An ultrasensitive and broadband transparent ultrasound transducer for ultrasound and photoacoustic imaging in-vivo

Ultrasound imaging (USI) and optical imaging (OI) are complementary, safe, and cost-effective modalities that benefit from sensor fusion. However, two main challenges hinder seamless integration: creating a single form factor that co-aligns ultrasound and optical paths, and maintaining state-of-the-art performance for both modalities. Transparent ultrasound transducers (TUTs) offer a route to seamless integration, but existing TUTs have inferior acoustic performance relative to conventional opaque ultrasound transducers (OUTs), largely due to acoustic impedance mismatches and inadequate backing layers that cause ringdowns. The authors identify three prerequisites for a practical, high-performance TUT comparable to OUTs: (1) a transparent front matching material with acoustic impedance of 7–9 MRayl to maximize transmission efficiency; (2) a transparent backing material with impedance >5 MRayl to balance acoustic and electrical Q factors and suppress ringdown; and (3) firm, gap-free bonding among all layers to preserve acoustic quality. The study aims to meet these criteria simultaneously by designing and fabricating transparent ceramic-epoxy composites that serve as both matching/backing layers and adhesives, optimizing acoustic impedance, optical transparency, and rheological properties suitable for direct curing on interfaces. The overall goal is to realize a TUT that provides OUT-like broadband sensitivity and enables high-performance dual-modal USI/PAI in vivo.

Prior work has demonstrated multi-modal fusion using transparent components, but with trade-offs: Park et al. achieved quadruple modality fusion with a TUT, and Chen et al. realized high-speed wide-field optical-resolution PAM with a TUT; however, their TUTs exhibited inferior acoustic performance compared with opaque transducers. Ma et al. used a transparent, focused optical detector for high-quality multiscale PAI, yet such optical sensors cannot generate ultrasound images and are thus unsuited for dual-modal PA/US imaging. Conventional OUTs leverage metal-epoxy composites for front matching and backing, meeting both acoustic impedance and processing viscosity requirements, but these materials are opaque. Previous TUTs often used epoxy backing with low acoustic impedance (~3 MRayl), providing insufficient damping and resulting in significant ringdowns. Collectively, the literature underscores the need for transparent materials that deliver OUT-like front load, proper backing impedance, and robust interlayer adhesion.

Design and simulation: The team used one-dimensional acoustic transmission line analysis and the KLM model (custom MATLAB implementation via ABCD matrices) to optimize layer impedances and predict pulse-echo responses and electrical input impedance. A double front matching structure was targeted: a 7.5 MRayl first quarter-wave layer and a ~2.36–2.4 MRayl second quarter-wave layer, along with a transparent backing architecture achieving an effective 4–6 MRayl back load via a 3.8 MRayl backside matching layer and low-loss clear backing. Simulations compared the proposed TUT with a conventional OUT and prior TUTs in terms of front-load magnitude/phase, transmission efficiency, ringdown, and bandwidth. Transparent composite development: Ceramic-epoxy 0-3 composites were engineered by jointly optimizing optical scattering and acoustic properties. Longitudinal velocity and acoustic impedance were modeled using effective medium theory (Devaney) with inputs of bulk/shear moduli and densities, and validated experimentally via travel-time measurements on 0.7–2 mm chips (n=12 per case). Densities were computed from precise chip dimensions (±1 µm) and weight. Rheology was characterized without hardener using a rheometer at 25 °C and fit to the Krieger–Dougherty model to estimate intrinsic viscosity and maximum packing fraction, guiding formulation to achieve adhesive-like viscosities (<100 McPs) for gap-free direct curing. Light transmission was simulated (Mie theory) and measured (UV-VIS), emphasizing high transparency at visible/NIR wavelengths for quarter-wave thicknesses. Transducer fabrication: The TUT stack comprised a square, ITO-sputtered lithium niobate (LNO) crystal (5.6 mm × 5.6 mm × 100 µm) sandwiched between quarter-wave matching layers. The first front matching layer was a 7.5 MRayl SiO2 micro-powder composite (31 µm); the second front layer was a clear urethane (2.4 MRayl, 18 µm). The backside matching layer was a 3.8 MRayl SiO2 composite (25 µm). ITO electrodes (~200 nm) were used front and back; connections were made via a brass housing ring (ground) and conductive epoxy ring (signal). The backing cavity was filled with clear urethane resin. Composite pastes were degassed (≤1 mTorr, 30 min), cured (48 h RT + 2 h at 45 °C), and lapped/polished to target thickness. Performance measurements: Electrical input impedance (20 Hz–60 MHz) and two-way pulse-echo responses were measured using standard instrumentation; X-cut quartz at 3 mm in water served as the reflector. Optical transmittance of fabricated layers and assembled TUTs was measured across visible–NIR. Imaging system and protocols: An integrated AR-PAM/US microscope used a 1064 nm nanosecond laser with fiber coupling and beam steering mirrors/prism to pass through the TUT. A parabolic mirror co-focused optical and acoustic beams. The imaging head operated in a water tank with plastic window and coupling gel. Raster scans used 50 µm steps at 1 kHz repetition (B-scan ≈1 Hz over 50 mm). Object-side pulse energy was ~220 µJ; measured focal diameter ~650 µm (66 mJ/cm², below ANSI limits). Data were acquired at 250 MS/s, high-pass filtered (≥1 MHz; 10 MHz for selected PA views), Hilbert demodulated, and visualized with log-scaled maps and 3D PHOVIS. Phantoms and in-vivo studies: Resolution targets were a 20 µm tungsten wire (USI) and 6 µm carbon fiber (PAI). SNR was assessed using 0.5 mm pencil leads embedded in chicken breast at 1 mm spacing (n=140 measurements). In-vivo animal imaging followed IACUC approval (BALB/c mice), with depilation, isoflurane anesthesia, temperature control, and post-procedure monitoring. Human palm imaging followed IRB approval and informed consent, with protective measures and positional stabilization. Dye perfusion study and ex-vivo validation: IR-1048 dye (100 µg/mL in PBS/DMSO) was intravenously injected (100 µL) in mice; PA imaging was performed up to 9 h post-injection. Liver and bladder signals were tracked over time; ex-vivo livers (30 min post-injection vs control) were imaged to validate in-vivo findings.

• Materials/Device: Developed transparent SiO2–epoxy matching/backing composites achieving target impedances (front: 7.5 MRayl; backside matching: 3.8 MRayl; effective backing 4–6 MRayl) with high optical transparency (>80%) and adhesive-like viscosity (<100 McPs). • Acoustic performance: Single-resonance, broadband TUT with 30 MHz center frequency and 63% −6 dB bandwidth. Pulse-echo sensitivity comparable to a conventional opaque ultrasound transducer (OUT). Compared with a conventional TUT, the proposed TUT showed ~3× higher peak-to-peak echo amplitude and ~3.9× shorter waveform length; simulations also predicted ~4.78× higher transmission efficiency at center frequency vs conventional TUT. • Spectral/impedance: Effective front-load magnitude/phase fell in the ideal zone similar to OUTs; backside effective impedance comparable to OUTs, minimizing ringdown relative to prior TUTs with ~3 MRayl backing. • Optical: Assembled TUT exhibited slightly lower visible transmittance than a conventional TUT but comparable NIR transmittance; device photo shows clear visualization through 100 µm gaps. • Imaging performance (microscopy system): Axial resolutions: 32.6 ± 1.9 µm (USI) and 40.4 ± 1.6 µm (PAI). Lateral resolutions: 110.4 ± 21.8 µm (USI) and 148.4 ± 13.5 µm (PAI). SNRs in chicken breast: 55 dB (USI) and 49 dB (PAI). Penetration depths: 16.6 mm (USI) and 13.6 mm (PAI) in chicken breast; ~15 mm PA depth in mouse in vivo; in human palm, 9.0 mm (USI) and 6.5 mm (PAI). • In-vivo imaging: Volumetric US/PA imaging in live mouse revealed thoracic and abdominal anatomy and vasculature with depth-to-resolution ratios (DRR) up to ~371 for PAI; human palm imaging distinguished epidermis, dermis, subcutaneous tissue, fat pad, muscles, and vasculature (subcutaneous vessels, veins, arteries) to 6.5 mm depth. • Dye perfusion kinetics: PA signal in mouse liver peaked at 30 min post-injection (5.58 ± 0.08 arb. units), remained elevated at 3 h (3.70 ± 0.18), and returned near baseline by 6 h (1.03 ± 0.19). Bladder signal peaked at 3 h (1.98 ± 0.10) and returned to baseline by 9 h (1.15 ± 0.13). Ex-vivo dyed livers showed ~3.65 ± 0.54× higher PA signal than controls (1.00 ± 0.30).

By engineering transparent ceramic-epoxy composites that simultaneously satisfy front matching, backing impedance, and adhesive bonding requirements, the authors achieved a TUT whose acoustic front-load and spectral phase closely match those of conventional opaque transducers. The improved backing (effective 4–6 MRayl) reduces ringdown compared with earlier TUTs that used low-impedance epoxy backing. Consequently, the proposed TUT delivers OUT-like pulse-echo sensitivity with broadband response (63% bandwidth at 30 MHz) and high optical transparency, enabling true dual-modal USI/PAI in a single, co-axial form factor. The system demonstrated high-definition, deep in-vivo imaging in mouse and human subjects, with best-in-class depth-to-resolution ratios for AR-PAM and comprehensive anatomical visualization beyond superficial vasculature. Compared to prior TUTs and optical detectors, the new TUT supports both generation and detection of ultrasound, facilitating simultaneous US/PA imaging without sacrificing performance. Comparative analyses indicate comparable or superior noise-equivalent pressure and efficiency to custom opaque devices, while avoiding the sidelobe and alignment drawbacks of ring-shaped transducers. These results validate the central hypothesis that properly designed transparent matching/backing layers can overcome the historical performance gap between TUTs and OUTs, unlocking practical, integrated US/PA systems for biomedical applications.

The study introduces a practical, ultrasensitive, and broadband transparent ultrasound transducer that integrates tailored transparent matching and backing layers, functioning also as adhesives for gap-free bonding. The device achieves OUT-like pulse-echo sensitivity with 63% bandwidth at 30 MHz and >80% optical transparency. When integrated into an AR-PAM/US system, it enables deep, high-resolution in-vivo dual-modal imaging in mice and humans, achieving leading depth-to-resolution ratios and robust visualization of anatomical structures and vasculature. The approach establishes a new design standard for TUTs and opens a pathway for seamless fusion of optical and acoustic sensing in medical research and industry. Future work may extend the technology to array formats, optimize acoustic NA and optics for further resolution/penetration gains, and translate to targeted clinical applications.

• The proposed TUT exhibits a small additional ringdown tail due to relatively lower effective back-load in peripheral spectral regions compared with an OUT. • The imaging system used a parabolic mirror, which limits the acoustic numerical aperture relative to a concave lens, potentially constraining lateral resolution. • Optical transmittance of the assembled TUT is slightly lower than a conventional TUT in the visible range (though comparable in NIR). • Penetration depth varies with sample attenuation; human palm imaging showed shallower depths due to stronger acoustic attenuation compared to mouse tissue. • Simulations for backside optimization focused on magnitude of effective impedance (phase not considered), and propagation losses were not included in pulse-echo simulations.