Distribution and Transport of Alpha Emitting Isotopes in Tin Packaging Materials Dr. Brett M. Clark Honeywell Electronic Materials

IEEE SER Workshop October 25, 2012

Outline • Introduction • Radiation Induced Soft Errors • Alpha Emission Beyond Secular Equilibrium • Hypothesized Diffusion • Implications to Packaging • Conclusions

2012 IEEE SER Workshop

2

Alpha Radiation and Soft Errors Ion path

20

80

15

60

10

40

5

20

0

0 0

2

4

6

8

Range ( m)

+ n doped - ++- + p doped substrate + +-

dQ/dx (fC/ m)

• Nuclear decay product - Helium nucleus with +2 charge - High kinetic energy (2-8 MeV) - 10-30 m range in silicon • Packaging Technology transition - Wafer Level / 3D Packaging • Alpha sources closer in proximity to sensitive nodes • Low alpha materials required for increasing number of applications

10

Particle Energy (MeV)

Understand alpha behavior in Pb-free materials 2012 IEEE SER Workshop

3

Paradigm Shift • Current Industry Understanding - Secular equilibrium is an inclusive explanation of time dependent alpha emission levels

• Secular Equilibrium Explanation Insufficient - Experimental data departs from theoretical prediction models • This is a concern… - Alpha emission from the packaging materials potentially higher than projected

Sn Alpha Emissivity vs Time 14

B

Counts/khr·cm2

12

C Sec. Equil. Model Sec. Equil. Model

10

8

6

4

2

0 0

100

200

300

400

500

Time (Days)

600

700

800

2012 IEEE SER Workshop

4

Experimental Design • Hypothesis: Microsegregation transports emitters towards the center of the mass during solidification  Elements partition between liquid and solid phases  Solute distribution is heterogeneous

• Select Sn materials with range of surface emissions - Measure alpha flux in sequential layers of the volume - Dissolve >15 μm from the surface with each iteration Alpha Flux Measurements

Initial Surface

Layers removed by acid dissolution 700 µm

Final Surface

2012 IEEE SER Workshop

5

Alpha Emissivity vs. Depth

Alpha Emission vs Depth: 100 μm 80

Linear Concentration

Counts∙khr-1·cm-2

70

Gradients

y = 0.86x + 12 R² = 0.98

60

A B

50

C

40

y = 0.22x + 8.4 R² = 0.99

30 20

y = 0.019x + 0.78 R² = 0.35

10 0 0

20

40

60

80

100

120

Depth from Surface (μm)

Alpha surface emissivity A>B>C 2012 IEEE SER Workshop

6

Microsegregation Discussion • Distribution curve slope is determined by emitter concentration in the material • Trace impurities transported into the center of mass - Alpha particles detected from only top ~ 20 μm - Initial alpha measurements may be erroneously low • Heterogeneous emitter distribution does not explain the time dependant alpha data - Surface alpha emitter concentration increasing

Additional time dependent mechanism at work 2012 IEEE SER Workshop

7

Polonium Diffusion Experiment • Hypothesis: - Alpha emission increasing over time from 210Po diffusion • Reproduce experimentally to confirm mechanism - Doped low alpha Sn with 210Po - Sample heated & compared against room temp control • Measure/Heat/Measure cycle repeated Alpha emissivity vs [210Po]

Alpha Spectrum ‐ Sn Heat Experiment Significant increase in 210Po

3

Pre heat

Alpha‐hr‐1

2.5

Heat Cycle 1  Control 

2 1.5 1 0.5 0

1

2

3

4

Energy (MeV)

5

6

1.0E+03

2 π Emission cts ∙khr‐1∙cm‐2

3.5

1.0E+02 1.0E+01 1.0E+00 5.8E‐01

5.8E+00

5.8E+01

5.8E+02

5.8E+03

1.0E‐01 1.0E‐02

210Po equivalent concentration (atoms/g Sn)

2012 IEEE SER Workshop

8

Emissivity vs. Time: 210Po doped Sn

Diffusion accelerated by temperature 2012 IEEE SER Workshop

9

Dynamic Alpha Emissivity • Heat conditioning drives majority of polonium to the surface - Heating on consecutive days did not change Po activity - Diffusion is accelerated ~240x at 200 °C vs. 23 °C • Surface alpha emission determined by Po concentration - Po concentration > 10-4 parts per trillion unacceptable (>2 α·khr1·cm-2)

Analyzed volume

Time elapse

210Po

diffusion to surface

-1·cm-2 Transport Alpha Emitters Confirmed Material of state after packaging >2 α·khr 2012 IEEE SER Workshop

10

Packaging implications • Microsegregation effects in packaging processes - Phase transition processes susceptible - Process temperature accelerated diffusion • Polonium mobility considerations - Static position of emitters oversimplified - Negate to some extent shielding assumptions made in design • Diffusion transports emitters in volume to surface Surface Po Activity vs Sphere Diameter  1.E+04 1.E+03

α∙khr‐1∙cm‐2

1.E+02 1.E+01 1.E+00 1.E‐01

1

50

100

150

1.E‐02 1.E‐03 1.E‐04

Po atoms/g 1E+04 1E+03

1.E‐05 1.E‐06

Sphere Diameter (μm) 2012 IEEE SER Workshop

11

Conclusions • Microsegregation and diffusion effects can result in increased alpha emissions • Both mechanisms should be considered when evaluating packaging processes in high reliability applications • Most direct solution is to selectively eliminated Po from feedstock materials

2012 IEEE SER Workshop

12

Acknowledgements • Michael Fields: Analytical data • Michael Paladin: Sample Preparation • Thomas Ploegman: Sample Preparation • Jack Long: Sample Preparation

2012 IEEE SER Workshop

13