A tunable wide-field birefringent element (Evans, 1949) provides an efficient means to separate the polrarized magnetic signal from the full width of a selected spectral line. The technique is applicable to observing either longitudinal or transverse magnetic fields.
At the Lockheed Solar Observatory we tested the method for the observation of longitudinal fields. Our device, which we will call the filter magnetograph, is simply a ~ filter tuned to the center of the FeI line at 5324 A, and properly employed in a 7" aperture refractor. We removed the exterior polaroid from the + β’ (half-width) tunable element. This is the thickest element and narrowest in bandwidth. The filter was then mounted so that this 1 A analyzer element faced the sun as the first birefringent element in the filter, immediately following a rotating 88 wave plate.
The zero position of the rotating 2/4 plate was set to tune the ~-A element to 2 + ~ A for left circularly polarized light and to 2-~-% A for right circularly polarized light. The 90 ~ position of the 2/4 plate produced reciprocal tuning of the 71-~ element to ~.-~ ~ for left circularly polarized light and ,~ + 3%4/~ for right circularly polarized light (Figure 1).
Rotating the ~-wave plate back and forth between 0 ~ and 90 ~ made the ~-~ birefringent element act as a filtering device for separating opposite magnetic polarities represented by the frequency-separated polarized components of the Zeeman pattern (Figure le). Use of a polarizing beam-splitter following the 2/4 place could provide simultaneous picture pairs.
The 71-~ birefringent element should provide best filtering for the frequency separation observed in the oppositely polarized components of strong solar magnetic fields in Fei at 5324 A. Our photographs, though under-exposed for such strong fields in sunspot umbrae, clearly show the weaker fields in the sunspot penumbrae and the photospheric network (Figure 2).
The two photographs with intensity differences (Ra+Lb vs La+ Rb in Figure 2) show up as magnetic regions when the two are subtracted (Leighton, 1959). Separating the opposite polarities of transverse fields with the same sensitivity would require use of a birefringent element twice as thick and the phasing of a 89 wave plate through 45 o