Searching for Spectroscopic Multiple-star Systems in the LAMOST Medium-Resolution Spectroscopic Survey

Abstract We employed the Gaussian decomposition package Gausspy+ to autonomously identify multipeak features in the cross-correlation functions (CCFs) of spectra from the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) Medium-Resolution Survey (MRS), enabling the detection of single-lined spectroscopic binaries (SB1s), double-lined spectroscopic binaries (SB2s), and spectroscopic triples (STs). By providing Gausspy+ with training spectra exhibiting single-, double-, and triple-peak CCFs, we derived optimal fitting parameters suited to the dataset. To ensure the reliability of the fits, we established selection criteria for exposures showing double- or triple-peak CCFs based on statistical analysis of the test sample. Combining visual screening with our testing sets, the accuracy for SB2 identification reached 95.7%, while the accuracy for ST identification reached 98.6%. Applying this method to nearly 450,000 LAMOST DR10 targets, we identified 15,887 SB​​​​​​2 candidates and 8771 ST candidates. In addition, we analyzed 903 targets that exhibit significant radial velocity (RV) variations but whose CCFs only show a single peak. For these objects, we performed orbital parameter inference using The Joker , and identified 294 SB1 candidates with periodic RV variations. Notably, among the identified candidates, 65.3% of SB2s, 93.9% of STs, and 93.9% of SB1s are newly discovered systems, substantially expanding the census of spectroscopic multiple systems in the LAMOST MRS dataset. Furthermore, we also provide orbital solutions for 334 SB2 systems. For ST systems, we identified 46 hierarchical triple systems and obtained their orbital solutions. These results further enrich the characterization of multiple star systems in the LAMOST MRS dataset.