Two epochs of globular cluster formation from deep fields luminosity functions: Implications for reionization and the Milky Way satellites

Authors: Katz & Ricotti
Link: here

In this paper the contribution of old globular clusters to reionization is discussed. Observations of globular clusters in the Milky Way (and a few extra-galactic ones) reveal two distinct categories: metal-poor and metal-rich. Age determinations of these two populations are highly uncertain (errors of 1 Gyr) and there is a significant spread in ages (1 Gyr for the metal-poor and 6 Gyr for the metal-rich population). However, the age gap between metal-rich and metal-poor globular clusters is greater than the age range within each population, suggesting that there are two distinct epochs of globular cluster formation. Likely 55% (worst case: 20%) of the globular clusters in the Milky Way have formed at z>4. The authors try to constrain the formation rate of globular clusters at hight redshift by computing synthesised luminosity functions.

The following assumptions are made on globular cluster formation:
1) At the time of formation, the globular clusters were 9.1 times more massive than at the time they were observed in the Milky Way
2) The globular cluster initial mass function is lognormal or a power law
3) From the frequency of globular clusters in galaxies in the local Universe, the mass density and the formation rate can be inferred
The unknown is the fraction of todays globular clusters that formed at high redshift

The most massive globular clusters can be observed with HST up to redshift 6, and with JWST up to redshift 7.5. At z>0.01 globular clusters are unresolved and their light will dominate a high-z dwarf galaxy for about 10 Myr after their formation.

The authors constrain the fraction of globular clusters forming as a function of redshift using the luminosity function and continuum slope. They use two approaches. First assuming only one globular forms per halo, which requires the least amount of assumptions but is unrealistic. Second, allowing for multiple globular clusters per halo, assuming a linear relationship between the halo mass and globular cluster mass, a Press-Schechter mass function and a minimum halo mass in which star formation takes place (independent of redshift).

With these upper limits on the formation rate of globular clusters the authors conclude there are two distinct formation epochs of globular clusters: near z~2.5 and around the epoch of reionisation. Although the authors can only give upper limits, they argue that these upper limits are close to the actual values. These formation rates of globular clusters during reionisation imply that if the escape fraction is close to 1, the number of ionising photons produced is sufficient for reionisation.

The Impacts of Ultraviolet Radiation Feedback on Galaxies during the Epoch of Reionisation

Authors: Hasegawa, K. & Semelin, B.
Paper: here

The goal of this paper is to assess the effectiveness of feedback from UV sources in the early Universe on high-redshift, low mass galaxies. They use a coupled radiative transfer and hydro SPH simulation to study the UV feedback in a cosmological simulation of 5 comoving Mpc. This code has high spatial and mass resolution, and so is able to include sources internal to a given galaxy, which while not the first code to do this, is one of the improvements over many other reionization simulations. This increased spatial resolution in the RT code, as well as the increased mass resolution (though a small volume because of this), allows them to follow the effects of the UV feedback on star formation between z = 6 - 15.

They find that including the UV feedback (photo-ionization of HI, HeI, HeII; photo-dissociation of H2 and H2+; photo-detachment of H-; self-shielding of H2 above a given column density) suppresses star formation, which is expected. However, they argue that the decrease in SF is larger than other simulations. The causes of this suppression depends on halo mass. For low mass halos (<10^9 Msun) the baryons are evacuated, leaving few baryons left to form stars. For higher mass halos, there is little mass loss, but the gas density is smoothed out, and the H2 molecule nearly destroyed. As their star formation algorithm is a strong function of density and temperature, this causes the drop of star formation in their simulation.

This work suffers from resolution effects, which the authors themselves note. The global star formation history has not converged for the runs with radiative transfer. This is partially due to resolving the lower mass galaxies, but also due to the fact that with increased resolution, stars form earlier, and hence the feedback effects start earlier. Thus, further studies with higher resolution are necessary.

Another problem with this work relates to the star formation algorithm. First, they have a very low density threshold required for star formation: a gas particle must have rho > 200*rho_ave, which includes basically every gas particle inside a halo. Second, they do not include any metal feedback from the supernovas in their box. The metals will cool the gas, bringing more gas below the temperature cut (T < 5000K) for star formation. Without the metals, only the gas which is strongly self-shielded against H2 dissociation can cool enough to form stars. Thus, further studies with metals are also necessary, and the authors note they will include this in future work.

Galaxy disks do not need to survive in the L-CDM paradigm

Full title: Galaxy disks do not need to survive in the L-CDM paradigm: the galaxy merger rate out to z~1.5 from morpho-kinematic data

Authors: Puech, M., et al
Paper: here

The main result of the paper is the derivation of the time evolution of the galaxy merger rate using a sample of intermediate mass (M_stellar=10^10-10^11 M_sun) emission-line galaxies for which they have both kinematic (IMAGES survey) and morphological (from HST) data. These galaxies are chosen because they are the most likely progenitors of local spirals (see justifications in the paper). A large fraction of these galaxies are not relaxed morphologically and/or kinematically and the authors demonstrate that the disturbances are most likely triggered by major mergers.

The authors then match as many as possible of these observed, disturbed galaxies to simulated major mergers by comparing the real and simulated velocity fields and morphology etc. This allows them to associate a time since the merger began with each observation and they group them into “pre-fusion”, “post-fusion” (the main starburst period) and “relaxation” phases. They plot the observed SFR (normalised by gas mass) versus the time since the start of merger to illustrate that the IMAGES galaxies sample all the phases of an average merger well. They can then derive a merger rate from: rate = ngal in merger phase/time period for this phase, and a corresponding redshift for this rate: z= average start time of mergers in this phase. This gives them merger rates of 5.5, 10.1, 11.1 %/Gyr at z=0.72, 1.12, 1.55 respectively. They show that these merger rate values are in agreement with those from the Hopkins et al. 2009 semi-empirical model to within a factor of 2-3.

The authors draw some additional conclusions, including:
· Given the high gas fractions they infer for the major mergers occurring at z>0.6, a significant disc should be able to reform by z=0.
· The merger rate from the semi-empirical model is not in conflict with the fraction of bulgeless galaxies observed locally.
· Since cold streams have been shown to be suppressed below z~1.5, gas-rich major mergers could potentially take over as a channel for massive, thin disk formation at this redshift.
· Overall: there is no disc survival problem, but a period in which discs could/should be rebuilt at z<1.5.