Measurements of Secondary Cosmic Microwave Background Anisotropies with the South Pole Telescope

TL;DR

This study presents the first CMB temperature power spectra from the South Pole Telescope for a 100 deg$^2$ field at 150 and 220 GHz, extending into the damping tail and toward the SZ-dominated regime. Using a multi-frequency, DSFG-subtracted analysis and extensive simulations, the authors detect SZ power at $\ell\approx3000$ with $4.2\pm1.5\ \mu$K$^2$, significantly lower than fiducial WMAP5-based predictions, suggesting either lower $\sigma_8$ or overestimated intracluster gas pressure in models. An MCMC framework incorporating the primary CMB, tSZ, kSZ, and DSFG components yields a constrained $\sigma_8$ around $0.773\pm0.025$ (with substantial theory uncertainty on SZ amplitudes), highlighting the dominant role of modeling uncertainties in SZ-based cosmology. The analysis demonstrates effective separation of DSFG foregrounds via frequency differencing and sets the stage for larger, multi-band SPT surveys to improve SZ/kSZ measurements and probe reionization history. Overall, the work provides an independent, small-scale CMB test of structure growth and motivates refined astrophysical modeling of intracluster gas and dusty galaxy populations.

Abstract

We report cosmic microwave background (CMB) power spectrum measurements from the first 100 sq. deg. field observed by the South Pole Telescope (SPT) at 150 and 220 GHz. On angular scales where the primary CMB anisotropy is dominant, ell ~< 3000, the SPT power spectrum is consistent with the standard LambdaCDM cosmology. On smaller scales, we see strong evidence for a point source contribution, consistent with a population of dusty, star-forming galaxies. After we mask bright point sources, anisotropy power on angular scales of 3000 < ell < 9500 is detected with a signal-to-noise > 50 at both frequencies. We combine the 150 and 220 GHz data to remove the majority of the point source power, and use the point source subtracted spectrum to detect Sunyaev-Zel'dovich (SZ) power at 2.6 sigma. At ell=3000, the SZ power in the subtracted bandpowers is 4.2 +/- 1.5 uK^2, which is significantly lower than the power predicted by a fiducial model using WMAP5 cosmological parameters. This discrepancy may suggest that contemporary galaxy cluster models overestimate the thermal pressure of intracluster gas. Alternatively, this result can be interpreted as evidence for lower values of sigma8. When combined with an estimate of the kinetic SZ contribution, the measured SZ amplitude shifts sigma8 from the primary CMB anisotropy derived constraint of 0.794 +/- 0.028 down to 0.773 +/- 0.025. The uncertainty in the constraint on sigma8 from this analysis is dominated by uncertainties in the theoretical modeling required to predict the amplitude of the SZ power spectrum for a given set of cosmological parameters.

Figures (11)

• Figure 1: Average beam functions and uncertainties for SPT. Left axis: The SPT beam function for $150\,$GHz ( red) and $220\,$GHz ( blue). Right axis: The $1\,\sigma$ uncertainties on the beam function for each frequency. The beam uncertainties shown here include only uncertainties on the main lobe beam width, $\sigma_b$, since the uncertainty of the sidelobe amplitude has been subsumed into the calibration uncertainty.
• Figure 2: Jack-knives for the SPT data set at $150\,$GHz ( blue circles) and $220\,$GHz ( black diamonds). For clarity, the $220\,$GHz jack-knives have been shifted to the right by $\Delta\ell = 100$. Top panel: Bandpowers of the "first half - second half" jack-knife compared to the expected error bars about zero signal. Disagreement with zero would indicate either a noise misestimate or a time-dependent systematic signal. Second panel: Power spectrum of the left-going minus right-going difference map. This test yields strong constraints on the accuracy of the detector transfer function deconvolution and on possible directional systematics. Third panel: Bandpowers for the difference map when the data is split based on azimuth. Signals fixed in azimuth such as side-lobe pickup from the nearby support building would produce non-zero power. We see no evidence for ground-based pickup. The cumulative probability to exceed the $\chi^2$ observed in these three tests is 76% at 150 GHz and 22% at 220 GHz. Bottom panel: The un-differenced SPT power spectra at each frequency for comparison.
• Figure 3: The SPT $150\,$GHz ( purple circles), $150\times220\,$GHz ( green diamonds) and $220\,$GHz ( blue triangles) bandpowers. The damping tail of the primary CMB anisotropy is apparent below $\ell = 3000$. Above $\ell = 3000$, there is a clear excess with an angular dependence consistent with point sources. These sources have low flux (as sources with $>6.4\,$mJy at $150\,$GHz have been masked) and a rising frequency spectrum, consistent with our expectations for Poisson distributed DSFGs. The point source population and resulting contributions to anisotropy power are discussed in more detail in H09.
• Figure 4: WMAP5 ( blue squares), ACBAR ( green triangles), QUaD ( turquoise diamonds) and the SPT ( black circles) DSFG-subtracted SPT bandpowers are plotted over the best-fit, lensed $\Lambda$CDM cosmological model ( dashed red line), best-fit tSZ power spectrum ( solid black line), homogeneous kSZ model ( dashed black line), and residual Poisson-distributed point source contribution ( solid orange line). The combined best-fit model is shown by the solid red line. The plotted SPT bandpowers have been multiplied by the best-fit calibration factor of 0.92. Point sources above $6.4\,$mJy at $150\,$GHz have been masked. The patchy kSZ template is also shown for reference ( dotted black line). The DSFG-subtracted bandpowers are normalized to preserve the amplitude of the primary CMB anisotropies.
• Figure 5: The SPT $150\,$GHz ( purple diamonds) and DSFG-subtracted ( black circles) bandpowers over-plotted on the best-fit models to the DSFG-subtracted bandpowers. The best-fit, lensed $\Lambda$CDM cosmological model for the primary CMB anisotropy is shown by the dashed red line, while the sum of the best-fit $\Lambda$CDM model, kSZ, tSZ and point source terms is shown by the solid red line. The primary CMB anisotropy alone is a poor fit to the SPT data. The uncertainties on the DSFG-subtracted bandpowers are larger for two reasons. First, the normalization convention inflates the uncertainties by a factor of $1/0.675^2$, and second, these bandpowers also include the more noisy $220\,$GHz data. Beam and calibration uncertainties are marked by a second blue error bar for the DSFG-subtracted bandpowers only. Note that the calibration and beam uncertainties are correlated between $\ell$-bins. The $150\,$GHz data has been shifted to the right by $\Delta\ell = 40$ for clarity. Point sources above $6.4\,$mJy at $150\,$GHz have been masked in this analysis.