Beyond simple telehealth: Biosensor frontiers

Although the secure transport of patient information is fundamental to telehealth, it is just the sprawling back-end of a technology ecosystem that is only as good as the sensors themselves.

The second telehealth wave requires a high level of technical sophistication: Sensor systems accurate enough for in-patient hospital monitoring will be integrated into telehealth devices worn permanently on the body because they are comfortable, convenient and unobtrusive.

A reliable, long-lasting power source is a top priority for patient wearability. Batteries may be fine for PDAs but they're not particularly people-compatible. Harvesting energy from the body has numerous advantages, from eliminating the need to change batteries to reducing the amount of toxins polluting the environment.

On average, heat flow from the human body generates a power density of about 20 mW/cm². Advanced thermoelectric generators (TEGs) can deliver an output voltage of 0.7 V at matched load, which is sufficient for power biosensors.

Because they perform the basic heat-to-electricity conversion, thermopiles are a key TEG element. But conventional thermopiles are too expensive because fabrication techniques can't be automated.

Micromachined thermopiles, on the other hand, meet cost and performance criteria and are already being integrated into commercial thermoelectric coolers. The Inter-university Microelectronics Centre (IMEC) has adapted this technology to biosensors. Its micromachined TEG, worn on the wrist, is expected to generate up to 30 µW/cm² and to exceed 4.0 V.

Maximum power generated depends on TEG size, ambient temperature and level of patient's metabolic activity. In a wristwatch configuration, IMEC researchers have found that, at 22°C, the TEG can deliver useful power of 0.2 mW to 0.3 mW during daily activity.

The sensors have the difficult task of acquiring and conditioning a variety of physiological, biological and neurological signals, all of which pose unique challenges in terms of the number of channels required, sensitivity and signal voltage levels. Noise sources such as electrode offset and interference from power distribution circuits makes things more difficult.

Biocompatibility is another concern, particularly when the sensor is embedded in the body. Most materials used in IC production are toxic, especially when the implant is intended to be long-term. Encapsulation provides only a partial answer because sensors use electrodes that, by their nature, must be in direct contact with human tissue. IMEC and other research organizations are developing materals that deposit biofriendly coatings during the fabrication process, much as polysilicon, metal and other layers are deposited. --J.S.