Fabrication of graphene: CdSe quantum dots/CdS nanorod heterojunction photodetector and role of graphene to enhance the photoresponsive characteristics

Integration of graphene with semiconducting quantum dots (QDs) provides an elegant way to access the intrinsic properties of graphene and optical properties of QDs concurrently to realize the high-performance optoelectronic devices. In the current article, we have demonstrated the high-performance photodetector based on graphene: CdSe QDs/CdS nanorod heterostructures. The resulting heterojunction photodetector with device configuration ITO/graphene: CdSe/CdS nanorods/Ag show excellent operating characteristics including a maximum photoresponsivity of 15.95 AW−1 and specific detectivity of 6.85 × 1012 Jones under 530 nm light illumination. The device exhibits a photoresponse rise time of 545 ms and a decay time of 539 ms. Furthermore, the study of the effect of graphene nanosheets on the performance enhancement of heterojunction photodetector is carried out. The results indicate that, due to the enhanced energy transfer from photoexcited QDs to graphene layer, light absorption is increased and excitons are generated led to the enhancement of photocurrent density. In addition to that, the graphene: CdSe QDs/CdS nanorod interface can facilitate charge carrier transport effectively. This work provides a promising approach to develop high-performance visible-light photodetectors and utilizing advantageous features of graphene in optoelectronic devices.


Introduction
Graphene nanosheets, a honeycomb structured carbon material, possess unique properties like high thermal stability, mechanical strength, and excellent electrical conductivity [1,2]. By its distinct characteristics comprising tunable optical and electronic properties via electrostatic or chemical doping, attention has been drawn to represent graphene in emerging novel devices that can be used in various applications spanned from electronics to the medical field [3][4][5]. Moreover, its gapless electronic structure leads to broadband optical absorption, and its high carrier mobility enables ultrafast response times. However, the primary drawback of graphene is zero-band gap nature along with low light absorption (2.3% in the visible and infrared region), ultra-fast recombination of photogenerated carriers, and ease of aggregation in solution, which limits its application to the optoelectronic devices effectively [6].
Meanwhile, semiconductor QDs are zero-dimensional materials exhibits unique optical and electrical properties due to their small size of 2-10 nm [7,8]. With the inherent fascinating properties, semiconducting QD has potential applications in photodetectors, solar cells, bio-imaging, and many more [9][10][11]. Nevertheless, like graphene, the semiconducting QDs also have disadvantages while considering their applications in optoelectronic devices. For instance, the decoherence and low carrier mobility limit the optical gain of QD materials [12]. Therefore, in recent years, there has been renewed interest in the integration of graphene with metals, metal oxides, quantum dots (QD), and will further broaden the application field with tunable properties [13][14][15]. In the graphene nanocomposite structure, the enhanced exciton dissociation and charge transfer at QDs/graphene interface promote the photocarrier generation. Therefore, an extraordinary photoconductive gain can be obtained, benefitting from long carrier lifetime of QDs and short charge transit time of graphene with high mobility [16]. Furthermore, it is possible to extend the QDs emission to longer wavelengths by combining the electronic properties of graphene with those of QDs in graphene: QD nanocomposites [17]. One of the important issues in the attachment of QDs on the surface of graphene is the monodispersity, which is essential for controlled charge transfer across the graphene layers.
Therefore, many efforts were devoted to synthesizing graphene-QD hybrid structures and designing photoelectric devices. For example, Geng et al demonstrated the noncovalent attachment of CdSe QDs to graphene to realize highly transparent semiconducting films [18]. Guo et al have developed a strategy to fabricate a solar cell using CdSe QDsgraphene and achieved the incident photon-to-charge-carrier conversion efficiency (IPCE) of 16%, a large improvement from the graphene-only and QD-only devices (IPCE<6%) [16]. The decoration of CdSe QDs on graphene sheets and the effect of graphene inclusion on optoelectronic properties of the QD-based device was investigated by Kim et al [19]. Sun et al have reported the CVD-grown monolayer graphene-PbS QDs based flexible infrared photodetector on plastic substrates and the device has shown enhanced photoresponsivity of 10 7 AW −1 [20]. Recently, Chen et al proposed a photodetector of intercalated graphene layers with thick PbS QDs films (i.e. alternating layers of QDs and graphene) and achieved the efficient charge collection over the spectral range from visible to IR region [21]. by adopting the advanced technique of ink-jet printing, Cook et al have fabricated ZnO/ graphene nanoplatelet bulk heterojunction UV photodetector and its photoresponsivity value reaches up to 2.2 AW −1 [22]. Furthermore, charge carrier dynamics of nanoparticle (QDs)/ graphene nanocomposites also have been studied experimentally and theoretically [23][24][25][26]. Especially, the photoluminescence of CdSe QDs in graphene nanohybrids is reduced both in intensity and lifetime due to strong interaction and energy transfer [27]. These works mentioned above highlight the importance of graphene's role to enhance the charge carrier dissociation and transport in the optoelectronic device. However, to the best of our knowledge, only a few studies have focused on the effect of graphene in heterojunction photodetectors but the results are promising. For example, Konstantatos et al demonstrated monolayer or bilayer graphene-PbS QDs heterojunction phototransistor, where trapped charges at the interface causing a photogate effect and enhance the device photoresponsivity up to 10 7 AW −1 which is significantly higher than graphene only device of 10 -2 AW −1 [28]. Therefore, there is an immediate need for studying the optoelectronic properties of graphene: QD nanocomposites and specifically, the role of graphene on performance metrics of optoelectronic devices. Furthermore, it should be noted that previous studies have indicated that heterostructures made by means of complex layer assemblies lead to ultrafast charge transfer and efficient exciton separation due to the occurrence of dissimilar band alignment [29]. More specifically, the realization of heterojunction by the assembly of 1D anisotropic material (nanorods, nanowires) with 2D layers or thin-film structure offering rather interesting surface interface phenomena at the interface via its unique morphological structure and subsequently increase the light absorption [30,31]. Therefore, instead of planar heterojunction, vertical heterojunction would provide the appropriate platform to analyze the charge carrier dynamics exclusively.
Herein, we present the hybrid heterojunction photodetector consists of graphene: CdSe QDs and CdS nanorod. The graphene: CdSe QDs nanocomposite was synthesized via low-temperature in situ synthesis technique and high-quality CdS nanorods were synthesized by the hydrothermal method. The PL and TRPL studies of graphene: CdSe QDs nanocomposite showed an obvious quenching effect compared to the pure CdSe QDs as the reason for fast separation and transfer of photo-induced charge carriers between CdSe QDs and graphene layers. Besides, CdS nanorod was utilized to enhance the light absorption in the visible region and outstanding charge transport layer. The as-fabricated hybrid heterojunction photodetector with device structure ITO/graphene: CdSe QDs /CdS/Ag shows enhanced photo responsivity of 15.95 AW −1 and specific detectivity of 6.85×10 12 Jones which is significantly higher than ITO/CdSe/CdS/Ag device. Our work demonstrates the great potential of graphene: QD nanocomposite as the photoactive layer in the high-performance heterojunction photodetector.

Preparation of graphene solution
For typical preparation of graphene solution, 5 mg of graphene powder was dissolved in 5 ml of 1-Octadecene (1 mg ml −1 ) and sonicated for 4 h. The resulting solution is centrifuged for 5 min at 2000 rpm. The supernatant was collected and the process is repeated at 4000, 6000, 8000, and 10 000 rpm to get single-layer graphene in ODE solution. The final concentration of graphene: ODE solution is 0.13 mg ml −1 .

In situ synthesis of CdSe QD-Graphene
The CdSe QDs were synthesized according to the reported literature with slight modification [32]. In a typical synthesis procedure, the precursors are prepared in the following way. Cadmium precursor: 4 mM of cadmium acetate dihydrate is taken in the round bottom flask and 10 ml of 1-octadecene, 0.250 ml of oleic acid, 1 ml of graphene solution, and 1 ml of n-octylamine is added. And then, the mixture is stirred for 30 min at 130 ℃. Selenium precursor: 2 mM metallic selenium powder is dissolved in 6 ml of tri-octyl phosphine (TOP) and 1.8 ml of toluene. For the CdSe QDs-graphene growth, 2 ml of selenium-TOP solution was injected into the reaction flask containing cadmium precursor, and the reaction mixture was kept at 170 ℃. For pristine CdSe QDs growth, the procedure is followed as the above, but without graphene solution added to the cadmium precursor. The samples were taken at different time intervals of 5, 10, 15, 20, and 30 min. The samples were purified with hexane/methanol solution and precipitated by centrifugation at 10,000 rpm with acetone. The precipitated colloidal powder is dried and dissolved in hexane, toluene, or chloroform.

Synthesis of CdS nanorods
In a typical procedure, 20 mM cadmium acetate was dissolved in 25 ml of DI water. Then 0.250 ml of TGA was added under vigorous stirring at room temperature. Further, sulfur source solution was prepared by dissolving 40 mM sodium sulfide dihydrate in 50 ml of DI water and added with cadmium source solution. The resulting mixture was stirred for 30 min and transferred into a 100 ml Teflon-lined stainless-steel autoclave. The autoclave was maintained at 180 ℃ for 8 h and then cooled to room temperature naturally. After cooling to room temperature, the precipitation was washed with DI water and absolute ethanol several times to remove the excess reactants and byproduct. Finally, the sample was dried in a vacuum oven at 40°C overnight.

Photodetector fabrication and characterization
Firstly, the pre-patterned ITO glass substrate was cleaned with detergent, deionized (DI) water, isopropyl alcohol, and acetone for 15 min using ultra-sonication. Further, substrates were treated with UV ozone to make the surface hydrophilic. The as-synthesized CdS NRs solution (10 mg ml −1 in n-butanol) was drop-casted on ITO glass substrates and dried at 80 ℃ for 5 min. The CdSe QDs or graphene-CdSe QDs in n-hexane (100 mg ml −1 ) was spin-coated at 1000 rpm for 6 s and 3000 rpm for 60 s and then samples were treated with ethanol washing twice. Finally, a vacuum evaporated Ag electrode (80 nm) was deposited to complete the device fabrication. The active device area is calculated to 1.539×10 -2 cm 2 .
The UV-visible absorption spectra of as-synthesized CdSe QDs and graphene: CdSe QDs samples were analyzed on UV/vis/NIR spectrophotometer (Shimadzu, UV-3600). The steady-state photoluminescence (PL) spectra were acquired by using a fluorescence spectrophotometer (Hitachi, F-4600). Raman spectra of samples were taken at room temperature using 514.5 nm incident photons from an Ar ion laser (JY LabRam) in backscattering geometry. The surface morphology of the sample was acquired by a ZEISS Sigma 500 field emission scanning electron microscopy (SEM) at 30 kV. A Philips CM200 tunneling electron microscope (TEM) operating at an accelerating voltage of 200 kV, with a Bruker SDD EDX system, was used for transmission electron microscopy studies. The FTIR spectra of the samples were recorded in a NICOLET 10 spectrometer operating in the range of 4000 to 400 cm −1 . All the electrical parameters of devices were measured by the Keithley 4200-SCS semiconductor characterization system assisted with a probe station. The irradiation was generated from monochromatic light-emitting diodes (530 nm), while the power of the incident radiation was tuned and measured with a power meter (Sanwa Mobiken LASER POWER METER LP1). All the measurements were done in air at room temperature.

Results and discussion
The as-synthesized CdSe QDs and graphene: CdSe QDs diluted in hexane solution under UV illumination are shown in figure S1 (supporting information) (available online at stacks.iop.org/NANO/32/315204/mmedia). Photographic image of the CdSe QDs and graphene: CdSe QDs nanocomposites synthesized with different reaction times kept under UV light irradiation at 365nm showed that the samples emit different colored light revealing the size-tunable formation of QDs with time variable. This was confirmed quantitatively by UV-vis absorption spectrum as shown in figures 1(a), (b). As seen from figure 1(a), CdSe QDs synthesized for 5 min show the absorption peak at 470 nm and the absorption peaks were shifted to higher wavelengths as the synthesis time increases, which indicated the increase of QD size. The size of QD is calculated from the first absorption peak based on the method reported in the literature and tabulated in table S2 [33]. It is of interest to notice that the size of pristine CdSe QDs is increased from 2.10 to 2.27 nm as the synthesis time increase from 5 to 30 min. On the other hand, the absorption spectra of graphene: CdSe QDs as in figure 1(b) showed a slight red shifting compared to pure CdSe QDs. It should however be noted that the graphene had no obvious absorption characteristics in the visible region; the visible light absorption was due to the contribution from CdSe QDs. Furthermore, the absorption edge of graphene: CdSe QDs nanocomposite was moved from 481 to 498 nm with sample synthesis time. Compared with bulk CdSe (1.78 eV, 698 nm), the obvious shifting of optical absorption edge in both pristine CdSe QDs and graphene: CdSe QDs nanocomposite clearly indicates the existence of quantum confinement effect [34]. Furthermore, with reference to pristine CdSe QDs, the absorption peak shifting of graphene: CdSe QDs strengthened the fact of charge carrier transfer between graphene and QDs [35].
The photoluminescence spectra of the pristine CdSe QDs and graphene: CdSe QD nanocomposites were displayed in figures 1(c), (d). As seen in figure 1(c), a prominent emission peak at 510 nm was observed for pristine CdSe QDs synthesized for 5 min and it shifts towards a higher wavelength as the synthesis time prolongs. At most, CdSe QDs synthesized for 30 min showed the emission peak at 535 nm. It can be understood that the shifting of emission peak might be related to the quantum confinement effect as the size of QD evolves and it is well agreed with the absorption spectra study. However, for graphene: CdSe QD nanocomposites, the emission peaks were observed at 512, 518, 526,533, and 543 nm for 5, 10, 15, 20, and 30 min samples respectively as shown in figure 1(d) and table S2. The considerable shifting of emission peak as compared with pristine CdSe QDs again supports the effective interaction between graphene and QDs. Moreover, as seen from figure S2, the comparative PL spectra showed that the PL intensity of graphene: CdSe QDs was downgraded than that of pristine CdSe QDs since graphene layers could provide additional energy-transfer pathways with intrinsic radiative channels for excited-state charge transfer [36]. Furthermore, Raman spectroscopy was used to characterize the ordered and disordered crystal structures of carbon materials before and after functionalization. To generalize the scheme, CdSe QDs-graphene samples synthesized for 5 and 15 min were characterized for Raman spectra in the sense of the different sizes of QDs on a graphene sheet. Raman spectra of pristine graphene sheets were also given for comparison. As indicated in figure 1(f), all three samples were exhibited the characteristic phonon modes of vibration on laser excitation. For pristine graphene sheet, two peaks at 1319.8 and 1580.4 cm −1 were observed and assigned to characteristic D and G bands of the two-dimensional carbon layer. Besides, a broad low-intensity peak also appears at 2500 cm −1 which represents the 2D band of graphene. However, for the CdSe QDs functionalized graphene layer, an obvious characteristic peak of D and G bands are appeared along with a clear peak at 403 cm −1 which represents the overtone LO mode termed as CdSe 2LO [37]. In general, the D band at 1357 cm −1 is the breathing mode of π-point phonons of A 1g symmetry attributed to local defects and disorders, particularly the defects located at the edges of graphene. And also, the G band is assigned to the E 2g phonon of sp2 bonds of carbon atoms. After functionalizing CdSe QDs on the graphene, significant red shifting of D and G bands were observed for graphene: CdSe QD hybrids synthesized at 5 min and 15 min. A redshift of D band by 55 cm −1 and 63 cm −1 were observed for 5 min and 15 min samples respectively. On the other hand, redshifts of G band by 277 cm −1 and 279 cm −1 were observed for 5 min and 15 min samples respectively. In the light of earlier reported findings, the G band shifting can be affected by the carrier doping levels, strain, and localized temperature [38][39][40].
Here, unlike the D band, the G band has shifted significantly accomplished that charge transfer from QD functionalization and subsequent strain occurred by QDs loading on the graphene layer which is consistent with the findings of the literature report [23]. In addition to the redshifting of characteristic peaks, the intensity ratio I D /I G as shown in figure S3(a) (in SI) also provides the degree of QDs functionalization on graphene layers. After QDs deposition on graphene, the I D /I G increases from 1.47 for pristine graphene to 4.69 for graphene: QDs of 15 min. Moreover, as shown in figure S3(b) (in SI), an enhancement of I 2D /I G ratio was observed, which suggests the presence of a QDs on the graphene layers.
The photo-induced kinetics of graphene: CdSe QDs are examined by time-resolved photoluminescence (TRPL) spectroscopy. The TRPL measurements were taken on pure CdSe QDs (5 min), graphene: CdSe QDs (5 min), and graphene: CdSe QDs (15 min) samples using 585 nm excitation wavelength (see figure S4 in the SI). A tri-exponential decay model was used to fit the decay curve: is the timedependent fluorescence intensity, A is the amplitude and τ is the lifetime. The emission lifetime of measured samples is summarized in table S1 (in SI). It should be noted that the average emission lifetime (τ ave ) of graphene: CdSe QDs nanocomposites was relatively shorter than that of the corresponding pristine CdSe QDs. The average lifetime of CdSe QDs is calculated to be 557.9 ns and graphene: CdSe QDs (5 min) is 128.34 ns. Here, one should consider the fact that the as-synthesized CdSe QDs are used for measurement and device application without the ligand-exchange process. Therefore, the long-chain hydrocarbon oleic acid ligand used for the initial synthesis process form an insulating layer around each QD, and subsequently, the organic ligands create an energetic barrier to charge transport [41]. However, the difference in the average lifetime between pure CdSe QDs and graphene: CdSe QDs indicates the existence of a nonradioactive pathway from the electronic interaction between QDs and graphene [42,43].
The TEM and HRTEM images of representative CdSe QDs and graphene: CdSe QDs synthesized for 15 min were depicted in figure 2. As seen in figure 2(a), CdSe QDs show a uniform size distribution with good crystalline quality. The size of the CdSe QD was in the range of 2-2.3 nm as calculated from the HRTEM image of the sample synthesized for 15 min, which is consistent with that of absorption spectra measurement. The emergence of clear lattice planes on the HRTEM as in figure 2(b) represents the good crystallization of CdSe QDs and is indexed to the cubic phase of CdSe crystal with the lattice constant a=0.62 nm. Interestingly, the synthesis of QDs in presence of graphene would show a more pronounced growth compared with pure QD synthesis.
As it is seen in figures 2(c), (d), the CdSe QDs are randomly attached and highly loaded on the surface of graphene. Furthermore, it is appeared to be clearer in shape with lattice fringes emphasized along the (100) plane. One can see clearly that the size distribution of CdSe QDs on graphene is more or less the same as that of pristine CdSe QDs. Therefore, the results support the argument that the redshifting of absorption spectra was not from the size tuning of QDs but the strong interaction between graphene and QDs due to charge delocalization [44].
The phase purity and the crystal structure of the CdS nanorods synthesized by the hydrothermal method were investigated by powder x-ray diffraction (PXRD) study. The XRD patterns in figure S5 (in SI) show that the as-prepared CdS nanorods are in the hexagonal phase with cell constants a=4.123 Å and c=6.686 A which are very close to the values in the reported literature [45,46]. The XRD pattern exhibits prominent, broad peaks at 2θ values of 31.5, 35.2°, 41.6°, 50.7°, and 60.4°which could be indexed to the scattering from 200, 102, 102, 110, 112, and 104 planes respectively. These structural characteristics have also been supported by Raman spectra as shown in figure S6 (in SI). As seen from Raman spectra, CdS nanorods exhibit 1LO and 2LO phonon modes observed at 298 cm −1 and 587 cm −1 respectively. In addition to the LO phonons and their replicas, other peaks are also found at 101 cm −1 , 162 cm −1 , 268 cm −1 , suggesting that the nanorods have better crystal quality [47].  nanorods consists of nanocrystals with a size of 5-6 nm, which is reflected in the SAED pattern of nanorod as shown in figure 2(h). Figure 2(g) shows clearly the lattice space of 0.33 nm, which is assigned to the (002) diffraction plane indicate that the growth direction is along with the [0001] direction of wurtzite CdS. According to the literature, the temperature at which a CdS nanocrystals changed from amorphous phase to the hexagonal wurtzite phase in the hydrothermal process is 160 ℃ and pure hexagonal phase is formed at 240 ℃ [48]. Therefore, the existence of both nanoparticles and nanorods is obvious in as-synthesized CdS products treated at 180 ℃.
To demonstrate the efficient charge transfer properties of nanohybrids, a heterojunction photodetector was fabricated with device structures as shown in figure 3(a). Two types of the device have been fabricated such as ITO/CdSe QDs/ CdS/Ag (Device A) and ITO/graphene: CdSe QDs /CdS/ Ag (Device B). The surface morphology of drop-casted CdS NRs on ITO/glass substrates is depicted in figure 3(b) and also, figure 3(c) shows FE-SEM image of graphene: CdSe QDs nanocomposite film on CdS NRs, which clearly shows densely packed and thick distribution of thin-film structure favorable for device fabrication. The photodetector device was characterized by apply bias sweep from +1.5 to −1.5 V. Figure 3(d) shows the I-V curves of device A in dark and under 530 nm laser illumination at varied optical power density from 100 μW cm −2 to 751 μW cm −2 . It is noted that a clear rise of the photocurrent with increasing incident power density was observed, which signifies the effective conversion of photon flux to photogenerated carriers. Moreover, the I-V curve shows slight nonlinear and asymmetrical behavior, corroborating the proper formation of heterojunction. However, it was found that device B shows more nonlinear characteristics compared to device A. It is worth noting here that, the dark current of device A and device B are 1.695×10 -8 A and 9.91×10 -6 A respectively, and indicates that the dark current has been increased with graphene addition due to the increment of conductivity. At the same time, photocurrent was also increased tremendously with graphene doping from 1.0515 mA (100 μW cm −2 ) to 2.799 mA (751 μW cm −2 ). Moreover, the ON-OFF switching of the device under +1 V bias under illumination with 530 nm and 100 μW cm −2 light intensity is shown in figure 3(e), displaying the stable and repetitive cycles, which demonstrate the photodetection reversibility.
Another important characteristic of the photodetector is the linearity of photocurrent upon illumination light intensity. To examine such linearity relationship, 530 nm LED has been chosen with power density variation from 100 μW cm −2 to 751 μW cm −2 at bias +1 V. A seen in figure S7 (in SI), by fitting the plot with the power-law equation I ph = AP α , the value of α is calculated to be 0.43 and 0.39 for device A and device B, respectively. The obvious deviation from the ideal value of 1 is attributed to the loss of photoexcited carriers through charge recombination. Both defects and impurities may act as charge recombination centers, which could be filled by photoexcited carriers as the light intensity increases. Moreover, the slight increase of α value for device B than device A indicates the inherent role of graphene as the electron collector and subsequent charge transportation medium.
One important technical aspect of the potential photodetector is its photoresponse time. Figures 4(a), (b) shows the rise and decay time upon V bias =+1 V for device A and device B under 530 nm light illumination and 100 μWcm −2 light intensity. The rise time is defined as the time gap between 10% of the 'off' state to 90% of the 'on' state and decay time is defined as the opposite. The decay time shows a slight decrement from 546 ms to 539 ms, while the rise time decreased from 656 ms to 545 ms after introducing graphene with CdSe QDs. This however may be explained by the fact that the charge mobility in QD films is limited by grain boundaries, the presence of graphene sheets provides an additional conducting channel for charge transport [47,48]. Therefore, device B responds to the light signal faster rather than device A.
Furthermore, the performance of photodetectors is evaluated in terms of key parameters such as photoresponsivity (R), Specific detectivity (D * ), and external quantum efficiency. The photoresponsivity R is expressed as R=I p /PS, where I p =I light -I dark , P is the incident power density and S is the effective device area. D * is defined in terms of responsivity R and its simplified form is where I p =I ph -I d , A is the active area, P is the incident power density, q is the coulombic charge. As shown in figure 4(c), the maximum photoresponsivity and specific detectivity for device A is calculated to be 18.57 mAW −1 and 9.24×10 11 Jones respectively. On the other hand, device B shows an enhanced photoresponsivity and specific detectivity of 15.95 AW −1 and 6.85×10 12 Jones respectively as depicted in figure 4(d). Moreover, the incident power-dependent R and D * indicate that R and D * have shown higher value at a weak light signal. Also, both R and D * decreased gradually with increasing the light intensity for both device A and device B. Therefore, the results strongly manifest the existence of considerable recombination loss of charge carriers in the device. This behavior has also been observed in other systems such as Gr/PbS, perovskite nanostructures, etc [20,49,50]. Figures 5(a), (b) shows a schematic diagram of the energy bands corresponding to the carrier transport mechanisms of the holes and electrons in dark and under light illumination for device A and device B. When a 530 nm light illuminates the photodetectors through the top electrode, the photons penetrate the CdSe QDs layer, resulting in the creation of the excitons. Because of the band alignment and potential difference between the band positions, the photogenerated electrons follow the heterojunction mechanism, where the electrons present in the CB of CdS NRs layer are transferred to the CB of CdSe QDs layer and to the Ag electrode. At the same time, holes in the VB of CdSe QDs are transferred to VB of CdS NRs and to the ITO ( figure 5(a)) [51]. On the other hand, in device B, charge excitons are generated in the CdSe QDs: graphene nanocomposite under light illumination and charge carrier separation is occurred under applied electric field, however, in presence of graphene, charge carrier separation is enhanced due to nano heterojunction formation at the interface of CdSe induces the charge carrier transportation effectively. Therefore, the electrons and holes are accumulated at the Ag and the ITO layers, respectively, resulting in the generation of the photocurrent in the photodetectors ( figure 5(d)). It is anticipated that the photodetector performance could be further improved by the proper ligand-exchange process to overcome the resistance provided by oleic acid and designing of the device structure. We also compared the performance of our device with other graphene nanocomposite-based photodetectors as shown in table 1.

Conclusion
In summary, we have successfully synthesized the graphene: CdSe QD nanocomposites at a low-temperature regime by a one-pot solvothermal method and subsequently demonstrated the heterojunction photodetector. In the course of this work, it was found that the synergistic interaction between CdSe QDs and graphene facilitating the separation of electron-hole pairs and prolong the lifetime of the charge carriers, which is favorable for high-performance optoelectronic device application. Interestingly, the fabrication of heterojunction photodetector Ag/graphene: CdSe QDs/CdS NRs/ITO exhibits higher photoresponsivity and detectivity up to 15.95 AW −1 and 6.85×10 12 Jones respectively. The appropriate structure and band alignment in graphene: CdSe QDs/CdS NR heterojunction benefits the visible light absorption and enhanced charge transfer properties. The present study reveals that the formation of nano junctions by introducing graphene into QD layers is a good strategy to improve the charge carrier separation and transportation that results in the enhancement of the photosensing capability of the photodetector.