Linear cosmological density perturbation

For pressureless subhorizon matter perturbations, a Fourier mode in conformal time obeyswhere $\mathcal{H}=a^{\prime }/a$.

These equations describe the small-amplitude phase of structure formation; nonlinear growth requires a more general treatment.

For pressure-dominated linear cosmological density perturbations, ${\ddot{\delta }}_k+2H{\dot{\delta }}_k+{\omega }^2{\delta }_k=0$. Writing ${\delta }_k=\chi /a$ gives $\ddot{\chi }+({\omega }^2-\ddot{a}/a)\chi =0$. In the WKB approximation, the oscillation amplitude is $1/(a\sqrt{\omega })$. For constant sound speed it is proportional to $a^{-1/2}$; the local frozen-frequency amplitude damping time $H^{-1}$ does not imply a universal global envelope $a^{-1}$.

Compare a flat background Friedmann equation with a nearly homogeneous curved region on a slicing of equal local Hubble parameter. To first order in its curvature parameter $K$, subtraction gives $\delta \rho =3K/(8\pi G{a}^2)$, or $\delta =K/(a^2{H}^2)$. Conserved curvature therefore gives $\delta \propto a^2$ in radiation domination and $\delta \propto a$ in matter domination. This construction specifies the slicing: two patches at equal proper time generally do not have the same Hubble parameter. For growing dust perturbations, the continuity equation gives $H_s-H=-\dot{\delta }/3$ to first order. These different time comparisons can share growth powers without sharing the same curvature-to-density coefficient.

For $a\propto t^{1/2}$ and $\rho =3/(32\pi G{t}^2)$, the specified equation $\ddot{\Delta }+2H\dot{\Delta }=(32\pi G\rho /3)\Delta$ reduces to $\ddot{\Delta }+t^{-1}\dot{\Delta }=t^{-2}\Delta$. Its two modes are $t$ and $t^{-1}$. This is a property of that equation; pressure-gradient and gauge choices must be specified before extending it to general radiation perturbations.

When the self-gravity source in the pressureless linear cosmological density perturbation equation is negligible, $\ddot{\delta }+2(\dot{a}/a)\dot{\delta }=0$. Multiplication by $a^2$ makes it $d(a^2\dot{\delta })/dt=0$, so $\delta =C_1+C_2\int dt/a^2(t)$. This explains how Hubble friction changes growth for an arbitrary positive scale factor. For $a\propto t^p$, the nonconstant mode is $t^{1-2p}$ when $p\ne 1/2$ and $\log t$ when $p=1/2$.

In a spatially flat universe containing separately conserved pressureless matter and a coasting fluid with $w=-1/3$, the ratio $\eta ={\rho }_s/{\rho }_m$ is proportional to the scale factor. If the coasting component is modeled by the barotropic fluid perturbation equations with no anisotropic stress, its contribution to the trace gravitational source vanishes because $1+3w=0$. The matter growing mode isA decaying solution is $D_{-}=\sqrt{1+\eta }/{\eta }^{3/2}$. The reduction of order formula with integrating factor ${\eta }^{3/2}\sqrt{1+\eta }$ produces the integral solution. It has $D=\eta [1-4\eta /7+O({\eta }^2)]$ during matter domination and tends to $5/2$ during coasting-fluid domination. Thus growth freezes rather than continuing in proportion to the scale factor. The background geometry is flat despite the same background expansion law as an open dust universe.

In a flat matter-dominated universe with $a\propto t^{2/3}$ and $P\propto {\rho }^{4/3}$, the squared sound speed scales as $t^{-2/3}$. Each comoving Fourier mode then satisfies an Euler-Cauchy equation, with gravitational growth at long wavelengths and logarithmic-time acoustic oscillations at short wavelengths.

For a flat Newtonian expanding background $R\propto t^{2/3}$ with polytropic equation of state $p=K{\rho }^{\gamma }$, a comoving density contrast satisfies $\ddot{\delta }+4\dot{\delta }/(3t)+[{\Lambda }^2{t}^{-2\nu -2}-2/(3t^2)]\delta =0$. Here $\Lambda =kc(t_0)t_0^{\gamma -1/3}$ and $\nu =\gamma -4/3$. For $\nu >0$, substitution $z=\Lambda /(\nu t^{\nu })$ and $\delta =t^{-1/6}y(z)$ gives the Bessel equation of order $\lambda$. A universally valid basis uses the Bessel function of the first kind and Bessel function of the second kind, $t^{-1/6}{J}_{\lambda }(z)$ and $t^{-1/6}{Y}_{\lambda }(z)$. When $\lambda$ is not an integer, $J_{-\lambda }$ can replace $Y_{\lambda }$. For $4/3<\gamma <5/3$, the early oscillatory envelope is $t^{(3\gamma -5)/6}$; late modes have leading powers $t^{2/3}$ and $t^{-1}$. Late growth requires a nonzero coefficient of the growing branch.

Put $\delta =t^{2/3}h(t)$ in the Bessel density modes of a polytropic expanding universe equation. It becomes $h^{\prime \prime }+8h^{\prime }/(3t)+{\Lambda }^2{t}^{-2\nu -2}h=0$. Substituting $h=1+a{\Lambda }^2{t}^{-2\nu }+\cdots$ gives $a=[2\nu (5/3-2\nu ){]}^{-1}$. Since $ck/R=\Lambda t^{-\nu -1}$, the fractional pressure correction is small when $(ck/R{)}^2{t}^2\ll 2\nu (5/3-2\nu )$. For $0<\nu <1/3$, this is the order-of-magnitude condition $ck/R\ll \sqrt{6\pi \nu G\rho }$. It describes approach to the dust-like growing branch, not an exact universal time at which every initial perturbation starts increasing. The instantaneous Jeans instability balance instead compares $(ck/R{)}^2$ with $4\pi G\rho$.

For a nonrelativistic polytropic equation of state $P=K{\rho }^{4/3}$ during matter domination, $c_s^2\propto a^{-1}$ and the comoving Jeans wavenumber is constant. The leading density equation becomes $\ddot{\delta }+4\dot{\delta }/(3t)+[2(q-1)/(3t^2)]\delta =0$, giving the displayed powers. A growing mode exists for $q<1$; at $q=1$ one mode is constant. At $q=25/24$ the repeated root has the second solution $t^{-1/6}\ln t$. For $q>25/24$ both real solutions oscillate in $\ln t$ with envelope $t^{-1/6}$. Thus the instability threshold and the onset of oscillatory solutions are distinct in an expanding background.

When the characteristic exponents of the density-mode Euler-Cauchy equation coincide, the second independent solution gains a logarithm. In the polytropic matter-era example the complete solution at this threshold is $t^{-1/6}(A+B\log t)$.

The peculiar gravitational potential is the Newtonian potential after subtracting the potential of the homogeneous expanding background. In comoving coordinates it obeys a Poisson equation sourced by the density contrast, so its gradient accelerates matter relative to the Hubble flow.

The density contrast is the fractional perturbation $\delta =(\rho -\bar{\rho })/\bar{\rho }$.

For a perfect fluid with $P=w\rho$, in Newtonian gauge in cosmology the density perturbation on fluid-comoving slices is the displayed combination of the fractional density perturbation and velocity potential. The Einstein energy and momentum constraints combine into ${\nabla }^2\varphi =4\pi G{a}^2\bar{\rho }\Delta$ when scalar anisotropic stress vanishes. This gives a relativistic Poisson constraint valid beyond the subhorizon limit.

During matter domination in a flat universe, the growing comoving-gauge density contrast is proportional to $a\propto {\eta }^2$. Its Newtonian potential is constant by the Poisson constraint. The momentum constraint gives $v=-2\varphi /(3\mathcal{H})=-\eta \varphi /3$. Thus the peculiar velocity potential grows as $a^{1/2}$ in the convention where the peculiar velocity is $\nabla v$.

For a flat single-fluid background with constant barotropic equation of state, $w\ne -1/3$, adiabatic perturbations and zero anisotropic stress, $a\propto {\eta }^{2/(1+3w)}$ and $a^2\bar{\rho }\propto {\eta }^{-2}$. The scalar potential obeys ${\varphi }^{\prime \prime }+[6(1+w)/(1+3w)]{\varphi }^{\prime }/\eta -w{\nabla }^2\varphi =0$. Substituting the Poisson constraint $\Delta \propto {\eta }^2{\nabla }^2\varphi$ gives the displayed equation. The homogeneous mode and the degenerate vacuum fluid $w=-1$ require separate interpretation.

For radiation in cosmology, the comoving-gauge density contrast satisfies ${\Delta }_k^{\prime \prime }+(k^2/3-2/{\eta }^2){\Delta }_k=0$. The displayed basis solves it. The first mode is regular and grows as $A{x}^2/3$ outside the sound horizon; the second diverges as $B/x$ toward the initial singularity. At large $x$, the regular mode approaches $-A\cos x$ and has constant-amplitude acoustic oscillations.

In matter domination with the constant growing potential, the full Newtonian-gauge energy constraint gives the stated density contrast. The second term dominates inside the Hubble radius and grows as $a$, because ${\mathcal{H}}^2\propto a^{-1}$. Outside the Hubble radius, the constant $-2\Phi$ term dominates and is gauge dependent. The comoving density combination ${\Delta }_c={\delta }_c+3\mathcal{H}{\theta }_c/k^2$ cancels it for ${\theta }_c=2k^2\Phi /(3\mathcal{H})$, leaving the cosmological Poisson equation relation. Matter-spectrum power laws must specify whether they refer to this comoving density or to subhorizon Newtonian-gauge density.

The smoothed matter density variance measures the root-mean-square linear density contrast after applying a window corresponding to a Lagrangian mass scale. For a homogeneous isotropic field it is ${\sigma }^2=(1/(2{\pi }^2))\int k^{2}P(k)|W(kR){|}^{2}dk$. The variance convention, growth epoch and smoothing mass-density relation must agree with the collapse barrier used in a halo abundance calculation.

A spherical top-hat window function averages a field uniformly inside a sphere of radius $R$ and gives zero weight outside it. Its normalized real-space kernel is $W_R(\mathbf{x})=3/(4\pi R^3)$ for $|\mathbf{x}|\le R$, and its Fourier transform is

The linear growth factor multiplies the growing mode of a linear cosmological density perturbation, so $\delta (\mathbf{x},t)=D(t)\delta (\mathbf{x},t_0)/D(t_0)$ when scale-independent growth applies.

For scale-independent linear growth $\delta (\mathbf{k},a)=D(a){\delta }_{\mathrm{ini}}(\mathbf{k})$, the logarithmic linear growth rate measures fractional amplification per logarithmic change of the scale factor. In an Einstein-de Sitter universe, $D\propto a$ and $f=1$. The cosmological Poisson equation for conserved pressureless matter gives $\psi \propto D/a$ at fixed comoving coordinate, and therefore ${\psi }^{\prime }=\mathcal{H}(f-1)\psi$. This relates suppression of matter growth to the Integrated Sachs-Wolfe effect.

For pressureless matter in a Friedmann-Lemaître-Robertson-Walker metric with negative spatial curvature of an FLRW universe and no cosmological constant, set $x=(1-{\Omega }_{m0})a/{\Omega }_{m0}>0$. A decaying linear growth factor is $D_{-}(x)=\sqrt{(1+x)/x^3}$. Reduction of order gives the regular growing linear growth factorIts Wronskian with $D_{-}$ is $x^{-3/2}(1+x{)}^{-1/2}$, so the modes are independent. The limits are $D_{+}\sim 2x/5$ during matter domination and $D_{+}\to 1$ in a curvature-dominated universe. Multiplying by $5/2$ normalizes the early mode to $x$. Negative spatial curvature of an FLRW universe thus stops late linear amplification even though the model contains no dark energy.

For pressureless linear cosmological density perturbations in an Einstein-de Sitter universe, $H=2/(3t)$ and $4\pi G{\bar{\rho }}_m=2/(3t^2)$. The growth equation is an Euler differential equation whose indicial equation is $(p-2/3)(p+1)=0$. Hence the independent modes are ${\delta }_{+}\propto t^{2/3}\propto a$ and ${\delta }_{-}\propto t^{-1}\propto a^{-3/2}$. The first is the growing linear growth factor, while the second is a decaying transient.

With physical matter and radiation densities, primordial amplitudes and spatial flatness fixed, extra smooth late dark energy increases $H_0$ but keeps ${\Omega }_m{H}_0^2$ fixed. Provided the dark energy is negligible at equality, the early cosmological transfer function and equality wavenumber in physical units are unchanged. Earlier accelerated expansion suppresses the linear growth factor, lowering present dimensional matter power by its squared growth ratio. Units involving $h$ can additionally relabel the horizontal and vertical axes.

For pressureless sub-Hubble perturbations, combine the linearized cosmological continuity equation, cosmological Euler equation and gravitational Poisson equation with matter as its source. The expansion contributes a $2H$ friction term and matter self-gravity contributes the negative coefficient of $\delta$. The background expansion can include radiation or dark energy; treating their perturbations as additional sources requires a more complete system.

Use a monotone linear growth factor $b(t)$ as time, $\mathbf{u}=\dot{\mathbf{x}}/\dot{b}$, and $\psi =a\phi /(Cb)$ with $C=3H_0^2{\Omega }_{m0}/2$ and $a_0=1$. The cosmological Poisson equation becomes ${\nabla }^2\psi =\delta /b$; the continuity equation becomes ${\partial }_b\delta +\nabla \cdot [(1+\delta )\mathbf{u}]=0$. The material derivative is $D_b={\partial }_b+\mathbf{u}\cdot \nabla$. Substitution into the pressureless cosmological Euler equation, followed by the growth equation and $\dot{b}=Hbf$, proves the displayed motion equation, with $f=d\ln b/d\ln a$. The density derivative in the conservative continuity equation is Eulerian, not material.

For a pure scalar growing density contrast with time-independent boundary conditions, ${\nabla }^2\psi ={\delta }_{\ast }$ fixes the scaled peculiar gravitational potential up to an irrelevant constant. The linearized continuity equation fixes the longitudinal scaled velocity to $-\nabla \psi$. The remaining homogeneous velocity is ${\mathbf{u}}_h=\mathbf{C}(\mathbf{x})/(a^2\dot{b})$; a divergence-free such field does not change the growing density at linear order. Excluding this decaying velocity component is necessary for time-independent $\mathbf{u}$. In an Einstein-de Sitter universe, $b=a$ and ${\mathbf{u}}_h\propto a^{-3/2}$, furnishing a counterexample if the pure growing-velocity hypothesis is omitted.

For dust, curvature and a constant cosmological constant, write $H^2=A{a}^{-3}+B{a}^{-2}+C$. Then $\dot{H}=-3A{a}^{-3}/2-B{a}^{-2}$ and $\ddot{H}=H(9A{a}^{-3}/2+2B{a}^{-2})$, which imply $\ddot{H}+2H\dot{H}-3A{a}^{-3}H/2=0$. Thus $H$ is a solution of the pressureless linear matter perturbation growth equation. In an Einstein-de Sitter universe, $H\propto t^{-1}$ is the decaying mode and the independent growing mode is $t^{2/3}$. Additional dynamical components need not preserve this identity.

For pressureless matter on subhorizon scales, the linear growth factor obeysThe second term is dilution by cosmic expansion, while the final term drives Jeans instability.

Replacing the self-gravity term in the matter growth equation by $(1-f)4\pi G{\bar{\rho }}_{m}D$, with $0<1-f<1$, weakens clustering while leaving the background expansion unchanged.

During matter domination, the growing mode under a reduced clustering strength $1-f$ isIt grows more slowly than the standard $t^{2/3}$ mode whenever $0<f<1$.

In a matter-dominated universe, the long-wavelength pressureless growth equation has modes

When matter self-gravity is negligible during radiation domination, $a\propto t^{1/2}$ and $\ddot{\delta }+2H\dot{\delta }=0$ gives $\delta =C_1+C_2\log a$.

In a radiation-dominated background with matter self-gravity retained, ${\delta }_{aa}+{\delta }_a/a-3\delta /(2a{a}_{\mathrm{eq}})=0$. A change to $x=\sqrt{6a/a_{\mathrm{eq}}}$ gives the Modified Bessel differential equation of order zero. The Modified Bessel function of the first kind increases as $1+3a/(2a_{\mathrm{eq}})+\cdots$; the Modified Bessel function of the second kind decreases with a leading negative logarithm. The early solution space has the constant/logarithmic behaviors, not the $a$ growth of matter domination. This background approximation must not be extrapolated through equality.

During curvature domination $a\propto t$, so a pressureless perturbation without a self-gravity term has modes $1$ and $a^{-1}$. The surviving mode is constant.

For a pressureless fluid withthe first-order density contrast $\delta =\delta \rho /\bar{\rho }$ obeys

The Jeans wavenumber separates pressure-supported modes from gravitationally unstable density modes.

Jeans instability is the gravitational growth of density perturbations whose self-gravity overcomes pressure support. For physical sound speed $c_s$, the comoving threshold isModes with $k\ll k_J$ grow gravitationally, whereas modes with $k\gg k_J$ undergo pressure-supported acoustic oscillations.

For a matter-dominated fluid with constant $c_s^2=w\ll 1$, combine stress-energy conservation in synchronous gauge with the scalar trace Einstein equation. Discard pressure corrections to the background gravitational and velocity-damping terms while retaining the potentially large pressure-gradient term $w{k}^2\delta$. This produces the displayed leading nonrelativistic equation. It is not an exact equation at finite $w$. At negligible pressure the growing and decaying matter-era solutions are ${\tau }^2$ and ${\tau }^{-3}$.

// Target: cosmology.bigb

In Synchronous gauge in cosmology with coordinates comoving with cold dark matter, the linearized Einstein field equations and matter conservation give the displayed density equation. Primes denote conformal time derivatives. The radiation term contains its pressure contribution and cannot be dropped on superhorizon scales merely because radiation oscillates after cosmological horizon entry.

During radiation domination, $a\propto \tau$ and $8\pi G{\rho }_R{a}^2=3/{\tau }^2$. The adiabatic initial conditions reduce the CDM density equation in a matter-radiation universe to ${\delta }_C^{\prime \prime }+{\delta }_C^{\prime }/\tau -4{\delta }_C/{\tau }^2=0$. Its Euler-Cauchy equation powers are two and minus two. The growing solution is the displayed one, in the stated CDM-comoving synchronous gauge.

In the leading radiation domination limit, neglect $k^2{\delta }_r$ and the matter contribution to gravity. The radiation equation gives $S^{\prime \prime }=0$. Substitution into ${\delta }_c^{\prime \prime }+{\delta }_c^{\prime }/\tau =3{\delta }_r/{\tau }^2$ givesThus adiabatic initial conditions select both $S_0=0$ and $S_1=0$, leaving the growing and decaying power-law modes. The four constants describe the leading supplied coupled system; regularity and additional initial Einstein field equations constraints can further restrict their physical interpretation.

With physical wavelength $2\pi /k$, the Jeans length is $2\pi /k_J$, where the Jeans wavenumber is $\sqrt{4\pi G\rho }/c_s$. Linear pressure support dominates gravity below this length, whereas longer wavelengths can grow. In an expanding background the growth also depends on the Hubble damping term, so the static threshold does not by itself give the time dependence.

The physical Jeans length is ${\ell }_J=c_s\sqrt{\pi /(G{\rho }_0)}$. A pressureless uniform sphere of initial density ${\rho }_0$ has free-fall time of a uniform sphere $t_{\mathrm{ff}}=\sqrt{3\pi /(32G{\rho }_0)}$, giving the title relation. Thus a full-wavelength sound-crossing time and this exact free-fall convention agree in scaling, not numerical value. Using the scale $k_J^{-1}$ instead gives $1/(c_s{k}_J)=1/\sqrt{4\pi G{\rho }_0}$, the inverse long-wave static Jeans growth-rate spectrum limit. At the threshold itself pressure balances gravity and the net growth rate vanishes; the pressureless free-fall time must not be confused with that marginal mode's infinite growth time.

When baryons are a small fraction of the gravitating matter, their perturbation is forced by the cold dark matter potential. Comparing pressure frequency $c_s^2{k}^2$ with $4\pi G{a}^2{\bar{\rho }}_C$ yields the scale ${\lambda }_J=c_s\sqrt{\pi /(G{\bar{\rho }}_C)}$ for following the growing matter mode. It is not an absolute statement that every sub-Jeans baryon perturbation is zero or cannot be forced: a slowly varying response is ${\delta }_B\simeq (k_J^2/k^2){\delta }_C$. In the pressureless limit the difference from CDM is a constant plus a decaying mode during matter domination, so baryons catch up in relative amplitude.

For the screened attraction $-Gm{e}^{-\alpha r}/r$ with $\alpha \ge 0$, $({\nabla }^2-{\alpha }^2)\varphi =4\pi G\rho$. A static homogeneous barotropic medium with sound speed $v_s>0$ has ${\omega }^2=k^2[v_s^2-4\pi G{\rho }_0/(k^2+{\alpha }^2)]$. A growing mode exists iff $4\pi G{\rho }_0>v_s^2{\alpha }^2$, with $0<k<\sqrt{4\pi G{\rho }_0/v_s^2-{\alpha }^2}$. For nonzero screening the growth rate tends to zero as $k\to 0$; the homogeneous mode is not an exponentially growing density fluctuation. Finite interaction range can therefore remove the instability entirely, while all sufficiently small nonzero wavenumbers remain unstable when the threshold is exceeded.

The Jeans swindle is the removal of the unperturbed uniform density's gravitational field when treating an infinite homogeneous Newtonian medium as static. Without such subtraction or an external supporting prescription, a nonzero constant density and zero background gravitational acceleration cannot satisfy Poisson's equation. Linear perturbations then obey ${\nabla }^2\delta \varphi =4\pi G\delta \rho$. This is a specified idealization, not an exact isolated infinite equilibrium.

In a static homogeneous universe, a density mode obeysModes with $k>k_J$ oscillate as sound waves, while modes with $k<k_J$ have one exponentially growing and one exponentially decaying solution.

Under the Jeans swindle, longitudinal density disturbances in a static uniform gas have ${\omega }^2=c_s^2{k}^2-4\pi G{\rho }_0$. For $0<k<k_J=\sqrt{4\pi G{\rho }_0}/c_s$, one solution grows as $e^{s(k)t}$ with the displayed rate. It decreases monotonically with wavenumber, so the longest available wavelengths grow fastest and the growth-rate supremum is $\sqrt{4\pi G{\rho }_0}$ as $k\to 0^{+}$. There is no preferred finite fastest wavelength in this model. The exact uniform $k=0$ density change is not an ordinary finite-potential Fourier perturbation of the fixed background, and finite-cloud boundaries restrict the available long modes.

Up to a convention-dependent numerical factor, a collisional fluid has comoving Jeans length ${\lambda }_{J,\mathrm{com}}=(c_s/a)\sqrt{\pi /(G\bar{\rho })}$.

Before cosmological recombination, photon-baryon sound speed controls the baryonic comoving Jeans length. The tightly coupled photon-baryon fluid has $c_s^2=c^2/[3(1+R_b)]$, where $R_b=3{\bar{\rho }}_b/(4{\bar{\rho }}_{\gamma })\propto a$. Using total background density in the instantaneous dynamical Jeans estimate, the scale grows as $a$ during radiation domination, approaches a plateau in strongly baryon-loaded matter domination, and drops when photon decoupling removes photon pressure support. Subsequently, adiabatic monatomic-gas cooling gives $T_b\propto a^{-2}$ and a comoving Jeans length proportional to $a^{-1/2}$. Residual Compton scattering can delay that cooling law.

For collisionless matter, velocity dispersion replaces sound speed in the Jeans estimate. After free decoupling, a nonrelativistic thermal velocity dispersion redshifts as $a^{-1}$.

An instantaneous kinetic suppression estimate replaces sound speed by the velocity dispersion of cold dark matter. Nonrelativistic particles in thermal contact with photons have $\sigma \propto a^{-1/2}$; after kinetic decoupling, their momenta redshift and $\sigma \propto a^{-1}$. A background dynamical estimate using total density is then constant during radiation domination and decreases as $a^{-1/2}$ during matter domination. Using only the species' own matter density instead makes the self-gravitating collisionless Jeans length decrease as $a^{-1/2}$ throughout the decoupled nonrelativistic era. Neither instantaneous estimate is the accumulated distance traveled by collisionless free streaming.

Collisionless free streaming is the ballistic motion of particles out of an overdensity. It erases perturbations below the comoving distance travelled by particles while their thermal velocities are appreciable.

Suppose a collisionless particle has momentum $p=T_r\propto a^{-1}$ during pure radiation domination. Since $a\propto t^{1/2}$, integrating $\chi ={\int }_0^{t}v(t^{\prime })d{t}^{\prime }/a(t^{\prime })$ gives the displayed comoving distance. The physical distance at observation is $a\chi =H^{-1}\mathrm{arsinh}(y)/y$, $y=m/T_r$. The printed expression without $a^{-1}$ is consequently a physical distance, or a comoving distance in coordinates normalized to $a=1$ at observation. Its ratio tends to one in the relativistic limit and to $\log (2y)/y$ in the nonrelativistic limit. The accumulated scale differs from an instantaneous velocity-over-expansion-rate estimate.

Spatial translation symmetry in a flat FRW metric conserves the covariant canonical momentum $q_i=m{a}^{2}d{x}^i/ds$ of a freely falling massive particle. An orthonormal spatial frame gives $p_{\mathrm{phys},i}=q_i/a$, so physical momentum redshifts as $a^{-1}$ even while the particle changes from relativistic to nonrelativistic motion. The local energy is $E=\sqrt{m^2+q^2/a^2}$ and the coordinate velocity is $d\mathbf{x}/dt=\mathbf{q}/[a\sqrt{q^2+m^2{a}^2}]$.

A freely falling particle's physical momentum is measured in the local orthonormal frame of a comoving observer. With proper peculiar velocity $\mathbf{v}=ad\mathbf{x}/dt$ and Lorentz factor $\gamma =(1-v^2{)}^{-1/2}$, it is $m\gamma \mathbf{v}$. It differs by one factor of the scale factor from the conserved comoving momentum. For a nonrelativistic particle, its peculiar velocity therefore decays as $a^{-1}$.

For a barotropic fluid, the squared sound speed is $c_s^2=\partial P/\partial \rho$. A Fourier density mode obeysin linear Newtonian perturbation theory.

For a barotropic perfect fluid with constant $w$, its cosmological adiabatic sound speed satisfies $c_s^2=w$. If $w<0$, the short-wavelength pressure term has the wrong sign for restoring oscillations: a mode grows as $\exp (\sqrt{-w}k\tau )$ when expansion can be neglected. For $w=-1/3$ in a coasting background with constant conformal expansion rate $K$, the exponents are $-K\pm \sqrt{K^2+k^2/3}$. This is a gradient instability, not ordinary positive-pressure Jeans instability. A network with tension or elastic anisotropic stress needs different constitutive perturbation equations; a negative background equation of state alone does not establish its perturbation sound speed.

The squared adiabatic sound speed is the ratio of background pressure and energy-density time derivatives. For a barotropic equation of state it is $dP/d\rho$. The non-adiabatic pressure perturbation is the difference $\delta P-c_a^2\delta \rho$; in a general field or multicomponent model this background quantity need not equal the rest-frame propagation speed.

If a smooth component dominates the expansion but does not cluster, pressureless matter still satisfiesAs ${\bar{\rho }}_m$ becomes negligible, Hubble friction leaves a constant mode and a decaying mode, so structure growth freezes.

When $H$ is constant and the matter source is negligible,so ${\delta }_m=C_1+C_2{a}^{-2}$. The growing matter-era mode therefore approaches a constant after cosmological-constant domination begins.

For constant positive Hubble parameter $\mu$ and a four-thirds polytropic equation of state, a linear cosmological density perturbation obeys $\ddot{\delta }+2\mu \dot{\delta }+{\eta }^2{e}^{-3\mu t}\delta =0$. Put $\delta =e^{-\mu t}y$ and $x=2|\eta |e^{-3\mu t/2}/(3\mu )$. The resulting Bessel differential equation has order $2/3$; use ordinary Bessel functions for ${\eta }^2>0$ and modified Bessel functions for ${\eta }^2<0$. Their small-$x$ expansions give one constant limiting mode and one $e^{-2\mu t}$ mode. The constant limit need not exceed the initial amplitude without an initial-velocity condition.

If ${\eta }^2=-s^2<0$, $\delta (0)>0$ and $\dot{\delta }(0)\ge 0$, the integrating factor identity $(e^{2\mu t}\dot{\delta }{)}^{\prime }=s^2{e}^{-\mu t}\delta$ proves that $\delta$ stays positive and increases. Its finite limit satisfies ${\delta }_{\infty }=\delta (0)+\dot{\delta }(0)/(2\mu )+s^2{\int }_0^{\infty }{e}^{-3\mu t}\delta (t)dt/(2\mu )>\delta (0)$. The initial sign condition is essential: the positive decaying modified Bessel function mode has limit zero.

In a matter-plus-cosmological-constant universe, changing the independent variable from cosmic time to $a$ gives

The growing solution of the matter perturbation equation in a spatially flat matter-plus-cosmological-constant universe isnormalized so that $D_{+}(a)\sim a$ during early matter domination.

When $H\to H_{\Lambda }>0$, the growth integral has an integrable $a^{-3}$ tail. Its product with $H$ tends to a constant, with leading correction proportional to $a^{-2}$ in a matter-plus-Lambda background. Both displayed basis functions can tend to constants; a linear combination cancels that constant and isolates the truly decaying late-time $a^{-2}$ behavior.

Changing the lower limit in $H(a){\int }_{a_i}^{a}d\tilde{a}/(\tilde{a}H{)}^3$ adds a constant multiple of $H(a)$, the independent matter-era decaying solution. Thus a finite lower limit defines a useful independent solution but can include a decaying admixture. A pure growing normalization is selected by an appropriate linear combination.