Intrinsic Alignments

Galaxies are not scattered randomly across the cosmos; their very shapes and orientations are sculpted by the invisible architecture of the universe, a phenomenon known as intrinsic alignment. This subtle coherence poses a significant challenge to modern cosmology, particularly for precision measurements like weak gravitational lensing, which assumes random galaxy orientations to map dark matter. Failure to account for these alignments can corrupt our understanding of the universe's fundamental properties.

This article delves into the fascinating world of intrinsic alignments. First, in "Principles and Mechanisms," we will explore the physics of how tidal forces and torques from the cosmic web shape and orient galaxies. Following this, "Applications and Interdisciplinary Connections" will examine the dual nature of intrinsic alignments—as a critical contaminant that must be mitigated and as a powerful, independent probe for testing gravity and uncovering secrets of the early universe.

Imagine you are out in deep space, far from any planet or star, holding a perfectly spherical ball of soft dough. In this lonely void, the dough remains perfectly spherical. Now, let's conjure up two massive objects, say two large planets, and place them on opposite sides of your dough ball, far away but aligned with it. What happens? The gravitational pull from each planet tries to tug the dough towards it. The side of the dough ball closer to planet A is pulled more strongly than its center, and the center is pulled more strongly than the far side. The same thing happens with planet B. The net effect is not to move the dough ball (since the forces are balanced), but to stretch it. Your once-perfectly spherical dough will be pulled into an elongated shape, an ellipsoid.

This stretching force is a ​​tidal force​​. It's the very same force that causes the ocean tides on Earth, exerted by the Moon and the Sun. What is remarkable is that this simple principle is the primary sculptor of galaxies on the largest scales. Galaxies are not isolated islands; they are embedded in a vast, intricate network of matter known as the ​​cosmic web​​. Denser regions of this web pull on the gas, stars, and dark matter that form a galaxy, tidally stretching and shaping it. This phenomenon, where a galaxy's orientation or shape is not random but is correlated with the surrounding large-scale structure, is what we call ​​intrinsic alignment​​.

To be a bit more precise, like any good physicist, we can describe this tidal effect mathematically. The stretching is governed by the second derivatives of the gravitational potential, $\Phi$. These derivatives form a mathematical object called the ​​tidal tensor​​, $\frac{\partial^2 \Phi}{\partial x_i \partial x_j}$. This tensor tells you, at any point in space, the direction and magnitude of the tidal "squeeze" and "stretch".

The simplest and most powerful idea to explain the alignment of massive, elliptical galaxies is the ​​Linear Alignment Model (LAM)​​. It proposes a beautifully simple relationship: the shape of a galaxy (or, more accurately, the dark matter halo in which it lives) is directly proportional to the local tidal tensor. A stronger tidal field means a more elongated galaxy. The galaxy's shape, which we can describe with a quantity called the quadrupole moment $s_{ij}$, is simply a copy of the tidal field at its location:

This model has been astonishingly successful. It tells us that by looking at the shapes of elliptical galaxies, we are indirectly looking at a map of the cosmic tidal fields. The galaxies themselves become tracers of the invisible gravitational scaffolding of the universe.

But nature, as always, has more than one trick up her sleeve. While massive elliptical galaxies are "squeezed" into alignment by the present-day tidal field, spiral galaxies tell a different story. Their shapes are dominated by their spin, giving them their characteristic disc structure. Their alignment is not about being stretched, but about being spun up. During their formation, the lumpy, irregular distribution of matter around a protogalaxy exerts gravitational tugs that are not perfectly balanced. This imbalance creates a twisting force, or a ​​tidal torque​​. This torque spins the protogalaxy, and its final axis of rotation (and thus the orientation of the galactic disk) retains a memory of these primordial torques. This process is more complex and dynamic, often described by ​​non-linear alignment models​​ that involve quadratic terms of the density field, reflecting the interaction between different parts of the collapsing structure.

So, we have two primary tales of alignment: the slow, persistent squeezing of giant ellipticals by the ambient tidal field, and the violent, formative spinning of spirals by primordial torques.

Now, you might be thinking this is a fascinating piece of astrophysics, a neat way to understand galaxy formation. And it is! But it also presents a formidable challenge for one of the most powerful tools in cosmology: ​​gravitational lensing​​.

Gravitational lensing is the bending of light by mass, as predicted by Einstein's theory of general relativity. When we look at distant galaxies, their light has traveled across billions of light-years and passed through the cosmic web. The gravity of this intervening matter distorts the path of the light, causing the images of the background galaxies to appear slightly stretched or sheared. This ​​cosmic shear​​ is an incredibly valuable signal, as it allows us to map the distribution of all matter—including dark matter—between us and the source galaxies.

The measurement is conceptually simple. We measure the observed ellipticity, $\epsilon^{\mathrm{obs}}$, of millions of distant galaxies. This observed shape is, to a very good approximation, the sum of the true gravitational shear, $\gamma$, and the galaxy's own ​​intrinsic ellipticity​​, $\epsilon^{\mathrm{int}}$:

If the intrinsic shapes of galaxies were completely random, we could simply average the ellipticities of many galaxies in a patch of sky. The random intrinsic shapes would average out to zero, $\langle \epsilon^{\mathrm{int}} \rangle = 0$, leaving us with the pure gravitational shear signal, $\langle \epsilon^{\mathrm{obs}} \rangle \approx \gamma$.

But as we've just learned, their shapes are not random. They are intrinsically aligned by the very same large-scale structure that is causing the gravitational lensing! This creates a pernicious form of contamination. The intrinsic alignment signal can mimic, or masquerade as, a gravitational lensing signal.

Imagine a simple but extreme scenario where a galaxy's intrinsic ellipticity is directly proportional to the local tidal shear, say $\epsilon^{\mathrm{int}} = C_I \gamma$, where $C_I$ is some constant. The observed ellipticity would then be a more complicated function than just the sum, systematically boosting (or reducing) the true lensing effect. The alignment effectively adds "fake" shear, corrupting our maps of the dark matter.

This contamination appears in our statistical measurements in two main forms:

1. ​​Intrinsic-Intrinsic (II) Correlation​​: Consider two galaxies that are physically close to each other in the cosmic web. They are both being shaped by the same local tidal field. Therefore, their intrinsic shapes will be correlated. If one is stretched horizontally, the other is likely to be stretched horizontally too. When we correlate the shapes of galaxy pairs in our analysis, this $\langle \epsilon^{\mathrm{int}}_{i} \epsilon^{\mathrm{int}}_{j} \rangle$ correlation adds a spurious signal that has nothing to do with lensing.

2. ​​Gravitational-Intrinsic (GI) Correlation​​: This one is more subtle. Consider a foreground galaxy and a distant background galaxy that lies behind it along the line of sight. The foreground galaxy's shape, $\epsilon^{\mathrm{int}}$, is aligned by the mass concentration it lives in. That same mass concentration acts as a gravitational lens, shearing the light from the background galaxy to produce a shear $\gamma$. Thus, the intrinsic shape of the foreground galaxy is correlated with the gravitational shearing of the background galaxy. This $\langle \gamma_i \epsilon^{\mathrm{int}}_j \rangle$ correlation is a cross-contamination between the lensing signal and the alignment signal.

Failing to account for these II and GI correlations can lead us to infer the wrong amount of dark matter or misunderstand its properties, which could completely derail our quest to understand the nature of dark energy and the ultimate fate of the universe.

So, how do we fight back against this contamination? We can't just throw away the data. The solution is to understand the enemy—to model the intrinsic alignments so well that we can cleanly separate them from the cosmic shear signal. And here, the physics of alignment gives us the very tools we need.

Since the Linear Alignment Model connects galaxy shapes to the underlying matter distribution, it predicts that the statistical properties of intrinsic alignments should be directly related to the statistical properties of matter itself. Cosmologists characterize the "lumpiness" of the matter distribution using the ​​matter power spectrum​​, $P_m(k)$, which tells us how much structure exists on different physical scales (represented by the wavenumber $k$). The LAM predicts that the power spectrum of the intrinsic alignments, $P_{II}(k)$, should be directly proportional to this matter power spectrum.

This is a powerful starting point. But the real world provides an even more distinctive signature. When we observe a galaxy survey, we don't measure true 3D positions; we measure positions on the sky (angles) and redshifts. A galaxy's redshift is primarily due to the expansion of the universe, but it also has a small contribution from its peculiar velocity—its motion through space. This is the famous ​​Doppler effect​​. A galaxy moving towards us will have its redshift slightly decreased (a blueshift), making it appear closer than it is. A galaxy moving away will have its redshift increased, making it appear farther. This effect, known as ​​Redshift-Space Distortions (RSD)​​, squashes structures along our line of sight.

The beauty is that this distortion depends on the angle between our line of sight and the direction of the galaxy pair we are looking at. The amount of squashing changes as we look along the line of sight versus perpendicular to it. It turns out that this same angular dependence gets imprinted on the intrinsic alignment signal. Because peculiar velocities are driven by gravity, and alignments are also driven by gravity, the two phenomena are intertwined. The cross-correlation between the positions of galaxies (which are affected by RSD) and the shapes of galaxies (the GI effect) has a unique and predictable dependence on $\mu$, the cosine of the angle to the line of sight. The exact form of this angular dependence, captured in expressions like the famous Kaiser formula term $(b_g + f\mu^2)$, provides a "smoking gun" signature that allows us to isolate and measure the strength of the GI contamination directly from the data.

Furthermore, we can use clever survey strategies. Since the II correlation only exists for galaxies that are physically close, we can mitigate it by avoiding correlating the shapes of galaxies that are at similar redshifts. By correlating galaxies in different tomographic "slices" of the universe that are well-separated, we ensure that the II term vanishes, leaving a cleaner signal to analyze.

The journey, of course, does not end here. The Linear Alignment Model, while powerful, is an approximation. The real universe is a messy, non-linear place. As mentioned, the physics of spiral galaxy alignment involves complex tidal torques. And even for ellipticals, on smaller scales, deep inside the chaotic environment of a galaxy cluster, the simple linear picture breaks down.

This is where the frontier of research lies. Cosmologists are developing more sophisticated non-linear and quadratic models to capture these richer physics. Moreover, they are turning to powerful theoretical frameworks like the ​​Effective Field Theory of Large-Scale Structure (EFT of LSS)​​. The EFT provides a rigorous way to handle our ignorance of the complex, "messy" physics happening on small scales. It tells us precisely which new parameters, or "counterterms," we need to add to our model to account for these unknown effects in a systematic way, ensuring our large-scale predictions remain robust.

Intrinsic alignments, once seen as just a frustrating contaminant, have thus transformed into a rich field of study in their own right. They are a direct probe of the tidal environment of the cosmic web and a window into the intricate physics of galaxy formation. By understanding and modeling this "contaminant," we not only clean our lensing maps to reveal the secrets of dark matter and dark energy, but we also learn a profound lesson about the deep and beautiful unity of the cosmos—where the grandest structures in the universe leave their indelible signature on the delicate shapes of the galaxies within them.

Now that we have explored the physical principles that cause galaxies to align with the cosmic web, a natural and pressing question arises: so what? Why should we care about this subtle, almost imperceptible coherence in the shapes of distant galaxies? The answer, as is so often the case in physics, is a story of a nuisance transformed into an opportunity. Intrinsic alignments (IA) represent one of the most significant challenges in modern precision cosmology, a frustrating contaminant that must be meticulously accounted for. Yet, by grappling with this challenge, we have discovered that these very alignments provide a powerful and entirely new kind of compass, one that can point us toward answers to some of the most profound questions in fundamental physics. Intrinsic alignments are both a curse and a blessing, a fly in the ointment and the key to a new recipe.

The primary reason intrinsic alignments rose from a niche curiosity to a central topic in cosmology is their ability to mimic and corrupt the signal of weak gravitational lensing. Weak lensing, or "cosmic shear," is one of our most powerful tools for mapping the distribution of dark matter. The logic is simple and elegant: as light from distant galaxies travels towards us, its path is bent by the gravity of all the matter it passes. This bending distorts, or "shears," the observed images of the background galaxies. By measuring the tiny, coherent distortions across millions of galaxies, we can reconstruct the invisible scaffolding of dark matter that holds the universe together. The entire technique relies on the assumption that, in the absence of lensing, the orientations of galaxies are random.

But as we have seen, they are not. Intrinsic alignments introduce a coherence that has nothing to do with the lensing of light from background sources. Worse, because the same gravitational fields that cause lensing also cause intrinsic alignments, the two signals are often correlated. Imagine a large cluster of galaxies. Its immense gravity will stretch nearby, physically associated galaxies into alignment with its tidal field (an intrinsic alignment). At the same time, that same gravitational pull will shear the images of much more distant galaxies located behind the cluster (a weak lensing effect). An observer sees two correlated effects: the foreground galaxies are aligned, and the background galaxies are sheared. The trouble begins when the foreground galaxies are mistaken for background ones, or when the statistical correlation between the two is not properly accounted for. This "Gravitational-Intrinsic" (GI) correlation is a direct physical link between the two phenomena and a major source of systematic error in our analysis of the cosmic web.

What happens if we are blissfully unaware of this contamination and interpret all the coherent alignment we see as a pure weak lensing signal? The consequences are severe. We would systematically miscalculate the masses of galaxy clusters and the density of the large-scale structure between us and the source galaxies. For instance, if intrinsic alignments tend to radially align galaxies with an overdensity, they produce a tangential shear pattern that opposes the gravitational lensing shear. Ignoring this effect would lead us to underestimate the lensing signal, and therefore underestimate the mass of the lensing structure. Since measurements of cluster masses are used to anchor our estimates of the total matter density of the universe, $\Omega_m$, this seemingly small effect can lead to a significant bias in our inferred value for one of the most fundamental parameters of our cosmos.

As our astronomical surveys, such as the Vera C. Rubin Observatory and the Euclid space telescope, reach unprecedented levels of precision, the problem gets deeper. It's not enough to account for the average contamination; we must also understand how it affects the uncertainty in our measurements. The statistical properties of our data are not as simple as a bell curve. Non-Gaussian features, captured by higher-order statistics like the trispectrum, become important. These features tell us about the intricate, web-like nature of cosmic structure. Intrinsic alignments contribute significantly to this non-Gaussianity, creating complex correlations in the data that, if ignored, can trick us into thinking our measurements are far more precise than they actually are. In short, IA not only biases our results, but it can also corrupt our confidence in them.

The reach of this contamination even extends across disciplinary boundaries within cosmology. A powerful method for analyzing spin-2 fields like cosmic shear is to decompose them into "E-modes" and "B-modes," a technique borrowed from the study of the cosmic microwave background (CMB) polarization. Gravitational lensing produces almost pure E-modes. A detection of a primordial B-mode signal on cosmic scales would be a monumental discovery, likely pointing to gravitational waves from the inflationary epoch of the early universe. However, the picture is complicated by intrinsic alignments. While the primordial IA field is expected to be a pure E-mode, the gravitational lensing of this E-mode field by foreground structures can twist it, generating a B-mode signal. Therefore, intrinsic alignments create a foreground B-mode signal that could contaminate and obscure the search for one of the most sought-after signals in all of science.

Faced with such a pervasive contaminant, the first task for cosmologists was to characterize and model it in order to remove it. And this effort has led to a wonderful reversal: by learning how to measure intrinsic alignments, we've learned how to use them as a cosmological probe in their own right. If we have a sample of galaxies for which we know the distances, we can measure the alignments directly by correlating the observed shapes of galaxies with their positions and the positions of other nearby galaxies. We can even study their higher-order statistics, like the bispectrum, to gain a more detailed understanding of how galaxies respond to their environment. This turns the problem on its head. Instead of being unwanted noise, the alignment signal becomes a rich source of information.

This newfound signal is remarkably sensitive to the underlying physics governing the universe. Because galaxy shapes are coupled directly to the local tidal field—a direct derivative of the gravitational potential—they are a more immediate probe of the gravitational environment than the integrated, line-of-sight effect of weak lensing. This opens a new window onto fundamental physics.

A New Window on Fundamental Physics

​​Weighing the Ghost Particle:​​ Neutrinos are the most elusive of the known fundamental particles. We know they have mass, but we don't know how much. Because they are so light and fast-moving, they resist clumping under gravity, spreading out across the cosmos as a smooth background. This means that on small scales, the "clustered" matter that forms galaxies (cold dark matter and baryons) is a slightly larger fraction of the total than on large scales. This provides a subtle way to weigh neutrinos. Weak lensing is sensitive to all matter, including the smooth neutrino component. Intrinsic alignments, however, are thought to be caused by the tidal fields of the clustered matter in which galaxies form. By comparing the IA signal to the weak lensing signal, we can sense the difference between the clustered matter and the total matter, a difference which depends directly on the total mass of the neutrinos. The alignments of galaxies, therefore, act as a scale for one of the lightest particles in the universe.

​​Echoes of the Big Bang:​​ Our leading theory of the universe's first moments, cosmic inflation, predicts that the initial density fluctuations were almost perfectly Gaussian. Detecting any deviation from this, known as primordial non-Gaussianity (PNG), would provide an invaluable clue about the specific physics at play in the primordial soup. Such primordial non-Gaussianity, often parameterized by a value $f_{\text{NL}}$, would leave a faint but characteristic imprint on the large-scale structure. It turns out that intrinsic alignments are a particularly sharp tool for finding this imprint. The presence of PNG introduces a unique, scale-dependent signature into how galaxies cluster and how they align. By measuring how the strength of galaxy alignments changes across vast cosmic scales, we can search for this signature and thereby test fundamental theories of the universe's birth.

​​Testing Einstein's Gravity:​​ On the largest scales, is General Relativity the complete story of gravity? Some theories propose that gravity behaves differently over cosmological distances to explain the accelerating expansion of the universe without dark energy. These "modified gravity" theories often change the relationship between the matter distribution and the gravitational potential it generates. Since intrinsic alignments are a direct response to the tidal field (and thus the potential), they are an exceptionally clean probe of this relationship. Different theories of gravity predict different strengths of alignment for a given amount of matter. By measuring the intrinsic alignment power spectrum, we can perform a direct test of General Relativity on scales where it has never been so precisely tested, potentially uncovering new gravitational physics.

In the end, the story of intrinsic alignments is a perfect microcosm of the scientific process. What began as a vexing source of error, a contaminant that threatened to undermine one of our primary cosmological tools, has been transformed through careful study into a tool of its own. It is a testament to the idea that in the universe, there is no such thing as mere "noise." There is only signal we do not yet understand. By learning to listen to the subtle whispers of aligned galaxies, we have not only sharpened our view of the dark matter universe but have also opened a new channel to hear the echoes of the Big Bang, feel the weight of the neutrino, and test the very foundations of gravity itself.