Analysis of Al2O3-parylene C bilayer coatings and impact of microelectrode topography on long term stability of implantable neural arrays

Objective. Performance of many dielectric coatings for neural electrodes degrades over time, contributing to loss of neural signals and evoked percepts. Studies using planar test substrates have found that a novel bilayer coating of atomic-layer deposited (ALD) Al2O3 and parylene C is a promising candidate for neural electrode applications, exhibiting superior stability to parylene C alone. However, initial results from bilayer encapsulation testing on non-planar devices have been less positive. Our aim was to evaluate ALD Al2O3-parylene C coatings using novel test paradigms, to rigorously evaluate dielectric coatings for neural electrode applications by incorporating neural electrode topography into test structure design. Approach. Five test devices incorporated three distinct topographical features common to neural electrodes, derived from the utah electrode array (UEA). Devices with bilayer (52 nm Al2O3 + 6 mu m parylene C) were evaluated against parylene C controls (N >= 6 per device type). Devices were aged in phosphate buffered saline at 67 degrees C for up to 311 d, and monitored through: (1) leakage current to evaluate encapsulation lifetimes (> 1 nA during 5VDC bias indicated failure), and (2) wideband (1-10(5) Hz) impedance. Main results. Mean-times-to-failure (MTTFs) ranged from 12 to 506 d for bilayer-coated devices, versus 10 to > 2310 d for controls. Statistical testing (log-rank test, alpha = 0.05) of failure rates gave mixed results but favored the control condition. After failure, impedance loss for bilayer devices continued for months and manifested across the entire spectrum, whereas the effect was self-limiting after several days, and restricted to frequencies < 100 Hz for controls. These results correlated well with observations of UEAs encapsulated with bilayer and control films. Significance. We observed encapsulation failure modes and behaviors comparable to neural electrode performance which were undetected in studies with planar test devices. We found the impact of parylene C defects to be exacerbated by ALD Al2O3, and conclude that inferior bilayer performance arises from degradation of ALD Al2O3 when directly exposed to saline. This is an important consideration, given that neural electrodes with bilayer coatings are expected to have ALD Al2O3 exposed at dielectric boundaries that delineate electrode sites. Process improvements and use of different inorganic coatings to decrease dissolution in physiological fluids may improve performance. Testing frameworks which take neural electrode complexities into account will be well suited to reliably evaluate such encapsulation schemes.