Measuring ethanol concentrations in liquid environments is crucial across various applications, including fermenters, the food industry, pharmaceuticals, medicine, and healthcare[1-6]. Ethanol is widely used as a solvent, preservative, and antibacterial agent[7], and it plays a key role in paints and varnishes[8], and as a primary ingredient in alcoholic beverages and renewable biofuels[9, 10]. Additionally, ethanol has been employed in the development of tunable water-based microwave-absorbing metamaterials by altering the dielectric properties of water[11]. It also serves as a propagation medium in test objects used for calibrating measurement facilities in ultrasound imaging systems[12]. The extensive use of ethanol across different applications shows the critical importance of accurately measuring ethanol concentrations in various solutions.

Different studies have been conducted to determine the concentration of ethanol in the WE mixture, particularly within dielectric tubes[13]. Common microwave methods, such as the open-ended coaxial probe, transmission lines, and resonant-type sensors, require either submersion of the sensor or the placement of a small amount of liquid on the sensing element[14-18]. While effective, these methods necessitate cleaning between measurements, which can be particularly problematic when dealing with corrosive substances[13]. Newly developed sensors incorporating microfluidic channels have gained popularity due to their precision, but they are limited in their ability to handle inhomogeneous media and small interaction regions[13-15]. In particular, sensor-based methods can only measure the concentration at a specific point, meaning that when the spatial distribution of concentration needs to be measured, they must rely on time-consuming scanning techniques. Optical techniques, including sensors and spectroscopic methods, are widely employed for measuring the spatial distribution of concentration, but they are limited by their inability to effectively penetrate thick or opaque materials, restricting their application to transparent or semi-transparent samples[19-21].

Recently, a new optical method called Thermoelastic Optical Indicator Microscopy (TEOIM) has been proposed[22]. This method has demonstrated its effectiveness in detecting concentrations of NaCl and glucose in aqueous solutions, as well as in identifying water droplets and air bubbles within dielectric tubes[23, 24]. By taking advantage of the ability of microwaves to penetrate dielectric media, this method characterizes the state of fluids inside optically opaque tubes when exposed to incident microwaves and measures changes in microwave near-field intensity. TEOIM incorporates all the benefits of optical microscopy, including a wide field of view, high measurement throughput, high spatial resolution, and non-contact, non-invasive measurements[23-26] As a result, TEOIM can offer a non-invasive and non-contact measurement of ethanol concentrations in a water-ethanol (WE) mixture, effectively overcoming the limitations of traditional microwave and optical techniques. In particular, one of the key advantages of TEOIM is its ability to quickly measure the spatial distribution of sample concentrations without the need for scanning, making it a promising solution for overcoming the limitations of conventional sensor technologies[27].

In this study, we measured the concentration of ethanol in a WE mixture within a dielectric tube using the TEOIM method. The magnetic microwave near-field (H-MWNF) distributions of dielectric tubes containing WE mixtures with varying ethanol concentrations were measured, and the intensities were subsequently analyzed. From the measurement results, it was shown that the ethanol concentration can be measured based on changes in H-MWNF intensity, which are caused by the permittivity variation of water due to ethanol concentrations.

Figure 1(a) shows the experimental setup of TEOIM for ethanol concentration measurement in a WE mixture in a dielectric tube. The optical indicator (OI) is composed of a glass substrate coated by an indium tin oxide (ITO) thin film. The ITO thin film has high electrical conductivity (sheet resistivity < 10 ohm/sq), and under microwave irradiation the ITO layer heats up due to the absorption of the magnetic component of the microwave electromagnetic field. The OI was attached to the alumina plate facing the ITO thin film layer towards the faces of the alumina plate. The dielectric tube is attached to the side of the alumina plate opposite to the side where the OI is mounted, and WE mixtures were injected into the dielectric tube using a syringe. The prepared samples were placed in front of an open-ended rectangular waveguide (X-band, TE mode, dimensions: 22.86 mm × 10.16 mm, length: 40 mm) for microwave irradiation. The microwave signal was generated using a signal generator and amplified to +34 dBm with a power amplifier. This amplified signal was transmitted through the waveguide connected to the microwave generation system. The waveguide was positioned 5 cm behind the dielectric tube, irradiating microwave signals to the tube.

Figure 1. (Color online) (a) Illustration of the measurement setup. The red line indicates the propagation path of the probing light, which is emitted from the surface light source (wavelength: 530 nm). The light is polarized to left circular polarization and incident on the optical indicator. The reflected light from the optical indicator changes to elliptical polarization due to the photo-elastic effect on the glass substrate. This light then passes through the analyzer and is imaged by a CCD camera. The microwave is generated and amplified by a microwave signal generator and power amplifier. The amplified microwave is transferred to an open-ended waveguide through coaxial cables and radiates from the open part of the waveguide to a dielectric tube. (b) Photographs of the backside view of a sample section of the measurement setup. (c) Photographs of a side view of the measurement setup. (d) photographs of the WE mixture samples.

When microwaves are radiated to the OI, localized heat is generated in the ITO thin film through the Joule heating process caused by the H-MWNF of the dielectric tube. This heat diffuses into the glass substrate, raising its temperature and inducing thermal stress. The heat source distribution causing stress on the glass substrate of the optical indicator was measured using a polarized light microscope system with circularly polarized probing light. When circularly polarized light is incident on the stressed glass, its polarization state becomes elliptically polarized upon reflection from the stressed medium. The CCD camera captured the reflected light with the analyzer orientations set at 0° and 45° to detect linear birefringent (LB) distribution images. The heat source distribution causing the stress was then calculated using the following formula[22]

Here, q(x,y) represents the density of the heat source, x and y are the Cartesian coordinates of the pixel, C is a constant related to the wavelength of the light and physical properties of the OI, and β1 and β2 are images related to the normal and shear stress distributions of the OI, respectively.

Figure 2 shows the visualized H-MWNF distributions for the WE mixture at different ethanol concentrations (0%, 20%, 40%, 60%, 80%, and 100%), where the percentage values indicate the volume ratio of water to ethanol. In this measurement, the H-MWNF was visualized across a frequency range from 7 GHz to 15 GHz. Additionally, because the electric field strongly couples with the liquid in the tube when the microwave polarization direction is parallel to the length of the tube, we chose this orientation for the microwave polarization[24]. The H-MWNF distribution measurement results show that with pure water (0% ethanol), the H-MWNF distribution is weak when the microwave frequency is above 9 GHz. As the ethanol concentration increased to 20%, a strong localized H-MWNF distribution around the dielectric tube appeared at 8.5 GHz and 11 GHz, indicating strong coupling with the incident microwave due to the presence of ethanol. As the ethanol concentration exceeds 20%, the intensity of the H-MWNF distribution at 8.5 GHz decreases, suggesting a change in the coupling behavior between the incident microwave and the WE mixture in the tube. On the other hand, at a microwave frequency of 11 GHz, a strong H-MWNF was consistently observed across ethanol concentrations from 20% to 100%. Additionally, at a microwave frequency of 13.5 GHz, a strong H-MWNF distribution appeared at concentrations above 40%, and the strength of the H-MWNF distribution exhibited a continuously increasing pattern as the ethanol concentration increased. These results indicate that at lower frequencies (from 8.5 GHz to 11 GHz), the strength of the H-MWNF distribution varies in a complex manner with concentration, whereas at higher frequencies (13.5 GHz), the H-MWNF distribution strength shows a more predictable and consistent pattern, increasing with ethanol concentration.

Figure 2. (Color online) The H-MWNF distribution images of WE mixture at varying concentrations and frequencies.

Figure 3 shows the H-MWNF intensity variation of a WE mixture at different concentrations across the frequency range of 7GHz to 15GHz. The measurement results showed that the intensity of the H-MWNF varies across different frequencies from 7 GHz to 15 GHz for various concentrations of WE mixture. These results indicate that the interaction between the microwave and WE mixture is frequency-dependent, as the intensity fluctuates significantly with changes in frequency. In addition, there are distinct peaks at specific frequencies, such as 8.5 GHz, 11 GHz, and 13.5 GHz, which suggest strong coupling between the microwave and the WE mixture. The intensity peak at 8.5 GHz for the 20% ethanol concentration indicates a potential resonance effect, where the microwave efficiently couples with the mixture. In addition, the intensity peak at 11 GHz is higher for concentrations of 40%, 60%, and 100%, indicating that this frequency is particularly responsive to changes in ethanol concentration across a broader range. The most significant increase in intensity is observed at 13.5 GHz, especially at 100% ethanol, suggesting that at higher frequencies, the coupling strength with the microwave field increases with ethanol concentration. This pattern can be attributed to enhanced microwave absorption effects in pure ethanol.

Figure 3. Intensity variation of WE mixtures at different concentrations across microwave frequencies of 7 GHz to 15 GHz. (a) at 0% (b) 20% (c) 40% (d) 60% (e) 80% (f) 100%.

In addition, experimental results also showed that at lower ethanol concentrations from 0% to 20%, as shown in Fig. 3(b), the intensity decreases with increasing frequency up to about 10 GHz, indicating weaker interaction with the microwave at these lower concentrations. In contrast, as ethanol concentration increases from 40% to 100%, the intensity becomes more pronounced at higher frequencies, particularly at 11 GHz and 13.5 GHz, which suggests that the dielectric properties of the mixture change in a way that enhances absorption or coupling with the microwave field at these frequencies. The peaks observed at 8.5 GHz, 11 GHz, and 13.5 GHz likely correspond to specific resonant frequencies where the dielectric properties of the mixture align with the microwave, leading to stronger coupling. These results imply that different frequencies can be used to probe specific concentration ranges of the ethanol-water mixture, with 8.5 GHz being sensitive to lower concentrations and 13.5 GHz to higher concentrations. This frequency-dependent behavior could be valuable in designing sensors for detecting specific ethanol concentrations based on the microwave response at various frequencies.

Figure 4 shows the variation in H-MWNF intensity at three different frequencies 8.5 GHz (a), 11 GHz (b), and 13.5 GHz (c) as the concentration of WE mixture changes within a dielectric tube. In Fig. 4(a), at 8.5 GHz, the H-MWNF intensity peaks at 20% ethanol concentration, with diminishing interaction observed at higher ethanol concentrations. In contrast, at 11 GHz in Fig. 4(b), the H-MWNF intensity increases steadily up to around 60% concentration and then fluctuates slightly. This implies a strong and relatively stable response across a wide range of ethanol concentrations. At 13.5 GHz in Fig. 4(c), the H-MWNF intensity consistently increases with ethanol concentration. This result suggests that the interaction between the microwave and the mixture is frequency and concentration dependent. Lower frequencies like 8.5 GHz are more sensitive to lower ethanol concentrations, while higher frequencies like 13.5 GHz show stronger interactions at higher ethanol concentrations.

Figure 4. Variation of intensity with WE concentration at different frequencies at (a) 8.5 GHz, (b) 11 GHz, and (c) 13.5 GHz in a dielectric tube.

The relationship between H-MWNF intensity and WE concentration at 13.5GHz was further analyzed by fitting the intensity data to a linear equation as shown in Fig. 5(b). The fitted equation, I=7×10-9C+2×10-7 demonstrates a clear linear relationship between the intensity and concentration of the WE mixture, where C represents the concentration of ethanol in the mixture, and I represent the measured H-MWNF intensity. From the fitting equation, we obtained a coefficient of determination of 0.95, indicating that the equation well describes the changes in H-MWNF intensity according to changes in WE concentration. This suggests that the H-MWNF intensity increases proportionally with ethanol concentration, providing a predictable and stable response across the tested range.

Figure 5. (Color online) The H-MWNF intensity of microwave field at 13.5 GHz and intensity as a function of ethanol concentration in water-ethanol mixtures (a) and (b) respectively.

In this study, we measured the ethanol concentration in the WE mixture in the dielectric tube using TEOIM. We measured the H-MWNF distribution over a frequency range from 7 GHz to 15 GHz and analyzed the H-MWNF intensity. Our results showed that at 13.5 GHz, the H-MWNF intensity increased consistently and linearly with ethanol concentration, indicating that this frequency is particularly sensitive to changes in ethanol levels and is suitable for precise quantification. In contrast, frequencies of 8.5 GHz and 11 GHz displayed varying sensitivity, with 13.5 GHz proving to be the most stable and predictable across different ethanol concentrations.

This research was supported by the 2024 scientific promotion program funded by Jeju National University.