Ultra-Broadband Photodetector for the Visible to TeraHertz

Ultra-Broadband Photodetector for the Visible to Terahertz Range by Self-Assembling Reduced Graphene Oxide-Silicon Nanowire Array HeterojunctionsY ang C ao ,J iayi Z hu ,J ia X u ,J unhui H e ,*J ia-Lin S un ,*Y ingxin W ang ,a nd Z iran Z hao graphene could extend the working spectrum of graphene photodetectors to the far-infrared (FIR) region, and even asfar as the terahertz (THz) region.[ 9,10 ] Band structure engi-neering of graphene has been widely investigated. A bandgap can be created in graphene by patterning graphene nanor-ibbon, nanomesh and quantum dot-like arrays structures bymethods including electron-beam lithography,[ 11 ]nanowire shadow mask etching,[ 12 ] block copolymer lithography, [ 13 ] unzipping of carbon nanotubes[ 14,15 ] and Titanium sacrifi -cial layer etching.[ 16 ] However, the fabrication processes for bandgap creation are complex. Recent studies indicate that the reduced graphene oxide (RGO ) obtained by chemical exfoliation from graphite has a natural energy gap withoutthose complex fabrication processes.[ 17 ] This natural energy gap can be tuned by controlling the degree of RGO reduc-tion. The apparent energy gap ranged from 10 to 50 meV (corresponding to the energy of 124–24.8 µm electromag-netic waves) and approached zero with extensive reduc-tion. The naturally narrow energy gap in RGO is suitable for detection of mid-infrared (MIR) (2.5–30 µm) or even THz (30–3000 µm) waves. Recently, Chitara et al. reported NIR (1.55 µm) photodetectors based on extensively reducedRGO.[ 18 ] However, few studies have addressed the develop-ment of RGO MIR/THz photodetectors. There are still some obstacles to the realization of high performance ultra-broad-band photodetectors. The limited optical absorbance of RGO is not conducive to achieving a high photodetector response, and the RGO energy gap of tens of meV is inappropriate for VIS and NIR light detection. However, Si nanowire (SiNW) arrays have two unique advantages that can compensate to an extent for the RGO defi ciencies. O ne is the Si bandgap of ∼1.12 eV , which makes SiNW arrays suitable for photo-detection in the VIS to NIR range. The other is the aligned nanowire array structure that can intensively suppress light refl ection over a wide spectral range from VIS to NIR. This can signifi cantly improve the light harvesting of SiNW arraybased devices.[ 19 ] Introducing SiNW arrays to RGO devices can not only enhance the device light harvesting properties, but can also enhance the photoresponse to light from the VIS to NIR range.T hus, here we integrated the advantages of RGO and SiNW arrays into a single photodetector by forming RGO -SiNW array heterojunctions through a simple, low-cost drop casting method followed by heat treatment. A broad photo-response from VIS (532 nm) up to the terahertz regionSilicon NanowiresDOI: 10.1002/smll.201303339Y . Cao, J. Zhu, J. HeF unctional Nanomaterials Laboratory and Key Laboratory of Photochemical Conversion and Optoelectronic MaterialsTechnical Institute of Physics and ChemistryC hinese Academy of Sciences (CAS) Z hongguancundonglu 29 ,H aidianqu ,B eijing 100190 ,C hina E-mail: j hhe@ J . XuB eijing Key Laboratory of Low-grade Energy Multiphase Flow and Heat Transfer, School of Renewable EnergyN orth China Electric Power University B eijing 102206 ,C hina J .-L.SunS tate Key Laboratory of Low-Dimensional Quantum Physics Department of Physics T singhua University B eijing 100084 ,C hina E-mail:j lsun@ J .-L.SunC ollaborative Innovation Center of Quantum Matter B eijing ,C hinaY . Wang, Z. ZhaoK ey Laboratory of Particle & Radiation Imaging (Tsinghua University) Ministry of EducationDepartment of Engineering Physics T singhua University B eijing 100084 ,C hinaH igh-performance broadband photodetectors have recently attracted signifi cant interest[ 1–5 ] because of their impor-tance to a variety of applications, including imaging, remote sensing, environmental monitoring, astronomical detection, photometers and analytical applications. Graphene is a prom-ising material for broadband photodetection applications because of its ability to absorb incident light over a wide wavelength range, from at least the visible (VIS) spectrum tothe infrared.[ 6] Recent works have demonstrated that zero-bandgap single- or few-layer graphene-based photodetec-tors based on a fi eld-effect transistor (FET) structure could operate in the near-infrared (NIR) and VIS parts of theelectromagnetic spectrum.[ 7,8 ] However, no working spectra have been demonstrated for these zero-bandgap graphene photodetectors in longer wavelength ranges. Theoretical cal-culations indicated that opening and varying the bandgap ofY. Cao et al. communications(2.52 THz, or a wavelength of 118.8 µm) was thus realized in a single RGO-based photodetector at room temperature.T he SiNW array and graphene oxide (GO) nanosheets were prepared by metal-assisted chemical etching [ 20 ]and an improved Hummers’ method, [ 21 ]respectively. The details are given in the Supporting Information. As Figure S1 shows, the SiNW array was hydrophobic, but had good ethanol wettability. The RGO-SiNW array heterojunction was thus fabricated by drop-casting an ethanol suspension of GO nanosheets on top of the SiNW array, which is favorable for GO nanosheets to be uniformly spread over the SiNW array surface during direct drop casting of the GO suspen-sion on top of the SiNW array. The heterojunction was then subjected to heat treatment at 550 °C in an Ar (95%)-H 2 (5%) atmosphere for 3 h. This heat treatment implemented reduction of the GO nanosheets to a degree while simulta-neously improving the contact between the SiNW array and the RGO nanosheets. Scanning electron microscopy (SEM) observation of the heterojunction ( F igure1)revealed three important features. First, the SiNW array, which was about 5 µm long, was oriented vertically to the wafer surface (Figure 1a). The high density and the uniformity of the array suggested excellent SiNW quality (Figure 1b). Second, a good heterojunction was formed by the SiNW array and the RGO nanosheet (Figure 1c). Although cracks (pointed by arrows in Figure 1d) existed in some RGO layers, the multilayer struc-ture of RGO sheets that overlapped and interconnected with each other (Figure 1d) ensured an electrically-conducting pathway. The multilayer structure of RGO nanosheets has similar transmittances of 22−25% for all the lights from the VIS (532 nm) up to the terahertz region (2.52 THz) (see Figure S2 and Table S1), indicating insignifi cant wavelength selectivity of light transmission by RGO nanosheets. Third, most RGO nanosheets on the SiNW array were spread out, and the existence of wrinkles and folds (pointed by arrows in Figure S3) showed that the two-dimensional nanosheetF igure 1. C ross-sectional-view (a) and top-view (b) SEM images of the SiNW array. Cross-sectional-view (c) and top-view (d) SEM images of the RGO-SiNW array heterojunction produced by heat-treatment at 550 °C. (e) Schematic illustration of the device confi guration. Inset: a digital photograph of the device.Ultra-Broadband Photodetector for the Visible to Terahertz Range by Self-Assembling Reducedwas fl exible (Figure S3). Figure 1 e shows a schematic of the fabricated RGO -SiNW array heterojunction photodetector in a two-pole structure, and the inset shows a digital photo-graph of the device. The device fabrication is described in the Experimental Section. All measurements were carried out at room temperature under ambient conditions. The forward bias is defi ned as a positive voltage applied to the RGO nanosheets. F igure 2 a , c, e and g show typical current–voltage ( I –V )characteristics of the RGO-SiNW array heterojunction photo-detector in the dark and under irradiation by 532 nm (VIS), 1064 nm (NIR), 10.6 µm (MIR), and 118.8 µm (2.52 THz)P h o t o c u r r e n t (m A )Time (s)P h o t o c u r r e n t (m A )Time (s)P h o t o c u r r e n t (m A )Time (s)P h o t o c u r r e n t (m A )Time (s)C u r r e n t (m A )Voltage (V)C u r r e n t (m A )Voltage (V)C u r r e n t (m A )Voltage (V)C u r r e n t (m A )Voltage (V)(c)(a) (b)(d)(e)(f)(g)F igure 2. P hotoelectrical properties of the RGO-SiNW array heterojunction photodetector produced by heat-treatment at 550 °C. (a, c, e and g) I –V characterizations, (b, d, f and h) photoresponses under illumination by (a, b) 532 nm (VIS), (c, d) 1064 nm (NIR), (e, f) 10.6 µm (MIR) and (g, h) 118.8 µm (THz) lasers. The insets in (a, c, e and g) show their respective I –V characteristics in detail around the origin.Y. Cao et al. communicationslasers, respectively. The dark I-V curve showed rectifying char-acteristics and confirmed that the RGO nanosheet formed a Schottky-like junction with the SiNW array. [ 22 ]Equilib-rium energy band diagram of the RGO/n-Si heterojunction in the dark is shown in Figure S4(a). In all four irradiation cases, from VIS to THz, the RGO-SiNW array heterojunc-tion yielded a significant photocurrent I ph, which is defi ned as the difference between the current under irradiation and the dark current. One phenomenon of note is that under irra-diation with both VIS and NIR lasers, approximately equiva-lent photocurrents can be found under both forward bias and reverse bias, whereas, under irradiation with MIR and THz lasers, the photocurrents under forward bias are much larger than those under reverse bias. A recent study on gapless gra-phene−silicon wafer heterojunctions showed I ph at reverse bias by illumination with visible light, because the photoex-citation only took place in Si and the graphene only worked as carrier collector. [ 23 ]When the current RGO-SiNW array heterojunction was irradiated with VIS or NIR lasers, how-ever, photoexcitation resides in both of SiNW array and RGO which has a minor bandgap, due to the higher photoenergies of VIS (∼2.33 eV) and NIR (∼1.17 eV) than the bandgaps of Si (∼1.12 eV) and RGO (tens of meV). The photoexcitation of SiNW array mainly contributed to the generation of photo-current at reverse bias, while that of the RGO predominantly contributed to the generation of photocurrent at forward bias (Figure 2a and c). The energy band diagram under a reverse bias, V r bias,is schematically shown in Figure S4(b), in which V r bias lifts the Fermi level of RGO (E f(RGO)) to a higher value than its original position in the dark. Thus, the Dirac cone of RGO opens up a large number of accessible states for injection of photo-induced holes from Si and allowing a com-plete collection of the injected holes. [ 23 ]In contrast, when the RGO-SiNW array heterojunction was irradiated with MIR and THz lasers, I ph was only attributed to the photoexcita-tion of RGO because the photoenergies of MIR (∼0.117 eV) and 2.52 THz (∼0.01 eV) are both lower than the bandgap of Si. Figure 2e and g again confi rm that the I ph ascribed to the photo e xcited carriers in RGO mainly occurred at forward bias. Figure S4(c) schematically represents the situation under a forward bias, V f bias. The photo-induced holes by MIR and THz lasers transfer from RGO to Si.T he insets in Figure 2a, c, e and g show their respective I–V characteristics in detail near the origin, and further prove the above explanation. From comparison of the insets, an open-circuit voltage ( V oc)of ∼0.15 V and a short-circuit cur-rent ( I sc)of ∼10 µA are found under irradiation with VIS and NIR lasers (see the insets in Figure 2a and c), which can be attributed to photoexcitation of the SiNW array. [ 22 ]However, for the cases of MIR and THz irradiation, V oc and I sc are absent because there is no photoexcitation inside the SiNW array (see the insets in Figure 2e and g). Thus, under MIR and THz irradiations, photogenerated currents in the device only come from the RGO.I n addition, as shown in Figure 2a and c, under irradia-tion with VIS and NIR lasers, the photocurrents at biases of 1 V and −1 V are approximately equivalent, but much lower dark currents at a bias of −1 V render a much higher O N/ OFF ratio (the ratio of current during irradiation to the dark current) than at a bias of 1 V. In the VIS irradiation case, the ON/OFF ratios at −1 V and at 1 V are 7.09 and 1.37, respec-tively; in the NIR irradiation case, the ON/OFF ratios at −1 V and 1 V are 11.27 and 1.57, respectively. Thus, for photodetec-tion of VIS and NIR light, −1 V should be used as the device’s working bias. For photodetection of MIR and THz light, as a result of relatively higher photocurrents under forward bias than under reverse bias, 1 V is chosen as the photodetector’s working bias.T he respective photoresponses to the VIS/NIR lasers at a bias of −1 V and to the MIR/THz lasers at a bias of 1 V are shown in Figure 2b, d, f and h. Signifi cant photocurrent response was observed for all the VIS, NIR, MIR and THz irradiations. Good photoresponse repeatability was observed in the cases of MIR and THz irradiation, but for the VIS and NIR irradiations, the photoresponse repeatability was weakened, as reflected in the increase of photocurrent with light-on and off cycles. This is mainly attributed to the thermal effect of Si. However, it is noted that the photore-sponse time remained constant in the light-on and off cycles. Thus, the RGO-SiNW array heterojunction photodetector can effectively perform ultra-broadband detection of light from VIS to THz. To the best of our knowledge, this is the broadest range reported to date for a single graphene-based photodetector device. For VIS and NIR irradiation, when the irradiation was ‘O N’ or ‘O FF’, the device produced a quick response (indicated by the green arrows in Figure 2b and d). For example, when NIR irradiation was “ON”, the photocur-rent increased to 65% of the maximum within 70 ms (the time interval between two collected data points). The rise time t r and decay time t d of the photodetector were calculated for the four irradiation cases according to the method in the literature, [ 24 ]and are summarized in Table S2. Clearly, under both VIS and NIR irradiation, the device shows shorter t r and t d values than under both MIR and THz irradiation, where both t r and t d are reduced by approximately 50%. This sug-gests that photoexcitation of the SiNW array can signifi cantly improve the photodetector response speed. For MIR and THz irradiation, the device response, i.e. the increase in the current, should only come from photoexcitation within the RGO, and thermal effects can be excluded because graphene heating by IR laser absorption is known to reduce graphene conductivity. [ 25–28 ]W e also investigated the effects of reduction degree of RGO nanosheets on the photodetector performance. Because the RGO reduction degree increases with increasing heat treatment temperature, the GO-SiNW array heterojunction was also heat-treated at 1000 °C for comparison. The Raman spectra of RGO-SiNW-1000 °C and RGO-SiNW-550 °C ( F igure3a) show two strong characteristic bands at 1358 cm −1 and 1597 cm −1, which are assigned to a D band corresponding to the structural disorder of the sp 2bonded carbon domain, and a G band corresponding to the C-C stretching in the graphitic lattice, respectively. The intensity ratio of the two bands ( I D/I G) decreases from 1.15 to 0.97 with an increase in the heat-treatment temperature from 550 °C to 1000 °C, indicating that the higher temperature heat-treatment caused more restoration of electronic conjugation within the RGO nanosheets and increase in the cluster size of nanographiteUltra-Broadband Photodetector for the Visible to Terahertz Range by Self-Assembling Reducedcrystallites. This would in turn enhance the RGO nanosheetconductivity.[ 29 ] Figure 3 b shows the I -V characteristics of the RGO -SiNW-1000 °C and RGO -SiNW-550 °C devices in the dark. Clearly, under the same bias, the dark current of the RGO-SiNW-1000 °C device is much higher than that of the RGO-SiNW-550 °C device. This can mainly be attrib-uted to the enhanced conductivity of the RGO nanosheets after heat-treatment at 1000 °C. The higher temperature treatment should also be benefi cial for enhanced contact between the SiNW array and the RGO nanosheet, which would enhance the charge transport.T o eliminate the photocurrent interference generated by the SiNW array, we selected the irradiation cases with MIR and THz lasers, which can only excite RGO, to investigate the effects of the reduction degree of RGO on the photodetector performance. Figure 3 c and d show the photoresponses of the RGO-SiNW-1000 °C device and the RGO-SiNW-550 °C device, respectively, measured under the same conditions upon MIR and THz irradiation. It is shown that the RGO -SiNW-1000 °C photodetector exhibited the same stability and reproducibility as the RGO -SiNW-550 °C. Under both MIR and THz irradiation, the steady-state photocurrent of the RGO -SiNW-1000 °C device is, however, much larger than that of the RGO -SiNW-550 °C device. The photocur-rents of the RGO -SiNW-1000 °C device are 5.45 times and 5.68 times as high as those of the RGO-SiNW-550 °C device under the MIR and THz irradiations, respectively. This means that the responsivities (defi ned as the photocurrent gener-ated per unit power of incident light on a photoconductor[ 30 ])of the RGO-SiNW-1000 °C photodetector are 5.45 times and5.68 times as high as those of the RGO-SiNW-550 °C photo-detector under the MIR and THz irradiations, respectively(see Table S3). According to the above methods, t r and t dof the RGO -SiNW-1000 °C photodetectors under the MIR and THz irradiations were calculated from the curves in Figure 3 c and d, and they are approximately equal to those of the RGO -SiNW-550 °C photodetector (see Table S3). Thus, the heat-treatment temperature only had a decisive effect on the responsivity of RGO -SiNW array heterojunc-tion photodetectors. These results suggest that the photode-tector responsivity can be signifi cantly affected by the degree of reduction of RGO.F rom Table S2, it is noted that the photodetector has the highest responsivity for MIR (10.6 µm) among the four irra-diation cases. Because the 8–14 µm MIR wave band is the widest infrared atmospheric window and it mainly stem from the thermal radiation of the object itself, the detection of 8–14 µm MIR is particularly important in medical, industrial and military applications, including detection of cancer or infection, thermal infrared remote sensing and night vision. Therefore, we studied the MIR (10.6 µm) detection perfor-mance of the RGO-SiNW-1000 °C device in detail. Figure S5a shows the photoresponses of the device when irradiated by lasers of different powers. Clearly, the RGO -SiNW-1000 °C photodetector exhibits excellent stability and reproducibility under different irradiation powers. Figure S5b, which shows the photocurrent versus the irradiation power for the photo-detector, indicates that the photocurrent increases linearly9001200150018002100RGO-SiNW-550o C I D /I G = ca. 1.15RGO-SiNW-1000o C I D /I G = ca. 0.97I n t e n s i t y ( A r b . U . )Raman Shift (cm -1)-1.0-0.8-0.6-0.4-0.20.00.20.40.60.8 1.0-11234RGO-SiNW-550°CRGO-SiNW-1000°CC u r r e n t (m A )Voltage (V)In darkP h o t o c u r r e n t (m A )Time (s)P h o t o c u r r e n t (m A )Time (s)(a)(b)(c)(d)F igure 3. C omparison of properties of RGO-SiNW array heterojunctions after heat-treatment at 550 °C (blue) and 1000 °C (red), respectively. (a) Raman spectra. (b) I –V characterizations in the dark. (c) and (d) show their photoresponses at a bias voltage of 1 V upon MIR (10.6 µm) irradiation at 354 mW and upon 2.52 THz (118.8 µm) irradiation at 98 mW, respectively.Y. Cao et al.communicationswith increasing irradiation power. The photodetector respon-sivity was calculated to be 9.0 mA W−1 at a bias voltage of 1 V .I t is well known that human body infrared radiation is in the MIR region, with a peak wavelength of approximately 9–10 µm. O ur RGO -SiNW-1000 °C photodetector showed high MIR (10.6 µm) radiation sensitivity and thus might be suitable for high sensitivity infrared radiation detection of the human body. Figure S6 shows that at the bias voltage of 1 V , when a human hand was close to the device, the RGO-SiNW-1000 °C photodetector could produce an average photo c urrent (ph ) of up to 0.18 mA. The RGO-SiNW array heterojunction photodetector would thus be potentially useful as human infrared sensors.I n summary, we integrated the advantages of RGO and SiNW array in a single photodetector by forming an RGO -SiNW array heterojunction via simple drop-casting of the suspension of GO nanosheets on top of the SiNW array fol-lowed by heat-treatment. The photodetector is photorespon-sive to the VIS (532 nm), NIR (1064 nm), MIR (10.6 µm), and 2.52 THz (118.8 µm) irradiations at room temperature. To our best knowledge, such a photoactivity range from visible (532 nm) up to terahertz (118.8 µm) is the broadest reported to date for graphene-based photodetectors. Upon VIS and NIR irradiations, photoresponse occurs from photo-excitation in both the SiNW array and the RGO , and the photoexcitation of the SiNW array can enhance the photo-detector response speed by 50%. Upon MIR and THz irra-diations, photoresponse occurrs from photoexcitation only in the RGO. Among the four irradiations, the RGO -SiNW array heterojunction photodetector demonstrates the highestresponsivity of 9 mA W−1 to MIR irradiation. The photode-tector can achieve high sensitivity detection of the infrared radiation from the human body, and thus would be poten-tially useful as human infrared sensors. In addition, The photo d etector responsivity can be signifi cantly affected by the reduction degree of RGO nanosheets. These results indi-cate that the RGO -SiNW array heterojunction has great potential for ultra-broadband photodetection applications, especially in the MIR region.E xperimental Section R GO-SiNW Array Heterojunction Fabrication and Characteri-zation : Preparations of GO and SiNW array see the Supporting Information. To fabricate the RGO-SiNW array heterojunction, the as-prepared SiNW array was washed with 5% HF to remove the oxide layer, and was then cleaned with deionized water and dried by an N 2 fl ow. A GO suspension (0.7 mg mL −1 ) was prepared by sonication (1 h) of GO nanosheets in ethanol to produce a reason-ably stable light-brown GO suspension. The GO suspension was then directly drop-casted on top of the SiNW array. The heterojunc-tion was dried in air at 60°C and subjected to further heat-treat-ment in an Ar (95%)-H 2 (5%) atmosphere at 550 or 1000 °C for 3 h. The heterojunction morphology was characterized by a SEM(S-4800, Hitachi, Japan). Transmittances of the RGO nanosheets for all incident wavelengths were measured by drop-casting theRGO nanosheets on a quartz substrate (VIS and NIR) or a single crystal silicon wafer substrate (MIR and THz) followed by 550 °Cheat treatment using UV-Vis-NIR spectrophotometer (Varian Cary 5000), Fourier transform infrared spectrometer (Varian 3100 FT-IR) and pyroelectric detector (SENSOR-und L ASERTECHNIK THz20), respectively. Raman spectra were obtained on a Raman spec-trometer (inVia-Refl ex, Renishaw, U.K.) with 532 nm wavelength incident laser light. The contact angles (CAs) on the SiNW array surface were measured at ambient temperature using a JC contact angle/interface system (Shanghai Zhongchen Digital Technique Apparatus Co., China).D evice Fabrication : The front contact, designed as a circular shape, was produced by depositing an Au ring with 3 mm inner diameter (larger than the light spot diameter) and 7 mm outer diameter using a mask on top of the RGO-SiNW array heterojunc-tion. The back ohmic contact was then formed by coating eutecticGa:In on the back side of the n-Si substrate. [ 31 ]P hotoresponse Measurement : VIS, NIR, MIR and THz irra-diations were achieved using a 532 nm semiconductor laser (218 mW; spot diameter: ∼2.00 mm), a 1064 nm laser (317 mW; spot diameter: ∼3.00 mm), a 10.6 µm CO 2 laser (the laser power can be adjusted continuously; spot diameter: ∼3.00 mm) and a 2.52 THz gas laser (∼100 mW; spot diameter: ∼3.00 mm; FIRL 100, Edinburgh Instruments Ltd.), respectively. The electrical character-istics were collected using a Keithley 2400 SourceMeter.S upporting Inf ormation Supporting Information is available from the Wiley Online Library or from the author.A cknowledgements This work was supported by National Natural Science Founda-tion of China (Grants No. 11104283, 21271177, 11174172, and 11004054), the National Basic Research Program of China (Grant No. 2010CB934103), the National High Technology Research and Development Program (“863” Program) of China(Grant no. 2011AA050525), the CAEP THz Science and Technology Founda-tion (Grant No. CAEPTHZ201211), National Science and TechnologySupport Program of China (Grant No. 2013BAK14B03) and Open Research Fund Program of the State Key Laboratory of Low-Dimen-sional Quantum Physics (Grant No. KF201207). [1]S .N anot ,A .W.C ummings ,C .L.P int ,A .I keuchi ,T .A kiho ,K .S ueoka ,R .H.H auge ,F .L éonard ,J .K ono ,S ci. Rep. 2013,DOI:10.1038/srep01335.[2]A .G assenq ,F .G encarelli ,J .V an Campenhout ,Y .S himura ,R .L oo ,G .N arcy ,B .V incent ,G .R oelkens ,O pt. Express 2012,20,27297.[3]D .S.T sai ,C .A.L in ,W .C.L ien ,H .C.C hang ,Y .L.W ang ,J .H.H e ,A CS Nano 2011,5,7748.[4]S .A dhikary ,Y .A ytac ,S .M eesala ,S .W olde ,A .G.U.P erera ,S .C hakrabarti ,A ppl. Phys. Lett. 2012,101,261114.[5]S .M.M enke ,R .P andey ,R .J.H olmes ,A ppl. Phys. Lett. 2012,101,223301.Ultra-Broadband Photodetector for the Visible to Terahertz Range by Self-Assembling Reduced[6]R .R.N air ,P .B lake ,A .N.G rigorenko ,K .S.N ovoselov ,T .J.B ooth ,T .S tauber ,N .M.R.P eres ,A .K.G eim ,S cience 2008,320,1308.[7]F .X ia ,T .M ueller ,Y .L in ,A .V aldes-Garcia ,P .A vouris ,N at. Nano-technol. 2009,4,839.[8]T .M ueller ,F .X ia ,P .A vouris ,N at. Photonics 2010,4,297.[9]V .R yzhii ,M .R yzhii ,N .R yabova ,V .M itin ,T .O tsuji ,J pn. J. Appl.Phys. 2009,48,04C144.[10]V .R yzhii ,V .M itin ,M .R yzhii ,N .R yabova ,T .O tsuji ,A ppl. Phys.Express 2008,1,063002.[11]M .Y.H an ,B .Özyilmaz ,Y .Z hang ,P .K im ,P hys. Rev. Lett. 2007,98,206805.[12]J .B ai ,X .D uan ,Y .H uang ,N ano. Lett. 2009,9,2083.[13]J .B ai ,X .Z hong ,S .J iang ,Y .H uang ,X .D uan ,N at. Nanotechnol.2010,5,190.[14]D .V.K osynkin ,A .L.H igginbotham ,A .S initskii ,J .R.L omeda ,A .D imiev ,B .K.D imiev ,J .M.T our ,N ature 2009,458,872.[15]L .J iao ,L .Z hang ,X .W ang ,G .D iankov ,H .D ai ,N ature 2009,458,877.[16]Y .Z hang ,T .L iu ,B .M eng ,X .L i ,G .L iang ,X .H u ,Q .J.W ang ,N at.Commun. 2013,4,1811.[17]G .E da ,C .M attevi ,H .Y amaguchi ,H .K im ,M .C hhowalla ,J . Phys.Chem. C 2009,113,15768.[18]B .C hitara ,L .S.P anchakarla ,S .B.K rupanidhi ,C .N.R ao ,A dv.Mater. 2011,23,5419.[19]K .P eng ,Y .X u ,Y .W u ,Y .Y an ,S .T.L ee ,J .Z hu .S mall 2005,1,1062.[20]M .L.Z hang ,K .Q.P eng ,X .F an ,J .S.J ie ,R .Q.Z hang ,S .T.L ee ,N .B.W ong ,J . Phys. Chem. C 2008,112,4444.[21]W .S.H ummers ,R .E.O ffeman ,J . Am. Chem. Soc. 1958,80,1339.[22]X .L i ,H .Z hu ,K .W ang ,A .C ao ,J .W ei ,C .L i ,Y .J ia ,Z .L i ,X . L i , D . W u,A dv. Mater. 2010,22,2743.[23]X .A n ,F .L iu ,Y .J.J ung ,S .K ar ,N ano Lett. 2013,13,909.[24]Z .L iu ,H .T.H uang ,B .L iang ,X .F.W ang ,Z .R.W ang ,D .C hen ,G .Z.S hen ,O pt. Express 2012,20,2982.[25]K .I.B olotin ,K .J.S ikes ,J .H one ,H .L.S tormer ,P .K im ,P hys. Rev.Lett. 2008,101,096802.[26]W .X u ,F .M.P eeters ,T .C.L u ,P hys. Rev. B 2009,79,073403.[27] A . A kturk ,N . G oldsman ,I nternational Conference on Simulationof Semiconductor P rocesses and Devices , H akone , J AP AN , SEP 09–11, 2008.[28]S .G hosh ,B .K.S arker ,A .C hunder ,L .Z hai ,S .I.K hondaker ,A ppl.Phys. Lett. 2010,96,163109.[29]T .K.H ong ,D .W.L ee ,H .J.C hoi ,H .S.S hin ,B .S.K im ,A CS Nano ,2010,4,3861.[30]H .M orkoç,A .D i Carlo ,P .C ingolani ,S olid-State Electron .2002,46,157.[31]X .S hen ,B .S un ,D .L iu ,S .T.L ee ,J . Am. Chem. Soc. 2011,133,19408.Received: O ctober 24, 2013 Revised: D ecember 26, 2013 Published online: M arch 7, 2014。