CHEMPHYSCHEM ARTICLES DOI: 10.1002/cphc.201402201
Effect of Electrical Contact on the Performance of Bi2S3 Single Nanowire Photodetectors Renxiong Li,[a] Juehan Yang,[a] Nengjie Huo,[a, b] Chao Fan,[a] Fangyuan Lu,[a] Tengfei Yan,[a] Zhongming Wei,*[a, c] and Jingbo Li*[a, b] Bi2S3 single-crystalline nanowires are synthesized through a hydrothermal method and then fabricated into single nanowire photodetectors. Due to the different contact barrier between the gold electrode and Bi2S3 nanowires, two kinds of devices with different electrical contacts are obtained and their photoresponsive properties are investigated. The non-ohmic contact devices show larger photocurrent gains and shorter response times than those of ohmic contact devices. Furthermore, the
influence of a focused laser on the barrier height between gold and Bi2S3 is explored in both kinds of devices and shows that laser illumination on the AuBi2S3 interface can greatly affect the barrier height in non-ohmic contact devices, while keeping it intact in ohmic contact devices. A model based on the surface photovoltage effect is used to explain this phenomenon.
1. Introduction In recent decades, one-dimensional or quasi-one-dimensional nanowires (NWs) have attracted tremendous attention due to their unique properties, such as large surface to volume ratio and quantum size effect. With ordered alignment and excellent crystallinity, NWs are an ideal choice to investigate the intrinsic transport properties of different materials and to obtain devices with high performance. Some NWs have super performances as field-effect transistors (FET).[1–3] Some NWs, which are very sensitive to light and gases, can be widely used in the fields of photodetectors, photovoltaics, photoelectricity, optical switches, optical interconnections, and gas sensors.[4–8] Today, more and more people are considering two-dimensional materials, which are also low-dimensional materials, and possess high optical and electronic properties.[9–11] However, their super performance depends greatly on the thickness of their flakes. Indirect to direct transformation in the band gap accompanies a decrease in the thickness of these 2D materials.[12] To the best of our knowledge, the most effective and frequently used method to obtain high-quality 2D materials is still micromechanical cleavage. Other methods, such as epitaxial growth, [a] R. Li, J. Yang, N. Huo, C. Fan, F. Lu, T. Yan, Dr. Z. Wei, Prof. J. Li State Key Laboratory for Superlattices and Microstructures Institute of Semiconductors, Chinese Academy of Sciences P.O. Box 912, Beijing 100083 (P.R. China) E-mail: [email protected] [email protected] [b] N. Huo, Prof. J. Li Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei, Anhui 230026 (P.R. China) [c] Dr. Z. Wei Nano-Science Center & Department of Chemistry University of Copenhagen, Universitetsparken 5 2100 Copenhagen (Denmark) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402201.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
chemical vapor deposition, and liquid exfoliation, are still challenged by uncontrollable traps and a large area of growth.[13, 14] With all of these methods, it is difficult to distinguish singlelayer from few layer samples.[15] This may greatly limit their application in industrial production. Owing to the ongoing mature advancement in NW growth and fabrication methodologies, many NW devices have been fabricated and many new device structures have been investigated.[16–20] So far, many works had been reported on NW photodetectors and a deeper understanding of NW photodetectors has been achieved.[21–23] Compared with their bulk counterparts, NW photodetectors show great advantages in photoconductivity gains, response times, and device sizes that are beneficial to their utilization in integrated circuits.[24, 25] Bi2S3, which is an important n-type semiconductor with a narrow band gap range from 1.3 to 1.7 eV,[26, 27] has been widely investigated in visible-light photocatalysis, thermoelectricity, and photodiode arrays.[28–30] Its good conductivity and relatively easy synthetic process further extend its application to photodetectors and solar cells.[31, 32] Previously, Bi2S3 nanorod photodetectors with a photoconductive gain of 1.1 and solution-processed Bi2S3 nanocrystalline photodetectors with a photoconductive gain of 10 and a temporal response of 10 ms have been reported.[33, 34] In previous work, we investigated the photoresponsive behavior of Bi2S3 NW films at low temperatures, which was distinctly different from that at room temperature.[35] Herein, we fabricated photodetectors based on long single-crystalline Bi2S3 NWs with different electrical contacts and explored their respective performance. We synthesized long Bi2S3 NWs and investigated their detailed electrical contact effects on the performance of NW photodetectors. We obtained Bi2S3 single NW photodetectors with high photoconductive gains of about 10 in ohmic contact and 100 in non-ohmic contact devices. The results showed that non-ohmic contact devices with rectifying behavior possessed higher photoconChemPhysChem 2014, 15, 2510 – 2516
2510
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
ductive gains and shorter response times than those of ohmic contact devices. Furthermore, by illuminating the different parts of the NW photodetector, we found that a focused laser could effectively influence the barrier height between gold and Bi2S3 in non-ohmic contact devices, which was distinctly different from the behavior of ohmic contact ones. A simple model of the surface photovoltage (SPV) effect is proposed to explain our experiment result.
Experimental Section The Bi2S3 NWs used for fabricating the metal–semiconductor–metal (MSM) photodetectors were synthesized by a hydrothermal process reported in the literature.[36] The obtained products were characterized by a series of techniques. X-ray diffraction (XRD) was performed by using Cu Ka irradiation with a Pgeneral XD-3 diffractometer. Field-emission scanning electron microscopy (FESEM) images were acquired on a NanoSEM650 microscope instrument by using an accelerating voltage of 10 kV. The Raman scattering properties were recorded on a micro-Raman spectrometer by using a l = 632.8 nm laser line as the excitation source. The Raman signals were collected at room temperature through a 100 objective with a background configuration. High-resolution transmission electron microscopy (HRTEM) images and Energy-dispersive X-ray spectroscopy (EDS) results were obtained by using a Tecnai G2 F20 STWIN transmission electron microscopy under a working voltage of 200 kV.
phology of as-synthesized NWs. In Figure 1 b, the observed lattice spacings of 0.358 and 0.398 nm can be indexed to (130) and (001) planes in the crystal structure, respectively. This demonstrated that Bi2S3 NWs grew predominantly along the c axis. The SAED pattern for the Bi2S3 NW is given in the inset of Figure 1 b and indicates the good crystallinity of the as-synthesized samples. The XRD pattern and EDS results were obtained to further characterize our sample. The XRD patterns of the assynthesized samples in Figure 1 c can be indexed to the orthorhombic phase of Bi2S3 (JCPDS 17-0320). The EDS results in Figure 1 d further confirm that our samples are mainly composed of bismuth and sulfur, in which the peak of copper comes from the copper grid used in the TEM test. The simulated crystal structure shown in Figure 1 e schematically illustrates the growth direction of as-synthesized Bi2S3 NWs, which agrees with the TEM results. A FESEM image of large-scale Bi2S3 NWs is presented in Figure S1 a in the Supporting Information. The lengths of the NWs range from several micrometers to tens of
The Bi2S3 MSM photodetectors were fabricated as follows: a small amount of Bi2S3 NWs were dispersed in ethanol under ultrasonication and dropped onto the SiO2/ Si substrate (300 nm thick SiO2) to acquire discrete, single NWs, and then a pair of Au electrodes were evaporated onto them with a gap of 25 mm. Photocurrent measurements were performed on an Agilent B2902 instrument with a focused green laser of l = 532 nm as the light source (light spot of 2 mm).
2. Results and Discussion 2.1. Fabrication and Characterization of Bi2S3 NWs and Photodetectors The morphology of Bi2S3 NW samples were investigated by HRTEM (Figure 1 a and b). Figure 1 a shows the typical mor-
Figure 1. HRTEM images of as-synthesized Bi2S3 NWs with a scale bar of a) 1 mm and b) 2 nm. Inset: the selectedarea electron diffraction (SAED) pattern. c) XRD patterns and d) EDS spectrum of as-synthesized Bi2S3 NWs. e) Simulated crystal structure of Bi2S3 NW with a (001) growth direction.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2511
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
micrometers and the diameters range from tens of nanometers to hundreds of nanometers. The crystal structure was further examined by Raman spectroscopy under different exciting powers (see Figure S1 b in the Supporting Information). The peaks located at 70, 93, 123, 186, 236, and 261 cm1 all fit well with Bi2S3.[37, 38] Among them, the peaks at 186, 236, and 261 cm1 are sensitive to the excitation powers. The peak at 123 cm1 emerged only at a high laser power of 0.1 mW; this was due to the laser heating effect.[39] All of these Raman characteristics confirmed that our Bi2S3 NWs were pure and of high quality. Figure 2 a shows an illustration of a typical MSM photodetector fabricated on the SiO2/Si substrate. Bi2S3 NWs were first
detectors. It is hard to precisely control the electrical contacts in our fabrication process. However, further exploration revealed that such rectifying behavior mainly originated from the distinct difference in contacts between the two opposite AuBi2S3 junctions. The thickness of gold electrodes (80 nm) evaporated on the NW devices was smaller than the diameter of Bi2S3 NW (500 nm), as shown in the AFM images of Figure S2 in the Supporting Information. The Bi2S3 NWs may not be in a strong contact with the gold electrodes, so this uncertainty could cause a large contact barrier at one side of metal–semiconductor (M–S) junction (Schottky contact) and a smaller contact barrier at the other side (ohmic or weak Schottky contact). Then the rectifying behavior emerged. Such incomplete cover caused the Au Bi2S3 contacts to be easily affected by post-treatment, such as the soaking treatment of methanol or ethanol, since methanol (ethanol) may easily permeate into the interface and change the interface states. Details of this exploration are presented in Figure S3 in the Supporting Information. Similar phenomena can be found in ZnO nanobelt (NW) dielectrophoresis experiments performed by Lao and co-workers.[42] Further analysis on the non-ohmic contact device is shown in Figure 2 c and d. In Figure 2 c (and inset), the enlarged rectifying behavior and corresponding I–V curve in logarithmic coordinates are presented. This rectifying characteristic was very similar to that of a Schottky diode and indicated that only one side of Figure 2. a) Schematic illustration of a Bi2S3 NW fabricated photodetector. b) I–V curves corresponding to different AuBi2S3 formed the large barricontact devices measured under dark conditions. Inset: a typical SEM image of the as-fabricated NW photodetecer. To analyze the transport betor. c) Enlarged I–V characteristics of a typical non-ohmic contact device. d) The I–V characteristics of the nonhavior, herein we use the Schottohmic contact device at forward bias. The gray line displays the best fit to Equation (1). ky diode model to describe the rectifying characteristics for claridropped onto SiO2/Si substrate and the gold electrodes were ty. The current can be defined from Equation (1):[43] evaporated onto them by using a gold wire of 25 mm in diameter as the shadow. Among all 20 devices we fabricated, I–V curves showed two typical kinds of electrical contacts, which V Vth I ¼ I ½exp e 1 ð1Þ 0 appeared randomly (Figure 2 b). Apart from the broken devinkT ces, six of them showed ohmic contacts and eight of them showed non-ohmic contacts. The non-ohmic devices always displayed an unsymmetrical I–V characteristic with a rectifying behavior (the maximum on–off current ratio was 10 in our exin which I0 is the saturation current, Vth is the threshold voltperiment). This result was also reported previously.[40, 41] The age, k is the Bolzmann constant, T is the absolute temperature, and n is the ideality factor that describes the deviation of the reason for the formation of such clear rectifying characteristics diode from an ideal Schottky diode. For our device, the ideality needs to be comprehensively identified. Defects, the interface factor was derived to be 2.46 from the results shown in Figstate, and the possibility of an amorphous layer outside the ure 2 d; this indicates a barrier has formed between gold and Bi2S3 NW are all parameters that could contribute to the large the Bi2S3 interface.[44] contact resistance in the AuBi2S3 junctions of our MSM photo 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2512
CHEMPHYSCHEM ARTICLES 2.2. Photoresponsive Properties of Non-Ohmic Contact Devices Figure 3 shows typical photoresponsive behavior of non-ohmic contact devices. In Figure 3 a, the I–V curves under different powers of green laser are displayed. The bias voltages are tested from 1 to + 1 V to avoid probable damage to the Bi2S3 NW devices. Figure 3 b shows the curves for photocurrent
www.chemphyschem.org tions easily. In this sense, the effect of trap states on prolonging the response time was screened and the photocurrent reduced rapidly with a high response speed. 2.3. Photoresponsive Properties of Ohmic Contact Devices
Besides the non-ohmic contact devices, the photoresponsive performance of ohmic contact devices is also displayed in Figure 4. In Figure 4 a, the I–V curves under different powers of green laser all show the nearlinear characteristics, which indicates good ohmic contacts. The corresponding I–V curves in the logarithmic coordinates are also shown in inset of Figure 4 a; this further reveals the good symmetric properties. However, regardless of which part of the Bi2S3 NWs were illuminated by the focused laser, all curves showed very good symmetrical shapes. The curves for photocurrent versus time are presented in Figure 4 b with an applied bias of 0.01 V, and the photocurrent curves fit well with a previous description of a typical twostage process:[47] rapid (t1) and slow (t2) increase. The slow increase and decrease stages in the photocurrent are caused by Figure 3. Measurements performed on non-ohmic contact devices. a) I–V curves measured under different laser the trap states in our samples. powers. Inset: the corresponding I–V curves in logarithmic coordinates. b) Photoresponsive curves of Bi2S3 NW The minimum response time in photodetectors measured by periodic irradiation of the laser at an interval of 10 s (applied bias 0.1 V). Response time of the photocurrent by c) turning on and d) turning off the laser (applied bias 0.1 V, laser power 300 mW). the ohmic devices was derived to be 1 s under illumination of 900 mW, from the results shown versus time by periodically turning on and off the laser. All in Figure 4 c and d; this much longer than that of non-ohmic curves are measured under the same conditions with a bias of devices because no barrier was formed at the AuBi2S3 inter0.1 V. The response time is calculated to be about 50 ms from face in the ohmic devices. Electrons can flow through the juncFigure 3 c and d. It is known that the response time for phototions freely; thus the electrons trapped during the illumination detectors is mainly controlled by the trap states in the samprocess can also escape from the trap states and flow through ples.[45, 46] For non-ohmic contact devices, there was a large barthe junctions. In this way, the photocurrent is reduced slowly with a long response time. rier at the AuBi2S3 interfaces. If the focused laser illuminated the AuBi2S3 interface, due to the SPV effect, many electrons in Bi2S3 were excited from the valence band to conduction band 2.4. Comparison of the Two Electrical Contact Devices and then separated with the help of built-in potential. In this sense, light caused the Bi2S3 energy band to bend at the Au Figure 5 shows photoconductivity gains under different laser powers. Clearly, both kinds of devices showed the same trend: Bi2S3 interface and reduced the contact barrier. It is easier for the photoconductivity gains increased with increasing laser electrons to move across the junctions and the photocurrent power; this indicates that higher laser power can excite more increased hugely (details of the SPV effect are discussed in Secelectron–hole (e–h) pairs in the NWs. Such an increase in the tion 2.5). However, after turning off the laser, the photoexcited photocurrent gains became saturated if the laser power was electrons are no longer generated. The contact barrier rehigh enough and then the photoconductivity gains maintained turned and the electrons were blocked again from flowing a fixed value.[48] However, regardless of the power at which the through the junctions. Photoexcited electrons captured by trap states during the illumination process slowly escaped laser illuminated, the photoconductivity gains were larger in from the trap states, but no longer flowed through the juncnon-ohmic contact devices than those in ohmic contact devi 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2513
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org hugely in the non-ohmic devices, whereas they were nearly the same in ohmic contact devices. The lower bias voltage means smaller dark currents and higher photoconductivity gains. This phenomenon was much clearer in the non-ohmic contact devices. Furthermore, the photocurrent versus illumination positions were also determined and are presented in Figure S4 in the Supporting Information; this indicated homogeneity in our Bi2S3 NWs. The slight difference between them may be caused by the laser heating effect, which generated some thermoelectric current.[49]
2.5. Effect of Laser Illumination on the Barrier Height of MSM Photodetectors Figure 4. Measurements performed on the ohmic devices. a) I–V curves measured under different laser powers. Inset: the corresponding I–V curves in logarithmic coordinates. b) Photoresponsive curves of Bi2S3 NW photodetectors measured by periodic irradiation of the laser at an interval of 10 s (applied bias 0.01 V). Response time of the photocurrent by c) turning on and d) turning off the laser (applied bias 0.01 V, laser power 900 mW).
We also compared the effect of the laser on the position of the zero-point voltage (V0) in both kinds of electrical contact devi-
Table 1. Photoconductivity gains under different bias voltages in nonohmic and ohmic devices (laser power of 300 mW).
Figure 5. Photoconductivity gains under different illumination powers (the applied bias in non-ohmic and ohmic contact devices is 0.1 and 0.01 V, respectively).
ces. This result implied that there was a better photoresponsive performance in non-ohmic contact devices, which can be explained by the smaller dark current in non-ohmic devices due to the existence of a large contact barrier. The influence of applied bias on photoconductivity gains was also studied and the results are presented in Table 1. All measurements were recorded under the same conditions with a laser power of 300 mW. The photoconductivity gains varied 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Applied bias
1 V
0.5 V
0.1 V
0.1 V
0.5 V
1V
non-ohmic ohmic
82 7.1
1328.4
222 9.8
133 9.4
94 7.6
70 6.1
ces. In these measurements, the laser was illuminated on different parts of the Bi2S3 NWs as shown in the insets of Figure 6 a and b. All data are derived from the enlarged I–V characteristics in the logarithmic coordinates given in Figures S5 and S6 in the Supporting Information. Clearly, the laser effectively influenced V0 in the non-ohmic contact device, whereas V0 remained intact in the ohmic contact device. In the nonohmic contact device, V0 moves toward negative voltage gradually with increasing laser power. For the non-ohmic contact devices, V0 is close to Vth. Under laser illumination, the reduced barrier height between the AuBi2S3 contacts caused Vth to shift. However, in the ohmic contact device, V0 remained intact at 0 V regardless of the power of the focused laser used. This is completely different from the non-ohmic contact devices. Because no barrier formed in the ohmic contact device, the SPV effect made no difference. Moreover, the V0 shifts with illumination position were also investigated and are presented in Figure 6 b (see also the Supporting Information). For both nonohmic and ohmic contact devices, V0 remained intact when ilChemPhysChem 2014, 15, 2510 – 2516
2514
CHEMPHYSCHEM ARTICLES
Figure 6. Shifts of zero-point voltage under different illumination powers in non-ohmic and ohmic contact devices upon illumination at the a) M–S interface and b) middle part of the NW in the gap. Left insets: schematic representations of the illumination positions of the Bi2S3 NWs; right insets: the corresponding optical images of the NW device.
luminated on the Bi2S3 NWs. This phenomenon was to be expected, since the SPV effect only took place at the M–S interface with the existence of a contact barrier. When illuminated at the local part of the Bi2S3 NWs, as shown in the inset of Figure 6 b, no electrons aggregated at the M–S interface and the barrier height changed slightly. Detail analysis of the SPV effect is presented as follows: The SPV can be defined as the illumination-induced change in the surface potential.[50, 51] If two materials with different Fermi energies are in contact, a contact barrier is generated at the interface. At the M–S interface, due to this built-in potential, there is a depletion region with width W (Figure 7 a). Under illumination, the photoexcited e–h pairs were separated at the depletion region, the holes moved towards the surface of the semiconductor, and induced electrons to aggregate in the metal at the side of the interface. In this way, the depletion region can be seen as a high-doping layer and the electron concentration is controlled by the light. Considering the image–force effect and the abrupt transition junction approximation, the variation of barrier height caused by the electron concentration at the depletion layer can be written simply as Equations (2) and (3):[52]
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 7. a) Schematic illustration of the cross section of the AuBi2S3 interface: D is the diameter of our NW and W is the width of the depletion region. b) Band alignment of the M–S interface with no illumination. c) Band alignment of the M–S interface under illumination. The dashed lines represent variation of the energy band. d) The barrier height variation versus laser power (illuminated at the M–S interface as shown in figure 6a).
rffiffiffiffiffiffi Q 4p
ð2Þ
Pinc GS hv
ð3Þ
q DF ¼ eS Q ¼ Rabs
in which Q is the sum of holes aggregated at the depletion region by the laser; Rabs, Pinc, hn, G, and S are the absorption rate of the semiconductor, the power of the laser, the photon energy of the laser, the generation rate of photons to create e–h pairs, and the separation rate of e–h pairs at the surface, respectively. This square-root relationship fits well with our results shown in Figure 7 d.
3. Conclusions Long, single-crystalline Bi2S3 NWs were successfully synthesized and fabricated into single NW photodetectors. During our fabrication process, two kinds of electrical contacts formed: nonohmic and ohmic. Detailed investigations on both kinds of electrical contact devices were performed to explore their reChemPhysChem 2014, 15, 2510 – 2516
2515
CHEMPHYSCHEM ARTICLES spective photoresponsive properties. Non-ohmic contact devices possessed higher photocurrent gains ( 100) and faster response speeds (50 ms), whereas ohmic contact devices possessed smaller photocurrent gains ( 10) and slow response speeds (1000 ms). Furthermore, if the laser illuminated one side of the AuBi2S3 contacts, the position of the zero-point voltage (V0) could be effectively influenced in non-ohmic contact devices. The variation of barrier height at the AuBi2S3 interface was believed to cause this phenomenon. Due to the SPV effect, photoexcited electrons could aggregate at the interface of AuBi2S3, change the barrier height, and lead to the shift in the zero-point voltage. These findings further expand our knowledge about the fabrication of NW photodetectors and may contribute to future photodetector designs.
Acknowledgements This work was supported by the National Natural Science Foundation of China under Grant No. 91233120 and the National Basic Research Program of China (2011CB921901). Keywords: bismuth · nanostructures · photochemistry · semiconductors · sulfur [1] X. Duan, Y. Huang, Y. Cui, J. Wang, C. M. Lieber, Nature 2001, 409, 66 – 69. [2] Y. Cui, Z. Zhong, D. Wang, W. U. Wang, C. M. Lieber, Nano Lett. 2003, 3, 149 – 152. [3] D. Chen, Z. Liu, B. Liang, X. Wang, G. Shen, Nanoscale 2012, 4, 3001 – 3012. [4] R. Yan, D. Gargas, P. Yang, Nat. Photonics 2009, 3, 569 – 576. [5] J. Wang, M. S. Gudiksen, X. Duan, Y. Cui, C. M. Lieber, Science 2001, 293, 1455 – 1457. [6] C. Li, D. Zhang, B. Lei, S. Han, X. Liu, C. Zhou, J. Phys. Chem. B 2003, 107, 12451 – 12455. [7] L. Peng, L. Hu, X. Fang, Adv. Mater. 2013, 25, 5321 – 5328. [8] R. M. Penner, Annu. Rev. Anal. Chem. 2012, 5, 461 – 485. [9] D. J. Late, B. Liu, H. S. S. Ramakrishna Matte, V. P. Dravid, C. N. R. Rao, ACS Nano 2012, 6, 5635 – 5641. [10] D. J. Late, B. Liu, J. Luo, A. Yan, H. S. S. Ramakrishna Matte, M. Grayson, C. N. R. Rao, V. P. Dravid, Adv. Mater. 2012, 24, 3549 – 3554. [11] J. Luo, D. Late, I. Wu, K. Biswas, M. Kanatzidis, M. Grayson, AIP Conf. Proc. 2011, 1416, 135-138. [12] D. J. Late, Y. K. Huang, B. Liu, J. Acharya, S. N. Shirodkar, J. Luo, A. Yan, D. Charles, U. V. Waghmare, V. P. Dravid, C. N. R. Rao, ACS Nano 2013, 7, 4879 – 4891. [13] R. V. Kashid, D. J. Late, S. S. Chou, Y. K. Huang, M. De, D. S. Joag, M. A. More, V. P. Dravid, Small 2013, 9, 2730 – 2734. [14] H. S. S. Ramakrishna Matte, A. Gomathi, A. K. Manna, D. J. Late, R. Datta, S. K. Pati, C. N. R. Rao, Angew. Chem. Int. Ed. 2010, 49, 4059 – 4062; Angew. Chem. 2010, 122, 4153 – 4156. [15] D. J. Late, B. Liu, H. S. S. Ramakrishna Matte, C. N. R. Rao, V. P. Dravid, Adv. Fun. Mater. 2012, 22, 1894 – 1906. [16] J. Hu, T. W. Odom, C. M. Lieber, Acc. Chem. Res. 1999, 32, 435 – 445. [17] C. M. Lieber, MRS Bull. 2003, 28, 486 – 491. [18] Z. Zhong, D. Wang, Y. Cui, M. W. Bockrath, C. M. Lieber, Science 2003, 302, 1377 – 1379.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [19] P. Fan, K. C. Y. Huang, L. Cao, M. L. Brongersma, Nano Lett. 2013, 13, 392 – 396. [20] B. Ryu, Y. T. Lee, K. H. Lee, R. Ha, J. H. Park, H. Choi, S. Im, Nano Lett. 2011, 11, 4246 – 4250. [21] H. Kind, H. Yan, B. Messer, M. Law, P. Yang, Adv. Mater. 2002, 14, 158 – 160. [22] Y. Gu, J. P. Romankiewicz, J. K. David, J. L. Lensch, L. J. Lauhon, Nano Lett. 2006, 6, 948 – 952. [23] A. Armstrong, G. T. Wang, A. A. Talin, J. Electron. Mater. 2009, 38, 484 – 489. [24] C. Soci, A. Zhang, X. Bao, H. Kim, Y. Lo, D. Wang, J. Nanosci. Nanotechnol. 2010, 10, 1430 – 1449. [25] Y. Li, F. Qian, J. Xiang, C. M. Lieber, Mater. Today 2006, 9, 18 – 27. [26] C. Ye, G. Meng, Z. Jiang, Y. Wang, G. Wang, L. Zhang, J. Am. Chem. Soc. 2002, 124, 15180 – 15181. [27] J. D. Desai, C. D. Lokhhande, Mater. Chem. Phys. 1995, 41, 98 – 103. [28] S. Liufu, L. Chen, Q. Yao, C. Wang, Appl. Phys. Lett. 2007, 90, 112106. [29] Z. Zhang, W. Wang, L. Wang, S. Sun, ACS Appl. Mater. Interfaces 2012, 4, 593 – 597. [30] B. D. Nayak, H. N. Acharya, G. B. Miitra, S. Tanhanker, Thin Solid Films 1983, 110, 165 – 170. [31] H. Li, J. Yang, J. Zhang, M. Zhou, RSC Adv. 2012, 2, 6258 – 6261. [32] A. K. Rath, M. Bernechea, L. Martinez, G. Konstantatos, Adv. Mater. 2011, 23, 3712 – 3717. [33] L. Li, R. Cao, Z. Wang, J. Li, L. Qi, J. Phys. Chem. C 2009, 113, 18075 – 18081. [34] G. Konstantatos, L. Levina, J. Tang, E. H. Sargent, Nano Lett. 2008, 8, 4002 – 4006. [35] R. Li, Q. Yue, Z. Wei, J. Mater. Chem. C 2013, 1, 5866 – 5871. [36] Y. Yu, C. H. Jin, R. H. Wang, Q. Chen, L. Peng, J. Phys. Chem. B 2005, 109, 18772 – 18776. [37] K. Trentelman, J. Raman Spectrosc. 2009, 40, 585 – 589. [38] Y. W. Koh, C. S. Lai, A. Y. Du, E. R. T. Tiekink, K. P. Loh, Chem. Mater. 2003, 15, 4544 – 4554. [39] Y. Zhao, K. T. E. Chua, C. K. Gan, J. Zhang, B. Peng, Z. Peng, Q. Xiong, Phys. Rev. B 2011, 84, 205330 – 205336. [40] H. Bao, X. Cui, C. Li, Y. Gan, J. Zhang, J. Guo, J. Phys. Chem. C 2007, 111, 12279 – 12283. [41] K. Yao, Z. Zhang, X. Liang, Q. Chen, L. Peng, J. Phys. Chem. B 2006, 110, 21408 – 21411. [42] C. Lao, J. Liu, P. Gao, L. Zhang, D. Davidovic, R. Tummala, Z. Wang, Nano Lett. 2006, 6, 263 – 266. [43] A. G. Milnes, D. L. Feucht, Heterojunctions and Metal – Semiconductor Junctions, Academic Press, New York, 1972, pp. 34 – 57. [44] B. L. Sharma, Metal – Semiconductor Schottky Barrier Junctions and Their Applications, Plenum Press, New York, 1984, pp. 2 – 17. [45] D. Han, D. C. Melcher, E. A. Schiff, M. Silver, Phys. Rev. B 1993, 48, 8658 – 8666. [46] R. P. Barclay, J. Phys. Condens. Matter 1994, 6, 9721 – 9728. [47] K. Schwarzburg, F. Willig, Appl. Phys. Lett. 1991, 58, 2520 – 2522. [48] C. Soci, A. Zhang, B. Xiang, S. A. Dayeh, D. P. R. Aplin, J. Park, X. Y. Bao, Y. H. Lo, D. Wang, Nano Lett. 2007, 7, 1003 – 1009. [49] B. Varghese, R. Tamang, E. S. Tok, S. G. Mhaisalkar, C. H. Sow, J. Phys. Chem. C 2010, 114, 15149 – 15156. [50] K. H. Yoon, J. K. Hyun, J. G. Connell, I. Amit, Y. Rosenwaks, L. J. Lauhon, Nano Lett. 2013, 13, 6183 – 6188. [51] Z. Zhang, J. T. Yates, Jr., Chem. Rev. 2012, 112, 5520 – 5551. [52] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, Wiley, Hoboken, 2007, pp. 111 – 117.
Received: February 7, 2014 Published online on July 30, 2014
ChemPhysChem 2014, 15, 2510 – 2516
2516
Effect of Electrical Contact on the Performance of Bi2S3 Single Nanowire Photodetectors Renxiong Li,[a] Juehan Yang,[a] Nengjie Huo,[a, b] Chao Fan,[a] Fangyuan Lu,[a] Tengfei Yan,[a] Zhongming Wei,*[a, c] and Jingbo Li*[a, b] Bi2S3 single-crystalline nanowires are synthesized through a hydrothermal method and then fabricated into single nanowire photodetectors. Due to the different contact barrier between the gold electrode and Bi2S3 nanowires, two kinds of devices with different electrical contacts are obtained and their photoresponsive properties are investigated. The non-ohmic contact devices show larger photocurrent gains and shorter response times than those of ohmic contact devices. Furthermore, the
influence of a focused laser on the barrier height between gold and Bi2S3 is explored in both kinds of devices and shows that laser illumination on the AuBi2S3 interface can greatly affect the barrier height in non-ohmic contact devices, while keeping it intact in ohmic contact devices. A model based on the surface photovoltage effect is used to explain this phenomenon.
1. Introduction In recent decades, one-dimensional or quasi-one-dimensional nanowires (NWs) have attracted tremendous attention due to their unique properties, such as large surface to volume ratio and quantum size effect. With ordered alignment and excellent crystallinity, NWs are an ideal choice to investigate the intrinsic transport properties of different materials and to obtain devices with high performance. Some NWs have super performances as field-effect transistors (FET).[1–3] Some NWs, which are very sensitive to light and gases, can be widely used in the fields of photodetectors, photovoltaics, photoelectricity, optical switches, optical interconnections, and gas sensors.[4–8] Today, more and more people are considering two-dimensional materials, which are also low-dimensional materials, and possess high optical and electronic properties.[9–11] However, their super performance depends greatly on the thickness of their flakes. Indirect to direct transformation in the band gap accompanies a decrease in the thickness of these 2D materials.[12] To the best of our knowledge, the most effective and frequently used method to obtain high-quality 2D materials is still micromechanical cleavage. Other methods, such as epitaxial growth, [a] R. Li, J. Yang, N. Huo, C. Fan, F. Lu, T. Yan, Dr. Z. Wei, Prof. J. Li State Key Laboratory for Superlattices and Microstructures Institute of Semiconductors, Chinese Academy of Sciences P.O. Box 912, Beijing 100083 (P.R. China) E-mail: [email protected] [email protected] [b] N. Huo, Prof. J. Li Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei, Anhui 230026 (P.R. China) [c] Dr. Z. Wei Nano-Science Center & Department of Chemistry University of Copenhagen, Universitetsparken 5 2100 Copenhagen (Denmark) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402201.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
chemical vapor deposition, and liquid exfoliation, are still challenged by uncontrollable traps and a large area of growth.[13, 14] With all of these methods, it is difficult to distinguish singlelayer from few layer samples.[15] This may greatly limit their application in industrial production. Owing to the ongoing mature advancement in NW growth and fabrication methodologies, many NW devices have been fabricated and many new device structures have been investigated.[16–20] So far, many works had been reported on NW photodetectors and a deeper understanding of NW photodetectors has been achieved.[21–23] Compared with their bulk counterparts, NW photodetectors show great advantages in photoconductivity gains, response times, and device sizes that are beneficial to their utilization in integrated circuits.[24, 25] Bi2S3, which is an important n-type semiconductor with a narrow band gap range from 1.3 to 1.7 eV,[26, 27] has been widely investigated in visible-light photocatalysis, thermoelectricity, and photodiode arrays.[28–30] Its good conductivity and relatively easy synthetic process further extend its application to photodetectors and solar cells.[31, 32] Previously, Bi2S3 nanorod photodetectors with a photoconductive gain of 1.1 and solution-processed Bi2S3 nanocrystalline photodetectors with a photoconductive gain of 10 and a temporal response of 10 ms have been reported.[33, 34] In previous work, we investigated the photoresponsive behavior of Bi2S3 NW films at low temperatures, which was distinctly different from that at room temperature.[35] Herein, we fabricated photodetectors based on long single-crystalline Bi2S3 NWs with different electrical contacts and explored their respective performance. We synthesized long Bi2S3 NWs and investigated their detailed electrical contact effects on the performance of NW photodetectors. We obtained Bi2S3 single NW photodetectors with high photoconductive gains of about 10 in ohmic contact and 100 in non-ohmic contact devices. The results showed that non-ohmic contact devices with rectifying behavior possessed higher photoconChemPhysChem 2014, 15, 2510 – 2516
2510
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
ductive gains and shorter response times than those of ohmic contact devices. Furthermore, by illuminating the different parts of the NW photodetector, we found that a focused laser could effectively influence the barrier height between gold and Bi2S3 in non-ohmic contact devices, which was distinctly different from the behavior of ohmic contact ones. A simple model of the surface photovoltage (SPV) effect is proposed to explain our experiment result.
Experimental Section The Bi2S3 NWs used for fabricating the metal–semiconductor–metal (MSM) photodetectors were synthesized by a hydrothermal process reported in the literature.[36] The obtained products were characterized by a series of techniques. X-ray diffraction (XRD) was performed by using Cu Ka irradiation with a Pgeneral XD-3 diffractometer. Field-emission scanning electron microscopy (FESEM) images were acquired on a NanoSEM650 microscope instrument by using an accelerating voltage of 10 kV. The Raman scattering properties were recorded on a micro-Raman spectrometer by using a l = 632.8 nm laser line as the excitation source. The Raman signals were collected at room temperature through a 100 objective with a background configuration. High-resolution transmission electron microscopy (HRTEM) images and Energy-dispersive X-ray spectroscopy (EDS) results were obtained by using a Tecnai G2 F20 STWIN transmission electron microscopy under a working voltage of 200 kV.
phology of as-synthesized NWs. In Figure 1 b, the observed lattice spacings of 0.358 and 0.398 nm can be indexed to (130) and (001) planes in the crystal structure, respectively. This demonstrated that Bi2S3 NWs grew predominantly along the c axis. The SAED pattern for the Bi2S3 NW is given in the inset of Figure 1 b and indicates the good crystallinity of the as-synthesized samples. The XRD pattern and EDS results were obtained to further characterize our sample. The XRD patterns of the assynthesized samples in Figure 1 c can be indexed to the orthorhombic phase of Bi2S3 (JCPDS 17-0320). The EDS results in Figure 1 d further confirm that our samples are mainly composed of bismuth and sulfur, in which the peak of copper comes from the copper grid used in the TEM test. The simulated crystal structure shown in Figure 1 e schematically illustrates the growth direction of as-synthesized Bi2S3 NWs, which agrees with the TEM results. A FESEM image of large-scale Bi2S3 NWs is presented in Figure S1 a in the Supporting Information. The lengths of the NWs range from several micrometers to tens of
The Bi2S3 MSM photodetectors were fabricated as follows: a small amount of Bi2S3 NWs were dispersed in ethanol under ultrasonication and dropped onto the SiO2/ Si substrate (300 nm thick SiO2) to acquire discrete, single NWs, and then a pair of Au electrodes were evaporated onto them with a gap of 25 mm. Photocurrent measurements were performed on an Agilent B2902 instrument with a focused green laser of l = 532 nm as the light source (light spot of 2 mm).
2. Results and Discussion 2.1. Fabrication and Characterization of Bi2S3 NWs and Photodetectors The morphology of Bi2S3 NW samples were investigated by HRTEM (Figure 1 a and b). Figure 1 a shows the typical mor-
Figure 1. HRTEM images of as-synthesized Bi2S3 NWs with a scale bar of a) 1 mm and b) 2 nm. Inset: the selectedarea electron diffraction (SAED) pattern. c) XRD patterns and d) EDS spectrum of as-synthesized Bi2S3 NWs. e) Simulated crystal structure of Bi2S3 NW with a (001) growth direction.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2511
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
micrometers and the diameters range from tens of nanometers to hundreds of nanometers. The crystal structure was further examined by Raman spectroscopy under different exciting powers (see Figure S1 b in the Supporting Information). The peaks located at 70, 93, 123, 186, 236, and 261 cm1 all fit well with Bi2S3.[37, 38] Among them, the peaks at 186, 236, and 261 cm1 are sensitive to the excitation powers. The peak at 123 cm1 emerged only at a high laser power of 0.1 mW; this was due to the laser heating effect.[39] All of these Raman characteristics confirmed that our Bi2S3 NWs were pure and of high quality. Figure 2 a shows an illustration of a typical MSM photodetector fabricated on the SiO2/Si substrate. Bi2S3 NWs were first
detectors. It is hard to precisely control the electrical contacts in our fabrication process. However, further exploration revealed that such rectifying behavior mainly originated from the distinct difference in contacts between the two opposite AuBi2S3 junctions. The thickness of gold electrodes (80 nm) evaporated on the NW devices was smaller than the diameter of Bi2S3 NW (500 nm), as shown in the AFM images of Figure S2 in the Supporting Information. The Bi2S3 NWs may not be in a strong contact with the gold electrodes, so this uncertainty could cause a large contact barrier at one side of metal–semiconductor (M–S) junction (Schottky contact) and a smaller contact barrier at the other side (ohmic or weak Schottky contact). Then the rectifying behavior emerged. Such incomplete cover caused the Au Bi2S3 contacts to be easily affected by post-treatment, such as the soaking treatment of methanol or ethanol, since methanol (ethanol) may easily permeate into the interface and change the interface states. Details of this exploration are presented in Figure S3 in the Supporting Information. Similar phenomena can be found in ZnO nanobelt (NW) dielectrophoresis experiments performed by Lao and co-workers.[42] Further analysis on the non-ohmic contact device is shown in Figure 2 c and d. In Figure 2 c (and inset), the enlarged rectifying behavior and corresponding I–V curve in logarithmic coordinates are presented. This rectifying characteristic was very similar to that of a Schottky diode and indicated that only one side of Figure 2. a) Schematic illustration of a Bi2S3 NW fabricated photodetector. b) I–V curves corresponding to different AuBi2S3 formed the large barricontact devices measured under dark conditions. Inset: a typical SEM image of the as-fabricated NW photodetecer. To analyze the transport betor. c) Enlarged I–V characteristics of a typical non-ohmic contact device. d) The I–V characteristics of the nonhavior, herein we use the Schottohmic contact device at forward bias. The gray line displays the best fit to Equation (1). ky diode model to describe the rectifying characteristics for claridropped onto SiO2/Si substrate and the gold electrodes were ty. The current can be defined from Equation (1):[43] evaporated onto them by using a gold wire of 25 mm in diameter as the shadow. Among all 20 devices we fabricated, I–V curves showed two typical kinds of electrical contacts, which V Vth I ¼ I ½exp e 1 ð1Þ 0 appeared randomly (Figure 2 b). Apart from the broken devinkT ces, six of them showed ohmic contacts and eight of them showed non-ohmic contacts. The non-ohmic devices always displayed an unsymmetrical I–V characteristic with a rectifying behavior (the maximum on–off current ratio was 10 in our exin which I0 is the saturation current, Vth is the threshold voltperiment). This result was also reported previously.[40, 41] The age, k is the Bolzmann constant, T is the absolute temperature, and n is the ideality factor that describes the deviation of the reason for the formation of such clear rectifying characteristics diode from an ideal Schottky diode. For our device, the ideality needs to be comprehensively identified. Defects, the interface factor was derived to be 2.46 from the results shown in Figstate, and the possibility of an amorphous layer outside the ure 2 d; this indicates a barrier has formed between gold and Bi2S3 NW are all parameters that could contribute to the large the Bi2S3 interface.[44] contact resistance in the AuBi2S3 junctions of our MSM photo 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2512
CHEMPHYSCHEM ARTICLES 2.2. Photoresponsive Properties of Non-Ohmic Contact Devices Figure 3 shows typical photoresponsive behavior of non-ohmic contact devices. In Figure 3 a, the I–V curves under different powers of green laser are displayed. The bias voltages are tested from 1 to + 1 V to avoid probable damage to the Bi2S3 NW devices. Figure 3 b shows the curves for photocurrent
www.chemphyschem.org tions easily. In this sense, the effect of trap states on prolonging the response time was screened and the photocurrent reduced rapidly with a high response speed. 2.3. Photoresponsive Properties of Ohmic Contact Devices
Besides the non-ohmic contact devices, the photoresponsive performance of ohmic contact devices is also displayed in Figure 4. In Figure 4 a, the I–V curves under different powers of green laser all show the nearlinear characteristics, which indicates good ohmic contacts. The corresponding I–V curves in the logarithmic coordinates are also shown in inset of Figure 4 a; this further reveals the good symmetric properties. However, regardless of which part of the Bi2S3 NWs were illuminated by the focused laser, all curves showed very good symmetrical shapes. The curves for photocurrent versus time are presented in Figure 4 b with an applied bias of 0.01 V, and the photocurrent curves fit well with a previous description of a typical twostage process:[47] rapid (t1) and slow (t2) increase. The slow increase and decrease stages in the photocurrent are caused by Figure 3. Measurements performed on non-ohmic contact devices. a) I–V curves measured under different laser the trap states in our samples. powers. Inset: the corresponding I–V curves in logarithmic coordinates. b) Photoresponsive curves of Bi2S3 NW The minimum response time in photodetectors measured by periodic irradiation of the laser at an interval of 10 s (applied bias 0.1 V). Response time of the photocurrent by c) turning on and d) turning off the laser (applied bias 0.1 V, laser power 300 mW). the ohmic devices was derived to be 1 s under illumination of 900 mW, from the results shown versus time by periodically turning on and off the laser. All in Figure 4 c and d; this much longer than that of non-ohmic curves are measured under the same conditions with a bias of devices because no barrier was formed at the AuBi2S3 inter0.1 V. The response time is calculated to be about 50 ms from face in the ohmic devices. Electrons can flow through the juncFigure 3 c and d. It is known that the response time for phototions freely; thus the electrons trapped during the illumination detectors is mainly controlled by the trap states in the samprocess can also escape from the trap states and flow through ples.[45, 46] For non-ohmic contact devices, there was a large barthe junctions. In this way, the photocurrent is reduced slowly with a long response time. rier at the AuBi2S3 interfaces. If the focused laser illuminated the AuBi2S3 interface, due to the SPV effect, many electrons in Bi2S3 were excited from the valence band to conduction band 2.4. Comparison of the Two Electrical Contact Devices and then separated with the help of built-in potential. In this sense, light caused the Bi2S3 energy band to bend at the Au Figure 5 shows photoconductivity gains under different laser powers. Clearly, both kinds of devices showed the same trend: Bi2S3 interface and reduced the contact barrier. It is easier for the photoconductivity gains increased with increasing laser electrons to move across the junctions and the photocurrent power; this indicates that higher laser power can excite more increased hugely (details of the SPV effect are discussed in Secelectron–hole (e–h) pairs in the NWs. Such an increase in the tion 2.5). However, after turning off the laser, the photoexcited photocurrent gains became saturated if the laser power was electrons are no longer generated. The contact barrier rehigh enough and then the photoconductivity gains maintained turned and the electrons were blocked again from flowing a fixed value.[48] However, regardless of the power at which the through the junctions. Photoexcited electrons captured by trap states during the illumination process slowly escaped laser illuminated, the photoconductivity gains were larger in from the trap states, but no longer flowed through the juncnon-ohmic contact devices than those in ohmic contact devi 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2510 – 2516
2513
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org hugely in the non-ohmic devices, whereas they were nearly the same in ohmic contact devices. The lower bias voltage means smaller dark currents and higher photoconductivity gains. This phenomenon was much clearer in the non-ohmic contact devices. Furthermore, the photocurrent versus illumination positions were also determined and are presented in Figure S4 in the Supporting Information; this indicated homogeneity in our Bi2S3 NWs. The slight difference between them may be caused by the laser heating effect, which generated some thermoelectric current.[49]
2.5. Effect of Laser Illumination on the Barrier Height of MSM Photodetectors Figure 4. Measurements performed on the ohmic devices. a) I–V curves measured under different laser powers. Inset: the corresponding I–V curves in logarithmic coordinates. b) Photoresponsive curves of Bi2S3 NW photodetectors measured by periodic irradiation of the laser at an interval of 10 s (applied bias 0.01 V). Response time of the photocurrent by c) turning on and d) turning off the laser (applied bias 0.01 V, laser power 900 mW).
We also compared the effect of the laser on the position of the zero-point voltage (V0) in both kinds of electrical contact devi-
Table 1. Photoconductivity gains under different bias voltages in nonohmic and ohmic devices (laser power of 300 mW).
Figure 5. Photoconductivity gains under different illumination powers (the applied bias in non-ohmic and ohmic contact devices is 0.1 and 0.01 V, respectively).
ces. This result implied that there was a better photoresponsive performance in non-ohmic contact devices, which can be explained by the smaller dark current in non-ohmic devices due to the existence of a large contact barrier. The influence of applied bias on photoconductivity gains was also studied and the results are presented in Table 1. All measurements were recorded under the same conditions with a laser power of 300 mW. The photoconductivity gains varied 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Applied bias
1 V
0.5 V
0.1 V
0.1 V
0.5 V
1V
non-ohmic ohmic
82 7.1
1328.4
222 9.8
133 9.4
94 7.6
70 6.1
ces. In these measurements, the laser was illuminated on different parts of the Bi2S3 NWs as shown in the insets of Figure 6 a and b. All data are derived from the enlarged I–V characteristics in the logarithmic coordinates given in Figures S5 and S6 in the Supporting Information. Clearly, the laser effectively influenced V0 in the non-ohmic contact device, whereas V0 remained intact in the ohmic contact device. In the nonohmic contact device, V0 moves toward negative voltage gradually with increasing laser power. For the non-ohmic contact devices, V0 is close to Vth. Under laser illumination, the reduced barrier height between the AuBi2S3 contacts caused Vth to shift. However, in the ohmic contact device, V0 remained intact at 0 V regardless of the power of the focused laser used. This is completely different from the non-ohmic contact devices. Because no barrier formed in the ohmic contact device, the SPV effect made no difference. Moreover, the V0 shifts with illumination position were also investigated and are presented in Figure 6 b (see also the Supporting Information). For both nonohmic and ohmic contact devices, V0 remained intact when ilChemPhysChem 2014, 15, 2510 – 2516
2514
CHEMPHYSCHEM ARTICLES
Figure 6. Shifts of zero-point voltage under different illumination powers in non-ohmic and ohmic contact devices upon illumination at the a) M–S interface and b) middle part of the NW in the gap. Left insets: schematic representations of the illumination positions of the Bi2S3 NWs; right insets: the corresponding optical images of the NW device.
luminated on the Bi2S3 NWs. This phenomenon was to be expected, since the SPV effect only took place at the M–S interface with the existence of a contact barrier. When illuminated at the local part of the Bi2S3 NWs, as shown in the inset of Figure 6 b, no electrons aggregated at the M–S interface and the barrier height changed slightly. Detail analysis of the SPV effect is presented as follows: The SPV can be defined as the illumination-induced change in the surface potential.[50, 51] If two materials with different Fermi energies are in contact, a contact barrier is generated at the interface. At the M–S interface, due to this built-in potential, there is a depletion region with width W (Figure 7 a). Under illumination, the photoexcited e–h pairs were separated at the depletion region, the holes moved towards the surface of the semiconductor, and induced electrons to aggregate in the metal at the side of the interface. In this way, the depletion region can be seen as a high-doping layer and the electron concentration is controlled by the light. Considering the image–force effect and the abrupt transition junction approximation, the variation of barrier height caused by the electron concentration at the depletion layer can be written simply as Equations (2) and (3):[52]
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 7. a) Schematic illustration of the cross section of the AuBi2S3 interface: D is the diameter of our NW and W is the width of the depletion region. b) Band alignment of the M–S interface with no illumination. c) Band alignment of the M–S interface under illumination. The dashed lines represent variation of the energy band. d) The barrier height variation versus laser power (illuminated at the M–S interface as shown in figure 6a).
rffiffiffiffiffiffi Q 4p
ð2Þ
Pinc GS hv
ð3Þ
q DF ¼ eS Q ¼ Rabs
in which Q is the sum of holes aggregated at the depletion region by the laser; Rabs, Pinc, hn, G, and S are the absorption rate of the semiconductor, the power of the laser, the photon energy of the laser, the generation rate of photons to create e–h pairs, and the separation rate of e–h pairs at the surface, respectively. This square-root relationship fits well with our results shown in Figure 7 d.
3. Conclusions Long, single-crystalline Bi2S3 NWs were successfully synthesized and fabricated into single NW photodetectors. During our fabrication process, two kinds of electrical contacts formed: nonohmic and ohmic. Detailed investigations on both kinds of electrical contact devices were performed to explore their reChemPhysChem 2014, 15, 2510 – 2516
2515
CHEMPHYSCHEM ARTICLES spective photoresponsive properties. Non-ohmic contact devices possessed higher photocurrent gains ( 100) and faster response speeds (50 ms), whereas ohmic contact devices possessed smaller photocurrent gains ( 10) and slow response speeds (1000 ms). Furthermore, if the laser illuminated one side of the AuBi2S3 contacts, the position of the zero-point voltage (V0) could be effectively influenced in non-ohmic contact devices. The variation of barrier height at the AuBi2S3 interface was believed to cause this phenomenon. Due to the SPV effect, photoexcited electrons could aggregate at the interface of AuBi2S3, change the barrier height, and lead to the shift in the zero-point voltage. These findings further expand our knowledge about the fabrication of NW photodetectors and may contribute to future photodetector designs.
Acknowledgements This work was supported by the National Natural Science Foundation of China under Grant No. 91233120 and the National Basic Research Program of China (2011CB921901). Keywords: bismuth · nanostructures · photochemistry · semiconductors · sulfur [1] X. Duan, Y. Huang, Y. Cui, J. Wang, C. M. Lieber, Nature 2001, 409, 66 – 69. [2] Y. Cui, Z. Zhong, D. Wang, W. U. Wang, C. M. Lieber, Nano Lett. 2003, 3, 149 – 152. [3] D. Chen, Z. Liu, B. Liang, X. Wang, G. Shen, Nanoscale 2012, 4, 3001 – 3012. [4] R. Yan, D. Gargas, P. Yang, Nat. Photonics 2009, 3, 569 – 576. [5] J. Wang, M. S. Gudiksen, X. Duan, Y. Cui, C. M. Lieber, Science 2001, 293, 1455 – 1457. [6] C. Li, D. Zhang, B. Lei, S. Han, X. Liu, C. Zhou, J. Phys. Chem. B 2003, 107, 12451 – 12455. [7] L. Peng, L. Hu, X. Fang, Adv. Mater. 2013, 25, 5321 – 5328. [8] R. M. Penner, Annu. Rev. Anal. Chem. 2012, 5, 461 – 485. [9] D. J. Late, B. Liu, H. S. S. Ramakrishna Matte, V. P. Dravid, C. N. R. Rao, ACS Nano 2012, 6, 5635 – 5641. [10] D. J. Late, B. Liu, J. Luo, A. Yan, H. S. S. Ramakrishna Matte, M. Grayson, C. N. R. Rao, V. P. Dravid, Adv. Mater. 2012, 24, 3549 – 3554. [11] J. Luo, D. Late, I. Wu, K. Biswas, M. Kanatzidis, M. Grayson, AIP Conf. Proc. 2011, 1416, 135-138. [12] D. J. Late, Y. K. Huang, B. Liu, J. Acharya, S. N. Shirodkar, J. Luo, A. Yan, D. Charles, U. V. Waghmare, V. P. Dravid, C. N. R. Rao, ACS Nano 2013, 7, 4879 – 4891. [13] R. V. Kashid, D. J. Late, S. S. Chou, Y. K. Huang, M. De, D. S. Joag, M. A. More, V. P. Dravid, Small 2013, 9, 2730 – 2734. [14] H. S. S. Ramakrishna Matte, A. Gomathi, A. K. Manna, D. J. Late, R. Datta, S. K. Pati, C. N. R. Rao, Angew. Chem. Int. Ed. 2010, 49, 4059 – 4062; Angew. Chem. 2010, 122, 4153 – 4156. [15] D. J. Late, B. Liu, H. S. S. Ramakrishna Matte, C. N. R. Rao, V. P. Dravid, Adv. Fun. Mater. 2012, 22, 1894 – 1906. [16] J. Hu, T. W. Odom, C. M. Lieber, Acc. Chem. Res. 1999, 32, 435 – 445. [17] C. M. Lieber, MRS Bull. 2003, 28, 486 – 491. [18] Z. Zhong, D. Wang, Y. Cui, M. W. Bockrath, C. M. Lieber, Science 2003, 302, 1377 – 1379.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [19] P. Fan, K. C. Y. Huang, L. Cao, M. L. Brongersma, Nano Lett. 2013, 13, 392 – 396. [20] B. Ryu, Y. T. Lee, K. H. Lee, R. Ha, J. H. Park, H. Choi, S. Im, Nano Lett. 2011, 11, 4246 – 4250. [21] H. Kind, H. Yan, B. Messer, M. Law, P. Yang, Adv. Mater. 2002, 14, 158 – 160. [22] Y. Gu, J. P. Romankiewicz, J. K. David, J. L. Lensch, L. J. Lauhon, Nano Lett. 2006, 6, 948 – 952. [23] A. Armstrong, G. T. Wang, A. A. Talin, J. Electron. Mater. 2009, 38, 484 – 489. [24] C. Soci, A. Zhang, X. Bao, H. Kim, Y. Lo, D. Wang, J. Nanosci. Nanotechnol. 2010, 10, 1430 – 1449. [25] Y. Li, F. Qian, J. Xiang, C. M. Lieber, Mater. Today 2006, 9, 18 – 27. [26] C. Ye, G. Meng, Z. Jiang, Y. Wang, G. Wang, L. Zhang, J. Am. Chem. Soc. 2002, 124, 15180 – 15181. [27] J. D. Desai, C. D. Lokhhande, Mater. Chem. Phys. 1995, 41, 98 – 103. [28] S. Liufu, L. Chen, Q. Yao, C. Wang, Appl. Phys. Lett. 2007, 90, 112106. [29] Z. Zhang, W. Wang, L. Wang, S. Sun, ACS Appl. Mater. Interfaces 2012, 4, 593 – 597. [30] B. D. Nayak, H. N. Acharya, G. B. Miitra, S. Tanhanker, Thin Solid Films 1983, 110, 165 – 170. [31] H. Li, J. Yang, J. Zhang, M. Zhou, RSC Adv. 2012, 2, 6258 – 6261. [32] A. K. Rath, M. Bernechea, L. Martinez, G. Konstantatos, Adv. Mater. 2011, 23, 3712 – 3717. [33] L. Li, R. Cao, Z. Wang, J. Li, L. Qi, J. Phys. Chem. C 2009, 113, 18075 – 18081. [34] G. Konstantatos, L. Levina, J. Tang, E. H. Sargent, Nano Lett. 2008, 8, 4002 – 4006. [35] R. Li, Q. Yue, Z. Wei, J. Mater. Chem. C 2013, 1, 5866 – 5871. [36] Y. Yu, C. H. Jin, R. H. Wang, Q. Chen, L. Peng, J. Phys. Chem. B 2005, 109, 18772 – 18776. [37] K. Trentelman, J. Raman Spectrosc. 2009, 40, 585 – 589. [38] Y. W. Koh, C. S. Lai, A. Y. Du, E. R. T. Tiekink, K. P. Loh, Chem. Mater. 2003, 15, 4544 – 4554. [39] Y. Zhao, K. T. E. Chua, C. K. Gan, J. Zhang, B. Peng, Z. Peng, Q. Xiong, Phys. Rev. B 2011, 84, 205330 – 205336. [40] H. Bao, X. Cui, C. Li, Y. Gan, J. Zhang, J. Guo, J. Phys. Chem. C 2007, 111, 12279 – 12283. [41] K. Yao, Z. Zhang, X. Liang, Q. Chen, L. Peng, J. Phys. Chem. B 2006, 110, 21408 – 21411. [42] C. Lao, J. Liu, P. Gao, L. Zhang, D. Davidovic, R. Tummala, Z. Wang, Nano Lett. 2006, 6, 263 – 266. [43] A. G. Milnes, D. L. Feucht, Heterojunctions and Metal – Semiconductor Junctions, Academic Press, New York, 1972, pp. 34 – 57. [44] B. L. Sharma, Metal – Semiconductor Schottky Barrier Junctions and Their Applications, Plenum Press, New York, 1984, pp. 2 – 17. [45] D. Han, D. C. Melcher, E. A. Schiff, M. Silver, Phys. Rev. B 1993, 48, 8658 – 8666. [46] R. P. Barclay, J. Phys. Condens. Matter 1994, 6, 9721 – 9728. [47] K. Schwarzburg, F. Willig, Appl. Phys. Lett. 1991, 58, 2520 – 2522. [48] C. Soci, A. Zhang, B. Xiang, S. A. Dayeh, D. P. R. Aplin, J. Park, X. Y. Bao, Y. H. Lo, D. Wang, Nano Lett. 2007, 7, 1003 – 1009. [49] B. Varghese, R. Tamang, E. S. Tok, S. G. Mhaisalkar, C. H. Sow, J. Phys. Chem. C 2010, 114, 15149 – 15156. [50] K. H. Yoon, J. K. Hyun, J. G. Connell, I. Amit, Y. Rosenwaks, L. J. Lauhon, Nano Lett. 2013, 13, 6183 – 6188. [51] Z. Zhang, J. T. Yates, Jr., Chem. Rev. 2012, 112, 5520 – 5551. [52] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, Wiley, Hoboken, 2007, pp. 111 – 117.
Received: February 7, 2014 Published online on July 30, 2014
ChemPhysChem 2014, 15, 2510 – 2516
2516
