Miniaturized CMOS Circuits and Measurement Techniques for Broadband Dielectric Spectroscopy

Abundant research reported in the literature has indicated that broadband dielectric spectroscopy (BDS), i.e., the measurement of material permittivity versus frequency, can serve a broad range of applications, including, but not limited to, biomedical, food, automotive, and agricultural industries. Adopting this technique in real-life application scenarios is directly dependent on the miniaturization of bulky measurement setups, currently in use for these (prototype) sensing systems. At the same time, a highly sensitive and precise permittivity readout is essential to distinguish between different materials or track variations in the material state composition. This work focuses on developing ultra-compact sensing elements, readout electronics, and measurement techniques to determine the localized complex permittivity with high accuracy, sensitivity, and spatial resolution at microwave operation frequencies.

Firstly, various sensing elements and high-resolution measurement setups are discussed for their compatibility with CMOS integration. Application scenarios are directed towards the characterization of low-loss materials, which often present much higher impedance than the currently 50-Ω oriented measurement setups. An I/Q-mixer-based interferometric technique is introduced to re-normalize the readout system reference impedance and improve the measurement sensitivity at high-impedance loads. Experimental results underline the potential of this technique. However, its compatibility with CMOS technology to enable small-factor systems is challenging at the intended frequencies of operation. Therefore, a double-balanced, RF-driven Wheatstone bridge with programmable branch impedance implemented in CMOS technology is proposed and analyzed for the high-resolution measurement of high-impedance loads (chapter 2).

Next, a high-sensitivity, ultra-compact BDS sensor system is introduced for localized permittivity sensing. As a sensing element, it utilizes a metal patch that performs the actual sensing by presenting permittivity-dependent admittance. This patch is best implemented on the top metallization layer of a CMOS technology such that it can directly interface with the material-under-test (MUT). High measurement sensitivity is achieved by embedding the patch in a double-balanced, RF-driven Wheatstone bridge followed by a frequency down-converting mixer. By driving the bridge with a square wave, permittivity information can be acquired at the fundamental and subsequent harmonics. This concept allows increasing the measurement speed and, at the same time, provides an extended measurement frequency range (chapter 3).

The measurement of the complex permittivity of materials is enabled by developing a dedicated calibration procedure for the patch-based BDS sensor. Measurement results of known liquids show good agreement with theoretical values in the literature, and the relative permittivity resolution in these measurements is better than 0.3 over a 0.1–10 GHz range. The proposed sensor implementation features a measurement speed of 1 ms and occupies an active area of only 0.15×0.3 mm^2, enabling the realization of very compact sensor arrays that can facilitate (real-time) 2-D dielectric imaging of permittivity contrast (chapter 4).

Such a real-time BDS sensor array has been implemented as a 5x5 array, illustrating the scalability of the proposed patch-based BDS concept. This matrix has been demonstrated for its functionality by resolving spatial permittivity variations in the sub-mm range (chapter 5).

Last, the findings and conclusions of this dissertation, and recommendations for future work, are discussed (chapter 6).