Article | Open | Published:

# Thickness-modulated metal-to-semiconductor transformation in a transition metal dichalcogenide

## Abstract

The possibility of tailoring physical properties by changing the number of layers in van der Waals crystals is one of the driving forces behind the emergence of two-dimensional materials. One example is bulk MoS2, which changes from an indirect gap semiconductor to a direct bandgap semiconductor in the monolayer form. Here, we show a much bigger tuning range with a complete switching from a metal to a semiconductor in atomically thin PtSe2 as its thickness is reduced. Crystals with a thickness of ~13 nm show metallic behavior with a contact resistance as low as 70 Ω·µm. As they are thinned down to 2.5 nm and below, we observe semiconducting behavior. In such thin crystals, we demonstrate ambipolar transport with a bandgap smaller than 2.2 eV and an on/off ratio of ~105. Our results demonstrate that PtSe2 possesses an unusual behavior among 2D materials, enabling novel applications in nano and optoelectronics.

## Introduction

Two-dimensional crystalline materials such as graphene, transition metal dichalcogenides (TMDCs), and black phosphorus (BP) have attracted considerable attention in recent years, due to a set of properties that make them relevant for both fundamental science and technological applications1,2,3. Due to the interaction between atomic layers, these materials exhibit dramatically different properties compared to their bulk forms4,5,6,7,8,9. While the field was opened with the discovery of graphene10, the presence of a bandgap in TMDCs has made this class of materials appealing for applications in electronics11. A key enabler in the development of two-dimensional electronics is the ability to control the properties of the constituent materials, in order to modulate or complement their natural characteristics. Among these, the bandgap is a crucial one, governing both optical and transport properties. For example, a configurable bandgap can be used to tune the adsorption spectrum to specific wavelengths12, while the bandgap is also an important factor determining the on-state current and on/off ratio in field-effect devices. Due to their atomic scale thickness, van der Waals materials are susceptible to several methods of bandgap modulation, involving strain, pressure, chemical or substitutional doping, and application of electrical fields13,14,15,16,17,18,19,20,21,22. However, it is difficult to implement most of these methods in devices without increasing their complexity. A stable and simple approach to bandgap engineering of van der Waals materials involves merely changing the number of layers in the crystal. Apart from exfoliation, this method could be exploited on a large scale by engineering the growth process or by selectively thinning the material after deposition. For example, in the case of MoS2 the 1.29 eV indirect bandgap in bulk form becomes a direct bandgap with a 1.8 eV optical resonance in monolayers4,5,6,7. Even more striking is the case of BP, where the bandgap modulation ranges from 0.35 eV in bulk to 1.73 eV in a monolayer23,24,25. However, in these systems we do not see a fundamental change of their semiconducting nature; we only see a change in the magnitude and type (direct vs. indirect) of the bandgap.

A much wider, qualitative change in the nature of electrical conductivity would extend the reach of two-dimensional materials further. It would, for example, allow the same material to be used both as an interconnect and as the semiconducting channel, greatly simplifying the realization of electronic circuits based on two-dimensional materials, as already proposed for PdS2 and hydrogenated graphene26,27. One material that could offer this is PtSe2, a group VIIIc transition metal dichalcogenide. It shares the same layered structure with other MX2 compounds, with each layer of the crystal made of a two-dimensional close-packed array of Pt atoms sandwiched between Se atoms in a 1T structure28. The electronic properties of bulk PtSe2 have been controversial, with experiments suggesting both a semiconductor29 and a semimetal30,31. More recently, it has been shown that in its bulk form PtSe2 is expected to be a semimetal28,32, with the opening of a bandgap in monolayers33,34,35. Recent studies have also shown that this material can be grown, and that it exhibits appreciable transport properties and interesting spin physics, motivating further studies of its transport properties34,36,37.

Here, we fabricate electronic devices based on PtSe2 and study electrical transport properties of crystals ranging in thickness from 14 to 2 nm. They show a transition from a metal to a semiconductor for thicknesses lower than 2.5 nm. Metallic samples show low contact resistance, while in semiconducting samples we observe ambipolar transport and estimate the size of the bandgap by employing electric-double-layer transistors (EDLTs).

## Results

### Device characterization

Ultrathin PtSe2 crystals with 1T-phase (Fig. 1a) were obtained by mechanical exfoliation on doped Si wafers covered with 270 nm of SiO2. AFM imaging was used to precisely measure the thickness of the resulting crystals, followed by evaporation of Pd contacts to obtain back-gated devices such as the one shown in Fig. 1b. Depending on the crystal size and geometry, simple two-terminal devices, as well as devices with variable electrode spacing needed for contact resistance determination using the transfer length method were prepared. Transfer characteristics of the devices were measured in vacuum at different temperatures between 300 and 10 K (see methods). We first discuss the case of the thickest sample (13.5 nm) we characterized, which is representative of our observations for samples thicker than 7 nm. As expected from theoretical predictions of bulk PtSe2 being a semimetal, the device shows a metallic behavior with almost negligible n-type current modulation by the back gate (dIds/dVg~10–10 S) and a perfectly linear bias (IdsVds) dependence, as shown in Fig. 2a, b. We calculate a sheet resistance of 400 Ω/sq at 10 K, which increases with temperature. Using the transfer length method, we extract a contact resistance as low as 70 Ω·µm at 10 K (Fig. 2c), significantly lower than the lowest reported numbers for semiconducting TMDCs with phase-engineered38 or degenerately doped39 contacts, indicating that the metallic PtSe2 could be an interesting material for realizing contacts to two-dimensional semiconductors40.

We now turn our attention to crystals thinner than 7 nm, measuring the conductance as a function of the back-gate voltage. In Fig. 3a, we present a summary of the conductance dependence on the back-gate voltage Vg measured at 10 K for devices with different thicknesses of PtSe2. Starting with an almost gate-independent conductance in bulk samples, indicating efficient screening of the gate voltage by the high density of charge carriers in the metallic sample, in samples with a thickness of 5 nm we clearly start to observe a significant non-linear modulation with the presence of a well-developed off-state and n-type conduction in the thinnest sample we measured (2 nm). This change in electrical transport properties is consistent with the presence of a finite number of states around the Fermi level in bulk PtSe2, which has been predicted by calculations28,34,35. The shrinking density of states with decreasing thickness results in a greater tunability of the conductance with the gate, up to the point where an appreciable bandgap is opened. This transition is summarized in Fig. 3b, where we observe a progressive decrease of the maximal conductance by more than five orders of magnitude compared to our thickest sample, from 2.5 to 15 nS in 2-nm-thick samples. This is accompanied by an increase in the relative current modulation that can be obtained, from ≈1 for the 13.5 nm thick sample to ≈105 (2 nm). The semiconducting behavior in the 2-nm-thick sample can be more clearly seen in Fig. 3c. We measure the modulation of the drain current Ids as a function of back-gate voltage Vg. The device turns off even at room temperature, while the conductivity increases with the temperature (Fig. 3c, inset), as expected from a semiconductor. We have also made devices based on monolayer PtSe2 (see Supplementary Fig. 1), for which no conduction could be detected. This follows the observed trend of the reduction of conductivity with reduced thickness, making few layers the optimal thickness range for transistor-based applications, where a compromise between the high on-state current and on/off ratio are required.

### PtSe2 EDLT characterization and bandgap measurement

To gain further insight into the behavior of PtSe2, we fabricate EDLTs based on the ([EMIM]-[TFSI]) ion gel41,42 and PtSe2 crystals of different thicknesses, with the device schematic shown in Fig. 4a. Because of the higher gate capacitance41, this approach allows us to explore electrical transport in a wider range of electrostatically induced doping, while in the same time we also benefit from a reduced barrier for charge injection at the contacts12. Figure 4b shows the conductance as a function of polymer electrolyte gate voltage for PtSe2 crystals in the 2.5–5 nm thickness range. In agreement with our observations on back-gated devices, the 4 and 5 nm thick samples show three orders of magnitude higher conductivity compared to the thinnest one, but a relative modulation of less than 50%. Together with the absence of an off-state in thicker crystals (also see Supplementary Fig. 2), we now observe that the modulation is indeed ambipolar. The presence of both electron and hole conduction branches allows us to rule out that the high conductivity we observe in thick PtSe2 is due to intrinsic doping.

We move to the characterization of the 2.5-nm-thick PtSe2 crystal, which was used to realize a two-terminal device. The inset of Fig. 4c shows the room-temperature conductance as a function of back-gate voltage before coating the device with the ion gel. Applying a drain-source bias of 0.1 V, we observe a threshold voltage of 1.5 V on the n-side for the PE gate, and can reach on-state conductance exceeding 10 µS. The on/off ratio is higher than 105, and we estimate FE mobility around 3 cm2V−1s−1. For comparison (Fig. 4c, inset), the back-gated operation resulted in an on-state conductance of 36 nS, threshold voltage of 48 V, and a maximum relative modulation around 104.

Since we already observe the presence of a bandgap in the sample, we subsequently cool it down to 280 K to be able to apply a higher gate voltage, thanks to the wider electrochemical window of the polymer electrolyte12. We assume the electrolyte capacitance at this temperature to be ~4 µF/cm2, from previous reports41. As can be seen from Fig. 4d, the device shows clear ambipolar behavior, with an on/off ratio around 105 and subthreshold swings (SS) of 83 and 106 mV/dec for holes and electrons, respectively.

The large electrostatic capacitance in ionic and polymer electrolyte-gated FETs allows us to determine the bandgap43,44,45,46,47,48 of PtSe2 in the semiconducting thickness range. The prerequisite for this is ambipolar operation. The value of the bandgap can then be extracted from the difference between the threshold voltages for electron and hole branches $${\mathrm{\Delta }}V_{{\mathrm{th}}} = |V_{{\mathrm{th}}}^{\mathrm{e}} - V_{{\mathrm{th}}}^{\mathrm{h}}|$$. Using a reference electrode to provide an accurate reading of the voltage drop across the gel Vref (see methods), the energy $$\mathrm{e}{\mathrm{\Delta }}V_{{\mathrm{th}}}$$ is in an ideal case, equal to the energy needed to add an electron to the bottom of the conduction band and to create a hole at the top of the valence band. To apply this technique, we sweep the gate voltage at different source-drain bias levels and extrapolate from a linear fit the two threshold voltages at zero bias for better precision (see Supplementary Fig. 3). We obtain $$V_{{\mathrm{th}}}^{\mathrm{e}} = 0.4{\ \mathrm{V}}$$,$$V_{{\mathrm{th}}}^{\mathrm{h}} = - 1.8{\ \mathrm{V}}$$, which gives a bandgap value of $$E_{\mathrm{G}} = 2.2 \pm 0.1$$ eV (where the error is estimated from the fit at different Vds).

Theoretical calculations of the PtSe2 band structure predict the bandgap of monolayers to be between 1.2 and 2.1 eV34,35 with a significant reduction in bilayers. No calculations for larger thicknesses are available at this time. Interestingly, even if our sample is definitely thicker than a bilayer, it still shows a sizeable bandgap allowing efficient current modulation at room temperature. This difference between the theoretical calculations and our results for the 2.5-nm-thick PtSe2 could be explained by a possible underestimation of the bandgap in DFT calculations. Our samples could also be covered by a disordered “dead layer”, implying that the actual thickness of the active channel is lower than the one inferred by AFM measurements. These points require further future investigation. We would finally like to note that our measured value represents the upper limit for the actual bandgap, since it has been shown that all non-idealities in the EDLT operation such as trap states (see Supplementary Fig. 4) or contact resistance46 can lead to an overestimated bandgap. However, our measurements show that a significant bandgap is already present in few-layer PtSe2.

## Discussion

Our results on PtSe2 highlight a new type of behavior in an atomic scale TMDC material, with room temperature electrical transport characteristics which can be tuned from metallic to semiconducting, allowing both high electrical conductivity and efficient transistor operation to be realized in the same material. By implementing ohmic contacts and boron nitride encapsulation, it would be further possible to access the pristine properties of this material49, where theory predicts mobility of up to ~1890 cm2V−1s−1 at room temperature50. Considering the air-stable nature of PtSe2 even in the monolayer limit (see Supplementary Fig. 1), these observations open new possibilities for the facile realization of atomic scale circuitry based on thickness-modulated TMDCs.

## Methods

### Device fabrication

PtSe2 crystals were obtained by mechanical exfoliation from bulk crystals (HQ Graphene) onto a doped Si substrate with 270 nm of SiO2. The substrate was imaged using an optical microscope (Olympus BX51M) equipped with a color camera. The thickness of selected crystals was confirmed using AFM topography imaging (Asylum Research Cypher). Metallic contacts were prepared using e-beam lithography and e-beam evaporation of Pd (60 nm). For PE gating, special gate electrodes were designed with a surface area much larger than that of the PtSe2 channel, in order to maximize the gate efficiency.

### Measurements

Devices were characterized at different temperatures inside a Janis cryostat under high vacuum (10–6 mbar). Drain currents were measured using a Keithley Sourcemeter 2450, while a Keithley Sourcemeter 2400 was used for the solid and PE gates. The drain voltage was kept at 100 mV for most measurements, while the source was grounded. Voltage drop across the polymer electrolyte was measured using a Keithley Electrometer 6415.

### EDLT operation

For PE gating, we used an ion gel made from an ethyl acetate solution of the triblock copolymer PS-PMMA-PS and the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMIM][TFSI]). The devices were spin coated with the gel (6000 rpm, 1 min) in Ar atmosphere. The voltage drop across the polyelectrolyte was measured by using Pd reference electrodes close to the device, in order to be able to refer the measurements to the voltage effectively acting on the channel, as shown in previous works43,47. As can be seen from the VrefVPE plot (Supplementary Fig. 5), the gate efficiency is around 60%. A small hysteresis can be observed in the dependence of Vref on VPE, as well as from the IdsVPE dependence. However, the Ids versus Vref curves are almost free of hysteresis. This confirms that the hysteretic behavior is due to the gate–PE interface, and not due to the device itself46. The measurements were repeated with two different reference electrodes for increased reliability, yielding almost indistinguishable results.

### Data availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## References

1. 1.

Wang, Q. H., Kalantar-Zadeh, K., Kis, A., Coleman, J. N. & Strano, M. S. Electronics and optoelectronics of two-dimensional transition metal dichalcogenides. Nat. Nanotechnol. 7, 699–712 (2012).

2. 2.

Koppens, F. H. L. et al. Photodetectors based on graphene, other two-dimensional materials and hybrid systems. Nat. Nanotechnol. 9, 780–793 (2014).

3. 3.

Fiori, G. et al. Electronics based on two-dimensional materials. Nat. Nanotechnol. 9, 768–779 (2014).

4. 4.

Lebegue, S. & Eriksson, O. Electronic structure of two-dimensional crystals from ab initio theory. Phys. Rev. B 79, 115409 (2009).

5. 5.

Splendiani, A. et al. Emerging photoluminescence in monolayer MoS2. Nano Lett. 10, 1271–1275 (2010).

6. 6.

Mak, K. F., Lee, C., Hone, J., Shan, J. & Heinz, T. F. Atomically thin MoS2: a new direct-gap semiconductor. Phys. Rev. Lett. 105, 136805 (2010).

7. 7.

Kuc, A., Zibouche, N. & Heine, T. Influence of quantum confinement on the electronic structure of the transition metal sulfide TS2. Phys. Rev. B 83, 245213 (2011).

8. 8.

Liu, K. et al. Evolution of interlayer coupling in twisted molybdenum disulfide bilayers. Nat. Commun. 5, 4966 (2014).

9. 9.

van der Zande, A. M. et al. Tailoring the electronic structure in bilayer molybdenum disulfide via interlayer twist. Nano Lett. 14, 3869–3875 (2014).

10. 10.

Novoselov, K. S. et al. Electric field effect in atomically thin carbon films. Science 306, 666–669 (2004).

11. 11.

Radisavljevic, B., Radenovic, A., Brivio, J., Giacometti, V. & Kis, A. Single-layer MoS2 transistors. Nat. Nanotechnol. 6, 147–150 (2011).

12. 12.

Allain, A. & Kis, A. Electron and hole mobilities in single-layer WSe2. ACS Nano 8, 7180–7185 (2014).

13. 13.

Ramasubramaniam, A., Naveh, D. & Towe, E. Tunable band gaps in bilayer transition-metal dichalcogenides. Phys. Rev. B 84, 205325 (2011).

14. 14.

Liu, Q. et al. Tuning electronic structure of bilayer MoS2 by vertical electric field: a first-principles investigation. J. Phys. Chem. C 116, 21556–21562 (2012).

15. 15.

Mouri, S., Miyauchi, Y. & Matsuda, K. Tunable photoluminescence of monolayer MoS2 via chemical doping. Nano Lett. 13, 5944–5948 (2013).

16. 16.

Conley, H. J. et al. Bandgap engineering of strained monolayer and bilayer MoS2. Nano Lett. 13, 3626–3630 (2013).

17. 17.

Desai, S. B. et al. Strain-induced indirect to direct bandgap transition in multilayer WSe2. Nano Lett. 14, 4592–4597 (2014).

18. 18.

Suh, J. et al. Doping against the native propensity of MoS2: degenerate hole doping by cation substitution. Nano Lett. 14, 6976–6982 (2014).

19. 19.

Manzeli, S., Allain, A., Ghadimi, A. & Kis, A. Piezoresistivity and strain-induced band gap tuning in atomically thin MoS2. Nano Lett. 15, 5330–5335 (2015).

20. 20.

Yu, Y. et al. Giant gating tunability of optical refractive index in transition metal dichalcogenide monolayers. Nano Lett. 17, 3613–3618 (2017).

21. 21.

Deng, B. et al. Efficient electrical control of thin-film black phosphorus bandgap. Nat. Commun. 8, 14474 (2017).

22. 22.

Kang, M. et al. Universal mechanism of band-gap engineering in transition-metal dichalcogenides. Nano Lett. 17, 1610–1615 (2017).

23. 23.

Li, L. et al. Direct observation of the layer-dependent electronic structure in phosphorene. Nat. Nanotechnol. 12, 21–25 (2017).

24. 24.

Carvalho, A. et al. Phosphorene: from theory to applications. Nat. Rev. Mater. 1, 201661 (2016).

25. 25.

Li, L. et al. Black phosphorus field-effect transistors. Nat. Nanotechnol. 9, 372–377 (2014).

26. 26.

Ghorbani-Asl, M., Kuc, A., Miró, P. & Heine, T. A single-material logical junction based on 2D crystal PdS2. Adv. Mater. 28, 853–856 (2016).

27. 27.

Son, J. et al. Hydrogenated monolayer graphene with reversible and tunable wide band gap and its field-effect transistor. Nat. Commun. 7, 13261 (2016).

28. 28.

Guo, G. Y. & Liang, W. Y. The electronic structures of platinum dichalcogenides: PtS2, PtSe2 and PtTe2. J. Phys. C Solid State Phys. 19, 995 (1986).

29. 29.

Hulliger, F. Electrical properties of some nickel-group chalcogenides. J. Phys. Chem. Solids 26, 639–645 (1965).

30. 30.

Soled, S., Wold, A. & Gorochov, O. Crystal growth and characterization of several platinum sulfoselenides. Mater. Res. Bull. 11, 927–932 (1976).

31. 31.

Kliche, G. Far-infrared and X-ray investigations of the mixed platinum dichalcogenides PtS2−xSex, PtSe2−xTex, and PtS2−xTex. J. Solid State Chem. 56, 26–31 (1985).

32. 32.

Dai, D. et al. Trends in the structure and bonding in the layered platinum dioxide and dichalcogenides PtQ2 (Q=O, S, Se, Te). J. Solid State Chem. 173, 114–121 (2003).

33. 33.

Huang, H., Zhou, S. & Duan, W. Type-II Dirac fermions in the PtSe2 class of transition metal dichalcogenides. Phys. Rev. B 94, 121117 (2016).

34. 34.

Wang, Y. et al. Monolayer PtSe2, a new semiconducting transition-metal-dichalcogenide, epitaxially grown by direct selenization of Pt. Nano Lett. 15, 4013–4018 (2015).

35. 35.

Zhuang, H. L. & Hennig, R. G. Computational search for single-layer transition-metal dichalcogenide photocatalysts. J. Phys. Chem. C 117, 20440–20445 (2013).

36. 36.

Zhao, Y. et al. High-electron-mobility and air-stable 2d layered PtSe2 FETs. Adv. Mater. 29, 1604230 (2017).

37. 37.

Yan, M. et al. High quality atomically thin PtSe2 films grown by molecular beam epitaxy. 2D Materials 4, 045015 (2017).

38. 38.

Kappera, R. et al. Phase-engineered low-resistance contacts for ultrathin MoS2 transistors. Nat. Mater. 13, 1128–1134 (2014).

39. 39.

Chuang, H.-J. et al. Low-resistance 2D/2D ohmic contacts: a universal approach to high-performance WSe2, MoS2, and MoSe2 transistors. Nano Lett. 16, 1896–1902 (2016).

40. 40.

Allain, A., Kang, J., Banerjee, K. & Kis, A. Electrical contacts to two-dimensional semiconductors. Nat. Mater. 14, 1195–1205 (2015).

41. 41.

Pu, J. et al. Highly flexible MoS2 thin-film transistors with ion gel dielectrics. Nano Lett. 12, 4013–4017 (2012).

42. 42.

Ovchinnikov, D. et al. Disorder engineering and conductivity dome in ReS2 with electrolyte gating. Nat. Commun. 7, 12391 (2016).

43. 43.

Braga, D., Gutiérrez Lezama, I., Berger, H. & Morpurgo, A. F. Quantitative determination of the band gap of ws2 with ambipolar ionic liquid-gated transistors. Nano Lett. 12, 5218–5223 (2012).

44. 44.

Chang, Y.-H. et al. Monolayer MoSe2 grown by chemical vapor deposition for fast photodetection. ACS Nano 8, 8582–8590 (2014).

45. 45.

Funahashi, K. et al. Large-area WSe2 electric double layer transistors on a plastic substrate. Jpn. J. Appl. Phys. 54, 06FF06 (2015).

46. 46.

Gutiérrez-Lezama, I., Reddy, B. A., Ubrig, N. & Morpurgo, A. F. Electroluminescence from indirect band gap semiconductor ReS2. 2D Materials 3, 045016 (2016).

47. 47.

Huang, J.-K. et al. Large-area synthesis of highly crystalline WSe2 monolayers and device applications. ACS Nano 8, 923–930 (2014).

48. 48.

Jo, S., Ubrig, N., Berger, H., Kuzmenko, A. B. & Morpurgo, A. F. Mono- and bilayer WS2 light-emitting transistors. Nano Lett. 14, 2019–2025 (2014).

49. 49.

Avsar, A. et al. Van der Waals bonded Co/h-BN contacts to ultra-thin Black Phosphorus devices. Nano Lett. 17, 5361–5367 (2017).

50. 50.

Zhang, W., Huang, Z., Zhang, W. & Li, Y. Two-dimensional semiconductors with possible high room temperature mobility. Nano Res. 7, 1731–1737 (2014).

## Acknowledgements

We thank Adrien Allain and Kolyo Marinov for fruitful discussions and help. Devices were fabricated at the EPFL Center for Micro/Nanotechnology (CMi). We acknowledge the help of Zdenek Benes (CMi) with e-beam lithography. We would like to acknowledge support by the European Research Council (ERC, Grant 682332) and Swiss National Science Foundation (Grant 157739). A.K. acknowledges funding from the European Union’s Horizon 2020 research and innovation program under grant agreement no. 696656 (Graphene Flagship).

## Author information

### Affiliations

1. #### Electrical Engineering Institute, École Polytechnique Fédérale de Lausanne (EPFL), 1015, Lausanne, Switzerland

• Alberto Ciarrocchi
• , Ahmet Avsar
• , Dmitry Ovchinnikov
•  & Andras Kis
2. #### Institute of Materials Science and Engineering, École Polytechnique Fédérale de Lausanne (EPFL), 1015, Lausanne, Switzerland

• Alberto Ciarrocchi
• , Ahmet Avsar
• , Dmitry Ovchinnikov
•  & Andras Kis

### Contributions

A.C., A.A. and D.O. fabricated the samples and performed the measurements. A.K. supervised the experiments. A.C. and A.K. worked on the manuscript preparation with input from all authors.

### Competing interests

The authors declare no competing interests.

### Corresponding author

Correspondence to Andras Kis.

## Electronic supplementary material

### DOI

https://doi.org/10.1038/s41467-018-03436-0

• ### 2D Materials: The thick and the thin

• Claire Ashworth

Nature Reviews Materials (2018)