4408

Letter

Vol. 40, No. 19 / October 1 2015 / Optics Letters

Electrically controllable extraordinary optical transmission in gold gratings on vanadium dioxide JUNHO JEONG,1,* ARASH JOUSHAGHANI,1 SUZANNE PARADIS,2 DAVID ALAIN,2

AND

JOYCE K. S. POON1

1

Department of Electrical and Computer Engineering, University of Toronto, 10 King’s College Road, Toronto, Ontario M5S 3G4, Canada Defence Research and Development Canada–Valcartier, 2459 Pie-XI Boulevard North, Quebec, Quebec G3J 1X5, Canada *Corresponding author: [email protected]

2

Received 10 July 2015; revised 28 August 2015; accepted 31 August 2015; posted 1 September 2015 (Doc. ID 245745); published 18 September 2015

Highly tunable optical transmission through one-dimensional gold gratings patterned on top of a film of the phase transition material, vanadium dioxide (VO2 ), is demonstrated. Dense electrical integration is enabled by grating features that also function as electrical contacts to the VO2 . Extraordinary optical transmission is observed in the VO2 insulator phase, and the optical transmission is extinguished by up to about 6 dB in a 170 nm thick VO2 film. Measurements of gratings with varying duty cycles demonstrate the dependence of the optical transmission and tuning on the device geometry. © 2015 Optical Society of America OCIS codes: (050.2770) Gratings; (160.2100) Electro-optical materials; (160.3130) Integrated optics materials. http://dx.doi.org/10.1364/OL.40.004408

Extraordinary optical transmission (EOT) is a phenomenon in which the optical transmission through subwavelength apertures in metallic thin films is enhanced [1–4]. The optical transmission can be several orders of magnitude greater than expected from Bethe diffraction theory, and the normalized-to-area transmission can exceed unity [1,3–5]. EOT occurs due to surface plasmons excited at the metal–insulator interfaces and Fabry–Perot resonances, which efficiently couple an incident wave of the appropriate wavelength, polarization, and angle of incidence to the output side of the apertures [2,6–8]. EOT has been observed in both two-dimensional metallic hole arrays and onedimensional gratings or slit structures [1,9–11]. A typical passive geometry for EOT consists of a thin film of metal, usually gold (Au) or silver (Ag), with a thickness greater than the skin depth, deposited on an insulator (e.g., SiO2 ) or suspended in air. Active control over EOT has potential applications in reconfigurable high-resolution imaging, tunable sensors, spatial light modulators, and highly efficient switching elements. To implement a tunable EOT device, the apertures can be formed in or on a tunable optical material. Tunable EOT has been demonstrated in gallium arsenide in the mid-infrared wavelength range [12], as well as in vanadium dioxide (VO2 ) [13]. VO2 is an especially promising material, as it is capable of a reversible phase transition between an insulating and a metallic state that can be initiated 0146-9592/15/194408-04$15/0$15.00 © 2015 Optical Society of America

thermally, electrically, or optically [14–17]. The insulator–metal phase transition is accompanied by a large change in the real and imaginary parts of the refractive index, which can be used to realize ultra-thin yet highly tunable optical surfaces [18–21]. For example, in our samples, at a wavelength of 1550 nm, the refractive index changes from about 3
0.4i in the insulating phase to about 2
3i in the metallic phase [22–24]. In this Letter, we demonstrate the modulation of EOT near wavelengths of 1550 nm by means of metallic surface gratings patterned atop a thin film of VO2 . In comparison to previous reports of EOT or optical surface modulation in VO2 , which have required heating of the entire substrate [13,18–21], the grating features here also functioned to inject electrical currents to initiate the phase transition. Consequently, the phase transition was localized to the grating region. EOT was demonstrated and directly imaged for the p-polarized light when the VO2 was in the insulator phase, and the transition into the metal phase led to strong optical absorption. Figures 1(a) and 1(b) show the top and cross-section view schematics of the devices under study. A 170 nm thick VO2 film was deposited using radio-frequency magnetron sputtering on a 3 μm thick thermally grown silicon dioxide (SiO2 ) on a silicon substrate at a temperature of 500°C. Material properties of the VO2 samples can be found in [15,16,22,24]. Our prior work has shown that the phase transition is initialized electronically by carrier injection followed by Joule heating [15,16,24]. Au gratings with a thickness, t, of 260 nm were patterned atop the VO2 using electron-beam lithography, thermal evaporation, and lift-off. Figure 1(c) shows scanning electron micrographs (SEMs) of a fabricated device at several magnifications. The wide Au regions on either ends of the grating visible in the bottommost SEM in Fig. 1(c) are the electrical contact pads for the probes. The grating is at the center of the image and consists of thin metallic strips that are alternately connected to the contact pads on the left and right sides. The devices were designed such that an applied current would initiate the phase transition in the VO2 film under the gratings via carrier injection and heating [15,16,24,25]. The gap between the grating lines has a width of w, and the grating period is denoted as P. Using the finite-difference time domain method, the zerothorder transmission spectra at normal incidence from the top

Letter

Vol. 40, No. 19 / October 1 2015 / Optics Letters

4409

Fig. 1. (a) Top and (b) cross-section view schematics of a VO2 film with Au surface gratings. (c) Scanning electron micrographs of a device at several magnifications.

side of the gratings are computed using the wavelengthdependent refractive index values measured by ellipsometry and given in [24]. The peak wavelength is determined by the period of the gratings, and the duty cycle changes the magnitude of the transmission and extinction ratio (ER). Simulations with perfectly matched layer boundary conditions did not exhibit EOT. w  350 nm and P  1 μm were found to maximize the ER for the p- (or y-) polarized light at normal incidence, assuming a uniform refractive index change in the VO2 film. Figure 2(a) shows the simulated transmission spectra in the insulating and metallic phases of VO2 for s- (or x-) polarized and p-polarized light for Au gratings with t  260 nm, w  350 nm, and P  1 μm and a 170 nm thick VO2 film. The computed spectra for the VO2 film in the absence of gratings are nearly flat despite the wavelength-dependent optical constants, since the VO2 film is thin. A thicker film would attenuate the transmission but increase the ER. The p-polarized transmission through the grating exhibits EOT when the VO2 is in the insulator phase, and exhibits an enhanced absorption, also called “extraordinary optical absorption” [26], when the VO2 is in the metallic phase. Note that the plot shows the absolute transmission, which is not normalized to the area of the apertures in the grating. Figure 2(b) shows the ER, which is the ratio between the transmissions when the VO2 is in the insulator and metallic phases. The ER of s-polarized light is higher than the p-polarized light at wavelengths longer than about 1510 nm, but the s-polarization transmission is about 11 to 14 dB lower than the p-polarization transmission when the VO2 is in the insulating phase. Due to EOT and absorption, the ER for p-polarized light is about 2 to 3 dB higher than an unpatterned VO2 film. Figures 2(c) and 2(d) show the normalized cross-section optical intensity profile when the VO2 film is in the insulating (left) and metallic (right) phase for p- and s-polarized incident light, respectively. In this design, EOT occurs only for the p-polarization, most likely due to the excitation of the surface plasmons in the Au-VO2 interface. As evidenced in Fig. 2(c), the optical intensity is concentrated under the grating slits with significant overlap with the VO2 . This concentration of electromagnetic energy can be explained by a weak Fabry–Perot resonance in the out-of-plane direction, and it leads to enhanced absorption when the VO2 is in the metallic phase [26]. However, for

Fig. 2. (a) Computed transmission spectra for the VO2 film by itself in the insulator (i) and metallic (m) phases, and the p- and s-polarized light through the Au gratings when the VO2 is in the insulator (i) and metallic (m) phases. The grating parameters are t  260 nm, w  350 nm, and P  1 μm, and the VO2 thickness is 170 nm. (b) The simulated ER spectra. (c) The normalized cross-section optical intensity profile when the VO2 is in the insulator (i) and metallic (m) phases for p-polarized light and (d) s-polarized light.

s-polarization, as shown in Fig. 2(d), most of the electromagnetic energy is reflected from the surface grating for both phases, which results in the low transmission evident in Fig. 2(a). The fabricated devices were measured using the setup illustrated in Fig. 3(a). Light from a tunable laser was launched through a long working distance objective from the topside of the gratings, and the transmitted light was collected with a second long working distance objective from the bottom. The applied currents were supplied by a precision sourcemeter with an 1 kΩ resistor connected in series. This resistor protected the device from being over-driven when the VO2 was in the low resistivity state, yet allowed sufficient Joule heating

Fig. 3. (a) Schematic of the experimental setup. (b) The IR image at 0 mA, with the grating region indicated by white dashed lines. The grating region is brighter than the surrounding region, indicating EOT. (c) Differential IR images using the 0 mA image as a reference at 0, 0.1, 0.25, and 1 mA. The circled numbers correspond to the bias points in Fig. 4(a).

4410

Vol. 40, No. 19 / October 1 2015 / Optics Letters

Letter

Fig. 4. (a) The voltage versus current and transmission versus current curves of the device measured with p-polarized light at wavelength 1550 nm. The optical transmission has been normalized to the transmission without the gratings. (b) Measured ER spectra for p-polarized and (c) s-polarized normal incident light for applied currents of 0.3, 1, 2, and 6 mA.

Fig. 5. (a) Simulated ER spectra for p-polarized and (b) s-polarized light at normal incidence for several gap widths in the Au gratings (w): w1  250 nm, w2  300 nm, and w3  350 nm. The grating thickness is t  260 nm, and the period is P  1 μm. The VO2 thickness is 170 nm. (c) The experimentally measured ER spectra for p-polarized and (d) s-polarized light at normal incidence for w1, w2 , and w3 for t  260 nm and P  1 μm.

to complete phase transition [15]. The incident polarization was set by a broadband squeezed fiber controller. To determine the incident polarization, we defocused the incident laser light onto the sample and observed the brightness in the grating region of the device with an InGaAs infrared (IR) camera. Since the p-polarization exhibited EOT, the grating region appeared brighter than the surrounding VO2 region under p-polarized illumination, as shown in Fig. 3(b). The s-polarized illumination did not exhibit this effect, so we were able to set the input polarization state by observing the contrast in the transmission when the VO2 was in the insulating phase. Next, we swept the applied current from 0 to 2.5 mA at increments of 5 μA, and we took IR images to locate the area in the grating region that underwent the insulator–metal transition. The images showed that the VO2 film under the grating did not undergo the phase transition uniformly. Due to the polycrystallinity of the sputtered VO2 film and the nonuniformity of the current distribution, the phase transition was initialized in a localized region of VO2 . Figure 3(c) shows the differential IR transmission images at several applied currents relative to a current of 0 mA. The dotted lines mark the location of the gratings, and the darkness indicates the degree of IR absorption. The images show that the VO2 insulator–metal transition was initially localized and expanded with increasing current, corroborating with the formation of thermal filaments in VO2 [27–30]. The fringes are due to diffraction from the iris spatial filters. Figure 4(a) shows the voltage versus current (VI) and transmission versus current for p-polarized incident light focused to a spot diameter of 3 μm at a wavelength of 1550 nm incident on the dark spot in Fig. 3(c). The transmission is normalized to the transmission of the film without the gratings. The incident power onto the chip was 13.1 μW, and we observed no change in the VI and spectral characteristics (in Figs. 4 and 5) when the power was varied from 13.1 to 655 μW. The circled numbers in Fig. 4(a) correspond to the bias conditions for the images in Fig. 3(c). The initial transmission through the gratings was about 20% higher than the transmission through an

unpatterned VO2 film, and the EOT effect became diminished as the applied current reached 2.5 mA. From the VI curve, the phase transition of the VO2 film can be seen to occur in two steps—the first at 0.10 mA [point (2)] and the second at 0.25 mA [point (3)]. This current-induced two-step phase transition in VO2 has previously been observed (e.g., [31,32]); the first step is suggestive of a carrier-initiated phase transition, followed by a Joule heating-assisted phase transition in the second step, which exhibits greater hysteresis. The VI curve in Fig. 4(a) shows abrupt changes in the optical transmission corresponding to the two steps in the phase transition as well as at 1.00 mA [point (4)]. During the first step of the VO2 phase transition (at currents less than 0.1 mA), the optical transmission only changed by about 10%. Most of the increase in optical absorption occurred after the full two-step phase transition, when heating would have led to the complete phase transition of the VO2 , and the expansion of the thermal filament and injection of free carriers would further increase the absorption. The drop in the transmission at point (4) corresponds to an abrupt expansion of the metallic region, as shown in Fig. 3(c). The slope of the VI curve remains roughly constant at this point, because the current would have flowed through the least resistive path, which would already be established near the center of the metallic region [30]. The optical hysteresis was more significant than the electrical hysteresis because the effective electrical width of the filament in our material samples saturates at about 2 μm [24,33], even as Joule heating slightly expands the optically absorptive region. The zeroth-order transmission of p- and s-polarized normal incident light was measured as the laser wavelength swept between 1480 and 1580 nm and the current was varied between 0 and 6 mA. The ERs for p- and s-polarizations were determined from the ratios between the transmissions at 0 mA and 6 mA and shown in Figs. 4(b) and 4(c), respectively. As expected, the spectral features and the ERs reached their maximum at 6 mA. For instance, at a wavelength of 1550 nm, the ERs of p- and s-polarized light were 5.6 dB and 4.3 dB,

Letter respectively. The measurements share qualitative agreement with the simulations in Fig. 2(b). For both the simulations and measurements, the ER for the p-polarization reached a minimum of around 1530 nm, and the ER for the s-polarization increased monotonically with wavelength, with a maximum difference of about 4 dB between 1480 nm and 1580 nm. However, the measured ER values were lower than the simulation by about 4 dB and 5 dB for the p- and s-polarizations, respectively. The reasons for the discrepancy are that, as evidenced in Fig. 3(c), only a portion of the film underwent the phase transition and only a few periods of the gratings were excited. Although the transition region expanded as the current increased, it could not cover the entire the grating region for the current values tested. Higher currents led to device failure. To investigate the dependence of the tunability of EOT on device geometry, we measured gratings with various duty cycles. For these devices, the period and thickness were kept to P  1 μm and t  260 nm, respectively, but the gap width, w, was varied. The VO2 film thickness remained at 170 nm. Figures 5(a) and 5(b) show the simulated ER spectra for p- and s-polarizations, respectively, for three values of w: w1  250 nm, w2  300 nm, and w3  350 nm. Using the same settings and conditions as before, the zeroth-order transmission spectra of p- and s-polarized normal incident light were measured, and the ER spectra were extracted at a current of 6 mA. The change in duty cycle did not affect the current required to induce the phase transition. The results are shown in Figs. 5(c) and 5(d), respectively. The measurements and simulations again share good qualitative agreement. Interestingly, the ERs of the p- and s-polarization have opposite dependences on the gap widths. The p-polarization ER spectrum was the highest for the largest gap (w3 ), whereas the s-polarization ER spectrum was the highest for the smallest gap (w1 ). The p-polarization ER spectra can be explained qualitatively by the grating strength: a wider gap leads to a higher grating strength, which increases the bandwidth of the grating, resulting in both a broader spectral bandwidth and increased interaction with and absorption by the VO2 film. For the s-polarization, the ER spectral shape did not exhibit much dependence on w. The transmission dependence on w can be seen as simply tending toward that of a VO2 film as w → ∞ [see Fig. 2(a)]. The measurement for w2 did not completely follow the trend, and the discrepancy may be due to nonuniformity of the VO2 film across the devices. In sum, we have demonstrated integrated electrical control of EOT using one-dimensional Au gratings patterned atop a phase transition material, VO2 . The gratings were also used for current injection to induce the phase transition in the VO2 to enable dense electrical integration. An extension would be to investigate the tunability of the optical phase of the transmitted light and switching speeds. The result presented here, combined with nanosecond electrical triggering of the insulator– metal phase transition in VO2 [34,35], shows the potential of creating high-speed spatial light modulators and dynamic beam steering metasurfaces using VO2 . Funding. Canada Research Chairs (Chaires de recherche du Canada); Natural Sciences and Engineering Research Council of Canada (NSERC) (Conseil de Recherches en Sciences Naturelles et en Génie du Canada).

Vol. 40, No. 19 / October 1 2015 / Optics Letters

4411

REFERENCES 1. T. W. Ebbesen, H. J. Lezec, F. Ghaemi, T. Thio, and P. A. Wolff, Nature 391, 667 (1998). 2. L. Martín-Moreno, F. J. García-Vidal, H. J. Lezec, K. M. Pellerin, T. Thio, J. B. Pendry, and T. W. Ebbesen, Phys. Rev. Lett. 86, 1114 (2001). 3. C. Genet and T. W. Ebbesen, Nature 445, 39 (2007). 4. F. J. García-Vidal, L. Martín-Moreno, T. W. Ebbesen, and L. Kuipers, Rev. Mod. Phys. 82, 729 (2010). 5. H. A. Bethe, Phys. Rev. 66, 163 (1944). 6. U. Schröter and D. Heitmann, Phys. Rev. B 58, 15419 (1998). 7. J. A. Porto, F. J. García-Vidal, and J. B. Pendry, Phys. Rev. Lett. 83, 2845 (1999). 8. H. Liu and P. Lalanne, Nature 452, 728 (2008). 9. F. J. García-Vidal, H. J. Lezec, T. W. Ebbesen, and L. Martín-Moreno, Phys. Rev. Lett. 90, 213901 (2003). 10. Y. Pang, C. Genet, and T. Ebbesen, Opt. Commun. 280, 10 (2007). 11. F. van Beijnum, C. Rétif, C. B. Smiet, H. Liu, P. Lalanne, and M. P. van Exter, Nature 492, 411 (2012). 12. E. A. Shaner, J. G. Cederberg, and D. Wasserman, Appl. Phys. Lett. 91, 181110 (2007). 13. J. Y. Suh, E. U. Donev, R. Lopez, L. C. Feldman, and R. F. Haglund, Appl. Phys. Lett. 88, 133115 (2006). 14. C. N. Berglund and H. J. Guggenheim, Phys. Rev. 185, 1022 (1969). 15. A. Joushaghani, J. Jeong, S. Paradis, D. Alain, J. S. Aitchison, and J. K. S. Poon, Appl. Phys. Lett. 104, 221904 (2014). 16. A. Joushaghani, J. Jeong, S. Paradis, D. Alain, J. S. Aitchison, and J. K. S. Poon, Appl. Phys. Lett. 105, 231904 (2014). 17. K. Appavoo, B. Wang, N. F. Brady, M. Seo, J. Nag, R. P. Prasankumar, D. J. Hilton, S. T. Pantelides, and R. F. Haglund, Nano Lett. 14, 1127 (2014). 18. M. A. Kats, R. Blanchard, P. Genevet, Z. Yang, M. M. Qazilbash, D. N. Basov, S. Ramanathan, and F. Capasso, Opt. Lett. 38, 368 (2013). 19. M. J. Dicken, K. Aydin, I. M. Pryce, L. A. Sweatlock, E. M. Boyd, S. Walavalkar, J. Ma, and H. A. Atwater, Opt. Express 17, 18330 (2009). 20. M. D. Goldflam, T. Driscoll, B. Chapler, O. Khatib, N. Marie Jokerst, S. Palit, D. R. Smith, B.-J. Kim, G. Seo, H.-T. Kim, M. D. Ventra, and D. N. Basov, Appl. Phys. Lett. 99, 044103 (2011). 21. S. K. Earl, T. D. James, T. J. Davis, J. C. McCallum, R. E. Marvel, R. F. Haglund, and A. Roberts, Opt. Express 21, 27503 (2013). 22. A. Joushaghani, B. A. Kruger, S. Paradis, D. Alain, J. S. Aitchison, and J. K. S. Poon, Appl. Phys. Lett. 102, 061101 (2013). 23. A. Joushaghani, J. Jeong, S. Paradis, D. Alain, J. S. Aitchison, and J. K. S. Poon, Opt. Express 23, 3657 (2015). 24. A. Joushaghani, “Micro- and nano-scale optoelectronic devices using vanadium dioxide,” Ph.D. dissertation (University of Toronto, 2014). 25. G. Stefanovich, A. Pergament, and D. Stefanovich, J. Phys. Condens. Matter 12, 8837 (2000). 26. J. S. White, G. Veronis, Z. Yu, E. S. Barnard, A. Chandran, S. Fan, and M. L. Brongersma, Opt. Lett. 34, 686 (2009). 27. C. Berglund, IEEE Trans. Electron Devices 16, 432 (1969). 28. J. Duchene, M. Terraillon, P. Pailly, and G. Adam, Appl. Phys. Lett. 19, 115 (1971). 29. A. Zimmers, L. Aigouy, M. Mortier, A. Sharoni, S. Wang, K. G. West, J. G. Ramirez, and I. K. Schuller, Phys. Rev. Lett. 110, 056601 (2013). 30. S. Zhang, M. A. Kats, Y. Cui, Y. Zhou, Y. Yao, S. Ramanathan, and F. Capasso, Appl. Phys. Lett. 105, 211104 (2014). 31. H.-T. Kim, B.-J. Kim, S. Choi, B.-G. Chae, Y. W. Lee, T. Driscoll, M. M. Qazilbash, and D. N. Basov, J. Appl. Phys. 107, 023702 (2010). 32. C. Aurelian, G. Julien, L. Jonathan, M. David, C. Corinne, O. JeanChristophe, C. Alain, and B. Pierre, Sci. Technol. Adv. Mater. 11, 065002 (2010). 33. A. Joushaghani, J. Jeong, S. Paradis, D. Alain, J. S. Aitchison, and J. K. S. Poon, Photonics 2, 916 (2015). 34. Y. Zhou, X. Chen, C. Ko, Z. Yang, C. Mouli, and S. Ramanathan, IEEE Electron Device Lett. 34, 220 (2013). 35. P. Markov, R. E. Marvel, H. J. Conley, K. J. Miller, R. F. Haglund, and S. M. Weiss, ACS Photonics 2, 1175 (2015).