Axino dark matter from thermal production

The thing to know: this is the paper that turned the thermal axino production rate into a real calculation rather than an order-of-magnitude estimate. Earlier work had a rate that depended on an arbitrary infrared cutoff; Brandenburg and Steffen do the HTL resummation properly, split hard and soft momenta, and get a finite, gauge-invariant answer that includes Debye screening. That is the contribution, and it holds up. The relic abundance that follows is smaller than the earlier estimates, which shifts the (m_axino, T_R) window. For 100 keV and 10⁶ GeV you land near the observed Omega h², and the abstract's 'could provide the dominant part' is the right hedge, not a dodge. The derivation is from the SUSY QCD Feynman rules plus resummation, not fitted to the abundance, so the circularity worry does not land here. What is not new is the cosmology: instantaneous reheating at T_R, radiation domination, no late entropy production, axino as stable LSP with R-parity. The paper states all of that. The axino mass and T_R are free inputs, and the PQ scale enters through the axino-gluino-gluon coupling, so this is a parameter-space statement, not a prediction of Omega_DM. That is normal for the literature and the paper is explicit about it. The soft spot is the one the stress-test note flags: the matching between hard and soft regions leaves a residual scale dependence, and the paper discusses it without turning it into a quantified error band. How much does it matter? The qualitative conclusion survives a factor-of-two shift in the rate; the specific quoted point does not. I would call that a moderate caveat, not a load-bearing flaw, and it is the kind of thing you check yourself if you plan to use the number. The citation pattern is standard, no padding. Who it is for: anyone working on non-WIMP dark matter, axion and axino cosmology, or thermal production in a SUSY plasma. It is a reference calculation, cited more than read cover to cover. Peer review: yes, this deserved referees in 2004 and would today. It is a careful technical paper with a real result.

The paper's central quantitative claim is that for m_axino = 100 keV and T_R = 10^6 GeV, thermally produced axinos give Ω h^2 ~ 0.1. This number is obtained by integrating the HTL-resummed rate over temperature. The rate is computed by splitting the phase space into hard (p ~ T) and soft (p ~ gT) regions, with the soft region treated via HTL resummation and the hard region via bare propagators. The matching between the two introduces a dependence on the matching scale, which the paper discusses but does not quantify. If the matching-scale uncertainty is, say, a factor of 2 in the rate, the quoted window shifts by a factor of 2 in T_R or m_axino. The paper's qualitative conclusion ('could provide the dominant part') is robust to such a shift, but the specific quoted point is not. This is a real technical caveat that the paper acknowledges but doesn't resolve.

The paper's technical contribution is real: the HTL resummation removes the IR cutoff dependence of earlier estimates and yields a finite, gauge-invariant rate. The relic abundance calculation is standard. The main caveat is the matching-scale dependence, which the paper discusses but doesn't quantify. The cosmological assumptions are explicitly stated. The axino mass is a free parameter. The paper's claim is appropriately hedged. So the verdict should remain ACCEPT, with the caveat that the quoted point should be understood as having a quantitative uncertainty from the matching-scale dependence. The reader's weakest_assumption (cosmological history) is a real concern but is acknowledged; the more specific technical concern is the matching-scale dependence, which the reader mentions in the rationale but not in the weakest_assumption field.

This is a careful and well-motivated calculation that repairs a known defect of earlier estimates — the IR divergence and gauge dependence of the naive axino production rate — with a standard and appropriate resummation. The resulting rate is a reusable ingredient for axino cosmology, and the (m_axino, T_R) window is a concrete target for SUSY/PQ parameter scans. The authors are explicit that the axino mass and reheating temperature are free inputs and that the axino is assumed to be the LSP with conserved R-parity, so the paper does not overclaim a unique prediction. The main strengths are the gauge-invariant treatment of the soft sector and the honest statement of the cosmological assumptions.

The paper asks whether the axino, the fermionic superpartner of the axion, can be the dark matter if it is the lightest supersymmetric particle and R-parity is unbroken. Its central move is to compute the axino production rate from scatterings in the hot quark-gluon-squark-gluino plasma of supersymmetric QCD rather than treating that rate as a free parameter. Hard thermal loop resummation is used to keep the result finite and gauge invariant while including Debye screening. Integrating that rate over a standard post-inflationary thermal history gives a relic abundance, and the authors find that a 100 keV axino with a reheating temperature of $10^6$ GeV can supply most of cold dark matter.

What carries the argument. The load-bearing object is the hard-thermal-loop-resummed axino production rate from two-to-two scatterings in the quark-gluon-squark-gluino plasma. Ordinary perturbation theory gives a rate that is sensitive to infrared physics; HTL resummation reorganizes the plasma propagators so that Debye screening cuts off the would-be divergences and the result is finite and gauge invariant. The paper then feeds this rate into the Boltzmann equation for the axino number density and integrates from reheating to the present to get $Omega_{a} h^2$.

Core claim. The paper claims that thermal scatterings after inflation can produce enough axinos to be the dominant cold dark matter, with a concrete example at $m_{a}=100$ keV and $T_R=10^6$ GeV. The enabling result is a finite, gauge-invariant axino production rate in supersymmetric QCD obtained with hard thermal loop resummation, which accounts for Debye screening in the plasma. The relic abundance follows from integrating that rate through the radiation-dominated era. This turns the axino from a merely plausible dark-matter candidate into one with a calculable thermal abundance in a specified cosmological history.

The premise it rests on. The relic abundance is computed for a standard thermal history: instantaneous reheating at $T_R$, radiation domination thereafter, no late entropy production, and an axino that is the stable lightest supersymmetric particle with R-parity conserved.

For an axino mass around 100 keV, a reheating temperature of order $10^6$ GeV gives a thermal axino abundance comparable to the cold dark matter density.

Because the production rate is now finite and gauge invariant, the relic abundance can be used to place a thermal-history-dependent upper bound on $T_R$ for a given axino mass.

The calculation identifies Debye screening in the quark-gluon-squark-gluino plasma as the infrared regulator that makes the supersymmetric-QCD production rate predictive.

The viable window depends on the axino mass and reheating temperature, so a measurement of either would test the scenario.

If the axino is the lightest supersymmetric particle, its thermal abundance competes with the usual gravitino problem: the same reheating temperature that makes axinos abundant can also make gravitinos, so the allowed $T_R$ window may be narrower than the axino-only calculation suggests.

A nonstandard thermal history with late entropy production would dilute the thermal axino abundance and shift the 100 keV and $10^6$ GeV point, making the scenario sensitive to the same physics that solves the gravitino problem.

The 100 keV mass is low enough that free-streaming may matter for small-scale structure, so a dedicated warm-dark-matter analysis would test whether the cold label is safe.

An improved HTL calculation including higher-order corrections or a fuller Boltzmann treatment could move the quoted abundance, so the window should be read as a leading-order result.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review