Home
Search
Collections
Journals
About
Contact us
My IOPscience
Novel aluminum near field transducer and highly integrated micro-nano-optics design for heatassisted ultra-high-density magnetic recording
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 295202 (http://iopscience.iop.org/0957-4484/25/29/295202) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 134.117.10.200 This content was downloaded on 13/05/2015 at 14:43
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 459501 (1pp)
doi:10.1088/0957-4484/25/45/459501
Corrigendum: Novel aluminum near field transducer and highly integrated micronano-optics design for heat-assisted ultrahigh-density magnetic recording (2014 Nanotechnology 25 295202) Lingyun Miao1, Paul R Stoddart2 and Thomas Y Hsiang1 1
Department of Electrical and Computer Engineering, University of Rochester, Rochester, NY 14627, USA 2 Centre for Quantum and Optical Science, Faculty of Science, Engineering and Technology, Swinburne, University of Technology, Hawthorn, VIC 3122, Australia E-mail: [email protected] Received 29 September 2014 Accepted for publication 29 September 2014 Published 20 October 2014
should be ‘In fact, the penetration depth into cobalt is quite small (less than 11 nm) at the optimum wavelength of 450 nm.’ No other typo was found in the published paper.
There was an error made on page 3 of the original manuscript, it reads ‘In fact, the penetration depth into cobalt is quite small (less than 1 nm) at the simulation wavelengths.’ It
0957-4484/14/459501+01$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
Nanotechnology Nanotechnology 25 (2014) 295202 (7pp)
doi:10.1088/0957-4484/25/29/295202
Novel aluminum near field transducer and highly integrated micro-nano-optics design for heat-assisted ultra-high-density magnetic recording Lingyun Miao1, Paul R Stoddart2 and Thomas Y Hsiang1 1
Department of Electrical and Computer Engineering, University of Rochester, Rochester, NY 14627, USA 2 Centre for Quantum and Optical Science, Faculty of Science, Engineering and Technology, Swinburne University of Technology, Hawthorn, VIC 3122, Australia E-mail: [email protected] Received 10 December 2013, revised 30 May 2014 Accepted for publication 6 June 2014 Published 1 July 2014 Abstract
Heat-assisted magnetic recording (HAMR) has attracted increasing attention as one of the most promising future techniques for ultra-high-density magnetic recording beyond the current limit of 1 Tb in−2. Localized surface plasmon resonance plays an important role in HAMR by providing a highly focused optical spot for heating the recording medium within a small volume. In this work, we report an aluminum near-field transducer (NFT) based on a novel bow-tie design. At an operating wavelength of 450 nm, the proposed transducer can generate a 35 nm spot size inside the magnetic recording medium, corresponding to a recording density of up to 2 Tb in−2. A highly integrated micro-nano-optics design is also proposed to ensure process compatibility and corrosion-resistance of the aluminum NFT. Our work has demonstrated the feasibility of using aluminum as a plasmonic material for HAMR, with advantages of reduced cost and improved efficiency compared to traditional noble metals. S Online supplementary data available from stacks.iop.org/NANO/25/295202/mmedia Keywords: aluminum, bow-tie, near field transducer, heat-assisted magnetic recording (Some figures may appear in colour only in the online journal) 1. Introduction
The biggest challenge for the HDD industry is that conventional perpendicular magnetic recording is about to reach its limit of recording density at 1 Tb in−2 [4, 5]. Among several promising future technologies such as bit-patterned media, heat-assisted magnetic recording (HAMR), and shingled magnetic recording, HAMR is considered one of the most flexible techniques that can be implemented in the conventional HDD industry for extending areal density [6–8]. HAMR utilizes a focused laser beam to define magnetic recording features. As illustrated in figure 1 [9], the recording density under HAMR is determined by the thermomagnetic recording mechanism. In general, a special recording medium which has good thermal stability at room temperature is used
A significant part of the world’s economy has been related to the IT and consumer electronics markets in recent years [1], and digital data is a key contributor to its growth. Hard disk drives (HDDs) have dominated the digital data storage market as the primary non-volatile memory device, despite strong competition from solid-state drives (SSDs) [2]. Surprisingly, however, a recent technology roadmap analysis shows that the older generation tape-based magnetic recording (TAPE) would probably outperform both HDD and SSD in the next six years in terms of areal density extendibility [3], provided that there is no breakthrough in the latter technologies. 0957-4484/14/295202+07$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
L Miao et al
Nanotechnology 25 (2014) 295202
Figure 2. Novel bow-tie aluminum antenna design (top view, not
drawn to scale, unit: nm). Figure 1. The use of a focused laser beam to define magnetic
recording features under the HAMR scheme (not drawn to scale). Reprinted with permission from [9]. Copyright 2012 The Institution of Engineering and Technology. The inset shows a HDD with a conventional read/write head over the recording surface.
in HAMR. When a focused laser beam is delivered onto the medium’s surface, the local medium is temporarily heated up to exceed its Curie temperature and lower its coercivity to allow magnetic recording. When the local medium cools down, it becomes stable again to retain data. Typically a laser power of about several tens of milliwatt is necessary to make this scheme work. In order to achieve ultra-high-density magnetic recording, it is necessary to deliver a highly focused laser beam onto the surface of the medium. For instance, an optical spot size of 50 nm or less is necessary for a recording density of 1 Tb in−2 [6]. Thus, metallic near-field transducers (NFTs) that use localized surface plasmon resonance (LSPR) to focus the optical field to a small sub-wavelength region have attracted attention for HAMR. Different NFTs have been proposed, such as bow-tie [10], E-antenna [8], and various types of apertures [11–13]. The reported NFT designs are exclusively based on noble metals, with gold being the most popular material platform. In this work, we report a novel bow-tie aluminum NFT design that can generate a 35 nm spot size inside the magnetic recording medium at an operating wavelength of 450 nm. This spot size corresponds to a recording density of up to 2 Tb in−2. A highly integrated micro-nano-optics design is also proposed to ensure process compatibility and stability, including corrosion-resistance of the aluminum NFT. Our work has demonstrated the feasibility of using an aluminum plasmonic device in HAMR, offering advantages of low cost and high efficiency over their traditional gold counterparts.
Figure 3. Cross-section schematic of the FDTD model (not drawn to scale, Y axis is perpendicular to the page).
process scheme is chosen. In addition to these obvious advantages, for the specific HAMR application discussed in this work, aluminum NFTs can also provide enhanced LSPR due to (a) the low screening of Al (ε∞ ≈ 1) relative to other noble metals such as Au (ε∞ ≈ 9) and Ag (ε∞ ≈ 4), and (b) its higher electron density, contributing 3 electrons per atom compared to 1 electron per atom for Au and Ag [20]. The novel bow-tie aluminum antenna design is shown in figure 2. Unlike the traditional bow-tie antenna NFT design, which has two trapezoidal metallic plates separated by a narrow gap [21], in our design we scale down the trapezoidal plates so that they can take advantage of the lightning rod effect [22, 23] to concentrate the local fields. At the same time, two larger rectangular plates are connected to the trapezoidal components. These plates function as charge reservoirs as discussed in the report of an E-antenna design by Stipe et al [8]. Detailed dimensions are provided in the schematic shown in figure 2. An important aspect of this design is that it emphasizes manufacturability, especially for the case of large-volume manufacturing. For instance, this shape is more favorable in terms of conventional photolithography processes than the E-antenna design at similar dimensions. Another example is that the 20 nm separation in the middle is also reasonably designed for fabrication: while reducing the gap would enhance the field intensity in the center, in reality a narrower separation between the two metallic plates would be more difficult to fabricate. The consideration of the overall dimensions is tightly related to the integrated optic design and hence the internal optical coupling efficiency, which will be discussed later in this paper. The NFT thickness was initially chosen to be 50 nm, and a thickness optimization method is also presented later in the paper. Three-dimensional finite-difference timedomain (FDTD) simulation was first performed on the NFT
2. Aluminum near field transducer Recently there has been increased interest in aluminum plasmonics in different applications other than HAMR [14–18]. As the third most abundant element available on earth [19], aluminum is a very competitive inexpensive plasmonic material. In most cases, there is a self-protecting passivation layer of aluminum oxide (about 3 nm thick) formed on the surface of aluminum. Hence, the stability of an aluminum plasmonic device is achievable as long as a suitable 2
L Miao et al
Nanotechnology 25 (2014) 295202 0.002
300
0.24
0.0018
0.16
y(nm)
Transmission
0.20
100
0.0016 0.0014
0.12 0.08
-100
0.0012
0.04 -300 -300
0.001
-100 100 x(nm)
300
0.00
0.0008 0.0006 0.0004 0.0002 300
350
400
450
500
550
600
650
700
750
800
lambda (nm)
Figure 5. Transmission versus wavelength plot, the inset shows the
E-field intensity for the 450 nm wavelength at 1 nm inside the Co layer overlapped with the NFT shape. Table 1. Summary of minimum spot size inside the recording
medium for different NFT thicknesses. NFT thickness (nm) Resonance wavelength (nm) FWHM_X (nm) FWHM_Y (nm)
40 400 38 43
50 450 35 35
60 500 42 41
When calculating the transmission using FDTD we put a monitor with 35 nm × 35 nm dimensions at 1 nm inside the cobalt film for data collection and integration. Note that the field intensity becomes weaker than that of the incident beam because the output field was collected inside the cobalt metal layer (the incident beam was placed in air). The field intensity inside a specific metal would mainly depend on the material properties at a given wavelength, as well as the depth into the metal where the field was collected. As mentioned above, we evaluated the field intensity at 1 nm inside the cobalt medium. This specific depth was chosen because it is a common reference depth into the recording medium (cobalt) for the application of HAMR [8, 25, 26]; hence we can have a fair comparison between this work and some other reported NFTs for use in HAMR. In fact, the penetration depth into cobalt is quite small (less than 1 nm) at the simulation wavelengths. The optimum wavelength for the strongest field confinement has a sensitive dependence on the geometry of a bow-tie NFT [25]. In this work, the X-Y dimensions were first optimized (as shown in figure 2) with the NFT thickness fixed at 50 nm. Then, additional simulations were performed at thicknesses of 40 nm and 60 nm to determine the optimum NFT thickness. The resonance wavelength shifted to 400 nm and 500 nm for 40 nm and 60 nm NFT thicknesses, respectively. The minimum spot size inside the medium at different NFT thicknesses is summarized in table 1. It is found that 50 nm is the optimum NFT thickness in order to obtain smallest spot size inside the top medium layer under this novel bow-tie antenna design.
Figure 4. Numerical FDTD simulation results: (a) E-field intensity
along X axis, 1 nm inside Co medium layer, and (b) E-field intensity along Y axis, 1 nm inside Co medium layer. Both plots have wavelength scan from 300 nm to 800 nm in steps of 50 nm.
structure to find the resonance wavelength using Lumerical FDTD Solutions [24]. The cross-section schematic of the FDTD model is shown in figure 3. The bow-tie antenna was embedded into alumina (Al2O3) for the sake of both process compatibility [9] and protection of the aluminum NFT. As mentioned earlier, the 3 nm thick surface native oxide of alumina was also considered in the model. Finally, in order to mimic HDD operation [8], the model included a 6 nm-thick air gap between the NFT structure and a 12 nm-thick cobalt layer on top of a 50 nm-thick gold layer. The cobalt layer and gold layer represent the magnetic recording medium and the heat sink, respectively. The incident Gaussian beam source has a linear polarization along the X axis (parallel to the axis of the bow-tie structure), and the wavelength was scanned from 300 nm to 800 nm in steps of 50 nm. The electric field (E-field) intensity plots at 1 nm inside the cobalt medium along X and Y directions (see figure 3 for definition of X and Y axes) at different wavelengths from the numerical FDTD simulation are summarized in figures 4(a) and (b), respectively. It can be seen that at 450 nm wavelength the central spot is most enhanced, and the smallest spot size inside the cobalt medium, or FWHM, is found to be 35 nm × 35 nm at 450 nm wavelength. This is consistent with the plot of transmission versus wavelength shown in figure 5. 3
L Miao et al
Nanotechnology 25 (2014) 295202
Table 2. Comparison of overall performance between the Al NFT and previously reported Au NFTs.
Items for comparison A1 NFT in this work Recently reported Au NFT [8] Early reported Au NFTs [25]
Device design Modified bow-tie E-antenna Circular aperture Rectangular aperture Bow-tie aperture C aperture Triangle antenna Beaked triangle Bow-tie antenna Canted bow-tie
Resonance wavelength (nm)
FWHM inside top medium layer (nm2)
Efficiency within FWHM
450
35 × 35
12.3%
19.4%
830
30 × 28
\
12.8%
650
113 × 142
\
0.14%
650
43 × 25
\
0.92%
725 700 650 725 650 750
59 × 56 34 × 39 55 × 54 43 × 41 39 × 36 31 × 36
\ \ \ \ \ \
1.7% 2.1% 1.1% 2.9% 1.4% 2.1%
Efficiency within 50 nm × 50 nm
1 Tb in−2. As previously discussed, it should also be noticed that the novel bow-tie structure actually provides better manufacturability compared with other high-efficiency NFT designs, such as the E-antenna at similar dimensions.
Besides the minimum spot size, another important performance metric of the NFT is its optical efficiency. More specifically, for HAMR applications the dissipated power within the minimum spot size inside the top medium layer is a direct indication of NFT efficiency. Similar to some previous approaches [25, 26], in this work we integrate the E-field intensity within an area defined inside the top medium layer and normalize it to that of the incident beam source to calculate the NFT efficiency from that region. However, unlike previously published works [8, 25] where the NFT efficiency was calculated within a 50 nm × 50 nm footprint, which corresponds to 1 Tb in−2 recording density, here we define the NFT efficiency within the FWHM region (35 nm × 35 nm) because this smaller spot size represents a recording density of up to 2 Tb in−2 that can be achieved based on HAMR. For the purpose of comparison, we also calculate the efficiency within the same 50 nm × 50 nm footprint. It is found that the 50 nm-thick aluminum NFT with dimensions specified in figure 2 has an efficiency of 12.3% within the FWHM region at 450 nm resonance wavelength inside the top cobalt medium layer. When expanding the region to a 50 nm × 50 nm footprint the efficiency reaches 19.4%. At this point, it is useful to compare the overall performance of this aluminum NFT with some other reported gold NFTs. This is summarized in table 2. It can be seen that this novel bow-tie aluminum NFT design offers significantly higher efficiency compared to other gold NFTs, which is of critical importance to HAMR applications. Within the same 50 nm × 50 nm footprint, or 1 Tb in−2 recording density, it offers efficiencies 51.6% higher than the E-antenna gold NFT [8], and nearly 14 times higher than the traditional bow-tie gold NFT [25]. When considering a higher recording density at 2 Tb in−2, the 35 nm × 35 nm FWHM region has an efficiency of 12.3%, which is almost the same as the E-antenna gold NFT efficiency for a 50 nm × 50 nm footprint (12.8%). Meanwhile, our novel bow-tie aluminum NFT offers comparable minimum spot size (35 nm × 35 nm) inside the top medium layer for high density magnetic recording beyond
3. Highly integrated micro-nano-optics design for HAMR In the past there have been concerns about the corrosion stability of aluminum components in HAMR [25]. In this work we propose a highly integrated micro-nano-optics design that can ensure both process compatibility and the corrosion-resistance of the aluminum NFT. The design is based on the integration of a previously reported on-wafer microfocal lens [9] and the bow-tie aluminum nano-antenna described above. Figure 6 illustrates the integration scheme. The microfocal lens couples the external laser beam and focuses it onto the aluminum NFT, which is placed at the center of the lens focal plane. The aluminum NFT then uses LSPR to deliver a highly focused spot to the recording medium layer. Similar to the FDTD model shown in figure 3, the aluminum NFT is completely embedded in the microfocal lens body, which is bulk alumina (Al2O3). Thus the aluminum NFT is isolated from its surroundings to avoid corrosion during operation. The integrated design can be achieved using conventional fabrication methods including thin-film deposition, photolithography, ion milling, and reactive ion etching (RIE). A study by Miao and Hsiang has demonstrated that the on-wafer microfocal lens design is robust in terms of manufacturing and performance [27]. Hence, the entire design could provide a highly manufacturable low-cost solution that is compatible with conventional industry fabrication processes. Three-dimensional FDTD simulation was again performed, based on the integrated model shown in figure 6. The microfocal lens works on the principle of first-order 4
L Miao et al
Nanotechnology 25 (2014) 295202
efficiency of 4.3% within this FWHM region (table 4). Referring to the data listed in table 2, it can be seen that the proposed integrated design provides very competitive efficiency without compromising in spot size, even compared with isolated gold NFTs that were previously reported.
4. Discussion and conclusion In order to analyze the simulation results for the integrated microfocal lens and transducer, it is helpful to first decouple the microfocal lens from the aluminum NFT and check the results from the isolated lens. Figure 7 shows the far-field projection near the lens focal plane, together with E-field intensity plots along the X and Y directions on the focal plane, and a summary of key data obtained from the simulation. Because the microfocal lens is circularly symmetric, different lens regions will see different incident polarizations given a linearly polarized incident beam [27]. Hence, the lens focal spot is no longer a perfect Gaussian shape. As summarized in figure 7(d), the focal plane central peak spot is 0.549 μm along the X axis and 0.484 μm along the Y axis. Similarly, the FWHM is 0.224 μm and 0.208 μm in the X and Y directions, respectively. Considering this focal spot distortion from the microfocal lens, it is understood that the final optical spot inside the top medium layer delivered by the integrated design is also distorted because the incident beam profile seen by the NFT is no longer a perfect Gaussian shape as assumed in the isolated NFT simulation. The integrated design delivers a spot of 36 nm × 40 nm, which is slightly different from the 35 nm × 35 nm result obtained from the simulation of the isolated NFT (table 1). The percentage difference between X and Y axes for the lens focal plane spot size (11.8%) agrees well with that for the final integrated optical spot size (10%). Based on the FDTD results, the external coupling efficiency is 40.9% for the isolated microfocal lens at 450 nm wavelength. The isolated NFT efficiency (within FWHM) is 12.3% at the same operation wavelength. Thus theoretically the integrated system would have a total efficiency of 5%. The simulation results from the integrated design give an overall efficiency of 4.3% (table 4). Considering the NFT dimensions as shown in figure 2, it has an area span of 520 nm × 250 nm. The lens focal spot has an area span of 549 nm × 484 nm with FWHM at 224 nm × 208 nm. Hence, most of the focal-spot intensity is delivered onto the NFT region. A rough estimate taking into consideration both area and intensity factors shows about 87% internal optical coupling efficiency between the lens and the NFT. Thus, the theoretical system efficiency is adjusted to be about 4.4%. This is very close to the 4.3% total efficiency obtained from the integrated design simulation. It is important to consider this internal optical-coupling efficiency when designing the overall dimension of the NFT, as mentioned earlier in section 2. Finally, it needs to be pointed out that due to the complexity of its shape, the reported novel bow-tie aluminum NFT may not achieve its fully optimized dimensions. Nevertheless, our work has demonstrated the feasibility of
Figure 6. Schematic of the proposed micro-nano-optics design for
use in HAMR.
Table 3. Key parameters of the microfocal lens used in this work.
Operation wavelength Lens thickness (f) Lens step height (h) Lens outer radius (R)
450 nm 40 μm 0.288 μm 40.229 μm
Table 4. Integrated design simulation results.
Resonance wavelength (nm)
Overall efficiency
FWHM inside top medium layer (nm2)
450
4.3%
36 × 40
diffractive optics, and the detailed design can be found in [9]. Table 3 summarizes key parameters of the microfocal lens used in this study. The NFT dimensions follow the optimum design discussed above, and the air gap and metallic layers are as described in figure 3. The simulation results based on this integrated design show a final-delivered spot inside the top medium layer of 36 nm × 40 nm at 450 nm wavelength, with an overall 5
L Miao et al
Nanotechnology 25 (2014) 295202 3500
x10
(b) 6.6
3000
(a)
3.0 2.5
2500 5.7
Z (um)
I E I ^2 along X
3
2000 1500
2.0 1.5
4.8
1.0
1000
3.9
0.5
500 0 -0.0
-0.4
-0.2
0
x (um)
0.2
0.4
3.0 -1.0
0.0
0.0 -0.5
-0.0
0.5
x (um)
1.0
3500
I E I ^2 along Y
(c) 3000
(d)
2500
Operation wavelength (nm)
2000 1500 1000
450
Focal spot
FWHM_X
0.224
FWHM (µm)
FWHM_Y
0.208
Central peak
Spot_X
0.549
spot size (µm)
Spot_Y
Lens external coupling efficiency
0.484 40.9%
500 0 -0.0
-0.4
-0.2
0
y (um)
0.2
0.4
0.0
Figure 7. Simulation results for an isolated microfocal lens: (a) far-field projection near the lens focal plane, (b) E-field intensity along the X axis, (c) E-field intensity along the Y axis, and (d) summary of key data obtained from the FDTD simulation.
using an aluminum plasmonic device in HAMR, which offers advantages of low cost and high efficiency over traditional gold counterparts. The reported aluminum NFT provides a competitive spot size of 35 nm with a much higher efficiency of 12.3% compared with previously reported gold NFTs. We also propose a highly integrated micro-nano-optics design that can ensure process compatibility and corrosion-resistance of the aluminum NFT. The integrated system can deliver a spot size of 36 nm × 40 nm inside the recording medium with an overall efficiency of 4.3%. Our work indicates that aluminum can be a promising inexpensive plasmonic material for the development of HAMR techniques, thereby providing a solution for the next generation of ultra-high density magnetic recording at or above 2 Tb in−2.
Lorenzo Rosa from Centre for Micro-Photonics at Swinburne University of Technology for useful discussions on FDTD simulation skills.
References [1] Consumer Electronics Industry: Market Research Reports, Statistics and Analysis, available online: http://reportlinker. com/ci02060/Consumer-Electronics.html [2] Kasavajhala V 2011 Solid State Drive vs Hard Disk Drive Price and Performance Study Dell Technical White Paper [3] Fontana R E Jr, Hetzler S R and Decad G 2012 Technology roadmap comparisons for TAPE, HDD, and NAND flash: implications for data storage applications IEEE Trans. Magn. 48 1692–6 [4] Weller D and Moser A 1999 Thermal effect limits in ultrahigh density magnetic recording IEEE Trans. Magn. 35 4423–39 [5] McDaniel T W, Challener W A and Sendur K 2003 Issues in heat-assisted perpendicular recording IEEE Trans. Magn. 39 1972–9 [6] Rottmayer R E et al 2006 Heat-assisted magnetic recording IEEE Trans. Magn. 42 2417–21 [7] Seigler M A et al 2008 Integrated heat assisted magnetic recording head: design and recording demonstration IEEE Trans. Magn. 44 119–24
Acknowledgments This material is based upon work supported by the National Science Foundation under Grant No. 1316918. One of the authors, Lingyun Miao, would like to thank Professor Saulius Juodkazis for his permission of running FDTD simulation on the workstation in the Centre for MicroPhotonics at Swinburne University of Technology, and 6
L Miao et al
Nanotechnology 25 (2014) 295202
[18] Jha S K, Ahmed Z, Agio M, Ekinci Y and Löffler J F 2012 J. Am. Chem. Soc. 134 1966–9 [19] USGS Commodity Statistics and Information 2011 available online: http://minerals.usgs.gov/minerals/pubs/commodity/ (accessed May 27) [20] Knight M W, Liu L, Wang Y, Brown L, Mukherjee S, King N S, Everitt H O, Nordlander P and Halas N J 2012 Aluminum plasmonic nanoantennas Nano Lett. 12 6000–4 [21] Grober R D, Schoelkopf R J and Prober D E 1997 Optical antenna: towards a unity efficiency near-field optical probe Appl. Phys. Lett. 70 1354–6 [22] Gersten J and Nitzan A 1980 J. Chem. Phys. 73 3023 [23] Liao P F and Woakun A 1982 J. Chem. Phys. 76 751 [24] Lumerical FDTD Solutions http://lumerical.com [25] Challener W A and Itagi A V 2009 Modelling and numerical simulations II Modern Aspects of Electrochemistry ed M Schlesinger (New York: Springer) [26] Miao L and Hsiang T Y 2014 Tapered waveguide design for heat-assisted magnetic recording applications IEEE Trans. Magn. 50 3100207 [27] Miao L and Hsiang T Y 2014 Feasibility analysis of on-wafer microfocal lens for optical coupling in heat-assisted magnetic recording systems IEEE Trans. Magn. 50
[8] Stipe B C et al 2010 Magnetic recording at 1.5 Pb m-2 using an integrated plasmonic antenna Nat. Photonics 4 484–8 [9] Miao L and Hsiang T Y 2012 Microfocal lens for energyassisted magnetic recording technology Micro & Nano Letters 7 1005–7 [10] Grober R, Bukofsky S and Selberg S 1997 Application of nearfield optics to critical dimension metrology Appl. Phys. Lett. 70 2368–70 [11] Shi X and Hesselink J 2002 Mechanisms for enhancing power throughput from planar nano-apertures for near-field optical data storage J. Appl. Phys. 41 1632–5 [12] Yin L et al 2004 Surface plasmons at single nanoholes in Au films Appl. Phys. Lett. 85 467–9 [13] Popov E et al 2005 Surface plasmon excitation on a single subwavelength hole in a metallic sheet Appl. Opt. 44 2332–7 [14] Zoric I, Zach M, Kasemo B and Langhammer C 2011 ACS Nano 5 2535–46 [15] Castro-Lopez M, Brinks D, Sapienza R and Van Hulst N F 2011 Nano Lett. 11 4674–8 [16] Andersson T, Zhang C F, Tchaplyguine M, Svensson S, Martensson N and Bjorneholm O J 2012 Chem. Phys. 136 204504 [17] Taguchi A, Saito Y, Watanabe K, Yijian S and Kawata S 2012 Appl. Phys. Lett. 101 081110
7
Search
Collections
Journals
About
Contact us
My IOPscience
Novel aluminum near field transducer and highly integrated micro-nano-optics design for heatassisted ultra-high-density magnetic recording
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 295202 (http://iopscience.iop.org/0957-4484/25/29/295202) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 134.117.10.200 This content was downloaded on 13/05/2015 at 14:43
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 459501 (1pp)
doi:10.1088/0957-4484/25/45/459501
Corrigendum: Novel aluminum near field transducer and highly integrated micronano-optics design for heat-assisted ultrahigh-density magnetic recording (2014 Nanotechnology 25 295202) Lingyun Miao1, Paul R Stoddart2 and Thomas Y Hsiang1 1
Department of Electrical and Computer Engineering, University of Rochester, Rochester, NY 14627, USA 2 Centre for Quantum and Optical Science, Faculty of Science, Engineering and Technology, Swinburne, University of Technology, Hawthorn, VIC 3122, Australia E-mail: [email protected] Received 29 September 2014 Accepted for publication 29 September 2014 Published 20 October 2014
should be ‘In fact, the penetration depth into cobalt is quite small (less than 11 nm) at the optimum wavelength of 450 nm.’ No other typo was found in the published paper.
There was an error made on page 3 of the original manuscript, it reads ‘In fact, the penetration depth into cobalt is quite small (less than 1 nm) at the simulation wavelengths.’ It
0957-4484/14/459501+01$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
Nanotechnology Nanotechnology 25 (2014) 295202 (7pp)
doi:10.1088/0957-4484/25/29/295202
Novel aluminum near field transducer and highly integrated micro-nano-optics design for heat-assisted ultra-high-density magnetic recording Lingyun Miao1, Paul R Stoddart2 and Thomas Y Hsiang1 1
Department of Electrical and Computer Engineering, University of Rochester, Rochester, NY 14627, USA 2 Centre for Quantum and Optical Science, Faculty of Science, Engineering and Technology, Swinburne University of Technology, Hawthorn, VIC 3122, Australia E-mail: [email protected] Received 10 December 2013, revised 30 May 2014 Accepted for publication 6 June 2014 Published 1 July 2014 Abstract
Heat-assisted magnetic recording (HAMR) has attracted increasing attention as one of the most promising future techniques for ultra-high-density magnetic recording beyond the current limit of 1 Tb in−2. Localized surface plasmon resonance plays an important role in HAMR by providing a highly focused optical spot for heating the recording medium within a small volume. In this work, we report an aluminum near-field transducer (NFT) based on a novel bow-tie design. At an operating wavelength of 450 nm, the proposed transducer can generate a 35 nm spot size inside the magnetic recording medium, corresponding to a recording density of up to 2 Tb in−2. A highly integrated micro-nano-optics design is also proposed to ensure process compatibility and corrosion-resistance of the aluminum NFT. Our work has demonstrated the feasibility of using aluminum as a plasmonic material for HAMR, with advantages of reduced cost and improved efficiency compared to traditional noble metals. S Online supplementary data available from stacks.iop.org/NANO/25/295202/mmedia Keywords: aluminum, bow-tie, near field transducer, heat-assisted magnetic recording (Some figures may appear in colour only in the online journal) 1. Introduction
The biggest challenge for the HDD industry is that conventional perpendicular magnetic recording is about to reach its limit of recording density at 1 Tb in−2 [4, 5]. Among several promising future technologies such as bit-patterned media, heat-assisted magnetic recording (HAMR), and shingled magnetic recording, HAMR is considered one of the most flexible techniques that can be implemented in the conventional HDD industry for extending areal density [6–8]. HAMR utilizes a focused laser beam to define magnetic recording features. As illustrated in figure 1 [9], the recording density under HAMR is determined by the thermomagnetic recording mechanism. In general, a special recording medium which has good thermal stability at room temperature is used
A significant part of the world’s economy has been related to the IT and consumer electronics markets in recent years [1], and digital data is a key contributor to its growth. Hard disk drives (HDDs) have dominated the digital data storage market as the primary non-volatile memory device, despite strong competition from solid-state drives (SSDs) [2]. Surprisingly, however, a recent technology roadmap analysis shows that the older generation tape-based magnetic recording (TAPE) would probably outperform both HDD and SSD in the next six years in terms of areal density extendibility [3], provided that there is no breakthrough in the latter technologies. 0957-4484/14/295202+07$33.00
1
© 2014 IOP Publishing Ltd Printed in the UK
L Miao et al
Nanotechnology 25 (2014) 295202
Figure 2. Novel bow-tie aluminum antenna design (top view, not
drawn to scale, unit: nm). Figure 1. The use of a focused laser beam to define magnetic
recording features under the HAMR scheme (not drawn to scale). Reprinted with permission from [9]. Copyright 2012 The Institution of Engineering and Technology. The inset shows a HDD with a conventional read/write head over the recording surface.
in HAMR. When a focused laser beam is delivered onto the medium’s surface, the local medium is temporarily heated up to exceed its Curie temperature and lower its coercivity to allow magnetic recording. When the local medium cools down, it becomes stable again to retain data. Typically a laser power of about several tens of milliwatt is necessary to make this scheme work. In order to achieve ultra-high-density magnetic recording, it is necessary to deliver a highly focused laser beam onto the surface of the medium. For instance, an optical spot size of 50 nm or less is necessary for a recording density of 1 Tb in−2 [6]. Thus, metallic near-field transducers (NFTs) that use localized surface plasmon resonance (LSPR) to focus the optical field to a small sub-wavelength region have attracted attention for HAMR. Different NFTs have been proposed, such as bow-tie [10], E-antenna [8], and various types of apertures [11–13]. The reported NFT designs are exclusively based on noble metals, with gold being the most popular material platform. In this work, we report a novel bow-tie aluminum NFT design that can generate a 35 nm spot size inside the magnetic recording medium at an operating wavelength of 450 nm. This spot size corresponds to a recording density of up to 2 Tb in−2. A highly integrated micro-nano-optics design is also proposed to ensure process compatibility and stability, including corrosion-resistance of the aluminum NFT. Our work has demonstrated the feasibility of using an aluminum plasmonic device in HAMR, offering advantages of low cost and high efficiency over their traditional gold counterparts.
Figure 3. Cross-section schematic of the FDTD model (not drawn to scale, Y axis is perpendicular to the page).
process scheme is chosen. In addition to these obvious advantages, for the specific HAMR application discussed in this work, aluminum NFTs can also provide enhanced LSPR due to (a) the low screening of Al (ε∞ ≈ 1) relative to other noble metals such as Au (ε∞ ≈ 9) and Ag (ε∞ ≈ 4), and (b) its higher electron density, contributing 3 electrons per atom compared to 1 electron per atom for Au and Ag [20]. The novel bow-tie aluminum antenna design is shown in figure 2. Unlike the traditional bow-tie antenna NFT design, which has two trapezoidal metallic plates separated by a narrow gap [21], in our design we scale down the trapezoidal plates so that they can take advantage of the lightning rod effect [22, 23] to concentrate the local fields. At the same time, two larger rectangular plates are connected to the trapezoidal components. These plates function as charge reservoirs as discussed in the report of an E-antenna design by Stipe et al [8]. Detailed dimensions are provided in the schematic shown in figure 2. An important aspect of this design is that it emphasizes manufacturability, especially for the case of large-volume manufacturing. For instance, this shape is more favorable in terms of conventional photolithography processes than the E-antenna design at similar dimensions. Another example is that the 20 nm separation in the middle is also reasonably designed for fabrication: while reducing the gap would enhance the field intensity in the center, in reality a narrower separation between the two metallic plates would be more difficult to fabricate. The consideration of the overall dimensions is tightly related to the integrated optic design and hence the internal optical coupling efficiency, which will be discussed later in this paper. The NFT thickness was initially chosen to be 50 nm, and a thickness optimization method is also presented later in the paper. Three-dimensional finite-difference timedomain (FDTD) simulation was first performed on the NFT
2. Aluminum near field transducer Recently there has been increased interest in aluminum plasmonics in different applications other than HAMR [14–18]. As the third most abundant element available on earth [19], aluminum is a very competitive inexpensive plasmonic material. In most cases, there is a self-protecting passivation layer of aluminum oxide (about 3 nm thick) formed on the surface of aluminum. Hence, the stability of an aluminum plasmonic device is achievable as long as a suitable 2
L Miao et al
Nanotechnology 25 (2014) 295202 0.002
300
0.24
0.0018
0.16
y(nm)
Transmission
0.20
100
0.0016 0.0014
0.12 0.08
-100
0.0012
0.04 -300 -300
0.001
-100 100 x(nm)
300
0.00
0.0008 0.0006 0.0004 0.0002 300
350
400
450
500
550
600
650
700
750
800
lambda (nm)
Figure 5. Transmission versus wavelength plot, the inset shows the
E-field intensity for the 450 nm wavelength at 1 nm inside the Co layer overlapped with the NFT shape. Table 1. Summary of minimum spot size inside the recording
medium for different NFT thicknesses. NFT thickness (nm) Resonance wavelength (nm) FWHM_X (nm) FWHM_Y (nm)
40 400 38 43
50 450 35 35
60 500 42 41
When calculating the transmission using FDTD we put a monitor with 35 nm × 35 nm dimensions at 1 nm inside the cobalt film for data collection and integration. Note that the field intensity becomes weaker than that of the incident beam because the output field was collected inside the cobalt metal layer (the incident beam was placed in air). The field intensity inside a specific metal would mainly depend on the material properties at a given wavelength, as well as the depth into the metal where the field was collected. As mentioned above, we evaluated the field intensity at 1 nm inside the cobalt medium. This specific depth was chosen because it is a common reference depth into the recording medium (cobalt) for the application of HAMR [8, 25, 26]; hence we can have a fair comparison between this work and some other reported NFTs for use in HAMR. In fact, the penetration depth into cobalt is quite small (less than 1 nm) at the simulation wavelengths. The optimum wavelength for the strongest field confinement has a sensitive dependence on the geometry of a bow-tie NFT [25]. In this work, the X-Y dimensions were first optimized (as shown in figure 2) with the NFT thickness fixed at 50 nm. Then, additional simulations were performed at thicknesses of 40 nm and 60 nm to determine the optimum NFT thickness. The resonance wavelength shifted to 400 nm and 500 nm for 40 nm and 60 nm NFT thicknesses, respectively. The minimum spot size inside the medium at different NFT thicknesses is summarized in table 1. It is found that 50 nm is the optimum NFT thickness in order to obtain smallest spot size inside the top medium layer under this novel bow-tie antenna design.
Figure 4. Numerical FDTD simulation results: (a) E-field intensity
along X axis, 1 nm inside Co medium layer, and (b) E-field intensity along Y axis, 1 nm inside Co medium layer. Both plots have wavelength scan from 300 nm to 800 nm in steps of 50 nm.
structure to find the resonance wavelength using Lumerical FDTD Solutions [24]. The cross-section schematic of the FDTD model is shown in figure 3. The bow-tie antenna was embedded into alumina (Al2O3) for the sake of both process compatibility [9] and protection of the aluminum NFT. As mentioned earlier, the 3 nm thick surface native oxide of alumina was also considered in the model. Finally, in order to mimic HDD operation [8], the model included a 6 nm-thick air gap between the NFT structure and a 12 nm-thick cobalt layer on top of a 50 nm-thick gold layer. The cobalt layer and gold layer represent the magnetic recording medium and the heat sink, respectively. The incident Gaussian beam source has a linear polarization along the X axis (parallel to the axis of the bow-tie structure), and the wavelength was scanned from 300 nm to 800 nm in steps of 50 nm. The electric field (E-field) intensity plots at 1 nm inside the cobalt medium along X and Y directions (see figure 3 for definition of X and Y axes) at different wavelengths from the numerical FDTD simulation are summarized in figures 4(a) and (b), respectively. It can be seen that at 450 nm wavelength the central spot is most enhanced, and the smallest spot size inside the cobalt medium, or FWHM, is found to be 35 nm × 35 nm at 450 nm wavelength. This is consistent with the plot of transmission versus wavelength shown in figure 5. 3
L Miao et al
Nanotechnology 25 (2014) 295202
Table 2. Comparison of overall performance between the Al NFT and previously reported Au NFTs.
Items for comparison A1 NFT in this work Recently reported Au NFT [8] Early reported Au NFTs [25]
Device design Modified bow-tie E-antenna Circular aperture Rectangular aperture Bow-tie aperture C aperture Triangle antenna Beaked triangle Bow-tie antenna Canted bow-tie
Resonance wavelength (nm)
FWHM inside top medium layer (nm2)
Efficiency within FWHM
450
35 × 35
12.3%
19.4%
830
30 × 28
\
12.8%
650
113 × 142
\
0.14%
650
43 × 25
\
0.92%
725 700 650 725 650 750
59 × 56 34 × 39 55 × 54 43 × 41 39 × 36 31 × 36
\ \ \ \ \ \
1.7% 2.1% 1.1% 2.9% 1.4% 2.1%
Efficiency within 50 nm × 50 nm
1 Tb in−2. As previously discussed, it should also be noticed that the novel bow-tie structure actually provides better manufacturability compared with other high-efficiency NFT designs, such as the E-antenna at similar dimensions.
Besides the minimum spot size, another important performance metric of the NFT is its optical efficiency. More specifically, for HAMR applications the dissipated power within the minimum spot size inside the top medium layer is a direct indication of NFT efficiency. Similar to some previous approaches [25, 26], in this work we integrate the E-field intensity within an area defined inside the top medium layer and normalize it to that of the incident beam source to calculate the NFT efficiency from that region. However, unlike previously published works [8, 25] where the NFT efficiency was calculated within a 50 nm × 50 nm footprint, which corresponds to 1 Tb in−2 recording density, here we define the NFT efficiency within the FWHM region (35 nm × 35 nm) because this smaller spot size represents a recording density of up to 2 Tb in−2 that can be achieved based on HAMR. For the purpose of comparison, we also calculate the efficiency within the same 50 nm × 50 nm footprint. It is found that the 50 nm-thick aluminum NFT with dimensions specified in figure 2 has an efficiency of 12.3% within the FWHM region at 450 nm resonance wavelength inside the top cobalt medium layer. When expanding the region to a 50 nm × 50 nm footprint the efficiency reaches 19.4%. At this point, it is useful to compare the overall performance of this aluminum NFT with some other reported gold NFTs. This is summarized in table 2. It can be seen that this novel bow-tie aluminum NFT design offers significantly higher efficiency compared to other gold NFTs, which is of critical importance to HAMR applications. Within the same 50 nm × 50 nm footprint, or 1 Tb in−2 recording density, it offers efficiencies 51.6% higher than the E-antenna gold NFT [8], and nearly 14 times higher than the traditional bow-tie gold NFT [25]. When considering a higher recording density at 2 Tb in−2, the 35 nm × 35 nm FWHM region has an efficiency of 12.3%, which is almost the same as the E-antenna gold NFT efficiency for a 50 nm × 50 nm footprint (12.8%). Meanwhile, our novel bow-tie aluminum NFT offers comparable minimum spot size (35 nm × 35 nm) inside the top medium layer for high density magnetic recording beyond
3. Highly integrated micro-nano-optics design for HAMR In the past there have been concerns about the corrosion stability of aluminum components in HAMR [25]. In this work we propose a highly integrated micro-nano-optics design that can ensure both process compatibility and the corrosion-resistance of the aluminum NFT. The design is based on the integration of a previously reported on-wafer microfocal lens [9] and the bow-tie aluminum nano-antenna described above. Figure 6 illustrates the integration scheme. The microfocal lens couples the external laser beam and focuses it onto the aluminum NFT, which is placed at the center of the lens focal plane. The aluminum NFT then uses LSPR to deliver a highly focused spot to the recording medium layer. Similar to the FDTD model shown in figure 3, the aluminum NFT is completely embedded in the microfocal lens body, which is bulk alumina (Al2O3). Thus the aluminum NFT is isolated from its surroundings to avoid corrosion during operation. The integrated design can be achieved using conventional fabrication methods including thin-film deposition, photolithography, ion milling, and reactive ion etching (RIE). A study by Miao and Hsiang has demonstrated that the on-wafer microfocal lens design is robust in terms of manufacturing and performance [27]. Hence, the entire design could provide a highly manufacturable low-cost solution that is compatible with conventional industry fabrication processes. Three-dimensional FDTD simulation was again performed, based on the integrated model shown in figure 6. The microfocal lens works on the principle of first-order 4
L Miao et al
Nanotechnology 25 (2014) 295202
efficiency of 4.3% within this FWHM region (table 4). Referring to the data listed in table 2, it can be seen that the proposed integrated design provides very competitive efficiency without compromising in spot size, even compared with isolated gold NFTs that were previously reported.
4. Discussion and conclusion In order to analyze the simulation results for the integrated microfocal lens and transducer, it is helpful to first decouple the microfocal lens from the aluminum NFT and check the results from the isolated lens. Figure 7 shows the far-field projection near the lens focal plane, together with E-field intensity plots along the X and Y directions on the focal plane, and a summary of key data obtained from the simulation. Because the microfocal lens is circularly symmetric, different lens regions will see different incident polarizations given a linearly polarized incident beam [27]. Hence, the lens focal spot is no longer a perfect Gaussian shape. As summarized in figure 7(d), the focal plane central peak spot is 0.549 μm along the X axis and 0.484 μm along the Y axis. Similarly, the FWHM is 0.224 μm and 0.208 μm in the X and Y directions, respectively. Considering this focal spot distortion from the microfocal lens, it is understood that the final optical spot inside the top medium layer delivered by the integrated design is also distorted because the incident beam profile seen by the NFT is no longer a perfect Gaussian shape as assumed in the isolated NFT simulation. The integrated design delivers a spot of 36 nm × 40 nm, which is slightly different from the 35 nm × 35 nm result obtained from the simulation of the isolated NFT (table 1). The percentage difference between X and Y axes for the lens focal plane spot size (11.8%) agrees well with that for the final integrated optical spot size (10%). Based on the FDTD results, the external coupling efficiency is 40.9% for the isolated microfocal lens at 450 nm wavelength. The isolated NFT efficiency (within FWHM) is 12.3% at the same operation wavelength. Thus theoretically the integrated system would have a total efficiency of 5%. The simulation results from the integrated design give an overall efficiency of 4.3% (table 4). Considering the NFT dimensions as shown in figure 2, it has an area span of 520 nm × 250 nm. The lens focal spot has an area span of 549 nm × 484 nm with FWHM at 224 nm × 208 nm. Hence, most of the focal-spot intensity is delivered onto the NFT region. A rough estimate taking into consideration both area and intensity factors shows about 87% internal optical coupling efficiency between the lens and the NFT. Thus, the theoretical system efficiency is adjusted to be about 4.4%. This is very close to the 4.3% total efficiency obtained from the integrated design simulation. It is important to consider this internal optical-coupling efficiency when designing the overall dimension of the NFT, as mentioned earlier in section 2. Finally, it needs to be pointed out that due to the complexity of its shape, the reported novel bow-tie aluminum NFT may not achieve its fully optimized dimensions. Nevertheless, our work has demonstrated the feasibility of
Figure 6. Schematic of the proposed micro-nano-optics design for
use in HAMR.
Table 3. Key parameters of the microfocal lens used in this work.
Operation wavelength Lens thickness (f) Lens step height (h) Lens outer radius (R)
450 nm 40 μm 0.288 μm 40.229 μm
Table 4. Integrated design simulation results.
Resonance wavelength (nm)
Overall efficiency
FWHM inside top medium layer (nm2)
450
4.3%
36 × 40
diffractive optics, and the detailed design can be found in [9]. Table 3 summarizes key parameters of the microfocal lens used in this study. The NFT dimensions follow the optimum design discussed above, and the air gap and metallic layers are as described in figure 3. The simulation results based on this integrated design show a final-delivered spot inside the top medium layer of 36 nm × 40 nm at 450 nm wavelength, with an overall 5
L Miao et al
Nanotechnology 25 (2014) 295202 3500
x10
(b) 6.6
3000
(a)
3.0 2.5
2500 5.7
Z (um)
I E I ^2 along X
3
2000 1500
2.0 1.5
4.8
1.0
1000
3.9
0.5
500 0 -0.0
-0.4
-0.2
0
x (um)
0.2
0.4
3.0 -1.0
0.0
0.0 -0.5
-0.0
0.5
x (um)
1.0
3500
I E I ^2 along Y
(c) 3000
(d)
2500
Operation wavelength (nm)
2000 1500 1000
450
Focal spot
FWHM_X
0.224
FWHM (µm)
FWHM_Y
0.208
Central peak
Spot_X
0.549
spot size (µm)
Spot_Y
Lens external coupling efficiency
0.484 40.9%
500 0 -0.0
-0.4
-0.2
0
y (um)
0.2
0.4
0.0
Figure 7. Simulation results for an isolated microfocal lens: (a) far-field projection near the lens focal plane, (b) E-field intensity along the X axis, (c) E-field intensity along the Y axis, and (d) summary of key data obtained from the FDTD simulation.
using an aluminum plasmonic device in HAMR, which offers advantages of low cost and high efficiency over traditional gold counterparts. The reported aluminum NFT provides a competitive spot size of 35 nm with a much higher efficiency of 12.3% compared with previously reported gold NFTs. We also propose a highly integrated micro-nano-optics design that can ensure process compatibility and corrosion-resistance of the aluminum NFT. The integrated system can deliver a spot size of 36 nm × 40 nm inside the recording medium with an overall efficiency of 4.3%. Our work indicates that aluminum can be a promising inexpensive plasmonic material for the development of HAMR techniques, thereby providing a solution for the next generation of ultra-high density magnetic recording at or above 2 Tb in−2.
Lorenzo Rosa from Centre for Micro-Photonics at Swinburne University of Technology for useful discussions on FDTD simulation skills.
References [1] Consumer Electronics Industry: Market Research Reports, Statistics and Analysis, available online: http://reportlinker. com/ci02060/Consumer-Electronics.html [2] Kasavajhala V 2011 Solid State Drive vs Hard Disk Drive Price and Performance Study Dell Technical White Paper [3] Fontana R E Jr, Hetzler S R and Decad G 2012 Technology roadmap comparisons for TAPE, HDD, and NAND flash: implications for data storage applications IEEE Trans. Magn. 48 1692–6 [4] Weller D and Moser A 1999 Thermal effect limits in ultrahigh density magnetic recording IEEE Trans. Magn. 35 4423–39 [5] McDaniel T W, Challener W A and Sendur K 2003 Issues in heat-assisted perpendicular recording IEEE Trans. Magn. 39 1972–9 [6] Rottmayer R E et al 2006 Heat-assisted magnetic recording IEEE Trans. Magn. 42 2417–21 [7] Seigler M A et al 2008 Integrated heat assisted magnetic recording head: design and recording demonstration IEEE Trans. Magn. 44 119–24
Acknowledgments This material is based upon work supported by the National Science Foundation under Grant No. 1316918. One of the authors, Lingyun Miao, would like to thank Professor Saulius Juodkazis for his permission of running FDTD simulation on the workstation in the Centre for MicroPhotonics at Swinburne University of Technology, and 6
L Miao et al
Nanotechnology 25 (2014) 295202
[18] Jha S K, Ahmed Z, Agio M, Ekinci Y and Löffler J F 2012 J. Am. Chem. Soc. 134 1966–9 [19] USGS Commodity Statistics and Information 2011 available online: http://minerals.usgs.gov/minerals/pubs/commodity/ (accessed May 27) [20] Knight M W, Liu L, Wang Y, Brown L, Mukherjee S, King N S, Everitt H O, Nordlander P and Halas N J 2012 Aluminum plasmonic nanoantennas Nano Lett. 12 6000–4 [21] Grober R D, Schoelkopf R J and Prober D E 1997 Optical antenna: towards a unity efficiency near-field optical probe Appl. Phys. Lett. 70 1354–6 [22] Gersten J and Nitzan A 1980 J. Chem. Phys. 73 3023 [23] Liao P F and Woakun A 1982 J. Chem. Phys. 76 751 [24] Lumerical FDTD Solutions http://lumerical.com [25] Challener W A and Itagi A V 2009 Modelling and numerical simulations II Modern Aspects of Electrochemistry ed M Schlesinger (New York: Springer) [26] Miao L and Hsiang T Y 2014 Tapered waveguide design for heat-assisted magnetic recording applications IEEE Trans. Magn. 50 3100207 [27] Miao L and Hsiang T Y 2014 Feasibility analysis of on-wafer microfocal lens for optical coupling in heat-assisted magnetic recording systems IEEE Trans. Magn. 50
[8] Stipe B C et al 2010 Magnetic recording at 1.5 Pb m-2 using an integrated plasmonic antenna Nat. Photonics 4 484–8 [9] Miao L and Hsiang T Y 2012 Microfocal lens for energyassisted magnetic recording technology Micro & Nano Letters 7 1005–7 [10] Grober R, Bukofsky S and Selberg S 1997 Application of nearfield optics to critical dimension metrology Appl. Phys. Lett. 70 2368–70 [11] Shi X and Hesselink J 2002 Mechanisms for enhancing power throughput from planar nano-apertures for near-field optical data storage J. Appl. Phys. 41 1632–5 [12] Yin L et al 2004 Surface plasmons at single nanoholes in Au films Appl. Phys. Lett. 85 467–9 [13] Popov E et al 2005 Surface plasmon excitation on a single subwavelength hole in a metallic sheet Appl. Opt. 44 2332–7 [14] Zoric I, Zach M, Kasemo B and Langhammer C 2011 ACS Nano 5 2535–46 [15] Castro-Lopez M, Brinks D, Sapienza R and Van Hulst N F 2011 Nano Lett. 11 4674–8 [16] Andersson T, Zhang C F, Tchaplyguine M, Svensson S, Martensson N and Bjorneholm O J 2012 Chem. Phys. 136 204504 [17] Taguchi A, Saito Y, Watanabe K, Yijian S and Kawata S 2012 Appl. Phys. Lett. 101 081110
7
