An optimized ultrasound detector for photoacoustic breast tomography

Photoacoustic imaging has proven to be able to detect vascularization-driven optical absorption contrast associated with tumors. In order to detect breast tumors located a few centimeter deep in tissue, a sensitive ultrasound detector is of crucial importance for photoacoustic mammography. Further, because the expected photoacoustic frequency bandwidth (a few MHz to tens of kHz) is inversely proportional to the dimensions of light absorbing structures (0.5 to 10+ mm), proper choices of materials and their geometries, and proper considerations in design have to be made for optimal photoacoustic detectors. In this study, we design and evaluate a specialized ultrasound detector for photoacoustic mammography. Based on the required detector sensitivity and its frequency response, a selection of active material and matching layers and their geometries is made leading to a functional detector models. By iteration between simulation of detector performances, fabrication and experimental characterization of functional models an optimized implementation is made and evaluated. The experimental results of the designed first and second functional detectors matched with the simulations. In subsequent bare piezoelectric samples the effect of lateral resonances was addressed and their influence minimized by sub-dicing the samples. Consequently, using simulations, the final optimized detector could be designed, with a center frequency of 1 MHz and a -6 dB bandwidth of ~80%. The minimum detectable pressure was measured to be 0.5 Pa, which will facilitate deeper imaging compared to the currrent systems. The detector should be capable of detecting vascularized tumors with resolution of 1-2 mm. Further improvements by proper electrical grounding and shielding and implementation of this design into an arrayed detector will pave the way for clinical applications of photoacoustic mammography.

💡 Research Summary

This paper presents the design, simulation, fabrication, and experimental validation of a highly sensitive, broadband ultrasound detector tailored for photoacoustic (PA) breast tomography. The authors begin by outlining the specific requirements of PA breast imaging: the need to detect weak acoustic signals generated by vascularized tumors located several centimeters beneath the skin, and the necessity for a detector that simultaneously offers high sensitivity, a wide acceptance angle, and a frequency response that matches the broad spectrum of PA signals (tens of kHz to a few MHz) arising from absorbing structures ranging from 0.5 mm to over 10 mm in size.

To meet these demands, the detector is designed with a square aperture of 5 mm × 5 mm, providing a large active area that reduces thermal noise and thus lowers the minimum detectable pressure (MDP). However, a large aperture typically introduces undesirable lateral resonances that distort the frequency response. The authors address this by sub‑dicing the piezoelectric element into an array of 0.9 mm × 0.9 mm sub‑elements, acoustically isolated by air kerfs (≈100 µm) but electrically connected in parallel, thereby suppressing lateral modes while preserving the overall capacitance advantage.

The active material selected is CTS 3203HD, a high‑coupling PZT ceramic offering superior dielectric constant and low loss compared to conventional PZT formulations. Front and back acoustic matching layers are introduced to bridge the impedance mismatch between the piezoelectric ceramic (≈30 MRayl) and water/tissue (≈1.5 MRayl). Their acoustic impedances are chosen according to Zm = √(Zw·Zp), and their thicknesses are optimized through iterative simulation.

Two complementary simulation tools are employed. A one‑dimensional KLM model provides rapid estimates of center frequency and bandwidth, guiding the initial choice of layer thicknesses. Detailed three‑dimensional finite‑element models (implemented in PZFlex) then capture the full electromechanical behavior, including the effects of sub‑dicing, matching layer geometry, and backing material. Mesh resolution is set to 1/30 of the wavelength at 1 MHz to ensure accurate representation of wave propagation.

Based on these simulations, three functional prototypes are fabricated. The first two prototypes validate the basic KLM‑predicted performance, showing a center frequency near 1 MHz and a fractional bandwidth of roughly 80 % (−6 dB from 0.4 MHz to 1.25 MHz). Electrical impedance measurements confirm the expected resonance behavior. The third, final prototype incorporates the optimized sub‑dicing pattern and refined matching layers, resulting in a markedly flatter frequency response and the suppression of lateral resonance peaks.

Experimental characterization employs two methods: (1) a transmission setup where the detector is driven by a broadband pulser and the emitted pressure field is recorded with a calibrated needle hydrophone, allowing extraction of the detector’s transfer function; and (2) a pulse‑echo configuration using a stainless‑steel reflector to assess the time‑domain response. Both methods corroborate the simulated bandwidth and confirm a minimum detectable pressure of 0.5 Pa—an improvement of two to three orders of magnitude over previously reported PVDF arrays (MDP ≈ 80 Pa). The −6 dB acceptance angle exceeds 30°, ensuring adequate coverage for a 360° rotational scan geometry.

In conclusion, the study demonstrates a systematic, iterative approach to designing a PA breast imaging detector that balances large aperture (for low noise) with broadband response (for resolution of structures from 0.5 mm to >10 mm). The key innovations—sub‑dicing to eliminate lateral resonances and precise matching‑layer optimization—enable a detector with a 1 MHz center frequency, ~80 % fractional bandwidth, and 0.5 Pa MDP. The authors suggest that further enhancements, such as improved electrical grounding, shielding, and integration of an acoustic lens, will facilitate the transition to multi‑element arrays suitable for clinical photoacoustic mammography.