High-coherence hybrid Si/III-V semiconductor lasers

The relentlessly increasing demand for network bandwidth, driven primarily by Internet-based services such as mobile computing, cloud storage and video-on-demand, calls for more efficient utilization of the available communication spectrum, as that afforded by the resurging DSP-powered coherent optical communications. Encoding information in the phase of the optical carrier, using multilevel phase modulationformats, and employing coherent detection at the receiver allows for enhanced spectral efficiency and thus enables increased network capacity. The distributed feedback semiconductor laser (DFB) has served as the near exclusive light source powering the fiber optic, long-haul network for over 30 years. The transition to coherent communication systems is pushing the DFB laser to the limits of its abilities. This is due to its limited temporal coherence that directly translates into the number of different phases that can be imparted to a single optical pulse and thus to the data capacity. Temporal coherence, most commonly quantified in the spectral linewidth Δν, is limited by phase noise, result of quantum-mandated spontaneous emission of photons due to random recombination of carriers in the active region of the laser.

In this work we develop a generically new type of semiconductor laser with the requisite coherence properties. We demonstrate electrically driven lasers characterized by a quantum noise-limited spectral linewidth as low as 18 kHz. This narrow linewidth is result of a fundamentally new laser design philosophy that separates the functions of photon generation and storage and is enabled by a hybrid Si/III-V integration platform. Photons generated in the active region of the III-V material are readily stored away in the low loss Si that hosts the bulk of the laser field, thereby enabling high-Q photon storage. The storage of a large number of coherent quanta acts as an optical flywheel, which by its inertia reduces the effect of the spontaneous emission-mandated phase perturbations on the laser field, while the enhanced photon lifetime effectively reduces the emission rate of incoherent quanta into the lasing mode. Narrow linewidths are obtained over a wavelength bandwidth spanning the entire optical communication C-band (1530-1575nm) at only a fraction of the input power required by conventional DFB lasers. The results presented in this thesis hold great promise for the large scale integration of lithographically tuned, high-coherence laser arrays for use in coherent communications, that will enable Tb/s-scale data capacities.

Liquid silicate equation of state : using shock waves to understand the properties of the deep Earth

The equations of state (EOS) of several geologically important silicate liquids have been constrained via preheated shock wave techniques. Results on molten Fe₂SiO₄ (fayalite), Mg₂SiO₄ (forsterite), CaFeSi₂O₆ (hedenbergite), an equimolar mixture of CaAl₂Si₂O₈-CaFeSi₂O₆ (anorthite-hedenbergite), and an equimolar mixture of CaAl₂Si₂O₈-CaFeSi₂O₆-CaMgSi₂O₆(anorthite-hedenbergite-diopside) are presented. This work represents the first ever direct EOS measurements of an iron-bearing liquid or of a forsterite liquid at pressures relevant to the deep Earth (> 135 GPa). Additionally, revised EOS for molten CaMgSi₂O₆ (diopside), CaAl₂Si₂O₈ (anorthite), and MgSiO₃ (enstatite), which were previously determined by shock wave methods, are also presented.

The liquid EOS are incorporated into a model, which employs linear mixing of volumes to determine the density of compositionally intermediate liquids in the CaO-MgO-Al₂O₃-SiO₂-FeO major element space. Liquid volumes are calculated for temperature and pressure conditions that are currently present at the core-mantle boundary or that may have occurred during differentiation of a fully molten mantle magma ocean.

The most significant implications of our results include: (1) a magma ocean of either chondrite or peridotite composition is less dense than its first crystallizing solid, which is not conducive to the formation of a basal mantle magma ocean, (2) the ambient mantle cannot produce a partial melt and an equilibrium residue sufficiently dense to form an ultralow velocity zone mush, and (3) due to the compositional dependence of Fe2+ coordination, there is a threshold of Fe concentration (molar X_Fe ≤ 0.06) permitted in a liquid for which its density can still be approximated by linear mixing of end-member volumes.

Deep tissue fluorescence imaging with time-reversed light

Advances in optical techniques have enabled many breakthroughs in biology and medicine. However, light scattering by biological tissues remains a great obstacle, restricting the use of optical methods to thin ex vivo sections or superficial layers in vivo. In this thesis, we present two related methods that overcome the optical depth limit—digital time reversal of ultrasound encoded light (digital TRUE) and time reversal of variance-encoded light (TROVE). These two techniques share the same principle of using acousto-optic beacons for time reversal optical focusing within highly scattering media, like biological tissues. Ultrasound, unlike light, is not significantly scattered in soft biological tissues, allowing for ultrasound focusing. In addition, a fraction of the scattered optical wavefront that passes through the ultrasound focus gets frequency-shifted via the acousto-optic effect, essentially creating a virtual source of frequency-shifted light within the tissue. The scattered ultrasound-tagged wavefront can be selectively measured outside the tissue and time-reversed to converge at the location of the ultrasound focus, enabling optical focusing within deep tissues. In digital TRUE, we time reverse ultrasound-tagged light with an optoelectronic time reversal device (the digital optical phase conjugate mirror, DOPC). The use of the DOPC enables high optical gain, allowing for high intensity optical focusing and focal fluorescence imaging in thick tissues at a lateral resolution of 36 µm by 52 µm. The resolution of the TRUE approach is fundamentally limited to that of the wavelength of ultrasound. The ultrasound focus (~ tens of microns wide) usually contains hundreds to thousands of optical modes, such that the scattered wavefront measured is a linear combination of the contributions of all these optical modes. In TROVE, we make use of our ability to digitally record, analyze and manipulate the scattered wavefront to demix the contributions of these spatial modes using variance encoding. In essence, we encode each spatial mode inside the scattering sample with a unique variance, allowing us to computationally derive the time reversal wavefront that corresponds to a single optical mode. In doing so, we uncouple the system resolution from the size of the ultrasound focus, demonstrating optical focusing and imaging between highly diffusing samples at an unprecedented, speckle-scale lateral resolution of ~ 5 µm. Our methods open up the possibility of fully exploiting the prowess and versatility of biomedical optics in deep tissues.

Electrical and optical interconnects for high-performance computing

Technology scaling has enabled drastic growth in the computational and storage capacity of integrated circuits (ICs). This constant growth drives an increasing demand for high-bandwidth communication between and within ICs. In this dissertation we focus on low-power solutions that address this demand. We divide communication links into three subcategories depending on the communication distance. Each category has a different set of challenges and requirements and is affected by CMOS technology scaling in a different manner. We start with short-range chip-to-chip links for board-level communication. Next we will discuss board-to-board links, which demand a longer communication range. Finally on-chip links with communication ranges of a few millimeters are discussed.

Electrical signaling is a natural choice for chip-to-chip communication due to efficient integration and low cost. IO data rates have increased to the point where electrical signaling is now limited by the channel bandwidth. In order to achieve multi-Gb/s data rates, complex designs that equalize the channel are necessary. In addition, a high level of parallelism is central to sustaining bandwidth growth. Decision feedback equalization (DFE) is one of the most commonly employed techniques to overcome the limited bandwidth problem of the electrical channels. A linear and low-power summer is the central block of a DFE. Conventional approaches employ current-mode techniques to implement the summer, which require high power consumption. In order to achieve low-power operation we propose performing the summation in the charge domain. This approach enables a low-power and compact realization of the DFE as well as crosstalk cancellation. A prototype receiver was fabricated in 45nm SOI CMOS to validate the functionality of the proposed technique and was tested over channels with different levels of loss and coupling. Measurement results show that the receiver can equalize channels with maximum 21dB loss while consuming about 7.5mW from a 1.2V supply. We also introduce a compact, low-power transmitter employing passive equalization. The efficacy of the proposed technique is demonstrated through implementation of a prototype in 65nm CMOS. The design achieves up to 20Gb/s data rate while consuming less than 10mW.

An alternative to electrical signaling is to employ optical signaling for chip-to-chip interconnections, which offers low channel loss and cross-talk while providing high communication bandwidth. In this work we demonstrate the possibility of building compact and low-power optical receivers. A novel RC front-end is proposed that combines dynamic offset modulation and double-sampling techniques to eliminate the need for a short time constant at the input of the receiver. Unlike conventional designs, this receiver does not require a high-gain stage that runs at the data rate, making it suitable for low-power implementations. In addition, it allows time-division multiplexing to support very high data rates. A prototype was implemented in 65nm CMOS and achieved up to 24Gb/s with less than 0.4pJ/b power efficiency per channel. As the proposed design mainly employs digital blocks, it benefits greatly from technology scaling in terms of power and area saving.

As the technology scales, the number of transistors on the chip grows. This necessitates a corresponding increase in the bandwidth of the on-chip wires. In this dissertation, we take a close look at wire scaling and investigate its effect on wire performance metrics. We explore a novel on-chip communication link based on a double-sampling architecture and dynamic offset modulation technique that enables low power consumption and high data rates while achieving high bandwidth density in 28nm CMOS technology. The functionality of the link is demonstrated using different length minimum-pitch on-chip wires. Measurement results show that the link achieves up to 20Gb/s of data rate (12.5Gb/s/$\mu$m) with better than 136fJ/b of power efficiency.

The deep ocean density structure at the Last Glacial Maximum : What was it and why?

The search for reliable proxies of past deep ocean temperature and salinity has proved difficult, thereby limiting our ability to understand the coupling of ocean circulation and climate over glacial-interglacial timescales. Previous inferences of deep ocean temperature and salinity from sediment pore fluid oxygen isotopes and chlorinity indicate that the deep ocean density structure at the Last Glacial Maximum (LGM, approximately 20,000 years BP) was set by salinity, and that the density contrast between northern and southern sourced deep waters was markedly greater than in the modern ocean. High density stratification could help explain the marked contrast in carbon isotope distribution recorded in the LGM ocean relative to that we observe today, but what made the ocean's density structure so different at the LGM? How did it evolve from one state to another? Further, given the sparsity of the LGM temperature and salinity data set, what else can we learn by increasing the spatial density of proxy records?

We investigate the cause and feasibility of a highly and salinity stratified deep ocean at the LGM and we work to increase the amount of information we can glean about the past ocean from pore fluid profiles of oxygen isotopes and chloride. Using a coupled ocean--sea ice--ice shelf cavity model we test whether the deep ocean density structure at the LGM can be explained by ice--ocean interactions over the Antarctic continental shelves, and show that a large contribution of the LGM salinity stratification can be explained through lower ocean temperature. In order to extract the maximum information from pore fluid profiles of oxygen isotopes and chloride we evaluate several inverse methods for ill-posed problems and their ability to recover bottom water histories from sediment pore fluid profiles. We demonstrate that Bayesian Markov Chain Monte Carlo parameter estimation techniques enable us to robustly recover the full solution space of bottom water histories, not only at the LGM, but through the most recent deglaciation and the Holocene up to the present. Finally, we evaluate a non-destructive pore fluid sampling technique, Rhizon samplers, in comparison to traditional squeezing methods and show that despite their promise, Rhizons are unlikely to be a good sampling tool for pore fluid measurements of oxygen isotopes and chloride.

Lightweight deformable mirrors for future space telescopes

This thesis presents a concept for ultra-lightweight deformable mirrors based on a thin substrate of optical surface quality coated with continuous active piezopolymer layers that provide modes of actuation and shape correction. This concept eliminates any kind of stiff backing structure for the mirror surface and exploits micro-fabrication technologies to provide a tight integration of the active materials into the mirror structure, to avoid actuator print-through effects. Proof-of-concept, 10-cm-diameter mirrors with a low areal density of about 0.5 kg/m² have been designed, built and tested to measure their shape-correction performance and verify the models used for design. The low cost manufacturing scheme uses replication techniques, and strives for minimizing residual stresses that deviate the optical figure from the master mandrel. It does not require precision tolerancing, is lightweight, and is therefore potentially scalable to larger diameters for use in large, modular space telescopes. Other potential applications for such a laminate could include ground-based mirrors for solar energy collection, adaptive optics for atmospheric turbulence, laser communications, and other shape control applications.

The immediate application for these mirrors is for the Autonomous Assembly and Reconfiguration of a Space Telescope (AAReST) mission, which is a university mission under development by Caltech, the University of Surrey, and JPL. The design concept, fabrication methodology, material behaviors and measurements, mirror modeling, mounting and control electronics design, shape control experiments, predictive performance analysis, and remaining challenges are presented herein. The experiments have validated numerical models of the mirror, and the mirror models have been used within a model of the telescope in order to predict the optical performance. A demonstration of this mirror concept, along with other new telescope technologies, is planned to take place during the AAReST mission.

Did galaxies reionize the universe?

The epoch of reionization remains one of the last uncharted eras of cosmic history, yet this time is of crucial importance, encompassing the formation of both the first galaxies and the first metals in the universe. In this thesis, I present four related projects that both characterize the abundance and properties of these first galaxies and uses follow-up observations of these galaxies to achieve one of the first observations of the neutral fraction of the intergalactic medium during the heart of the reionization era.

First, we present the results of a spectroscopic survey using the Keck telescopes targeting 6.3 < z < 8.8 star-forming galaxies. We secured observations of 19 candidates, initially selected by applying the Lyman break technique to infrared imaging data from the Wide Field Camera 3 (WFC3) onboard the Hubble Space Telescope (HST). This survey builds upon earlier work from Stark et al. (2010, 2011), which showed that star-forming galaxies at 3 < z < 6, when the universe was highly ionized, displayed a significant increase in strong Lyman alpha emission with redshift. Our work uses the LRIS and NIRSPEC instruments to search for Lyman alpha emission in candidates at a greater redshift in the observed near-infrared, in order to discern if this evolution continues, or is quenched by an increase in the neutral fraction of the intergalactic medium. Our spectroscopic observations typically reach a 5-sigma limiting sensitivity of < 50 AA. Despite expecting to detect Lyman alpha at 5-sigma in 7-8 galaxies based on our Monte Carlo simulations, we only achieve secure detections in two of 19 sources. Combining these results with a similar sample of 7 galaxies from Fontana et al. (2010), we determine that these few detections would only occur in < 1% of simulations if the intrinsic distribution was the same as that at z ~ 6. We consider other explanations for this decline, but find the most convincing explanation to be an increase in the neutral fraction of the intergalactic medium. Using theoretical models, we infer a neutral fraction of X_HI ~ 0.44 at z = 7.

Second, we characterize the abundance of star-forming galaxies at z > 6.5 again using WFC3 onboard the HST. This project conducted a detailed search for candidates both in the Hubble Ultra Deep Field as well as a number of additional wider Hubble Space Telescope surveys to construct luminosity functions at both z ~ 7 and 8, reaching 0.65 and 0.25 mag fainter than any previous surveys, respectively. With this increased depth, we achieve some of the most robust constraints on the Schechter function faint end slopes at these redshifts, finding very steep values of alpha_{z~7} = -1.87 +/- 0.18 and alpha_{z~8} = -1.94 +/- 0.23. We discuss these results in the context of cosmic reionization, and show that given reasonable assumptions about the ionizing spectra and escape fraction of ionizing photons, only half the photons needed to maintain reionization are provided by currently observable galaxies at z ~ 7-8. We show that an extension of the luminosity function down to M_{UV} = -13.0, coupled with a low level of star-formation out to higher redshift, can fit all available constraints on the ionization history of the universe.

Third, we investigate the strength of nebular emission in 3 < z < 5 star-forming galaxies. We begin by using the Infrared Array Camera (IRAC) onboard the Spitzer Space Telescope to investigate the strength of H alpha emission in a sample of 3.8 < z < 5.0 spectroscopically confirmed galaxies. We then conduct near-infrared observations of star-forming galaxies at 3 < z < 3.8 to investigate the strength of the [OIII] 4959/5007 and H beta emission lines from the ground using MOSFIRE. In both cases, we uncover near-ubiquitous strong nebular emission, and find excellent agreement between the fluxes derived using the separate methods. For a subset of 9 objects in our MOSFIRE sample that have secure Spitzer IRAC detections, we compare the emission line flux derived from the excess in the K_s band photometry to that derived from direct spectroscopy and find 7 to agree within a factor of 1.6, with only one catastrophic outlier. Finally, for a different subset for which we also have DEIMOS rest-UV spectroscopy, we compare the relative velocities of Lyman alpha and the rest-optical nebular lines which should trace the cites of star-formation. We find a median velocity offset of only v_{Ly alpha} = 149 km/s, significantly less than the 400 km/s observed for star-forming galaxies with weaker Lyman alpha emission at z = 2-3 (Steidel et al. 2010), and show that this decrease can be explained by a decrease in the neutral hydrogen column density covering the galaxy. We discuss how this will imply a lower neutral fraction for a given observed extinction of Lyman alpha when its visibility is used to probe the ionization state of the intergalactic medium.

Finally, we utilize the recent CANDELS wide-field, infra-red photometry over the GOODS-N and S fields to re-analyze the use of Lyman alpha emission to evaluate the neutrality of the intergalactic medium. With this new data, we derive accurate ultraviolet spectral slopes for a sample of 468 3 < z < 6 star-forming galaxies, already observed in the rest-UV with the Keck spectroscopic survey (Stark et al. 2010). We use a Bayesian fitting method which accurately accounts for contamination and obscuration by skylines to derive a relationship between the UV-slope of a galaxy and its intrinsic Lyman alpha equivalent width probability distribution. We then apply this data to spectroscopic surveys during the reionization era, including our own, to accurately interpret the drop in observed Lyman alpha emission. From our most recent such MOSFIRE survey, we also present evidence for the most distant galaxy confirmed through emission line spectroscopy at z = 7.62, as well as a first detection of the CIII]1907/1909 doublet at z > 7.

We conclude the thesis by exploring future prospects and summarizing the results of Robertson et al. (2013). This work synthesizes many of the measurements in this thesis, along with external constraints, to create a model of reionization that fits nearly all available constraints.

Strong atom-light interactions along nanostructures: Transition from free-space to nanophotonic interfaces

An exciting frontier in quantum information science is the integration of otherwise "simple'' quantum elements into complex quantum networks. The laboratory realization of even small quantum networks enables the exploration of physical systems that have not heretofore existed in the natural world. Within this context, there is active research to achieve nanoscale quantum optical circuits, for which atoms are trapped near nano-scopic dielectric structures and "wired'' together by photons propagating through the circuit elements. Single atoms and atomic ensembles endow quantum functionality for otherwise linear optical circuits and thereby enable the capability of building quantum networks component by component. Toward these goals, we have experimentally investigated three different systems, from conventional to rather exotic systems : free-space atomic ensembles, optical nano fibers, and photonics crystal waveguides. First, we demonstrate measurement-induced quadripartite entanglement among four quantum memories. Next, following the landmark realization of a nanofiber trap, we demonstrate the implementation of a state-insensitive, compensated nanofiber trap. Finally, we reach more exotic systems based on photonics crystal devices. Beyond conventional topologies of resonators and waveguides, new opportunities emerge from the powerful capabilities of dispersion and modal engineering in photonic crystal waveguides. We have implemented an integrated optical circuit with a photonics crystal waveguide capable of both trapping and interfacing atoms with guided photons, and have observed the collective effect, superradiance, mediated by the guided photons. These advances provide an important capability for engineered light-matter interactions, enabling explorations of novel quantum transport and quantum many-body phenomena.

Paranormal operator on a Hilbert space

In this thesis an extensive study is made of the set P of all paranormal operators in B(H), the set of all bounded endomorphisms on the complex Hilbert space H. T ϵ B(H) is paranormal if for each z contained in the resolvent set of T, d(z, σ(T))//(T-zI)^-1 = 1 where d(z, σ(T)) is the distance from z to σ(T), the spectrum of T. P contains the set N of normal operators and P contains the set of hyponormal operators. However, P is contained in L, the set of all T ϵ B(H) such that the convex hull of the spectrum of T is equal to the closure of the numerical range of T. Thus, N≤P≤L.

If the uniform operator (norm) topology is placed on B(H), then the relative topological properties of N, P, L can be discussed. In Section IV, it is shown that: 1) N P and L are arc-wise connected and closed, 2) N, P, and L are nowhere dense subsets of B(H) when dim H ≥ 2, 3) N = P when dimH ˂ ∞ , 4) N is a nowhere dense subset of P when dimH ˂ ∞ , 5) P is not a nowhere dense subset of L when dimH ˂ ∞ , and 6) it is not known if P is a nowhere dense subset of L when dimH ˂ ∞.

The spectral properties of paranormal operators are of current interest in the literature. Putnam [22, 23] has shown that certain points on the boundary of the spectrum of a paranormal operator are either normal eigenvalues or normal approximate eigenvalues. Stampfli [26] has shown that a hyponormal operator with countable spectrum is normal. However, in Theorem 3.3, it is shown that a paranormal operator T with countable spectrum can be written as the direct sum, N ⊕ A, of a normal operator N with σ(N) = σ(T) and of an operator A with σ(A) a subset of the derived set of σ(T). It is then shown that A need not be normal. If we restrict the countable spectrum of T ϵ P to lie on a C²-smooth rectifiable Jordan curve G_o, then T must be normal [see Theorem 3.5 and its Corollary]. If T is a scalar paranormal operator with countable spectrum, then in order to conclude that T is normal the condition of σ(T) ≤ G_o can be relaxed [see Theorem 3.6]. In Theorem 3.7 it is then shown that the above result is not true when T is not assumed to be scalar. It was then conjectured that if T ϵ P with σ(T) ≤ G_o, then T is normal. The proof of Theorem 3.5 relies heavily on the assumption that T has countable spectrum and cannot be generalized. However, the corollary to Theorem 3.9 states that if T ϵ P with σ(T) ≤ G_o, then T has a non-trivial lattice of invariant subspaces. After the completion of most of the work on this thesis, Stampfli [30, 31] published a proof that a paranormal operator T with σ(T) ≤ G_o is normal. His proof uses some rather deep results concerning numerical ranges whereas the proof of Theorem 3.5 uses relatively elementary methods.

Deep near-infrared spectroscopy of high-redshift galaxies: the physical growth of passive systems

The assembly history of massive galaxies is one of the most important aspects of galaxy formation and evolution. Although we have a broad idea of what physical processes govern the early phases of galaxy evolution, there are still many open questions. In this thesis I demonstrate the crucial role that spectroscopy can play in a physical understanding of galaxy evolution. I present deep near-infrared spectroscopy for a sample of high-redshift galaxies, from which I derive important physical properties and their evolution with cosmic time. I take advantage of the recent arrival of efficient near-infrared detectors to target the rest-frame optical spectra of z > 1 galaxies, from which many physical quantities can be derived. After illustrating the applications of near-infrared deep spectroscopy with a study of star-forming galaxies, I focus on the evolution of massive quiescent systems.

Most of this thesis is based on two samples collected at the W. M. Keck Observatory that represent a significant step forward in the spectroscopic study of z > 1 quiescent galaxies. All previous spectroscopic samples at this redshift were either limited to a few objects, or much shallower in terms of depth. Our first sample is composed of 56 quiescent galaxies at 1 < z < 1.6 collected using the upgraded red arm of the Low Resolution Imaging Spectrometer (LRIS). The second consists of 24 deep spectra of 1.5 < z < 2.5 quiescent objects observed with the Multi-Object Spectrometer For Infra-Red Exploration (MOSFIRE). Together, these spectra span the critical epoch 1 < z < 2.5, where most of the red sequence is formed, and where the sizes of quiescent systems are observed to increase significantly.

We measure stellar velocity dispersions and dynamical masses for the largest number of z > 1 quiescent galaxies to date. By assuming that the velocity dispersion of a massive galaxy does not change throughout its lifetime, as suggested by theoretical studies, we match galaxies in the local universe with their high-redshift progenitors. This allows us to derive the physical growth in mass and size experienced by individual systems, which represents a substantial advance over photometric inferences based on the overall galaxy population. We find a significant physical growth among quiescent galaxies over 0 < z < 2.5 and, by comparing the slope of growth in the mass-size plane dlogR_e/dlogM_∗ with the results of numerical simulations, we can constrain the physical process responsible for the evolution. Our results show that the slope of growth becomes steeper at higher redshifts, yet is broadly consistent with minor mergers being the main process by which individual objects evolve in mass and size.

By fitting stellar population models to the observed spectroscopy and photometry we derive reliable ages and other stellar population properties. We show that the addition of the spectroscopic data helps break the degeneracy between age and dust extinction, and yields significantly more robust results compared to fitting models to the photometry alone. We detect a clear relation between size and age, where larger galaxies are younger. Therefore, over time the average size of the quiescent population will increase because of the contribution of large galaxies recently arrived to the red sequence. This effect, called progenitor bias, is different from the physical size growth discussed above, but represents another contribution to the observed difference between the typical sizes of low- and high-redshift quiescent galaxies. By reconstructing the evolution of the red sequence starting at z ∼ 1.25 and using our stellar population histories to infer the past behavior to z ∼ 2, we demonstrate that progenitor bias accounts for only half of the observed growth of the population. The remaining size evolution must be due to physical growth of individual systems, in agreement with our dynamical study.

Finally, we use the stellar population properties to explore the earliest periods which led to the formation of massive quiescent galaxies. We find tentative evidence for two channels of star formation quenching, which suggests the existence of two independent physical mechanisms. We also detect a mass downsizing, where more massive galaxies form at higher redshift, and then evolve passively. By analyzing in depth the star formation history of the brightest object at z > 2 in our sample, we are able to put constraints on the quenching timescale and on the properties of its progenitor.

A consistent picture emerges from our analyses: massive galaxies form at very early epochs, are quenched on short timescales, and then evolve passively. The evolution is passive in the sense that no new stars are formed, but significant mass and size growth is achieved by accreting smaller, gas-poor systems. At the same time the population of quiescent galaxies grows in number due to the quenching of larger star-forming galaxies. This picture is in agreement with other observational studies, such as measurements of the merger rate and analyses of galaxy evolution at fixed number density.

Advanced Monte Carlo Simulation and Machine Learning for Frequency Domain Optical Coherence Tomography

Optical Coherence Tomography(OCT) is a popular, rapidly growing imaging technique with an increasing number of bio-medical applications due to its noninvasive nature. However, there are three major challenges in understanding and improving an OCT system: (1) Obtaining an OCT image is not easy. It either takes a real medical experiment or requires days of computer simulation. Without much data, it is difficult to study the physical processes underlying OCT imaging of different objects simply because there aren't many imaged objects. (2) Interpretation of an OCT image is also hard. This challenge is more profound than it appears. For instance, it would require a trained expert to tell from an OCT image of human skin whether there is a lesion or not. This is expensive in its own right, but even the expert cannot be sure about the exact size of the lesion or the width of the various skin layers. The take-away message is that analyzing an OCT image even from a high level would usually require a trained expert, and pixel-level interpretation is simply unrealistic. The reason is simple: we have OCT images but not their underlying ground-truth structure, so there is nothing to learn from. (3) The imaging depth of OCT is very limited (millimeter or sub-millimeter on human tissues). While OCT utilizes infrared light for illumination to stay noninvasive, the downside of this is that photons at such long wavelengths can only penetrate a limited depth into the tissue before getting back-scattered. To image a particular region of a tissue, photons first need to reach that region. As a result, OCT signals from deeper regions of the tissue are both weak (since few photons reached there) and distorted (due to multiple scatterings of the contributing photons). This fact alone makes OCT images very hard to interpret.

This thesis addresses the above challenges by successfully developing an advanced Monte Carlo simulation platform which is 10000 times faster than the state-of-the-art simulator in the literature, bringing down the simulation time from 360 hours to a single minute. This powerful simulation tool not only enables us to efficiently generate as many OCT images of objects with arbitrary structure and shape as we want on a common desktop computer, but it also provides us the underlying ground-truth of the simulated images at the same time because we dictate them at the beginning of the simulation. This is one of the key contributions of this thesis. What allows us to build such a powerful simulation tool includes a thorough understanding of the signal formation process, clever implementation of the importance sampling/photon splitting procedure, efficient use of a voxel-based mesh system in determining photon-mesh interception, and a parallel computation of different A-scans that consist a full OCT image, among other programming and mathematical tricks, which will be explained in detail later in the thesis.

Next we aim at the inverse problem: given an OCT image, predict/reconstruct its ground-truth structure on a pixel level. By solving this problem we would be able to interpret an OCT image completely and precisely without the help from a trained expert. It turns out that we can do much better. For simple structures we are able to reconstruct the ground-truth of an OCT image more than 98% correctly, and for more complicated structures (e.g., a multi-layered brain structure) we are looking at 93%. We achieved this through extensive uses of Machine Learning. The success of the Monte Carlo simulation already puts us in a great position by providing us with a great deal of data (effectively unlimited), in the form of (image, truth) pairs. Through a transformation of the high-dimensional response variable, we convert the learning task into a multi-output multi-class classification problem and a multi-output regression problem. We then build a hierarchy architecture of machine learning models (committee of experts) and train different parts of the architecture with specifically designed data sets. In prediction, an unseen OCT image first goes through a classification model to determine its structure (e.g., the number and the types of layers present in the image); then the image is handed to a regression model that is trained specifically for that particular structure to predict the length of the different layers and by doing so reconstruct the ground-truth of the image. We also demonstrate that ideas from Deep Learning can be useful to further improve the performance.

It is worth pointing out that solving the inverse problem automatically improves the imaging depth, since previously the lower half of an OCT image (i.e., greater depth) can be hardly seen but now becomes fully resolved. Interestingly, although OCT signals consisting the lower half of the image are weak, messy, and uninterpretable to human eyes, they still carry enough information which when fed into a well-trained machine learning model spits out precisely the true structure of the object being imaged. This is just another case where Artificial Intelligence (AI) outperforms human. To the best knowledge of the author, this thesis is not only a success but also the first attempt to reconstruct an OCT image at a pixel level. To even give a try on this kind of task, it would require fully annotated OCT images and a lot of them (hundreds or even thousands). This is clearly impossible without a powerful simulation tool like the one developed in this thesis.