Cruz Juan Pablo Nicolás, Garzón Carlos Mario, Recco Abel André C
Departamento de Física, Universidad Nacional de Colombia, Av. Cra. 30 # 45-03, Ed. 404. Ciudad Universitaria, 111321 Bogotá D.C., Colombia.
Departamento de Física, Universidade do Estado de Santa Catarina, Rua Paulo Malschitzki, 200, Campus Universitário Prof. Avelino Marcante, 89219710 Joinville, SC, Brazil.
Microsc Microanal. 2023 Jun 9;29(3):938-952. doi: 10.1093/micmic/ozad051.
In this study, a methodology for assessing the thickness of titanium nitride (TiN) coatings by energy dispersive X-ray spectroscopy (EDS) in the scanning electron microscope is explored. A standardless method is applied, where the film thickness (th) is related to the microscope accelerating voltage (V0), the type of substrate and the ratio between the more intense peaks in the EDS spectrum, arising from both the substrate and the coating (afterwards called the I-ratio, IR). Three different substrates covered with TiN were studied, namely, silicon, glass, and stainless steel. Monte Carlo simulations enabled to state an analytic equation, which allows assessing the coating thickness as follows:th=thcr⋅exp[-βIR1/n]where IR = Iksubstrate/Ikcoating, thcr (critical thickness) is the largest coating thickness, which is assessable at a fixed V0, β is a multiplication factor, and n is an exponent, where thcr, β and n are assessable from V0 and substrate type. Interpolation via the equation presented, using reference thicknesses, allowed thickness predictions with around 80% of datapoints differing less than around 2% from the reference value. A procedure for detecting variations as low as 1.0% in coating thickness regarding the nominal thickness is presented.
在本研究中,探索了一种在扫描电子显微镜中通过能量色散X射线光谱(EDS)评估氮化钛(TiN)涂层厚度的方法。应用了一种无标样方法,其中膜厚(th)与显微镜加速电压(V0)、基底类型以及EDS光谱中由基底和涂层产生的更强峰之间的比率(此后称为I比率,IR)相关。研究了三种覆盖有TiN的不同基底,即硅、玻璃和不锈钢。蒙特卡罗模拟得出了一个解析方程,该方程可以如下评估涂层厚度:th = thcr⋅exp[-βIR1/n],其中IR = Iksubstrate/Ikcoating,thcr(临界厚度)是在固定V0下可评估的最大涂层厚度,β是一个乘法因子,n是一个指数,其中thcr、β和n可根据V0和基底类型进行评估。通过使用参考厚度,利用所提出的方程进行插值,可对厚度进行预测,约80%的数据点与参考值的差异小于约2%。本文还提出了一种检测涂层厚度相对于标称厚度低至1.0%变化的程序。