De-Embedding Technology for Active and Passive Device Testing Used in the Microwave and Terahertz Bands: A Review: An Overview of Different De-Embedding Techniques

With the recent rapid improvement of wireless communication rates [1], [2], the cutoff frequency of semiconductor devices has gradually increased [3], [4], and products with large signal bandwidth circuits have become a hot topic [5], [6], [7]. Port-to-port signal transmission and coupling mechanisms are also becoming more complex with the increases in frequency, presenting a challenge for high-frequency on-chip measurement [8], [9]. An accurate calibration must be applied to eliminate deviations between the probe terminal and the internal port of the device [10]. Unlike device under test (DUT), the reference plane for calibration of a vector network analyzer is from A to A’, as shown in Figure 1. This inevitably leads to an interconnection structure being utilized to address the mismatch between A to B and A’ to B’ [11], [12]. In response, the de-embedding technique has been developed to eliminate this extra step from the measurement data [13], [14].