Paper detail

From spherical to periodic symmetry: the analog of orbital angular momentum for semiconductor crystals

The angular momentum formalism provides a powerful way to classify atomic states. Yet, requiring a spherical symmetry from the very first line, this formalism cannot be used for periodic systems, even though cubic semiconductor states are commonly classified according to atomic notations. Although never noted, it is possible to define the analog of the orbital angular momentum, by only using the potential felt by the electrons. The spin-orbit interaction for crystals then takes the $\mathbfcal{\hat{L}}\cdot \hat{\vS}$ form, with $\mathbfcal{\hat{L}}$ reducing to $\hat{\vL}=\vr\times\hat{\vp}$ for spherical symmetry. This provides the long-missed support for using the eigenvalues of $\mathbfcal{\hat{L}}$ and $\mathbfcal{\hat{J}}=\mathbfcal{\hat{L}}+\hat{\vS}$, as quantum indices to label cubic semiconductor states. Importantly, these quantum indices also control the phase factor that relates valence electron to hole operators, in the same way as particle to antiparticle, in spite of the fact that the hole is definitely not the valence-electron antiparticle. Being associated with a broader definition, the ($\mathbfcal{\hat{L}},\mathbfcal{\hat{J}}$) analogs of the $(\hat{\vL},\hat{\vJ})$ angular momenta, must be distinguished by names: we suggest "spatial momentum" for $\mathbfcal{\hat{L}}$ that acts in the real space, and "hybrid momentum" for $\mathbfcal{\hat{J}}$ that also acts on spin, the potential symmetry being specified as "cubic spatial momentum". This would cast $\hat{\vJ}$ as a "spherical hybrid momentum", a bit awkward for the concept is novel.