For T ≠ 0, the partition function Z of a hydrogen atom includes contributions from an infinite number of energy levels, with degeneracies growing as 2n². This leads to a divergence because the terms e^(−βE_n) do not decay quickly enough as n increases. Confinement introduces a high-energy cutoff by restricting the atom to finite spatial dimensions, reducing available states and thus truncating the series, preventing divergence. This cutoff effectively limits contributions to the partition function from very high energy levels, which would otherwise dominate in an unconfined system .

The Sackur-Tetrode equation for the entropy of a gas indicates an entropy expression, S = Nk_B (5/2 - ln(nλ³_th)), that is extensive, meaning it scales with the system size, N. However, for distinguishable particles, the derived entropy S = Nk_B (3/2 - ln(λ²_th/V)) is not extensive, as it does not scale linearly with volume. This discrepancy relates to Gibbs paradox, highlighting that the assumption of distinct particle labels leads to unphysical predictions. Recognizing this non-extensivity assisted in formulating the principle of indistinguishability in quantum mechanics, making correct statistical descriptions possible .

For N non-interacting spin-1/2 particles in a magnetic field, the divergence of the partition function implies an unrealistic prediction of infinite thermodynamic quantities at low temperatures. This divergence arises due to the assumption of infinite potential energy states contributing to the sum in quantum systems. Confinement to physical or realistic bounds, like finite particle numbers or discrete energy level limits, prevents this divergence, ensuring that calculations of internal energy, U = −NµBB tanh(µBB/k_BT), and related properties such as magnetization remain finite and measurable. This constraint is pivotal for accurately modeling real-world magnetic systems using statistical mechanics .

The derivation starts by maximizing the Gibbs-Shannon entropy under given constraints for the system's probability distributions and average energy. Using the method of Lagrange multipliers, one introduces multipliers for these constraints and differentiates to find the stationary points. This yields probability distributions Pi = e^(-βEi)/Z, where Z is the partition function. The entropy expression S/k_B = ln Z + βU comes from substituting these probability distributions back into the Gibbs-Shannon formula, reflecting the dual dependence of entropy on the partition function and internal energy .

The partition function for a two-level atom, Z_atom = g1 + g2e^(-βΔ), depends on the degeneracies of its energy states. With non-zero degeneracies (g1 for the ground state, g2 for the excited state), the number of effective ways to occupy each energy level increases, thus impacting the partition function's value and thermodynamic properties. The heat capacity contribution, Cv = g1g2Δ²e^(-βΔ) / (k_BT²(g1 + g2e^(-βΔ))²), reflects how these degeneracies influence energy distribution and fluctuations. Greater degeneracy typically lowers energy quantization effects, leading to smoother transitions in thermodynamic quantities .

The moment of inertia I is crucial in determining the rotational energy levels, E_j = ħ²/(2I) j(j + 1), for a diatomic molecule. It affects the spacing of these energy levels, which determines at what temperatures transitions between them significantly affect the partition function. Small moments of inertia lead to closely spaced levels, affecting specific heat at low temperatures as the population of excited states becomes thermally accessible. At high temperatures, the specific heat approaches a classical limit, but at low temperatures, the quantum mechanical spacing significantly alters the temperature dependence of Cv .

Negative temperatures occur in systems with bounded energy states, such as ideal paramagnets in strong magnetic fields, where entropy can decrease as energy increases. Here, unlike typical systems where adding energy increases disorder, increasing energy aligns more spins, hence decreasing entropy. This gives rise to a negative derivative in entropy with respect to energy, effectively defining negative temperature. Physically, a negative temperature system is 'hotter' than any positive temperature system because it will give up energy if brought into contact with a positive temperature system, causing unique thermodynamic behaviors such as absorbing energy rather than dissipating it upon contact with higher positive temperature systems .

The Gibbs expression for entropy, S = −kB Σi Pi ln Pi, is foundational in statistical mechanics as it provides a way to quantitatively describe entropy in terms of probability distributions. To derive the entropy of mixing, one applies this expression to a system of two or more different ideal gases. The process involves calculating the change in entropy when these gases, initially separated, are mixed. It shows that entropy increases due to the increase in possible microstates when the system is mixed, quantitatively described by the change in the sum of compositions times the natural logarithm of their compositions .

At high temperatures, where k_BT ≫ ħω, the energy spacing between quantum states becomes negligible compared to thermal energy, leading to a behavior similar to a classical system. In this regime, the partition function for a harmonic oscillator can be approximated as Z ≈ (βħω)^(-1), indicating a simplification due to reduced quantum effects. The specific heat, Cv = k_B, reaches the classical limit, indicating equal partition of energy across degrees of freedom. This demonstrates how quantum systems revert to classical statistical mechanics at sufficiently high temperatures .

The partition function encapsulates all the statistical properties of a system by summing over contributions from all possible energy states. In a magnetic material at high temperatures, the thermal energy k_BT easily overcomes small energy differences caused by external magnetic fields, simplifying the partition function. The system then follows classical statistical distributions, and the average magnetization can be predicted from thermal averages, where magnetization, M = mB/k_BT, becomes proportional to the applied field, reflecting the alignment tendency of magnetic moments that is described by the partition function .