A single-pixel wireless contact lens display

3 downloads 0 Views 1MB Size Report
Jan 3, 2016 - design and fabrication of micro-Fresnel lenses on a contact lens are ... from vertical cavity surface emitting lasers could be used to .... IEEE standard and a recent ICNIRP review [21] point toward .... contained a small hole drilled through the center and exiting ... electrical charging during EBL. ... ∼8 μC cm.
Home

Search

Collections

Journals

About

Contact us

My IOPscience

A single-pixel wireless contact lens display

This content has been downloaded from IOPscience. Please scroll down to see the full text. 2011 J. Micromech. Microeng. 21 125014 (http://iopscience.iop.org/0960-1317/21/12/125014) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 129.15.109.254 This content was downloaded on 01/03/2016 at 07:07

Please note that terms and conditions apply.

IOP PUBLISHING

JOURNAL OF MICROMECHANICS AND MICROENGINEERING

doi:10.1088/0960-1317/21/12/125014

J. Micromech. Microeng. 21 (2011) 125014 (8pp)

A single-pixel wireless contact lens display A R Lingley1 , M Ali2 , Y Liao1 , R Mirjalili1 , M Klonner2 , M Sopanen2 , S Suihkonen2 , T Shen3 , B P Otis1 , H Lipsanen2 and B A Parviz1 1

Department of Electrical Engineering, University of Washington, 185 Stevens Way, Seattle, WA 98195-2500, USA 2 Department of Micro- and Nanosciences, Aalto University, Tietotie 3, 02150 Espoo, Finland 3 Department of Ophthalmology, University of Washington, 325 9th Ave., Seattle, WA 98104, USA E-mail: [email protected]

Received 9 June 2011, in final form 19 September 2011 Published 22 November 2011 Online at stacks.iop.org/JMM/21/125014 Abstract We present the design, construction and in vivo rabbit testing of a wirelessly powered contact lens display. The display consists of an antenna, a 500 × 500 μm2 silicon power harvesting and radio integrated circuit, metal interconnects, insulation layers and a 750 × 750 μm2 transparent sapphire chip containing a custom-designed micro-light emitting diode with peak emission at 475 nm, all integrated onto a contact lens. The display can be powered wirelessly from ∼1 m in free space and ∼2 cm in vivo on a rabbit. The display was tested on live, anesthetized rabbits with no observed adverse effect. In order to extend display capabilities, design and fabrication of micro-Fresnel lenses on a contact lens are presented to move toward a multipixel display that can be worn in the form of a contact lens. Contact lenses with integrated micro-Fresnel lenses were also tested on live rabbits and showed no adverse effect. (Some figures in this article are in colour only in the electronic version)

Prior work has demonstrated different types of contact lens functionalization. Contact lens mounted biosensors have been developed to measure eye movement [4, 5], tear glucose concentration [6–10], corneal temperature [11], blood oxygen [12] and intraocular pressure [13]. Although it is not contact lens based, an implanted intraocular vision aid (IOVA) is similar in concept to this project in that it projects images onto the retina from a system of light emitting diodes (LEDs) and microlenses [14]. Integrating displays on contact lenses present several challenges. First, as the display is standalone, power must be provided and stored for system operation without a wired connection. Second, the lens system must be biocompatible and meet regulations for radiofrequency (RF) radiation. A third challenge is mechanical and electrical integration of heterogeneous micrometer-scale components (e.g. silicon integrated circuitry and III–V semiconductor optoelectronics) on a polymer substrate. Furthermore, all components must fit within the volume of standard contact lenses, ∼1.0 cm2 in area with thicknesses of 200 μm or smaller [15]. Additionally, the human eye, with its minimum focal distance of several centimeters, cannot resolve objects on a contact lens. We envision two solutions to this problem: light from LEDs could be focused using subsidiary lenses, or alternatively, beams

1. Introduction Wearable computing will likely provide new ways to manage information and interact with the world. For example, personal see-through devices could overlay computer-generated visual information on the real world, providing immediate, handsfree access to information [1]. The first step toward alwaysaccessible information and superimposed, data-rich imagery will most likely be implemented in see-through eyeglass displays. Ultimately, a much less cumbersome display for augmented reality may take the form of a contact lens. In the future, contact lens systems may receive data from external platforms (e.g. mobile phones) and provide real-time notification of important events. As contact lensbased biosensors advance [2], they may alert the wearer of physiological anomalies, such as irregular glucose or lactate levels. With more colors and increased resolution, contact lenses may display text, be used with gaming devices, or offer cues from navigation systems [3]. Our long-term goal is to create a display that can be comfortably worn in the form of a contact lens, which will include a pixel array, focusing optics, an antenna, and circuitry for power harvesting, radio communication, and pixel control (figure 1). 0960-1317/11/125014+08$33.00

1

© 2011 IOP Publishing Ltd

Printed in the UK & the USA

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

2.2. Blue micro-LEDs (a)

(b)

We previously investigated the fabrication [16] and deployment of red LEDs on contact lenses [17]. To move toward a full color display, we chose to fabricate blue micro-LEDs. GaN and its alloys were deemed suitable due to nontoxicity, high efficiency and appropriate achievable emission wavelength [18]. The present micro-LED design, with peak intensity at ∼475 nm, is adequate to illuminate the retina. A 50 mm diameter, 435 μm thick c-plane sapphire wafer (Roditi International Corporation Ltd) was used as a substrate for the InGaN/GaN-based micro-LEDs. A vertical flow, close-coupled showerhead (CCS), metal organic vapor phase epitaxy (MOVPE) reactor system was used for the epitaxial growth. Ammonia gas was used as a nitrogen precursor, and trimethylgallium (TMGa) and trimethylindium (TMI) were used as precursors for gallium and indium, respectively. Doping of p- and n-type layers was achieved with Biscyclopentadienylmagnesium (Cp2Mg) and silane (SiH4). We used a two-step technique to grow a GaN buffer layer on the sapphire substrate. This was followed by growth of a 2 μm thick n-doped GaN layer at 1030 ◦ C. A multiple quantum well (MQW) stack consisting of five pairs of 3 nm thick InGaN (QWs) and 20 nm thick GaN barriers was grown on the ndoped GaN layer. InGaN growth was performed at 760 ◦ C to ensure sufficient indium incorporation in the wells, whereas the GaN barriers were grown at higher temperatures to achieve acceptable surface morphology. A 50 nm GaN capping layer was grown on the QW stack before the temperature was increased to 970 ◦ C to grow 250 nm of p-GaN. The capping layer prevents desorption of indium from the QWs during higher temperature steps. Finally the LED structure was annealed inside the MOVPE reactor for p-GaN activation. The process of forming electrical contacts involved three lithographic processes for lift-off, three metallizations and plasma etching. AZ5214E photoresist (Capitol Scientific, Inc.) was used in image reversal mode to achieve resist sidewall undercut for lift-off. First, because of its excellent etch selectivity during chlorine-based inductively coupled plasma etching, Ni was evaporated as an etch mask to define pixel size and shape. Chlorine-based plasma was used to etch 800 nm through the p-GaN and MQW stack to expose the n-GaN layer. P-metal contacts (Ni/Au/Ag/Au, 5/5/70/ 200 nm) were evaporated on the p-type mesa structures after a second photolithography. Ag was used to enhance light reflection from the contact side of the LED structure. Next, the n-contact metal (Ni/Al/Au, 20/860/200 nm) was deposited. The n- and p-metals were annealed between 400 and 500 ◦ C to improve the ohmic character of the contacts. Finally, the wafers were thinned to 250 μm by using lapping and grinding processes (CORWIL Technology Corporation) and diced into 750 × 750 μm2 chips prior to assembly on the contact lens platform. Figure 2 shows the LED growth stack and post-growth processing.

(c)

Figure 1. Conceptual rendition of a multipixel contact lens display. (a) A contact lens display comprising a multipixel light emitting diode (LED) chip (1), power-harvesting/control circuitry (2), antenna (3), and interconnects (4). These subsystems are encapsulated in a transparent polymer (5), creating a system to project virtual images (6) perceivable by the eye of the wearer. (b) LED chip with 100 pixels. LED active layers can be grown atop a transparent substrate. Emitted light travels through the substrate and is reimaged using planar Fresnel lenses. (c) Magnified view with one pixel activated, showing Fresnel lenses opposite each LED pixel.

from vertical cavity surface emitting lasers could be used to create an image. In this paper, we describe an unfocused single-pixel contact lens display that does not require complex optics. In the following sections we report how we designed, built and tested each subsystem of a single-pixel contact lens display. We describe how the subsystems were integrated into a fully functional system that was turned on safely on a live eye. We present our measurement results, including preliminary in vivo rabbit studies, and address our goal of creating a multipixel display with fabrication and testing of micro-Fresnel lenses and multi-pixel LED chips.

2. Design and fabrication 2.1. System operation This single-pixel contact lens display comprises a light emitting diode, an antenna, an integrated circuit (IC) for power harvesting and LED control, and a polymer substrate with electrical interconnects. An external transmitting antenna emits RF radiation that is collected by the on-lens antenna. The integrated circuit rectifies and stores this energy, and duty cycles power to the LED at a frequency sufficiently high to give the appearance of continual light emission.

2.3. Antenna Antenna design was constrained by contact lens size and eye physiology, and therefore a 5.0 mm radius loop antenna was 2

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

Figure 2. Schematic depicting the micro-LED growth stack (left) and postgrowth fabrication process (right).

power consumption, to be shorter on the eye than in free space. However, as contact lens and transmitting antenna separation decreases, near field interactions begin to dominate and more power is available. These simulation results confirm that only small amounts of power are available for operating electronic and photonic systems on-lens. Developing more realistic RF models for the eye will help in optimizing the transmitted power to the system. Impedance matching between chip and antenna can have a significant impact on system performance. Typically, wireless systems have passive matching networks between chip and antenna to achieve maximum power transfer [19], improve power amplifier efficiency, or to provide voltage gain that improves the sensitivity of voltage-limited rectifiers. For onlens systems, surface mount parts are prohibitively large, so the chip and antenna must be directly connected without an external matching network. Multiple rectifier stages are used in the on-lens chip to reduce the necessary RF input power needed to activate the LED. In our case, adequate antenna-to-chip matching is possible without an external matching network. Simulations show an antenna self-resonance of ∼3.6 GHz. The antenna can be capacitive or inductive, depending on whether it is operated above or below self-resonance, respectively. However, chip size limitations require that the power harvesting IC is capacitive. Hence, when operating below 3.6 GHz, chip capacitance (negative imaginary impedance) and antenna inductance (positive imaginary impedance) cancel and maximum power is transferred to the chip. The chip impedance is approximately 7  + 1.2 pF and the on-lens antenna has an inductance of 22 nH below the resonant frequency. Thus, the resonant frequency of the antenna and chip system is approximately 980 MHz. It is necessary to adhere to biological safety limits for exposure to nonionizing radiation provided by the Institute for Electrical and Electronics Engineers (IEEE) and the International Committee on Non-ionizing Radiation Protection (ICNIRP). The ICNIRP set the basic requirement for exposure of the head and trunk for the general public at 2 W kg−1 (up to 10 GHz), or a plane wave power density of 10 W m−2 (2–300 GHz) [20]. The IEEE Standard C95.12005 places exposure limits at 2 W kg−1 (100 kHz–3 GHz)

Figure 3. Simulated received power from a 1.0 W source at a distance of 1.0 m on contact lens placed in free space (dashed line), and also with 5.0 μm of saline above the antenna and 1.0 cm of saline below the antenna to model the eye (solid line).

used to harvest RF energy but leave the pupil unobstructed. Two factors determine the frequency of operation: (1) power received by the on-lens loop antenna, and (2) antenna-to-chip impedance matching. To determine power received by the on-lens antenna, we began by simulating gain for a loop antenna having a 5.0 mm radius, 0.5 mm width and 5.0 μm thickness (Advanced Design System, Momentum 3D Planar EM Simulator, Agilent Technologies, Inc.). The peak on-lens antenna received power was determined using simulated gain data and the Friis transmission equation, with perfect antenna–chip matching and values for safe transmit power, maximum likely antenna separation and minimum transmit antenna gain. Figure 4 shows power received with 1.0 m separation using a 1.0 W power source and transmitting antenna gain of 1.76 dBi. In free space, modeled received power peaked at ∼117 μW at approximately 2.0 GHz (figure 3). At lower frequencies, received power is limited by on-lens antenna efficiency, while path loss dominates at higher frequencies. We also simulated the eyeball and the tear film with 1.0 cm and 5.0 μm of saline solution (relative permittivity = 74, conductivity = 0.47 S m−1 ) below and above the antenna, respectively. Antenna received power is significantly reduced in this more realistic operational environment. Here, the simulated maximum received power is ∼12 μW (figure 3). Therefore, we expect operating distances, assuming equivalent 3

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

First, 100 mm wafers with major and minor flats were cut from 25.4 cm × 25.4 cm × 100 μm sheets of polyethylene terephthalate (POLICROM, Inc.) using a CO2 laser cutter. The wafers were cleaned with sonication in acetone, isopropyl alcohol (IPA), deionized water and then spin-rinse-dried. Next, AZ4620 (MicroChem Corp.) was spun (spread: 500 rpm, 100 rpm s−1 , 10 s, spin: 4500 rpm, 1000 rpm s−1 , 30 s) and soft baked (35 ◦ C, ramped to 70 ◦ C, held for 10 min, and ramped back to 35 ◦ C). Ultraviolet exposure was done in an ABM contact aligner for 35 s with I-line at 13 mW cm−2 and H-line at 25 mW cm−2 . Development was carried out in AZ400K (MicroChem Corp.) for 90 s. The wafers were plasma cleaned in a Branson barrel etcher for 3 min at 300 W and immediately placed in an electron beam evaporator. A layer of Cr/Ni/Au (20, 80, 350 nm) was evaporated at 5 × 10−6 Torr. Acetone and sonication for ∼30 s facilitated metal lift-off. After lift-off, the wafers were rinsed in IPA, deionized water and put through a spin-rinse-dry. Next, a ∼2 μm layer of SU-8 2 was spun (spread: 500 rpm, 100 rpm s−1 , 10 s, spin: 3000 rpm, 300 rpm s−1 , 40 s), baked (35 ◦ C, ramped to 70 ◦ C, held for 10 min, and ramped back to 35 ◦ C), exposed for 20 s and developed for 90 s in SU-8 with agitation. Again, the wafers were rinsed in IPA, and sent through the spin–rinse– dry. Next, a 40 nm seed layer of Au was deposited using electron beam evaporation, and AZ4620 was patterned using the aforementioned method. Pulsed electroplating (1 ms on, 1 ms off) was used to plate ∼5 μm of Au using a cyanideR based Pur-A-Gold 401 formulation (Enthone-OMI Inc.). This electroplating step is necessary in order to produce more efficient antennas with thicker metal than is normally achievable with e-beam evaporation or sputtering. The photoresist was removed in acetone, and the seed layer was etched in 5:1 (vol:vol) deionized H2 O:Gold Etch TFA (Transene Company, Inc.). SU-8 25 was spun (spread: 500 rpm, 100 rpm s−1 , 10 s, spin: 2000 rpm, 300 rpm s−1 , 40 s), exposed for 30 s, and developed for 4 min in SU-8 developer. Blue dicing tape was placed on the back of the wafers, and individual contact lenses were cut out with a CO2 laser. Aluminum oxide sand paper (5 μm grit) were used to polish the contact lens edges after laser cutting.

(a)

(b)

(c)

(d )

Figure 4. Optical micrographs as well as electrical and optical characteristics of an LED chip. (a) View of the micro-LED chip assembled onto a polymer template, taken looking through the sapphire substrate. The scale bar is 375 μm. Upon application of a bias voltage (right), the micro-LED pixel emits light. (b) Measured current versus bias voltage. (c) Measured light emission spectrum. (d) Measured efficiency versus bias current for a 260 μm circular pixel. Peak emission is at 475 nm (blue) while operating at 3.0 V. In typical indoor light levels, the micro-LED is bright and visible to the naked eye at ∼2.6 V and ∼60 μA.

2.5. Component assembly and molding We connected the LED and CMOS circuit to the lens in a manner similar to that presented in [22]. Prior to solder coating, the custom silicon IC aluminum pads (from MOSIS) were electrolessly plated with Ni and Au (CV Inc.). Leadfree, 60 ◦ C eutectic solder (Indium Corp., Indalloy 19) was used to coat the exposed pads on the template, microchips and micro-LEDs. The solder was melted in a 100 mL beaker while immersed in ethylene glycol (EG). Next, 60 μL of HCl was added to the ethylene glycol to clean the solder surface. While the solder was heating, a solution of 25 mL of EG and 10 μL of HCl was prepared in a petri dish. Prior to coating, the EG used during heating was removed from the solder by pipette, and 10 mL of fresh EG was added along with 60 μL of HCl. Next, the IC, micro-LED chip and template were coated with solder using a pipette and then placed in the prepared

or 10 W m−2 (2 GHz–100 GHz). Research contained in the IEEE standard and a recent ICNIRP review [21] point toward thermally induced cataracts as the primary biological concern of RF exposure to the eye. Our aim is to remain well below the thresholds set by these standards during the normal operation of the contact lens system. 2.4. Contact lens template fabrication The antenna, electrical interconnects, electrical isolation and pads for solder coating were fabricated directly on the contact lens. Polyethylene terephthalate (PET) was chosen as a substrate because of good chemical resistance, thermal stability during photolithography and transparency. 4

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

petri dish. The silicon microchip and micro-LED chip were placed on the contact lens in corresponding SU-8 25 wells using a microscope for observation. The petri dish was heated to melt the solder, at which time components self-aligned and connected to the contact lens interconnects and antenna. The petri dish was removed from the hot plate approximately 30 s after chip alignment. After cooling for 5 min, the contact lenses were soaked in IPA for 20 min, then rinsed in IPA and dried with nitrogen. We used a hemispherical aluminum mold (radius = 7.1 mm) to obtain the desired lens curvature. The mold contained a small hole drilled through the center and exiting the side, which was connected to a vacuum line. We heated the mold and a metal ring on hotplate set to 200 ◦ C. The vacuum was pulled, and the contact lens was centered on the mold. We then placed the mold on an insulating surface, and the ring, having an inner diameter slightly less than the contact lens, was used to slowly press and shape the lens without disturbing the components. After cooling, the contact lens was removed from the mold, cleaned with IPA and dried with nitrogen. Lastly, we deposited a layer of parylene-C (∼10 μm, Specialty R Coating Systems, PDS 2010 Labcoater 2) conformally over the contact lens for the purposes of biocompatibility and mechanical strength. The contacts lenses were sterilized in ethanol for 10 min and stored in a saline solution prior to testing on animals.

zones. Next, negative EBL resist AR-N 7700 (ALLRESIST GmbH) was spun onto the substrate for 45 s at 6000 rpm and soft baked at 85 ◦ C for 1 min. Following exposure at ∼8 μC cm−2 and baking at 100 ◦ C for 2 min, the samples were developed in AR 300-46 (ALLRESIST GmbH) for 60 s. Phosphoric acid etchant 80-16-4(65) was used for ∼4.5 min to etch the aluminum. After aluminum etching and resist stripping, a thin layer of BGL-GZ-83 (PROFACTOR GmbH) was spun onto the samples to prevent scratching during tests.

3. Results 3.1. Blue micro-LEDs Micro-LEDs were tested using an electrical probe station and a light collection system comprising an optical fiber, integrating sphere and photospectrometer. Figure 4 shows LED off and on states, measured current–voltage characteristics, measured spectral output and measured electrical-to-optical power conversion efficiency. The micro-LED light emission shape can be controlled by using a patterned metal p-contact. 3.2. CMOS integrated circuit The stringent low power consumption and small footprint requirements prohibited the use of commercially available integrated circuits. Thus, incorporation of wireless electronics onto a contact lens made it necessary for us to design a custom IC (fabricated through MOSIS using a 130 nm CMOS process) to perform control and radio functions. This chip (figure 5) performs the following tasks: conversion of RF power received by the on-lens antenna to a dc-supply voltage, on-chip energy storage sufficient to excite the micro-LED, and generation of an on-chip clock source to duty cycle an LED driver switch. The power management circuitry consists of a ring oscillator, a pulse generator and a passive level shifter. The ring oscillator is powered from the second stage of an eight-stage cascaded rectifier. Each inverter in the three-stage oscillator uses stacked high-threshold devices for low leakage and low-power operation. The power consumption of the ring oscillator is less than 500 nW at 1.0 MHz. The chip is described in more detail in [17].

2.6. Fresnel lens fabrication To extend the utility of the contact lens display in the future, we plan to develop multipixel systems using matrix-addressable micro-LED arrays and micro-Fresnel lenses to focus images onto the retina. We have developed micro-LED pixel arrays containing five or eight 80 μm pixels on 750 × 750 μm2 chips and tested their functionality. However, work by other groups verifies the possibility of producing higher density arrays, for example with 22 μm pitch [23]. This suggests that it would be possible to place hundreds of pixels on a 500 × 500 μm2 LED array chip. To investigate the integration of subsidiary lenses on contacts, we fabricated and tested micrometer-scale Fresnel lenses to focus LED output onto the retina. Fresnel lenses could be positioned on an LED chip opposite the active layers (i.e. on the back side of the handle wafer, as shown in figure 1(c)) or on contact lens templates for the same effect. In order to facilitate characterization, we fabricated Fresnel lenses on contact lens templates. Fresnel lens parameters (e.g. focal length, zone radii, zone width) were calculated using modified Fresnel diffraction approximations, and simulations were performed to evaluate lens effectiveness. Fresnel microlenses were fabricated using electron beam lithography (EBL) to pattern opaque metal rings on transparent PET. PET poses two sets of problems: first, PET can melt during fabrication, and second, it can generate adverse electrical charging during EBL. The following fabrication process was developed to eliminate these effects. First, the substrate was cleaned and 70 nm of Al was evaporated to dissipate charge during EBL and also to form opaque Fresnel

3.3. Testing the contact lens system First, we demonstrated wireless activation of an on-lens LED using a 1.05 GHz transmission frequency in free space. The transmit antenna was an SAS-571 double ridge guide horn antenna (A.H. Systems, 700 MHz–18 GHz) fed with a +25 dBm signal. The contact lens was held using polymer tweezers and oriented with the maximum directivity facing the horn antenna. For a small loop antenna, the maximum directivity and loop are coplanar. Under these conditions, the LED was brightly lit and visible to the naked eye at ∼10 cm. In free space using a +35 dBm signal, we could activate the LED at distances of ∼1.0 m, as measured from the horn antenna mouth. 5

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

Table 1. Contact lens test result summary. Test platform

Transmit antenna

Input power

Frequency (GHz)

Operating distance

Free space Free space Rabbit cadaver eye, w/o saline film Rabbit cadaver eye, with saline film Rabbit cadaver eye, with saline film Live rabbit eye

Horn Horn Horn Horn Dipole Dipole

+25 dBm +35 dBm +35 dBm +35 dBm +35 dBm +35 dBm

1.05 1.05 1.05 1.8 ∼0.8−2.0 ∼0.8−2.0

10 cm 1m 10 cm 1 cm 2 cm 2 cm

(a)

(a)

(b)

(b)

(c)

Figure 5. Custom designed and fabricated power harvesting and LED driver circuit. (a) Schematic of the CMOS IC architecture and connectivity. (b) Photograph of assembled chip with rectifier and storage capacitor outlined in white, viewed through the polymer substrate. Scale bar is 250 μm. Our assembly procedure facilitates accurate alignment of the chip and contact lens.

Tests on a rabbit cadaver eye showed activation of the LED from ∼10 cm with the horn antenna and +35 dBm input power, again at 1.05 GHz. However, when a saline film was applied to the top of the lens, the frequency changed significantly (∼1.05 GHz to ∼1.8 GHz) and the operational range decreased. To see the LED brightly lit, we used a dipole antenna from close range (∼2.0 cm). In this setup, we could activate the LED in a larger frequency range (∼0.8 to 2.0 GHz). The larger frequency range and closer operational distance suggests a more inductive link. Table 1 summarizes test conditions and results. All in vivo studies were performed according to the guidelines of the National Institutes of Health for use of laboratory animals, and with the approval of the Institute of Animal Care and Use Committee of the University of Washington (Protocol UW4139-01). All exams were conducted in the University of Washington vivarium designed with maximal comfort for the rabbits, and all examination equipment was portable to the rabbit ‘bed-side’. All in vivo

Figure 6. Testing the contact lens display on a live rabbit. (a) Photograph of a completed contact lens system. (b) The contact lens display is placed on the eye of a live rabbit and powered by a dipole antenna, showing bright emission from the on-lens pixel. (c) Subsequent tests using fluorescein showed no corneal epithelial damage.

evaluations were performed under general anesthesia. Female New Zealand White Rabbits (mean wt. 2.5 kg) underwent general anesthesia with 5% induction of isoflurane and oxygen, which were maintained at 2% during the in vivo studies. Artificial tears were applied frequently to insure hydration. The prototype contact lens was well fitted to the rabbit ocular surface (figure 6(b)). The micro-LED indicator was activated periodically and remained in vivo for up to 40 min. We evaluated the effect of wearing the contact lens display on the rabbit cornea using portable slit lamp biomicroscopy 6

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

and pachymetry. Additionally, topical fluorescein was applied to the corneal surface using a dropper and the eye was visually examined for potential corneal abrasion, thermal burn or corneal edema as a result of in vivo testing. Such damage would result in corneal surface roughness, causing the fluorescein to accumulate. Figure 6(c) demonstrates a healthy corneal epithelial layer after testing, showing no fluorescein streaks or cloudiness.

(a)

(b)

3.4. Fresnel lenses (c)

To characterize the Fresnel lenses, a Nikon TE2000 Eclipse inverted optical microscope and a Hamamatsu ORCA-ER monochrome CCD camera were used. The Fresnel lenses were placed on a glass slide with the rings toward the microscope objective, and a 200 mm pinhole was placed 56 mm above the lens. The microscope was focused on the lens surface, and the focal plane was slowly adjusted until the brightest spot was observed. Distances were measured using the focus knob on the microscope. For a lens with a diameter of 108 μm, the focal length was measured to be 62 μm. Figures 7(a) and (b) show the simulated and measured intensity distributions, respectively, at the focal plane for a lens we constructed. After fabrication and characterization, we molded and parylene coated the lenses (figure 7(c)), and then tested them in vivo on rabbits using procedures outlined above (figures 7(d) and (e)).

(d )

(e)

Figure 7. Micro-Fresnel lenses. (a) Simulated intensity profile at the focal plane of the lens. The z-axis units are arbitrary. (b) Measured intensity profile at the focal plane of the lens. (c) Optical microscope image of a micro-Fresnel lens integrated on a contact lens. (d) Photograph of a contact lens with fifteen micro-Fresnel lenses placed on an anesthetized rabbit. No adverse effects were observed on the eye. (e) Close-up view of 3 × 5 array of Fresnel lenses on rabbit.

4. Discussion We have demonstrated the operation of a contact lens display powered by a remote radiofrequency transmitter in free space and on a live rabbit. This verifies that antennas, radio chips, control circuitry and micrometer-scale light sources can be integrated into a contact lens and operated on live eyes. Although our display has only a single controllable pixel, we have provided the first proof-of-concept technology demonstrations for producing multipixel and in-focus images using a contact lens by producing multipixel micro-LED array chips on transparent substrates and micrometer-scale Fresnel lenses that can be integrated into a contact lens. The demonstration of Fresnel lenses on contact lenses points toward the potential of integrating other passive and active micro-optical components on a contact lens for vision correction and enhancement. Significant improvements are necessary to produce fully functional, remotely powered, high-resolution displays. Although we could power our system in free space from more than a meter, operating distances on the rabbit eye were reduced to the cm range. Matching, interface and absorption losses are likely causes of the limited operational distance. We are working to improve matching losses, to ensure that power received by the contact lens is maximized at the frequency of best antenna-to-chip matching, and to optimize LED efficiency and duty cycling to reduce power consumption of individual pixels. Although only microwatts of power is available on the contact lens, most light generated by the optical components directly enters the eye. Thus, the

display could efficiently generate an image while consuming little power. PET has been our contact lens substrate of choice thus far due to the ease of performing some microfabrication processes; however, PET has poor oxygen permeability and its extended use could lead to lactate build-up and corneal swelling [24]. It will be necessary to adapt our processes to more common rigid gas permeable or hydrogel contact lens materials. This work is currently in progress in our laboratory. We have demonstrated the integration of blue and red micrometer-scale light sources on contact lenses. The integration of green emitting micrometer-scale light sources must be achieved in order to fully extend the color gamut. A display with a single controllable pixel could be used in gaming, training, or giving warnings to the hearing impaired. We also believe it is possible to develop systems with better resolution, color, range and computing power. Displays with a handful of pixels could be used to provide directional information, and displays with hundreds of pixels used to read short emails or text messages. Although high resolution, full-color, stand-alone contact lens displays might be many years away, the technological demonstrations to date depict a clear path containing a number of useful intermediate devices that can be feasibly produced in the near to medium 7

J. Micromech. Microeng. 21 (2011) 125014

A R Lingley et al

terms. If such displays were successfully deployed, they would fundamentally change the nature of interaction between humans and visual information.

[10] Chu M et al 2011 Soft contact lens biosensor for in situ monitoring of tear glucose as non-invasive blood sugar assessment Talanta 83 960–5 [11] Kinn J B and Tell R A 1973 A liquid-crystal contact-lens device for measurement of corneal temperature IEEE Trans. Biomed. Eng. 20 387–8 [12] Scott V, Tarassenko L, Glynn C and Hill A 1996 Contact lens oximetry: a valid concept? Pulse Oximetry: A Critical Appraisal, IEE Colloquium (3 January–3 March) [13] Leonardi M, Pitchon E M, Bertsch A, Renaud P and Mermoud A 2009 Wireless contact lens sensor for intraocular pressure monitoring: assessment on enucleated pig eyes Acta Ophthalmol. 87 433–7 [14] Szurman P et al 2005 Experimental implantation and long-term testing of an intraocular vision aid in rabbits Arch. Ophthalmol. 123 964–9 [15] Phillips A J 2007 Contact Lenses 5th edn ed A J Phillips and L Speedwell (New York: Elsevier) p 218 [16] Saeedi E, Kim S and Parviz B A 2007 Self-assembled inorganic micro-display on plastic IEEE 20th Int. Conf. on MEMS pp 755–8 [17] Pandey J et al 2010 A fully integrated RF-powered contact lens with a single element display IEEE Trans. Biomed. Circuits Syst. 4 454–61 [18] Nakamura S 2002 III-V nitride-based LEDs and lasers: current status and future opportunities Electron Devices Meeting, 2000. IEDM Technical Digest. International pp 9–11 [19] Barnett R E, Liu J and Lazar S 2009 A RF to DC voltage conversion model for multi-stage rectifiers in UHF RFID transponders IEEE J. Solid-State Circuits 44 354–70 [20] Ahlbom A et al 1998 Guidelines for limiting exposure to time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz) Health Phys. 74 494–522 [21] Swerdlow A 2010 Exposure to High Frequency Electromagnetic Fields, Biological Effects and Health Consequences (100 kHz–300 GHz) ed P Vecchia et al (Oberschleissheim, Germany: International Committee on Non-Ionizing Radiation Protection) pp 208–11 [22] Stauth S A and Parviz B A 2006 Self-assembled single-crystal silicon circuits on plastic Proc. Natl Acad. Sci. 103 13922–7 [23] Choi H W, Jeon C W and Dawson M D 2004 High-resolution 128 × 96 nitride microdisplay IEEE Electron Device Lett. 25 277–9 [24] Sweeney D F, Stretton S, Fonn D, Swarbrick H A and Holden B A 2007 Contact Lenses 5th edn ed A J Phillips and L Speedwell (New York: Elsevier) pp 273–4

Acknowledgments We would like to acknowledge the National Science Foundation (the EFRI program), the Defense Advanced Research Projects Agency (MTO), the Nokia Foundation, the Aalto MIDE program, and Research to Prevent Blindness for their financial support. We would also like to thank Roman Khakimow and Igor Shavrin for simulation and characterization of Fresnel lenses, and the staff at the Washington Technology Center Microfabrication Laboratory for their help in process development.

References [1] Billinghurst M and Starner T 1999 Wearable devices: new ways to manage information Computer 32 57–64 [2] Lahdesmaki I, Shum A and Parviz B A 2010 Possibilities for continuous glucose monitoring by a functional contact lens Instrum. Meas. Magazine IEEE 13 14–7 [3] Parviz B A 2009 For your eye only IEEE Spectr. 46 36–41 [4] Robinson D 1963 A method of measuring eye movement using a scleral search coil in a magnetic field Biomed. Electron., IEEE Trans. 10 137–45 [5] Kim K, Lee S and Kim H 2004 A wireless measurement system for three-dimensional ocular movement using the magnetic contact lens sensing technique IEMBS ’04. 26th Annual Inter. Conf. of the IEEE vol 1 pp 2287–9 [6] Badugu R, Lakowicz J R and Geddes C D 2004 Noninvasive continuous monitoring of physiological glucose using a monosaccharide-sensing contact lens Anal. Chem. 76 610–8 [7] March W F, Mueller A and Herbrechtsmeier P 2004 Clinical trial of a noninvasive contact lens glucose sensor Diabetes Technol. Ther. 6 782–9 [8] Domschke A, March W F, Kabilan S and Lowe C 2006 Initial clinical testing of a holographic non-invasive contact lens glucose sensor Diabetes Technol. Ther. 8 89–93 [9] Yao H, Shum A, Cowan M, Lahdesmaki I and Parviz B A 2011 A contact lens with embedded sensor for monitoring tear glucose level Biosens. Bioelectron. 26 3290–6

8