Since the invention of the laser and its first application, there has been an almost continuous
stream of new applications - many of which require specific laser sources. These
applications often require a laser source with a specific power, pulse duration, energy
and wavelength. In some cases these demands are easily met, whilst in others they have
proven rather more difficult to achieve. Traditionally, wavelength versatility has been
limited to the regions for which rare earth or gas gain media are available. These lasers
themselves can be used to generate other wavelengths through the nonlinear processes of
second and third harmonic generation, as well as sum frequency generation. Despite all
of this, there still exists a significant section of the visible and infrared spectrum for which
no convenient sources exist. This thesis is concerned with the development of sources in
these regions along with broadband sources covering significant portions of the spectrum
by themselves.
These new wavelengths are generated in a variety of speciality fibres using either
nonlinear processes or new gain media doped into standard silica fibres. Three types
of speciality fibre are used: low concentration bismuth doped fibre which provides gain
in the 1.0-1.4 μm region; photonic crystal fibres; and very high (75%) concentration
germanium fibres to generate a laser source at 2.1 μm based upon stimulated Raman
scattering. Photonic crystal fibres provide high nonlinearities and controllable dispersion
which enables the generation of broadband supercontinuum sources based upon the
interaction of many nonlinear effects. Each source will be described in depth, with
particular attention given to the underlying physics that gives rise to the source. Previous
and current limitations will be examined and an outlook of the future development of
such sources will be discussed.