We derive the flux from a simple "hopping" picture.
Setup: Carriers hop left or right by a mean free path ℓ every collision time τ. Let vth=ℓ/τ be thermal speed. Consider a plane at x0.
Step 1 — Count carriers that can cross.Why this step? Only carriers within one mean free path ℓ of the plane can reach it before their next collision. Half of them (by symmetry) head toward the plane.
From the left slab (centered at x0−ℓ/2): flux rightward =21n(x0−2ℓ)vth
From the right slab (centered at x0+ℓ/2): flux leftward =21n(x0+2ℓ)vth
Step 2 — Net particle flux F (rightward positive).F=21vth[n(x0−2ℓ)−n(x0+2ℓ)]Why this step? Net flow = (going right) − (going left).
Step 3 — Taylor expand around x0.n(x0±2ℓ)≈n(x0)±2ℓdxdn
So the bracket becomes −ℓdxdn, giving
F=−21vthℓdxdn≡−DdxdnWhy this step? The difference of the two Taylor terms isolates the slope, proving flux depends on the gradient. We named D=21vthℓ the diffusion coefficient.
Step 4 — Turn flux into current. Multiply by charge.
Why this step? Current density J=(charge)×(particle flux).
Sketch of why: In equilibrium, diffusion current exactly cancels drift current (J=0). Setting qDndxdn=−qnμnE and using n∝eqV/kT (Boltzmann) forces D/μ=kT/q.
What quantity does diffusion current depend on? ⇒ the gradient dn/dx, not n.
Why is there a minus sign in Fick's law? ⇒ flow is down the gradient.
Why do Jn and Jp have opposite signs? ⇒ opposite carrier charge.
What links D and μ? ⇒ Einstein relation D/μ=kT/q.
Recall Feynman: explain to a 12-year-old
Imagine a classroom where all the kids are squished into one corner and the rest of the room is empty. Nobody is ordered to move — but as kids randomly wander, more of them wander OUT of the crowded corner than wander back IN, simply because the corner is packed. Slowly the room fills up evenly. If the kids are wearing charged badges, this wandering is an electric current — diffusion current. The bigger the difference in crowding between two spots (the steeper the "crowdedness slope"), the faster the flow.
Dekho, diffusion current ka concept bilkul simple hai. Socho ek jagah pe electrons bahut zyada crowded hain aur paas wali jagah pe kam. Random thermal motion ke wajah se har electron idhar-udhar bhaag raha hai — koi force nahi laga raha. Lekin kyunki ek side pe zyada electrons hain, us side se zyada electrons doosri side jaate hain bajaye wapas aane ke. Iss net flow ko hi hum diffusion current kehte hain, aur ye chalti hai concentration gradient (dn/dx) ki wajah se — electric field ki zaroorat hi nahi.
Sabse important baat: diffusion current n (kitne carriers hain) pe depend nahi karti, balki slopedn/dx pe karti hai. Agar carriers uniform hain (slope zero), toh chahe density kitni bhi ho, diffusion current zero hogi. Fick's law kehta hai flux F=−Ddn/dx — minus sign isliye kyunki flow hamesha high se low taraf jaata hai. Phir current banane ke liye charge se multiply karo: electrons ke liye Jn=+qDndn/dx aur holes ke liye Jp=−qDpdp/dx. Signs alag isliye hain kyunki electron ka charge negative hai, holes ka positive.
Ek aur cheez yaad rakho — Einstein relationD/μ=kT/q=VT≈26 mV (300 K pe). Kyunki D aur mobility μ dono same random collisions se aate hain, isliye ye ek doosre se juda hue hain. Ye relation diode current samajhne mein bahut kaam aati hai.
Ye topic itna important kyun hai? Kyunki PN junction diode ka current mostly diffusion se hi banta hai — minority carriers junction ke paas gradient banate hain aur diffuse karte hain. Toh agar tumhe diode, BJT, ya solar cell samajhna hai, diffusion current ki foundation solid honi chahiye.