pubs.acs.org/NanoLett
Carbon Nanotube-Based Dual-Mode Biosensor for Electrical and Surface Plasmon Resonance Measurements Jeseung Oh,†,⊥ Young Wook Chang,‡,⊥ Hyung Joon Kim,† Seunghwan Yoo,‡ Dong Jun Kim,§ Seongil Im,‡ Young June Park,| Donghyun Kim,†,§ and Kyung-Hwa Yoo*,†,‡ †
Nanomedical Graduate Program, ‡ Department of Physics, and § School of Electrical and Electronic Engineering, Yonsei University, Seoul, 120-749, Republic of Korea and | School of Electrical Engineering, Seoul National University, Seoul, 151-742, Republic of Korea ABSTRACT We have developed carbon nanotube-based dual-mode biosensors with a metal semiconductor field effect transistor structure on a quartz substrate. DNA hybridization occurring on the Au top gate can be detected by simultaneously measuring the change in the electrical conductance and the surface plasmon resonance (SPR). Since electrical and SPR measurements offer high sensitivity and reliability, respectively, this dual-mode biosensor is expected to provide both of these features. KEYWORDS Dual-mode biosensor, carbon nanotube, metal semiconductor field effect transistor, electrical conductance, surface plasmon resonance
R
ecently, biosensors based on a carbon nanotube (CNT) with a field effect transistor (FET) structure have attracted considerable attention because of their merits, such as label-free detection, real-time monitoring, ultrahigh sensitivity, and simplicity of apparatus.1-8 When a target analyte binds to receptors immobilized on the CNT surface, the electrical conductance of the CNT-FET changes because of electrostatic gating9,10 or Schottky barrier effects.11,12 Thus, various biological interactions, such as DNA hybridization,4-6 antibody-antigen interactions,7,8 and so forth,1,2 have been monitored by measuring the conductance change. In spite of these successful demonstrations, there are several problems to be solved, such as sensor-to-sensor variation and unspecific binding, to realize practical applications. To overcome the above problems, we have proposed a dual-mode biosensor that enables the detection of biological events by simultaneously measuring changes in both electrical conductance and SPR. The dual-mode biosensor has the structure of a CNT metal semiconductor field effect transistor (CNT-MESFET) (Figure 1). A long CNT is grown or deposited on a transparent substrate, and a gold (Au) strip is fabricated in the middle of the CNT. The Schottky barrier forms at an interface between the Au strip and the CNT; thus, the Au strip plays the role of a top gate. As a result, charged biomolecules, which bind to receptors immobilized on the surface of the Au strip, may be detected by measuring the change
in the electrical conductance of the CNT. In addition, the transparent substrate allows the CNT-MESFET to detect SPR occurring at the interface between the Au top gate and the transparent substrate when an incident beam of polarized light of a given wavelength is directed toward the surface at a given angle through a prism (Figure 1). Biological events can be detected and quantified by measuring the changes in SPR. In comparison with CNT-FET biosensors, SPR sensors provide better reliability, although their sensitivity is relatively low.13,14 On the other hand, CNT-FET biosensors offer ultrahigh sensitivity, although their performance is dependent on the sensor. For the dual-mode biosensor, SPR and conductance measurements may complement each other, so that both high sensitivity and reliability may be obtained. Furthermore, the chemical modification of the Au surface used in the SPR biosensors is well-known, so unspecific binding would be more easily inhibited in the dual-mode
* To whom correspondence should be addressed. Tel: +82-2-2123-3887. Fax: +82-2-312-7090. E-mail:
[email protected]. ⊥
These authors contributed equally to this work. Received for review: 01/14/2010 Published on Web: 07/06/2010 © 2010 American Chemical Society
FIGURE 1. Schematic diagram of a dual-mode CNT-MESFET biosensor. 2755
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755–2760
FIGURE 2. (a) I-V curves measured using source (S), drain (D), and top gate (TG) electrodes. (b) I-VBG curves measured at V ) 1 V. ISD is the current measured between the source and drain electrodes, ISTG is the current measured between the source and top gate electrodes, and IDTG is the current measured between the drain and top gate electrodes.
biosensor than in the CNT-FET biosensors. In this study, we have fabricated dual-mode biosensors on quartz substrates and we demonstrated that DNA hybridization can be detected by measuring the change in electrical conductance and SPR at the same time. Prior to fabricating the CNT-MESFET on the quartz, we fabricated CNT-MESFETs on SiO2/Si substrates to characterize them (inset of Figure 2a and Supporting Information Figure S1). Single-walled CNTs (SWNTs) were grown by the patterned-catalyst growth technique. Fe/Mo-supported alumina catalysts were patterned on the heavily doped Si substrate with a 200 nm thermally grown SiO2 layer. Then, SWNTs were grown by chemical vapor deposition (CVD) using a mixture of methane and argon as the carbon source at 900 °C. Source (S) and drain (D) electrodes were made by depositing Cr/Pd (5 nm/50 nm) and using a lift-off technique, followed by rapid thermal annealing at 450 °C for 30 s in a vacuum to form good contacts. To determine whether the CNT connected between source and drain electrodes was metallic or semiconducting, the source/drain current (ISD) was measured as a function of the bottom gate voltage (VBG) for all devices. The heavily doped Si substrate was used as the bottom gate and electrical measurements were carried out using a semiconductor parameter analyzer (Keithley 4200-SCS). The top gate was fabricated by depositing Cr/Au (1 nm/5 nm) only on the middle of the semiconducting CNT, as shown in Supporting Information Figure S1. Using three electrodes, source (S), drain (D), and top gate (TG), the I-V curves were measured under ambient conditions (Figure 2a). The ISD-VSD curve measured between the source and drain electrodes was linear, whereas the ISTG-VSTG and IDTG-VDTG curves exhibited diode-like behavior, where ISTG (or IDTG) stands for the current measured between the source (or drain) and top gate electrodes. These results suggest that Schottky contacts formed between the top gate electrode and the CNT, while ohmic contacts formed between the source/drain electrodes and the CNT because of the heat treatments. Figure 2b shows the ISD-VBG, ISTG-VBG, and IDTG-VBG transfer curves measured at V ) 1 V. As expected from the measured I-V curves in Figure 2a, ISTG and IDTG were lower than ISD. However, all curves exhibited p-type semiconducting behavior with nearly identical thresh© 2010 American Chemical Society
old voltages, Vth, although Vth was found at a higher VBG in comparison with the case of the ISD-VBG curve measured before depositing the top gate. From these results, we concluded that the semiconducting properties of the CNT were not significantly affected by depositing the top gate. Most biological interactions occur under aqueous conditions, so biosensors usually operate in an aqueous environment. To investigate the effects of a liquid gate voltage (VLG) on the electrical properties of the CNT-MESFET, we passivated the source/drain electrodes and the CNT parts uncovered by the top gate by depositing a SiO2 thin film, leaving only the Au top gate exposed to the solution and then mounted a polydimethylsiloxane (PDMS) well over the CNTMESFET (Figures 3a and Supporting Information S2a). Subsequently, we filled the PDMS well with a 10 mM solution of the phosphate buffered saline (PBS, pH 7.4) and inserted an Ag/AgCl reference electrode used as the liquid gate (LG). Figure 3b depicts the ISD-VLG curve measured at VSD ) 10 mV and the leakage current (ISLG) between the source and the liquid gate. The p-type semiconducting behavior was clearly observed with negligible ISLG. This result indicated that the ISD of the CNT-MESFET could be modulated by applying VLG in spite of the shielding effect from the Au top gate. For comparison, we also measured the ISD-VLG curve at VSD ) 10 mV for the CNT-FET without the top gate (Figure 3c). A clear difference was seen in a subthreshold swing defined as S ) dVLG/d(log ISD).15 The values of S were estimated to be approximately 0.1 and 0.23 V/decade for the CNT-MESFET and the CNT-FET, respectively, which were smaller than S ≈ 1.7 V/decade estimated from the ISD-VBG curves for both devices (Figure 2b). These findings implied that the CNT-MESFET and CNT-FET had different values of liquid gate capacitance, whereas their bottom gate capacitance was nearly identical. In the case of the liquid gate, the gate capacitance was approximately given by the double layer capacitance, Cdl ) εA/xOHP, where ε is the dielectric constant of the electrolyte, A is the area of the surface exposed to the electrolyte, and xOHP is the distance to the outer Helmholtz plane.16,17 Since the CNT-MESFET might have a larger A than the CNT-FET, the steeper transition of ISD or the smaller S observed for the CNT-MESFET may 2756
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755-–2760
FIGURE 3. (a) Schematic diagram of a CNT-MESFET with a PDMS well. (b) ISD-VLG and ISLG-VLG curves measured at VSD ) 10 mV for the CNTMESFET. (c) ISD-VLG curves measured at VSD ) 10 mV for the CNT-FET and CNT-MESFET. The inset shows a schematic diagram of the CNT-FET with the PDMS well. (d) ISD-VLG curves measured at VSD ) 10 mV for the CNT-MOSFET, CNT-FET, and CNT-MESFET. The inset shows a schematic diagram of the CNT-MOSFET with the PDMS well.
possibly be ascribed to the larger liquid gate capacitance of the CNT-MESFET. To further clarify the effects of gate capacitance on the ISD-VLG transfer curves, we also fabricated a CNT-MOSFET (metal oxide semiconductor field effect transistor) with a 50 nm thick Al2O3 dielectric between the CNT channel and the Au top gate (inset of Figure 3d). Figure 3d shows the ISD-VLG curve of the CNT-MOSFET measured at VSD ) 10 mV using the Ag/AgCl reference electrode. As VLG increased, ISD of the CNT-MOSFET decreased more slowly than ISD of the CNTMESFET or the CNT-FET. The plot of dISD/dVLG versus VLG for the three devices, CNT-MESFET, CNT-FET, and CNT-MOSFET, shows clear differences (Supporting Information Figure S2b). Since the liquid gate capacitance of the CNT-MOSFET was smaller than that of the CNT-MESFET or the CNT-FET because of the capacitance of the Al2O3 dielectric layer connected in series, this result supported the idea that a larger gate capacitance induced a more rapid decrease of ISD with increasing VLG. To monitor the DNA hybridization in real time, the CNTMESFET was covered with a PDMS microfluidic channel instead of the PDMS well (inset of Figure 4a). Then, probe single-stranded DNA (ssDNA) molecules with thiol groups at one end were immobilized on the surface of the Au top gate. Immobilization was achieved via Au-S bonds by introducing a 1 µM solution of 30-mer probe ssDNA molecules (5′-SHAAGTCAGTTATACGCGTCTAGTACCGTTTG-3′, purchased from M. Biotech Inc.) in PBS into the microfluidic channel. Following immobilization, the CNT-MESFET was washed thoroughly by flowing PBS into the channel to remove the excess probe DNA. For the hybridization experiments, a solution of 30-mer target ssDNA (5′-CAAACGGTACTAGACGCGTATAACTGA CTT-3′), which was completely complementary © 2010 American Chemical Society
to the probe ssDNA, was added into the microfluidic channel. While solutions of complementary target ssDNA molecules with different concentrations flowed into the microfluidic channel, the ISD of the CNT-MESFET was recorded at VBG or VLG ) 0 V as a function of time (Figure 4a). As reported for the CNT-FET biosensors,4,6 ISD decreased after DNA hybridization and further decreased with increasing concentrations of target ssDNA. To investigate the origin of the decreased conductance, the ISD-VBG and ISD-VLG curves were measured before and after DNA hybridization (Figure 4(b)). After DNA hybridization, the ISD-VBG curve was shifted to lower VBG and the curve shape was nearly unchanged. On the other hand, the ISD-VLG curve was also shifted to lower VBG, but ISD was significantly reduced at negative voltages of VLG (inset of Figure 4b). This difference might be ascribed to the different gate capacitances of VBG and VLG,18 since DNA hybridization on the Au top gate presumably resulted in the change in the liquid gate capacitance, whereas it led to no change in the bottom gate capacitance. For both the ISD-VBG and ISD-VLG curves, ISD decreased at VBG or VLG ) 0 V after DNA hybridization. This decrease in ISD may possibly be understood in terms of the energy band diagrams shown in Figure 4c. DNA hybridization occurring on the Au surface is known to decrease the effective work function of Au.19,20 After hybridization, the potential of the CNT channel covered with the Au top gate is lowered, while the potential of the uncovered CNT portion is unchanged. As a result, we can imagine the band diagram of Figure 4c; the hybridization-induced depletion (as indicated by the arrows) draws a large channel resistance, causing the decrease in ISD and the shift in Vth. When we plotted the measured conductance normalized by the conductance before hybridization (G/G0) as a function 2757
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755-–2760
FIGURE 4. (a) Real-time monitoring of ISD at VSD ) 0.1 V during DNA hybridization in PBS. Arrows indicate the injection time of target ssDNA solutions with different concentrations. The inset shows a schematic diagram of the CNT-MESFET covered with a PDMS microfluidic channel. (b) ISD-VBG curves of CNT-MESFET biosensor measured at VSD ) 50 mV before and after DNA hybridization. The inset shows the ISD-VLG curves measured at VSD ) 50 mV before and after DNA hybridization. (c) Schematic energy band diagrams for the CNT-MESFET after DNA immobilization (dashed curve) and hybridization (solid curve). (d) G/G0 versus the concentration of target ssDNA molecules for the CNTMESFET and CNT-FET. (e) G/G0 versus the concentration of MMP-9 for the CNT-MESFET and CNT- FET. (f) Real-time monitoring of ISD at VSD ) 0.1 V while a 1 µM solution of one-base mismatch DNA solution was added, followed by the introduction of a 1 µM solution of target DNA solution.
of the target ssDNA concentration (Figure 4d), we observed a linear relationship between G/G0 and the logarithm of the target ssDNA concentration. For comparison, we have also displayed the plot of G/G0 versus target DNA concentration measured using the CNT-FET in Figure 4d, where 1-pyrenebutanoic acid succinimidyl ester and amine-modified ssDNA (5′-NH2-AAGTCAGTTATACGCGTCTAGTACCGTTTG-3′) were used as a linker and probe DNA molecules, as reported by others.7 In comparison with the CNT-FET, the CNT-MESFET exhibited a higher sensitivity. In addition to DNA hybridization, we also detected matrix metallopeptidase-9 (MMP-9) using both CNT-MESFET and CNT-FET and compared their sensitivity, where anti-MMP-9 was used as a receptor of MMP-9 (see Supporting Information for the experimental details). As in DNA hybridization, ISD decreased upon the binding of MMP-9 to anti-MMP-9 immobilized on the top gate or on the CNT surface. Figure 4e shows the plots of G/G0 versus the logarithm of MMP-9 concentration measured for CNT-MESFET and CNT-FET. The CNT-MESFET exhibited a steeper slope, supporting that the CNT-MESFET biosensor © 2010 American Chemical Society
offers higher sensitivity than the CNT-FET. This higher sensitivity can probably be attributed to the band bending in the bulk CNT, as illustrated in Figure 4c. To determine whether it is possible to discriminate between hybridization with complementary and one-basemismatch DNA using the CNT-MESFET, we measured the conductance as a function of time while one-base-mismatch DNA (5′- CAAACGGTACGAGACGCGTATAACTGACTT-3′) was added. Unlike the completely complementary target ssDNA, the addition of one-base-mismatch DNA did not lead to a decrease in ISD. Depending on the sensor, ISD was restored almost to the initial value (Supporting Information Figure S3) or slightly increased (Figure 4f). However, the subsequent introduction of target DNA molecules resulted in a decrease in ISD (Figure 4(f)), demonstrating that complementary and one-base-mismatch DNA molecules would be distinguished. Finally, the CNT-MESFET was fabricated on the quartz substrate for a dual-mode biosensor (Figure 5a). Since the large top gate was necessary for the SPR measurements, ultralong SWNTs (>18 mm) were grown on the quartz 2758
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755-–2760
FIGURE 5. (a) FESEM image of the CNT-MESFET fabricated on the quartz substrate. (b) |∆G|/G0 versus the concentration of target ssDNA solution for three different CNT-MESFETs on SiO2/Si substrates. (c) SPR curves measured for the CNT-MESFET on the quartz substrate as-prepared, after DNA immobilization and hybridization. The inset is an expanded view of SPR minima. (d) SPR curves measured for two different CNTMESFETs on the quartz substrate as-prepared (black curves), after DNA immobilization (red curves) and hybridization (blue curves). (e) Realtime response of |∆G|/G0 and normalized reflectance simultaneously measured during DNA hybridization. The SPR reflectance was measured at 38°.
substrate by thermal CVD. Briefly, 10 mM of FeCl3 solution was deposited on the edge of the quartz substrate by microcontact printing or dipping. The quartz substrate was placed in the middle of a 1 in. quartz tube and gradually heated to 950 °C in an argon/hydrogen gas atmosphere (200 scm/60 scm), then held at 950 °C for 10 min to form Fe nanoparticles. Then, a flow of methane (100 scm) and hydrogen gas (60 scm) at 950 °C for 1 h led to the growth of ultralong SWNTs.21-23 For the electrical measurements, Au/Cr electrodes were prepared and rapid thermal annealing was performed at 450 °C for 30 s under a vacuum. Subsequently, the Au top gate of 2 × 21 mm2 was fabricated in the middle of the CNTs by photolithography and lift-off techniques. We measured the ISD-VSD curves after DNA immobilization and hybridization (Supporting Information Figure S4). As expected from the data in Figure 4a, the conductance was reduced by immobilization of probe ssDNA molecules on the surface of the top gate and further reduced by subsequent hybridization, even though several CNTs were connected between the source and drain electrodes. Figure 5b depicts the plots of G/G0 versus the logarithm of the target ssDNA concentration obtained from the three different CNT-MESFETs on SiO2/Si substrates. For the three different sensors, we observed nearly parallel lines, although their intercepts © 2010 American Chemical Society
were different. These findings suggest that sensor-to-sensor variation might be calibrated if a complementary property, independent of the sensor, is measured concurrently, and the sensor-dependent intercept can be determined. We therefore measured SPR characteristics using a custom optical setup based on an angle-scanning scheme and implemented with dual concentric motorized rotation stages (URS75PP, Newport, Irvine, CA). Light from a He-Ne laser (36 mW, λ ) 632.8 nm, Melles-Griot, Carlsbad, CA) is TM polarized and intensity-modulated by an optical chopper. The light is then incident through an SF10 prism substrate on the sample surface that is index-matched to the prism. The reflected light is detected by a photodiode (818-UV, Newport, Irvine, CA). The angular resolution of the rotation stage is 0.0002° (nominal). The measured signal is taken by a low noise lock-in amplifier (Model SR830, Stanford Research Systems, Sunnyvale, CA). The detection procedure is automated by LabView (National Instruments, Austin, TX). The SPR curves, which were simultaneously measured with the ISD-VSD curves (Supporting Information Figure S4), are presented in Figure 5c. Immobilization of probe ssDNA molecules on the Au top gate led to an angle displacement of 0.12° in SPR reflectance minimum. After DNA hybridization caused by the addition of a 1 µM solution of target ssDNA molecules, the SPR angle was further shifted by 2759
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755-–2760
Supporting Information Available. Experimental details and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org.
0.10°, although the SPR angle was unchanged by adding a 1 µM solution of noncomplementary ssDNA molecules. These results are similar to the results reported by others.14,24 To evaluate whether different sensors yield the same displacement of SPR angle for a given concentration, we measured the SPR curves for another sensor prepared similarly (Figure 5d). After a 1 µM solution of target ssDNA molecules was added to induce DNA hybridization, the SPR angle was shifted by approximately 0.10°, indicating that the SPR measurements provided reproducible data independently of the sensor. Figure 5e displays the real-time response of conductance and SPR reflectance measured simultaneously while the DNA molecules were hybridized. From these simultaneously measured conductance and SPR reflectance, the value of the sensor-dependent intercept (Figure 5b) is possibly estimated, which is necessary to calibrate the sensor-to-sensor variation. In summary, we have fabricated a CNT-based biosensor with a MESFET structure to develop a dual-mode biosensor for electrical and SPR measurements. The Au strip, which was deposited on the middle of the CNT channel, acted as the top gate because of the Schottky contact between the Au strip and the CNT. For comparison, we measured the ISD-VLG transfer curves for CNT-MESFET, CNT-FET, and CNTMOSFET. The CNT-MESFET exhibited the steepest transition of ISD because it had the largest liquid gate capacitance among the three devices. DNA hybridization occurring on the Au top gate of CNT-MESFET decreased the ISD of the CNTMESFET, which can probably be attributed to the change in the effective work function of Au, leading to band bending in the bulk CNT. To measure ISD and SPR at the same time, CNT-MESFET with ultralong CNTs was fabricated on a quartz substrate. DNA hybridization on the Au top gate resulted in a shift in the SPR reflectance minimum as well as the decrease in the ISD. Since SPR measurements provided reproducible results independently of the sensor, these results demonstrated that the sensor-to-sensor variation of conductance measurements might be calibrated using the SPR data.
REFERENCES AND NOTES (1) (2) (3) (4)
(5) (6) (7)
(8)
(9) (10) (11)
(12) (13) (14) (15) (16) (17) (18) (19) (20) (21)
(22)
Acknowledgment. This work was supported by NRF through the National Core Research Center for Nanomedical Technology (R15-2004-024-00000-0) of Yonsei University and the Nano System Institute National Core Research Center of Seoul National University.
© 2010 American Chemical Society
(23) (24)
2760
Besteman, K.; Lee, J. O.; Wiertz, F. G. M.; Heering, H. A.; Dekker, C. Nano Lett. 2003, 3, 727. So, H. M.; Won, K.; Kim, Y. H.; Kim, B. K.; Ryu, B. H.; Na, P. S.; Kim, H.; Lee, J. O. J. Am. Chem. Soc. 2005, 127, 11906. Maehashi, K.; Katsura, T.; Kerman, K.; Takamura, Y.; Matsumoto, K.; Tamiya, E. Anal. Chem. 2007, 79, 782. Dong, X. C.; Lau, C. M.; Lohani, A.; Mhaisalkar, S. G.; Kasim, J.; Shen, Z. X.; Ho, X. N.; Rogers, J. A.; Li, L. J. Adv. Mater. 2008, 20, 2389. Martinez, M. T.; Tseng, Y. C.; Ormategui, N.; Loinaz, I.; Eritja, R.; Bokor, J. Nano Lett. 2009, 9, 530. Star, A.; Tu, E.; Niemann, J.; Gabriel, J. C. P.; Joiner, C. S.; Valcke, C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 921. Li, C.; Curreli, M.; Lin, H.; Lei, B.; Ishikawa, F. N.; Datar, R.; Cote, R. J.; Thompson, M. E.; Zhou, C. W. J. Am. Chem. Soc. 2005, 127, 12484. Chen, R. J.; Bangsaruntip, S.; Drouvalakis, K. A.; Kam, N. W. S.; Shim, M.; Li, Y. M.; Kim, W.; Utz, P. J.; Dai, H. J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 4984. Star, A.; Gabriel, J. C. P.; Bradley, K.; Gruner, G. Nano Lett. 2003, 3, 459. Boussaad, S.; Tao, N. J.; Zhang, R.; Hopson, T.; Nagahara, L. A. Chem. Commun. 2003, 13, 1502. Chen, R. J.; Choi, H. C.; Bangsaruntip, S.; Yenilmez, E.; Tang, X. W.; Wang, Q.; Chang, Y. L.; Dai, H. J. J. Am. Chem. Soc. 2004, 126, 1563. Byon, H. R.; Choi, H. C. J. Am. Chem. Soc. 2006, 128, 2188. Nelson, B. P.; Grimsrud, T. E.; Liles, M. R.; Goodman, R. M.; Corn, R. M. Anal. Chem. 2001, 73, 1. Su, X. D.; Wu, Y. J.; Knoll, W. Biosens. Bioelectron. 2005, 21, 719. Sze, S. M. Physics of Semiconductor Devices, 2nd ed.; Wiley: New York, 1981. Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications, 2nd ed.; Wiley: New York, 2001. Heller, I.; Kong, J.; Williams, K. A.; Dekker, C.; Lemay, S. G. J. Am. Chem. Soc. 2006, 128, 7353. Heller, I.; Janssens, A. M.; Mannik, J.; Minot, E. D. S. G.; Lemay, S. G. Nano Lett. 2008, 8, 591. Hansen, D. C.; Hansen, K. M.; Ferrell, T. L.; Thundat, T. Langmuir 2003, 19, 7514. Tang, X. W.; Bansaruntip, S.; Nakayama, N.; Yenilmez, E.; Chang, Y. L.; Wang, Q. Nano Lett. 2006, 6, 1632. Zheng, L. X.; O’Connell, M. J.; Doorn, S. K.; Liao, X. Z.; Zhao, Y. H.; Akhadov, E. A.; Hoffbauer, M. A.; Roop, B. J.; Jia, Q. X.; Dye, R. C.; Peterson, D. E.; Huang, S. M.; Liu, J.; Zhu, Y. T. Nat. Mater. 2004, 3, 673. Hong, B. H.; Lee, J. Y.; Beetz, T.; Zhu, Y. M.; Kim, P.; Kim, K. S. J. Am. Chem. Soc. 2005, 127, 15336. Huang, L. M.; White, B.; Sfeir, M. Y.; Huang, M. Y.; Huang, H. X.; Wind, S.; Hone, J.; O’Brien, S. J. Phys. Chem. B 2006, 110, 11103. Su, X. D.; Wu, Y. J.; Robelek, R.; Knoll, W. Langmuir 2005, 21, 348.
DOI: 10.1021/nl100125a | Nano Lett. 2010, 10, 2755-–2760