Abstract
The ANTARES collabo ration is deploying a neutrino telescope under the
Mediterranean Sea [2]. A 3D array of photomultipliers (PMTs) will detect the
Cherenkov light emitted by the muons produced in neutrino interactions. Since
the reconstruction effciency and pointing accuracy of the detector will depend
on the correct de termination of the arrival times of the Cherenkov light to the
PMTs, the time calibration of the detector is a key issue. Intrinsic time
fluctua-
tions in the detector comp onents indicate that an accuracy of σ ~ 0.5 ns in the
relative time calibration among PMTs is an adequate goal. On the other hand,
physics considerations indicate that an absolute time calibration of σ ~ 1ms is
enough for all practical purposes. In this presentation the different sources of
time uncertainties in the detector are briefly reviewed. The methods to reach
the
desired level of accuracy in the time calibration are shortly described
emphasising
the system based on optical beacons, i.e. well-controlled pulsed light sources.
1. Introduction
The ANTARES detector will consist of 12 strings, with 25 storeys each. A
storey consists of 3 optical modules(OMs) looking downward at 45°and separated
120°horizontally from each other. An optical module [1] is basically a
pressure-
resistant sphere housing a 10” PMT, its base and a pulsed LED for calibration
purposes. An angular re solution for the muon tracks of 0.2? is expected to be
reached at high energy, provided the relative timing resolution (RTR) between
OMs is achieved at the required level of precision. The RTR is limited by the
transit time spread (TTS) of the signal in the PMTs , σ ~ 1.3 ns, and by the
fluctuations due to the transmission phenomena of light in seawater, σ ~ 1.5 ns.
Therefore, a calibration aiming at a RTR of σ ~ 0.5 ns is suffcient (one standard
deviation will be understood in the fol lowing when referring to time jitters,
unless
stated otherwise).
Apart from the relative timing among OMs, an absolute time calibration,
i.e. the appropiate correlation of the detector’s clock system with Universal
Time,
will be required. An absolute time accuracy of~1ms is enough for any conceivable
physics goal (e.g. correlation with gamma ray bursts or supernova events).
2. Sources of time delays and uncertainties
The delay between the impingement of light on the PMT photocathode and
the arrival of the signal to the time to voltage converter(TVC) of the readout chip
(the Ana logue Ring Sampler, ARS) is ~100 ns. This delay is dominated by the
transit time (TT) in the PMT with additional contributions from the cables that
link the PMT base to the LCM (the local control module where the electronics of
each storey is located) and from the delay in the ARS. Laboratory measurements
indicate that variations of this delay with respect to the values calibrated in
the
laboratory and time drifts once the detector is in the sea will be small
(estimated
to be <5 ns and <2 ns, respectively). The output signal from the TVC is
digitised
relative to the clock reset time stamp issued by the clock card in the LCM of
each
storey . The delay between the TVC output and the localclock card is of the order
of 10 ns and stable (<0.2 ns). An unavoidable event to event fluctuation at this
level is the spread in the transit time (TTS) of the signal in the PMT, ~1.3 ns.
The TTS itself has been checked to be stable, without no sign of degradation due
to the ageing of the PMT and with a small dependence on the photocathode to
dynode voltage (~2 ps/V).
The clock signal on-shore is distributed at the seabed to the different
strings by a passive splitter located at a junction box, from which all the link
cables to the strings depart and where the electro-optical cable to the shore ar-
rives. The delay between this splitter and the clock cards in each LCM will be
2to4 μs, depending on the position of the corresponding storey. The lab-to-sea
difference in this delay is expected to be ~1 ns and variations with time smaller
than 0.5 ns. The jitter is better than 100 ps as determined in the laboratory.
The
delay between the splitter and the on-shore Master clock signal is expected to
be
~200 μs, mainly due to the fibre optics responsible of the signal transmission.
The stability of the signal in the sea cable should be better than <0.2 ns. The
land cable, though, may induce variations of up to ~1 ns due to temperature
changes . Nevertheless, this variations will only affect the absolute time calibra-
tion relative to the GPS standard time and therefore are negligible with respect
to the ~1 ms goal established on the basis of physics requirements. A summary of
the expected time delays and jitters is given in table 1. Finally, the
uncertainty in
the propagation time of the Cherenkov photons from the muon track to the PMT
due to scattering and to dispersion –through the wavelength dependence of the
group velocity of light in water– gives an estimated σ ~ 1.5 ns at 50 m (a
distance
larger than the most probable value at any energy, according to simulations).
Table 1. Summary of ANTARES time offsets and uncertainties.
| |
PMT cathode →
ARS |
ARS ←
LCM clock |
LCM clock →
Junction Box |
Junction Box →
Master Clock |
Delay
Stability
Jitter |
~ 100 ns
~ 2ns
1.3 ns |
~ 10 ns
<0.2 ns
<0.5 ns |
2–4 μs
<0.5 ns
<0.1 ns |
200 μs
~ 2ns
<0.1 ns |
3. Time calibration systems
A relative time calibration of a string is first performed in a dedicated dark
room using: 1) the clock calibration system; 2) the LEDs located in the OMs and
3) a laser-fibre system [3]. Once in the sea, four different but complementary
systems will be used. The first two are the same as before. In addition, a
series of
optical beacons located throughout the strings provide a series of well
controlled
pulsed light sources that will allow a relative time calibration including all
the
relevant effects. Finally, the several thousand downward going muon tracks that
are expected to be reconstructed per day in the detector will allow an
additional
time alignment.
Bi-directional components in the clock system will allow an integrated
echo-based time calibration. The clock stability is σ<0.05 ns. The precision of
the clock system relative to the GPS standard time has been measured in tests
to be σ ~ 1.3 ns. This system can be used to calibrate from the master clock
up to the clock cards in each LCM storey. The remaining components (notably
the ARS, the PMT and the relevant transmission cables) should be calibrated
otherwise. In each OM, a LED driven by a pulser circuit based on the design
given in [4] can be used for calibration. The LEDs emit light at 470 nm with a
FWHM of 15 nm.
4. The Optical Beacons
The optical beacons will allow relative time calibration by means of in-
dependent pulsed light sources. Two kinds of beacons will be distributed within
the detector: the LED beacons and the Laser beacons. The former will be placed
every 7 storeys to a total of 4 per string and located in the upper part of the
frame
that supports the OMs. The latter will be placed at the bottom of the string.
The LED beacon (fig.1., left) consists of six faces fixed to an hexagonal
cylinder mounted inside a cylindrical glass container. Each face has six LEDs,
five looking horizontally outward and the sixth facing upward. Each LED beacon
can produce between 3 × 107 and 3 × 109 photons per pulse depending on the
configuration of LEDs selected and theamplitude of the driving voltage. The light
Fig. 1. Left: internal mounting of an LED beacon holding the 6 LED faces.
Right:
blow-up of a laser beacon. The internal electronics, the laser and the external
titanium container with the quartz rod on its upper surface can be seen.
pulse of the LED beacon has a rise time of 2 ns and a width between 4.5 ns
and
6.5 ns depending on amplitude. A small fast photo multiplier within each beacon
allows a precise determination of the exact time of the light flash. The laser
beacon (fig.1., right) is a much more powerful device that uses a diode pumped
Q-switched Nd-YAG laser to produce pulses of ~1μJ with a FWHM of ~0.8 ns
at a wavelength of 532 nm. The laser is housed in a cylindrical titanum
container
and points upward. The beam is widened by a diffuser which spreads the light out
to a cosine distribution. A quartz cylinder is bonded to the upper surface of
the
diffuser, so that the light exits through the vertical walls of the cylinder,
where
sedimentation is negligible. A built-in fast photodiode gives the actual time of
emission of the flash. Detailed Monte Carlo studies indicate that the system of
optical beacons will be able to calibrate with an accuracy of <0.5 ns the
relative
time between all the PMTs of the detector.
5. Conclusions
The systems described to calibrate in time the ANTARES detector will
allow a relative timing precision between OMs of the order of ~0.5 ns, while
absolute time will be determined with an accuracy better than 1 ms.
