Effects of primordial magnetic fields on 21 cm multifrequency angular power spectra

This research investigates how primordial magnetic fields, present before decoupling, influence the 21 cm line signal and its resulting multifrequency angular power spectra (MAPS). By simulating these effects in cosmological models, the study explores a new avenue to constrain parameters of primordial magnetic fields through observations from instruments like uGMRT, MeerKAT, and SKA1-MID.

Theoretical Framework for Magnetic Field Effects

Primordial magnetic fields are modeled as stochastic with different characteristics depending on their origin (e.g., during inflation or phase transitions). These fields can influence the 21 cm line signal through two primary mechanisms:

1. Structure formation enhancement on small scales by increasing the amplitude of the total linear matter power spectrum due to the Lorentz term in the baryon velocity equation, which is defined by a term like (2.1) involving LB = 2/3 (1/3ΠB − ∆B).

2. Magnetic field dissipation in the post-recombination universe via decaying MHD turbulence and ambipolar diffusion, which can lead to additional heating of matter.

The study focuses on the effect of primordial magnetic fields on the linear matter power spectrum and subsequently on 21 cm intensity maps. The models considered include:

: the adiabatic, primordial curvature mode as well as the compensated magnetic mode.

Methodology for Generating MAPS

The core methodology involves simulating temperature maps using modified linear matter power spectra incorporating magnetic fields with specific parameters (e.g., B0 = 4 nG and spectral indices nB = −2.9 and nB = −2.5). These simulated maps are then used as initial matter power spectra in the crime numerical code to obtain the 21 cm brightness temperature maps.

The resulting multifrequency angular power spectra (MAPS) are calculated by cross-correlating these 21 cm signal maps at different frequencies. The frequency dependence allows for studying the evolution of the signal with varying frequency or redshift due to the light-cone effect, which is particularly important during reionization. The MAPS are defined as:

: Cl(ν1, ν2) = halm(ν1)a∗lm(ν2) (3.3).

Observational Prospects and Signal-to-Noise Ratios (SNR)

The prospects for constraining magnetic field parameters are estimated using signal-over-noise ratios, calculated by taking into account only system noise after assuming perfect foreground removal. The SNR is estimated using the noise power spectrum given by equation (3.6).

The study presents numerical results for various observational setups:

1. uGMRT Band 3 (central frequency νc = 432.8 MHz, redshift z = 2.28).

2. MeerKAT L band configuration (1) (central frequency νc = 986 MHz, redshift z = 0.44).

3. MeerKAT L band configuration (2) (central frequency νc = 1077.5 MHz, redshift z = 0.32).

4. The prospective SKA1-MID Wide Band 1 Survey at the corresponding central frequencies for uGMRT, MeerKAT L-band (1), and (2).

The results show that Signal-over-noise ratios larger than one can be obtained even for larger frequency separations for larger multipoles in the future SKA1-MID survey.

Key Findings on MAPS Evolution

A significant finding is that "a significant decrease in amplitude for ∆ν > 1 MHz can be observed in figure 1 from the multifrequency angular power spectra Cl(νc,conf, ∆ν)." This indicates that the frequency cross-correlated 21 cm signal decorrelates rapidly with growing differences between frequencies, ∆ν. The rate of this decorrelation depends on the multipole (l), being slower on larger scales (smaller l).

Conclusion and Future Work

The results from this first study seem to be promising to constrain parameters of a primordial magnetic field with multifrequency angular power spectra using 21 cm line intensity mapping observations. However, future work is needed to include additional aspects such as the effect of magnetic field decay via MHD simulations, use multifrequency cross correlations to probe evolution, and perform a full numerical cosmological parameter estimation. This requires improved foreground modeling and more precise estimates of telescope sensitivities.

Signal-over-Noise Ratio Enhancement

The cumulative signal-over-noise ratio (cum S/N) analysis shows that for the MeerKAT L band configurations, the cumulative S/N becomes larger than 1 for growing multipole for magnetic spectral indices nB = −2.9 and nB = −2.5 under current specifications. For the proposed SKA1-MID Wide Band 1 Survey, "at least the auto correlation power spectra of the 21 cm line signal, i.e.

As a fastidious and diligent researcher, I have analyzed this paper, Effects of primordial magnetic fields on 21 cm multifrequency angular power spectra, focusing on its methodology, theoretical framework, and observational prospects.

The following improvements are recommended for AI systems (specifically those involved in cosmological parameter estimation, large-scale structure analysis, and signal processing) based on the capabilities outlined or implied by this research:

) 1. Implementation of Magnetic Field Parameter Sensitivity Analysis within Simulation Pipelines:

The paper explicitly shows how different magnetic field parameters (e.g., B0 = 4 nG, spectral indices nB = −2.9 and nB = −2.5) lead to distinct Multi-Frequency Angular Power Spectra (MAPS).

• Specific Improvement: Develop AI/ML algorithms capable of performing rapid, high-dimensional sensitivity mapping over the parameter space of primordial magnetic fields (amplitude B0 and spectral index nB). The AI should be trained on the theoretical MAPS outputs (Equation 3.5) to predict the fractional change in MAPS between different models.

• What it can do: This improved system can rapidly determine which regions of the primordial magnetic field parameter space are most constrained by future observations, effectively acting as an automated parameter prior selection tool before expensive full numerical simulations are run.

) 2. Advanced Foreground Mitigation and Map Reconstruction using Blind/Non-Blind Techniques:

The paper emphasizes that foreground removal is crucial, suggesting the use of techniques like Principal Component Analysis (PCA), Independent Component Analysis (ICA), and line-of-sight polynomial fitting.

• Specific Improvement: Implement deep learning architectures (e.g., Variational Autoencoders or Generative Adversarial Networks) specifically tailored for blind foreground removal in 21 cm intensity maps, moving beyond traditional linear methods described in [26]. The AI should be trained on simulated maps contaminated with known galactic and extragalactic synchrotron/free-free emissions to learn complex spectral and spatial correlations.

• What it can do: This system can produce cleaner 21 cm intensity maps from observational data (like MeerKAT or SKA) faster and more robustly than current pipelines, directly increasing the achievable Signal-to-Noise Ratio (SNR) in the final MAPS calculation.

) 3. Automated Cross-Correlation and MAPS Calculation Engine:

The core of the study involves cross-correlating temperature maps at different frequencies to derive MAPS, a process that is computationally intensive and relies on accurate redshift estimation.

• Specific Improvement: Build an AI module optimized for the efficient calculation of Equation (3.3) and (3.4), specifically handling the frequency dependence and redshift evolution along the Line-of-Sight (LOS). This module should incorporate learned priors about the expected correlation structure based on cosmological models (like Planck 2018 best fits).

• What it can do: It can rapidly generate MAPS for a vast grid of hypothetical observational configurations (different telescope arrays, different bandwidths, and various magnetic field models) in seconds, drastically accelerating the exploration of parameter space.

) 4. Predictive Signal-to-Noise Ratio (SNR) Forecasting:

The paper provides detailed simulations showing the cumulative SNR (Figure 3), indicating that for future surveys like SKA1-MID, S/N can exceed unity even for non-zero frequency separations at larger multipoles.

• Specific Improvement: Develop a predictive model that estimates the achievable cumulative SNR, incorporating not just instrument noise (Equation 3.6) but also learned effects from improved foreground modeling and mode filtering techniques (as suggested in the conclusion). This model should be trained on simulations of various observational strategies.

• What it can do: This system will provide a realistic feasibility map for future telescope proposals, telling researchers exactly which combination of observing time, bandwidth, and frequency separation is required to achieve a specific constraint on the magnetic field parameters.

) 5. Integration of Non-Linear Effects into Linear Power Spectrum Modeling:

The paper notes that primordial magnetic fields primarily enhance the linear matter power spectrum on small scales via the Lorentz term (Equation 2.1).

• Specific Improvement: Develop an AI layer that can intelligently incorporate higher-order, non-linear corrections to the matter power spectrum, specifically those induced by magnetic field effects (like the term proportional to k 4/L B(k)), into the initial conditions used for 21 cm intensity map simulations.

• What it can do: This allows AI systems to move beyond purely linear approximations and provide more accurate predictions of the small-scale power spectrum, which is where magnetic field effects are most pronounced, leading to tighter constraints on B0 and nB.