Letter pubs.acs.org/NanoLett
A Highly Tunable and Fully Biocompatible Silk Nanoplasmonic Optical Sensor Myungjae Lee,† Heonsu Jeon,†,‡ and Sunghwan Kim*,§,∥ †
Department of Physics and Astronomy and Inter-University Semiconductor Research Center and ‡Department of Biophysics and Chemical Biology, Seoul National University, Seoul 151-747, Republic of Korea § Department of Physics and ∥Department of Energy Systems Research, Ajou University, Suwon 443-749, Republic of Korea S Supporting Information *
ABSTRACT: Novel concepts for manipulating plasmonic resonances and the biocompatibility of plasmonic devices offer great potential in versatile applications involving real-time and in vivo monitoring of analytes with high sensitivity in biomedical and biological research. Here we report a biocompatible and highly tunable plasmonic bio/ chemical sensor consisting of a natural silk protein and a gold nanostructure. Our silk plasmonic absorber sensor (SPAS) takes advantage of the strong local field enhancement in the metal− insulator−metal resonator in which silk protein is used as an insulating spacer and substrate. The silk insulating spacer has hydrogel properties and therefore exhibits a controllable swelling when exposed to water−alcohol mixtures. We experimentally and numerically show that drastic spectral shifts in reflectance minima arise from the changing physical volume and refractive index of the silk spacer during swelling. Furthermore, we apply this SPAS device as a glucose sensor with a very high sensitivity of 1200 nm/RIU (refractive index units) and high relative intensity change. KEYWORDS: Silk fibroin hydrogel, surface plasmon, biocompatible device, plasmonic sensor, metal−insulator−metal resonators
F
to couple incident light to metallic surfaces.13−15 It is wellknown that the figure of merit (FOM), defined as the RI sensitivity divided by the resonance line-width, of LSPR sensors are 1 to 2 orders of magnitude lower than PSPR sensors.16,17 However, even though the FOM of LSPR sensors has approached that for PSPR sensors in a recent study, the FOM values for SPR sensors cannot overcome a theoretical upper limit of around 100, which corresponds to a 10 nm resonance shift of the 10 nm line-width spectral dip/peak for a 0.01 RI change.18 Considering that obtaining an extremely narrow spectral response in SPR resonators is impossible due to the intrinsic loss of metals, the limited FOM values means that highly sensitive and stable measurement systems are required to detect traces of analytes. Herein, we report the design and demonstration of a highly tunable LSPR sensor by utilizing the concept of a metal− insulator−metal (MIM) absorber. The thin insulator layer of the MIM sensor is composed of biocompatible and biodegradable silk fibroin hydrogel, which can accommodate water molecules by up to 60% in volume. Because the optical response of the proposed MIM structure is drastically sensitive to changes in thickness and RI of the thin insulator layer, the hydrogel properties of the silk insulator layer facilitate a high sensitivity to analytes through the control of the swelling ratio (thickness of the layer) and the mean RI of the silk hydrogel.
or point-of-care clinical evaluation, next generation opticsbased sensors are ideally suited as biomicroelectromechanical systems (BioMEMS), particularly in terms of microfluidics.1 In place of traditional MEMS materials such as silicon and polydimethylsiloxane (PDMS), the use of degradable biopolymers such as gelatin, poly(L-lactic acid) (PLA), and alginate allow for implantable BioMEMS devices to satisfy the growing demand for in vivo applications.2−4 Silk fibroin, the natural protein extracted from the Bombyx mori cocoon, is an interesting material in the bio-optics field due to its biocompatibility and unique mechanical and optical characteristics.5,6 Silk fibroin has been applied as a substrate material for biodegradable devices and a carrier for biodopants such as enzymes and antibodies with a biological function.7 The favorable combination of noble metals (especially gold) and silk fibroin promises a fully biocompatible plasmonic bio/chemical sensor platform that can be implanted in living tissue with no immune response. The surface plasmon resonances (SPRs) in this combination of materials also have many interesting features relevant for bio/chemical sensing. SPR-based sensors have found wide use in medical diagnosis, food safety, and environmental monitoring, due to their low cost and high sensitivity.8−10 Commercially available SPR sensors are based on propagating surface plasmon resonances (PSPRs), which can be generated on a flat metallic surface.11,12 Localized surface plasmon resonances (LSPRs) carried by metal nanostructures are a means to measure local refractive index (RI) changes at the nanoscale with a simplified measurement system without the need for prisms or gratings © XXXX American Chemical Society
Received: February 18, 2015 Revised: March 24, 2015
A
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 1. A fully biocompatible SPAS. (a) Working principle of the SPAS. The top layer is a 2D gold disk array and the bottom layer is a gold mirror. The diameter, thickness, and lattice constant of the gold disks are 200, 16, and 520 nm, respectively. The two layers are separated by a silk spacer layer with a thickness of 20 nm. The silk spacer absorbs the environmental liquid, thereby inducing the swelling. The changed volume and refractive index of the swelled silk layer affects the surface plasmon resonance behavior that results in the absorption of the incident wave. (b) The fabrication process of the SPAS. An arbitrary substrate such as a free-standing silk film and rigid silicon can be used to support the SPAS. (c) The SPAS (diffracted colors) on a chicken breast tissue. An optical fiber is adjacent to the SPAS surface for measurements. (d,e) Scanning electron microscopy and atomic force microscopy images of the fabricated SPAS. Scale bars represent 2 μm for (c) and 1 μm for (d).
Such an effect is not expected in conventional SPR sensors, which only use RI changes at metal−dielectric interfaces. We achieve a resonance shift of over 100 nm between a water environment and an isopropyl alcohol (IPA) environment despite the very small RI difference of these two liquids. Furthermore, the resonance shift exhibits a composition dependence. In addition, a network of silk polymer chains can act as a fluidic channel for the flow of analytes in water through a nanometer-sized layer. The proposed silk plasmonic structure is applied as a glucose sensor with higher sensitivity than other plasmonic glucose sensors, even though the aqueous glucose solution only affects the mean RI of the silk insulator spacer. Figure 1a shows a schematic diagram and the operating principle of the proposed silk plasmonic absorber sensor (SPAS). A two-dimensional (2D) gold disk array is stacked above a 200 nm thick gold film. The two gold layers are separated by a 20 nm silk spacer, as depicted in Figure 1a. Each gold nanodisk forms a MIM resonator with a silk spacer and a bottom gold layer. The coupled resonance in the periodic MIM induces a strong absorption of incident light and allows for the detection of absorption features in reflectance measurements.
Such a measurement is possible because the transmittance of the structure is totally eliminated due to the thick gold mirror, along with the polarization independence in the x- and ydirections at normal incidence.19 It is noteworthy that the core element of the SPAS is the silk hydrogel spacer. A simple methanol or water vapor treatment can ultimately increase the β-sheet crystallinity by causing hydrogen bond formation between silk molecules, thereby making the silk film a hydrogel.20,21 When we take advantage of the hydrogel property of the silk spacer (swelling by water-uptake), the silk insulator nanogap of the MIM structure changes its physical volume and RI and these induce drastic changes in the resonance mode. Figure 1b shows a diagram of the fabrication process for the SPAS. A free-standing silk film was used as a substrate. A thick gold film was subsequently deposited and the diluted silk solution was spin-coated to form a thin silk spacer. The gold nanodisk array was defined with a positive resist using standard electron beam lithography followed by a lift-off process. The finalized free-standing SPAS can be affixed on various surfaces including biological tissues, as shown in Figure 1c, for sensing. The substrate of the SPAS can be replaced as necessary with B
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 2. Measurements and simulations of the SPAS. (a) Schematic showing the butt-end fiber coupling measurement setup using a 1 × 2 multiplexing multimode fiber coupler. (b) Experimental reflectance spectra when the SPAS is immersed in air (black solid), IPA (red solid), and water (blue solid). (c) Simulated electric field intensity distributions for the reflection dip at the linear scale. The electric field is strongly localized at the silk spacer. (d) Simulated reflectance spectra for the structure in air (black curve), IPA (red curve), and water (blue solid). To emphasize the effect of the swelling on the resonance behavior, a simulated spectrum (blue dashed) in which the swelling is not considered is plotted in (d).
fiber coupling setup utilized by a 1 × 2 multiplexing multimode fiber coupler, as shown in Figure 2a.23,24 A white light source was fed into the input port and reached the SPAS through the cleaved fiber-tip. The reflected signal was collected through the same fiber tip and sent to the output port to be recorded. The implementation of a fiber coupler allows for a large alignment tolerance and avoids the geometrical restriction involved with adapting free-space optics when adjoined vertically emitted/ absorptive optical devices such as surface-emitting lasers and the SPAS.24 Figure 2b shows the measured reflectance-spectra
rigid materials such as silicon and glass. Figure 1d,e shows scanning electron microscopy (SEM) and atomic force microscopy (AFM) images of a fabricated sample with a 520 nm lattice constant (Λ), a 200 nm radius, and a 16 nm thick gold cap. All these dimensions were measured from the measured images. The strong adhesion between amino acids in the silk fibroin and gold further promotes a reliable fabrication method for integration with plasmonic nanostructures.22 A simple and efficient measurement scheme would make LSPR-based sensors more valuable. We employed a butt-end C
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
In Figure 3, we show spectral changes when the SPAS is immersed in IPA−water mixtures with varying compositions
of a proof-of-principle experiment. To confirm the tuning of the plasmonic resonance by the swelling of the silk hydrogel spacer, reflectance spectra of the SPAS were measured in air (n = 1.00), IPA (n = 1.37), and water (n = 1.32). As expected, the reflectance dip for the air-environment was red shifted when immersed in IPA due to the increasing RI of the environmental medium. When immerged in water, however, the plasmonic resonance reveals an unprecedented phenomenon, a blue shift of ∼100 nm, which is not expected from the RI change. This means that additional physical quantities affect the plasmonic resonance behavior. To investigate the optical response of the SPAS, we performed numerical simulations using a finite-difference time-domain (FDTD) method. Numerical structural input data from AFM topography data such as the radius and the height of the gold caps were used in FDTD simulations to compensate for any fabrication imperfections. As depicted in Figure 2c, the electromagnetic field is strongly localized in the intermediate silk spacer, thereby giving rise to a noticeable reflection dip in the spectrum with nearly zero intensity (Figure 2d). One interesting feature of the localized plasmonic mode is that the position of the reflection minimum is insensitive to the incident angle.19 This observation implies that even when laying on a curved surface, the SPAS will possess a reflection minimum at the same wavelength and therefore can be used as an attachable biosensor for living tissue. It is noteworthy that the experimental reflection minimum does not exhibit a nearly zero intensity, estimated in the simulations, because the excitation wave through the cleaved fiber tip has mode profiles different from the plane wave used in the simulations. For air and IPA environments, the simulated and experimental spectra agree well with each other with respect to the prediction for the position of the reflection minimum. The wavelength of the reflection minimum for IPA in the experimental spectrum is a little shorter than that in the simulated spectrum. This difference is because the polymer network of the silk spacer can also absorb a small amount of IPA molecules, as we will discuss later. For a water environment medium, we used two simulated spectra to examine the effect of the swelling on the plasmonic resonance. The spectrum in which the effect of the swelling is not taken into account is shown by the blue dotted curve. The blue solid curve in the same figure corresponds to the spectrum with swelling taken into account. In the simulation, a 30% thicker and a 60% water-containing silk spacer was considered, corresponding to a reduced RI of 1.4, from a previous study for silk inverse opal.25 In addition, we examined the effects of other individual physical quantities on the plasmonic resonance (see Supporting Information). Three physical quantities, the RI of the environment medium, the RI of the silk spacer, and the thickness of the silk spacer, were considered. Other quantities were fixed when a particular physical quantity was changed in the simulation. Regarding the RI of the silk spacer, the simulated sensitivity in terms of wavelength shift per refractive index unit (RIU) is 626 nm/ RIU, more than double that of the RI change of the environment medium (269 nm/RIU). Along with RI changes, the increase of the thickness of the silk spacer leads to a drastic resonance shift of 10.7 nm per 1 nm increase in thickness. These results underpin the fact that the tuning of the resonator itself can tremendously affect the resonance behavior, allowing the demonstration of a bio/chemical sensor with high sensitivity.
Figure 3. Plasmonic resonance behavior of the SPAS for water−IPA mixtures. (a) Reflectance spectra of the SPAS immersed in water−IPA mixture solutions with varying compositions. (b) Relations between the wavelength of the reflection minimum and the volume fraction of water, ϕ. The error bar represents the standard deviation calculated from five data points measured at each concentration.
and therefore different swelling ratios. With increasing water concentration, reflection minima undergo a blue shift and become narrower while becoming stronger absorbers. The stronger and narrower absorption can be described with a simple model by considering impedance matching between the environmental medium and the SPAS.19 The expansion and lower RI of the water-absorbed silk spacer make its complex impedance closer to that of the water environment, thereby resulting in enhanced absorption. Figure 3b shows the variation of the spectral position of the reflection minimum against the volume fraction of water, ϕ. Swelling of the silk spacer starts at ϕ = 0, thereby inducing a blue shift in the reflection minimum. An interesting and baffling property is that the spectral shift is reversed to a small red shift after ϕ = 0.5. This phenomenon cannot be explained by the RI change of the mixture with increasing ϕ because the RI of water is lower than that of IPA, and therefore indicates that the swelling of the silk spacer at ϕ = 0.5 (water/IPA = 1:1) is slightly larger than that at ϕ = 1.0 (pure water). The swelling ratios of the silk hydrogel for specific solvents were correlated with the relative magnitude of the polarity of the solvent. Table 1 shows the spectral shifts D
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
using water−methanol mixtures, which reveals a higher crystallinity in the mixture relative to that in pure methanol.27 According to this explanation for increasing expansion in IPA− water mixtures, the extra expansion should be inversely proportional to the polarities of the organic solvents. Supporting Information Figure S4 shows the variation of the spectral position of the reflection minimum against ϕ in aqueous acetone solutions, indicating that liquids with a lower polarity lead to a higher expansion. To investigate the potential for plasmonic bio/chemical sensing, the SPAS was employed to monitor glucose concentrations, one of the major challenges in the management of diabetes.28 In the present case, the physical volume of the silk spacer is not influenced by glucose concentrations because there is no causal factor for glucose molecules to stretch the network of pure silk proteins. However, the absorption of the glucose solution in the silk spacer increases the RI of the insulator layer, thereby leading to higher sensitivity relative to that for a similar structure consisting of rigid materials. Figure 4a shows the experimental reflected intensity spectra. An increasing glucose concentration enhances the RI of the aqueous solution and results in a red-shifted spectrum, along with relative intensity changes. The performance of the glucose sensor was analyzed by plotting the experimental wavelength shift (Δλ) as a function of glucose concentration in Figure 4b. A wavelength sensitivity of 1200 nm/RIU was obtained using the converted RI values,29 a value 4 times higher than the simulated value (269 nm/RIU) for the same structure with the rigid spacer-layer. The SPAS has reusability and reversibility, which are essential for their useful and practical sensing applications (see Figure 5S in the Supporting Information). Interestingly, as shown in Supporting Information Figure 5S, the wavelength sensitivity could almost double when the diameter of gold nanodisks became larger. Our future work will be to optimize the parameters of the SPAS for the highest sensitivity. This combination of relative intensity change and very high wavelength sensitivity is useful for monitoring extremely small analyte quantities, such as glucose in saliva.19,30
relative to the spectrum measured in air and polarities for various solvents (water, methanol, IPA, and acetone) to support this correlation. Table 1. Solvent Polarity Dependence of the Resonant Wavelength Shift (Δλ) of the SPASa
a
solvent
relative polarity
resonant shift (Δλ) [nm]
water methanol IPA acetone
1.00 0.762 0.546 0.355
−37 −13 74 80
The quantity Δλ is defined relative to air.
We need to address how the 50:50 mixture of solvents with weak (IPA) and strong (water) polarity can induce a larger volume-expansion relative to that for the pure solvent with strong polarity. Organic solvents are attracted by water through hydrogen bonding interactions and can then induce a reentrant transition in the hydrogels that have a hydrophobic interaction in which water and organic solvents are both good solvents for swelling.26 The strength of the hydrogel collapse by the reentrant in the mixture with water is inversely proportional to the polarities of organic solvents, which means that lower polarity of the solvent might be able to induce stronger hydrogen bonding to water. Although the silk spacer expands rather than contracts in organic solvent mixtures with water, we believe that the hydrogen-bonding between water and organic solvent, for which the strength is related to the polarity, will aid in the understanding of the observed properties. As depicted in Supporting Information Figure S3, the diffusion rate of low polar liquids should be sensibly low. However, water can promote the penetration of low polar liquids since water molecules can transport low polarity molecules to the molecular network of the silk hydrogel. In turn, this penetration leads to additional swelling because of the relatively strong hydrogen bonding between water and organic solvent molecules. Our results are consistent with previous observations of the annealing treatment of amorphous silk membranes
Figure 4. Glucose detection. (a) Representative reflectance spectra of the same SPAS immersed in glucose aqueous solution at different concentrations. (b) Measured wavelength shifts as a function of glucose concentration. The line is a linear fit with the refractive index sensitivity determined to be 1200 nm/RIU. E
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters Additionally, flexible optical sensors can be adversely affected by fluctuations of signals with respect to the detecting position due to different curvatures and stretched/constricted lattice constants. We confirmed that the SPAS exhibited the incident angle- and lattice-constant-insensitive resonance (see Figure 6S in the Supporting Information), which provide a great advantage for detecting consistent signals from flexible or curved sensors in simulations and experiments. More importantly, reforming the silk molecules by tagging stimuliresponsive molecules (for example, glucose oxidase for glucose) will enable the controllable swelling of the silk spacer, thereby inducing a drastic change in plasmonic resonances.31 In conclusion, we reported the fabrication and characterization of a novel plasmonic sensor consisting of a gold resonating nanoabsorber and a silk biopolymer. Absorbed analytes in the silk spacer change various physical quantities and lead to the strong tunability of plasmonic resonances. Water−alcohol mixtures were used as stimuli to control the swelling ratios and the RIs of the silk spacer. The resulting spectral shifts were analyzed using numerical simulations. In addition, we investigated the spectral response of the SPAS when immersed in glucose aqueous solutions with varying concentrations. Although the volume expansion of the silk spacer was not expected here, the experimental results exhibited a very high wavelength sensitivity of 1200 nm/RIU, valuable for a highly sensitive glucose biosensor. The combination of silk protein and gold plasmonic structures enables fully biocompatible plasmonic biochemical sensors and an improved detection ability from the unique properties of silk. Such a SPAS demonstrated here can be used for real-time, extremely sensitive, and in vivo monitoring of analytes in body fluid. In addition, the observation of an incident angle- and latticeconstant-insensitive resonance would be useful for consistent sensing when implanting or attaching the device on curved tissues.
■
Korea. S.K. is also supported by the TJ Park Science Fellowship of the POSCO TJ Park Foundation.
■
ASSOCIATED CONTENT
S Supporting Information *
Detailed descriptions of the experimental procedures, hydrogel properties of silk fibroin, simulation results for effects of physical quantities on the plasmonic resonance, solventdependence of the resonance-shift, reusability and reversibility of the resonator, and curvature-insensitive feature of the resonance. This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
(1) Bashir, R. Adv. Drug Delivery Rev. 2004, 56, 1565−1586. (2) Yang, L.-J.; Lin, W.-Z.; Yao, T.-J.; Tai, Y.-C. Sens. Actuators, A 2003, 103, 284−290. (3) Grayson, A. C. R.; Choi, I. S.; Tyler, B. M.; Wang, P. P.; Brem, H.; Cima, M. J.; Langer, R. Nat. Mater. 2003, 2, 767−772. (4) Cabodi, M.; Choi, N. W.; Gleghorn, J. P.; Lee, C. S. D; Bonassar, L. J.; Stroock, A. D. J. Am. Chem. Soc. 2005, 127, 13788−13789. (5) Omenetto, F. G.; Kaplan, D. L. Nat. Photonics 2008, 2, 641−643. (6) Hwang, S.-W.; Tao, H.; Kim, D.-H.; Cheng, H. Y.; Song, J.-K.; Rill, E.; Brenckle, M. A.; Panilaitis, B.; Won, S. M.; Kim, Y.-S. Science 2012, 337, 1640−1644. (7) Pritchard, E. M.; Dennis, P. B.; Omenetto, F. G.; Naik, R. R.; Kaplan, D. L. Biopolymers 2012, 97, 479−498. (8) Shankaran, D. R.; Gobi, K. V.; Miura, N. Sens. Actuators, B 2007, 121, 158−177. (9) Brolo, A. G. Nat. Photonics 2012, 6, 709−713. (10) Zeng, S.; Baillargeat, D.; Ho, H.-P.; Yong, K.-T. Chem. Soc. Rev. 2014, 43, 3426−3452. (11) Liedber, B.; Nylander, C.; Lundstrom, I. Sens. Actuators 1983, 4, 299−304. (12) Mullett, W. M.; Lai, E. P. C.; Yeung, J. M. Methods 2000, 22, 77−91. (13) Pryce, I. M.; Kelaita, Y. A.; Aydin, K.; Atwater, H. A. ACS Nano 2011, 5, 8167−8174. (14) Kabashin, A. V.; Evans, P.; Pastkovsky, S.; Hendren, W.; Wurtz, G. A.; Atkinson, R.; Pollard, R.; Podolskiy, V. A.; Zayats, A. V. Nat. Mater. 2009, 8, 867−871. (15) Zhang, S. P.; Bao, K.; Halas, N. J.; Xu, H. X.; Nordlander, P. Nano Lett. 2011, 11, 1657−1663. (16) Sherry, L. J.; Chang, S. H.; Schatz, G. C.; Van Duyne, R. P.; Wiley, B. J.; Xia, Y. N. Nano Lett. 2005, 5, 2034−2038. (17) Ameling, R.; Langguth, L.; Hentschel, M.; Mesch, M.; Braun, P. V.; Giessen, H. Appl. Phys. Lett. 2010, 97, 253116. (18) Shen, Y.; Zhou, J. H.; Liu, T. R.; Tao, Y. T.; Jiang, R. B.; Liu, M. X.; Xiao, G. H.; Zhu, J. H.; Zhou, Z.-K.; Wang, X. H. Nat. Commun. 2013, 4, 2381. (19) Liu, N.; Mesch, M.; Weiss, T.; Hentschel, M.; Giessen, H. Nano Lett. 2010, 10, 2342−2348. (20) Hu, X.; Shmelev, K.; Sun, L.; Gil, E.-S.; Park, S.-H.; Cebe, P.; Kaplan, D. L. Biomacromolecules 2011, 12, 1686−1696. (21) Jin, H.-J.; Park, J.; Karageorgiou, V.; Kim, U. J.; Valluzzi, R.; Kaplan, D. L. Adv. Funct. Mater. 2005, 15, 1241−1247. (22) Park, J.; Choi, Y.; Lee, M.; Jeon, H.; Kim, S. Nanoscale 2015, 7, 426−431. (23) Kim, S.; Lee, J.; Jeon, H.; Kim, H.-J. Appl. Phys. Lett. 2009, 94, 133503. (24) Kim, S.; Park, Y.; Hwang, K.; Lee, J.; Jeon, H.; Kim, H. J. J. Opt. Soc. Am. B 2009, 26, 1330−1333. (25) Kim, S.; Mitropoulos, A. N.; Spitzberg, J. D.; Tao, H.; Kaplan, D. L.; Omenetto, F. G. Nat. Photonics 2012, 6, 818−823. (26) Orakdogen, N.; Okay, O. Polymer 2006, 47, 561−568. (27) Tsukada, M.; Gotoh, Y.; Nagura, M.; Minoura, N.; Kasai, N.; Freddi, G. J. Polym. Sci., Part B: Polym. Phys. 1994, 32, 961−968. (28) Schmelzeisen-Redeker, G.; Staib, A.; Strasser, M.; Müller, U.; Schoemaker, M. J. Diabetes Sci. Technol. 2013, 7, 808−814. (29) Muhmood, W.; Yunus, M.; Rahman, A. Appl. Opt. 1988, 27, 3341−3343. (30) Feng, J.; Siu, V. S.; Roelke, A.; Mehta, V.; Rhieu, S. Y.; Palmore, G. T. R.; Pacifici, D. Nano Lett. 2012, 12, 602−609. (31) Kost, J.; Horbett, T. A.; Ratner, B. D.; Singh, M. J. Biomed. Mater. Res., Part A 1985, 19, 1117−1133.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors acknowledge support from the Basic Science Program (NRF-2014R1A1A1008080) and Nano-Material Technology Development Program (2009-0082580) through the National Research Foundation (NRF) and the NRF Grant (2008-0061906) funded by the MSIP and MEST, Republic of F
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
A Highly Tunable and Fully Biocompatible Silk Nanoplasmonic Optical Sensor Myungjae Lee,† Heonsu Jeon,†,‡ and Sunghwan Kim*,§,∥ †
Department of Physics and Astronomy and Inter-University Semiconductor Research Center and ‡Department of Biophysics and Chemical Biology, Seoul National University, Seoul 151-747, Republic of Korea § Department of Physics and ∥Department of Energy Systems Research, Ajou University, Suwon 443-749, Republic of Korea S Supporting Information *
ABSTRACT: Novel concepts for manipulating plasmonic resonances and the biocompatibility of plasmonic devices offer great potential in versatile applications involving real-time and in vivo monitoring of analytes with high sensitivity in biomedical and biological research. Here we report a biocompatible and highly tunable plasmonic bio/ chemical sensor consisting of a natural silk protein and a gold nanostructure. Our silk plasmonic absorber sensor (SPAS) takes advantage of the strong local field enhancement in the metal− insulator−metal resonator in which silk protein is used as an insulating spacer and substrate. The silk insulating spacer has hydrogel properties and therefore exhibits a controllable swelling when exposed to water−alcohol mixtures. We experimentally and numerically show that drastic spectral shifts in reflectance minima arise from the changing physical volume and refractive index of the silk spacer during swelling. Furthermore, we apply this SPAS device as a glucose sensor with a very high sensitivity of 1200 nm/RIU (refractive index units) and high relative intensity change. KEYWORDS: Silk fibroin hydrogel, surface plasmon, biocompatible device, plasmonic sensor, metal−insulator−metal resonators
F
to couple incident light to metallic surfaces.13−15 It is wellknown that the figure of merit (FOM), defined as the RI sensitivity divided by the resonance line-width, of LSPR sensors are 1 to 2 orders of magnitude lower than PSPR sensors.16,17 However, even though the FOM of LSPR sensors has approached that for PSPR sensors in a recent study, the FOM values for SPR sensors cannot overcome a theoretical upper limit of around 100, which corresponds to a 10 nm resonance shift of the 10 nm line-width spectral dip/peak for a 0.01 RI change.18 Considering that obtaining an extremely narrow spectral response in SPR resonators is impossible due to the intrinsic loss of metals, the limited FOM values means that highly sensitive and stable measurement systems are required to detect traces of analytes. Herein, we report the design and demonstration of a highly tunable LSPR sensor by utilizing the concept of a metal− insulator−metal (MIM) absorber. The thin insulator layer of the MIM sensor is composed of biocompatible and biodegradable silk fibroin hydrogel, which can accommodate water molecules by up to 60% in volume. Because the optical response of the proposed MIM structure is drastically sensitive to changes in thickness and RI of the thin insulator layer, the hydrogel properties of the silk insulator layer facilitate a high sensitivity to analytes through the control of the swelling ratio (thickness of the layer) and the mean RI of the silk hydrogel.
or point-of-care clinical evaluation, next generation opticsbased sensors are ideally suited as biomicroelectromechanical systems (BioMEMS), particularly in terms of microfluidics.1 In place of traditional MEMS materials such as silicon and polydimethylsiloxane (PDMS), the use of degradable biopolymers such as gelatin, poly(L-lactic acid) (PLA), and alginate allow for implantable BioMEMS devices to satisfy the growing demand for in vivo applications.2−4 Silk fibroin, the natural protein extracted from the Bombyx mori cocoon, is an interesting material in the bio-optics field due to its biocompatibility and unique mechanical and optical characteristics.5,6 Silk fibroin has been applied as a substrate material for biodegradable devices and a carrier for biodopants such as enzymes and antibodies with a biological function.7 The favorable combination of noble metals (especially gold) and silk fibroin promises a fully biocompatible plasmonic bio/chemical sensor platform that can be implanted in living tissue with no immune response. The surface plasmon resonances (SPRs) in this combination of materials also have many interesting features relevant for bio/chemical sensing. SPR-based sensors have found wide use in medical diagnosis, food safety, and environmental monitoring, due to their low cost and high sensitivity.8−10 Commercially available SPR sensors are based on propagating surface plasmon resonances (PSPRs), which can be generated on a flat metallic surface.11,12 Localized surface plasmon resonances (LSPRs) carried by metal nanostructures are a means to measure local refractive index (RI) changes at the nanoscale with a simplified measurement system without the need for prisms or gratings © XXXX American Chemical Society
Received: February 18, 2015 Revised: March 24, 2015
A
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 1. A fully biocompatible SPAS. (a) Working principle of the SPAS. The top layer is a 2D gold disk array and the bottom layer is a gold mirror. The diameter, thickness, and lattice constant of the gold disks are 200, 16, and 520 nm, respectively. The two layers are separated by a silk spacer layer with a thickness of 20 nm. The silk spacer absorbs the environmental liquid, thereby inducing the swelling. The changed volume and refractive index of the swelled silk layer affects the surface plasmon resonance behavior that results in the absorption of the incident wave. (b) The fabrication process of the SPAS. An arbitrary substrate such as a free-standing silk film and rigid silicon can be used to support the SPAS. (c) The SPAS (diffracted colors) on a chicken breast tissue. An optical fiber is adjacent to the SPAS surface for measurements. (d,e) Scanning electron microscopy and atomic force microscopy images of the fabricated SPAS. Scale bars represent 2 μm for (c) and 1 μm for (d).
Such an effect is not expected in conventional SPR sensors, which only use RI changes at metal−dielectric interfaces. We achieve a resonance shift of over 100 nm between a water environment and an isopropyl alcohol (IPA) environment despite the very small RI difference of these two liquids. Furthermore, the resonance shift exhibits a composition dependence. In addition, a network of silk polymer chains can act as a fluidic channel for the flow of analytes in water through a nanometer-sized layer. The proposed silk plasmonic structure is applied as a glucose sensor with higher sensitivity than other plasmonic glucose sensors, even though the aqueous glucose solution only affects the mean RI of the silk insulator spacer. Figure 1a shows a schematic diagram and the operating principle of the proposed silk plasmonic absorber sensor (SPAS). A two-dimensional (2D) gold disk array is stacked above a 200 nm thick gold film. The two gold layers are separated by a 20 nm silk spacer, as depicted in Figure 1a. Each gold nanodisk forms a MIM resonator with a silk spacer and a bottom gold layer. The coupled resonance in the periodic MIM induces a strong absorption of incident light and allows for the detection of absorption features in reflectance measurements.
Such a measurement is possible because the transmittance of the structure is totally eliminated due to the thick gold mirror, along with the polarization independence in the x- and ydirections at normal incidence.19 It is noteworthy that the core element of the SPAS is the silk hydrogel spacer. A simple methanol or water vapor treatment can ultimately increase the β-sheet crystallinity by causing hydrogen bond formation between silk molecules, thereby making the silk film a hydrogel.20,21 When we take advantage of the hydrogel property of the silk spacer (swelling by water-uptake), the silk insulator nanogap of the MIM structure changes its physical volume and RI and these induce drastic changes in the resonance mode. Figure 1b shows a diagram of the fabrication process for the SPAS. A free-standing silk film was used as a substrate. A thick gold film was subsequently deposited and the diluted silk solution was spin-coated to form a thin silk spacer. The gold nanodisk array was defined with a positive resist using standard electron beam lithography followed by a lift-off process. The finalized free-standing SPAS can be affixed on various surfaces including biological tissues, as shown in Figure 1c, for sensing. The substrate of the SPAS can be replaced as necessary with B
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 2. Measurements and simulations of the SPAS. (a) Schematic showing the butt-end fiber coupling measurement setup using a 1 × 2 multiplexing multimode fiber coupler. (b) Experimental reflectance spectra when the SPAS is immersed in air (black solid), IPA (red solid), and water (blue solid). (c) Simulated electric field intensity distributions for the reflection dip at the linear scale. The electric field is strongly localized at the silk spacer. (d) Simulated reflectance spectra for the structure in air (black curve), IPA (red curve), and water (blue solid). To emphasize the effect of the swelling on the resonance behavior, a simulated spectrum (blue dashed) in which the swelling is not considered is plotted in (d).
fiber coupling setup utilized by a 1 × 2 multiplexing multimode fiber coupler, as shown in Figure 2a.23,24 A white light source was fed into the input port and reached the SPAS through the cleaved fiber-tip. The reflected signal was collected through the same fiber tip and sent to the output port to be recorded. The implementation of a fiber coupler allows for a large alignment tolerance and avoids the geometrical restriction involved with adapting free-space optics when adjoined vertically emitted/ absorptive optical devices such as surface-emitting lasers and the SPAS.24 Figure 2b shows the measured reflectance-spectra
rigid materials such as silicon and glass. Figure 1d,e shows scanning electron microscopy (SEM) and atomic force microscopy (AFM) images of a fabricated sample with a 520 nm lattice constant (Λ), a 200 nm radius, and a 16 nm thick gold cap. All these dimensions were measured from the measured images. The strong adhesion between amino acids in the silk fibroin and gold further promotes a reliable fabrication method for integration with plasmonic nanostructures.22 A simple and efficient measurement scheme would make LSPR-based sensors more valuable. We employed a butt-end C
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
In Figure 3, we show spectral changes when the SPAS is immersed in IPA−water mixtures with varying compositions
of a proof-of-principle experiment. To confirm the tuning of the plasmonic resonance by the swelling of the silk hydrogel spacer, reflectance spectra of the SPAS were measured in air (n = 1.00), IPA (n = 1.37), and water (n = 1.32). As expected, the reflectance dip for the air-environment was red shifted when immersed in IPA due to the increasing RI of the environmental medium. When immerged in water, however, the plasmonic resonance reveals an unprecedented phenomenon, a blue shift of ∼100 nm, which is not expected from the RI change. This means that additional physical quantities affect the plasmonic resonance behavior. To investigate the optical response of the SPAS, we performed numerical simulations using a finite-difference time-domain (FDTD) method. Numerical structural input data from AFM topography data such as the radius and the height of the gold caps were used in FDTD simulations to compensate for any fabrication imperfections. As depicted in Figure 2c, the electromagnetic field is strongly localized in the intermediate silk spacer, thereby giving rise to a noticeable reflection dip in the spectrum with nearly zero intensity (Figure 2d). One interesting feature of the localized plasmonic mode is that the position of the reflection minimum is insensitive to the incident angle.19 This observation implies that even when laying on a curved surface, the SPAS will possess a reflection minimum at the same wavelength and therefore can be used as an attachable biosensor for living tissue. It is noteworthy that the experimental reflection minimum does not exhibit a nearly zero intensity, estimated in the simulations, because the excitation wave through the cleaved fiber tip has mode profiles different from the plane wave used in the simulations. For air and IPA environments, the simulated and experimental spectra agree well with each other with respect to the prediction for the position of the reflection minimum. The wavelength of the reflection minimum for IPA in the experimental spectrum is a little shorter than that in the simulated spectrum. This difference is because the polymer network of the silk spacer can also absorb a small amount of IPA molecules, as we will discuss later. For a water environment medium, we used two simulated spectra to examine the effect of the swelling on the plasmonic resonance. The spectrum in which the effect of the swelling is not taken into account is shown by the blue dotted curve. The blue solid curve in the same figure corresponds to the spectrum with swelling taken into account. In the simulation, a 30% thicker and a 60% water-containing silk spacer was considered, corresponding to a reduced RI of 1.4, from a previous study for silk inverse opal.25 In addition, we examined the effects of other individual physical quantities on the plasmonic resonance (see Supporting Information). Three physical quantities, the RI of the environment medium, the RI of the silk spacer, and the thickness of the silk spacer, were considered. Other quantities were fixed when a particular physical quantity was changed in the simulation. Regarding the RI of the silk spacer, the simulated sensitivity in terms of wavelength shift per refractive index unit (RIU) is 626 nm/ RIU, more than double that of the RI change of the environment medium (269 nm/RIU). Along with RI changes, the increase of the thickness of the silk spacer leads to a drastic resonance shift of 10.7 nm per 1 nm increase in thickness. These results underpin the fact that the tuning of the resonator itself can tremendously affect the resonance behavior, allowing the demonstration of a bio/chemical sensor with high sensitivity.
Figure 3. Plasmonic resonance behavior of the SPAS for water−IPA mixtures. (a) Reflectance spectra of the SPAS immersed in water−IPA mixture solutions with varying compositions. (b) Relations between the wavelength of the reflection minimum and the volume fraction of water, ϕ. The error bar represents the standard deviation calculated from five data points measured at each concentration.
and therefore different swelling ratios. With increasing water concentration, reflection minima undergo a blue shift and become narrower while becoming stronger absorbers. The stronger and narrower absorption can be described with a simple model by considering impedance matching between the environmental medium and the SPAS.19 The expansion and lower RI of the water-absorbed silk spacer make its complex impedance closer to that of the water environment, thereby resulting in enhanced absorption. Figure 3b shows the variation of the spectral position of the reflection minimum against the volume fraction of water, ϕ. Swelling of the silk spacer starts at ϕ = 0, thereby inducing a blue shift in the reflection minimum. An interesting and baffling property is that the spectral shift is reversed to a small red shift after ϕ = 0.5. This phenomenon cannot be explained by the RI change of the mixture with increasing ϕ because the RI of water is lower than that of IPA, and therefore indicates that the swelling of the silk spacer at ϕ = 0.5 (water/IPA = 1:1) is slightly larger than that at ϕ = 1.0 (pure water). The swelling ratios of the silk hydrogel for specific solvents were correlated with the relative magnitude of the polarity of the solvent. Table 1 shows the spectral shifts D
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
using water−methanol mixtures, which reveals a higher crystallinity in the mixture relative to that in pure methanol.27 According to this explanation for increasing expansion in IPA− water mixtures, the extra expansion should be inversely proportional to the polarities of the organic solvents. Supporting Information Figure S4 shows the variation of the spectral position of the reflection minimum against ϕ in aqueous acetone solutions, indicating that liquids with a lower polarity lead to a higher expansion. To investigate the potential for plasmonic bio/chemical sensing, the SPAS was employed to monitor glucose concentrations, one of the major challenges in the management of diabetes.28 In the present case, the physical volume of the silk spacer is not influenced by glucose concentrations because there is no causal factor for glucose molecules to stretch the network of pure silk proteins. However, the absorption of the glucose solution in the silk spacer increases the RI of the insulator layer, thereby leading to higher sensitivity relative to that for a similar structure consisting of rigid materials. Figure 4a shows the experimental reflected intensity spectra. An increasing glucose concentration enhances the RI of the aqueous solution and results in a red-shifted spectrum, along with relative intensity changes. The performance of the glucose sensor was analyzed by plotting the experimental wavelength shift (Δλ) as a function of glucose concentration in Figure 4b. A wavelength sensitivity of 1200 nm/RIU was obtained using the converted RI values,29 a value 4 times higher than the simulated value (269 nm/RIU) for the same structure with the rigid spacer-layer. The SPAS has reusability and reversibility, which are essential for their useful and practical sensing applications (see Figure 5S in the Supporting Information). Interestingly, as shown in Supporting Information Figure 5S, the wavelength sensitivity could almost double when the diameter of gold nanodisks became larger. Our future work will be to optimize the parameters of the SPAS for the highest sensitivity. This combination of relative intensity change and very high wavelength sensitivity is useful for monitoring extremely small analyte quantities, such as glucose in saliva.19,30
relative to the spectrum measured in air and polarities for various solvents (water, methanol, IPA, and acetone) to support this correlation. Table 1. Solvent Polarity Dependence of the Resonant Wavelength Shift (Δλ) of the SPASa
a
solvent
relative polarity
resonant shift (Δλ) [nm]
water methanol IPA acetone
1.00 0.762 0.546 0.355
−37 −13 74 80
The quantity Δλ is defined relative to air.
We need to address how the 50:50 mixture of solvents with weak (IPA) and strong (water) polarity can induce a larger volume-expansion relative to that for the pure solvent with strong polarity. Organic solvents are attracted by water through hydrogen bonding interactions and can then induce a reentrant transition in the hydrogels that have a hydrophobic interaction in which water and organic solvents are both good solvents for swelling.26 The strength of the hydrogel collapse by the reentrant in the mixture with water is inversely proportional to the polarities of organic solvents, which means that lower polarity of the solvent might be able to induce stronger hydrogen bonding to water. Although the silk spacer expands rather than contracts in organic solvent mixtures with water, we believe that the hydrogen-bonding between water and organic solvent, for which the strength is related to the polarity, will aid in the understanding of the observed properties. As depicted in Supporting Information Figure S3, the diffusion rate of low polar liquids should be sensibly low. However, water can promote the penetration of low polar liquids since water molecules can transport low polarity molecules to the molecular network of the silk hydrogel. In turn, this penetration leads to additional swelling because of the relatively strong hydrogen bonding between water and organic solvent molecules. Our results are consistent with previous observations of the annealing treatment of amorphous silk membranes
Figure 4. Glucose detection. (a) Representative reflectance spectra of the same SPAS immersed in glucose aqueous solution at different concentrations. (b) Measured wavelength shifts as a function of glucose concentration. The line is a linear fit with the refractive index sensitivity determined to be 1200 nm/RIU. E
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters Additionally, flexible optical sensors can be adversely affected by fluctuations of signals with respect to the detecting position due to different curvatures and stretched/constricted lattice constants. We confirmed that the SPAS exhibited the incident angle- and lattice-constant-insensitive resonance (see Figure 6S in the Supporting Information), which provide a great advantage for detecting consistent signals from flexible or curved sensors in simulations and experiments. More importantly, reforming the silk molecules by tagging stimuliresponsive molecules (for example, glucose oxidase for glucose) will enable the controllable swelling of the silk spacer, thereby inducing a drastic change in plasmonic resonances.31 In conclusion, we reported the fabrication and characterization of a novel plasmonic sensor consisting of a gold resonating nanoabsorber and a silk biopolymer. Absorbed analytes in the silk spacer change various physical quantities and lead to the strong tunability of plasmonic resonances. Water−alcohol mixtures were used as stimuli to control the swelling ratios and the RIs of the silk spacer. The resulting spectral shifts were analyzed using numerical simulations. In addition, we investigated the spectral response of the SPAS when immersed in glucose aqueous solutions with varying concentrations. Although the volume expansion of the silk spacer was not expected here, the experimental results exhibited a very high wavelength sensitivity of 1200 nm/RIU, valuable for a highly sensitive glucose biosensor. The combination of silk protein and gold plasmonic structures enables fully biocompatible plasmonic biochemical sensors and an improved detection ability from the unique properties of silk. Such a SPAS demonstrated here can be used for real-time, extremely sensitive, and in vivo monitoring of analytes in body fluid. In addition, the observation of an incident angle- and latticeconstant-insensitive resonance would be useful for consistent sensing when implanting or attaching the device on curved tissues.
■
Korea. S.K. is also supported by the TJ Park Science Fellowship of the POSCO TJ Park Foundation.
■
ASSOCIATED CONTENT
S Supporting Information *
Detailed descriptions of the experimental procedures, hydrogel properties of silk fibroin, simulation results for effects of physical quantities on the plasmonic resonance, solventdependence of the resonance-shift, reusability and reversibility of the resonator, and curvature-insensitive feature of the resonance. This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
(1) Bashir, R. Adv. Drug Delivery Rev. 2004, 56, 1565−1586. (2) Yang, L.-J.; Lin, W.-Z.; Yao, T.-J.; Tai, Y.-C. Sens. Actuators, A 2003, 103, 284−290. (3) Grayson, A. C. R.; Choi, I. S.; Tyler, B. M.; Wang, P. P.; Brem, H.; Cima, M. J.; Langer, R. Nat. Mater. 2003, 2, 767−772. (4) Cabodi, M.; Choi, N. W.; Gleghorn, J. P.; Lee, C. S. D; Bonassar, L. J.; Stroock, A. D. J. Am. Chem. Soc. 2005, 127, 13788−13789. (5) Omenetto, F. G.; Kaplan, D. L. Nat. Photonics 2008, 2, 641−643. (6) Hwang, S.-W.; Tao, H.; Kim, D.-H.; Cheng, H. Y.; Song, J.-K.; Rill, E.; Brenckle, M. A.; Panilaitis, B.; Won, S. M.; Kim, Y.-S. Science 2012, 337, 1640−1644. (7) Pritchard, E. M.; Dennis, P. B.; Omenetto, F. G.; Naik, R. R.; Kaplan, D. L. Biopolymers 2012, 97, 479−498. (8) Shankaran, D. R.; Gobi, K. V.; Miura, N. Sens. Actuators, B 2007, 121, 158−177. (9) Brolo, A. G. Nat. Photonics 2012, 6, 709−713. (10) Zeng, S.; Baillargeat, D.; Ho, H.-P.; Yong, K.-T. Chem. Soc. Rev. 2014, 43, 3426−3452. (11) Liedber, B.; Nylander, C.; Lundstrom, I. Sens. Actuators 1983, 4, 299−304. (12) Mullett, W. M.; Lai, E. P. C.; Yeung, J. M. Methods 2000, 22, 77−91. (13) Pryce, I. M.; Kelaita, Y. A.; Aydin, K.; Atwater, H. A. ACS Nano 2011, 5, 8167−8174. (14) Kabashin, A. V.; Evans, P.; Pastkovsky, S.; Hendren, W.; Wurtz, G. A.; Atkinson, R.; Pollard, R.; Podolskiy, V. A.; Zayats, A. V. Nat. Mater. 2009, 8, 867−871. (15) Zhang, S. P.; Bao, K.; Halas, N. J.; Xu, H. X.; Nordlander, P. Nano Lett. 2011, 11, 1657−1663. (16) Sherry, L. J.; Chang, S. H.; Schatz, G. C.; Van Duyne, R. P.; Wiley, B. J.; Xia, Y. N. Nano Lett. 2005, 5, 2034−2038. (17) Ameling, R.; Langguth, L.; Hentschel, M.; Mesch, M.; Braun, P. V.; Giessen, H. Appl. Phys. Lett. 2010, 97, 253116. (18) Shen, Y.; Zhou, J. H.; Liu, T. R.; Tao, Y. T.; Jiang, R. B.; Liu, M. X.; Xiao, G. H.; Zhu, J. H.; Zhou, Z.-K.; Wang, X. H. Nat. Commun. 2013, 4, 2381. (19) Liu, N.; Mesch, M.; Weiss, T.; Hentschel, M.; Giessen, H. Nano Lett. 2010, 10, 2342−2348. (20) Hu, X.; Shmelev, K.; Sun, L.; Gil, E.-S.; Park, S.-H.; Cebe, P.; Kaplan, D. L. Biomacromolecules 2011, 12, 1686−1696. (21) Jin, H.-J.; Park, J.; Karageorgiou, V.; Kim, U. J.; Valluzzi, R.; Kaplan, D. L. Adv. Funct. Mater. 2005, 15, 1241−1247. (22) Park, J.; Choi, Y.; Lee, M.; Jeon, H.; Kim, S. Nanoscale 2015, 7, 426−431. (23) Kim, S.; Lee, J.; Jeon, H.; Kim, H.-J. Appl. Phys. Lett. 2009, 94, 133503. (24) Kim, S.; Park, Y.; Hwang, K.; Lee, J.; Jeon, H.; Kim, H. J. J. Opt. Soc. Am. B 2009, 26, 1330−1333. (25) Kim, S.; Mitropoulos, A. N.; Spitzberg, J. D.; Tao, H.; Kaplan, D. L.; Omenetto, F. G. Nat. Photonics 2012, 6, 818−823. (26) Orakdogen, N.; Okay, O. Polymer 2006, 47, 561−568. (27) Tsukada, M.; Gotoh, Y.; Nagura, M.; Minoura, N.; Kasai, N.; Freddi, G. J. Polym. Sci., Part B: Polym. Phys. 1994, 32, 961−968. (28) Schmelzeisen-Redeker, G.; Staib, A.; Strasser, M.; Müller, U.; Schoemaker, M. J. Diabetes Sci. Technol. 2013, 7, 808−814. (29) Muhmood, W.; Yunus, M.; Rahman, A. Appl. Opt. 1988, 27, 3341−3343. (30) Feng, J.; Siu, V. S.; Roelke, A.; Mehta, V.; Rhieu, S. Y.; Palmore, G. T. R.; Pacifici, D. Nano Lett. 2012, 12, 602−609. (31) Kost, J.; Horbett, T. A.; Ratner, B. D.; Singh, M. J. Biomed. Mater. Res., Part A 1985, 19, 1117−1133.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors acknowledge support from the Basic Science Program (NRF-2014R1A1A1008080) and Nano-Material Technology Development Program (2009-0082580) through the National Research Foundation (NRF) and the NRF Grant (2008-0061906) funded by the MSIP and MEST, Republic of F
DOI: 10.1021/acs.nanolett.5b00680 Nano Lett. XXXX, XXX, XXX−XXX
