Design, development and evaluation of a high spatial density CMOS chip for bidirectional communication with electrogenic cells
- Entwurf, Entwicklung und Evaluation eines hochauflösenden CMOS-Chips für die bidirektionale Kommunikation mit elektrisch aktiven Zellen
Yegin, Ugur; Mokwa, Wilfried (Thesis advisor)
Aachen : Publikationsserver der RWTH Aachen University (2012)
Dissertation / PhD Thesis
Aachen, Techn. Hochsch., Diss., 2012
In this project, we present 2 CMOS chips designed in our institute with the aim of establishing a bio-electronic interface with electrically active cells. A measurement and control setup consisting of electronic components responsible for the control of the chips as well as the digitizazion and processing of the data they provide has also been developed. An intuitive LabView program faciliates the user interaction by allowing the setting of all analog bias voltages and digital controls, which are transmitted to the measurement and control setup by a Dynamic Link Library (DLL) (written in C) using a USB interface. The prospect of using a fingerprint sensor chip as an alternative bio-electrocnic device was also investigated in addition to the development of the said chips and their peripheral systems. Similarly, a measurement and controls setup as well as another LabView software were also implemented for the fingerprint sensor. The 2 afromentioned, in-house designed chips are called Calibur and Zulfiqar, with the latter being the one developed later. Calibur was manufactured with a 3 metal 2 poly (3M2P) 0.5μm technology from ON Semiconductor, while a 5 metal 1 poly (5M1P) 0.35μm technology was chosen for Zulfiqar. Both chips were first simulated using the Cadence IC 5.1.41 software on the circuit level to test their electrical features. Later, their individual layouts were created and evaluated using the Calibre tool (Mentor Graphics). The final fabrication of both chips took place in the cleanroom facillities of ON Semiconductor through the Europractice membership of our institute. The first chip, Calibur, was developed by Dr. Mathias Schindler during his PhD. A 64x64 pixel (12.5μm pitch) array was implemented at the center of the chip, followed by a bank of 64 amplifiers (1 per row of the array) to ist right hand side. These amplifiers carried out the on-chip amplification of the analog signals recorded from electrogenic cells, prior to their transfer to the measurement setup. The central array was flanked by 2 decoders to its left and bottom side: A row and column decoder. These digital circuits were responsible for the activation and calibration of entire columns, as well as for the selection of individual pixels for electrical stimulation. Further down on the signal path, a bank of 8 8-to-1 multiplexers were implemented on the right hand side of the 64 amplifiers to reduce the number of simulatenously read-out channels from 64 to 8. The outputs of the multiplexers were connected to I/O pads, transferring the signals off-chip. Zulfiqar was implemented using a similar floor plan. The use of a different technology however, madet the re-design of the entire circuitry and layout along with it necessary. The greatest challenge in the Zulfiqar project was the integration of a gain stage directly into each pixel in the central array to amplify the analog signals directly at their source, prior to their amplification by the 64 off-pixel amplifiers. To this end, a differential common-source amplifier was implemented in each pixel. The avenue of using the pixels bi-directional, meaning both for read-out and stimulation was preserved by adapting the circuitry. The neccessity to keep the pixel size and pitch as small as possible imposed restrictions on the complexity and the dimensions of the circuit implemented. After the passing through initial gain stage in the pixels, the analog signals were again amplified in the 64 amplifiers on the right hand side of the array. This concept greatly increased the sensitivity of Zulfiqar in comparison to Calibur, enabling the detection of much lower signals. Subsequent to the fabrication, the chips were coated with a thin film high-k material. A great deal of effort was also spent on the optimization of this step, as it was considered to be crucial to improve the chips performances. Different methods of deposition and materials (and combinations of those materials) were tested and characterized in order to achieve a surface passivation with the highest electrical permittivity value possible. As a result, a 3-stack double layer of TiO2 / HfO2 and a 4-stack double layer of Al2O3 / Ta2O5 deposited using a chemical vapor deposition (CVD) based method called atomic layer deposition (ALD) proved acceptable for our purposes. Finally, the fingerprint sensor also underwent electrical characterization to evaluate the prospect of using it as an alternative, cheap and off-the-shelf solution to the custom designed IC’s. Though the proof of principle could be shown by detecting signals delivered to the chip through an electrolyte (much like the characterization of Calibur and Zulfiqar), the range of signals still detectable was well over the "electrophysiologically relavant" region. Thus the fingerprint sensor cannot be used as an alternative for the custom designed IC’s as it is now, after improvements to its sensitivity however, this avenue should be re-investigated.
- Chair of Materials in Electrical Engineering I and Institute of Materials in Electrical Engineering