Many commercial wavefront-measurement interferometers rely on optical phase shifting to provide wavefront information at every point in the interferogram (the old-fashioned method of deriving wavefront shape from interference-fringe positions provides less information, especially for high-spatial-frequency wavefront components). Optical phase shifting is also important for other forms of interferometry, including stellar interferometry and phase-measuring profilometry.
Phase shifters usually are based on mirrors that are mechanically translated, often by piezoelectric actuators; other types of phase shifters rely on different principles but also require a mechanically moved optical element. Researchers at Tianjin University (Tianjin, China) and the University of Wollongong (Wollongong, Australia) have now developed an acousto-optic (AO) phase shifter that requires no position shifting of optical elements.1
Acousto-optic crystals are often used in optical systems to modulate, frequency shift, or diffract a laser beam. In the case of frequency shifting, the beam interacts with an acoustic wave that moves inside the crystal, Bragg-reflecting from the wave. The frequency of the reflected beam is the sum of the frequency of the original beam and the frequency of the acoustic wave. Depending on its direction of motion, the acoustic wave can contribute either a positive or negative frequency component.
In the new AO phase shifter, the beam to be phase shifted reflects first from one AO crystal that adds a radio-frequency (RF) component, then from a second AO crystal that subtracts an RF component of the same magnitude, restoring the beam’s original frequency. The phase delays between the two RF signals can be varied, adding a controllable phase shift to the beam.
Fringes are produced by an interferometric setup that includes an acousto-optic phase shifter. A phase shift is introduced (bottom) that is 90° from the starting phase (top).
In the experimental setup, a 5-mW beam from a HeNe laser was split into two interferometric paths. One of the resulting beams was reflected by two lead molybdate AO crystals; the first crystal induced a positive frequency shift, while the second caused a negative shift. The first-order-diffraction (Bragg-reflection) efficiency of the crystals was about 80%. The two beams were then made to interfere by passing them through a converging lens; a microscope objective projected the resulting interference fringes onto a CCD detector (see figure). Measuring the location of the fringe centers provided a measurement of phase shift.
The 65-MHz frequency of the RF signals resulted in a 15.8-ns period (the relative time delay needed to produce a 360° phase shift). A programmable digital-delay line with an incremental adjustment of 0.25 ns produced an optical phase shift adjustable in increments of approximately 6°. “The resolution can be easily improved by designing an electronic device to generate two-channel RF signals with finer control of the phase delay between the two signals,” says Enbang Li, one of the researchers.
Optical-path variations in the span between the crystals, induced by thermal or other fluctuations, could introduce errors in the phase shift; although the researchers have not yet quantified the setup’s long-term stability, experiments are under way. Li notes that his group has already used the AO setup for phase-measuring profilometry of 3-D shapes.
REFERENCE
1. Enbang Li et al., Optics Lett. 30, 189 (2005).