Early Causal Tensor Power Spectrum

• Early Causal Tensor (ECT) power spectrum is defined by a universal k^3 infrared scaling from causal, sub-horizon gravitational wave sources in the early universe.
• The spectrum is parameterized by a dimensionless tensor amplitude r_c that directly links observational CMB B-mode features to fundamental early-universe source parameters.
• Observational constraints from CMB B-mode data impose model-independent bounds on r_c, impacting diverse sources such as phase transitions, cosmic defects, and scalar-induced gravitational waves.

The Early Causal Tensor (ECT) power spectrum characterizes a universal phenomenology for primordial gravitational waves (GWs) sourced by post-inflationary, causality-limited, and sub-horizon mechanisms in the early universe. Such sources include first-order cosmological phase transitions, topological defect networks, and scalar-induced GWs, all exhibiting a white-noise–like $k^3$ scaling at low comoving wavenumber. This infrared behavior emerges generically whenever the source correlation length is finite, yielding a distinct observational imprint on the cosmic microwave background (CMB) $B$-mode polarization and a specific spectral shape in the stochastic GW background. The ECT formalism provides parameterized, model-independent constraints on early-universe physics by mapping observational data directly to fundamental source parameters.

1. Theoretical Foundation and Motivation

Inflationary gravitational waves arise from quantum vacuum fluctuations stretched to super-horizon scales, yielding an approximately scale-invariant tensor power spectrum: $P_t^{\rm inf}(k) \propto k^{n_t}$ with $n_t \simeq 0$. In contrast, any gravitational wave background generated post-inflation—such as from bubble collisions, magnetohydrodynamic (MHD) turbulence, cosmic string networks, or large-amplitude small-scale scalar fluctuations—is subject to causal limits: no correlations can exist on scales exceeding the horizon at the time of generation. This causal bound imposes a generic infrared form, first shown in Cai et al. (2019), wherein the tensor power spectrum scales as $P_t(k) \propto k^3$ for $k \ll k_s$, with $k_s$ corresponding to the source’s maximal correlation scale (Greene et al., 28 Jan 2026).

If the source is operative and subsequently shuts off at redshift $z_p \gg 10^4$, well before matter–radiation equality, the $k^3$ regime extends across all observable CMB scales due to Silk damping at high $k$. These sources are collectively referred to as Early Causal Tensors (ECTs) (Zebrowski et al., 28 Jan 2026, Greene et al., 28 Jan 2026).

2. Universal Parameterization of the ECT Power Spectrum

On observational scales relevant to the CMB, the ECT power spectrum is parameterized in terms of the dimensionless tensor power spectrum: $$\mathcal{P}_h(k) = \frac{k^3}{2\pi^2} P_h(k) = r_c\,A_s\,\left(\frac{k}{k_0}\right)^3,$$ where:

• $k_0 = 0.01\,\mathrm{Mpc}^{-1}$ is the pivot scale, selected to align the ECT B-mode amplitude at $\ell \sim 100$ with that of an inflationary spectrum of equal amplitude.
• $A_s = 2.1 \times 10^{-9}$ is the scalar amplitude at $k_0$, as measured from the CMB.
• $r_c$ is the ECT amplitude, defined by $$r_c \equiv \mathcal{P}_h(k_0)/A_s = P_h(k_0)/P_s(k_0)$$.

This construction maintains direct analogy with the standard inflationary tensor-to-scalar ratio $r$, but for the characteristic $k^3$ ECT spectral shape (Zebrowski et al., 28 Jan 2026).

3. Physical Origin of the $k^3$ Infrared Scaling

For a generic causal, sub-horizon source with spacetime correlation length $R_*$ and finite duration $\Delta\tau$, the real-space correlation function $\xi(r)$ vanishes for $|r| > R_*$. The Fourier transform yields: $$P_h(k) = \int d^3r\,e^{-i \vec{k} \cdot \vec{r}}\,\xi(r) \propto c_0 + c_2\,k^2 + O(k^4),$$ implying $\mathcal{P}_h(k) \propto k^3$ at $k \ll k_* \equiv R_*^{-1}$.

A more general derivation via unequal-time correlators of the GW source stress-energy $\Pi(\tau, k)$ confirms this scaling in the small-$k$ limit whenever the source is local and of finite duration (Greene et al., 28 Jan 2026). All such sources, when completed at $z \gg 10^5$, display their $k^3$ regime over the full CMB window ($k_{\mathrm{min}} \sim 10^{-3}\,\mathrm{Mpc}^{-1}$ to $k_{\mathrm{max}} \sim 0.16\,\mathrm{Mpc}^{-1}$).

4. CMB $B$-Mode Angular Power Spectrum from ECTs

Tensor perturbations generate $B$-mode polarization through their dynamical impact on the photon-baryon fluid prior to recombination. For a given tensor power spectrum, the angular $B$-mode power is: $$C_\ell^{BB} = \int_0^\infty \frac{dk}{k} P_h(k)\, [\Delta_\ell^B(k)]^2,$$ where $\Delta_\ell^B(k)$ incorporates the transfer function for tensor-to-$B$-mode projection (Greene et al., 28 Jan 2026). Numerical evaluation using Boltzmann solvers such as CLASS is required for precision modeling on all angular scales.

The distinct ECT $k^3$ spectral form yields enhanced $C_\ell^{BB}$ at high multipoles and suppressed power at low multipoles, leading to a blue-tilted $B$-mode spectrum with a characteristic peak at $\ell \sim 10^3$, strong suppression at $\ell \lesssim 10$, and recombination-scale features at $\ell \sim 100$. This morphology provides a clear target for separating ECT signals from inflationary, scale-invariant primordial tensor backgrounds (Zebrowski et al., 28 Jan 2026, Greene et al., 28 Jan 2026).

5. Observational Constraints on the ECT Amplitude

A contemporary joint likelihood analysis of BICEP/Keck, SPT-3G, SPTpol, Planck, and WMAP $B$-mode data, encompassing $\ell\sim20$–2300, fits simultaneously for the inflationary $r$ and ECT $r_c$ amplitudes, as well as lensing and foreground nuisance parameters. Under uniform priors $r, r_c \in [0,1]$, the data yield no significant detection of the ECT component. The derived $95\%$ confidence upper limit is: $r_c < 0.0077 \quad (95\%~\mathrm{CL})$ (Zebrowski et al., 28 Jan 2026). This bound directly constrains the allowed amplitude of any ECT contribution, independent of microphysical source details, as all such sources yield the same $B$-mode spectral shape on CMB scales.

6. Mapping ECT Constraints to GW Energy Density and Early-Universe Models

The primordial ECT spectrum, with amplitude $r_c$, implies a present-day stochastic gravitational wave background (SGWB) with spectral density: $$\Omega_\mathrm{GW}(f) h^2 = \frac{h^2}{3} \left(\frac{k}{H_0}\right)^2 \mathcal{P}_h(k) \frac{1}{k^2} \propto f^3 \qquad (f \ll f_\mathrm{eq}),$$ for comoving $k=2\pi f$ today and for modes re-entering the horizon during radiation domination. The $r_c$ limit implies: $$\Omega_\mathrm{GW} h^2(f) \lesssim \mathrm{few} \times 10^{-20} \left(\frac{f}{10^{-19}~\mathrm{Hz}}\right)^3$$ at CMB frequencies, far below the reach of terrestrial and pulsar-timing GW experiments (Zebrowski et al., 28 Jan 2026).

A variety of early-universe models map to this constraint, as summarized below:

Model Class	Key Parameter(s)	ECT Constraint Mapping
First-order PT	$\alpha$, $\beta/H$, $v_w$	$$\Omega_{\rm GW}^{\rm PT} \sim 10^{-6}(H/\beta)^2(\alpha/0.1)^2(v_w/0.7)^3$$; $r_c$ limit constrains $$(H/\beta)^2(\alpha/0.1)^2 \lesssim \mathcal{O}(10^{-2})$$
Scalar-induced GW	$A_\zeta$	$$r_c \simeq A_\zeta^2 / A_s \implies A_\zeta \lesssim 10^{-2}$$, impacting PBH formation scenarios
Cosmic strings/defects	$G\mu$	$r_c \simeq C_s (G\mu/10^{-6})^2$ with $C_s \sim 0.1$–1; $G\mu \lesssim \mathrm{few} \times 10^{-7}$
Other causal mechanisms	coupling constants	$r_c < 0.0077$ involves bounds $\lesssim\mathcal{O}(10^{-1}-10^{-2})$ on model-specific couplings

The amplitude $r_c$ captures all model dependence; the spectral shape is universally set by causality (Zebrowski et al., 28 Jan 2026, Greene et al., 28 Jan 2026).

7. Distinctions from Inflationary and Loop-Induced Tensor Spectra

Inflationary models yield a nearly scale-invariant ($n_t \simeq 0$) spectrum, in contrast to the blue-tilted ECT form. Theoretical calculations of quantum corrections (loop- or open EFT–based) to the inflationary tensor power spectrum introduce subleading corrections of the form $$\Delta_t^2(k) = 2H^2/(\pi^2 M_{pl}^2)\left[1 + \mathcal{O}(H^2/M_{pl}^2)\ln(H/\mu)\right]$$, but do not mimic a causal $k^3$ scaling (Brahma et al., 2022). The ECT ansatz therefore isolates non-inflationary, early-universe GW sources via their distinct causal spectral shape.

• First constraints on causal sources of primordial gravitational waves from BICEP/Keck, SPTpol, SPT-3G, Planck and WMAP $B$-mode data (Zebrowski et al., 28 Jan 2026)
• A Universal CMB $B$-Mode Spectrum from Early Causal Tensor Sources (Greene et al., 28 Jan 2026)
• Quantum corrections to the primordial tensor spectrum: Open EFTs & Markovian decoupling of UV modes (Brahma et al., 2022)