1. Francisco Salesa Greus IFIC (CSIC–Universitat de València, Spain) Representing the KM3NeT Consortium 4 th International Workshop on Very Large Volume Neutrino Telescopes for the Mediterranean Sea (VLVnT09) Athens, Greece, 13-15 October 2009

2. The future KM3NeT detector. Time calibration requirements. Time calibration systems for KM3NeT: – Laboratory calibration. – Internal clock calibration. – Optical calibration system: The Nanobeacon. The Laser Beacon. – Cross-check methods: K40 and atmospheric muons. Summary. 10/15/20092F. Salesa - VLVNT09

3. KM3NeT will be a NT of at least 1 km 3 of sea water (Mediterranean Sea) and a deep-sea infrastructure for earth and sea science. Formed by 41 institutions from 10 countries. To be deployed after 2013. 10/15/2009F. Salesa - VLVNT093   c =43°

4. The angular resolution is critical in a NT. A good angular resolution provides the needed point spread function to resolve cosmic neutrino sources from the atmospheric background. The angular resolution relies on a good positioning and time calibration. Working within the specifications the attainable angular resolution of KM3NeT is better than 0.1° for E > 100 TeV. absolute time resolution The absolute time resolution (provides a specific time for each neutrino event w.r.t. UT) depends on: GPS timing. Detector electronic paths. In order to obtain correlations with the physics phenomena (e.g. GRB) an accuracy of 1 ms is enough. Relative time resolution Relative time resolution (among OMs) depends on: OM transit time spread (TTS), typically  ~1-1.5 ns. Optical properties of the sea water: light scattering + chromatic dispersion (  ~1-1.5 ns) for light coming from a distance of 50 m. Electronics (  <0.5 ns). All these components give an overall unavoidable time spread of 1-2 ns. The determination of the calibration constants with a  ≤1 ns fulfils the requirements for relative time resolution. 10/15/20094F. Salesa - VLVNT09

5. Before the deployment the time calibration constants are determined in the laboratory. The calibration consists of two main parts: one to obtain the clock-phase delay, another to obtain the intrinsic OM time offset: – The clock-phase delay is given very precisely by the internal clock calibration. – A special setup should be designed for the intrinsic time offset computation. The special setup can also be used for additional calibrations (e.g. electronics & charge). The internal clock calibration is repeated in situ. 10/15/20095F. Salesa - VLVNT09 Optical fibres Signal splitter Laser STARTSTOP GPS E/O/ E STARTSTOP On-shore Station Time Digital Converter In-situ Echo-based system Junction Box optical splitter Special calibration setup

6. The experience from the previous projects shows that a system of external light sources with a known emission time ensures the time calibration and provides measurements of the optical water properties (c.f. H. Yepes talk). Decoupling the intra/inter detection unit (DU) calibration seems the best solution: – The calibration in the same storey and in the same DU will be performed by a group of Nanobeacons. – The calibration among DU will be performed by several Laser Beacons. The calibration constants are obtained putting all the information together. 10/15/20096F. Salesa - VLVNT09 Laser LED

7. One Nanobeacon inside each OM (can be used for TTS monitoring). The OM housing the nanobeacon provides the reference time. Designed to illuminate storeys in the same DU (points upwards). Based on the ANTARES LED Beacon, but with improvements: – Less expensive. – High redundancy. – Avoid cumbersome synchronization process, only one (or two) LED. – Not triggered by the clock which can induce electronic noise. Laboratory tests performed: angular distribution, pulse shape, etc. 10/15/20097F. Salesa - VLVNT09 LED modelRise time (ns) (nm) FWHM (º)Intensity (pJ) CB112.547014150 CB302.24722890 NSPB520S3.247051170 AB872.447014130

8. The circuit works at a nominal 24 V. The flashing rate depends on the feeding voltage (no external trigger). It is 25 kHz at 24 V and requires only an on/off interface. One single LED is expected to reach 350 m (90 pJ ~2 x 10 8 photons). In ANTARES one LED reach ~200 m. First tests performed mounting the Nanobeacon in an OM from ANTARES in water. 10/15/2009F. Salesa - VLVNT098 RMS ~ 1 ns Trigger signal w.r.t. OM signal PRELIMINARY

9. In the peak region (± 10°) the KM3NeT LED emits 1.5 orders of magnitude more than the ANTARES LED. An opening angle of 15° is sufficient to illuminate OMs placed above, even in a perpendicular arrangement (NuONE DU). The pre-selected model is the Avago HLMP-CB30 LED. 10/15/2009F. Salesa - VLVNT099 50m 6m - KM3Net LED model (uncleaved) - ANTARES LED model (cleaved) - KM3Net LED model (uncleaved) - ANTARES LED model (cleaved)

10. Based on the ANTARES Laser Beacon (diode pumped Q-switched Nd-YAG laser). Some Laser Beacons will be deployed at the bottom of several DU. Alternatively to the previous green ANTARES laser ( = 532 nm). There is the possibility of working with a blue laser ( = 473 nm). The amount of light emitted can be tuned by means of a voltage controlled optical attenuator. An internal photodiode reads the signal and provides its timestamp. 10/15/2009F. Salesa - VLVNT0910 Laser Properties Blue (473 nm)ANTARES Green (532 nm) New Green (532 nm) Average power 20 mW~0.8 mW45 mW RatekHz range Pulse energy 5  J1  J45  J Rise time< 1.5 ns~ 0.6 ns~0.15 ns Liquid Crystal Retarder Polarizing Beam-Splitter Laser Head Polarizing cube beam-splitter Liquid Crystal Head Variable Voltage

11. 10/15/2009F. Salesa - VLVNT0911 Calibration constants correction RMS ~ 0.7 ns Time difference between a LED OB and an OM Electronics contribution less than 0.5 ns RMS ~ 2 ns RMS ~ 0.6 ns Only 15% larger than 1 ns Retuning of feeding HV can be corrected with the OBs 00 0 PRELIMINARY σ ~ 0.4 ns

12. Background signal can be used to cross-check the time calibration. For configurations with adjacent OMs in the same storey, the K40 present in the salt water can be used for charge and intra-storey time calibration of the detector. Atmospheric muons (both up-going and down-going) can be used for the calibration among and in the same DU. 10/15/2009F. Salesa - VLVNT0912 40 K 40 Ca  e - (  decay)  Cherenkov Gaussian peak on coincidence plot OM 0 OM 1 OM 2 Taking differences by pairs

13. The check of the time offsets measured by the K40 shows the improvement using the calibration constants computed by the OB system. The calibration constants computed with intense light sources (OBs) are still valid at lower intensities (K40). 10/15/2009F. Salesa - VLVNT0913 RMS=0.71 ns RMS=0.50 ns Laboratory Calibration constants OB calibration constants

14. The time calibration of KM3NeT based on the previous NT projects experience. An absolute time calibration of 1 ms is enough and a relative time calibration at the nanosecond level is desirable. A first calibration will be performed in the laboratories. An optical calibration system will be used for the in situ time calibration (results from ANTARES encourages that system). Decoupling of intra/inter DU calibration. Optical and muon background can be used as a cross-check. Thanks to the time calibration systems, KM3NeT will be able to achieve an angular resolution better than 0.1° for E > 100 TeV. 10/15/200914F. Salesa - VLVNT09

15. This system is used to calibrate the path travelled by the signal starting at the PMT photocathode up through read-out electronics. It can be achieved by using an LED pulser mounted close to the PMT and capable of illuminating the photocathode, or via an optical fibre illuminated with a laser or LED outside the OM. 10/15/200915F. Salesa - VLVNT09 Example of a similar system used in the ANTARES OMs. LED Blue LED

16. 10/15/2009F. Salesa - VLVNT0916 MAXIMUM INTENSITY 24

17. K40 10/15/2009F. Salesa - VLVNT0917 LED OB LED OB – K40 40 K 40 Ca e - (  decay) Cherenkov light  K40 The offsets calculated with the LED OBs are validated by the K40 (independent calibration procedure) The K40 test is done by pairs because we know neither the position of the source nor the time emission of the light. With the LED Beacon we can reproduce this test with more precision and with a more intense light source. ANTARES line

18. 10/15/2009F. Salesa - VLVNT0918 Without alignment With alignment The laser is fixed on the anchor. Therefore it can be used to check the line movements. If the line is considered rigid and straight, the time difference distribution has a RMS ~ 2.3 ns. Taking into account the shape of the line the values are distributed within RMS ~ 0.6 ns. σ=0.6 ns

19. 10/15/2009F. Salesa - VLVNT0919 NuONE DU MEDUSA DUSeaWiet DU