Design and simulation of artificial retinal stimulation IC with switched capacitor using Si nanowire optical properties

Abstract

This study introduces an approach for converting the current from a sensor into controllable voltage. To this end, a switched-capacitor structure was integrated to provide efficient current-to-voltage conversion. The generated voltage was further regulated by an operational amplifier current source, enhancing stability and precision. An n-type metal oxide semiconductor field-effect transistor structure under an H-bridge was integrated into the system to achieve fine-tuned control over current stimulation. This component contributed to voltage regulation and enabled bi-directional control of current flow, offering versatility in adjusting current amplitudes using working and counter electrodes. This dynamic control mechanism was pivotal for effectively controlling the intensity of current stimulation. We applied Verilog-A modeling to simulate the optical characteristics of Si nanowires. The proposed system efficiently converted sensor-derived current into voltage using a switched-capacitor structure. Simultaneously, the precision was enhanced via operational amplifier regulation and n-type metal-oxide-semiconductor field-effect transistor-based H-bridge control. The simulation showed a current stimulus amplitude ranging from 2 to 13 μA for a variable photocurrent of Si nanowires (Rex: 10 kΩ, pulse: 100 Hz, 1 ms). The ability to finely control current stimulation intensity holds promise for diverse applications requiring accurate and adjustable current manipulation. This study contributes to the growing field of sensor technology by offering a unique perspective on the integration of nanostructures and electronic components for an enhanced control and functionality.

Introduction

Retinal prostheses are a promising technology for restoring vision in individuals who have lost their eyesight because of retinal degenerative diseases. These devices are categorized on the basis of the location of electrode insertion and method of receiving visual information from the front.^1,2 The position of the electrode insertion varies with the target cell for stimulation; thus, several types of retinal prostheses can be obtained. Various methods, including the modulation of electrical pulses to control the delivery of charge to the retinal tissue, have been employed in these prostheses. Techniques such as pulse amplitude modulation (PAM), pulse width modulation, and pulse frequency modulation are commonly used in medical applications to precisely control the amount of charge delivered for electrical stimulation in retinal prostheses. ³ Intentional current injection is a common practice in such medical applications. In retinal prosthetics, ensuring stable operation and minimizing device size are crucial. Therefore, the integration of technologies that provide reliable and compact solutions for electrical stimulation is crucial. The Argus 2 system was the first commercially available product to process visual information through a camera and apply current stimulation patterns to electrodes inserted into the eye. ⁴ For PRIMA system, visual information is processed using a camera and current stimulation patterns are provided by irradiating the photodiode array inserted into the eye using a short-wavelength laser. ⁵ For Alpha AMS, current stimulation is applied using a microphotodiode array. ⁶ Although each has its advantages and disadvantages, Argus 2 and PRIMA use a camera. By contrast, Alpha AMS does not use a camera, but the system is reported to be bulkier than those employed in other methods and has a problem when inserted. Therefore, we aimed to simulate whether a system such as the Alpha AMS can be implemented using the parameters of a Si nanowire photodetector, which can be implemented in a smaller size than that of a microphotodiode array. Nanowires are known for their high surface- to-volume ratio, which significantly enhances their photoresponsivity. The high surface area of nanowires allows for greater interaction with incident light, improving the efficiency of photodetection. Recent advancements in this field involve the integration of sensors for visual information, enabling the implantation of devices that receive visual input internally rather than relying on external cameras. This integration significantly enhances the capabilities of the retinal prostheses. These prosthetic devices offer users a natural and responsive visual experience by eliminating reliance on external cameras. Reliance on cameras often creates a gap between artificial and natural vision, such as the lack of eye movement when not turning one's head. Therefore, this development bridges the gap between artificial and natural vision and enables advancement toward user-friendly retinal prosthetics without the need for external devices. These advancements show significant potential for improving the quality of life of individuals with vision loss, creating possibilities for integrated and immersive visual perception that closely aligns with the natural functioning of the human eye.^6,7

Circuit design

Retinal prostheses are classified on the basis of the location of electrode insertion and method used to perceive visual information. The placement of electrodes varies with the target cells used for stimulation. Some designs incorporate both electrodes and visual information reception within the eye, typically by capturing visual data from an external camera and converting them to stimulate cells. A design has been developed to enhance this approach by integrating a sensor that receives visual information directly, along with electrodes, into the eye. For this purpose, a circuit that combines Si nanowire photodetectors and biphasic stimulators has been designed. This integration aims to merge the reception of visual information with electrodes within the eye. The circuit design was developed using a TSMC 180 nm process design kit (Figure 1). ⁸

Figure 1.

Overview of the proposed implantable retinal prostheses.

Design of Verilog-A for Si nanowire

For the Verilog-A model of Si nanowires, we fabricated nanowires and measured their currents. The width of the fabricated nanowires was 150 nm with a spacing of 2 μm, and measurements were taken from 250 Si nanowires.^9,10 During the fabrication process, SiO₂ and Si were etched following wet oxidation and deep reactive ion etching to produce Si nanowires. ¹¹ The method of measurement involved altering the lighting conditions (specifically white light), as illustrated in Figure 2(a), and recording the resultant changes in the current flowing through the Si nanowire. This experiment confirmed the optical properties of the Si nanowires, enabling the design of a Verilog-A model for Si nanowires. A broader range was considered for the simulation to accommodate potential variations owing to differences in the number of wires or fabrication errors. A broad range was considered in the simulation. As shown in Figure 2, the resistance of the Si nanowire photodetector varied from hundreds of kilohms to 10 MΩ. Consequently, the sweep of the simulation ranged from 10 kΩ to 11 MΩ, including the modeling of the variable resistance as shown in Figure 2(b). Therefore, a capacity of 50 pF was used because the capacitance must be sufficiently large to obtain a distinguishable charging voltage when the photocurrent of the Si nanowire is increased proportionally with the amount of light.

Figure 2.

(a) Si nanowire IV characteristics ¹⁰ and (b) Si nanowire Verilog-A model.

Design of switched capacitor structure

Si nanowire sensors are light-responsive current sources capable of adjusting current on the basis of light intensity. These sensors dynamically alter their current output in response to varying light conditions, rendering them highly suitable for applications requiring sensitivity to light intensity. To effectively regulate the current flowing through the working electrode (WE) and counter electrode (CE), a suitable I–V (current-to-voltage) converter was incorporated into the H-bridge configuration. To this end, we employed a switched capacitor structure. This structure was designed to convert the current, which was modulated on the basis of the light intensity detected by the Si-nanowire sensor, into a corresponding voltage. Three specific signals were employed to convert the current into voltage in the switched-capacitor structure: ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}}$ , ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ , and a Verilog-A model input voltage signal. Each signal corresponds to a distinct function. ¹² The Verilog-A model input signal, determined by the resistance of the Si nanowire, was used to measure light intensity. The ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}}$ signal initiates the storage of the current passing through the nanowire in the capacitor, thereby converting it into voltage. ¹³ Subsequently, the ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ a signal was used to reset and discharge the accumulated charge in the capacitor. This signal triad facilitated current conversion, through the Si nanowire into voltage, which varied with light conditions. This voltage was then relayed to the next stage, that is, the stimulator. The charging of the capacitor was governed by specific equations designated as equations (1), (2), and (3). By referencing these equations, the corresponding voltage can be calculated on the basis of the charge accumulated in the capacitor.

Additionally, in the designed switched capacitor structure, the resistance of the silicon nanowires and the circuit capacitance function as a low-pass filter, significantly reducing noise before it affects current control in the low dropout regulator via the amplifier. Noise beyond the bandwidth of amplifier is naturally attenuated, and the parallel RC configuration of the WE and CE further filters out switching noise. Considering the millisecond-scale refractory period of the cells, our system operates without high-speed switching, reducing the occurrence of typical switching noise (Figure 3).

$$\mathit{P}\mathit{h}\mathit{o}\mathit{t}\mathit{o}\mathit{c}\mathit{u}\mathit{r}\mathit{r}\mathit{e}\mathit{n}\mathit{t}[\mathit{A}]\times {\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}}[\mathit{t}]=\mathit{Q}[\mathit{C}]$$

(1)

$${\mathit{V}}_{\mathit{c}\mathit{a}\mathit{p}}={\displaystyle \frac{\mathit{Q}}{\mathit{C}}\left[{\displaystyle \frac{\mathit{C}}{\mathit{F}}}\right]}$$

(2)

$$\mathit{V}(\mathit{t})=\left(1-{\mathit{e}}^{-\frac{\mathit{t}}{\mathit{R}\mathit{C}}}\right)$$

(3)

Figure 3.

Schematic of switched capacitor structure.

Design of amplifier

In a previous study, a two-stage operational amplifier (OPAMP) was used to control the gate of an NMOS. ¹⁴ However, the output voltage range obtained by this approach was limited. A folded cascode structure was considered to address this issue, owing to its ability to expand the voltage swing despite a reduction in gain compared with that of the standard cascode. Consequently, a rail-to-rail folded cascode amplifier was developed to address the limitation. ¹⁵ In the sensor section, the current was converted into voltage using a switched capacitor structure, ensuring that the amplifier transformed a stable voltage supply into a variable current source at the NMOS gate. This amplifier was designed explicitly with a rail-to-rail input, enabling the utilization of the voltage input range from 0 to V_dd. ¹⁶ Additionally, the output stage was configured as a single-end output, in contrast to a conventional two-stage bulk-drive folding cascode amplifier, to obtain high gain and performance characteristics (Figure 4).^17,18

Figure 4.

Schematic of rail-to-rail folded cascode amplifier.

Design of biphasic stimulator

A biphasic stimulator was implemented using the widely used H-bridge circuit. Generally, a high-voltage (HV) process injects a high-current amplitude into a stimulator circuit.^19,20 However, because a specific process is limited, the amount of charge required for cell stimulation can be satisfied by using a PCM method and standard process. This setup enables the direction of current flow to the WE and CE to be controlled via the on/off switching of MOSFETs.^21,22 In this arrangement, the H-bridge determines the direction of the current, and the amplitude of the current is controlled using an N-type MOSFET situated beneath the H-bridge, in conjunction with the current source connected to the OPAMP.^23–25 Consequently, the amplitude of the injected current is influenced by the voltage stored in the capacitor, reflected in the output voltage of the OPAMP, and further modulated by R_ex. To effectively simulate this system, we assumed a current pathway through identical electrodes and solutions on both sides, with the solution resistance set at 500Ω between the two RC parallel electrode models. ²⁶ We obtained a max stimulus current of 13 µA. An HV process was optimized using the PAM stimulation method. In existing studies, the HV process design applies a max stimulus current ranging from 0.5 to 12.75 mA.^27–29 This PAM method can cause cell damage problems. Therefore, in this study, we designed a standard process instead of an HV process. This was effective in reducing the maximum current for stimulus. PCM was used to optimize and reduce the current required for cell stimulation. The amount of charge required for cell stimulation was controlled by the number of pulse waves as shown in Figure 5(c).

Figure 5.

(a) Electrode-solution model and (b) model parameter (c) PCM method.

Simulation result

Result of switched capacitor structure

The simulation required four signals: two H-bridge-enabled signals (as shown by ${\mathit{P}}_1$ and ${\mathit{P}}_2$ in Figure 6), ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}},$ and ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ . Each signal was separated into phases of 90° and 100 Hz. The sequence of enable signals proceeded as follows: ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ , ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}}$ , ${\mathit{P}}_1$ , and ${\mathit{P}}_2$ . As expressed in equation (3), the charging time is determined by the size of the capacitor and resistance of the Si nanowire sensor. The Si-nanowire photodetector sensor detected resistance changes based on light intensity, showing an inverse relationship. Therefore, the RC value (time constant) was determined, and the charging graph is shown in Figure 7. The times required to charge the capacitor and the current at which the photocurrent flows through the sensor in response to optical stimuli are inversely proportional. As the light intensity increased, the photocurrent generated by the sensor increased, and additional charges accumulated in the capacitor, causing the voltage to increase.

Figure 6.

Schematic of top cell.

Figure 7.

Charge/discharge capacitor voltage according to the Si nanowire model

Result of amplifier

As mentioned previously, the designed amplifier was a rail-to-rail folded cascode amplifier. Its performance metrics, including the unit gain bandwidth, gain margin, and phase margin, are presented in Table 1. In this design, the accumulated voltage, which was derived from the current flowing through the nanowire, in the switched capacitor structure was used as the input for the amplifier. Subsequently, the voltage was regulated by adjusting the output of the amplifier. The adjusted output was connected to the gate of the MOSFET, which regulated the current flowing into the biphasic stimulator.

Table 1.

Specification of rail-to-rail cascode amplifier.

	Phase margin	Gain margin	Loop gain
Deg	132.414	–	–
Frequency (MHz)	15.053	90.638	–
dB	–	15.033	127

Table 2.

Comparative analysis of retinal prostheses.

Photoreceptor	Camera	Microphotodiode array	Si nanowire photodetector (This work)
Data processing	Necessary	Some necessary 33	Not necessary
Photo sensor placement	Extraocular	Intraocular	Intraocular
Electrode implantation	Optic nerve, subretinal, epiretinal	Epiretinal, subretinal	Epiretinal, subretinal
Reference	30, 31, 32	33, 34	9, 10

Result of biphasic stimulator

The combined operation of a switched capacitor, amplifier, and other components drove the biphasic stimulator. The amplitude of the current flow between the WE and CE was controlled by connecting them to the MOSFET gate. In a biphasic stimulator, the direction of current flow was controlled by the ${\mathit{P}}_1$ and ${\mathit{P}}_2$ signals from the H-bridge, enabling the current to flow in both directions between the WE and CE. Furthermore, under identical conditions of Si-nanowire current regulation, the amplitude of the current flowing through the electrode is shown in Figure 8. During the simulation, current data were not visible because of a glitch during the switching operation; therefore, the simulation value was used for the plot by adjusting the sampling rate to remove the glitch (Figure 9 and Table 1).

Figure 8.

Current amplitude flow in WE-CE according to the capacitor charging voltage. CE: counter electrode; WE: working electrode.

Figure 9.

System layout result.

Conclusion

In this study, we simulated a circuit to drive a stimulator to display its responsiveness to the photoreactivity of Si nanowires using the optical characteristics of the nanowires. This demonstrated the potential use of Si nanowires in retinal prostheses by translating variations in current influenced by the optical properties of the Si nanowires into changes in the current amplitude of the biphasic stimulator. The switched capacitor structure is essential in efficiently converting sensor-derived current into voltage, thereby regulating current flow through the WE and CE. When ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ , ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}},{\mathit{P}}_1$ , and ${\mathit{P}}_2$ had a 100 Hz frequency and phase shift of 90° of each signal and when the pulse width was 1 ms and R_ex was 10 kΩ, WE and CE were regulated from 2 to 13 µA depending on the current flow through the Si nanowires. This was activated by injecting a charge of 2.5 nC in the case of nerve cells, indicating that the simulated circuit could be stimulated with an injection amount ranging from 2 to 13 nC. ³⁵ In addition, as shown in Figures 7 and 8, because the capacitor charging voltage was maintained from after ${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{i}\mathit{t}}$ to before ${\mathit{P}}_{\mathit{r}\mathit{s}\mathit{t}}$ , PCM could be implemented by providing ${\mathit{P}}_1$ and ${\mathit{P}}_2$ signals several times. The amplitude of the injected current was finely tuned by the dynamic interaction of the Si-nanowire sensor, switched capacitor structure, amplifier, and biphasic stimulator. The simulation results validated the efficacy of the proposed circuit, with the switched-capacitor structure exhibiting prompt charge and discharge responses, the amplifier maintaining the desired specifications, and the biphasic stimulator controlling the current amplitude under various conditions. Given the millisecond-scale refractory period of the cells, our system does not require high-speed operation, which should result in minimal power consumption. However, a comprehensive analysis of power consumption will be conducted in future work once the chip has been fabricated and fully tested. This integrated approach emphasizes the characteristics of Si nanowires and contributes to the development of retinal prostheses, presenting a route toward refined and adaptable current manipulation for enhanced functionality in medical applications.