SPT-3G D1 Confirms 6.2σ Hubble Tension: A High-Resolution CMB Power Spectrum Analysis

Introduction and Background

The SPT-3G Instrument and D1 Main Field Survey

1. Focal Plane Architecture and SQUID Multiplexing

   The third-generation camera on the 10-meter South Pole Telescope (SPT-3G) represents a monumental leap in bolometric detector technology. The focal plane is populated with approximately 16,000 transition-edge sensor (TES) bolometers, arranged into an array of trichroic, horn-coupled pixels [cite:103]. Each individual pixel simultaneously couples to three distinct frequency bands—95, 150, and 220 GHz—enabling the exquisite multi-frequency foreground separation required for precision cosmology. Operating at sub-Kelvin temperatures, the massive scale of the TES array introduces significant cryogenic readout challenges. To manage the thermal load and wire count, the SPT-3G instrument employs a frequency-domain multiplexing (fMUX) architecture [cite:104]. Specifically, a 68× fMUX scheme utilizes Superconducting Quantum Interference Devices (SQUIDs) to read out 68 distinct bolometer channels on a single pair of coaxial cables. This extreme multiplexing factor dramatically reduces parasitic heat leaks while preserving the low-noise environment essential for measuring microkelvin-level CMB fluctuations.

2. The 2019-2020 D1 Main Field Dataset

   The observational foundation of this study relies on the D1 "Main field" dataset, acquired during the 2019 and 2020 austral winters. Situated at the geographic South Pole, the telescope benefits from exceptionally stable atmospheric conditions, allowing for continuous, high-efficiency integration over the targeted 1500-square-degree patch of the southern sky [cite:105]. Although this field constitutes roughly 4% of the total sky area, the sheer integration time and detector density yield the deepest CMB TT, TE, and EE maps ever produced. The dual-season scanning strategy was designed to rigorously cross-link the maps, suppressing large-scale atmospheric noise and ensuring high-fidelity signal recovery down to arcminute resolutions. The resulting high-signal-to-noise ratio in the polarization channels is the primary driver behind the improved constraints on the primordial scalar spectral index and the physical baryon density, effectively unlocking the high-multipole damping tail for precision parameter estimation.

Power Spectrum Analysis and Pipeline

1. Formalism and Silk Damping

   The core of the cosmological parameter extraction lies in the estimation of the angular power spectra, C_ℓ, from the raw time-ordered data. The Boltzmann-acoustic formalism dictates that the observed anisotropies are a projection of the primordial density perturbations as they evolve through the pre-recombination plasma [cite:106]. The characteristic angular scale of the acoustic peaks, θ★, is determined by the ratio of the comoving sound horizon at recombination, r_s, to the comoving angular diameter distance, D_A.

   θ★ = r_s / D_A = ∫_z★^∞ (c_s(z) / H(z)) dz ÷ ∫_0^z★ (c / H(z)) dz

   At high multipoles (ℓ > 1000), the acoustic oscillations are exponentially suppressed by Silk damping, caused by the imperfect coupling between photons and baryons. The precise mapping of this damping envelope allows cosmologists to break critical degeneracies that natively plague measurements restricted to larger angular scales. Specifically, isolating the high-multipole behavior of the primordial density field strongly constrains the physical baryon density and the primordial scalar spectral index [cite:111]. By executing this observation with the multi-frequency capability of the D1 focal plane, the SPT-3G pipeline effectively neutralizes foreground contamination that otherwise obscures the deep Silk damping tail, yielding parameter constraints that rival the full-mission Planck dataset.

2. EE and TE Precision at High Multipoles

   Following the rigorous methodology outlined by Camphuis et al. (2025), the data analysis pipeline applies pseudo-C_ℓ estimators to the cross-spectra of independent map bundles [cite:107]. This technique explicitly avoids auto-spectra, thereby nullifying the contribution of uncorrelated instrumental noise. The distinguishing feature of the 2019–2020 analysis is the unprecedented precision achieved in the E-mode auto-power (EE) and temperature-E-mode cross-power (TE) spectra across the multipole range of ℓ = 1800 to 4000. In this regime, galactic foregrounds such as thermal dust and synchrotron emission are meticulously modeled and subtracted utilizing the instrument's trichroic capabilities. The resulting bandpowers are sample-variance limited in TT up to ℓ ≈ 3000, and critically, the EE and TE spectra exhibit statistical uncertainties up to a factor of two smaller than previous generation experiments. This high-ℓ precision is precisely what anchors the subsequent cosmological parameter fits, rendering them highly resistant to large-scale anomalies.

Systematics and Null Tests

Cosmological Parameters and the Hubble Tension

1. The 6.2σ Hubble Tension Confirmation

   The most profound implication of the D1 Main field analysis is its impact on the Hubble tension. By feeding the pristine high-ℓ TE and EE bandpowers into the standard ΛCDM Markov Chain Monte Carlo framework, Camphuis et al. (2025) derived an extraordinarily precise measurement of the present-day expansion rate: H₀ = 66.66 ± 0.60 km/s/Mpc [cite:109]. This measurement is completely independent of the Planck satellite data, relying solely on the physics of the acoustic peaks and the Silk damping tail observed by SPT-3G. When contrasted with the definitive local distance ladder measurement from the SH0ES collaboration (Riess et al., 2022), which reports H₀ = 73.04 ± 1.04 km/s/Mpc, the discrepancy balloons to a staggering 6.2σ significance. This level of statistical divergence effectively eliminates systematic errors within a single experiment as a viable explanation, cementing the Hubble tension as a genuine physical phenomenon requiring new theoretical frameworks.

2. Combined CMB Constraints and DESI DR2 Comparisons

   To fully exploit the available cosmological information, the SPT-3G data was subsequently combined with the latest releases from the Atacama Cosmology Telescope (ACT) and the Planck legacy maps. This joint CMB likelihood yields a consensus early-Universe Hubble constant of H₀ = 67.19 ± 0.38 km/s/Mpc, alongside a tightly constrained amplitude of matter fluctuations, σ₈ = 0.8137 [cite:110]. While this solidifies the early-versus-late expansion rate dichotomy, it also illuminates a secondary, emerging discrepancy. When comparing the joint CMB predictions for structure growth against the recent DESI DR2 (Dark Energy Spectroscopic Instrument Data Release 2) baryon acoustic oscillation measurements, a 2.8σ tension materializes in the matter density and growth rate plane.

   S_8 = σ₈ (Ω_m / 0.3)^0.5 ≈ 0.832 ± 0.011

   The DESI DR2 data favors a slightly lower matter density and suppressed structure growth compared to the CMB extrapolation. This secondary tension suggests that modifications to ΛCDM designed to resolve the H₀ crisis, such as Early Dark Energy or self-interacting neutrinos, must simultaneously navigate the strict bounds imposed by the DESI large-scale structure clustering data.

Conclusion