Defect-induced multiferroicicy in bulk solid solutions of WSe$_2$ and WTe$_2$

Abstract

Transition metal dichalcogenides provide a versatile platform for tunable ferroic phenomena at the atomic scale owing to their reduced dimensionality. Here we investigate the structural, magnetic, and ferroelectric properties of bulk solid solution W(Se1-xTex)2(1-delta) single crystals synthesized by chemical vapor transport. The room temperature behavior is analyzed as a function of tellurium concentration (x) and chalcogen defect fraction (delta). X ray diffraction and Raman spectroscopy reveal lattice expansion and symmetry reduction with increasing x, consistent with a 2H to 1Td structural transition above a critical composition xc about 18 percent. Piezoresponse force microscopy identifies piezoelectricity near stoichiometric compositions (delta less than 5 percent) and switchable ferroelectricity in the chalcogen deficient regime (delta greater than 20 percent). Magnetometry measurements show a corresponding evolution from paramagnetic to ferromagnetic behavior with increasing delta. Near stoichiometric Te poor samples exhibit piezoelectric and paramagnetic responses, whereas multiferroic states characterized by the coexistence of ferroelectric and ferromagnetic responses emerge at high vacancy concentrations. The performed characterizations indicate that x primarily governs structural symmetry, while delta controls the emergence of both ferromagnetic and ferroelectric responses. These trends are summarized in a configurational phase diagram highlighting the cooperative influence of dopants and defects on ferroic behavior. Overall, controlled stoichiometry and vacancy engineering offer an effective strategy to tailor ferroic responses in transition metal dichalcogenides.

Figures (7)

• Figure 1: (a) Crystal structure of a 2H-polytype of Te-doped WSe$_{2}$ with space group P63/mmc ($\#$ 194). (b) Diagram of a configurational space representing the possible room-temperature solid solutions with formula W[Te$_x$Se$_{1-x}$]$_{2(1-\delta)}$ . The $x$- and $y$-axes represent the number of moles of tellurium and selenium relative to tungsten, $n_{Te}/n_{W}$ and $n_{Se}/n_{W}$, respectively. The pure components 2H-WSe$_{2}$ and 1T$_d$-WTe$_2$ (with zero defects, $\delta=0$) are represented by the $(0,2)$ and $(2,0)$ coordinates in the diagram. (c) Crystal structure of a 1T$_d$-polytype of WTe$_{2}$ with space group Pnm21.
• Figure 2: (a) Powder x-ray diffraction patterns of undoped and Te-doped samples. The observed peaks correspond to the ${00\ell}$ reflections of the 2H polytype and align with the reported positions of stoichiometric 2H-WSe$_2$ (dotted vertical lines) for compositions with $x \le 17\%$, spanning both $\delta>0$ and $\delta<0$ regimes. (b)–(d) Expanded views of representative reflections [(002), (006), and (008)] showing a systematic shift toward lower $2\theta$ with increasing Te content, consistent with an expansion of the $c$-axis lattice parameter. In the close-up views, patterns are vertically offset and independently scaled in intensity for clarity. (e) Evolution of the extracted $c$-axis lattice parameter as a function of Te content. (f) Dependence of the lattice parameter on the chalcogen defect fraction $\delta$. (g) Schematic representation of structural symmetry across the compositional $x$–$\delta$ phase space.
• Figure 3: (a) Raman spectra of W[Te$_x$Se$_{1-x}$]$_{2(1-\delta)}$ single crystals with varying composition showing a systematic shift toward lower wavenumbers with increasing Te content. Dashed lines mark the peak positions of the undoped ($x=0$) compound. (b) Evolution of the extracted $E^1_{2g}$ and $A^1_g$ peak positions, their separation $(A^1_g-E^1_{2g})$, and relative intensities as functions of Te content. (c) Corresponding evolution as a function of the chalcogen defect fraction $\delta$.
• Figure 4: Magnetization hysteresis loops at 300 K for W[Te$_x$Se$_{1-x}$]$_{2(1-\delta)}$ samples with different tellurium fractions $x$, for both (a) $\delta < 0$ and (b) $\delta > 0$. Diamagnetic contributions have been subtracted. Panels (c) and (d) display the coercive field $H_c$, saturation magnetization $M_s$, and remanent magnetization $M_R$ as functions of the chalcogen defect parameter $\delta$ and tellurium fraction $x$, respectively.
• Figure 5: SS-PFM piezoresponse phase and amplitude loops at 300 K for W[Te$_x$Se$_{1-x}$]$_{2(1-\delta)}$ samples with varying tellurium fractions $x$. (a)$\delta < 0$ and (b)$\delta > 0$. (c) DART-PFM amplitude normalized by the simple harmonic oscillator (SHO) model for $x=7\%, \delta=-2\%$. (d) Effective $d_{33}$ piezoelectric coefficient from DART-PFM amplitude measurements for $x=7\%, \delta=-2\%$ and $x=9\%, \delta=-1\%$. Error bars correspond to the standard deviation across each $500$ nm measurement area for each voltage.