To clarify the whole picture of the valence-band structures of prototype ferromagnetic semiconductors (III,Mn)As (III: In and Ga), we perform systematic experiments of the resonant tunneling spectroscopy on [(In_0.53Ga_0.47)_1-x Mn_x]As (x=0.06-0.15)
and In_0.87Mn_0.13As grown on AlAs/ In_0.53Ga_0.47As:Be/ p+InP(001). We show that the valence band of InGaMnAs almost remains unchanged from that of the host semiconductor InGaAs, that the Fermi level exists in the band gap, and that the p-d exchange splitting in the valence band is negligibly small in (InGaMn)As. In the In0.87Mn0.13As sample, although the resonant peaks are very weak due to the large strain induced by the lattice mismatch between InP and InMnAs, our results also indicate that the Fermi level exists in the band gap and that the p-d exchange splitting in the valence band is negligibly small. These results are quite similar to those of GaMnAs obtained by the same method, meaning that there are no holes in the valence band, and that the impurity-band holes dominate the transport and magnetism both in the InGaMnAs and In_0.87Mn_0.13As films. This band picture of (III,Mn)As is remarkably different from that of II-VI-based diluted magnetic semiconductors.
We comment on the recent paper Reconciling results of tunnelling experiments on (Ga,Mn)As arXiv:1102.3267v2 by Dietl and Sztenkiel. They claimed that the oscillations observed in the d2I/dV2-V characteristics in our studies on the resonant tunneling
spectroscopy on GaMnAs, are not attributed to the resonant levels in the GaMnAs layer but to the two-dimensional interfacial subbands in the GaAs:Be layer. Here, we show that this interpretation is not able to explain our experimental results and our conclusions remain unchanged.
