Observing the First Billion Years

Observing the First Billion Years

Knowledge Ark · Universe · Chapter 02 / Article 10

Observing the first billion years

Ancient light is arriving now. By gathering it, separating its wavelengths, and reading what is missing, astronomers reconstruct the universe’s earliest galaxies.

Deep observationsCosmic dawnFrom light to evidence
Conceptual deep-field view with faint galaxies and gravitational-lensing arcs. Positions, brightnesses, and lensing geometry are illustrative; this is not a survey image.
Find the sourceDeep images reveal faint objects and the colors that make them promising candidates.
Establish its epochSpectral features help determine how much the light has stretched during its journey.
Test the explanationModels connect that light to stars, gas, dust, and the source’s surroundings.
The final article in this chapter

How do we know what happened so long ago?

This chapter has followed matter from small density fluctuations to stars, galaxies, and active black holes. The evidence for that story reaches us as light and other measurable signals.

During the universe’s first billion years, early stars began shining, galaxies assembled, and radiation started transforming intergalactic hydrogen. Cosmic dawn describes the emergence of those first luminous sources; reionization was the extended, uneven process that followed.[1]

Observations show individual systems at particular moments. By comparing many sources across different epochs, astronomers investigate how that early population changed.

01
A distant image records an earlier moment

How can a telescope look back in time?

Light travels at a finite speed. When we observe a distant galaxy, we see the state it was in when that light began its journey.

Space also expands during the journey, stretching the light toward longer wavelengths. Astronomers describe that change with redshift, written z. Much of the ultraviolet and visible light emitted by very early galaxies now reaches us in the infrared.

Redshift becomes a cosmic age only after an expansion model is specified. Light-travel time and present-day distance are different quantities in an expanding universe.[2]

Is the first billion years a fixed redshift range?

In standard cosmology, z ≈ 6 corresponds to a universe roughly 0.9–1.0 billion years old. Higher redshifts describe earlier epochs, but a survey’s practical reach depends on source brightness and instrument sensitivity.[14]

The most distant galaxy detected in a survey is not necessarily the first galaxy that formed. Earlier or fainter systems may lie beyond the observations.

02
The absence of light in one band can be a useful clue

How do astronomers find a faint early galaxy?

Deep exposures collect more photons, while images through different filters measure brightness across wavelength bands. A promising distant galaxy may appear in longer-wavelength images while becoming undetectable in shorter-wavelength ones.

This is the Lyman-break, or dropout, method. At very high redshift, intervening hydrogen strongly suppresses the light shortward of the redshifted Lyman-alpha wavelength, originally about 121.6 nanometers. The position of that drop helps estimate the redshift.[3]

Illustrative example · A galaxy at z = 10

One source, different wavelength bands

0.90 μmStrongly suppressed
1.15 μmStrongly suppressed
1.50 μmLight can reach us
2.00 μmLight can reach us
The Lyman-alpha break lies near 1.34 μm in this example. These are idealized wavelength samples, not real telescope images or named filter passbands. Detection also depends on the source’s spectrum and exposure depth.

Dusty foreground galaxies, cool nearby objects, unusually strong emission lines, and noise can imitate parts of this pattern. A photometric redshift is therefore a model-based estimate, often with competing possible solutions.[3]

Foreground light, detector noise, and overlapping images of nearby sources make the search harder. Researchers measure backgrounds carefully and check that a detection persists across independent exposures.

How is this different from the 91.2-nanometer Lyman limit?

The Lyman limit is the threshold for a photon to ionize ground-state hydrogen. Lyman-alpha is a different transition, at 121.6 nanometers.

In the very distant universe, cumulative absorption makes the practical dropout boundary lie near redshifted Lyman-alpha. Using 91.2 nanometers for that boundary would place the example at the wrong observed wavelength.

03
A spectrum reveals structure that broad-band colors cannot

What turns a candidate into a secure detection?

Spectroscopy spreads light into a much more detailed sequence of wavelengths. Emission lines, absorption features, and the shape of a continuum break can then test the candidate’s proposed redshift.

Several consistent features are especially persuasive. A well-measured break can also establish a high redshift when bright lines are absent; early JADES observations demonstrated this with galaxies beyond z = 10.[7]

One weak feature still needs scrutiny. Researchers check alternative line identifications, detector artifacts, foreground contamination, and whether the rest of the spectrum fits the same explanation.

Why follow-up matters

CEERS-93316 initially appeared consistent with z ≈ 16. Spectroscopy placed it at z = 4.9: dust and strong emission lines had created misleading colors. Its revised distance made the population census more reliable.[21]

The same galaxy can require different infrared capabilities
Spectral feature At emission Observed at z = 10
Hydrogen Lyman-alpha 0.1216 μm 1.34 μm
[O III] optical oxygen line 0.5007 μm 5.51 μm
Hydrogen H-alpha 0.6563 μm 7.22 μm

Calculated from the wavelength-stretching relation, with rounded values. Wavelength coverage does not guarantee that a line is bright enough to detect.

Webb’s NIRSpec covers approximately 0.6–5.3 μm. At z = 10, the two optical lines above have shifted beyond that range. MIRI reaches longer wavelengths, although faint-source sensitivity and observing time still limit what can be measured.[5], [6]

04
Each instrument samples a different part of the evidence

Why do we need more than one observatory?

An image reveals where the light is; a spectrum helps explain what produced it. Neither alone describes every component of an early galaxy.

Find and resolve faint infrared sources

Webb imaging

NIRCam records deep images through multiple filters, revealing colors, compact structures, and star-forming clumps. MIRI adds longer infrared wavelengths that can constrain older starlight or dust, depending on redshift.[4], [6]

Separate the light into wavelengths

Webb spectroscopy

NIRSpec can observe many targets through its microshutters or obtain spectra across a small resolved region. These modes connect redshift, gas properties, and spatial structure.[5]

Add legacy depth and wider searches

Hubble and ground-based telescopes

Hubble’s optical and near-infrared fields remain valuable. Facilities such as Subaru, Keck, and the VLT contribute surveys and spectroscopy, including searches through narrow filters for strong emission lines.[3]

Measure cool dust and gas tracers

ALMA

Millimeter and submillimeter observations can detect redshifted atomic lines and dust emission. The [C II] carbon and [O III] oxygen lines trace different gas conditions; they are distinct from molecular lines such as CO.[10], [19]

Matching these views helps separate stars, dust, gas, and possible active black holes. Upper limits also matter: a missing dust or line detection constrains what could be present, given the observation’s sensitivity.

05
Foreground gravity can make a distant source easier to study

How does gravitational lensing extend our reach?

A massive foreground cluster bends light from galaxies behind it. This gravitational lensing can stretch a source into an arc, magnify it, or create several images of the same galaxy.

Magnification increases apparent area and total received flux while conserving surface brightness. In favorable cases, a faint galaxy becomes detectable or its internal regions become easier to resolve.[20]

Hubble’s Frontier Fields used this natural magnification to study intrinsically faint galaxies. Webb builds on that approach with more sensitive infrared observations.[8], [20]

Blank fields and lensed fields therefore complement one another: the former provide a simpler view of a population, while the latter can expose otherwise inaccessible detail.

06
The discoveries are already reshaping models of early galaxies

What have we learned from the light?

A distant galaxy confirmed with Webb

MoM-z14

z = 14.44

In a study published in 2026, Naidu and colleagues reported this luminous galaxy seen about 280 million years after the Big Bang.

NIRSpec measured a sharp Lyman-alpha break, supported by several faint ultraviolet emission features. The observation extends the period in which galaxies can be studied through their spectra.[9]

A complementary measurement with ALMA

JADES-GS-z14-0

Oxygen at z ≈ 14.18

Schouws and colleagues detected the redshifted [O III] 88 μm line, obtaining a precise redshift for a galaxy seen about 300 million years after the Big Bang.

Oxygen shows that stellar processing had already enriched its gas. The line also gives information about ionized material that complements Webb’s starlight measurements.[10]

Brightness is only the beginning

Stellar mass, age, and star-formation rate are inferred by comparing the measured spectrum and colors with models. Young massive stars can dominate the light while representing only part of the total stellar mass.

Dust, stellar evolution, binary stars, and the assumed distribution of stellar birth masses—the initial mass function, or IMF—affect those inferences. A bright ultraviolet source is not automatically an exceptionally massive galaxy.[11]

Simulations by Sun and colleagues showed how bursts of star formation can increase the number of ultraviolet-bright galaxies without requiring the same increase in stellar mass. This is one explanation being tested alongside changes in star-formation efficiency and other galaxy physics.[12]

Does an oxygen line give a direct metallicity measurement?

A metal line establishes that the emitting element is present. Turning its strength into an abundance requires additional information: gas temperature and density, ionization state, the radiation source, and often other lines.

Different combinations of these conditions can produce similar spectra. Conversely, failing to detect a metal line does not establish that the gas contains no metals.[19]

07
Galaxies illuminate—and transform—the gas between them

Can we observe reionization as it unfolds?

Reionization proceeded in patches as radiation from early sources ionized surrounding hydrogen. Researchers combine several kinds of evidence to investigate when those regions grew and how they connected.

A spectral view along a line of sight

Lyman-alpha transmission

Neutral hydrogen strongly scatters Lyman-alpha photons. Whether a galaxy’s line reaches us depends on surrounding gas, the line’s velocity shift, and the galaxy’s own production and escape of radiation.

Deeper observations reported in 2026 reaffirmed Lyman-alpha emission from GS-z13-1-LA at z ≈ 13.1. The result supports very early ionizing activity; the size and origin of the surrounding ionized region remain model-dependent.[13]

A radio view of neutral hydrogen

The 21-centimeter signal

Neutral hydrogen has a weak radio transition emitted at a wavelength of about 21 centimeters. Cosmic expansion shifts it to longer radio wavelengths.

At z = 8, its frequency falls from roughly 1,420 MHz to 158 MHz. Its contrast against the background depends on neutral fraction, density, temperature, and gas motions—not simply on where hydrogen exists.[22]

What is measured today · Research published in 2026

HERA’s Phase II results report upper limits on fluctuations in the cosmic-dawn and reionization signal. They constrain possible histories of the gas rather than provide a confirmed three-dimensional image of ionized bubbles.[18]

SKA-Low is designed to pursue more sensitive measurements and, ultimately, maps of this changing gas. The challenge is separating a very faint cosmological signal from much brighter radio foregrounds and instrumental effects.[23]

How do a sky average, a fluctuation measurement, and a map differ?

A sky average measures the mean radio spectrum. A fluctuation measurement measures how strongly the signal varies across spatial scales. A map aims to locate those variations across the sky and in redshift.

They answer related questions with different observing and calibration requirements. A statistical constraint can be valuable well before detailed imaging becomes possible.[18], [22], [23]

The Cosmic Microwave Background adds another constraint: scattering by free electrons leaves an integrated imprint on its polarization. That helps constrain the overall ionization history, while individual galaxy spectra probe particular places and times.[14]

These measurements build on the processes described in Reionization: Ending the Dark Ages.

08
Better observations test the population as well as the exceptional objects

What would make the next discoveries more decisive?

Reaching an earlier epoch is one goal. Equally valuable is learning how representative the detected galaxies are: which objects a survey misses, which bright phases it favors, and how much the result varies from one field to another.

Researchers insert simulated sources into images and test which ones their detection methods recover. They also compare independent fields and selection techniques. These checks turn a catalog of detections into a better estimate of the underlying population.[3], [8]

A broader census

Survey more sky

Roman launched on 30 August 2026. Its planned wide-field infrared surveys will help place detailed observations of individual sources within much larger samples.[17]

More detailed spectra

Resolve more of the physics

Continued Webb observations and the future 39-meter Extremely Large Telescope aim to probe fainter sources and more detail in early galaxies.[5], [16]

Connect sources to their surroundings

Compare galaxies and gas

Combining galaxy positions with radio statistics can test how luminous sources heated and ionized their environment.[23]

What would identify the first, metal-free stars?

A convincing Population III interpretation needs several consistent clues, including a very energetic radiation field and stringent limits on metal enrichment.

A 2026 study reported strong helium emission without detected metal lines near GN-z11 and argued that Population III stars provide the most plausible explanation. The observational signal and its physical interpretation should still be distinguished: a metal-line nondetection is a sensitivity limit, not a direct measurement of exactly zero metals.[15], [19]

Progress comes from explaining the observations together: the source counts, spectra, gas conditions, and changing environments. Each additional method makes the reconstruction of early cosmic history more demanding—and more informative.

Chapter 02 complete · The Emergence of Large-Scale Structures

The early universe leaves evidence we can read.

A faint source, a missing band of light, an oxygen line, or a constraint on radio fluctuations each reveals a different part of the story. Together, they connect the first luminous objects to the larger structures that followed.

The next chapter turns to galaxy formation and evolution: how galaxies acquired their structures, changed through interactions, and became the diverse systems we observe across cosmic time.

Sources and further reading

Observations and mission status reviewed in September 2026. Cosmic ages are approximate; inferred galaxy properties depend on the observations and physical models used.

  1. NASA — Early UniverseThe questions behind observations of the first stars and galaxies.
  2. Hogg (1999) — Distance measures in cosmologyRedshift, cosmic distances, lookback time, and cosmological dimming.
  3. Finkelstein (2016) — Observational Searches for Galaxies at z > 6The principles of distant-galaxy selection and the limits of imaging searches.
  4. STScI — JWST Near Infrared CameraNIRCam imaging and observing capabilities.
  5. STScI — JWST Near Infrared SpectrographNIRSpec wavelength coverage and spectroscopic modes.
  6. STScI — MIRI Medium Resolution SpectroscopySpectroscopy at longer infrared wavelengths.
  7. Curtis-Lake et al. (2023) — Spectroscopic confirmation of four metal-poor galaxies at z = 10.3–13.2An early demonstration of Webb’s ability to confirm high-redshift candidates.
  8. Livermore, Finkelstein & Lotz (2017) — Directly Observing the Galaxies Likely Responsible for ReionizationUsing cluster magnification to extend searches toward fainter galaxies.
  9. Naidu et al. (2026) — A Cosmic Miracle: A Remarkably Luminous Galaxy at zspec = 14.44 Confirmed with JWSTMoM-z14, its spectrum, and the abundance of very early luminous galaxies.
  10. Schouws et al. (2025) — Detection of [O III] 88 μm in JADES-GS-z14-0 at z = 14.1793An ALMA oxygen-line measurement of a galaxy about 300 million years after the Big Bang.
  11. Conroy (2013) — Modeling the Panchromatic Spectral Energy Distributions of GalaxiesInferring stellar masses, star formation, dust, and stellar populations from light.
  12. Sun et al. (2023) — Bursty Star Formation Naturally Explains the Abundance of Bright Galaxies at Cosmic DawnA simulation-based explanation for part of the high-redshift brightness puzzle.
  13. Witstok et al. (2026) — An OASIS of Lyman-α within a neutral intergalactic desertDeeper observations of Lyman-alpha emission from GS-z13-1-LA.
  14. Planck Collaboration (2020) — Planck 2018 results VI: Cosmological parametersCMB constraints on the expansion history and the integrated effects of reionization.
  15. Maiolino et al. (2026) — The search for Population III: Confirmation of a He II emitter with no metal lines at z = 10.6A reported helium-emitting region near GN-z11 and its Population III interpretation.
  16. ESO — ELT timelineThe construction and planned scientific capabilities of the Extremely Large Telescope.
  17. NASA (2026) — Nancy Grace Roman Space Telescope launchesThe August 2026 launch and the mission’s planned wide-field observations.
  18. HERA Collaboration / Abdurashidova et al. (2026) — First Results from HERA Phase IIStatistical upper limits on the redshifted 21-centimeter signal.
  19. Gutkin, Charlot & Bruzual (2016) — Modelling the nebular emission from primeval to present-day star-forming galaxiesHow gas conditions and stellar populations affect emission-line diagnostics.
  20. Zitrin (2026) — Strong Gravitational Lensing with the James Webb Space TelescopeMagnification, surface brightness, and the interpretation of lensed sources.
  21. Arrabal Haro et al. (2023) — Confirmation and refutation of very luminous galaxies in the early universeSpectroscopy resolves the misleading colors of candidate CEERS-93316.
  22. Furlanetto, Oh & Briggs (2006) — Cosmology at Low Frequencies: The 21 cm Transition and the High-Redshift UniverseThe physics of hydrogen’s radio signal and its temperature dependence.
  23. SKAO — Probing the Cosmic DawnThe goals of future radio studies of neutral gas and reionization.
All articles in this chapter
  1. Gravitational Clumping and Density Fluctuations
  2. Population III Stars: The Universe’s First Generation
  3. Early Mini-Halos and Protogalaxies
  4. Supermassive Black Hole Seeds
  5. Primordial Supernovae: Element Synthesis
  6. Feedback Effects: Radiation and Winds
  7. Merging and Hierarchical Growth
  8. Galaxy Clusters and the Cosmic Web
  9. Active Galactic Nuclei in the Young Universe
  10. Observing the First Billion Years — you are here
Back to blog