Appl. Phys. A 91, 229–233 (2008)
Applied Physics A
DOI: 10.1007/s00339-008-4398-1
Materials Science & Processing
Disordered multiwalled carbon nanotube mat for light spot position detecting
jia-lin sunu jia xu jia-lin zhuu baolei li
Department of Physics and Key Laboratory of Atomic and Molecular Nanosciences of Education Ministry, Tsinghua University, Beijing 100084, P.R. China
Received: 13 December 2007/Accepted: 14 December 2007 Published online: 15 February 2008 • © Springer-Verlag 2008
Photovoltaic effects in a disordered multiwalled carbon nanotube mat/nickel heterostructure have been investigated. It is found that the photovoltage in the whole mat between two nickel electrodes under the irradiation of a 532-nm laser shows a single-valued function dependence on the light spot position, while it shows an almost linear dependence on light intensity. Based on the ‘position effect’, the prototype of the light spot position detector has been constructed for the applications of the disordered multiwalled carbon nanotubes in precision measurement.
ABSTRACT
PACS 85.60.-q;
1
72.40.+w; 73.63.-b; 81.07.De; 78.67.-n
Introduction
Position detectors are vital elements in precision optical alignment, such as biomedical applications, robotics, process control, and position information systems generally [1–4]. After Wallmark first introduced the position-sensitive detectors [5], they have been much investigated and their structures and theories are varied. But, most commercial position detectors tend to be fairly smallarea and expensive. Since the discovery of carbon nanotubes (CNTs) [6], their remarkable physical properties have received extensive attention due to the reduced dimensionality structure. In recent years, CNTs’ optical properties such as nonlinear optical and limiting properties [7], Raman scattering [8, 9], luminescence [10], photon-induced molecular adsorption, and desorption properties [11], and so on have been intensively studied, and some promising applications and sensor prototypes have been proposed [12, 13]. However, most of these devices focus on nanoscale samples. If the macroscopic samples of CNT materials also have these remarkable optical properties, they can be explored in a much wider range of applications and the cost will reduce. In micro- and macroscale devices, the mechanism of contact between CNTs and metal electrodes is a critical issue. It is generally believed that a barrier is formed between metal u Fax: +86-10-62781604, E-mail:
[email protected] u Fax: +86-10-62781604, E-mail:
[email protected]
electrodes and CNTs, such as a Schottky barrier at semiconducting CNT/metal junctions [14, 15], and a barrier-like behavior is shown at metallic CNT/metal junctions [16, 17]. More understanding is needed if the sample is macroscopic, because the band structure of CNTs is complex and there are more defects. Recently, increasing interests are focused on their properties with the optical effects. Several experimental and theoretical works on individual and bundles of singlewalled carbon nanotubes (SW CNTs) and double-walled carbon nanotubes (DW CNTs) demonstrate the generation of a photocurrent when the CNT/metal junctions are illuminated [18–20]. In contrast to SW CNTs, understanding photoelectronic properties of multiwalled carbon nanotubes (MW CNTs), which consist of a number of stacked shells, has proceeded more slowly because of their extremely complex structure. A recent experiment pointed out that a macroscopically long bundle of ordered MW CNTs also shows such a photocurrent at MW CNT/metal junctions [18]. However, MW CNTs prepared by catalytic chemical vapor deposition in a nanoagglomerate fluidized bed reactor often highly entangle with each other in agglomerates. If the sample is a macroscopic mat or film, some novel effect must be introduced by the interaction between the disordered CNTs [21]. In this paper, we present experimental studies of photovoltaic effects in disordered MW CNT mat/nickel heterostructures. The MW CNT mat, which consists of many disordered MW CNTs interconnected with each other by van der Waals force, is manufactured using a pressure method. The photovoltage in the whole mat between two MW CNT/nickel heterojunctions under the irradiation of a 532-nm laser shows a single-valued function dependence on the light spot position. Based on the ‘position effect’, a prototype of the light spot position detector is constructed. 2
Experiment
The disordered MW CNTs were synthesized by the chemical vapor deposition method and purified using thermal annealing and acid treatment [22]. Our sample is made through the following steps. Purified disordered MW CNT raw material was dispersed in ethanol and agitated by ultrasonics to make a uniform suspension. When a small amount of disordered MW CNT material had been pressed by a large pressure, the nanotubes would like to attract each other through van der Waals force and form a mat easily. So, we
230
Applied Physics A – Materials Science & Processing
FIGURE 1 (a) Schematic illustration of disordered MW CNT mat/Ni heterostructures and optical path. Dotted line is the optical path after modulating the reflector. (b) Top view of the circuit in the position detector. A and B are two interfaces. O is the position without photovoltage. The circular areas in the MW CNT mat show the laser spot. (c) SEM image of disordered MW CNT mat
first transferred the MW CNT suspension to a mould and then compressed it into a mat on a plastic substrate at a pressure of 30 MPa when the ethanol solution had dried. Figure 1c shows the scanning electron microscope (SEM) image of the disordered MW CNT mat, which is manufactured by the method above. It can be seen that the diameters of the nanotubes are in the range of 20 – 40 nm and the mat consists of many nanotubes accommodated in random directions and positions with multiple interconnections. The thickness and the resistance of the mat can be varied easily by controlling the amount of disordered nanotubes used. The typically used mat is about 80-µm thick. Then, we cut the MW CNT mat into long strips of 3 mm × 10 mm in size. A pair of pure nickel foils clamped the two ends of the mat strip respectively to form the disordered MW CNT mat/Ni heterostructures, as schematically shown in Fig. 1a. The distance between the two electrodes is about 7 mm. In our experiments, a 532-nm laser with a spot diameter of 0.2 mm is employed as the illumination source. There is a reflector in the optical path. The reflected laser beam can illuminate an arbitrary position of the sample between interfaces A and B by modulating the angle θ . The whole circuit including a galvanometer, an adjustable resistor, and the sample was connected using aerial wires, as shown in Fig. 1a and b. 3 3.1
Results and discussion Photovoltaic effects
Figure 2a plots the electrical current through the MW CNT mat versus bias voltage before laser illumination, clearly showing that in the range of bias used its resistance steadies at 224 Ω. That means that there is an ohmic contact between Ni electrodes and the MW CNT mat. When a laser is used to illuminate the disordered MW CNT mat/Ni heterostructures, there is a current in the circuit. We fix the light spot position to the interface A and adjust the adjustable re-
(a) Volt–ampere characteristics of disordered MW CNT mat/Ni heterostructures. (b) Photovoltage as a function of all circuit resistance
FIGURE 2
sistor meanwhile. Current is recorded as a function of the resistance. Then, we can calculate the bias voltages of the adjustable resistor, the CNT mat–Ni heterostructure, and the whole circuit, which are marked as Uout , Uin , and Uall , respectively, as shown in Fig. 2b. It is found surprisingly that the whole circuit voltage Uall turns out to be constant. Similar results are exhibited when the light spot moves to other positions on the MW CNT mat between interfaces A and B. This is the photovoltaic effect of the detector. So, the photovoltage is chosen as the mark to describe the photoelectronic properties of the detector. In previous studies, individual semiconducting carbon nanotubes and metal contacts have received much attention. The observed phenomena are in accordance with the presence of Schottky barriers at the contacts, which is well documented from studies of the operation mechanism of tube FETs [23–25]. Previous experiments on the photoexcitation spectrum of a SW [26, 27] and a DW [28] CNT film using a displacement current technique demonstrated that there were excitonic states in both semiconducting and metallic CNTs. As far we know, there have been no photoexcitation spectrum studies performed on MW CNT films so far. Based on our experiment and the similarity between MW and DW CNTs, we believe that when a MW CNT mat is illuminated by the laser, hole–electron pairs are generated at the same time. For some hole–electron pairs which diffuse near the contact, it is much easier for hot electrons to enter the metal through Schottky barriers or other barriers formed at the contact by tunneling or emission before combining with holes than to go back from metal to CNTs. Barriers play the role of separating the carriers effectively. Then, the photovoltaic
SUN et al.
Disordered multiwalled carbon nanotube mat for light spot position detecting
effects can be observed. From the research we have done previously [18], another point of view had been presented. Compared with three-dimensional metals, the allowed momenta of Bloch electrons and phonons in CNTs are strongly quantized, and the density of states shows distinct distributions. It also may be the reason for the photoinduced current. This is called the heterodimension effect in short. Based on many experiments, it is found that the net transport direction of the photoexcited electrons is from the lower-dimensional CNTs to higher-dimensional CNTs or metal electrodes. 3.2
Position effect
Figure 3 shows the photovoltage as a function of the illuminating position defined as the distance between the laser spot and the interface A. Here, the laser intensity is 150 mW. The direction of the voltage is defined as positive if the current flows through the MW CNT mat from interface A to interface B. It can be seen that the photovoltage reaches the maximum when the interfaces are illuminated. The whole shape of the function curve is single-valued and of nearly point symmetry. As the light spot moves from interface A to interface B, the photovoltage decreases first, and becomes zero at one point defined as O, which is approximately the geometric center of the strip, and then increases inversely. Repeated experiments on different samples in our experiment exhibit similar results. We call this the ‘position effect’ of the detector. The position effect has not been reported in the research on individual carbon nanotubes. We firstly explored it in the research done on SW CNT/metal, DW CNT/metal, and ordered MW CNT/metal heterostructures [18]. The work in this paper shows that such macroscopic and disordered MW CNT mat/Ni heterostructures also can exhibit the position effect despite multiple interconnections in CNTs. It can be deduced that the position effect of the photocurrent or photovoltage can only be witnessed when the macroscopically long CNT material and metal electrodes are in contact. In other words, the two electrodes in CNT/metal heterostructures must be separated far enough in order to have no or little
FIGURE 3 Photovoltage as a function of the laser spot position relative to interface A. The solid line is a polynomial fitting of the corresponding data
231
interaction with each other. Compared to the position effect curves of SW CNT/metal and DW CNT/metal heterostructures, the curve of disordered MW CNT mat/Ni heterostructures exhibits a much more satisfactory shape. Ordered MW CNT/metal heterostructures seem much more propitious to be used as detectors as the result of their high degree of linearity between photovoltage and position, but the cost of manufacturing macroscopically long ordered MW CNTs is very large and it is impossible for the commercial scale of the detectors. So, we prefer disordered MW CNT mat/Ni heterostructures to act as a light spot detector. All position effect curves collected from different samples in our experiment have the same characters: single-valued, monotone, and without singular points. For every sample, a one to one mapping from photovoltage to light spot position can be established based on the experimental data: x = f(U) .
(1)
Here, x and U are the light spot position and photovoltage, respectively. Mapping f can serve as a working function of our position detector. Once the working function is obtained, a realistic measurement can be implemented. For instance, in our setup (Fig. 1a), the angle θ can be measured. Before modulating the reflector, U1 is obtained from the galvanometer and resistance; then, the corresponding x 1 can be obtained from the working function through interpolation. The light spot position after modulating the reflector x 2 can be calculated in the same way. Therefore, the angle θ is measured by our position detector since the differential of the light spot position ∆x geometrically corresponds to θ . In a realistic measurement, the change of the angle θ could originate from many factors, such as electromagnetic or thermal force, oscillation, deformation, and so on. So, lots of measurements can be achieved by monitoring the photovoltage of our sample. Two parts are included in precision analysis of this position detector prototype. Firstly, the error introduced by interpolation cannot be neglected. There are two methods to decrease the error of interpolation. One is perfecting the device, including enhancing the uniformity of the MW CNT mat and the symmetry of the two Ni electrodes. This is because the density of the CNTs under illumination determines the generation ratio of carriers, and the contacts of the two heterostructures affect the symmetry of the position effect curves. Moreover, the stability of the laser intensity obviously affects the error of the collected data. The other is to increase the density of experimental data upon the MW CNT mat. The second part in precision analysis is spatial resolution, which indicates the minimum distance that can be measured when the light spot is moved from one position to another. Obviously, the precision of the galvanometer affects the spatial resolution of our detector. Because the position effect is nonlinear from interface A to interface B, the whole sample is divided into three sections, AC, CD, and DB (Fig. 3). It can be easily deduced that section CD asks for a higher accuracy of the galvanometer. We estimate it using a linear method. Section CD is from 2.25 mm to 5.25 mm, and its corresponding photovoltage is from about 95.5 µV to − 146.5 µV. If we want to measure a length of 10 µm, it can be calculated that 3.6 nA must be enumerated by the galvanometer. Actually, the precision of the galvanometer
232
Applied Physics A – Materials Science & Processing
in our experiment is 0.1 nA. So, without question our detector prototype can distinguish a length of 10 µm. This spatial resolution is comparable with other typical Schottky barrier position sensitive detectors [29]. We acknowledge that the work function of this detector is required before measurement on account of the nonlinearity of the position effect. And, this may cause a little trouble compared with other position sensitive detectors. 3.3
Response of photovoltage to light intensity and switch
In order to further study the photoelectronic properties of the detector, the laser beam spot is fixed to a certain place along the mat. The photovoltage is measured as a function of the laser intensity. It is found that in the intensity range used there is a linear-like relationship between them, as displayed in Fig. 4a. It means that the generation and separation of the carriers in the MW CNT mat manufactured by the pressure method is almost linearly proportional to the intensity of the laser. When a laser pulse is used to excite it, the corresponding photoelectronic response is also recorded, as shown in Fig. 4b. The speed of the shutter used in the dynamic response measurement is 1 m/s. So, the time delay introduced by the mechanism is 2 ms. The photoelectronic response must be a combined process of optical, thermal, electrical, and elastic effects [30]. For the morphology of our sample, the elastic effects can be neglected. Figure 5a and b respectively describe the photoelectronic response with respect to the light being on and off;
FIGURE 5 Photoelectronic response: (a) light on; (b) light off. The solid lines are the exponential fittings of the corresponding data
furthermore, mathematically fitted curves are shown in both figures. It is found that the experimental data fit well into an equation of the form t t U = U0 + U1 exp − + U2 exp − . (2) t1 t2 The constants U0 , U1 , and U2 in (2) are, respectively, U0 = y0 + A1 + A2 , U1 = −A1 , U2 = −A2 for Fig. 5a and U0 = y0 , U1 = A1 , U2 = A2 for Fig. 5b. There are two characteristic response times, t1 and t2 , in the photoelectronic response. The values of t1 and t2 are ∼ 0.72 s, ∼ 6.20 s and 0.79 s, 6.54 s, respectively, as shown in Fig. 5a and b. The CNTs’ photoconductivity has been investigated, and a similar simple kinetic model has been used to interpret the processes of photocarrier generation and relaxation [31, 32]. So, the shorter one t1 originates from the photoconductivity excitation of the MW CNT mat, which is shorter than in [31]. The much longer one t2 can be attributed to the thermocouple effect as the result of the macroscopic sample, and it is the major factor in the limiting speed of this detector. 4
(a) Dependence of photovoltage on the light intensity. The solid line is a linear fitting of the corresponding data. (b) Response of photovoltage to light switch
FIGURE 4
Conclusions
In conclusion, disordered MW CNT mat–Ni heterostructures have been fabricated and the phenomena of the photovoltage have been explored. It is found that the photovoltage exhibits the ‘position effect’ along the whole sample. Based on the ‘position effect’, the prototype of the light spot position detector has been constructed for the new applications of the disordered multiwalled carbon nanotubes in precision measurement.
SUN et al.
Disordered multiwalled carbon nanotube mat for light spot position detecting
ACKNOWLEDGEMENTS Financial support from the National Natural Science Foundation of China (Grant Nos. 10 374 057 and 10 474 048), the Key Grant Project of the Chinese Ministry of Education (No. 306 020), and the ‘973’ Programme of China (No. 2005CB623606) is gratefully acknowledged.
14 15 16 17 18
REFERENCES 19 1 J. Zhang, T. Yamaguchi, K. Iwata, H. Kikuta, C.S. Park, Appl. Opt. 42, 5661 (2003) 2 W.S. Park, H.S. Cho, Opt. Eng. 41, 860 (2002) 3 D.W. Buhler, T.R. Oxland, L.P. Nolte, Med. Eng. Phys. 19, 187 (1997) 4 J.K. Kim, M.S. Kim, J.H. Bae, J.H. Kwon, H.B. Lee, S.H. Jeong, Appl. Opt. 39, 2584 (2000) 5 J.T. Wallmark, Proc. IRE 45, 474 (1957) 6 S. Iijima, Nature 354, 56 (1991) 7 P. Chen, X. Wu, X. Sun, J. Lin, W. Ji, K.L. Tan, Phys. Rev. Lett. 82, 2548 (1999) 8 J.M. Holden, P. Zhou, X.X. Bi, P.C. Eklund, S. Bandow, R.A. Jishi, K.D. Chowdhury, G. Dresselhaus, M.S. Dresselhaus, Chem. Phys. Lett. 220, 186 (1994) 9 A.M. Rao, E. Richter, S. Bandow, B. Chase, P.C. Eklund, K.A. Williams, S. Fang, K.R. Subbaswamy, M. Menon, A. Thess, R.E. Smalley, G. Dresselhaus, M.S. Dresselhaus, Science 275, 187 (1997) 10 M. Freitag, V. Perebeinos, J. Chen, A. Stein, J.C. Tsang, J.A. Misewich, R. Martel, P. Avouris, Nano Lett. 4, 1063 (2004) 11 M. Shiraishi, T. Takenobu, H. Kataura, M. Ata, Appl. Phys. A 78, 947 (2004) 12 Y.N. Xia, P.D. Yang, Y.G. Sun, Y.Y. Wu, B. Mayers, B. Gates, Y.D. Yin, F. Kim, Y.Q. Yan, Adv. Mater. 15, 353 (2003) 13 V.N. Popov, Mater. Sci. Eng. R 43, 61 (2004)
20 21 22 23 24 25 26 27 28 29 30 31 32
233
A.A. Odintsov, Phys. Rev. Lett. 85, 150 (2000) F. Leonard, J. Tersoff, Phys. Rev. Lett. 84, 4693 (2000) K. Kong, S. Han, J. Ihm, Phys. Rev. B 60, 6074 (1999) K. Balasubramanian, M. Burghard, K. Kern, M. Scolari, A. Mews, Nano Lett. 5, 507 (2005) J.L. Sun, J. Wei, J.L. Zhu, D. Xu, X. Liu, H. Sun, D.H. Wu, N.L. Wu, Appl. Phys. Lett. 88, 131 107 (2006) D.H. Lien, W.K. Hsu, H.W. Zan, N.H. Tai, C.H. Tsai, Adv. Mater. 18, 98 (2006) K. Balasubramanian, Y.W. Fan, M. Burghard, K. Kern, M. Friedrich, U. Wannek, A. Mews, Appl. Phys. Lett. 84, 2400 (2004) C. Schonenberger, A. Bachtold, C. Strunk, J.P. Salvetat, L. Forro, Appl. Phys. A 69, 283 (1999) Y. Wang, J. Wu, F. Wei, Carbon 41, 2939 (2003) S. Heinze, J. Tersoff, R. Martel, V. Derycke, J. Appenzeller, P. Avouris, Phys. Rev. Lett. 89, 106 801 (2002) Z.H. Chen, J. Appenzeller, J. Knoch, Y.M. Lin, P. Avouris, Nano Lett. 5, 1497 (2005) J. Appenzeller, J. Knoch, M. Radosavljevic, P. Avouris, Phys. Rev. Lett. 92, 226 802 (2004) A. Mohite, S. Chakraborty, P. Gopinath, G.U. Sumanasekera, B.W. Alphenaar, Appl. Phys. Lett. 86, 061 114 (2005) A. Mohite, J.T. Lin, G. Sumanasekera, B.W. Alphenaar, Nano Lett. 6, 1369 (2006) A. Mohite, G.U. Sumanasekera, K. Hirahara, S. Bandow, S. Iijima, B.W. Alphenaar, Chem. Phys. Lett. 412, 190 (2005) J. Henry, J. Livingstone, IEEE Sens. J. 6, 6 (2006) Y. Zhang, S. Iijima, Phys. Rev. Lett. 82, 3472 (1999) I.A. Levitsky, W.B. Euler, Appl. Phys. Lett. 83, 1857 (2003) A. Fujiwara, Y. Matsuoka, H. Suematsu, N. Ogava, K. Miyano, H. Kataura, Y. Maniwa, S. Suzuki, Y. Achiba, Japan. J. Appl. Phys. Part 2 40, L1229 (2001)
