**4. Wireless interface for implantable sensor**

From the perspective of a circuit designer, employing of short-channel transistors is an appealing and efficient method to reduce the SoC chip area. The added advantage of these transistors is that the power supply voltage also downscales proportionally with the reduction of the transistor channel length. Thus hot electron effect and time-dependent dielectric breakdown (TDDB) [4] do not deteriorate the robustness and reliability of the devices. However, the threshold voltage, VT of the MOSFET does not scale down as aggressively as the channel length or the power supply voltage [5], which puts a constraint on the number of transistors that can be cascaded in a given process. Some of the techniques for low-power circuit design include bulk-or body-driven technique, floating gate technique, subthreshold biasing scheme of the transistors etc. The following sections summarize various low-power

The bulk (or body-)-driven scheme is a technique that allows circuit designers to implement ultra-low voltage and ultra-low power system. In the bulk-driven circuit design scheme, the bulk or the body terminal of a MOSFET is enabled as an AC input. The gate of the device is kept at a certain potential so that the transistor is 'on' during the entire operation. This technique eliminates the threshold voltage limitation mentioned earlier and helps achieve ultra-low supply voltage requirement while cascading a good number of transistors. Circuit designers have used this technique to design a low-voltage, low-power amplifier with a supply voltage of as low as 1 V [3]. A major drawback of this approach is the lower body transcon‐ ductance (*g*mb). In addition, the bulk-driving voltage needs to be kept within a certain range so that the body diodes of the MOSFET are reverse-biased during the variation of the AC input signal of the body terminal. Therefore, this technique can only be applied to a limited number

Although the implementation of the floating gate technique is mostly seen in the integrat‐ ed memory cell applications, it can also be used for designing low-power circuits for implantable sensors. Floating gate is the polysilicon gate of the MOSFET that is surround‐ ed by silicon dioxide (SiO2). Once the charge has been deposited on the floating gate, it can be stored permanently. Therefore this technique is suitable for flash memory cells. The amount of this charge can be adjusted by an ultraviolet (UV) light or a large gate volt‐ age. The stored charge on the floating gate can be used to reduce the threshold voltage of the transistor. Thus this technique helps reduce the DC supply voltage requirement as well

Another method that is being used to implement the low-power circuit is the subthreshold design technique. In subthreshold region (or weak inversion region) design, the gate-to-source voltage of the MOSFET is biased below the threshold voltage (*V*GS *≤ V*TH) of the transistors. This level of gate-to-source voltage can weakly invert the transistor channel underneath the gate.

circuit design techniques reported in literature.

**3.1. Bulk-driven technique**

154 Advances in Bioengineering

of applications [6].

**3.2. Floating gate technique**

as the total power consumption [7].

**3.3. Subthreshold design technique**

For tether-less operation and to avoid skin infections, the data signal from the signal processor needs to be wirelessly transmitted to the outside environment. Several research works have been reported in recent years to meet the design requirements of wireless operation in biomedical applications. These systems should be miniaturized, light-weight, low-power and reliable for long term operation. The radiated power needs to be less than the limit set by FCC (Federal Communications Commission) for wireless telemetry. Wireless communication is one of the most prevailing means of data transmission for biomedical sensors. Wireless transmit‐ ters and receivers can be found everywhere from short-distance medical endoscopic applica‐ tions [8] to long-distance cell phone communication. Technological advances in silicon manufacturing have made it possible to design low-power, low-cost integrated circuit for biomedical sensing application. Examples include electroencephalography (EEG), electrocar‐ diography (ECG) and biometric information sensing for early detection of diseases such as tumor, cancer, and Alzheimer. Most of these applications require low data rates of a few Hz to a few kHz. ECG monitoring typically needs 12-bit resolution of ADC with 250 Hz data sampling rate for a transmission data rate of 3 kbps [9]. However the minimum energy per bit requirements for implantable sensor put a constraint on the transmitter power consumption to extend the battery life time. Since it would not be feasible to change the battery of the sensor often for an implantable sensor, low-power circuit design is an essential requirement. Con‐ sidering most of the power harvesting techniques as well as battery storage capacities, a power budget of only 100 μW could be available for each sensor node [10]. Although one can relax the power constraints by larger battery/energy-harvester size, the importance of a low-power communication scheme cannot be overlooked for a compact design of the sensor node. Despite the urge for the design of a low-power low-data rate transceiver with traditional narrowband architecture, the best transceiver design in this domain for implantable sensors still consumes about 500 μW of power [11].

The first demonstration of a fully customized mixed-signal silicon chip that had most of the attributes required for a wearable or implantable BSN was described in [12]. The system blocks include low-power analog sensor interface for temperature and pH sensing, a data multiplex‐ ing and conversion module, a digital platform based around an 8-b microcontroller, data encoding for spread-spectrum wireless transmission, and an RF section requiring very few offchip components as shown in Figure 2. A programmable direct-sequence spread-spectrum (DS-SS) transmitter is integrated into the SoC in order to improve the reliability of the wireless transmission [13]. The transmitter is comprised of a data encoder and an RF section. The minimum data rate from the encoder is approximately 3.67 kbps. The amplification stage of the RF section is designed to be a near-class-E RF power amplifier driven by the digital output of the encoder. The gain budget of the amplifier enables it to maintain high gain and linearity while limiting the total current. The on-chip RF section uses a relatively low frequency carrier for modulation. An 800μm × 300μm on-chip spiral inductor transmits the signal that is detectable at a range of 0.5 m in air using a Winradio receiver with a conventional whip antenna at a data rate of up to 5 kbps. Even though the on-chip inductor is less efficient than an external antenna, it demonstrates the feasibility of integrated antennas on silicon [14], [15]. A dataacquisition device detects the signal from the SoC.

**Figure 2.** Schematic diagram of the system-on-chip architecture for body sensor networks [12].

An integrated CMOS ultra-wideband, high duty cycled, non-coherent wireless telemetry transceiver for wearable and implantable medical sensor applications was reported in [16]. A prototype wireless capsule for endoscopy was designed using the proposed transceiver and it demonstrated *in vivo* image transmission of 640 × 480 resolution at a frame rate of

2.5 fps with 10 Mb/s data rate. This transceiver supports scalable data rate of up to 10 Mbps with energy efficiency of 0.35 nJ/bit and 6.2 nJ/bit for transmitter and receiver, respective‐ ly. The block diagram of the transmitter is shown in Figure 3. The transceiver uses On/Off keying (OOK) modulation scheme where a binary "1" is represented by a short pulse and binary "0" is represented by no pulse transmission. For improved performance of anten‐ nas with miniaturized size, UWB frequency band in 3–5 GHz is selected. The UWB pulse is generated by a fast on/off voltage-controlled oscillator (VCO) controlled by the TX data. A driving amplifier provides voltage amplification and isolation between the VCO and the antenna. Both the VCO and the driving amplifier consume power during pulse genera‐ tion only. At the receiver end, the weak signal is first amplified by a variable gain lownoise amplifier (LNA) followed by a squarer performing the energy detection. A variable gain amplifier (VGA) amplifies the squarer output further and a slicer digitizes the final signal. The digital baseband provides synchronization, error correction coding, and interfaces with the external sensors. The transceiver chip consumes a die area of 3 mm × 4 mm when designed in a standard 0.18 μm CMOS process. The transmitter draws an average power of 0.35 mW with the energy per bit of 0.35 nJ/bit for up to 10 Mbps. The receiver average power consumption is as low as 6.1 mW with duty cycling under 1 Mbps data rate and with the energy efficiency maintained at 6.2 nJ/bit.

**Figure 3.** Block diagram of the ultra-wideband wireless telemetry transceiver [16].

budget of only 100 μW could be available for each sensor node [10]. Although one can relax the power constraints by larger battery/energy-harvester size, the importance of a low-power communication scheme cannot be overlooked for a compact design of the sensor node. Despite the urge for the design of a low-power low-data rate transceiver with traditional narrowband architecture, the best transceiver design in this domain for implantable sensors still consumes

The first demonstration of a fully customized mixed-signal silicon chip that had most of the attributes required for a wearable or implantable BSN was described in [12]. The system blocks include low-power analog sensor interface for temperature and pH sensing, a data multiplex‐ ing and conversion module, a digital platform based around an 8-b microcontroller, data encoding for spread-spectrum wireless transmission, and an RF section requiring very few offchip components as shown in Figure 2. A programmable direct-sequence spread-spectrum (DS-SS) transmitter is integrated into the SoC in order to improve the reliability of the wireless transmission [13]. The transmitter is comprised of a data encoder and an RF section. The minimum data rate from the encoder is approximately 3.67 kbps. The amplification stage of the RF section is designed to be a near-class-E RF power amplifier driven by the digital output of the encoder. The gain budget of the amplifier enables it to maintain high gain and linearity while limiting the total current. The on-chip RF section uses a relatively low frequency carrier for modulation. An 800μm × 300μm on-chip spiral inductor transmits the signal that is detectable at a range of 0.5 m in air using a Winradio receiver with a conventional whip antenna at a data rate of up to 5 kbps. Even though the on-chip inductor is less efficient than an external antenna, it demonstrates the feasibility of integrated antennas on silicon [14], [15]. A data-

> Microcontroller

Data Encoder RF Oscillator

On-Chip Spiral Inductor Resistor

SAW Resonator

Digital Divider

Resistor text

RF Relaxation Oscillator

ADC

An integrated CMOS ultra-wideband, high duty cycled, non-coherent wireless telemetry transceiver for wearable and implantable medical sensor applications was reported in [16]. A prototype wireless capsule for endoscopy was designed using the proposed transceiver and it demonstrated *in vivo* image transmission of 640 × 480 resolution at a frame rate of

about 500 μW of power [11].

156 Advances in Bioengineering

acquisition device detects the signal from the SoC.

Sensor Interface

MUX

**Figure 2.** Schematic diagram of the system-on-chip architecture for body sensor networks [12].

pH Sensor

Temperature Sensor

pH Sensor

Temperature Sensor

Cleven *et al.* [17] presented a novel fully implantable wireless sensor system implanted into the femoral artery with computed tomography angiography intended for long-term monitor‐ ing of hypertension patients. The system was employed to measure intra-arterial pressure at a sampling rate of 30 Hz and an accuracy of ±1.0 mmHg over a range of 30–300 mmHg, and consumed up to 300μW power. The implant consists of two functional components: the pressure sensor tip and the transponder unit for communicating with the external readout station. Both the components are linked by a data cable. The full length of the sensor system is approximately 22 cm. The telemetric unit has a diameter of approximately 2 cm and a thickness of approximately 4 mm. The telemetry chip schematic including external compo‐ nents necessary for telemetric mode is presented in Figure 4. The analog output signal from the pressure sensor ASIC is digitized by the sensor readout block. At the same time, the bidirectional data pads provide the offset and the gain settings. The digital component of the chip, a state machine, provides the protocol for data transmission of the measured values. The HF front-end controls the telemetry components while generating the controlled supply voltage required for sensor readout. Using a transmission frequency of 133 kHz, the digitalized information is sent by telemetry to the receiver coil of the external readout electronics.

**Figure 4.** Schematics of transponder ASIC and external circuitry for long-term monitoring of hypertension patients [17].

A transcutaneous two-way communication and power system for wireless neural recording was reported in [18]. Figure 5(a) shows a schematic of the power and bidirectional data transfer system. Wireless powering and 1.25 Mbps forward data transmission (into the body) are achieved using a frequency-shift keying modulated class-E converter. The carrier frequency for reverse telemetry (out of the body) is generated using an integer-N phase-locked loop which provides the necessary wideband data link to support simultaneous reverse telemetry from multiple implanted devices on separate channels. The physical arrangement of the coils is illustrated in Figure 5(b). For the implanted device, coil 1 (L1) represents the external power coil, coil 2 (L2) is the implanted power coil, coil 3 (L3) is regarded as one of the external differential data coils, and coil 4 (L4) is the implanted data coil. A large AC current is generated in coil 1 using a class-E converter to transfer power to the implant. An AC current proportional to the coupling coefficient between the external and the implanted power coils is induced in coil 2. The resulting AC voltage is rectified and supplied to the application-specific integrated circuit (ASIC) to power it up. Frequency-Shift Keying (FSK) modulation of the 5 MHz power carrier at a data rate of 1.25 Mbps is performed to achieve forward data transfer and to send control data to the ASIC. The reference clock is multiplied up by an integer-N PLL in the ASIC circuitry to generate a reverse telemetry carrier between 50 and 100 MHz. The reverse telemetry uses either Amplitude-Shift Keying (ASK) or Binary-Phase-Shift Keying (BPSK) modulation scheme. To generate the reverse telemetry signal, the on-chip driver circuitry induces current in coil 4. Data is received by one of the two external differential data coils, coil 3. The purpose of a differential coil configuration is to cancel both the large power signal at its fundamental frequency and harmonics generated by the class-E converter that fall within the frequency range of the reverse telemetry.

the pressure sensor ASIC is digitized by the sensor readout block. At the same time, the bidirectional data pads provide the offset and the gain settings. The digital component of the chip, a state machine, provides the protocol for data transmission of the measured values. The HF front-end controls the telemetry components while generating the controlled supply voltage required for sensor readout. Using a transmission frequency of 133 kHz, the digitalized information is sent by telemetry to the receiver coil of the external readout electronics.

State

Sensor Readout

Transponder ASIC

EEPROM

Machine HF

Front End

**Figure 4.** Schematics of transponder ASIC and external circuitry for long-term monitoring of hypertension patients [17].

A transcutaneous two-way communication and power system for wireless neural recording was reported in [18]. Figure 5(a) shows a schematic of the power and bidirectional data transfer system. Wireless powering and 1.25 Mbps forward data transmission (into the body) are achieved using a frequency-shift keying modulated class-E converter. The carrier frequency for reverse telemetry (out of the body) is generated using an integer-N phase-locked loop which provides the necessary wideband data link to support simultaneous reverse telemetry from multiple implanted devices on separate channels. The physical arrangement of the coils is illustrated in Figure 5(b). For the implanted device, coil 1 (L1) represents the external power coil, coil 2 (L2) is the implanted power coil, coil 3 (L3) is regarded as one of the external differential data coils, and coil 4 (L4) is the implanted data coil. A large AC current is generated in coil 1 using a class-E converter to transfer power to the implant. An AC current proportional to the coupling coefficient between the external and the implanted power coils is induced in coil 2. The resulting AC voltage is rectified and supplied to the application-specific integrated circuit (ASIC) to power it up. Frequency-Shift Keying (FSK) modulation of the 5 MHz power carrier at a data rate of 1.25 Mbps is performed to achieve forward data transfer and to send

L1

158 Advances in Bioengineering

**Figure 5.** (a) Schematic of bidirectional data transfer system for wireless neural recording, (b) Physical diagram of dual inductive link coils [18].

Cao *et al.* prototyped a device for gastroesophageal reflux disease (GERD) monitoring in [19]. The system consists of an implantable, battery-less and wireless transponder with integrated impedance and pH sensors and a wearable, external reader that wirelessly powers up the transponder and interprets the transponded radio-frequency signals. The total size of the transponder implant is 0.4 cm × 0.8 cm × 3.8 cm and it harvests radio frequency energy to operate dual-sensor and load-modulation circuitry. The system is designed in a way that it can store data in a memory card and/or transmit data to a base station wirelessly. Figure 6 shows the block diagram of the sensor system. The coil antennas and the tuning capacitors form the resonant circuits. Relaxation oscillators are used as the frequency converters in the transpond‐ er. The system is designed to operate at 1.34 MHz since the recommended maximum permis‐ sible exposure of magnetic fields is the highest in the frequency range of 1.34 MHz to 30 MHz [20]. A coil antenna is made using a 34-AWG magnet wire wound around the printed circuit board. The energy harvesting circuit consists of a series of diodes and capacitors (100 pF) in a voltage multiplier circuitry that builds up the DC voltage from the received RF signals [21], [22], [23]. To maintain a constant DC level of 2.5 V for biasing the circuits, a voltage regulator is used.

**Figure 6.** Block diagram of the gastroesophageal reflux disease (GERD) monitoring system [19].

Cheong *et al.* [24] presented an inductively powered implantable blood flow sensor microsys‐ tem with bidirectional telemetry. The microsystem is comprised of silicon nanowire (SiNW) sensors with tunable piezoresistivity, an ultra-low-power application-specific integrated circuit (ASIC), and two miniature coils that are coupled with a larger coil in an external monitoring unit to form a passive wireless link. The implantable microsystem operates at 13.56- MHz carrier frequency. It receives power and command from the external unit and backscat‐ ters digitized sensor readout through the coupling coils. Cheong *et al.* fabricated the ASIC in a standard 0.18-μm CMOS process and the chip occupied an active area of 1.5 × 1.78 mm2 while consuming only 21.6 μW of power. The overall system architecture consisting of an implant‐ able wireless sensor microsystem and an external hand-held device is shown in Figure 7. The ASIC consists of a sensor interface circuit, an analog-to-digital converter (ADC), a digital baseband (DBB), a low-dropout (LDO) regulator, and front-end circuits for wireless powering and bidirectional telemetry. The external monitoring unit needs to be placed in close proximity to the implant microsystem to initiate the passive sensing operation. The RF power is trans‐ mitted by the external unit through the carrier at 13.56 MHz. The parallel resonant LC tanks and the rectifiers convert the received RF signal to a DC signal, and the LDO regulator powers the ASIC with regulated DC supply. Following the demodulation of the incoming modulated carrier, it is de-spread by the DBB to configure the system parameters such as integration time, amplifier gain, selection between two sensors, resonance tuning, and modulation index. At the same time, the clock is extracted from the incoming carrier and is provided to the DBB. Once the system parameters are set according to the received commands, the sensing operation ensues. A successive approximation register (SAR) ADC converts the analog voltage output from the sensor interface circuit into digital data. The digital data is spread and formatted by the DBB and is sent to the load modulator that backscatters the incoming RF carrier according to the sensor data bit stream from DBB.

**Figure 7.** Architecture of the implantable blood flow monitoring system [24].

The above discussion presents an overall picture of the recent evolution in wireless technology for implantable sensors for monitoring of various physiological parameters. The background information and the current trends for the design of a wireless transmitters and receivers are also discussed. For a low-power, non-invasive and unobtrusive performance of an implantable sensor, wireless power transfer is mandatory. In the following sections, a brief discussion on various wireless power transfer and energy scavenging techniques are presented in terms of application requirements, available resources and radiation constraints.
