Fast visible light photoelectric switch based on ultralong single crystalline V 2 O 5 nanobelt

A photoelectric switch with fast response to visible light (<200μs), suitable photosensitivity and excellent repeatability is proposed based on the ultralong single crystalline V2O5 nanobelt, which are synthesized by chemical vapor deposition and its photoconductive mechanism can well be explained by small polaron hopping theory. Our results reveal that the switch has a great potential in next generation photodetectors and light-wave communications. ©2012 Optical Society of America OCIS codes: (230.0250) Optoelectronics; (250.6715) Switching; (040.5150) Photoconductivity. References and links 1. Y. Jiang, W. J. Zhang, J. S. Jie, X. M. Meng, X. Fan, and S. T. Lee, “Photoresponse properties of CdSe singlenanoribbon photodetectors,” Adv. Funct. Mater. 17(11), 1795–1800 (2007). 2. E. Comini, G. Baratto, G. Faglia, M. Ferroni, A. Vomiero, and G. Sberveglieri, “Quasi-one dimensional metal oxide semiconductors: preparation, characterization and application as chemical sensors,” Prog. Mater. Sci. 54(1), 1–67 (2009). 3. T. Zhai, X. Fang, M. Liao, X. Xu, H. Zeng, B. Yoshio, and D. Golberg, “A comprehensive review of onedimensional metal-oxide nanostructure photodetectors,” Sensors (Basel Switzerland) 9(8), 6504–6529 (2009). 4. L. Q. Mai, X. Xu, L. Xu, C. H. Han, and Y. Z. Luo, “Vanadium oxide nanowires for Li-ion batteries,” J. Mater. Res. 26(17), 2175–2185 (2011). 5. T. Y. Zhai, H. M. Liu, H. Q. Li, X. S. Fang, M. Y. Liao, L. Li, H. S. Zhou, Y. Koide, Y. Bando, and D. Golberg, “Centimeter-long V2O5 nanowires: from synthesis to field-emission, electrochemical, electrical transport and photoconductor properties,” Adv. Mater. (Deerfield Beach Fla.) 22(23), 2547–2552 (2010). 6. B. Yan, L. Liao, Y. M. You, X. J. Xu, Z. Zheng, Z. X. Shen, J. Ma, L. M. Tong, and T. Yu, “Single crystalline V2O5 ultralong nanoribbon waveguides,” Adv. Mater. (Deerfield Beach Fla.) 21(23), 2436–2440 (2009). 7. N. V. Joshi, Photoconductivity: Art, Science, and Technology (Marcel Dekker, 1990), Chap. 1. 8. L. C. Hsu, Y. P. Kuo, and Y. Y. Li, “On-chip fabrication of an individual α-Fe2O3 nanobridge and application of ultrawide wavelength visible-infrared photodetector/optical switching,” Appl. Phys. Lett. 94(13), 133108 (2009). 9. L. Peng, J. L. Zhai, D. J. Wang, P. Wang, Y. Zhang, S. Pang, and T. F. Xie, “Anomalous photoconductivity of cobalt-doped zinc oxide nanobelts in air,” Chem. Phys. Lett. 456(4–6), 231–235 (2008). 10. Y. J. Chen, C. L. Zhu, M. S. Cao, and T. H. Wang, “Photoresponse of SnO2 nanobelts grown in situ on interdigital electrodes,” Nanotechnology 18(28), 285502 (2007). 11. X. Y. Xue, T. L. Guo, Z. X. Lin, and T. H. Wang, “Individual core-shell structured ZnSnO3 nanowires as photoconductors,” Mater. Lett. 62(8–9), 1356–1358 (2008). 12. J. B. K. Law and J. T. L. Thong, “Simple fabrication of a ZnO nanowire photodetector with a fast photoresponse time,” Appl. Phys. Lett. 88(13), 133114 (2006). 13. J. Muster, G. T. Kim, V. Krstic, J. G. Park, Y. W. Park, S. Roth, and M. Burghard, “Electrical transport through individual Vanadium Pentoxide nanowires,” Adv. Mater. (Deerfield Beach Fla.) 12(6), 420–424 (2000). 14. T. M. Searle and B. Bowler, “Optical study of the excited state of the V′ centre in MgO,” J. Phys. Chem. Solids 32(3), 591–602 (1971). 15. N. F. Mott and A. M. Stoneham, “The lifetime of electrons, holes and excitons before self-trapping,” J. Phys. C Solid State Phys. 10(17), 3391–3398 (1977). 16. D. Berben, K. Buse, S. Wevering, P. Herth, M. Imlau, and Th. Woike, “Lifetime of small polarons in iron-doped lithium–niobate crystals,” J. Appl. Phys. 87(3), 1034–1041 (2000).


Introduction
Next generation electronic and photoelectronic systems require the elements with smaller size, higher efficiency, and less energy consumption.Photoelectric switch as a key element has been paid more and more attention due to its potential applications in imaging techniques and light-wave communications [1].Generally, a good photoelectric switch should meet requirements in three aspects of photosensitivity, photoresponse speed and repeatability.Due to large surface-to-volume ratio and superior sensitivity to light compared with the traditional film materials, nanostructure materials have been considered to be one of most promising switch materials [2].So far, various one-dimensional (1D) metal-oxide nanostructures such as ZnO, SnO 2 , Ga 2 O 3 , Cu 2 O, Fe 2 O 3 , In 2 O 3 , CdO and CeO 2 have been used to fabricate photoelectric switch for different wavelengths, especially for visible light because of the easiness to get economical light source, however, the above 1D metal-oxide nanostructures under the visible light irradiation reach only a photoresponse time from 20 millisecond (ms) to 500 second (s) [3], limiting their practical application to a large extent.In addition, difficulty to get a macro scale nanostructure is another obstacle because the enough light irradiation area and contacting parts are needed for a practical device.Therefore, it is necessary to explore higher-performance photoelectric switch with shorter response time to visible light, based on new nanostructure materials with a desired size.
Recently, V 2 O 5 nanowires have been extensively studied because of their promising applications in lithium-ion batteries, field-emitters, waveguide and so on [4][5][6].Here we report a photoelectric switch with fast response time to visible light (<200μs), based on highquality ultralong single crystalline V 2 O 5 nanobelt (20-500nm wide, several centimeters long; aspect ratio >10 5 ) fabricated by using chemical vapor deposition (CVD), in which Vanadium powder is initially used instead of conventional VO 2 or V 2 O 5 powder to make the fabrication easier, more timesaving and less harmful.Moreover, a switching test circuit, including a LED light, displays the practicality of the switch (Fig. 1, Single-frame excerpts from Media 1).

Experiments
V 2 O 5 nanobelts were grown by CVD.Vanadium powder (0.5 g, 99.9%) was placed at the center of a horizontal vacuum tube (vacuumized to 4.6 × 10 −3 Pa) furnace as a source.By directly heating the source to 1000 °C at 1000 Pa with an Ar/O 2 gas flow of 4.8 SCCM (standard cubic centimeter per minute) and 0.2 SCCM, respectively, for duration of 5h, V 2 O 5 nanobelts were grown at Si substrate downstream in a lower temperature region 16 cm from the source.The synthesized products were characterized by a field emission scanning electron microscope (FE-SEM)(Hitachi S4800) and a transmission electron microscope (Tecnai G 2 F20 U-TWIN) equipped with selected-area electron diffraction (SAED).For electrical transport measurements, the ultralongV 2 O 5 nanobelts were first dispersed on a SiO 2 (200nm thick)/Si wafer with a desired density.Two electrodes together with their bonding pads were patterned by UV lithography.After the development, a Ti/Au (10 nm/100 nm) film was deposited on the structure followed by a lift-off process.The single V 2 O 5 nanobelt device had a channel length of 2 mm.
The spectral responses for different wavelengths (365-680 nm) were recorded at room temperature by measuring a DC current, by means of using a 50W xenon lamp (the relative xenon lamp light distribution is considered and normalized) and a visible light spectrometer.The bias voltage is 1V.
A red laser with a 671 nm wavelength in continuous wave operation (and a green laser with a 532 nm wavelength was used as substitute) and an Acoustic Optic Modulator with 30 ns pulse were used to test the photoresponse speed.The circuit was supplied by 2 V constant voltage source.To detect the current changing, V 2 O 5 nanobelt was in series with a 1 KΩ resistor which was far less than the resistance of the nanobelt.Thus, the V 2 O 5 nanobelt was under the constant bias of 2V.A low-noise voltage preamplifier (SR 560 with 1MHz bandwidth) was used to amplify the voltage of the 1 KΩ resistor, and connected to the oscilloscope (Tektronix DPO 4104 with 1GHz bandwidth) to acquire the response curve.The whole test was at room temperature in air.The photograph of the V 2 O 5 nanobelts is shown in Fig. 2(a).At the edge of the substrate, most of them are with a length up to several centimeters, which suits for tailoring and constructing practical switch.

The performance of the photoelectric switch based on V 2 O 5 nanobelt
As mentioned earlier, photosensitivity, photoresponse speed and repeatability are often used to judge performance of a photoelectric switch.We fabricated the metal contacts by photolithography to make a photoelectric switch based on a V 2 O 5 nanobelt for these measurements.where R λ refers to the photosensitivity, I ph the current under the light irradiation, I D the dark current, I the photocurrent, P λ the light power density, and S the laser facula area.Figure 3(a) is the testing scheme of the spectra response for different wavelengths (365-680 nm).The result as shown in Fig. 3(b) indicates that the photosensitivity of the nanobelt increases with an increase in wavelength.Since measurements were performed at a constant light power density, the photon density decreases as the wavelength decreases.Consequently the photoninduced carrier concentration also decreases.However, the photosensitivity normalized for constant photon density drops to about zero at short wavelengths, suggesting that photons of higher energies are preferentially adsorbed at or near the semiconductor surface where the recombination rate is much higher [7]. Figure 3(c) shows the current I ph under laser irradiation increases by 1.85 times (671nm) and 1.65 times (532nm) than the dark current I D for a laser power density of 30.57mW cm −2 .Radiation-intensity response experiment shows from Fig. 3(d) that photocurrent (I) is nearly saturated when power density exceeds 10 mW cm −2 , indicating a wide power density range for working.

Photoresponse speed
The next factor assessing photoelectric switch is the photoresponse speed, which is mainly measured by photoresponse time (rising feature time and decay feature time).1.In contrast with the red laser, the relevant photoresponse time of the V 2 O 5 nanobelts under green laser (wavelength 532 nm) irradiation is also shown in Table 1, respectively.Generally, although the photoresponse time under red laser irradiation is a little shorter than that under the green laser irradiation, they are of the same magnitude (100 μs to 300 μs).The time is shorter by 2-6 orders of magnitude than that of the existing 1D metal-oxide nanostructures under the visible light irradiation [8][9][10][11].Even though we break through wavelength limit, the rapidest response time, so far, based on a ZnO Nanowire photodetector can only reach 0.4 ms under UV illumination [12].It should also be noted that the photoresponse speed is meaningful and efficient only when photoresponse time can follow switching frequencies.Additionally, decrease of photocurrent for an excellent switch should not be lower than 50% for a desired switching frequency.The photoelectric switch is found to work with excellent stability in a wide frequency range.For a certain light power density, the photocurrent (I) decreases by <10% when the frequency changes from 10 [Fig.

Photoconductive mechanism
On photoconductive mechanism of V 2 O 5 , J. Muster et al have demonstrated that the dominant conduction mechanism is governed by small polaron hopping [13], and this means that the photocurrent generated by the photons should be closely related to the polaron excitation.The typical value of the lifetime of small polaron is ranged from 1 to 1000 μs [14][15][16].This is in accordance with our relaxation time.Therefore, we believe the response time of V 2 O 5 nanobelts should be determined by the lifetime of excited small polaron.It is interesting to note that the time constant for falling edges is always shorter than that for rising edges, which results in asymmetric curves.This phenomenon is also observed by Y. Jiang et al [1].A common explanation is that traps are assumed to distribute with varying concentration in the bandgap.Therefore we believe that traps and other defect states are involved in this process.Specifically, this difference can be attributed to small polaron self-trap effect.The excited small polaron may firstly fill the traps and then reach the maximum after all the traps are saturated, which induces a delay in reaching the steady photocurrent.In addition, because for the rising time the nanobelt is transiting from highly resistive to relatively conducting, i.e. with decreasing resistance; while for the recovery the nanobelt is from conducting to resistive, with increasing resistance.Therefore, during the rising the RC time constant is larger, which gives a slower response.

Conclusion
In summary, a visible light photoelectric switch has been investigated based on the centimeter-scale single crystalline V 2 O 5 nanobelts fabricated by CVD, in which the Vanadium powder was used for making the fabrication easier, timesaving and less harmful.The excellent photosensitivity, photoresponse speed and repeatability of the switch indicate that it is a good candidate of photoelectric switches for photodetectors and light-wave communications.Particularly, photoresponse time of the photoelectric switch can be shorter than 200 μs, which is 2-6 orders of magnitude shorter than other 1D metal-oxide nanostructures under visible light irradiation.

Fig. 1 .
Fig. 1.Single-frame excerpts from video recordings (Media 1) of V2O5 nanobelt photoelectric switching test.(a) and (b) The LED is turned on/off when the red laser (671nm) irradiates on/off the switch.(c) and (d) The LED is turned on/off when the green laser (532nm) irradiates on/off the switch.

Fig. 2 .
Fig. 2. (a) The centimeter-scale single crystalline V2O5 nanobelts.(b) SEM image of the V2O5 nanobelts.The inset is the cross section of a V2O5 nanobelt.(c) TEM image of a V2O5 nanobelt.(d) HRTEM image of a V2O5 nanobelt.The insert shows selected area electron diffraction pattern indexed with the [010] zone axis.
Figure 2(b) and its inset exhibit the FE-SEM images which demonstrate that the V 2 O 5 nanobelts have very smooth surfaces, which correspond to specific crystallographic planes.An individual nanobelt in Fig. 2(c) was further investigated by high resolution transmission electron microscopy (HRTEM) as shown in Fig. 2(d).The well-resolved lattice fringes of (200) planes of the orthorhombic V 2 O 5 with the interplanar distance of 0.58 nm reflect that the V 2 O 5 nanobelt is single crystallinity one grown along the [001] direction.The inset of Fig. 2(d) depicts the SAED pattern of the nanobelt, the diffraction spots could be indexed to the {200}, {101}, { 101} families of orthorhombic V 2 O 5 structure.#159655 -$15.00USD Received 7 Dec 2011; revised 27 Feb 2012; accepted 6 Mar 2012; published 12 Mar 2012 (C) 2012 OSA

Fig. 3 .
Fig. 3. (a) Testing scheme of spectra response.(b) The spectra response of the V2O5 nanobelt at different wavelengths (365-680 nm).(c) I-V curves of nanobelts unirridiated and irradiated with constant laser (671 nm and 532 nm) power density of 30.57mW cm −2 .(d) Curves of photocurrent versus light power density at deferent wavelength.The photosensitivity is usually described by spectra response and radiation-intensity response, and it is defined as ( ) ph D R I I P S I P S λ λ λ =− = (1)

Figure 4 Figures 4
photoelectric switch.The photocurrent has relationship with the rising feature time τ r and decay feature time constant τ d as max [1 exp( )] r

4
Figures 4(b) and 4(c) reveal the photocurrent (the data are normalized to the highest photocurrent under laser irradiation) of the red laser (671 nm) can repeatedly switch at different light power density.Both the rising and falling edges of the V 2 O 5 nanobelt can be well fitted by the relationship mentioned above.The photocurrent of the V 2 O 5 nanobelt also depends on light power density.Figures 4(b)-(d) show the photocurrent of the nanobelt irradiated with light (wavelength 671 nm) power density for 2.55, 5.10, 10.19, 15.29, 20.38 and 30.57mW cm -2 , respectively.The photocurrent increases by small margin as the light power density rises up, especially for light power density of above 10.19 mW cm −2 .Fitting the data from Fig. 4(b)-(c), the photoresponse time of the V 2 O 5 nanobelt photoelectric switch under different light power density (red laser with a wavelength of 671 nm) is shown in Table1.In contrast with the red laser, the relevant photoresponse time of the V 2 O 5 nanobelts under green laser (wavelength 532 nm) irradiation is also shown in Table1, respectively.Generally, although the photoresponse time under red laser irradiation is a little shorter than that under the green laser irradiation, they are of the same magnitude (100 μs to 300 μs).The time is shorter by 2-6 orders of magnitude than that of the existing 1D metal-oxide nanostructures under the visible light irradiation[8][9][10][11].Even though we break through wavelength limit, the rapidest response time, so far, based on a ZnO Nanowire photodetector can only reach 0.4 ms under UV illumination[12].It should also be noted that the photoresponse speed is meaningful and efficient only when photoresponse time can follow switching frequencies.Additionally, decrease of photocurrent for an excellent switch should not be lower than 50% for a desired switching frequency.The photoelectric switch is found to work with excellent stability in a wide frequency range.For a certain light power density, the photocurrent (I) decreases by <10% when the frequency changes from 10 [Fig.4(b)] to 1000 Hz [Fig.4(c)].As depicted in Fig. 4(d), the device still function well ((I 5000Hz -I 10Hz ) / I 10Hz > 70%) for a switching frequency up to 5000 Hz.
Figures 4(b) and 4(c) reveal the photocurrent (the data are normalized to the highest photocurrent under laser irradiation) of the red laser (671 nm) can repeatedly switch at different light power density.Both the rising and falling edges of the V 2 O 5 nanobelt can be well fitted by the relationship mentioned above.The photocurrent of the V 2 O 5 nanobelt also depends on light power density.Figures 4(b)-(d) show the photocurrent of the nanobelt irradiated with light (wavelength 671 nm) power density for 2.55, 5.10, 10.19, 15.29, 20.38 and 30.57mW cm -2 , respectively.The photocurrent increases by small margin as the light power density rises up, especially for light power density of above 10.19 mW cm −2 .Fitting the data from Fig. 4(b)-(c), the photoresponse time of the V 2 O 5 nanobelt photoelectric switch under different light power density (red laser with a wavelength of 671 nm) is shown in Table1.In contrast with the red laser, the relevant photoresponse time of the V 2 O 5 nanobelts under green laser (wavelength 532 nm) irradiation is also shown in Table1, respectively.Generally, although the photoresponse time under red laser irradiation is a little shorter than that under the green laser irradiation, they are of the same magnitude (100 μs to 300 μs).The time is shorter by 2-6 orders of magnitude than that of the existing 1D metal-oxide nanostructures under the visible light irradiation[8][9][10][11].Even though we break through wavelength limit, the rapidest response time, so far, based on a ZnO Nanowire photodetector can only reach 0.4 ms under UV illumination[12].It should also be noted that the photoresponse speed is meaningful and efficient only when photoresponse time can follow switching frequencies.Additionally, decrease of photocurrent for an excellent switch should not be lower than 50% for a desired switching frequency.The photoelectric switch is found to work with excellent stability in a wide frequency range.For a certain light power density, the photocurrent (I) decreases by <10% when the frequency changes from 10 [Fig.4(b)] to 1000 Hz [Fig.4(c)].As depicted in Fig. 4(d), the device still function well ((I 5000Hz -I 10Hz ) / I 10Hz > 70%) for a switching frequency up to 5000 Hz.