Breakdown of compensation and persistence of nonsaturating magnetoresistance in gatedWTe2thin flakes

The recently discovered large nonsaturating magnetoresistance in semimetal $\mathrm{WT}{\mathrm{e}}_{2}$ may result from near-perfect electron-hole compensation, however recent reports question whether the compensation is adequate to explain the observations. Experiments on significantly uncompensated $\mathrm{WT}{\mathrm{e}}_{2}$ are needed. We measure magnetoresistance ${\ensuremath{\rho}}_{xx}(H)$, Hall effect ${\ensuremath{\rho}}_{xy}(H)$, and an electrolyte gating effect in thin (100 nm) exfoliated $\mathrm{WT}{\mathrm{e}}_{2}$. We observe ${\ensuremath{\rho}}_{xy}(H)$ linear in $H$ at low $H$ consistent with near-perfect compensation, however ${\ensuremath{\rho}}_{xy}(H)$ becomes nonlinear and changes sign with increasing $H$, implying a breakdown of compensation. We break compensation more significantly by using an electrolytic gate for highly electron-doped $\mathrm{WT}{\mathrm{e}}_{2}$ with Li. In gated $\mathrm{WT}{\mathrm{e}}_{2}$ the nonsaturating ${\ensuremath{\rho}}_{xx}(H)$ persists to $H=14\phantom{\rule{0.16em}{0ex}}\mathrm{T}$, even with significant deviation from perfect electron-hole compensation $(p/n=0.84)$ where the two-band model predicts a saturating ${\ensuremath{\rho}}_{xx}(H)$. Our results indicate electron-hole compensation is not the mechanism for extremely large magnetoresistance in $\mathrm{WT}{\mathrm{e}}_{2}$; alternative explanations are needed.