9.4 High-Redshift Galaxies and Cosmic Evolution

High-Redshift Galaxy Types

High-redshift galaxies let us observe the universe as it was billions of years ago. By identifying galaxies at redshifts $z > 2$, astronomers can directly study the conditions of early star formation, chemical enrichment, and black hole growth. The main populations used for this work each probe different physical regimes of early galaxy evolution.

Lyman-Break and Lyman-Alpha Galaxies

Lyman-break galaxies (LBGs) are identified through a sharp drop in flux blueward of the Lyman limit (912 Å in the rest frame). This break arises because neutral hydrogen, both within the galaxy and in the intervening intergalactic medium (IGM), absorbs nearly all photons with energies above 13.6 eV. As the galaxy's redshift increases, this spectral break shifts into observable optical bands, making it detectable with broadband photometric filters.

The technique works like this:

1. Image a field through multiple broadband filters spanning the optical range.
2. Look for objects that appear in redder filters but "drop out" of bluer ones.
3. The filter where the dropout occurs constrains the redshift. For example, a $U$-band dropout corresponds to $z \sim 3$, while an $i$-band dropout corresponds to $z \sim 6$.
4. Follow up spectroscopically to confirm the redshift and measure physical properties.

This dropout technique has been the workhorse method for building large samples of galaxies at $z \gtrsim 2.5$.

Lyman-alpha emitters (LAEs) are selected instead by a strong emission line at a rest wavelength of 1216 Å. This line is produced when ionized hydrogen in active star-forming regions recombines and the resulting cascade produces a Lyman-alpha photon. Because Lyman-alpha is a resonance line, it scatters extensively through neutral gas, so its detection is sensitive to the geometry and kinematics of the interstellar and circumgalactic medium. LAEs tend to be lower-mass, less dusty systems compared to LBGs, and narrowband filter surveys tuned to specific redshifts can efficiently find them.

Together, LBGs and LAEs provide complementary views of the early galaxy population. LBGs sample the UV-bright, moderately massive end, while LAEs probe lower-mass, less obscured systems. Both populations are essential for constraining the faint-end slope of the UV luminosity function and calibrating cosmic star formation rate estimates.

Submillimeter Galaxies and Quasars

Submillimeter galaxies (SMGs) are detected at wavelengths around 850 μm, where thermal emission from warm dust peaks for galaxies at $z \sim 2\text{–}5$. These are among the most intensely star-forming systems in the universe, with star formation rates reaching $\sim 100\text{–}1000 \; M_\odot \; \text{yr}^{-1}$. Their heavy dust obscuration means they are often faint or invisible in rest-frame UV surveys, so they represent a population that LBG selection largely misses.

A useful property of submillimeter observations is the negative K-correction: the steep rise of the dust spectral energy distribution on the Rayleigh-Jeans side means that an SMG at $z \sim 4$ can appear nearly as bright at 850 μm as one at $z \sim 1$. This makes submillimeter surveys roughly equally sensitive to dusty starbursts across a wide redshift range.

Quasars are the most luminous active galactic nuclei (AGN), powered by accretion onto supermassive black holes at rates approaching or exceeding the Eddington limit. At high redshift, they serve two roles:

• Probes of early black hole growth. The existence of quasars with black hole masses $\gtrsim 10^9 \; M_\odot$ at $z > 6$ places strong constraints on seeding mechanisms and early accretion histories.
• Backlights for the IGM. Absorption features in quasar spectra (the Lyman-alpha forest, damped Lyman-alpha systems, and the Gunn-Peterson trough) reveal the density, temperature, ionization state, and metallicity of intervening gas along the line of sight.

SMGs and quasars together fill in the picture that UV-selected samples miss: SMGs trace the dust-obscured component of cosmic star formation, while quasars trace the most extreme episodes of black hole accretion and provide sightlines through the evolving IGM.

Galaxy Evolution and Assembly

Galaxy Luminosity Function and Downsizing

The galaxy luminosity function $\phi(L)$ describes the number density of galaxies per unit luminosity per unit comoving volume. It is the most fundamental statistical description of the galaxy population at any epoch.

The standard parameterization is the Schechter function:

$\phi(L) \, dL = \phi^* \left(\frac{L}{L^*}\right)^\alpha e^{-L/L^*} \frac{dL}{L^*}$

• $\phi^*$ is the normalization (number density at $L^*$).
• $L^*$ is the characteristic luminosity where the function transitions from a power law to an exponential cutoff. Galaxies brighter than $L^*$ are exponentially rare.
• $\alpha$ is the faint-end slope. A more negative $\alpha$ means a steeper rise in the number of faint galaxies.

Tracking how $\phi^*$, $L^*$, and $\alpha$ evolve with redshift reveals how the galaxy population grows and transforms over time. For instance, the rest-frame UV luminosity function at $z \sim 4\text{–}8$ shows a steepening faint-end slope ($\alpha \sim -1.6$ to $-2.0$), indicating that faint, low-mass galaxies increasingly dominate the total UV luminosity density at earlier times.

Downsizing is the empirical observation that the most massive galaxies completed the bulk of their star formation earlier and on shorter timescales than lower-mass galaxies. Evidence comes from multiple directions:

• Massive ellipticals at $z \sim 0$ have old, $\alpha$-enhanced stellar populations, indicating short formation timescales ($\lesssim 1$ Gyr).
• The characteristic mass above which galaxies are predominantly quenched shifts to lower masses at lower redshifts.
• The specific star formation rate (star formation rate per unit stellar mass) at a given epoch is lower for more massive galaxies.

This trend runs counter to the simplest predictions of hierarchical structure formation, where massive halos assemble later from smaller ones. Resolving this tension requires invoking mass-dependent feedback, particularly AGN feedback in massive halos that shuts down star formation early.

Cosmic Star Formation History and Galaxy Assembly

The cosmic star formation rate density (SFRD) traces the total mass of stars formed per unit time per unit comoving volume as a function of redshift. Its shape is now well established:

• SFRD rises steeply from early times, reaching a broad peak around $z \sim 1.5\text{–}3$ (often called "cosmic noon", roughly 2 to 4 billion years after the Big Bang).
• It then declines by roughly an order of magnitude from $z \sim 2$ to $z = 0$.

Different tracers sample different components of star formation:

• Rest-frame UV luminosity traces unobscured star formation from young, massive stars.
• Far-infrared/submillimeter emission traces dust-reprocessed light from obscured star formation.
• Radio continuum emission (synchrotron from supernova remnants) provides a dust-independent tracer.

Combining UV and IR measurements is essential for a complete census, since dust obscuration accounts for roughly half or more of the total star formation at cosmic noon.

Galaxy assembly proceeds through several channels:

• Major mergers (mass ratios $\lesssim 3{:}1$) can trigger intense starbursts, funnel gas to the nucleus (potentially igniting AGN), and transform disk morphologies into spheroids.
• Minor mergers (mass ratios $\gtrsim 10{:}1$) build up stellar halos and contribute to gradual mass growth without dramatically altering the central structure.
• Cold-mode accretion of gas along cosmic web filaments feeds star formation in a more continuous fashion, particularly important at high redshift where gas fractions are high.

Feedback processes regulate this growth and are critical for reproducing observed galaxy properties in simulations:

• Stellar feedback (radiation pressure, stellar winds, supernovae) is most effective in low-mass halos, driving galactic outflows that reduce star formation efficiency and enrich the circumgalactic medium with metals.
• AGN feedback operates in two modes: radiative ("quasar mode") feedback during high-accretion episodes can expel gas from the central regions, while kinetic ("maintenance mode") feedback from radio jets in massive ellipticals prevents cooling of hot halo gas, keeping these galaxies quenched.

These feedback mechanisms are central to explaining the galaxy mass-metallicity relation, the bimodal color distribution of galaxies, and the sharp exponential cutoff at the bright end of the luminosity function.

Cosmic Structure and Reionization

Reionization and the Cosmic Web

Reionization is the phase transition during which the intergalactic medium went from being almost entirely neutral (after recombination at $z \sim 1100$) to almost entirely ionized. Current constraints place this transition broadly between $z \sim 6$ and $z \sim 15\text{–}20$, with the bulk of reionization completing by $z \sim 6$.

The primary sources of ionizing photons are thought to be star-forming galaxies, particularly faint, low-mass systems that dominate the UV luminosity density at high redshift. AGN contribute as well, though their number density at $z > 6$ appears too low to drive reionization alone.

Key observational probes of reionization include:

• Gunn-Peterson absorption in quasar spectra. Complete absorption of Lyman-alpha flux (a Gunn-Peterson trough) in quasars at $z \gtrsim 6$ indicates a substantially neutral IGM.
• CMB polarization. Thomson scattering of CMB photons off free electrons produced during reionization generates a large-angle polarization signal. Planck measurements yield an optical depth $\tau \approx 0.054$, corresponding to a midpoint of reionization around $z \sim 7.7$.
• 21-cm cosmology. The hyperfine transition of neutral hydrogen at 21 cm, redshifted to meter wavelengths, can in principle map the three-dimensional distribution of neutral gas throughout reionization. Experiments like HERA and the SKA aim to detect this signal.

Reionization did not proceed uniformly. It was patchy, with ionized bubbles forming first around overdense regions with clustered sources, then expanding and eventually overlapping. This topology connects directly to the large-scale distribution of matter.

The cosmic web is the large-scale structure that emerged from gravitational amplification of primordial density fluctuations. Dark matter collapsed first into halos connected by filaments, with sheets and voids filling the remaining volume. Baryonic matter followed the dark matter potential wells, accumulating in halos and along filaments. Galaxies preferentially form at the nodes and along the filaments of this web, while voids remain largely empty.

Large-Scale Structure and Cosmological Implications

The distribution of galaxies on large scales encodes information about both the initial conditions of the universe and the physics governing structure growth.

Baryon acoustic oscillations (BAO) are a particularly clean probe. Sound waves in the pre-recombination plasma imprinted a characteristic scale of $\sim 150$ Mpc (comoving) in the matter distribution. This scale appears as a bump in the galaxy two-point correlation function (or equivalently, oscillations in the power spectrum) and serves as a standard ruler. Measuring the BAO scale at different redshifts constrains the expansion history $H(z)$ and the angular diameter distance $d_A(z)$, providing tight constraints on dark energy parameters.

Galaxy clustering statistics more broadly constrain the matter density $\Omega_m$, the amplitude of fluctuations $\sigma_8$, and the growth rate of structure, all of which test the $\Lambda$CDM model.

Major redshift surveys have mapped the three-dimensional galaxy distribution in progressively greater detail:

• The Sloan Digital Sky Survey (SDSS) mapped over a million galaxy redshifts, providing the first high-precision BAO measurements.
• The 2dF Galaxy Redshift Survey independently confirmed the BAO signal and measured the growth rate of structure.
• Current and upcoming surveys (DESI, Euclid, the Vera Rubin Observatory's LSST) will extend these measurements to higher redshifts and fainter galaxies.

Weak gravitational lensing offers a complementary approach. The gravitational field of foreground mass concentrations introduces small, coherent distortions (shear) in the shapes of background galaxies. By statistically measuring these shape distortions across large areas of sky, you can reconstruct the projected mass distribution, including dark matter that emits no light. This provides a direct measurement of the matter power spectrum, independent of assumptions about how galaxies trace mass (the galaxy bias).

Extending these measurements to higher redshifts is a major goal of current observational programs. Comparing the amplitude and growth of structure at $z \sim 1\text{–}3$ with predictions from $\Lambda$CDM tests whether the standard model holds or whether modifications (evolving dark energy, modified gravity) are needed.