Study of ferroelectric materials in confined geometries

Nature exhibits an enormous array of physical processes which exist on various length scales. The length scale can significantly alter the physical properties of a material. As an example, materials which are ferroelectric in the bulk may not be ferroelectric below a critical size at the nanoscale. In addition, the size can also affect the Curie temperature, the temperature above which a material is no longer ferroelectric.Structures can exist in zero dimensions, such as nanoparticles, in one dimension, such as nanorods or crystal lines in glass, in two dimensions, such as thin films, and in three dimensions, such as bulk single crystals. Creating structures in each of these dimensions allows a wide range of capabilities. Nanocrystals have found use in creating hybrid photovoltaics. Other applications of low dimensional structures include the use of thin films as strain sensors or crystal lines in glass as optical waveguides. In this work, the femtosecond (fs) laser-induced growth of crystals in glass has been explored due to the potential use of these crystals in glass for optical data transmission. In particular, lithium niobate (LiNbO$_3$) is of special interest due to its favorable properties. Bulk LiNbO$_3$ substrates have found widespread applications in the creation of optical modulators, frequency converters, and acousto-optic filters. LiNbO$_3$ crystals in glass have previously been fabricated on both the surface of glass via continuous wave (CW) laser irradiation and within glass via fs laser irradiation. Previous works in forming LiNbO$_3$ crystals in glass, however, have shown that a significant problem remains where the fs laser precipitated crystals in glass possess a noticeably non-uniform, polycrystalline structure. These results point to a need to precisely control the interplay of nucleation and growth of crystals within the dynamic heating profile induced by the laser. It requires a careful optimization of fs laser parameters and glass composition. Taking into consideration the material constraints and desired application for the crystal, one could then select between various crystal growth modes such as an all-solid-state glass $\rightarrow$ crystal transformation produced by a heat profile that yields a convex crystal growth front ahead of the scanning laser focus or a melt $\rightarrow$ solid transformation that occurs with a concave crystal growth front behind the scanning laser focus.This work demonstrates the successful formation of single crystal LiNbO$_3$ in lithium niobosilicate (LNS) glass of effectively unlimited length, which is constrained only by the homogeneity of the starting glass. The specific growth dynamics and the confined nature of the crystal growth method lead to new phenomena in terms of crystal orientation that can be studied in this high quality, highly oriented single crystal line: (1) After nucleation, the crystal lattice rotates until its c-axis is oriented along the laser scanning direction. (2) Once the crystal is oriented in this way, there is a gradually varying misorientation of the crystal axis that is symmetric in regards to the center of the crystal line cross-section. The latter observation indicates that the parameters have been controlled such that the growth occurs upon heating in a convex growth front ahead of the scanning laser focus. In addition, the latter observation is expanded upon by detailing the potential application of these crystal lines as graded index crystal waveguides in glass.The lattice misorientation and rotation within the crystals in glass were systematically analyzed with regard to laser power, laser scanning speed, and composition. There is an indication that the misorientation rate changes with laser scanning speed, which can be attributed to the elongation of the temperature profile with faster scanning speed. The rotation rate was correlated more weakly to the processing parameters. There may be an indication of the glass composition with the least glass former showing the highest proclivity towards nucleating a crystal near c-axis orientation. A systematic study of the effect of glass composition on fs laser-induced crystal growth revealed that modifying the glass composition had a considerable impact on the crystal growth rate. With increasing amount of glass former, the crystal growth rate reduced substantially. This was evidenced by a polycrystalline growth mode, where the high nucleation rate of the glass allowed the formation of a crystal, but the crystal growth rate could not keep up with the laser scanning speed, resulting in a polycrystalline line with a series of crystallites on the order of microns. In the glass with the higher amount of glass former, this polycrystalline growth mode was reached at significantly slower laser scanning speeds.Combined excitation emission spectroscopy revealed that Er incorporated into fs laser-induced LiNbO$_3$ in Er-doped LNS glass. This result shows that both fs and CW laser-induced crystallization of Er-doped LNS glass can lead to the formation of Er-doped LiNbO$_3$ crystals. This is a promising step towards the realization of multi-functional single crystals in glass within integrated optical devices. A zero-dimensional system was also studied in this work. In exploring batches of lithium niobate nanocrystals made from different initial Li to Nb ratios ($\rho$) in the synthesis step, a considerable variation in the final nanocrystal product as a function of $\rho$ was observed. For values of $\rho < 50 \%$, nanocrystals hardly formed, resulting in large non-uniform clumps. The nanocrystal product improved with increasing $\rho$, becoming a more homogeneous collection of spherical nanocrystals. The nanocrystal stoichiometry was found to depend on the initial $\rho$ in the synthesis step. This was identified by the peak widths of a Raman mode in the nanocrystals relative to a calibration from bulk single crystal Raman spectra. Batches made with $\rho \geq 55 \%$ were found in this way to be stoichiometric LiNbO${_3}$ (that is, the Li to Nb ratio is $1$ to $1$). Thus, nanocrystals from these batches have presumably the least defect content as would result from lithium deficiency. Batches with $\rho \geq 50 \%$ and $\rho < 55 \%$ were found to have a composition similar to that of congruent bulk single crystal lithium niobate (i.e. somewhat Li deficient). The Raman spectra also indicated the formation of an additional phase, thought to be LiNb${_3}$O${_8}$, in batches with $\rho \geq 55 \%$. Raman spectroscopy also allowed for the identification of the predominant orientation of the liquid suspended nanocrystals within the optical trap. Nanocrystals principally oriented with their c-axis parallel to the laser propagation direction. This is due to the strong interaction of the laser polarization with the dipoles induced in the direction perpendicular to the c-axis.