Infrared reflectance analysis is facilitated via the comparison of spectra recorded in situ to a databank of actual or synthetic infrared reflectance spectra. It has recently been shown that reference spectra corresponding to the many different morphological forms of the same chemical can be generated synthetically using the imaginary, k, and real, n, components of the complex refractive index, n ∼ = n + ik. One method to obtain the n and k vectors is infrared ellipsometry, which measures the changes in amplitude, tan Ψ, and phase, Δ, of polarized light reflected from the sample both as a function of wavenumber and angle of incidence. The method requires specularly reflected light, so best results are usually obtained with polished planar samples of large surface area. Due to the difficulties of obtaining such samples, however, we investigate the possibility of pressing powders of neat materials and obtaining the corresponding optical constants from the pellets. In this paper, variability in the sample pellet and preparation method is investigated, as is variability in the fitting procedure for the derived optical constants. The n/k vectors are derived from the measured ellipsometric parameters, tan ψ and Δ, as they are fit by an oscillator model which yield n( ν ∼ ) and k( ν ∼ ) vectors as a function of wavenumber, ν ∼ . Construction of the oscillator model is not automatic and depends on significant input from the analyst as well as the sample’s physical characteristics. For pellet pressing, the experimental variability was found to be minimized for size-selected powdered samples as gauged by the minimal variance in ψ and Δ for three different pellets; similarly, the analytical precision for multiple measurements of the same pellet was also quite good, suggesting that a pressed pellet is a viable sample preparation method. Experimental variabilities were comparatively small; the greatest variability came in the analytic fitting procedure with differences in the k-peak values up to 10% for only the sharpest bands arising from four different fits to the same data set. The final ellipsometric n/k data are compared to literature values obtained from crystalline ammonium sulfate ((NH4)2SO4) samples as well as single-angle reflectance measurements that also used pressed pellets. Comparison with the previous literature values shows generally good agreement, although larger k-values are observed for the independent sets of data derived from pressed pellets. These data are suggested as an improved set of optical constants for (NH4)2SO4.

中文翻译：

---

EXPRESS：来自压制颗粒的改进的红外光学常数：II。(NH4)2SO4 的椭圆光度 n 和 k 值与变异分析

通过将原位记录的光谱与实际或合成红外反射光谱的数据库进行比较，促进了红外反射分析。最近已经表明，可以使用复折射率 n ∼ = n + ik 的虚部 k 和实部 n 分量合成生成对应于同一化学物质的许多不同形态形式的参考光谱。获得 n 和 k 向量的一种方法是红外椭偏法，它测量从样品反射的偏振光的振幅、tan Ψ 和相位 Δ 的变化，这两者都是波数和入射角的函数。该方法需要镜面反射光，因此通常使用大表面积的抛光平面样品获得最佳结果。然而，由于获得此类样本的困难，我们研究了压制纯材料粉末并从颗粒中获得相应光学常数的可能性。在本文中，研究了样品颗粒和制备方法的可变性，以及衍生光学常数拟合程序的可变性。n/k 矢量来自测量的椭偏参数 tan ψ 和 Δ，因为它们通过振荡器模型拟合，该模型产生 n( ν ∼ ) 和 k( ν ∼ ) 矢量，作为波数 ν ∼ 的函数。振荡器模型的构建不是自动的，它取决于分析人员的重要输入以及样本的物理特性。对于压丸，通过三种不同颗粒的 ψ 和 Δ 的最小方差来衡量，发现选择尺寸的粉末样品的实验变异性最小；相似地，同一颗粒的多次测量的分析精度也非常好，表明压制颗粒是一种可行的样品制备方法。实验变量相对较小；最大的可变性出现在分析拟合程序中，对于同一数据集的四种不同拟合，k 峰值差异高达 10%。最终的椭偏 n/k 数据与从结晶硫酸铵 ((NH4)2SO4) 样品获得的文献值以及同样使用压制颗粒的单角反射率测量值进行比较。与以前的文献值的比较显示出总体上良好的一致性，尽管从压制颗粒中获得的独立数据集观察到了更大的 k 值。