On-Chip Narrowband Thermal Emitter for Mid-IR Optical Gas Sensing

Apr 17, 2017 - On-Chip Narrowband Thermal Emitter for Mid-IR Optical Gas Sensing .... Journal of the Optical Society of America B 2018 35 (7), 1490 ...
0 downloads 0 Views 1MB Size
Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI

Article

On-Chip Narrowband Thermal Emitter for Mid-IR Optical Gas Sensing Alexander Lochbaum, Yuriy Fedoryshyn, Alexander Dorodnyy, Ueli Koch, Christian Hafner, and Juerg Leuthold ACS Photonics, Just Accepted Manuscript • DOI: 10.1021/acsphotonics.6b01025 • Publication Date (Web): 17 Apr 2017 Downloaded from http://pubs.acs.org on April 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Photonics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

On-Chip Narrowband Thermal Emitter for Mid-IR Optical Gas Sensing

Alexander Lochbaum*, Yuriy Fedoryshyn, Alexander Dorodnyy, Ueli Koch, Christian Hafner, and Juerg Leuthold

Institute of Electromagnetic Fields, ETH Zurich, 8092 Zurich, Switzerland

*Email: [email protected]

Abstract Efficient light sources compatible to complementary metal oxide semiconductor (CMOS) technology are key components for low-cost, compact mid-infrared gas sensing systems. In this work we present an on-chip narrowband thermal light source for the mid-infrared wavelength range by combining microelectromechanical system (MEMS) heaters with metamaterial perfect emitter structures. Exhibiting a resonance quality factor of 15.7 at the center wavelength of 3.96 μm and an emissivity of 0.99, the demonstrated emitter is a spectrally narrow and efficient light source. We show temperaturestable (resonance wavelength shift 0.04 nm/°C) and angular-independent emission characteristics up to angles of 50°, and provide an equivalent circuit model illustrating the structure’s resonance behavior. Owing to its spectrally tailored, non-dispersive emission, additional filter elements in a free-space optical gas sensing setup become obsolete. In a proof-of-concept demonstration of such a filter-free gas sensing system with CO2 concentrations in the range of 0 - 50000 ppm, we observe a 5-fold increase in relative sensitivity compared to the use of a conventional blackbody emitter. Our light source is fully compatible with standard CMOS processes and tunable in emission wavelength through the midinfrared wavelength band. It paves the way for a new class of highly integrated, low-cost optical gas sensors.

Keywords mid-infrared sensing, metamaterials, thermal emission, optical sensing, CMOS technology Mid-infrared (mid-IR) optical gas sensing is a robust chemical sensing method with numerous applications in industrial process control, building automation and environmental sensing1. Many molecules exhibit strong absorption features at mid-IR wavelengths, which yield characteristic absorption line patterns2-3. Currently, emerging technologies like indoor air quality monitoring for home automation, and CO2-refrigerated air-conditioning systems for automotive drive the demand for sensitive, low-cost, and highly compact sensing systems. However, their development is hindered by the lack of CMOS-compatible, spectrally efficient, and compact mid-IR light sources. Mid-IR light sources for sensing applications range from simple microbulbs4 to high-end quantum cascade lasers (QCL)5. Within the last decade, QCL developed into the standard for light sources in mid-IR spectroscopy due to their high spectral density and single-mode operation, which allows for

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

probing of individual absorption lines to quantify chemical species6-7. Nevertheless, the complex and costly fabrication of these outstanding mid-IR light sources limits their viability for low cost optical gas sensing applications in near future. Fortunately, for most of the aforementioned applications, it is sufficient to distinguish between entire absorption bands in a non-dispersive infrared (NDIR) configuration, which relaxes the requirements on the light source emission bandwidth1. For that reason, membrane heaters based on microelectromechanical system (MEMS) technology recently came up as compact, integrated thermal light sources8-10. MEMS heaters are proven to be energy efficient8, allow for rapid modulation owing to their low thermal mass9, and are compatible with standard CMOS foundry processes10. Unfortunately, standard CMOS materials exhibit inherently low emissivity values and broadband emission characteristics following Planck’s law, which makes additional blackening layers and (discrete) filter elements necessary. This impairs the source’s compatibility with standard fabrication processes and increases the overall size and cost of the sensing system. To overcome these limitations, thermal emission engineering has been proposed as a means to enhance the brightness and tailor the emission spectrum of thermal light sources11-12. By deliberate choice of the thermal emitter’s structure and its constituent materials, optical resonances within the emitter are excited that enhance its emissivity over a limited wavelength range to values close to unity. Various concepts for thermal emission and absorption engineering have been demonstrated in recent years. Among these are plasmonic gratings13-15, photonic crystals16, multi-quantum well structures17-18, and metamaterial emitters19-20. With regard to optical sensing applications, metamaterial emitters are of particular interest, as they enable for single- and multi-band21-24, as well as polarization- and angle-independent25-27 emission while being compatible with low-cost fabrication processes28-29. Yet, a narrowband on-chip thermal light source, which is CMOS-compatible and thermally stable up to typical application temperatures of 350 °C, is still missing. In this work, we demonstrate a highly efficient, narrowband on-chip thermal emitter for the mid-IR wavelength range. We show selectively enhanced thermal emission at a wavelength of 3.96 μm with emissivity values up to 0.99 over a narrow spectral bandwidth (FWHM = 252 nm, 𝑄 = 15.7 ) by combining MEMS heater technology with metamaterial emitter structures. The high resonance quality factor in comparison to other metamaterial emitters is enabled by an optimized resonator geometry in combination with low-loss constituent materials. An intuitive equivalent circuit model of the metamaterial structure provides insights into the emitter’s resonance behavior, and experimental characterization results match very well with our numerical findings. Finally, we demonstrate the emitter’s performance as a light source for filter-free NDIR gas measurements in a proof-of-concept experiment.

On-Chip Metamaterial Thermal Emitter The structure of the metamaterial mid-IR light source is illustrated in an exploded view in Figure 1a. From bottom to top, it consists of a CMOS substrate, a MEMS hotplate, a metamaterial perfect emitter (MPE) structure30, and an Al2O3 sealing layer. The CMOS substrate, comprising integrated electronic components, is realized in a commercially available foundry process. On top of the CMOS substrate, the MEMS hotplate is realized as a circular, dielectric membrane with integrated metal heater. The employed design allows for resistive heating of the hotplate up to temperatures of 450 °C in the center of the membrane, while the substrate remains at room temperature. On top of the MEMS hotplate, the circular MPE area is structured for wavelength-selective emissivity enhancement. The metamaterial multilayer structure is enclosed by an Al2O3 sealing layer to prevent oxidation of the metal structures ACS Paragon Plus Environment

Page 2 of 20

Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

upon repeated thermal cycling. The structure is operated in ambient atmosphere without additional packaging.

Figure 1. Concept of metamaterial perfect emitter (MPE) light source. (a) Illustration of the MPE light source in an exploded view (dimensions not to scale). The MPE structure is deposited in a series of post-processing steps on top of a MEMS hotplate. (b) Illustration of the MPE unit cell with indicated field lines at resonance (electric field E: red arrows, magnetic field H: blue arrows). For illustration purposes, the cross-shaped top resonator is cut along the 𝑥𝑧-plane, and the front part is solely sketched as a wireframe. Critical structure dimensions are outlined in the illustration. The intensity plots of the magnetic field (c), the electric field (d), as well as the power loss density (e) at the resonance wavelength illustrate the highly localized character of the MPE resonance.

The cross-section profile of the MPE unit cell is shown in Figure 1b. The unit cell comprises a crossshaped top resonator (Cu), which is separated from a metal backplane (Cu) by means of a dielectric spacer layer (Al2O3). The metamaterial's optical properties can be accessed by taking a closer look at the resonance behavior of its constituent unit cell. At the resonance frequency, the magnetic field H couples inductively to the metal-insulator-metal (MIM) structure and is strongly confined under the center of the top resonator, see Figure 1c. The resonantly enhanced magnetic field induces antiparallel currents in the top resonator and the metal backplane. The resulting current loop is closed by the electric field E, which drops off in 𝑧-direction across the dielectric spacer, see Figure 1d. Hence, the top resonator forms together with the backplane a symmetric version of a split-ring-resonator (SRR)31 - an artificial magnetic atom32. Applying effective medium theory, the macroscopic response of the metamaterial to electromagnetic waves can be homogenized, and the MPE is modeled as a quasihomogeneous slab with effective material parameters 𝜀eff and 𝜇eff 33. The inductively-coupled resonance described above results in a resonant effective permeability 𝜇eff . Similarly, the top resonator entails a resonant effective permittivity 𝜀eff . Hence, the unit cell geometry defines the effective material parameters of the metamaterial, which is used to tailor its emission properties. ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 20

To maximize the emissivity of the structure, one can conversely maximize its absorptivity by virtue of Kirchoff’s law. The absorptivity 𝐴 = 1 − 𝑅 − 𝑇 approaches unity if both the transmissivity 𝑇 and the reflectivity 𝑅 vanish. Because the thickness of the metal backplane is large with respect to the penetration depth at mid-IR frequencies, the transmissivity through the metamaterial is 𝑇 = 0. The reflectivity is minimized by adapting the structure geometry such that 𝜇eff and 𝜀eff are equal, in which case the metamaterial’s wave impedance at the resonance frequency is matched to free space, ′′ (𝜔 𝜇eff (𝜔res ) 𝜇′ (𝜔res ) + 𝑗𝜇eff res ) 𝑍(𝜔res ) = √ = √ eff = 𝑍0 , ′ ′′ 𝜀eff (𝜔res ) 𝜀eff (𝜔res ) + 𝑗𝜀eff (𝜔res )

(1)

and hence 𝑅 = 0. To further investigate the underlying loss mechanism, the dissipated power density d𝑃d ⁄d𝑉 = 𝜎(𝐄 ⋅ 𝐄∗ )/2 within the metamaterial unit cell is depicted in Figure 1e. The majority of the energy dissipation takes place in the metal layers, at the interface to the dielectric spacer layer (Supporting Information). Thus, the quality factor of the structure resonance is primarily influenced by the loss characteristics of the constituent metal, and copper is chosen as a low-loss, CMOS-compatible option.

Results and Discussion Design and simulation The metamaterial unit cell dimensions have been optimized for maximum emissivity at the resonance wavelength 𝜆res = 4.04 μm by full-wave FEM simulations with periodic boundary conditions. As optimal device parameters, we calculated a periodicity 𝑎 = 2.243 μm, top resonator length 𝑙 = 963 nm, top resonator width 𝑤 = 187 nm, and dielectric thickness 𝑡diel = 116 nm for fixed thicknesses of the metal and sealing layers, 𝑡metal = 𝑡sealing = 100 nm. The simulated reflection, transmission and absorption characteristics are shown in Figure 2a. As expected, the transmissivity is zero across the investigated frequency range. The absorptivity is resonantly enhanced to 𝐴 = 0.99 at the resonance wavelength 𝜆res with a full width at half maximum FWHM = 254 nm, which corresponds to a quality factor 𝑄 = 𝜆res⁄FWHM = 15.9. To the best of our knowledge, this is the first time a metamaterial absorber or emitter approaches bandwidths that are sufficiently narrow for non-dispersive gas sensing applications. The comparably narrowband resonance in comparison to typical MIM quality factors 𝑄 ≈ 10 - 1121, 23 can be attributed to a combination of low-loss constituent materials and an efficient top resonator geometry34.

ACS Paragon Plus Environment

Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Figure 2. Simulated structure characteristics. In all plots, real and imaginary parts of complex-valued quantities are denoted by solid and dashed lines, respectively. The resonance wavelength 𝜆res = 4.04 μm is marked by a vertical, dashed line. (a) Simulated absorption, reflection and transmission characteristics of the MPE unit cell. At the resonance wavelength, the structure exhibits a narrowband absorption resonance with peak absorptivity 𝐴 = 0.99 and full width at half maximum FWHM = 254 nm. (b) Relative effective permittivity 𝜀eff and permeability 𝜇eff of the MPE structure, as derived from the parameter retrieval procedure. (c) Complex impedance of MPE structure. The purple and green lines denote the results derived from the parameter retrieval

procedure 𝑍PR ,

and

from the equivalent

circuit model

𝑍RLC ,

respectively.

(d) Equivalent circuit model of the MPE unit cell, representative of the structure’s resonance characteristics.

To further investigate the resonant behavior of the metamaterial structure and its matching to free space, the effective material parameters 𝜀eff and 𝜇eff as well as the structure impedance 𝑍PR have been retrieved from the numerical simulation results by inversion of the scattering parameters33. The retrieved complex material parameters are plotted in Figure 2b. Both parameters exhibit Lorentzian-shaped resonances with shifted center frequencies, matching their real and imaginary parts to each other at the resonance frequency as demanded from Eq. 1. In the wavelength range between 3.96 μm and 4.05 μm, the MPE exhibits a negative permeability, Re 𝜇eff < 0, indicative of the metamaterial’s large magnetic response. In Figure 2c, the retrieved structure impedance 𝑍PR is shown. It possesses an inductive character over most of the analyzed frequency range, Im 𝑍PR < 0. Close to the resonance frequency, the structure

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

impedance exhibits a resonance and matches almost ideally to free 𝑍(𝜔res ) = √𝜇eff (𝜔res )⁄𝜀eff (𝜔res ) = (1.14 + 𝑖0.17) ⋅ 𝑍0 is sufficiently close to 𝑍0 .

Page 6 of 20

space,

i.e.

Alternatively, the structure’s impedance can be expressed by an equivalent circuit model, which is sketched into the metamaterial unit cell in Figure 2d. The induced loop currents in top resonator and backplane are modeled by a series connection of an inductor and a resistor (𝐿1 , 𝑅1 and 𝐿2 , 𝑅2 ), also taking into account the energy dissipation within the metal layers. The electric field, which is strongly enhanced between top resonator and backplane, is modeled by two identical capacitors 𝐶12 connecting the inductive elements. Due to the highly localized nature of the structure resonance, coupling to neighboring unit cells is negligible in this simplified model. Numerical values for the circuit elements were obtained by a nonlinear least-squares fit of the total equivalent circuit impedance 𝑍RLC to the retrieved structure impedance 𝑍PR . The extracted parameters result to 𝑅1 = 220 mΩ, 𝐿1 = 17.9 fH, 𝐶12 = 0.49 fF, 𝑅2 = 12.7 mΩ, and 𝐿2 = 1.3 fH. As expected from the high field intensities around the top resonator element, its impedance (denoted by 𝑅1 and 𝐿1 ) dominates over the respective ground plane values (𝑅2 and 𝐿2 ). The imbalance between the inductive elements 𝐿1 and 𝐿2 results in the overall inductive character of the MPE impedance within the investigated frequency range. A very good agreement of the fit results with the retrieved structure impedance confirms the model’s validity. It should be noted that the model is only valid up to first-order resonances. Higher order modes and grating resonances need to be taken into account by additional circuit elements34. Fabrication Figure 3a displays the metamaterial thermal light source (total size 1.7 mm × 1.7 mm), bonded to a TO-39 header for testing purposes. The MPE structure has been fabricated in a series of post processing steps on a MEMS hotplate. Figure 3b shows a micrograph of the fabricated thermal emitter; a magnified detail of the MPE area is depicted in Figure 3c. The MPE area (𝑑MPE = 400 μm) is centered on the suspended membrane (𝑑mem = 600 μm), incorporating ~ 2.5 × 104 MPE unit cells. The heater structure is embedded into the dielectric membrane and centered underneath the MPE structure. In Figure 3d, a scanning electron microscope (SEM) image of 2 × 2 MPE unit cells is shown. Actual structure dimensions have been measured from the SEM micrographs and verified against the nominal values. In addition to the MEMS hotplate emitters, three MPE samples on silicon substrates were fabricated for the purpose of optical characterization; each with an MPE area of 4 mm × 4 mm, containing ~ 3.2 × 106 unit cells.

Figure 3. (a) MPE light source bonded to TO-39 header for device tests. (b) Micrograph of circular MEMS membrane (𝑑mem = 600 μm) with centered MPE area (𝑑MPE = 400 μm). (c) Detail micrograph of MEMS membrane and MPE area, where individual MPE unit cells can be identified. (d) Scanning electron microscope (SEM) picture of 2 × 2 MPE unit cells with structure geometry parameters.

ACS Paragon Plus Environment

Page 7 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Characterization Absorptivity and emissivity measurements have been performed to access the metamaterial’s optical properties. Measurement results presented in Figures 4-5 have been acquired from silicon substrate samples. Room-temperature reflectivity measurements were carried out in the wavelength range from 3 μm to 9 μm by Fourier-transformed infrared (FTIR) spectroscopy in combination with a mid-IR microscope. From the results, absorptivity curves were calculated and compared to the simulated absorption spectra.

Figure 4. (a) Measured absorptivity and emissivity characteristics of the MPE structure (solid lines) in very good agreement to simulation results (dashed line). (b) Narrowband spectral exitance of MPE (solid lines) approaching the limit of a reference blackbody (dashed lines) for all investigated temperatures. Gray lines: spectral exitance at peak emission wavelength versus peak emission wavelength. The peak emission wavelength of the MPE is nearly independent of temperature (i.e. the solid gray line is nearly vertical), whereas the position of the blackbody’s peak emission exhibits a considerable temperature dependence (dashed gray line). (c) Comparison of MPE absorptivity and emissivity curves at different temperatures, confirming its temperature-independent emission characteristics.

In Figure 4a, the measured absorptivity is shown together with the simulated absorption spectrum. The measured absorptivity reaches a maximum value 𝐴 = 0.99 at 𝜆res = 3.96 μm with a FWHM of 252 nm, corresponding to a quality factor 𝑄 = 15.7. Relaxing the imposed thermal stability requirements, quality factors up to 21.9 have been measured for samples without adhesion layer (Supporting Information). Outside the resonance, the structure maintains a very low background absorptivity (𝐴 < 0.04), rendering it ideal as a selective emitter. The experimental results are in very good agreement with the simulated values, with a shift of the resonance wavelength of only Δ𝜆res = 80 nm

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 20

(Δ𝜆res /𝜆res = 2 %) towards shorter wavelengths. The discrepancy between simulated and measured absorption spectrum might be attributed to the influence of the adhesion layer (5 nm Cr), which has been omitted in the simulation model. To the best of our knowledge, this is the first CMOS-compatible MPE structure with experimentally confirmed quality factors high enough for applications in NDIR gas sensing, where absorption bands in the mid-IR have FWHMs of ~ 100 - 200 nm1, 3. In view of the MPE’s use as a thermal light source, it is of central interest to directly characterize its emissivity at elevated temperatures. A temperature-controlled sample holder with a ceramic heating element was used to guarantee homogeneous heating of the samples. In Figure 4a, the measured emissivity of the MPE structure at 330 °C is shown. Its peak emissivity and resonance wavelength show very good agreement to the room-temperature absorptivity curve (𝐸 = 0.99, 𝜆res = 3.96 μm), but with a slightly larger resonance bandwidth, FWHM = 325 nm (𝑄 = 12.2). A possible explanation lies in the resistivity increase of the MPE’s metal portions with temperature, leading to increased losses and thus to a larger resonance bandwidth. To investigate the MPE structure’s thermal stability, its absorption spectrum after 10 hours of repeated thermal cycling between room temperature and 350 °C has been measured. A very good overlap with the initial absorptivity can be observed, confirming the MPE’s stability upon repeated thermal cycling. For a direct comparison of the MPE emission to an ideal blackbody, the spectral exitance of the MPE structure at various temperatures is plotted in Figure 4b together with the spectral exitance of a blackbody reference sample. For each temperature, the MPE structure reaches the blackbody exitance with a temperature-independent peak emissivity 𝐸 = 0.99. The resonance wavelength of the MPE structure is solely defined by its unit cell geometry and thus should be nearly independent of the sample temperature. For selective, narrowband thermal emission from a membrane heater, this aspect is of prime importance, as considerable temperature fluctuations across MEMS hotplates are inevitable35. In Figure 4b, the observable temperature dependence of the MPE’s resonance wavelength is indeed close to negligible, and a resonance wavelength shift Δ𝜆res = 6 nm over the investigated temperature range of 150 °C is extracted (Supporting Information). The resonance wavelength against temperature curve can be fitted linearly, which yields a resonance wavelength shift coefficient 𝜉λres ,𝑇 = 42 pm/°C. Division by the resonance wavelength at 350 °C yields a normalized value 𝜎λres =

𝜉λres ,T 𝜆res,350°C

= 1.07 × 10−5 °C −1 .

(2)

As the linear thermal expansion coefficient of Cu, 𝛼Cu = 1.65 × 10−5 °C−1 36, is very similar to this value, the temperature dependence of the resonance emission wavelength might be attributed to the thermal expansion of the MPE top resonator. In comparison, the blackbody reference sample exhibits over the investigated temperature range a peak wavelength shift Δ𝜆peak,BB = 1.47 μm in accordance with Wien’s displacement law. In Figure 4c, the extracted emissivity curves are compared against the measured absorptivity spectra for each temperature. Emissivity and absorptivity curves exhibit an excellent agreement over the investigated temperature range, validating Kirchhoff’s Law. Both the absorptivity and the emissivity curves exhibit a resonance bandwidth broadening with increasing temperature. In analogy to Eq. 2, one can calculate a normalized coefficient for the change of resonance bandwidth, equating for the absorptivity to 𝜎Δλabs = 𝜉Δλabs ,𝑇 / Δ𝜆abs,350°C = 5.8 × 10−4 °C −1 , and for the emissivity to 𝜎Δλemis = 𝜉Δλemis,𝑇 / Δ𝜆emis,350°C = 5.4 × 10−4 °C−1, respectively. A very good

ACS Paragon Plus Environment

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

agreement of the linear fit with the measured resonance bandwidths suggests the nature of the resonance broadening to be temperature-related. Angular-dependent emissivity For the intended applications in a free-space optical gas sensing system, an omnidirectional (Lambertian) emitter with spectrally selective, angular-independent emission characteristics is favorable to maximize the signal-to-noise ratio of the sensing system.

Figure 5. (a) Simulated angular-dependent emissivity of the MPE structure for TM polarization. Narrowband, angularindependent emission characteristics up to large emission angles can be observed, rendering the structure attractive for freespace gas sensing applications. A (dispersive) grating resonance with center wavelength 𝜆g = 𝑎(1 + sin(𝜃)) is identified, exhibiting an anti-crossing with the fundamental MPE resonance at 𝜃 ≈ 50°. (b) Experimental angular-dependent emissivity of the MPE structure, confirming the simulation results with excellent agreement.

In Figure 5, the MPE’s angular emission characteristics have been investigated, proving its polarizationand angle-independent, narrowband emissivity enhancement. In Figure 5a, the simulated angular emissivity for transverse magnetic (TM) polarization is displayed. Corresponding simulations for TE polarization yield similar results and can be found in the Supporting Information. The structure resonance at 𝜆res = 4.04 μm retains high emissivity values and a constant resonance wavelength up to emission angles of 𝜃 ≈ 50°. The reason for the dispersion-free emission characteristics lies in the strongly localized nature of the inductively-coupled MPE resonance25-27 (Figures 1c-d): for TM polarization, the magnetic field maintains its orientation with respect to the inductive loop even under oblique angles, resulting in the angular-independent emission characteristics observed in Figure 5a. At 𝜃 ≈ 50° a coupling of the dispersion-free, localized MPE resonance with the first-order grating resonance can be observed, leading to an anti-crossing behavior37-38. The grating resonance is dispersive and has its origins in the periodicity of the MPE structure. The center wavelength of a first-order grating ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 20

resonance is given by 𝜆g = 𝑎(1 + sin(𝜃))39, and the numerical results in Figure 5a show good agreement with the analytical expression. The grating resonance can be shifted out of the wavelength band of interest by appropriate choice of the unit cell periodicity 𝑎. Dispersion-free emission characteristics up to 𝜃 = 80° have been simulated employing a smaller periodicity, which comes at the price of a larger resonance bandwidth due to a stronger coupling between neighboring unit cells (Supporting Information). At 𝜆res,2 ≈ 3.25 μm, the second-order resonance of the MPE structure can be observed at oblique angles. Even-order resonances cannot be excited under normal incidence, as the current distribution within the MPE structure is symmetric, and the averaged magnetic moment25 across the structure is zero. At oblique angles the current distribution within the structure becomes asymmetric, which leads to a non-zero averaged magnetic moment across the structure, allowing for the excitation of even-order resonances. To verify the simulation results, emissivity measurements between 2.5 μm and 8 μm were performed at the external port of a modified FTIR spectrometer. The measurement results in Figure 5b show an excellent agreement with the numerical simulations. The angular-independent emission characteristics of the MPE structure could be experimentally verified, and even details like the grating resonance and the second-order MPE resonance at oblique angles were resolved. Tunability and robustness In order to assess the practicability of the proposed MPE sources, its resonance wavelength tunability as well as its robustness against fabrication tolerances have been investigated in an experimental parameter study. 11 × 11 MPE samples covering 𝑙 = 1003 nm +/- 50 nm and 𝑤 = 155 nm +/- 40 nm at 𝑎 = 2.432 μm have been fabricated. The resonance key figures as extracted from FTIR measurements are shown in Figure 6. By varying the top resonator length 𝑙, the MPE resonance wavelength shown in Figure 6a can be tuned across the mid-infrared wavelength range, which is of practical relevance for targeting different molecular absorption bands in the so-called “fingerprint region"3. The MPE emitter concept proves to be robust in peak absorptivity, Figure 6b, and quality factor, Figure 6c, with regard to systematic deviations in length and width of the top resonator on the order of 50 nm. This is of particular importance for the future transfer of the fabrication process to highvolume compatible lithography techniques with according tolerances, like deep-UV or nano-imprint lithography40.

Figure 6. Extracted resonance key figures for experimental parameter sweep. (a) The resonance wavelength scales linearly with the top resonator length. (b) And (c) peak absorptivities and quality factors retain high values 𝐴 > 0.92 and 𝑄 > 15.4, respectively, across the investigated parameter range.

ACS Paragon Plus Environment

Page 11 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Emission from membrane heater In the previous experiments, the MPE’s emission and absorption characteristics at elevated temperatures have been acquired from homogeneously heated silicon substrate samples. To investigate the influence of temperature variations across the resistively-heated MEMS hotplate on the MPE‘s emission characteristics, the total emission spectrum from a MPE-covered membrane emitter has been measured and verified against its room-temperature absorption characteristics. The temperature distribution across the resistively-heated MPE hotplate has been obtained from thermographic measurements. The temperature map of the MPE hotplate during continuous-mode operation with a heating power 𝑃hotplate = 49.4 mW is shown in Figure 7a. A maximum membrane temperature 𝑇max = 322 °C is measured in the center of the MPE area, point 1, with a temperature decrease to 𝑇 = 215 °C at the edge of the MPE area, point 2. The acquired spectral exitance of the MPE hotplate, together with the roomtemperature absorptivity of the same structure, is displayed in Figure 7b. Both graphs exhibit a very good qualitative overlap with a resonance wavelength shift of Δ𝜆res = 𝜆res,exit − 𝜆res,abs = 11 nm. The estimated relative coefficient Δ𝜆res /Δ𝑇 ≈ 42.6 pm/°C correlates well with results obtained during the emissivity / absorptivity comparison in Figure 4. The almost negligible temperature dependence of the MPE emission spectrum is of large practical relevance for MEMS hotplate emitters, as it considerably relaxes the temperature uniformity requirements across the membrane. In addition to the main resonance, a spectrally broad background emission with a local maximum around 𝜆 ≈ 9 μm can be observed in Figure 7b. This is attributed to the thermal emission of the dielectric membrane in areas outside of the MPE region (200 μm < 𝑟 ≤ 300 μm), which shows non-negligible mid-IR emission in Figure 7a as well. The main constituent material of the membrane, SiO2, exhibits a resonance around 9 um41, in support of this conclusion. The observed background emission could be minimized in future iterations by an optimized thermal design of the membrane.

Figure 7. (a) Temperature map of the resistively-heated MPE hotplate, derived from thermographic measurements. Temperature values are solely valid on the MPE surface, marked by the color-coded area. A maximum temperature 𝑇max = 322 °C

is

measured

in

the

center

of

the

membrane,

together

with

a

temperature

variation

Δ𝑇 = 107 °C across the MPE area. (b) Measured spectral exitance from MPE hotplate, compared to its room-temperature absorptivity. Both curves exhibit a very good overlap, confirming the temperature-independent, narrowband emission characteristics despite the large temperature variations found on MEMS hotplates.

Proof-of-concept experiment In a concluding system experiment, two key features of the metamaterial light source, on-chip integration and high spectral efficiency, are exploited for demonstration of millimeter-scale, filter-free NDIR gas sensing at the example of CO2. A schematic representation of the experimental setup is provided in Figure 8a. ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 20

Figure 8. (a) Schematic of the millimeter-scale, filter-free NDIR gas sensing setup. Inset: absorptivity spectrum of the MPE employed in the system experiment together with the attenuation spectra of atmospheric CO2 and H2O for 𝑐CO2 = 400 ppm, rH = 40%, 𝑇 = 25 °C, and 𝑃 = 1013 hPa. (b) Normalized voltage difference Δ𝑉/ 𝑉0ppm for the MPE and the BB system together with the applied CO2 concentration profile. (c) Normalized voltage difference against applied CO2 concentration for both systems. A 5-fold increase in relative sensitivity can be observed for the MPE system. (d) The measured sensitivity of the MPE system to water vapor decreases by a factor of 9 compared to the reference system employing a conventional blackbody.

The setup comprises a thermal light source (our MPE with 𝜆res = 4.26 μm or a blackbody emitter as a conventional light source) and a single-channel thermopile detector. Both elements are separated at a distance 𝑑 = 4 mm, which corresponds to a reduction in absorption path length by a factor > 10 compared to conventional NDIR systems1. On the detector side, the dielectric filter element becomes dispensable, as the spectral filtering capability is readily integrated into the emitter (see Figure 8a inset for the absorption profile of the employed MPE with respect to the atmospheric attenuation spectra of CO2 and H2O). Hence, it can be omitted and the simplified system configuration is referred to as filterfree. Emitter and detector are both located in a measurement container, which is connected to a gas mixing system allowing for variable CO2 and humidity levels. Reference sensors for total mass flow (MF), temperature (T) and relative humidity (RH) are connected to the downstream side of the measurement container. The source is driven by a periodic, rectangular voltage signal (𝑓rect = 5 Hz,

ACS Paragon Plus Environment

Page 13 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

𝑉 = 1.3 V, duty cycle 55 %), allowing for lock-in amplification of the pre-amplified thermopile signal (pre-amplification factor 𝐺 = 4300, lock-in model: Stanford Research SR830, time constant 𝜏 = 3 s, sensitivity 𝑆 = 500 mV). We estimate a peak membrane temperature of ~ 320 °C. In Figure 8b, the normalized voltage difference Δ𝑉/ 𝑉0ppm of the MPE system in comparison to a system employing a conventional light source (blackbody emitter, “BB”) is shown for a measurement cycle in dry air with CO2 concentrations ranging from 0 ppm to 50000 ppm. Here, Δ𝑉 is defined as the difference between the lock-in voltage at the respective CO2 concentration and the lock-in voltage at 0 ppm CO2, Δ𝑉 = 𝑉lock-in – 𝑉0ppm. The normalized voltage difference Δ𝑉/ 𝑉0ppm corresponds to the fractional absorbance of the sensing system3 and correlates well with the applied CO2 concentration profile. A maximum signal variation Δ𝑉50000ppm / 𝑉0ppm = 4.8 % can be determined over the course of a measurement cycle for the MPE system, in comparison to 0.8 % for the BB emitter system. The overall sensing system is considered to be detector-limited, as the measured voltage noise (55 nV⁄√Hz for the MPE system) is dominated by the specified detector noise (45 nV⁄√Hz). In Figure 8c, the normalized voltage difference Δ𝑉⁄𝑉0ppm is plotted against the applied CO2 concentration for both systems (MPE and conventional BB emitter). An exponential decrease can be fitted to both data sets in accordance with the Beer-Lambert law2. To quantify the relative sensitivity increase due to the employment of a metamaterial-enhanced emitter, the relative sensitivity in the linear regime of the system at the CO2 concentration 𝑐CO2 = 0 ppm is calculated, 𝑠r =

d Δ𝑉⁄𝑉0ppm | d 𝑐CO2 𝑐

.

(3)

CO2 =0 ppm

Comparing the relative sensitivity of the MPE system, −4 −5 𝑠r,MPE = 1.7 × 10 %/ppm, to the value of the blackbody system, 𝑠r,BB = 3.4 × 10 %/ppm, an increase in relative sensitivity by a factor of 5 can be determined. This increase can be attributed to the larger fractional absorbance of the IR signal due to the narrowband emission characteristics of the MPE light source. Furthermore, the cross-sensitivity of the MPE system to humidity (H2O) in comparison to the system employing a conventional light source (blackbody emitter, “BB”) has been investigated. As shown in Figure 8a, atmospheric humidity levels lead to considerable absorption in the vicinity of the CO2 absorption band. Besides exhibiting a high sensitivity to the target gas (CO2), the cross-sensitivity of the sensing system to H2O should be minimal to effectively distinguish the target gas against the atmospheric background. Both sensing systems (MPE, BB) have been investigated at various relative humidity levels between 0 % and 50 % in nitrogen atmosphere (𝑐CO2 = 0 ppm). From the absolute signal changes with humidity and the CO2 sensitivity calibration in Figure 8c, a (parasitic) equivalent CO2 concentration change (in units ppm) can be attributed to each humidity level (Supporting Information). The results are shown in Figure 8d. From a linear fit to both series, a cross-sensitivity to H2O of 𝑠H2O,MPE = 41.5 ppm/%rH can be extracted for the MPE system, in comparison to 𝑠H2O,BB = 379.6 ppm/%rH for the system employing the conventional light source. This correlates to a 9-fold reduced humidity cross-sensitivity for the MPE system in comparison to a blackbody system.

Conclusion In summary, we demonstrate an efficient, narrowband mid-infrared light source by combining MEMS heater technology with metamaterial perfect emitter (MPE) structures. The MPE exhibits a resonantlyACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 20

enhanced emissivity of 0.99 with a quality factor of 15.7 at the center wavelength of 3.96 μm, and features temperature-stable (resonance wavelength shift 0.04 nm/°C) and angular-independent (up to angles of 50°) emission characteristics. Its properties prove to be robust against fabrication tolerances, and render the light source attractive for highly integrated, free-space optical gas sensing applications. Employing the MPE light source in a proof-of-concept demonstration of such a sensing system, we measured for CO2 concentrations up to 50000 ppm and an absorption length of 4 mm a 5-fold increase in relative sensitivity compared to the use of a conventional blackbody emitter. By explicit optimization of the structure geometry for smallest resonance bandwidths, the sensitivity could potentially even be further increased. With total dimensions of only 1.7 mm × 1.7 mm and full compatibility to standard CMOS materials and processes, the presented light source potentially opens the way for a new generation of highly integrated, low-cost NDIR gas sensors. Furthermore, due to the reciprocity of the underlying working principle, also thermal detectors could vastly benefit from MPE structures.

ACS Paragon Plus Environment

Page 15 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Methods Numerical simulations A commercial FEM package has been used for the 3D, full wave optical simulations of the metamaterial structure42. Frequency-dependent material constants for Cu and Al2O3 were taken from literature41, 43. We employed periodic (Bloch) boundary conditions to account for phase shifts over the unit cell. For the incident field, a TEM plane wave with the electric field linearly polarized in 𝑥-direction was chosen. The unit cell’s reflection and transmission coefficients were derived from its scattering parameters. Angular-resolved emission characteristics were simulated using a self-implemented Fourier-ModalMethod (FMM)44. Up to 91 modes per dimension were used to discretize the electromagnetic fields. The number of modes was chosen to be uneven between the periodicity dimensions in accordance with the polarization of incident light in order to reduce to simulation time. Device fabrication For the fabrication of the MPE structure on top of the MEMS hotplate, first a continuous Cu backplane has been deposited by evaporation, using Cr as an adhesion layer. In the next step, the Al 2O3 spacer layer has been deposited by atomic layer deposition (ALD). A standard e-beam lift-off process has been used for the definition of the top resonators. In a subsequent process step, the sealing layer has been deposited by ALD. For the definition of the circular MPE area, a negative tone resist has been structured, followed by dry etching. The fully processed sample has been wire-bonded to TO-39 headers. Optical characterization Reflectivity measurements have been performed using a Fourier-transform infrared (FTIR) spectrometer (BioRad FTS 375C, KBr beam splitter) coupled to an infrared microscope (MCT detector, 15x Cassegrain objective). Measurements were performed with a resolution of 2 cm-1 and every spectrum was averaged over 32 measurements. The MPE structure was mounted in a custom build sample holder, which allowed for sample temperatures between room temperature and 350°C with temperature deviations from the setpoint smaller than 1°C. The area on the MPE sample used for characterization was restricted by an aperture to 200 μm x 200 μm. To calculate the structure absorptivity, reflectivity spectra were normalized to the reflectivity spectrum of an uncoated gold mirror. Emissivity and spectral exitance measurements were performed at the external port of a FTIR spectrometer (Bruker V70, MCT detector, KBr beam splitter). Measurements were performed with a resolution of 2 cm-1 and every spectrum was averaged over 100 measurements. Angular-dependent emissivity measurements were performed by rotating the sample with respect to the FTIR port up to emission angles of 60°. An aperture (𝑑AP1 = 4 mm) in the image plane of the FTIR was used to limit the area of collection on the MPE surface to a spot with diameter 𝑑coll ≈ 2.2 mm. A second aperture (𝑑AP2 = 3 mm) in the Fourier plane of the FTIR defined the angular resolution of the system to 𝑑AP2 Δ𝜙 = 2 atan ( 2𝑓 ) ≈ 1.72° (𝑑AP2 = 3 mm, 𝑓 = 100 mm). Spectral emissivity curves of MPE structures were calculated by normalization of the sample spectrum to a blackbody reference spectrum at respective temperatures, which was collected from a carbon black-coated reference sample heated to the same temperature as the MPE structure. Thermographic measurements

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 20

Thermographic measurements were conducted on a resistively heated MPE hotplate (driving voltage 𝑉 = 1.3 V). With a calibrated infrared camera (FLIR A6703sc), the emitted radiance from the sample was acquired. The temperature on the MPE area was calculated using the experimental absorptivity spectrum of the MPE structure and a blackbody emitter model. Hence, the obtained temperature values are strictly valid only on the MPE area, marked by the color-coded area in Figure 7a. For the calculation of the relative resonance wavelength shift Δ𝜆res /Δ𝑇, an average membrane temperature 𝑇mem,avg ≈ 280 °C was obtained from Supporting Information Figure S9 (ambient temperature 𝑇ambient ≈ 22 °C).

Supporting Information Supporting Information Available: Details on the volume loss in materials; validation of parameter retrieval procedure; silicon substrate and blackbody reference sample photo-/micrographs; key figure maps for samples without adhesion layers; resonance wavelength and -bandwidth versus temperature curves; angular-dependent simulations for TE polarization and for 𝑎 = 1.5 μm structure; temperature profile across MPE hotplate. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsphotonics.XXXXXXX.

Acknowledgments This work was carried out at the Binnig and Rohrer Nanotechnology Center. Grant ETH-45 14-2 is acknowledged for funding of this project. The authors would like to thank Martin Lanz for packaging of the MPE hotplate samples and Sebastian Volk for support with the FTIR measurements.

References 1. Hodgkinson, J.; Tatam, R. P., Optical gas sensing: a review. Meas. Sci. Technol. 2013, 24. 2. Ingle Jr, J. D.; Crouch, S. R., Spectrochemical analysis. Prentice Hall: London, 1988. 3. Banwell, C. N.; McCash, E. M., Fundamentals of molecular spectroscopy. McGraw-Hill: London, 1972. 4.

Ilic, O.; Bermel, P.; Chen, G.; Joannopoulos, J. D.; Celanovic, I.; Soljačić, M., Tailoring high-

temperature radiation and the resurrection of the incandescent source. Nat. Nanotechnol. 2016, 11, 320324. 5. Faist, J.; Capasso, F.; Sivco, D. L.; Sirtori, C.; Hutchinson, A. L.; Cho, A. Y., Quantum cascade laser. Science 1994, 264, 553-556. 6. Villares, G.; Hugi, A.; Blaser, S.; Faist, J., Dual-comb spectroscopy based on quantum-cascade-laser frequency combs. Nat. Commun. 2014, 5, 5192. 7. Szedlak, R.; Harrer, A.; Holzbauer, M.; Schwarz, B.; Waclawek, J. P.; MacFarland, D.; Zederbauer, T.; Detz, H.; Andrews, A. M.; Schrenk, W.; Lendl, B.; Strasser, G., Remote Sensing with Commutable Monolithic Laser and Detector. ACS Photonics 2016, 3, 1794-1798. 8.

Barritault, P.; Brun, M.; Gidon, S.; Nicoletti, S., Mid-IR source based on a free-standing

microhotplate for autonomous CO2 sensing in indoor applications. Sensors and Actuators A: Physical 2011, 172, 379-385. ACS Paragon Plus Environment

Page 17 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

9. Guha, P. K.; Ali, S. Z.; Lee, C.; Udrea, F.; Milne, W.; Iwaki, T.; Covington, J. A.; Gardner, J., Novel design and characterisation of SOI CMOS micro-hotplates for high temperature gas sensors. Sensors and Actuators B: Chemical 2007, 127, 260-266. 10. Ali, S. Z.; De Luca, A.; Hopper, R.; Boual, S.; Gardner, J.; Udrea, F., A Low-Power, Low-Cost Infra-Red Emitter in CMOS Technology. IEEE Sens. J. 2015, 15, 6775-6782. 11. Greffet, J. J.; Carminati, R.; Joulain, K.; Mulet, J. P.; Mainguy, S.; Chen, Y., Coherent emission of light by thermal sources. Nature 2002, 416, 61-64. 12. Inoue, T.; De Zoysa, M.; Asano, T.; Noda, S., Realization of narrowband thermal emission with optical nanostructures. Optica 2015, 2, 27-35. 13. Ikeda, K.; Miyazaki, H. T.; Kasaya, T.; Yamamoto, K.; Inoue, Y.; Fujimura, K.; Kanakugi, T.; Okada, M.; Hatade, K.; Kitagawa, S., Controlled thermal emission of polarized infrared waves from arrayed plasmon nanocavities. Appl. Phys. Lett. 2008, 92, 1117. 14. Pusch, A.; De Luca, A.; Oh, S. S.; Wuestner, S.; Roschuk, T.; Chen, Y.; Boual, S.; Ali, Z.; Phillips, C. C.; Hong, M.; Maier, S. A.; Udrea, F.; Hopper, R. H.; Hess, O., A highly efficient CMOS nanoplasmonic crystal enhanced slow-wave thermal emitter improves infrared gas-sensing devices. Sci. Rep. 2015, 5, 17451. 15. Qu, Y. R.; Li, Q.; Gong, H. M.; Du, K. K.; Bai, S. A.; Zhao, D.; Ye, H.; Qiu, M., Spatially and Spectrally Resolved Narrowband Optical Absorber Based on 2D Grating Nanostructures on Metallic Films. Adv. Opt. Mater. 2016, 4, 480-486. 16.

Chan, D. L.; Soljačić, M.; Joannopoulos, J., Thermal emission and design in one-dimensional

periodic metallic photonic crystal slabs. Phys. Rev. E 2006, 74, 016609. 17. De Zoysa, M.; Asano, T.; Mochizuki, K.; Oskooi, A.; Inoue, T.; Noda, S., Conversion of broadband to narrowband thermal emission through energy recycling. Nat. Photonics 2012, 6, 535-539. 18. Inoue, T.; De Zoysa, M.; Asano, T.; Noda, S., Realization of dynamic thermal emission control. Nat. Mater. 2014, 13, 928-931. 19. Liu, X.; Starr, T.; Starr, A. F.; Padilla, W. J., Infrared spatial and frequency selective metamaterial with near-unity absorbance. Phys. Rev. Lett. 2010, 104, 207403. 20. Watts, C. M.; Liu, X.; Padilla, W. J., Metamaterial electromagnetic wave absorbers. Adv. Mater. 2012, 24. 21. Miyazaki, H. T.; Kasaya, T.; Iwanaga, M.; Choi, B.; Sugimoto, Y.; Sakoda, K., Dual-band infrared metasurface thermal emitter for CO2 sensing. Appl. Phys. Lett. 2014, 105, 121107. 22. Chen, K.; Adato, R.; Altug, H., Dual-band perfect absorber for multispectral plasmon-enhanced infrared spectroscopy. ACS Nano 2012, 6, 7998-8006. 23. Liu, X. L.; Tyler, T.; Starr, T.; Starr, A. F.; Jokerst, N. M.; Padilla, W. J., Taming the blackbody with infrared metamaterials as selective thermal emitters. Phys. Rev. Lett. 2011, 107, 045901. 24. Hao, J. M.; Zhou, L.; Qiu, M., Nearly total absorption of light and heat generation by plasmonic metamaterials. Phys. Rev. B 2011, 83, 165107.

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

25.

Page 18 of 20

Liu, N.; Mesch, M.; Weiss, T.; Hentschel, M.; Giessen, H., Infrared perfect absorber and its

application as plasmonic sensor. Nano Lett. 2010, 10, 2342-2348. 26.

Lee, B. J.; Wang, L. P.; Zhang, Z. M., Coherent thermal emission by excitation of magnetic

polaritons between periodic strips and a metallic film. Opt. Express 2008, 16, 11328-11336. 27.

Todorov, Y.; Tosetto, L.; Teissier, J.; Andrews, A. M.; Klang, P.; Colombelli, R.; Sagnes, I.;

Strasser, G.; Sirtori, C., Optical properties of metal-dielectric-metal microcavities in the THz frequency range. Opt. Express 2010, 18, 13886-13907. 28.

Bagheri, S.; Strohfeldt, N.; Sterl, F.; Berrier, A.; Tittl, A.; Giessen, H., Large-Area Low-Cost

Plasmonic Perfect Absorber Chemical Sensor Fabricated by Laser Interference Lithography. ACS Sensors 2016, 1, 1148-1154. 29.

Miyazaki, H. T.; Kasaya, T.; Oosato, H.; Sugimoto, Y.; Choi, B.; Iwanaga, M.; Sakoda, K.,

Ultraviolet-Nanoimprinted Packaged Metasurface Thermal Emitters for Infrared CO2 Sensing. Sci. Technol. Adv. Mater. 2016. 30. Lochbaum, A.; Fedoryshyn, Y.; Hafner, C.; Leuthold, J., High-Q Metamaterial Mid-IR Emitter on a Membrane Heater for Gas Sensing Applications. In Optical Sensors, Optical Society of America: 2016; p SeW2F. 6. 31. Linden, S.; Enkrich, C.; Wegener, M.; Zhou, J.; Koschny, T.; Soukoulis, C. M., Magnetic response of metamaterials at 100 terahertz. Science 2004, 306, 1351-1353. 32. Soukoulis, C. M.; Kafesaki, M.; Economou, E. N., Negative-Index Materials: New Frontiers in Optics. Adv. Mater. 2006, 18, 1941-1952. 33. Smith, D. R.; Schultz, S.; Markos, P.; Soukoulis, C. M., Determination of effective permittivity and permeability of metamaterials from reflection and transmission coefficients. Phys. Rev. B 2002, 65, 195104. 34. Costa, F.; Genovesi, S.; Monorchio, A.; Manara, G., A Circuit-Based Model for the Interpretation of Perfect Metamaterial Absorbers. IEEE Trans. Antennas Propag. 2013, 61, 1201-1209. 35. De Luca, A.; Cole, M.; Fasoli, A.; Ali, S.; Udrea, F.; Milne, W., Enhanced infra-red emission from sub-millimeter microelectromechanical systems micro hotplates via inkjet deposited carbon nanoparticles and fullerenes. J. Appl. Phys. 2013, 113, 214907. 36. Gale, W. F.; Totemeier, T. C., Smithells metals reference book. Butterworth-Heinemann: 2003. 37. Raether, H., Surface plasmons on smooth surfaces. Springer: Berlin Heidelberg, 1988. 38. Jouy, P.; Todorov, Y.; Vasanelli, A.; Colombelli, R.; Sagnes, I.; Sirtori, C., Coupling of a surface plasmon with localized subwavelength microcavity modes. Appl. Phys. Lett. 2011, 98, 021105. 39. Saleh, B. E.; Teich, M. C., Fundamentals of photonics. Wiley: New York, 1991; Vol. 22. 40.

Ahn, S. H.; Guo, L. J., Large-area roll-to-roll and roll-to-plate nanoimprint lithography: a step

toward high-throughput application of continuous nanoimprinting. ACS Nano 2009, 3, 2304-2310. 41. Kischkat, J.; Peters, S.; Gruska, B.; Semtsiv, M.; Chashnikova, M.; Klinkmuller, M.; Fedosenko, O.; Machulik, S.; Aleksandrova, A.; Monastyrskyi, G.; Flores, Y.; Masselink, W. T., Mid-infrared

ACS Paragon Plus Environment

Page 19 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

optical properties of thin films of aluminum oxide, titanium dioxide, silicon dioxide, aluminum nitride, and silicon nitride. Appl. Opt. 2012, 51, 6789-6798. 42. CST Microwave Studio 2015. https://www.cst.com/. 43. Rakic, A. D.; Djurisic, A. B.; Elazar, J. M.; Majewski, M. L., Optical properties of metallic films for vertical-cavity optoelectronic devices. Appl. Opt. 1998, 37, 5271-5283. 44. Li, L. F., New formulation of the Fourier modal method for crossed surface-relief gratings. J. Opt. Soc. Am. A 1997, 14, 2758-2767.

ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For Table of Contents Use Only:

ACS Paragon Plus Environment

Page 20 of 20