Supermassive black holes in galaxy centres can grow by the accretion of gas, liberating energy that might regulate star formation on galaxy-wide scales 1-3 . The nature of the gaseous fuel reservoirs that power black hole growth is nevertheless largely unconstrained by observations, and is instead routinely simplified as a smooth, spherical inflow of very hot gas 4 . Recent theory 5-7 and simulations 8-10 instead predict that accretion can be dominated by a stochastic, clumpy distribution of very cold molecular clouds -a departure from the 'hot mode' accretion model -although unambiguous observational support for this prediction remains elusive. Here we report observations that reveal a cold, clumpy accretion flow towards a supermassive black hole fuel reservoir in the nucleus of the Abell 2597 Brightest Cluster Galaxy (BCG), a nearby (redshift z = 0.0821) giant elliptical galaxy surrounded by a dense halo of hot plasma [11][12][13] . Under the right conditions, thermal instabilities can precipitate from this hot gas, producing a rain of cold clouds that fall toward the galaxy's centre 14 , sustaining star formation amid a kiloparsec-scale molecular nebula that inhabits its core 15 . The observations show that these cold clouds also fuel black hole accretion, revealing 'shadows' cast by the molecular clouds as they move inward at about 300 kilometres per second towards the active supermassive black hole in the galaxy centre, which serves as a bright backlight. Corroborating evidence from prior observations 16 of warmer atomic gas at extremely high spatial resolution 17 , along with simple arguments based on geometry and probability, indicate that these clouds are within the innermost hundred parsecs of the black hole, and falling closer towards it. We observed the Abell 2597 Brightest Cluster Galaxy (Fig. 1) with the Atacama Large Millimeter/submillimeter Array (ALMA), enabling us to create a three-dimensional map of both the location and motions of cold gas at uniquely high sensitivity and spatial resolution. The ALMA receivers were sensitive to emission from the J = 2 − 1 rotational line of the carbon monoxide (CO) molecule. CO(2-1) emission is used as a tracer of cold (∼ 10 − 30 K) molecular hydrogen, which is vastly more abundant, but not directly observable at these low temperatures.The continuum-subtracted CO(2-1) images (Fig. 2) reveal that the filamentary emission line nebula that spans the galaxy's innermost ∼ 30 kpc (Fig. 1b) consists not only of warm ionised gas [18][19][20] , but cold molecular gas as well. In projection, the optical emission line nebula is cospatial and morphologically matched with CO(2-1) emission detected at a significance between > ∼ 3σ (in the outer filaments) and > ∼ 20σ (in the nuclear region) above the background noise level. The warm ionised nebula is therefore likely to have a substantial molecular component, consistent with results for other similar galaxies 21 . The total measured CO(2-1) line flux corresponds to a molecular hydrogen gas mass of M H 2 = (1.8 ± 0.2) × 10 9...

The universe's largest galaxies reside at the centers of galaxy clusters and are embedded in hot gas that, if left unchecked, would cool prodigiously and create many more new stars than are actually observed. 1-5 Cooling can be regulated by feedback from accretion of cooling gas onto the central black hole, but requires an accretion rate finely tuned to the thermodynamic state of the hot gas. 6,7 Theoretical models in which cold clouds precipitate out of the hot gas via thermal instability and accrete onto the black hole exhibit the necessary tuning. [8][9][10] We have recently presented observational evidence showing that the abundance of cold gas in the central galaxy increases rapidly near the predicted threshold for instability. 11 Here we present observations showing that this threshold extends over a large range in cluster radius, cluster mass, and cosmic time, and incorporate the precipitation threshold into a comprehensive framework of theoretical models for the thermodynamic state of hot gas in galaxy clusters. According to that framework, precipitation regulates star formation in some giant galaxies, while thermal conduction prevents star formation in others, if it can compensate for radiative cooling and shut off precipitation.Our framework can be expressed in terms of the time t cool required for the hot gas to radiate an amount of energy equivalent to its current thermal energy. If intracluster gas were unable to cool, cosmological structure formation via hierarchical merging would produce galaxy clusters with radial cooling-time profiles similar to a baseline profile t base (r) that can be computed with numerical simulations. 12,13 Massive galaxy clusters are observed to converge to this baseline profile at large radii, 14 but radiative cooling cannot be ignored at smaller radii, where t cool can be much shorter than the age of the universe. Gas at small radii must either cool and condense or cooling of that gas must trigger thermal feedback that compensates for the radiative losses. 15Thermal conduction is capable of compensating for cooling in cluster gas with t cool > 1 Gyr. 16, 17 Our framework therefore includes a locus of conductive balance, t cond (r), along which thermal conduction exactly balances radiative cooling. 18 The locus itself is unstable, because conduction outcompetes cooling if t cool is above that locus but cannot compete below it. 19 Conduction should therefore drive gas above the locus toward an isothermal core profile t iso (r) identical to the baseline profile at large radii but with a constant temperature equal to the peak temperature of the baseline profile at smaller radii. Clusters in an isothermal core state have central cooling times exceeding ~1 Gyr, and so mergers with other galaxy clusters, which occur on timescales of several Gyr, can compete with cooling and further raise t cool in the cores of those objects. Once t cool exceeds the 14 Gyr age of the universe, radiative cooling can no longer lower t cool , and this threshold corresponds to the "no cooling" profi...

Recent observations show that the star formation rate (SFR) in the Phoenix cluster’s central galaxy is ∼500 M⊙ yr−1. Even though Phoenix is a massive cluster (M200 ≈ 2.0 × 1015 M⊙; z ≈ 0.6) such a high central SFR is not expected in a scenario in which feedback from an active galactic nucleus (AGN) maintains the intracluster medium in a state of rough thermal balance. It has been argued that either AGN feedback saturates in very massive clusters or the central supermassive black hole is too small to produce enough kinetic feedback and hence is unable to quench the catastrophic cooling. In this work, we present an alternate scenario wherein intense short-lived cooling and star formation phases followed by strong AGN outbursts are part of the AGN feedback loop. Using results from a 3D hydrodynamic simulation of a standard cool-core cluster (M200 ∼ 7 × 1014 M⊙; z = 0), scaled to account for differences in mass and redshift, we argue that Phoenix is at the end of a cooling phase in which an AGN outburst has begun but has not yet arrested core cooling. This state of high cooling rate and star formation is expected to last for ≲100 Myr in Phoenix.

There are (at least) two unsolved problems concerning the current state of the ther- mal gas in clusters of galaxies. The first is to identify the source of the heating which onsets cooling in the centres of clusters with short cooling times (the 'cooling-flow' problem). The second to understand the mechanism which boosts the entropy in cluster and group gas. Since both of these problems involve an unknown source of heating it is tempting to identify them with the same process, particularly since active galactic nuclei heating is observed to be operating at some level in a sample of well-observed 'cooling-flow' clusters. Here we show, using numerical simulations of cluster formation, that much of the gas ending up in clusters cools at high redshift and so the heating is also needed at high redshift, well before the cluster forms. This indicates that the same process operating to solve the cooling-flow problem may not also resolve the cluster-entropy problem.

No abstract