Confidence Intervals · Challenge problem

Two independent studies each report a $95\%$ confidence interval of the form estimate $\pm\, 1.96 \times$ standard error, with normal sampling distributions: $(10,\ 14)$ for $\mu_1$ and $(13,\ 17)$ for $\mu_2$. Using the same two estimates and standard errors, find the LOWER end of the $95\%$ confidence interval for $\mu_2 - \mu_1$. Express your answer in simplest radical form.