Open Access
Add to BibSonomyAbstract
We investigate the optoelectronic properties of a carbon nanotube (CNT)–metal heterostructure in the terahertz range. On the basis of terahertz time-domain spectroscopy characterization of a double-walled CNT (DWNT) film, we present and analyze the photocurrent measurement for a DWNT–nickel heterojunction illuminated by continuous-wave terahertz radiation. A significant current across the junction directly induced by terahertz excitation is observed and a negative photoconductivity behavior is found to occur in the device. The photocurrent shows a linear response to the bias voltage and the illumination power within the examined range. These phenomena support the feasibility of using CNT–metal heterojunctions as novel terahertz detectors.
© 2014 Optical Society of America
1. Introduction
Terahertz (THz) technology has attracted considerable research interests in recent years due to its promising applications in a variety of fields such as biology, medicine, astronomy, security, environment monitoring, industrial inspection and fundamental science [1]. In order to apply this technology in real-world environments, especially for imaging requirements [2], developing high performance terahertz detectors is vital, but remains a challenging task. Room-temperature, sensitive and multipixel sensors are highly desired in the terahertz frequency range [3]. Conventional terahertz thermal detectors [4], including bolometers, pyroelectric sensors and Golay cells, either need cooling, or offer poor sensitivity, or are not easily integrated into an array. Complementary metal oxide semiconductor-based sub-millimeter wave focal plane arrays [5, 6] have been demonstrated to exhibit good performance for real-time terahertz imaging, but they usually operate well at frequencies below 1 THz. More recently, nanostructured materials and devices have been expected to bring new opportunities for terahertz radiation detection, for example, carbon nanotube (CNT) [7], semiconductor nanowire [8] and graphene [9] transistors.
As a class of nearly ideal quasi-one-dimensional materials, CNTs possess unique electronic and optical properties, making them good candidates for active elements in photodetectors [10]. Different kinds of nanodevices based on CNTs have been proposed for photodetection. Among them, CNT–metal heterojunctions (HJs) show significant photocurrent response even at zero bias in the optical range [11–14] and have potential applications for light sensing [12]. The efficient electron transport in the HJ is attributed to the photoelectric and thermoelectric effects induced by the optical fields. When the photon energy is absorbed by CNTs, the electrons in them might have sufficient energy to enter the metal to produce a photocurrent [13]. Terahertz radiation has relatively low photon energy (4.1 meV at 1 THz). Therefore, at terahertz frequencies, questions concerning whether such HJ structures exhibit similar response behavior and whether they could be useful alternatives as terahertz detectors are required to be investigated. Recently, Fedorov et al. reported the sub-terahertz response of asymmetric CNT devices containing a junction formed by the suspended and non-suspended parts of a dense network of CNTs [15]. However, to the best of our knowledge, research on CNT-metal heterodimension junctions exposed to terahertz radiation has not been explored so far.
In this paper, we characterize the photocurrent response properties of a CNT–metal HJ under terahertz illumination at room temperature. We experimentally investigate the temporal features of the induced current and its dependence on the bias voltage and the incident radiation intensity. The physical mechanisms responsible for photocurrent generation are analyzed and the potentials of CNT-metal HJ for terahertz detection are also discussed.
2. Experimental methods
Fabrication of the HJ device can be illustrated as Figs. 1(a) and 1(b). We chose a planar glass slide as the insulating substrate. An annealed nickel (Ni) foil (size 39 mm × 19 mm, thickness 0.2 mm) with one corner cut off, as marked by the orange dotted frame in Fig. 1(b), was first laid on the substrate, and then a few layers of double-walled CNT (DWNT) films (size ∼ 23 mm × 19 mm) were laid partly on the Ni foil and partly on the glass slide. In order to improve the utilization efficiency of terahertz radiation, the DWNT films over the truncated corner of the Ni foil were concentrated into a narrow strip. The large-scale DWNT films were synthesized by a catalytic chemical vapor deposition method [16]. One single-layered DWNT film has a thickness of about 50–100 nm and it consists of many CNT bundles which highly tangle with each other in a network, as seen from the scanning electron microscope (SEM) image of Fig. 1(c). The overlapping area between DWNTs and Ni forms a CNT–metal HJ. Two separated electrodes made of conductive silver paint were deposited on the two ends of the HJ sample and connected to a sourcemeter (Keithley 2400) used for bias voltage supply and photocurrent measurement. Continuous-wave terahertz radiation with a wavelength of 118.8 μm (frequency 2.52 THz) and an average power of about 60 mW, generated from a far-infrared gas laser (FIRL 100, Edinburgh Instruments Ltd.), was focused onto the junction interface with a focal spot diameter of approximately 1.5 mm. The junction interface is indicated by a green dashed line in Figs. 1(a) and 1(b). By scanning the sample around this interface, an optimal beam spot position corresponding to a maximum current response would be found and it was fixed during the experiment to avoid the influence of position effect [11], as marked by the red solid circle in Fig. 1(b).
Fig. 1 Schematic diagram (a) and optical photograph (b) of a CNT–metal HJ device. The green dashed lines in (a) and (b) indicate the junction interface irradiated by terahertz radiation. The orange dotted frame and the red solid circle in (b) indicate the contour of the Ni foil and the terahertz spot position, respectively. (c) SEM micrograph of the DWNT film. (d) Single-layered DWNT film adhering to a Si wafer. The lower part shows an AFM image of a small area near the film edge. (e) Height profile along the blue solid line marked in the AFM image of (d).
The terahertz excitation on HJ is contributed mainly by the interaction of the incident radiation with CNTs. Consequently, it is necessary to evaluate the terahertz optical and electrical properties of DWNTs individually. For this purpose, we fabricated another sample consisting of a single-layered DWNT film (size ∼ 12 mm × 7 mm) adhering to a high-resistivity silicon (Si) wafer substrate which is transparent to terahertz radiation, as shown in Fig. 1(d). To transfer the nanotube thin film onto the substrate, it was first suspended in deionized water, expanded with the aid of some ethanol and then picked up directly by the Si wafer immersed into the water [17]. The lower part of Fig. 1(d) displays an atomic force microscopy (AFM) image of a 10 μm × 10 μm area near the film edge. Figure 1(e) plots the height profile along the blue line in the AFM image. The actual average film thickness was determined to be about 200 nm from AFM data. This value is larger than expected, probably owing to wrinkle formation of nanotube layers during the transfer process. We used a conventional terahertz time-domain spectroscopy (THz-TDS) system [18] to measure the transmission spectra of the DWNT film. All the terahertz experiments were performed at room temperature.
3. Results and discussion
3.1. THz-TDS characterization of DWNT film
In THz-TDS experiments, the terahertz pulse transmitted through the Si substrate was recorded as the reference signal and the pulse transmitted through the DWNT film and the substrate was recorded as the sample signal. The inset of Fig. 2(a) plots the measured reference and sample signals. Deconvolution of these two temporal waveforms allows us to obtain the frequency dependent complex transfer function of the film under test. Figure 2(a) gives the magnitude of this function within a useful frequency range of 0.1–2.55 THz. We can observe that the transmission amplitude decreases with increasing frequency in this range and has an average value of 40%, almost consistent with a previous report for a 200-nm-thick free-standing DWNT film [19]. The spectral artifacts at relatively high frequencies may result from water vapor absorption. Under the assumption of a homogeneous and planar medium, a theoretical formula for the transfer function of the thin film could be derived as follows based on Fresnel equations [20, 21]:
where ω is the angular frequency, t and r are the Fresnel transmission and reflection coefficients with the subscripts 0, 1, and 2 representing air, Si, and DWNTs, respectively, ñ and d are the complex refractive index and thickness of the DWNT film, respectively, and c is the speed of light in air. Here we approximate the refractive index of air to be 1. Minimizing the difference between the experimental data and the theoretical model enables the optical parameters [refractive index n and absorption coefficient α, related to ñ by ñ = n + jαc/(2ω)] of the sample to be extracted. The open circles and squares in Fig. 2(b) indicate the extracted parameters, whose absolute values are on the same order as previous measurements [19] and that for a single-walled CNT film [22].Fig. 2 Terahertz spectra of DWNT film. (a) Transmission amplitude spectrum. Inset: temporal waveforms of the terahertz pulses transmitted through the Si substrate (solid curve) and the DWNT film on Si substrate (dashed curve). (b) Refractive index (open circles) and absorption coefficient (open squares) extracted from the complex transfer function, where the dashed and solid curves are their fits to the Drude–Lorentz model, respectively.
Actually, the response of the DWNT–Ni HJ will be examined at a frequency of 2.52 THz where the SNR is very low for our THz-TDS system due to the limited spectral range. Therefore, we should employ a physical model to describe the optical or electrical properties of DWNTs and fit the measured data to this model to get their parameters at 2.52 THz. Because CNTs contain both metallic and semiconducting tubes, the dielectric function of CNTs is usually described by the Drude–Lorentz oscillator model [19, 22, 23]:
where ε∞ is the high-frequency dielectric constant, ωp is the Drude plasma frequency, Γ is the Drude damping rate, and ωpi, ω0i, and Γi are the oscillator strength, center frequency, and spectral width of each Lorentz oscillator, respectively. According to the relationship between the complex dielectric constant ε̃ and the complex refractive index ñ that ε̃ = ñ2, we fit the experimental points of n and α with the above model and present the results as the dashed and solid curves in Fig. 2(b). Here, ε∞ is fixed to be 2.5. The initial values of other unknown parameters in Eq. (2) are taken from Ref. [19] and their fitting results are listed in Table 1. Different from Ref. [17], only the Drude part and the broad resonance centered around 1.45 THz are considered. The other two Lorentzian features around 0.45 and 0.75 THz are too weak for our sample and would lead to large errors if they were included in the fitting model. Our derived Drude and Lorentz parameters do not agree very well with that of Ref. [19], which is likely to originate from different tube diameters and densities of the DWNT samples. From the fitted curves, we can estimate that n = 9.2 and α = 2.2 × 106 m−1 at 2.52 THz for our DWNT film. Using the following relation [24]: where σ̃ represents the complex conductivity and ε0 is the free-space permittivity, the real and imaginary conductivity at this frequency are calculated to be 5.3 × 104(Ωm)−1 and 4.7 × 104(Ωm)−1, respectively.
Table 1. Fitted parameters for the Drude–Lorentz oscillator model of the dielectric function.
3.2. Terahertz response of DWNT–Ni HJ
First, we measured the current–voltage (I–V) characteristics of the HJ sample without and with terahertz illumination, as shown by the solid and dashed lines in Fig. 3. The current is defined as positive when its flow direction through the HJ is from metal to CNTs. It is clear that the electrical current has a linear dependence on the bias voltage, demonstrating an ohmic behavior of the HJ. The total resistance of the sample is about 5.6 Ω. In addition, terahertz excitation yields a slight decrease in current. We define the difference between the illuminated current Ip and the dark current Id as ΔI = Ip − Id and plot it by the dotted line in Fig. 3. The current difference also follows a nearly linear relationship with the voltage. At a bias of −1 V, ΔI has a value of 4.2 mA, and the relative difference ΔI/Id is 2.4%, which almost keeps constant except around the zero bias. Note that at zero bias ΔI > 0, implying that there exists a small current flow across the HJ excited by terahertz radiation even without a bias electric field. According to the current direction, it is found that the excited electrons are transported from DWNTs to the Ni foil, consistent with previous observations for optical illumination [11]. The photovoltaic responsivity of our device is estimated to be about 22 mV/W from the bias voltage corresponding to the zero-current crossing of the I–V curve with terahertz radiation on. This value would be improved by using a coupling antenna and the lock-in technique to collect data [15].
Fig. 3 I–V characteristics of the DWNT–Ni HJ. The solid and dashed lines denote the measurement results without and with terahertz illumination (Id and Ip), respectively, and the dotted line denotes their difference ΔI = Ip − Id.
To further investigate the terahertz induced photocurrents in such HJ, we tested the time-dependent response of the sample by periodically blocking and unblocking the incident beam (120 s period, 50% duty cycle) under zero and 100 mV bias voltages. The temporal photocurrent response curves are shown in Figs. 4(a) and 4(b). When the terahertz beam was turned on or off, the current underwent an initial jump and then a gradual change to a steady state. At zero bias, a positive current on the order of 10 μA was produced by terahertz irradiation, coinciding with the results of I–V measurements. The discrepancy in the exact value of this current arises from that the voltage was not strictly zero. The generation process of this photocurrent can be considered as that the electrons in CNTs absorb the terahertz photon energy and then are transported from the lower dimensional CNTs to the higher dimensional metal owing to the heterodimension effect of the HJ [11]. Using the optical parameters acquired in section 3.1, we calculate an energy absorptivity of 11% at 2.52 THz for a 200-nm-thick DWNT film on Si substrate. Mainly two mechanisms are possibly responsible for the terahertz absorption of CNTs, including photoelectric excitation and thermal coupling. The photoelectric effect involves the intrinsic excitation of semiconducting CNTs and the surface plasmon polariton (SPP) excitation on metallic CNTs. However, since terahertz radiation has relatively low photon energy in comparison with optical waves, the intrinsic excitation may not dominate the photoelectric process. The existence and propagation of SPPs on CNT film have been experimentally verified at terahertz frequencies [25]. These two kinds of photoelectric effects are all fast processes. The thermal effect is common in the photoresponse of CNTs and is a slow process with time constant on the order of several seconds [13, 26]. It is easy to see that the DWNT–Ni HJ reveals photovoltaic behavior in the terahertz range and has the potential to serve as a terahertz detector. Indeed, the SNR of the photocurrent, or in other words, the responsivity of the device, is not high, which can be attributed to the disorder of DWNT bundles in the sample. Compared to the work reported by Fedorov et al. [15], the CNTs in our device are all non-suspended and are in direct contact with either the metal foil or the glass substrate. Thus, the difference in heat transfer between the two sides of the HJ of our device is less pronounced, giving rise to the decrease in the contribution of thermal effect to the photoresponse. Consequently, our device has a relatively low responsivity. The advantage of our device is that its fabrication is easier.
Fig. 4 Temporal photocurrent response of the DWNT–Ni HJ with terahertz irradiation on and off under zero (a) and 100 mV (b) bias voltages. (c) and (d) are fits (dashed lines) of the falling and rising edges (open circles) of the curve in (b), respectively.
For the case of 100 mV bias, the current dropped by about 140 μA while terahertz radiation was turned on. Based on the I–V measurements, we can deduce that this current change would be more remarkable with increasing bias. A better SNR was observed here because of the high dark current. The temporal response curve can be expressed as
where I1 and I2 are negative for THz on and positive for THz off, and τ1 and τ2 are the time constants. We fit the first falling and rising edges of the curve in Fig. 4(b) with this function and give the results in Figs. 4(c) and 4(d). The two time constants are obtained to be 0.23 s and 8.09 s in average, corresponding to a fast and a slow response process, respectively, which agree with our above analysis on photocurrent generation mechanisms. Different from the case of zero bias, current reduction occurred after terahertz illumination, demonstrating the negative photoconductivity behavior of the HJ [27]. This effect originates from the resistance increase of the sample. The thermal heating of CNTs upon terahertz absorption could lead to the sample temperature to rise and thus the resistance increases. Besides, scattering of electrons by SPPs may also cause the conductivity to change dramatically.It is worth to note that the optimal beam spot position is located at the junction interface indicated by the green dashed line in Fig. 1(b). In fact, in our experiments, we have tested the case that the terahertz beam was incident to the large-area DWNTs on top of Ni foil and found the photocurrent was much smaller. The reason is that for the latter case, the electron transport needs to overcome the resistance of the large-area DWNT film so as to form a loop to generate a photocurrent. While for the former case, the excited electrons enter into the Ni foil immediately and do not need to pass through the DWNT film above Ni foil to form a loop. Because Ni has a lower resistance compared to the DWNT film, the photocurrent is larger in this case.
Finally, we examined the relation between the photocurrent and terahertz radiation power. The incident beam power was adjusted by two polyethylene grid polarizers and its absolute value was monitored via a calibrated pyroelectric detector. As usual, a 100 mV bias voltage was applied to the HJ. Figure 5 shows measurements of the photocurrent change between terahertz on and off at different illumination intensities. It is obvious that the current difference linearly varies with the received power, as expected for the application to terahertz detection.
Fig. 5 Terahertz induced photocurrent as a function of incident power. The filled squares with error bars correspond to the experimental data and the solid line is a linear fit.
4. Conclusion
In conclusion, we have fabricated a CNT–metal heterodimension junction formed by DWNT film and Ni contacts and characterized its photocurrent response upon terahertz radiation. THz-TDS measurements of the individual DWNT film can provide values of optical parameters for quantitative evaluation of the interaction between terahertz radiation and DWNTs. A current difference relative to the dark current in the HJ has been observed and it exhibits an almost linear dependence on the bias voltage and the illumination power. The dynamic photocurrent response of the HJ to terahertz irradiation reveals a fast transient process and a slow gradual process, suggesting the existence of the combined effects of photoelectric excitation and thermal coupling. How much each effect contributes to the photoresponse needs to be further studied. Our work has demonstrated the possibility of CNT–metal HJ for terahertz detection. However, improvement of the sensitivity, response speed and integration level of this device remains to be explored.
Acknowledgments
This work was supported by the National Science and Technology Support Program of China (Grant No. 2013BAK14B03), the National Natural Science Foundation of China (Grant No. 11174172), and the CAEP THz Science and Technology Foundation (Grant No. CAEPTHZ201211).
References and links
1. M. Tonouchi, “Cutting-edge terahertz technology,” Nat. Photonics 1, 97–105 (2007). [CrossRef]
2. W. L. Chan, J. Deibel, and D. M. Mittleman, “Imaging with terahertz radiation,” Rep. Prog. Phys. 70, 1325–1379 (2007). [CrossRef]
3. F. Sizov, “THz radiation sensors,” Opto-Electron. Rev. 18, 10–36 (2010). [CrossRef]
4. Y. Wang, Z. Zhao, Z. Chen, and K. Kang, “Calibration of a thermal detector for pulse energy measurement of terahertz radiation,” Opt. Lett. 37, 4395–4397 (2012). [CrossRef] [PubMed]
5. R. A. Hadi, H. Sherry, J. Grzyb, Y. Zhao, W. Föster, H. M. Keller, A. Cathelin, A. Kaiser, and U.R. Pfeiffer, “A 1 k-Pixel Video Camera for 0.7–1.1 Terahertz Imaging Applications in 65-nm CMOS,” IEEE J. Solid-State Circuit 47, 2999–3012 (2012). [CrossRef]
6. S. Boppel, A. Lisauskas, A. Max, V. Krozer, and H. G. Roskos, “CMOS detector arrays in a virtual 10-kilopixel camera for coherent terahertz real-time imaging,” Opt. Lett. 37, 536–538 (2012). [CrossRef] [PubMed]
7. Y. Kawano, T. Uchida, and K. Ishibashi, “Terahertz sensing with a carbon nanotube/two-dimensional electron gas hybrid transistor,” Appl. Phys. Lett. 95,083123 (2009). [CrossRef]
8. M. S. Vitiello, L. Viti, L. Romeo, D. Ercolani, G. Scalari, J. Faist, F. Beltram, L. Sorba, and A. Tredicucci, “Semiconductor nanowires for highly sensitive, room-temperature detection of terahertz quantum cascade laser emission,” Appl. Phys. Lett. 100,241101 (2012). [CrossRef]
9. L. Vicarelli, M. S. Vitiello, D. Coquillat, A. Lombardo, A. C. Ferrari, W. Knap, M. Polini, V. Pellegrini, and A. Tredicucci, “Graphene field-effect transistors as room-temperature terahertz detectors,” Nat. Mater. 11, 865–871 (2012). [CrossRef] [PubMed]
10. P. Avouris, M. Freitag, and V. Perebeinos, “Carbon-nanotube photonics and optoelectronics,” Nat. Photonics 2, 341–350 (2008). [CrossRef]
11. J.-L. Sun, J. Wei, J.-L. Zhu, D. Xu, X. Liu, H. Sun, D.-H. Wu, and N.-L. Wu, “Photoinduced currents in carbon nanotube/metal heterojunctions,” Appl. Phys. Lett. 88,131107 (2006). [CrossRef]
12. J. Wei, J.-L. Sun, J.-L. Zhu, K. Wang, Z. Wang, J. Luo, D. Wu, and A. Cao, “Carbon nanotube macrobundles for light sensing,” Small 2, 988–993 (2006). [CrossRef] [PubMed]
13. S. Lu and B. Panchapakesan, “Photoconductivity in single wall carbon nanotube sheets,” Nanotechnology 17, 1843–1850 (2006). [CrossRef]
14. D.-H. Lien, W.-K. Hsu, H.-W. Zan, N.-H. Tai, and C.-H. Tsai, “Photocurrent Amplification at Carbon Nanotube–Metal Contacts,” Adv. Mater. 18, 98–103 (2006). [CrossRef]
15. G. Fedorov, A. Kardakova, I. Gayduchenko, I. Charayev, B. M. Voronov, M. Finkel, T. M. Klapwijk, S. Morozov, M. Presniakov, I. Bobrinetskiy, R. Ibragimov, and G. Goltsman, “Photothermoelectric response in asymmetric carbon nanotube devices exposed to sub-terahertz radiation,” Appl. Phys. Lett. 103,181121 (2013). [CrossRef]
16. J. Wei, H. Zhu, Y. Li, B. Chen, Y. Jia, K. Wang, Z. Wang, W. Liu, J. Luo, M. Zheng, D. Wu, Y. Zhu, and B. Wei, “Ultrathin single-layered membranes from double-walled carbon nanotubes,” Adv. Mater. 18, 1695–1700 (2006). [CrossRef]
17. J. Wei, Y. Jia, Q. Shu, Z. Gu, K. Wang, D. Zhuang, G. Zhang, Z. Wang, J. Luo, A. Cao, and D. Wu, “Double-walled carbon nanotube solar cells,” Nano Lett. 7, 2317–2321 (2007). [CrossRef] [PubMed]
18. W. Gao, Y. Wang, and Z. Zhao, “Development of the data acquisition and control system for a portable THz time-domain spectrometer,” Proc. SPIE 8909, 89090O (2013). [CrossRef]
19. S. Kumar, N. Kamaraju, A. Moravsky, R. O. Loutfy, M. Tondusson, E. Freysz, and A. K. Sood, “Terahertz time domain spectroscopy to detect low-frequency vibrations of double-walled carbon nanotubes,” Eur. J. Inorg. Chem. 2010, 4363–4366 (2010). [CrossRef]
20. T. D. Dorney, R. G. Baraniuk, and D. M. Mittleman, “Material parameter estimation with terahertz time-domain spectroscopy,” J. Opt. Soc. Am. A 18, 1562–1571 (2001). [CrossRef]
21. I. Pupeza, R. Wilk, and M. Koch, “Highly accurate optical material parameter determination with THz time-domain spectroscopy,” Opt. Express 15, 4335–4350 (2007). [CrossRef] [PubMed]
22. S. Maine, C. Koechlin, S. Rennesson, J. Jaeck, S. Salort, B. Chassagne, F. Pardo, J.-L. Pelouard, and R. Haïar, “Complex optical index of single wall carbon nanotube films from the near-infrared to the terahertz spectral range,” Appl. Opt. 51, 3031–3035 (2012). [CrossRef] [PubMed]
23. I. Maeng, C. Kang, S. J. Oh, J.-H. Son, K. H. An, and Y. H. Lee, “Terahertz electrical and optical characteristics of double-walled carbon nanotubes and their comparison with single-walled carbon nanotubes,” Appl. Phys. Lett. 90,051914 (2007). [CrossRef]
24. C.-S. Yang, C.-H. Chang, M.-H. Lin, P. Yu, O. Wada, and C.-L. Pan, “THz conductivities of indium-tin-oxide nanowhiskers as a graded-refractive-index structure,” Opt. Express 20, A441–A451 (2012). [CrossRef] [PubMed]
25. T. D. Nguyen, S. Liu, M. D. Lima, S. Fang, R. H. Baughman, A. Nahata, and Z. V. Vardeny, “Terahertz surface plasmon polaritons on freestanding multi-walled carbon nanotube aerogel sheets,” Opt. Mater. Express 2, 782–788 (2012). [CrossRef]
26. I. A. Levitsky and W. B. Euler, “Photoconductivity of single-wall carbon nanotubes under continuous-wave near-infrared illumination,” Appl. Phys. Lett. 83, 1857–1859 (2003). [CrossRef]
27. J.-L. Zhu, G. Zhang, J. Wei, and J.-L. Sun, “Negative and positive photoconductivity modulated by light wavelengths in carbon nanotube film,” Appl. Phys. Lett. 101,123117 (2012). [CrossRef]
