We present a study of Lyot style ( i.e. , classical , band-limited , and Gaussian occulter ) coronagraphy on extremely large , highly-segmented telescopes .
We show that while increased telescope diameter is always an advantage for high dynamic range science ( assuming wavefront errors have been corrected sufficiently well ) , segmentation itself sets a limit on the performance of Lyot coronagraphs .
Diffraction from inter-segment gaps sets a floor to the achievable extinction of on-axis starlight with Lyot coronagraphy .
We derive an analytical expression for the manner in which coronagraphic suppression of an on-axis source decreases with increasing gap size when the segments are placed in a spatially periodic array over the telescope aperture , regardless of the details of the arrangement .
A simple Lyot stop masking out pupil edges produces good extinction of the central peak in the point-spread function ( PSF ) , but leaves satellite images caused by inter-segment gaps essentially unaffected .
Masking out the bright segment gaps in the Lyot plane with a reticulated mask reduces the satellite images ’ intensity to a contrast of 5 \times 10 ^ { -9 } on a 30 m telescope with 10 mm gaps , at the expense of an increase in the brightness of the central peak .
The morphology of interesting targets will dictate which Lyot stop geometry is preferable : the reticulated Lyot stop produces a conveniently uni-modal PSF , whereas a simple Lyot stop produces an extended array of satellite spots .
A cryogenic reticulate Lyot stop will also benefit both direct and coronagraphic mid-IR imaging .