High-z Growth of Galaxies - 20-24 Jan 2020
Arrived Wednesday midday
- Sources of dust: core-collapse SNe (Schneider+ 2014; Marassi+ 2014, 2015; Bocchio+ 2016). In the Galaxy, the mass of SN dust enriching the ISM is much less than the observed SN remnants younger than 10 kyr (1987A, CasA, Crab, N49).
- (Marassi+ 2019; Schneider+ 2020) New grid of SN yields (13-120
$M_\odot$ , 1e-3 - 1$Z_\odot$ , and three rotation rates) - (Dell'Agli+ 2019) AGB stars: Silicates are produced by
$>3 M_\odot$ stars with yields increasing with metallicity. No silicates are produced below$0.07 Z_\odot$ . Carbon dust is produced by$<3 M_\odot$ and is independent of metallicity. Below$5 \times 10^{-3} Z_\odot$ , no dust is produced. - AGB contribution to the total dust budget is >30% and becomes dominant after the stellar population is older than 500 Myr.
- (Ginolfi+ 2018) Using a closed box model to explain the dust content of the Galaxy, but it comes up short. Suggests that grain growth and production is important.
- (Schneider+ 2016) The differences in observed dust masses in galaxies may be caused by the growth (accretion) timescale being dependent on metallicity.
- (Mancini+ 2016) Uses a dust model to correct mock observations from a simulation that better matches observations at the bright-end. Young (<15 Myr) stars suffer from strong attenuation.
- (Graziani+ 2020) dustyGADGET uses the same model as above but in a cosmological (30 Mpc/h) simulation where the dust evolution is computed within the simulation. The re-processing of dust within galaxies is important and increases the dust mass by an order of magnitude in galaxies with
$> 3 \times 10^8 M_\odot$ . - (Zhu+ and Li+ in prep) Full RT zoom-in simulations with GIZMO with an improved BH seeding and accretion model. Focusing on massive galaxies with
$5-30 \times 10^{10} M_\odot$ in stellar mass and SFRs between 30-1000$M_\odot$ /yr. Using ART$^2$ to match a massive galaxy at$z \sim 8$ (MACS 0416_v1) but needed to artificially boost the metallicity by 10 (needed solar metallicity). Dust temperatures are as high as 90 K in the center.
- Review of many different galaxy relationships and kinematics and their evolution with redshift
$z = 0-3$ .- In the main sequence (of galaxies), the SFR
$\propto$ stellar mass at stellar masses below$10^{10} M_\odot$ . - (e.g. review by Combes+ 2018) Mass-metallicity relation increases in normalization with decreasing redshift, and molecular gas fraction decreases with increasing stellar mass and redshift.
- In the main sequence (of galaxies), the SFR
- Future issues: origin of scatter in scaling relations; causes of high gas turbulence; origin of high gas fractions; mass loading and energetics of outflows; quenching processes; relations of low-mass galaxies...
- Need more resolved kinematics, cold gas distributions, multi-phase gas probes, progenitor populations, and substructure down to ~100 pc.
- See Buat+ (2019; arXiv)
- Dusty galaxies have very compact emission regions and clumpy UV emission
- Use CIGALE (Boquien+ 2019; link) for SED fitting from the UV to submm.
- There are several choices for attenuation curves (e.g. Salmon+ 2016 for photo-z data; Trayford+ 2019 for simulations; Lo Faro+ 2017 for ULIRGs). Which one to use?
- Calzetti law gives the best fits for their sample of 17 (except for 1)
- (Malek+ 2018) Galaxies with high sSFRs have a large difference (too much) in the predicted UV-NIR luminosity from submm emission.
- For the complete SED (UV->submm) fit, sources are best fit with different laws (7/17 for Calzetti; 6/17 for Charlot & Fall 2000; 4/17 for Lo Faro+ 2017).
- Takeaway: SFR measurements are accurate only if IR (submm) data are available.
- Why is the cosmic SFR higher at high-z? High gas content? or SF efficiency?
- (Decarli+ 2019; Magnelli+ 2020) Cosmic molecular gas density (
$M_\odot$ / cMpc$^{-3}$) tracers the cosmic SFR... but the$L_{\rm CO} - M_{\rm gas}$ conversion is uncertain. - Looking at the cosmic dust mass density as a function of stellar mass, it flattens at low-mass, and
$M_\star < 10^9 M_\odot$ galaxies only contribute 10-20% to the total density. - The availability of gas for SF is the SF prime driver, and the variation in SF efficiency only plays a secondary role.
- (Yue+ 2015; Dumitru+ 2017) Because the smallest galaxies during the Epoch of Reionization are too dim for even JWST to detect them, we can use intensity mapping in [CII].
- CCAT-Prime (6-m submm telescope with a 1-degree FoV) will be commissioned at the end of 2021, which is perfect for line intensity mapping to obtain many statistics.
- (Hummels+ 2019; Suresh+ 2019; Peeples+ 2019) Lack of cold gas in the CGM is numerical because of a lack of resolution at low densities.
- (Cicone+ 2019 white paper) Probing H2 gas is currently unfeasible at z=0 (even with 100s of hours with ALMA) because the large-scale (several arcmins) emission is filtered out by interferometers.
- Massive molecular outflows are ubiquitous in ULIRGs, QSOs, and starbursts, but they are also present in low-mass galaxies.
- There are three main scenarios: (1) Galactic fountain, (2) outflow gas escapes from the virial radius, or (3) outflows are trapped within the halo (CGM)
- (Maiolino+ 2012; Cicone+ 2015) Molecular CGM at a z=6.4 QSO with a detection of [CII] extended up to 30 kpc. Also detected at z = 3-6 (e.g. Ginoifi+ 2017).
- (Emonts+ 2016, 2018) Protocluster at z=2.2 (containing the Spiderweb galaxy) Molecular gas is detected out to 70 kpc in the diffuse CGM. Metallicities are similar to the ISM of SF galaxies. Supports that galaxies grow through recycled cold dense (>100 cm$^{-3}$) gas.
- (Decarli+ 2019) Targeting CO lines as a tracer of H2 mass, dust continuum, and other lines (as a bonus) in the HUDF field. 150 hours.
- (Gonzalez-Lopez+ 2020; Magnelli+ 2020) Several HUDF detections. All have optical counterparts. Found flattening at <100 micro-Jy. Only detected 3 more sources at these faint fluxes. Recovers 93% of the EGB at 1mm. A similar flattening is expected at 850 microns.
- (Popping+ 2020) Using the UniverseMachine (Behroozi+ 2019), they found that this flattening is associated with evolution in galaxy formation, in particular, the flattening of the galaxy LF at z=1-2.
- (Aravena+ 2020) Galaxies are above/at/below main sequence. Depletion times ~1 Gyr. Gas/stellar mass ratio is ~unity.
- (Decarli+ in prep) The cosmic molecular gas density evolves similarly to the cosmic SFR, whereas the atomic gas density decreases. The total (net) accretion is around 0.1 M$_\odot$ / yr / Mpc$^{-3}$, which can be directly compared to DM expectations.
- (Franco+ 2018) ALMA survey detects SMGs out to z~5 and unveils a population of optically dark galaxies (~17%) in GOODS-South
- (Wang+ 2020 submitted next week) SED fitting with a modified blackbody, stellar, AGN and other (nb: missed them) components
- (Franco+ 2020) Derived properties of galaxy population
- Presenting results of Gomez-Guijarro+ (2019)
- Detects compact UV and dust continuum emission.
- Stellar mass and SFR much larger in FIR-detected stellar components, suggesting minor mergers and burstiness. Shows the smooth transition from SF to quiescent galaxies.
- They find that the most extended SFGs become more compact before they quench, thus older SFGs are compact. Main sequence is dominated by extended SFGs.
- (Cullen+ 2019) Even in the local universe, there are uncertainties in gas metallicity calibrations from line ratios. So they use stellar metallicities as a high-z probe, but it requires deep observations to detect the continuum (Sommariva+ 2012).
- At high-z, the mass-metallicity relation increases monotonically with mass but the normalization is lower (in agreement with simulations)
- They're also looking for alpha-enhancements at z>2.5 (prior work: Troncosco+ 2014, Steidel+ 2016)
- (in prep) Stellar metallicities are correlated with observed Lya and CIII] EWs. Further evidence that stronger ionizing luminosities of low-Z stars causes high EWs.
Next-generation 21cm cosmology - 27-31 Jan 2020
- (Barry+ 2019ab) Open-source analysis pipeline that calculates power spectra that is described in this talk
- Github repository
- MWA Ultralow band provides the Epoch of X-ray Heating. This band is important in preparation for SKA-LOW
- Many possible physical explanations for the absorption feature
- Three main concerns within the EDGES team
- Instrumental systematics
- Foreground systematic
- Inappropriate analysis
- (in prep) They are performing additional analysis with a different band. Some of their results have troughs that vary between -0.2 and -0.9 K, where the original result was -0.5 K.
- EDGES Analysis code is on Github
- EDGES-3 is testing right now in Oregon and should be deployed late this year.
- Post-processing reionization calculations of MassiveBlackII, considering stars, BHs, XRBs, and ISM shocks (see this paper)
- BH, ISM, and XRBs emissivities are sub-dominant to stars. At z<7, BH contribution raises to ~10% of the stellar emissivity.
- Can check the emissivities so that it doesn't overproduce the X-ray background. At z>5, the XRB is 2.9e-12 erg/s/cm2/deg2 (Cappelluti+ 2017) from Chandra.
- He II emissivity is mostly caused by BH emission
- Reionizes from 10% at z=8 to nearly unity at z~6
- (Kakiichi+ 2017; Ma+ 2020) Can use QSO ionized regions to constrain the late stages of reionization and compare them to simulations, which can generate brightness temperatures for 21-cm observations. QSOs enlarge the HII region by about 1 arcmin (radius) and should be detectable with SKA.
- Many simulations ranging from 6.4 Mpc/h to 500 Mpc/h, both N-body (+ RT post-processing) and AMR+RT (i.e. CoDa)
- (Watson+ 2013) Halo HMFs are different at z>6. Modified fits are provided in the paper.
- (Ahn+ 2016) Enhanced scatter in small-scale structure is caused by sub-halo clustering, well-fit with a log-normal distribution. Can be used to create a subgrid model (Nasirudin+ 2020)
- (Bianco+ in prep) Clumping factor determination
$C(\delta | \mu, \sigma)$ from multi-scale simulations to make a subgrid model for enhanced recombinations in reionization calculations, where$\mu$ and$\sigma$ are the mean and scatter of the density field (I think). - (Ocvirk+ 2020) CoDa2: The luminosity functions at z=6-10 matches observations.
- The SFR decreases slightly (~10%) with RT (compared to w/o) in halos above
$10^{10}$ solar masses, which is probably caused by mergers of smaller galaxies with suppressed SF. - 25% in
$10^9 M_\odot$ halos and by an order of magnitude in$10^{8-9} M_\odot$ halos.
- The SFR decreases slightly (~10%) with RT (compared to w/o) in halos above
- (Hosein+ 2020) tSZ predictions from CoDa simulations
- (Sullivan+ 2018, in prep) 5.7 Mpc box with RAMSES-RT with ~20 comoving pc resolution. Also used ANNs to obtain relations between halo mass and other quantities to obtain a subgrid model.
- (Sullivan+ in prep) Also calculating UV escape fraction, which is very variable, but its time-average is around a few percent.
- DRAGONS model (Mutch+ 2015; Qin+ 2018): semi-analytic model coupled with N-body simulations and 21cmFAST for spatially-dependent ionization fields. Run in an MCMC environment to explore parameter space.
- (Qiu+ 2019) Constraints on the galaxy LF -- requires strong SN feedback, regardless on the dust extinction employed when calculating M$_{\rm UV}$.
- Can use clustering of galaxies to check whether the
$M_\star - M_{\rm UV}$ relation reproduces the observed clustering.
- Can use clustering of galaxies to check whether the
- (Liu+ 2017) There is a deviation from linear in the
$M_\star - M_{\rm UV}$ relation because of inefficient SF in low-mass galaxies, causing a turnover in the luminosity between$M_{\rm UV}$ = -12 and -14. - In their models while fitting th eLF, neutral fraction, and CMB optical depth, the most likely overall
$f_{\rm esc}$ is 12-15% if constant in time. - (Geil+ 2015) Correlations betwene galaxy formation models and the 21-cm power spectrum. If low-mass halos contribute more to reionization, there will be a larger number of smaller HII regions, i.e. more power at higher k.
- BlueTides (Di Matteo+) used a spatially dependent UV background with the Battaglia+ (2013) method.
- (Davies+ in prep) The brightest galaxies are mostly contained within isolated HII regions, but they aren't necessarily at the center because of their neighbors.
- (Davies+ in prep) Predictions for 10-100s of galaxies brighter than
$M_{\rm UV} = -22$ (i.e. GN-z11, Oesch+ 2016) with WFIRST HLS within a SKA field of view and dz~1. - (Davies+ in prep) Stack the brightest galaxies in 21cm. Can be detected with SKA in about 1000 hr. They stack galaxies with similar UV luminosities (in different bins) and find that SKA can detect that smaller HII regions are associated with fainter galaxies (
$M_{\rm UV} = -19$ ).
- (Park, Mesinger+ 2020) Forward modeling: Using MCMC with 21cmFAST (8 parameters) with mock JWST observations to see how the constraints on galaxy formation changes, especially when considering lensing errors at low-luminosities
- Predicts that galaxy LFs flatten above
$M_{\rm UV} = -13$ .
- Considering the observables of stars forming in minihalos, whether it be metal-free or metal-enriched.
- Invoking smooth mass cutoffs to emulate feedback effects.
- These minihalos will be too dim for JWST (or just barely), so 21-cm might be the best way to detect them.
- (Villaescusa-Navarro+ 2014) HI intensity mapping. Volume-filling fraction versus column density. Filaments -> CGM -> Halos. Need a model for the HI distribution, which is obtained from simulations.
- (Villaescusa+ 2015; Obuljen+ 2018) Can probe neutrino masses with intensity mapping (SKA)
- (Villaescusa+ 2017) Also can probe BAO with intensity mapping with SKA
- (Villaescusa+ 2018) Used IllustrisTNG to probe HI optical depths within halos, which is sensitive to feedback. Can inspect as a function of halo mass, but there is a large scatter between halos.
- (Sadoun+ 2017) Inspecting the CGM and Ly$\alpha$ emission in simulations(?), showing that the CGM becomes more neutral at z>5.
- (Weinberger+ 2019) Modeling LAEs from simulations at high-z.
- Inspecting the transmission factor and showing why their number densities decrease at higher redshifts as the Universe becomes more neutral.
- Favors a late reionization (z=7-7.5 midpoint), however their results aren't too consistent with the timing because LAEs are biased and exist in overdense regions.
- Also looking at a cross-correlation between LAEs and 21cm.
- (Keating+ 201?) Using Ly$\beta$ out to z=7 after Ly$\alpha$ becomes saturated.
- (Dayal & Ferrara 2018) The SFRD is uncertain (both theoretical and observational) at z>8, and JWST will be crucial in constraining it.
- (Dayal+ 2010, 2012; Hutter+ 2014, 2015, 2016) Clustering is important in determining whether a galaxy is a LAE or LBG. Confirmed afterwards observationally (see Tivi+ 2020).
- (Hutter+ 2017) LAE-21cm cross-correlations. Within ~10 cMpc, it decreases gradually as the neutral decreases from 50% to zero.
- (Dayal+ 2019, 2020) DELPHI semi-analytic model that models galaxy formation and reionization
- Star formation dominates the photon budget. Using observational constraints, there is an
$f_{\rm esc} - M_{\rm halo}$ relation for AGN. They assume a halo mass independent$f_{\rm esc}$ for stars. - At z>6, halos with masses above
$10^{10} M_\odot$ have more ionizing photons coming from AGN than stars, up to 10-25% of the total budget. SF in high-mass halos provide 15% of the photon budget. - (Chatterjee+ 2019) The EDGES absorption feature rules out <2 keV WDM particles because of the z=17 timing. Need enough structure to produce radiation by then.
- Star formation dominates the photon budget. Using observational constraints, there is an
- Uses a subgrid model (Ahn+ 2015) for minihalos in a 500 Mpc/h 3184^3 (?) N-body simulation.
- Calculates the LW flux in this volume, using the picket-fence modulation (Haiman+ 2000). They find oscillations in the LW background at early times (z>20) because of it.
- Their minihalo models build up enough Ly$\alpha$ to decouple the 21cm signal from the CMB. Smaller Pop III stars are able to build up a stronger Ly$\alpha$ background as they don't ionize significantly.
- What's missing from models of reionization to explain the late time transmission (and fluctuations) of LyC at z=5-6?
- Chardin+ (2017), Meiksin (2020), Kulkarni+ (2019), Keating+ (2020), Davis & Furlanetto (2016), Nasir & D'Aloisio (2019) all give different explanations
- Suggests to the simulators to run down to z=5 and compare/normalize to observables, especially the neutral fraction (i.e. mean opacity).
- (Meyer+ 2019, 2020) Small-scale Ly$\alpha$, i.e. Ly$\alpha$-LAE cross-correlations.
- (Meyer+ 2020) Recover the average
$f_{\rm esc}$ at z=6 that's ~10% and decrease with time. Much freedom in galaxies' contribution to reionization ($M_UV$ ,$f_esc$ , clustering, and redshift dependence) to which the reionization topology is sensitive. - Future: Use ML to train continuum fitting at low-z to use at high-z to obtain an optical depth over the full Ly$\alpha$ forest. However, there is a danger of QSO evolution, i.e. average continuum changes with redshift.
- (Davies+ 2018; Wang+ in prep) Using PCA to train a model on z~2 QSO continua. They use this PCA model on z>6 QSOs to determine the damped Ly$\alpha$ wings, which constrains the neutral fraction.
- With Euclid surveys and JWST follow-ups, they could possible discover z=6-8 QSOs which would constrain the neutral fraction to 15-20% in the same redshift range. Complementary to 21-cm observations.
- (Mazzaucchelli+ 2017; Bañados+ 2018) z~6 SMBH growth from seeds.
- (Davies+ 2019) Local Reionization Soltan Argument that relates the QSO Stroemgren radius to SMBH mass growth.
- With their models, they predict that an epsilon of 10% (radiative efficiency) is way too high. Constrained to an upper limit of 1.7%.
- This results in a much shorter e-folding time and less difficulty to grow the SMBHs from seeds.
- If the radiative efficiency is 10%, the obscuration rate has to be >82%, which is very high when compared to observations.
- (Davies+ in prep) Comparing how the baryons react to a flash reionization (i.e. Jeans smoothing) and their observable properties in QSO proximity zones.
- More saturated Ly$\alpha$ lines in the IGM that wasn't previously heated. Possibly can extract thermal history from proximity zones.
- Doing the same analysis for streaming velocities, X-ray heating, and WDM cosmology.
- In a power spectrum analysis, can distinguish between (1) WDM (<10 keV) and CDM, and (2) X-ray heated (>500 K) models and CDM. Assuming 10 QSOs at z=7.5 with S/N=30, R=10000.
Departed Wednesday evening
Multi-messenger Cosmological Black Holes - 10-14 Feb 2020
- (Woods+ 2017) Using the Kepler stellar evolution code, including a post-Newtonian correction (e.g. Fuller+ 1986) but without rotation.
- Initial condition: 10
$M_\odot$ with a central density$10^{-3}$ g/cm$^3$ and constant accretion rates. - At 1
$M_\odot$ /yr, there's a fully convective core, where the nuclear burning occurs, and a radiative envelope. The core grows to 50,000$M_\odot$ over 250 kyr. - The GR instability only happens at 250,000
$M_\odot$ when it's still H-burning. Very little escaping (enriched) matter with most of it collapsing into a DCBH.
- Initial condition: 10
- (Haemmerle+ 2018) Adding rotation. SMS need to be slow rotators,
$v_{\rm surf} < 10-20%$ of the critical rotation rate.- Preliminary results: For 1
$M_\odot$ /yr with pulsational mass losses of$10^{-3} M_\odot$ /yr, you need a$B \sim 10$ kG field for magnetic braking to lose angular momentum. Consistent with the upper end of observed B-fields. Gravitational torques and viscoity will help further.
- Preliminary results: For 1
- DCBH observational prospects
- (Hartwig+ 2018) merger rates for LISA predictions
- (Surace+ 2018, 2019) SMS spectra
- (Woods+ 2020, submitted) SMS lifetimes: nearly mass-indepedent of ~1 Myr.
- (Beckmann+ 2019) Using must-refine particles to have 0.01pc resolution around the SMBH.
- No feedback to probe the upper limit.
- As the resolution increases, the accretion rates descres and goes from a chaotic to episdoic pattern because the BH forms a (not an accretion) large-scale (~1pc) disk around it.
- These episodes are controlled by infalling clumps that disturb the nuclear disc andredistribute the angular momentum, driving a strong inflow.
- (Beckmann+ in prep) SMBH doesn't grow greatly to z=11. Grows very quickly from 250
$M_\odot$ to$10^6 M_\odot$ and doesn't grow more than a factor of a few after 200 Myr. - The accretion rates are around
$10^{-3} - 10^{-4}$ of Eddington because it doesn't exist in the center of galaxy. It typically exists between spiral arms of the nuclear disk. - There's no correlation between the gas supply at 10pc and the BH accretion growth. However, there is a very good correlation between 1pc and it.
- BH growth is dependent on the mass transfer from large-scale to sub-pc scales.
- (Habouzit+ 2019) AGN feedback in IllustrisTNG (and other large-scale simulations) helps suppress SF in massive galaxies to match observations (i.e. luminosity function).
- How does the choice of SF/BH subgrid models affect galaxy/BH properties? What diagnostics to rule out or improve models?
- Analysis and comparison of Illustris, TNG100, TNG300, Horizon-AGN, EAGLE, and SIMBA.
- Scaater of observations not fully reproduced (low mass BHs in galaxies)
- Simulations have low-mass galaxies with BHs that are more massive than observed and galaxies with stellar masses above what's observed.
- (Ding+ 2019) M-sigma relation may evolve at high redshift (z > 2.5), where it's pretty constant below that redshift
- (Habouzit+ in prep) Differences between Illustris and TNG
- Higher seed masses in TNG
- Stronger SN feedback in low-mass galaxies and at high-z in TNG, leading to the initial growth phase being delayed and then grow efficiently at higher masses, whereas Illustris has linearly growth.
- Stronger low accretion mode of AGN feedback (winds) in TNG, which suppresses growth.
- All simulations produce too many high-z (z~4) AGN but is tuned to match z=0 QSO LFs
- (Aird+ 2018) Observations of constraints on AGN fraction
- (Zhu+ 2020, in prep) Suite of 18 zoom-in simulations with Gizmo of a single z=6 QSO host, using
- Three different seeding models (stellar-mass, IMBH, DCBH)
- Different accretion models: Bondi-Hoyle, Eddington-limited, chaotic cold accretion, super-Eddington models
- Exploring both thermal and kinetic feedback models. Thin disk model controls the radiative efficiency
$\epsilon$ .
- Many differences in the SMBH/galaxy growth history, depending on the subgrid models.
- Exploring effects of substructures on the morphology of merging galaxies and how this affects binary SMBHs
- Torques from bars can be more efficient than dynamical friction in the evolution of SMBH orbits
- Small scale orbital decay: 3-body interactions, gas-BH interactions through a circumbinary disk, and 3-MBH interactions after some time (Bonetti+ 2016)
- (Biava+ 2019) Orbital decay timescales are very long (10 Gyr) for IMBH (
$< 10^5 M_\odot$ ) binaries. Need to get above$10^7 M_\odot$ to merge within 1 Gyr. - (Duffell+ 2019; Mackenzie+ 2019; Munoz+ 2019) However in these models, it's assumed that the gas at the edge of the gap doesn't flow inwards to the BHs. But this is a simplification that isn't seen in simulations. If gas is transffered to the BHs into "minidisks," then the orbit can shrink at a faster rate.
- Sesana & Khan (2015) post-process the results of SAMs with recipes for stellar- and gas-driven evolution, exploring the effects of triple MBH systems
- (Tremmel+ 2017) Most of the BH seeding happens in the first 2 Gyr, whereas in Illustris it's drawn out over 6 Gyr.
- (Tremmel+ 2018ab) Galaxy-scale orbital evolution prolongs the close-pair formation timescale when compared to SAMs
- When a host satellite galaxy survives orbital interactions (i.e. dense stellar cluster and high mass ratios), it can accelerate the mergers.
- When the galaxy is tidally destroyed, only dynamical friction is important that can result in a higher percentage in close pairs.
- Minor mergers result in common "lonely" BHs with the host galaxies disrupted.
- (Ricarte+ in prep) The number of lonely SMBHs increase with halo mass, going from 1 at
$10^{10} M_\odot$ to $\sim$100 at$10^{13} M_\odot$ . No dependence on large-scale environment, but central galaxies in clusters have more lonely SMBHs at high (z>4) redshifts, whereas there's more mass in the central SMBH than the lonely ones at z=0. - Because merger timescales are long, dual AGN aren't necessarily associated with ongoing mergers.
- Dynamical friction can explain dual BHs and the stellar mass dominates the gravitational potential at kpc-scales
- There's a big difference in the orbital evolution in force resolution. Showed between 1 and 20 pc resolution. Need to resolve the BH gravitational influence radius.
- (Pfister+ 2019) Implemented the Tremmel+ (2015) model into RAMSES
- (Pfister+ in prep)
$3 \times 10^9 M_\odot$ (stellar?) halo at z=6 with a$10^5 M_\odot$ seed BH with full stellar/BH feedback. Resolution of 9 pc. Dynamical friction model.- When studying the orbital evolution, the orbital decay timescale matches with the actual merger timescale in the simulation.
- However when the seed BH is decreased to
$10^4 M_\odot$ , it can't sink into the central because the dynamical friction is smaller and stalls at ~1 kpc. - The BHs are displaced by inhomogeneities in the potential and can have a hard time growing afterwards.
- Analysis of zoom-in Ponos simulation (Fiacconi+ 2017).
- Focusing on progenitors (up to z=6) of a
$10^{13} M_\odot$ z=0 halo.
- Focusing on progenitors (up to z=6) of a
- Comparing the estimated and simulated orbital decay timescales, where the latter is a factor of a few smaller. The difference is caused by inhomogeneities, i.e. mostly bars.
- The timescales in their 6 cases are ~100 Myr.
- Even if both BHs exist outside the central regions, the bar can drag both of them into the center.
- (Bonetti+ 2020) Using scattering experiments to investigate the binary MBH evolution
- Found that low mass ratio binaries have a larger than expected eccentricity evolution (i.e. Rasskazov+ 2019)
- Could be connected to the ejection of retrograde stars
- GR (post-Newtownian) decreases the fraction of binaries that merge
- In a SAM, they considered three scenarios: (1) fast coalesence (standard channels), (2) delayed coalesence (3-body interactions), and (3) no coalesence.
- (Bonetti+ 2019) Depending on the scenario, eLISA can detect (at S/N > 8) between 40 and 300 in four years. S/N, even at z~10, can be as high as 100.
- A stochastic GW background is detectable with PTA and SKA.
- (Bhowmick+ 2019) Inspecting MassiveBlackII for quasar pairs. Major mergers are triggers of the brightest (g > 21) AGN pairs.
- They also calculated the number of AGN in a system within 1 Mpc of each other. In a deg$^2$ there are ~10 (100) triples or quadrapules with the DESI survey (Vera Rubin Observatory).
- (Bhowmick+ in prep) Now inspecting TNG simulations and comparing them with MBII, searching for pairs within 0.1 and 1 Mpc/h. There are ~600 pairs, ~100 triples, and ~10 quads within 0.1 Mpc/h of each other in TNG100.
- At fixed halo mass and stellar mass, AGN with higher Eddington ratios are more likely to have companions.
- (Valiante+ 2011, 2012, 2016) Using a SAM that's data-constrained
- considers SF in minihalos and atomic cooling halos.
- BH mergers: Form binaries after a few Myr. Then BH binaries are stalled until next merger when a triple BH encounter happens using a probability from Bonetti+ (2018ab)
- Seed BHs with light and heavy seeds
- BH growth through accretion and mergers
- Stellar/BH mechanical, chemical, and radiative feedback
- (Valiante+ 2016) This model assumed instantaneous BH mergers after a halo merger. Now their current model (in prep) greatly suppresses BH growth from mergers. When they don't include accretion, the BH mass is
$10^6 M_\odot$ with little growth from mergers after z~10. - (Valiante+ 2018) Mock spectra
- DCBH hosts from when they grow from
$10^5 M_\odot$ to$10^7 M_\odot$ within 250 Myr. The QSO dominates over the stellar component. - With light seeds, the galaxy is starburst dominated. The BH only grows from
$80 M_\odot$ to$800 M_\odot$ over 400 Myr.
- DCBH hosts from when they grow from
- (in prep) Most of the mergers in the light seed model occur at z>6, whereas with heavy seeds, about 2/3 happen at z>6.
- Using merger trees from N-body simulations (Millennium, XXL, and Millennium II) in their SAM for BH/galaxy evolution
- With the spatial information from the simulations, they can calculate the LW flux and metal diffusion from the halos. They use that information to determine which halos form seeds
- They include a treatment of BH spin evolution with accretion and mergers. BHs in ellipticals tend to have a spin distribution favoring lower values.
- (in prep) BH kicks affect the occupation fraction in pseudobulges the most, but not so much in ellipticals (~60-80% reduced masses for a fixed stellar mass). Orphan BHs depends highly on the seed population.
-
J-PAS have 54 narrow-band filters (
150 Å), creating a rough (R50) spectra, and imaging. The field of view is very big (deg$^2$?).- Well-suited for QSO surveys: ID, photo-z, statistics, clustering, properties, environment, and host galaxy properties
- (Habouzit+ 2017) Need SN feedback to limit runaway BH growth after z~6 so that they agree with observed QSO LFs
- (Lupi+ 2019) Constrainted cosmological simulation of a
$3 \times 10^{12} M_\odot$ halo at z=6.- There's a dichotomy between the SFR and BH accretion rate above 100
$M_\odot$ /yr, where there's a jump by a factor of 10 in BH accretion rate. - Observations never measure stellar mass because the QSO overwhelms the stellar component. The dynamical mass is measured through the QSO emission lines, and the molecular gas is measured with ALMA.
- Therefore, they calculate the [CII] intensity maps of the simulated host galaxy and convolute it with the ALMA PSF
- This galaxy aligns with the present-day M-sigma relation, however the smoothing over the PSF tends to decrease the stellar/dynamical mass by a factor of 2-10.
- There's a dichotomy between the SFR and BH accretion rate above 100
- (Trebitsch+ 2020) Obelisk simulation. Using RAMSES-RT, they simulate a protocluster down to z=3. All the usual RAMSES-RT subgrid physics, including BPASS stellar models
- (20 Mpc)$^3$, dx = 35 pc,
$M_{\rm DM} = 10^6 M_\odot$ ,$M_{\rm BH} = 3 \times 10^4 M_\odot$ seeds. Zoom-in within Horizon-AGN. - Galaxy LF flattens above
$M_{\rm UV} = -16$ , although this samples a highly biased region - Also found that AGN activity are more likely in biased regions
- Within 100 pc of the BHs, stars dominate the overall matter density. Used these densities to estimate the BH merger times, which suggests that mergers occur frequently in these galaxies.
- With the higher resolution (than the Horizon simulations), it captures many more BH mergers. The biased regions host approximately 3-5x more mergers.
- (20 Mpc)$^3$, dx = 35 pc,
(di Matteo) Multimessenger Astrophysics with MBHs: large volume hydrodynamic cosmological simulations
- (Feng+ 2015) Bluetides simulation. 500 pc resolution with P-GADGET with a (400 Mpc/h)$^3$ boxsize
- The large volume gives plenty of statistics to construct galaxy UVLF between
$M_{\rm UV}$ = -24 and -16. - There are a few rare SMBHs that grow very rapidly from
$10^7$ to$10^9 M_\odot$ during z=8-10. Their host galaxies are not disky galaxies. - (di Matteo+ 2017) The differences between a halo that hosts a very massive SMBH compared to a lesser one (difference of 100) depends on the tidal tensor (measured at 1 Mpc/h). This measures how filamentary the LSS is. The biggest SMBH exist in halos with mostly radial inflows from filaments.
- (Tennati+ 2018) Mock observations of a z~7 QSO/galaxy. Compared with observations, it matches well. The galaxy is very compact and massive.
- (DeGraf+ in prep) Using the Illustris simulation, SMBHs can grow in pairs when their mergers are delayed by dynamical processes
- Suppression of BH merger rates at z > 2 with no differences at later times.
- Host galaxies of BH mergers/pairs show distrurbed morphologies. Can use Gini vs.
$M_{20}$ observational relation to isolate these galaxies.
- After 4 years of LISA operation, we can expect 100+ detections, 100+ systems with localization to 10 deg$^2$, 100+ systems with individual masses determined to 1%, 50 systems with primary (secondary) spin determinations to 0.01 (0.1), 50 systems with spin directions within 10 degrees, 30 events with final spin determined to 0.1
- (Sesana+ 2011) SAM with different BH seeding (light vs heavy), metallicity feedback (metal-free vs. metal-enriched), accretion efficiencies (
$f_{\rm Edd}$ ), and accretion geometries (coherent vs. chaotic) - (D'Ascoli+ 2018; Tang+ 2017, 2018) Simulations of mass accretion onto binary BHs during and after merger and how this connects to the EM counterpart.
- (McGee+ 2020) Athena + LISA joint detections. One of the figures shows a "waterfall" plot overlaying the sky localization (~0.4 deg$^2$) with LISA and the necessary exposure time with Athena (10 ks) for detection. Sweet spot is between
$10^5 - 10^7 M_\odot$ . - With a BBH with
$10^6 M_\odot$ at z=0.5, the sky localization is 10 deg$^2$ about 2-3 days before merger, which is about the VRO field of view.
- (Burke-Sporaer+ 2011) There's a single robust binary (7.8 pc separation; gravitationally bound) detected with VLBI
- (Foord+ 2019) Possible dual AGN observed with Chandra that showed some oblong emission. If it's a binary, they'd be separated by 700pc.
- (Foord+ 2020) BAYMAX is their code that identify dual AGN in flux-limited observations, using a Bayesian framework. They used it to determine that their dual AGN candidate (2019 paper) is a single AGN.
- (in prep) They're currently analyzing triple AGN with BAYMAX, and determining the dual AGN fraction as a function of time, which they can compare with simulations (i.e. EAGLE).
- (Whitley+ submitted) Gapped-disk simulation of accretion onto a BH within a binary.
- IC:
$10^8 M_\odot$ BH primary with q=0.01, 100$R_g$ separation. - There are accretion "arms" between (1) the circumbinary disk and the minidisk, and (2) the two minidisks that appear in opposite phases (i.e. 180 degrees).
- A good observational measure of such a system is the evolution of
$L_{\rm UV} / L_{\rm X-ray}$ that has major/minor modes that correspond to each arm. The orbital period in this system is 0.1 yr.
- IC:
- (Burke-Spolaor+ 2017) Recoiling BH in Abel 2261 BCG? But the VLA and HST data cannot prove this.
- X-ray occupation fractions are >50% (and consistent with 100%) of massive BHs in low-mass galaxies (
$M_\star = 10^9 - 10^{10} M_\odot$ ) - Interesting implications for TDEs and EMRIs and seeding mechanisms
- AGN in dwarf galaxies. Example papers
- Optical: Reines+ (2013), Baldassare+ (2018) that used variability signatures
- UV/X-ray: Gallo+ (2010), Satori (2015), Kaviraj+ (2019)
- Mid-IR: Satyapal+ (2014), Kaviraj+ (2019)
- Radio: Reines+ (2010), Bellovary+ (2019)
- (Woo+ 2019) Detection of
$10^4 M_\odot$ BH in a bulgeless dwarf. It lies on the M-sigma relation even though it doesn't have a bulge. Does this suggest secular evolution? - (Gallo & Sesana 2019) BH mass function below
$10^6 M_\odot$ , using BH-stellar mass relation (Reines & Volonteri 2015), GAMA galaxy mass function, and X-ray constraints
- SKA will be constructed in 4 phases, where it will become more sensitive 4 years after construction start (2025) than LOFAR and MeerKAT.
- IR-radio SFR correlation: (Magnelli+ 2015, Delhaize+ 2017, Calistro-Rivera+ 2017) SFR ~
$10^{q_{\rm IR}} L_{1.4}$ , where$q_{\rm IR}$ depends on redshift and is given in the references and$L_{1.4}$ is given at 1.4 GHz.
- XRBs dominate over AGN at z>6. Possible high-z X-ray heating sources: XRBs, thermal emission from galaxies, BHs, DM annihilation, CRs, B-fields
- (in prep) Using a binary evo code (Binary_c) to estimate the number and properties of high-z XRBs, especially in metal-poor stars. Modeling host absorption with X-ray radiative transfer. Goal: source model population
- (Mirocha+ 2017; Cohen+ 2017) 21-cm signals, both the power spectrum and the global signal, are very sensitive to star formation/feedback modeling
- (Cohen+ 2019; Monsalve+ 2019) Trained an ANN with 30k models to emulate global signals
- Requires SF in minihalos, excludes inefficient heating
$f_X < 0.0042$ , favors soft X-rays$E_{\rm min} > 2.3$ keV, exclude high optical depth$\tau > 0.08$
- Requires SF in minihalos, excludes inefficient heating
- (Hickox+ 2009; Magliocchetti+ 2020) Overlap of emission properties (i.e. strong in X-ray, radio, and/or IR) in AGNs and how it depends on the host galaxy
- (Magliocchetti+ 2014-2018) Determining the observational criteria for AGN/SF division in galaxies
- No quenching down to z=1, suggesting that either (1) AGN and SF aren't connected to each other, or (2) AGN activity stimulates SF (positive feedback)
- Calculating the 2-pt correlation function, they determine a clustering radius of ~11 cMpc. Several groups find no evolution over redshift. When used with abundance matching, the minimum host halo mass is ~$10^{13.5} M_\odot$. Doing the same analysis for SF galaxies, they find
5 Mpc ($10^{11.5} M_\odot$ ).
- Using ML to classify SF galaxies and AGN through many different ways, i.e. far-IR, emission lines, and also how to determine photo-z
- Have constructed an ML team at University of Miami, called BlackBase, to construct a ML framework for AGN classification.
- Training set includes ~150 catalogs from the literature
- (Cappelluti+ 2013, 2017, Mitchell-Wynne+ 2016) Analyzed the far-IR and X-ray background excess. It was first attributed to the first stars and their remnants. However Ricarte+ (2019) found that it can't be explained by BH seeds (Pop III or DCBH).
- What can cause it? Missing moderate redshift AGN?
- No evidence for changes in accretion physics out to z=6.
- (Cowie+ 2020) New observational limits for AGN number density out to z~8
- Athena will detect ~300 AGN at z=6-8 down to L* that can begin to constrain models
- AXIS and/or Lynx needed to probe up to z~10 and higher and down to luminosities low enough to probe seeds