Internal
One of the issues to be resolved for the detectors to be used at the European XFEL is their radiation hardness. Although it is at this stage not entirely clear what experiments and measurements precisely will be performed, it is clear that radiation levels impinging on the detector can become severe. Therefore a dedicated radiation damage study campaign has been launched. This study looks at the radiation induced effects in Silicon diodes, as well as ASICs. The project, which is serving all detector developments, is lead and coordinated by the University of Hamburg, and uses a dedicated irradiation setup at the F4 beamline at DORIS, Desy.
A set-up to study the radiation hardness of silicon sensors for the XFEL
E. Fretwurst>sup>1, F. Januschek2, R. Klanner1, H. Perrey1
1Institute of Experimental Physics, Hamburg University, Luruper Chaussee 149, D-22761 Hamburg, Germany
2DESY, Notkestraße 85, D-22607 Hamburg, Germany
Imaging experiments at the XFEL pose unprecedented requirements to the detectors in terms of radiation tolerance: Photon fluxes up to 1016 photons/cm2 corresponding to approximately 109 Gy in silicon, are expected. An irradiation station has been set up in the DORIS beam line F3, silicon test structures have been irradiated, and first results on the radiation tolerance of Si-structures obtained.
Following effects of high dose irradiation of silicon sensors and electronics may occur: (1) bulk damage, (2) charge build-up at the Si-SiO2-interface and (3) current generation at the Si-SiO2-interface. Given that the photon energy of 12 keV is well below the threshold for bulk damage (~300 keV), no effects are expected for (1). (2) and (3) have been observed at lower doses (~100 kGy) with some evidence for saturation of both effects with increasing dose.
Figure 1: Irradiation set-up in F3 beam
The beam profiles, measured by setting the horizontal/vertical slit to 0.5 mm are: horizontal boxshaped with 5 mm width; vertical parabolic with a full width of ~4 mm. The photon energy spectrum has been calculated from the synchrotron spectrum and the absorbers (50 μm Al and 250 μm Be). The energy of the photons absorbed in the silicon peaks at 10 keV with a full width at half height of about 6 keV. Without chopper the dose rate is ~150 kGy/sec. With chopper it can be reduced ~0.5 kGy/sec. A pneumatic beam shutter allows a minimum irradiation time of ~1 sec.
To study the charge build-up and the current generation at the Si-SiO2 interface, gate controlled diodes as shown in Fig. 2 have been used [1]. They are fabricated on a 285 μm thick n-doped Si-substrate with a central p+- doped diode and five Al gate on top of about 350 nm of SiO2-Si3N4. Standard C-V measurements in the frequency range 10-800 kHz on gate rings 2 and 3 are used to determine the flat band voltage from which NOx, the charge density at the interface, is obtained. Fig. 3 shows NOx- versus irradiation dose obtained from the 10 kHz data. Starting at ~3•1011/cm2 for un-irradiated diodes, NOx increases to a maximum of ~4•1012/cm2 (flat-band voltage shift of about 75 V) and then decreases to about 2•1012/cm2. Note that the dose scale is logarithmic. Strong annealing is observed: the measurement results depend on the time between irradiation and measurement. The decrease above 5 MGy is probably due to an insufficient cooling of the substrate for irradiations without the chopper. It is obvious that the large increase of oxide charges has to be taken into account when designing sensors and read-out electronics. Nevertheless, it is comforting that NOx appears to saturate.
The surface current from the Si-SiO2-interface is determined from the change in current when driving the Si-SiO2-interface from accumulation to depletion. Whereas at low irradiations the I/V curves follow the classical gate controlled diode characteristics [1], they get more complex at higher doses. This still has to be understood. The results on the surface generation current density versus dose are shown in Fig 4 . From an initial current of 6•10-9 A/cm2 for un-irradiated detectors it rises to about 10-6 A/cm2 at a few MGy, and then decreases at higher doses. Again, we assume that annealing due to insufficient cooling for the irradiations without chopper is the cause. But there is clear evidence that the increase in surface generation current density is limited.
To summarize: A high dose (up to 150 kGy/s) irradiation stand has been set up in the F3 beam line at DORIS. Gated diode test structures have been irradiated up to doses of 1 GGy. A significant increase in charge density and in generation current at the Si-SiO2-interface has been observed, which appears to saturate at doses above a few MGy. Other groups have already used the stand.
We are very thankful to H. Graafsma for the interest in the work, to T. Wroblewski and A. Rothkirch for the strong support when setting up and performing the irradiations and to W. Gärtner for designing and building the irradiation set-up.
References
- A.S. Grove, Physics and Technology of Semiconductor Devices, John Wiley Sons, 1967;
- J. Wüstenfeld, Characterisation of the Ionisation-Induced Surface Effects for the Optimisation of Silicon Detectors for Particle Physics, Dissertation University of Dortmund, 2001.
