Low-x contribution to the Bjorken sum rule within double logarithmic $ln^2x$ approximation

Abstract

The small-$x$ contributions to the Bjorken sum rule within double logarithmic $ln^2x$ approximation for different input parametrisations $g_1^{NS}(x,Q_0^2)$ are presented. Analytical solutions of the evolution equations for full and truncated moments of the unintegrated structure function $f^{NS}(x,Q^2)$ are used. Theoretical predictions for $\int_{0}^{0.003} g_1^{NS}(x,Q^2=10) dx$ are compared with the SMC small-$x$ data. Rough estimation of the slope $\lambda$, controlling the small-$x$ behaviour of $g_1^{NS}\sim x^{-\lambda}$ from the SMC data is performed. Double logarithmic terms $\sim (\alpha_s ln^2x)^n$ become leading when $x\to 0$ and imply the singular behaviour of $g_1^{NS}\sim x^{-0.4}$. This seems to be confirmed by recent experimental SMC and HERMES data. Advantages of the unified $ln^2x$+LO DGLAP approach and the crucial role of the running coupling $\alpha_s=\alpha_s(Q^2/z)$ at low-$x$ are also discussed.