N.J.C.SPOONER, J.W.ROBERTS, D.R.TOVEY Department of Physics, University of Sheffield, Hicks Building, Hounsfield Road, Sheffield S3 7RH

The anisotropic scintillation response to the direction of nuclear recoils in trans-stilbene makes it possible to propose a dark matter detector sensitive to the predicted forward-back asymmetry in the flux of weakly interacting massive particles (WIMPs) that arises from the motion of the Earth through the Galactic dark matter halo. As a first step to assessing this possibility it is necessary to know in detail the form of the anisotropy for the carbon recoils at low energy and the form of the carbon recoil scintillation efficiency (carbon being the target in this case). We present here first measurements of these quantities below 100 keV using neutrons from a monoenergetic neutron beam.

1 Introduction.

First limits on Weakly Interacting Massive Particles have been set using low background ionization (Ge and Si) or scintillation detectors (NaI or CaF2) by estimating what maximum possible nuclear recoil energy spectrum from WIMPs of a given mass could be compatible with the residual gamma (beta) background in the detectors (see this proceedings for references and latest results). More advanced detectors in which the keV-range nuclear recoils of interest can be distinguished from the gamma background are now being planned by several groups using, for instance, simultaneous detection of ionization and phonons in a semiconductor [ 1], or have recently started giving limits, for instance using pulse shape discrimination in NaI [ 2]. These detectors hold out the prospect of lowering dark matter limits, possibly into the range where a positive signal may be expected. The need for discrimination techniques is clear since intrinsic detector backgrounds will most likely never be reduced sufficiently (mainly due to cosmogenic activity). However, the question remains, even for detectors with recoil discrimination, how to be sure that a perceived excess of events above a detector's residual background, if such occurs, is due to a positive signal from dark matter or due to a systematic error? One possibility is to make use of the motion of the Earth through the dark matter. This is expected to produce a forward/back asymmetry in the scattering events in the laboratory frame so that measurement of the direction of nuclear recoils would provide a potentially powerful means of identifying real dark matter events. It would also provide an additional means of discriminating against background electron recoils, since these occur isotropically. Certain organic scintillators are known to show anisotropic scintillation light output and pulse shape properties correlated to the direction of nuclear recoils occurring within them. The effect is likely due to directional variation in the quenching mechanism assumed responsible for the non-linear energy response of organic scintillators to heavy ions. In particular anthracene and stilbene have been previously investigated for this effect [3] and a study of anthracene as a directional dark matter detector has been reported by other investigators [4]. Other organic crystals, including carbyzol, naphthalene and p-terphenyl are also known to show anisotropic behaviour [3]. In this work we concentrate on stilbene, chosen here because it has low toxicity and is available in single crystals >100s gram at low cost. It is possible in principle to assess the sensitivity of a stilbene (or other organic crystal) detector to the direction of WIMP recoils by calculating the dark matter limit that would result if no scintillation anisotropy were observed in such a detector. An intrinsic gamma background comparable, for instance, to existing scintillator dark matter detectors needs to be assumed for this. More importantly the calculation requires a value for the conversion efficiency ec of carbon recoil energy into scintillation photons vs. energy and a measurement of the form of the anisotropic effect, i.e. the dependence of ec on the recoil direction so that this response can be folded with the recoil directions. [The later can be calculated from the angular dependence of the differential nuclear recoil spectrum, dR/dER, for WIMP elastic scattering with respect to laboratory recoil angle, y. Where dR/dER can be approximated by:

In these equations c1 and c2 are fitting parameters [6], dR/dER is the differential event rate (keV-1kg-1d-1 or Òdru"), Eo, = 1/2 MDVo2 for WIMP mass MD (GeV c-2) with Vo = 230 kms-1. VE is the target (Earth) velocity and R0 (kg-1d-1 or "tru''). the event rate for VE = 0. r = 4MTMD/(MT+MD)2 where MT = 0.932 A is the mass of the target nucleus in GeV c-2]. From equ. 2. above it is found that the ratio of forward/back scattering becomes >100:1 for sufficiently high WIMP mass [7]. It is this large forward/backward asymmetry that potentially can provide a powerful means of discrimination. Other investigators have studied H (proton) recoils in stilbene [8]. However, in this case we are more interested in the possibility of utilizing the carbon recoil, which would be a better target for coherent interactions. To our knowledge carbon recoils in stilbene at low energy (<100 keV) have not been studied before.

2 Experimental Technique

Measurements were performed using a 25 mm diameter x 25 mm trans-stilbene crystal manufactured by Amcrys-H. This crystal was wrapped in PTFE reflective tape and mounted directly onto a 2 inch Electron Tubes Ltd. PMT type 9266A. This apparatus was mounted in the line of a collimated 5.5 MeV monoenergetic neutron beam at the Birmingham University radiation centre using an arrangement previously described [9]. NE213 neutron counters, set at various angles to the crystal at 1 m distance, were used to provide a coincidence signal thereby defining the energy of carbon recoil events in the stilbene. Electronics as described in [10] were used to allow pulse shape information to be recorded as well as time of flight and coincidence information. A gimbals arrangement was used to allow the crystal to be rotated in the beam about the cylindrical axis and perpendicular to that axis. The characteristics of the direction anisotropy in organic scintillators takes two forms, a dependence of the pulse height on the direction of the recoil track and a related dependence of the pulse shape. In this initial work we considered only the former, defined as AL = 2(Lmax - Lmin)/(Lmax + Lmin) where the direction for Lmax and Lmin are assumed independent of energy. In the previous measurements of H recoils [8] the recoil directions corresponding to Lmin and Lmax were related to the crystal axes using x-ray analysis, the crystal axes being designated by directions a,b, and c where a is perpendicular to b and b is perpendicular to c but lies at an angle of 124o. For stilbene the maximum light output was found to occur for recoil tracks parallel to the b axis while the minimum occurs in a direction perpendicular to the ab plane defined by an artificial axis c'. In the present tests it was not possible to obtain accurate crystal orientation due to the large size of the crystal. Therefore all angles quoted are relative to each other measured as the angle between the recoil direction and the axis of rotation (either the cylindrical axis or that perpendicular to it).

3 Experimental Results

Fig. 1 shows results for an electron calibration of the crystal taken at arbitrary orientation using electron recoils resulting from interactions by x-rays from a fluorescence x-ray source. It can be seen that the response is linear and well fitted by a straight line through the origin.

Fig. 2 shows the result of measurement of carbon recoil scintillation efficiency relative to the electron efficiency vs. recoil energy.

The points in this case correspond to different neutron scatter angles with the crystal maintained with cylindrical axis always perpendicular to a line bisecting the neutron scatter angle and with the other crystal axis maintained perpendicular to the beam-crystal-NE213 plane. It can be seen that at high energy the carbon recoil efficiency is around 3%. this is broad agreement with other measurements in organic scintillators [11]. However, there is an apparent increase in efficiency at low energy, the efficiency rising to > 10% below 100 keVrecoil. Fig. 3a and b show preliminary results of the carbon recoil scntillation response vs. crystal orientation for an example carbon recoil energy of 48 keV. Fig. 3a shows the result vs. rotation about the cylindrical axis and Fig. 3b for rotation about the axis perpendicular to the cylindrical axis. The errors in each case are dominated by poor statistics at these low energies (low numbers of photoelectrons collected). Nevertheless, there is evidence for a systematic shift in recoil efficiency with angle for both orientations at about the 15% level.

4 Conclusions and Future Plans

These preliminary measurements of the carbon recoil scintillation efficiency in stilbene seem to indicate that the H recoil anisotropy previously measured in stilbene at high energy is also present for carbon recoils and is maintained at low energies (<100 keV) at about the 15% level. Further measurements are needed to confirm this and to relate this anisotropy to crystal orientation before a full prediction of the direction sensitivity for WIMPs can be assessed. These are planned for the future with a new neutron beam at Sheffield University. The measurement of the carbon recoil efficiency vs. recoil energy indicates improving efficiency at lower energy. This effect has recently been observed in other scintillators, in particular for Cs and I recoils in CsI(Tl) [12] and in CaF2(Eu) [13] and may be due to reduced energy loss to phonons as the recoil energy becomes closer to the electron excitation energy. This is potentially important for scintillation dark matter detectors since it indicates the prospect of improved detection sensitivity at low energy, regardless of any directional effects.


  1. R. Gaitskell et al., this proceedings
  2. P.F. Smith et al., Phys. Lett. B379, 299 (1996)
  3. J.B. Birks, Theory and practice of scintillation counting (Pergamon, Oxford, 1964)
  4. P. Belli et al., Nuovo Cimento C. Vol. 15, No. 4, 475 (1992)
  5. D.N. Spergel, Phys. Rev. D, Vol. 37, No. 6, 1353 (1988)
  6. J.D. Lewin and P.F. Smith, Astrop. Phys. Vol. 6. No.1, 87 (1996)
  7. P.F. Smith and J.D. Lewin, Phys. Rep. 187, 203 (1990)
  8. F.D. Brooks and D.T.L. Jones, Nucl. Instr. Meth. 121, 69 (1974)
  9. N.J.C. Spooner et al., Phys. Lett. B321, 156 (1994)
  10. N.J.C. Spooner et al., Astrop. Phys., Vol. 5, No. 3 58 (1996)
  11. G.J. Davies et al., Phys. Lett. B322 159 (1994)
  12. G. Gerbier, 2nd RESCEU Symposium, Tokyo, 26-28th Nov. 1996, tbp
  13. T. Kashimoto, 2nd RESCEU Symposium, Tokyo, 26-28th Nov. 1996, tbp