Multilayer WSe₂/ZnO heterojunctions for self-powered, broadband, and high-speed photodetectors

Photodetectors, which convert incident photons into electrical signals, play an important role in many applications including biological/medical imaging [1], optical telecommunications [2], environmental monitoring [3], astronomical observations [4], and integrated circuits [5]. Traditional photodetectors require an external bias to produce a detectable photocurrent, limiting their performance in terms of dark current [6], circuit noise [7], and power consumption [8]. On the other hand, self-powered photodetectors (SPPDs) can detect the optical signals without applying an external bias voltage based on three mechanisms: (I) photovoltaic effect in the p–n or the Schottky junction-based photodetectors [9, 10]; (II) pyro-phototronic effect in the pyroelectric-based photodetectors [4]; (III) tribo-phototronic effect in the triboelectric nanogenerators [11]. SPPDs have a wide range of applications such as solar cell technology [9], gas sensors [12, 13], and remote sensing [14].

Recently, SPPDs based on 2D transition metal dichalcogenides (TMDs) and their heterostructures have attracted great attention due to their ultrahigh photoresponsivity [15], fast response [16], and broadband detection [1]. These device properties can be attributed to the high absorption coefficient, suitable bandgap for photovoltaic application, and high electron mobility of TMDs [9, 17]. The p–n junction is the building block of photovoltaic SSPDs [18]. In the p–n junction-based SPPDs, the photo-generated electron/hole carriers can be separated by the built-in electric field at the interface of p-type and n-type 2D materials [19]. The p–n junction can be formed in a single 2D material using chemical doping [20] and double electrostatic gating [21]. However, creating a p–n homojunction with high stability using these methods is challenging and complicated [20, 22]. Alternatively, various p–n heterojunctions based on two different 2D materials, such as WSe₂, and other nanostructures can be fabricated by transfer techniques [23] with desirable band alignments and optoelectronic performance [1]. For instance, Chen et al [24] has reported a SPPD based on WSe₂/GaAs 2D/1D with a responsivity of ∼123 mA W⁻¹ at an incident light wavelength of 532 nm. A SPPD device with a maximum photoresponsivity of ∼100 mA W⁻¹ based on WSe₂/graphene 2D/2D structure has also been reported [25]. In another work, Zhang et al [26] has reported a broadband SPPD based on WSe₂/Si 2D/3D vertical heterojunction with a maximum photoresponsivity of 294 mA W⁻¹.

Zinc oxide (ZnO), an n-type semiconductor, is also a suitable material for UV photodetectors due to its wide band gap, high optical transmission, and fast switching time [27]. High-performance broadband photodetectors can be achieved by combining the photodetection performance of 2D TMD materials and ZnO nanostructures. For example, a 2D/1D heterostructure photodetector based on MoSe₂/ZnO nanorods with a responsivity of 330 mA W⁻¹ at a bias voltage of −2 V has been recently reported [28]. Moreover, a MoSe₂ (n-type)/ZnO (n-type) thin film heterostructure photodetector showed a fast photoresponse of 40 μs [27]. Nonetheless, the photodetection performance of a p–n heterojunction photodetector between a 2D material and thin film ZnO has not been explored yet.

Here, by taking advantage of a built-in electric field formed between p-type 2D WSe₂ and n-type ZnO thin film, a self-powered, broadband, and high-speed photodetector has been developed. To form the vertically stacked WSe₂/ZnO heterostructure, mechanically exfoliated WSe₂ was transferred on the ZnO thin film deposited by the atomic layer deposition (ALD) process. This structure shows a low dark current of ∼22 fA, a maximum photoresponsivity as high as ∼131 mA W⁻¹ and a specific detectivity of ∼3.92 × 10¹⁰ Jones under an incident light wavelength of λ = 300 nm at zero bias. It also shows a high cut-off frequency of ∼300 Hz with response rise/fall times of about ∼496/∼515 μs, making it suitable for high-speed self-powered optoelectronic applications. Under reverse bias, this photodetector demonstrates an ultrahigh steady-state responsivity of ∼7160 mA W⁻¹ and a high detectivity of ∼1.18 × 10¹¹ Jones at −5 V bias voltage. Hence, the p-WSe₂/n-ZnO 2D/2D heterojunction is proposed as an excellent candidate for high-performance, self-powered, and broadband photodetectors.

2.1. Fabrication of the WSe₂/ZnO heterostructures

2.2. Material characterization

2.3. Electrical/optical characterization

A p-type WSe₂ crystal was transferred on the square shaped 40 nm thick n-type ZnO thin film. Two Cr/Au metal electrode contacts were deposited, one on the ZnO thin film and the other on the 2D WSe₂. Figure 1(a) shows the schematic illustration of the WSe₂/ZnO heterostructure on the SiO₂/Si substrate and the optical image of the fabricated WSe₂/ZnO heterostructure is shown in figure 1(b). The contact area between WSe₂ and ZnO was calculated to be ∼637 μm². The scanning electron microscopy (SEM) image of the junction is also shown in the inset of figure 1(b). The energy dispersive x-ray spectroscopy was performed to confirm the elemental composition of ZnO and WSe₂ materials (figure S2). The thickness of multilayer WSe₂ flake and ZnO thin film are ∼55 nm and ∼40 nm, respectively, as determined by atomic force microscopy (AFM) (figure 1(c)). Monolayer WSe₂ has a direct bandgap, which makes it attractive for optoelectronic applications, but its low thickness hinders light absorption [33, 34]. In contrast, although multilayer WSe₂ has an indirect bandgap, the larger thickness can lead to high optical absorption with values approaching unity when coupled with a back reflector, making it suitable for photodetection applications [35]. The current–voltage (I–V) characteristic of the WSe₂/ZnO heterojunction under dark condition is illustrated in figure 1(d), showing a high current rectification ratio, defined as the ratio of current at ±1 V, of ∼10³ without gate bias voltage. The diode rectification behavior can be attributed to the built-in electric field at the WSe₂/ZnO interface [36]. Note that the Cr/Au electrode connected to the ZnO thin film is grounded during the I–V measurements.

Zoom In Zoom Out Reset image size

The strong optical absorption of WSe₂ [37] in the visible-near-infrared ranges and ZnO at UV range [38] and the diode rectifying nature of the WSe₂/ZnO heterojunction device makes it suitable for photodetection applications. The I–V characteristics of the WSe₂/ZnO heterostructure device under dark and light illumination are demonstrated in figure 2(a). The current increases about four orders of magnitude upon light illumination at zero bias voltage (figure 2(a)) [4]. The self-powered collection of the photo-generated carriers at zero bias could be attributed to the strong built-in electric field at the WSe₂/ZnO interface and the photovoltaic effect [39]. The inset of figure 2(a) shows the schematic illustration of the WSe₂/ZnO heterojunction under light illumination. In figure 2(a), the photocurrent of ∼0.01 μA at the light intensity of 1.5 mW cm⁻² and the device area of S = 3480 μm², corresponds to a photocurrent density of ∼300 μA cm^–2 for this device, which is relatively high compared to the previously reported 2D materials-based photodetectors [40, 41]. The small offset in the I–V characteristic at zero voltage may be due to the adsorption of oxygen and water molecules on the device surface [42, 43] or the existence of a parasitic capacitance in parallel with the device resistance [44]. Nevertheless, this deviation does not affect the performance of the device since the current–time (I–t) measurements show a constant dark current of ∼22 fA without deviation. Figure S3 shows that the open-circuit voltage (V_oc) and short-circuit current (I_sc) of the WSe₂/ZnO heterojunction device becomes as high as 350 mV and 15.6 nA, respectively, under UV light (λ = 365 nm) illumination at a light intensity of 4.58 mW cm⁻². A high I_photo/I_dark ratio of 10⁵ was also measured using a light intensity of 4.58 mW cm⁻², which is higher than that of previously reported WSe₂-based heterojunctions [45, 46].

Zoom In Zoom Out Reset image size

To evaluate the self-powered photodetection performance of the WSe₂/ZnO device, the dynamic current–time (I–t) characteristics of the device was measured under monochromatic light illumination at different wavelengths ranging from λ = 300 to 850 nm, each with the same light intensity of 42 μW cm^–2, under zero bias voltage (figure 2(b)). The highest photocurrent of ∼12 pA is achieved at λ = 300 nm, which is in agreement with the light absorption spectrum of WSe₂ and ZnO [37, 38]. The photoresponsivity and external quantum efficiency (EQE) of the WSe₂/ZnO device under illumination over a wide range wavelength of λ = 300–850 nm are shown in figure 2(c). The maximum and minimum responsivities of 45 and 0.67 mA W⁻¹ are obtained at λ = 300 nm and 850 nm, respectively. Also, this photodetector has a responsivity of 17 mA W⁻¹, 6.54 mA W⁻¹, and 1.12 mA W⁻¹ at the wavelengths of 700, 750, and 800 nm. Despite its low responsivities in the 700–850 nm range, it is still classified as a broadband detector due to its ability to detect light across the UV (300–380 nm), visible (380–700 nm), and NIR (700–850 nm) spectra [4, 47]. The maximum EQE of ∼18% was also achieved at λ = 300 nm at the light intensity of 42 μW cm^–2. The I–t characteristics of the WSe₂/ZnO device at λ = 300 nm over several light intensities of ∼2 to ∼42 μW cm^–2 at zero bias are illustrated in figure 2(d). With increasing light intensity from 2 to 42 μW cm⁻², the photocurrent amplitude increases linearly (figure 2(e)), which is consistent with earlier findings [9, 10]. The photoresponsivity (R) and detectivity (D*) of the WSe₂/ZnO heterojunction versus light intensity at λ = 300 nm are shown in figure 2(f). Both responsivity and detectivity decrease as the light intensity increases which may be due to higher carrier recombination at higher light intensities [10]. The maximum photoresponsivity and detectivity are ∼131 mA W⁻¹ and ∼3.92 × 10¹⁰ Jones at the wavelength of λ = 300 nm with a light intensity of ∼2 μW cm⁻² at zero bias. The EQE was also calculated to be ∼54.3% at the lowest light intensity of ∼2 μW cm⁻², which is much higher than the EQE of ∼18% at the highest light intensity of ∼42 μW cm⁻². The observation of the higher EQE at low incident light intensity than at high light intensity has also been shown in WSe₂ photodetectors [48]. The peak in responsivity and EQE at λ = ∼700 nm could be attributed to the excitonic absorption peak in WSe₂ [49, 50]. The dark current of this photodetector device is also remarkably low, ∼22 fA, which may be suppressed by the depletion region at the WSe₂/ZnO interface [51]. The main photodetector parameters including responsivity (R), detectivity (D*), and EQE values were calculated using the equations (1)–(4) [52, 53].

$$\begin{eqnarray}&&{I}_{\mathrm{ph}}={I}_{\mathrm{light}}-{I}_{\mathrm{dark}}\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&R=| {I}_{\mathrm{ph}}/\left(P\times S\right)| \end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&\mathrm{EQE}=\frac{hcR}{e\lambda }\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}&&{D}^{* }=\frac{{S}^{1/2}R}{{\left(2e{I}_{\mathrm{dark}}\right)}^{1/2}},\end{eqnarray} \tag{ 4 }$$

where I_ph, I_dark, and I_light are the photocurrent, dark current and current under illumination, respectively; P is the light intensity and S is the effective surface area and h, c, e, and λ are Planck's constant, speed of light, electron charge constant, and wavelength, respectively [52].

To investigate the collection/separation of photo-generated carriers under different reverse voltage conditions, the photocurrent was measured under several reverse bias voltages at a constant light intensity of 2 μW cm⁻² and the wavelength of λ = 300 nm. Figure 3(a) shows the amplitude of the generated photocurrent under reverse bias voltages of 0, −0.5, −1, −2, −3, −5, −7, and −10 V. A schematic illustration of the measurement setup is shown in the inset of figure 3(a). Small reverse bias voltages (e.g. −0.5 V and −1 V) result in a significant increase in photocurrent due to the facilitation of charge collection. The effect of reverse bias on photocurrent nearly saturates at values ≥ −5 V (figure 3(a)) [4, 54]. The maximum photoresponsivity of 11.98 A W⁻¹ was obtained at the voltage bias of −10 V (figure 3(b)). Note that both dark current and photocurrent increase with increasing reverse bias voltage (figure S4). The responsivities of 1.74, 2.99, 4.37, 7.16, and 8.77 A W⁻¹ were achieved for the reverse biases of −1, −2, −3, −5, and −7 V, respectively (figure 3(b)), which are much higher than the responsivity at the zero bias (0.131 A W⁻¹). Interestingly, as the applied reverse bias voltage across the WSe₂/ZnO heterojunction increases, the responsivity increases sublinearly but the detectivity increases up to the peak value of ∼1.39 × 10¹¹ Jones at −2 V and thereafter decreases with further increase in applied reverse bias voltage [55]. The detectivity highly depends on the responsivity and the dark current of the device (equation (4)). At the bias voltages higher than −2 V, the dark current increase rate is surpassing the responsivity growth rate, resulting in a decrease in detectivity. Therefore, the optimum performance of the WSe₂/ZnO photodetector is at a reverse bias voltage of −2 V. The EQE of the device increases significantly to a value of ∼4958% at the bias voltage of −10 V (figure S5). The EQE cannot exceed 100% at zero bias without photomultiplication [4]. With the help of the reverse bias, the EQEs can enhance to values much higher than 100% due to the photoconductivity gain effect [4]. Note that the whole device area (S = 3480 μm²) is considered active for the calculation of the responsivities/detectivities at the non-zero bias voltages.

Zoom In Zoom Out Reset image size

The self-powered frequency-dependent photoresponse of the WSe₂/ZnO photodetector was investigated in figure 4. A 470 nm LED powered by a function generator was used to generate light pulses with different frequencies and the WSe₂/ZnO device was connected to an oscilloscope through a resistor (15 MΩ) to measure the time-dependent photocurrent (figure 4(a)). The normalized I–t characteristics of the device with zero applied voltage under the light frequencies of 10, 300, and 500 Hz are shown in figure (b)–(d), respectively. The normalized photoresponses in frequencies of 50, 150, and 800 Hz are also illustrated in figures S6(a)–(c). The photocurrent amplitude as a function of light frequency is shown in figure S6(d), illustrating a cut-off frequency of ∼300 Hz. The amplitude of the photoresponse at 300 Hz is about ∼0.74 which is about 70% of its maximum value at the frequency of 1 Hz. At higher frequencies, the photocurrent amplitude drops significantly. The cut-off frequency is the frequency at which the amplitude of the photocurrent signal drops by a factor of 70% [4]. The response time and fall times of ∼496 and ∼515 μs were measured for a square wave optical input signal with a frequency of 300 Hz (figure 4(e)), demonstrating the suitability of this device for fast self-powered optoelectronic applications. Figure 4(f) illustrates the stability of the photoresponse of the WSe₂/ZnO photodetector over 5000 consecutive light pulses, showing a square wave photoresponse without a notable degradation in the signal amplitude.

Zoom In Zoom Out Reset image size

The photodetection performance of the recently published 2D WSe₂- and ZnO-based photodetectors is compared to our WSe₂/ZnO heterojunction photodetector in table 1. The WSe₂/ZnO heterojunction photodetector is broadband, highly responsive, fast, and offers high responsivities and detectivities, compared to the previous works listed in table 1. The responsivity of the WSe₂/ZnO device under reverse bias voltages is almost two orders of magnitudes higher than those of other WSe₂- and ZnO-based photodetectors.

To investigate the mechanism of the self-powered photodetection in the WSe₂/ZnO heterojunction, three devices based on the ZnO thin film, WSe₂ crystal, and WSe₂/ZnO heterojunction were fabricated. The schematic of the Cr-ZnO-Cr, Cr-WSe₂-Cr, and Cr-WSe₂/ZnO-Cr structures under light illumination are shown in figures 5(a)–(c), respectively. Figures 5(d)–(f) illustrate the energy band diagram of the devices under light illumination. The electron affinity of ZnO is ∼4.3–4.5 eV [66] and the work function of Cr is ∼4.5 eV [9], resulting in two equal Schottky barriers with the Cr electrodes (figure 5(d)) [67]. These equal Schottky barriers lead to a net photocurrent of about zero (∼5 fA) at zero bias when the device is under light illumination (figure 5(g)) [9, 68]. However, the photocurrent is non-zero when applying a bias voltage to the device due to the separation of carriers by the applied electric field [69]. Similarly, the WSe₂ crystal, with an electron affinity of ∼4 eV [70] and a work function of ∼4.8 eV [71], forms two equal Schottky barriers with Cr electrodes (figure 5(e)) [71], which results in a net photocurrent of about zero (∼0.03 pA) when the whole device is under light illumination (figure 5(h)). This small photocurrent could be due to the unwanted WSe₂/Cr contact area asymmetry at two electrodes [9, 72]. Therefore, the symmetric metal-semiconductor-metal device structure is not suitable for SSPD application [9]. However, when contacting the WSe₂ and ZnO, a built-in electric field is formed at the interface due to the energy band bending, leading to a high current rectification ratio of ∼10³ under dark conditions and a large net photocurrent (∼45 pA) under light illumination with zero bias (figure 5(i)). Therefore, the high V_oc = ∼400 mV and I_sc = ∼45 pA in the WSe₂/ZnO heterojunction is attributed to the presence of a strong built-in electric field at the WSe₂/ZnO interface and the separation/acceleration of electron/hole carriers by the electric field in the depletion region [73]. The ideality factor (n) and barrier height $\left({\phi }_{B}\right)$ of the WSe₂/ZnO diode are calculated to be n = ∼2.83 and ${\phi }_{B}=\sim 0.87\,V$ (Section I, Supporting Information), which are comparable to the previously reported ZnO/Si heterostructure diode [74].

Zoom In Zoom Out Reset image size

One important consideration for SPPDs is large-area scalability. In this work, the WSe₂ flakes were mechanically exfoliated and transferred on the patterned ZnO thin film to demonstrate proof-of-concept WSe₂/ZnO photodetection devices. The ZnO thin film is scalable to large areas by ALD and large area 2D materials can be prepared by chemical vapor deposition (CVD) [75] or wet chemical processes [76]. Therefore, it is possible to scale up this proof-of-concept photodetector to an imaging array of WSe₂/ZnO photodetectors on a chip. Hence, the fast response, broadband detection, high-speed, and high responsivity of the WSe₂/ZnO heterostructure make it an excellent candidate for self-powered optoelectronic applications.

A self-powered broadband photodetector based on p-WSe₂ and n-ZnO thin film was developed. The WSe₂/ZnO p–n heterojunction photodetector showed a broadband response (300–850 nm) with maximum responsivity and detectivity of ∼131 mA W⁻¹ and ∼3.92 × 10¹⁰ Jones under λ = 300 nm at zero bias, respectively. The responsivity and detectivity of the device increase to ∼7160 mA W⁻¹ and ∼1.18 × 10¹¹ Jones, respectively, under a reverse bias voltage of −5 V, which is due to the facilitation of charge collection under bias voltage. The photodetector also exhibited a cut-off frequency of ∼300 Hz, a fast response time of ∼496 μs, and a remarkable photocurrent stability over 5000 light pulse cycles. Hence, the p-WSe₂/n-ZnO based photodetector is a candidate for self-powered, high-speed, and broadband photodetection technology.

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), Canada Foundation for Innovation (CFI), British Columbia Knowledge Development Fund (BCKDF), Western Economic Diversification Canada (WD), and Simon Fraser University. The authors thank B Kim for maintenance of the SFU Engineering Science cleanroom facility. The authors acknowledge CMC Microsystems and 4D LABS shared facilities that facilitated this research.

The data cannot be made publicly available upon publication because they are not available in a format that is sufficiently accessible or reusable by other researchers. The data that support the findings of this study are available upon reasonable request from the authors.