Active Mie-like resonance for noninvasive glucose detection

An ideal toolset to diagnose diabetes consists of an in-hand portable device that can effectively detect glucose concentration in the blood in real time without having to perforate the skin. However, existing glucose monitoring systems are invasive, and those that are noninvasive have low accuracy. Here, we explore an alternative approach for noninvasive monitoring of the glucose concentration in blood. We excite Mie-like resonances within highly lossy media by introducing active loads, uniformly applied within the inner circumference of a dielectric ring structure. Our study opens a possible way of developing portable glucose monitoring systems that can detect the glucose concentration with a noninvasive approach and improved accuracy compared with existing techniques.

Figure 1

(a) A finite model of the finger, simplified to a concentric cylindrical structure composed of bone, capillary, and skin. The radius of each cylinder is defined as $r_s,r_g,r_b$ from the center and the finger height is $h_f$. (b) Illustration of the ring integrated with active loads. The upper and lower portions of the ring are grounded with $\mathrm{Cu}$ plates.

Figure 2

(a) Simulated model of Figs. 1 and 1 in comsol Multiphysics 6.0. The ring is worn on the finger. The transmit port (blue arc) is applied inside the ring (blue area). The copper plates are positioned to sandwich the ring to support the resonance. Here, area “a” is the bone, “b” is the capillary and “c” is the skin. The entire finger model is defined with length $h_f$ with two $h_{\text{pml}}$ layers; scattering boundary condition (BC) and PML condition are applied properly to model wave propagation. (b) (Upper panel) Relative permittivity of skin and bone. (Lower panel) Relative permittivity of glucose-water solutions from 0 to 4 wt %.

Figure 3

(a) Mie-like resonance with lossless finger wearing the ring without active loads. The azimuthal modes are excited with respect to the frequency axis. $E_z$ field plotted when the height of the ring $h_r$ is (i) 10 mm, (ii) 12 mm, (iii) 10 mm, (iv) 14 mm, calculated at the mid-height of the ring. (b) Mie-like resonance with material loss in the finger. Now, the ring is integrated with active loads around the inner circumference of the ring. The $E_z$ field is plotted when the height of the ring $h_r\mathrm{is}$ (i) 15 mm, (ii) 12 mm, (iii)10 mm.

Figure 4

(a) Radiated power $P_{\text{rad}}$ comparison between the case with $1/Y_{s,z}=-40\mathrm{\Omega }$ (solid line) and that with $1/Y_{s,z}=-38\mathrm{\Omega }$ (dashed line). The admittance value of active loads differs from the calculated results in (b). The height of the ring is h_r = 5, 10, 7, 15 mm. (b) Radiated power $P_{\text{rad}}$ with different active loads, with $1/Y_{s,z}$ from −38 to −42 Ω. Here, the height of the ring $h_r$ is 10 mm. The green line is the same as in (a) when $h_r$ is 10 mm. (c) Radiated power $P_{\text{rad}}$ when the thicknesses of capillary, bone, and skin layers are changed. (d) Magnitude of electric and magnetic fields at point d in (b). The field is calculated at the mid-height of the ring.

Figure 5

(a) Radiated power $P_{\text{rad}}$ around $\text{T}{\text{M}}_{11,{15}_{}\text{mm}}$ mode when $1/Y_{s,z}=-38\mathrm{\Omega }$. (b) Radiated power $P_{\text{rad}}$ around $\text{T}{\text{M}}_{21,15\text{mm}}$ mode when $1/Y_{s,z}=-38\mathrm{\Omega }$. (c) Comparison of frequency resolution of $\text{T}{\text{M}}_{11,{15}_{}\text{mm}}$ and $\text{T}{\text{M}}_{21,{15}_{}\text{mm}}$ with respect to glucose concentration. (d) Radiated power $P_{\text{rad}}$ around $\text{T}{\text{M}}_{21,{10}_{}\text{mm}}$ mode when $1/Y_{s,z}=-38\mathrm{\Omega }$. (e) Radiated power $P_{\text{rad}}$ around $\text{T}{\text{M}}_{31,{10}_{}\text{mm}}$ mode when $1/Y_{s,z}=-40\mathrm{\Omega }$. (f) Comparison of frequency resolution of $\text{T}{\text{M}}_{21,{10}_{}\text{mm}}$ and $\text{T}{\text{M}}_{31,{10}_{}\text{mm}}$ with respect to glucose concentration.