The first absolute experimental determinations of the differential cross-sections for the formation of ground-state positronium are presented for He, Ar, H2 and CO2 near 0 ○ . Results are compared with available theories. The ratio of the differential and integrated cross-sections for the targets exposes the higher propensity for forward-emission of positronium formed from He and H2.PACS numbers: 36.10. Dr, 34.80.Bm, 34.80.Uv The formation of positronium (Ps, the bound state of an electron and a positron) is an important channel in the scattering of positrons from atoms and molecules e.g. [1][2][3], accounting for up to 50% of the total cross-section, with experimental and theoretical investigations of its integrated formation cross-sections available for a wide range of atoms and simple molecules, e.g. [3][4][5]. Recent experimental studies also include its formation in an excited state [6] or accompanied by ionic excitation [7]. However, whilst theoretical predictions for the differential Ps formation cross-section ( dQPs dΩ ) are available for atomic [8][9][10][11][12][13][14][15][16] and molecular [17,18] hydrogen, the noble gases [9,16,[19][20][21][22][23][24][25][26][27][28][29] Details of the experimental arrangement employed at UCL for producing a beam of Ps atoms, together with a review of recent advances, may be found in [41]. In brief, the Ps beam is produced by charge-exchange of positrons (e + ) with a target gas (A), i.e. e + + A → Ps + A + , and detected downstream by a channel-electronmultiplier (CEM or CEMA) in coincidence with one or more γ-ray detectors (e.g. CsI or NaI) where it has been found to be composed predominantly of groundstate atoms [42,43].Depending on the relative spin orientation of its constituents, ground-state Ps may be formed in an ortho-( 3 S 1 ) or para-( 1 S 0 ) state. The two are characterized by lifetimes differing by three orders of magnitude (142 ns and 125 ps, respectively) and different annihilation modes (dominantly 3-γ and 2-γ, respectively). Only ortho-Ps reaches the detection region.In order to determine dQPs dΩ (a measure of the probability that Ps is emitted within a solid angle dΩ = 2πsinθdθ), we have measured the number of Ps atoms (ε where τ Ps is the lifetime of ortho-Ps, t its flight-time to the detector and ǫ Ps the 'true' Ps beam production efficiency. In Eq.(1), ǫ m Ps may be seen to depend on the (energy-dependent) ratio of the positron to positronium detection efficiencies R d also determined by our group [38][39][40].By studying the variation of ǫ Ps with gas pressure, optimum beam operating conditions may be determined for a given target and Ps energy [39,40,44,45]. An example is shown in Figure 1 for production of 20 eV Ps from CO 2 . Here ǫ Ps may be seen to increase and then decrease with increasing pressure. This variation may be expressed as: