1: \begin{abstract}

2: We calculate the width $2\Delta_{\text{CT}}$ and intensity of the charge-transfer peak 

3: (the one lying at the on-site energy $E_d$) in the impurity

4: spectral density of states as a function of $E_d$ in the SU($N$) impurity

5: Anderson model (IAM). We use the dynamical density-matrix renormalization group (DDMRG)

6: and the noncrossing-approximation (NCA) for $N$=4, and a 1/$N$ variational

7: approximation in the general case. In particular, while for $E_d \gg \Delta$, 

8: where $\Delta$ is the resonant level half-width, $\Delta_{\text{CT}}=\Delta$ as

9: expected in the noninteracting case, for $-E_d \gg N \Delta$ one has 

10: $\Delta_{\text{CT}}=N\Delta$. In the $N$=2 case, some effects of the variation of $%

11: \Delta_{\text{CT}}$ with $E_d$ were observed in the conductance through a quantum

12: dot connected asymmetrically to conducting leads at finite bias [J.

13: K\"onemann \textit{et al.}, Phys. Rev. B \textbf{73}, 033313 (2006)]. More

14: dramatic effects are expected in similar experiments, that can be carried

15: out in systems of two quantum dots, carbon nanotubes or other, realizing the

16: SU(4) IAM.

17: \end{abstract}