Permittivity sensing has wide applications in a variety of industries such as petroleum, agriculture,food production, etc.. The complex permittivity serves as an effective indicator of the
valuable information of the sample material such as their constitute, moisture or quality, which
are critical in the process of product development, production, and treatment. The nondestructive
nature of the permittivity sensing makes it highly cost-effective and enables it for real time material
characterization. This calls for the adoption of CMOS technologies to achieve a high level
of integration of signal processing capabilities for cost-effective and ubiquitous applications without
sacrificing precision and accuracy. Therefore, numerous works featuring CMOS permittivity
sensor have been developed. Among all aspects of CMOS permittivity sensing systems, high resolution
sensing is of particular importance, which enables CMOS sensing in applications such as
high precision biosensing, precision medicine, etc. High resolution sensing enables high throughput
measurement, capturing fast process in real time, etc. To achieve high resolution, the prior art
focuses on two key aspects: permittivity sensor’s sensitivity boost and system noise reduction.
In the optical community, ultra-high sensitivity is achieved by utilizing resonator based sensors,
whose resonant frequency is shifted by the permittivity of different MUT’s. Sharp curvature and
steep slope of the transmission coefficient of the high Q resonator sensors result in high sensitivity.
Whispering gallery mode (WGM) resonators show particularly high sensitivity due to its high Q
and the strong field interaction with the MUT’s. However, the expensive and bulky tunable optical
sources and the limited signal processing and readout capabilities severely constrain the applications
of the high sensitivity optical sensors.
To advance the state of the arts of the high resolution permittivity sensors, this dissertation dedicates
to the theory, design and implementation of sub-THz WGM resonator based CMOS permittivity
sensor. Thanks to its short wavelength, the WGM resonator sensor at sub-THz possesses
a compact form-factor, achieving a high level of integration with the CMOS transceivers (TRX).
Like its optical counterparts, the compactness of the sub-THz sensor also enables high sensitivity
detection while requiring a smaller amount of MUT samples than sensors working at GHz frequency
range. The CMOS TRX generates the sub-THz signal to excite the WGM sensor and readout the
signal to obtain the permittivity information.
The dissertation analyzes the EM mechanisms of the WGM resonator sensor, including the loss
analysis and coupling condition analysis, lays theoretical foundations for the sensitivity optimization
for high resolution sensing. A novel mechanism to detect complex permittivity using the
WGM resonator sensor is proposed for the first time in the permittivity sensing and resonator
sensor community. A low power permittivity sensing system at 160 GHz is prototyped to verify
the permittivity sensing capability of the WGM resonator sensor integrated with the CMOS TRX
IC, which demonstrates a permittivity sensing resolution of 0.098 for a integration time of 10 us
with a power consumption of only 9 mW. A high resolution complex permittivity sensing system
at 160 GHz is implemented that adopts a band-stop WGM disk resonator based sensor structure
and multi-fold noise suppression techniques, which demonstrates a record of 0.05% complex sensing
resolution within 14-us integration time and consumes 54 mW of DC power.
In the process of designing the high resolution permittivity sensing system, mathematical derivations
are developed to predict the noise suppression effect of the three noise reduction techniques:
the phase noise suppression by the coherent phase noise cancellation, the flicker noise suppression
by the chopping scheme and the thermal noise suppression by the integrator. Since these three noise
reduction techniques are commonly used in sensing / imaging systems, the developed derivation in
this dissertation can be readily applied to predict the noise suppression effect numerically and to
guide the design of the noise reduction circuitry.
