La consommation d’énergie liée aux technologies de l’information augmente trèsrapidement et dans la mesure où la société a besoin d’être toujours plus connectée tout ens’appuyant sur des solutions durables, les technologies actuelles ne suffisent plus. La photoniqueintégrée s’impose dès lors comme une alternative à l’électronique pour réaliser du traitementdu signal économe en énergie. Au cours de cette thèse, j’ai étudié des structures sub-longueurd’onde en semiconducteur, les cristaux photoniques, qui présentent des propriétés non linéairesimpressionnantes. Plus précisément, le confinement fort et la propagation en lumière lente permettentun traitement sur puce de signal ultra-rapide tout optique, soit à partir de mélange àquatre ondes ou d’auto-modulation de phase. L’originalité est l’utilisation de nouveaux matériauxsemi-conducteurs ayant moins d’absorption non linéaires et par porteurs libres, effets qui limitentla pleine exploitation des effets non linéaires dans les structures photoniques en Silicium. Dansma thèse, des semiconducteurs III-V ont été utilisés pour développer des guides et des cavitéscristal photonique de grande qualité qui sont en mesure de supporter des densités de puissanceoptiques extrêmement élevées ainsi que de grands niveaux de puissance moyenne. J’ai amélioré laconductivité thermique des guides d’ondes grâce à l’intégration hétérogène de membranes GaInPavec du dioxyde de silicium. Cette plateforme permettra à terme de démontrer de l’amplificationsensible à la phase dans le régime continu que j’ai déjà démontré dans le régime pulsé en utilisant des membranes suspendues en GaInP. En parallèle, j’ai démontré des cristaux photoniques de grande qualité dans du Gallium phosphure, qui est un matériau très prometteur en raison de lagrande bande interdite et de la très bonne conductivité thermique. Les résultats préliminaires ontpermis la réalisation d’un régime non linéaire intense (mini-peigne de fréquence, compression etfission de soliton ...). / The energy consumption of the whole ICT ecosystem is growing at a fast paceand in a global context of the search for an ever more connected yet sustainable society, a technologicalbreakthrough is desired. Here, integrated nonlinear photonics will help by providingnovel possibilities for energy efficient signal processing. In this PhD thesis, I have been investigatingsub-wavelength semiconductor structures, particularly photonic crystals, which have shownremarkable nonlinear properties. More specifically the strong confinement and slow light propagationenables on-chip ultra-fast all-optical signal processing, either based on four-wave-mixingor self-phase modulation. The main point here is the use of novel semiconductor materials withimproved nonlinear properties with respect to Silicon. In fact, it has now been acknowledgedthat the nonlinear and free-carriers absorption in Silicon integrated photonic structures is anissue hindering the full exploitation of nonlinear effects. In my thesis, wide-gap III-V semiconductorshave been used to develop high quality photonic crystal waveguides and cavities whichare able to sustain extremely high optical power densities as well as large average power levels.I have demonstrated PhC waveguides with much improved thermal conductivity through heterogeneousintegration of GaInP membranes with silicon dioxide. This will allow continuous wave phase-sensitive amplification, which I already demonstrated in the pulsed regime using GaInPself-suspended membranes. In parallel, I have demonstrated high quality PhC in Gallium Phosphide,which is a very promising material because of the large bandgap and the very good thermalconductivity. Preliminar results demonstrate the achievement of extremely large nonlinear regime(mini-comb, soliton compression and fission ...).

<div>Biological dynamics have been studied by many methods. Fluorescence dynamic microscopy and optical coherence tomography provided fundamental understandings of biological systems. However, their high NA optics only represent local characteristics. Biodynamic imaging (BDI) technique implements a low NA optics and acquires the statistical average of Doppler shifts that occurred by dynamic light scattering with biological dynamic subsystems provided globally averaged dynamic characteristics. </div><div>BDI is used for this study to investigate biomedical applications. The chemotherapy efficacy measurement by BDI demonstrated a good agreement between the Doppler spectral phenotypes and the preclinical outcomes. Also, dynamic responses of microbiomes by chemical stimuli demonstrated featured Doppler characteristics. The bacterial infection of epithelial spheroids showed consistent spectral responses and antibiotic-resistant E. coli infection treatment with a sensitive and resistive antibiotic showed a dramatic contrast. Furthermore, the phase-sensitive characteristics of BDI provided a clue to understanding the characteristics of the random process of biological systems. Levy-like heavy-tailed probability density functions are demonstrated and </div><div>the shape changed by infection will be discussed. </div>

No description available.

Dans ce travail de thèse, je considère deux phénomènes distincts, tous deux liés aux interactions non-linéaires entre la lumière et des atomes. La première partie est dédiée à du mélange à 4 ondes basé sur des degrés de liberté internes d’atomes d’hélium à température ambiante, et l’utilise pour des processus d’amplification et de la génération d’états comprimés. Le second phénomène étudié est basé sur des degrés de liberté externes d’atomes de césium froids et est utilisé pour du stockage de lumière et la génération d’un champ conjugué en phase par mélange d’ondes. J'ai expérimentalement observé et caractérisé de l'amplification sensible à la phase par mélange à quatre ondes dans de l'hélium métastable à température ambiante. J'ai obtenu un gain maximum d'environ 9 dB avec une bande passante d'environ 300 kHz. Les fonctions de transfert phase/phase obtenues ont montré une forte compression de phase, indiquant que le phénomène était presque exempt de processus indésirables. Dans la seconde partie, j'explique comment les résonances de recul, dues à un transfert de quantité de mouvement entre un photon et un atome, peuvent être utilisées pour du stockage de lumière. J'explique également comment ce phénomène peut conduire à la génération d’un champ conjugué, et pourquoi la théorie existante ne permet pas de modéliser le creux qui apparaît dans le spectre de génération du champ conjugué lorsqu’on augmente la puissance optique. Pour reproduire ce nouvel élément, j’ai effectué un développement jusqu’au 5e ordre, qui démontre qu’il dépend de la cohérence qui est excitée entre des niveaux de moments atomiques différents. Je montre ensuite qu'un modèle plus simple, basé sur trois niveaux atomiques définis par des degrés de liberté interne et externe de l'atome, peut expliquer le phénomène observé. / In this PhD work, two distinct phenomena are considered, which are both related to non-linear interactions between light and atoms. The first part of the thesis is dedicated to four wave mixing based on the internal degrees of freedom of room temperature helium atoms and uses it for amplification processes and generation of squeezed light. The second studied process is based on external degrees of freedom of cold cesium atoms and used for light storage and phase conjugate field generation through multi-wave mixing. I experimentally observed and characterized phase sensitive amplification via four-wave mixing in metastable helium at room temperature. I have obtained about 9 dB of maximum gain with a bandwidth of about 300 kHz. The obtained phase transfer functions showed a strong phase squeezing, indicating that the phenomenon was almost free of unwanted processes. In the second part, I explain how recoil induced resonances, which are due to the transfer of momentum between a photon and an atom, can be used to store light. I also explain how this phenomenon can lead to generation of a phase conjugate field, and why the existing theory fails to model the dip, which appears in the phase conjugate generation spectrum when the field power is increased. I extend the model to the fifth order so that it can reproduce this new feature and demonstrate that it depends on the decay rate of the coherence, which is excited between atomic levels of different momenta. I then show that a simpler model, which is based on three levels defined by internal and external degrees of freedom of the atom, can explain the observed phenomenon.

Magnetic Resonance Imaging (MRI) is a non-ionizing form of medical imaging which has practical uses in diagnosing, characterizing, and studying diseases in vivo. Current clinical practice utilizes a highly trained radiologist to view MR images and qualitatively diagnose, characterize, or study a disease. There is no easy way to compare qualitative data. That is why developing quantitative measures in MRI show promise. Quantitative measures of disease can be compared across a population, MRI sites, and over time. Osteoarthritis is one disease where those who have it may benefit from the development of quantitative MRI measures. Those benefits may include earlier diagnosis and treatment of the disease or treatment which may halt or even reverse the damage from the disease.The work presented in this dissertation focuses on analyzing and developing new methods of radiofrequency (B1) field mapping to improve quantitative MRI measures. The dissertation opens with an introduction and a brief primer on MRI physics, followed by an introduction to B1 and flip-angle mapping in MRI (Chapters 1-3). Chapter 4 presents a careful statistical analysis of a recent and popular B1 mapping method, the Bloch-Siegert shift (BSS) method, along with a comparison of the technique to other common B1 mapping methods. The statistical models developed in chapter 4 are verified using both Monte Carlo simulation and actual MRI experiments in phantoms. Chapter 5 analyzes and details the potential errors introduced in B1 mapping when a 3D slab-selective excitation is employed. A method for correcting errors introduced by 3D slab-selective B1 mapping is then introduced in chapter 6, along with metrics to quantify the error involved. The thesis closes with a summary of other scientific contributions made by the author in chapter 7. The chapters comprising the bulk of the presented research (4-7) are briefly summarized below. Chapter 4, the statistical analysis of B1 mapping methods, demonstrates the effectiveness of deriving the B1 estimate from the phase of the MR image. These techniques are shown to perform particularly well in low signal-to-noise ratio (SNR) applications. However, there are benefits and drawbacks of each B1 mapping technique. The BSS method deposits a significant amount of radiofrequency (RF) power into the patient, causing a concern that tissue heating may occur. The Phase-Sensitive (PS) method of B1 mapping outperforms the other techniques in many situations, but suffers from significant sensitivity to off-resonance. The Dual-Angle (DA) method is very simple to implement and the analysis is straightforward, but it can introduce significant mean bias in the estimate. No B1 mapping technique performs well for all situations. Therefore, the best B1 mapping method needs to be determined for each situation. The work in chapter 4 provides guidance for that choice. Many B1 mapping techniques rely on a linear relationship between flip angle and transmit voltage. That assumption breaks down when a 3D slab-selective excitation is used. 3D slab-selective excitation is a common technique used to reduce the field-of-view (FOV) in MRI, which can directly reduce scan time. The problem with slab-selective excitation in conjunction with B1 mapping has been documented, but the potential errors in B1 estimation have never been properly analyzed across different techniques. The analysis in chapter 5 demonstrates that the errors introduced in B1 mapping using a slab-selective excitation in conjunction with the ubiquitous DA B1 mapping method can be significant. It is then shown that another B1 mapping technique, the Actual Flip Angle Imaging (AFI) method, doesn't suffer from the same limitation. The analysis presented in Chapter 6 demonstrates that some errors introduced by 3D slab-selective B1 mapping may be modeled and corrected allowing the use of 3D slab-selective excitation to reduce field-of-view, and potentially reduce scan time. The errors are modeled and corrected with a general numerical method using Bloch simulations. The general method is applied to the DA method as an example, but is general and could easily be extended to other methods as well. Finally, a set of metrics are proposed and briefly explored that can be used to better understand the topology and severity of errors introduced into B1 mapping methods. With a better understanding of the errors introduced, the need for correction can be determined. Chapter 7 details other significant ancillary contributions made by the author including: (1) presentation of a new B1 mapping method, the decoupled RF-pulse phase-sensitive B1 mapping method, which has potential for parallel transmit MRI; (2) demonstration of an ultra-short TE method which has potential for imaging Alzheimers brain lesions in vivo; (3) introduction of a new steady-state diffusion tensor imaging technique; (4) phase-sensitive B1 mapping in sodium is demonstrated, a feat not previously demonstrated; (5) a comparison between a dual-tuned and single-tuned sodium coil; (6) introduction of a water- and fat-separation technique using multiple acquisition SSFP; (7) an inter-site and inter-vendor quantitative MRI study is introduced; (8) a relaxation and contrast optimization for laryngeal imaging at 3T is introduced; and (9) diffusion imaging with insert gradients is introduced.