Besides the large surveys with statistical value to explore galaxy evolution with redshift, the discovery of special cases, pointed observations of high-z quasars, and the study of over-densities, have provided a wealth of information.
5.1. Starburst and quasar associations
Objects already known as SMG with single dish continuum detectors were easily detected with ALMA, like this association of three LBG at z = 5.3 studied by Riechers et al (2014). From the lines of [CII] and OH detected, an SFR surface density of 530 M⊙/yr/kpc2 was derived, implying a disk approaching the Eddington limit for radiation pressure on dust. Since OH is slightly blue-shifted with respect to [CII], this might indicate a molecular outflow due to SN feedback. Swinbank et al (2014) have made a survey with ALMA in the Extended Chandra Deep Field South (ECDFS) of 99 SMG, and found that they are all ULIRGs with SFR ∼ 300 M⊙/yr and dust temperature of 32K. The contribution of these SMG to the cosmic star formation is about 20% over z = 1-4.
High redshift starbursts can be detected serendipitously, like the bright z = 5.24 lensed submillimeter galaxy in the field of Abell 773 (Combes et al, 2012), as part of the Herschel Lensing Survey (HLS, Egami et al 2010). This project surveyed a series of nearby galaxy clusters at z ∼ 0.1-0.5 playing the role of gravitational telescopes, amplifying background galaxies. These were selected by their very red SPIRE colours, implying a high redshift. Follow-up at millimeter wavelengths allows to discover the spectroscopic redshift, with the help of at least two detected lines. In this case several CO lines up to CO(7-6), CI, H2O and the [NII] 205 µm lines were discoverd, and allowed to constrain the variations of fundamental constants (Levshakov et al, 2012). With ALMA, spectroscopic redshifts are obtained routinely.
An hyperluminous quasar at z = 4.4 slected from WISE-SDSS was detected with ALMA by Bischetti et al (2018) in dust continuum and [CII] line. It is at the center of a proto-cluster, merging with two close companions. The quasar is actively forming stars with SFR ∼ 100 M⊙/yr, and the host galaxy will increase its stellar mass more rapidly than its black hole mass, which is observed 2 orders of magnitude too massive, with respect to local relations.
Venemans et al (2017) have detected several CO lines, CI and [CII] in z ∼ 7 quasars, and shown that these lines have excitation compatible with photodissociation regions, but not X-ray dominated regions. The properties of the molecular gas and dust in these quasars are dominated by an important star-formation activity, confirming that intense starbursts are co-existing with AGN activities.
5.2. Lensed high-z galaxies, high spatial resolution and GMC studies
ALMA can have very high spatial resolution in its extended configuration, up to 15-20 milli-arcsec (mas). A remarquable object was observed to demonstrate these capabilities, with baselines up to 15km: SDP.81 (ALMA Partnership et al, 2015). This gravitationally lensed galaxy at z = 3.042 was discovered by the Herschel survey H-ATLAS (Eales et al, 2010, Negrello et al, 2010). Its redshift was determined through CO lines; the lensing galaxy is at z = 0.299, and the amplification factor is µ = 11 (Bussmann et al, 2013). With a beam of 25 mas at 1mm (180pc at z = 3.042), the ALMA continuum map reveals the two gravitational arcs with unprecedented sharpness. Figure 9 show a tapered version of the maps in continuum, CO and H2O lines; the images have been tapered to lower resolution to gain more signal to noise. The two arcs are part of an Einstein ring, of radius 1.5". The forefround lensing galaxy is invisible on these images, except for a weak continuum source at the center of the ring, which has a spectral index consistent with synchrotron emission. The lensing galaxy is a massive elliptical (3.6 1011 M⊙ inside the Einstein ring of 1.5" = 6.7 kpc, at z = 0.299, and no AGN is detected optically. But the 1.4 GHz flux is compatible with the mm spectral index, and corresponds to an AGN radio core. The continuum from the arcs comes from dust emission in the high-z star forming galaxy, with an SFR = 527 M⊙/yr. The three CO lines detected (from J = 5, 8 and 10) show regions in the galaxy of different excitation, implying a complex structure. The H2O emission comes from a thermal line, which ratio with the CO lines is rather weak, may be due to differential lensing. The wealth of details acquired in ∼ 30h of telescope time in early science with only 22 to 36 antennae is quite impressive.
 |
Figure 9. ALMA images with high resolution (CO lines and continuum, 100-170 mas) or lower resolution (H2O line, 900 mas). Top: CO J = 5 - 4, 8 - 7, and 10 - 9 integrated intensity. Middle: 2.0, 1.3, and 1.0 mm continuum. Bottom: Band 6 and 7 spectral index, 1.14 mm continuum (combined Band 6 and 7 data, and H2O integrated intensity. The beam sizes are indicated by the black ellipses at the bottom of the panels. Image reproduced with permission from ALMA Partnership et al (2015), copyright by AAS. |
Given the enhanced spatial resolution due to lensing, it is possible to explore the resolved Kennicutt-Schmidt relation (KS) in these high-z galaxies. The surface densities of the molecular gas and star formation rate have been compared in different regions of SDP.81 (z ∼ 3) (Sharda et al, 2018). There is much more SFR than predicted from the linear KS relation, and the authors propose another relation between gas and SFR, taking into account the free-fall time of the clouds. Since the observed turbulence in the cloud is much higher than for local galaxies, based on the observed high gas velocity dispersion, a model of multifreefall based on turbulence (Salim et al, 2015) is in better agreement with observations. Note that another attempt to derive the resolved KS relation for distant galaxies, even without any lensing, has given results compatible with the linear KS relation (Freundlich et al, 2013).
ALMA is now able to detect normal galaxies at z ∼ 7 (e.g. Maiolino et al, 2015). With the [CII] line and continuum dust emission, the detection of Lyman break galaxies have been successful, implying SFR of 5-15 M⊙/yr. A spatial offset of the order of 4 kpc has been observed between the [CII] emission and the Lyα line and far UV, suggesting that stellar feedback rapidly destroys/disperses the molecular clouds.
5.3. Black hole mass estimation at high-z
A large number of quasars have been detected now at z ∼ 6 in molecules, and their CO / [CII] kinematics can be used to derive the central dynamical mass. From their broad lines detected in optical, and a widely known relation between Broad Line Region (BLR) luminosity and radius calibrated from reverberation mapping (e.g. Bentz et al, 2013), it is possible to compare the black hole mass (MBH), and the host dynamical mass. The high-z quasars appear all with a much higher black-hole mass than expected from their dynamical mass and the local M-σ relation (Wang et al, 2013, Venemans et al, 2016). This surprising result could come from the large uncertainties of the mass estimation. It is not possible to distinguish bulge and disk, so the MBH is compared to the total host dynamical mass, but this is precisely conservative. The inclination of the rotating molecular disk is not well known, and the result is valid only statistically. The MBH estimation has also a fudge factor for inclination. In most systems, it is assumed that the [CII] or CO lines are centered on the systemic velocity, and can be reliably used to derive the dynamical mass of the central stellar bulge. However, the optical MgII broad emission lines are systematically blue-shifted. The average blue-shift is of ∼ 500 km/s, but can be found up to 1700 km/s. This is interpreted as an outflow due to the central AGN, given that the symmetrical red-shifted region behind is too obscured to be seen.
Even if the dynamical mass is not well estimated, it is possible to have a lower limit for it with the mass of the gas, estimated from both the lines and the dust emission. In these high-z systems, the gas fraction is often larger than 50%, and the dynamical mass cannot be more underestimated than by a factor 2. The derived black hole masses are then robustly 3-4 higher than expected from the local relation (see Figure 10).
 |
Figure 10. Black hole mass versus the dynamical mass of z ∼ 6 quasar host galaxies and the bulge mass of local galaxies. The black diamonds are values obtained for local galaxies (Kormendy and Ho, 2013). Their MBH – Mbulge relation is represented by the solid line and the shaded area. The large and colored symbols are the high-z quasars. The green stars are the z > 6.5 quasars from Venemans et al (2016). For a given bulge mass, the high-redshift quasars have a more massive black hole than local galaxies. From the quasar luminosity (linked to its accretion rate), and from the observed star formation rate, it is possible to extrapolate the trajectory of the points (arrows) during the next 50 Myr. Image reproduced with permission from Venemans et al (2016), copyright by AAS. |
5.4. Ly-alpha blobs and proto-clusters
Protoclusters are overdensities in the early universe, where the growth of structures and their accompanying black holes are accelerated. They are not yet virialised into clusters, but are precious to understand why black holes might start growing very quickly, and AGN feedback might shape the first galaxies. Narrowband imaging at rest-frame Ly-α have revealed accumulation of Ly-α emitters (LAE), but also extended (> 30 kpc) Ly-α emission (often termed Lyman-Alpha Blobs, LAB) (e.g. Steidel et al, 2000, Matsuda et al, 2004). In these protoclusters, X-ray observations have revealed a significantly higher (by a factor ∼ 5) fraction of AGN (e.g. Lehmer et al, 2009), suggesting a longer duty-cycle for black hole accretion in galaxies of rich environments. Alexander et al (2016) observed with ALMA such AGN in proto-clusters and obtained a high detection rate, implying SFRs of 200-400 M⊙/yr, somewhat enhanced with respect to the field. This enhanced star formation may explain the extended Ly-α emission of the LAB, given a reasonable escape fraction for the continuum ionizing photons.
Proto-clusters can also be the site of a colder gas phase, which is extended as a circumgalactic medium (CGM) around the main galaxies. A striking example is the Spiderweb, a conglomerate of merging galaxies at z = 2.2 (Emonts et al, 2016). Several CO lines and CI were observed with ALMA and ATCA, showing an extended network of clumps and filaments, with gas excitation similar to that of the Milky Way (see Figure 11). The gas is metal enriched and dense, and most of it must come from tidal and ram-pressure stripping from the interactng/merging galaxies in the central region of the proto-cluster (Emonts et al, 2018).
 |
Figure 11. Left: One channel map (V = -447 km/s, 90 km/s wide) of the [CI] emission obtained with ALMA towards the Spiderweb proto-cluster of galaxies (Emonts et al, 2018). The blue contours are from [CI]3P1 - 3P0 emission, and the red contours from the 36 GHz radio continuum, both overlaid on the HST image. Right: Overlay of CO(4-3) contours on the HST F160W image of the Candels-5001 proto-cluster at z = 3.47 (Ginolfi et al, 2017). The red bar is the HST PSF, at 0.34 µm in the rest frame. Images reproduced with permission from Emonts et al (2018) and Ginolfi et al (2017). |
Molecular gas structures elongated on scales of ∼ 40 kpc are not rare in high-z proto-clusters, and a molecular mass of 2-6 × 1011 M⊙ has been detected in CO(4-3) and dust emission, with clumping and relatively high metallicity, at z = 3.5 (Ginolfi et al, 2017). The extended structure is compatible with a tidal/ram pressure origin, but could also be fueled by some cold gas accretion (see Figure 11).