Visual walkthrough — Statistical mechanics — microstate, macrostate

2.4.8 · D2 · Physics › Thermodynamics & Statistical Mechanics (Advanced) › Statistical mechanics — microstate, macrostate

Hum maante hain aapko kuch bhi nahi pata siwaaye counting ke. Har symbol pehle kamaaya jaata hai, tab use hota hai.

---

Step 1 — "State" kya hota hai? Do words jo hum mix nahi kar sakte

KYA HAI. $N$ coins ek row mein rakh lo. Har coin ya to H (heads) hai ya T (tails). Do bilkul alag sawaal:

• "Mujhe exact pattern do" — jaise H T T H. Woh poora ordered list ek microstate hai.
• "Bas batao kitne heads hain" — ek number $n$. Woh ek macrostate hai.

Yahan $N$ = coins ki sankhya, aur $n$ = heads ki sankhya. Abhi kuch aur nahi.

Inhe alag kyun karein? Kyunki ek real gas mein aap sirf doosri tarah ki cheez measure kar sakte ho (kuch bulk numbers). Pehli tarah ki cheez aap kabhi likh nahi sakte. Toh physics ko yeh poochhna padta hai: kitne microstates ek hi macrostate share karte hain?

PICTURE. Left column ek microstate hai (poora pattern). Right box kaafi saare patterns ko ek single label "2 heads" ke neeche group karta hai.

---

Step 2 — Har microstate count karo: $2^N$ kyun

KYA HAI. $N$ coins ke liye kitne complete patterns exist karte hain? Coin 1 ke paas 2 choices hain; unme se har ek ke liye, coin 2 ke paas 2 choices hain; aur aise hi aage.

Multiply kyun karein? Yeh fundamental counting principle hai: independent choices multiply hoti hain, kyunki coin 1 ki har choice coin 2 ki har choice ke saath pair ho sakti hai. Do coins → $2\times 2=4$ patterns; teen coins → $2\times 2\times 2=8$.

${\Omega }_{\text{total}}=\underset{N\text{coins}}{\underbrace{2\times 2\times \cdots \times 2}}=2^N$

• $2$ — options ki sankhya har coin ke liye (H ya T).
• $N$ — kitne coins hain, yaani hum kitni baar multiply karte hain.
• ${\Omega }_{\text{total}}$ — distinct microstates ka grand total.

PICTURE. Ek branching tree: har level par leaves ki sankhya double ho jaati hai. Sabse neeche leaves count karo — woh $2^N$ hai.

Recall 2's ko add kyun nahi karte?

Add karna ($2+2+\dots$) toh choices akele ek ek coin ke liye count karega. Lekin ek pattern ko har coin ke liye ek saath choice chahiye — choices stack hoti hain, aur independent choices ko stack karne ka matlab hai multiply karna. ::: multiply, kyunki har coin ka option doosre sab ke options ke saath pair hota hai.

---

Step 3 — Ek macrostate count karo: $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ kyun

KYA HAI. Macrostate ko "exactly $n$ heads" fix karo. Ise kitne patterns realise karte hain? Hume choose karna hoga ki $N$ slots mein se kaun se $n$ slots heads hain.

Combination kyun, permutation kyun nahi? Kyunki heads identical hain — do heads ko swap karne se same pattern milta hai. Humein sirf kaun se slots heads hain, yeh important hai, naa ki humne unhe kis order mein select kiya. "$N$ mein se $n$ slots choose karo, order irrelevant" — yeh exactly combination hai (dekho Binomial and Gaussian distributions):

$\Omega (n)=\left(\genfrac{}{}{0ex}{}{N}{n}\right)=\frac{N!}{n!(N-n)!}$

Har term padhte hain:

• $N!$ — saare $N$ coins ko har order mein line up karo ($N$ factorial $=N\times (N-1)\times \cdots \times 1$).
• $n!$ — lekin $n$ heads interchangeable hain; unke $n!$ reorderings ek hi pattern dete hain, toh unhe divide out karo.
• $(N-n)!$ — similarly $N-n$ tails bhi interchangeable hain; unke reorderings bhi divide out karo.
• $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ — padho "$N$ choose $n$": $n$ heads wale distinct patterns ki sankhya.

PICTURE. Neeche: $N$ slots boxes ki tarah; hum $n$ ka chosen subset shade karte hain. Alag alag shadings jo same slot-set use karti hain woh same macrostate hain — isliye hum divide karte hain.

---

Step 4 — Poori distribution: $\Omega (n)$ plot karo

KYA HAI. $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ har $n$ ke liye $0$ se $N$ tak compute karo aur plot karo. $N=10$ ke liye values Pascal's triangle ki row $1,10,45,120,210,252,210,120,45,10,1$ hain.

Plot kyun karein? Ek picture punchline turant dikha deti hai: counts flat nahi hain. Extremes ($n=0$ ya $n=N$) mein $\Omega =1$ hai; middle ($n=N/2$) sab se upar tower karta hai.

PICTURE. $\Omega (n)$ vs $n$ ki bars. Green centre bar red edge bars ko pachhaad deta hai. Woh height difference hi wajah hai ki aap kabhi all-heads nahi dekhte.

---

Step 5 — Woh forecast jo sach honi hi chahiye: ${\sum }_n\left(\genfrac{}{}{0ex}{}{N}{n}\right)=2^N$

KYA HAI. Saare macrostates ki multiplicities add karo. Agar humare do counts (Step 2 aur Step 3) consistent hain, toh yeh sum har microstate ko exactly ek baar rebuild karna chahiye.

${\sum }_{n=0}^N\left(\genfrac{}{}{0ex}{}{N}{n}\right)=2^N$

• $\sum _{n=0}^N$ — $n$ ko har macrostate $0,1,\dots ,N$ par sweep karo.
• $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ — uss ek macrostate mein microstates.
• $2^N$ — Step 2 se grand total.

Yeh kyun kaam karta hai. Binomial theorem kehta hai $(a+b{)}^N={\sum }_n(\genfrac{}{}{0ex}{}{N}{n})a^n{b}^{N-n}$. $a=b=1$ rakho: har term sirf $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ ban jaata hai, aur left side $(1+1{)}^N=2^N$ hai. Toh microstates ko macrostate ke hisaab se partition karne par koi kho nahi jaata.

PICTURE. Step 4 ki bars, end-to-end stack ki gayi hain, exactly total $2^N$ fill karti hain — na overlap, na gaps.

---

Step 6 — Counts se probability tak, aur degenerate edges

KYA HAI. Fundamental postulate of statistical mechanics ke zariye — har accessible microstate equally likely hai — ek macrostate ki probability uska total mein hissa hai:

$P(n)=\frac{\Omega (n)}{2^N}=\frac{1}{2^N}\left(\genfrac{}{}{0ex}{}{N}{n}\right)$

• $\Omega (n)$ — $n$ heads paane ke tarike (favourable microstates).
• $2^N$ — saare microstates (postulate ke hisaab se equally likely).
• $P(n)$ — woh probability jis par aap macrostate $n$ measure karte ho.

Edge / degenerate cases — koi skip nahi:

• $n=0$ (saare tails) ya $n=N$ (saare heads): $\Omega =1$, toh $P=1/2^N$. Sabse rare allowed macrostates — yeh exist karte hain lekin almost kabhi nahi dikhai dete.
• $n=N/2$ (sirf tab jab $N$ even ho): sabse bada $\Omega$, sabse probable macrostate.
• $N$ odd: exact middle nahi; do central values $n=\frac{N-1}{2},\frac{N+1}{2}$ peak ke liye tie karte hain.
• $N=1$: sirf do macrostates, har ek $\Omega =1$, har ek $P=1/2$ — split invisible hai (sharp peak ke liye kaafi coins chahiye).

PICTURE. Large $N$ ke liye probability curve $P(n)$, red mein tiny edge bars marked aur green mein towering peak — plus ek arrow dikhata hai peak width $N$ ke relative kitni shrink ho rahi hai.

---

Step 7 — Width kahan se aati hai? $\sqrt{N}$ derive karna

Isse pehle ki hum "width shrinking" ki baat karein, hume clearly kehna hoga "width" ka matlab kya hai aur yeh $\sqrt{N}$ kyun hai — sirf cite nahi karenge.

KYA HAI aur KYUN ($\sqrt{N}$ appear karne ki trick). Head-count ko har coin ke ek chhote counter ke sum ki tarah likho. Coin $i$ ke liye $x_i=1$ agar woh heads hai aur $x_i=0$ agar tails. Tab

$n=x_1+x_2+\cdots +x_N$

• $x_i$ — $i$-ve coin ka contribution: $1$ ya $0$.
• $n$ — total heads, bas unka sum.

Ek fair coin ke liye mean $\bar{x}=\frac{1}{2}$ hai aur uski variance hai

$\text{Var}(x_i)=p(1-p)=\frac{1}{2}\cdot \frac{1}{2}=\frac{1}{4}$

• $p=\frac{1}{2}$ — probability ki coin heads dikhayega.
• $p(1-p)$ — ek single 0/1 coin ki variance (sabse zyada jab $p=\frac{1}{2}$, zero jab coin rigged ho).

Key rule (variances kyun add hoti hain, $\sigma$'s nahi). Independent pieces ke liye, variances add hoti hain — squared strays cross-terms ke bina accumulate hote hain kyunki ek coin ki kismet doosre coin ke baare mein kuch nahi batati:

${\sigma }^2=\text{Var}(n)=\underset{\text{coins}}{\underbrace{N}}\times \underset{\text{each }=1/4}{\underbrace{p(1-p)}}=\frac{N}{4}$

Square root lo width khud paane ke liye:

$\boxed{\sigma =\sqrt{\frac{N}{4}}=\frac{1}{2}\sqrt{N}}\Rightarrow \sigma \sim \sqrt{N}$

• $N$ — independent coins ki sankhya; yeh variance ke andar ek baar, linearly aata hai.
• $\sqrt{\cdot }$ — variance ko waapas width mein convert karna exponent ko half kar deta hai, toh $N\to \sqrt{N}$.

PICTURE. Left: ek coin ki variance, $p(1-p)$, $p=\frac{1}{2}$ par peak karti hui. Right: $N$ independent strays squares ki tarah add hote hain (Pythagoras-style), toh total width $\sqrt{N}$ lambi hoti hai, $N$ nahi.

---

Step 8 — $10^{23}$ coins matlab hamesha ek hi jawab

KYA HAI. $P(n)$ ka peak centre $\bar{n}=N/2$ par baithta hai. Step 7 se uski width (standard deviation) $\sigma =\frac{1}{2}\sqrt{N}$ hai, jabki uski position $N/2$ hai. Toh relative spread — width poori cheez ke fraction ke roop mein — hai

$\frac{\text{width}}{\text{position}}=\frac{\sigma }{\bar{n}}=\frac{\frac{1}{2}\sqrt{N}}{N/2}=\frac{1}{\sqrt{N}}\underset{N\to \infty }{\to }0$

• $\sigma =\frac{1}{2}\sqrt{N}$ — Step 7 mein derive ki gayi width.
• $\bar{n}=N/2$ — peak kahan baithta hai.
• $1/\sqrt{N}$ — fractional fluctuation, large $N$ ke liye vanish karta hua.

Yeh kyun matter karta hai. $N\sim 10^{23}$ ke liye, $1/\sqrt{N}\sim 10^{-11}$: heads ka fraction $50\%$ par pin ho jaata hai, approximately $10^{11}$ mein $1$ part tak. Measurable macrostate simply hilti hi nahi. Woh frozen, most-probable macrostate equilibrium hai — aur uska logarithm entropy hai.

PICTURE. Teen curves $N=10,100,1000$ heads-fraction $n/N$ par rescaled: jaise $N$ badhta hai bell $0.5$ par ek spike mein collapse ho jaati hai.

---

Ek-picture summary

Sab ek saath: branching tree (saare $2^N$ microstates) bar chart mein jaata hai ($\Omega (n)$ har macrostate ke liye), jiska total $2^N$ pe re-sum hota hai, jiska normalised version $P(n)$ hai, jiska width $\sigma =\frac{1}{2}\sqrt{N}$ $N$ ke relative sharpen hokar equilibrium ban jaata hai.

Recall Feynman retelling — poora walkthrough plain words mein

$N$ coins line up karo. Har coin do taraf flip hoti hai, toh $2^N$ complete patterns hain — yeh saare microstates hain. Lekin tum, experimenter, sirf ek number dekhte ho: kitne heads hain, woh macrostate. Yeh jaanne ke liye ki kitne patterns ek certain head-count $n$ dete hain, tum choose karte ho ki kaun se $n$ slots heads hain aur order ignore karte ho — woh $\left(\genfrac{}{}{0ex}{}{N}{n}\right)$ hai. Un counts ko plot karo aur ek pahaad aata hai: "all heads" par ek akela pattern, "half heads" par ek towering pile. Saari piles waapas add karo aur exactly $2^N$ milta hai (kuch kho nahi jaata). Kyunki har pattern equally likely hai, ek macrostate ki probability bas uska pile-height $2^N$ se divide karke milti hai. Ab bade numbers ka jaadu: har coin ko ek chhota $1$-ya-$0$ counter do; har ek variance mein $\frac{1}{4}$ se wobble karta hai, aur kyunki independent wobbles squares ki tarah add hoti hain, total wobble $\sigma =\frac{1}{2}\sqrt{N}$ hoti hai — ek random-walk $\sqrt{N}$, seedha $N$ nahi. Toh pahaad ki width $\sqrt{N}$ ki tarah badhti hai lekin uska centre $N$ ki tarah badhta hai, use uske size ke relative razor-thin bana deta hai. $10^{23}$ coins par yeh exactly half heads par ek needle hai. Woh immovable, overwhelmingly-favoured macrostate hi woh hai jise hum equilibrium kehte hain — aur heat, temperature, aur irreversibility yeh sab is counting story ke alag alag kapde hain.

---

Connections

• Parent: 2.4.08 Statistical mechanics — microstate, macrostate (Hinglish)
• Entropy and Boltzmann's relation S = k ln Ω
• Fundamental postulate of statistical mechanics
• Binomial and Gaussian distributions
• Microcanonical ensemble
• Second law and irreversibility
• Phase space and Liouville's theorem
• Temperature as ∂S/∂E