Background and Motivation
High Intensity Focused Ultrasound (HIFU) has become popular in recent years for therapeutic treatment of medical conditions such as arrhythmias and cancers. Like radio-frequency (RF), microwave, and laser therapeutic treatments, HIFU causes local heating to necrose pathological tissues. HIFU has distinct advantages over these other therapies because ultrasound can be focused at distance onto a region of interest without harming intervening tissues. This allows ultrasound to be used more effectively as a non-invasive or minimally invasive therapeutic tool; reduction of incisions allows faster patient recovery and reduces the chances of infection, mortality, and morbidity.
While piezoelectrics have been traditionally used for HIFU ablation, fabrication flexibility and improved perfomance in CMUT technology have made them highly competitive for HIFU applications. Using CMUT technology, we develop a non-invasive, MR-guided [1, 2] HIFU probe for treatment of lower abdominal cancers, such as metastatic colorectal cancer. Resection of these cancers increases 5-year survival from 8% to 30-35%, but only 20% of patients are suitable for resection. For 80% of patients, MR-guided noninvasive treatment is needed and provided perfectly with HIFU.
Feasibility of CMUT for HIFU
Typical imaging CMUTs have demonstrated potential for HIFU applications; they have shown output surface pressures as high as 1 MPa peak-to-peak and survived CW operation for several hours.
A typical 1.8 by 0.66 mm imaging CMUT was excited by 2MHz CW input when immersed in an oil tank. The applied DC voltages were swept from 100-180 V, and AC voltages were selected so the overall voltage did not exceed 250 V to prevent dielectric breakdown of the oxide. The pressure measured by a Z44_0400 hydrophone (Onda Corporation, Sunnyvale, CA) 1 cm from the transducer’s surface was corrected for attenuation and diffraction, and plotted against applied voltage (Fig. 1).
Though these CMUTs were designed for imaging, they could output pressures as high as 1 MPa peak to peak. When the CMUT is biased near the collapse voltage and excited with a large enough AC voltage, the membrane can displace nearly the entire extent of the gap, which maximizes the output pressure.
The same device was operated in CW operation for 25 min (Fig. 1) with a continuous output pressure of 0.77 MPa peak to peak. The major limitations of high power CW operation are the breakdown of the insulating layers from the high voltages and the failure of metal traces from high current densities. Increasing the thickness of the insulation layers and interconnects can greatly improve performance in the next-generation CMUT membranes .
FIGURE 1. A wafer-bonded device designed for imaging applications at 7 MHz has been demonstrated to produce output pressures as high as 1 MPa (left). The device also achieved 0.7 MPa peak-to-peak consistently in continuous wave mode with 80 V bias voltage and 130 Vpp AC voltage (right). The slight decrease in output pressure with time is probably caused by oxide charging.
Next-Generation HIFU/Imaging Transducers
New membrane designs have been developed and designed. In order to achieve the increased pressures needed for HIFU operation, these transducers have larger gap size, 0.2 - 0.3 mm, to generate larger displacements and velocities. In addition, a heavy mass, formed by either a double wafer bonding process or electroplating of a thick gold electrode serves to increase the output pressure. The mass allows a more piston-like transducer behavior, which increases the average displacement of the membrane. It also allows separate control of the mass and spring constant of the membrane. We used an axisymmetric model in ANSYS to simulate a CMUT designs shown in Fig. 5. The CMUT membrane and electrode are modeled by PLANE82 elements; TRANS126 elements transfer electrical to mechanical energy. A fluid column of FLUID29 elements was constructed, and the average pressure was calculated to be at least 2 MPa peak-to-peak. The simulated frequency response of this transducer exhibited good imaging qualities with an output of 7.5 kPa/V and 90% fractional bandwidth at 80% of the collapse voltage.
FIGURE 2. Picture of the two ANSYS models used to simulate the two different HIFU membrane designs. The first design (left) uses a large gold electrode formed by electroplating to serve as the piston-like mass. The second design (right) uses a double wafer bonding process to forma thick silicon mass inside the membrane.
FIGURE 3. Sample output pressure transient measurement and harmonic response of a simulated piston design.
The next-generation CMUT membranes are currently being fabricated and measured. After comparison with the model and optimization of the processing and design, a final HIFU design can be reached.
 Chen L, Bouley D, Yuh E, d'Arceuil H, Butts K, Study of the behavior of focused ultrasound lesions with MRI and histology, JMRI 1999;10:146-153.
 Duerk J, Butts K, Hwang K, Lewin J, Pulse Sequences for Interventional Magnetic Resonance Imaging, Topics in Magnetic Resonance Imaging 2000; 11(3): 147-62.
 S. H. Wong, I. O. Wygant, D. T. Yeh, X. Zhuang, B. Bayram, M. Kupnik, Ö. Oralkan, A. S. Ergun, G. G. Yaralioglu, and B. T. Khuri-Yakub, “Capacitive micromachined ultrasonic transducer arrays for integrated diagnostic/therapeutic catheters,” AIP Conference Proceedings, Therapeutic Ultrasound: 5th International Symposium on Therapeutic Ultrasound, vol. 829, no. 1, pp. 395-399, May 8, 2006.
This project is funded by National Institutes of Health under grant 5R01CA077677-08 (Project title: iMRI Methods for Cancer Diagnosis and Treatment, Principal Investigator: Prof. Kim R. Butts)