Abstract Cobalt60 and 10 MeV proton irradiations have been carried out on n-buried channel frame transfer CCDs in order to study changes in charge transfer efficiency (CTE) and dark current for room tem rature, 1ps/pixel, readout conditions。 Bias dependence and post-annealing effects were observed for iORiZatiOn damage。 CTE effects are explained in terms of ca]2tufe aRd emission from deep level traps。 Temporal instabilities (RTS signals) were observed in the proton induced dark current。84872

I。INTRODUCTION

This paper presents results from a radiation test program aimed at assessing the performance of large area (780 x 1152 pixel) CCDs for a spacebome, Earth imaging instrument, which uses the CCDs in a frame transfer mode at near-room temperature and at a readout rate of approximately I qs/pixel。 It will be seen that CTE effects are not the same as previously reported for low temperature, slow scan applications, such as x-ray astronomy (e。g。 Holland et ai [I])。 Also the near room tetllperature Operation implies that minimization of dark current by operating the CCD surface in inversion is a useful technique。 We report on the use of appropriate clocking ScholTtes (known as dither clocking or dynamic dark current suppression [23) whilst operating in a radiation environment。 The tested devices were all manufactured by EEV Ltd (Chelmsford, UK), however many of the effects can be expected to be displayed by other n-channel CCDs。

Six thinned, back illuminated CCDs will be used in MERIS (Medium Resolution Imaging Spectrometer), which will be a core instrument on ESA's first POlar Earth 0bSOrvatiOR MiSSiOtl (POEM-I) which is scheduled for launch in 1998 into a 800 km Sun synchronous orbit。 The designed life is 4 years during which the CCDs will be exposed to a total dose of approximately I 。4 krad(Si); arising mainly from trapped protons in the South Atlantic Anomaly。 lonizauon damage was tested using a cobalt60 source because initial electron-hole recombination effects are similar to those for high energy protons [3]。 Testing for displacement damage was primarily carried out using 10 MeV protons since, at that energy, predictions for any proton environment can be made using the curve for non-ionizing energy loss (NIEL) [4)。

' The work reported in this paper was conducted under contract to the European Space Agency, ESTEC

The flight CCDs for MERIS are at present under development and a large area back illuminated device was not available for this work。 Accordingly, a variety of commercially available devices was used in order to compare different design features。 A total of 16  devices were irradiated of which 4 had 780x1152 pixels (type CCD05) and the remainder, 385x576 pixels (type CCD02); two of these small chips being thinned and back illuminated。

One of the larger CCDs was fabricated on 100 icm silicon, which is baselined for MERIS because the increased depletion depth (-10pm) gives reduced crosstalk between pixels compared with standard CCDs, which have 20 lcm silicon (4Jun dcplction depth)。 However this device had a thicker epitaxial layer (l00pin rather than 20-25pm which is standard, and also baselined for MERIS)。 All  the devices were   fabricated   using   a   conventional    3-phase,   3-level

polysilicon  process  with  a dual  oxide/SijN4 gain t (each  with  a thickness  of  R5nm)。    With  EEV  devices the

buried channel is kept away from the channel  stop boundaries, where electric fields and interface state concentrations tend to be high。 The pixel area for the larger CCD05 devices is 22。5x22。Sum' and for the CCD02 devices is 22x22pm*。  The channel stop width is 6qm in both cases。

II。EXPERIMENTAL  DETAILS

Five devices (3 powered, 2 not powered) were given up to 4 krad(Si) of Co60 i rradiation at a dose rate of 1。73 krad/hour and 9 were given fluences of up to 7。 2x1@9 normally incident