Monolithically fabricated flexible graphene-based active implant

Abstract

Treatment of diseases, illnesses or disorders are always sought without any undesired side-effects and complications. Neurological disorders, such as Epilepsy and Parkinson’s, are currently treated using pharmaceuticals. However, drugs have a low specificity and lose their effectiveness after several years. In addition, prolonged drug usage often leaves patients with many side-effects and complications. Development in neuroscience indicate the treatment of such neurological disorders via neuro-stimulation. Currently, research in understanding the mechanisms behind neuronal stimulation is carried out by research institutes all around the globe. There are three modes for neuronal stimulation, electrical, optical and lastly the use of ultrasonic waves. As optogenetics provides a more specific technique in neuronal stimulation and electrical recording can provide high spatial resolution data, it is sought to combine this two modes.
This work reports the development of a monolithically fabricated active implant, with optogenetic compatibility using transfer-free graphene electrodes. To achieve this desired goal, a microfabrication process is developed, which is reproducible and scalable with modern day microfabrication technology. In order to be compatible with optogenetics, graphene electrodes are used, as these are transparent and allow for neuronal stimulation using light. The graphene electrodes are grown on a pre-defined molybdenum catalyst, allowing for a transfer-free chemical vapor depostion (CVD) of graphene. Microelectrode-arrays (MEAs) are developed, using graphene electrodes, that include different sized electrodes and different amount of electrodes. The amount being constrained by the available space on the cortex of a mouse.
The aim of this work was to be able to control or read out these MEAs using electronics developed alongside the electrodes. As graphene is grown at a temperature above the melting point of aluminium, conventional CMOS technology can not be used in combination with graphene electrodes. The process developed, did not include any materials which would be damaged during the graphene growth. The metal gate is replaced by a polysilicon gate and the metal interconnects connecting the active devices and the MEAs are defined after the growth of graphene. Resulting in an active implant which is monolithically fabricated in combination with a transfer-free graphene process.
Monolithically fabricating an active graphene based implant, comes with complications. Delamination of the passivisation layer occurred whilst trying to open contact openings to and from the active devices. This was resolved using a more appropriate chemical etchant.
A fully monolithically fabricated active graphene-based implant was obtained. Electrical measurements showed that the active devices did not behave as expected. Revisiting simulations, it was established that there was leakage of dopants from the gate to the channel. This results in the NMOS devices to be always on and the PMOS devices to have a higher threshold voltage.
However, this particular wafer, had many high temperature steps after the doping of the polysilicon. Reduction of these high temperature steps result in less doping getting into the channel. Alternatively, a thicker gate oxide can be used, to serve as a more robust barrier. A last proposed solution is the reduction of the doping concentration and doping energy of the polysilicon, which would result in less dopant being near the gate oxide interface and thus less dopant leaking into the channel.
There are wafers included in this work with less high temperature steps. However, due to time limitation it was not possible to finish production of these wafer in order to confirm it experimentally. Nevertheless, it is believed that with minor adjustments, this research project is a viable foundation for active graphene based implants and structures.