At present, the effect of alkali metals (AM) on the electronic properties of polycrystalline silicon (PS) [1,2] and its adsorption properties can be considered studied well enough from both experimental and the oretical points of view [2][3][4]. The interaction of AMs with radiation defects with energy levels of ≈0.15, 0.17, ~0.3, and ~0.36 eV, which results in a decrease in the density of states at grain boundaries (GB) and an increase in the radiation resistance of solar cells (SC), is reliably identified [1,2]. The potential barrier increases during the diffusion and adsorption of AM atoms along GBs [4]. As for the effect of AMs on the electronic properties of GBs on the PS surface during doping, it is still an unsolved problem, although the effect of such impurities on PS's electrical properties [1,2] and the effect of adsorption on the electrical properties of GBs in the PS bulk [4] are relatively well studied. An urgent problem concerns study of the effect of AMs on the electrical properties of GBs on the PS surface during the development of SCs and other semiconductor devices on it. This work is devoted to study of the effect of AMs on the electrical properties of GBs on the PS surface.It is known that, among the methods for determin ing a GB's electrical properties, carrier concentration variation by doping (see, e.g., [5] and references therein) is the most attractive when studying polycrys talline elemental semiconductors. This method makes it possible to obtain information in a wide energy range, especially, when using n type and p type samples.Samples were prepared as polished wafers 300 μm thick from cast p type PS with the resistivity ρ ≈ 1 Ω cm and grain sizes of 100-300 μm [6]. Rapidly diffusing AMs, i.e., Li, Na, K, and Cs were chosen as dopants. Doping was performed by ion implantation (II). After II, samples were annealed in vacuum of 6-10 Torr in the temperature range from 300 to 1000 K. To estimate the effect of AMs on the GB's electrical properties, the sample conductivity σ was measured by the four point probe and Van der Pauw methods before and after II and after each heat treatment performed with a step of 50 K.It is known [5] that the conductivity expanded in Taylor series is given by (1) where e is the electron charge, k is the Boltzmann con stant, 〈a〉 is the grain size, A is the effective Richardson constant, T is the temperature, and ϕ is the potential barrier height at the GBs. The quantity ϕ is related to depletion layer thicknesses W on both sides of the GB and the concentration of electrically active dopants N G . If grains are not completely depleted, the potential barrier height is written as [5] (2) where ε is the permittivity.It is known from the doping method [5] that the concentration N G of the electrically active dopant is related to the concentrations N D of implanted impuri ties. The implanted impurity concentration N D decreases during heat treatment either due to direct evaporation or segregation on oxide layers which are etched in the next stages. Then N G can become ...