Noninvasive Glucose Sensing in Aqueous Solutions Using an Active Split-Ring Resonator

• Published in: IEEE sensors journal 2021-09, Vol.21 (17), p.18742-18755
• Main Authors: Abdolrazzaghi, Mohammad, Katchinskiy, Nir, Elezzabi, Abdulhakem Y., Light, Peter E., Daneshmand, Mojgan
• Format: Article
• Language: English
• Subjects: active microwave sensor; Aqueous solutions; Biomedical materials; Biosensing; Biosensors; Blood; Blood plasma; Clarke analysis; Dynamic range; Error detection; Frequency measurement; Glucose; Glucose monitoring; Microwave sensors; Monitoring; Q factors; Resonators; Selectivity; Sensors; Spectroscopy; split ring resonator
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2021.3090050

It is shown here that microwave sensors can be used to monitor glucose in serum concentration with minimum detectable as well as resolution of 1 mMol \cdot \text{L}^{{-1}} ( \approx 18 mg \,\cdot dL −1 ). The ultrasensitive detection technique relies on a split ring resonator, operating at the frequency of 1.156 GHz, as the core of the sensor where its loss is compensated to enhance the quality factor from \sim 190 (passive mode) to \sim 3850 (active mode) to enable high resolution (modified frequency detection error from ±12 kHz down to ±2.5 kHz) frequency-shift sensing. Initially, glucose concentrations of 100-1000 mMol \,\cdot \text{L}^{{-1}} (1800-18000 mg \,\cdot dL −1 ) in water were detected within 250 kHz of dynamic range (between two spectrum ends). Selectivity of the sensor to glucose is verified with respect to common interstitial fluid ingredients with biological levels. Finally, to enhance the resolution of the proposed sensor, its loss-compensation is further improved leading to increased accuracy of measuring glucose samples in a 0.9 % NaCl solution containing 10 % horse serum that closely resembles blood plasma and interstitial fluid. This allows exploration of lower concentrations in the physiological range 1-30 mMol \,\cdot \text{L}^{{-1}} (18-540 mg \,\cdot dL −1 ) with improved frequency detection error down to ±0.75 kHz for two cases of with/without serum solutions with dynamic range of 30 kHz/38 kHz. The highly accurate glucose monitoring technique could be utilized for developing noninvasive glucose sensors for biomedical applications in real-time glucose monitoring.