---
title: Primordial Curvature Power Spectrum
url: https://www.emergentmind.com/topics/primordial-curvature-power-spectrum
type: topic
---

# Primordial Curvature Power Spectrum

The primordial curvature power spectrum quantifies the two-point statistical properties of scalar curvature perturbations generated during cosmic inflation. It acts as the central observable linking quantum fluctuations of the inflaton to structure formation, cosmic microwave background (CMB) anisotropies, and a vast array of phenomenology across scales from kiloparsecs to sub-parsec domains. Its form, amplitude, and possible features are key in testing inflationary models, constraining early-Universe physics, and interpreting observational results from the CMB, large-scale structure, pulsar timing arrays (PTA), and primordial black hole (PBH) surveys.

## 1. Definition, Physical Origin, and Mathematical Formalism

The power spectrum of primordial curvature perturbations, denoted $\mathcal{P}_\mathcal{R}(k)$ or $P_\mathcal{R}(k)$, is defined as the Fourier-space two-point function of the gauge-invariant comoving curvature perturbation $\mathcal{R}$:
\[ \langle \mathcal{R}(\mathbf{k})\mathcal{R}(\mathbf{k}') \rangle = (2\pi)^3 \delta^3(\mathbf{k}+\mathbf{k}') \, \frac{2\pi^2}{k^3} \mathcal{P}_\mathcal{R}(k) \]
In canonical single-field slow-roll inflation, this is computed by quantizing the Mukhanov–Sasaki variable $v_k$ described by
\[ v_k'' + [k^2 - z''/z] v_k = 0 \]
with $z = a \sqrt{2\epsilon}M_\text{Pl}$, $a$ the scale factor, and $\epsilon$ the first slow-roll parameter. The Bunch–Davies initial vacuum is adopted, and after horizon exit the curvature perturbation $\mathcal{R}_k$ freezes out, yielding the celebrated slow-roll result:
\[ \mathcal{P}_\mathcal{R}(k) = \frac{H^2}{8\pi^2\epsilon M_\text{Pl}^2} \bigg|_{k=aH} \]
Observationally, this spectrum is tightly constrained near the pivot scale $k_* = 0.05\,\mathrm{Mpc}^{-1}$ with amplitude $A_s \simeq 2.1 \times 10^{-9}$ and spectral index $n_s \simeq 0.965$ [1302.2995, 1705.09924].

## 2. Extraction and Features from CMB and Large-Scale Structure

On cosmological scales ($10^{-4}\lesssim k \lesssim 3\,\mathrm{Mpc}^{-1}$), the power spectrum is reconstructed via a linear mapping from $\mathcal{P}_\mathcal{R}(k)$ to observables:
- CMB temperature/polarization: $C_\ell^{XY} = 4\pi \int \frac{dk}{k} \,\mathcal{P}_\mathcal{R}(k)\Delta_\ell^X(k)\Delta_\ell^Y(k)$
- Galaxy power spectra: $P_\text{gal}(k,z) = b^2(z)T_m^2(k,z)\frac{2\pi^2}{k^3}\mathcal{P}_\mathcal{R}(k)$

Regularized non-parametric inversion (Tikhonov, GCV, Mallows $C_p$) reveals robust features:
- Infrared cutoff at $k \lesssim 5\times10^{-4}\,\mathrm{Mpc}^{-1}$, corresponding to suppressed low multipoles (quadrupole, octupole) in the CMB
- Small-scale oscillations or dips with marginal $\sim$2–3$\sigma$ significance at $k\sim0.002$, $0.04$, and $0.056\,\mathrm{Mpc}^{-1}$ [1308.2317, 1510.03338].
Below $k\sim0.1\,\mathrm{Mpc}^{-1}$, the spectrum agrees with a near–scale-invariant power law [1705.09924], and no high-significance features are confirmed.

## 3. Theoretical Impact of Spatial Curvature

Inflation in universes with nonzero spatial curvature ($K=\pm1$) modifies both the background evolution and the Mukhanov–Sasaki equation:
\[
v_k'' + [k^2+K-z''/z]v_k=0
\]
with Laplace–Beltrami eigenvalues $k^2+K$ ($k\in\mathbb{Z}$ for $K=+1$). Curvature induces:
- Infrared cut-off in $\mathcal{P}_\mathcal{R}(k)$ for $k\lesssim\sqrt{|K|}$ 
- Superimposed oscillatory features for $k \sim (2–5)\sqrt{|K|}$
- Asymptotic recovery of the flat-universe power law for $k \gg \sqrt{|K|}$ [1907.08524, 2009.05573]
For Planck 2018 best-fit $\Omega_K\approx-0.009$, the cut-off is at $k_\text{cut}\approx 0.002\,\mathrm{Mpc}^{-1}$, corresponding to the observed low-$\ell$ suppression. The curved-universe template improves $\chi^2$ fit by $\Delta\chi^2\approx -6$ to $-9$ compared to standard $\Lambda$CDM [1907.08524].

## 4. Constraints from Primordial Black Hole Production and Evaporation

PBH non-detections and Hawking evaporation yields direct constraints on $\mathcal{P}_\mathcal{R}(k)$ at very small scales ($k \gtrsim 10^{15}\,\mathrm{Mpc}^{-1}$). The collapse fraction $\beta(M)$ is exponentially sensitive to the mass variance $\sigma^2(M)$, related (in radiation domination) by
\[
\sigma^2(M) = \int \frac{dk}{k} W^2(kR) \frac{16}{81}(kR)^4 \mathcal{P}_\mathcal{R}(k)
\]
Non-detection of sub-solar-mass PBHs in BBN and CMB imposes bounds:
- $k\sim 3.7\times10^{17}\,k_{0.05}$: $\mathcal{P}_\mathcal{R}(k) \lesssim 3\times10^{-3}$
- $k\sim7.4\times10^{18}\,k_{0.05}$: $\mathcal{P}_\mathcal{R}(k) \lesssim 4\times10^{-3}$
Evaporating PBH relic and neutrino/γ-ray backgrounds further constrain $\mathcal{P}_\mathcal{R}(k) \lesssim 2\times10^{-2}$ for $4.5\times10^{18}\lesssim k\lesssim 1.8\times10^{21}\,\mathrm{Mpc}^{-1}$ [1812.09807, 2411.18887, 1904.10971]. These bounds fill gaps in previously unconstrained windows and eliminate broad blue-tilted enhancements or features on ultra-small scales.

## 5. Model-Independent Reconstruction in PTA and GW Observations

PTA collaborations (NANOGrav, PPTA, EPTA, CPTA) have detected a common-spectrum stochastic GW signal, interpretable as scalar-induced GWs sourced by enhanced $\mathcal{P}_\mathcal{R}(k)$ at $k\sim 10^7\!-\!10^9\,\mathrm{Mpc}^{-1}$. Bayesian model-independent reconstruction approaches (linear interpolation over free nodes, polynomial, lognormal, broken power-law) consistently favor:
- Broad or narrow peaks in $\mathcal{P}_\mathcal{R}(k)$ centered at $k_p \sim 10^6\!-\!10^9\,\mathrm{Mpc}^{-1}$
- Peak amplitudes $\mathcal{P}_\mathcal{R}(k_p)\sim 2\times10^{-2}$, seven orders above CMB normalization
- No current discrimination between single-peak, double-peak, broken power law, lognormal, or box shapes (Bayes factors $<1$ between models)
Best-fit reconstructions are consistent with BBN, μ-distortion, and PBH constraints, requiring the peak to be narrow ($\Delta\ln k < 1$) to avoid overproduction of PBHs [2310.17199, 2308.05632, 2307.04419, 2407.15501].

| Source              | $k$-range [$\mathrm{Mpc}^{-1}$] | Bound or Detection ($\mathcal{P}_\mathcal{R}(k)$) |
|---------------------|------------------------------|-------------------------------------------|
| CMB (Planck)        | $10^{-4}$–$3$                | $2.1\times10^{-9}$                        |
| PTA (NANOGrav)      | $10^6$–$10^9$                | $\sim10^{-2}$ (peak)                      |
| PBH evaporation     | $10^{15}$–$10^{20}$          | $\lesssim 2\times10^{-2}$                 |

## 6. Theoretical Mechanisms for Small-Scale Enhancements

Transient features in the inflationary potential (inflection points, ultra-slow-roll phases) or nontrivial sound speed evolution can produce large peaks or steep growth in $\mathcal{P}_\mathcal{R}(k)$:
- Ultra-slow-roll inflation ($\eta \simeq -6$): $k^4$ growth
- Sudden drop in sound speed: $k^2$ rise, further steepening with simultaneous $\eta$ and $c_s$ changes, realizing $k^6$ envelope [2308.09286]
- All viable PBH-producing models must generate a narrow enhancement with $\mathcal{P}_\mathcal{R}(k_p)\lesssim 10^{-2}$, width $\Delta(\ln k)\lesssim1$, and slopes $\alpha,\beta\lesssim 4$ in broken power-law templates [2210.03641, 2212.05516].

## 7. Implications, Open Questions, and Outlook

The aggregate constraints demand near–scale invariance at cosmological scales, with only tightly localized peaks at small scales. Any inflationary scenario with broad or persistent blue tilts, slow-roll violation, or non-Gaussian statistics is subject to strong bounds from PBHs and spectral distortions. Detection of a cutoff or oscillatory features from spatial curvature would point toward “just-enough” inflation and nonflat initial conditions [1907.08524, 2009.05573]. Current and future GW/PTA experiments promise to decisively probe the relevant $\mathcal{P}_\mathcal{R}(k)$ window, test the origin of BBH mergers, and further elucidate inflationary microphysics on scales previously inaccessible to cosmology.

Source: https://www.emergentmind.com/topics/primordial-curvature-power-spectrum