Reorganize perturbation theory for soft thermal modes

Consider a stable real scalar field in four spacetime dimensions with interaction $\mathcal L_{\mathrm{int}}=-\lambda\phi^4/4!$ and positive weak coupling. Subtract or renormalize the vacuum contribution separately. For temperature large compared with the vacuum mass, the leading one-loop thermal tadpole gives a positive mass-squared correction. The factor one half multiplying the loop integral is tied to this precise quartic normalization. A reference using a quartic coefficient divided by four instead of factorial four will quote a different numerical coupling.

For massless internal modes at leading order, the thermal integral contains $n_B(p)/p$. Its angular integration and the elementary Bose integral give a quantity proportional to $T^2$. Combining the integral with the tadpole factor yields $\delta m_T^2=\lambda T^2/24$. This result is an equilibrium high-temperature leading approximation for this scalar theory. It is not a universal thermal mass for every boson, and it does not determine a collision width or a transport relaxation rate.

The Euclidean static propagator contains an inverse denominator $k^2+m_0^2+\Pi_T$. Expanding it in insertions of the thermal self-energy produces successive powers of $\Pi_T/(k^2+m_0^2)$. At hard momentum of order T this ratio is of order lambda. At soft momentum of order square root lambda times T, the denominator can be the same size as the thermal correction, so every insertion matters. The failure is a scale problem, even when the underlying interaction coupling remains small.

Resumming the leading insertions replaces the bare soft propagator by one with the thermal mass retained in its denominator. A systematic reorganized expansion adds that term to the quadratic action and subtracts it as an interaction counterterm. The subtraction removes contributions already accounted for by the new free propagator. Merely changing every mass in a diagram and then also adding the original mass correction would double count. Specify the expansion order and the regions being treated before attributing a numerical improvement to resummation.

A static screening mass describes the spatial falloff of an equilibrium perturbation. A pole mass or dispersion relation follows from a real-time propagator, while a width comes from its dissipative structure. Their leading values can coincide in a simple scalar approximation without being identical definitions. A momentum-dependent self-energy can separate them. Gauge theories require additional consistency conditions and collective response rather than a scalar mass substitution for every component; the later hard-thermal-loop lesson treats that distinction.

Resummation does not automatically control a critical region. If the effective curvature becomes very small, long-wavelength fluctuations can invalidate the reorganized loop expansion. Large occupations, another small denominator, or missing interaction corrections may also change the power counting. Useful checks include comparing successive orders, testing dependence on matching choices, and identifying the remaining scale hierarchy. A finite-looking resummed formula can still lie outside its regime, so removing a divergence is not by itself proof of a quantitatively controlled prediction.