Multi-layered MoS₂ phototransistors as high performance photovoltaic cells and self-powered photodetectors

Abstract

An optoelectronic diode based on a p–n junction is one of the most fundamental device building blocks with extensive applications. Compared with graphene, layered transition-metal dichalcogenides demonstrate promising applications in novel valley-electronics and opto-electronics. Here we reported the fabrication and optoelectronic properties of a single multilayer MoS₂ sheet. Our results indicate that the thin MoS₂ shows a linear transport property while thick MoS₂ shows diode characteristics with well-defined current rectification behavior. We assign that the rectification behavior is due to the formation of a p–n junction in the single multilayer MoS₂ piece. The intrinsic defects in MoS₂ can change the conduction polarity, such as: sulfur vacancies contribute to the n-type behavior while sulfur interstitials and molybdenum vacancies contribute to the p-type conduction. The variation of intrinsic defects and stoichiometry is obvious over the micrometer range in thick MoS₂. The fabricated MoS₂ transistors were assessed under bias and gate voltage modulation when exposed to red, green and UV light under vacuum. The multilayer MoS₂ shows dominant p-type behavior under dark conditions while its shows dominant n-type conduction under light illumination. In addition, this MoS₂ phototransistor shows an evident photovoltaic effect. The open-circuit voltage (V_oc) and short-circuit current (I_sc) are observed to be −0.48 V and 494 nA under red illumination. These results demonstrate the potential application of a single multilayer MoS₂ sheet in optoelectronics, such as light-emitting diodes (LEDs), field-effect photovoltaic cells and photodetectors.

---

1. Introduction

The photovoltaic effect in silicon has opened the era of solar cells.¹ The principle of field-effect solar cells is using a transparent top gate to modulate the potential barrier and electric field in a p–n junction. A multitude of semiconductor materials, such as, TiO₂, CdS/CdTe, CuIn_xGa_1−xSe₂ (CIGS) and perovskites etc, have been investigated for use in photovoltaic cells.^2–4 A photodetector, which may be applied in imaging techniques, signal processing and sensing, is another important optoelectronics device. Among a multitude of various photodetectors, the low dimensional material-based photodetectors are very attractive because of their longer lifetime of photo-generated carriers and higher photosensitivity due to the large surface-to-volume ratio.⁵ The large photocurrent in solar cells can be used to fabricate self-powered photodetectors without the application of an external bias voltage, which can help their application in independent systems.⁶ Next generation photodetectors and photovoltaic devices will require robust candidate materials with characteristics superior to those current ones.

Recently, two-dimension (2D) layered crystallines have attracted a significant amount attentions.^7,8 For example, graphene shows great potential applications in next-generation electronics due to the high carrier mobility and high optical transparency.⁹ However, the low photoresponsivity and poor external quantum efficiency of pristine graphene, which originate from the zero bandgap, fast carrier transfer, low light absorption coefficient and short photo-induced carrier lifetime, may inhibit its practical optoelectronic applications.¹⁰ Fortunately, the disadvantages of graphene can be overcome by 2D layered semiconductors with finite band gap and relative high carrier mobility. These 2D crystals show promising applications in novel electronics, valley-electronics and optoelectronics.^11,12 Among these 2D semiconductors, molybdenum disulfide (MoS₂), a member of transition metal dichalcogenides (TMDCs) family, has received particular attention.¹³ MoS₂ field-effect transistors (FETs) with high mobility (200 cm² V⁻¹ s⁻¹) and on/off ratio of 10⁸ have been recently reported.¹⁴ Andras Kis et al. demonstrated ultrasensitive monolayer MoS₂ phototransistors, which showed a photoresponse in the 400–680 nm range and a maximum external photoresponsivity of 880 A W⁻¹ at wavelength of 561 nm.¹⁵ The complexes of monolayer MoS₂@NPG (nanoporous gold), MoS₂/graphene, TiO₂/MoS₂/graphene and nMoS₂/CdS (n ≥ 1) electrode exhibits excellent high catalytic activities, which can be used for high efficiency electrochemical and photoelectrochemical hydrogen production.^16–20 In addition, MoS₂ films can be used as flexible electrodes for supercapacitor devices²¹ or as an anode materials for lithium ion batteries.^22,23 By modified with Au nanostructures, few layer MoS₂ nanosheet phototransistor or photoelectrochemical cell showed enhanced photodetection and photocatalytic water splitting.^24,25 The thermal stability up to 1100 °C also makes MoS₂ suitable for device application. The more remarkable properties of MoS₂ lie in its band-gap and band-structure, which depend on the thickness (number of atomic layer) strongly. Specially, MoS₂ shows crossover from indirect bandgap (1.3 eV) in bulk to direct bandgap (1.9 eV) in monolayer, as a result of quantum confinement.^26,27

Compared with single layer MoS₂, multi-layer MoS₂ has also been extensively studied and showed interesting properties because the multilayer MoS₂ can carry higher drive current due to its lower band gap, multiple conduction channels and triple of density of states at the conduction band minimum.^28,29 The specific thickness of MoS₂ can be adjusted to optimize the electrical properties. Such as, both n-type and p-type conductivities have been reported in different MoS₂ thin films deposited on SiO₂ while bulk MoS₂ almost shows n-type character.^30,31 This current polarity (p- or n-type) modulation by number of atomic layer is benefit for the assembly of p–n junction, which is useful particularly for optoelectronic devices. The formation of p–n junctions in MoS₂ has been reported recently by substrate and dielectric modulation, chemical doping, electrostatic doping, plasma treatment, ionic liquid gating and electric double layer gating.^32–37 Contact barrier engineering also play an important role on the creation of p–n diodes in TMDC based layered materials.^38,39 The polarity is governed by the Schottky barrier (SB) height and width for electrons and holes at the source electrode. A low SB height to the conduction band yields electron injection into the channel, thereby resulting in n-type device operation. Similarly, a low SB height to the valence band yields p-type characteristics. In this paper, we demonstrate the formation of p–n junction in single multilayer MoS₂ due to the defect dominated doping. Our device shows a well-defined current rectification behavior. The fabricated MoS₂ transistors were assessed under bias and gate voltage modulation with exposed to red, green, UV light or dark condition at vacuum. This multilayer MoS₂ diode shows different photo-response under various light irradiation. The photoresponsivity of MoS₂ photodetector is about 107 A W⁻¹ under red irradiation operating at zero bias voltage and could be modulated by the back gate voltage. The optical gain and response time are determined to be 2.05 × 10⁴% and 260 ms. This MoS₂ phototransistor shows evident photovoltaic effect. The open-circuit voltage (V_oc) and short-circuit current (I_sc) are observed to be −0.48 V and 494 nA under red illumination. We suggest a qualitative mechanism based on p–n junction energy band diagrams to explain these phenomena. These results demonstrate a potential application of the multilayer MoS₂ in optoelectronics, such as rectifiers, light-emitting diodes (LEDs), photodetectors, and photovoltaic cells.

2. Experimental details

The multilayer MoS₂ films were exfoliated on SiO₂ (300 nm)/Si substrate from bulk MoS₂ single crystal (3–5 mm in size obtained from SPI supplies) by using scotch-tape based mechanical exfoliation method. Optical microscopy (OM) and scanning electron microscopy (SEM) were used to observe the morphology of exfoliated MoS₂ and fabricated devices. The thickness of multilayer MoS₂ was determined by atomic force microscopy (AFM). The multilayer MoS₂ nanodevices were fabricated by following lithography procedure, thermal evaporation and lift-off process. The thickness of contact electrode Au film was about 120 nm. The room-temperature electrical transport measurement was carried out with keithley source meter 2602 (Keithley Instruments Inc.). To measure the photovoltaic effect and photoresponse property, the whole nanodevice was exposed to red, green and UV light (2 mW cm⁻²) perpendicularly at room temperature.

3. Results and discussion

Fig. 1a is a typical optical microscopy image of multilayer MoS₂. The strong color contrast between sample and substrate demonstrates the multi-layers in thickness. Fig. 1b is SEM image of fabricated multilayer MoS₂ device. The channel thickness of device is determined by AFM. According to the AFM image and corresponding height profile (Fig. 1c), the channel thickness is about 164 nm.

Fig. 1 (a) Optical microscopy image of single multilayer MoS₂. (b) Scanning electron microscopy of multilayer MoS₂ device. (c) The corresponding atomic force microscopy and height profile (Inset).

Fig. 2 shows the current–voltage curves (I_ds–V_ds) of multilayer MoS₂ devices with different thickness in atmosphere under dark condition. All these devices were fabricated with the same configuration and parameter. Our results demonstrate that the multilayer MoS₂ devices with thickness of 150–165 nm show large rectification behavior (Fig. 2d) while the MoS₂ devices with reduced thickness in the range of 70–120 nm show small rectification (Fig. 2b and c) or even linear behavior (Fig. 2a). The rectification ratio is about 7.2 × 10² at ±1 V (Fig. 2d). This rectification characteristic is due to the formation of p–n junction within MoS₂, and is not due to the Au/MoS₂ Schottky barriers. This is reasonable because of the thickness dependent band structure and defect density in MoS₂. The linear I_ds–V_ds (Fig. 2a) indicates the Au/MoS₂ contact can be ohmic, although the difference between the work function of Au and the electron affinity of MoS₂ is large.¹⁴

Fig. 2 (a–d) I_ds–V_ds of MoS₂ devices with different thickness.

Generally, unintentional doping by adsorbates can modulate the conductivity of MoS₂ across from n-type to p-type and contribute to the formation of p–n junction.⁴⁰ We execute the I_ds–V_ds measurement in vacuum to exclude the influence of adsorbates. The I_ds–V_ds in vacuum (black curve in Fig. 3a) shows almost same with that at atmosphere (red curve in Fig. 3a), which demonstrates the surface adsorbates in MoS₂ play a minor role on the formation of p–n junction. McDonnell S. et al. also observed the formation of p–n junction within single piece MoS₂ (SPI supplies) contacted with symmetrical Au electrodes.⁴¹ They assigned that intrinsic defects as well as defect-chemistry-related variations in MoS₂ dominate the metal/MoS₂ contact resistance. The MoS₂ with stoichiometry of 1.8 : 1 (S/Mo) tends to show n-type behavior (sulfur vacancies) while with stoichiometry of 2.3 : 1 (sulfur interstitials or molybdenum vacancies) tends to show p-type conduction. The variation in the electronic behavior can even occur obviously over nanometer range. In fact, the MoS₂ rectifying diodes formed by electrostatic or plasma doping may also due to the introduction of asymmetric carrier density or defect state during the gate modulation or plasma treatment process.^33,34 We measure the energy dispersive spectroscope (EDS) of multilayer MoS₂ randomly (Fig. S1†), which demonstrate clearly that the stoichiometry changes from 1.6 to 2.4 and the exfoliated multilayer MoS₂ is not uniform. Therefore, our micrometer-sized metal contacts will inevitably contact both the perfect and defective MoS₂. The combination of metal work function and defect-like work function cause asymmetrical contact and the observed rectification behavior. The reverse dark current increases exponentially following the equation I ∼ exp(αV), which is usually observed in the p–n diodes due to the recombination tunneling mechanism.^42,43 The constant α was evaluated to be −3.76 V⁻¹ by fitting the experimental data. Fig. 3b and c are output characteristics at atmosphere and vacuum. The current increases slightly under positive V_gs while increases greatly at negative V_gs, indicating that this multilayer MoS₂ device is mainly p-type FET. Generally, MoS₂ films on SiO₂/Si substrates are heavily n-doped by defects, charged impurities and adsorbates.⁴⁴ The present p-type conduction may come from the molybdenum vacancies⁴¹ and increased thickness.³² The current increases greater in vacuum (Fig. 3c) than that at atmosphere under negative V_gs, demonstrating the p-type conduction is more obvious and there are less n-type adsorbates dopant. To investigate the more intrinsic electrical transport properties, all the following experiments are measured in vacuum. Fig. 3d is transfer curves under positive and negative bias voltage, which confirm the dominant p-type transport behavior. The Fermi level approaches to the valance band. An apparent ambipolar behavior is observed under negative bias voltage although the p-type conduction is stronger than the n-type. The charge neutral point voltage (V_cnp) is about 7 V at V_ds = −0.5 V (Fig. S2†), indicating the multilayer MoS₂ is p-doped.

Fig. 3 (a) I_ds–V_ds of multi-layer MoS₂ device under dark condition at atmosphere and vacuum. (b and c) Output curves of MoS₂ device at atmosphere and vacuum. (d) Transfer curves of MoS₂ device under different bias voltages.

Fig. 4a shows the I_ds–V_ds curves of multilayer MoS₂ device under dark condition as well as light illumination with different wavelength. Compared with dark condition, the I_ds–V_ds curves under illumination show an evident feature that there is large I_ds at zero bias voltage, which is typical photovoltaic effect. The generation of photovoltaic current is due to the opposite movement of free electrons and holes, which come from the separation of photogenerated electron–hole pairs by the built-in electric field formed at the p–n junction. The photocurrent increase from UV to green, and red, which is consist with the photoresponse of MoS₂ nanomembrane.⁴⁵ The important parameters characterizing the photovoltaic effect, such as: open-circuit voltage (V_oc) and short-circuit current (I_sc) are observed to be −0.48 V and 494 nA under red illumination. The large I_sc would due to the high absorption coefficient of incident light in multilayer MoS₂. The large V_oc is due to the increasing quasi-Fermi level energy difference between electrons and holes under illumination. Although the diode like I–V characteristics is observed in thick MoS₂ devices, they show a weak photovoltaic response (i.e., V_oc: 0.1–0.15 V, J_sc: 0.62–4.92 mA cm⁻², FF: 0.21–0.33), which are worse than the present results.⁴⁶ Fig. 4b shows the extracted electrical power (P_el = I_ds × V_ds) from the device as a function of bias voltage. The maximum power can reach 116 nw. The voltage V_m and current I_m at the maximum photo-generated power are determined to be −0.34 V and 337 nA. In practical application of photovoltaic effect, a “square” like I–V profile, which is quantified by fill factor (FF), is desirable. The FF is defined as the ratio V_mI_m/V_ocI_sc. As seen from Fig. 4a, the I–V profiles approach “square” like and the FF is found to reasonably as high as 0.48 under red illumination. Fig. S3† shows the I_ds–V_ds and electrical power (P_el) of multi-layer MoS₂ device under illumination of a commercial white light LED with 1 W power, which also demonstrates evident photovoltaic effect. This large photocurrent at zero bias voltage can be used to fabricate self powered photodetector. Fig. 4c shows the time-dependent photoresponse of multilayer MoS₂ device with the UV, green, red light switched on and off at zero bias voltage. The photocurrent reaches rapidly to a saturation value under illumination. When the light is turned off, the photocurrent decreases rapidly. The “on” and “off” states for tens of cycles almost keep the same current level, indicating the excellent reversibility and stability of the self powered photodetector. The analysis from a magnified photoresponse process containing one rise and one reset (Fig. 4d) confirms the fast response speed of the device. The photocurrent rise time and reset time are key factors in determining the photodetector's sensitivity to a fast varying optical signal. The rise time was defined as the time needed to reach 90% of the photocurrent from dark current value after light illumination and the reset time was defined as the time needed to reach 10% of the photocurrent after switching off the light illumination. The present rising and reset time are less than 260 ms and 267 ms (limited by the time response rate of the measurement apparatus), respectively. The responsivity (R_λ) and optical gain (G) are another two critical parameters for photodetector. The G is related to the number of electron–hole pairs excited by one absorbed photon. R_λ is defined as the photocurrent generated per unit power of incident light on the effective area of the photodetector. R_λ and G can be calculated according to the following equations: R_λ = ΔI_λ/(P_λS), G = hcR_λ/(eλ), where ΔI_λ = I_λ − I_dark, P_λ is the incident light intensity, S is the effective illuminated area, h is Planck's constant, c is the light velocity, e is the electronic charge and λ is the incident light wavelength.⁴⁷ In the present nanodevice, the R_λ and G were estimated to be about 107 A W⁻¹ and 2.05 × 10⁴%, respectively.

Fig. 4 (a) I_ds–V_ds of multi-layer MoS₂ device under dark and illumination of UV (405 nm), green (532 nm), red (650 nm). (b) Electrical power P_el = I_ds × V_ds as a function of V_ds. (c) Time-resolved current of multi-layer MoS₂ device under illumination of UV, green, red at V_ds = 0 V. (d) Magnified portions of one rise and reset process of the self-powered photodetector.

Fig. 5a is transfer curve of multilayer MoS₂ device under UV illumination at V_ds = 0 V, which shows dominant electron conductivity and indicates the n-type transport behavior of photo-generated carriers. Inset is corresponding I_gs–V_gs. The leakage current is less than 0.5 nA even at V_gs = −40 V. This phototransistor shows an ambipolar behavior at 0.01 V bias voltage (Fig. 5b). Compared with the zero bias voltage, the positive V_ds is benefit for the hole tunneling due to the high electric field along the channel caused by bias voltage. The V_g(th) at V_ds = 0 V and 0.01 V shows a shift from −23 V to −18 V. Such right shift demonstrates p-type doping or improvement of p-channel conduction under positive bias voltage. Compared with dark condition, the V_g(th) shifts to more negative value under UV illumination, indicating the photoexcited holes are injected into external circuit while the photoexcited electrons are trapped in the MoS₂. Min Sup Choi et al. also demonstrated that the MoS₂ acts as an effective charge-trapping layer.⁴⁸ The transfer of holes is reasonable since the MoS₂ shows p-type conduction at dark condition. The negative charges in MoS₂ exclude the accumulation of electrons and thus higher gate voltage is needed to obtain the chare neutrality point, which is consist with the photo-gating effect of grapheme-MoS₂ hybrid phototransistor.⁴⁹ We further measure the transport properties of MoS₂ at negative V_ds and illumination (Fig. 5c). The current at reverse V_ds is opposite with the photocurrent and the total current is the sum of bias current and photocurrent. At low negative bias (such as −0.02 V), the transfer curve only shows n-type characteristic due to the dominant influence of illumination. The increasing reverse bias voltage can result in ambipolar behavior (Fig. 3d). On increasing the V_ds to −0.04 and −0.05 V, the reverse bias current increases greatly (Fig. 4a). At negative V_gs, the bias current shows p-type conduction and the increasing of bias current is larger than the photocurrent. Therefore, the reverse bias contributes more importance role to the transfer behavior. At positive V_gs, both the reverse current and photocurrent increase with the gate voltage. However, the contribution of reverse bias current is larger than the photocurrent at V_gs > 16 V and V_ds = −0.05 V (Fig. 5c). The role of reverse bias voltage is more evident when it is increased to −0.33 V (Fig. 5d), which shows typical ambipolar behavior and is similar with the dark condition. Therefore, the electrical properties of MoS₂ is dominated by photocurrent at positive bias (off state) because of the low tunneling probability while is dominated by bias current at negative bias (especially for large bias) due to the low density of states of photo-generated carriers under UV illumination (Fig. 4a).

Fig. 5 The modulation of gate voltage (V_gs) and bias voltage (V_ds) on the transport behavior under UV illumination: (a) I_ds–V_gs at V_ds = 0 V. (b) I_ds–V_gs at V_ds = 0.01 V. (c) I_ds–V_gs at V_ds = −0.02 V, −0.04 V, −0.05 V. (d) I_ds–V_gs at V_ds = −0.33 V.

Fig. 6a is transfer characteristic of multilayer MoS₂ device under green and red illumination at V_ds = 0 V. Inset is enlarged portion of the p-type conduction at large negative gate voltage. Compared with the transfer curve under UV illumination, the multilayer MoS₂ shows weaker p-type characteristic and more dominant n-type polarity under green and red illumination, which is reasonable since green and red illumination can produce more photo-generated carriers. In addition, the V_g(th) shifts to more negative value (−30 V for red, −24 V for green and −22 V for UV), which indicates that more photo-generated electrons are trapped in MoS₂ and higher gate voltage is needed to obtain the chare neutrality point. Fig. 6b shows the transfer behavior under negative V_ds (−0.1 V). At negative V_gs, the increasing of bias current is larger than the increasing of photocurrent and the role of bias current is dominant over the photocurrent, which is similar with the UV illumination. However, the contribution of photocurrent is more importance than the reverse bias current at positive V_gs even at large bias (V_ds = −0.2 V, Fig. S4†) due to the large amount of photo-generated carriers under green and red illumination. This situation is contrast with the UV illumination. Under red illumination, the MoS₂ shows unipolar n-type transfer at V_ds = −0.05 V (Fig. S5†). Fig. 6c shows transfer curves at various positive V_ds under green illumination, which show apparent ambipolar behavior. Moreover, compared with UV illumination, the V_g(th) shifts to negative value from −18 V to −20 V at V_ds = 0.01 V under green illumination, which is consist with the fact that green illumination can produce more photo-generated carriers. The positive V_ds can improve the p-type conduction, which is agree with the conclusion obtained from dark and UV illumination. The V_g(th) shift to right from −20 V to −7 V with increasing V_ds from 0.01 V to 0.5 V, indicating the apparent hole doping effect. Similar with green illumination, the positive V_ds enhance the p-type conduction under red illumination (Fig. 6d). The ambipolar behavior is evident under positive V_ds and the V_g(th) shift from −23 V to −12 V with increasing V_ds from 0.1 V to 0.4 V.

Fig. 6 The modulation of V_gs and V_ds on the transport behavior under green and red illumination: (a) I_ds–V_gs at V_ds = 0 V; (b) I_ds–V_gs at V_ds = −0.1 V; (c and d) I_ds–V_gs at different positive V_ds under green and red illumination, respectively.

To understand the influence of bias, gate voltage and illumination, we sketch the energy band diagram of p-MoS₂/n-MoS₂ in Fig. 7, where the conduction band, valance band and bandgap of multilayer MoS₂ were assumed to be 4.5 eV, 5.8 eV and 1.3 eV, respectively.⁴⁴ The Fermi level of p-MoS₂ and n-MoS₂ are set at 5.3 eV and 4.7 eV.⁴⁶ The work function of Au is about 5.1 eV. We assigned that both Au/p-MoS₂ and Au/n-MoS₂ are quasi-ohmic contacts with a very shallow Schottky barrier (Fig. 2).⁴⁶ Fig. 7a shows the band alignment of p-MoS₂ and n-MoS₂ after contact at zero bias voltage. The p–n junction shows a built-in potential (ΔΦ_in) of about 0.6 eV, which contributes to the strong diode-like transport characteristic. Such a rectifying diode effectively suppresses the reverse dark current and increases the shunt resistance of the MoS₂ photovoltaic diode, resulting in the enhancement of fill factor. Fig. 7b is the band alignment of MoS₂ p–n junction at negative bias voltage under modulation of gate voltage and illumination. The applied negative bias reduces the barrier and therefore large current is observed. According to the barrier modulation theory, the junction barrier can be tuned capacitively by gate voltage. When V_gs > 0, the electrons are attracted to the interface of MoS₂/SiO₂ to form an accumulation layer (as indicated by the green dashed line). The Fermi level of n-MoS₂ shifts toward to conduction band, which will increase the electron current greatly. The electrons are repelled from the interface if the negative gate voltage is applied (V_gs < 0). The Fermi level of p-MoS₂ downshifts and approaches more to the valence band, which will enhance the p-type transfer behavior (as indicated by the blue dashed line) as compared to the zero gate voltage. On the contrary, the junction barrier increase and little current is observed when positive bias voltage is applied (Fig. 7c). The gate voltage can modulate the hole current effectively. However, the modulation of electron current is not very obvious due to the large junction barrier. In addition, the carrier transport is strongly affected by illumination environment. MoS₂ is sensitive to light illumination and can absorb incident light efficiently to produce photo-excited electron–hole pairs. These electron–hole pairs are separated to free carriers by the internal build-in electric field in the MoS₂ channel, which would reduce the electron–hole recombination rate and increase the carrier lifetime, resulting in an increase of the free carrier concentration. These free electrons and holes move toward opposite directions (Fig. 7b) to generate photovoltaic current. This movement of photogenerated carriers is also responsible for the increased open-circuit voltage. Due to the native p-type behavior, the photo-generated holes can tunnel easily while photo-generated electrons are trapped to form electron doping. The electron doping shifts the Fermi level to conduction band and then the MoS₂ device shows n-type behavior. Increasing the bias voltage can tune the barrier height to modulate the carrier collection, especially for the hole. Such drain-induced barrier lowering (DIBL) effect is often observed in MoS₂ FET.⁵⁰ The photocurrent increases as the wavelength is changed from 405 nm, 532 nm to 650 nm (Fig. 4a). The 650 nm photons had an energy of 1.91 eV, which corresponds well to the absorption data of MoS₂, attributed to the “A” and “B” exciton transitions.^27,34,51 Compared with photons having larger energy (such as 405 nm and 532 nm), an incident photon with energy comparable to the excitonic transitions is efficiently absorbed with low energy loss and gives rise to larger photocurrent.⁵² In this work, the photocurrent at the source-channel and drain-channel interfaces are canceled each other because the temperature differences between the two interfaces are expected to be similar due to global light illumination. So, the photo-thermoelectric effect⁵³ contributes little to the photocurrent in the present work.

Fig. 7 Energy band diagrams of MoS₂ p–n junction: (a) the band alignment of p-MoS₂ and n-MoS₂ after contact at zero bias voltage. (b and c) The band alignment of MoS₂ p–n junction under modulation of gate voltage and illumination at negative and positive bias voltage, respectively.

Since the gate voltage can be used to tune the transfer behavior, we measure the output characteristic of multilayer MoS₂ photovoltaic cell by gate voltage modulation under red illumination (Fig. 8a). It can be seen that the negative gate voltage can reduce the open-circuit voltage (V_oc) and short-circuit current (I_sc) greatly. The positive gate voltage can improve the I_sc greatly although the V_oc almost keep the same. This enhanced I_sc can improve the electrical power output. Fig. 8b is electrical power extracted from the date in Fig. 8a. The maximum electrical power at V_gs = 40 V is about two times than that at V_gs = 0 V. This gate-tunable property is important for the applications of field-effect photovoltaic cell.

Fig. 8 Optimized photovoltaic and photocurrent detection performance under red illumination: (a) output curves at different V_gs. (b) Corresponding electrical power (P_el) as a function of V_ds.

4. Conclusion

In conclusion, the thick MoS₂ tends to show excellent diode characteristic with well-defined current rectification behavior in our experiments. The rectification ratio is about 7.2 × 10² at ±1 V. However, the thin MoS₂ shows linear transport property, which indicates the contact of Au/MoS₂ can be ohmic. According to the EDS, intrinsic defect and stoichiometry results, we assigned that the rectification characteristic is due to the formation of p–n junction within single MoS₂ piece, and not due to the Au/MoS₂ Schottky barriers. This native formation of p–n junction is superior than the doping treatment, which will introduce inevitably impurity scattering and then result in degradation of transport properties. The current of multilayer MoS₂ transistor increases drastically when exposed to red, green, UV light illumination under vacuum. The multilayer MoS₂ shows dominant p-type transfer behavior under dark condition while shows dominant n-type conduction under light illumination. In addition, This MoS₂ phototransistor shows evident photovoltaic effect. The open-circuit voltage (V_oc) and short-circuit current (I_sc) are observed to be −0.48 V and 494 nA under red illumination. The bias and gate voltage can be used to modulate the photovoltaic and transport properties. The bias voltage can increase the p-type conduction. The present results demonstrate that the thick multilayer MoS₂ can be potentially used as rectifiers, FETs, LEDs, self-powered photodetectors and photovoltaic cells.

Acknowledgements

We thank the NSF of China (term nos 51102091, 91121010, 91233203), Research Fund for the Doctoral Program of Higher Education of China (no. 20114306120003) and Scientific Research Fund of Hunan Provincial Education Department (no. 12C0191) for financial support.

References

1. M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C. Putnam and E. L. Warren, et al., Nat. Mater., 2010, 9, 239.
2. Z. Y. Fan, H. Razavi, J. W. Do, A. Moriwaki, O. Ergen and Y. L. Chueh, et al., Nat. Mater., 2009, 8, 648.
3. W. Wang, Y. W. Su and C. H. Chang, Sol. Energy Mater. Sol. Cells, 2011, 95, 2616.
4. V. W. Bergmann, S. A. L. Weber, F. J. Ramos, M. K. Nazeeruddin, M. Grätzel and D. Li, et al., Nat. Commun., 2014, 5, 5001.
5. L. Peng, L. F. Hu and X. S. Fang, Adv. Mater., 2013, 25, 5321.
6. Q. Yang, Y. Liu, Z. T. Li, Z. Y. Yang, X. Wang and Z. L. Wang, Angew. Chem., 2012, 124, 1.
7. R. Mas-Ballesté, C. Gómez-Navarro, J. Gómez-Herrero and F. Zamora, Nanoscale, 2011, 3, 20.
8. M. S. Xu, T. Liang, M. M. Shi and H. Z. Chen, Chem. Rev., 2013, 113, 3766.
9. A. H. C. Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov and A. K. Geim, Rev. Mod. Phys., 2009, 84, 109.
10. F. Bonaccorso, Z. Sun, T. Hasan and A. C. Ferrari, Nat. Photonics, 2010, 9, 611.
11. H. L. Zeng, J. F. Dai, W. Yao, D. Xiao and X. D. Cui, Nat. Nanotechnol., 2012, 7, 490.
12. Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman and M. S. Strano, Nat. Nanotechnol., 2012, 7, 699.
13. R. Ganatra and Q. Zhang, ACS Nano, 2014, 8, 4074.
14. B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti and A. Kis, Nat. Nanotechnol., 2011, 6, 147.
15. O. Lopez-Sanchez, D. Lembke, M. Kayci, A. Radenovic and A. Kis, Nat. Nanotechnol., 2013, 8, 497.
16. Y. W. Tan, P. Liu, L. Y. Chen, W. T. Cong, Y. Ito and J. H. H. Han, et al., Adv. Mater., 2014, 26, 8023.
17. K. Chang, Z. W. Mei, T. Wang, Q. Kang, S. X. Ouyang and H. H. Ye, ACS Nano, 2014, 8, 7078.
18. Q. J. Xiang, J. G. Yu and M. Jaroniec, J. Am. Chem. Soc., 2012, 134, 6575.
19. X. Li, J. G. Yu, J. X. Low, Y. P. Fang, J. Xiao and X. B. Chen, J. Mater. Chem. A, 2015, 3, 2485.
20. K. Chang, M. Li, T. Wang, S. X. Ouyang, P. Li and L. Q. Liu, et al., Adv. Energy Mater., 2015 DOI:10.1002/aenm.201402279.
21. Y. Yang, H. L. Fei, G. D. Ruan, C. S. Xiang and J. M. Tour, Adv. Mater., 2014, 26, 8163.
22. L. Zhang, H. B. Wu, Y. Yan, X. Wang and X. W. Lou, Energy Environ. Sci., 2014, 7, 3302.
23. Y. C. Liu, Y. P. Zhao, L. F. Jiao and J. Chen, J. Mater. Chem. A, 2014, 2, 13109.
24. J. S. Miao, W. D. Hu, Y. L. Jing, W. J. Luo, L. Liao and A. L. Pan, et al., Small, 2015 DOI:10.1002/smll.201403422.
25. Z. Y. Yin, B. Chen, M. Bosman, X. H. Cao, J. Z. Chen and B. Zheng, et al., Small, 2014, 10, 3537.
26. T. Cheiwchanchamnangij and W. R. L. Lambrecht, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 85, 205302.
27. K. F. Mak, C. Lee, J. Hone, J. Shan and T. Heinz, Phys. Rev. Lett., 2010, 105, 136805.
28. S. Kim, A. Konar, W. S. Hwang, J. H. Lee, J. Lee and J. Yang, et al., Nat. Commun., 2012, 3, 1011.
29. Y. J. Zhang, J. T. Ye, Y. Matsuhashi and Y. Iwasa, Nano Lett., 2012, 12, 1136.
30. Y. H. Lee, X. Q. Zhang, W. J. Zhang, M. T. Chang, C. T. Lin and K. D. Chang, et al., Adv. Mater., 2012, 24, 2320.
31. Y. J. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan and J. Lou, Small, 2012, 8, 966.
32. W. Z. Bao, X. H. Cai, D. Kim, K. Sridhara and M. S. Fuhrer, Appl. Phys. Lett., 2013, 102, 042104.
33. M. K. Chen, H. Nam, S. Wi, L. Ji, X. Ren and L. F. Bian, et al., Appl. Phys. Lett., 2013, 103, 142110.
34. S. Sutar, P. Agnihotri, E. Comfort, T. Taniguchi, K. Watanabe and J. U. Lee, Appl. Phys. Lett., 2014, 104, 122104.
35. L. Zhen, S. W. Chang, C. C. Chen and S. B. Cronin, Nano Res., 2014, 7, 973.
36. Y. J. Zhang, J. T. Ye, Y. Yomogida, T. Takenobu and Y. Iwasa, Nano Lett., 2013, 13, 3023.
37. M. S. Choi, D. Qu, D. Lee, X. C. Liu, K. Watanabe and T. Taniguchi, et al., ACS Nano, 2014, 8, 9332.
38. M. Fontana, T. Deppe, A. K. Boyd, M. Rinzan, A. Y. Liu and M. Paranjape, et al., Sci. Rep., 2013, 3, 1634.
39. S. Chuang, C. Battaglia, A. Azcatl, S. McDonnell, J. S. Kang and X. T. Yin, et al., Nano Lett., 2014, 14, 1337.
40. K. Dolui, I. Rungger and S. Sanvito, Phys. Rev. B: Condens. Matter Mater. Phys., 2013, 87, 165402.
41. S. McDonnell, R. Addou, C. Buie, R. M. Wallace and C. L. Hinkle, ACS Nano, 2014, 8, 2880.
42. D. L. Shao, M. P. Yu, J. Lian and S. Sawyer, Appl. Phys. Lett., 2012, 101, 211103.
43. W. C. Zhou, Y. Zhou, Y. H. Peng, Y. Zhang, Y. L. Yin and D. S. Tang, Nanotechnology, 2014, 25, 445202.
44. Z. Y. Yin, H. Li, L. Jiang, Y. M. Shi, Y. H. Sun and G. Lu, et al., ACS Nano, 2012, 6, 74.
45. M. Shanmugam, C. A. Durcan and B. Yu, Nanoscale, 2012, 4, 7399.
46. S. Wi, H. Kim, M. K. Chen, H. Nam, L. J. Guo and E. Meyhofer, et al., ACS Nano, 2014, 8, 5270.
47. Z. X. Wang, M. Safdar, C. Jiang and J. He, Nano Lett., 2012, 12, 4715.
48. M. S. Choi, G. H. Lee, Y. J. Yu, D. Y. Lee, S. H. Lee and P. Kim, et al., Nat. Commun., 2013, 4, 1624.
49. H. Xu, J. X. Wu, Q. L. Feng, N. N. Mao, C. M. Wang and J. Zhang, Small, 2014, 10, 2300.
50. H. M. Li, D. Y. Lee, M. S. Choi, D. Qu, X. C. Liu and C. H. Ra, et al., Sci. Rep., 2014, 4, 4041.
51. A. Splendiani, L. Sun, Y. B. Zhang, T. S. Li, J. Kim and C. Y. Chim, et al., Nano Lett., 2010, 10, 1271.
52. S. H. Lee, D. Lee, W. S. Hwang, E. Hwang, D. Jena and W. J. Yoo, Appl. Phys. Lett., 2014, 104, 193113.
53. N. Perea-López, Z. Lin, N. R. Pradhan, A. Iñiguez-Rábago, A. L. Elías and A. McCreary, et al., 2D Mater., 2014, 1, 011004.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra05434f

---