Cosmic Microwave Background Anisotropies with Mixed Isocurvature Perturbations

Roberto Trotta, Alain Riazuelo, Ruth DurrerPhysical 2001 · DOI · arXiv

This paper, published in 2001 during the "golden age" of Cosmic Microwave Background (CMB) physics, serves as a critical cautionary tale about the assumptions underlying precision cosmology. While the community was celebrating the first high-resolution maps of the early Universe, Trotta et al. asked a fundamental question: Are we measuring the Universe's properties, or are we just seeing the reflection of our own assumptions?

TL;DR

The authors demonstrate that the "precise" cosmological parameters derived from CMB data (like the Hubble constant and baryon density) depend heavily on the assumption of purely adiabatic initial conditions. By allowing for "mixed" isocurvature perturbations, the allowed range for these parameters expands dramatically, suggesting that early-2000s CMB data alone couldn't uniquely determine the composition of our Universe without making strong theoretical assumptions about inflation.

What Problem Does This Paper Address?

In the early 2000s, experiments like BOOMERanG and MAXIMA provided the first clear look at the "acoustic peaks" in the CMB—the ringing of the early Universe's plasma. To turn these bumps and wiggles into numbers like the expansion rate ($h$) or the density of atoms ($\omega_b$), researchers typically assumed adiabatic initial conditions. This means that in the very early Universe, all components (photons, dark matter, neutrinos, etc.) fluctuated together in density.

However, many theoretical models of the early Universe (like multi-field inflation) allow for isocurvature perturbations, where different components fluctuate relative to one another without changing the total energy density. The authors wanted to see if the "precision" of cosmology was an illusion created by ignoring these isocurvature possibilities.

Methods

The authors moved beyond the standard single-mode model to a more generic framework:

  • The Correlation Matrix: They used a $5 \times 5$ positive-definite matrix ($M$) to represent the amplitudes and cross-correlations of five possible perturbation modes: Adiabatic (AD), Baryon Isocurvature (BI), CDM Isocurvature (CI), Neutrino Isocurvature Density (NID), and Neutrino Isocurvature Velocity (NIV).
  • Parameter Search: They varied the physical parameters ($h$ and $\omega_b$) alongside the 10 independent components of the correlation matrix.
  • Optimization: Because the 14-dimensional parameter space is riddled with local minima, they used a "downhill simplex" method initiated from 15,000 random starting points to ensure they found the global best fit.
  • Marginalization: To see the true uncertainty in physical parameters, they "marginalized" (integrated) over all possible initial conditions to see what range of $h$ and $\omega_b$ remained viable.

Datasets

The study utilized the best available data of the era:

  • COBE: For large-scale anisotropies.
  • BOOMERanG: For the high-resolution data of the first and second acoustic peaks.
  • The authors also mention DASI and MAXIMA-1 as part of the broader context of high-precision CMB data that motivated the study.

Key Insights and Ideas

  • The Degeneracy Problem: Isocurvature modes can shift the positions and heights of the acoustic peaks in ways that mimic changes in physical parameters. For instance, a neutrino isocurvature mode can shift peaks to the right, potentially masking a different value for the Hubble constant.
  • The "Aha" Moment: Figure 2 in the paper is the "smoking gun." In the purely adiabatic case, the likelihood contours for $h$ and $\omega_b$ are tight, small circles. Once mixed perturbations are allowed, these contours "open up" and cover almost the entire parameter space.
  • Baryon Density Freedom: Under adiabatic assumptions, high baryon density is ruled out because it creates a massive asymmetry between even and odd acoustic peaks. However, the authors found that isocurvature modes can "cancel out" this asymmetry, making high-baryon universes perfectly consistent with the data.

Key Results

  • Dramatic Uncertainty: The allowed range for cosmological parameters "enlarges dramatically" when isocurvature is considered. For example, while adiabatic models favored $\omega_b \approx 0.02$, the mixed model could easily accommodate $\omega_b = 0.042$ with a better fit ($\chi^2 = 0.77$ vs $4.32$).
  • Best Fit Shifts: The "best fit" values for a mixed model are often significantly different from the adiabatic ones, suggesting that our "standard model" of cosmology is highly sensitive to the physics of the first $10^{-35}$ seconds of the Universe.
  • Residual Constraints: Despite the massive increase in uncertainty, one region remained excluded: the upper-left corner of the $(h, \omega_b)$ plane (high expansion rate, low baryon density). This is because even with isocurvature modes, there wasn't enough power in the secondary peak region to fit that specific combination.

Limitations and Future Work

  • Fixed Parameters: To keep the computation manageable, the authors fixed the total density ($\Omega_{tot} = 1$) and the cosmological constant ($\Omega_\Lambda = 0.7$). Varying these would likely have expanded the uncertainty even further.
  • Breaking the Degeneracy: The authors conclude that CMB temperature data alone is insufficient. They point toward CMB polarization as the key "tie-breaker." Polarization is sensitive to the photon velocity (dipole) rather than just density, which responds differently to adiabatic vs. isocurvature starts.
  • Complementary Data: They suggest that combining CMB data with the matter power spectrum (from galaxy surveys) would help break these degeneracies, as isocurvature modes evolve differently over cosmic time.

Summary updated

Abstract

In the light of the recent high quality data of the cosmic microwave background anisotropies, several estimations of cosmological parameters have been published. We study to what extent these estimations depend on assumptions about the initial conditions of the cosmological perturbations, which are usually supposed to be adiabatic. We show that, for more generic initial conditions, not only the best fit values are very different but the allowed parameter range enlarges dramatically. This raises the question which cosmological information (matter content of the Universe vs physics of inflation) can be reliably extracted from these data.
BibTeX
@inproceedings{trotta2001cosmic,
  author = {Roberto Trotta and Alain Riazuelo and Ruth Durrer},
  title = {{Cosmic Microwave Background Anisotropies with Mixed Isocurvature Perturbations}},
  booktitle = {Physical},
  year = {2001},
  doi = {10.1103/physrevlett.87.231301},
}
Figure