Spectroscopic measurements of the D α and H α line profiles emitted within the edge region of a tokamak plasma, have revealed the existence of a cold central component, broadened mainly by the Zeeman effect arising from the confining magnetic field. Evaluation of the Doppler broadening suggests that the cold component is probably produced by electron impact-induced molecular dissociation, dissociative excitation being one of the few mechanisms which can explain the formation of atoms of kinetic energy around 0.2 eV against a background of comparatively hot electrons and ions. Further analysis of these line profiles, observed along different directions in the equatorial plane and under various tokamak discharge conditions, reveals, in addition to this effective 'cold temperature', an effective 'lukewarm temperature', which we explain in terms of an appreciable collisional heating mechanism. Estimates of the rates for ion-induced dipole and ioninduced quadrupole collisions with excited atoms, yield values of the correct order of magnitude for this observed 'lukewarm temperature'. In addition, measurements of Balmer-α line profiles, radiated from a gas discharge in a magnetic field of similar magnitude, are analysed and their shapes compared with those from the tokamak plasma.

Profiles of the Balmer lines D,(H,), Db(Hs) and D, (H,) have been measured in the scrape-off layer and within the edge of the TEXTOR (upgrade) plasma, under Ohmic conditions and with neutral-beam injection. Each line profile shows a strong Zeeman effect in the vicinity of line centre, and a marked central dip when mainly the ( I components are observed. The line core evidently originates from cold atoms in the edge plasma, excited in the course of molecular dissociation, while the broad pedestal on which the core rests is radiated by excited atoms produced through chargeexchange recombination of deuterons (protons), transported outwards from the much hotter plasma interior, and by atoms heated directly by collisions with the deuterons (protons). Core temperatures of about 0.5 eV and less are obtained from line profile analysis.

Spectroscopic measurements of low‐n Balmer line profiles of atomic hydrogen and deuterium, emitted within the edge region of the TEXTOR‐94 tokamak plasma, have revealed the existence of a class of cold excited atoms, whose probable origin has been ascribed to electron impact‐induced molecular dissociation. Associated with these cold radiators are a second group of ‘lukewarm’ atoms, i.e. atoms heated by elastic collisions with hot protons (deuterons) diffusing outward from the plasma interior, as well as a third group of ‘hot’ atoms, produced in the corresponding excited states directly by charge‐exchange recombination between protons (deuterons) and boundary region atoms. A mechanism recently proposed to explain the heating process quantitatively, in terms of elastic atom‐ion collisions, is applied and discussed in this paper.

Experimental results from TEXTOR are presented to provide strong evidence for the feasibility of the "cold radiative plasma mantle", a concept which might be a possible solution for the energy exhaust problem in a fusion reactor. The concept is compared with the high density divertor. The compatibility to other constraints, limitations and open problems are discussed, in particular the issues of stationarity (feed-back control, thermal instabilities, q=2), energy confinement. Heexhaust and fuel dilution.

Examples are presented of Doppler broadening measurements on singlet, doublet and triplet emission spectra from the following impurity ions found in the boundary layer of the TEXTOR tokamak: CII, CIII, CIV, Sin, and Sim. The shapes of these spectral lines are significantly influenced by the confining magnetic field of some 2T, in some cases exhibiting an appreciable Paschen-Back effect which complicates their appearance. It is shown that reliable values for the particular ion temperature can be obtained from the Doppler widths of the various Zeeman components, when the presence of the magnetic field is properly accounted for. Such temperatures derived from partially ionised impurity species should, however, be cautiously interpreted, as the ions in question probably do not exist for long enough in the particular ionisation stage to achieve thermal equilibrium with the background deuterons and protons. This interpretation of our results is supported by a simple one-dimensional model of ionisation and collisional heating processes in the plasma boundary.