**4. Results**

The measurement setup is shown in Figure 9. It consists of a VNA (which is used for characterization of the TLs and generation of the input signal), proposed sensors with the power divider, the phase detector, and the MSP430FR6989 LaunchPad Development Kit [28] for measuring the output voltage and the syringe pump.

**Figure 9.** Photograph of the measurement setup.

Firstly, the designed sensor was characterized using a three-port measurement of the transmission signal. Figure 10a shows the transmission characteristics of the RH and LH sections in the frequency range of interest for air and water placed in the microfluidic reservoir and a comparison with the results of the simulation. Good agreemen<sup>t</sup> was obtained between the measured and simulated results, except for the small shift in the frequency. It should be noted that the measured amplitude of the signals was several dB larger than in the simulations. However, the maximum difference between the signal through the CLRH and RH sections was still within the 10 dB range. Increased losses arose as a result of imperfections in the fabrication process and were also related to the low conductivity of the

aluminum tape used for the conductive layers. Due to the small frequency shift, the phase difference was reduced for 9.5% and the maximal detected phase shift was 112.3 degrees.

**Figure 10.** Comparison between simulated and measured phase responses: (**a**) phase responses for air (<sup>ε</sup>*r* = 1) and water (<sup>ε</sup>*r* = 80.1) inside the microfluidic reservoir; (**b**) phase shift versus dielectric constant in the microfluidic reservoir for different fluids.

In the next step, the phase detector was used at the output to measure the phase difference for different fluids inside the microfluidic reservoir. Figure 11 shows the variation of the output voltage versus the dielectric constant of the fluid in the reservoir. The output voltage almost linearly depended on the dielectric constant of the fluid in the reservoir with a regression factor of R<sup>2</sup> = 0.94. The limit of detection of the output voltage of the proposed device was determined by the accuracy of the AD8302 integrated circuit and was equal to 10 mV. It can be mentioned that the imaginary part of the complex permittivity does not affect the phase shift, but can influence the amplitude of the signal and insertion losses [1]. The imaginary part of the complex permittivity cannot be directly calculated from the phase shift and therefore the measurement of the amplitudes of two signals should be taken into account. Since the used phase comparator allows a measurement of the amplitude difference between two signals, it can be used for estimating the imaginary part of the complex permittivity, or for mitigation of the comparator measurement error of the phase shift.

**Figure 11.** Output voltage as a function of the dielectric constant in the reservoir.

The proposed sensor was used to measure the cell concentration (i.e., biomass) inside the microfluidic bioreactor. MRC-5 human fibroblasts were grown in DMEM with 4.5% glucose supplemented with 10% fetal calf serum and antibiotic/antimycotic solution. The cell density (number of cells per unit volume) and the percentage of viable cells were determined before the measurement using the proposed sensor. The measurement by the sensor was performed immediately after seeding the cells into the reservoir while still free-floating in suspension. This is done for proof-of-concept purposes. Figure 12 shows the output voltage as a function of the number of cells in 1 mL of medium solution. The number of cells influences the effective permittivity of the fluid in the reservoir. Therefore, the effective permittivity of the substrate under the CLRH TL changes and consequently the output voltage changes. Although the change in output voltage is relatively small due to the high permittivity of the medium solution, the measurement response possesses better linearity in terms of the variation of the number of cells with a regression factor of R<sup>2</sup> = 0.98.

**Figure 12.** Output voltage as a function of the cell concentration in the microfluidic bioreactor.

## **5. Discussion**

The metamaterial CLRH transmission line approach was used to make a novel microfluidic sensor for the characterization of fluid flowing in the microfluidic reservoir. The proposed sensor comprises a power divider, microstrip lines, and a phase detector designed as a low-cost microfluidic platform using hybrid fabrication technology combining laser micromachining process, xurography, and lamination techniques. The CLRH TL fabricated above the channel is used to improve the sensitivity of the conventional TL in the phase comparator. The TL method was used, since it represents a fast and simple method for determination of the dielectric properties of the material, as well as allowing characterization at a single frequency with a high degree of integration with the sensor elements. The complete read-out detection circuit for determination of permittivity based on the phase shift is suitable for in-field measurement, and has been designed to operate at a frequency of 1.275 GHz.

The measurement results of the fabricated sensor confirm that a change in the permittivity of fluids in the micro reservoir from 1 to 80.1 (from air to water) results in a phase shift of almost 115 degrees. Using the phase advance phenomena of the CLRH line, the sensitivity of the sensor can be improved more than 10-fold compared with the conventional RH line with the same length. The phase difference for the same sensor realized with the RH TL instead of the CLRH one is only 12 degrees. In addition, the designed read-out detection circuit is relatively simple and allows measurement at a single frequency. The output voltage is approaching the linear function of the dielectric constant in the microfluidic reservoir with a regression factor of R<sup>2</sup> = 0.94.

Proof of concept for the potential application was demonstrated by implementing the proposed sensor for measurement of the cell concentration in a suspension-like cell culture in a microfluidic bioreactor. The experimental test confirmed that phase difference linearly changes with cell concentration. The signal was relatively low due to the very high dielectric constant of the cell suspension (i.e., the cells floating in the medium). However, the sensor was able to detect even such a low signal intensity.

The proposed device based on a metamaterials-based CLRH sensor presents a low-cost detection solution characterized by relatively high sensitivity and linearity, and therefore it can be used for monitoring small concentrations of specific fluids in di fferent mixtures. On the other hand, the proposed sensor is suitable for a number of biomedical applications utilizing suspension cell cultures, or for fluid characterization.

**Author Contributions:** V.R. proposed the idea, carried out the experiments and data analysis, and prepared the original drafts. G.K. contributed ideas, methodology for characterization and testing, and developed the phase detector. S.B. and I.P. performed simulations and manufactured the sensor, performed validation, as well as contributed to writing. M.D. and I.G.S. performed experimental verification with cell culture and microfluidic bioreactor and contributed to the manuscript review and editing.

**Funding:** This work was funded in the framework of project III66004, Development of new information and communication technologies, based on advanced mathematical methods, with applications in medicine, telecommunications, power systems, protection of national heritage and education, Ministry of Education, Science and Technological Development (R Serbia); and REALSENSE1: Monitoring of cell culture parameters using sensors for biomass and nutrients/metabolites in media: Lab-on-a-Chip (LOC) approach, Good Food Institute 2018 Competitive Grant Program.

**Conflicts of Interest:** The authors declare no conflict of interest.
