LETTER pubs.acs.org/NanoLett
A Reel-Wound Carbon Nanotube Polarizer for Terahertz Frequencies Jisoo Kyoung,† Eui Yun Jang,‡ Marcio D. Lima,§ Hyeong-Ryeol Park,† Raquel Ovalle Robles,§ Xavier Lepro,§ Yong Hyup Kim,‡ Ray H. Baughman,§ and Dai-Sik Kim*,† †
Center for Subwavelength Optics and Department of Physics and Astronomy and ‡School of Mechanical and Aerospace Engineering and the Institute of Advanced Aerospace Technology, Seoul National University, Seoul 151-747, Korea § Alan G. MacDiarmid NanoTech Institute, University of Texas at Dallas, Richardson, Texas 75083, United States
bS Supporting Information ABSTRACT: Utilizing highly oriented multiwalled carbon nanotube aerogel sheets, we fabricated micrometer-thick freestanding carbon nanotube (CNT) polarizers. Simple winding of nanotube sheets on a U-shaped polyethylene reel enabled rapid and reliable polarizer fabrication, bypassing lithography or chemical etching processes. With the remarkable extinction ratio reaching ∼37 dB in the broad spectral range from 0.1 to 2.0 THz, combined with the extraordinary gravimetric mechanical strength of CNTs, and the dispersionless character of freestanding sheets, the commercialization prospects for our CNT terahertz polarizers appear attractive. KEYWORDS: Carbon nanotube, THz polarizer, nanotube dischroism, aligned nanotube sheets, wound nanotube sheets, nanotube processing
T
he high electrical and optical anisotropy that results from their large aspect ratios and small diameters makes carbon nanotubes (CNTs) a nearly ideal quasi-one-dimensional system.1 The polarization dependence of the optical absorption of single-walled carbon nanotubes (SWCNT) has been reported, which enables determination of the nematic order parameters of samples.2,3 Polarization-dependent Raman scattering measurements of substantially aligned SWCNTs show dramatic reduction of Raman intensities for perpendicular excitation compared to parallel excitation, while the relative intensities of radial and tangential modes remain nearly constant.4 The generation and detection of radial breathing mode coherent phonon oscillations in a bulk film of highly aligned SWCNTs are strongly affected by the polarization direction of pump and probe beam with respect to the nanotube orientation direction.5 Because of quasi-onedimensional properties, carbon nanotubes can be regarded as optical antennas in terms of radiation patterns, polarization dependence, and antenna length effect.6,7 On the basis of this optical anisotropy, various recent effort have been made to use CNT sheets as linear polarizers for the visible frequency range,8 as well as for longer wavelengths.9,10 For example, in the important publication by Lei Ren et al.10 a highly aligned SWCNT film on a sapphire substrate shows a near vanishing absorption when terahertz (THz) polarization is perpendicular to the nanotubes axis and the reported extinction ratio (T///T^) is about 101 (=10 dB) at 1.8 THz for a 2 μm thick film. These authors predict that the extinction ratio can be easily enhanced in the future by just increasing film thickness, keeping negligible absorption for the perpendicular polarization due to the total absence of THz response. However, a CNTbased polarizer with high extinction ratios, say over 25 dB in a r 2011 American Chemical Society
wide band to compete with wire-grids,11,12 has not been reported in the millimeter and submillimeter wavelength regime, and the relatively achieved low extinction ratio has blocked the commercialization of CNT sheets as a polarizer even though CNTs have many advantages. Herein, we explore the potential of thick MWCNT (multiwalled carbon nanotube) sheets as an easily fabricatable, high-extinction coefficient, freestanding polarizer working in the technologically important THz frequency regime. The previously reported extinction ratios are low either because of small sheet thicknesses and/or nanotube misalignment. Simultaneous control of orientation and increase of CNT thickness is difficult using the previously described fabrication method of patterned CNT growth and transfer.10 In the present work, the thickness limit was easily surmounted by virtue of a newly developed solid-state draw method for highly aligned CNT sheet formation, which requires neither e-beam lithography process nor H2O etching.1315 A free-standing MWCNT sheet up to 9 μm thick was successfully manufactured, which acted as an excellent THz linear polarizer, providing giant extinction ratios of ∼37 dB in a broad spectral range (0.1 to ∼2.0 THz). The thicknesses of the MWCNT sheets can be freely adjusted and production of much thicker samples can be readily accomplished. We compared our new type of polarizer with a commercial wire-grid polarizer and demonstrate that the CNT polarizer provides better performance than the commercial product. Figure 1a schematically illustrates our method for manufacturing CNT polarizers. A MWCNT forest was synthesized by Received: June 30, 2011 Revised: August 17, 2011 Published: August 23, 2011 4227
dx.doi.org/10.1021/nl202214y | Nano Lett. 2011, 11, 4227–4231
Nano Letters
Figure 1. (a) Schematic diagram for the fabrication of a freestanding CNT polarizer using U-shaped PE (polyethylene) frame and highly aligned CNT sheet, drawn from a sidewall of a MWCNT forest. When the PE reel is rotated once, two layers of the CNT sheets are stacked (see Supporting Information movie 1 for details). (b) Photographic and (c) SEM images of our 75-layer CNT sheet polarizer.
catalytic chemical vapor deposition. The CNT sheet was drawn from a sidewall of the forest and can be longer than 3 m. The draw rate of the sheet reached 10 m/minute. Since each centimeter of forest length produced about a 510 m sheet length, very little forest area was required to make a polarizer. Additional details about the preparation of drawable MWCNT forests, sheet draw, and the structure and properties of CNT sheets are described elsewhere.13,16 We used a U-shaped polyethylene (PE) support to make freestanding sheet stacks that are polarizers. A motor was symmetrically attached at the base of this PE reel and rotated at 15 rpm to simultaneously draw the sheet from the forest and wrap it onto the reel to form the polarizer. Supporting Information movie 1 shows this convenient method for polarizer fabrication. The thickness of the sheet stack was determined by the number of rotations. When the motor rotated once, two layers of the CNT sheets were accumulated. Even though the PE reel is flexible, it is sufficiently mechanically rigid to support the CNT sheets during and after the fabrication process. Figure 1b and c shows photographic and scanning electron microscope (SEM) images, respectively, of a polarizer comprising a 75-layer freestanding CNT stack. THz time domain spectroscopy (THz-TDS) (Figure 2a) was used to evaluate polarizer performance.17 The polarization direction of the incident THz wave was fixed to the x-axis (parallel to the surface of the optical table) and our 75-layer CNT sheet was rotated from θ = 0 to 360 using intervals of 15 in a clockwise direction. The transmitted THz wave having both x- and ycomponents of electric fields was measured using a conventional electro-optic detection method.18 Since our detector crystal (ZnTe) was oriented in the Æ110æ direction, the electric fields of x-component (Ex) and y-component (Ey) were independently measured by rotating the crystal 90.19,20 Figure 2b represents the angle dependent transmission of Ex in the time domain. The gray line in Figure 2b is for the reference signal without any CNT sheet. Maximum transmission appeared when the CNT orientation axis was perpendicular to the THz electric field (0 and 180 according to the scheme depicted in Figure 2a) and nearly zero (in the noise limit) for 90 and 270. Note that although the angles and therefore the amplitudes were changed, the shape and the peak position were unaltered, meaning our CNT sheet causes no dispersion in the spectral region of interest.21 For the case of Ey, the maximum transmission occurred at four different angles
LETTER
Figure 2. (a) Experimental scheme for the polarization state measurements. The incident THz beam is polarized along the x-axis and the CNT sheet polarizer was rotated from θ = 0 to 360 in 15 interval. The 75-layer CNT sheet is used as a polarizer and 0 (90) corresponds to the nanotube orientation direction being perpendicular (parallel) to the polarization of the incident THz waves. (b) Time traces for x-component (Ex) electric fields for the different polarizer angles (0, 45, and 90). The reference signal (gray line) is measured without any CNT polarizer. More than 2 orders of magnitude extinction are observed between the perpendicular (0) and parallel (90) orientation.
(45, 135, 225, and 315), while the minimum transmission occurred at 0 , 90, 180, and 270o (not shown). To extract the spectral features of our CNT sheet, we applied fast Fourier transform of the time domain signals. Figure 3a shows the phase information of the reference signal (without CNT sheet, black line) and the transmitted Ex field at 0 (open black circle). Clearly, there was no measurable phase difference (Δϕ) between the two data implying our CNT sheet was indeed freestanding. Figure 3b shows the measured transmission, defined by T(θ) = Ex2 + Ey2, as a function of frequency (0.12.0 THz) for different polarizer angles from θ = 0 to 90. The average transmission at 0 is over 50% while 4 orders of magnitude annihilation was observed at 90. Since transmission through an ideal polarizer follows Malus’s law (T(θ) = T(0)cos2 θ), it is worthwhile to verify how well our measured data, T(θ), fits a cos2 θ dependence.22 Figure 3c shows the measured transmitted intensity T(θ) as a function of polarizer angle θ at f = 1.0 THz (open black circle) and the corresponding cosine-square leastsquares fit to cos2 θ (black line). As clearly seen, the experimental data and Malus’s law agree, which is one measure of the high quality of our CNT polarizer. The most important characteristic of a polarizer is the degree of polarization (DOP), which is defined by22 T max T min DOP ¼ ð1Þ T max þ T min where Tmax and Tmin are the maximum and minimum intensity of the light transmitted through a linear polarizer, respectively, when it is rotated through the complete range of 360. The DOP equals one for an ideal polarizer, since it creates fully polarized light. Figure 3d shows the degree of polarization of our 75-layer CNT sheet polarizer as a function of frequency. Over most of the spectral range, the DOP is very close to one (0.999), meaning our new type of polarizer works nearly ideally. This high performance is intriguing because of the existence of non-negligible deviations from the perfect alignment on the 100 nm scale. The likely explanation is these local deviations are reversed on larger dimensional scales, so on the ∼50 μm scale (about 10% of the nanotube length) the nanotube bundles appear largely straight and parallel (Figure 1c). 4228
dx.doi.org/10.1021/nl202214y |Nano Lett. 2011, 11, 4227–4231
Nano Letters
LETTER
Figure 3. (a) Phase information for transmission signals through the reference aperture (black line) and the 75-layer CNT polarizer at 0 (open black circle). No measurable phase shift between the two measurements is observed. (b) Transmission spectra for various polarizer angles from θ = 0 to 90. The transmission signals are normalized to the reference signal (without a CNT sheet polarizer). Averaged between 0.1 to 2.0 THz, over 50% transmission is shown for 0, while only 104 transmission is observed at 90. (c) The transmission spectrum as a function of polarizer angle θ at 1 THz and the corresponding fitting curve to cos2 θ (from Malus’s law). (d) The averaged degree of polarization (DOP) of the 75-layer CNT sheet in our spectral region of interest (0.1 to 2.0 THz) is nearly one.
Figure 4. (a) The extinction ratio and (b) degree of polarization (DOP) are measured as a function of the number of CNT sheet layers, each of which is averaged over the 0.12 THz range. (The solid lines are guides for the eye.).
Two important physical quantities for a polarizer, the extinction ratio (10 log10(Tmax/Tmin)) and DOP, were measured as a function of the number of sheet layers (1075) in the CNT polarizer. Each of these quantities was averaged over the 0.12 THz range. The resulting extinction ratios are proportional to the number of layers, as shown in Figure 4a. For the thickest sheet stack, which contained 75 layers, a 37 dB extinction ratio was observed, which is nearly equal to the highest values obtainable for a wire grid or a liquid crystal polarizer.12,23 Figure 4b plots the averaged DOP as a function of number of layers. The DOP continuously increased as the sheet stack became thicker and was close to one when the number of sheet layers increased to 50. Wire-grids have governed the polarizer market in the millimeter wavelength regime. We compare the performances of our CNT polarizer with those of a commercial freestanding wire-grid
polarizer to show that our CNT polarizers can provide commercially attractive performance. A commercial polarizer made of 10 μm tungsten wire with 30 μm spacing was employed to obtain the best performance in our spectral range. Figure 5a shows the ratio of the transmission minimum and maximum (Tmin/Tmax) for the wiregrid (blue line) and 75-layer CNT sheet polarizer (red line), respectively. The commercial polarizer showed around 103 (extinction ratio, 30 dB) in the low frequency regime, while only 102 (extinction ratio, 20 dB) was achieved above 1 THz. In stark contrast, an order of magnitude-enhanced extinction ratios were obtained over the wide spectral band using our CNT polarizer. Moreover, the extinction ratio shows little spectral dependence. This means that the performance of the CNT polarizer is very well preserved up to the high-THz frequency regime, unlike for the tungsten wire-grids. The crumpled feature observed in the CNT 4229
dx.doi.org/10.1021/nl202214y |Nano Lett. 2011, 11, 4227–4231
Nano Letters
LETTER
Figure 5. Comparison between the performances of a commercial freestanding wire-grid polarizer and the 75-layer freestanding CNT sheet polarizer. (a) The ratios of the minimum and the maximum transmission (Tmin/Tmax) and (b) the degree of polarization.
polarizer’s ratio spectrum demonstrates that the minimum transmission (Tmin) went completely to noise level in our THz-TDS system; the performance of our CNT polarizer can be even better. Comparison of the DOP spectra also demonstrates that the CNT polarizer has attractive performance features compared with commercial polarizers (Figure 5b). The DOP spectrum of the CNT polarizer (Figure 3d) is plotted again in Figure 5b in order to provide convenient comparison. Our polarizer exhibited uniform DOP spectrum, being very close to one in overall frequencies. On the other hand, the DOPs of the commercial wire-grid polarizer were mostly lower than for our polarizer and gradually decreased as frequency increased. For the wire-grids, the intrinsic loss (= 1 transmission at zero degree = absorption + reflection) is almost negligible (∼2%, not shown), while an average 50% loss occurred for the 75-layer CNT sheet polarizer. Large intrinsic loss might cause problems in using some “poor” sources (low power and low SNR (signal-tonoise ratio)) for experiments. However, even in that case, we have a solution. The number of layers of the film can be continuously tuned and when we make 25 layers, the transmission at 0 is now recovered to 85% (see Supporting Information Figure 1a for more information). Remarkably, the average extinction ratio remains around 28 dB (Supporting Information Figure S1a), still being able to compete with commercial wiregrids, considering the rapid, reliable fabrication process for our CNT polarizers. Also, transmission as a function of angle follows Malus’s law almost perfectly (Supporting Information Figure S1b) showing less dense film works as ideally as higher density film. A freestanding wire-grid polarizer is usually made by mechanically winding thin metallic strings (commonly, tungsten wire) on a rigid frame under high tension.11,24 When making a freestanding wire-grid polarizer, it is quite important to keep the winding tension, the diameter of each wire, and the period constant in order to obtain the highest quality. However, these constraint conditions are very difficult to realize. Furthermore, freestanding samples are often fragile because the strength of a thin metal wire is too weak mechanically and tungsten wire is easily oxidized. More seriously, since the conductivity of each wire is finite and the spacing between wires must be smaller than the wavelength of the incident beam for large extinction ratios, freestanding wire-grid type polarizers are not suitable for the high THz frequency regime, as seen in Figure 5. An alternative approach for overcoming the weaknesses of freestanding wire-grid polarizers is to fabricate these polarizers on a substrate.12,25,26 Such types
of wire-grid polarizers are relatively robust and work quite well in the high THz frequency range because making fine pitch becomes possible due to recently developed nanopatterning techniques such as lithography or nanoimprinting.27 Nevertheless, substrates give intrinsic loss and dispersion of the incident waves and there exists an unwanted Fabry-Perot effect. Moreover, nanopatterning techniques can be expensive and time-consuming. Our novel approach, winding MWCNT sheet on a U-shaped polyethylene frame, allows fabrication of freestanding THz polarizers that have many advantages such as: (1) Rapid, reliable fabrication. Our method demands neither e-beam lithography nor any chemical etching processes which make it quite difficult and complex to prepare ordinary THz polarizers. (2) It operates in a broad spectral region (from 0.12 THz) with little dispersion. (3) The extinction ratio of ∼37 dB, being comparable to the highest values obtained by wire-grids, is sufficiently high for most applications. (4) Our freestanding polarizers are much more mechanically robust than freestanding wire grid polarizers and free from oxidization at room temperature, ensuring semipermanent use (as long as they are not touched). (5) We can freely adjust the size of the polarizers from less than a coin to bigger than a hand (see Supporting Information Figure S2). Because of these strengths, the commercialization prospects for these nanotube polaizers are attractive.
’ ASSOCIATED CONTENT
bS
Supporting Information. Additional information, figures, and movie. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Phone: +82-2-880-8174. Fax: +82-2-884-3002. E-mail: dsk@ phya.snu.ac.kr.
’ ACKNOWLEDGMENT This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (SRC, No. R11-2008-095-01000-0) (Nos. 20100029648, 2011-0000318, 2011-0001293, and 2011-0018905), KICOS (GRL, K20815000003), Hi Seoul Science/Humanities Fellowship from Seoul Scholarship Foundation, Air Force Office 4230
dx.doi.org/10.1021/nl202214y |Nano Lett. 2011, 11, 4227–4231
Nano Letters
LETTER
of Scientific Research Grant FA9550-09-1-0537, and Robert A. Welch Foundation Grant AT-0029.
’ REFERENCES (1) deHeer, W. A.; Bacsa, W. S.; Ch^atelain, A.; Gerfin, T.; HumphreyBaker, R.; Forro, L.; Ugarte, D. Science 1995, 268, 845–847. (2) Islam, M. F.; Milkie, D. E.; Kane, C. L.; Yodh, A. G.; Kikkawa, J. M. Phys. Rev. Lett. 2004, 93, 037404. (3) Murakami, Y.; Einarsson, E.; Edamura, T.; Maruyama, S. Phys. Rev. Lett. 2005, 94, 087402. (4) Gommans, H. H.; Alldredge, J. W.; Tashiro, H.; Park, J.; Magnuson, J.; Rinzler, A. G. J. Appl. Phys. 2000, 88, 2509–2514. (5) Booshehri, L. G.; Pint, C. L.; Sanders, G. D.; Ren, L.; Sun, C.; Haroz, E. H.; Kim, J. H.; Yee, K. J.; Lim, Y. S.; Hauge, R. H.; Stanton, C. J.; Kono, J. Phys. Rev. B 2011, 83, 195411. (6) Wang, Y.; Kempa, K.; Kimball, B.; Carlson, J. B.; Benham, G.; Li, W. Z.; Kempa, T.; Rybczynski, J.; Herczynski, A.; Ren, Z. F. Appl. Phys. Lett. 2004, 85, 2607–2609. (7) Kempa, K.; Rybczynski, J.; Huang, Z.; Gregorczyk, K.; Vidan, A.; Kimball, B.; Carlson, J.; Benham, G.; Wang, Y.; Herczynski, A.; Ren, Z. F. Adv. Mater. 2007, 19, 421–426. (8) Xu, G. B.; Zheng, M.; Mclean, R. S. Carbon nanotube optical polarizer. U.S. Patent 2006/0121185 A1, June 8, 2006. (9) Jeon, T. I.; Kim, K. J.; Kang, C.; Maeng, I. H.; Son, J. H.; An, K. H.; Lee, J. Y.; Lee, Y. H. J. Appl. Phys. 2004, 95, 5736–5740. (10) Ren, L.; Pint, C. L.; Booshehri, L. G.; Rice, W. D.; Wang, X.; Hilton, D. J.; Takeya, K.; Kawayama, I.; Tonouchi, M.; Hauge, R. H.; Kono, J. Nano Lett. 2009, 9, 2610–2613. (11) Costley, A. E.; Hursey, K. H.; Neill, G. F.; Ward, J. M. J. Opt. Soc. Am. 1977, 67, 979–981. (12) Yamada, I.; Takano, K.; Hangyo, M.; Saito, M.; Watanabe, W. Opt. Lett. 2009, 34, 274–276. (13) Zhang, M.; Fang, S.; Zakhidov, A. A.; Lee, S. B.; Aliev, A. E.; Williams, C. D.; Atkinson, K. R.; Baughman, R. H. Science 2005, 309, 1215–1219. (14) Liu, K.; Sun, Y.; Chen, L.; Feng, C.; Feng, X.; Jiang, K.; Zhao, Y.; Fan, S. Nano Lett. 2008, 8, 700–705. (15) Feng, C.; Liu, K.; Wu, J.-S.; Liu, L.; Cheng, J.-S.; Zhang, Y.; Sun, Y.; Li, Q.; Fan, S.; Jiang, K. Adv. Funct. Mater. 2010, 20, 885–891. (16) Kuznetsov, A. A.; Fonseca, A. F.; Baughman, R. H.; Zakhidov, A. A. ACS Nano 2011, 5, 985–993. (17) Grischkowsky, D.; Keiding, S.; Vanexter, M.; Fattinger, C. J. Opt. Soc. Am. B 1990, 7, 2006–2015. (18) Kyoung, J. S.; Seo, M. A.; Park, H. R.; Ahn, K. J.; Kim, D. S. Opt. Commun. 2010, 283, 4907–4910. (19) Gallot, G.; Grischkowsky, D. J. Opt. Soc. Am. B 1999, 16, 1204–1212. (20) Seo, M. A.; Park, H. R.; Koo, S. M.; Park, D. J.; Kang, J. H.; Suwal, O. K.; Choi, S. S.; Planken, P. C. M.; Park, G. S.; Park, N. K.; Park, Q. H.; Kim, D. S. Nat. Photonics 2009, 3, 152–156. (21) George, P. A.; Strait, J.; Dawlaty, J.; Shivaraman, S.; Chandrashekhar, M.; Rana, F.; Spencer, M. G. Nano Lett. 2008, 8, 4248–4251. (22) Fowles, G. R. Introduction to Modern Optics, 2nd ed.; Dover: New York, 1975. (23) Hsieh, C.-F.; Lai, Y.-C.; Pan, R.-P.; Pan, C.-L. Opt. Lett. 2008, 33, 1174–1176. (24) Ade, P. A. R.; Costley, A. E.; Cunningham, C. T.; Mok, C. L.; Neill, G. F.; Parker, T. J. Infrared Phys. 1979, 19, 599–601. (25) Ma, Y.; Khalid, A.; Drysdale, T. D.; Cumming, D. R. S. Opt. Lett. 2009, 34, 1555–1557. (26) Kondo, T.; Nagashima, T.; Hangyo, M. Jpn. J. Appl. Phys. 2003, 42, L373. (27) Seh-Won, A.; et al. Nanotechnology 2005, 16, 1874.
4231
dx.doi.org/10.1021/nl202214y |Nano Lett. 2011, 11, 4227–4231