Fig. 1. Optical storage for two-channel VBs in atomic ensembles. (a) Experimental setup. A cigar cold Rb87 atomic ensemble obtained from a two-dimensional (2D) magneto-optical trap (MOT) acts as the storage medium. The two probe VBs with topological charge of 1 (top) and 2 (bottom) are modulated into four-channel beams using a quarter-wave plate (QWP), half-wave plates (HWPs), and beam displacers (BDs). The beams are then focused by a lens (L1) and mapped into the atoms for storage. After the storage process, the state projection measurement system consists of a QWP, a PBS, a spatial light modulator (SLM), Fabry-Perot cavities (FPCs, with a bandwidth of 500 MHz), and single-mode fibers (SMFs). (b) The simplified energy level structure of the optical storage based on EIT; |g⟩ and |s⟩ correspond to |5S1/2,F=1⟩ and |5S1/2,F=2⟩, which are the two hyperfine ground states of Rb87D1-line, while |e⟩ is the excited state |5P1/2,F=2⟩. The frequencies of the control (blue) and probe (red) beams are tuned resonantly to couple two ground states with an excited state, creating a coherent path that allows the probe beam to pass through the atoms without absorption. (c) The experimental sequence including the cooling, repump, magnetic coils, Zeeman pump, control, and probe field temporal profiles. The storage process is achieved by adiabatically turning off the control beam, which converts the probe photon into an atomic collective excitation. Subsequently, an identical photonic mode to the probe beam is retrieved by reactivating the control light after a specified time interval.

Fig. 2. Storage performance of the two-channel VBs. (a1), (b1) The intensities and phases of the four basis vectors. (a2), (b2) The temporal waveforms of the input probe beams and the retrieval signals after 500 ns storage time.

Fig. 3. (a1), (a2) Storage efficiencies for four basis vectors of channel 1 and channel 2 as a function of storage time delay; the blue and red dots represent the horizontal and vertical basis vectors of two-channel VBs, respectively. (b1), (b2) Interference curves of two-channel VBs after storage.

Fig. 4. The reconstructed density matrices of the two-channel VBs before (input) and after (output) storage. (a) State |H⟩|l=1⟩+|V⟩|l=−1⟩; (b) state |H⟩|l=2⟩+|V⟩|l=−2⟩.

Fig. 5. (a) The experiment setup and energy level of residual magnetic field compensation with microwave spectrum. (b) The microwave spectrum before compensation and after compensation.

Fig. 6. The absorption spectra of four basis vectors for two-channel VBs versus the probe beam detuning from the atomic resonance |5S1/2,F=1⟩→|5P1/2,F=2⟩. (a1), (a2) Horizontal and vertical basis vectors for channel 1. (b1), (b2) Horizontal and vertical basis vectors for channel 2 (the input states for the two paths are |H⟩|l=1⟩+|V⟩|l=−1⟩ for channel 1 and |H⟩|l=2⟩+|V⟩|l=−2⟩ for channel 2).

Fig. 7. The projection vector based on polarization and OAM. (a) The selection of polarization and OAM basis vectors for channel 1; (b) the selection of polarization and OAM basis vectors for channel 2.