NANO LETTERS
Sensing Mechanisms for Carbon Nanotube Based NH3 Gas Detection
2009 Vol. 9, No. 4 1626-1630
Ning Peng,† Qing Zhang,*,† Chee Lap Chow,† Ooi Kiang Tan,† and Nicola Marzari‡ Microelectronics Center, School of Electrical and Electronic Engineering, Nanyang Technological UniVersity, 639798 Singapore, and Department of Materials Science and Engineering, Massachusetts Institute of Technology, 77 Massachusetts AVenue, Cambridge, Massachusetts 02139-4307 Received December 30, 2008; Revised Manuscript Received January 30, 2009
ABSTRACT There has been an argument on carbon nanotube (CNT) based gas detectors with a field-effect transistor (FET) geometry: do the response signals result from charge transfer between adsorbed gas molecules and the CNT channel and/or from the gas species induced Schottky barrier modulation at the CNT/metal contacts? To differentiate the sensing mechanisms, we employed three CNTFET structures, i.e., (1) the entire CNT channel and CNT/electrode contacts are accessible to NH3 gas; (2) the CNT/electrode contacts are passivated with a Si3N4 thin film, leaving the CNT channel open to the gas and, in contrast, (3) the CNT channel is covered with the film, while the contacts are open to the gas. We suggest that the Schottky barrier modulation at the contacts is the dominant mechanism from room temperature to 150 °C. At higher temperatures, the charge transfer process contributes to the response signals. There is a clear evidence that the adsorption of NH3 on the CNT channel is facilitated by environmental oxygen.
The carbon nanotube (CNT)1 has a great potential in miniaturized chemical/biological sensing applications. The small size, large surface to volume ratio, and highly sensitive electrical properties make CNTs arguably the ultimate candidate for nanosensors. Kong et al. demonstrated the first CNT gas sensors in 2000.2 Since then, many techniques have been developed to improve the performance of CNT gas sensors, including polymer functionalization, metal nanoparticle decoration, etc.3-10 Although tremendous progress has been achieved, the underlying sensing mechanism still remains unclear. Previously proposed mechanisms include the indirect interaction through the hydroxyl group on SiO2 substrate2 or preadsorbed water layer,11 adsorption of gas molecules at the interstitial sites in the CNT bundle,12 direct charge transfer from the adsorbed gas molecules to CNT,13 and modulation of the Schottky barrier (SB) at CNT/metal contacts,14 etc. Until now, there is no unifying work able to identify the mechanisms. Furthermore, in order to optimize the CNT sensor for practical applications, it is important to understand whether the sensing signals result from the CNT channel and/or the CNT/metal contacts. Using a shortchannel device with passivated CNT/metal contacts by thermally evaporated SiO, Bradley et al. found good sensitivity to NH3 and suggested that NH3 mainly interacts with the CNT channel.15 Zhang et al. argued that when the passivation length was comparable to the depletion length * Corresponding author,
[email protected]. † Microelectronic Centre, Nanyang Technogical University. ‡ Department of Materials Science and Engineering, MIT. 10.1021/nl803930w CCC: $40.75 Published on Web 03/12/2009
2009 American Chemical Society
in the CNT, the contacts could be indirectly affected. In their work, poly(methyl methacrylate) (PMMA) was applied to protect the CNT/metal contacts from NO2 exposure and their devices became insensitive after contact passivation.16 Interestingly, Liu et al. also employed PMMA as a passivation layer. They observed changes in the transfer characteristics upon exposure to NH3 and NO2 for both contact-passivated and channel-passivated devices, suggesting that both the CNT channel and the CNT/metal contacts play a role in the detection process.17 The obvious ambiguity in those reports could arise from the permeable passivation materials used. Moreover, as the experiments were carried out at room temperature and air ambient only, exclusive identification of the sensing mechanisms is not possible. In this paper, we differentiate the sensing mechanisms using a selective Si3N4 passivation technique. The sensing signals from the CNT channel and CNT/metal contacts are truly distinguished. Strikingly distinct sensing performance at various testing conditions is observed. From our results, a clear understanding of gaseous interactions in a CNT sensor with field-effect transistor (FET) geometry is obtained. Single-walled CNTs (SWNTs) were aligned between Ti/ Au source and drain electrodes predefined on a p-type silicon wafer using an ac dielectrophoresis (DEP) technique,18,19 which is simple and cost-effective, suitable for CNT sensor fabrications. Note that the CNTs in this work are on top of Au electrodes and the contact regions are fully accessible to the ambient. A heavily doped Si with a 200 nm thick
Figure 1. Schematics for (a) device 1, as-prepared CNTFET; (b) device 1A, the contacts passivated by Si3N4, and (c) device 2, the central CNT channel passivated by Si3N4.
thermally grown SiO2 layer on top was used as the gate. These devices are typically the Schottky-barrier CNTFETs (SB-CNTFETs). As illustrated in Figure 1, three device structures were employed in our experiments: (1) an asprepared CNTFET with the exposed CNT channel and CNT/ Au contacts, (2) the contacts passivated only, and (3) the channel passivated only. In the latter two, a 500 nm Si3N4 layer was selectively deposited using plasma-enhanced chemical vapor deposition to passivate either the CNT channel or CNT/Au contacts, respectively.20 For each structure, more than three devices were tested and representative results are presented here. Dry air was used as the background gas with a flow rate of 500 sccm in the following experiments unless otherwise stated. NH3 gas was selected as the detecting species to study the sensing mechanisms of the CNT sensors. An as-prepared CNTFET (device 1) showed a sensitive response to small concentrations of NH3 at room temperature (see Figure 2a). It is seen that, under a positive gate voltage, both the sensitivity and reversibility were much higher than those under a negative one, consistent with our previous findings.21 Our results demonstrate a tunable gas sensor through adjusting the gate voltages. In order to experimentally differentiate whether the sensing responses are from the CNT channel and/or the CNT/Au contacts, we passivated the CNT/Au contacts of device 1 with a Si3N4 thin film, leaving the CNT channel open. After the passivation, we found that the device (device 1A) did not respond to NH3 at room temperature, even at a concentration up to 500 ppm, as shown in Figure 2b. For comparison, we only passivated the CNT channel with Si3N4 thin film in another CNTFET (device 2), but uncovering the CNT/Au contacts. Interestingly, device 2 showed a high sensitivity at room temperature (see Supporting Information for details). Therefore, we can unambiguously conclude that NH3 gas induced SB modulation is a dominant mechanism for our CNT gas sensors at room temperature. Actually, PMMA was widely employed as a passivation material to protect the CNT/metal contact regions for gas16,17 and protein sensing.22 However, two major problems exist Nano Lett., Vol. 9, No. 4, 2009
Figure 2. Real-time detection of NH3 at room temperature under various gate voltages (a) before (device 1) and (b) after the contacts passivation (device 1A), respectively. Inset: an atomic force microscope (AFM) image of device 1A after the passivation.
due to the polymer nature of PMMA. First, PMMA is not dense enough to fully passivate the contacts. For example, NO2 was found to penetrate the 2.2 µm thick SU-8/PMMA layer in 30 min.16 Thus, the CNT/metal contacts are inevitably affected by the gradual diffusion of the detecting species through the PMMA layer, so that the role of the contact in the detection could not be eliminated. Second, PMMA is thermally unstable above 100 °C. This is a critical limitation as the adsorptions of some biomolecules and gas molecules on CNTs are enhanced at high temperatures. In contrast, Si3N4 is much denser and it can completely insulate the contacts from chemical environment.23 Meanwhile, its thermal stability allows for high-temperature sensing experiments, as shown later. Device 1A did not show a detectable change upon NH3 exposure when the operating temperatures were below 150 °C. However, the transfer curve started to shift toward negative gate voltage after NH3 exposure at 150 °C; see Figure 3a. Since the contacts were fully isolated from NH3, this parallel shift could result from (1) gating effect from NH3 interaction with SiO2 or (2) charge transfer from NH3 to the CNT channel. With a control sample, i.e., a CNTFET fully covered with SiO2, no response was observed at even 200 °C, suggesting that NH3 could adsorb on the CNT wall and donate electrons to the CNT. Consequently, the Fermi level of the CNT moves toward the conduction band edge so that the threshold voltage VTH becomes more negative. When the testing temperature reached 200 °C, this phenom1627
Figure 3. The transfer characteristics for device 1A with the contacts passivated before and after exposure to NH3 at (a) T ) 150 °C and (b) T ) 200 °C, respectively.
Figure 4. Extracted sensitivities for (a) device 1 and device 1A at T ) 25 °C and (b) device 2 and device 1A at T ) 150 °C, respectively.
enon became more prominent. Progressive shift of the transfer curve in accordance with NH3 concentrations is shown in Figure 3b. Under a first-order estimation, the total charge transferred
reflected in the source-drain current. Our device structure with a CNT on top of metal electrodes could also enhance the SB modulation effect. Once the contact passivation was in place, device 1A essentially does not respond to NH3, implying that the CNT channel is not active to NH3 at room temperature. Figure 4b compares the sensitivities for device 1A and device 2 at T ) 150 °C. At small NH3 concentrations, a low coverage of NH3 on the CNT channel and poor charge transfer efficiency result in a small sensitivity in device 1A. When the NH3 concentration is increased, the sensitivity for device 2 becomes saturated, probably due to limited interaction area in the CNT/Au contacts.
∆Q ) CG∆VTH
where the gate capacitance Cg )
2πεε0L cosh-1(h/r)
the SiO2 dielectric constant ε ) 3.9, and thickness h ) 200 nm. For a SWNT bundle with a length L ∼ 5 µm and a radius r ∼ 5 nm, CG ≈ 0.25 fF. Thus, ∆Q is approximately 0.625 fC or about 4000 electrons, at 200 °C with ∆VTH ) 2.5 V for 500 ppm NH3. If the cross-sectional area of NH3 A ∼ 0.13 nm2, the length of the exposed CNT channel L′ ∼ 3 µm, and the coverage θ ∼ 0.07 (interpolated from the Langmuir plot for 500 ppm NH3 on CNT11), the charge transfer rate f ≈ ∆QA/qθπrL’ is about 0.02 electron. This value is reasonably consistent with typical theoretical predictions. The extracted sensitivities, S ) ∆R/R0, of device 1 and device 1A under three gate voltages at T ) 25 °C are shown in Figure 4a. In device 1, very high sensitivity and significant gate modulation were observed. When the gate voltage is varied from negative to positive, the dominant carrier injection process switches from tunneling to thermionic emission, and the source-drain current becomes very sensitive to the SB height. A small change in the contact SB height due to NH3 adsorption will be prominently 1628
Theoretical studies predict weak interaction between NH3 and pristine CNTs with little charge transfer.2,24,25 The existence of a large activation barrier prevents adsorption of NH3 on perfect CNTs even at high temperatures. However, the adsorption of the gas molecules on defective CNTs could be much easier.26,27 In addition, the adsorption barrier of NH3 on a defective CNT can be further lowered by predissociated oxygen atoms, leading to an enchanted charge transfer rate, as pointed out by Andzelm et al.28 In order to examine the effect of oxygen on our NH3 sensors, we used pure N2 as the background gas and then annealed device 1A at 350 °C for 2 h to degas the adsorbed oxygen. As shown in Figure 5, device 1A became insensitive to NH3 in N2 environment even at T ) 200 °C, with negligible VTH shift in the transfer characteristic. This is a clear evidence that NH3 adsorption on the CNT channel is facilitated by oxygen. Nano Lett., Vol. 9, No. 4, 2009
Figure 5. Response of ISD to various concentrations of NH3 in N2 at T ) 200 °C for device 1A.
From our results, we are able to rule out several possibilities of indirect interactions between NH3 and CNT. First, as the testing environment was totally dry, NH3 adsorption through the water layer is not applicable here. Second, if NH3 could interact through the SiO2 substrate or adsorb inside the CNT bundles, a reduced sensitivity should have been observed after contact passivation. However, our observation that device 1A became totally insensitive to NH3 at room temperature does not support this hypothesis. In fact, charge transfer and SB modulation are the two relevant mechanisms in our CNT sensors. At room temperature, the weak adsorption of NH3 on the CNT wall does not induce any measurable effect on the source-drain current. The sensing signal mainly arises from the contacts. When NH3 molecules are adsorbed on the CNT/ Au interface, the work function of the Au electrode is reduced29 and/or the electrostatic charge balance between the CNT and Au is disturbed by the dipoles of NH3 molecules,14 leading to an increased SB for hole injection. The sensitivity reflected in the source-drain current are, however, gate voltage dependent. As illustrated in Figure 6b, when a negative gate voltage bends the energy band of the CNT upward, the SB width becomes very narrow and holes could tunnel through the barrier, even when the SB height is increased by NH3. In contrast, at a positive gate voltage, the SB width is too thick for tunneling process. Thus, hole injection is only through thermionic emission over the SB height. The source-drain current is then expressed as30 IDS ∼ T2 exp(-qφB /KBT)
where φB is the Schottky barrier height, KB is the Boltzmann’s constant, and T is the temperature in Kelvin. In this case, the sensitivity for the SB modulation is S ) ∆R/R0 ≈ exp(q∆φB /KBT) - 1
Owing to the exponential dependence on the SB height change ∆φB, extremely high sensitivity can be achieved. Our results are consistent with Yamada’s theoretical prediction that SB modulation is most significant when CNT is operating in the depletion mode.14 One disadvantage of the SB dominated sensors is a typical long recovery process at room temperature. Once the operating temperature is increased, the sensitivity degrades shapely due to its exponential dependence with 1/T, as we Nano Lett., Vol. 9, No. 4, 2009
Figure 6. Schematic energy band diagram for: device 1 (a) before and (b) after NH3 exposure (An intrinsic CNT is considered. The work function of the source/drain electrodes is initially near the valence band edge of the CNT and is reduced after NH3 exposure.); device 1A (c) before and (d) after the NH3 exposure (After passivation, the work function of electrodes aligns near the midgap of CNT. The Fermi level of the exposed central CNT channel shifts upward due to electron-doping from NH3.). Legend: red dotted for VGS < 0; green solid for VGS ) 0; blue dashed for VGS > 0.
Table 1. Comparisons between Sensors Operated by SB Modulation and Charge Transfer mechanism sensitivity gate dependence operating temperature reversibility
SB modulation extremely high in depletion mode strong room temperature low, can be improved at positive gate voltage
charge transfer low weak T > 150 °C good
observed previously.21 The nanoscale CNT/metal contacts could also be prone to subtle environmental noises, causing fluctuations during detection. In contrast, using the CNT channel as a sensing element can avoid the problems. In addition, the large surface to volume ratio of CNT offers another merit. At T > 150 °C, when the adsorption barrier is further lowered by oxygen, NH3 adsorption on the defect sites of the CNT becomes favored. Upon the consequent charge transfer, the Fermi level of the exposed central CNT channel moves upward and the energy band shifts downward with respect to those passivated parts. A potential barrier is therefore created and impedes the current flow, as illustrated in Figure 6d.31 Here, we would like to emphasize that the charge transfer effect can be studied only when the contacts are fully protected. Otherwise, as temperature increases, the sensitivity enhancement from the charge transfer effect and the degradation due to SB modulation counteract each other. Our device structure provides a reliable platform for studying the interactions between adsorbed species and CNT channel at various temperatures. Table 1 compares the SB modulation and charge transfer mechanisms. The sensors with the SB modulation usually demonstrate a high sensitivity at room temperature. Their sensing performance can be adjusted through the gate voltages. In contrast, the sensors under the charge transfer mechanism require high working temperatures, showing a good reversibility. Functionalization technique3,26 could be applied in order to enhance the sensitivity for this type of sensors. 1629
In summary, sensing mechanisms for CNT-based NH3 sensors have been systematically studied on a FET platform. We clearly show that the SB modulation at the CNT/metal contacts dominates the sensing performance at room temperature, and the sensor exhibits high sensitivity with good tunability under appropriate gate voltages. At higher temperatures, say 150 °C or above, NH3 molecules start to adsorb on the CNT wall so that the charge transfer process contributes to the sensing signal. NH3 adsorption is confirmed to be facilitated by environmental oxygen. Acknowledgment. We thank the generous support from MOE AcRF Tier 2 Funding (ARC17/07, T207B1203) and the Singapore-MIT Alliance. N.P. would like to acknowledge the Chartered-NTU graduate research scholarship. Supporting Information Available: Descriptions of SWNT suspension, device fabrication, gas sensor system, transfer characteristics of device 1A, N2 background testing, and channel passivated device. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Oberlin, A.; Endo, M.; Koyama, T. J. Cryst. Growth 1976, 32, 335. (2) Kong, J.; Franklin, N. R.; Zhou, C.; Chapline, M. G.; Peng, S.; Cho, K.; Dai, H. Science 2000, 287, 622. (3) Qi, P.; Vermesh, O.; Grecu, M.; Javey, A.; Wang, Q.; Dai, H.; Peng, S.; Cho, K. J. Nano Lett. 2003, 3, 347. (4) An, K. H.; Jeong, S. Y.; Hwang, H. R.; Lee, Y. H. AdV. Mater. 2004, 16, 1005. (5) Bekyarova, E.; Davis, M.; Burch, T.; Itkis, M. E.; Zhao, B.; Sunshine, S.; Haddon, R. C. J. Phys. Chem. B 2004, 108, 51. (6) Zhang, T.; Mubeen, S.; Bekyarova, E.; Yoo, B. Y.; Haddon, R. C.; Myung, N. V.; Deshusses, M. A. Nanotechnology 2007, 18, 165504. (7) Kong, J.; Chapline, M. G.; Dai, H. AdV. Mater. 2001, 13, 1384. (8) Star, A.; Joshi, V.; Skarupo, S.; Thomas, D.; Gabriel, J-C. P. J. Phys. Chem. B 2006, 110, 42. (9) Sun, Y.; Wang, H. H. AdV. Mater. 2007, 19, 2818.
1630
(10) Sayago, I.; Terrado, E.; Aleixandre, M.; Horrillo, M. C.; Fernandez, M. J.; Lozano, J.; Lafuente, E.; Maser, W. K.; Benito, A. M.; Martinez, M. T.; Gutierrez, J.; Munoz, E. Sens. Actuators, B 2007, 122, 75. (11) Bradley, K.; Gabriel, J-C. P.; Briman, M.; Star, A.; Gruner, G. Phys. ReV. Lett. 2003, 91, 218301. (12) Zhao, J.; Buldum, A.; Han, J.; Lu, J. P. Nanotechnology 2002, 13, 195. (13) Chang, H.; Lee, D. J.; Lee, S. M.; Lee, Y. H. Appl. Phys. Lett. 2001, 79, 3863. (14) Yamada, T. Appl. Phys. Lett. 2006, 88, 083106. (15) Bradley, K.; Gabriel, P.; Star, A.; Gruner, G. Appl. Phys. Lett. 2003, 83, 3821. (16) Zhang, J.; Boyd, A.; Tselev, A.; Paranjape, M.; Barbara, P. Appl. Phys. Lett. 2006, 88, 123112. (17) Liu, X.; Luo, Z.; Han, S.; Tang, T.; Zhang, D.; Zhou, C. Appl. Phys. Lett. 2005, 86, 243501. (18) Yamamoto, K.; Akita, S.; Nakayama, Y. Jpn. J. Appl. Phys. 1996, 35, L917. (19) Peng, N.; Zhang, Q.; Li, J.; Liu, N. J. Appl. Phys. 2006, 100, 024309. (20) Peng, N.; Zhang, Q.; Tan, O. K.; Marzari, N. Nanotechnology 2008, 19, 465201. (21) Peng, N.; Zhang, Q.; Lee, Y. C.; Tan, O. K.; Marzari, N. Sens. Actuators, B 2008, 132, 191. (22) Heller, I.; Janssens, A. M.; Mannik, J.; Minot, E. D.; Lemay, S. G.; Dekker, C. Nano Lett. 2008, 8, 591. (23) Kaminishi, D.; Hirokazu, O.; Ohno, Y.; Maehashi, K.; Inoue, K.; Matsumoto, K.; Seri, Y.; Masuda, A.; Matsumura, H. Appl. Phys. Lett. 2005, 86, 113115. (24) Bauschlicher, C. W.; Ricca, A. Phys. ReV. B 2004, 70, 115409. (25) Peng, S.; Cho, K. Nanotechnology 2000, 11, 57. (26) Feng, X.; Irle, S.; Witek, H.; Morokuma, K.; Vidic, R.; Borguet, E. J. Am. Chem. Soc. 2005, 127, 10533. (27) Robinson, J. A.; Snow, E. S.; Badescu, S. C.; Reinecke, T. L.; Perkins, F. K. Nano Lett. 2006, 6, 1747. (28) Andzelm, J.; Govind, N.; Maiti, A. Chem. Phys. Lett. 2006, 421, 58. (29) Bilic, A.; Reimers, J. R.; Hush, N. S.; Hafner, J. J. Chem. Phys. 2002, 116, 8981. (30) Appenzeller, J.; Radosavljevic, M.; Knoch, J.; Avouris, Ph. Phys. ReV. Lett. 2004, 92, 048301. (31) According to our previous findings in ref 20, we consider the effect of oxygen is to modify the energy level alignment at the CNT/Au contacts. Therefore, the Fermi level of Au electrodes aligns near the valence band edge of CNT in air and close to midgap after Si3N4 passivation.
NL803930W
Nano Lett., Vol. 9, No. 4, 2009
