Biocompatible Low-Cost CMOS Electrodes for Neuronal Interfaces, Cell Impedance and Other Biosensors

  • Tony Graham

Student thesis: Doctoral ThesisPhD


The adaptation of standard integrated circuit (IC) technology for biosensors in drug discovery pharmacology, neural interface systems, environmental sensors and electrophysiology requires electrodes to be electrochemically stable, biocompatible and affordable. Unfortunately, the ubiquitous IC technology, complementary metal oxide semiconductor (CMOS), does not meet the first of these requirements. For devices intended only for research, modification of CMOS by post-processing using cleanroom facilities has been achieved by others. However, to enable adoption of CMOS as a basis for commercial biosensors, the economies of scale of CMOS fabrication must be maintained by using only low-cost post-processing techniques. The scope of this work was to develop post-processing methods that meet the electrochemical and biocompatibility requirements but within the low-cost constraint. Several approaches were appraised with the two most promising designs taken forward for further investigation. Firstly, a process was developed whereby the corrodible aluminium is anodised to form nanoporous alumina and further processed to optimise its impedance. A second design included a noble metal in the alumina pores to enhance further the electrical characteristics of the electrode. Experiments demonstrated for the first time the ability to anodise CMOS metallisation to form the desired electrodes. Tests showed the electrode addressed the problems of corrosion and presented a surface that was biocompatible with the NG108-15 neuronal cell line. Difficulties in assessing the influence of alumina porosity led to the development of a novel cell adhesion assay that showed for the first time neuronal cells adhere preferentially to large pores rather than small pores or planar aluminium. It was also demonstrated that porosity can be manipulated at room temperature by modifying the anodising electrolyte with polyethylene glycol. CMOS ICs were designed as multiple electrode arrays and optimised for neuronal recordings. This utilised the design incorporating a noble metal deposited into the porous alumina. Deposition of platinum was only partially successful, with better results using gold. This provided an electrode surface suitable for electric cell-substrate impedance sensors (ECIS) and many other sensor applications. Further processing deposited platinum black to improve signal-to-noise ratio for neuronal recordings. The developed processes require no specialised semiconductor fabrication equipment and can process CMOS ICs on laboratory or factory bench tops in less than one hour. During the course of electrode development, new methods for biosensor packaging were assessed: firstly, a biocompatible polyethylene glycol mould process was developed for improved prototype assembly. Secondly, a commercial ‘partial encapsulation’ process (Quik-Pak, U.S.) was assessed for biocompatibility. Cell vitality tests showed both methods were biocompatible and therefore suitable for use in cell-based biosensors. The post-processed CMOS electrode arrays were demonstrated by successfully recording neuronal cell electrical activity (action potentials) and by ECIS with a human epithelial cell line (Caco2). It is evident that these developments may provide a missing link that can enable commercialisation of CMOS biosensors. Further work is being planned to demonstrate the technology in context for specific markets.
Date of Award1 Jun 2010
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorJohn Taylor (Supervisor) & Chris Bowen (Supervisor)


  • Biosensor
  • Biocompatibility
  • Integrated circuit
  • CMOS
  • Neuron

Cite this