Irradiation of optical components

LEDs & VCSELs

LEDs fron GEC and ABB HAFFO have been irradiates with 24 GeV/c protons and ~1 MeV neutrons at ISIS and at the Birmingham Dynamitron. Results on packaged GEC LEDs are avilable in INDET-NO-137. Results on bare GEC LEDs and bare VCSELs from SANDIA are available in INDET-NO-183 .

Results from the irradiation of two Truelight (Taiwan) VCSELs are available here.

A sample of four MITEL 1A444 ST packaged VCSELs were irradiated to a fluence of 3 10¹⁵n/cm² at ISIS. A further sample of 20 MITEL 1A444 ST packaged VCSELs have been irradiated at ISIS to a fluence of 4 10¹⁴ n/cm²and were measured at Wuppertal . Annealing studies were carried out and a report is available here. An accelerated aging study was performed by heating the VCSELs while operating them. A report on these measurements is available here. A very high fluence study will be performed with 20 MITEL 1A4444 ST packaged VCSELs at ISIS.

New Gamma Irradiation tests of Truelight VCSELs show good results up to 60 MRad. A short report is available here (pdf file).

PINs

Bulk Si and epitaxial Si PIN diodes have been irradiates with 24 GeV/c protons and ~1 MeV neutrons at ISIS and the Birmingham Dynamitron. The bulk Si PIN diodes died very quickly with radiation but the epitaxial PINs show excellent radiation hardness results. The plots from the analysis are available here.

The speed of response for an epitaxial Si PIN diode irradiated with neutrons and protons has been measured as a function of the PIN bias voltage. A fast response is obtained if the bias voltage is 5V or greater. Click here to see the plot.

The results are now available as an ATLAS note INDET-200.

Results from the irradiation of PIN diodes to the levels required by the Pixel detector are now available. The PINs show no further decrease in responsivity and no PIN has died as a result of the radiation. The responsivity plot is available here.

The results of an accelaerated ageing test for the irradiated and non-irradiated PIN diodes were very good. No failures of the PINs were oberserved during the quiavlent LHC operation period of 96 years. See the transparencies of G. Mahout form more details.

The speed of response for non-irradiated and irradiated epitaxial Si PIN diodes have been measured as a function of bias voltages. A very fast response can be obtained for a bias voltage greater than about 5V, even for devices irradiated to a fluence of 10¹⁵ n/cm². A short report is available here.

The radiation hardness of a sample of 3 Honeywell ST packaged epitaxial Si PIN dides has also been measured. They were measured before and after exposure to a neutron fluence of 4 10**14 n/cm**2 (at Ljubljana). The plots from the analysis are available here.

The radiation hardness of Truelight silicon PIN diodes has been measured. A short report is available Word doc or pdf

Single Event Upset (SEU) Studies

MIPs

Tests have been made to check for SEU caused by particles hitting the PIN diode (such an effect was reported by HERA-B and it caused loss of lock of the PLL) or DORIC. A 81 MBq Sr⁹⁰ source was placed at a distance of 11mm above the Si PIN diode in the GEC package. The TTC link was run and a BER measurement was performed while the source was in place. This source gives a particle flux of ~ 6 10⁶ cm^-2 s^-1,which is significantly greater than the expected flux at the inner radius of the SCT of ~2 10⁶ cm^-2 s^-1 . There was no significant source of errors and the BER was less than 2 10^-12 at 90% c.l. No problems with loss of lock with the DLL were observed in about 20 hours.

In order to simulate the higher fluxes expected in the Pixel detector, a 660 MBq Sr⁹⁰source was used and the distance between source and PIN was 1.1 cm. This gives a flux of 4.3x10⁷ cm^-2 s^-1 which is slightly greater than the expected fluence in the worst case in ATLAS of 3.5x10⁷ cm^-2 s^-1 in the Pixel B-layer. During an overnight run no errors occured. Another test over 64 hours without a source indicated an overall BER of 3 x10^-12. The errors occurred in bursts and were probably due to spikes on power lines. In conclusion there is no evidence for any SEU effects even at the highest flux of MIPs.

Heavy Particles

NPL tests with neutrons

Tests have been made to check for SEU effetcs with a neutron beam at the Birmingham Cyclotron. No significant problems were found with a neutron fluence equal to the particle flux expected in the innermost layer of the SCT. A short report is avaialable here. SEU tests of the links have been performed at rates up to and beyond those expected for the B Pixel layer using a 14 MeV neutron beam at the National Physical Laboratory in London. A significant BER was measured with the beam on. The error rate increases linearly with beam flux and decreases strongly as the optical power is increased. The summary plots are available here. The errors are due to reactions depositing sufficient charge in the active region of the PIN diode to create a signal above the DORIC threshold. The solution to the problem will involve using larger optical signals for the TTC for the B pixel layer.

PSI Tests with pions/protons

SEU rates were measured with pion/proton beams in the momentum range 300 MeV/c to 465 MeV/c at PSI. The results are published in NIM Preprint available here. The SEU induced BC counter errors were measured with optical links and a K5 hybrid at the T7 beam at the CERN PS. The results of this analysis and an analysis of the implications for ATLAS operation are available here.

Send comments about these pages to Tony Weidberg