High-Sensitivity 2D MoS₂/1D MWCNT Hybrid Dimensional Heterostructure Photodetector

Abstract

A photodetector based on a hybrid dimensional heterostructure of laterally aligned multiwall carbon nanotubes (MWCNTs) and multilayered MoS₂ was prepared using the micro-nano fixed-point transfer technique. Thanks to the high mobility of carbon nanotubes and the efficient interband absorption of MoS₂, broadband detection from visible to near-infrared (520–1060 nm) was achieved. The test results demonstrate that the MWCNT-MoS2 heterostructure-based photodetector device exhibits an exceptional responsivity, detectivity, and external quantum efficiency. Specifically, the device demonstrated a responsivity of 3.67 × 10³ A/W (λ = 520 nm, V_ds = 1 V) and 718 A/W (λ = 1060 nm, V_ds = 1 V). Moreover, the detectivity (D*) of the device was found to be 1.2 × 10¹⁰ Jones (λ = 520 nm) and 1.5 × 10⁹ Jones (λ = 1060 nm), respectively. The device also demonstrated external quantum efficiency (EQE) values of approximately 8.77 × 10⁵% (λ = 520 nm) and 8.41 × 10⁴% (λ = 1060 nm). This work achieves visible and infrared detection based on mixed-dimensional heterostructures and provides a new option for optoelectronic devices based on low-dimensional materials.

1. Introduction

Low-dimensional semiconductor materials are a class of materials that emerged rapidly in the 20th century with the development of nanoscience and technology [1]. Typical low-dimensional semiconductor materials can be classified by dimensionality into two-dimensional superlattices and quantum wells, one-dimensional nanowire materials, and zero-dimensional quantum dots. In addition, various 1D nanostructures, such as semiconductor nanowires, nanoribbons, nanorods, and nanotubes, have been successfully manufactured. As a result, there has been a global surge in research activity focused on exploring the unique properties and potential applications of these materials [2,3,4,5,6,7]. Numerous studies have demonstrated the potential of these one-dimensional (1D) nanomaterials and structures for applications in nano-optoelectronics and electronic devices [8,9,10,11,12,13,14,15,16]. In recent years, carbon nanotubes (CNTs) have garnered considerable attention from the scientific community, owing to their remarkable intrinsic properties. With their ultrathin mass (1–3 nm), CNTs are seen as an ideal material for use as a channel or active material in nanoelectronics and optical electronics. The material’s ballistic transport, high stability, and mobility further contribute to its suitability for these applications. These unique properties of CNTs have made them a material of great interest for scientists and engineers, who are constantly exploring their potential applications in various fields [17,18,19,20]. Furthermore, how CNTs are convoluted by graphene sheets strongly determines their electronic properties [21]. Carbon nanotubes (CNTs) exhibit unique electrical behavior that can be either metallic or semiconducting depending on their chirality angle or helicity index [22]. Multi-walled carbon nanotubes (MWCNTs) are often characterized by a lack of regularity in terms of tube shape, as well as an unequal distribution of layers and disorderly arrangement. In the realm of carbon-nanotube-based photodetectors, thin films have been the material of choice. However, it is important to note that the orientation, uniformity, and surface defects of these films have a direct impact on the performance of the resultant devices. Therefore, achieving a high degree of control over the morphology and structure of MWCNTs and carbon nanotube thin films is of critical importance to the development of high-performance photodetectors. Two-dimensional (2D) materials offer distinct advantages due to the absence of surface hanging bonds. As a result, the complexification of surface structures can be effectively suppressed, leading to an enhanced stability and reproducibility of their physical and chemical properties. This makes 2D materials highly attractive for a broad range of applications, including electronics, photonics, energy storage, and catalysis. Furthermore, the unique mechanical, optical, and electronic properties of 2D materials also make them highly promising candidates for use in novel nanoscale devices and systems [23,24,25,26]. MoS₂ is one of the most widely studied materials in 2D layer TMDs, and exhibits unique electronic, optical, and mechanical properties due to its atomically thin structure. However, the characteristics of MoS₂ limit visible light absorption, and the limitation of the separation of photogenerated carriers by the significant exciton effect of two-dimensional materials makes parameters such as the optical responsivity of MoS₂ photodetectors more limited, which hinders its development in the direction of high-performance photodetectors. Fortunately, although light absorption is an inherent property of the material, the construction of heterojunctions can effectively complement the light absorption of a single MoS₂ material and enhance the overall light absorption. In addition, the carrier mobility can be effectively improved by constructing heterojunctions to enhance the optical response rate. Thus, it is a promising material for forming heterostructures with CNTs. In recent years, numerous studies have been aimed at enhancing the efficiency of MoS₂ photodetectors by means of a systematic design of van der Waals heterostructures [27,28,29,30].

Through van der Waals force binding, a one-dimensional (1D) and a two-dimensional (2D) material can form unique mixed-dimensional heterostructures that inherit the unique physical properties of van der Waals 2D/2D heterostructures, but also have rich stack configurations, providing new manipulable freedom to adjust the structure and properties of heterostructures [31,32,33,34,35,36]. The versatility of heterostructure fabrication provides researchers with a new level of control over the resulting structure and properties. Such heterostructures have the potential to enable the creation of new and unique electronic and optoelectronic devices. Considering the previous reports of CNT/MoS₂ heterostructures for photodetectors [37,38,39,40], most of them are based on randomly aligned carbon nanotube networks. However, few studies have used high-quality carbon nanotubes with highly directional alignment, most of the detection range is limited to visible light, and no broad spectrum detection from visible to near infrared has been achieved. The implementation of highly aligned carbon nanotubes in heterostructure fabrication could provide significant advantages, such as an improved structural stability and increased device performance. The ability to tailor the structure and properties of these heterostructures could lead to a further optimization of their detection capabilities for specific applications. Furthermore, achieving broad-spectrum detection from visible to near-infrared wavelengths would be highly desirable for many applications, including imaging and sensing technologies. Thus, there is a significant opportunity for further research in the development and optimization of CNT/MoS₂ heterostructures for photodetectors.

This paper reports a photodetector device based on a lateral hybrid vdW heterostructure composed of directionally aligned multi-walled CNT bundles and MoS₂. The introduction of CNTs in the vdW heterostructure reduces the contact area to the nanometer scale. The MWCNT-MoS₂ planar heterostructures were characterized by Raman spectroscopy, Raman mapping, and AFM. In addition, the MWCNT-MoS₂ back-gate FETs have rectification characteristics. The increased light absorption by the heterostructure and the integrated electrical field contribute to improving the photoresponse and detection rate of the photodetector. The device exhibits a distinct photoresponse, and the photocurrent magnitude grows with power density under laser irradiation at 520 nm and 1060 nm, respectively. This work provides the basis for the study of the electrical and optoelectronic properties of hybrid vdW heterostructure devices, which demonstrate the possibility of manufacturing 1D–2D hybrid structures for future nanoelectronics and nanophotonics.

2. Materials and Methods

2.1. Preparation of 1D CNT Bundles

The height of the CVD-grown CNT arrays was approximately 20 μm. The as-obtained material was prepared into a thin film structure by rolling the sample with a smooth roller under certain pressure conditions and then transferring it onto a tape. In this way, it can be transformed into a horizontal alignment from vertical alignment, transferred onto a substrate (Si substrate with 285 nm SiO₂ layer) using a dry method, and then transferred onto a slide affixed with PDMS using a dry method [41,42]. The fabrication process for CNT bundles in a laterally aligned formation is shown in Figure 1.

2.2. Preparation of Thin Layers of Molybdenum Disulfide

First, a block of molybdenum disulfide material with good monocrystalline was pre-pared, a flatter, smoother, glossy piece of material was glued from the block with the help of blue tape, and a thin layer of molybdenum disulfide was obtained by mechanical peeling method, which was then transferred to a soft polydimethylsiloxane (PDMS) substrate on a transparent slide to find the right thickness of material under the microscope.

2.3. Materials Characterization

The SEM images were obtained by ZEISS Sigma 300 Field Emission Scanning Electron Microscope, Raman imaging, and Raman spectra were achieved by LabRAM ODYSSEY using a 532 nm laser as an excitation source. Polarized Raman spectra were obtained at the 514.5 nm laser equipped with an objective of 100× as the excitation source. Material thickness and surface roughness were obtained by AFM. The phase of the MWCNT and MoS₂ materials was determined through ultra-high resolution X-ray diffraction (XRD) measurements using a Bruker-AXS instrument.

2.4. Devices Fabrication and Measurement

The fabrication procedure for lateral vdW heterostructure is shown in Figure 2. A suitable thickness of MoS₂ was transferred to MWCNTs using a micro-nano spot transfer technique [43], and then a standard lithography technique was used to etch a painted electrode pattern onto the target wafer by electron beam exposure (EBL), followed by electron beam deposition of Cr/Au (thickness: 20 nm/80 nm) and a subsequent peeling process to prepare gate FETs based on MoS₂ and MWCNT-MoS₂ heterostructures. The electrical and optoelectronic tests were performed using Lake Shore TTPX probe station and Keithley 4200 semiconductor device parameter analyzer. The incident light had wavelengths of 520 nm and 1060 nm and the spot size was approximately 1 mm². All optoelectronic tests were measured at room temperature in the air.

3. Results and Discussion

The obtained MWCNTS were first characterized by SEM—the SEM images in Figure 3a–c show a highly ordered arrangement of CNTs—and then we tested them with Raman polarization spectroscopy. The polar diagram of the Raman G-peak maximum intensity of the CNT bundles given in Figure 3d similarly confirms the orientation of the CNT that we obtained [44]. To confirm the phases of the vertically grown multi-walled carbon nanotube (MWCNT) arrays and MoS₂ material, X-ray diffraction (XRD) tests were carried out, and the results are shown in Figure 3e,f. The obtained XRD pattern of the CNT conforms to the standard card PDF #22-1069, where the main peaks of the bare CNT appear at 2θ = 23.9° and 43.0°, corresponding to its (201) and (304) planes, respectively. As for the MoS₂ bulk material, the XRD pattern exhibits peaks at 14.5°, 29.3°, 44.5°, 60.5°, and 77.9°, which can be ascribed to the (003), (006), (009), (113), and (0015) planes, respectively, when compared to the standard card PDF #17-0744. Figure 3g shows the image of the obtained heterogeneous structure under scanning electron microscopy (SEM) at different magnifications. Figure 3j shows the AFM characterization of this sample, and the results show that the surface of the sample is relatively homogeneous, where the thickness of MoS₂ is approximately 20 nm for multilayer MoS₂, whereas the thickness of MWCNT is approximately 200 nm.

To demonstrate the construction of the MWCNT-MoS₂ heterostructure, a series of characterizations of the obtained samples were carried out. First, the Raman mapping of the samples is given in Figure 4, where Figure 4a shows the SEM photographs of the test samples and Figure 4b–d show the Raman peak intensity mapping of the samples at 380 cm⁻¹, 404 cm⁻¹, and 1590 cm⁻¹ wave numbers, respectively. The peak intensity of the sample is at the 404 cm⁻¹ wave number (representing the MoS₂ characteristic peak) mapping signal, and the peak of MoS₂ appears in the region and MWCNT-MoS₂ region. There are two peaks located at 380 cm⁻¹ and 404.9 cm⁻¹ (Figure 4e), corresponding to the in-plane E¹_2g and A_1g vibration of MoS₂, respectively, indicating that the MoS₂ is multilayer [45]. In addition, the intensity ratio of the MWCNT Raman G-peak to the D-peak is slightly increased, which is attributed to the formation of the heterostructure. Combined with the SEM image results, the MWCNT-MoS₂ van der Waals heterostructure was successfully constructed in the central region of the sample. In addition, the Raman mapping of the sample has a more uniform color lining, which indicates that the sample has good homogeneity in the two-dimensional plane.

In order to investigate the effect of heterogeneous interfaces on charge transport behavior, gate FETs based on a bare MoS₂ and MWCNT-MoS₂ heterostructure were constructed using standard photolithography techniques. The optical photograph of the device is shown in Figure 5b, where the MWCNT-MoS₂ heterostructure FET is constituted between electrodes 1 and 3 or 2 and 4, and the bare MoS₂ FET is constituted between electrodes 3 and 4.

Figure 5c gives the transfer characteristic curves of the device and Figure 5d shows the corresponding logarithmic curve of its transfer curve. The transfer curves of the device show that, when the source-drain voltage V_ds of the device is negative, the I_ds of the device change very little with an increasing voltage, whereas, when the V_ds of the device is positive, the value of the I_ds of the device increases rapidly with an increasing voltage, which shows that the current rectification characteristic is similar to that of a diode.

To further demonstrate that the rectification characteristics are caused by the p-n heterojunction, the electrical properties of MoS₂ devices were tested. Figure 5e gives the transfer characteristic curves of MoS₂ devices at different gate voltages in the source/drain bias voltage range −1~1 V. The current increases with an increasing V_g and the curves exhibit a typical n-type behavior. In contrast, the transfer curves of heterojunction FETs exhibit typical p-n junction conductivity behavior. Therefore, the rectification characteristics of the CNT-MoS₂ heterojunction FET are due to the p-n heterojunction, since the device forms a p-n heterojunction that leads to forward conductivity and good correction characteristics.

To further determine the optical response of the device, a broad-beam laser with an adjustable power density was used as the light source to measure the current-time (I-T) characteristics under different power density irradiation. We compared the photoresponse variations in the bare MoS₂ photodetector and heterostructure devices and all experiments were performed in a room-temperature environment. The response time (defined as the increase in photocurrent from 10% to 90% of the peak current) is shown in Figure 6d for the bare MoS₂ photodetector, from which, it can be seen that the photoresponse of the bare MoS₂ photodetector is not very significant and the response time is slow, with a rise time of 182 ms and a fall time of 177 ms. R_I is the most commonly used parameter for characterizing the sensitivity of a photodetector, defined as the photocurrent induced per unit power irradiated on the photosensitive surface of the photodetector, and its equation can be expressed as:

R_I = I_ph/P_in ∗ S

(1)

where P_in is the optical power density, S is the effective area of the device under illumination, and I_ph is the corresponding photocurrent. The effective light area of the MWCNT-MoS₂ photodetector and the bare MoS₂ photodetector are approximately 0.6 μm² and 6 μm², respectively. By calculating, the photocurrent responsivity of the MWCNT-MoS₂ and bare MoS₂ photodetector is 183 A/W and 0.11 A/W (P = 2.27 mW/cm², V_ds = 0.01 V), respectively. The responsivity value of the MWCNT-MoS₂ photodetector is much higher than those reported of the photodetectors (

Table 1. Comparison of the responsivity values of MoS₂ and CNT-MoS₂ photodetectors.

Device Structure	Measurement Condition	Responsivity	Reference
Multilayer MoS2	λ = 532 nm, Vds = 1.2 V, P = 1.69 mW/cm2	59 A/W	[46]
s-SWCNT (network)/SL-MoS2	λ = 650 nm, Vds = −5 V, Vg = −40 V, P = 280 μW	>0.1 A/W	[37]
MWCNT (powder)/MoS2 core–shell	λ = 532 nm, Vds = 2 V, P = 1 mW	24 m A/W	[38]
MoS2 (fewer-layer)/SWCNTs network	λ = 532 nm, Vds = 3 V, P = 4 μW/cm2	8 × 103 A/W	[39]
SWCNT/MoS2 (bi- or tri-layer) as-grown heterostructures	λ = 532 nm, Vds = 0.1 V, P = 0.2 × 10−3 mW/cm2	300 A/W	[40]
SWCNTs/multilayer MoS2/ITO	λ = 532 nm, Vds = 1 V, P = 10 nW	2008.3 A/W	[47]
MWCNT (horizontally aligned)/multilayer MoS2	λ = 520 nm, Vds = 1 V, P = 2.27 mW/cm2	3670 A/W	This work
	λ = 1060 nm, Vds = 1 V, P = 10.9 mW/cm2	718 A/W

). The data presented in Table 1 clearly illustrate that the choice of the heterojunction structure, material synthesis method, material thickness, and measurement conditions can significantly impact the magnitude of the R. It is worth noting that the present study employs a method that is both straightforward and practical, enabling the achievement of remarkable R-values in MWCNT-MoS₂ heterostructures.

In our heterostructure device, the maximum photocurrent responsivity is 3.67 × 10³ A/W at a wavelength of 520 nm (P = 2.27 mW/cm², V_ds = 1 V), and the value is much higher at lower powers. D* indicates the ability to detect weak light signals, and can be expressed as:

$$D*=R\sqrt{S/2q{I}_{dark}}$$

(2)

Based on this equation, D* = 1.2 × 10¹⁰ Jones is calculated for a heterostructure device at an optical power density of 2.27 mW/cm² at a bias voltage of 1 V. EQE is the ratio of the number of photogenerated carriers to the number of incident photons, and can be expressed as:

$$EQE=R_{\lambda }\frac{hc}{e · \lambda }\times 100\%$$

(3)

where h is Planck’s constant and e and c are the fundamental charge and the speed of light, respectively. λ is the wavelength of the incident light. Thus, when R_I = 3.67 × 10³ A/W (λ = 520 nm, P = 2.27 mW/cm², V_ds = 1 V), the corresponding EQE is up to 8.77 × 10⁵%, indicating an excellent performance improvement in our MWCNT-MoS₂ photodetector.

By increasing the power of the incident light at 520 nm, the number of photogenerated carriers in the channel is increased correspondingly, so the magnitude of the photocurrent shows a linear enhancement when increasing the incident light power (Figure 7c), which is consistent with the relationship between I_ph and incident power:

I_ph = AP^⍺

(4)

The value of ⍺ obtained by fitting is approximately 0.91, which is very close to the ideal value of 1, indicating that the process is very efficient for electron–hole pair separation and complexation. The optical responsiveness R_I and detection D* of the MWCNT-MoS₂ heterostructure photodetector with laser power density for V_ds = 1 V are given in Figure 7d, and both the optical responsiveness R_I and detection D* decrease with an increasing laser power. By switching the laser periodically, the device can switch between ON and OFF states effectively, and there is no significant current decay, indicating that the MWCNT-MoS₂ photodetector has good stability. From the device response and reset curves for a single cycle given in Figure 7f, the rising-edge response time τ_rise for the detector is 42 ms, and the falling-edge response time τ_fall is 40 ms.

The optoelectronic response of a 1060 nm wavelength laser was also studied, as shown in Figure 8b–e. The maximum photocurrent responsivity is 718 A/W at a wavelength of 1060 nm (P = 10.9 mW/cm², V_ds = 1 V), and EQE = 8.41 × 10⁴%, D* = 1.5 × 10⁹ Jones. To better understand the heterostructure photodetector response mechanism in the visible–NIR, a schematic diagram of the device’s energy band under light illumination is given in Figure 8f. Under laser irradiation, electron–hole pairs are generated within the MoS₂ layer. Then, the electrons can be transferred to the surface of MWCNT. Compared with bare MoS₂ FETs, the heterostructure provides an enhanced photon absorption, and when the laser irradiates the heterostructure, the photogenerated electron–hole pairs are separated at the MWCNT-MoS₂ heterojunction interface due to the presence of the built-in electric field, generating photocurrents and promoting the capture of photogenerated carriers by the electrodes, speeding up the response so that the device can operate at V_ds = 0 V. When a forward bias is applied to the device, the p-n junction is in the open state and the superposition of the internal and external electric fields can separate the photogenerated electron–hole pairs more effectively, enabling the device to obtain a larger photocurrent. However, the device can obtain a higher photocurrent at V_ds = 1 V with a concomitant increase in the dark current. Therefore, when a bias voltage of V_ds = 1 V is introduced, the detectivity of the device decreases significantly, although the responsiveness of the photodetector can be enhanced. Overall, the performance of the MWCNT-MoS₂ photodetector is significantly improved compared to the bare MoS₂ photodetector, and this improved heterostructure system optimizes the performance of the photodetector. Moreover, current instability is observed in both dark and illumination conditions. It is possible that the MWCNT-MoS₂ heterostructure has an imperfect material contact interface that leads to carrier capture and release, which can result in jitter, such as the capacitive effect [48]. This phenomenon has also been observed in some hybrid dimensional heterostructure photodetectors, as reported in a previous study [49,50]. In the case of the MWCNT-MoS₂ heterostructure, the carbon nanotube is one-dimensional, and the interface with the two-dimensional material may be less than ideal. However, testing the device in a vacuum environment could significantly reduce the current instability as it would minimize the presence of impurities and other contaminants that may contribute to the carrier capture and release. Nevertheless, further investigation is needed to determine the root cause of the current instability and develop strategies to improve the device’s performance.

4. Conclusions

In conclusion, we demonstrated a hybrid dimensional heterostructure based on directionally aligned MWCNTs with MoS₂ for a high-sensitivity photodetector, which was successfully prepared using a micro-nano fixed-point transfer technique. We found that the heterostructure device can observe stable optical response waveforms in a wide wavelength range from visible (520 nm) to near-infrared (1060 nm), demonstrating a nanomaterial-based heterostructure in which the advantageous properties of the components are combined to substantially improve the device properties. The photodetector provides a clear optical response when irradiated by laser light in the environment, and shows light-sensing capability at various power intensities and wavelengths. The device fabrication method is convenient and easy to control, which makes it a promising candidate for low-cost photodetection.