First constraints on causal sources of primordial gravitational waves from BICEP/Keck, SPTpol, SPT-3G, Planck and WMAP B B -mode data
Jessica A. Zebrowski, Aurora Ireland, Christian L. Reichardt, Kylar Greene, Gordan Krnjaic, Yuhsin Tsai, François R Bouchet · 2026
Read First constraints on causal sources of primordial gravitational waves from BICEP/Keck, SPTpol, SPT-3G, Planck and WMAP B B -mode data by Jessica A. Zebrowski, Aurora Ireland, Christian L. Reichardt, Kylar Greene, Gordan Krnjaic, Yuhsin Tsai, François R Bouchet (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-20958.txt.
Introduction
Jessica Zebrowski and six co-authors set the first limits on a non-inflationary source of cosmic ripples. Their 2026 paper uses B-mode polarization data from five experiments. BICEP/Keck, SPTpol, SPT-3G, Planck, and WMAP all contribute. B-modes are swirl patterns in the cosmic microwave background. They can carry the imprint of primordial gravitational waves. Non-inflationary sources predict a causality-limited spectrum at low frequencies.
Causality caps how large the ripples can be on the biggest scales. That cap gives the spectrum a universal shape to test. No such constraints existed before this work. The combined data now bound how strong causal sources could have been. The result trims the space of early-universe theories. It shows the value of pooling many sky surveys into one test.
Catalog record
| CREATOR | Jessica A. Zebrowski, Aurora Ireland, Christian L. Reichardt, Kylar Greene, Gordan Krnjaic, Yuhsin Tsai, François R Bouchet |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-20958
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
First constraints on causal sources of primordial gravitational waves from BICEP/Keck, SPTpol, SPT-3G, Planck and WMAP B B -mode data Jessica A. Zebrowski, Aurora Ireland, Christian L. Reichardt, Kylar Greene, Gordan Krnjaic, Yuhsin Tsai, François R Bouchet Abstract Non-inflationary sources of gravitational waves in the early Universe generically predict causality-limited tensor power spectra at low frequencies. We report the first-ever constraints on such sources based on cosmic microwave background (CMB) B B -mode polarization measurements. Using data from BICEP/Keck, SPTpol, SPT-3G, Planck , and WMAP , we constrain the amplitude of an early causal tensor (ECT) power spectrum parameterized by r ect r_{\rm ect} , the ratio of causal tensor power to total scalar power at k = 0.01 k~=~0.01 Mpc -1 , and obtain a 95% CL upper limit of r ect < r_{\rm ect}{}< 0.0077 0.0077 . Since r ect r_{\rm ect} can easily be related to the parameters of a given theory, our bound robustly constrains a broad class of well-motivated gravitational wave sources in the early universe, including first-order cosmological phase transitions, enhanced small-scale density perturbations, and various topological defects. Finally, we translate our limit into a bound on the present-day energy density in gravitational waves at ultra-low frequencies otherwise inaccessible to traditional gravitational wave detection strategies, including pulsar timing arrays, interferometers, and resonant cavities. FERMILAB-PUB-25-0523-T I Introduction Tensor perturbations in the early universe imprint a distinctive B -mode pattern in the polarization of the cosmic microwave background (CMB), offering a powerful observational probe of fundamental physics at energy scales beyond the reach of terrestrial gravitational wave (GW) experiments 18 . While this connection has long been appreciated in the context of inflation, it has only recently been recognized that non-inflationary sources of primordial tensor perturbations can also yield observable B modes, potentially complicating interpretations of future measurements 13 ; 16 ; 33 . Examples of such sources include first-order cosmological phase transitions 19 ; 8 , enhanced scalar density perturbations 10 , dark-photon perturbations from axion rolling 12 , primordial gravitational vector modes 20 , and various topological defects 21 ; 26 ; 36 ; 15 . Unlike inflation, which stretches sub-horizon tensor perturbations to super-horizon scales and predicts a nearly scale-invariant power spectrum, post-inflationary sources are fundamentally limited by the impossibility of super-horizon correlations, and therefore universally yield white-noise spectra on sufficiently large scales. For all such sources, the dimensionless tensor power spectrum exhibits a characteristic \mathcal{P}_{h}(k)\propto k^{3} scaling at small comoving wavenumber k . This behavior is a robust, model-independent consequence of causality and finite correlation lengths, and it holds universally for length scales larger than the physical extent of the source, provided that the latter is bounded in frequency and time 6 . The characteristic k^{3} scaling will turn over at some k related to the correlation length of the source. However if the tensor source is before z\sim 7\times 10^{4} , this turn over will be hidden in the CMB by Silk damping – the horizon size at z\sim 7\times 10^{4} matches the Silk damping scale. Any non-inflationary source that generates GWs exclusively before this time necessarily predicts a white-noise power spectrum on CMB observable scales; this characteristic scaling yields a universal shape for the resulting B -mode angular power spectrum. We refer to such phenomena as early causal tensor (ECT) sources. Compared to inflationary predictions, ECT sources yield more power on small scales and less power on large scales. This striking shape difference allows for clear discrimination between the signals from inflation and ECT sources, given sufficiently precise measurements across a range of angular scales. While the ECT B -mode angular spectrum coincidentally resembles corresponding contributions from lensing, this degeneracy can be broken with external priors on lensing B modes. Furthermore, since all ECT sources predict the same B -mode spectral shape, the only parameter that distinguishes sources in this universality class is the overall amplitude, which we parametrize as r_{\rm ect} , the ratio of ECT tensor power to measured scalar power at the reference comoving wavenumber k_{\rm ref}=0.01 Mpc -1 . In this Letter , we report the first constraints on the causal tensor power spectrum using contemporary CMB B -mode observations. A companion work 14 explores models which produce such causal tensors. Marginalizing over foreground and cosmological signals, we place the first upper limit on the amplitude of the causal tensor spectrum of r_{\rm ect}{}< 0.0077 at 95% CL. The limit constrains a broad class of theoretical models that source GWs in the post-inflationary universe (see Fig. 1 ). We also translate this limit to a model-dependent bound on the GW energy density, \Omega_{\rm GW} , at ultra-low frequencies. II General Framework We briefly review the ECT framework presented by 14 , leaving details to that work. A Fourier mode of the transverse, traceless tensor perturbation h_{ij}(\tau,\vec{k}) evolves according to the wave equation h^{\prime\prime}_{ij}+2{\cal H}h^{\prime}_{ij}+k^{2}h_{ij}=16\pi Ga^{2}\,\Pi_{ij}\,, (1) where \tau is conformal time, ′ denotes differentiation with respect to \tau , a is the scale factor, and {\cal H}=a^{\prime}/a is the conformal Hubble rate. The source \Pi_{ij}(\tau,\vec{k}) is the transverse, traceless projection of the anisotropic stress tensor, whose functional form is model-dependent. We take \Pi_{ij} to be bounded in frequency and non-vanishing only transiently in the early universe for some characteristic time interval \tau\in[\tau_{i},\tau_{\star}] . We decompose h_{ij} and \Pi_{ij} into the two polarizations \lambda=(+,\times) , which are uncorrelated and have the same spectra. The solutions can be written as h^{\lambda}(\tau,\vec{k})\equiv h^{\lambda}_{\star}(\vec{k})\mathcal{T}(\tau,\vec{k}) where h^{\lambda}_{\star}(\vec{k})\equiv h^{\lambda}(\tau_{\star},\vec{k}) is the initial amplitude, and \mathcal{T}(\tau,\vec{k}) is the tensor transfer function, which solves the sourceless version of Eq. ( 1 ) and governs the deterministic evolution of these perturbations in an expanding universe. The initial amplitude depends on the source and is obtained by solving Eq. ( 1 ), h_{\star}^{\lambda}(\vec{k})=16\pi Ga^{2}\int_{\tau_{i}}^{\tau_{\star}}d\tau^{\prime}\,G_{k}(\tau,\tau^{\prime})\Pi_{\lambda}(\tau^{\prime},\vec{k}), (2) where G_{k}(\tau,\tau^{\prime})=\sin[k(\tau-\tau^{\prime})]/k is the Green’s function for Eq. ( 1 ), neglecting Hubble expansion over the short duration of source activity. Using this solution, we define the ( dimensionless ) tensor power spectrum \mathcal{P}_{h} through \left\langle h_{\star}^{\lambda}(\vec{k})\,h_{\star}^{\lambda^{\prime}}\!(\vec{k}^{\prime})^{*}\right\rangle=\frac{\delta_{\lambda\lambda^{\prime}}}{2}\frac{2\pi^{2}}{k^{3}}\mathcal{P}_{h}(k)(2\pi)^{3}\delta^{(3)}(\vec{k}-\vec{k}^{\prime})\,, (3) where \langle\dots\rangle denotes an ensemble average and * is complex conjugation. For any temporally-bounded source active on some characteristic sub-horizon scale k_{s} during radiation domination, \mathcal{P}_{h}(k)\propto k^{3} for all k\ll k_{s} 6 ; 14 , which reflects the causality requirement that sufficiently-separated spatial points must be uncorrelated. Thus, the power spectrum of any causal non-inflationary source must decay as white noise on super-horizon length scales. Given this, we parametrize the input power spectrum as \mathcal{P}_{h}(k)\equiv r_{\rm ect}{}A_{s}\left(\frac{k}{k_{\rm ref}}\right)^{3}\,, (4) where k_{\rm ref}=0.01 Mpc -1 is a reference wavenumber, A_{s}=2.1\times 10^{-9} is the measured amplitude of scalar fluctuations 25 , and r_{\rm ect} is the amplitude of the white-noise tensor power spectrum at k_{\rm ref} in units of A_{s} . This notation follows the long-standing convention to use the tensor-to-scalar ratio r=A_{T}/A_{s} as a measure of inflationary tensor modes. For the chosen k_{\rm ref}=0.01 Mpc -1 , the amplitude of the B -mode signals from inflationary and causal GWs approximately match at the recombination peak ( \ell\sim 100 ) when r=r_{\rm ect}{} . Figure 1 : Tensor power spectra \mathcal{P}_{h}(k) for select ECT models presented by 14 : a first-order cosmological phase transition (red), scalar-induced GWs (green), and cosmic strings (purple). The gray “linear