Electrodeposition Growth of Nanowire Arrays with Height Gradient

Apr 15, 2011 - 78216 San Luis Potosí, SLP, Mexico. Nowadays the quest for novel synthesis methods to produce nano-objects that provide control over ...
0 downloads 0 Views 2MB Size
LETTER pubs.acs.org/NanoLett

Electrodeposition Growth of Nanowire Arrays with Height Gradient Profiles for Microwave Device Applications Catalina E. Carreon-Gonzalez,† Joaquín De La Torre Medina,*,† Luc Piraux,‡ and Armando Encinas†,§ †

Instituto de Física, Universidad Autonoma de San Luis Potosí, Av. Manuel Nava 6, Zona Universitaria, 78290 San Luis Potosí, SLP, Mexico ‡ Institute of Condensed Matter and Nanosciences (ICMN), Universite Catholique de Louvain, Place Croix du Sud 1, B-1348, Louvain-la-Neuve, Belgium § Division de Materiales Avanzados, Instituto Potosino de Investigacion Científica y Tecnologica A. C., Caminio a la Presa 2055, 78216 San Luis Potosí, SLP, Mexico ABSTRACT: A simple and nonexpensive adapted dip-coating technique is presented and used to fabricate arrays of magnetic nanowires with a linear varying height profile. This approach allows controlling the wire height from tenths of nanometers up to several micrometers. Furthermore, the main parameters of this height gradient can be controlled, such as the maximum wire height and the lateral span of the wire array, which can be predicted with excellent accuracy using a proposed analytical model. Moreover, we show that by sequential electrodeposition with dip-coating, arrays of these height varying wires can be grown. This technique represents a novel method to fill porous templates with controlled spatial growth, leading to the fabrication of novel structures and providing control over structural features on the nanoscale level. In particular, the use of these asymmetrically loaded magnetic nanowired substrates to obtain improved microwave nonreciprocal behavior is shown for a microwave phase shifter. KEYWORDS: Height gradient profile, microwave devices, dip coating, nanowires, electrodeposition

N

owadays the quest for novel synthesis methods to produce nano-objects that provide control over specific structural, geometrical, and spatial features is one of the main topics of interest in nanosciences and nanotechnology since it gathers different research fields with the ultimate goal of producing novel and more functional materials and devices. In particular, electrodeposition is a nonexpensive, very versatile, and fast technique for the production of large arrays of metallic wires into porous templates where the diameter can be varied from 10 nm to several hundred of micrometers. For instance, arrays of nanowires (NWs) have received particular interest in recent years for their application as microwave devices,16 field emission,710 and pH, bio-, and gas sensors.1114 The performance of these systems depends on the materials from which the NWs are made of and on their size. Control on the NWs growth to obtain specific topographical designs by only varying the NWs height is crucial for the fabrication of devices with improved or novel functionalities. Indeed, it has been shown that magnetic nanowired substrates (MNWS) in microstrip topologies and filled asymmetrically with NWs exhibit microwave nonreciprocal behavior3 and a dependence of field emission on the diameter and on the NWs height has recently been evidenced.15,16 Nevertheless, control on the topography defined by the NWs height using electrodeposition is difficult to achieve since the porous template is exposed completely to the electrolyte which fills all the pores, so all the NWs grow nearly at the same speed and have, in average, the same height. Control on the NWs topography has r 2011 American Chemical Society

recently been achieved by using electronic lithography,11 lithographically patterned NW electrodeposition,17 and combinational templated assisted techniques.18 However these techniques are complex, require specialized equipment, and have inherent limitations on their capabilities, in particular, a monotonous growth and then height gradient profiles of NWs are not foreseen. In this work we present a novel and low cost technique based on electrodeposition and dip-coating of the template, which allows control of the NWs height and leads to specific topographical designs of NWs into the porous template. With this technique the NW growth presents a transversal height gradient profile along a direction parallel to the solution level. The geometrical features of such height gradient profiles, this is their width and maximum height, are predicted with remarkeable accuracy using a phenomenological model based on simple mass conservation arguments of the solution that leaves the electrodeposition cell. Finally, growing NW arrays in this way leads to an asymmetrically loaded MNWS which has a direct application in microwave devices such as phase shifters with improved nonreciprocal behavior. Nanowire array fabrication is carried out by electrodeposition19,20 and is based on the approach developed for growing NWs21 into commercial 60 μm thick porous anodized alumina (Al2O3) templates with average diameter and porosity of 150 nm Received: February 1, 2011 Revised: April 11, 2011 Published: April 15, 2011 2023

dx.doi.org/10.1021/nl2003939 | Nano Lett. 2011, 11, 2023–2027

Nano Letters

Figure 1. Schematics of the transverse view of the combined twoprobe electrodeposition and dip-coating setup. The cell is filled with an electrolyte and contains the porous alumina template inside a sample holder which is in contact with the cathode and the solution.

and 40%, respectively. Electrodeposition is done at a constant potential of 1.3 V using a Ni plating solution (Watts Nickel Pure, Technic). Prior to deposition a thin layer of indium gallium eutectic (Aldrich) is applied to one surface of the membrane to serve as a cathode. The painted side was placed in contact with a thin copper plate as shown in Figure 1. The deposition rate νdep is determined from the time necessary to completely fill the porous membrane, which is done by using an Ag/AgCl reference electrode for Ni plating at 1.3 V. Then all samples are fabricated by growing the nanowires at a desired height without the need of using the reference electrode since the nanowires growth is controlled with the deposition time. Microwave transmission measurements have been done using a microstrip transmission line. For this purpose, after electrodeposition the InGa eutectic is wiped out using isopropanol and a Cr(20 nm)/Au(600 nm) layer is evaporated onto the wiped side of the membrane to serve as ground plane for the microwave measurements. Next, a 500 μm wide and 1 cm long microstrip line is evaporated onto the side of the membrane opposite to the ground plane. Scanning electron microscopy (SEM) is used for the characterization of the NWs height inside the alumina membrane. Measurements of absorption spectra and differential phase shift in frequency swept mode are performed using a vector network analyzer (VNA) after saturating the samples to 10 kOe and at different applied fields. During electrodeposition a gradient of NWs heights along a specific in-plane direction is induced. This is done by using a dipcoating technique while electrodeposition is performed, which consists of submerging vertically the Al2O3 template into a cell containing an electrolyte and to slowly drain part of the solution as seen schematically in Figure 1. As a result, the level of the solution drops and the NWs growth continues only into the pores that are in contact with the solution. As shown in Figure 1 the Al2O3 membrane is embedded into a holder with the purpose that only the side of the membrane without InGa eutectic is exposed to the electrolyte. In this way, electrodeposition takes place into the pores that are in contact with the solution, including those beneath the meniscus formed between the solution and the membrane holder. Electrodeposition and

LETTER

Figure 2. (a) Schematic view of the geometrical parameters of a NW array with height gradient profile. SEM micrographs of (b) 820 μm wide, (c) 390 μm wide, and (d) 700 μm wide NW arrays with height gradient profile. (e) NW array with a 1100 μm wide sawlike NW height gradient profile, where the dotted line is a guide for the eye. In all figures the scale bars represent 100 μm.

the solution leakage are stopped until the wire heights, at the maximum of the NWs height gradient, reach a desired value hf below the membrane thickness. This leads to a height gradient profile with width Δy along a length L parallel to the solution level, where the geometrical parameters Δy and hf (see Figure 2a) depend on the solution leakage rate via the droplet leakage rate νl. In the dip-coating experiment the net solution volume that leaves the cell Vout and the net solution volume decrease in the cell Vdec are related through the equation Vout = Vdec, where Vout = ndV = νltV and Vdec = AΔy. In these expressions nd is the number of pendant droplets, νl = nd/t is the number of pendant droplets per time t or the average droplet leakage rate, V = 16.7 mm3 is the volume of a single pendant droplet, A = 778 mm2 is the effective horizontal area of the solution, and Δy corresponds to the width of the height gradient profile. On the other hand, in our experiment the NWs growth rate νdep is known and is equal to 1.5 μm 3 min1 for a deposition potential of 1.3 V. The deposition time t = hf/νdep corresponds to the time needed to grow NWs at the height hf for the known NWs growth rate νdep = 1.5 μm 3 min1. Therefore, from the previous analysis, we can determine geometric parameters of the height gradient profile like the maximum NWs height    AΔy νdep ð1Þ hf ¼ V νl and the initial angle θ = tan1(hf/Δy) (see Figure 2a) of the height gradient profile   Aνdep ð2Þ θ ¼ tan1 V νl Equation 1 allows the determination of hf once the droplet leakage rate νl is given for a specific value of the width Δy of the height gradient profile. Besides as expected, the initial angle θ depends only on νl since all the other quantities in eq 2 are known constant values. Particularly, the height gradient profiles of panels b, c, and d of Figure 2 have been grown using specific values of the droplet leakage rate νl and Δy, which are equal to 0.23 droplets 3 min1 and 820 μm; 1.33 droplets 3 min1 and 2024

dx.doi.org/10.1021/nl2003939 |Nano Lett. 2011, 11, 2023–2027

Nano Letters

Figure 3. Measured (symbols) and calculated (lines) (a) hf vs νl and (b) θ vs νl for height gradient profiles with average widths Δy of 480 μm (squares and dashed line) and 910 μm (circles and dotted line).

390 μm; and 2.67 droplets 3 min1 and 700 μm, respectively. This leads to height gradient profiles with specific hf values of 60 μm (Figure 2b), 19.5 μm (Figure 2c), and 19.3 μm (Figure 2d). Notice from Figure 2b and Figure 2d that different hf values are obtained for nearly the same value of Δy which is done just by varying the droplet leakage rate. Varying νl allows control of θ via eq 2, which is closely related to hf. Particularly, increasing νl leads to a decrease on θ as suggested by eq 2 and from Figure 2b and Figure 2d. Furthermore, our technique is suitable to fabricate more complex topographic designs by repeating the electrodeposition and dipcoating process several times onto the same membrane. This is carried out by fabricating a first height gradient profile by following the electrodeposition and dip-coating technique mentioned above. Then, a second line is painted with the InGa eutectic next to where the first height gradient profile has been grown. Electrodeposition of the second height gradient profile is done by placing the membrane with the previous height gradient profile outside the electrolyte and the second one into the electrolyte. This process can then be repeated to fabricate a desired number of adjacent height gradient profiles. Indeed, following this procedure has permitted fabrication of sawlike NW height profiles with total width of about 1100 μm as the one shown in the SEM micrograph of Figure 2e. In order to validate eq 1 and eq 2 for hf and θ, respectively, several height gradient profiles have been fabricated by varying νl and the deposition time. Samples with similar Δy have been gathered into a set of samples, so the parameters mentioned above are compared with the analytical model which is calculated using the average Δy value for each set of samples as input parameter. Figure 3a shows the variation of the final nanowires height hf as a function of the solution leakage rate νl for average Δy values of 480 μm (squares) and 910 μm (circles). As observed, lower values of hf are expected for higher νl for a constant Δy (dotted and dashed lines). This means that for a specific Δy value, the height gradient profile is difficult to be observed for high droplet leakage rates because hf is low compared to the membrane thickness of 60 μm. Further, at a fixed νl one obtains height gradient profiles with lower hf if Δy is lowered as suggested by eq 1 where a linear dependence of hf on Δy is observed. Notice that the geometrical parameters of the height

LETTER

gradient profile like Δy and hf are related to each other through the tangent of the initial angle θ of the microstructure, so these quantities can be gathered into a single and more general curve. Figure 3b shows the dependence of θ on νl, where we observe that θ decreases as νl increases, which is consistent with what is observed in Figure 3a. This is increasing significantly the droplet leakage rate leads to a faster drop in the solution level into the electrolytic cell, which in turn prevents the NWs from growing with significantly different heights along the lateral span of the NW array. As a consequence, the height gradient profile will present a low initial angle θ and a low maximum wires height. Deviations of the experimental hf and θ from those calculated using eq 1 and eq 2, are provided by the error bars in the experimental data in both Figure 3a and Figure 3b. Error bars for both hf and θ are given by their standard deviations, which have been determined from the standard deviation of Δy for both set of samples with average Δy = 480 μm (squares) and Δy = 910 μm (circles). All in all a very good agreement between the experimental values and the analytical model is observed in all figures, which validates the equations discussed above. On the other hand, these microstructures are interesting for the development of NW based sensors and field emission devices with NW height gradient profiles or specific topographic designs. However, a direct application of these NW arrays is focused on nonreciprocal microstrip line (NRML) in order to improve the microwave nonreciprocal behavior reported in ref 3. An asymmetrical loading along the width of the microstrip with nanowires is at the basis of the nonreciprocal behavior, which, in the case of ref 3, this is achieved by loading the microstrip with nanowires as a stairway-like height gradient profile defined in three adjacent zones. Figure 4a shows the schematic view of a NRML, which consists of a microstrip line evaporated onto an Al2O3 membrane filled asymmetrically with NWs. Recently, it has been demonstrated that the effective permittivity of a MNWS under a microstrip line depends on the NWs height.22 This means that the effective permittivity of the NRML shown in Figure 4a progressively changes along the width of the microstrip, in contrast to what is reported in ref 3, where the permittivity variations are discrete. Indeed, in the previous NRML reported in ref 3, both the effective permittivity and the ac electric field change from one zone to the next one; however they remain unchanged within each zone where the nanowires height is constant, limiting the nonreciprocal behavior and the differential phase shift characteristics. Therefore, varying continuously the nanowires height leads to a continuous variation of the permittivity, which combined with the solution proposed by Hines for the edge guided mode23 may yield an improved microwave nonreciprocal operation of the device. As a consequence, the nonreciprocal behavior and the isolation and differential phase shift characteristics may be maximized in NRML with linear height gradient profiles with hf closer to the porous membrane thickness and high θ values. Using the combined electrodeposition and dip-coating technique described above has allowed fabrication of NRML as the one shown in Figure 4a. Figure 4b displays the transverse view of a NRML with width Δy = 500 μm and length L = 1 cm, which clearly shows the height gradient profile along the transverse direction to the microstrip line. Indeed, the insertion losses Sþ and S respectively for the forward (þ) and backward () propagation directions (see Figure 4c), measured using a VNA at 4 and 8 kOe, show a nonreciprocal behavior since the absorption depth at the ferromagnetic resonance is different. The difference 2025

dx.doi.org/10.1021/nl2003939 |Nano Lett. 2011, 11, 2023–2027

Nano Letters

Figure 4. (a) Schematic view of the NRML on a MNWS. The NWs under the microstrip line present a height gradient profile along the width of the microstrip line and (b) SEM micrograph of the transversal section of a 500 μm wide and 1 cm long NRML, filled asymmetrically with NWs as shown in (a). Measured (c) S(/L at 4 and 8 kOe, (d) ΔS/L at 0, 2, 4, 6, and 8 kOe and (e) the corresponding Δj/L for the NRML of (b). The scale bar in (b) represents 100 μm, and the field values denoted as numbers in (c), (d), and (e) are given in kOe.

in absorption depths in both propagation directions leads to an isolation per unit length ΔS/L = (Sþ  S)/L that increases with the applied field as seen from Figure 4d (numbers are in kOe). Despite the isolation, performances are still poor because the maximum isolation is at the ferromagnetic resonance frequency fr, the isolation level is about three times higher than that of the NRML in ref 3. Besides, Figure 4e shows the differential phase shift recorded at the same field values, after saturating the sample at 9 kOe, as in Figure 4d, which is defined as the difference between phase shifts of the forward and backward propagation directions, this is Δj = jþ  j. As seen from this figure, Δj 6¼ 0 for frequencies in the vicinity and higher than fr for all applied field values and its behavior is similar to that reported in ref 3, which is attributed to the nonreciprocal operation of the microstrip line. As seen, Δj/L ≈ 40 deg 3 cm1 is obtained

LETTER

at 40 GHz and at 8 kOe for frequencies above fr where FMR losses and particularly ΔS/L vanish. This differential phase shift value is of about twice the value reported in ref 3 at high field values; however its value at zero field is lower than the reported value of 10 deg 3 cm1. As known, the nonzero differential phase shift at zero field arises from the self-demagnetizing field of the NWs which make them single domain, so they possess a self-biasing field. Therefore, as expected in this case the Δj/L value is lower than the previously reported one3 due to their large diameter and density.24,25 The results shown in panels d and e of Figure 4 demonstrate that a height gradient profile of the NWs improves drastically the performances of NRMLs in comparison to an asymmetrically loading of the membranes with NWs having a stairway-like height gradient profile. Despite the figure of merit of the present device, defined as the ratio between the differential phase shift and the average insertion losses, is still low because it is nearly the same as for the device of ref 3, the Δj/L and ΔS/L have been significantly improved. Finaly, the performances of our device are comparable to those of conventional devices based on ferrites, which exhibit Δj/L in the range 2053 deg 3 cm1 for frequencies as high as 10 GHz.2629 In conclusion we have developed a novel and low-cost combined electrodeposition and dip-coating technique for the fabrication of arrays of NWs with a linear varying height profile which leads to specific topographic NW designs. This technique allows controlling the main parameters of the height gradient profiles such as the maximum wire height and the lateral span of the wire array. Furthermore these parameters can be predicted with very good accuracy using a proposed analytical model based on mass conservation arguments of the solution that leaves the electrolytic cell. The resulting designs present continuous variations of the NWs height which makes them interesting for nonreciprocal microwave applications. Particularly, the fabricated nonreciprocal microstrip lines present improved isolation and differential phase shift compared with those having asymmetrical fillings of NWs with stairway-like height gradient profiles.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This work was partly supported by the Interuniversity Attraction Poles Program (P6/42)-Belgian State-Belgian Science Policy. The authors specially thanks Rosa Lina Tovar Tovar for providing the SEM micrographs. J. De La Torre thanks CONACYT for financial support through Grant No. 62292 and A.E. thanks CONACYT for financial support through Grant No. 105568 as well as UCL. ’ REFERENCES (1) Kuanr, B. K.; Veerakumar, V.; Marson, R.; Mishra, S. R.; Camley, R. E.; Celinski, Z. J. Nonreciprocal microwave devices based on magnetic nanowires. Appl. Phys. Lett. 2009, 94, 202505. (2) Darques, M.; De la Torre Medina, J; Piraux, L.; Cagnon, L.; Huynen, I. Microwave circulator based on ferromagnetic nanowires in an alumina template. Nanotechnology 2010, 12, 145208. 2026

dx.doi.org/10.1021/nl2003939 |Nano Lett. 2011, 11, 2023–2027

Nano Letters (3) De La Torre Medina, J.; Spiegel, J.; Darques, M.; Piraux, L.; Huynen, I. Differential phase shift in nonreciprocal microstrip lines on magnetic nanowired substrates. Appl. Phys. Lett. 2010, 96, 072508. (4) Saib, A.; Vanhoenacker-Janvier, D.; Huynen, I.; Encinas, A.; Piraux, L.; Ferain, E.; Legras, R. Magnetic photonic band-gap material at microwave frequencies based on ferromagnetic nanowires. Appl. Phys. Lett. 2003, 83, 2378. (5) Allaeys, J. F.; Marcilhac, B.; Mage, J. C. Influence of track-etching on polycarbonate membrane permittivity. J. Phys. D: Appl. Phys. 2007, 40, 3714–3717. (6) Boucher, V.; Menard, D. Effective magnetic properties of arrays of interacting ferromagnetic wires exhibiting gyromagnetic anisotropy and retardation effects. Phys. Rev. B 2010, 81, 174404. (7) Vila, L.; Vincent, P.; Dauginet-De Pra, L.; Pirio, G.; Minoux, E.; Gangloff, L.; Demoustier-Champagne, S.; Sarazin, N.; Ferain, E.; Legras, R.; Piraux, L.; Legagneux, P. Growth and field-emission properties of vertically aligned cobalt nanowire arrays. Nano Lett. 2004, 4, 521–524. (8) Dangwal, A.; Pandey, C. S.; M€uller, G.; Karim, S.; Cornelius, T. W.; Trautmann, C. Field emission properties of electrochemically deposited gold nanowires. Appl. Phys. Lett. 2008, 92, 063115. (9) Navitski, A.; Muller, G.; Sakharuk, V.; Cornelius, T. W.; Trautmann, C.; Karim, S. Efficient field emission from structured gold nanowire cathodes. Eur. Phys. J.: Appl. Phys. 2009, 48, 30502. (10) Baranova, L. A.; Baryshev, S. V.; Gusinskii, G. M.; Konnikov, S. G.; Nashchekin, A. V. Nickel field-emission microcathode: Art of fabrication, properties, and applications. Nucl. Instrum. Methods Phys. Res., Sect. B 2010, 268, 1686–1688. (11) Antohe, V. A.; Radu, A.; Matefi-Tempfli, M.; Attout, A.; Yunus, S.; Bertrand, P.; Dutu, C. A.; Vlad, A.; Melinte, S.; Matefi-Tempfli, S.; Piraux, L. Nanowire-templated microelectrodes for high-sensitivity pH detection. Appl. Phys. Lett. 2009, 94, 073118. (12) Shi, A. W.; Qu, F. L.; Yang, M. H.; Shen, G. L.; Yu, R. Q. Amperometric H2O2 biosensor based on poly-thionine nanowire/ HRP/nano-Au-modified glassy carbon electrode. Sens. Actuators, B 2008, 129, 779–783. (13) Walter, E. C.; Penner, R. M.; Liu, H.; Ng, K. H.; Zach, M. P.; Favier, F. Sensors from electrodeposited metal nanowires. Surf. Interface Anal. 2002, 34, 409–412. (14) Aravamudhan, S.; Kumar, A.; Mohapatra, S.; Bhansali, S. Sensitive estimation of total cholesterol in blood using Au nanowires based micro-fluidic platform. Biosens. Bioelectron. 2007, 22, 2289–2294. (15) Wang, S.; He, Y.; Fang, X.; Zou, J.; Wang, Y.; Huang, H.; Costa, P. M. F. J.; Song, M.; Huang, B.; Liu, C. T.; Liaw, P. K.; Bando, Y.; Golberg, D. Structure and Field-Emission Properties of Sub-Micrometer-Sized Tungsten-Whisker Arrays Fabricated by Vapor Deposition. Adv. Mater. 2009, 21, 2387–2392. (16) Zhang, X.-G.; Pantelides, S. T. Screening in Nanowires and Nanocontacts: Field Emission, Adhesion Force, and Contact Resistance. Nano Lett. 2009, 9, 4306–4310. (17) Kim, H.; Taggart, D. K.; Xiang, C.; Penner, R. M.; Potma, E. O. Spatial control of coherent anti-stokes emission with height-modulated gold zig-zag nanowires. Nano Lett. 2008, 8, 2373–2377. (18) Chong, M. A. S.; Zheng, Y. B.; Gao, H.; Tan, L. K. Combinational template-assisted fabrication of hierarchically ordered nanowire arrays on substrates for device applications. Appl. Phys. Lett. 2006, 89, 233104. (19) Strijkers, G. J.; Dalderop, J. H. J.; Broeksteeg, M. A. A.; Swagten, H. J. M.; de Jonge, W. J. M. Structure and magnetization of arrays of electrodeposited Co wires in anodic alumina. J. Appl. Phys. 1999, 86, 5141–5145. (20) Nielsch, K.; Wehrspohn, R. B.; Barthel, J.; Kirschner, J.; G€osele, U. Hexagonally ordered 100 nm period nickel nanowire arrays. Appl. Phys. Lett. 2000, 79, 1360–1362. (21) Bentley, A. K.; Farhoud, M.; Ellis, A. B.; Lisensky, G. C.; Nickel, A.-M.; Crone, W. C. Template synthesis and magnetic manipulation of nickel nanowires. J. Chem. Educ. 2005, 82, 765–768. (22) Spiegel, J.; De La Torre, J.; Darques, M.; Piraux, L.; Huynen, I. Permittivity model for ferromagnetic nanowired substrates. IEEE Microwave Wireless Compon. Lett. 2007, 17, 492–494.

LETTER

(23) Hines, M. E. Reciprocal and nonreciprocal modes of propagation in ferrite stripline and microstrip devices. IEEE Trans. Microwave Theory Tech. 1971, MMT-19, 442-451. (24) Ferre, R.; Ounadjela, K.; George, J. M.; Piraux, L.; Dubois, S. Magnetization processes in nickel and cobalt electrodeposited nanowires. Phys. Rev. B 1997, 56, 14066–14075. (25) Hertel, R. Micromagnetic simulations of magnetostatically coupled Nickel nanowires. J. Appl. Phys. 2001, 90, 5752–5758. (26) Wen, C. P. Coplanar waveguide: A surface strip transmission line suitable for nonreciprocal gyromagnetic device applications. IEEE Trans. Microwave Theory Tech. 1969, MMT-17, 1087–1090. (27) Mielewski, J.; Buda, A. Analysis of the nonreciprocal ferrite phase shifter with nonuniform cross-section. Proc. 12th Int. Conf. Microwave Radar (MIKON ’98) 1998, 2, 509–513. (28) Zafar, J.; Gibson, A. A. P.; Zafar, H. R. High power ferrite phase shifter for beam steering applications. Proc. 3rd Eur. Conf. Antenna Prop. (EUCAP 2009) 2009, 3029–3033. (29) Junding, W.; Xiong, Y.-Z.; Shi, M.-J.; Chen, G.-F.; Yu, M.-D. Analysis of twin ferrite toroidal phase shifter in grooved waveguide. IEEE Trans. Microwave Theory Tech. 1994, MTT-42, 616–621.

2027

dx.doi.org/10.1021/nl2003939 |Nano Lett. 2011, 11, 2023–2027