Transport properties in single-crystalline rutile TiO2

APPLIED PHYSICS LETTERS 99, 222107 (2011)

Transport properties in single-crystalline rutile TiO2 nanorods R. S. Chen,1,a) C. A. Chen,2 W. C. Wang,2 H. Y. Tsai,3 and Y. S. Huang2,3 1

Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei 10607, Taiwan 2 Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan 3 Graduate Institute of Electro-Optical Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan

(Received 14 September 2011; accepted 13 November 2011; published online 30 November 2011) Electronic transport properties of the single-crystalline titanium dioxide (TiO2) nanorods (NRs) with single rutile phase have been investigated. The conductivity values for the individual TiO2 NRs grown by metal-organic chemical vapor deposition are in the range of 110 X1 cm1. The temperature-dependent measurement shows the presence of two shallow donor levels/bands with activation energies at 8 and 28 meV, respectively. On the photoconductivity (PC), the TiO2 NRs exhibit the much higher normalized PC gain and sensitive excitation-power dependence than the polycrystalline nanotubes. The results demonstrate the superior photoconduction efficiency and distinct mechanism in the monocrystalline one-dimensional TiO2 nanostructures in comparison to C 2011 American Institute of Physics. the polycrystalline or nanoporous counterparts. V [doi:10.1063/1.3665635] Quasi one-dimensional (1D) nanostructures of titanium dioxide (TiO2) have attracted substantial attention due to their high aspect ratio in geometry and great potential in electronic and optoelectronic devices. Among them, singlecrystalline nanorods (NRs) or nanowires (NWs) and polycrystalline nanotubes (NTs) are the most studied nanomaterials as building blocks in the dye-sensitized solar cell (DSSC),1 photochemical,2–4 and sensors5,6 devices. The interesting geometric shape renders very high surface area and at the same time an anisotropic transport route for carrier. In comparison to polycrystalline 1D structure, monocrystalline NR and NW are able to avoid high-degree boundary scattering during charge transport and exhibit a relatively predictable and stable surface, which are crucial properties for the device applications. Nevertheless, the understandings for the fundamental transport properties, especially the photoconductivity (PC) in the monocrystalline 1D nanomaterial of TiO2 are still very limited.7–9 To date, only the dark transport properties defined by field-effect transistor (FET) measurements have been demonstrated for the undoped8 and doped9 TiO2 NWs. Investigation of the photoconduction properties is urgently required to enable the control of TiO2 NR or NW-related photodevices.7 In this letter, we report on the studies of temperature-dependent conductivity and power-dependent PC in the single-crystalline TiO2 NRs with pure rutile phase. The high photoconduction efficiency of the TiO2 NRs is quantitatively defined and compared to the polycrystalline NTs.10 The significant differences in the excitation power dependence behavior and the underneath mechanism between the TiO2 NR and NT are also discussed. The rutile TiO2 NRs used for this study were grown by cold-wall metal-organic chemical vapor deposition (MOCVD) using sapphire (100) as the substrate at the deposition tempera)

Author to whom correspondence should be addressed. Electronic mail: [email protected]. Tel.: þ886-2-2730-3754. Fax: þ886-2-2730-3733.

0003-6951/2011/99(22)/222107/3/$30.00

ature of 550 C and oxygen pressure of 1.5 mbar.11,12 The morphological and structural properties of the as-grown NRs were characterized using field-emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM), selected-area electron diffractometry (SAD), and x-ray diffractometry (XRD). The two-terminal single NR devices were fabricated using focused-ion beam (FIB) deposition and platinum (Pt) as the contact metal. Individual NRs were dispersed on the insulating Si3N4(200 nm)/n-Si template with prepatterned Ti/Au microelectrodes prior to the FIB deposition. Usually, the ohmic contact property of the FIB-fabricated devices can be achieved after a post annealing treatment at the temperature below 250 C under the atmospheric oxygen ambience. FIB deposition electrical measurements were carried out on an ultralow current leakage cryogenic probe station (LakeShore Cryotronics TTP4). A semiconductor characterization system (Keithley 4200-SCS) was utilized to source DC bias and to measure current. The He-Cd laser with 325 nm wavelength was used as the excitation source for the PC measurements. The value of incident power of laser was measured by a calibrated power meter (Ophir Nova II) with a photodiode head (Ophir PD300-UV). The typical micrographs of the vertically aligned TiO2 NRs grown by MOCVD on the sapphire (100) substrate are illustrated in Figs. 1(a) and 1(b). The well-aligned TiO2 NRs with the pure rutile phase and predominantly h001i preferred orientation along their long-axis are verified by the XRD pattern, Fig. 1(c). The single diffraction peak also indicates the single-crystalline quality of the as-grown TiO2 NRs. The h001i long-axis orientation and the monocrystalline rutile structure of the individual TiO2 NRs are further confirmed from the TEM image and its corresponding SAD pattern as shown in Fig. 1(d). The dark current (id) versus applied bias (V) curves measured at different temperatures (T) of the single TiO2 NR with d of 300 nm are depicted in Fig. 2(a). The inset shows a representative micrograph of the FIB-fabricated single-NR

99, 222107-1

C 2011 American Institute of Physics V

Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp

222107-2

Chen et al.

Appl. Phys. Lett. 99, 222107 (2011)

FIG. 1. (Color online) (a) The tilt-viewed and (b) the cross-sectional viewed FESEM images and (c) the XRD pattern of the well aligned rutile TiO2 NRs grown by MOCVD. (d) The TEM image and its SAD pattern focusing on the individual TiO2 NR.

device. The linear dependence of the id-V curves measured in the temperature range of 77–320 K indicates the good ohmic contact property of the NR device. The dark conductivity (r) values and their temperature dependence are obtained through the curve fitting, as depicted in Fig. 2(b). The room temperature (T ¼ 300 K) r of the selected NR is around 1.9 6 0.2 X1 cm1. As the r shows slightly sample dependence, the statistic r values from six samples are in the range of 1–10 X1 cm1. In comparison to other TiO2 1D nanomaterials, the values are, respectively, two and eight orders of magnitude higher than the single-crystalline rutile NWs (r ¼ 1.4 102 X1 cm1) (Ref. 8) and the polycrystalline NTs (r 1 108 X1 cm1) (Ref. 13). Although the NTs could also have high r due to the very high donor defect density,10 the polycrystalline nanostructure still frequently suffer several decades lower mobility and thus r due to the significant grain boundary scattering. The conductive nature could imply the low degree of scattering and relatively high carrier concentration in the boundary-free TiO2 NRs. Moreover, from the rT plot, Fig. 2(b), a semiconducting transport behavior is observed for the TiO2 NR. While lowering down temperature from 320 to 77 K, the r decreases from 2.2 to 0.5 X1 cm1. By the Arrhenius plot (the inset in Fig. 2(b), two different slopes can be roughly differentiated and the corresponding values of activation energy (Ea) are calculated as Ea1 ¼ 28 and Ea2 ¼ 8 meV, respectively. The values are smaller than those of TiO2 NWs (Ea ¼ 58 meV) (Ref. 8) and high-resistivity NTs (Ea ¼ 870 meV),13 indicating the presence of shallow donors in the TiO2 NRs. The origin resulting in the different Ea in TiO2 nanomaterials is still not clear. In principle, in addition to the different kinds of donor defects, a broader defect band due to the higher donor density could also make a smaller Ea in the TiO2 NR than the NWs. In addition to the dark conductivity, photoconduction property of the TiO2 NR is also investigated. Figure 3(a) illustrates the photocurrent (ip) value as a function of light intensity (I) of the TiO2 NR with d of 300 nm under the excitation wavelength of 325 nm at the bias of 0.1 V. The data points of the ip-I curve for the polycrystalline TiO2 NT taken

FIG. 2. (Color online) (a) The selected id–V curves measured in the temperature range of 77–320 K for the single TiO2 NR with 300 nm in diameter. The inset shows the representative micrograph of a FIB-fabricated TiO2 NR device. (b) The calculated conductivity versus temperature and its corresponding Arrhenius plot (the inset) of the single TiO2 NR with 300 nm in diameter.

from Ref. 10 is also plotted for the comparison. A typical photoresponse curve of the NR at I ¼ 40 Wm2 is depicted in the inset of Fig. 3(a). Although only five data points can be obtained from Ref. 10, the overall ip level of the NR is higher than the NT in this studied I region. In addition, less sensitive power dependence of ip is also observed for the TiO2 NR compared to the NT. As ip is determined by multiple parameters including the incident power (P) on the projected area (A ¼ dl) of the NR, quantum efficiency (g), and gain (C), ip is expressed as14–16 ip ¼

q PgC; E

(1)

V sl; l2

(2)

C¼

where q is the unit electron charge, E is the excitation photon energy, s is the photocarrier lifetime, and l is the mobility. To compare the photoconduction efficiency between different nanomaterials, the influences of experimental and geometric parameters should be excluded. Hence, a PC parameter, named normalized gain (Cn), defined as10,17 Cn ¼ gsl;

(3)

is estimated for the TiO2 NRs, and NTs. According to the definition, Cn is simply the function of material properties including g, s, and l without the dependence on the


222107-3

Chen et al.

Appl. Phys. Lett. 99, 222107 (2011)

trapping effect in the intrinsic n-type semiconductors. Usually, a transformation from the trap-dominant mechanism to the bimolecular recombination can be observed after a trap-filling intensity.18–20 The power-insensitive behavior of Cn can persist in the whole high power region actually reflects the presence of high-density hole trap in the polycrystalline TiO2 NTs. On the other hand, as the excitation intensities in this study are relative high, the power-sensitive Cn and its powerlaw dependence in the TiO2 NRs are similar to the previous observations for the single-crystalline oxide semiconductor NWs, such as ZnO (Ref. 20) and SnO2 (Ref. 21). This suggests a surface-dominant PC mechanism in the TiO2 NRs. In conclusion, the fundamental transport properties including dark conductivity (r ¼ 1–10 X1 cm1) and two shallow activation energies (Ea1 ¼ 28 and Ea2 ¼ 8 meV) in the boundary-free rutile TiO2 NRs have been defined in this study. The TiO2 NRs also shows the higher value and powersensitivity of the normalized gain than the nanoporous NTs. According to the quantitative study, the results consistently point the superior transport properties under the dark and ultraviolet light surroundings in the monocrystalline TiO2 NRs in comparison to the polycrystalline NT counterparts.

FIG. 3. (Color online) The light intensity dependences of (a) photocurrent and (b) normalized gain for the monocrystalline TiO2 NR (blue open circle) and the polycrystalline NTs (green open square). The data points of the TiO2 NT are taken from Ref. 10. The inset in Fig. 3(a) illustrates a typical photocurrent response curve at the bias of 0.1 V and I ¼ 40 Wm2.

experimental parameters and thus can be an index determining the combined efficiency of photocarrier generation and transport in a photoconductor. According to Eqs. (1)–(3), Cn can be calculated from the measured ip using the relationship, Cn ¼

E l2 ip : qVP

(4)

After taking the experimental parameters, includes E, V, l, and P, into account, the Cn values as a function of I for the TiO2 NR and NT are depicted in Fig. 3(b). It is interesting that the single-crystalline NR reveals significant power dependence of Cn following a power-law of Cn ! I0.87. While I increases from 1.6 to 8 102 Wm2, Cn decrease for over two orders of magnitude from 3.0 106 to 1.5 108 m2 V1. Meanwhile, the polycrystalline NT shows a weak power dependence of Cn ! I0.19 in the I range of 5.4 101–3.3 103 Wm2. A data point at I ¼ 3.3 103 Wm2, does not follow the same trend, is excluded in this fitting. From the distinct Cn-I relationship, it is clear that the TiO2 NR exhibits higher Cn than the NT in the broader I region of 1.6 to 4 102 Wm2. Once the I goes above 4 102 Wm2, the Cn of NRs becomes lower than that of NTs gradually. As Cn is linear dependent on l, the significantly higher scattering degree induced by the highly dense grain boundary and defects in the polycrystalline NTs could make a several orders of magnitude lower l and thus Cn while compared to the boundary-free monocrystalline NRs of TiO2.10 In addition, the near constant Cn (or C) is frequently attributed to the hole-

The author RSC would like to thank the financial supports of the Taiwan National Science Council (Grant Nos. NSC 992112-M-011-001-MY3 and NSC 99-2738-M-011-001) and the National Taiwan University of Science and Technology. 1

Q. Zhang and G. Cao, Nanotoday 6, 91 (2011). G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, and C. A. Grimes, Nano Lett. 5, 191 (2005). 3 J. H. Park, S. Kim, and A. J. Bard, Nano Lett. 6, 24 (2006). 4 G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang, and Y. Li, Nano Lett. 11, 3026 (2011). 5 O. K. Varghese, D. Gong, M. Paulose, K. G. Ong, and C. A. Grimes, Sens. Actuators B 93, 338 (2003). 6 I. D. Kim, A. Rothschild, B. H. Lee, D. Y. Kim, S. M. Jo, and H. L. Tuller, Nano Lett. 6, 2009 (2006). 7 A. Dessombz, C. R. Pasquier, P. Davidson, and C. Chaneac, J. Phys. Chem. C 114, 19799 (2010). 8 J. M. Baik, M. H. Kim, C. Larson, X. Chen, S. Guo, A. M. Wodtke, and M. Moskovits, Appl. Phys. Lett. 92, 242111 (2008). 9 Y. H. Lee, J. M. Yoo, D. H. Park, D. H. Kim, and B. K. Ju, Appl. Phys. Lett. 86, 033110 (2005). 10 C. Fabrega, F. Hernandez-Ramirez, J. D. Prades, R. Jimenez-Diaz, T. Andreu, and J. R. Morante, Nanotechnology 21, 445703 (2010). 11 C. A. Chen, Y. M. Chen, A. Korotcov, Y. S. Huang, D. S. Tsai, and K. K. Tiong, Nanotechnology 19, 075611 (2008). 12 C. A. Chen, Y. M. Chen, Y. S. Huang, D. S. Tsai, P. C. Liao, and K. K. Tiong, J. Alloys Compd. 480, 107 (2009). 13 R. Asmatulu, A. Karthikeyan, D. C. Bell, S. Ramanathan, and M. J. Aziz, J. Mater. Sci. 44, 4613 (2009). 14 R. S. Chen, H. Y. Chen, C. Y. Lu, K. H. Chen, C. P. Chen, L. C. Chen, and Y. J. Yang, Appl. Phys. Lett. 91, 223106 (2007). 15 P. Bhattacharya, Semiconductor Optoelectronic Devices (Prentice-Hall Inc., New Jersey, 1997), Chap. VIII, pp. 346–351. 16 M. Razeghi and A. Rogalski, J. Appl. Phys. 79, 7433 (1996). 17 J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, L. FernandezRomero, T. Andreu, A. Cirera, A. Romano-Rodriguez, A. Cornet, J. R. Morante, S. Barth, and S. Mathur, J. Phys. Chem. C 112, 14639 (2008). 18 K. S. Stevens, M. Kinniburgh, and R. Beresford, Appl. Phys. Lett. 66, 3518 (1995). 19 F. Binet, J. Y. Duboz, E. Rosencher, F. Scholz, and V. Harle, Appl. Phys. Lett. 69, 1202 (1996). 20 C. Soci, A. Zhang, B. Xiang, S. A. Dayeh, D. P. R. Aplin, J. Park, X. Y. Bao, Y. H. Lo, and D. Wang, Nano Lett. 7, 1003 (2007). 21 C. H. Lin, R. S. Chen, T. T. Chen, H. Y. Chen, Y. F. Chen, K. H. Chen, and L. C. Chen, Appl. Phys. Lett. 93, 112115 (2008). 2
