* Corresponding author. Tel.: 49-371-531-3108; fax: 49-371- 531-8049. E-mail address: s.jahn@physik.tu-chemnitz.de (S. Jahn) Nuclear Instruments and Methods in Physics Research A 438 (1999) 452 } 459 Monte-Carlo simulation of a neutron Brillouin scattering spectrometer Sandro Jahn*, Jens-Boie Suck TU Chemnitz, Institute of Physics, Materials Research and Liquids, D-09107 Chemnitz, Germany Received 20 May 1999; received in revised form 16 August 1999 Abstract A Monte-Carlo simulation was done to optimise the setup of a new type of time-of-#ight spectrometer, a thermal neutron Brillouin scattering (NBS) spectrometer. After collimation of the incident white neutron beam "ve incident energies between 20 and 138 meV can be obtained from four di!erent monochromator crystal faces. The monochromatic beam is split into nine pencil-like separate beams to improve the Q-resolution of the NBS-spectrometer. Including the time resolution of the Fermi chopper (after a background chopper) an energy resolution for elastic scattering between 2% and 7% and a momentum transfer resolution of 1.3}2.0% are obtained.  1999 Elsevier Science B.V. All rights reserved. PACS: 02.70.Lq; 07.05.Fb; 29.30.Hs Keywords: NBS; TOF; Monte-Carlo simulation 1. Introduction The atomic dynamics in condensed matter can be very e!ectively investigated by inelastic neutron scattering experiments. For topologically dis- ordered condensed matter these experiments are most sensitive to the collective atomic vibrations at small momentum transfers, Q, because the struc- ture of the inelastic part of the spectra (Brillouin peaks) is more distinct there [1] and for #uids they do not overlap as much with the central line as they do at larger Q's [2]. For this reason a new type of time-of-#ight (TOF) spectrometer adapted to well- resolved neutron inelastic scattering at small scat- tering angles, a neutron Brillouin scattering (NBS) spectrometer, was proposed [3]. As presently no spectrometer of this type is available at a thermal neutron source, a new NBS instrument called BRISP was proposed to be built at the HFR in Grenoble [4]. A good energy and momentum transfer resolu- tion is important in the case of the investigation of non-crystalline matter because one wants to ana- lyse the linewidths of the collective excitations which are broadened due to their short lifetimes and to follow the dispersion of these excitations in Q-space. In the measured spectra both the line broadening from the "nite lifetime of the excita- tions and from the resolution function of the 0168-9002/99/$ - see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 9 9 ) 0 0 8 3 8 - 4