- Email: [email protected]
polypyrrole addressable intra-connects
Synthetic Metals 159 (2009) 462–466
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Electrical and optical properties of carbon nanotube/polypyrrole addressable intra-connects Seon Woo Lee a , Haim Grebel a,∗ , Avi Kornblit b , Daniel Lopez b a b
Department of Electrical and Computer Engineering, Electronic Imaging Center at New Jersey Institute of Technology, 161 Warren St., Newark, NJ 07102, United States New Jersey Nanotechnology Consortium (NJNC), Alcatel-Lucent Technologies Bell Labs, Murray Hill, NJ 07974, United States
a r t i c l e
i n f o
Article history: Received 20 February 2008 Received in revised form 7 August 2008 Accepted 10 November 2008 Available online 7 January 2009 Keywords: Polypyrrole Carbon nanotube intra-connects Electrochemical polymerization
a b s t r a c t Carbon nanotube (CNT) intra-connects (bridges spanning across in-plane electrodes) were electroplated with polypyrrole (PPy), an electrically conductive polymer (ECP). Sharp metal electrodes initiated the CNT growth at pre-selected locations. The CNT bridge was then used as an electrode for conductive polymer electro-deposition. The samples were characterized by Raman spectroscopy and current–voltage measurements. We found that current–gate voltage (Ids –Vgs ) characteristics changed dramatically for the electroplated structures when the polymer exceeded a threshold thickness, in the order of 80 nm. In addition, the CNT/PPy structures exhibited large sensitivity to UV radiation: the current substantially reduced upon irradiation with moderate UV intensity values. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Carbon nanotubes (CNT) and electrically conductive polymer (ECP) have attracted much interest in recent years. Since its first discovery [1], CNT have been studied extensively as biosensors, field effect transistors and single electron devices [2–4]. CNT have shown remarkable electrical, optical, chemical and mechanical properties [5–8]. However, growth of CNT at pre-designated positions remains a challenge. Metal contacts for CNT-based devices are mostly postfabricated after dispersing the CNT on a substrate. Random growth of CNT between electrodes has been demonstrated, too [9,10]. We have demonstrated [11] a reliable growth of CNT intra-connect between pre-fabricated electrodes: the electrode tips were made sharp enough to initiate growth of a CNT channel between them at a yield of 30%. Here we use this technique to fabricate CNT intra-connects, which are further electroplated with conductive polymer. Polypyrrole (PPy) is a widely used electrically conductive polymer (ECP) for electronic, optical and biological purposes. Its properties are controllable by adjusting the doping level and type of dopant [12–16]. Polymeric-based and all-polymer transistor have been realized as well ([17–18] and references therein). However, most of the CNT/PPy structures thus far, have been realized in a bulk or thin film forms, portraying a complex charge hopping between the dispersed CNT and the backbones of PPy. It has been shown recently [19] that the detection of molecules is dra-
∗ Corresponding author. E-mail address: [email protected] (H. Grebel). 0379-6779/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.synthmet.2008.11.015
matically improved when employing single channel field effect transistors (FET). One, therefore, may aim at functionalized low dimensional polymeric channels as sensitive biosensors. Growth of low dimension polypyrrole channels is very challenging. Photoresist materials, used in the process of nano-device fabrication, add undesired surface states to the structure and its removal poses a substantial difficulty [20]. A different method would be to fabricate CNT channel(s) first; then, use the resultant bridge as an electrode for further electro-deposition of the ECP. Yet, the question to be asked is: what would be a desirable polymeric thickness on top of that CNT ‘electrode’? To this end we opt to fabricate a platform for extremely thin wire transducers made of CNT/PPy complexes. As we shall see below, despite the apparent imperfection of the CNT intra-connect (multi-channels made of either SWCNT or MWCNT), the overall device response has been found to be independent of the number of separated CNT channels involved. As will also be apparent below, the devices exhibited a sharp characteristics transition as the polymer sheath became larger than 80 nm. 2. Experiments The intra-connects have been fabricated between a layout of metal electrode tips (Fig. 1) using chemical vapor deposition (CVD). A detailed description of the process and the electrodes is provided elsewhere [20]. Typical distance between the two electrode tips was 1 m though the electrode layout had patterns of co-aligned and laterally shifted tips (Fig. 1 inset) as well. The morphology, electrical conductivity, photo-conductivity, optical properties of CNT intra-connects were then studied by the use of scanning elec-
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
Fig. 1. SEM image of the metal electrodes. The distance between the two tips was 1 m. Electrode configuration included lateral tip displacement as well (inset). The CNT intra-connect was fabricated by the use of CVD and later was electroplated with PPy. The polymer was coating the electrodes as well as the bridge but was confined only to the conductive area (Fig. 2(a)).
463
trodes and the CNT bridge alike yet, was limited to only conductive surfaces. The sample was later cleaned with deionized water and let dry out under nitrogen gas. Randomly dispersed film experiments employed single-wall carbon nanotubes (SWCNT). The tubes were purchased from CarboLex Co. with 60–70% purity, purified and dispersed by use of a sonicator in ethanol for a few hours. As-purchased tubes display a distribution of diameters. However after purification and functionalization, the majority of tubes were of (11,9) type having a diameter of 1.37 nm, as determined by their low frequency Raman spectra. The tubes were functionalized with either poly(vinyl pyrrolidone) (PVP) of molecular weight (MW) 40,000 and poly(ethylene imine) (PEI) of MW 630,000, in order to obtain wrapped tubes either in small bundles or, as individuals, p- or n-type, respectively. The ratio of the polymer and SWCNT was fixed at 2:1 for both cases resulting in uniform films. Wrapping was helpful in minimizing the tube agglomeration. SEM images of CNT intra-connects before and after the electro-deposition are shown in Fig. 2(a) and (b), respectively. High-resolution field emission scanning electron microscopy (FESEM, LEO 1530VP) has been used. 3. Results and discussion
tron microscopy (SEM), atomic force microscope (AFM), Raman spectroscopy, current–voltage (Ids –Vds ) and current–gate voltage (Ids –Vgs ) characteristic measurement. Polypyrrole (PPy) was synthesized by electrochemical oxidation of pyrrole. A 273 EG&G Princeton Applied Research Potentiostat/Galvanostat was used for the electro-polymerization process. The electro-polymerization process was carried out in a threeelectrode-cell configuration. The cell contained aqueous solution of 0.5 M pyrrole and 0.5 M potassium chloride (KCl) (Sigma–Aldrich) without further purification. The CNT intra-connects were used as working electrodes. A platinum wire and Ag/AgCl electrode were used as a counter and a reference electrode, respectively. A constant potential bias of 0.8 V was applied to enable the deposition of PPy. The film thickness was determined by the deposition time, typically on the order of 30 s. The black film covered the metal elec-
3.1. Raman spectroscopy Raman spectroscopy was used to evaluate the intra-connects. The electrodes were imaged in the far field and the beam of Ar ion laser at 514.5 nm was focused accurately in-between the tips. The tip construction made it very easy to identify the CNT intra-connect under test. A double spectrometer (0.25 cm) and a cooled CCD array were used to detect the scattered signals. The background signal was subtracted and the experimental data was fitted with several Guassian distributions. Results for the high-frequency spectra are shown in Fig. 2(c). By fitting, one can identify three major peaks, as expected for both CNT and PPy [21]. These are: CNT-only: 1350, 1585, 1619; PPy only: 1330, 1370, 1584 and the complex CNT/PPy: 1357, 1585 cm−1 , respectively. The relative peak position of the com-
Fig. 2. SEM images of multiple (three) CNT intra-connects before (a) and after (b) electro-polymerization with PPy. (c) High-frequency Raman scattering from only CNT intra-connect, PPy electroplated on conductive glass and CNT/electroplated-PPy intra-connects. The peaks for each component was, only CNT: 1350, 1585, 1619 cm−1 , only PPy: 1330, 1370, 1584 cm−1 by use of fitting. The electroplated bridges exhibited peaks at 1357, 1585 cm−1 , respectively.
464
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
plex changed: the two CNT peaks at 1600 cm−1 merged into one, implying a weak interaction between the two layers. From these results and the lack of low frequency Raman spectra we deduce that these tubes were mostly MWCNT. No such signature was obtained when the laser spot was irradiating spots away from the electrode region. 3.2. Current–voltage measurements Current–voltage (I–V) measurements were taken (Fig. 3(a)). The current range was on the order of 10−6 A at 1 V for individual and multiple CNT intra-connects alike. When exposed to white light (150 mW/cm2 , > 400 nm) the CNT intra-connect exhibited small photo-conductance effect. Upon exposure to UV light (4 mW/cm2 , = 355 nm) the conductance decreased by approximately 8%, probably because of oxygen desorption (see below). While it seems that some transformation of the non-conducting surfaces occurred after the electroplating step, careful conductance measurements revealed that these regions did not contribute to the I–V curves. We used sharp electrode probes, which were mounted on threedimensional translational stages to reach every point on the substrate surface. After polymerization with PPy, the channel conductivity has been enhanced by more than ten times: for example, current values measured at 1 V were 2.97 A for CNT-only bridges Fig. 3(a); it was 32.18 A for the CNT/PPy bridge complex with accuracy of nA. The CNT/PPy complex exhibited little photo-conductance Fig. 3(b); photo-conductance was therefore, attributed to only the CNT component since PPy is not sensitive to white light. The CNT/PPy structure was very sensitive to UV light exposure (4 mW/cm2 , = 355 nm) though Fig. 3(c); in many cases, the current dropped to zero and the bridge became open in less than a minute (not shown). It recovered in less than 1 min to its previous state when the UV source was removed. UV irradiation effects on CNT may be attributed to oxygen desorption through reduction of hole carriers [22,23]. As we shall see below, the effect on the PPy may be attributed to deep level impurities. As-grown carbon nanotubes (CNT) are naturally p-type and require negative gate voltage to operate in a gated configuration. So are the characteristics of electroplated polypyrrole. The drain-
source voltage was fixed, Vds = 1 V, while the gate voltage varied between −3 V to +3 V. Silicon substrate was used as a back-gate electrode with 20 nm thick oxide layer (Fig. 4(a)). Fig. 4(b) shows a typical Ids –Vgs characteristic of a as-grown CNT intra-connect. The current abruptly increased when a negative gate voltage was applied. The large saturation behavior may be due to weak gate dependence of the metallic-like MWCNT and the large gate capacitance (see below). We also note that the position of the threshold voltage at which the transition occurs. The threshold value is small yet negative (Vg < 0) for the mostly MWCNT multiple and separated channels of this study. The threshold gate value is also negative however, much larger (on the order of −2 V to −5 V) for single SWCNT intra-connects reported elsewhere [11]. Such observation accentuates the difference between SWCNT and MWCNT p-type channels. The intra-connects were simulated by commercial computer aided design (CAD) software (PSPICE) and by a direct implementation of current equations for each element (Fig. 4e) using MathCad: we assumed that the intra-connect is a p-channel contacting the electrodes through leaky back-to-back diodes. p-Channels are characterized by a constant, C1 = C0 (W/L): here, is the channel mobility, C0 is the gate capacitance and W/L is the ratio between the channel width to the channel length. While W/L is very small in our case (W ∼ a few nanometers; L ∼ 1 m), the mobility of the channel as well as, the gate capacitance is large. This implies a relatively large value for the constant C1 and is manifested as a step behavior in the Ids –Vgs curve (curve 2). As the thickness of the polymeric sheath on the CNT intra-connect increases, W/L increases only slightly; the capacitance of the structure C0 remains basically unaffected however; the overall channel mobility is substantially reduced—by two orders of magnitude due to the low mobility value of the polymer (CNT ∼ 5 × 104 cm2 /V-s, Ref. [24]; PPy ∼ 500 cm2 /V-s, Ref. [25]). As a result, the Ids –Vgs curve becomes more linear (curves 1 and 3). The difference between the current amplitudes of Fig. 4b–d is attributed to the decrease in the overall circuit resistance since the polymers coats the metal electrodes, leading to the intra-connect, as well. When the polymer coats the electrodes and the intra-connect, the contact barrier between the CNT channel and the electrodes is substantially reduced. The seemingly 2:1 factor between the ON and OFF states of the bridge depends on the p-channel and the con-
Fig. 3. (a) I–V characteristics before and after polymerization of CNT intra-connects. (b) Effect of light on only CNT intra-connects. White light intensity was 0.15 W/cm2 and the UV intensity was 4 mW/cm2 . (c) Effect of light on the CNT/PPy intra-connect. The bridges were largely insensitive to white light but very sensitive to the UV light at 355 nm.
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
465
Fig. 4. (a) Schematic configuration of the electronic circuit. (b) Ids –Vg characteristics of an intra-connect with only CNT. (c) Ids –Vg characteristics for CNT/PPy structure with 80 nm thick PPy. (d) Ids –Vg characteristics for CNT/PPy structure with thicker polymeric layers, assessed at 360 nm and 580 nm, respectively, by the use of AFM. (e) Equivalent circuit and simulations: the p-channel was attached to the contacts through two leaky diodes. (1) R = 1 M, C1 = 10−5 . (2) R = 1 M, C1 = 10−4 . (3) R = 100 K, C1 = 10−5 . The threshold voltage was small, VP = 0.01 V.
tact diodes’ properties. Ratios of 3:1 for these CNT intra-connects have been experimentally observed and corroborated by simulations. 3.3. UV measurements The effect of UV radiation on the bridge encouraged us to conduct several more experiments in both air and in vacuum. The experiments were conducted on randomly dispersed films: the CNT were functionalized with PVP (p-type) or PEI (n-type). Contacts to the films were made with spring-loaded copper strips. The films were mostly made of SWCNT. As found before for multiwalled films (MWCNT) [26], the resistance of films increased in vacuum. So were the results for SWCNT/PVP electroplated with PPy. The resistance of the latter further increased upon irradiation with UV (wavelength 355 nm, intensity 4 mW/cm2 at 15 cm from the target). Neither p-type nor, n-type films showed substantial resistance change after irradiation with UV light in vacuum. Under UV radiation in air though, the resistance of n-type films
has reduced and the resistance of p-type samples (including ptype/PPy) has increased as compared to non-irradiated samples. All samples were very slow to recover to their original resistance value while in air. For either type of SWCNT samples, the UV experiments point to the effect of oxygen on the conductance of CNT. Generally, ECP exhibit low-mobility and large concentrations of deep impurities. Experiments on electroplated-PPy samples imply further UV effect. Pure PPy is fairly stable under UV radiation [27] so we postulate that the increase resistance for pure PPy is related to the excitation of deep level impurities, which in turn, impeded the carrier hoping mechanism. Despite long recovery time, the samples did not show sustained damage, exhibited by repetitive experiments. The resistance increase under UV illumination rules out heat effects since in general, pure PPy films exhibit increased conductivity at increasing temperatures [28]. To conclude, experiments with UV light corroborated the experiments on PPy electroplated CNT bridges: the effect on CNT-only bridges were relatively small and their resistance increased due to
466
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
desorption of oxygen. The effect of UV on PPy is by far larger and may be attributed to excitations of deep level impurities. The CNT/PPy complex channel exhibits resistance increase from the effect on both components. 4. Conclusions In summary, we have successfully demonstrated the fabrication of CNT/PPy low dimensional gated channels for potential electrooptic and bio-sensor applications. For the latter, a large response to a binding event is desirable [29]. Such response is enabled if the channel mobility dramatically changes from a very large value (channel saturation, Fig. 4e curve (2)) to a very small value (where initial linear behavior is observed, Fig. 4e curve (1)). This means that the PPy thickness over the CNT electrode ought to be kept below 80 nm. References [1] S. Iijima, Nature 354 (1991) 56. [2] H. Boo, R.A. Jeong, S. Park, K.S. Kim, K.H. An, Y.H. Lee, J.H. Han, H.C. Kim, T.D. Chung, Anal. Chem. 78 (2006) 617. [3] J.A. Misewich, R. Martel, Ph. Avouris, J.C. Tsang, S. Heinze, J. Tersoff, Science 300 (2003) 783. [4] H.W.Ch. Postma, T. Teepen, Z. Yao, M. Grifoni, C. Dekker, Science 293 (2001) 76. [5] E.W. Wong, P.E. Sheehan, C.M. Lieber, Science 277 (1997) 1971. [6] M.F. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F. Kelly, R.S. Ruoff, Science 287 (2000) 637. [7] J.W.G. Wildoer, L.C. Venema, A.G. Rinzler, R.E. Smalley, C. Dekker, Nature 391 (1998) 59–62.
[8] M.S. Dresselhaus, G. Dresselhaus, P. Avouris, Carbon Nanotubes: Synthesis, Structure, Properties and Applications, Springer-Verlag, Berlin, 2001. [9] X. Liu, J. Ly, S. Han, D. Zhang, A. Requicha, M.D. Thompson, C. Zhou, Adv. Mater. 17 (2005) 2727. [10] J. Kong, H.T. Soh, A.M. Cassell, C.F. Quate, H. Dai, Nature 395 (1998) 878. [11] D. Katz, D. Lopez, A. Kornblit, H. Grebel, J. Nanosci. Nanotechnol. 8 (2008) 1. [12] F. Garnier, B. Bouabdallaoui, P. Srivastava, B. Mandrand, C. Chaix, Sens. Actuators B: Chem. 123 (2007) 13. [13] A.K. Wanekaya, Y. Lei, E. Bekyarova, W. Chen, R. Haddon, A. Mulchandani, N.V. Myung, Electroanalysis 18 (2006) 1047. [14] C.C. Bof Bufon, T. Heinzel, Appl. Phys. Lett. 89 (2006) 012104. [15] J. Wang, Y.L. Bunimovich, G. Sui, S. Savvas, J. Wang, Y. Guo, J.R. Heath, H.R. Tseng, Chem. Commun. 29 (2006) 3075. [16] K.H. An, S.Y. Jeong, H.R. Hwang, Y.H. Lee, Adv. Mater. 16 (2004) 1005. [17] H.S. White, G.P. Kittlesen, M.S. Wrighton, J. Am. Chem. Soc. 106 (1984) 5375. [18] F. Garnier, R. Hajlaoui, A. Yassar, P. Srivastava, Science 265 (1994) 1684. [19] P.R. Naira, M.A. Alam, Appl. Phys. Lett. 88 (2006). [20] D. Katz, D.S.-W. Lee, A. Lopez, H. Kornblit, J. Grebel, Vac. Sys. Technol. B 25 (2007) 1191. [21] T.M. Wu, S.H. Lin, J. Polym. Sci.: Part A: Polym. Chem. 44 (2006) 6449. [22] R.J. Chen, N.R. Franklin, J. Kong, J. Cao, T.W. Tombler, Y. Zhang, H. Dai, Appl. Phys. Lett. 79 (2001) 2258. [23] L. Valentini, I. Armentano, J. Kenny, S. Bidali, A. Mariani, Thin Sold Films 476 (2005) 162. [24] T. Dürkop, T. Brintlinger, M.S. Fuhrer, in: H. Kuzmany, J. Fink, M. Mehring, S. Roth (Eds.), Structural and Electronic Properties of Molecular Nanostructures, 242–6 (AIP Conference Proceedings, New York, 2002). [25] E. Garcia-camarero, F. Arjona, C. Guillien, E. Fatas, C. Montemayor, J. Mater. Sci. 25 (1990) 4914. [26] Y.T. Jang, S.I. Moon, J.H. Ahn, Y.H. Lee, B.K. Ju, Sens. Actuators B 99 (2004) 118. [27] Q. Fang, D.G. Chetwynd, J.W. Gardner, Sens. Actuators A 99 (2002) 74, and references therein. [28] P. Stallinga, H.L. Gomes, H. Rost, A.B. Holmes, M.G. Harrison, R.H. Friend, F. Biscarini, C. Taliani, G.W. Jones, D.M. Taylor, Physica B 923 (1999) 273. [29] A. Star, J.C.P. Gabriel, K. Bradley, G. Grulner, Nano Lett. 3 (2003) 459.
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Electrical and optical properties of carbon nanotube/polypyrrole addressable intra-connects Seon Woo Lee a , Haim Grebel a,∗ , Avi Kornblit b , Daniel Lopez b a b
Department of Electrical and Computer Engineering, Electronic Imaging Center at New Jersey Institute of Technology, 161 Warren St., Newark, NJ 07102, United States New Jersey Nanotechnology Consortium (NJNC), Alcatel-Lucent Technologies Bell Labs, Murray Hill, NJ 07974, United States
a r t i c l e
i n f o
Article history: Received 20 February 2008 Received in revised form 7 August 2008 Accepted 10 November 2008 Available online 7 January 2009 Keywords: Polypyrrole Carbon nanotube intra-connects Electrochemical polymerization
a b s t r a c t Carbon nanotube (CNT) intra-connects (bridges spanning across in-plane electrodes) were electroplated with polypyrrole (PPy), an electrically conductive polymer (ECP). Sharp metal electrodes initiated the CNT growth at pre-selected locations. The CNT bridge was then used as an electrode for conductive polymer electro-deposition. The samples were characterized by Raman spectroscopy and current–voltage measurements. We found that current–gate voltage (Ids –Vgs ) characteristics changed dramatically for the electroplated structures when the polymer exceeded a threshold thickness, in the order of 80 nm. In addition, the CNT/PPy structures exhibited large sensitivity to UV radiation: the current substantially reduced upon irradiation with moderate UV intensity values. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Carbon nanotubes (CNT) and electrically conductive polymer (ECP) have attracted much interest in recent years. Since its first discovery [1], CNT have been studied extensively as biosensors, field effect transistors and single electron devices [2–4]. CNT have shown remarkable electrical, optical, chemical and mechanical properties [5–8]. However, growth of CNT at pre-designated positions remains a challenge. Metal contacts for CNT-based devices are mostly postfabricated after dispersing the CNT on a substrate. Random growth of CNT between electrodes has been demonstrated, too [9,10]. We have demonstrated [11] a reliable growth of CNT intra-connect between pre-fabricated electrodes: the electrode tips were made sharp enough to initiate growth of a CNT channel between them at a yield of 30%. Here we use this technique to fabricate CNT intra-connects, which are further electroplated with conductive polymer. Polypyrrole (PPy) is a widely used electrically conductive polymer (ECP) for electronic, optical and biological purposes. Its properties are controllable by adjusting the doping level and type of dopant [12–16]. Polymeric-based and all-polymer transistor have been realized as well ([17–18] and references therein). However, most of the CNT/PPy structures thus far, have been realized in a bulk or thin film forms, portraying a complex charge hopping between the dispersed CNT and the backbones of PPy. It has been shown recently [19] that the detection of molecules is dra-
∗ Corresponding author. E-mail address: [email protected] (H. Grebel). 0379-6779/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.synthmet.2008.11.015
matically improved when employing single channel field effect transistors (FET). One, therefore, may aim at functionalized low dimensional polymeric channels as sensitive biosensors. Growth of low dimension polypyrrole channels is very challenging. Photoresist materials, used in the process of nano-device fabrication, add undesired surface states to the structure and its removal poses a substantial difficulty [20]. A different method would be to fabricate CNT channel(s) first; then, use the resultant bridge as an electrode for further electro-deposition of the ECP. Yet, the question to be asked is: what would be a desirable polymeric thickness on top of that CNT ‘electrode’? To this end we opt to fabricate a platform for extremely thin wire transducers made of CNT/PPy complexes. As we shall see below, despite the apparent imperfection of the CNT intra-connect (multi-channels made of either SWCNT or MWCNT), the overall device response has been found to be independent of the number of separated CNT channels involved. As will also be apparent below, the devices exhibited a sharp characteristics transition as the polymer sheath became larger than 80 nm. 2. Experiments The intra-connects have been fabricated between a layout of metal electrode tips (Fig. 1) using chemical vapor deposition (CVD). A detailed description of the process and the electrodes is provided elsewhere [20]. Typical distance between the two electrode tips was 1 m though the electrode layout had patterns of co-aligned and laterally shifted tips (Fig. 1 inset) as well. The morphology, electrical conductivity, photo-conductivity, optical properties of CNT intra-connects were then studied by the use of scanning elec-
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
Fig. 1. SEM image of the metal electrodes. The distance between the two tips was 1 m. Electrode configuration included lateral tip displacement as well (inset). The CNT intra-connect was fabricated by the use of CVD and later was electroplated with PPy. The polymer was coating the electrodes as well as the bridge but was confined only to the conductive area (Fig. 2(a)).
463
trodes and the CNT bridge alike yet, was limited to only conductive surfaces. The sample was later cleaned with deionized water and let dry out under nitrogen gas. Randomly dispersed film experiments employed single-wall carbon nanotubes (SWCNT). The tubes were purchased from CarboLex Co. with 60–70% purity, purified and dispersed by use of a sonicator in ethanol for a few hours. As-purchased tubes display a distribution of diameters. However after purification and functionalization, the majority of tubes were of (11,9) type having a diameter of 1.37 nm, as determined by their low frequency Raman spectra. The tubes were functionalized with either poly(vinyl pyrrolidone) (PVP) of molecular weight (MW) 40,000 and poly(ethylene imine) (PEI) of MW 630,000, in order to obtain wrapped tubes either in small bundles or, as individuals, p- or n-type, respectively. The ratio of the polymer and SWCNT was fixed at 2:1 for both cases resulting in uniform films. Wrapping was helpful in minimizing the tube agglomeration. SEM images of CNT intra-connects before and after the electro-deposition are shown in Fig. 2(a) and (b), respectively. High-resolution field emission scanning electron microscopy (FESEM, LEO 1530VP) has been used. 3. Results and discussion
tron microscopy (SEM), atomic force microscope (AFM), Raman spectroscopy, current–voltage (Ids –Vds ) and current–gate voltage (Ids –Vgs ) characteristic measurement. Polypyrrole (PPy) was synthesized by electrochemical oxidation of pyrrole. A 273 EG&G Princeton Applied Research Potentiostat/Galvanostat was used for the electro-polymerization process. The electro-polymerization process was carried out in a threeelectrode-cell configuration. The cell contained aqueous solution of 0.5 M pyrrole and 0.5 M potassium chloride (KCl) (Sigma–Aldrich) without further purification. The CNT intra-connects were used as working electrodes. A platinum wire and Ag/AgCl electrode were used as a counter and a reference electrode, respectively. A constant potential bias of 0.8 V was applied to enable the deposition of PPy. The film thickness was determined by the deposition time, typically on the order of 30 s. The black film covered the metal elec-
3.1. Raman spectroscopy Raman spectroscopy was used to evaluate the intra-connects. The electrodes were imaged in the far field and the beam of Ar ion laser at 514.5 nm was focused accurately in-between the tips. The tip construction made it very easy to identify the CNT intra-connect under test. A double spectrometer (0.25 cm) and a cooled CCD array were used to detect the scattered signals. The background signal was subtracted and the experimental data was fitted with several Guassian distributions. Results for the high-frequency spectra are shown in Fig. 2(c). By fitting, one can identify three major peaks, as expected for both CNT and PPy [21]. These are: CNT-only: 1350, 1585, 1619; PPy only: 1330, 1370, 1584 and the complex CNT/PPy: 1357, 1585 cm−1 , respectively. The relative peak position of the com-
Fig. 2. SEM images of multiple (three) CNT intra-connects before (a) and after (b) electro-polymerization with PPy. (c) High-frequency Raman scattering from only CNT intra-connect, PPy electroplated on conductive glass and CNT/electroplated-PPy intra-connects. The peaks for each component was, only CNT: 1350, 1585, 1619 cm−1 , only PPy: 1330, 1370, 1584 cm−1 by use of fitting. The electroplated bridges exhibited peaks at 1357, 1585 cm−1 , respectively.
464
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
plex changed: the two CNT peaks at 1600 cm−1 merged into one, implying a weak interaction between the two layers. From these results and the lack of low frequency Raman spectra we deduce that these tubes were mostly MWCNT. No such signature was obtained when the laser spot was irradiating spots away from the electrode region. 3.2. Current–voltage measurements Current–voltage (I–V) measurements were taken (Fig. 3(a)). The current range was on the order of 10−6 A at 1 V for individual and multiple CNT intra-connects alike. When exposed to white light (150 mW/cm2 , > 400 nm) the CNT intra-connect exhibited small photo-conductance effect. Upon exposure to UV light (4 mW/cm2 , = 355 nm) the conductance decreased by approximately 8%, probably because of oxygen desorption (see below). While it seems that some transformation of the non-conducting surfaces occurred after the electroplating step, careful conductance measurements revealed that these regions did not contribute to the I–V curves. We used sharp electrode probes, which were mounted on threedimensional translational stages to reach every point on the substrate surface. After polymerization with PPy, the channel conductivity has been enhanced by more than ten times: for example, current values measured at 1 V were 2.97 A for CNT-only bridges Fig. 3(a); it was 32.18 A for the CNT/PPy bridge complex with accuracy of nA. The CNT/PPy complex exhibited little photo-conductance Fig. 3(b); photo-conductance was therefore, attributed to only the CNT component since PPy is not sensitive to white light. The CNT/PPy structure was very sensitive to UV light exposure (4 mW/cm2 , = 355 nm) though Fig. 3(c); in many cases, the current dropped to zero and the bridge became open in less than a minute (not shown). It recovered in less than 1 min to its previous state when the UV source was removed. UV irradiation effects on CNT may be attributed to oxygen desorption through reduction of hole carriers [22,23]. As we shall see below, the effect on the PPy may be attributed to deep level impurities. As-grown carbon nanotubes (CNT) are naturally p-type and require negative gate voltage to operate in a gated configuration. So are the characteristics of electroplated polypyrrole. The drain-
source voltage was fixed, Vds = 1 V, while the gate voltage varied between −3 V to +3 V. Silicon substrate was used as a back-gate electrode with 20 nm thick oxide layer (Fig. 4(a)). Fig. 4(b) shows a typical Ids –Vgs characteristic of a as-grown CNT intra-connect. The current abruptly increased when a negative gate voltage was applied. The large saturation behavior may be due to weak gate dependence of the metallic-like MWCNT and the large gate capacitance (see below). We also note that the position of the threshold voltage at which the transition occurs. The threshold value is small yet negative (Vg < 0) for the mostly MWCNT multiple and separated channels of this study. The threshold gate value is also negative however, much larger (on the order of −2 V to −5 V) for single SWCNT intra-connects reported elsewhere [11]. Such observation accentuates the difference between SWCNT and MWCNT p-type channels. The intra-connects were simulated by commercial computer aided design (CAD) software (PSPICE) and by a direct implementation of current equations for each element (Fig. 4e) using MathCad: we assumed that the intra-connect is a p-channel contacting the electrodes through leaky back-to-back diodes. p-Channels are characterized by a constant, C1 = C0 (W/L): here, is the channel mobility, C0 is the gate capacitance and W/L is the ratio between the channel width to the channel length. While W/L is very small in our case (W ∼ a few nanometers; L ∼ 1 m), the mobility of the channel as well as, the gate capacitance is large. This implies a relatively large value for the constant C1 and is manifested as a step behavior in the Ids –Vgs curve (curve 2). As the thickness of the polymeric sheath on the CNT intra-connect increases, W/L increases only slightly; the capacitance of the structure C0 remains basically unaffected however; the overall channel mobility is substantially reduced—by two orders of magnitude due to the low mobility value of the polymer (CNT ∼ 5 × 104 cm2 /V-s, Ref. [24]; PPy ∼ 500 cm2 /V-s, Ref. [25]). As a result, the Ids –Vgs curve becomes more linear (curves 1 and 3). The difference between the current amplitudes of Fig. 4b–d is attributed to the decrease in the overall circuit resistance since the polymers coats the metal electrodes, leading to the intra-connect, as well. When the polymer coats the electrodes and the intra-connect, the contact barrier between the CNT channel and the electrodes is substantially reduced. The seemingly 2:1 factor between the ON and OFF states of the bridge depends on the p-channel and the con-
Fig. 3. (a) I–V characteristics before and after polymerization of CNT intra-connects. (b) Effect of light on only CNT intra-connects. White light intensity was 0.15 W/cm2 and the UV intensity was 4 mW/cm2 . (c) Effect of light on the CNT/PPy intra-connect. The bridges were largely insensitive to white light but very sensitive to the UV light at 355 nm.
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
465
Fig. 4. (a) Schematic configuration of the electronic circuit. (b) Ids –Vg characteristics of an intra-connect with only CNT. (c) Ids –Vg characteristics for CNT/PPy structure with 80 nm thick PPy. (d) Ids –Vg characteristics for CNT/PPy structure with thicker polymeric layers, assessed at 360 nm and 580 nm, respectively, by the use of AFM. (e) Equivalent circuit and simulations: the p-channel was attached to the contacts through two leaky diodes. (1) R = 1 M, C1 = 10−5 . (2) R = 1 M, C1 = 10−4 . (3) R = 100 K, C1 = 10−5 . The threshold voltage was small, VP = 0.01 V.
tact diodes’ properties. Ratios of 3:1 for these CNT intra-connects have been experimentally observed and corroborated by simulations. 3.3. UV measurements The effect of UV radiation on the bridge encouraged us to conduct several more experiments in both air and in vacuum. The experiments were conducted on randomly dispersed films: the CNT were functionalized with PVP (p-type) or PEI (n-type). Contacts to the films were made with spring-loaded copper strips. The films were mostly made of SWCNT. As found before for multiwalled films (MWCNT) [26], the resistance of films increased in vacuum. So were the results for SWCNT/PVP electroplated with PPy. The resistance of the latter further increased upon irradiation with UV (wavelength 355 nm, intensity 4 mW/cm2 at 15 cm from the target). Neither p-type nor, n-type films showed substantial resistance change after irradiation with UV light in vacuum. Under UV radiation in air though, the resistance of n-type films
has reduced and the resistance of p-type samples (including ptype/PPy) has increased as compared to non-irradiated samples. All samples were very slow to recover to their original resistance value while in air. For either type of SWCNT samples, the UV experiments point to the effect of oxygen on the conductance of CNT. Generally, ECP exhibit low-mobility and large concentrations of deep impurities. Experiments on electroplated-PPy samples imply further UV effect. Pure PPy is fairly stable under UV radiation [27] so we postulate that the increase resistance for pure PPy is related to the excitation of deep level impurities, which in turn, impeded the carrier hoping mechanism. Despite long recovery time, the samples did not show sustained damage, exhibited by repetitive experiments. The resistance increase under UV illumination rules out heat effects since in general, pure PPy films exhibit increased conductivity at increasing temperatures [28]. To conclude, experiments with UV light corroborated the experiments on PPy electroplated CNT bridges: the effect on CNT-only bridges were relatively small and their resistance increased due to
466
S.W. Lee et al. / Synthetic Metals 159 (2009) 462–466
desorption of oxygen. The effect of UV on PPy is by far larger and may be attributed to excitations of deep level impurities. The CNT/PPy complex channel exhibits resistance increase from the effect on both components. 4. Conclusions In summary, we have successfully demonstrated the fabrication of CNT/PPy low dimensional gated channels for potential electrooptic and bio-sensor applications. For the latter, a large response to a binding event is desirable [29]. Such response is enabled if the channel mobility dramatically changes from a very large value (channel saturation, Fig. 4e curve (2)) to a very small value (where initial linear behavior is observed, Fig. 4e curve (1)). This means that the PPy thickness over the CNT electrode ought to be kept below 80 nm. References [1] S. Iijima, Nature 354 (1991) 56. [2] H. Boo, R.A. Jeong, S. Park, K.S. Kim, K.H. An, Y.H. Lee, J.H. Han, H.C. Kim, T.D. Chung, Anal. Chem. 78 (2006) 617. [3] J.A. Misewich, R. Martel, Ph. Avouris, J.C. Tsang, S. Heinze, J. Tersoff, Science 300 (2003) 783. [4] H.W.Ch. Postma, T. Teepen, Z. Yao, M. Grifoni, C. Dekker, Science 293 (2001) 76. [5] E.W. Wong, P.E. Sheehan, C.M. Lieber, Science 277 (1997) 1971. [6] M.F. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F. Kelly, R.S. Ruoff, Science 287 (2000) 637. [7] J.W.G. Wildoer, L.C. Venema, A.G. Rinzler, R.E. Smalley, C. Dekker, Nature 391 (1998) 59–62.
[8] M.S. Dresselhaus, G. Dresselhaus, P. Avouris, Carbon Nanotubes: Synthesis, Structure, Properties and Applications, Springer-Verlag, Berlin, 2001. [9] X. Liu, J. Ly, S. Han, D. Zhang, A. Requicha, M.D. Thompson, C. Zhou, Adv. Mater. 17 (2005) 2727. [10] J. Kong, H.T. Soh, A.M. Cassell, C.F. Quate, H. Dai, Nature 395 (1998) 878. [11] D. Katz, D. Lopez, A. Kornblit, H. Grebel, J. Nanosci. Nanotechnol. 8 (2008) 1. [12] F. Garnier, B. Bouabdallaoui, P. Srivastava, B. Mandrand, C. Chaix, Sens. Actuators B: Chem. 123 (2007) 13. [13] A.K. Wanekaya, Y. Lei, E. Bekyarova, W. Chen, R. Haddon, A. Mulchandani, N.V. Myung, Electroanalysis 18 (2006) 1047. [14] C.C. Bof Bufon, T. Heinzel, Appl. Phys. Lett. 89 (2006) 012104. [15] J. Wang, Y.L. Bunimovich, G. Sui, S. Savvas, J. Wang, Y. Guo, J.R. Heath, H.R. Tseng, Chem. Commun. 29 (2006) 3075. [16] K.H. An, S.Y. Jeong, H.R. Hwang, Y.H. Lee, Adv. Mater. 16 (2004) 1005. [17] H.S. White, G.P. Kittlesen, M.S. Wrighton, J. Am. Chem. Soc. 106 (1984) 5375. [18] F. Garnier, R. Hajlaoui, A. Yassar, P. Srivastava, Science 265 (1994) 1684. [19] P.R. Naira, M.A. Alam, Appl. Phys. Lett. 88 (2006). [20] D. Katz, D.S.-W. Lee, A. Lopez, H. Kornblit, J. Grebel, Vac. Sys. Technol. B 25 (2007) 1191. [21] T.M. Wu, S.H. Lin, J. Polym. Sci.: Part A: Polym. Chem. 44 (2006) 6449. [22] R.J. Chen, N.R. Franklin, J. Kong, J. Cao, T.W. Tombler, Y. Zhang, H. Dai, Appl. Phys. Lett. 79 (2001) 2258. [23] L. Valentini, I. Armentano, J. Kenny, S. Bidali, A. Mariani, Thin Sold Films 476 (2005) 162. [24] T. Dürkop, T. Brintlinger, M.S. Fuhrer, in: H. Kuzmany, J. Fink, M. Mehring, S. Roth (Eds.), Structural and Electronic Properties of Molecular Nanostructures, 242–6 (AIP Conference Proceedings, New York, 2002). [25] E. Garcia-camarero, F. Arjona, C. Guillien, E. Fatas, C. Montemayor, J. Mater. Sci. 25 (1990) 4914. [26] Y.T. Jang, S.I. Moon, J.H. Ahn, Y.H. Lee, B.K. Ju, Sens. Actuators B 99 (2004) 118. [27] Q. Fang, D.G. Chetwynd, J.W. Gardner, Sens. Actuators A 99 (2002) 74, and references therein. [28] P. Stallinga, H.L. Gomes, H. Rost, A.B. Holmes, M.G. Harrison, R.H. Friend, F. Biscarini, C. Taliani, G.W. Jones, D.M. Taylor, Physica B 923 (1999) 273. [29] A. Star, J.C.P. Gabriel, K. Bradley, G. Grulner, Nano Lett. 3 (2003) 459.