A large thermal energy reservoir in the nascent intracluster medium at a redshift of 4.3

Most baryons in present-day galaxy clusters exist as hot gas ($\boldsymbol{\gtrsim10^7,\rm}\mathrm{K}$), forming the intracluster medium (ICM). Cosmological simulations predict that the mass and temperature of the ICM rapidly decrease with increasing cosmological redshift, as intracluster gas in younger clusters is still accumulating and being heated. The thermal Sunyaev-Zeldovich (tSZ) effect arises when cosmic microwave background (CMB) photons are scattered to higher energies through interactions with energetic electrons in hot ICM, leaving a localized decrement in the CMB at a long wavelength. The depth of this decrement is a measure of the thermal energy and pressure of the gas. To date, the effect has been detected in only three systems at or above $z\sim2$, when the Universe was 4 billion years old, making the time and mechanism of ICM assembly uncertain. Here, we report observations of this effect in the protocluster SPT2349$-$56 with Atacama Large Millimeter/submillimeter Array (ALMA). SPT2349$-$56 contains a large molecular gas reservoir, with at least 30 dusty star-forming galaxies (DSFGs) and three radio-loud active galactic nuclei (AGN) in a 100-kpc region at $z=4.3$, corresponding to 1.4 billion years after the Big Bang. The observed tSZ signal implies a thermal energy of $\mathbf{\sim 10^{61},\mathrm{erg}}$, exceeding the possible energy of a virialized ICM by an order of magnitude. Contrary to current theoretical expectations, the strong tSZ decrement in SPT2349$-$56 demonstrates that substantial heating can occur and deposit a large amount of thermal energy within growing galaxy clusters, overheating the nascent ICM in unrelaxed structures, two billion years before the first mature clusters emerged at $\mathbf{z \sim 2}$.

💡 Research Summary

In this paper the authors present the first detection of the thermal Sunyaev‑Zeldovich (tSZ) effect in a protocluster at redshift z = 4.3, the object SPT‑2349‑56. The system is an extreme environment: within a 100 kpc region it hosts at least thirty dusty star‑forming galaxies (DSFGs) and three radio‑loud active galactic nuclei (AGN), producing a total star‑formation rate of roughly 5 000 M⊙ yr⁻¹ and residing in a dark‑matter halo of order 10¹³ M⊙.

Deep ALMA Band‑3 observations (combined 12‑m and ACA 7‑m arrays) were obtained over more than 50 hours of on‑source integration. After careful calibration, the authors removed the continuum emission from the DSFGs using a Fourier‑space subtraction technique that accounts for the steep spectral slope of dust. Imaging the short‑baseline visibilities (uv < 10 kλ) revealed a strong, spatially extended decrement centred on the protocluster core. The signal peaks at 8.4 σ in the image plane and 10.4 σ in Fourier space, with an integrated flux density of –157 ± 16 µJy.

From the measured decrement the Compton‑y parameter is (5.6 ± 0.8) × 10⁻⁶, corresponding to a Compton‑Y of (2.0 ± 0.2) × 10⁻⁶ arcmin². Using the standard relation between Y and the line‑of‑sight integrated electron pressure, the authors infer a total thermal energy of the intracluster medium (ICM) of ≈10⁶¹ erg. This value exceeds the thermal energy expected for a virialized ICM in a 10¹³ M⊙ halo by roughly an order of magnitude.

The authors compare the result with the well‑established self‑similar Y–M scaling relation observed for lower‑redshift clusters. After correcting for the redshift dependence (E(z)²⁄³), SPT‑2349‑56 lies a factor of five above the relation, a deviation that is statistically significant (≈6.4 σ). They also examine predictions from the TNG‑Cluster cosmological simulation suite, which reproduces the self‑similar evolution at z ≲ 3 but predicts a systematic decline in Y/M at higher redshift, reflecting a cooler, less massive ICM in protoclusters. The observed Y for SPT‑2349‑56 is therefore in strong tension with these simulations.

Two broad explanations are considered. First, the halo mass could be substantially larger than current estimates (by a factor ≳ 3), which would bring the system onto the scaling relation. However, such a mass increase would conflict with independent dynamical and lensing constraints. The second, more plausible scenario invokes intense pre‑heating of the ICM by the combined energy output of the vigorous star formation and the three powerful AGN. At z > 4 the ambient gas density and pressure are high, confining AGN jets and enhancing the coupling of mechanical energy to the surrounding medium. This leads to an over‑pressurized, overheated ICM that retains a large fraction of the injected energy, producing the strong tSZ signal observed.

The authors argue that current hydrodynamical simulations, even those with sophisticated galaxy‑formation models, fail to reproduce such extreme heating because their sub‑grid AGN feedback prescriptions are not calibrated for the high‑redshift, high‑density environment of protoclusters. Consequently, the discovery points to a missing physical ingredient in models of early cluster assembly.

Finally, the rarity of SPT‑2349‑56 (the only object of its kind found in a 2 500 deg² survey) raises the question of how common such overheated nascent ICMs are. The authors suggest that either this system is an outlier or it represents a short‑lived but crucial phase of cluster formation that has been largely overlooked. They call for larger high‑z tSZ surveys and complementary high‑resolution X‑ray or molecular line observations to assess the prevalence of this phenomenon and to refine feedback models in the early Universe.