Beam-size Measurements on Pep-ii Using Synchrotron-light Interferometry

PEP-II transverse profiles are measured by imaging visible synchrotron light from dipoles in the two rings. The images are broadened by surface errors on the primary extraction mirrors, which have a complex design due to high beam currents. To improve vertical beam-size measurements, we installed a synchrotron-light interferometer, following Mitsuhashi at KEK. In a two-slit interferometer, fine fringes modulate the single-slit pattern from a point source. As the source size increases, the fringe depth decreases, providing a sensitive tool. Because the slits pass light from two stripes along the mirror, we can select the better parts of its surface. By adding a cylindrical lens to image the mirror in the direction perpendicular to the fringes, we can further select short segments of these stripes. In 2000, we put an interferometer on the low-energy ring, at a point 30 m from the BaBar detector, where the beam ellipse is tilted as we compensate for rotation in BaBar’s solenoid. Our interferometer turns to measure the beam tilt. More recently we added a second unit for the high-energy ring. All optics are in the PEP tunnel and allow remote adjustment of the focusing, slit width and separation. 1 SYNCHROTRON-LIGHT INTERFEROMETRY When monochromatic light from a point source diffracts through two parallel slits, the broad fringe pattern from each slit is modulated by fine-scale fringes. Next consider incoherent emission at wavelength 0 λ from a large source S. The emitting points in S have uncorrelated phases, and so the fringes from different points overlap and disappear. Between this case and the point source is a transition region known as partial coherence, in which the modulation depth increases from 0 to 100%. The modulation can then be used to measure the source size [1]. A century ago, Michelson used this technique to measure the angular diameter of a star. In recent years, Toshiyuki Mitsuhashi applied it to measuring beam sizes in storage rings at KEK [2]. To find the extent of this transition, consider the light at points 1 P and 2 P on the two slits. If for all typical points S in the source, the difference in path length is small compared to a wavelength, 1 2 0 SP SP λ − << , (1) then 1 P and 2 P remain correlated. This remains true for a source that is not strictly monochromatic, provided that this path difference is small compared to the coherence length for the bandwidth ν ∆ :