Tuning the electrical transport of type II Weyl semimetal WTe2 nanodevices by Ga+ ion implantation

Here we introduce lattice defects in WTe2 by Ga+ implantation (GI), and study the effects of defects on the transport properties and electronic structures of the samples. Theoretical calculation shows that Te Frenkel defects is the dominant defect type, and Raman characterization results agree with this. Electrical transport measurements show that, after GI, significant changes are observed in magnetoresistance and Hall resistance. The classical two-band model analysis shows that both electron and hole concentration are significantly reduced. According to the calculated results, ion implantation leads to significant changes in the band structure and the Fermi surface of the WTe2. Our results indicate that defect engineering is an effective route of controlling the electronic properties of WTe2 devices.

Firstly, we rule out the possibility of interference from samples aging or damage during GI and annealing. The AFM images of sample E under different conditions is shown in Fig. S1. Table S1 show the parameters related to thickness and the roughness of the sample`s surface.
There is oxidation on the top surface of the WTe2 sample. But the oxidation phenomenon is limited to top surface of sample, and there are no obvious Raman peak position shifts during the degradation in both 2L and 3L WTe2 as reported in another work 1 . The oxidation products such as WOx and TeO2 will passivate the WTe2 surface and prevent oxygen from further diffusing into inside of lattice and protect inner layer of WTe2, which has been reported in previous work 1 . The thickness of the samples we used for Raman or transport measurement in this work are from 8.5nm to 25.3nm (12-36 layers), so the effect of surface oxidation is very limited. In short, oxidation has no significant observable effect in our study.
The Ga+ ion we used for ion implantation is a single ion, and the caused damage by each ion is very limited. GI can only cause damage in the atomic level, and will not cause obvious changes in appearance and roughness as shown in Fig S1. Though some difference are observed in the parameters, the changes of thickness, root mean square and average deviation are not larger than the resolution (~100 pm) of our AFM.
So we can rule out the possibility of interference from samples damage during GI and annealing.
Here, the dependence of the implantation depth and the energy is discussed. SRIM is used to acquire the energy dependent projected range shown in Table S2. Projected range is the depth of the maximum Ga+ concentration. Figure S2 show the Raman spectra from sample C with thickness about 17.2 nm.
The parameters extracted from the Raman spectra are shown in Table S3. It is obvious that all the Raman vibrational modes (except the weakest mode 9 1 A ) originating from the relative movements of Te atoms weaken dramatically after GI (irradiated to total doses of  -2 0.44μ C cm ) relative to 2 1 A . Combining the calculated results of the formation energy in the main text, the Te Frenkel defect can also be determined as the dominant defect in sample C with GI. Figure S3 shows the transport measurement data from sample D with thickness about 15.7 nm. The inset of Fig. S3(b) show the low field part of the MR, and obvious weak anti-localization (WAL) can be observed, which suggest the presence of disorder in the crystal. Parameters used for GI and that derived from the two band-model fitting are shown in Table S4. Ga+ ion implantation lead to a significant reduction in carrier concentration and mobility. However, unlike sample B, ion implantation seems to break the carrier balance in sample D, with n/p changing from 0.941 to 0.824. Figure S4 shows the distribution of high symmetry points in the first Brillouin zone.
The changes in Fermi level caused by various crystal defects are recorded in Table   S5.