Shuru karne se pehle, ek figure us picture ko fix karta hai jis par hum baar baar laute hain: two-level system, hamara workhorse. Isse neeche Figure 1 label kiya gaya hai aur solutions mein usi naam se refer kiya jaata hai.
Figure 1 — The two-level system.
Do horizontal levels dekho: energy 0 (ground) aur energy ε (excited). Kam temperature pe system neeche baithta hai; jaise-jaise heat daali jaati hai, upper level bharne lagta hai. Neeche ka har exercise ya toh yahi picture hai, iska copy hai, ya ek smooth (oscillator/gas) analogue hai.
Goal: kya aap sahi formula shelf se utha ke numbers plug in kar sakte ho?
Recall Solution
KYA: master bridge F=−kBTlnZ directly apply karo.
KYUN: kuch derive nahi karna — Fdefine hi hota hai −kBTlnZ se jab Z mil jaaye.
F=−kBTln4=−kBT(1.3863)≈−1.386kBT.
Sign negative hai kyunki Z>1 matlab lnZ>0 hai, aur aage ek minus hai. Bahut accessible states wale system ki free energy low (bahut negative) hoti hai — nature ise pasand karta hai.
Recall Solution
KYA: do states pe Boltzmann weights e−βEi ka sum karo.
KYUN: definition Z=∑ie−βEi — har state ek exponential contribute karta hai.
Z=e−β⋅0+e−βε=1+e−βε.βε=1 pe: Z=1+e−1=1+0.3679=1.3679.
Picture (Figure 1): "1" ground level hai jo hamesha full weight se count hota hai; e−βε excited level ka shrunk weight hai — shrunk isliye kyunki isme energy ε lagti hai.
Goal: lnZ differentiate karo aur real thermodynamics nikalo.
Recall Solution
KYA:lnZ ka β-derivative lo.
KYUN tool:Uaverage energy ∑ipiEi hai; parent note ne dikhaya ki yeh weighted sum −∂βlnZ ke barabar hai — derivative har exponent se ek Ei "pull down" karta hai aur Z se division (jo ln ke andar chhupa hai) averaging automatically kar deta hai.
lnZ=ln(1+e−βε),∂β∂lnZ=1+e−βε−εe−βε.U=−∂β∂lnZ=1+e−βεεe−βε=eβε+1ε.βε=1 pe: U/ε=1/(e1+1)=1/3.7183=0.2689.
Cases mein sanity:β→∞ (thanda) ⇒U→0 (ground state). β→0 (garam) ⇒U→ε/2 (dono states equally likely). Hamari value 0.269ε beech mein hai, jaisa hona chahiye.
Recall Solution
KYA: parent derivation se assembled relation S=kBβU+kBlnZ use karo.
KYUN: yeh U (jo pehle se mila) aur Z (jo pehle se mila) reuse karne deta hai — koi nayi sums nahi.
βU=β⋅eβε+1ε=eβε+1βε=3.71831=0.2689.
lnZ=ln1.3679=0.3133.
S/kB=0.2689+0.3133=0.5822.Cross-check: ek 2-state system ki maximum possible entropy ln2=0.693 hai (dono states equally likely). Hum usse neeche hain kyunki finite T pe system abhi bhi thoda ground state favour karta hai. ✓
Goal: extensivity, limits, aur systems ko combine karna.
Recall Solution
KYA: single-system partition functions multiply karo, phir log lo.
KYUN tool: independent systems ke exponent mein energies additive hoti hain, aur exponent mein add karne ka matlab hai exponentials ko multiply karna — isliye Ztot=Z1N. Factorisation kyun hold karta hai iske liye Boltzmann distribution and microstates dekho.
Ztot=Z1N=(1+e−βε)N.F=−kBTlnZ1N=−NkBTln(1+e−βε).βε=1 pe: F/(NkBT)=−ln(1.3679)=−0.3133.
Lesson (extensivity):F∝N. Logarithm ne productZ1N ko sumNlnZ1 mein badal diya. Yahi reason hai ki free energy — Z itself nahi — natural thermodynamic potential hai. Helmholtz vs Gibbs free energy dekho.
Recall Solution
KYA: har temperature extreme mein ln(1+e−βε) examine karo.
KYUN: limits reveal karti hain ki formula physically behave kar raha hai ya nahi — jo formula 0 ya ∞ par misbehave kare woh suspect hai.
T→0 (β→∞): e−βε→0, toh ln(1+0)=0, deta hai F→0. Physically system ground state mein freeze ho gaya hai (energy 0, entropy 0), isliye F=U−TS→0. ✓
T→∞ (β→0): e−βε→1, toh ln2, deta hai F→−kBTln2→−∞. Entropy term −TS dominate karta hai (S→kBln2), aur F neeche girta jaata hai. ✓
Figure 2 mein F/ε, U/ε, aur TS/ε ko kBT/ε ke against plot kiya gaya hai taaki aap dekh sako ki F=U−TST badhne pe neeche kheenchta jaata hai.
Figure 2 — F, U aur TS temperature ke against, ek two-level system ke liye.
Goal: ek continuous system ke liye puri thermodynamics derive karo.
Recall Solution
KYA: do Gaussian integrals karo, phir lnZ differentiate karo.
KYUN Gaussian tool: ek quadratic Hamiltonian ka Boltzmann weight e−β(…)p aur x mein Gaussian hai; standard result ∫−∞∞e−ax2dx=π/a is sawaal ka jawab deta hai ki "sabhi positions aur momenta pe total weight kya hai?"
Z=h1∫e−βp2/2mdp∫e−βmω2x2/2dx=h1β2πmβmω22π=hβω2π=ℏβω1=ℏωkBT.
(h=2πℏ use kiya.)
Energy:lnZ=−lnβ+const, isliye
U=−∂β∂lnZ=−(−β1)=β1=kBT.
Yeh equipartition hai: do quadratic degrees of freedom (p aur x), har ek 21kBT contribute karta hai.
Heat capacity:CV=∂T∂U=∂T∂(kBT)=kB.
Free energy:F=−kBTln(kBT/ℏω).
Recall Solution
KYA:S=−∂F/∂T use karo.
KYUN: yeh thermodynamic identity dF=−SdT−pdV+μdN hai jo fixed V,N pe padhi jaati hai; Legendre transforms in thermodynamics dekho.
Maano x=kBT/ℏω, toh F=−kBTlnx. Phir
S=−∂T∂F=kBln(ℏωkBT)+kBT⋅T1=kBln(ℏωkBT)+kB.Check F=U−TS:U=kBT ke saath,
U−TS=kBT−T[kBlnx+kB]=kBT−kBTlnx−kBT=−kBTlnx=F.✓U ka kBT−TS ke andar ke +kBT ko cancel kar deta hai, sirf −kBTlnZ bachta hai — wahi "U cancels" magic jaisi parent derivation mein hai.
Goal: naye results build karo aur frameworks ko connect karo.
Recall Solution
KYA:U ko T ke respect mein differentiate karo, β=1/kBT use karke.
KYUN:C=∂U/∂T measure karta hai ki system har degree mein kitni energy soak karta hai — ek two-level ("Schottky") system ka fingerprint.
Likho U=ε(eβε+1)−1. Chain rule β se: dTdβ=−kBT21=−kBβ2.
∂β∂U=ε⋅(−(eβε+1)2εeβε)=−(eβε+1)2ε2eβε.C=∂T∂U=∂β∂UdTdβ=(−(eβε+1)2ε2eβε)(−kBβ2)=kB(βε)2(eβε+1)2eβε.βε=2 pe: e2=7.389, denominator (8.389)2=70.38.
C/kB=4⋅70.387.389=4⋅0.10498=0.4199.Physics:C→0 dono T→0 (frozen) aur T→∞ (aur absorb karne ki jagah nahi) par, beech mein peak karta hai — Schottky anomaly. Entropy — Gibbs and Boltzmann definitions dekho.
Recall Solution
KYA: har state ko e−βE ki jagah e−β(E−μN) se weight karo.
KYUN naya tool: ab particle number change ho sakta hai, isliye hum reservoir ko μ per particle "pay" karne dete hain. Yeh grand canonical recipe hai.
Step 1 — Z likho. Do states hain: empty (N=0,E=0) aur occupied (N=1,E=ε):
Z=e−β(0−μ⋅0)+e−β(ε−μ⋅1)=1+e−β(ε−μ).Step 2 — grand potential. Definition directly apply karo:
Φ=−kBTln(1+e−β(ε−μ)).Step 3 — average occupation. Trick ⟨n⟩=β1∂μ∂lnZ kaam karti hai kyunki μ ke respect mein differentiate karne se har exponent se ek factor βN neeche aata hai, exactly number-weighted average build karta hai:
⟨n⟩=β1∂μ∂lnZ=1+e−β(ε−μ)e−β(ε−μ)=eβ(ε−μ)+11.Interpretation. Yeh Fermi–Dirac distribution hai — two-level energy jaisi same algebraic shape, lekin ab yeh particles count karta hai. μ=ε set karne par (kisi bhi T pe) ⟨n⟩=e0+11=21 milta hai: chemical potential exactly woh energy hai jis par ek level half-filled hoti hai.
Recall Solution
KYA:p=−∂F/∂V fixed T,N pe use karo; Ideal gas from the partition function dekho.
KYUN: pressure ke liye sirf Z1 ki V-dependence matter karti hai, aur Z1∝V hai, isliye lnZ1=lnV+(V-independent terms).
F=−NkBTlnZ1=−NkBT(lnV−3lnλ).p=−∂V∂F=NkBT∂V∂lnV=NkBT⋅V1=VNkBT.
Dono sides ko volume V se multiply karne par ideal-gas law milta hai:
pV=NkBT.✓
Poori ideal-gas law sirf lnV ka derivative hai. (1/N! Gibbs correction F ko ek V-independent amount se shift karta hai, isliye p unchanged rehta hai — yeh entropy extensivity fix karta hai, pressure nahi.)
Recall Ladder ka ek-line summary
L1 F=−kBTlnZ mein plug karo ::: L2 U,S ke liye lnZ differentiate karo ::: L3 Z's multiply karo, log F ko extensive banata hai ::: L4 oscillator ke liye Gaussian integrals + equipartition ::: L5 heat-capacity peak, grand ensemble, ideal-gas law — sab ek lnZ se.