We report the composition and back-gate voltage tuned transport properties of ternary compound Bi₂(Te_1-xSe_x)₃ nanowires synthesized by chemical vapor deposition (CVD). It is found that the population of bulk carriers can be suppressed effectively with increasing the Se concentration x. In Bi₂(Te_1-xSe_x)₃ nanowires with x = 25% ± 5%, the ambipolar surface conduction associated with tuning the Fermi energy across the Dirac point of topological surface states is induced by applying a back-gate voltage. Importantly, we find that while the magneto-resistance (MR) follows the weak antilocalization (WAL) behavior when the Fermi level is tuned away from the Dirac point, MR is enhanced in magnitude and turns more linear in the whole magnetic field range (between ±9 T) near the Dirac point. The observation of the enhanced linear magneto-resistance (LMR) and crossover from WAL to LMR, near the Dirac point provides a deeper insight into understanding the nature of topological insulator’s surface transport and the relation between these two widely observed magneto-transport phenomena.

In chalcogenide topological insulator materials, two types of magnetoresistance (MR) effects are widely discussed: a sharp MR dip around zero magnetic field, associated with the weak antilocalization (WAL) effect, and a linear MR (LMR) effect that generally persists to high fields and high temperatures. We have studied the MR of thin films of the topological insulator Bi₂Te₃ from the metallic to semiconducting transport regime. In the metallic samples, the WAL is difficult to identify owing to the low magnitude of the WAL compared to the samples' conductivity. Furthermore, the sharp WAL dip in the MR is clearly present in samples with a higher resistivity. To correctly account for the low-field MR with the quantitative theory of the WAL according to the Hikami–Larkin–Nagaoka (HLN) model, we find that the classical (linear) MR effect should be taken into account in combination with the WAL quantum correction. Otherwise, the WAL fitting alone yields an unrealistically large coefficient α in the HLN analysis. This work clarifies the WAL and LMR as two distinct effects and offers an explanation for the overly large α in the WAL analysis of topological insulators in some studies.