Temperature-Corrected Fluidic Glucose Sensor Based on Microwave Resonator

Abstract

In this paper, a fluidic glucose sensor that is based on a complementary split-ring resonator (CSRR) is proposed for the microwave frequency region. The detection of glucose with different concentrations from 0 mg/dL to 400 mg/dL in a non-invasive manner is possible by introducing a fluidic system. The glucose concentration can be continuously monitored by tracking the transmission coefficient S 21 as a sensing parameter. The variation tendency in S 21 by the glucose concentration is analyzed with equivalent circuit model. In addition, to eradicate the systematic error due to temperature variation, the sensor is tested in two temperature conditions: the constant temperature condition and the time-dependent varying temperature condition. For the varying temperature condition, the temperature correction function was derived between the temperature and the variation in S 21 for DI water. By applying the fitting function to glucose solution, the subsidiary results due to temperature can be completely eliminated. As a result, the S 21 varies by 0.03 dB as the glucose concentration increases from 0 mg/dL to 400 mg/dL.

Keywords: complementary split-ring resonator; electromagnetic biosensor; fluidic glucose sensor; microwave; non-invasive detection; temperature correction.

Figure 1

Electrical properties of glucose solution. (a) dielectric constant for various concentration at $25.5{}^{\circ }\mathrm{C}$; (b) loss tangent for various concentration at $25.5{}^{\circ }\mathrm{C}$; (c) dielectric constant of deionized water for various temperature; (d) loss tangent of DI water for various temperature; (e) change in dielectric constant at 2.9 GHz; (f) change in loss tangent at 2.9 GHz.

Figure 2

Simulation results of the proposed sensor. (a) electric field distribution at the resonant frequency ($\lambda$/15 above the substrate); (b) modelling of proposed sensor and fluidic channel as experimental condition; (c) transmission characteristic with and without the fluidic channel.

Figure 3

The fabricated sensor and its resonant characteristics. (a) signal line of the complementary split-ring resonator (CSRR); (b) ground plane of the CSRR and zoom-in of the resonant part; (c) measurement result of the transmission characteristic.

Figure 4

The equivalent circuit model of the proposed sensor. (a) complementary split-ring resonator; (b) simplified circuit model with port termination for transmission coefficient.

Figure 5

Measurement setup for temperature control and testing the performance of the proposed sensor.

Figure 6

Measured result of the transmission coefficient variation under stable temperature condition. (a) $20{}^{\circ }\mathrm{C}$; (b) $30{}^{\circ }\mathrm{C}$; (c) $40{}^{\circ }\mathrm{C}$.

Figure 7

Relationship between the temperature and the transmission coefficient of the proposed sensor. (a) variation in temperature increases by $5{}^{\circ }\mathrm{C}$ from $20{}^{\circ }\mathrm{C}$ to $40{}^{\circ }\mathrm{C}$ every 2 min; (b) variation in transmission coefficient under varying temperature condition when the DI water flows through the fluidic channel; and (c) variation in transmission coefficient for the other concentrations of the glucose solution under varying temperature condition.

Figure 8

Measured results of variation in the transmission coefficient for varying temperature condition. (a) temperature set arbitrarily to address both room temperature ($20{}^{\circ }\mathrm{C}$) and body temperature ($36{}^{\circ }\mathrm{C}$); (b) variation in transmission coefficient before the temperature correction; (c) variation in transmission coefficient after the temperature correction; (d) statistical result for variation in the transmission coefficient after temperature correction obtained from repeated experiment.