Observations of the sky at millimeter wavelengths are of critical importance to modern cosmology. Measurements of the primary cosmic microwave background (CMB) anisotropies, its secondary anisotropies, and foreground objects all contribute to our understanding the universe. As the oldest light in the universe, the CMB contains imprints from much of cosmic history. Measurements of these secondary anisotropies provides a window into the evolution of the universe, including the growth of the largest structures in the universe, and the process of reionization by early stars and galaxies.

These measurements are only possible with a high-resolution telescope and extremely precise calibration. This dissertation is primarily concerned with the design and calibration of the SPT-3G camera on the South Pole Telescope. In particular, the calibration of the raw detector data, and the measurement of the telescope beams are described in detail. These pieces are essential for making a precise measurement of the mm-wave sky, and properly accounting for the effects of the telescope optics. In the last chapter, we show how all of this comes together to produce a catalog of galaxy clusters discovered through the thermal Sunyaev Zel’dovich effect. We report 89 candidates, with 81 confirmed by optical and infrared follow-up observations. 29 are reported for the first time. This provides a mass-limited catalog of galaxy clusters over 100 square degrees.

Fast radio bursts (FRBs) are bright, millisecond flashes of radio radiation detected at frequencies between 800 MHz and 8 GHz. We present the first rate constraints on millisecond duration celestial transients like FRBs at 150 GHz. Using a year of data from the South Pole Telescope we searched for signals with a similar phenomenology at 150 GHz and found no evidence for a celestial population. The 0.9 upper confidence limit is $1.1 \times 10^{5} \text{sky}^{-1}\text{ day}^{-1}$ for 1 ms signals with fluence above 10 Jy ms. The search sensitivity is expected to increase dramatically with future microwave telescopes due to increases in spectral information and a reduced background rate.

The first four chapters are dedicated to detailing how we convert a measured optical signal into a measurement of the power spectrum of the CMB. Particular emphasis is placed on the transition edge sensor bolometer readout, the map making and the power spectrum estimation from these maps. Chapter five focuses on how we can analyze the data to look for fast radio bursts.

The Cosmic Microwave Background, the oldest light in the universe, provides a powerful tool to test the theory of inflation and constrain the standard ΛCDM cosmological model. The theory of inflation describes a period of rapid, exponential growth in the early Universe. A measurement of r, the tensor-to-scalar ratio of CMB polarization, would provide evidence for inflation and measure its energy scale. However, this is an extremely difficult measurement to make due to the faintness of the signal. Historically, Small Aperture Telescopes (SATs) have been used to constraint inflation. However, there is immense power that can be gained by combining the low-ℓ (large angular scale) measurements from Large Aperture Telescopes (LATs), such as the South Pole Telescope (SPT), with SATs.

The first part of this dissertation describes the design, characterization, and deployment of the SPT-3G receiver on the SPT. There is special emphasis placed on the development and characterization of the frequency-multiplexed readout, and characterization of system crosstalk. The second part of this dissertation uses data from the 2019 and 2020 austral observing season over a 1500 deg² patch of sky to make low-noise maps of the B-mode polarization of the CMB. This delves into techniques used for mitigating low-ℓ noise sources, including polarized atmosphere. These include weighting schemes, and an unbiased method to remove this noise source from data utilizing co-pointing detectors of different frequencies and the fixed spectral scaling of polarized atmosphere in accordance with Rayleigh scattering. This work serves both as demonstration of a LAT achieving high sensitivity at low-ℓ and as a study of atmospheric noise that will inform the design and operation of future LAT and SAT telescopes. This shows a path for producing competitive r constraints from a LAT optimized for high-ℓ science, and is a proof-of-concept for using LAT data in the search for primordial gravitational waves with next-generation CMB experiments.

Over the past several decades, measurements of the Cosmic Microwave Background (CMB) have been a major driving force in our understanding of cosmology. Measurements of the CMB on large angular scales places tight constraints on the parameters of the Lambda-CDM cosmological model. Measurements of the CMB at smaller angular scales constrains secondary anisotropies, such as the thermal and kinetic Sunyaev Zel'dovitch (tSZ and kSZ) effects, which constrain the structure of the universe at later times. The CMB is also polarized, and the polarization signal encodes information about both the inflationary era and late-time structure formation in our universe. In the first part of this dissertation, I introduce the CMB and discuss the measurements that have been made so far and what we can learn from them.

The South Pole Telescope (SPT) is a 10 meter telescope that is dedicated to measuring the CMB down to small angular scales. So far, the SPT has housed two instruments, SPT-sz and SPT-pol. SPT-sz was sensitive to temperature and completed a 2540 square degree survey of the southern sky from 2008-2011. SPT-pol is a polarization sensitive camera that was deployed in 2012, and has been conducting a polarization survey since that time. These instruments both use photon-noise limited superconducting Transition Edge Sensors (TESes) in large numbers to obtain their high sensitivities. TES design and fabrication in large format arrays was critical to the success of these instruments. The second part of this dissertation focuses on the instrumentation of the SPT. I present the theory behind TES bolometer design with a focus on detector stability and optimal performance. Finally, I discuss the design and performance of the two instruments, with a focus on the detector development for the SPT-pol instrument.

In the final part of this dissertation, I use the full 2540 square degree SPT-sz survey to measure the power spectrum from 1850 < l < 11000 at 90, 150, and 220 GHz and constrain small scale CMB anisotropies. We use a multifrequency cross-spectrum analysis to fit the data using a model that includes the lensed primary CMB, secondary CMB anisotropies (tSZ and kSZ effects), and foregrounds. The foreground power consists of a poisson component from radio sources, poisson and clustered components from dusty point sources, galactic cirrus, and tSZ-CIB correlations. These data represent the most sensitive measurement of the small angular scale CMB power spectrum to date. The data constrain the amplitude of the tSZ and kSZ signals, from which we can learn about late time structure and the epoch of reionization.

The South Pole Telescope (SPT) is a dedicated 10-meter diameter telescope optimized for mm-wavelength surveys of the Cosmic Microwave Background (CMB) with arcminute resolution. The first instrument deployed at SPT features a 960 element array of horn-coupled bolometers. These devices consist of fully lithographed spider-web absorbers and aluminum-titanium bilayer transition edge sensors fabricated on adhesive-bonded silicon wafers with embedded metal backplanes. The focal plane is cooled using a closed cycle pulse-tube refrigerator, and read-out using Frequency Domain Multiplexed Superconducting Quantum Interference Devices (SQUIDs.) Design features were chosen to optimize sensitivy in the atmospheric observing bands available from the ground, and for stability with the frequency domain multiplexed readout system employed at SPT, and performance was verified with a combination of laboratory tests and field observations.

In 2008 the SPT surveyed 200 square-degrees at 150 and 220 GHz. These data have been analyzed using a cross-spectrum analysis and multi-band Markov Chain Monte Carlo parameter fitting using a model that includes lensed primary CMB anisotropy, secondary thermal (tSZ) and kinetic (kSZ) Sunyaev-Zel'dovich anisotropies, unclustered synchrotron point sources, and clustered dusty point sources. In addition to measuring the power spectrum of dusty galaxies at high signal-to-noise, the data primarily constrain a linear combination of the kSZ and tSZ anisotropy contributions (at 150 GHz and ell=3000): D^tSZ₃₀₀₀ + 0.5 D^kSZ₃₀₀₀ = 4.5 ± 1.0 &mu K², and place the lowest limits yet measured on secondary anisotropy power.

The South Pole Telescope is a 10m millimeter-wavelength telescope for finding

galaxy clusters via the thermal Sunyaev-Zel'dovich (tSZ) effect. This thesis is divided into two parts. The

first part describes the development of the kilopixel SPT-SZ receiver and the frequency-domain multiplexor

(fMUX). The second part describes the first SPT power spectrum measurement and the first detection of the

tSZ power spectrum.

The SPT-SZ focal plane consists of 960 spiderweb coupled transition-edge sensors. Due to strong electro-

thermal feedback, these devices have good sensitivity and linearity, though risk spontaneous oscillations.

Adding heat capacity to these devices can ensure stability, so long as the loopgain, $\mathcal{L}$, is less

than $G_\textrm{int}/G_0$, the ratio between the thermal conductances linking the TES to the heat capacity

and linking the heat capacity to the bath. I describe as experimental technique for measuring the

internal thermal structure of these devices, allowing for rapid sensor evaluation.

The fMUX readout system reduces wiring complexity in this receiver by AC-biasing each sensor at a unique

frequency and sending signals from multiple bolometers along one pair of wires. The Series SQUID Arrays

(SSAs) used to read changes in bolometer current are notably non-linear and extremely sensititve to ambient

magnetic fields. The SSAs are housed in compact magnetic shielding modules which reduces their effective

area to 80 $\textrm{m}\Phi_0/\textrm{gauss}$. The SSA are fedback with a flux-locked loop to improve their

linearity and dynamic range, and decrease their input reactance. The FLL is bandwidth of 1 MHz with a

measured loopgain of 10. In the current implementation, this bandwidth is limited between the SQUID input

coil and other reactances, which I study in Chapter \ref{chap:fllstab}.

In the second part of the thesis I present power spectrum measurements for the first 100~deg$^2$ field observed by the SPT. On angular scales where the primary CMB anisotropy is dominant,

$\ell \lesssim 3000$, the SPT power spectrum is consistent with the standard

$\Lambda$CDM cosmology. On smaller scales, we see strong evidence for

a point source contribution, consisteThe South Pole Telescope is a 10m millimeter-wavelength telescope for finding

galaxy clusters via the thermal Sunyaev-Zel'dovich (tSZ) effect. This thesis is divided into two parts. The

first part describes the development of the kilopixel SPT-SZ receiver and the frequency-domain multiplexor

(fMUX). The second part describes the first SPT power spectrum measurement and the first detection of the

tSZ power spectrum.

The SPT-SZ focal plane consists of 960 spiderweb coupled transition-edge sensors. Due to strong electro-

thermal feedback, these devices have good sensitivity and linearity, though risk spontaneous oscillations.

Adding heat capacity to these devices can ensure stability, so long as the loopgain, $\mathcal{L}$, is less

than $G_\textrm{int}/G_0$, the ratio between the thermal conductances linking the TES to the heat capacity

and linking the heat capacity to the bath. I describe as experimental technique for measuring the

internal thermal structure of these devices, allowing for rapid sensor evaluation.

The fMUX readout system reduces wiring complexity in this receiver by AC-biasing each sensor at a unique

frequency and sending signals from multiple bolometers along one pair of wires. The Series SQUID Arrays

(SSAs) used to read changes in bolometer current are notably non-linear and extremely sensititve to ambient

magnetic fields. The SSAs are housed in compact magnetic shielding modules which reduces their effective

area to 80 $\textrm{m}\Phi_0/\textrm{gauss}$. The SSA are fedback with a flux-locked loop to improve their

linearity and dynamic range, and decrease their input reactance. The FLL is bandwidth of 1 MHz with a

measured loopgain of 10. In the current implementation, this bandwidth is limited between the SQUID input

coil and other reactances, which I study in Chapter \ref{chap:fllstab}.

In the second part of the thesis I present power spectrum measurements for the first 100~deg$^2$ field observed by the SPT. On angular scales where the primary CMB anisotropy is dominant,

$\ell \lesssim 3000$, the SPT power spectrum is consistent with the standard

$\Lambda$CDM cosmology. On smaller scales, we see strong evidence for

a point source contribution, consistent with a population of dusty, star-forming galaxies.

I 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\,$$\pm$$\,1.5\,

\mu\rm{K}^2$, which is significantly lower than the power predicted by a

fiducial model using WMAP5 cosmological parameters. t with a population of dusty, star-forming galaxies.

I 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\,$$\pm$$\,1.5\,

\mu\rm{K}^2$, which is significantly lower than the power predicted by a

fiducial model using WMAP5 cosmological parameters.

The question of how exactly the universe began is the motivation for this work. Based on the discoveries of the cosmic expansion and of the cosmic microwave background (CMB) radiation, humans have learned of the Big Bang origin of the universe. However, what exactly happened in the first moments of the Big Bang? A scenario of initial exponential expansion called "inflation" was proposed in the 1980s, explaining several important mysteries about the universe. Inflation would have generated gravitational waves that would have left a unique imprint in the polarization of the CMB. To search for this evidence for inflation, a team gathered in 2002 to design a telescope experiment called BICEP. Sited at the South Pole, BICEP was a novel 25-cm aperture refractor with 49 pairs of polarization-sensitive bolometers. We completed 3 years of successful observations from February 2006 to December 2008. To constrain the amplitude of polarization resulting from inflation, expected to be at least 7 orders of magnitude fainter than the 3 K CMB intensity, precise control of systematic effects is essential. A crucial challenge is preventing systematic errors from introducing false polarization anisotropy signal at the level corresponding to ~0.1 μK in amplitude. One main focus of this thesis is the characterization of systematic effects for BICEP. We developed a simulation framework for propagating instrumental systematic effects to the final polarization results. Based on these simulations, we established benchmarks for the characterization of critical instrumental properties including bolometer relative gains, beam mismatch, polarization orientation, telescope pointing, sidelobes, thermal stability, and timestream noise model. Guided by these benchmarks, we carefully measured these properties and have shown that we have characterized the instrument adequately to ensure that systematic errors do not limit BICEP's current cosmology results. We have analyzed the first 2 years of data, lowering the upper limits on the gravitational-wave induced polarization by an order of magnitude over all previous experiments. The systematic error analysis has identified what future refinements are likely necessary to probe CMB polarization down to levels corresponding to inflationary energy scales below 2 × 10¹⁶ GeV.

The South Pole Telescope (SPT) is an instrument designed to survey galaxy clusters using the Sunyaev-Zel'dovich effect. In this thesis I describe the telescope, its first-generation receiver, and its readout and control systems. I also present Sunyaev-Zel'dovich measurements of 15 massive X-ray selected galaxy clusters obtained with the SPT. The cluster signals are measured at 150 GHz, and concurrent 220 GHz data are used to reduce astrophysical contamination. Radial profiles are computed using a technique that takes into account the effects of the beams and filtering. In several clusters, significant SZ decrements are detected out to a substantial fraction of the virial radius. The profiles are fit to the -model and to a generalized NFW pressure profile, and are scaled and stacked to probe their average behavior. The best-fit model parameters are consistent with previous studies: β = 0.86 and r_core/r₅₀₀ = 0.20 for the β-model, and (α, β, γ, c₅₀₀)=(1.0, 5.5, 0.5, 1.0) for the generalized NFW model. Both models fit the SPT data comparably well, and both are consistent with the average SZ profile out to the virial radius. The integrated Compton-y parameter Y_SZ is computed for each cluster using both model-dependent and model-independent techniques, and the results are compared to X-ray estimates of cluster parameters. Y_SZ is found to scale with Y_X and gas mass with low scatter. Since these observables have been found to scale with total mass, these results point to a tight mass-observable relation for the SPT cluster survey.

We describe the development of an antenna-coupled bolometer array for use in a Cosmic Microwave Background polarization experiment. Prototype single pixels using double-slot dipole antennas and integrated microstrip band defining filters have been built and tested. Preliminary results of optical testing and simulations are presented. A bolometer array design based on this pixel will also be shown and future plans for application of the technology will be discussed.