Influence of a temperature-dependent shear viscosity on the azimuthal asymmetries of transverse momentum spectra in ultrarelativistic heavy-ion collisions

Bottom line: this is a solid, careful viscous-hydro paper, but the headline claim in the abstract is a restatement of the authors' own PRL (Ref. [9]), and they say so. The genuinely new material is the temperature-localization scan of Sec. VII, the τπ-versus-η/s degeneracy in Fig. 15, and the v4 analysis. Those are the parts worth your time; the RHIC-versus-LHC dominance story you already know from their earlier work. The paper is published in PRC and is a standard reference for temperature-dependent η/s, so it is not going anywhere, but read it for the scan, not the abstract. What it does well: the transport coefficients are given in closed form (Eqs. 9, 10, 11), the initial-state parameters are tabulated (Table I), the EoS is a named lattice parameterization, and the freeze-out is standard Cooper-Frye with the 14-moment δf. The authors are candid about their own weak points — the δf ansatz, the smooth optical-Glauber initial conditions, the LHC low-pT v2 undershoot. The τπ degeneracy result (Fig. 15: raising η/s_min from 0.08 to 0.16 can be largely compensated by cτ = 10) is a real caution for anyone doing Bayesian extraction of η/s from v2 alone, and I would cite it for that. Soft spots, in proportion. First, no uncertainty quantification anywhere: every curve is one deterministic run, and the centrality-dependence claims rest on visual comparison. Standard for 2012, but worth flagging. Second, the per-parameterization multiplicity normalization of Sec. III C is a tuned knob that couples to the effect under study. Third, and this is the one that matters for the headline: the Sec. VII scan never states whether the initial energy density is renormalized to fixed multiplicity per Ti, whereas Sec. III C explicitly does so for the four main parameterizations. The stress-test concern lands as an under-specification, not a fatal flaw — the bump is narrow (δT = 10 MeV), and the observed peak at 150–200 MeV sits far below Tmax = 313/430/504 MeV, where entropy-production contamination would be largest, so a pure global artifact is unlikely. But with δv2/v2 of only 2–8%, a few-percent systematic could shift the peak by 20–30 MeV, and that shift is the quantitative content of abstract:s4. A referee should ask for the per-Ti normalization to be stated and the scan shown to be robust to it. Who this is for: hydro practitioners and anyone fitting η/s to flow data. It deserves a serious referee.

The scan of Sec. VII modifies the baseline η/s = 0.08 by a factor of roughly three over a ±10 MeV window around Ti (Eq. 16 with δT = 10 MeV, Δ = 1.5 MeV), then reads off δvn/vn at each Ti. The problem is that a bump placed at high Ti sits in the early, high-gradient stage where, as the paper itself says in Sec. III C, 'most of the entropy is produced during the early stages of the collision'; a bump placed at low Ti sits in the late, dilute stage. Total entropy production therefore varies systematically and monotonically with Ti across the scan, dragging the fireball size, lifetime, and radial flow with it. Sec. III C explicitly compensates normalization between the four main parameterizations ('We compensate this using different overall normalizations'), but the scan uses a different baseline (constant 0.08 plus a narrow bump) and the text never says whether the same compensation is applied per Ti. If it is not, part of every δvn point is a global size/lifetime effect that grows with Ti, which would bias the extracted 'most sensitive temperature' upward and could in principle manufacture a peak. The effect sizes are small (δv2/v2 of order 2–8% in Fig. 19), so a few-percent systematic is not negligible relative to the signal. There is partial internal counter-evidence: the observed peak sits at 150–200 MeV, far below Tmax = 313/430/504 MeV (Table I), where entropy-production contamination would be largest, so a pure global artifact is unlikely. But the peak's location and width are the quantitative content of abstract:s4, and a 20–30 MeV shift would change the 'narrow band around the transition' reading. This is the one place where the headline claim's method is under-specified.

The reader's weakest_assumption identifies exactly the right soft spot: the Sec. VII scan is the sole support for abstract:s4, and the text is silent on whether the initial energy density is renormalized per Ti, in contrast to the explicit compensation described in Sec. III C for the four main parameterizations. That asymmetry in bookkeeping is the load-bearing gap, and it is worth flagging because the scan's signal is only a few percent in δv2/v2. I would not, however, move the verdict on it. The paper's own Fig. 19 shows the peak at 150–200 MeV rather than near Tmax, which is the opposite of what a monotonic entropy-production artifact would produce; the qualitative conclusion therefore has internal support independent of the missing statement. The reader's other three points are weaker as load-bearing concerns. The absence of uncertainty quantification is real but is a general practice issue in this literature rather than a defect specific to the argument, and the paper's claims are ordinal (which phase dominates, where the sensitivity peaks) rather than quantitative fits. The per-parameterization normalization of Sec. III C is a tuned knob, but the paper discloses it and shows in Fig. 13 that the grouping of v2(pT) curves survives a Navier-Stokes shear initialization, which is a genuine robustness check. The overlap with Ref. [9] is disclosed in the introduction and in Sec. V, so the novelty is correctly located in the scan and the relaxation-time degeneracy of Fig. 15, not in the RHIC-versus-LHC dominance statement. One correction to the record: the classifier reports 'limitations the authors state: none found', but the text does acknowledge several — the 14-moment δf is called 'merely an ansatz' in Sec. II C, the smooth optical-Glauber initial conditions are flagged in Sec. IV A and Sec. VII, the LHC low-pT v2 undershoot is noted in Sec. IV B, and the Ti = 110 MeV v4 point is explicitly excluded as a freeze-out δf artifact in Sec. VII. Those acknowledgments should be credited rather than re-raised. The remaining unresolved item is the multiplicity matching of the scan, and the test above settles it cheaply. Verdict unchanged: CONDITIONAL, with the Sec. VII bookkeeping as the specific item to confirm.

This is a careful and systematic study that extends the authors' earlier work with a detailed decomposition of the temperature regions probed by v2 and v4 at three collision energies. The finding that flow data constrain η/s mainly in a narrow band around the QCD transition, and that higher harmonics are sensitive to the hadronic viscosity and to the δf correction rather than to the high-temperature viscosity, is a useful and falsifiable prediction for the field. The paper is transparent about its approximations (14-moment δf, smooth initial conditions, optical Glauber, isentropic low-temperature equation of state) and checks several of them explicitly. The main caveat is that the Sec. VII temperature scan does not state whether the initial energy density is renormalized to fixed multiplicity per Ti, which is load-bearing for the quantitative peak location.

When atomic nuclei collide near the speed of light, the resulting fireball of quarks and gluons expands and cools in about 10^-23 seconds. How freely that fireball flows depends on one number, the ratio of shear viscosity to entropy density, written as eta over s. A small value means a nearly perfect fluid. Experiments infer this number indirectly, from how the debris is pushed sideways, a pattern called elliptic flow. This paper asks a sharper question: viscosity is not a single number but a function of temperature, and the fireball sweeps through a wide temperature range. Which part of that range actually controls the observed flow? The authors run a viscous hydrodynamic simulation of the collision with four different temperature profiles for eta over s, and compare the resulting particle spectra and flow patterns against data from RHIC and the LHC. The answer: at RHIC energies the flow is set mostly by viscosity in the cooler hadron phase, below the transition to ordinary matter. At the higher LHC energies the hot quark-gluon phase starts to matter. Across all energies, the temperature window around 150 to 200 MeV, right at the transition, has the largest single influence on both elliptic flow and its fourth harmonic. The authors also show that changing the relaxation time of the viscous stress can mimic a change in eta over s itself, so the two cannot be separated from flow data alone.

Core claim. The abstract states: 'the shear viscosity near the QCD transition region has the largest impact on the collective flow of the system' (abstract:s4), together with 'The elliptic flow coefficient turns out to be dominated by the hadronic viscosity at RHIC energies' (abstract:s2). Concretely, Sec. VII scans a Gaussian-in-temperature bump in eta over s across Ti and finds that the relative change in v2 and v4 peaks near T = 150 to 200 MeV at all three collision energies, with the peak shifting slightly upward for v2 as energy rises. If correct, this says flow data constrain eta over s mainly in a narrow band around the transition, not uniformly across the fireball's temperature history.

The premise it rests on. The Sec. VII temperature scan assumes that the change in vn produced by a bump in eta over s at temperature Ti can be attributed to the local viscosity at Ti. But the bump also raises total entropy production and therefore the final multiplicity, and the text does not state that the initial energy density is renormalized to fixed multiplicity for this scan, unlike the four main parameterizations where Sec. III C explicitly says 'We compensate this using different overall normalizations.' If the scan is not multiplicity-matched, part of each delta-vn point reflects a global change in the fireball's size and lifetime rather than a local viscosity effect at Ti, which would blur the claimed temperature localization.

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