A Miniature, Broadband Focused PVDF-TrFE PMUT with Interface ASIC for High-Resolution IVUS Imaging

Abstract

Intravascular ultrasound (IVUS) is widely used for high-resolution imaging of vascular walls and plaques. This work presents a focused 0.8-mm aperture IVUS piezoelectric micromachined ultrasonic transducers (PMUTs) based on a PVDF-TrFE piezoelectric copolymer, offering inherently broad bandwidth. A novel fabrication approach enabled spherical focusing on a freestanding PVDF-TrFE piezopolymer film which was fixed in shape using conductive and acoustically inert epoxy. The PMUT, integrated with a high-voltage-tolerant Analog Front End Application Specific Integrated Circuits (AFE-ASIC) on a 1.5 mm width tower-shaped PCB, achieved center frequencies of 40 MHz and greater with −6 dB bandwidth of up to 92%. Pulse-echo and beam scanning confirmed the achievement of axial resolution 20 μm and lateral resolution up to 75 μm at 2.4 mm focal depth. IVUS images acquired from stent phantoms and vascular tissue clearly resolved vessel layers and stent struts, demonstrating 5 mm penetration depth and the system’s suitability for miniaturized, high-resolution IVUS applications.

Graphical Abstract

I. INTRODUCTION

Cardiovascular disease (CVD), causing heart failure and stroke, is the leading cause of death globally [1]. Atherosclerosis, the primary contributor to CVD, involves the buildup of plaques within arterial walls, leading to vessel narrowing, reduced blood flow, and increased risk of stenosis and thrombosis. Plaque rupture can result in thrombus formation, which may block critical arteries or embolize downstream, causing stroke or sudden death. Early assessment of plaque characteristics, e.g. burden, fibrous cap thickness, and composition, is essential [2]. Intravascular imaging is a key tool enabling the evaluation of plaque morphology and supports procedure planning and stent placement. Optical coherence tomography (OCT), has excellent resolution to resolve stent struts and vulnerable plaques but has poor tissue penetration and requires flushing the blood from the coronary artery. IVUS offers superior penetration depth with less resolution than OCT, enabling visualization of deeper components like the necrotic core, lumen narrowing, wall thickening, atheroma burden, and to a lesser extent stent deployment [3]. An intravascular imaging method that has good penetration depth and high resolution can be achieved using a focused broadband ultrasonic transducer.

The evolution of IVUS technology has introduced micromachined transducer platforms, notably capacitive (CMUTs) and piezoelectric micromachined ultrasonic transducers [1]. CMUTs are well-suited for array integration and advanced imaging modes like volumetric scanning, but their operating frequencies are typically limited to <10 MHz, reducing resolution. Additionally, their low acoustic impedance leads to significant signal transmission loss and compromised imaging depth. Piezopolymer PMUTs offer better acoustic coupling and operate efficiently at IVUS-relevant frequencies, making them more suitable for high-resolution imaging.

Like all imaging techniques, IVUS has tradeoffs in imaging depth, resolution, and cost, among other factors. High-frequency IVUS (>40 MHz), has high axial and lateral resolution, but reduced penetration depth. Low-frequency IVUS trades off resolution for penetration depth [2]. Focused transducers, concentrate the ultrasound energy into a narrow beam, enhancing lateral resolution and providing higher signal-to-noise ratios (SNR). Some approaches use combination transducers operating at different frequencies to achieve both high resolution and penetration depth with a single IVUS catheter [1]. Newer techniques, such as coded excitation using wide-band CMUT arrays have also been shown to improve imaging SNR [4].

In our work, we adopted the focused poly [(vinylidenefluoride-co-trifluoroethylene] (PVDF-TrFE) PMUTs, which offer broad excitation bandwidths compared to single-crystal PMUTs[5]. The broad bandwidth of polymer PMUTs operating in thickness mode enables both high resolution and good penetration depth from a single transducer. However, piezopolymers have reduced sensitivity and high intrinsic impedance and cannot be directly coupled to the typical 50-Ω instrumentation used for IVUS. Also, due to the lower sensitivity, polymer PMUTs benefit from high-voltage pulsing (beyond 100 V_PP typically) to emit sound pressures suitable for imaging.

Here, we present a miniaturized 0.8-mm aperture focused PVDF-TrFE copolymer PMUT for IVUS fabricated directly on a low-cost PCB. The piezopolymer film was pressurized to form a spherical shape to create a focused transducer. Conductive epoxy was applied to maintain the spherical shape, provide backside contact, and acoustic backing. This focused design, along with the broad intrinsic bandwidth of PVDF-TrFE, enabled both high axial and lateral resolution from a single transducer. A custom analog front end (AFE) ASIC from our previous work [6], [7] was co-located on a 1.5 mm width PCB to address the low sensitivity and high electrical impedance of PVDF-TrFE. Experimental results demonstrated a 40 MHz center frequency, 6 dB fractional bandwidth of 92%, and 20 μm axial resolution and 75 μm lateral resolution at 2.4 mm focal depth. IVUS images acquired from stent phantoms and vascular tissue clearly resolved vessel layers and stent struts, validating both high-resolution and sufficient penetration depth.

II. FABRICATION AND EXPERIMENT

A. PVDF-TrFE PMUT Fabrication

The cross-sectional view of the focused PVDF-TrFE PMUTs is shown in Fig. 1(e). This membrane formed a spherical section, whose f-number can be determined by $f_{\text{\#}}=r/A$ , where, r is the radius curvature of the spherical section, and A is the diameter of the aperture. The focal point of the transducer is located at point O. The f-number can be governed by the central deflection of the clamped circular membrane under differential pressure, which can be quantitatively described using the nonlinear pressure-deflection model for circular membranes in [8].

Fig. 1.

Process flow used to fabricate the proposed focused PVDF-TrFE transducers: (a) A 200 nm-thick gold film and a 9 um-thick PVDF-TrFE copolymer film are applied on the PMUT Printed circuit board, which has two arrays of 0.8 mm apertures (plane view). Epoxy adhesive is used to seal the periphery of the films and PCB to prevent air. (b) then the PMUT board is clamped into a custom Jig (cross-section). (d) After tightening bolts, the silicone gasket pressurized the periphery, forming an air-tight seal, air pressure is applied to the face of the PVDF film to deflect it into the desired spherical shape. Conductive epoxy is then placed into the aperture via the back hole. The pressure is maintained until the epoxy cures. (c) and (e) presents the cross-sectional structure of a single transducer in the PMUT array before and after applying air pressure.

A freestanding 9 μm-thick PVDF-TrFE film, coated with 200 nm of gold on one side (PolyK Technologies, LLC, State College, PA), was used as the piezoelectric polymer material, according to the quarter wavelength theory[9] and previous work[5]. This film was securely mounted onto a custom-designed printed circuit board (PCB) that featured two rows of 0.8 mm diameter apertures and two 0.2 mm-diameter castellation vias [Fig. 1(a)]. To ensure a proper seal between the film and the PCB, epoxy adhesive was applied around the periphery [Fig. 1(a)].

The assembled PCB was then clamped into a custom jig [Fig. 1(b)]. The jig had a transparent acrylic for visual monitoring of PMUT deformation. A silicone gasket was used to create an airtight seal. Differential air pressure was then applied to spherically shape the film through the air hole to obtain a 2.4 mm radius of curvature to create an f₃ transducer [Fig. 3(d)]. To maintain the spherical shape, establish electrical contact on the backside and provide acoustic backing, electrically conductive epoxy was injected into the aperture through the back hole of the jig [Fig. 1(d)]. Once the backing epoxy cured at room temperature, excess film was trimmed away, and individual PMUTs were depaneled via laser cutting. The 0.8-mm PMUT was modeled as a 2.2 pF capacitor with 30 kΩ leakage resistance [6].

Fig. 3.

IVUS pulse-echo and imaging test set up.

B. Tower Board Setup

A tower-shaped PCB was developed to evaluate the imaging capabilities of PMUTs connected with the AFE- ASIC (Fig. 2). The ASIC had 13 dB gain, 6-100 MHz bandwidth, 33 dB signal to noise ratio, and could withstand 200 V_pp pulsing [7]. The tower board tapered to a 1.5 mm wide tip that housed the 0.8 mm PMUT, connected via 0.2 mm diameter wires threaded through the castellation vias [Fig. 2(a)]. The AFE ASIC was flip-chip bonded next to the PMUT to amplify the echo signal and buffer the transducer output to a 50 Ω system [Fig. 2(b)].

Fig. 2.

(a) A tower-shaped PCB was designed to house the flip-chip AFE ASIC, and a PMUT mounted at the narrow tip. (b) The 1.5mm-wide portion contained all imaging components to simulate catheter integration. (c) A stainless-steel hypo tube protected the PCB and PMUT while leaving an acoustic window. (d) The catheter was driven by an external rotational motor to generate a 360° IVUS image.

To protect the tower tip, a 1.5 mm inner diameter, grounded, stainless steel 13-gauge hypo tube (2.4 mm OD) was used. To provide a catheter like form-factor. This tube provided structural rigidity and shielded against some RF interference. Laser-cut windows enabled the PMUT to emit and receive ultrasound signals [Fig. 2(c)]. While the full tower-shaped PCB was used for testing and basic imaging, the tip can be precisely cut off for integration with an IVUS catheter using a micro coaxial cable [Fig. 2(c)(d)].

C. Characterization Setup

The complete experimental setup used to characterize the PMUT-AFE system is shown in Fig. 3. The setup comprised a 40 MHz, 100 V_pp monocycle pulser (Avtech AVB2-TB-C), a DC power supply (HP E3630A), oscilloscope (Agilent Infiniium), and a 12-bit ADC (Gage Applied Technologies) operating at a 1 GS/s sampling rate. The pulser signal was delivered to the PMUT and monitored through a 30 dB attenuator via the oscilloscope. Echo signals amplified by the AFE ASIC were collected by the ADC and oscilloscope for analysis.

The tower-mounted PMUT probe was attached to a 3-axis motion control system (Newport ESP300). The motion stage supported 2-axis scanning and 360° rotation of the PMUT face (Fig. 3). This mechanical flexibility enabled axial and lateral beam profile evaluation and rotational image acquisition.

III. RESULTS AND DISCUSSIONS

A. Pulse echo and 2D Radiational results

To evaluate the transducer, axial and 2D radiation tests were conducted by pulse echo in deionized (DI) water. For axial radition patterns [Fig. 4(e)], a flat polished stainless steel reflector was used as the target and the transducer was translated axially the pulse echo response was recorded as a function of round-trip time from the reflector. The maximum pulse echo occurred at 2.4 mm, corresponding to an f-number of 3, as expected [Fig. 4(a)]. The signal showed minimal ringing and high SNR of 42.6 dB. The power spectrum [Fig. 4(b)] confirmed a center frequency of 40 MHz with a 6 dB fractional bandwidth of 92%. This broad bandwidth is enabled by the PVDF-TrFE’s low acoustic impedance and high mechanical loss. The 20 μm axial resolution is determined by $R_{\mathit{axial}}=c/(2\cdot f_c\cdot FBW)$, where c is the speed of sound, f_c is the center frequency, FBW is the 6 dB fractional bandwidth.

Fig.4.

(a)The first reflection showed 42.6 dB SNR and (b) approximately 92% 6-dB fractional bandwidth around 40 MHz, which determined 20 μm axial resolution with a f-number of 3. (c)The 2D beam profile from a different sample is shown when imaging a 20 μm wire exhibiting a lateral resolution of 75 μm. Spatial resolution test setup(d) (e).

Another PMUT sample was mechanically moved in the azimuthal and axial directions to capture the pulse-echo signal reflected from a 20 μm diameter stainless steel wire. The wire was fixed on a stand orthogonal to the scanning plane [Fig 4(d)]. The captured result showed a 2 mm long focal region with −6 dB lateral beamwidth of 75 μm (Fig. 4(c)). This is smaller than the theoretical lateral resolution [1] of 110 μm for a continuous wave at the 49.4 MHz center frequency for a circular aperture with a rectangular apodization.

Broadband transducers have been shown to achieve higher resolution [9]. To demonstrate this, concave transducers with center frequencies from 20–60 MHz (1 MHz step) were simulated in Field II (Fig. 5), all co-located and excited with the same 40 MHz pulse. Although not representing the actual PMUT, this simplified model illustrates that a broadband PMUT provides improved lateral resolution and reduced sidelobes compared to the theoretical continuous-wave resolution at the central frequency. In this analysis, phase was ignored and each frequency-dependent beam profile was normalized to the same area.

Fig. 5.

A set of co-located, concave transducers with center frequencies of 20, 40, and 60 MHz were simulated using Field II assuming equal area under the amplitude profiles and ignoring relative phase. It demonstrated an arrangement where broadband signals led to improved resolution and sidelobe levels.

Table I demonstrated the proposed PMUT compared to recently published lead-based ceramic IVUS PMUTs. Despite operating at the same center frequency, this work demonstrated the benefit in wide bandwidth to achieve high lateral imaging resolution.

Table I.

Comparison of Proposed Work With Other IVUS PMUTs

Ref.	Piezoelectric Material	Aperture Size/mm	Center Frequency/MHz	−6dB Fractional Bandwidth	Axial Res./um	Lateral Res./um
This work	PVDF-TrFE	0.8	40	92%	20	75
[9]	PMN - PT	0.5	45	61%	41.6	214.7
[10]	PNN-PZT	0.33	40	79%	36	141
[11]	PIN-PMN-PT	0.5	40	86%	43	226
[12]	PZT-5H	0.5	50	56.9%	26.7	120

B. Stent and Tissue Imaging Results

A 6-mm stent was deployed inside a 6.7-mm silicone tube anchored in paraffin wax to form a stent phantom for imaging [Fig. 6(c)]. A coronary ostium tissue sample was similarly prepared and anchored for stability [Fig. 6(d)]. Rotational IVUS imaging was performed by rotating the tower-mounted transducer in 1° steps, acquiring 360 Alines with 5,120 points per line over a 7.9 mm radius, and using DI water as a medium. The stent images clearly resolved struts and tube edges [Fig. 6(a)], with minimal shadowing artifacts as compared to IVOCT. In tissue imaging, vessel layers including intima, media, and adventitia were distinguishable [Fig. 6(b)], with 50 dB dynamic range and 5 mm penetration depth.

Fig.6.

A rotational ultrasonic image of the stent phantom (a) and the tissue (b) obtained by recording A-lines while rotating the tower at one degree of rotation per capture.

IV. CONCLUSION

In this work, we demonstrated a miniaturized focused PVDF-TrFE IVUS PMUT integrated with an AFE ASIC, achieving high spatial resolution. The presented low-temperature, scalable fabrication method is compatible with catheter integration. Characterization results showed a 40 MHz center frequency, 92% −6 dB bandwidth, 20 μm axial and 75 μm lateral resolution. Rotational IVUS imaging resolved stent struts and vessel layers with penetration depth up to 5 mm. Future work will pursue sub-millimeter integration for smaller vessel IVUS catheters.