The & l-has-van Alphen (dI-ivA)signal of the organic superconductor 13''-(BEDlTF)#SFsCHjCF$ 03O3 shows an inverse-sawtooth wave form which proves the existence of an ideal two-tlmensiond (2D) Fermi surface. The dHvA wave shape cartalmost perfectly be describedby a 2D theosy assuming a constantchemicatpotential. TMseither impIiesthe existence of the predictedquasi-one-dimensional band with an exceptionallylarge densi[yof statesor the chemical potentialmaybe pinneddue to locatized states nearthe Fermienergy.Keywords: Organic superconductors. M~gnetic measurements Organic metals of the type {BEDT-TTF)zX, where BEDT-TTF is bkethyletdthio-tetrathiafulvalene and X represents a monovalent anion, are characterized by their pronounced two-dimensional (2D) electronic structure with, often, negligible dispersion perpendicular to the highly conducting planes [1]. Nevertheless. the de Haas-van Aiphen (dHvA) oscillations in most of these high-quality metals exhibit only minor deviations from the conventional 3D Lifshitz-Kosevich (LK) theory [2]. This is in contrast to magnetoresistance oscillations where in a number of organic metals strong apparent deviations of the Shubnikovde Haas (SdH) signal fkom IX behavior have beersreported (see references in [3]). In transportdata, however, complicated scattering mechanisms are involved and a non-trivial tensor inversion is needed to extract the true SdH signaL Here, we report on true thermodynamic dHvA data which provide experimental evidenee for the existence of a perfeet 2D Fermi liquid in a bulk 3D organic metal.The highquaiity $-(BEDT-T'IT9ZSF5CH2CFZS03 single crystals were grown by cketrocrysb]]ization [4]. The magnetization Ml perpendicular to the applied magnetic field B was measured by means of a cmtilever torque magnetome- tory. The torque is given by ? = M x E which is proportional to B2 for an anisotropic field-independent susceptibility x of the sample. In general, x is a tensor and the absolute value of the torque has th$ form r = AXB2sin229, where AZ is the susceptibility difference between two principal axes of the x tensor and r9is the angle between 1? and one of the prittcipal tensor axes. For a 2D Fermi surface the torque dHvA signal is proportional to dFhiQ, with F = FJCOSQand the angie~between B and the normal to the conducting planes. Fig. l(a) shows the torque signal at Q = -0.7" where the dHvA amplitude is very small and any nonlinearities of the torque signal can be excluded. The dHvA oscillations are sitting on a torque background which follows a B2 behavior up to the highest field. This background signal is up to T = 75 K almost temperature independent and shows a sin2fi dependence as expected for a static field-independent susceptibility (Fig. 2). One principal axis of the x tensor lies at @ = 30". The occurrence of this anisotropic susceptibility is rather unusual for the organic metals and the origin of the signa~in the present material is unclear. It might be a remnant of the local magnetism of the molecular bands.At low fields and high temperat...