在以多级微反射镜为核心器件的静态傅里叶变换红外光谱仪中,由于准直系统距离红外光源较近,光源的热辐射会导致其局部温度升高,从而引起材料折射率发生改变,使得由准直系统出射的光束存在一定的发散角,进而影响光谱仪系统复原光谱所能达到的分辨率水平。本文研究了光谱仪系统正常工作状态下准直系统各区域的温度分布情况,由此计算出了相应的离焦量。通过计算准直光束发散角在光程差采样区域内的分布,分析了由此引入的附加光程差对光谱复原的影响。通过计算光谱结构误差随准直系统温度的变化,得到了准直系统温度控制的合理范围。最后,对基于SiC光源的光谱仪进行了实验,结果显示制冷光源复原光谱的光谱结构误差与非制冷光源的光谱结构误差相比有明显改善。因此,降低光源温度对减小准直系统热光学效应的影响是非常有效的。本文的研究结果将为解决同类问题提供参考。
In a stepped-mirror-based static Fourier transform infrared spectrometer, the collimation lens is located adjacent to the light source and the thermal radiation would lead to the partial temperature increase, and the refractive index of the infrared material unavoidably changes. Then the light beam passing through the collimation lens will induce a divergence angle, directly affecting the resolution of the recovered spectrum. Meanwhile, the angular divergence results in a displacement in the interference signal, making increasing di?culties in the interferogram processing and the spectrum recovery. In this paper, the distribution of temperature in different areas of the collimation lens is studied under the working condition, and the defocusing value of the collimation lens is 0.153 mm that is caused by the refractive index gradient of the infrared material along with the temperature. In addition, the divergence angle induced by defocusing is calculated, its distribution being nonuniform but symmetrical within the sample area. Moreover, the divergence angle brings about additional optical path difference, its effect on the recovered spectrum is analyzed. Compared with the ideal recovered spectrum, much noise emerges and the peak value is reduced in the real recovered spectrum. The spectrum-construction error of the real recovered spectrum is 18.72%, indicating that the recovered spectrum is seriously distorted, and the resolution at the center wavelength in the ideal recovered spectrum and the real spectrum are 4.71 cm?1 and 5.57 cm?1, respectively. This indicates that the spectrum resolving power is weakened. Furthermore, the reasonable temperature range is obtained by analyzing the curve of spectrum-construction error versus temperature of the collimation lens. When a spectrum-construction error of less than 5% is demanded, the temperature difference between the front len of collimation and the ambient must be less than 8 ?C. Finally, experiments are performed using a SiC rod as the light source, and in