1. Introduction

Glaucoma is a group of eye diseases that impact about 8.4 million people worldwide^[1], and which are characterized by elevated and large fluctuation intraocular pressure (IOP)^[2]. Normally, pressure levels in the eye range from 10 to 21 mmHg, but they can be as high as 50 mmHg in a diseased eye. Continuous measurement of the IOP in glaucoma patients can help in better disease diagnosis, monitoring, and management, which can be achieved with an implanted monitor. Hence, there has been an increasing amount of interest in a wireless implanted intraocular pressure monitoring system^[3-7].

IOP monitoring systems can have two kinds of structure: an LC resonant-based structure and a system on chip (SOC) structure^[8]. The LC resonant-based structure uses an inductance and a capacitive pressure sensor, and the resonance frequency changes with the capacitance of the sensor. The advantages of this structure are that it is simple, robust, and somewhat small. The main drawbacks are that it has a very limited functionality and requires a nearby external data acquisition unit. The SOC structure consists of an IC, a pressure sensor, and an antenna. The great advantage of this structure presented is that it provides more reliability and is able to work over a longer distance.

Although great progress has been made in SOC structure of the implanted system, there are some challenges for achieving high-resolution measurement, reliable wireless communication, and ultralow power. The power dissipations of Refs. [4, 5, 11] are 47 mW, 202.43 $\mu$W and 210 $\mu$W, respectively, which are too high for long-term implant monitoring. In this paper, we describe in detail the design of a wireless implantable IOP monitoring system, including an ASIC chip, a microelectromechanical systems (MEMS) capacitive pressure sensor (E1.3N, microFAB Bremen), and an antenna. The system is battery-free, allowing small size and long lifespan. Power is transmitted by electromagnetic wave. We use intermediate-frequency (IF) modulated backscatter for uplink communication. The signal is encoded by FM0 to improve the reliability of the communication. A low power capacitance-to-frequency (C-to-F) converter is designed to read the pressure sensor. An antenna and rectifier are used to finish the radio-frequency (RF) electromagnetic energy harvesting. In this work, the operating frequency is in the unlicensed 915 MHz industrial, scientific and medical (ISM) band.

2. System design

The ASIC is designed on the SMIC 180 nm technology and consists of a sensor interface, voltage regulators and references, an RF rectifier for remote powering, an amplitude shift keying (ASK) modulator, oscillator, and digital conversion, as shown in Fig. 1. The oscillator (OSC) provides the clock signal for the digital module, the low dropout regulator (LDO) provides a stable supply voltage for other modules, and the power-on reset (POR) takes the reset signal to the OSC and digital module. The MEMS pressure sensor used in our system, the E1.3N, has a sensitivity of 1.6 fF/mmHg in the pressure range of 10-50 mmHg^[9], and the capacitance of sensor changes within the scope of 6.0-6.1 pF. The C-to-F converter converts the capacitance to the frequency of pulse signal and then a frequency divider is used to amplify the period of the pulse signal. The digital logic counts the amplified pulse width, encodes the count results, and then modulates backscatter to the external reader. The FM0 code can reduce the mismatch time of ASK modulator and then improve the energy obtained from RF. The antenna and rectifier convert RF energy to DC energy to power the total system. Compared to the previous systems^[3-5], the proposed system has a reduced number of building blocks and power, and it improves the power conversion efficiency and the reliability of communication by FM0 encoding.

3. Antenna

Several previous designs on intraocular devices used inductive coupling in the kHz to MHz range, requiring a large, multi-turn coil inductor in the implant^{[10, 11]}, and the larger device size requires a larger incision. In contrast to inductive coupling, full-wave electromagnetic analysis considering tissue loss shows that the optimal frequency for power transfer to a size-constrained antenna is in the GHz range^{[12, 13]}. Other research shows that microwave (above 2.4 GHz) radiation induces irreversible damage to the lens epithelial cells. According to the analysis above, the 915 MHz band is considered to be the most suitable to be chosen as the wireless transmission frequency.

The antenna is a dipole-like structure, as Figure 2(a) shows. Bio-compatible polymethyl methacrylate (PMMA) plastic is used as the substrate material. The central hole of the loop antenna is designed to avoid a decrease in the quantity of light into the pupil. Normally, as the diameter of pupil changes from 2.5 to 4 mm, the diameter of the central hole is 5 mm. Given that the diameter of anterior chamber is about 12.11 mm, the diameter of the antenna is designed to be 11.4 mm. The thickness of the antenna is about 0.3 mm. This antenna is modeled using Ansoft's high frequency structural simulator (HFSS) in eye tissue, Figure 2(b) shows that the simulated gain of the antenna is about -17 dBi, and the $S_{11}$ parameter is about $-29$ dB, as Figure 2(c) shows.

4. Application-specific integrated circuit

4.1 Power transfer and power management

The IOP monitoring system is wirelessly powered by RF electromagnetic energy using an external reader operating at 915 MHz. The on-chip CMOS rectifier picks up the RF electromagnetic energy on the antenna to provide power to the chip. The 2-stage rectifier employs a cross-coupled differential CMOS configuration with a bridge structure, as Figure 3 shows. The power conversion efficiency of the rectifier depends on the turn-on resistance and the leakage power of the MOSFET. The length of MOSFET employs the minimum length of technology, which is 180 nm, and the width of MOSFET is expected to be a compromise between the turn-on resistance and the leakage power.

As the temperature of body almost keeps constant, a bandgap reference circuit is unnecessary for implantable medical devices. A bias circuit working in the subthreshold region is used to provide three stable reference voltages of 1.2, 0.9 and 0.3 V. The total current of the bias circuit is about 200 nA. A low dropout regulator is used to isolate the supply noise and the regulator employs a common PMOS with operational amplifier feedback topology, which has good stability and PSRR without external capacitors. The reference voltage of the regulator is 1.2 V and the output voltage is equal to the reference voltage. The static current of the regulator is about 100 nA.

4.2 Capacitance-to-frequency converter

The C-to-F circuit consists of 555 timer and 100 nA current source, as Figure 4 shows. The 555 timer controls the charge and discharge of the capacitance under test. There is a linear relationship between the sensing capacitor and the period of $T_{\rm signal}$. The relationship is given by

$\begin{equation} T=\frac{C_{\rm X} (V_{\rm H} -V_{\rm L} )}{I_{\rm charge} }+\frac{C_{\rm X} (V_{\rm H} -V_{\rm L} )}{I_{\rm discharge} }. \end{equation}$

(1)

This proportionality holds even when channel length modulation effect is taken into consideration, where $V_{\rm H}$ = 0.9 V, $V_{\rm L}$ = 0.3 V, and $I_{\rm charge}$ = $I_{\rm discharge}$ = 100 nA, when $C_{\rm X}$ = 6 pF, the period of $T_{\rm signal}$ is 72 $\mu$s. When $C_{\rm X}$ increases by 1 fF, the period of $T_{\rm signal}$ would increase by 12 ns. It is hard to count by the digital unit, so 12 ns is amplified to 3.072 $\mu$s by an 8-bit binary frequency divider.

4.3 Digital conversion

The main modules of the digital conversion are the 14-bit counter and FM0 coding. The counter is controlled by a clock, which is generated using a current-starved ring-oscillator topology, whose frequency is determined based on the pulsewidths, specifications of the MEMS sensor, desired pressure range of 50 mmHg, and required sensitivity of at least 0.5 mmHg. Since the output pulsewidth of the frequency divider has a sensitivity of 1.536 $\mu$s/fF, and the sensitivity of the sensor is 1.6 fF/mmHg, the total sensitivity is 2.4576 $\mu$s/mmHg. Therefore, the frequency of the clock is chosen to be 1.28 MHz. The timing diagram of the digital conversion is shown in Fig. 5. The pulse controls the clocking of the binary counter, and the falling edge of the pulse is used to stop and reset the counter and prepare it for the next conversion. After the count sequence, the value of the counter is fed into a parallel-to-serial converter, which is designed using a standard 14-bit shift register topology, to serially stream out the digital data to FM0 encoder. Then, the FM0 data is sent to an ASK modulator. The code of 1010 in Fig. 5 is the result of the counter. In the FM0 encoder, a logical one is encoded in a particular bit-cell by a zero-crossing transition only at the end of the cell, while a logical zero is encoded by adding an additional transition at mid-cell, the encoding result of 1010 is shown in Fig. 5.

In order to reduce the power dissipation of proposed system, most of the modules are worked in the sub-threshold condition. The simulation results of power dissipation of main modules in the ASIC are shown in Table 1. The total power dissipation is about 6 $\mu$A.

5. Experimental results

The ASIC of the IOP monitoring system was fabricated in SMIC 0.18 $\mu$m 2P4M CMOS technology, and it occupies 0.4 × 0.35 mm$^{2}$ silicon area, including the pads, as shown in Fig. 6. The measurement results are summarized in the following section.

5.1 Wireless powering test

The input impedance of the IOP monitoring system was matched to 50 $\Omega$ by the PNA-X network analyzer N5242A, the $S_{11}$ is about -25 dB. Then, the output DC voltage and power conversion efficiency is measured at different RF powers, as shown in Fig. 7. When the supply voltage is above 1.24 V, the total current is 6.1 $\mu$A. The input RF sensitivity is -13 dBm (minimum power needed to allow chip operation). When the supply voltage is 1.246 V, the power conversion efficiency is 15% with a -13 dBm RF input. The total power conversion efficiency depends on the matched-degree of the rectifier's input impedance matching to 50 $\Omega$ and on the rectifier's efficiency. The input impedance of the rectifier changes with the input RF power, the maximal matched-degress happened on -13 dBm RF input power. In addition, the output voltage keeps constant if the input RF power is higher than a certain level because the leakage power of the rectifier increases with the input power, which also causes the reduction of power conversion efficiency.

5.2 Pressure measurement

The ASIC is integrated with the MEMS capacitor and tested with the wireless enabled by the RFID test platform NI-100. Our empirical tests involved enclosing the fully functional wireless prototypes inside a custom-built pressure chamber. We then kept the temperature constant at 37 ℃. Process variations would result in the discrepancy between different tested sample, a two point calibration can be taken on the external reader to get an accurately measured result. Figure 8 shows the measured pressure sensing linearity from 0 to 60 mmHg, the slope is 3.25 $\mu$s/mmHg, and the average error of the measured pressure is about $\pm$1.5 mmHg.

A summary of the IOP monitoring system and its comparison with other state-of-the-art implementations is shown in Table 2. The RF sensitivity of the proposed system is the lowest, but the pressure resolution is higher than other mentioned system, which is caused by the phase noise of the clock of the counter. The measured power consumption of Ref. [3] is 2.3 $\mu$W, which is lower than the proposed system, the reason is that on-chip digitalization circuitry is removed in Ref. [3] and the result of the capacitance-to-frequency converter is sent to ASK modulator directly, hence the count is achieved in the external reader. Wireless transmission deteriorates the measured resolution, so the pressure resolution of 0.9 mmHg is hard to achieve in a practical application. The high pressure resolution of other mentioned system is achieved at the expense of power.

5.3 Biological test results

In order to verify the operation of the IOP monitoring system, the system which contains antenna, ASIC and pressure sensor was tested in pig eyes. The RFID test platform with an outside antenna powers the implanted system and receives the test result. Figure 9 shows the received results on the external reader, which is a demodulated signal. This is the original FM0 code, and the corresponding binary code is 011000000000010, the decimal value is 12290, the corresponding pressure is about 70 mmHg.

6. Conclusion

An ultralow power wireless IOP monitoring system is presented in this work, which contains an antenna, an ASIC, and a pressure sensor. The measurements results of the RF powering system show a low RF sensitivity of $-13$ dBm, which is the lowest of the systems that have so far been reported. The average error of pressure measurements is $\pm$1.5 mmHg. Finally, a complete system is demonstrated, showing successful transmission of captured pressure data.