No abstract

Abstract-This paper presents an investigation into total ionizing dose (TID) effects on I-V and noise characteristics of MOS transistors manufactured in a 0.18 µm CMOS Image Sensor (CIS) process. The CIS are intended for use in space science missions experiencing harsh radiation environments, such as ESA's forthcoming JUICE mission. Devices were therefore irradiated to various TID levels up to 1 Mrad. Following irradiation, significant leakage current and threshold voltage modification was observed, and this was found to be more severe for devices with small channel geometries. Noise spectral density measurements were also performed at the different irradiation steps. Noise in the smaller geometry devices was found to increase following irradiation, whereas for larger devices it was not significantly affected. These findings enable future assessment of the effects of TID on functional and electro-optical characteristics of high performance CIS designs for use in space. Index Terms-CMOS image sensor (CIS), metal oxide semiconductor (MOS) transistor, total ionizing dose (TID).

Pixellated monolithic silicon detectors operated in a photon-counting regime are useful in spectroscopic imaging applications. Since a high energy incident photon may produce many excess free carriers upon absorption, both energy and spatial information can be recovered by resolving each interaction event. The performance of these devices in terms of both the energy and spatial resolution is in large part determined by the amount of diffusion which occurs during the collection of the charge cloud by the pixels. Past efforts to predict the X-ray performance of imaging sensors have used either analytical solutions to the diffusion equation or simplified monte carlo electron transport models. These methods are computationally attractive and highly useful but may be complemented using more physically detailed models based on TCAD simulations of the devices.Here we present initial results from a model which employs a full transient numerical solution of the classical semiconductor equations to model charge collection in device pixels under stimulation from initially Gaussian photogenerated charge clouds, using commercial TCAD software. Realistic device geometries and doping are included. By mapping the pixel response to different initial interaction positions and charge cloud sizes, the charge splitting behaviour of the model sensor under various illuminations and operating conditions is investigated.Experimental validation of the model is presented from an e2v CCD30-11 device under varying substrate bias, illuminated using an Fe-55 source.

In this paper we present the results from a pilot project at e2v technologies to examine the performance of CMOS Active Pixel Sensors for scientific applications. We describe the characterisation of two prototype 128 × 128 pixel imaging devices with scanning circuitry, as well as 5 × 5 pixel test structures with further variation in pixel design. The main variation in the design is the type of photodiode. In this process two types of diode were available, a 'shallow' n+/p-well diode and 'deep' n-well/p-substrate diode. The characterisation includes the use of photon transfer curves to measure output responsivity and we quantify dark signal variations between pixel structures and reset noise levels. A source of additional dark signal is found to be light emission from the in-pixel transistors. We also present results from an optical characterisation of the stand alone devices, including QE response, MTF and PSF measurements. Finally we outline the considerations to produce such a device using a more advanced process with a smaller feature size.