R&D programme

(Some pictures of the boats used for test and deployment)

In order to ensure the success of the deployment of a large-scale detector in an uncontrollable environment such as the deep sea, it is necessary to perform an extensive programme of site evaluation and prototype testing.

The selection of a suitable site for a neutrino telescope requires consideration of water transparency, optical background, fouling of optical surfaces, strength of the deep sea currents, meteorological conditions, depth, on-shore support, infrastructure and pier availability.

Site evaluation mooring lines

A detailed programme of in situ measurements has been undertaken since October 1996, most of the data being taken at a site near Toulon (42 $^{\circ}$ 50' N, 6 $^{\circ}$ 10' E) at a depth of 2400 m (See global and detailed site maps on the right). Optical background data were also taken 20 nautical miles off Porto, Corsica (42 $^{\circ}$ 22' N, 8 $^{\circ}$ 15' E) at a depth of 2700 m.

Global Site Map : From Marseilles to Fréjus.Excerpt from map "N0 7204 P" Reproduced with authorisation N0 892/97 from " Service Hydrographique et Océanographique de la marine (S.H.O.M) - Brest (France) "

Autonomous mooring lines to measure the site parameters have been developed and more than 20 deployments and recoveries have been successfully performed.

Three generic test lines exist, as illustrated below. Each test setup is incorporated in a mooring line anchored at the sea bed and vertically supported by a buoy. The measuring system, which can be up to 50 m long, is placed around 100 m above the sea bed. A further 30 m of cable and instrumentation sits between the measuring system and the top of the test line. The electronics, data acquisition and the detectors are powered by a set of lithium battery packs. A 500 cm³ pack delivers 8 A-h at 26 V. These compact power sources have a lifetime which permits stand-alone tests for periods as long as one year.

The test lines have been deployed using a standard horizontal deployment method. The buoy at the top of the line is immersed first, so that the line is held under tension during the entire operation. The boat continues to move slowly away from the buoy as the rest of the line is paid out from the boat. The anchor is the last element to be deployed after which the line falls to the sea bed.

To date, the measurements carried out confirm that the properties of this site satisfy the constraints of the ANTARES physics programme.


Test line to study light attenuation	Test line to study optical background	Test line to study optical fouling.

It is planned to explore a wider area in the Mediterranean sea, for long periods, in order to study the variation of parameters as a function of site, depth and season. For specific parameters such as water transparency, measurements at several wavelengths are also foreseen. The results obtained so far have been instrumental in steering the design of the telescope.

Optical properties of the site

Optical background

Some pictures of light emitters in the deep ocean

The behaviour of the optical background on site places constraints on the trigger logic and the electronics as well as the mechanical layout of the optical modules. A setup has been devised for studying the time dependence of the background as well as its spatial extent and its correlation with deep sea current. The relevant test line has been immersed ten times in total, for periods spanning from hours to months.In this case 8-inch photomultiplier tubes were used.

Up to three optical modules have been used on the same line, two of which, A and B, were 0.5 m to 1.5 m apart, while the third, C, was 10 m to 40 m away. A current-meter was installed below the optical modules. Data consisted of measurements of singles rates for all three optical modules and coincidence rates for modules A and B within a time window of 100 ns. In order to sample long-term variations, the system was enabled for a few hours three times a week.

Time dependence of the counting rate.

An example of the observed counting rate is presented in the figure above which exhibits two distinct components: a low level background around 40 kHz, and, superimposed on this, rapid ( $\sim$ 1 s) excursions of up to several MHz. The low level background varies from 17 kHz to 47 kHz over a time scale of a few hours. This rate changes simultaneously on all optical modules even when they are 40 m apart. The peak activity is correlated with the current speed and is limited in spatial extent: peaks are seen simultaneously by optical modules when they are less than 1.5 m apart, but not when they are more than 20 m apart.

It was also shown that the bioluminescence activity is correlated with the current velocity.

Optical fouling

When exposed to sea water, the surfaces of optical modules are fouled by the combination of two processes: living organisms, mostly bacteria, grow on the outer surface, and sediments fall on the upward-looking surfaces. While the bacterial growth is expected to be almost transparent, sediments will adhere to it and make it gradually opaque, thus diminishing the sensitivity of the detector. This phenomenon is expected to be site-dependent as the bacterial growth decreases with depth and the sedimentation rate depends on local sources of sediments such as nearby rivers. A series of measurements has been performed in order to quantify these phenomena.

Light of wavelength 470 nm from a blue LED source in a glass sphere was normally incident on a set of five PIN diodes placed at 50 $^{\circ}$ to 90 $^{\circ}$ from the vertical axis of the sphere inside a second sphere 1 m away. Measurements of the light transmission and current velocity were made twice daily.

Light transmission as a function of time and polar angle. Figure on the right shows the light transmission monitored for 240 days with this configuration. For the horizontally-looking PIN diode (90 $^{\circ}$ ), the light source is affected by the same fouling as the detector, thus doubling the effect. In this case, a transmission loss of 1.2% per surface is observed after 8 months of exposure. All optical modules in the ANTARES detector will point downward, reducing the loss even further.

In addition, a cradle holding glass slides mounted on a horizontal cylinder with various orientations around the axis of the cylinder was incorporated into the mooring line. This apparatus was developed by IFREMER.

Two series of measurements have been performed with the apparatus being immersed for long periods (3 months and 8 months respectively) at 2400 m depth on the same site as the other tests. After recovery, a biochemical analysis of the slides has been performed by several laboratories demonstrating the dependence of the biofouling on the orientation of the slides. In those cases, where the slides point downward, saturation is observed. According to marine biologists, these numbers are several orders of magnitude below what is observed at shallow depths.

The optical modules are housed in pressure-resistant glass spheres produced by the Benthos company. The mooring line which measured the fouling of Benthos spheres also included a sediment trap in order to determine the vertical flux and the composition of sediments at the ANTARES site. Particulate matter contributes to the scattering of light in sea water as well as the fouling of the Benthos spheres. The sediment trap collected samples from July to December 1997 on a weekly basis. The analysis of the samples was performed in the CEFREM (Centre de Formation et de Recherche sur l'Environnement Marin) laboratory by the team of Professor A. Monaco. The total mass flux of sediments substantially increases from October onwards, probably as a result of heavy rainfall draining sediment from the shore. Indeed, the composition of the sediments shows a large contribution of material originating from continental river bed. This study will be complemented by the analysis of the sediment cores and of the water samples collected during the ANTARES site survey campaign of December 1998.

Transmission properties of the water

The water transparency affects the muon detection efficiency, while the amount of scattered light determines the limit on the angular resolution of the detector. These two parameters therefore influence the detector design and are necessary data for the Monte Carlo simulations from which the detector response is calculated. They must be measured in situ, as water samples may be degraded when brought to the surface. Two different experimental setups have been constructed to measure these parameters.

In December 1997, measurements were performed with a 33 m long rigid structure holding a collimated and continuous LED source located at a variable distance from an optical module. For each selected distance D between the source and the detector, the LED luminosity $\Phi_{\rm LED}$ was adjusted so as to yield a constant current $I_{\rm PMT}$ on the photomultiplier tube. The set-up was calibrated with a similar experiment done in air. The emitted and detected intensities in water being related by

$\begin{displaymath}\nonumber I_{\rm PMT} \propto \Phi_{\rm LED} /D^2 \times \exp(-D/\lambda_{\rm att.\,eff}) \end{displaymath}$

this test makes it possible to estimate the effective attenuation length from the dependence of the required LED intensity with the distance. The agreement of the data with a decrease following the formula given above was excellent and yielded an effective attenuation length of

$\begin{displaymath}\nonumber \lambda_{\rm att.\,eff} = 41\pm 1\; ({\rm stat.}) \pm 1\; ({\rm syst.}) {\rm\: m\;\;\;(December\; 1997)} \end{displaymath}$

This attenuation length results from a combination of absorption and scattering. The experimental set-up was unable to separate these, and the long rail made deployment difficult. The experiment was therefore redesigned so that it used a pulsed source, to facilitate scattering measurements, and a flexible structure to improve ease of deployment.

Distribution of arrival times of photons for two distances between the detector and the source. The 24 m distribution is normalized to 1, and the 44 m distribution is normalized relative to the first one. Arrival time distributions for 24 m, 44 m, and a calibration in air, with Monte Carlo curves superimposed.

In July 1998 and March 1999, measurements were performed with a set-up consisting of a pulsed isotropic LED source located at a distance of either 24 m or 44 m from a 1'' fast photomultiplier tube. An 8-bit TDC measured the distribution of the arrival times of the photons. The overall time resolution was $\sigma=4.5$ ns. Because of this, photons in the tail of the distribution have scattered with an angle at least $\sim 35^\circ$ for the 24 m spectrum, and $\sim 25^\circ$ for 45 m. Therefore, the scattering properties of the water are being measured for large scattering angles.

The time distributions recorded exhibit a peak stemming from direct photons, and a tail extending to larger delays due to scattered photons. For the 24 m (44 m) spectrum, 95% (90%) of the photons are collected within 10 ns. Scattering is thus a small effect at the ANTARES site.

An effective attenuation length could be determined from the ratio of the integrated spectra measured at the two distances, yielding:

$\begin{displaymath}\nonumber \lambda_{\rm att,\,eff} = \left\vert \begin{array}{... ...rm stat.})\:{\rm m} & ({\rm March} \;1999) \end{array} \right. \end{displaymath}$

A systematic uncertainty of a few metres might affect these estimates due to the fact that the LED luminosity is not monitored and yet assumed to be the same for the time distributions collected at the two distances. These measurements indicate more significant attenuation in March than in July. The difference between these and the December 1997 measurement is also significant, although it is partly accounted for by the use of a collimated source in 1997. A more detailed analysis can be done by fitting the data with a Monte Carlo distribution obtained by photon tracking. The data are well described using an absorption length in the range 55-65 m, a scattering length at large angles greater than 200 m and a roughly isotropic scattering angle distribution. This is consistent with the effective attenuation length deduced from the ratio of integrated spectra.

Sea conditions

Suitable sea conditions for periods of up to a few consecutive days are required to perform deployment and recovery operations. These conditions depend both on the nature of the operations and on the characteristics of the ship. For the single string deployment and recovery operations of June-September 1998 with the Castor, a wave height less than 1.5 m and wind speed less than 25 knots (5 on the Beaufort scale) were specified.

A study has been made incorporating data from a number of sources, namely:

data on wave height collected by an instrumented buoy moored 4 nautical miles south of Porquerolles Island from May 1992 to September 1995, which should be representative of the conditions in the ANTARES site;
data on the wind speed and direction as recorded by the Porquerolles Island Signal Station analysed for the period from January 1992 to March 1996;
additional information on sea conditions provided by satellite measurements in the area from 1992 onwards.

Preliminary analysis of these data has been performed by the Meteomer company. Periods of three consecutive days with favourable sea conditions occur less than five times per month from October to April, and more than five times per month from May to September.

Site survey

The strength and direction of the underwater currents need to be taken into account in the mechanical design of the detector. All the measurements of the deep sea current gathered during the test immersions have shown a maximum current of 18 cm/s. This value is accommodated in the mechanical design of the strings.

A visual and bathymetric survey of the sea floor was performed in December 1998, using the Nautile submarine. In the area selected as a potential ANTARES site, the sea floor is flat and displays no topographical anomalies such as steps or rocks, as can be seen in figure below. During the same series of dives, core samples from the sea floor were collected. They consist of solid mud which is a satisfactory substrate to support the detector.

Site survey map