An evaluation method for quantifying the influence of lift-off variations on nano-scale metal film thickness measurement using the eddy current technique

The eddy current method can be applied to in situ nano-scale metal film thickness measurements with a high precision. Generally, the lift-off distance should be kept constant in the measurement process and its variation would significantly affect the accuracy of an eddy current sensor. However, it is quite difficult to control the lift-off distance precisely in practice. Therefore, how to evaluate the influence of micro-scale lift-off distance variation on the measurement accuracy quantitatively is a critical issue which should be addressed especially in the field of in situ nano-scale detection. In this study, a simulation model of copper film thickness measurement using the eddy current method was established, which couples an electromagnetic field and an LC circuit module, with a film thickness range of up to 1500 nm. The lift-off benchmarks were set at 1, 2 and 3 mm, and the maximum vibration amplitude was 150 μm. Based on the established model, the coil parameters under each lift-off benchmark were optimized first for a good sensitivity. Then, coil impedance and output voltage in each lift-off vibration state were calculated sequentially. According to the calculation results, the influence of lift-off benchmark and variation amplitude on the thickness measurement error, as well as the sensitivity and linearity of output voltage, were revealed respectively. And a quantitative relationship between the measurement error and lift-off distance variation was further proposed. Meanwhile, a series of experiments have been carried out to further verify the above theoretical analysis and a quantitative evaluation method was established by nonlinear surface fitting. Finally, it was demonstrated that the greater the measured copper film thickness, the greater the measurement error caused by lift-off variations. Specifically, a lift-off variation of 1 μm would result in a thickness error of approximately 6 nm under the experiment conditions.