Pictures of the Future      Spring 2008

From Silicon to Ultrasound

The future of ultrasound may lie in devices that can generate three-dimensional high-resolution images in real time. Such devices will rely on transducers that consist of tiny vibrating membranes that are applied to a silicon wafer.

A doctor places a small scan head on a patient’s breast and pushes a button, instantly creating an image. The resulting data is transmitted to a computer, which scans the image for suspicious millimeter-sized objects that could develop into breast tumors.

Although this might sound like mammography, the system requires no X-rays, yet provides high-resolution 3D images. Even though this examination method is not yet available, the Silicon Ultrasound technology developed by Siemens could make it a standard procedure in just a few years.

Ultrasound images are generated using the echoes of sound waves that are reflected or scattered at the boundaries between different types of tissue within the body. In most of the systems used today, piezoelectric ceramic elements in the ultrasound scan head transmit short directed signals and record the resulting echoes. The depth of the reflecting structure is computed from the signal’s travel time. "As is the case with all examinations using waves, resolution is dependent on wavelength," says Peter-Christian Eccardt, who for many years has been responsible for the development of new types of micromechanical ultrasonic transducers at Siemens Corporate Technology. "The shorter the wavelength, the higher the frequency and, consequently, the greater the detail. But the piezoelectric ceramic elements, which are up to 250 µm in size, can’t be made much smaller. It’s also difficult to arrange them in the two-dimensional arrays that are required for the generation of three-­dimensional ultrasound images," adds Eccardt.

The ultrasound devices in doctors’ offices use frequencies between 2 and 15 MHz and work in two dimensions. They use up to 250 separate piezoceramic elements arranged next to each other on a scan head. The elements take vertical measurements of the section of the body located directly beneath them. This requires the doctor to move the scan head across the patient’s body to measure each individual section. "But in the future we’ll no longer use a linear scan head to measure a two-dimensional section. Instead, we’ll use a flat head to make a 3D snapshot of a complete volume in real time," predicts Eccardt.

In addition to largely automating measurements and evaluations, this would improve the quality of scans because they would no longer depend on how the doctor moved the scan head across a patient’s body. The new technology would thus speed up volume measurements so much that it would even be possible to produce images of moving organs such as the heart. Because this process incorporates the dimension of time, it is also referred to as 4D ultrasound. The system uses a sophisticated algorithm to offset continuous interfering motions such as those caused by the flow of blood.

However, 4D ultrasound would be almost impossible to achieve using piezoceramic elements. It would also require the integration of electronics into the scan head, since each individual element needs its own wiring. As a result, a scan head with 250 elements to a side would require more than 60,000 lines, creating a cable as thick as an arm. Even 2,000 lines would stretch the maneuverability of such a device to the limit.

Silicon instead of Piezoceramics. To make extremely high-resolution 3D and 4D ultrasound examinations possible, Siemens has therefore developed a completely new kind of technology: Silicon Ultrasound. "Since 1996, we’ve been exploring the idea of using semiconductor materials instead of piezoceramics. That’s because it was clear that doing so would allow us to achieve dimensions of less than one micrometer while at the same time enabling us to make use of inexpensive semiconductor production methods. What’s more, this approach lets us integrate part of the evaluation electronics into the scan head," explains Eccardt.

To produce Silicon Ultrasound systems, vibration membranes measuring between 50 and 60 µm are created on the surface of silicon wafers. The membranes are arranged in line with the needs of the finished scan head (see box). Because each element can be controlled by means of a line-column addressing system, a surface containing n columns and m lines no longer requires n × m wires, as is the case with piezoceramics, but only n + m wires. At the same time, the smaller dimensions of the sound-generating elements make it possible to reduce wavelengths when needed.

In addition, the new approach offers greater flexibility in controlling ultrasound wavelengths and scan head sound fields. Because the ultra-thin membranes are acoustically better adapted to the human body than are piezoceramics, the individual elements can be used across a greater range of frequencies. What’s more, improved configuration and addressing options make it possible to link the elements in almost any desired combination. This is particularly useful because lower-frequency sound waves penetrate further into tissue, making it possible to examine deeper layers. However, since the individual sound generators in a Silicon Ultrasound system can be very small, very high frequencies can be achieved as well, allowing higher resolutions than with piezoceramic elements.

This fact opens up entirely new areas of use for the technology. "Besides its potential uses for screening tests and the early recognition of breast cancer, ultrasound could be employed to detect prostate cancer and thyroid tumors. Other extremely promising applications for the system include the early recognition of diseases of the cardiovascular system and the heart muscle," says Klaus Hambüchen, head of the Ultrasound Business Unit at Siemens Medical Solutions in Mountain View, California.

Another advantage of Silicon Ultrasound technology is that it would make very small ultrasound catheter probes possible, which would substantially expand the system’s range of possible applications in hospitals. Such ultrasound probes would improve the visualization of heart functions and help doctors recognize plaques and obstructions.

To speed up its work on Silicon Ultrasound technologies, Siemens acquired Sensant Corporation in 2005. Based in San Leandro, California, Sensant was established in 1998 and has primarily focused on such technologies. In combination with Siemens’ existing ultrasound systems and the developments made by Siemens Corporate Technology, the experience gained by Sensant might enable Siemens to launch the first product on the market as early as next year.

"Our engineers are making steady progress. The first clinical results in breast and thyroid imaging show that we can expect Silicon Ultrasound technology to boost spatial and contrast resolution by a factor of ten," says Hambüchen. However, the aim is not only to create better images. Because the new technology also automates imaging to a certain extent, it makes images more comparable, as their quality is no longer dependent on the skill of the doctor guiding the probe. For Hambüchen this means "that doctors and hospitals will be able to raise their quality standards while at the same time cutting costs. All of this will ultimately benefit the patients."
                                                                                                                Bernhard Gerl