Real-Time Reactive Ion Etch Metrology Techniques to Enable In Situ Response Surface Process Characterization

Abstract

The industrial semiconductor community widely accepts that new process development is both costly and time consuming. In particular for plasma etch processing, multiple plasma input parameters such as gas flow, pressure, and radio frequency (rf) forward power directly affect desired wafer output parameters such as etch rate, uniformity, and material selectivity. Therefore, a high number of etch experiments must be performed to accurately quantify these input and output (I/O) relationships for each process. Moreover, in situ wafer monitoring in general is not currently widely pursued in the industry. 1 Therefore, the typical characterization procedure can only obtain data for a single given set point per process run, where a process set point refers to the nominal input parameters used for a particular plasma condition. Thus, money and time are spent etching only one wafer for each set point required to build the input/output response surface models for a typical design of experiments (DOE). With increasing process demands and the increasing cost of larger and larger individual wafers, these test wafer costs for model design continue to become more and more acute.This project leverages expertise from four primary research fields to achieve our results. Semiconductor process knowledge is needed to formulate the problem of new process characterization. Optical metrology background is used to obtain accurate material refractive index properties and to measure film thickness changes during processing. The thin-film properties help determine accurate film thickness and etch-rate information, which are used as inputs to the empirical process models. Nonlinear systems theory techniques are used to filter real-time optical data and estimate high speed thickness and etch-rate information. The high-speed filtering allows multiple set points to be explored per process run. Finally, response surface modeling is employed to generate empirical I/O predictions using plasma generation parameters as the model inputs, and film thickness, etch rate, and wafer uniformity as model outputs.The goal of this research is to maximize the number of set points one can observe during a single run of the etching process by monitoring output data in real-time. This decreases the number of runs required for a complete response surface mapping, effectively reducing the number of wafers and development time required to model a new process. We show accurate and instantaneous etch-rate information (on the order of tens to hundreds of milliseconds time scales) requiring less than 100 Å of etched material; 2 therefore, we obtain empirical I/O relationships at multiple set points in a short amount of time and on a single test wafer.In this paper, we demonstrate this in situ characterization mapping concept on two reactive ion etch (RIE) test-beds to show the versatility of the concept. One is an RIE chamber designed for flat panel display processing, where large area etch uniformity of an amorphous silicon (a-Si) film on metal and glass is the chosen m...

Stationary inverted Lyman population formed from incandescently heated hydrogen gas with certain catalysts

“…The population density of the n = 3 hydrogen excited state N 3 was determined from the absolute intensity of the Balmer α (6563 Å) line measured using the calibrated spectrometer. The spectrometer response was determined to be approximately flat in the 4000-7000 Å region by ion etching [17] and with a tungsten intensity calibrated lamp. The absolute intensities of n = 3-6 were determined from the absolute intensity of Balmer α (n = 3) and the relative intensity ratios.…”

Optical Diagnostics for Thin Film Processing

“…Real-time SE has also been used to control thickness and etch rate in a commercial ICP reactor using careful in situ design of experiments methods (ISDOE) and EKF filtering (243,244). Other groups used SE for real-time feedback control of thermal chlorine etching of GaAs and plasma etching of silicon nitride films (245), and control of wet etching of GaAs/AlGaAs (246).…”