Abstract
Primordial non-Gaussianity induces a scale-dependent bias in large scale structure (LSS) data, proportional to $f_{\rm NL}/k^2$ for the exact local ansatz. Recent work has shown that models of inflation that predict a large squeezed limit bispectrum, such as multi-field models and single field inflation with a modified initial state, typically give rise to a generalized local ansatz, with the scale-dependent bias now proportional to ${\cal A}_{\rm NL}/k^\alpha$. We use photometric measurements of the angular power spectrum of luminous red galaxies and quasars in the Sloan Digital Sky Survey Data Release Eight (SDSS DR8) with the above parameterization to constrain the amplitude ${\cal A}_{\rm NL}$ and scale-dependence $\alpha$. We find that the marginalized upper limit on $\alpha$ is 2.0 at the $95\%$ confidence level, consistent with the local ansatz. We also present Fisher forecasts for a survey of the same size as DR8 to assess the role of systematics in current photometric LSS data. Moreover, we present analytic results on the expected mass dependence of ${\cal A}_{\rm NL}$ for different inflationary models, which can be an important observable for future surveys, if primordial non-Gaussianity is non-zero.
Introduction
The distribution of galaxies and quasars in the modern universe serves as a "fossil record" of the first fractions of a second after the Big Bang. In the standard cosmological model, the seeds of all cosmic structures were tiny quantum fluctuations generated during a period of rapid expansion known as inflation. While the simplest inflationary models predict these fluctuations follow a Gaussian distribution, more complex theories—involving multiple fields or modified initial states—predict primordial non-Gaussianity (PNG). Detecting PNG would provide a "smoking gun" for physics beyond the simplest inflationary scenarios.
The Signature of Non-Gaussianity
PNG leaves a specific signature in the Large Scale Structure (LSS) of the universe by inducing a scale-dependent bias in how dark matter halos (and the galaxies within them) cluster on very large scales. Previous studies often assumed the "local ansatz," where the bias is proportional to $k^{-2}$. This paper generalizes this form to $\Delta b \propto \mathcal{A}_{NL}/k^\alpha$, where $\mathcal{A}_{NL}$ is the amplitude and $\alpha$ is the scale-dependence. This generalization allows researchers to test whether the data prefers a different "shape" of non-Gaussianity, which would point to distinct inflationary mechanisms.
Data and Methodology
The study utilizes photometric samples from the Sloan Digital Sky Survey Data Release Eight (SDSS DR8), focusing on nearly a million Luminous Red Galaxies (LRGs) and Quasars. These objects are analyzed in several redshift "slices" to measure their angular power spectrum.
The researchers employed a Markov-Chain Monte-Carlo (MCMC) analysis, using a modified version of CosmoMC, to compare theoretical models against the observed data. This process simultaneously accounts for standard cosmological parameters, such as dark matter density, using baseline data from WMAP 9-year and Union 2 Supernova observations. Redshift distributions were calibrated using the Baryon Oscillation Spectroscopic Survey (BOSS).
The Squeezed Limit Diagnostic
A central focus of the paper is the "squeezed limit" of the bispectrum—a three-point correlation function. This limit occurs when one very long-wavelength fluctuation interacts with two short-wavelength ones. This specific interaction is the most effective way to distinguish between different inflationary models using LSS data. By measuring how the bias changes on the largest observable scales (small $k$), researchers can probe the physics of the "squeezed" configuration, which is highly sensitive to the presence of multiple fields during inflation.
Key Results and Constraints
The analysis confirms that current LSS data is perfectly consistent with the standard "local" form of non-Gaussianity. The marginalized upper limit on the scale-dependence parameter was found to be $\alpha \leq 2.0$ at the 95% confidence level.
For the standard local model ($\alpha=2$), the amplitude was constrained to $f_{NL} = 2^{+65}_{-66}$ (at 68% confidence). While this is a null result—meaning no clear detection of non-Gaussianity was found—it significantly narrows the parameter space for inflationary models.
The Impact of Systematics
A critical finding of this work, supported by Fisher forecasts, is that current constraints are limited by observational systematics rather than the volume of data. Factors such as galactic dust or variations in telescope sensitivity can mimic the signal of primordial non-Gaussianity. The forecasts suggest that to improve these limits, future efforts must focus on better understanding and mitigating these "nuisance" signals.
Conclusion and Future Outlook
The paper also derives new analytic predictions for the mass-dependence of $\mathcal{A}_{NL}$. The authors demonstrate that the amplitude of the non-Gaussian signal should vary depending on the mass of the observed objects (e.g., heavy galaxies versus lighter quasars). This mass-dependence provides a powerful diagnostic tool for upcoming surveys like DESI or Euclid. If these missions detect a non-zero signal, measuring its variation with galaxy mass will be key to identifying the specific physics of the inflationary epoch.
Summary updated