A neural probe process enabling variable electrode configurations

Up till now, silicon neural probes have been produced using fixed lithographic mask sets, which is straightforward but inflexible and costly with respect to redesigns for small production volumes. We demonstrate a method to vary the recording site distribution on neural probes with a maskless finishing process. The concept is based on the use of direct write laser lithography (DWL) in one mask layer, thus, enabling on-demand processing of wafers with semi-custom designs at a reasonable cost and lead time. We use the DWL to define windows in the top isolation layer of the device, thus, selecting which electrodes, out of a standardised electrode array, should be active. In addition the active electrode area can be varied. The concept is evaluated using a 64-site neural probe design and manufacturing process. Impedance characterisation is made on active and inactive electrodes and on electrodes with varying active area. The results show ∼15 times lower impedance for active compared to inactive electrodes at 1 kHz, which is considered sufficient for signal discrimination.