Optical methods can provide useful tissue characterization tools. For this project, two-photon microscopy and polarized light examinations (polarimetry) were used to assess the organizational state of myocardium in healthy, infarcted, and stem-cell regenerated states. Two-photon microscopy visualizes collagen through second-harmonic generation and myocytes through two-photon excitation autofluorescence, providing information on the composition and structure/organization of the tissue. Polarimetry measurements yield a value of linear retardance that can serve as an indicator of tissue anisotropy, and with a dual-projection method, information about the anisotropy axis orientation can also be extracted. Two-photon microscopy results reveal that stem-cell treated tissue retains more myocytes and structure than infarcted myocardium, while polarimetry findings suggest that the injury caused by temporary ligation of a coronary artery is less severe and more diffuse that than caused by a permanent ligation. Both these methods show potential for tissue characterization.

Biomedical imaging requires high resolution to see the fine features of a sample and fast acquisition to observe live cells that move. Optical coherence tomography (OCT) is a powerful technique which uses optical interference for non-invasive high resolution 3D imaging in biological samples. The resolution of OCT is determined by the length over which the light used will in- terfere. Unfortunately, dispersion hurts the imaging resolution by broadening interference features. A technique called quantum-OCT (QOCT)[1] is immune to dispersion but re- quires entangled photon pairs. The need for entanglement drastically reduces the number of photons available for imaging, making QOCT too slow to be practical. Chirped-pulse interferometry (CPI) is also immune to dispersion. A chirped pulse is one where the fre- quency, or colour, of the light changes from red to blue from one end of the pulse to the other. CPI relies on frequency correlations created by applying different chirps to two sep- arate pulses. This method had the disadvantage of being limited to a single predetermined chirp rate, and discarded 50% of the power. However CPI has better resolution than OCT, automatic dispersion cancellation, and 10,000,000 times the signal strength of QOCT [13]. A new, much more flexible and efficient method of CPI will be demonstrated by creating the frequency correlations entirely in a single pulse. This new method is referred to as non- linear chirped pulse interferometry (NL-CPI). The non-linear chirp required in NCPI is very difficult to produce using only conven- tional optics. In this thesis we document the construction and characterization of a new method of creating the desired chirp using a programmable pulse-shaper (PS). We build a PPS and then demonstrated its functionality by compressing a 105nm FWHM bandwidth pulse to under 17f s, near its transform limited time duration. We also show that the values given to the PPS for dispersion are accurate by calculating and then compensating the dispersion caused by various optical elements in the CPI interferometer. Conventional OCT systems are immune to dispersion common to both arms of the interferometer. Non-linear interferometers experience broadening due to this dispersion, making them more difficult to use with fibre based interferometers common in conventional OCT. We show that NL-CPI can compensate for dispersion common to both arms of the interferometer, making NL-CPI more appealing as a replacement for conventional OCT. In this thesis we experimentally implement and demonstrate a prototype setup using non-linear CPI for dispersion-cancelled imaging of a mirror, with a resolution comparable to conventional OCT systems. We then use the system to produce 2-D cross sectional images of a biological sample, an onion. Q-OCT has previously been used to image an onion[16], but required treating the onion with gold nano particles to achieve a useful signal. The onion we used had no special treatment. In addition our axial scanning rate is also 10000 times faster than Q-OCT.

This dissertation reports the results of optical studies of epitaxial (e), polycrystalline (p) and amorphous (a) Ge2Sb2Te5 (GST) thin films. The dynamic properties of GST films in all three (e/p/a) phases were investigated by a time-resolved optical pump-probe technique in which a femtosecond pump pulse of 55 fs duration was used to excite the sample. The intensity and polarization of the reflected probe beam respectively provide information about the transient reflectance (R) and anisotropic reflectance (AR) induced in the sample, that in turn provide the information about the crystal structure, phonon spectrum, and ultimately phase transitions within the sample. The study of an epitaxial sample provides an opportunity to explore the character of the modes within the phonon spectrum. The epitaxial GST film was grown upon a homoepitaxial layer of GaSb grown upon a GaSb wafer. We observed a 6.7 THz coherent optical phonon (COP) in GaSb(001). The dependence of the signal strength upon the pump and probe polarization was explained in terms of a model that considered both Transient stimulated Raman Scattering (TSRS) and the action of a Surface Space-Charge (SSC) field. The presence of the 6.7 THz transverse COP in the AR channel and its four fold dependence on pump and probe polarization suggests a three-dimensional T2 character. The COP amplitude was maximum when the probe was polarized parallel to the cube edge (GaSb[100]) and the pump polarization was set parallel to a face diagonal (GaSb[110]). The results were fully understood using a microscopic model of selective bond breaking. The AR response of e-GST/GaSb(001) reveals the presence of a 3D 3.4 THz transverse optical phonon. The mode amplitude was independent of pump polarization indicating that the mode is excited by a SSC field. This SSC field could exist within the GST, if the distorted rock-salt structure of GST lacks inversion symmetry, or GaSb, which has the non-centrosymmetric zincblende structure, leading to impulsive excitation of phonons at the GST/GaSb interface. The mode in GST was inferred to be T2-like. The observation of a T2-like phonon mode confirms that GST is cubic in structure and challenges previous studies where 1D or 2D character was assigned to the 3.4 THz mode. While pump-probe measurements displayed the presence of a 3D 3.4 THz mode in the AR response of e-GST/GaSb(001), a 4.5 THz mode was observed in both R and AR channels for p-GST(37 nm)/Si(001) and a-GST(57 nm)/Si(001). The mode character was identified to be either of A or E type by comparing the frequency with frequencies reported in the literature. Additional Raman microscope measurements confirmed the presence of the modes observed in the pump-probe measurements and also revealed additional frequencies. The differences in the frequencies observed from the different samples are quite small suggesting the presence of similar bonds that are modified to some extent by the different structural environment found within each sample. After exposure to high pump fluence the original modes disappeared and were replaced by new modes with frequencies at 4.2 THz and 3.1 THz in e-GaSb, 4.2 THz in e-GST, 3.5 THz in p-GST and 3.6 THz in a-GST. The difference in the final frequencies observed for p and a-GST sample may result from the difference in stack structure affecting the time-dependent temperature profile in each sample. The dependence of the temperature profile on the sample stack was understood from an experimental study of the phase transition between the amorphous and crystalline states induced by exposure to a series of amplified laser pulses. The dependence of the crystalline area and its reflectivity upon the number of pulses and fluence was described using a simple algebraic model. The results justify the assumption of one-dimensional heat flow. The growth velocity of the crystalline region was calculated to be 7-9 m/s. Apparatus and methods were developed to extend the time-resolved optical studies described previously. Firstly, an apparatus was constructed for the measurement of the wavelength dependent sample reflectance with a white-light pulse. A reference arm was employed to allow normalization and hence removal of the intensity noise arising in the laser regenerative amplifier system. Secondly an electrical measurement apparatus was constructed to allow combined electro-optical measurements in future. Switching of GST vertical memory cells was successfully demonstrated. The cells were fabricated on a borosilicate substrate with TiW top and bottom electrodes. A DC voltage of 4.5 to 6 V was required to induce switching, while in pulsed measurements, the device demonstrated switching in response to a pulse with minimum duration of 100 ns.

Biomedical imaging requires high resolution to see the fine features of a sample and fast acquisition to observe live cells that move. Optical coherence tomography (OCT) is a powerful technique which uses optical interference for non-invasive high resolution 3D imaging in biological samples. The resolution of OCT is determined by the length over which the light used will in- terfere. Unfortunately, dispersion hurts the imaging resolution by broadening interference features. A technique called quantum-OCT (QOCT)[1] is immune to dispersion but re- quires entangled photon pairs. The need for entanglement drastically reduces the number of photons available for imaging, making QOCT too slow to be practical. Chirped-pulse interferometry (CPI) is also immune to dispersion. A chirped pulse is one where the fre- quency, or colour, of the light changes from red to blue from one end of the pulse to the other. CPI relies on frequency correlations created by applying different chirps to two sep- arate pulses. This method had the disadvantage of being limited to a single predetermined chirp rate, and discarded 50% of the power. However CPI has better resolution than OCT, automatic dispersion cancellation, and 10,000,000 times the signal strength of QOCT [13]. A new, much more flexible and efficient method of CPI will be demonstrated by creating the frequency correlations entirely in a single pulse. This new method is referred to as non- linear chirped pulse interferometry (NL-CPI). The non-linear chirp required in NCPI is very difficult to produce using only conven- tional optics. In this thesis we document the construction and characterization of a new method of creating the desired chirp using a programmable pulse-shaper (PS). We build a PPS and then demonstrated its functionality by compressing a 105nm FWHM bandwidth pulse to under 17f s, near its transform limited time duration. We also show that the values given to the PPS for dispersion are accurate by calculating and then compensating the dispersion caused by various optical elements in the CPI interferometer. Conventional OCT systems are immune to dispersion common to both arms of the interferometer. Non-linear interferometers experience broadening due to this dispersion, making them more difficult to use with fibre based interferometers common in conventional OCT. We show that NL-CPI can compensate for dispersion common to both arms of the interferometer, making NL-CPI more appealing as a replacement for conventional OCT. In this thesis we experimentally implement and demonstrate a prototype setup using non-linear CPI for dispersion-cancelled imaging of a mirror, with a resolution comparable to conventional OCT systems. We then use the system to produce 2-D cross sectional images of a biological sample, an onion. Q-OCT has previously been used to image an onion[16], but required treating the onion with gold nano particles to achieve a useful signal. The onion we used had no special treatment. In addition our axial scanning rate is also 10000 times faster than Q-OCT.

Στην παρούσα εργασία μελετάται η τρίτης τάξης μη γραμμική απόκριση νανοσωματιδίων παλλαδίου σε συμπολυμερή όπως επίσης και κβαντικών ψηφίδων ιωδιούχου μολύβδου. Ο προσδιορισμός της μη γραμμικής επιδεκτικότητας τρίτης τάξης και η υπερπολωσιμότητα δεύτερης τάξης επιτυγχάνεται με την βοήθεια των τεχνικών Z-scan και οπτικού φαινομένου Kerr (OKE) χρησιμοποιώντας παλμούς λέιζερ 35 psec και 4 nsec με μήκος κύματος διέγερσης 532 nm και 1064 nm. / In this work the third order nonlinear optical response of palladium nanoparticles encapsulated in block copolymers has been investigated as well as lead iodide quantum dots. The third-order nonlinear susceptibility χ(3) and the second hyperpolarizability γ are determined by employing Z-scan and Optical Kerr Effect (OKE) techniques using 35 ps and 4 ns laser pulses at excitation wavelengths of 532 and 1064 nm.

I developed a high power blue laser for use in scientific and technical applications (eg. precision spectroscopy, semiconductor inspection, flow cytometry, etc). It is linearly polarized, single longitudinal and single transverse mode, and a convenient fiber coupled continuous wave (cw) laser source. My technique employs external cavity frequency doubling and provides better power and beam quality than commercially available blue diode lasers. I use a fiber Bragg grating (FBG) stabilized infrared (IR) semiconductor laser source with a polarization maintaining (PM) fiber coupled output. Using a custom made optical and mechanical design this output is coupled with a mode matching efficiency of 96% into the doubling cavity. With this carefully designed and optimized cavity, measurements were carried out at various fundamental input powers. A net efficie ncy of 81 % with an output power of 680 mW at 486 nm was obtained using 840 mW of IR input. Also I report an 87.5 % net efficiency in coupling of blue light from servo locked cavity into a single mode PM fiber. Thus I have demonstrated a total fiber to fiber efficiency of 71% can be achieved in our approach using periodically poled potassium titanyl phosphate (PPKTP). To obtain these results, all losses in the system were carefully studied and minimized.

This thesis presents the experimental work performed on the spectrin protein. The aim of the work was to study the direct interactions of spectrin, the cytoskeleton of RBCs, with membrane lipid to determine its effects on the mechanical properties of the lipid bilayer. Motivation for this work came from a lack of unanimity in the field of spectrin, and the hypothesized potential of the protein to perforate giant unilamellar vesicles. The work aimed to investigate and determine how spectrin-lipid interactions influence membrane mesoscopic morphology and biophysics in ways that could ultimately be important to cellular function. For this purpose, a protocol was implemented to take into account the different aspects of the binding. Direct visualisation of the spectrin-lipid interaction and distribution was achieved using confocal fluorescence microscopy. Changes in the mechanical properties of the membrane were investigated using the micropipette aspiration technique. Finally the thermodynamics of the interaction were considered with isothermal titration calorimetry experiments. This allowed evaluation of the protein-lipid interaction in a complete and coherent manner. Experiments were also performed on another elastic protein, alpha-elastin, for comparison. In addition to its similarities with spectrin (both possess hydrophobic domains and entropy elasticity), elastin is auto-fluorescent which makes it an attractive model protein. Elastin was also used as a sample model to implement new techniques using nonlinear optics microscopy.

The aim of the project described in this thesis is to synthesise new functional molecules which interact with light for neurophysiological applications. In particular, I describe a family of amphiphilic porphyrins with large first hyperpolarisabilities which are used as SHG contrast agents and voltage-sensitive probes. In addition I detail a methodological microscopy tool and a novel caged form of a neuronal ion-channel antagonist. Chapter 1 introduces the key concepts underlying the use of dyes as SHG contrast agents. In particular it focuses on aspects of molecular design, covering both the amphiphilicity and nolinearity required by the target molecule. It covers quantification of the nonlinear properties of SHG stains, then surveys a number of examples which showcase the flexibility of SHG imaging as a biomedical technique. Chapter 2 describes a family of amphiphilic porphyrins with large first hyperpolarisabilities. Working from the structure-property relationships identified in Chapter 1, we fully characterise these dyes and demonstrate that they can be used in SHG imaging. We demonstrate that these molecules may also be tuned by complexation of a metal ion which can modulate their photophysical and solubility behaviour. Chapter 3 provides a description of how to determine the orientational distribution of dipolar dyes in a membrane by multiphoton microscopy. We measure the signal intensity of the dye in a model membrane system then find distributional moments which lead to the distribution itself. Chapter 4 explores whether off-axis contributions to the first hyperpolarisability tensor can significantly augment the dominant on-axis contribution from the main dipolar charge-transfer band. We synthesise and characterise a series of cis-donor cis-acceptor porphyrin compounds and explore their biophysical characteristics. Chapter 5 is the culmination of this project and after discussing method development, goes on to show how we measure the voltage sensitivity of an amphiphilic porphyrin SHG dye. We compare the archetypal porphyrin dye chromophore with three commercially available styryl dyes and demonstrate that our dye has greater sensitivity and a more rapid response. Chapter 6 describes a side project, the use of a photolabile cage to protect MK801, a neuronal ion-channel antagonist. By developing a water soluble photolabile cage using molecular design techniques, we are able to release MK801 in neurons with precise spatiotemporal control, allowing us to pinpoint the locus of two key neurophysiological processes.

L'interaction d'une impulsion laser intense (~10¹⁴ W /cm²) et de courte durée (femtoseconde) avec un gaz rare induit une polarisation hautement non-linéaire dans le domaine spectral XUV; les harmoniques d'ordre élevés. En raison des propriétés spécifiques du rayonnement harmonique et de ses applications, cette thématique est particulièrement riche et fertile. La production efficace d'harmoniques d'ordres élevés repose à la fois sur la réponse non-linéaire de l'atome unique et un comportement collectif.Le fil directeur des études présentées dans cette thèse est la compréhension et le contrôle de l'accord de phase ou du quasi accord de phase en présence d'une ionisation substantielle du gaz générateur. Dans ce contexte, nous montrons l'importance de la longueur de cohérence sur l'accord de phase en génération d'harmoniques. Nous étudions sa dépendance en fonction de la focalisation du laser, de la pression mais aussi sa dépendance temporelle liée à l'ionisation, effet que nous avons mis en évidence lorsqu'on a cherché à optimiser une double impulsion harmonique. Le travail de développement, sur la station LASERIX, de la source à double impulsion harmonique générée à partir d'un même milieu gazeux et avec un délai picoseconde variable est présenté. Cette source possède un véritable potentiel d'applications scientifiques, injectée dans un milieu amplificateur plasma qu'on appelle laser X, la double impulsion permettra de sonder la réponse temporelle de ce type de milieu. Par ailleurs, des expériences et des simulations menées sur la génération d'harmoniques en propagation guidée visent ainsi à étendre les spectres harmoniques vers les courtes longueurs d'ondes, zone spectrale pour laquelle le laser X à plasmas est émis. Ceci donnera l'accès à une source offrant des caractéristiques complémentaires des lasers X, sources développées en parallèle sur la station LASERIX. / The interaction of an intense laser pulse of short duration with a rare gas induces a highly non-linear polarization in the XUV spectral range: the high order harmonics. Due to the specific properties of the harmonic radiation and its applications, this issue is particularly rich and fertile. The efficient production of high order harmonics is based both on the non-linear response of the single atom and on collective behavior.The principle of the research presented in this thesis is the understanding and control of phase matching or quasi-phase matching in the presence of substantial ionization in the generating gas. In this context, we show the importance of the coherence length on the phase matching in High harmonic generation. We study its dependence on laser focusing, pressure but also its time dependence related to ionization. Moreover, experiments and simulations aim at extending harmonic spectra towards shorter wavelengths, a spectral range for which the X Ray Laser is emitted. This will give access to a source with complementary characteristics as regards to X-ray lasers. This source shall be developed in parallel on the LASERIX station or injected in soft X-ray laser amplifiers.

En complément aux méthodes traditionnelles d’observation et de stimulation en neuroscience, l’optogénétique, combinant l’expression ciblée de protéines photosensibles dans les neurones et l’utilisation de nouvelles techniques de microscopies, a connu un essor important ces dernières années. Ce nouveau procédé permet d’enregistrer de manière non invasive les signaux fonctionnels de circuits intacts tels que les changements de potentiel de membrane ou de concentration intracellulaire de calcium mais également de moduler l’excitabilité des neurones. Pour illuminer ces protéines photosensibles, de nouvelles méthodes de microscopie ont été développées. En particulier, afin d’obtenir une résolution spatiale optimale au sein d’un tissu biologique, il devient nécessaire d’utiliser l’illumination bi-photonique et d’utiliser des techniques permettant la mise en forme du faisceau lumineux pour s’adapter à la morphologie des circuits ou même des neurones étudiés.Au cours de ma thèse, j’ai développé une combinaison de méthodes optiques (associant le contraste de phase généralisé avec la focalisation temporelle) afin d’activer le canal cationique channelrhodopsin-2 en excitation bi-photonique. Ce travail a démontré, pour la première fois, l’activation simultanée de potentiels d’action dans plusieurs cellules tout en conservant une résolution axiale à l’échelle cellulaire (~10 μm).La mise en forme du faisceau lumineux semble très avantageuse pour améliorer la spécificité de l’activation. Il restait à démontrer que les faisceaux ainsi modulés conservaient leur intégrité spatiale en se propageant à l’intérieur de tissus biologiques diffusants. J’ai donc étudié la propagation de faisceaux lasers modulés par les techniques du contraste de phase généralisé et de l’holographie numérique en combinaison avec la focalisation temporelle. L’utilisation de la focalisation temporelle permet aux volumes d’excitation de rester confinés sur l’axe de propagation comme observé précédemment, mais aussi de reconstruire un profil d’excitation en profondeur dans le tissu, qui correspond au profile généré sans milieu diffusant. Cet effet est plus important pour le contraste de phase généralisé que pour l’holographie numérique et se dégrade en fonction de la profondeur à laquelle l’activation a lieu. J’ai démontré pour la première fois, l’activation en profondeur (> 200 μm) de neurones grâces à ces méthodes.Enfin, j’ai testé les mêmes techniques d’illumination sur d’autres protéines photosensibles, telles que la C1V1 et l’halorhodopsin. Après avoir établi les spectres d’activation afin de trouver la longueur d’onde optimale pour l’activation bi-photonique, j’ai exprimé ces protéines dans des tranches de cerveaux. Les deux protéines requièrent une activation à 1040 nm à la limite du laser Ti:Sapphire utilisé dans de nombreux laboratoires biologiques. La C1V1 a généré des courants similaires à la ChR2 en terme d’amplitude tout en conservant la lente cinétique de fermeture caractéristique de ce canal. L’halorhodopsin, quant à elle, reste difficile à activer avec de faibles courants et ne permet pas une inhibition sélective de trains de potentiels d’action. Ce problème est probablement dû à un faible taux d’expression observé dans les neurones étudiés et serait peut-être résolu en changeant de construction virale. / Optogenetics relies on the genetically targeted expression of light sensitive proteins in specific cell populations. This novel field has had a large impact in neuroscience, allowing both monitoring and stimulating the activity of specific neuronal populations, in intact brain preparations. Optogenetic tools have been used to record functional signals, such as changes in membrane potential or intracellular calcium concentration, as well as to modulate the excitability of neurons. To fully exploit the potentiality of optogenetics, new microscopy techniques have been developed to optimize illumination of photo-active compounds in situ. In particular, an important effort has been directed towards improving the spatial and temporal resolution of light stimulation, in order to match the dynamics of physiological processes. In this frame, the use of two-photon excitation becomes necessary to ensure penetration of light in scattering biological tissues, as well as confining the excitation volume and improve the specificity of illumination. My thesis was dedicated to the development and use of advanced optical methods for two-photon excitation of optogenetic tools. In a first project, we combined optical approaches (generalized phase contrast and temporal focusing) to perform two-photon activation of neurons expressing the light-sensitive cationic channel channelrhodopsin-2 (ChR2). Our work demonstrated for the first time the simultaneous generation of action potentials in multiple neurons, while maintaining a micrometric axial and lateral resolution. These results pointed out the advantages of light sculpting to increase both the specificity and the flexibility of photo-stimulation.In order to investigate the potential of this technique for efficient in-depth stimulation, we therefore studied the propagation through scattering biological media of laser beams generated by two different light patterning techniques, generalized phase contrast and digital holography in combination with temporal focusing. We demonstrated that temporal focusing enabled the excitation volumes to maintain micrometric axial confinement, as well as to maintain well defined patterns deep inside tissues. We also demonstrated for the first time the activation of ChR2 at depth over 200 μm.Finally, the last part of my PhD was focused on testing light patterning methods for the activation of two other photosensitive proteins, the excitatory channel C1V1 and the inhibitory pump, halorhodopsin.