Atomic and molecular clouds are the objects of great interest, because their physical and chemical properties determine the initial conditions for star formation process [Snow&McCall 2006]. One of the key parameters determining the effectiveness of star formation in the galaxy is the characteristic timescales of formation of molecular hydrogen (i.e., timescale of the formation of molecular clouds). Moreover, observations show that some molecules are more abundant in diffuse clouds than predicted by models, with the values of abundances typical for cold dense clouds. Existing chemical models of diffuse clouds fail to explain those abundances and rate of conversion of H to H2 [e.g., Hollenbach & Salpeter, 1970; Pirronello et al., 1999, 1997; Cazaux & Tielens, 2004]. [Tsytovich V.N. Et al., 2014] proposed a theoretical model that predicts the existence of an earlier unknown instability in a diffuse medium, leading to inhomogeneities of dust distribution over the clouds on small spatial scales (<30 AU). This project is aimed to study the possible effect of gas-dust clumps, possibly formed in the interstellar medium due to that instability, on the chemical evolution of the interstellar diffuse medium. Observational basis of this project is observational data of molecular abundances obtained in different spectral ranges with leading instruments (IRAM, etc.), and a theoretical description of the instability in a diffuse medium, proposed in [Tsytovich V. N., 2014]. It is planned to systematize observational data of molecules and their abundances in the diffuse interstellar medium, perform a numerical modelling of chemical evolution for typical conditions of the diffuse interstellar medium, and study the effect of small-scale inhomogeneities of the dust distribution in the diffuse medium on the characteristic timescales of conversion of atomic hydrogen to molecular, and on the abundances of other molecules observed in diffuse clouds.