Confocal microscopes utilize point illumination and pinhole detection to reject out-of-focus light. Because of the point illumination and detection pinhole, confocal microscopes typically utilize point scanning for imaging, which limits the overall acquisition speed. Due to the excellent optical sectioning capabilities of confocal microscopes, they are excellent tools for the study of three-dimensional objects at the microscopic scale. Fluorescence confocal microscopy is especially useful in biomedical imaging due to its high sensitivity and specificity. However, all designs for confocal microscopes must balance tradeoffs between the numerical aperture (NA), field of view (FOV), acquisition speed, and cost during the design process. In this dissertation, two different designs for an array confocal microscope are proposed to significantly increase the acquisition speed of confocal microscopes. An array confocal microscope scans an array of beams in the object plane to parallelize the confocal microscope to significantly reduce the acquisition time. If N beams are used in the array confocal microscope, the acquisition time is reduced by a factor of N. The first design scans an array of miniature objectives over the object plane to overcome the trade-off between FOV and NA. The array of objectives is laterally translated and each objective scans a small portion of the total FOV. Therefore, the number of objectives used in the array limits the FOV, and the FOV is increased without sacrificing NA. The second design utilizes a single objective with a high NA, large FOV, and large working distance designed specifically for whole brain imaging. This array confocal microscope is designed to speed up the acquisition time required for whole brain imaging. Utilizing an objective with a large FOV and scanning using multiple beams in the array significantly reduces the time required to image large three-dimensional volumes. Both array confocal microscope designs use beam-splitting gratings to efficiently split one laser beam into a number of equal energy outgoing beams, so this dissertation explores design methods and analyses of beam-splitting gratings to fabrication errors. In this dissertation, an optimization method to design single layer beam-splitting gratings with reduced sensitivity to fabrication errors is proposed. Beam-spitting gratings are typically only designed for a single wavelength, so achromatic beam-splitting grating doublets are also analyzed for possible use in array confocal microscopes with multiple excitation wavelengths. An analysis of the lateral shift between grating layers in the achromatic grating doublet proves grating profiles with constant first spatial derivatives are significantly less sensitive than continuous phase profiles. These achromatic grating doublets have designed performance at two wavelengths, but the diffraction angles at the two wavelengths differ. To overcome that limitation, scale-invariant achromatic gratings are designed, which not only provide designed performance at two wavelengths, but also equal diffraction angles at two wavelengths.