To map the invisible universe, cosmologists rely on weak gravitational lensing. As light from distant galaxies travels across the cosmos, its path is bent by the gravitational pull of intervening dark matter. This bending subtly distorts—or shears—the observed shapes of those background galaxies. By measuring how the shapes of millions of galaxies are correlated across the sky, researchers can reconstruct the distribution of dark matter and measure the dark energy driving cosmic expansion, as reviewed by R. G. Bower in Surveying the dark side.
However, this technique hinges on a crucial assumption: before their light is deflected, the orientations of galaxies are entirely random. If their initial orientations are not random, any inherent correlation in their shapes will masquerade as a gravitational lensing signal.
This is the core challenge of intrinsic galaxy alignment (IA). Galaxies are not dropped randomly into space; they are born and evolve within shared, large-scale gravitational environments that physically force them into alignment.
The Physical Mechanism: How the Cosmic Web Orients Galaxies
Why do galaxies align themselves? The answer lies in the gravitational tidal fields of the cosmic web.
When a galaxy forms inside a massive dark matter halo or along a cosmic filament, it experiences differential gravitational forces across its structure. For elliptical galaxies, these tidal forces stretch the stellar distribution, physically orienting the galaxy's major axis toward local matter concentrations. For spiral galaxies, tidal forces exert a torque on the infalling gas, dictating the angular momentum and the resulting tilt of the galactic disk.
Without Intrinsic Alignment:
[ Galaxy 1: / ] --- (Random initial orientations) --- [ Galaxy 2: \ ]
↓ Light passes through dark matter halo ↓
[ Observed: | ] ====================================== [ Observed: | ] -> Coherent alignment = True Lensing
With Intrinsic Alignment:
[ Galaxy 1: | ] --- (Physically aligned by local filament) --- [ Galaxy 2: | ]
↓ No intervening lensing occurs ↓
[ Observed: | ] ====================================== [ Observed: | ] -> Coherent alignment = False Lensing
Hydrodynamic simulations and observational surveys confirm that this alignment is a large-scale structural phenomenon. Yen‐Chi Chen et al. demonstrated in Investigating galaxy-filament alignments in hydrodynamic simulations using density ridges that galaxies tend to align their major axes parallel to nearby cosmic filaments.
In a subsequent observational study using the Sloan Digital Sky Survey, Yen‐Chi Chen et al. showed in Detecting galaxy–filament alignments in the Sloan Digital Sky Survey III that this parallel alignment is statistically significant and extends across massive distances of up to 30–40 Mpc. This directional skeleton of the universe ensures that physically associated pairs or groups of galaxies exhibit highly correlated shapes prior to any lensing effects, as detailed by Sandrine Codis et al. in Statistical exploration of halo anisotropic clustering and intrinsic alignments with the mass-Peak Patch algorithm.
The Systematic-Cosmological Degeneracy
Because intrinsic alignment creates coherent shape correlations that mimic the geometric distortions of cosmic shear, it introduces a severe systematic bias into cosmological measurements. If a survey observes two galaxies with correlated ellipticities, it cannot immediately distinguish whether that correlation was caused by a foreground dark matter lens or by the local tidal field acting on both galaxies, a problem highlighted early on by Rachel Mandelbaum et al. in Ellipticity of dark matter haloes with galaxy-galaxy weak lensing.
This creates a dangerous degeneracy between astrophysical nuisance parameters and fundamental physics:
- Masking New Physics: The DES Collaboration et al. showed in Dark Energy Survey Year 3 results: Constraints on extensions to $\Lambda$CDM with weak lensing and galaxy clustering that potential signals of modified gravity or extensions to the standard cosmological model ($\Lambda$CDM) are often statistically indistinguishable from intrinsic alignment modeling choices.
- Biasing Parameter Fits: If researchers rely on physically unrealistic alignment parameters to force a model to fit the data, it can obscure underlying systematic errors within a survey. G. Efstathiou demonstrated in Statistical inconsistencies in the KiDS-450 data set that unconstrained or poorly motivated intrinsic alignment models frequently lead to biased cosmological parameters.
Mathematical Frameworks: From NLA to TATT
To prevent these biases, cosmologists incorporate intrinsic alignment directly into their theoretical models. Instead of treating galaxy shapes as purely random, they parameterize how shapes respond to the surrounding matter density field.
The simplest widely used framework is the Linear Alignment (LA) model, and its small-scale extension, the Non-Linear Alignment (NLA) model. The NLA model assumes that a galaxy's intrinsic ellipticity is linearly proportional to the local tidal gravitational potential at the time of the galaxy's formation.
However, as survey precision has increased, the linear assumption has proven insufficient. It fails to capture higher-order physical processes like tidal torquing, which governs the shapes of disk-dominated spiral galaxies. To address this, M. Gatti et al. noted in Dark Energy Survey Year 3 results: cosmology with moments of weak lensing mass maps that high-precision surveys must transition to more sophisticated frameworks.
The current state-of-the-art framework is the Tidal Alignment and Tidal Torque (TATT) model. TATT is a five-parameter model that combines both linear alignment (quadratic in the density field for ellipticals) and tidal torquing (which scales differently and affects spirals). By mapping these distinct physical dependencies, the authors of The Dark Energy Survey Year 3 and eBOSS: constraining galaxy intrinsic alignments across luminosity and colour space (S. Samuroff et al.) successfully isolated alignment amplitudes across varying galaxy luminosities and colors. This comprehensive parameterization ensures that astrophysical shape noise is cleanly separated from the underlying cosmic shear signal, as demonstrated by Bhuvnesh Jain et al. in Dark energy survey year 3 results: High-precision measurement and modeling of galaxy-galaxy lensing.
Strategies for Disambiguation
How do we break the degeneracy completely? Modern cosmology employs three main strategies to isolate the purely geometric lensing signal from local astrophysical contamination.
1. Cross-Correlation with the CMB
The most robust way to eliminate intrinsic alignment is to cross-correlate galaxy shapes with the lensing of the Cosmic Microwave Background (CMB). CMB photons originate from the early universe ($z \approx 1100$) and are lensed by foreground structures on their way to our telescopes. Because the CMB source plane is physically separated from foreground galaxies by billions of light-years, CMB photons cannot be physically aligned by the local tidal fields shaping those galaxies.
As pioneered by Christopher M. Hirata et al. in Correlation of CMB with large-scale structure. II. Weak lensing, any correlated signal between CMB lensing maps and galaxy shapes isolates the purely geometric lensing effect. This cross-survey approach was recently utilized by C. Chang et al. in Joint analysis of Dark Energy Survey Year 3 data and CMB lensing from SPT and Planck. II. Cross-correlation measurements and cosmological constraints to cleanly disentangle local tidal physics from large-scale gravitational distortions.
2. Lensing Shear Ratios
Another geometric approach exploits the redshift evolution of lensing. By measuring the ratio of galaxy-galaxy lensing signals from a shared lens population across multiple source galaxy bins at different distances, researchers can isolate purely geometric distance ratios. Bhuvnesh Jain et al. demonstrated in Dark Energy Survey Year 3 results: Exploiting small-scale information with lensing shear ratios that these "shear ratios" are highly sensitive to geometry but insensitive to the underlying matter power spectrum, providing a clean method to self-calibrate and constrain intrinsic alignment parameters within the data.
3. Non-Gaussian Map Statistics
Standard weak lensing analyses rely on two-point correlation functions, which average out spatial patterns. By moving toward map-based, non-Gaussian statistics—such as mass map moments, peak counts, and Minkowski functionals—researchers can extract higher-order structural information from the cosmic web.
As shown by D. Zürcher et al. in Cosmological forecast for non-Gaussian statistics in large-scale weak lensing surveys and Towards a full wCDM map-based analysis for weak lensing surveys, these advanced statistical methods break parameter degeneracies, providing tighter, more robust cosmological constraints that are significantly less vulnerable to intrinsic alignment contamination.
Key Foundational Work
Surveying the dark side (2006) — contributed 1 concept elements to this direction.
Correlation of CMB with large-scale structure. II. Weak lensing (2008) — contributed 1 concept elements to this direction.
Ellipticity of dark matter haloes with galaxy-galaxy weak lensing (2006) — contributed 1 concept elements to this direction.