South by High Redshift Conference | Computational Cosmology at Georgia Tech

Conference website: Austin, TX.

(Day 1) Wednesday, 01 April 2015

Jump to: (Day 2) (Day 3)

Stacy – Influence of Minihalo Environment on the Growth of Pop III Clusters

Exploring ~10 Pop III star formation sites in a 1 Mpc/h box, focusing on multiplicity of Pop III clusters.
Total mass in sink particles are in the range 20-120 solar masses.
Binary fraction of 36%, flat IMF (alpha = -0.17) looking after 5000 years.
Some sink particles are ejected from the inner cluster and may have low masses (< 5 solar masses). AGB stars?

Johnson – Have we found Population III?

(Yoshii 1981; Iben 1983; Frebel+ 2009; Komiya+ 2010) Has been long recognized that Pop III stars could accrete metal-enriched IGM, masking their true nature.
(Johnson 2014) Accretion onto a Pop III star. Dust grains can absorb more stellar radiation than gas. As gas accretes, dust can be expelled by radiation pressure.
Expected chemical signature would not have elements that were depleted into grains. High Zn and low Ti abundances.
However, a stellar wind and magnetosphere could suppress accretion.

Yoshida – Formation of massive primordial stars and black holes

(Aoki+ 2014) Detection of a 2nd generation star? Low alpha/Fe, Co/Fe. Signatures of a pair-instability SN?
100 First Stars (Hirano+ 2014): There are several wide binaries. Top-heavy IMF, favoring >100 solar masses. Highly rotating cores result in low-mass (40 solar masses) stars; low-spinning cores result in high-mass (300 solar masses) stars, rapidly accreting.
1540 Primordial Stars (Hirano+ 2015): Pop III.2D stars are clouds that are exposed to a LW background
Direct Collapse BH through Burst Accretion: Possible to get direct collapse from the supermassive stars in rapid accretion events.

Schneider – Decoding the stellar fossils of dusty star formation at the highest redshifts

(Schneider+ 2014) Sources of dust: dust production versus progenitor masses.
(Ventura+ 2012ab, 2013, 2014): Dust yields from AGB stars between 1-8 solar masses at low metallicity. Uncertainities from mass loss and convection borders.
AGB contribute up to 50% of the dust in 300 Myr.
(Schneider+ 2015) Inspecting the dust-to-gas ratio to metallicity relation, which has a jump from the effect of grain growth increases at some critical metallicity around [Z/H] ~ -1.5.
(Mancini+ in prep) Studying the dust mass in z ~ 7 “normal” star forming galaxies, especially the dust evolution and contributions from SN and AGB stars.
Dust mass depends sensitively on ISM density, so simulations with dust are required to shed further light on dust mass evolution.

Safranek-Shrader – Resolving Star Formation in the Early Universe

Coupled simulations, extracting a halo from a GADGET simulation and re-simulating with FLASH.
Gas cools down to the CMB temperature. The dust and gas temperatures are coupled above 10^8 cm^-3. The gas becomes optically thick around 10^10 cm^-3, suppressing fragmentation.
Accretion rates are variable in high mass sink particles. Low-mass sinks are terminated by dynamical processes.
Hints at a power-law IMF?

Jeon – Formation of the First Galaxies under Stellar Feedback

Stellar feedback evacuates gas in halos with masses < 10^8 solar masses.
20-30% of the volume is ionized by z ~ 10. Most of the radiation has come from Pop III stars at this time.
Metallicity floor of [Z/H] ~ -4. No primordial gas exists in the target halo.
CC SNe are the main metal contributors in the target galaxy. Pop II and III stars can coexist, but Pop II stars dominate.

Zackrisson – Searching for Population III Galaxies at High Redshifts

(Zackrisson+ 2015) Extreme lensing (µ ~ 1000) will allow for the detection of tiny stellar systems of M ~ 10^4 solar masses. Possible to get IMF constraints.
(Zackrisson+ 2011) Yggdrasil code. Population synthesis code, including Pop III stars.
(Rydberg+ 2015ab) Inspecting the CLASH field for Pop III galaxy candidates.

Griffen – The Caterpillar Project

Focusing on 24 MW-like halos at z=0 with a mass resolution of 3 x 10^4 solar masses.
(Snyder+ 2015) There is some morphology dependence (bulge-dominated vs. disks) on merger history. Not all major mergers are created equally, depends on the cold gas fraction and orientation of the merger.
Interactive tool to create initial conditions and submit them to clusters.
Convergence down to V_max ~ 5 km/s, resolving down to 10^6 solar mass halos.
Infall times and merger history – delineates between being quenched before or after infall.

Natarajan – Formation of SMBHs at High Redshift

Stressing that an SMBH outlier at z = 6 isn’t a problem with LCDM because it may be a rare object. Once you find a large population of SMBHs, then it becomes a problem.
(Natarajan+ 2015) Timing challenge on rapid accretion to grow the biggest SMBHs within 1 Gyr.
(Alexander & Natarajan 2014) Super-Eddington accretion when stellar-mass BHs are orbiting in a dense gas cloud.
(Volonteri & Natarajan 2012; Natarajan+ 2015) Lower BH occupation fraction in low-mass galaxies; dramatic differences with respect to seed masses.

Glover – Forming black holes by direct collapse: what determines the critical UV flux?

Astrophysical uncertainty for J_crit: Strength of the X-ray background, SED of the background, halo-to-halo scatter.
Modeling uncertainties: Composition of chemical network, rate coefficients, treatment of H_2 self-shielding
(Glover+ 2015) Determined the minimal chemical network for direct collapses. A key reaction that was missing was the collisional ionization of H by H:
- H + H -> H+ + e- + H
- Dominates when the electron fraction is small. Omitting this reaction changes J_crit by a factor of two.
(Sugimura+ 2014) Investigated different H2 self-shielding prescriptions. Draine & Bertoldi (1996) assume rotationally cold H2; Wollcott-Green+ 11 assumes LTE. How to calculate the H2 column density?
Open questions: Influence of incident SEDs? Treatment of self-shielding?

Agawaral – Current State of Affairs: The Direct Collapse BH Scenario

J_crit depends on the SED of the background – determines the photo-detactment of H- and photo-dissociation of H2.
Parameter study of J_crit by exploring different rate coefficients, searching for the rates that can produce a DCBH. The galaxy that’s providing the incident radiation 5 kpc away.
Exploring different SEDs, finds that a galaxy with 10^8.5 solar masses with an age of 1 Myr can provide enough radiation to prevent H2 cooling (J_crit > 700 for Starburst99).
No “unique” value of J_crit: depends on the age, metallicity, SFH of the nearby galaxy.

Whalen – The Evolution of Supermassive Black Holes Over Cosmic Time

(Whalen & Fryer 2012) Low-mass BHs (core-collapse) often have kicks from an asymmetric explosions, further increasing the difficulty of subsequent accretion.
Using a (100 Mpc)^3, identified a high-sigma peak (10^9 solar mass halo) at z=18. Considered X-ray feedback with a monochromatic spectrum of 1 keV. 10 AMR levels (300 pc resolution). Subgrid alpha-disk model of accretion.
X-ray feedback suppresses mass accretion by a factor of six.

Trump – Revealing Black Hole Seeds by the Fossil Record in Local Dwarfs

Did the M-sigma relation form at high-z or through mergers (mean limit; Janske+ 2011)?
(Sun+ 2015) The relationship is already in place by z=2 in massive galaxies. QSO duty cycle ~10%.
(Wilcott+ 2015) M-sigma is mostly in place at z=6 with a large scatter in massive galaxies (V_rms > 100 km/s).
(Trump+ 2015) Using line ratio AGN selection techniques to search for AGN in low-mass galaxies in SDSS (320,000 galaxies).
Coeval SMBH growth and star formation in massive systems.
Below a stellar mass of 10^10 solar masses, the SMBH occupation fraction drops from unity to ~10%.
(Trump+ 2011, 2014) Use spatially-resolved BPT to separate the nuclear AGN from the extended HII regions.
Q&A: M_BH – M_star relation may form before M_BH – sigma.

Ferrara – Ultra Faint Galaxies

(Ali+ 2015) Paper-64 result: constraints that reionization is not cold (must be photo-heated and photo-ionized).
(Mesinger+ 2013) Three peaks in the 21-cm power spectrum at k = 0.1 Mpc/h, corresponding to the peaks in absorption, strong absorption, and emission.
(Mesinger+ 2015) Can use the number density evolution of Lyman-α emitters to constrain reionization; however, one must consider the co-evolution of galaxy properties and reionization.
(Yue+ 2014) Estimated number of galaxies in Abell 2744 with lensing. 80% of the ionizing photons coming from halos <10^9 solar masses. May have already detected 20-30 of these galaxies in the Frontier Fields.
Cosmic IR background provides another constraint on the first galaxies.

Becker – Patchy Reionization and the Ionizing Photon Budget

(Roberston+ 2015) Reconciles lower τ with galaxies forming later.
Extending a simple reionization model to z = 2.5, one finds 19 LyC photons / atom / Gyr from stars only, however observations only show 4 LyC photons / atom / Gyr from stars and AGN. Suggests that galaxies are evolving.
To count photons in the background, need to measure the opacity (Becker+ 2013), temperature (Becker+ 2011) and mean free path (Worseck+ 2014) of the IGM.
(Becker & Bolton 2013) Ionizing photon density doesn’t evolve from z= 2.5 to 5.0. However, the SFRD is decreasing with redshift. However, pushing this to z=6, then the UVB drops by a factor of ten.
At z ~ 6, there is large scatter in Lyman-alpha opacity from the density field alone because it’s biased toward overdense regions with QSOs.
(Becker+ 2015) ULAS J0148+0600 (z=5.98): Huge GP trough between z = 5.52 to 5.88 (110 Mpc/h). But there is Lyβ transmission. Mainly this shows that there is great variation in the UVB at this time. Can this originate from density variations alone? No! Need a factor of 3 variations.
With a fixed mean free path and galaxy sources in a 100 Mpc/h simulation, there is a ratio between UVB with different mfps. No simple model fits the data.
Upper limit of the mfp at z=5.5 is ~40 Mpc/h with 4 photons / atom / Gyr. To explain the opaque QSO lines of sight, a mfp of 10 Mpc/h fits the data but with a higher ionizing photon density (19 photons / atom / Gyr).

Bullock – Did faint galaxies reionize the universe?

(Boylan-Kolchin+ 2014) High-z galaxy LF must break at M_UV ≥ -14 from near-field constraints.
Through abundance matching, one can recover a relationship between UV magnitudes and halo mass. Simulations (Wise+; Finlator+) lie on these relationships.
(BK+ 2014) ~170 atomic cooling halos at z=8 in a LG Lagrangian volume. Over-predict galaxies with >2 x 10^5 solar masses of ancient stars.
Note: z=8 progenitors of MW/M31 have M_UV ~ –16 from abundance matching.

Dressler – The Contribution of Faint LAEs to Reionization

LAEs alone appear to be supplying a large fraction of the LyC photons in reionization. LyC escape = 20% (Gronke+, Garel+ 2015).

Alavi – Understanding The Likely Sources of Reionization by Studying The Faint Lensed Star-forming Galaxies at 1<z<3

Investigating the number density of dwarf galaxies at z = 1-2. The faint-end slope is around -1.5 with no turnover down to M_UV = -12.
There is some general trend to bluer spectra (beta – UV slope) in fainter galaxies. The intrinsic scatter still exists in low-luminosity galaxies at z ~ 1.
(Alavi+ in prep) Using the FIRE simulations, they calculate the UV slope versus M_UV for several galaxies. The scatter is created by bursty star formation (i.e. redder/bluer during quiescent/active periods) even without dust attenuation.

Tilvi – Probing the Epoch of Reionization

Studying the evolution of Lyman-alpha equivalent width. Missing high EW galaxies at z ≥ 7. Neutral IGM absorption? Need to move to other probes, such as C III].
Only ~20% of galaxies have Lyα EW > 25Å at z > 6.
(Treu+ 2012, 2013) Using toy models, the data favors models number density evolution, i.e. only the brightest sources can penetrate through a neutral IGM.

(Day 2) Thursday, 02 April 2015

Jump to: (Day 1) (Day 3)

Shapiro – Simulating Cosmic Reionization and Its Observable Consequences

Now using the CLUES simulations to conduct rad-hydro runs. Questions: Does reionization help with the missing satellites problem? Was the Local Group ionized internally or externally?
Using RAMSES, coupled with ATON to perform the simulations.
Reionization occurs at z ~ 5, which only includes radiating sources within the LG. However, this goes against the previous work of Ocvirk+.

Park – Impact of Cosmic Reionization on Small-Scale Structures : Clumping Factor

(Park+ in prep) Examining the clumping factors on scales smaller than a comoving Mpc.
Use a 200 kpc/h box with 50 Msun DM resolution to study the clumping factor (cf. Emberson+ 2013) with an external UVB with J_21 = 0.1, 1, turned on at z = 10, 9, and 8.
After 200 Myr, the clumping factor decreases from 10 to 3.
Use the number of recombinations (relative to the mean rate) as a function of density as a subgrid model in large-scale reionization simulations.

Dijkstra – Constraints on Reionization at z=6-7 from Lya Emitting Galaxies

The Lya fraction and LAE luminosity function both have reductions at z ≥ 6. Why?
Lya transfer is a multi-scale problem (100 Mpc -> galaxy scales -> scattering in a dusty medium). The gas kinematics are very important in Lya transmission. Must consider neutral absorbers within large-scale HII regions.
To explain the reductions, models need an ionized fraction of 50% at z=7.
(Gronke+ 2015) Calculated the faint-end slopes for LAEs and LBGs, and they differ by 0.4. The LAEs are probing fainter galaxies.
Are Lya and LyC escape fractions correlated? The slope of the LAE LF could indicate the emissivity of LyC photons.

Pober – A Lower Limit on the z = 8.4 IGM Temperature From 21 cm Power Spectrum Observations

In the cold reionization scenario, a small spin temperature can dominate the power spectrum.
Using 21cmFAST simulation to explore spin temperature and ionized fraction parameter space. Rules out a cold reionization scenario.
PAPER has finished observations and work has started on HERA.

Mutch – Modelling galaxy formation and the EoR with DRAGONS

Coupling a semi-analytic galaxy formation model with a semi-numeric model (21cmFAST) of reionization.
Calibrate the SA model against star mass functions at z=5-7 (Duncan+ 2014)
Reionization at z = 8 ± 1, given a factor of two uncertainty in escape fraction.
Matches galaxy LFs very well at z=6 and z=8, even though it wasn’t calibrated to do so.
Will include AGN feedback in the near future.

Finkelstein – A Review of Observational Galaxy Studies at z ≥ 4

Current space programs: HUDF, (S-)CANDELS, CLASH, Frontier Fields, Spitzer, GOODS, SEDS
Ground-based: ULTRAVISTA/VIDEO (VISTA), HUGS (VLT), SXDF (Subaru), CFHTLS (CFHT), UKIDSS (UKIRT)
What have we learned?
- The good news: 2-sigma agreement between different groups on LFs (Finkelstein+ 2014)
- A surprise! The knee in the Schechter function does not evolve, but the faint-end slope becomes steeper and the number densities decrease with increasing redshift.
- Very little data at the bright end at z=7 and z=8. Cannot determine whether it is a power-law or Schechter function.
(Finkelstein+ 2015) Same luminosity galaxies exist in smaller halos at higher redshift. Why?
- Increased gas density; more cold gas; effects of feedback lesser at high-z.
- Halo masses = 10^11.4 (z=7) and 10^11.9 (z=4) solar masses
Full analysis of the first-year Frontier Fields data does not show a sudden drop-off in cosmic SFR density (McLeod+ in prep)

Bowler – Rapid evolution in the bright end of the galaxy luminosity function between z = 5-7

Using a survey field that is 8x the CANDELS field down to AB magnitudes of 25-27, depending on the filter.
There has been disagreement at the bright-end between space- and ground-based surveys.
This was caused by cosmic variance between the fields by a factor of three. Showing the need of >degree-scale fields to probe the bright-end of the LF.
Models indicate strong dust obscuration may be necessary to reproduce the observations (Bowler+ 2014).

Curtis-Lake – Can we tell if Lyman Break Galaxies are getting bigger between redshifts 8 and 4?

(Ono+ 2013; Shibuya+ 2015) Size evolution of LBGs from z = 2 to 8 from 1 kpc by a factor of ~3.
(Curtis-Lake+ 2014) Looks that size scales somewhere between the halo circular velocity and halo mass. The peak of the size distribution doesn’t change with redshift, but the large-end tail builds up over time.
However, using a null hypothesis of redshifting the z=4 galaxies to greater distances, a size evolution of (1+z)^n of both n=-1 and 0 cannot be ruled out.

Bradač – Pushing the Frontiers of Galaxy Formation with Spitzer and Cluster Lenses as Cosmic Telescopes

(Wang+ in prep) There can be some discrepancies between photometric and spectroscopic redshifts in lensed multi-imaged systems.
Use Spitzer at z ~ 8 to break degeneracies in ages to measure the Balmer break. New survey called SURFS-UP covering 10 clusters. Spitzer observations are important for preparation for JWST.
(Hoag+ 2015) z = 9.5 galaxy with an estimated stellar mass of 10^8.8 Msun, SFR = 1 Msun/yr, ~350 Myr ages. There are several other detections with stellar masses between 10^8 and 10^9 solar masses. Also found sSFRs around 10 Gyr^-1 with spread (and errors) of an order of magnitude. No trend with SFR.

Livermore – Pushing the Frontiers: uncovering the earliest galaxies in the Hubble Frontier Fields

Use wavelet decomposition to remove the light from the clusters.
- Convolve the image with a large wavelet, subtract it, and repeat with progressively smaller wavelets.
Found 95 galaxies in the Abell 2744 between z = 6-9, mostly focusing on z=7 galaxies.
Finding a possible faint-end turnover at M_1500 == -16 at z=7 in one cluster. Warned that there are many uncertainties at these scales because of incompleteness and magnification errors.

Song – Probing stellar mass build-up in galaxies at z=4-8 with CANDELS and S-CANDELS

Using HST + Spitzer data (sample size ~ 7000 with ~300 at z=7) to constrain the stellar mass function.
Finds a constant scatter of 0.4-0.5 dex in the stellar mass – UV luminosity relation. No evolution in slope with redshift.
At z=4, the stellar mass function is depleted from the halo mass function, but at higher (z ≥ 6) redshifts, it’s closer to the halo mass function. Feedback in action?
Stellar mass function is in agreement with the cosmic SFRD.

Salmon – The SFR and Stellar Mass evolution at z>4 from CANDELS

UV faint, high-mass objects with significant reddening.
The SFR – stellar mass relation evolves little in slope and declines in scale at z > 5.
The scatter in SFR at a given mass is small at all redshifts (0.2-0.3 dex). Favors smooth, cosmological accretion into galaxies.
The SFR function steepens with redshift, becoming increasingly like a power law at z ≥ 5. This could possibly mean the mode of SFR is changing going into reionization.

Sajina – Massive galaxies at z~1-6 as seen in Spitzer-SERVS and Vista-VIDEO data

Most of the detected galaxies are dusty and actively forming stars.
They find 8 z=4-5 galaxies with stellar masses >10^11 solar masses. Need spectroscopic confirmation of all of the detected galaxies. Where are the dusty galaxies at these earlier times because they exist at z = 2-4.

Greiner – Exploring hidden star formation over 3<z<5 with Gamma-Ray Burst Host Galaxies

There are over 400 GRBs with a known redshift with 65 in z = 3-5 and 21 with a host galaxy detected.
The UV LF is inconsistent with GRB data. They find no evidence for selection effects, such as metallicity bias. To match the LF, need a model with little metallicity dependence.

Note: Missed talks by Rigby (The James Webb Space Telescope and the High Redshift Universe) and Illingworth (The First Billion Years: The Growth of Galaxies in the Reionization Epoch).

Oesch – Stellar Mass Measurements at the Cosmic Dawn: Insights from Ultra-Deep HST and Spitzer Observations

(Gonzalez+ 2011, in prep) Stellar Mass – M_UV relationship: shallow when considering a simple SED fit. Highest mass systems are not the brightest in the UV because the dust saturates the UV light.
Recently awarded GREATS survey with 733-hour exposure time.
(Oesch+ 2015) Spec. confirmed galaxy at z = 7.7 with Lya.
New survey (HDUV) to provide rest-frame UV for z=0.5-3 galaxies, possibly analogs of high-z galaxies.

Bouwens – Rest-frame Optical Nebular Emission Lines in z~4-8 Galaxies

(van der Wal+ 2011) There have been extreme emission line galaxies with EWs ~1000Å at z ~ 1.
With H-alpha, H-beta, [OIII] fluxes, you can measure SFR, metallicities, and dust extinction.
See Capak+ (2015) for recent data on dust corrections.
Found 11 bright z=6-9 galaxies in the CANDELS field to follow up spectroscopically. Looked at one, and it was, lo and behold, a z=7.7 galaxy (mentioned in Oesch’s talk).

Schmidt – GLASS: Probing the state of the reionization epoch through HST spectroscopy

Select dropouts at z ≥ 7 in six of the Frontier Fields and classified the objects according to several selection rules. Using the Treu+ (2012) framework with equivalent widths.
NIR flux limits are around 10^-18 erg/s/cm^2, resulting in 100 EW limits at z ≥ 7 from GLASS. This is evidence for strong evolution as the universe is reionized.

Sobral – Exploring the z~6-9 Universe with the largest Lyman-alpha narrow-band surveys

(Sobral+ 2015; Matthee+ 2015) Detection of HeII 1640Å at z = 6.6, and the emission is very compact. HeII / Lya = 0.3 with Lya EW = 230Å, however there is no CIII] or OIII]. No other lines!
The Lya/HeII is coming from one blue clump, and there is a redder separate component.

(Day 3) Friday, 03 April 2015

Sorry, but I had to go back to Atlanta due to other commitments!