During the project, several device samples with sets of acoustic resonators were fabricated using maskless photolithography and magnetron sputtering to study the effect of liquid properties on the behavior of backward acoustic waves. The periods of interdigital transducers (IDTs) were selected based on theoretical analysis of dispersion relations for the phase velocity of backward waves, obtained using the transfer matrix method. This approach made it possible to cover the section of the dispersion curve corresponding to the backward wave, approach the point of zero group velocity, and compare with the behavior of a forward wave. The experimentally reconstructed dispersion relation of the backward wave was in full agreement with the calculated one. To optimize the device parameters, a qualitative assessment of IDT efficiency was performed for transducers with different spatial periods (0.9–2.7 mm). The depth of the dip in the frequency dependence of the S₁₁ parameter was used as the criterion; the most efficient conversion (dip depth of –12.27 dB) was achieved by the transducer with a period of 1.5 mm, near the zero group velocity point. Based on this transducer, a compact acoustoelectronic device was fabricated on a YX-LiNbO₃ plate measuring 13.5×15 mm, comprising a single IDT (period 1.5 mm, 6 electrode pairs, aperture 8.5 mm) and a measurement cell. The problem of spurious reflections from the plate edges was solved by combining a serrated edge profile with a multilayer damping coating having a gradient acoustic impedance, which substantially reduced the noise level.
An automated experimental setup was developed based on a vector network analyzer, a thermostable enclosure with temperature control, a 16-channel transducer switching unit, and a personal computer. A software package for recording, processing, and analyzing experimental data was developed to automate data acquisition. The effect of liquid conductivity in the range of 0.0034–8 S/m on the properties of backward waves was investigated. The behavior of the backward A₁ wave was evaluated by comparison with the forward SH₁ wave in the same plate. Analysis of the resonance peak behavior in the frequency dependence of the S₁₁ parameter showed that the backward wave is sensitive to conductivity across the entire studied range, whereas for the forward wave, conductivity-induced shifts in the resonance peak frequency are only detectable in the low-conductivity region of 0.0034–0.5 S/m. Analysis of the effect of liquid viscosity (0.9–934 cP) showed that for the backward A₁ wave in a YX-LiNbO₃ plate, contact with liquid leads to a sharp decrease in the resonance peak depth — by more than an order of magnitude compared to air — yet the wave remains detectable, confirming its suitability for sensing applications. The effect of viscosity on the A₁ wave characteristics is moderate due to the dominance of normal displacements, which limits sensitivity. Nevertheless, the backward wave exhibits a monotonic and reproducible response in the low-viscosity range (0.9–14 cP), making it applicable in multiparameter sensing systems in combination with other wave types. It is shown that sensors based on backward acoustic waves have broad prospects for measuring the electrical and mechanical properties of liquids.
To support this, the effect of liquids with different permittivities (ε) and acoustic impedances (Z) on the characteristics of the backward A₁ wave in a lithium niobate plate was studied. Experiments were carried out with five media (air, water, acetone, dichloroethane, gasoline) on a sample containing seven interdigital transducers. Additionally, dispersion curves were calculated using the transfer matrix method, and frequency dependences of the S₁₁ parameter were computed using the finite element method. It was established that the shift in resonance frequency is governed primarily by the permittivity of the liquid (the frequency decreases monotonically as ε increases), while changes in Q-factor and S₁₁ dip depth are associated mainly with the acoustic impedance Z. A dedicated FEM model with fixed mechanical properties of air confirmed that when ε is varied, the frequency shift is preserved while the amplitude change is nearly eliminated, demonstrating the feasibility of two-parameter measurement of medium properties. Comparison of phase velocities obtained by the transfer matrix method, FEM, and experiment showed discrepancies of less than 1%. It was found that a buffer layer with low acoustic impedance reduces radiation losses for the A₁ wave while maintaining high sensitivity to ε.
Based on the completed studies, a set of recommendations for the design of sensing devices based on backward waves was formulated. For temperature measurement, it is recommended to use the backward wave near the zero group velocity point, where temperature sensitivity is significantly higher than for forward waves. For liquid conductivity detection, the backward A₁ wave offers a wide measurement range; however, due to response ambiguity, the use of an auxiliary channel based on the forward SH₁ wave is advisable. For liquid identification, the backward wave enables separate measurement of permittivity (via frequency shift) and acoustic impedance (via Q-factor). To enhance sensitivity, a combination of a strong piezoelectric with an anisotropic dielectric is recommended, and to reduce losses, a low-impedance buffer layer is advised. For device miniaturization, combining a serrated edge profile with damping coatings along the plate edges is effective. For multiparameter analysis, multimode registration of the full acoustic wave spectrum is proposed. When fabricating acoustoelectronic devices based on backward waves in plates, strict control of plate thickness (tolerance ±2 µm) is essential due to the high sensitivity of backward waves to geometry.