Nonplanar High-Frequency Wideband Ultrasonic Transducer Arrays

描述(由申请人提供)：几十年来，与高频(&gt；20 MHz)换能器阵列相关的技术困难阻碍了高分辨率成像系统的广泛采用。最近，用于小动物成像的高频阵列和工作在20-70 MHz范围内的支持系统已经被商业化开发。然而，该系统极其昂贵的成本(50万美元)和相关的传感器探头(约200K美元)仍然是这一临床前工具的广泛使用以及基于阵列的高分辨率成像扩展到眼科、皮肤科和心血管医学临床应用的巨大障碍。发展高频阵列的主要障碍与用于压电换能器的传统制造技术有关：1)为了高频工作，压电晶体的厚度需要达到几十微米。使用传统的技术，如研磨和抛光，很难将晶体稀释到所需的厚度。2) 阵列的间距应该在波长的数量级上(对于30 MHz，~50m)。当使用标准的10到15�m刀片切割晶体时，有效区域的很大一部分被浪费在元素分离上。因此，应该采用等离子体刻蚀或激光加工等方法来切割晶体。3)根据缝隙的填充方式，可以观察到元件之间的交叉耦合增加。4)较小的元件尺寸导致较高的阻抗，并使得使用外部电子设备驱动阵列变得更加困难。在过去的十年中，电容式微机械超声换能器(CMUT)阵列为高频阵列的实施带来了巨大的希望：1)CMUT的工作频率由薄振板的宽度和厚度设置，该振动板可以精确地定义为亚微米级的特征，从而实现高效的高频超声转换。2)通过光刻图形，可以将密集布置的阵列元件彼此隔离，而不需要划片。3)实验证明，相邻CMUT元件之间的串扰可低至-39分贝。4)电子电路可以方便地与换能器阵列集成在同一衬底上，也可以通过芯片到芯片的键合来实现高信号质量、低噪声和宽带宽。CMUT有可能进一步扩展工作频率，实现高频2-D阵列和具有其他几何形状的阵列，并通过利用批量微制造来降低阵列制造成本。为了演示高频宽带仰角聚焦CMUT，我们确定了两个具体目标：具体目标1：设计、实施和测试中心频率为40 MHz和60 MHz、部分带宽大于100%的256元一维线性CMUT阵列。具体目标2：开发一种在薄衬底上沿仰角方向绘制一维线性阵列的工艺，并演示仰角聚焦，而无需在阵列前面使用光泽透镜。

DESCRIPTION (provided by applicant): Technological difficulties associated with high-frequency (>20 MHz) transducer arrays have prevented wide adoption of high-resolution imaging systems for many decades. Recently, high-frequency arrays and a supporting system operating in the 20 - 70 MHz range have been commercially developed for small animal imaging. However, the extremely expensive cost of this system (>$500 k) and associated transducer probes (~$20 k) is still a huge barrier for widespread use of this preclinical tool as well as the expansion of array-based high-resolution imaging to clinical applications in ophthalmology, dermatology, and cardiovascular medicine. The major hurdles for the development of high-frequency arrays have been related to the conventional manufacturing techniques used for piezoelectric transducers: 1) The thickness of the piezoelectric crystal needs to be tens of microns for high-frequency operation. Using conventional techniques such as lapping and polishing it is very difficult to thin crystals down to required thicknesses. 2) The pitch of the array should be on the order of a wavelength (~50 m for 30 MHz). When the crystal is diced using a standard 10 to 15-�m blade, significant portion of the active area is wasted for element separation. Therefore, methods such as plasma etching or laser machining should be employed for dicing the crystal. 3) Depending on how the kerfs are filled, increased cross coupling could be observed between elements. 4) The small element size results in high electrical impedance and makes it more difficult to drive the array using external electronics. Capacitive micro machined ultrasonic transducer (CMUT) arrays have demonstrated over the last decade that they hold a great promise for the implementation of high-frequency arrays: 1) The frequency of operation of CMUTs is set by the width and thickness of a thin vibrating plate, which can be precisely defined with sub-micron features to enable efficient high-frequency transduction of ultrasound. 2) By lithographic patterning, densely placed array elements can be isolated from each other with no need for dicing. 3) It has been experimentally demonstrated that crosstalk in neighboring CMUT elements can be as low as -39 dB. 4) Electronic circuits can be conveniently integrated with transducer arrays on the same substrate or by chip-to-chip bonding to achieve a high signal quality, low noise, and wide bandwidth. CMUTs have the potential to further extend the frequency of operation, enable high-frequency 2-D arrays and arrays with other geometries, and lower the cost of array manufacture by taking advantage of batch microfabrication. To demonstrate high-frequency wideband elevation-focused CMUTs we have identified two specific aims: Specific Aim 1: Design, implement, and test 256-element, 1-D linear CMUT arrays operating at 40 MHz and 60 MHz center frequencies with a fractional bandwidth greater than 100%. Specific Aim 2: Develop a process to curve 1-D linear arrays on thin substrates in the elevation direction and demonstrate elevation focusing without using a glossy lens in front of the array.