Page 29 - Tunable Lasers Handbook
P. 29

12      F. J. Duarte

                    where the output is coupled via the reflection losses of the grating, at a cost of
                    higher amplified spontaneous emission (ASE) levels [2,18,24]. In addition to the
                    information given in Table 1, this class of oscillator design has also been applied
                    to optical parametric oscihtors [32] (see Chapter 6).
                        Class I1  oscillators incorporating multiple-prism beam  expanders are,  in
                    general, more efficient than  pure  grazing-incidence designs but  they  are  also
                    more complex. In Fig. 2, MPL oscillators using multiple-prism beam expanders
                    deployed in (+,+,+,-)  and (+,-,+,-)  configurations are illustrated. In a (+,+,+,-)
                    configuration, the first three prisms are deployed in  an  additive configuration
                    with the fourth prism deployed in a compensating mode to neutralize the cumu-
                    lative dispersion of the first three prisms. In a (+,-,+,-)   configuration, two pairs
                    of compensating prisms are utilized [ 1,2]. These configurations are used to yield
                    zero dispersion at a wavelength of  design thus reducing beam deviations due to
                    (an/&") factors and leaving the tuning characteristics of  the oscillator dependent
                    on the grating. Extensive details on multiple-prism design have been given by
                    Duarte  [l] and relevant mathematical formulas are given in a later section on
                    intracavity dispersion. The main  design constraint is to provide the  necessary
                    beam expansion to achieve total illumination of the grating at a maximum trans-
                    mission efficiency and a minimum intracavity length.
                        The intrinsic intracavity dispersion of  a grazing-incidence grating design is
                    higher than the dispersion achieved by an MPL grating configuration.  A configu-
                    ration that provides higher intracavity dispersion than MPL designs and higher
                    conversion efficiency than pure grazing-incidence cavities is the HMPGI grating
                    cavity mentioned earlier [20,24] (Fig. 2c). In HMPGI oscillators the grating is
                    deployed in a near grazing-incidence configuration that is far more efficient than
                    a pure grazing-incidence configuration [24] (see Section 9). Further, because the
                    required intracavity beam expansion is far less than that typical of MPL oscilla-
                    tors, efficient and compact multiple-prism expanders can be readily designed to
                    provide the necessary intracavity preexpansion. Today, HMPGI oscillators are
                    widely used in research and commercial tunable laser systems.
                        Improved oscillator designs use a polarizer output coupler rather than a tra-
                    ditional mirror as the output coupler [23,33] (see Fig.  2). The output-coupler
                    polarizer is  made of  a Glan-Thompson  polarizer with  an antireflection-coated
                    inner surface and an outer surface that is coated for partial reflection. Dispersive
                    oscillators  incorporating  multiple-prism grating  assemblies yield  strongly p-
                    polarized narrow-linewidth emission [1,2,20]. In this context, the function of the
                    output-coupler polarizer is to suppress single-pass unpolarized ASE in high-gain
                    lasers. Thus, the use of a polarizer output coupler in dispersive dye laser oscilla-
                    tors has yielded extremely low levels of ASE in the  10-7  to 10-9 range [22,23].
                    The Glan-Thompson polarizer output coupler is illustrated in Fig. 3.
   24   25   26   27   28   29   30   31   32   33   34