AMANDA: status report from the 1993-94 campaign and optical properties of the South Pole ice

~lIII ! ~'_'1:1~111.1[1~1

PROCEEDINGS SUPPLEMENTS EI~SEVIER

Nuclear Physics B (Proc. Suppl.) 38 (1995) 287-292

AMANDA: Status Report froill the 1993-94 Campaign and Optical Properties of the South Pole Ice Presented by A. Goobar on behalf of the AMANDA collaboration: P.Askebjer*, S.W.Barwickt, L.Bergstrgm*, A.Bouchta*, S.Carius *+, A.Coulthard~, K.Engel5, B.Erlandsson*, A.Goobar*, L.Gray§, A.Hallgren~, F.Halzen~, P.O.Hulth*, J.Jacobsen~, S.Johansson* ¶, V.Kandhadai§, I.Liubarsky ~, D.Lowderll, T.Millerll**, P.Mock t, R.Morseg, R.Porrata t, P.B.Pricell, A.Richardsll, H.Rubinstein ++, E.Schneider l, Q.Sun ~, S.Tilav~, C.Walck* &; G.Yodh t We report the first results of the AMANDA detector. During the antarctic summer 1993-94 four strings were deployed between 0.8 an 1 km depth, each equipped with 20 photomultiplier tubes (PMTs). A laser source was used to investigate the optical properties of the ice in situ. W e find that the ice is intrinsically extremely transparent. The measured absorption length is 59 4- 3 m, i.e. comparable with the quality of the ultra-pure water used in the IMB and Kamiokande proton-decay and neutrino experiments [1,2] and more than two times longer than the best value reported for laboratory ice [3]. Due to a residual density of air bubbles at these depths, the motion of photons in the medium is randomized. For spherical, smooth bubbles we find that, at 1 km depth, the average distance between collisions is about 25 cm. The measured inverse scattering length on bubbles decreases linearly with increasing depth in the volume of ice investigated.

1. I n t r o d u c t i o n

2. D e t e c t o r d e p l o y m e n t

The AMANDA (Antarctic Muon And Neutrino Detector Array) project was conceived to exploit polar ice as a transparent and sterile detection medium to attain large detection volumes for muons and neutrinos from astrophysical sources. PMTs deployed in the South Pole ice sense the Cherenkov light emitted by highly relativistic muons. Downgoing muons originate from cosmic-ray showers, and nearly isotropic muons result from interactions between neutrinos and nucleons in the ice and rock around the detector. The polar ice, unlike ocean water, is free of bioluminescent organisms and natural radioactive isotopes such as 4°K. In addition, the ice forms a rigid and totally adaptive support structure for the detector elements.

Contrary to the drilling experience gained from the 1991-92 deployment of 2 small strings, deploying A M A N D A strings using a new drill was surprisingly easy. The drill was designed to operate in cold South Pole ice unlike the 1991-92 drill which had operated in the Arctic where the ice is 20°C warmer. The winches are now computer assisted. The hot water drilling technique has been firmly established: • Four holes drilled to 1 km depth, 60 cm in diameter. • Water at 9O°C was pumped down the holes at ~ 160 liters/rain. • Drilling took four days/hole (12000 liters of fuel). • Deployment; one d a y / A b l A N D A string.

*Stockholm University, Sweden t University of California, lrvine, USA Uppsala University, Sweden § University of Wisconsin, Madison, USA ¶ Currently at Jb'nk@ing University, Sweden II University of California, Berkeley, USA **Currently at Bartol Research Institute, Delaware, USA 0920-5632/95/$09.50 © 1995 Elsevier Science B.V. All rights rcsclwed. SSDI 0920-5632(94)00757-8

Four strings with 20 optical modules (OMs) each were installed between 800 and 1000 meters. The OMs consist of EMI 9353/9351 PMTs, 20 cm in diameter, surrounded by pressure vessels manufactured by Benthos. The PMTs are

288

AMANDA collaboration~Nuclear Physics B (Proc. Suppl.) 38 (1995) 287-292

.:?f'

' ~

"' '" -" '-[

O m

N i k k o suplx)rt cables and carabiners Foam. p a d d i n g Electt-onic

II\ Breakout from ~t*,-rr~fin cable -#J~t~\ to t~eloen c o a x / ) ~ \ , _ . Optical fiber / / N~\ P M T base "~ ~ \'~ . with //~.,t "~k/" voltage

t-~edth,,,ugh-_~~ Benthos ~ hemisphe,~_ ~T i

s p a ~ e , ,'ing

,

:

"iL\~t !~

. "I~\ Thorn EMI 93:51JI/IT

•

Aluminiuffl ~___.lVlT b .:

divida

N ~ J [

~ 711

\*~--'~----~"

'J

1-.20,~m

Benthos " " ~ / ¢ / transparent \t / .~ | hemisphere - " ' ~ d ...1_ Nylon ~][/ sphere

Fignre 1. The AMANDA detector. The four strings were deployed between December 1993 and January 1994. There are 20 modules in each string, spaced by 10 meters. Each P M T is contained inside a pressure vessel. An optical fiber carries laser light from the ice surface to a diffusive nylon sphere suspended below each PMT. mainly sensitive to light coming from below, see figure 1. Three of the eighty OMs were damaged during deployment. Four other, although operating, are to a varying degree problematic. Most of these problems were associated with the first string. The deployment procedures were adjusted accordingly and, except for a single OM, strings 3 and 4 are fully functioning since J a n u a r y 1994. At the low temperature of the medium,-55°C, the dark noise rates observed average below 2 kHz per PMT.

3. P r o p e r t i e s o f p o l a r ice One of the characteristic features of shallow polar ice is the abundance of air bubbles [4]. These are formed as pockets of air get trapped between grains of snow [5]. Microscopic examinations of ice-core samples from Greenland and Antarctica [6,7] show that, at first, the size of air bubbles decreases with increasing depth (i.e. pressnre). When the hydrostatic pressure exceeds the formation pressure of air hydrates, a transi-

tion occurs where the ltumber of bubbles start to decrease. The hydrate crystals that form have a refractive index only 0.4 % larger than that of pure ice [8], making scattering on these crystals negligible. At the Russian Antarctic station of Vostok, where the temperature of the ice is similar to that of the South Pole ice (the accumulation rate of snow and the altitude are, however, different) bubbles were t, ot seen below 1280 m depth, setting an upper limit to the number density of air bubbles of 0.5 cm - a [7]. [n the ice cores extracted from the Antarctic Byrd station, bubble-free ice was found below 1100 meters depth [5]. To study the properties of the ice at the South Pole we used the calibration setup of the AMANDA detector. Along with the main signal cable carrying the high voltage and the P M T signals, optical fibers carry light from a laser at the surface to each optical module. Each optical fiber terminates in a diffusive nylon sphere placed about 20-30 cm from the PMT. By sending laser pulses to individual nylon spheres and measuring the photon travel-time distributions

AMANDA collaboration~Nuclear Physics B (Proc. Suppl.) 38 (1995) 287-292 between modules (time resolution ~5 ns), the optical properties of the nledium can be derived. The intensities of the laser pulses coining out of the spheres are high enough that photons are detected by PMTs in neighboring strings. The calibration pulses have a mean wavelength of 475 nm (~25 nm FWHM). The travel times of photons between the emitting nylon sphere and the receiving optical modules were found to be many times longer than what would be expected from a straight path. Figure 2 shows two examples of such measurements. The geometrical distance between the source and the detector is 21 and 32 meters in figures 2a and 2b, respectively. With no obstacles, the signals would have arrived after 91 ns and 142 ns, respectively, indicating that photons encounter many scattering centra between the source and the detector. If the light is detected at the P M T after a large number of scatterings the situation can be described as a random walk (diffusive) process. The arrival-time function for the radiative transport of a spherically symmetric light pulse emitted at a distance d = 0 at time t = 0 can then be described by [9,10]: [

,u(d,t)

-

(4zrDt)a/2

_d 2

e,~,

Ci~bub a ( 1 - T)

In figure 2 we show a comparison of the experimental data, the results of our Monte Carlo simulations, and the analytical formula of Eq. (1). The agreement between the Monte Carlo simulations, the data and the analytical fit is remarkable. We thus conclude that the physics behind the P M T calibration data is well understood, and that we have at our disposal a tool to measure properties of the ice such as Aa and Ably.

25O

~

200 15O IO0

58 0

50O

1000

750O (o)

2000 2500 orrivol time [ns]

m 250 m 1 200

--c~t

e ,~

,

(1)

where /~a is the absorption length, )kb,ub is the average distance that a photon travels between collisions with bubbles, ci = c / n is the velocity of light in the ice, n = 1.31, given in [3]. The constant of diffusion D is given by

D -

289

(2)

where 7- = is the average of the cosine of the scattering angle. The distributions of arrival times were also calculated in a Monte Carlo simulation. Thanks to the existence of widely separated length scales, d --, A~ > > > > AL, where tL is the wavelength of the laser light, a fairly simple optical treatment is adequate for this analysis. With our approximations we find for spherical, smooth bubbles that r in Eq.(2) takes a vMue around 0.75 (corresponding to an average scattering angle of 41°).

5OO

1000

15OO (b)

2000 a r r i v a l rJme I n s ]

Figure 2. Arrival time distributions for two different source-detector separations. The data points (dots) are compared to the Monte Carlo simulation (histogram) and the fits of Eq. (1) (solid line). The emitting and receiving PMTs are 21 and 32 ineters apart in figures (a) and (b) respectively.

A global X.2 minimization of the time data was pertbrmed where the relative vertical distances between the strings were left as free parameters, Aa was assumed to be a constant and 1/Ab,,b was expressed as a linear function of depth. For the present analysis we have selected time distributions where at least 1000 pulses were registered

290

AMANDA collaboration~Nuclear Physics B (Proc. Suppl.) 38 (1995) 287-292

by the receiving PMT. The time measured is set by the arrival of the first photon. In order to reproduce the true time distribution given by the different photon paths in the medium, we selected distributions where the probability for more than one photon detected at the receiving PMT for each trigger is less than 6 %. A total of 36 time distributions like the ones shown in figure 2 were used in the global fit. The optical properties of the medium are parameterized as described in Eq. (1). Only statistical uncertainties in the number of detected photons were considered. The relative vertical positions of the four strings were determined to about 0.2 meter precision. For the absorption length and the inverse scattering length on air inclusions, the global fit yields: 3,~ = 59 + 1 Ii1, and

1 Ab~,b --

(28.6-b 1.1) -- (0.025::~: 0.001)Z 4(1--r)

-1 m

,

(3)

where z is the vertical coordinate, defined to be zero at the surface and increasing downwards. We have used Monte Carlo simulations to estimate the systematic uncertainties in the global fit from the contamination of nmltiple-photon events, background events, and limited sample sizes. We find that the various effects tend to cancel each other. A conservative estimate of the total uncertainty in ~ is: A~ = 59 + l(stat) 4- a(syst) meters. Figure 3(a) shows the observed linear dependence of 1 on depth. This depth dependence is steeper than would be expected if the bubble number density were constant and only the bubble sizes decreased under the hydrostatic pressure. The shaded area in Fig. 3(b) corresponds to the uncertainty in the bubble shapes, i.e. in the average scattering angle. Also shown in the figure are results from the measurements on the Vostok and Byrd ice cores.

4. S P A S E - A M A N D A coincident triggers Coincidences with the SPASE 1 surface air shower array have been observed. These events, at a rate of more than 100 per day for a 10fold trigger, are now used as a calibration tool. A sample of coincident muon events is shown in Fig.4. The background is consistent with random cosmic ray muons triggering the detector within the SPASE-AMANDA coincidence time window. The 4-string AMANDA detector is the largest ninon detector ever built. The muon threshold energy is ~200 GeV. The data can be used to study the response of ice to a triggered muon beam. As the direction of the muon is known the data can be exploited to finetune the procedures we have developed for reconstructing muon direction in the presence of residual bubbles. Muon bundles contained in SPASE showers are observed. Although A M A N D A does not directly measure muon-energy, the coincidence rate can be studied as a flmction of trigger multiplicity. The data are being studied with the goal to make a qualitatively improved determination of the elemental cosmic ray composition in the region of the "knee" in the primary spectrum. Heavy nuclei of high energy prodnce more muons than protons do, hence the sensitivity to composition of the primary cosmic ray beam. It is well-known that this represents one of the most flmdamental measurements in cosmic ray astrophysics. 5. C o n c l u s i o n s We have successfldly instrumented four AMANDA strings in the South Pole ice. After six months of operation, all the 73 modules that were not damaged during deployment are still flmctioning well and are being used to analyze muon events detected in coincidence with the SPASE array. The examined ice volunm, 0.8-1 km below surface, has an extremely long absorption length, suggesting that the ice cap is indeed an ideal medium for a neutrino telescope. If the absorption length does not deteriorate with depth, the volume of a future muon and neutrino det Bartol-Leeds

291

AMANDA collaboration~Nuclear Physics B (Proc. Suppl.) 38 (1995) 28~292

10

. . . .

I

.

.

.

.

I

. . . .

I

.

.

.

.

I

'

100

'

9

T ,<~ 80

5 5 2 1 0800

,

,

,

t 8501 ,

,

,

, 900J '

'

'

, 9/9

,

,

,

, 10100 ,

L

Depth z (rr)

(o) 14 12

~%-..

Byrd ~ce core d a t a .........Vostok ice core d o t o .....

*'., " ' - - ,

"E 10 8

4o

,.,. ::::::::::::::::::::::: ...... =================================== ......~ . ~ . ~

a

k};:::::::::::: .h;;~~4~.::::: :::: :::

2 o

600

800

1000

(b)

1200

0800

Depth z (m)

850

900

950

1000

Depth 7 (m)

Figure 3. Left: Inverse scattering length X%ZT. 1 baS a function of depth, z. The data points in (a) are AMANDA measurements shown together with the best fitted linear dependence of Eq. (3) for r = 0.75 (i.e. spherical bubbles). The fitted line extrapolates to zero at ~1150 meters. The shaded area in (b) shows the AMANDA results allowing for r = 0 0.75 (r = 0 corresponds to isotropic scattering angle). Also shown are the Byrd and Vostok ice core measurements [5,7] where we have computed an equivalent value of ka, b from their quoted bubble number density (A[) and average bubble radius squared (Of_,), 1 J~71"~O2. Right: Measured absorption length, A~, between 0.8 and 1 km depth. The solid line shows ~b,tb -the best fitted value. teetor to be deployed in the bubble-free region can be made significantly larger than previously anticipated since the PMTs can be spaced fl, rther apart. A linear extrapolation of our data would yield that bubbles vanish at ~1150 meters at the South Pole. The data from Vostok and Byrd (Fig 3(b)) show, however, that the rate of bubble disappearance becomes somewhat slower towards the end of the transition process. During the antarctic snmmer 1995-96 new strings will be deployed at depths reaching down to 2 km below the surface.

REFERENCES

1. Becker-Szendy, R. et al. Nuclear Instruments and Methods (1994) m press. 2. Indue, K. "Neutrinos and Anti-Neutrinos from tile Sun", Ph.D. Thesis, Institute of Cosmic Ray Research, University of Tokyo (1993). 3. Grenfell, T. C. & Perovieh, D. K. J. Geophys. Res. 86, 7447 (1981); Warren, S. G. Applied Optics 23, 1206 (1984). 4. Shoji, H. & Langway Jr, Ch.C. Nature 298, 548 (1982), and references therein. 5. Gow, A. J. & Williamson, T.Jour. of Geophys. Res. 80, 5101 (1975). 6. Shoji, tI. & Langway Jr, Ch. C. Jour. de Phys. Supp 3 . 4 8 , CI-551 (1987). 7. Barkov N. [. &; Lipenkov V. Ya. Materialy Glyasiologicheskikh Issledovanii 51, 178 (1985).

292

AMANDA collaboration~Nuclear Physics B (Proc. Suppl.) 38 (1995) 28~292

I

T

-

r

.

i

'

I

'

40

:

J

~7-

~

!

~

'

'

a

•

,

r

;

~

,

,

'

'

I

I 3oo

T

J

r

~

I ~5o

'

.~

I o

r

.

,

I so

r

.

~

I ~ 1 oo

~

.

I 1 so

~

, p~lr

,

I 200

,

,

,

,

,

,

,

2so

,

J ~

20

10

0

10

20

thela

30 dmtnbulkon

40

50

60

I

'°

t 0

i

i

i

l'ltL 50

|l' 1 oo

15o pill

. n*__~ i , , i , , 200 Olstltbutlo n

......... i i i 2so

i

i

!

i

. • ,',,Ljm i i i ~

i . . . . i , ,

i

,

Figure 4. Sample of AMANDA muon triggers (10 or more) coincident with SPASE showers. The data is plotted in terms of the 2 polar angles defining the direction of the air showers as determined by SPASE. As viewed fi'om tire air shower array the AMANDA detector is positioned at zenith angles 0 _~ 45 ° and ~ 1200 .

8.

Hondoh, T. et al. J. Inclusion Phenomena. 8, 17 (1990); Uchida, T. private communication. 9. Chandrasekhar, S. Rev. Mod. Phys. 15, 1 (1943). 10. Ishimaru, A. J.Opt.Soc.Am 68,8 (1978); Patterson, M. S., Chance, B. & Wilson, B.C. Applied Optics 28, 12 (1989). ACKNOWLEDGMENTS: We thank B. Koci and the entire PICO organization for their excellent support with the ice drilling. This work was supported by the National Science Foundation, USA, the K.A. Wallenberg Foundation, Sweden, the Swedish Natural Science Researcti Council tile G. Gustafsson Foundation, Sweden, Swedish Polar Research, Sweden, the Graduate School of the University of Wisconsin, Madison. USA, and the University of California at Irvine,'USA.