WHY it matters: As transistors shrink, wires get narrower but often carry similar or more current, so current density J grows. EM lifetime falls steeply with J, so EM sets a hard design rule: a max allowed current per wire width.
Direct field forceFfield=qE — the electric field pulling the positive ion toward the cathode.
Electron-wind forceFwind — moving electrons collide with ions, transferring momentum toward the anode (electron flow direction).
In good conductors the wind dominates. We bundle both into an effective chargeZ∗:
F=Z∗qE=Z∗qρJ
Why this step? Because E=ρJ (Ohm's law in local form, ρ = resistivity), we express the driving force directly in terms of the current density J — the thing designers control.
This force drives an atomic flux. Using the Einstein relation (mobility =D/kBT):
Jatom=kBTNDF=kBTNDZ∗qρJ
where N = atom density, D=D0e−Ea/kBT is diffusivity. A divergence in this flux (∇⋅Jatom=0) is what accumulates mass → voids/hillocks.
Combine: failure occurs when a critical amount of mass has moved. Rate ∝Jatom∝Je−Ea/kBT. Empirically Black found the current dependence is a power n, giving Median-Time-To-Failure:
Why e+Ea/kT (positive)? Higher T → faster diffusion → shorter life, so MTTF must drop as T rises. e+Ea/kT decreases with T. ✔ Sanity check.
A wire is a hallway full of tiny metal marbles. When you push a strong stream of electric "wind" through it, the wind knocks marbles down the hall. Slowly, one spot runs out of marbles and a hole forms → the wire snaps (open). Another spot gets a pile-up → touches a neighbor wire (short). Push harder (more current) or make it hotter, the marbles move faster and it breaks sooner. But if the hallway is very short, the marbles pile at the end and push back so hard they stop moving — that wire never breaks.
Dekho, electromigration ka matlab hai ki metal wire ke andar jo electrons ka strong "wind" behta hai, wo actual mein metal ke atoms ko dhakka de kar hilaa deta hai. Yeh dhakka electron flow ki direction mein hota hai, yani anode ki taraf — isliye ek side pe atoms khatam ho jaate hain (void, wire open ho jaati hai) aur doosri side pe pile-up ho jaata hai (hillock, short ho sakta hai). Yeh slow process hai, mahino-saalon mein chip ko fail karta hai — isliye ise wear-out reliability failure kehte hain.
Kitni jaldi fail hoga, yeh Black's equation batati hai: MTTF=AJ−neEa/kBT. Yaad rakho — current density J badhao to life girti hai (agar n=2, to J double karne se life 1/4 ho jaati hai), aur temperature badhao to bhi life girti hai (kyunki garmi se atoms tez diffuse karte hain). Isliye exponential mein +Ea/kT hai — jaise-jaise T badhta hai, yeh term chhota hota hai, MTTF girta hai. Yeh baat modern chips mein aur serious hai kyunki wires patli hoti jaa rahi hain lekin current utna hi — matlab J high, EM ka risk high.
Ek pyaari trick bhi hai: Blech effect. Agar wire bahut chhoti ho, to anode side pe atoms ka pile-up ek back-pressure (back-stress) banata hai jo electron wind ko rok deta hai. Is condition ko J⋅L product se check karte hain — agar J⋅L critical value se kam hai, wire immortal ho jaati hai, kabhi fail nahi hoti. Isliye chhote jumper wires "free" maane ja sakte hain design mein.
Practical takeaway: designer ke liye EM ek design rule ban jaata hai — har wire width ke liye maximum current fix. Copper aluminium se better hai (higher Ea), lekin immune nahi — wo apni surface/interface ke along migrate karta hai, isiliye cap aur barrier layers zaroori hain.
