Cosmic Distance-Duality Relation

• Cosmic Distance-Duality Relation is a foundational concept linking luminosity distance and angular-diameter distance via a (1+z)² scaling under metric gravity and photon conservation.
• Observed deviations in the duality function η(z) can indicate new physics, photon conservation violations, or systematic issues in astronomical measurements.
• Parameterizations like η₁(z)=1+η₀z and η₂(z)=1+η₀z/(1+z) enable precise tests, with recent methods reaching sub-per-mille constraints on potential discrepancies.

The cosmic distance-duality relation (CDDR), also known as Etherington's reciprocity theorem, is a foundational result in modern observational cosmology. It provides a precise, model-independent link between two distinct cosmological distance measures: the luminosity distance, $d_L(z)$, and the angular-diameter distance, $D_A(z)$, as a function of redshift $z$. The standard form is $d_L(z) = (1+z)^2 D_A(z)$, or equivalently, the duality function $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$ under minimal assumptions. This relation is critical for cosmological inference using sources such as supernovae, galaxy clusters, gravitational lensing systems, baryon acoustic oscillations, and gravitational-wave standard sirens. Confirmation of CDDR validates the metric nature of gravity, null geodesic photon propagation, and photon-number conservation; observational violation would demand new physical mechanisms or emergent systematics.

1. Theoretical Foundation and Mathematical Formulation

The CDDR arises in any metric theory of gravity where photons (or, analogously, gravitational waves) propagate along unique null geodesics and their number is conserved during transit. Etherington's theorem (1933) proves that for a source at redshift $z$, the observed flux and solid angle subtended by the source obey

$d_L(z) = (1+z)^2 D_A(z),$

with the factor $(1+z)^2$ reflecting the combined effects of redshift on photon energy and the cosmological time dilation of arrival rate. The resultant duality function,

$\eta(z) \equiv \frac{(1+z)^2 D_A(z)}{d_L(z)},$

equals unity so long as the underlying assumptions—metric spacetime, unique null geodesics, and photon number conservation—hold exactly.

Any deviation, $\eta(z) \neq 1$, must arise from one of the following: photon–axion or photon–chameleon conversion (breaking photon number conservation), non-metric theories (e.g., emergent gravity, torsion), or environmental effects such as grey dust (opacity) not accounted for in modeling. This provides a robust null test for fundamental physics (Jhingan et al., 2014, Yuan et al., 24 Mar 2026, Melia, 2018).

2. Parameterizations and Phenomenological Extensions

To probe and quantify possible CDDR violations, numerous studies introduce phenomenological parameterizations of $D_A(z)$0. The two most prevalent forms are

$D_A(z)$1

where $D_A(z)$2 is a dimensionless parameter constrained by data: $D_A(z)$3 recovers the standard relation. The first (linear) form diverges at high redshift, while the second saturates, avoiding pathological behavior at large $D_A(z)$4. Other one- and two-parameter generalizations include polynomial, logarithmic ($D_A(z)$5), and power-law structures (Santos-da-Costa et al., 2015, Li et al., 18 Jul 2025, Zhang et al., 22 Jun 2025, Qin et al., 2021).

The table below summarizes common parameterizations and their key properties:

Parameterization	Functional Form	High-$D_A(z)$6 Behavior
Linear	$D_A(z)$7	Diverges ($D_A(z)$8)
Scale-factor (bounded)	$D_A(z)$9	Tends to $z$0
Logarithmic	$z$1	Grows logarithmically
Power law	$z$2	Power-law increase

Empirical fits are typically performed using Markov Chain Monte Carlo (MCMC), Bayesian evidence, or weighted least squares, employing model comparison frameworks to select among competing forms (Silva et al., 2020, Zhang et al., 22 Jun 2025).

3. Observational Strategies and Methodological Advances

Testing the CDDR empirically requires comparing independent measurements of $z$3 and $z$4 at matched or overlapping redshifts:

• Supernovae and Galaxy Clusters: The canonical approach combines Type Ia supernova luminosity distances with angular-diameter distances from galaxy clusters, using Sunyaev–Zel'dovich effect plus X-ray surface brightness (Jhingan et al., 2014, Holanda et al., 2010, Santos-da-Costa et al., 2015).
• Baryon Acoustic Oscillations (BAO) and SNe Ia: BAO provides model-agnostic $z$5 via the sound-horizon scale, while SNe Ia yield $z$6; matching is enforced via interpolation or machine-learning techniques such as neural networks (Li et al., 18 Jul 2025, Zhang et al., 22 Jun 2025, Avila et al., 9 Sep 2025).
• Strong/Lensed Gravitational Lensing and Time Delays: The Einstein radius and time-delay distances in strong lensing systems reconstruct $z$7 independent of flux or standard candles, often combined with supernovae or quasars for $z$8 (Tang et al., 2024, Rana et al., 2017, Ruan et al., 2018, Qin et al., 2021).
• Gravitational-Wave Standard Sirens and Lensing: In the most recent advances, joint observations of strongly lensed gravitational waves (SLGW) allow simultaneous, single-source determination of $z$9 (from lensing observables) and $d_L(z) = (1+z)^2 D_A(z)$0 (from waveform amplitude), entirely free from electromagnetic opacity effects. Precision improves substantially with dual space-based detector networks (e.g., Taiji+LISA), with population-level constraints reaching half-widths as small as $d_L(z) = (1+z)^2 D_A(z)$1 in optimal cases (Yuan et al., 24 Mar 2026, Huang et al., 2024, Liao, 2019).

Non-parametric approaches—such as Gaussian Process regression—enable direct model-independent reconstruction of $d_L(z) = (1+z)^2 D_A(z)$2, avoiding biases from imposed parametric forms (Mukherjee et al., 2021, Avila et al., 9 Sep 2025).

4. Current Constraints and Empirical Status

Multiple datasets—using disjoint methodologies—converge on the remarkable consistency of the CDDR with $d_L(z) = (1+z)^2 D_A(z)$3:

• Supernovae + Galaxy Clusters: Real cluster data analyzed under an elliptical β-model, when paired with Union2 or Pantheon SNe Ia, yield best-fit $d_L(z) = (1+z)^2 D_A(z)$4 consistent with zero at $d_L(z) = (1+z)^2 D_A(z)$5 for all standard parameterizations; mock clusters under simplistic spherical assumptions tend to prefer $d_L(z) = (1+z)^2 D_A(z)$6 at 2–3\sigma, indicating sensitivity to cluster morphology systematics (Jhingan et al., 2014, Holanda et al., 2010).
• BAO + SNe Ia: Combined analysis using SDSS/DESI BAO and PantheonPlus/DESY5 SNe gives $d_L(z) = (1+z)^2 D_A(z)$7 (1\sigma) for all tested forms. Results depend on the treatment of the SN absolute magnitude $d_L(z) = (1+z)^2 D_A(z)$8: when left free or fixed to recent calibrations (e.g., $d_L(z) = (1+z)^2 D_A(z)$9), no CDDR violation is found; fixing to other values can induce up to a $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$0 spurious deviation, underscoring the influence of SN calibration (Li et al., 18 Jul 2025, Zhang et al., 22 Jun 2025).
• Strong Lensing + SNe/Quasars: Model-independent Gaussian Process reconstructions of $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$1 using time-delay lensing and SNe/quasar luminosity distances confirm agreement within 1–2\sigma up to $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$2, with larger uncertainties at high redshift (Tang et al., 2024, Rana et al., 2017, Qin et al., 2021, Ruan et al., 2018).
• Gravitational-Wave Lensing: Simulations of strongly lensed GW events, including state-of-the-art detector sensitivities, project sub-per-mille constraints on $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$3. No statistically significant deviation is found at the $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$4 level, with typical constraints $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$5 in joint Taiji+LISA analysis for $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$6, and $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$7 for $\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$8 (Yuan et al., 24 Mar 2026).

An exception appears at high redshift ($\eta(z) \equiv (1+z)^2 D_A(z)/d_L(z) = 1$9) in recent BAO–SNe analyses, where Lyman-$z$0 BAO results produce $z$1 outliers in certain null tests; these may indicate residual systematics in the BAO measurement or SN extrapolation, rather than genuine CDDR violation (Wang et al., 15 Jun 2025). Multiple-testing corrections and careful systematics control are required to clarify these outliers.

5. Systematics, Model Dependencies, and Limitations

The precision of CDDR tests is limited by systematic uncertainties inherent to the observables:

• Cluster Morphology: The astrophysical modeling of the intracluster medium (elliptical vs. spherical β-models) can bias $z$2 estimates and hence spurious violations of the CDDR (Jhingan et al., 2014, Holanda et al., 2010).
• BAO and Supernova Calibration: The BAO sound horizon scale and SN absolute magnitude $z$3 are degenerate with CDDR parameters. The choice of prior or external calibration on $z$4 and $z$5 can shift inferred $z$6 at the 2–3σ level—this is particularly sensitive to the ongoing Hubble tension between SH0ES and Planck calibrations (Zhang et al., 22 Jun 2025, Li et al., 18 Jul 2025).
• Lensing Profile Assumptions: Uncertainties in the lens mass-density profile (e.g., departures from isothermal spheres, external convergence) in strong lensing propagate directly into inferred $z$7 and hence $z$8 (Rana et al., 2017, Tang et al., 2024, Ruan et al., 2018).
• Opacity and Photon Conservation: Only gravitational-wave and lensing time-delay methods are demonstrably free from electromagnetic opacity effects. Even minor, unrecognized astrophysical dimming (e.g., from grey dust) would manifest as apparent CDDR violation using electromagnetic distance indicators (Liao, 2019).

Non-parametric and simulation-based approaches have improved the robustness of modern tests, but proper error propagation and sample variance at high redshift remain challenging.

6. Prospects for Future Improvements

Ongoing and next-generation surveys (e.g., DESI, LSST, Euclid, Roman Space Telescope, Taiji, LISA) will substantially enhance the precision and redshift reach of CDDR tests:

• Strong Lensing: Forthcoming samples of thousands of lensed quasars will reduce statistical errors and allow hierarchical Bayesian marginalization over lens models, anticipated to shrink $z$9 to the percent level (Tang et al., 2024).
• Space-Based GW Observations: Multi-band GW detector networks can break degeneracies inherent in single-observatory data, achieving $d_L(z) = (1+z)^2 D_A(z),$0 sensitivity to CDDR deviations at $d_L(z) = (1+z)^2 D_A(z),$1 in simulated populations (Yuan et al., 24 Mar 2026, Huang et al., 2024).
• Non-EM Probes: GW standard sirens and time-delay distances provide fully opacity-free, model-independent tests, fundamentally immune to electromagnetic extinction and absorption (Liao, 2019).
• High-Redshift Reach: Incorporating quasars, gamma-ray bursts, and lensed GW events expands coverage to $d_L(z) = (1+z)^2 D_A(z),$2, where few traditional indicators are available (Avila et al., 9 Sep 2025).
• Systematic Mitigation: Rigorous propagation and control of observational, astrophysical, and modeling uncertainties—including machine-learning interpolation, compressed-point methods, and non-parametric reconstructions—are mandatory for interpreting sub-percent deviations as physical.

The table below summarizes projected precision from leading methodologies:

Method	Typical $d_L(z) = (1+z)^2 D_A(z),$3	Redshift Range	Systematic Control
Taiji+LISA GW Lensing	$d_L(z) = (1+z)^2 D_A(z),$4	$d_L(z) = (1+z)^2 D_A(z),$5	Opacity-free, full waveform modeling
BAO+SNe (ANN, GPR)	$d_L(z) = (1+z)^2 D_A(z),$6	$d_L(z) = (1+z)^2 D_A(z),$7	M_B, r_d priors crucial
SNe+Clusters (Ellipt.)	$d_L(z) = (1+z)^2 D_A(z),$8	$d_L(z) = (1+z)^2 D_A(z),$9	Cluster morphology
Time-delay Lensing+SNe	$(1+z)^2$0	$(1+z)^2$1	Lens model, sample size

7. Cosmological Implications and Summary

The cosmic distance-duality relation serves as a cornerstone consistency check for the standard cosmological model. Its persistent validation across independent datasets and methods—spanning supernovae, clusters, BAO, strong lensing, and gravitational waves—places strong empirical limits on possible departures from metric gravity, photon conservation, and standard photon geodesic propagation. No statistically significant CDDR violation has been detected at the $(1+z)^2$2 level over $(1+z)^2$3, except for localized anomalies at high redshift potentially attributable to systematic effects (Yuan et al., 24 Mar 2026, Wang et al., 15 Jun 2025, Li et al., 18 Jul 2025).

The rapid improvement of gravitational-wave-based and high-z lensing techniques promises per-mille-level tests in the near future, with the ultimate objective of probing or excluding subtle new physics—such as dark-sector-induced opacity or fundamental breakdowns of the metric framework. The CDDR remains a pivotal tool in the quest to validate the axioms of cosmological distance inference.