Structure-property relationships in the optimization of polysilicon thin films for electrical recording/stimulation of single neurons

Abstract

We had earlier demonstrated the use of polysilicon microelectrodes for recording electrical activity from single neurons in vivo. Good machinability and compatibility with CMOS processing further make polysilicon an attractive interface material between biological environments on one hand and MEMS technology and digital circuits on the other hand. In this study, we focus on optimizing the polysilicon thin films for (a) electrical recording and (b) stimulation of single neurons by minimizing its electrochemical impedance spectra and maximizing its charge storage/injection capacity respectively. The structure-property relationships in ion-implanted (phosphorus) LPCVD polysilicon thin films under different annealing and doping conditions were carefully assessed during this optimization process. A 2D model of the polysilicon thin film consisting of 4 grains and 3 grain boundaries was constructed and the effect of grain size and grain boundaries on dc resistivity was simulated using device simulator ATLAS. Optimal processing conditions and doping concentrations resulted in a 10-fold decrease in electrochemical impedance from 1.1 kOmega to 0.1 kOmega at 1 kHz (area of polysilicon interface = 4.8 mm(2)). Subsequent characterizations showed that evolution of secondary grains within the polysilicon thin films at optimal doping and annealing conditions (10(21)/cm(3) of phosphorus and annealed at 1200 degrees C) was responsible for decreasing the impedance. Cyclic voltammetry studies demonstrated that charge storage properties of low doped (10(15)/cm(3)) thin films was 111.4 microC/cm(2) in phosphate buffered saline which compares well with platinum wires (approximately 50 microC/cm(2)) and the double-layered capacitance (C(dl)) could be sustained between -1 to 1 V before breakdown and hydrolysis. We conclude that polysilicon can be optimized for recording and stimulating single neurons and can be a valuable interface material between neurons and CMOS or MEMS devices.

A Materials Roadmap to Functional Neural Interface Design

“…[46] These grain sizes dramatically impact the conductivity, making them more conductive than their amorphous material counterparts. [47–49] Polycrystalline silicon (Polysilicon) is a highly pure, polycrystalline form of silicon that is easily fabricated into microelectrodes and is compatible with complementary metal-oxide-semiconductor (CMOS) [50] and MEMS [51] technology. Additionally, the conductive properties of polysilicon can be modulated by varying processing conditions or dopant densities.…”

Highly Doped Polycrystalline Silicon Microelectrodes Reduce Noise in Neuronal Recordings In Vivo

“…9) and a constant phase element (CPE in Fig. 9) [16]. The resistance offered by the bulk polycrystalline silicon is represented by R bulk in series with CPE and can be modeled using (1).…”

“…In a typical metal microelectrode based recording system, most of the neuronal signal and noise is captured by the internal resistance of the amplifier due to its relatively large value compared to the electrical impedance of the interface and R bulk . For the polycrystalline silicon microelectrode, the interface impedance is mostly capacitive with a magnitude of approximately 1 MΩ at 1 kHz [16] and the bulk resistance, R bulk for the neuronal signal (action potential amplitudes larger than 15 μ V) is 9 kΩ as obtained from the measured I–V characteristics. However, for noise amplitudes typically in the range of 0–15 μ V in vivo , bulk resistance R bulk of the polycrystalline silicon can now be estimated using (1) with doping concentration of 10 21 /cm 3 , estimated trap density of 1.6 × 10 9 /cm 2 , the dimensions of the microelectrode (50 μ m × 4 μ m cross-section), and the measured grain size of approximately 60 nm and is found to be 1.1 MΩ at 5 μ V. For voltages smaller than 5 μ V, the bulk resistance increases exponentially beyond the estimated 1.1 MΩ.…”

Integrated circuit amplifiers for multi-electrode intracortical recording

“…A more nuanced model that captures resistive and frequency-dependent effects is shown in figure 7(b). Although there are loose associations of physical mechanisms to each element, the model is empirical (McAdams andJossinet 1996, Yoshida andStruijk 2004), and the parameters are found by curve fitting (Weiland and Anderson 2000, Otto et al 2006, Saha and Muthuswamy 2007. The model and its elements have subtle variations and many names 6 .…”