Tantalum diselenide

On the left, how tantalum and selenium atoms are stacked in the 2H phase. On the right, how selenium atoms are bounded to the central tantalum atom. Red is selenium while light blue is tantalum.

In the 1T phase, selenium atoms show an octahedral symmetry[5] and the relative orientation of the selenium atoms in the topmost and bottommost layers is opposed. On a macroscopic scale, the sample shows a gold colour.[4] The lattice parameters are a = b = 3.48 Å,[5] while c = 0.627 nm.[4] Depending on the temperature, it shows different types of charge density waves (CDW): an incommensurate CDW (ICDW) between 600 K and 473 K[4] and a commensurate CDW (CCDW) below 473 K.[4] In the commensurate CDW, the resulting superlattice shows a ${\displaystyle {\sqrt {13}}\times {\sqrt {13}}}$ reconstruction[6] often addressed as star of David (SOD),[6] with respect to the lattice parameter (a = b) of non distorted TaSe₂ (above 600 K). Film thickness can influence as well the CDW transition temperature: the thinner the film, the lower the transition temperature from ICDW to CCDW.[4]

In the 1T phase the single trilayers are stacked always in the same geometry, as shown in the corresponding image.

The 2H phase is based on a configuration of selenium atoms characterized by a trigonal prismatic symmetry[5] and an equal relative orientation in the topmost and bottommost layers. The lattice parameters are a = b = 3.43 Å,[5][7] while c = 1.27 nm.[7] Depending on the temperature, it shows different types of charge density wave: an incommensurate CDW (ICDW) between 122 K and 90 K and a commensurate CDW (CCDW) below 90 K.[4][8] The lattice distortion below 90 K gives rise to a CCDW that makes a 3 × 3 reconstruction[9] with respect to the non-distorted lattice parameter (a = b) of 2H TaSe2 (above 122 K).

In the 2H phase the single trilayers are stacked one opposed to others, as shown in the relative image. Through molecular beam epitaxy it is possible to grow one single trilayer of 2H TaSe₂, also known as 1H phase. Basically, the 2H phase can be seen as the stacking of 1H phase with opposed relative orientation with respect to each others.[5]

In the 1H phase the ICDW transition temperature is raised to 130 K.[9]

TaSe₂ exhibits different properties according to the polytype (2H or 1T), even if the chemical composition remains unchanged.

The resistivity at low temperature is similar to that of a metal, but it starts decreasing at higher temperatures. A peak is exhibited at approximately 473 K, which resembles the behavior of semiconductors.[10] 1T phase has almost two orders of magnitude higher resistivity than to the 2H phase.[10]

The magnetic susceptibility of the 1T phases has no peaks at low temperature and remains always nearly constant until 473 K temperature is reached (ICDW temperature transition), when it jumps to slightly higher values.[10] 1T phase is diamagnetic.[11]

Resistivity linearly depends on the temperature when the latter exceeds 110 K.[12] On the opposite, below this threshold it shows a non-linear behaviour. This abrupt variation of R(T) at 110 K might be related to the formation of some kinds of magnetic ordering in TaSe₂: ordered spins scatter electrons in a less efficient way. This increases electrons mobility and yields a faster drop in resistivity than that ideally corresponding to a linear trend.

The magnetic susceptibility of the 2H polytype slightly depends on the temperature and peaks in the range 110-120 K. The trend is linearly ascending or descending below and above 110 K, respectively.[12] This maximum in the 2H phases is related to the formation of the CCDW at 120 K.[10] The 2H phase is Pauli paramagnetic.[11]

Coulored scheme of the disposition of tantalum atoms in the 1T tantalum diselenide charge density wave depending on their site.

The Hall coefficient ${\displaystyle R_{H}}$ is almost independent of the temperature above 120 K, a threshold below which it instead starts to drop to lastly reach a value of zero at 90 K. In the range included between 4 and 90 K, the coefficient ${\displaystyle R_{H}}$ is negative, its minimum being experienced at approximately 35 K.[12]

Bulk 1T TaSe₂ is metallic,[5] while single monolayer (trilayer Se-Ta-Se in octahedral symmetry) is observed to be insulating[5][6] with a band gap of 0.2 eV,[5] in contrast with theoretical calculation which expected to be metallic as the bulk.[5]

Bulk 2H TaSe₂ is metallic[5] and so the single monolayer[5] (trilayer Se-Ta-Se in trigonal prismatic symmetry), which is also known as the 1H phase.[5]

Investigating the non-linear refractive index of tantalum diselenide can be pursued preparing atomically thin flakes of TaSe₂ with the liquid phase exfoliation method. Since this technique requires using alcohol, the refractive index of tantalum diselenide can be retrieved through the Kerr's law:[13]

${\displaystyle n=n_{0}+n_{2}I}$

where n₀=1.37 represents the linear refractive index of ethanol,[13] ${\displaystyle n_{2}}$ is the non-linear refractive index[13] of TaSe₂ and ${\displaystyle I}$ is the incident intensity of the laser beam.[13] Using different light wavelengths, in particular λ=532 nm and λ=671 nm, it is possible to measure both ${\displaystyle n_{2}}$ and ${\displaystyle \chi ^{(3)}}$, the third order nonlinear susceptibility.[13]

Both these quantities depend on ${\displaystyle I}$ because the higher the intensity of the laser, the higher the samples are heated up, which results in a variation of the refractive index.[13]

For ${\displaystyle \lambda =532}$ nm, ${\displaystyle n_{2}=8\times 10^{-7}cm^{2}/W}$[13] and ${\displaystyle \chi ^{(3)}={1.37}\times 10^{-7}(e.s.u)}$[13]

For ${\displaystyle \lambda =671}$ nm, ${\displaystyle n_{2}={3.3}\times 10^{-7}cm^{2}/W}$[13] and ${\displaystyle \chi ^{(3)}={1.58}\times 10^{-7}(e.s.u)}$[13]

Bulk 2H TaSe₂ has been demonstrated to be superconductive below a temperature of 0.14 K.[9] However, the single monolayer (1H phase) can be associated with a critical temperature increased by an increment that can range up to 1 K.[9]

Despite the 1T phase typically does not show any superconductive behaviour,[14] formation of TaSe_2−xTe_x compound is possible through doping with tellurium atoms. The former compound superconductive character depends on the fraction of tellurium (x can varies in the range ${\displaystyle 0<x<2}$ ).[14] The superconductive state arises when the fraction of Te ranges within ${\displaystyle 0.5<x<1.3}$:[14] the optimal configuration is achieved at ${\displaystyle x=0.6}$[14] and in correspondence of a critical temperature ${\displaystyle T_{c}=1.6}$ K.[14] In the optimal configuration, the CDW is totally suppressed by the presence of tellurium.[14]

Opposite to MoS₂, which is largely employed as a lubricant in many different mechanical application, TaSe₂ has not shown the same properties, with an average friction coefficient of 0.15.[15] Under friction tests, like the Barker pendulum, it shows an initial friction coefficient of 0.2-0.3,[15] which quickly increases to larger values as the number of oscillations of the pendulum increases (while for MoS₂ it is almost constant during all the oscillations.)[15]

In general, TMDs can be synthesized through a chemical vapor transport technique accordingly to the following chemical equation:[16]

${\displaystyle {\Bigl (}{\frac {n-1}{n}}{\Bigr )}M+{\frac {1}{n}}MCl_{5}+2X\longrightarrow MX_{2}+{\frac {5}{2n}}Cl_{2}}$

where ${\displaystyle M}$ is the chosen transition metal (Ta, Mo, etc.) and ${\displaystyle X}$ represents the chosen chalcogen element (Se, Te, S etc.). The parameter ${\displaystyle n}$, which governs the crystal growth, can vary between 3 and 50, and can be selected appropriately so that the crystal growth is optimized.[16] During such growth, which might last for 2 – 7 days, the temperature is initially increased within a range between ${\displaystyle T_{h}=}$600 °C - 900 °C.[16] Then, it is cooled down to ${\displaystyle T_{c}=}$ 530 °C - 800 °C.[16] After the growth completion, the crystals are cooled down to room temperature. Depending on the value of ${\displaystyle T_{c}}$, either the 2H or the 1T phase can be stabilized: in particular, using tantalum and selenium with ${\displaystyle T_{c}<}$ 800 °C, only the 2H phase is stabilized. For the 1T phase, ${\displaystyle T_{c}}$ must be larger.[16] This allows to selectively grow the desirable phase of the chosen TMD.[16]

Using powder of TaCl₅ and Se as precursors, and a gold substrate, the 2H phase can be stabilized. The gold substrate has to be heated up to 930 °C, while TaCl₅ and Se can be heated to 650 °C and 300 °C, respectively.[17] Argon and hydrogen gases are used as carriers. Once the growth is complete, the sample is cooled down to room temperature.

Since the single trilayers are kept together only by weak Van der Waals forces, atomically thin layers of tantalum diselenide can be easily separated by using scotch/carbon tape on the bulk TaSe₂ crystals.[18] With this method it is possible to isolate few layers (or even a single layer) of TaSe₂.[18] Then, the isolated layers can be deposited above other substrates, such as SiO₂,[18] for further characterizations.

Pure tantalum is directly sublimated on a bilayer of graphene inside a selenium atmosphere.[5] Depending on the temperature of the substrate ${\displaystyle T_{s}}$ (graphene bilayer), the 1T or the 2H phase can be stabilized: in particular, if ${\displaystyle T_{s}=450}$ °C the 2H is favoured, while at ${\displaystyle T_{s}=560}$ °C the 1T is stabilized.[5] This growth method is suitable only for atomically thin/few layers, but not for bulk crystals.

Bulk crystals of TaSe₂ (or any other TMDs) are put in a solution of pure alcohol. The mixture is then sonicated in an ultrasonic device with a power of at least 450 W for 15 hours.[13] In this way it is possible to overcome the Van der Waals forces that keep the single monolayers of TaSe₂ together, resulting in the formation of atomically thin flakes of tantalum diselenide.[13]

Since 2H TaSe₂ has been found to feature very large optical absorption and emission of light at approximately 532 nm,[19] it might be used for the development of new devices. In particular, the possibility of transferring energy between TaSe₂ and other TMDs, especially MoS₂, has been proved. This process can be accomplished in a non-radiative resonant way by exploiting the large coupling between the TaSe₂ emission and the excitonic absorption of TMDs.[19]

Moreover, it is a promising material that may be used for the injection of hot carriers in semiconducting materials and other non-metallic TMDs[19] due to the high lifetime of the generated photoelectrons.[19]

Exploiting the dependence of the non linear effects of TaSe₂ by the intensity ${\displaystyle I}$ of the incident laser beam,[13] it is possible to build an all-optical switch by means of two lasers which operate at different wavelengths and intensities. In particular, a high intensity laser at ${\displaystyle \lambda _{1}=671}$ nm is used to modulate a low intensity signal at ${\displaystyle \lambda _{2}=532}$ nm.[13] Since there is a minimum value of ${\displaystyle I}$ in order to trigger the non-linear effects, the low intensity signal cannot excite alone.[13] On the contrary, when the high intensity beam (${\displaystyle \lambda _{1}}$) is coupled with the low intensity signal (${\displaystyle \lambda _{2}}$), non-linear effects at both ${\displaystyle \lambda _{1,2}}$ arise.[13] So, it is possible to trigger the non-linear effects on the low intensity signal (${\displaystyle \lambda _{2}}$) by operating on the high intensity one (${\displaystyle \lambda _{1}}$).

Exploiting the coupling between ${\displaystyle \lambda _{1}}$ and ${\displaystyle \lambda _{2}}$ enables transferring information from the high intensity beam to the low intensity one. With this method, the delay time for transferring the information from ${\displaystyle \lambda _{1}}$ to ${\displaystyle \lambda _{2}}$ is around 0.6 s[13]

Usually spin-orbit torque and spin to charge devices are built by interfacing a ferromagnetic layer with a bulk heavy transition metal, such as platinum.[20] However, these effects take mainly place at the interface rather than in the platinum bulk, which introduces heat dissipation due to ohmic losses.[20] Theoretical and DFT simulations suggest that interfacing a 1T TaSe₂ monolayer with cobalt might lead to higher performances with respect to the usual platinum-based devices.[20]

Recent experiments showed that the spin-orbit scattering length of TaSe₂ is around ${\displaystyle L_{so}=17}$ nm,[21] which is highly comparable with the one of platinum, ${\displaystyle L_{so}=12}$ nm.[21] This suggests the possible implementation of tantalum diselenide for the development of new 2D spintronic devices based on the Spin Hall effect.[21]

DFT and AIMD simulations suggest that the stacking of flakes of both TaSe₂ and TaS₂ in a disordered way could be used for the development of a new efficient and cheaper cathode that might be used for the extraction of H₂ from other chemical compounds.[22]