---
title: Einstein–de Sitter Spacetime Models
url: https://www.emergentmind.com/topics/einstein-de-sitter-spacetime
type: topic
---

# Einstein–de Sitter Spacetime Models

Einstein–de Sitter spacetime, as treated in the cited arXiv literature, is an expanding Robertson–Walker or generalized Robertson–Walker background on \((0,\infty)\times \mathbb R^3\) whose analytic significance lies in the singular behavior of its wave operator at \(t=0\), its non-Minkowskian propagation geometry, and the rigidity properties of spacelike hypersurfaces. In semilinear wave problems it is encoded by operators of the form \(\partial_t^2-t^{-4/3}\Delta+2t^{-1}\partial_t\) or, more generally, \(\partial_t^2-t^{-2k}\Delta+\mu t^{-1}\partial_t\), while in hypersurface theory it appears as a Lorentzian warped product with distinguished cosmological time slices [1612.09536][1401.0632][2009.04388][2009.05372][2101.12626].

## 1. Geometric realizations and model forms

The cited papers present Einstein–de Sitter spacetime through closely related Robertson–Walker-type descriptions. In the semilinear wave analysis, the Einstein–de Sitter universe is presented as a particular Friedmann–Robertson–Walker metric with scale factor
\[
a_{sc}(t)=t^{2/3},
\]
and line element
\[
ds^2=-dt^2+a_{sc}^2(t)\left[\frac{dr^2}{1-Kr^2}+r^2\,d\Omega^2\right].
\]
In the Einstein–de Sitter case this becomes
\[
ds^2=-dt^2+t^{4/3}\left[\frac{dr^2}{1-Kr^2}+r^2\,d\Omega^2\right].
\]
For the spatially flat case \(K=0\), this is described as the familiar matter-dominated cosmological model [1612.09536].

In the hypersurface rigidity study, the ambient spacetime is the \(4\)-dimensional Einstein–de Sitter model
\[
\bar M=(0,\infty)\times \mathbb R^3
\]
with Lorentzian metric
\[
\bar g=-dt^2+t^{2/3}(dx_1^2+dx_2^2+dx_3^2),
\]
equivalently a warped product with warping function \(f(t)=t^{1/3}\). In that formulation Einstein–de Sitter is a Robertson–Walker, indeed a generalized Robertson–Walker, spacetime with flat fiber \(\mathbb R^3\) [1401.0632].

The generalized PDE literature abstracts the same expanding background through a time-dependent propagation speed
\[
a_k(t)=t^{-k},\qquad k\in[0,1)\ \text{or}\ k\in(0,1),
\]
and corresponding operators
\[
\partial_t^2-t^{-2k}\Delta+2t^{-1}\partial_t
\quad\text{or}\quad
\partial_t^2-t^{-2k}\Delta+\mu t^{-1}\partial_t.
\]
For \(k=\frac23\) and \(n=3\), these reduce to the standard Einstein–de Sitter semilinear wave equation; in the damped model the classical Einstein–de Sitter case is \(k=\frac23\), \(\mu=2\), \(n=3\) [2009.04388][2009.05372][2101.12626].

| Source | Geometric or operator form | Standard EdS identification |
|---|---|---|
| [1612.09536] | \(ds^2=-dt^2+t^{4/3}\left[\frac{dr^2}{1-Kr^2}+r^2\,d\Omega^2\right]\) | Spatially flat \(K=0\) case emphasized |
| [1401.0632] | \(\bar g=-dt^2+t^{2/3}(dx_1^2+dx_2^2+dx_3^2)\) | GRW spacetime with flat fiber \(\mathbb R^3\) |
| [2009.04388], [2009.05372], [2101.12626] | \(\partial_t^2-t^{-2k}\Delta+2t^{-1}\partial_t\) or \(\partial_t^2-t^{-2k}\Delta+\mu t^{-1}\partial_t\) | \(k=\frac23\), and in the damped case \(\mu=2\), \(n=3\) |

A recurrent feature in all of these formulations is that the cosmological background is not treated as passive decoration. It alters either the metric coefficients directly or the wave operator through time-dependent degeneracy and singular lower-order terms.

## 2. Wave operators, singular initial geometry, and finite propagation

For the Einstein–de Sitter semilinear wave equation, the covariant model is
\[
\square_g \psi=\lambda |\psi|^{p-1}\psi,
\]
and the Einstein–de Sitter d’Alembertian takes the form
\[
\square_{EdeS}\psi = -\partial_t^2\psi -2t^{-1}\partial_t\psi +t^{-4/3}A(x,\partial_x)\psi,
\]
where \(A(x,D_x)\) is a second-order elliptic operator, equal to \(\Delta\) in the physically standard flat case. The resulting equation is
\[
\psi_{tt}-t^{-4/3}A(x,D_x)\psi+2t^{-1}\psi_t=F(\psi), \qquad t>0,\ x\in \mathbb R^n.
\]
The two analytically distinctive features identified in the paper are a non-Fuchsian singularity at \(t=0\) and a finite propagation radius governed by the optical distance
\[
\phi(t)=3t^{1/3},
\]
rather than the Minkowski \(t\)-cone [1612.09536].

Because \(t=0\) is singular, standard Cauchy data \(\psi(x,0)\), \(\psi_t(x,0)\) are not used. The weighted initial data are
\[
\lim_{t\to 0^+} t\psi(x,t)=\varphi_0(x),
\]
\[
\lim_{t\to 0^+}\left(t\psi_t(x,t)+\psi(x,t)+3t^{-1/3}A(x,D_x)\varphi_0(x)\right)=\varphi_1(x),
\]
with limits taken in \(H^1(\mathbb R^n)\) and \(L^2(\mathbb R^n)\), respectively. The singular initial hypersurface therefore forces a renormalized notion of data [1612.09536].

A key simplification is the Liouville-type transform
\[
\psi=t^{-1}u,
\]
based on the identity
\[
t^{-1}\circ {\mathcal S}\circ t={\mathcal L},
\]
where
\[
{\mathcal L}:=\partial_t^2-t^{-4/3}A(x,D_x)+2t^{-1}\partial_t,\qquad
{\mathcal S}:=\partial_t^2-t^{-4/3}A(x,D_x).
\]
The transformed equation is
\[
u_{tt}-t^{-4/3}A(x,D_x)u=t^{1-p}|u|^p,
\]
with transformed initial conditions
\[
\lim_{t\to0^+}u(x,t)=\varphi_0(x),\qquad
\lim_{t\to0^+}\left(u_t(x,t)+3t^{-1/3}A(x,D_x)\varphi_0(x)\right)=\varphi_1(x).
\]
Analytically, this removes the singular damping term and yields a semilinear generalized Tricomi-type equation with singular time-dependent propagation coefficient and source \(t^{1-p}|u|^p\) [1612.09536].

Finite propagation is an essential structural property. In the singular problem, support estimates depend on the transformed operator \({\mathcal S}=\partial_t^2-t^{-4/3}A(x,D_x)\), and for the Einstein–de Sitter case the propagation radius grows like
\[
3T^{1/3}s_A,
\]
with
\[
s_A=\max_{x,\ |\xi|=1}\frac1{a(x)}\sum_{|\alpha|=2}a_\alpha(x)\xi^\alpha.
\]
In the generalized models posed for \(t>1\), the corresponding light-cone function is
\[
\phi_k(t)=\frac{t^{1-k}}{1-k},
\qquad
A_k(t)=\int_1^t \tau^{-k}\,d\tau=\phi_k(t)-\phi_k(1),
\]
and support satisfies
\[
\operatorname{supp}u(t,\cdot)\subset B_{R+A_k(t)}.
\]
This expresses the curved propagation geometry induced by the expanding background [1612.09536][2009.04388][2009.05372][2101.12626].

## 3. Power nonlinearities: finite lifespan and blow-up thresholds

For the massless self-interacting scalar field with power nonlinearity, the central finite-lifespan theorem states that if \(F(\psi)=|\psi|^p\), \(A\) satisfies the ellipticity and compact-perturbation assumptions, and
\[
1<p<p_{cr}(n),
\]
then for every \(\varepsilon>0\) and every Sobolev index \(s\), there exist \(\varphi_0,\varphi_1\in C_0^\infty(\mathbb R^n)\) with arbitrarily small \(H_{(s)}\)-norm such that the corresponding solution of the weighted Cauchy problem blows up in finite time. The blow-up criterion is
\[
\lim_{t\nearrow T}\int_{\mathbb R^n} a(x)\psi(x,t)\,dx=\infty
\]
for some \(T<\infty\). Thus finite lifespan is expressed through divergence of a weighted moment functional, not through an \(H^1\)-norm criterion [1612.09536].

The threshold is defined by
\[
p_{cr}(n):=\max\left\{p_0(n),\,1+\frac6n\right\},
\]
where \(p_0(n)\) is the positive root of
\[
(n+3)p^2-(n+13)p-2=0.
\]
In the physically relevant case \(n=3\),
\[
p_{cr}(3)=3,
\]
so the theorem yields finite-time blow-up for all
\[
1<p<3.
\]
The same paper notes that \(p=3\) corresponds to the \(\varphi^4\) model in the equation \(\square_g\psi=\lambda |\psi|^{p-1}\psi\), since the exponent in the PDE is \(p\), so \(p=3\) yields a quartic potential [1612.09536].

For positive solutions and the gauge-invariant nonlinearity \(F(\psi)=|\psi|^{p-1}\psi\), the corollary states that in dimension \(n=3\), if \(1<p<3\), then one can choose arbitrarily small smooth compactly supported data so that the positive solution has finite lifespan. A direct implication is that global sign-preserving small-data solutions do not exist in this range [1612.09536].

The same work studies a shifted-time problem,
\[
\psi_{tt}-t^{-2k}A(x,D_x)\psi+2t^{-1}\psi_t=|\psi|^p,\qquad t>1,
\]
with standard data at \(t=1\). In that setting blow-up occurs if either
\[
1<p<1+\frac{2}{n(1-k)},
\]
or
\[
1<p\le 2+\frac{2k}{n+1-kn}
\quad\text{and}\quad
p<p_0(n,k),
\]
where \(p_0(n,k)\) is the positive root of
\[
p^2(n+1-kn)-p(2k+n+3-kn)-2(1-k)=0.
\]
For the Einstein–de Sitter value \(k=\frac23\), \(n=3\), this again gives blow-up for
\[
1<p<3.
\]
This rules out the interpretation that finite lifespan is merely an artifact of the singular initial hypersurface \(t=0\); the blow-up persists even when the singularity is removed by shifting the initial time to \(t=1\) [1612.09536].

The proof architecture combines a moment method with low-frequency test functions and Kato-type differential inequalities. The basic functional
\[
F(t)=\int_{\mathbb R^n}a(x)u(x,t)\,dx
\]
satisfies
\[
F''(t)=t^{1-p}\int_{\mathbb R^n}a(x)|u(x,t)|^p\,dx\ge 0,
\]
so \(F\) is convex. A sharper argument introduces an elliptic eigenfunction \(\varphi\) satisfying \(A(x,D_x)\varphi=\varphi\), a time profile
\[
\lambda(t)=\bigl(3t^{1/3}+1\bigr)e^{-3t^{1/3}},
\]
and the projected functional
\[
F_1(t)=\int_{\mathbb R^n}a(x)u(x,t)\lambda(t)\varphi(x)\,dx.
\]
The lower bounds for \(F_1\), combined with support estimates and Hölder inequalities, yield the improved differential inequalities needed to reach the threshold \(p_{cr}(n)\) [1612.09536].

## 4. Generalized Einstein–de Sitter models, critical exponents, and lifespan estimates

A generalized Einstein–de Sitter spacetime is encoded in the wave equation
\[
\phi_{tt}-t^{-2k}\Delta \phi + 2 t^{-1}\phi_t = |\phi|^p,
\qquad k\in[0,1),
\]
or, after the transformation \(u=t\phi\),
\[
u_{tt}-t^{-2k}\Delta u=t^{1-p}|u|^p.
\]
The natural solution class in the lifespan analysis is the energy solution
\[
u\in C([1,T),H^1(\mathbb{R}^n))\cap C^1([1,T),L^2(\mathbb{R}^n))\cap L^p_{\mathrm{loc}}([1,T)\times \mathbb{R}^n),
\]
with compactly supported nonnegative data prescribed at \(t=1\) [2009.04388].

Two exponents govern the blow-up theory:
\[
p_0(n,k)\ \text{is the positive root of}\ 
((1-k)n+1)p^2-((1-k)n+3+2k)p-2(1-k)=0,
\]
and
\[
p_1(n,k)=1+\frac{2}{(1-k)n}.
\]
The dominant critical threshold is
\[
p_{\mathrm{crit}}=\max\{p_0(n,k),\,p_1(n,k)\}.
\]
The comparison of \(p_0\) and \(p_1\) is organized by
\[
N(k)=\frac{1-2k+\sqrt{4k^2-4k+3}}{2(1-k)},
\]
so that \(p_0(n,k)>p_1(n,k)\) if and only if \(n>N(k)\), while \(p_0(n,k)\le p_1(n,k)\) if and only if \(n\le N(k)\) [2009.04388].

In the Strauss-type critical case \(p=p_0(n,k)\) with \(n>N(k)\), every sufficiently small nonnegative compactly supported energy solution blows up in finite time, and
\[
T(\varepsilon)\le \exp\!\big(C\varepsilon^{-p(p-1)}\big).
\]
In the Fujita-/ODE-type critical case \(p=p_1(n,k)\) with \(n\le N(k)\), finite-time blow-up again occurs, with
\[
T(\varepsilon)\le \exp\!\big(C\varepsilon^{-(p-1)}\big).
\]
For
\[
1<p<\max\{p_0(n,k),p_1(n,k)\},
\]
the paper gives polynomial upper bounds, including
\[
T(\varepsilon)\lesssim \varepsilon^{-\left(\frac{2}{p-1}-(1-k)n\right)^{-1}}
\]
in the \(p_1\)-subcritical region and
\[
T(\varepsilon)\lesssim \varepsilon^{-\theta(p,n,k)},
\]
with
\[
\theta(p,n,k)=1-k +\left((1-k)\frac{n+1}{2}+\frac{3k}{2}\right)p -\left((1-k)\frac{n-1}{2}-\frac{k}{2}\right)p^2
\]
in the \(p_0\)-subcritical region [2009.04388].

The critical cases are analytically marginal. The paper handles them by an iteration argument with slicing, in which successive lower bounds improve logarithmically on subintervals such as \([\ell_j t,t]\), where
\[
\ell_j = 2-2^{-(j+1)}.
\]
In the Strauss-type critical regime the method is built around a weighted functional
\[
U(t)=t^{-k/2}\int_{\mathbb{R}^n}u(t,x)\,\xi_q(t,t,x;k)\,dx,
\]
while in the Fujita-type critical regime it uses the spatial average
\[
U(t)=\int_{\mathbb{R}^n}u(t,x)\,dx.
\]
The key linear auxiliary ODE
\[
\partial_t^2 y_j(t,s;\lambda,k)-\lambda^2 t^{-2k} y_j(t,s;\lambda,k)=0
\]
is solved explicitly through modified Bessel functions after the change of variable \(\tau=\lambda\phi_k(t)\) [2009.04388].

A further generalization introduces singular damping:
\[
u_{tt}-t^{-2k}\Delta u+\mu t^{-1}u_t=|u|^p.
\]
In that setting the paper identifies a Fujita-type exponent
\[
p_1(k,n)=1+\frac{2}{(1-k)n}
\]
and a generalized Strauss-type root given by
\[
\frac{n-1}{2}p^2+\frac{\mu-k}{2(1-k)}p^2 -\frac{n+1}{2}p-\frac{\mu+3k}{2(1-k)}p -1=0.
\]
The conjectured critical exponent is
\[
p_{\mathrm{crit}}=
\max\left\{
p_0\!\left(k,n+\frac{\mu}{1-k}\right),\
p_1(k,n)
\right\}.
\]
The paper interprets
\[
n\mapsto n+\frac{\mu}{1-k}
\]
as a formal shift in dimension, recovering the flat damped-wave shift when \(k=0\), the undamped generalized Tricomi threshold when \(\mu=0\), and the Einstein–de Sitter model at \(k=\frac23\), \(\mu=2\), \(n=3\) [2009.05372].

## 5. Derivative-type nonlinearities and Glassey-type thresholds

The generalized Einstein–de Sitter operator also supports a derivative-type semilinear problem,
\[
\partial_t^2 u - t^{-2k}\Delta u + 2t^{-1}\partial_tu = |\partial_tu|^p,
\qquad t>1.
\]
For \(n>1\), \(k\in(0,1)\), nonnegative compactly supported data, and local solutions with finite propagation
\[
\operatorname{supp}u(t,\cdot)\subset B_{R+A_k(t)},
\]
the blow-up range is
\[
1<p\le p_{\mathrm{Gla}\big((1-k)n+2k+2\big),
\]
where
\[
p_{\mathrm{Gla}(m)=
\begin{cases}
\dfrac{m+1}{m-1}, & m>1,\\[1mm]
\infty, & m=1.
\end{cases}
\]
Thus the geometry and damping shift the effective Glassey dimension from \(n\) to \((1-k)n+2k+2\) [2101.12626].

The lifespan estimate is
\[
T(\varepsilon)\le
\begin{cases}
C\varepsilon^{-\left(\dfrac{1}{p-1}-\dfrac{(1-k)n+2k+1}{2}\right)^{-1}},
& 1<p<p_{\mathrm{Gla}\big((1-k)n+2k+2\big),\\[2mm]
\exp\!\big(C\varepsilon^{-(p-1)}\big),
& p=p_{\mathrm{Gla}\big((1-k)n+2k+2\big).
\end{cases}
\]
For the standard Einstein–de Sitter parameters \(n=3\), \(k=\frac23\), the upper-bound exponent becomes
\[
p_{\mathrm{Gla}\!\left(\frac{13}{3}\right)=\frac85.
\]
The comparison Tricomi-type model
\[
\partial_t^2 u - t^{-2k}\Delta u = |\partial_tu|^p
\]
has threshold
\[
p_{\mathrm{Gla}\big((1-k)n+2k\big),
\]
so the Einstein–de Sitter damping term produces a shift by \(2\) in the effective dimension [2101.12626].

The method combines a one-dimensional integral representation formula for the linear problem with Yagdjian’s integral transform approach and a nonlinear characteristic-line argument. In one dimension, the linear Cauchy problem
\[
\partial_t^2u-t^{-2k}\partial_x^2u+\mu t^{-1}\partial_tu+\nu t^{-2}u=g(t,x)
\]
admits an explicit representation with kernels \(E\), \(K_0\), \(K_1\) written in terms of hypergeometric functions. For the Einstein–de Sitter specialization \((\mu,\nu)=(2,0)\), the kernel satisfies
\[
E(t,x;b,y;2,0,k)=\frac{b}{t}E(t,x;b,y;0,0,k).
\]
After integrating over transverse variables and using finite propagation, the proof reduces to a nonlinear integral inequality for a renormalized characteristic quantity \(\mathcal U(z)\), from which the Glassey-type threshold is extracted [2101.12626].

A common source of ambiguity is the phrase “nonexistence of global solutions.” In this setting it means finite-time blow-up or failure of global-in-time continuation for local classical solutions with the stated sign and support assumptions; it does not mean absence of local solutions [2101.12626].

## 6. Spacelike hypersurfaces, hyperbolic angle, and rigidity

In the geometric analysis of hypersurfaces, Einstein–de Sitter spacetime is the Lorentzian warped product
\[
\bar M=(0,\infty)\times \mathbb R^3,\qquad
\bar g=-dt^2+t^{2/3}(dx_1^2+dx_2^2+dx_3^2),
\]
and spacelike hypersurfaces are \(3\)-dimensional immersed manifolds \(x:S\to\bar M\) with induced Riemannian metric \(g\). Choosing the future-pointing unit timelike normal \(N\) satisfying
\[
\bar g(N,-\partial_t)<0,
\]
one has
\[
\bar g(N,\partial_t)=\cosh\theta,
\]
where \(\theta\) is the hyperbolic angle between \(N\) and \(-\partial_t\). The restriction of cosmological time to the hypersurface is
\[
\tau=\pi_I\circ x,
\]
with
\[
\nabla\tau=-\partial_t^\top,\qquad
g(\nabla\tau,\nabla\tau)=\sinh^2\theta.
\]
Hence \(\sinh^2\theta\) measures deviation from being a slice \(t=\mathrm{const}\) [1401.0632].

The shape operator \(A\) and mean curvature \(H\) are defined by
\[
AX=-\bar\nabla_XN,\qquad
H=-\frac13\operatorname{trace}(A).
\]
For a slice \(t=t_0\),
\[
A=\frac{f'(t_0)}{f(t_0)}\,I,\qquad
H=-\frac{f'(t_0)}{f(t_0)}=-\frac{1}{3t_0}\neq 0.
\]
Therefore Einstein–de Sitter spacetime has no maximal slices. A hypersurface lies between two spacelike slices when
\[
x(S)\subset [t_1,t_2]\times \mathbb R^3,
\]
equivalently when \(\tau\) is bounded above and below by positive constants [1401.0632].

The distinguished timelike vector field
\[
K:=(\pi_I)^{1/3}\partial_t
\]
is conformal, since
\[
\bar\nabla_X K=\frac13 (\pi_I)^{-2/3}X.
\]
Using \(K\), the time function \(\tau\), and the warped-product curvature identities, the paper derives formulas for \(\Delta\tau\), \(\Delta\tau^{1/3}\), \(\nabla\cosh\theta\), and \(\Delta\cosh\theta\), culminating in the inequality
\[
\cosh\theta\,\Delta\cosh\theta \ge \sinh^2\theta\left(\frac{H^2}{2}+\frac{1}{3\tau^2}\cosh\theta\right) +3\sinh^2\theta\left(\frac{1}{4\tau^2}-\frac{H^2}{3}\right) +\frac{1}{2\tau^2}\sinh^4\theta.
\]
This is the analytical engine of the rigidity argument [1401.0632].

A further ingredient is the intrinsic Ricci lower bound. For a constant mean curvature spacelike hypersurface \(S\) in Einstein–de Sitter spacetime, the paper proves
\[
\operatorname{Ric}_S(X,X)\ge -\frac{9H^2}{4}|X|^2
\]
for every tangent vector \(X\in TS\). This supplies the hypothesis for the Liouville-type theorem: if a complete Riemannian manifold has Ricci curvature bounded from below and a nonnegative smooth function \(u\) satisfies
\[
\Delta u\ge cu^2
\]
for some \(c>0\), then \(u\equiv 0\). Taking
\[
u=\sinh^2\theta,
\]
and using boundedness between two slices so that \(1/\tau^2\) has a positive lower bound, one obtains
\[
\Delta(\sinh^2\theta)\ge c\,(\sinh^2\theta)^2.
\]
Hence \(\sinh^2\theta\equiv 0\), so \(\theta\equiv 0\), and the hypersurface must be a spacelike slice [1401.0632].

The two main consequences are precise. First, there are no complete maximal hypersurfaces in Einstein–de Sitter spacetime lying between two spacelike slices. Second, the only complete constant mean curvature spacelike hypersurfaces lying between two spacelike slices are the slices themselves. In this sense, completeness, constant mean curvature, and confinement between two cosmological times rigidly determine the hypersurface geometry [1401.0632].

Source: https://www.emergentmind.com/topics/einstein-de-sitter-spacetime