wire manufacturers provide the copper wire and palladium coated copperwire used for evaluation. Bill of Material (BOM) is chosen from the recommendation from the assembly house.Process parameters of EFO current, bonding force and time are determined based on the preliminary processcharacterization study. Using these process parameters, bondability tests are performed for wire pull strength, ballshear IMC coverage on both BGA and QFN. The measurement data shows the similar results with acceptable levelwhich prove that process characterization for all wire types can be established with acceptable level. Reliability testis underway and Temperature Cycling and HAST test for BGA is discussed in this paper.1. INTRODUCTION2. OUTCOME FROM PHASE1 PROJECT1Copper (Cu) bond wires are increasingly being used fora wide variety of components from consumerapplications to high-reliability electronic products, whichdrive cost merit as well as electrical and thermalperformance improvement. Despite these positiveimpacts of the improvements, reliability still needs to becollectively assessed by the industry in a quantitativemanner. Furthermore, for component qualificationpurposes, standard reliability test methods and durationsestablished for Gold (Au) wire device need to review forCu wire bonded devices.58 individuals from over 40 global leading companiesresponded to the survey, covering a wide spectrum ofthe industry, ranging from material suppliers to OEMs.Main concerns for OEMs about factors that would slowerthe implementation progress were reliability andunproven historical performance. Biased HAST wasrated as the most difficult reliability test to pass. Testduration, temperature range, and whether currentJEDEC standard can properly detect corrosion were themajor concerns.iNEMI launched a collaborative project on Cu WireBonding Reliability in 2010 to firstly have conducted asurvey on the industry-wide conversion status, as well askey reliability concerns for Cu wire bonding. Based onthe findings from this survey, the team moves to nextstep in late 2011, which project is to understand theeffects of key factors such as packaging materialselection and process condition for environmentalreliability performance.Two package types are selected for the evaluation,which is QFN (Qual Flat Non-Leaded) and BGA (BallGrid Array). Two bonding wires are used, which are bareCu wire and Palladium (Pd) coated Cu wire, and thewires used are supplied from two companies. Theprocess characteristics studies were performed tooptimize process condition, and process parameters aredetermined for this project experiment. The acceleratedenvironmental tests are performed on temperaturecycling (TC), high temperature storage (HTS) and highlyaccelerated stress test (HAST) with 5 different conditionsand duration for both BGA and QFN.Fig.1 Concerns about Cu wire Bonding from SurveyOverall, there was strong consensus that the reliabilityperformance of bare Cu and Cu/Pd wire bonding neededto be further evaluated with experimental work.
Table3 Mold Compound Property for BGA3. EXPERIMENTAL3.1 Test Vehicle Information:Unit- Bonding Wire: Table1 shows the key characteristics ofbonding wires. Two bonding wire types of bare Cu andPd coated Cu (hereinafter referred to as "Cu/Pd") areused for the evaluation. These wires are supplied by twowire manufacturers (shown as Co-B and Co-C in thetable). 4N Gold (Au) wire is used as control, which Auwires are chosen from the proven quality history from thevolume production at the assembly house.Table1 Wire Type for EvaluationCo-BCuCo-CCu/PdCu/PdWire Dia. (mil)0.70.80.70.8Breaking Load(gf)3-84 - 103-84 - 10Elongation (%)7 - 157 - 159 - 1511 - 178- 148– 1485 - 9585 - 95 95 - 105 90 - 105 50- 80Hardness (Hv)WireResistivity (RT)μΩԨ1.71.71.71.70.70.84.5 - 8.5 5.5 - 10.5Die Pad OpeningDie Pad PitchPackage SizeMax Wire LengthBall Size/ Pitch-510 /degC210 N/mm2TypeC1.04.023012.0Table4. Wire and Mold Compound for BGA CellsLegWire typeWireMakerBGA1AuMoldFormingCompound gas typeFlow rate(LPM)Co. Atype ANoNoBGA2 Bare CuCo. Btype BN2H20.5BGA3Co. Ctype BN2H20.5Cu/PdBGA4 Bare CuCo. Btype CN2H20.5Cu/PdCo. Ctype CN2H20.550– 80BGA6Cu/PdCo. Ctype BN20.51.9BGA7Cu/PdCo. Btype BN2H20.51.9Table2. Package Outline used for evaluationDie .24.31808.0BGA5- Daisy Chain Die: Table2 shows the test die andpackage information. Daisy chain die is used forpackage assemblies, which packages are 32 leads QFNand 384 i/o BGA. This wafer metallization of Al/TiNthickness is 500/6000Å(angstrom).QFN 56ld2.5 x 2.5 x0.02mm0.038mm0.045mm5.0 x 5.0 x0.65mm1.482mm- /0.50mmCTETypeA0.83.42402.40BGA 384i/o5.1 x 5.1 x0.02mm0.049mm0.06mm14 x 14 x1.16mm3.75mm0.04/ 0.65mm3.2 Design of Experiment:- BGA:0.8mil wire is bonded on aluminum (Al) plated pad with49um pad opening and 60um pad pitches of daisy chaindie and electrolytic Ni/Au plated on substrate bondingfingers. Cu plating thickness is 18.88um average forsubstrate Cu traces and Ni/Au plating thickness is7.928/0.486um. K&S Icon bonder is used. Au wire isbonded based on the qualified process parameters, andBare Cu and Cu/Pd wire bonding parameters aredetermined by preliminary process characterizationstudy. 95%N25%H2 forming gas is used except one cellto see any difference from 100%N2 gas. Forming gasflow rate is fixed with 0.5 LPM for all cells. Green moldcompound is used for all cells, and different mold resinsare used for Au, Bare Cu and Cu/Pd wire bondedpackage encapsulation. Table3 shows the key propertyof BGA mold compound and Table4 shows the matrix ofcell material information.- QFN:0.7mil wire is bonded on Al plated pad with 38um padopening and 45um pad pitches of daisy chain and silver(Ag) plated on the leads of copper leadframe. Thisleadframe is etched copper metal with Ag plated of70 350 um thickness. K&S Icon bonder is used.95%N25%H2 forming gas is used for bare Cu wirebonding and both forming gas and N2 gas are used forCu/Pd wire bonding. Forming gas flow rate is chosenbetween 07 – 0.9LPM. QFN#5 for Cu/Pd is set with 0.3 –0.5 LPM. Only one mold resins is used for all cells. Itskey characteristics are CTE alpha1 with 0.9 10-5/degC-5and apha2 with 3.9 10 /degC by Tg 130degC. FlexuralModulus at 25degC is 24000N/mm2 and 750N/mm2 at260degC. Table5 shows the Cell matrix.Table5. Wire and Mold Compound for QFN CellsLegWiretypeWiremakerMoldCompoundQFN1AuForming Flow rateGas(LPM)Co. DType DN/AN/AQFN2 Bare CuCo. BType DN2H20.7 0.9QFN3Cu/PdCo. CType DN20.7 0.9QFN4Cu/PdCo. CType DN2H20.7 0.9QFN5Cu/PdCo. BType DN20.3 0.53.3 Bonding Parameters:- BGA:Electric Flame-off (EFO) current at 1st bonding for BareCu wire is approx. 10% higher than Au wire bondinglevel and approx. 20% higher for Cu/Pd wire bonding.Same EFO current is used for both Bare Cu and Cu/Pdwire bonding which is approx. 25% higher than Au wirebonding.stBonding force at 1 bonding for Bare Cu and Cu/Pd isthe same, which is higher by approx. 80% and half forceat 2nd bonding, comparing with Au wire bonding. Bonding
Table6 Measurement Data of BGA Bondingcapillary is PECO granular finish type and it haschanged every time when new cell is started bonding.- QFN:EFO current for Bare Cu wire is used the same range inN2H2 forming gas and approx. 20% higher for Cu/Pd wirebonding at 1st bonding, but slightly different value is usedndbetween N2 gas and forming gas. 2 bonding EFOcurrent is half for both Bare Cu and Cu/Pd wire bonding,stcomparing with Au wire bonding. Bonding force for 1ndbond is the same range, but approx. 30% higher at 2bonding for both Bare Cu and Cu/Pd wire bonding butslightly lower force for Bare Cu wire. Bonding capillary isPECO granular type and it has changed every timewhen new cell is used for bonding.Ball Size Ball Shear Wire Pull Stitch Pull(um)(g)(g)(g)Cu wire/CriteriaMinAuBare CuCu/PdCo-C4. WIRE BONDING STUDY RESULT:4.1 Process Monitoring Plan:Bill of Material (BOM) is chosen from the qualifiedmaterials at the assembly house. These include theleadframe material for QFN and substrate for BGA.Below-mentioned are key monitoring items at eachassembly process step.Cu/PdCo-B38-42um 13.5g 3g ax40.917.928.507.17Avg39.416.408.276.45- Die Saw: Chipping and silicon dust on bond pads arevisually checked to make sure that die surface conditiondoes not impact the wire bondability.- Die Attach: The items of bond Line thickness, Die tilt,epoxy void, filet height and die shift/shear are measuredand inspected with no issues. Since items are impactedbondability or final package reliability, these verificationsare made to reduce the potential items to influence thepackage quality.- Wire Bond: Bonding parameters of bond temperature,st1 bond force and ultrasonic power, bonding time, wirefloor age, and capillary age are recorded during theassembly process. The items of wire pull, ball shear, ballndsize and ball thickness, cratering, pad splash, 2 bondwire pull and Intermetallic compound (IMC) coverage aremeasured. These items will be discussed by this paper.Fig.2 1st Bonding IMC Coverage on BGAFig.3 shows the appearances on the wire bonded on diepad and substrate bonding fingers. No abnormalities arefound for all bonding.- Encapsulation: Wire sweep is checked by x-ray.- Singulation & Open Short test: All units are inspectedand only good units are sent for reliability testing.4.2 Key Characteristics Measurement:- BGATable6 shows the measurement results of Ball Size, BallShear, Wire Pull at 1st bonding on die pad and 2ndbonding on substrate bonding fingers. There are nosignificant differences on all items among these 4different bonding wires.Fig.2 shows IMC coverage on the die pad. More than80% coverage ratio is used for the acceptance range.Company-B Cu/Pd wire bonding is shown out of theacceptance range. This is resulted in smaller ball sizeand ball shear shown in table6. However wire pullstrength is no significant difference with other bodings.Fig.3 Bonding Performances on BGAFig.4 shows the cross section of 1st bonded ball whichare taken after the wire bonding process. Max 1.5umaluminum splash was found on Bare Cu wire bonding.
stndBonding appearance on the 1 bond on die pad and 2bond on the leadframe bonding tip are not significantlydifference with Au wire bonding.Fig.4 Cross Section of Smashed Bonding on BGA- QFNTable7 shows the Ball Size, Shear, Wire Pull at 1stbonding on die pad and 2nd bonding on leadframe. Thereare no significant differences on all items for these 4different bonding wires.Fig. 6 Bonding Appearance on QFNFAB (free air ball) is not asymmetry for all bondingcondition which proved EFO condition for Bare Cu andCu/Pd wire bonding. No abnormalities are also found atthe bonded layer by cross section observation.Table7 Measurement Data for QFNBall Size Ball Shear Wire Pull Stitch PullCu wire/ CriteriaAuCo-BCu/PdN232-36um 8g 3g 24Avg34.515.638.364.56Bare CuN2 16.059.275.02Avg34.914.908.614.30Co-CCu/PdN2 6.588.954.56Fig.5 shows 1st bond IMC coverage on the die pad. Thecoverage ratio of all Bare Cu and Cu/Pd wire bonding ismet with the acceptance range.Fig. 7 FAB and Cross Section Photo5. SURFACE FINISH BONDABILIYTY STUDY:Bondability study was additionally performed on theelectroless plating surface on the BGA substrates, whichare ENEPIG (Electroless Nickel Electroless PalladiumImmersion Gold) and ENEP (Electroless NickelElectroless Palladium). Two chemical manufacturersplated on the BGA substrates. Two different ENEPIGthickness are tested, which one is Ni 6.642um, Pd0.211um and Au 0.113um and another is Ni 3.627um, Pd0.818um and Au 0.310um in average. ENEP thickness isNi 5.844um and Pd 0.233um and another is Ni 3.588umand Pd 0.8666um.1st trial with same bonding condition of NiAu finishshowed that stitch bonding on both of ENEPIG andENEP was not made within the acceptable level. Afteradditional characterization of bonding condition, thereare buildings some unit for bondability Test.ENEPIG surface by Cu/Pd wire bonding shows theacceptable range with additional scrubbing mode. Whencomparing the wire type, Cu/Pd bonded wire is slightlyhigher pull strength on ENEPIG, and Bare Cu bondedwire is higher on ENEP. There are no significantdifference between the chemical manufactures’ platedsurface and ENEPG and ENEP.Fig. 5 1st Bond IMC Coverage on QFN6. RELIABILITY RESULT:6.1 Reliability Test Matrix:
All units of BGA and QFN are performed open/shorttesting after assembly process and these units arescreened by x-ray inspection for reliability testing. Theseunits are baked under 125degC for 24 hours beforepreconditioning test of 30degC and 60RH% for 192hours which is Jedec level3. Units are sent for solderreflow with 260degC for 3 times.Same Reliability test are performed for BGA and QFN,which 5 different condition of HAST (130degC/85%RH,110degC/85%RH, 85degC/85%RH, 130degC/55%RH,110degC/55%RH), HTS (175degC storage temperaturecondition) and TC (-55degC/125degC, 2x/hours). Thereason for conducting 5 HAST conditions is to identifywhich can prove the detection of corrosion.6.2 Reliability Test Result:Reliability testing is still ongoing, so this paper providesthe result for TC -55degC/125degC and HAST130degC/85%RH with 5.5V bias test results for BGApackage. Team continues for reliability tests.- Temperature Cycling for BGATC (-55degC/125degC, 2x/hours) is performed for 1500cycles. BGA3 is survived until 1500 cycle and all otherlegs were failed at 500 cycles. Failure analysis isunderway to analyze.Table12 TC -55/125degC 5.5V bias Result for 860/840/1120/811500cy0/620/850/860/840/1120/81- Highly Accelerated Stress Test for BGA:HAST 130degC/85%RH test is performed for 384 hours.Failures are found on BGA# 2 to 5 at 384 hours, but noopen failure at Au wire bonded package and BGA#7.BGA7 is used same mold resin and forming gas withBGA3, and only Cu/Pd wire type is different amongthese cells. Failure analysis is underway to analyze.Fig14. HAST 130degC/85%RH Test Result for 0/330/340/340/96384hrs0/241/331/332/341/340/967. ConclusionAfter process parameters are characterized, bondabilityof Bare Cu and Cu/Pd bonding wires are evaluated onthe various plating surface finishes on Cu metalleadframe and organic substrates. It concludes thatprocesses can be optimized to meet the processmonitoring criteria. Following are key findings from theexperiments.- Surface Finish: Cu/Pd wire is bonded on ENEPIGsurface finish and bare Cu wire is bonded on some cellsof ENEP surface finished. But bonding process windowis very narrow and bonding machine easily stopped dueto no stick on the substrate bond fingers if theparameters are not optimized.- Process parameters: For BGA, same parameters forstEFO current, Force and Time can be used except 1bond for Bare Cu wire with slightly lower EFO currentand longer bonding time. For QFN, Bare Cu bonding isrequired lower EFO current and higher bonding Force,comparing with Cu/Pd wire bonding.- Bondability: Process parameters can be determined bythe feasibility study and these studies lead to nosignificant difference of process monitoring results.- BGA Reliability Test: TC -55degC/ 125degC resultshows that BGA3 (Type C Cu/Pd) is not failed for 1500cycles, and BGA7 (Type B Cu/Pd) is survived HAST130degC/85%RH test for 384 hours. These cells usesame mold resin with forming gas environment.8. Next StepTeam has been working on continuous Reliability testingfor both BGA and QFN, and will perform the failureanalysis to identify the defects causes.Team plans to move to phase3 to evaluate the effects ofgreen mold compound to Cu wire bonding reliability.AKNOWLEGEMENTThe authors acknowledge the great contributions ofCopper Wire Bonding Reliability Phase 2 project teammembers, who are; Amkor, ASE, Atotech, Cisco, DowChemical, Heraeus, Huawei, IBM, IST, Lenovo, NipponMetal, Plexus, and STATSChipPAC. We especially thankIntel Malaysia reliability lab team, who voluntarilyperforms Temperature Cycling and High TemperatureStorage test.Reference[1] Grace O'Malley, Peng Su, Haley Fu, Martin Bayesand Masahiro Tsuriya, “Current Industry Adoption ofFine-Pitch Cu Wire Bonding and Investigation Focusat iNEMI” EMPC 2011
comparing with Au wire bonding. Bonding force for 1st bond is the same range, but approx. 30% higher at 2nd bonding for both Bare Cu and Cu/Pd wire bonding but slightly lower force for Bare Cu wire. Bonding capillary is PECO granular type and it has changed every time when new cell is used for bonding
the Cu wire bonding technology presents a promising alternative. 2.2. Cu wire bonding In general, to judge the wire bonding ability, we control the bonding ball shear and the bonding wire pull (Fig. 2.3). Value set-tings differ according to the type of assembly package. Moreover, wire diameters also differ depending on whether Au or Cu was used.
The earliest look at the iNEMI 2019 Roadmap. A free copy of the contributors chapter ( 500 value) or (if chair/co-chair) the entire iNEMI 2019 Roadmap ( 3,500 value) The opportunity to help drive the industry over the 10-year horizon of the iNEMI Roadmap and insure your company's needs are represented. Contributing Team Member's .
the gold bonding wire and gold board metallization. Power and time relate to the ultrasonic generator settings used to “ultrasonically soften,” the bonding wire. Tool force is the amount of weight applied to the bonding wire to mechanically couple the bonding wire to the bonding pad surface. D
black black white white white white red green blue black Wire 3 Wire 3 Wire 4 Wire 4 Wire 5 Wire 6 Wire 8 Wire 7 Wire 2 e 2 Wire-number Wire color Wire-number e color For multicolor configurations For all other configurations * internal pull-up resistor to vcc 3.3 V Dimension 15.15
Wire bonding was conducted on a gold ball bonder at a bonding temperature of about 155-160 C on the PCB surface using 25um 4N gold wire. The bonding tool used for the wire bonding process was for off-the-shelf 80um pitch applications with capillary tip dimensions of about 100um. Wire pull tests were
(Co-sponsored by IEEE’s CPMT Society, Fraunhofer IZM, IMAPS Europe) 16. Thursday, 17 November. Implementing the iNEMI System-in-Package Technical Plan. iNEMI Conference Room A 61, 9:00 - 12:00 . This meeting will review and prioritize gaps identified in the iNEMI Technical Plan and develop projects to close the gaps. Update on Pb-Free Soldering Optimization. Productronica Forum, Hall A6, 13 .
Wire Size Wire Size 20.75mm 1.50mm2 E1010RD EN2502 (10mm) (12mm) Wire Size Wire Size 21.00mm2 2.50mm E1510BK EN4012 (10mm) (12mm) Wire Size 21.50mm 4.00mm2 E2512GY EN6012 (12mm) (12mm) Wire Size Wire Size 2.50mm2 26.00mm E4012OR EN10012 (12mm) (12mm) Wire Size Wire Size 4.00mm2 10.00mm E6012GR EN61018 (12mm) Wire Size Wire Size 26.00mm2 16.00mm .
American Revolution Activity Book This Activity Book contains activity pages that accompany the lessons from the Unit 6 Teacher Guide. The activity pages are organized and numbered according to the lesson number and the order in which they are used within the lesson. For example, if there are two activity pages for Lesson 4, the first will be numbered 4.1 and the second 4.2. The Activity Book .
