Task Group 8-81f Self-mitigation PERM #33 Update David Pinsky This document does not contain technology or technical data controlled under either the U.S. International Traffic in Arms Regulations or the U.S. Export Administration Regulations.

Background Components with pure tin finished terminations are at risk for tin whisker growth and potential unreliability. The addition of lead (Pb) is an effective means to mitigate tin whisker growth. Therefore, eutectic tin lead solder does not pose a meaningful risk of tin whisker growth provided the lead content is at least 3% by weight. Those portions of pure tin terminations that are replaced by eutectic tin lead solder during SMT processing no longer poses a risk of whisker formation. Therefore, components where eutectic tin lead solder has fully replaced all tin plating on the terminations are fully mitigated against tin whisker risks. Parts that can be mitigated in this fashion are said to be "self mitigating“. 11/20/201811/20/2018

Examples Solder covers the entire termination: self-mitigating Solder does not cover the entire termination: not self-mitigating 11/20/2018

Previous Work A study was performed in 2010 by Raytheon, the results of which have been published, concluding that: Component termination geometry could be used to predict reliable self mitigation for components soldered to HASL finished boards, using a specific (fairly typical) SMT process The study did not address how these results may or may not apply to components soldered to boards using different manufacturing processes, surface finishes, or pad geometries 11/20/201811/20/2018

Aim of the Present Study If the conditions under which self mitigation is achieved reliably can be determined, it will be possible to use certain tin terminated parts without introducing the risk of whisker shorts, and without employing any other mitigating techniques Task Group 8-81F was formed under the PERM Council (Committee 8-81) to investigate This study was performed to evaluate the conditions under which typical SMT components will achieve self mitigation 11/20/2018

Design of Experiment Many potential factors for consideration were considered for inclusion in the study. The four factors shown in the table below were selected. The layout of the board and the components reused from the prior study, to permit direct comparison between the results of the two studies. Experimental Factor Settings Component Packages 16 different part numbers (details below) Board finish OSP and Sn Pb HASL Pad size Per initial study and 25% smaller Manufacturing Process Seven different locations 11/20/2018

Design of Experiment Each of the seven assembly locations were provided with a kit containing eight bare boards (Four types, two replicants) and all of the components necessary to populate them. Each assembler chose process conditions as they would deem appropriate to achieve compliance with J-STD-001, Class 3 requirements HASL finish/large pads HASL finish/small pads OSP finish/large pads OSP finish/small pads 11/20/2018

Assembled Test Vehicle 11/20/2018

Status Phase I results presented at APEX 2016 and published in SMT Magazine in September Almost finished with Phase II Results presented at the IPC Reliability Forum: Manufacturing High-Performance Products, Chicago, IL April 26-27 Final report completion this fall Presentation planned at APEX 2018

Phase I Conclusions Package geometry is a significant factor Some package styles mitigate across all processes Some do not mitigate for any process Solder process is a significant factor Pad size is significant, but not important Board finish is not significant

Termination Length (mils) Termination Height (mils) Components Used Part Number Package Style Quantity per vehicle Termination Length (mils) Termination Height (mils) 06035C103KAT2A 0603 chip 10 31 13 0603YD225KAT2A A3PN030-ZVQG100 TQFP100-14mm 1 40 16 ADM213EARSZ SSOP28-5.3mm 2 59 EPM7032AETC44-10N A-TQFP44-10mm-.8mm IR2156SPBF SO14G-3.8mm 41 23 LM2901DG LTC3703EG_PBF MBRM140T1G DO-216AA 5 49/17 20 MC9S08GT16AMFBE QFP44-.8mm 66 36 MC9S08QE4CLC LQFP32-7mm-.8mm 42 25 OP482GSZ 46 19 PZT2222AT1G SOT223 69/73 STAC9200X5TAEB1X LQFP48-7mm-.5mm W3L1YC474MAT1AF 0612 chip 12 39 XC9572XL-5TQG100C LQFP100-14mm-.5mm

Assembly Process Details Setting Process A Process B Process C Process D Process E Process F Process G Reflow type Vapor Phase Convection Oven Convection Oven Flux ROL0 ORM0 63/37 Sn62 RMA No-Clean No clean Tacky Flux Stencil thickness 5 mils 4 mils Laser 5 mil Stencil aperture 1:1 Time above Liquidus 90s 60-75s 66s 45 sec 60s 70s Peak temperature 218C 215C 220C 225C 220 ͦC 213C Atmosphere Air Nitrogen Number of reflow cycles 1 2 Rework None Yes

XRF Data Summary Probability of self-mitigation by package and process 80% Confidence Interval Probability of self-mitigation by package and process

Phase II Results XRF Very old parts Cross-sections Consistency across multiple leads of a single package Corroboration with measurements using newer, more advanced instrument Check on very “thin” readings Very old parts Cross-sections

Phase II XRF Two packages indicated as self-mitigating were re-examined, with most leads on all four sides measured, not just two leads as in Phase I All terminations verified as self-mitigated

Example of XRF results (Board 46, U11, Lead 1, Egress) Raytheon Example of XRF results (Board 46, U11, Lead 1, Egress) Spectrum and Fit Target at lead egress Results 11/20/2018

Phase II XRF Some measurements were repeated using a new, more powerful instrument (M4 Tornado) All of board 46 from assembler “A” Result: All assembler “A” components previously identified as self-mitigating by MXR300 were also identified as self-mitigating by the M4 Tornado Specific terminations suspected of very thin coatings or with borderline Pb content Result: Thin coating indications also provided by M4 Tornado, consistent measurements for borderline Pb readings

Phase II- old parts One and a half boards were populated with parts that had been stored in an uncontrolled environment (Tom Hester’s office) for more than five years. Results of XRF were compared between the old and new parts assembled by the same process at the same time on exact replicant boards

Phase II- old parts Old parts assembled per Process G for packages with >.98 probability with a 90% confidence interval of full coverage for new parts Package Style Measured Mitigated 0603 Chip 62 10 SO-14G 21 19 LQFP-48 4 TQFP-100 8 LQFP-100 Drastic loss of Mitigation Still Mitigating?

Phase II- Sections Results from SEM/EDS consistent with XRF data for devices with leads, some inconsistency for chip devices Transitions between solder flow and original tin are abrupt Undersides of gull-wing leads are often fully covered even when the top sides are not

50X 750X 2000X Abrupt transition between solder and tin plating

Example on fully mitigated lead Element Weight% Atomic% Sn L 91.47 94.93 Pb M 8.53 5.07 Totals 100.00   750X 2000X

Example on non-mitigated lead Bd 45 U73 left side. Bd 45 U73 top left spectra. Bd 45 U73 top left side. 2/25/2017 CER 46949 Jim Andorf

U1 – Lead 2 50X 750X 2000X Upper side of lead not fully mitigated

Under side of lead fully mitigated U1 – Lead 2 50X Element Weight% Atomic% Cu L 1.54 2.89 Sn L 93.70 94.36 Pb M 4.76 2.75 Totals 100.00   750X 2000X Under side of lead fully mitigated

0603 XRF vs EDS Conflict EDS reading at this site reported 98.5 wt% Sn Raytheon 0603 XRF vs EDS Conflict Bd4 C9 right top spectra. Bd4 C9. EDS reading at this site reported 98.5 wt% Sn (Small amounts of Pb phase Is seen in image) XRF reading at this site reported 66.3 wt% Sn Bd4 C9 right top. 11/20/2018

Raytheon XRF v EDS conflict Investigation into cause of the conflict is in progress- outcome is TBD There is high confidence that the instruments were properly measuring the composition of the regions that were excited by the beams Possible causes under investigation Mis-alignment of the XRF spot relative to the termination Effect of Sn and Pb segregation in joint 11/20/201811/20/2018

XRF Data for 0603 Further detailed analysis is being performed on the XRF results for the 0603 components Represent a “borderline” part Could provide insight into effect of process variables Considered: Oven reflow v Vapor Phase reflow Parts with longest storage before assembly v other oven-reflowed parts

XRF Data for 0603

Raytheon Results from SEM/EDS Abrupt solder/tin transition verifies that XRF readings are meaningful for mitigation Bottom surfaces of leads exhibit a higher degree of mitigation than do the upper surfaces For leaded devices- those indicated as certain to mitigate by XRF was seen to mitigate by SEM/EDS For chip devices- cause(s) of some disagreements needs resolution 11/20/2018

Conclusions Results from Phase I validated Consistency across leads of large devices XRF data confirmed by cross-section for devices with leads XRF results duplicated on alternate instrument Results for chip devices under investigation Uncontrolled long term aging of parts can compromise self-mitigating behavior

Plans What next?- Let’s discuss Repeat experiment with SAC305? Consider other parameters of interest?