After the discovery of the high temperature superconductor YBa2Cu307_y (YBCO), much research effort was devoted to the effect of cationic [1] and anionic [2] substitutions on the superconductivity. One of the most important areas of study was substituting oxygen by fluorine by (i) solid state reaction [3], (ii) ionic implantation [4] and (iii) solid-gas reaction [5]. After extensive investigation, most researchers agreed that the fluorination did not make any important improvement to the superconductivity under such solid state conditions and relatively low temperature. In this letter, we report the influence of fluorine ion as additive on the crystal growth from high temperature solution. A small amount of BaF2 greatly improved the crystal growth and the superconducting transition temperature (Te) by reducing aluminium contamination of Cu-O chains from the alumina crucible. The conductance spectra for Cu-O chains by scanning tunnelling microscopy (STM) showed that such contamination made the chain layer more insulating. No Fwas found in the crystal. A growth model with a monomolecular layer spiral dislocation mechanism is proposed, which enables to understand the function of F-. High quality crystals with sharp superconducting transition at 94 K have been obtained in a yttria crucible by using BaF2 as additive. The starting materials for the crystal growth were YBCO, BaCO3, CuO and BaF2. The crystal growth was carried out by spontaneous nucleation from a self-flux system with a small amount of BaF2 as additive. In order to select an appropriate system for the crystal growth, phase diagrams [6] of the flux system and the growth system were explored by improved thermogravimetric analysis (ITGA) and differential thermal analysis (DTA), from which the eutectic composition of 0.28 BaO + 0.72 CuO was chosen as solvent and the solute concentration and growth temperature were introduced from the liquidus range of the growth system. Alumina or yttria crucibles were used as containers for the crystal growth. In fact, most compounds, including BaF2, are not suitable as flux for the crystal growth of YBCO [7] because of its low stability. The crystal could be grown mainly from the self-flux system. However, we could reduce the concentration of BaF2 in the self-flux to such a low level as an additive that YBCO remains the major crystalline phase. Sometimes small amounts of a particular additive can dramatically affect the nucleation and crystal growth, resulting in improvements of crystal size or quality a n d modification of the habit. We have used B203, BaF2, Bi203 and BaC12 as additives for the crystal growth. Only BaF2 achieved improvements of both crystal growth and superconductivity [8]. Table I shows the effects of various concentrations of BaF2 on the cyrstal growth under the same solute concentration. The optimum amount of BaF2 appears