Microstates, Configurations, and Statistical Weight

Take one mole of a gas at room temperature. Its molecules constantly collide, exchange energy, and rearrange among available energy levels. At any instant, the collection has a definite but fleeting arrangement: some molecules sit in the ground state, others in excited states. Statistical thermodynamics asks a bold question. Can we predict the most probable arrangement without tracking any single molecule? The answer lies in counting. We count how many distinct ways, or microstates, can produce the same overall distribution, called a configuration. This article builds the counting toolkit from scratch. You will meet microstates and configurations. You will derive the statistical weight formula. You will see why one configuration overwhelmingly dominates all others for large numbers of molecules, setting up the Boltzmann distribution.

Key takeaways

A microstate specifies exactly which molecule occupies which energy level at one instant.
A configuration is a set of occupation numbers, how many molecules share each energy level, ignoring molecular identity.
The statistical weight $W$ counts the number of microstates consistent with a given configuration.
$W$ is computed using the multinomial formula from combinatorics.
For a mole of particles, one configuration has a $W$ so large it dominates all others essentially completely.
Key idea: Systems evolve toward the configuration with the largest statistical weight, because it can be realized in overwhelmingly more microscopic ways.

The core insight

A configuration describes only how many molecules occupy each energy level; a microstate describes which specific molecules occupy each level, and the statistical weight $W$ counts how many microstates share one configuration, making $W$ the combinatorial bridge between molecular detail and macroscopic probability.

The physical picture

Imagine four distinguishable molecules, labeled A, B, C, D, distributed between two energy levels: a ground state at energy $0$ and an excited state at energy $\varepsilon$ . Suppose the total energy available is fixed at $\varepsilon$ , so exactly one molecule must occupy the excited level and three must occupy the ground level. There are four microstates that satisfy this: A excited, B excited, C excited, or D excited (with the other three always in the ground state). Each microstate is a distinct, fully specified arrangement. But all four microstates describe the same configuration: three molecules in the ground state, one in the excited state, written as the occupation list $(3, 1)$ . The statistical weight of this configuration is $W = 4$ , because four microstates realize it. If instead two molecules were excited (impossible here under our fixed-energy constraint, but useful for illustration), the configuration $(2, 2)$ would have a different, generally larger, weight. This distinction, many microstates funneling into one configuration, is the entire foundation of statistical thermodynamics. Molecules are physically indistinguishable in reality, but in classical statistical thermodynamics we temporarily label them (A, B, C, D) purely as a bookkeeping device to count arrangements. This labeling trick lets combinatorics do the work; the labels are removed conceptually once we recognize that only configurations, not microstates, are experimentally distinguishable.

Figure

Four microstates collapsing into one configuration (3,1)

The picture changes dramatically once we scale up. For $N = 100$ molecules distributed over several levels, the number of microstates per configuration can reach astronomical values, while nearby configurations differ in weight by enormous factors. This sharp peaking is what allows us to speak confidently about "the" equilibrium distribution rather than a blur of possibilities.

Figure

Statistical weight W rises steeply for configurations nearer equal spreading

The core relations

W = \frac{N!}{n_0! \, n_1! \, n_2! \cdots}

where: $N$ is the total number of molecules, and $n_0, n_1, n_2, \ldots$ are the occupation numbers of successive energy levels, satisfying $\sum_i n_i = N$ .

\sum_i n_i = N, \qquad \sum_i n_i \varepsilon_i = E

where: the first constraint fixes total particle number and the second fixes total energy $E$ available to the system.

W_{\text{total configurations}} = \sum_{\{n_i\}} W(\{n_i\})

where: the sum runs over every configuration consistent with the constraints; the dominant term controls the system's observable behavior.

Derivation ACounting microstates: the multinomial weight formula

Goal. Show that the number of distinct ways to distribute $N$ distinguishable molecules into levels with occupation numbers $n_0, n_1, n_2, \ldots$ is $W = N!/(n_0!\,n_1!\,n_2!\cdots)$ .

Step 1. Place molecules one at a time into labeled "slots." Imagine lining up all $N$ molecules in some order and reading them off into the energy levels: the first $n_0$ molecules read off go into level $0$ , the next $n_1$ into level $1$ , and so on. Every permutation of the $N$ molecules gives one such reading-off, so there are $N!$ orderings in total.

Step 2. Recognize overcounting within each level. Two orderings that merely permute molecules within the same level produce the identical physical arrangement, since a configuration only records occupation numbers, not who is where. If level $0$ contains $n_0$ molecules, there are $n_0!$ ways to permute them among themselves without changing the configuration.

Step 3. Divide out the redundant permutations for every level. The total overcounting factor is the product of internal permutations across all levels,

n_0! \times n_1! \times n_2! \times \cdots

because permutations within different levels are independent of one another.

Step 4. Combine Steps 1 and 3. The number of distinct configurations-consistent microstates is the total orderings divided by the overcounting factor,

W = \frac{N!}{n_0!\, n_1!\, n_2! \cdots}

Step 5. Check limiting behavior. If all $N$ molecules occupy a single level ( $n_0 = N$ , all others zero), then

W = \frac{N!}{N!} = 1

consistent with there being only one way to put everyone in the same box. If each molecule occupies its own distinct level ( $n_i = 1$ for $N$ levels), then $W = N!$ , matching the fact that any permutation gives a genuinely different microstate.

\boxed{W = \dfrac{N!}{\displaystyle\prod_i n_i!}}

Worked examples

Example 1: A small foundational system. Consider $N = 6$ molecules distributed over three energy levels with occupation numbers $n_0 = 3$ , $n_1 = 2$ , $n_2 = 1$ (this satisfies $3+2+1=6$ ). Compute $W$ . Step 1. Write the factorials needed:

6! = 720, \quad 3! = 6, \quad 2! = 2, \quad 1! = 1

Step 2. Apply the formula:

W = \frac{720}{6 \times 2 \times 1} = \frac{720}{12} = 60

So there are 60 distinct microstates (distinguishable labelings of which specific molecules sit where) that all correspond to the single configuration $(3,2,1)$ . Sanity check: compare with the maximally spread configuration $n_0=n_1=2, n_2=2$ isn't valid here since we fixed different occupation numbers, but consider $(2,2,2)$ instead, keeping $N=6$ :

W = \frac{720}{2\times2\times2} = \frac{720}{8} = 90

This confirms the general trend: configurations spreading molecules more evenly across levels tend toward larger $W$ than lopsided ones, foreshadowing why very uneven configurations are statistically disfavored (though energy constraints ultimately decide which configurations are even accessible).

Example 2: Comparing weights at larger N — an exam-level check. A system has $N = 20$ molecules distributed over two energy levels (ground and excited). Compare the statistical weights of configuration A, $(n_0, n_1) = (18, 2)$ , against configuration B, $(n_0, n_1) = (10, 10)$ . Step 1. Compute $W_A$ :

W_A = \frac{20!}{18!\,2!}

Using $20! / 18! = 20 \times 19 = 380$ , and $2! = 2$ :

W_A = \frac{380}{2} = 190

Step 2. Compute $W_B$ . This requires $20!$ and $(10!)^2$ .

20! = 2.4329 \times 10^{18}

10! = 3{,}628{,}800, \quad (10!)^2 = 1.3168 \times 10^{13}

W_B = \frac{2.4329 \times 10^{18}}{1.3168 \times 10^{13}} = 1.848 \times 10^{5}

Step 3. Compare.

\frac{W_B}{W_A} = \frac{1.848\times10^5}{190} \approx 973

The evenly spread configuration B is realized in roughly 973 times more microstates than the lopsided configuration A, even though $N$ here is a modest 20. Sanity check: as $N$ grows toward Avogadro's number, this ratio explodes into numbers so large that only configurations extremely close to the maximum- $W$ configuration contribute meaningfully to any measurable property. This is precisely the seed of the Boltzmann distribution: nature "selects" the configuration of overwhelming weight, subject to fixed total energy and particle number.

Limiting cases & common traps

Situation	Behavior of $W$	Common pitfall
All molecules in one level ( $n_0=N$ )	$W=1$	Forgetting $0!=1$ , misreporting $W=0$
Each molecule in a separate level	$W=N!$	Confusing this with the total number of configurations, not just one
Even spreading across levels	$W$ near maximum	Assuming even spreading is always most probable regardless of the energy constraint
Large $N$ (macroscopic)	One configuration's $W$ dominates all others combined	Treating all configurations as equally important contributors to observable averages

The chemist behind it

Ludwig Boltzmann pioneered the idea that the properties of bulk matter emerge from counting the arrangements of countless molecules. His insight, that entropy is intimately connected to the number of microstates consistent with a macroscopic state, gave physical meaning to the statistical weight $W$ introduced here. The equation carved on his gravestone in Vienna, $S = k \ln W$ , cements this counting exercise as the microscopic root of the second law of thermodynamics, linking directly forward to the Boltzmann distribution.

Origins · 1 minLudwig Boltzmann and how this was discovered

Ludwig Boltzmann

1844-1906 · Austrian physicist who founded statistical mechanics by linking entropy to microstate counting

Ask the tutor: How do we maximize $W$ subject to the constraints on $N$ and $E$ to derive the Boltzmann distribution itself?

Why do we treat molecules as distinguishable when counting microstates, if they are physically identical?

Labeling molecules is a temporary combinatorial device, not a physical claim. It lets us apply straightforward permutation counting to derive $W$ . Because experiments cannot actually track individual identical molecules, only configurations (occupation numbers) are physically meaningful outcomes. The labeling trick correctly reproduces the counting of distinguishable arrangements consistent with a given configuration; a full quantum treatment of indistinguishability introduces corrections (like the Gibbs factor $N!$ in ideal gas partition functions) that are addressed separately in later, more advanced treatments.

Does every configuration have the same statistical weight?

No. The weight $W = N!/\prod_i n_i!$ depends strongly on how occupation numbers are arranged. Highly uneven configurations, such as all molecules crowded into one level, give small $W$ (even $W=1$ in the extreme). Configurations that spread molecules more evenly across accessible levels typically yield much larger $W$ . This disparity, which grows explosively with $N$ , is exactly why one configuration ends up dominating all others in macroscopic systems.

What role does total energy play in restricting which configurations are allowed?

Not every occupation-number set is physically permitted. Only configurations satisfying both $\sum_i n_i = N$ and $\sum_i n_i \varepsilon_i = E$ are candidates, where $E$ is the fixed total energy of the isolated system. Among this restricted set of energy-conserving configurations, we then compare statistical weights. The configuration with the largest $W$ subject to these two constraints turns out, after maximization, to be the Boltzmann distribution.

How does statistical weight connect to entropy?

Boltzmann's relation, $S = k \ln W$ , converts the purely combinatorial quantity $W$ into a thermodynamic state function with units of energy over temperature. Because $W$ for the dominant configuration is astronomically larger than for any other accessible configuration, using $\ln W_{\text{max}}$ in place of $\ln \left(\sum W\right)$ introduces negligible error for macroscopic samples. This equivalence bridges microscopic counting and classical thermodynamic entropy.