Thin Solid Films 485 (2005) 194 – 197 www.elsevier.com/locate/tsf
Photon detector composed of metal and semiconductor nanoparticles Atsuo Takahashi1, Norihiko Minoura*, Isao Karube National Institute of Advanced Industrial Science and Technology, Research Center of Advanced Bionics, Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan Received 12 August 2004; accepted in revised form 24 March 2005 Available online 10 May 2005
Abstract Applying the function of the single electron transistor, a novel photon detector consisting of a self-assembled structure of metal and semiconductor nanoparticles and an organic insulating layer was developed. It showed coulomb blockade behavior under dark conditions and remarkable increase in current corresponding to light intensity under light irradiation. Ultraweak photon emission of about 600 counts per second in the ultraviolet region could be detected at room temperature by this photon counter. D 2005 Elsevier B.V. All rights reserved. PACS: 73.23.Hk; 85.30.Vw; 85.60.Gz Keywords: Nanostructures; Optoelectronic devices; Tunneling
1. Introduction In recent years, elements for controlling single electrons have been developed with the full use of nanotechnology. Single electron transistor (SET) [1– 3] is one of those elements, and its application to single photon counters has been tried [4,5]. The performance of SET is restricted in the region that the charging energy of a quantum dot e 2 / 2C dominates over the thermal energy k BT, where C is the total capacitance of tunneling junctions. To operate the SET at room temperature, the area of tunneling junctions which compose the SET should be on a few-nanometer scale [3]. However, since in top-down approaches such as electronbeam lithography it is difficult to manipulate the structure of a few-nanometer scale, SETs or SET-based single photon counters fabricated by such methods can work only at very low temperature [1,2,4,5], and moreover, integration or bulk production is difficult because of their laborious fabrication process. * Corresponding author. Tel.: +81 29 861 2987; fax: +81 29 855 3833. E-mail address:
[email protected] (N. Minoura). 1 Present address: Tokyo Institute of Technology, Department of Organic and Polymer Material, Meguro-ku, 2-12-1 Ookayama, Tokyo 152-8552, Japan. 0040-6090/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2005.03.034
On the contrary, bottom-up approaches using metal or semiconductor nanoparticles have been expected as a method to build such small structures more easily. Indeed, Coulomb blockade was easily observed at room temperature in systems which consist of metal nanoparticles and organic insulating layers [6– 9]. In this report, we demonstrate a novel photon detector consisting of a self-assembled structure of metal and semiconductor nanoparticles and an organic insulating layer.
2. Experimental details As shown by the cross-sectional schematic diagram of the photon detector in Fig. 1(a), the photon detector is composed of two kinds of electrodes (upper side is the ITO (indium tin oxide) transparent electrode and lower side is the gold electrode the surface of which was coated by the organic insulating layer) and two kinds of nanoparticles (gold and TiO2) which were coated by the organic insulating layer and intervene between these electrodes. Gold nanoparticles and TiO2 nanoparticles were dispersed planarly so that these disparate nanoparticles adjoin respectively. A brief fabrication method of photon detector is as follows. The gold nanoparticles of approximately 11 nm in
A. Takahashi et al. / Thin Solid Films 485 (2005) 194 – 197
195
particles and amino groups of SAM-coated gold nanoparticles. After the vacancies of nanoparticles were filled with the amino-undecanethiol, ITO film was sputterdeposited onto those constructed layers of nanoparticles and insulator. The whole structure is equivalent with SETs which were connected in parallel, regarding the gold electrode as the source, the ITO electrode as the drain, the gold nanoparticle as the quantum dot and the TiO2 nanoparticle as the gate. The gold nanoparticles form a tunnel junction with two electrodes, while the polymeric multilayer surrounding the TiO2 nanoparticles is too thick for an electron to tunnel through it. The mechanism of detection of photon is shown in Fig. 1(b). The voltage V sd, the value of which is within the Coulomb blockade, i.e., V sd o e / 2C, is applied between the source and the drain. The electron and the hole in the TiO2 nanoparticle excited by the photon drift to the upper side and lower side of the TiO2 nanoparticle, respectively, due to the electric field between the source and the drain. The electrostatic energy of the tunnel junctions is changed by the hole drifted to the lower side, thereby, the tunneling of the electron from the source to the quantum dot and from the quantum dot to the drain can occur.
Fig. 1. (a) Cross-sectional schematic diagram of the photon detector. (b) Schematic diagram of a brief mechanism of the detection of photons. The hole which was photo-generated and drifted to the lower side in the TiO2 nanoparticle changes electrostatic energy stored in the tunnel junctions between the gold nanoparticle and two electrodes (source and drain), by attracting the charge in the gold nanoparticle. Subsequently, an electron is allowed to tunnel into and out of the gold nanoparticle due to the breakup of the Coulomb blockade.
diameter were prepared by the reduction of HAuCl4 by trisodium citrate dehydrate and tannic acid [10]. Then the colloidal gold nanoparticles were dropped into an N-9fluorenylmethyloxycarbonyl (Fmoc)-aminohexanethiol/ N,N-Dimethylformamide solution so that a self-assembled monolayer (SAM) of the N-Fmoc-aminohexanethiol was formed on the gold nanoparticles. TiO2 nanoparticles of approximately 25 nm in diameter (P25, Degussa Co., Germany) were provided from Nippon Aerosil Co., Ltd. and coated by polymeric layers of about 16 nm in thickness by the method of layer-by-layer adsorption of poly(sodium 4-styrenesulfonate) (PSS) and poly(allylamine hydrochloride) (PAH) [11]. The SAM of the carboxy-pentanethiol was formed on the sputter-deposited gold electrode. The SAMcoated gold nanoparticles from which Fmoc was removed, were conjugated sparsely (not to be packed) with the carboxy-pentanethiol on the gold electrode using N-hydroxysuccinimide and N-ethyl-N¶-dimethylaminopropylcarbodiimide. Subsequently, polyelectrolyte-coated TiO2 nanoparticles were immobilized next to conjugated gold nanoparticles by the electrostatic interaction between the PSS layer adsorbed at the most outside of TiO2 nano-
3. Results and discussion At room temperature and under dark conditions, I – V characteristic between the source and the drain was measured (solid line in Fig. 2). In the forward bias case, the current was almost constant and hardly flowed, while in the reverse bias case, coulomb blockade was observed. This commutation may be due to the fact that a part of the electron path, that is, gold nanoparticle-SAM of aminohexanethiol-ITO electrode, forms the structure of the MIS (Metal-Insulator-Semiconductor) diode. In the MIS junction, the n-type semiconductor surface is in deep depletion by the applied positive bias, thereby the Fermi level in the metal cannot exceed the conduction band minimum of the semiconductor and the tunneling of electrons hardly occurs [12 –15]. On the other hand, when the negative bias is applied to the semiconductor, large reverse current flows
Fig. 2. I – V curves between source and drain measured at room temperature. The drain electrode is the ITO (solid line) and the gold (dashed line).
196
A. Takahashi et al. / Thin Solid Films 485 (2005) 194 – 197
because the tunneling barrier at the semiconductor surface is eliminated [12 –15]. This suppression of current in the positive bias region is broken by the irradiation of light, enabling the tunneling current to flow [12,16]. When the gold was deposited as the drain instead of ITO, clear coulomb blockade was observed in both positive and negative bias regions (dashed line in Fig. 2). Response in the current between the source and the drain I sd to the irradiation of light was measured at room temperature for three different light intensities, i.e., ultra violet (UV) lamp, fluorescent lamp and fluorescent lamp with superimposed optical filters. The band gap energy of TiO2 is about 3.0 eV, which means that only the photons with wavelength of below about 400 nm can be adsorbed by TiO2. UV intensities of each light source were about 1.0 AW/cm2, 6.9 10 5 AW/cm2 and 1.6 10 7 AW/cm2, respectively, which were measured by UV intensity meter. The number of photons which arrive at the detection area for each light source was then 4 109 counts per second (cps), 3 105 cps and 6 102 cps, respectively, where the detection area was 0.2 mm2 and the wavelength of the photons was assumed to be 365 nm uniformly. As shown in Fig. 3, when lights were irradiated to the photon detector, I sd was increased immediately and reached saturated values gradually corresponding to UV intensities. After the irradiation was stopped, I sd was decreased and reverted to the values before irradiating. The relationship between the irradiated number of photons and the saturated value of I sd was plotted double logarithmically in Fig. 4. It may be due to the structure of the photon detector in which numbers of
Fig. 3. Time evolution of the I sd measured with 0.5 V bias voltage when lights of different intensities were irradiated for a certain time (shadowed region). Number of photons the wavelength of which were assumed to be 365 nm are (a) 4 109, (b) 3 105 and (c) 6 102 cps.
Fig. 4. Double logarithmic plot of the number of the incident photons N p versus the saturated values of I sd in Fig. 3.
SET were connected in parallel, that linearity was realized in the double logarithmic plot over a remarkably wide range of N p. If the photon detector is composed of one SET, the increasing ratio of current by irradiation of light will not be so large because of the pulsed current versus gate bias [1,2]. As shown in Fig. 3, after the irradiation was stopped, long tailing time was required to return to the initial value. This may be due to the inhibition of the recombination of the electron and the hole excited in the TiO2 nanoparticles. Once the electron and the hole were generated and were separated to the upper side and the lower side of TiO2 nanoparticle, respectively, they hardly recombine even if the irradiation of light was stopped and/or the voltage between the source and the drain was cut off, because electrostatic attraction between the electron (the hole) in the TiO2 nanoparticle and the hole in the ITO electrode (the electron in the gold electrode) is stronger than that between the electron and the hole in the TiO2 nanoparticle. These were estimated using the thickness of the PSS/PAH multilayer of about 16 nm and the dielectric constant of that of about 3– 5 [17,18], and the diameter of TiO2 nanoparticles of 25 nm and the dielectric constant of that of about 23 [19]. To let those electrons and holes in TiO2 nanoparticles tunnel through the thick polymeric layer, high voltage was applied between the source and the drain. In Fig. 5, an example of time evolution of I sd when applying high voltage for 1.5 min
Fig. 5. Time evolution of the I sd measured with 10 mV bias voltage. After the irradiation of light (same condition with Fig. 3(b)) was stopped, bias voltage was elevated to 62 V for 1.5 min to discharge the TiO2 nanoparticles.
A. Takahashi et al. / Thin Solid Films 485 (2005) 194 – 197
after stopping irradiation is indicated. I sd successfully dropped after application of high voltage. In Fig. 3, the rise time until I sd reaches a saturated value was also long as well as the tailing time. This may be due to the low efficiency of charge separation in the TiO2 nanoparticles. The recombination kinetics of the photo-excited electron and hole in the TiO2 nanoparticles is very fast (in the order of picoseconds) [20], and the applied voltage (0.5 V) may be insufficient to separate the electron and hole completely. Conjugation of dyes to the TiO2 nanoparticles is suggested to be effective to separate the charge efficiently. It is known that when a metal complex dye or an organic dye conjugated to the TiO2 nanoparticles is photo-excited, the generated electron transfers rapidly to the TiO2, in the order of picoseconds, and recombination of the electron and the hole abided in the dye hardly occurs [21,22]. Consequently, improvement in the speed of the rise time, and further, improvement in sensitivity and expansion of the wavelength range of detectable photons, are expected.
4. Conclusion In conclusion, a novel photon detector fabricated by the bottom-up method has been demonstrated. Although it still has some problems which should be solved as mentioned above, it could detect ultraweak photons (6 102 cps). Taking advantage of the features of being easily producible, miniaturizable, working with low voltage and at room temperature, the photon detector will be applicable to various devices such as hand-held biosensors, highly sensitive telescopes and quantum cryptography devices.
Acknowledgement The authors gratefully acknowledge Dr. Yasushi Abe and Kazuya Itoda for valuable discussion.
197
References [1] C. Kurdak, A.J. Rimberg, T.R. Ho, J. Clarke, J.D. Walker, Physica, E 5 (2000) 274. [2] D.H. Kim, S.K. Sung, J.S. Sim, K.R. Kim, J.D. Lee, B.G. Park, B.H. Choi, S.W. Hwang, D. Ahn, Appl. Phys. Lett. 79 (2001) 3812. [3] M.I. Lutwyche, Y. Wada, J. Appl. Phys. 75 (1994) 3654. [4] A.N. Cleland, D. Esteve, C. Urbina, M.H. Devoret, Appl. Phys. Lett. 61 (1992) 2820. [5] S. Komiyama, O. Astafiev, V. Antonov, T. Kutsuwa, H. Hirai, Nature 403 (2000) 405. [6] N.K. Chaki, M. Aslam, T.G. Gopakumar, J. Sharma, R. Pasricha, I.S. Mulla, K. Vijayamohanan, J. Phys. Chem., B 107 (2003) 13567. [7] P. Jiang, Z.F. Liu, S.M. Cai, J. Appl. Phys. 90 (2001) 2039. [8] S.J. Huang, G. Tsutsui, H. Sakaue, S. Shingubara, T. Takahagi, J. Vac. Sci. Technol., B 18 (2000) 2653. [9] M. Dorogi, J. Gomez, R. Osifchin, R.P. Andres, R. Reifenberger, Phys. Rev., B 52 (1995) 9071. [10] G. Tsutsui, S.J. Huang, H. Sakaue, S. Shingubara, T. Takahagi, Jpn. J. Appl. Phys. Part 1—Regul. Pap. Short Notes Rev. Pap. 40 (2001) 346. [11] F. Caruso, H. Lichtenfeld, E. Donath, H. Mohwald, Macromolecules 32 (1999) 2317. [12] H.A. Lin, R.J. Jaccodine, M.S. Freund, Appl. Phys. Lett. 73 (1998) 2462. [13] H.A. Lin, R. Jaccodine, M.S. Freund, Appl. Phys. Lett. 72 (1998) 1993. [14] A. Turut, B. Bati, A. Kokce, M. Saglam, N. Yalcin, Phys. Scr. 53 (1996) 118. [15] M.B. Johnson, J.M. Halbout, J. Vac. Sci. Technol., B 10 (1992) 508. [16] T. Takahashi, M. Yoshita, Appl. Phys. Lett. 70 (1997) 2162. [17] L.J. Kepley, D.D. Sackett, C.M. Bell, T.E. Mallouk, Thin Solid Films 208 (1992) 132. [18] M.F. Durstock, M.F. Rubner, Langmuir 17 (2001) 7865. [19] G. Garcia-Belmonte, V. Kytin, T. Dittrich, J. Bisquert, J. Appl. Phys. 94 (2003) 5261. [20] N. Serpone, D. Lawless, R. Khairutdinov, E. Pelizzetti, J. Phys. Chem. 99 (1995) 16655. [21] X.J. Dang, J.T. Hupp, J. Am. Chem. Soc. 121 (1999) 8399. [22] T. Rajh, O. Poluektov, A.A. Dubinski, G. Wiederrecht, M.C. Thurnauer, A.D. Trifunac, Chem. Phys. Lett. 344 (2001) 31.