Broadband sources have been suggested for use in the interferometric measurement for which cannot easily generate interference fringes with good contrast unless the optical path difference (OPD) is very small. This leads to another problem: how to get the equivalent wavelength of a broadband light source. For a monochromatic source, the OPD between adjacent fringes is one wavelength; but for a broadband light source which has a variety of wavelengths of light, we need to know which wavelength the OPD between adjacent fringes is equal to. The distribution of the Power Spectral Density (PSD) of the broadband source has a great influence on the value of the equivalent wavelength. For a symmetrical PSD, the equivalent wavelength is its center value; for a non-symmetrical PSD, the equivalent wavelength is not a fixed value which is related to the fringe order. For high-precision interferometry, the equivalent wavelength of such a source must be calculated precisely. In this paper, a formula for the equivalent wavelength of a non-symmetrical PSD of a broadband source is deduced. Because of the complexity of the formula, the relationship between optical path difference and equivalent wavelength is not very intuitive so that a lot of simulation calculations have been done. According to those simulation calculations: the equivalent wavelength of the zero-order fringe is the centroid wavelength of the broadband spectrum; within a range of OPD and with the increase of the fringe order, if the peak wavelength of the spectrum is greater than the centroid wavelength, the equivalent wavelength will increase; if the peak wavelength is less than the centroid wavelength, the equivalent wavelength will decrease; the narrower the spectral bandwidth is, the less obvious the change is.
Aluminum reflector, especially OAP (Off-Axis Parabolic) reflector, has been widely used in terahertz and infrared systems for its low cost, lightweight, good machinability, small size, simple structure, and having the same thermal expansion and contraction with the system structure which makes it have a wide temperature adaptability. Thorlabs, Daheng and other large optical components companies even have Aluminum OAP sold on shelf. Most of the precision Aluminum OAP is fabricated by SPDT (single point diamond turing). Affected by intermittent shock, the roughness of aluminum OAP mirrors through conventional single-point diamond lathes is around 7 nm which limits the scope of application for aluminum mirrors, like in the high power density terahertz/infrared systems and visible/UV optical systems. In this paper, a continuous process frock is proposed, which effectively reduces the influence of turning impact on the mirror roughness. Using this process, an off-axis parabolic aluminum reflector with an effective diameter of 50 mm, off-axis angle of 90 degree is fabricated, and the performances are validated. Measurement by VEECO NT1100 optical profiler with 20× objects, the surface roughness achieves 2.3 nm, and the surface figure error is within λ/7 RMS (λ= 632.8 nm) tested by FISB Aμ Phase laser interferometer with the help of a standard flat mirror. All these technical specifications are close to the traditional glass-based reflectors, and make it possible for using Aluminum reflectors in the higher LIDT (laser induced damage threshold) systems and even for the micro sensor of ionospheric for vacuum ultraviolet micro nano satellites.
This paper analyzes the influence on the centroid detection accuracy by several parameters, including the signal-to-noise ratio, frame rate and spot diameter. It provides the selection basis for the camera used in the centroid detection system. The diameter of the spot has little influence on the centroid detection accuracy within a wide range. Meanwhile, with the same signal-to-noise ratio, when the frame rate of the camera becomes higher, the centroid detection accuracy can be improved through the method of superposing more frames. The measurement software uses the specific multi-frame superposition denoising algorithm based on high speed CMOS camera which solves the conflict of improvement of accuracy and time-consuming of the algorithm. The repeatability of the centroid can reach up to 0.0023 pixel with the measuring speed of 62.5fps, the same as the frame rate of the camera.