Research Article

Adaptive infrared-reflecting systems inspired by cephalopods

See allHide authors and affiliations

Science  30 Mar 2018:
Vol. 359, Issue 6383, pp. 1495-1500
DOI: 10.1126/science.aar5191

Now you see it, now you don't

Thermal vision cameras detect differences in temperature by sensing infrared wavelengths. If a coating could be developed that showed dynamic tuning of the effective temperature, it might be possible to hide objects from infrared sensing. Xu et al. started with a basic Bragg reflector made up of multiple layers of alternating materials with varying refractive index. The authors designed structures that were wavy to begin with so that they could be flattened out by electrical activation. This changed the infrared reflectivity and, thus, the effective temperature of the object observed in its infrared profile.

Science, this issue p. 1495

Abstract

Materials and systems that statically reflect radiation in the infrared region of the electromagnetic spectrum underpin the performance of many entrenched technologies, including building insulation, energy-conserving windows, spacecraft components, electronics shielding, container packaging, protective clothing, and camouflage platforms. The development of their adaptive variants, in which the infrared-reflecting properties dynamically change in response to external stimuli, has emerged as an important unmet scientific challenge. By drawing inspiration from cephalopod skin, we developed adaptive infrared-reflecting platforms that feature a simple actuation mechanism, low working temperature, tunable spectral range, weak angular dependence, fast response, stability to repeated cycling, amenability to patterning and multiplexing, autonomous operation, robust mechanical properties, and straightforward manufacturability. Our findings may open opportunities for infrared camouflage and other technologies that regulate infrared radiation.

Materials and systems that statically reflect radiation in the short- to long-wavelength infrared region of the electromagnetic spectrum have been studied for decades and critically underpin the performance of many entrenched technologies, including building insulation (1), energy-conserving windows (2), spacecraft components (3), electronics shielding (4), container packaging (5), protective clothing (6), and camouflage platforms (7). One highly desirable but not easily attainable property for infrared-reflecting materials (and related technologies) is on-demand adaptability: precise and sensitive real-time dynamic responsiveness to changes in the surrounding environment. To date, only a limited number of adaptive infrared systems, such as infrared camouflage platforms, have been reported because they must satisfy stringent and demanding technical performance criteria, making the development of these technologies extremely challenging (a comparative overview for adaptive camouflage is provided in table S1). For example, thermochromic materials can display substantial infrared emissivity changes because of thermally induced phase transitions but feature high operating temperatures, hysteresis during cycling, and structural characteristics that are difficult to control (812). In addition, infrared electrochromic devices have optical properties that can be altered with electrochemical redox reactions but struggle with precise spectral tunability and necessitate the use of inert noble metals in order to achieve functionality and stability (1316). Moreover, certain metamaterial-based systems can spatiotemporally modulate their emissivity through photogenerated carrier doping but use ultraviolet light for actuation, need elevated working temperatures for adequate contrast, and display long recovery times (17). Furthermore, thermal cloaking platforms modify the infrared signatures of other objects by manipulating the heat flow in the surroundings but require large temperature gradients and demand complete object immersion, restricting implementation (18, 19). Last, soft robots with integrated microfluidics can alter their thermal appearance and infrared patterning via the pneumatic injection of heated or cooled liquids but rely on the continuous flow of multiple liquids maintained at different temperatures and have slow response times dictated by their thermal conductivities (20). Consequently, in order to enable a broad range of practical applications (such as camouflage), the “ideal” adaptive infrared-reflecting system would simultaneously possess a simple actuation mechanism, low working temperature, tunable spectral range, weak angular dependence, fast response, stability to repeated cycling, amenability to patterning and multiplexing, autonomous operation, robust mechanical properties, and straightforward manufacturability (table S1).

Biomimetic system design

As an active color-changing system, the skin of coleoid cephalopods (such as squid, octopuses, and cuttlefish) represents an exciting source of inspiration (2123). The patterning and coloration of cephalopod skin can be altered autonomously and repeatedly for the purposes of concealment or signaling, as illustrated for the squid in Fig. 1A. Such remarkable feats of camouflage are enabled by the sophisticated architecture of the squid’s soft and flexible skin, in which innervated dermal layers contain chromatophore pigment cells (as part of larger chromatophore organs) and reflective cells called iridocytes (Fig. 1B) (21, 24, 25). The two cell types operate in tandem but perform distinct optical functions (adaptive chromatophores are present in many cephalopods, but adaptive iridocytes have only been identified in a few squid species to date) (21, 2427). The adaptive chromatophore pigment cells contain pigment granule–packed internal sacculi, which are expanded and contracted through the mechanical action of radial muscle cells (Fig. 1C). These yellow, red, and brown cells feature response times of hundreds of milliseconds and function as size-variable biological spectral filters that absorb and reflect visible light of specific wavelengths (21, 26). The adaptive iridocytes contain alternating arrangements of membrane-enclosed nanostructured protein layers and extracellular space, for which the geometries and refractive index differences are altered via a biochemical signaling cascade (Fig. 1D). These iridescent cells feature response times of tens of seconds and function like reconfigurable biological Bragg stacks that reflect visible light of variable wavelengths (21, 27). In part because of the properties of these cells, the skin of squid (and by extension, other cephalopods) behaves as a dynamic bioelectronic display, with the functionality summarized in table S1.

Fig. 1 Biological inspiration and system design.

(A) Camera images of a squid changing its appearance in front of a rocky background. (B) An image of squid skin. The yellow, red, and brown circular regions are chromatophores, and the underlying bright iridescent regions are iridocytes. (C) A schematic of a cephalopod chromatophores organ, in which a central chromatophore pigment cell is ringed by muscle cells. The adaptive chromatophore pigment cells contain internal sacculi packed with pigment granules (inset), and they are expanded and contracted through the mechanical action of the muscle cells. The chromatophore pigment cells function as size-variable biological spectral filters that absorb and reflect visible light of specific wavelengths. (D) A schematic of a squid iridocyte. The adaptive iridocytes contain alternating arrangements of membrane-enclosed protein layers and extracellular space (inset), for which the geometries and refractive index differences are altered via a biochemical signaling cascade. The iridocytes function like reconfigurable biological Bragg stacks that reflect visible light of variable wavelengths. (E) A generic schematic of the top view of our infrared-reflecting platform (left) before and (right) after actuation. Described is the areal change for the active region upon actuation. (F) A generic schematic of the side view of our infrared-reflecting platform (left) before and (right) after actuation. Described is the surface morphology change for the active region upon actuation. [The images in (A) are reproduced from a video by H. Steenfeldt under the YouTube Creative Commons Attribution license, and the image in (B) was obtained by G. Hanlon and reproduced with permission of R. Hanlon.]

The remarkable capabilities of cephalopod skin and its components have inspired the engineering of various adaptive artificial optoelectronic devices (2832). For example, dielectric elastomer actuators have been leveraged for cephalopod-inspired color-changing systems that function within the visible region of the electromagnetic spectrum (3032). In the most basic incarnation, such devices consist of an elastomer membrane sandwiched between two electrodes, where the application of a voltage between the electrodes induces electrostatic pressure, leading to a decrease in the membrane’s thickness and an increase in the overall electrodes’ areas (3335). More generally, such actuators, which essentially translate electrical stimuli into mechanical outputs, have been explored for a variety of applications, including artificial muscles, pneumatic automation, energy generation, tactile displays, and adaptive optics (3335). However, the technological viability of dielectric elastomer actuators has been limited by challenges associated with a requirement for both high operating voltages and electrodes with a demanding combination of properties (3337). For infrared camouflage applications, the latter requirement is particularly daunting because the electrode materials must simultaneously demonstrate straightforward processability into thin freestanding films, amenability to surface modification and patterning, excellent flexibility and compliance, high conductivity that does not drop upon deformation, transparency over a broad spectral range, lack of degradation under variable humidity, and stability to repeated cycling.

We designed a platform inspired by squid skin’s highly evolved bio-optical components while leveraging the technical foundation established for dielectric elastomer actuators. We conceptualized a device in a parallel plate capacitor–type configuration, which consists of a proton-conducting bottom electrode, a dielectric elastomer membrane, a proton-conducting top electrode, and an infrared-reflecting coating (Fig. 1, E and F). Before actuation, the devices feature relatively small but size-variable active areas (Fig. 1E, left), analogous to cephalopod chromatophores (Fig. 1C, left) (26), with the surfaces covered by a dense but geometrically reconfigurable arrangement of reflective microstructures (Fig. 1F, left), analogous to squid iridocytes (Fig. 1D, left) (27). After actuation, the devices expand their active areas to modulate the amount of absorbed incident infrared light (Fig. 1E, right)—again, analogous to cephalopod chromatophores (Fig. 1C, right) (26)—as well as alter the geometry of their active areas’ microstructured surfaces to modulate the relative intensity of the reflected incident infrared light (Fig. 1F, right)—again, analogous to squid iridocytes (Fig. 1D, right) (27).

System fabrication, characterization, and testing

We fabricated the cephalopod-inspired adaptive infrared-reflecting systems as detailed in the supplementary materials. We adapted protocols used for the lithographic fabrication of various dielectric elastomer actuators (3335) and chose to manufacture devices with centimeter-scale active areas, allowing for benchtop assembly and simplifying spectroscopic characterization. For the compliant electrodes, we prepared films from a sulfonated pentablock copolymer that features exceptional protonic conductivity, with the size and shape of the films defining the device’s active area (38, 39). For the electroactive layer, we mounted an acrylate dielectric elastomer membrane within a size-adjustable holder and equiaxially stretched the membrane. Subsequently, to facilitate experiments within different infrared wavelength regimes, we produced two types of devices featuring chemically and structurally distinct infrared-reflecting coatings. In order to fabricate broadband infrared-reflecting devices, we used electron-beam evaporation to deposit a thin film of aluminum (Al) metal onto a proton-conducting top electrode, which was then laminated onto a mounted acrylate membrane already outfitted with an unmodified bottom electrode (fig. S1A). In order to fabricate narrowband infrared-reflecting devices, we used electron-beam evaporation to deposit alternating layers of titanium dioxide (TiO2) and silicon dioxide (SiO2) directly onto the top electrode of a mounted acrylate membrane, which was already outfitted with both the top and bottom electrodes (fig. S1B). For both device types, we mechanically contracted the holder to release some of the tension introduced during the initial mounting of the membrane and thus introduced microstructuring (wrinkles) to the surfaces of the devices’ active areas. Last, when necessary for characterization, we transferred electrical lead-modified devices to rigid support frames. The overall scalable procedure furnished infrared-reflecting systems with the general architecture depicted in Fig. 1, E and F.

We investigated the effect of mechanical actuation (equiaxial strain) on the properties of devices with a broadband infrared reflectance (Fig. 2A). Before mechanical actuation, our devices possessed relatively small active areas (Fig. 2B) and relatively large thicknesses, with their surfaces covered by a dense three-dimensional network of micrometer-sized wrinkles, as revealed with optical microscopy (Fig. 2C) and scanning electron microscopy (Fig. 2D). The corresponding infrared spectra showed that the unactuated devices featured a high average total reflectance of 71 ± 3% (Fig. 2A, bottom left), a low average total transmittance of <1% (fig. S2), and a moderate average total absorptance of 28 ± 2% (fig. S2). For the wrinkled surfaces, the total reflectance featured a weak average specular component of 23 ± 1% and a dominant average diffuse component of 48 ± 2%, in a ratio of ~0.5 (Fig. 2A, bottom left). By contrast, upon mechanical actuation, our devices possessed larger active areas (Fig. 2B) and smaller thicknesses, with the wrinkles flattened into a quasi two-dimensional network of irregular domains, as revealed with optical microscopy (Fig. 2C) and scanning electron microscopy (Fig. 2D). The corresponding infrared spectra showed that the actuated devices featured an increased average total reflectance of 96 ± 1% (Fig. 2A, bottom right), a low average total transmittance of <1% (fig. S2), and a low average total absorptance of 3 ± 1% (fig. S2). For the flattened surfaces, the total reflectance featured a much larger average specular component of 88 ± 3% and a smaller average diffuse component of 8 ± 2%, in a ratio of ~11 (Fig. 2A, bottom right). For our devices, mechanical actuation dynamically modulated both the specular-to–diffuse reflectance ratios (owing to the change in the morphology) and the total absorptance (owing to the change in the thickness) by approximately an order of magnitude. Moreover, the specular component of the reflectance exhibited a weak angular dependence in the devices’ unactuated and actuated states (fig. S3). The changes in the devices’ infrared-reflecting properties were consistent and fully reversible upon repeated actuation, with no physical delamination and only minor performance degradation observed after 75 and 750 cycles (fig. S4). Thus, actuation of our devices induced a change in the microstructure of their active areas and enabled concomitant reversible, angle-independent, and stable modulation of the broadband reflectance (and absorptance) within the short- to long-wavelength infrared region.

Fig. 2 Mechanical modulation of the broadband and narrowband reflectance.

(A) (Top) Schematics of an aluminum-modified device (left) before and (right) after mechanical actuation in an adjustable holder, illustrating the change in the surface morphology and the reflection of infrared light. (Bottom) The infrared reflectance spectra of an aluminum-modified device (left) before and (right) after mechanical actuation. The total reflectances (solid orange traces) are shown along with their specular (dotted yellow traces) and diffuse (dotted blue traces) components. (B) Digital camera images of an aluminum-modified device’s active area (left) before and (right) after mechanical actuation. The yellow dashed square indicates the active area. (C) Optical microscope images of the surface morphology of an aluminum-modified device’s active region (left) before and (right) after mechanical actuation. (D) Scanning electron microscopy images of the surface morphology of an aluminum-modified device’s active region (left) before and (right) after mechanical actuation. (E) (Top) Schematics of a TiO2/SiO2 Bragg stack–modified device (left) before and (right) after mechanical actuation in an adjustable holder, illustrating the change in the surface morphology and the reflection of infrared light. (Bottom) The infrared reflectance spectra of a TiO2/SiO2 Bragg stack–modified device, with a peak reflectance intensity at 3 μm, (left) before and (right) after mechanical actuation. The total reflectances (solid orange traces) are shown along with their specular (dotted yellow traces) and diffuse (dotted blue traces) components. (F) The infrared spectra of three unactuated devices that have been designed to feature peak reflectance wavelengths of 3 μm (red trace), 4 μm (green trace), and 5 μm (blue trace). (G) Plots of the total (orange squares), specular (yellow triangles), and diffuse (blue circles) peak reflectances of the TiO2/SiO2 Bragg stack–modified devices as functions of the applied length strain. The lines represent linear fits of the data.

We investigated the effect of mechanical actuation (equiaxial strain) on the properties of devices with a narrowband infrared reflectance (Fig. 2E). To obtain reflectances that featured peak wavelengths (λpeak) in the mid-wavelength infrared region of the electromagnetic spectrum, we used infrared-reflecting coatings composed of Bragg stacks, alternating TiO2 and SiO2 layers with thicknesses of λpeak/(4*nTiO2) and λpeak/(4*nSiO2) (where nTiO2 and nSiO2 are the refractive indices). This straightforward approach furnished devices with peak reflectance intensities at wavelengths of 3, 4, and 5 μm (Fig. 2F). For such narrowband-reflecting devices, the optical properties were responsive to the applied strain, analogous to their broadband-reflecting counterparts. As a specific example, before mechanical actuation, the infrared spectra obtained for representative microstructured (wrinkled) and relatively thicker devices featured peak total reflectance intensities of 34 ± 3% at a wavelength of 3 μm, with a weak specular component of 8 ± 1% and a dominant diffuse component of 26 ± 2% in a ratio of ~0.3 (Fig. 2E, bottom left). However, after mechanical actuation, the infrared spectra of the now flattened and relatively thinner devices featured increased peak total reflectance intensities of 55 ± 7% at a wavelength of 3 μm, with a much larger specular component of 29 ± 5% and a nearly unchanged diffuse component of 26 ± 2% in a ratio of ~1.1 (Fig. 2E, bottom right). In general, the devices’ total reflectances at this specific wavelength increased as a function of the strain (whereas the corresponding total absorptances presumably decreased) (Fig. 2G). The reflectances’ specular components likewise increased with the strain, but the diffuse components remained relatively unaffected (Fig. 2G). Moreover, the changes in the devices’ infrared-reflecting properties were fully reversible, with only minor performance degradation observed after 100 actuation cycles (fig. S5). Therefore, actuation of our devices directly induced dynamic modulation of their reflectance within a specific narrow wavelength range of the infrared region.

We studied electrical actuation of our devices in a standard dielectric elastomer actuator configuration (Fig. 3A). In agreement with the mechanical actuation experiments, the devices possessed a relatively small active area with a wrinkled surface before electrical actuation, but a larger active area with a flattened surface after electrical actuation (Fig. 3A and movie S1). We found that the areal strain demonstrated an exponential dependence on the applied voltage, with a voltage of ~3.5 kV, resulting in a maximum strain of ~230% (Fig. 3B). In addition, the areal strain exhibited a distinct dependence on the frequency of the applied voltage; for example, a variable-frequency square waveform (minima of 0 kV and maxima of 3.2 kV) induced a drop in the maximum areal strain from 181 ± 11% to 85 ± 5% between the frequencies of 0.05 and 2 Hz, followed by an increase in the maximum areal strain to 110 ± 5% at a frequency of 16 Hz (Fig. 3C). Moreover, for the aforementioned square waveform at a frequency of 0.5 Hz, the devices’ response time (defined as the rise time from 10 to 90% of the strain change) was 720 ± 50 ms because of the rapid areal expansion and shrinkage rates of 55.8 ± 1.6 and 56.8 ± 0.8% per second, respectively (Fig. 3D), and the energy associated with device actuation during one typical cycle was estimated to be >8 J/m2 (supplementary materials). These systems’ figures of merit were comparable with or even exceeded those reported for analogous acrylate dielectric elastomer–based devices with conventional ionic hydrogel, carbon grease, or conductive nanowire composite electrodes (36, 37).

Fig. 3 Electrical characteristics of single devices.

(A) (Top) Schematics of an aluminum-modified device (left) before and (right) after electrical actuation in a rigid frame, illustrating the areal change for the active region. (Bottom) Digital camera images of an aluminum-modified device’s active region (left) before and (right) after electrical actuation. (B) A plot of the areal strain as a function of the applied voltage for aluminum-modified devices. (C) A plot of the maximum areal strain as a function of the frequency for square waveforms with identical minima of 0 kV and maxima of 3.2 kV. (D) A plot of the areal strain as a function of time for three consecutive cycles of a square waveform with a frequency of 0.5 Hz, as well as minima of 0 kV and maxima of 3.2 kV.

We showed electrical modulation of our systems’ infrared-reflecting properties in multiple configurations (Fig. 4). To this end, we prepared not only single devices but also three-by-three arrays in which such devices served as independently addressable “pixels” and then visualized both platforms with a thermal infrared camera under an incident heat flux (supplementary materials). We focused our experiments on aluminum-coated devices because of most commercial infrared cameras’ spectral range of 7.5 to 14 μm. As an example, a representative single device featured an apparent temperature difference of ~3.6°C between its aluminum-coated active area and the surrounding membrane before electrical actuation (Fig. 4A), presumably because of the dominance of the reflectance’s diffuse component in the unactuated state. By contrast, the same device featured an increased apparent temperature difference of ~6.8°C between its aluminum-coated active area and the surrounding membrane after electrical actuation (Fig. 4A), presumably because of the dominance of the reflectance’s specular component in the actuated state. The observed changes in the apparent temperature were rapid, stable, and fully reversible over numerous distinct on/off cycles for such devices (movie S2). As another example, the devices in a representative array featured analogous modulation of the apparent temperature differences between their active areas and the surrounding membranes upon device-specific, independent electrical actuation (fig. S6). Multiple devices in the arrays also readily exhibited tandem changes in their active areas’ local apparent temperatures (Fig. 4B). Indeed, electrical actuation of different combinations of seven multiplexed devices enabled distinct representative arrays to “spell out” the letters “U,” “C,” and “I” (Fig. 4B). In principle, the general fabrication approach is amenable to the fabrication of advanced displays with higher pixel densities and/or submillimeter pixel sizes (35, 40).

Fig. 4 Electrical modulation of the infrared appearance for single and multiplexed devices.

(A) (Top) Schematics of an aluminum-modified device under a constant thermal flux (left) before and (right) after electrical actuation in a rigid frame, illustrating the change in the surface morphology and the reflection of infrared light. (Bottom) Infrared camera images of a representative aluminum-modified device’s active region (left) before and (right) after electrical actuation. The change in apparent temperature and thermal appearance is represented by a change in color. (B) (Top) Schematics of three-by-three arrays of multiplexed devices (left) before electrical actuation and (right) after selective electrical actuation of different device (pixel) combinations as the letters “U,” “C,” and “I.” (Bottom) Infrared camera images of representative three-by-three arrays of multiplexed devices (left) before electrical actuation and (right) after selective electrical actuation of different device (pixel) combinations as the letters “U,” “C,” and “I.” The changes in apparent temperature and thermal appearance are represented by changes in color.

We demonstrated that our systems could operate without input from an external operator (Fig. 5A). We connected electrically controlled devices (with aluminum-coated active areas) to a remotely positioned, independently heated temperature sensor, which transduced thermal information from a distal environment while avoiding undesired cross-talk. At a sensor temperature of ~26°C, a representative autonomous system possessed a wrinkled active area and an apparent temperature difference between the active area and surrounding unmodified membrane of ~3.4°C, as revealed with digital and infrared camera imaging, respectively (Fig. 5A). Upon a change in the temperature of the sensor to ~34°C, the system’s active area flattened and expanded by 18 ± 5% (relative to the initial value), and its apparent temperature difference increased to ~3.7°C with respect to the unmodified membrane (Fig. 5A). After a further rise in the sensor’s temperature to ~42°C, the active area flattened and expanded further by 35 ± 6% (relative to the initial value), and its apparent temperature difference increased to ~4.0°C with respect to the unmodified membrane (Fig. 5A). With an additional rise in the sensor’s temperature to ~48°C, the active area flattened and expanded even more by 74 ± 3% (relative to the initial value), and its apparent temperature difference increased to ~4.4°C with respect to the unmodified membrane (Fig. 5A). Over a temperature window of >20°C for the sensor, our autonomous system exhibited reproducible and stable changes in the size of its active area (and apparent temperature), albeit in nonlinear fashion (fig. S7).

Fig. 5 Autonomous operation and reversible concealment of devices in the infrared.

(A) (Top) Schematics of a sensor-connected, aluminum-modified device. The remote sensor is positioned in a different local thermal environment, with increasing temperatures of ~26, ~34, ~42, and ~48°C (from left to right). (Middle) Digital camera images of a representative device’s active region at the different sensor temperatures. (Bottom) Infrared camera images of the same device’s active region at the different sensor temperatures. Shown are the expansion and change in thermal appearance for the autonomous device’s active region, with increasing local temperature for the sensor. (B) (Top) Schematics of a squid silhouette–shaped, aluminum-modified device maintained under a constant thermal flux and positioned above a warm surface (left) before and (right) after electrical actuation. (Middle) Digital camera images of a representative squid silhouette–shaped, aluminum-modified device’s active region (left) before and (right) after electrical actuation. (Bottom) Infrared camera images of the same squid silhouette–shaped, aluminum-modified device’s active region (left) before and (right) after electrical actuation. The device alters its thermal appearance and becomes distinguishable from the background only after actuation.

Last, we evaluated the ability of our infrared camouflage systems to conceal themselves under infrared visualization (Fig. 5B). We prepared devices with aluminum-coated active areas in the shape of a swimming squid’s silhouette and then imaged them with a thermal infrared camera under an incident heat flux above a surface with a locally elevated temperature (Fig. 5B and supplementary materials). Before electrical actuation, a representative squid silhouette–shaped device featured a relatively small microstructured (wrinkled) active area (Fig. 5B) and a negligible apparent temperature difference with the immediate surroundings that effectively made it invisible above a warm surface with a temperature of ~35°C, as revealed with infrared camera imaging (Fig. 5B). By contrast, after electrical actuation, the squid silhouette–shaped device featured a larger flattened active area (Fig. 5B) and an apparent temperature difference of ~2°C with the immediate surroundings that made it stand out as a specific shape, as revealed with infrared camera imaging (Fig. 5B). In general, the observed changes in appearance under infrared visualization were again rapid, stable, and fully reversible over numerous distinct on/off cycles (movie S3).

Conclusion

By drawing inspiration from cephalopods, we have developed and validated a new class of adaptive infrared-reflecting materials and devices with an unprecedented combination of properties. First, our artificial platforms translate many of the key natural capabilities of cephalopods from the visible to the infrared regions of the electromagnetic spectrum (although in principle they could be adapted for functionality within the visible). Second, our presented systems feature capabilities and figures of merit that exceed or match the state of the art for adaptive infrared camouflage technologies (table S1). Third, the systems have been intrinsically designed for ready manufacturability and straightforward integration, potentially facilitating applications in dielectric elastomer–based artificial muscles, pneumatic automation, energy generation, and adaptive optics (as well as in other areas). Last, the systems may enable new autonomous portable or wearable thermoregulatory technologies, although such applications will likely require optimization of mechanical actuation strategies or the reduction of the associated operating voltages. Ultimately, the described materials and devices may afford new possibilities for the many modern technologies that rely on controlling the transfer of thermal radiation and thus may help transform many aspects of daily life.

Supplementary Materials

www.sciencemag.org/content/359/6383/1495/suppl/DC1

Materials and Methods

Figs. S1 to S7

Table S1

References (4149)

Movies S1 to S3

References and Notes

Acknowledgments: The authors thank I. Gorodetskaya and L. Bagge for invaluable discussions and A. Yee’s laboratory for the use of an optical microscope. The authors thank G. Hanlon and R. Hanlon for the use of the image in Fig. 1B. The authors thank Kraton Polymers for their support and assistance, by providing samples of NEXARpolymers for use in the project, and for their appreciation of the benefits from university research. Funding: The authors are grateful to the Advanced Research Projects Agency–Energy (cooperative agreement DE-AR0000534), the Defense Advanced Research Projects Agency (cooperative agreement HR0011621411), and the Air Force Office of Scientific Research (grant FA2386-14-1-3026) for their financial support. Author contributions: C.X. and A.A.G. conceived, designed, and interpreted the experiments. C.X. carried out the experiments. G.T.S. aided with the interpretation of the spectroscopic measurements. C.X. and A.A.G. wrote the manuscript. Competing interests: C.X., G.T.S., and A.A.G. are listed as inventors on a provisional U.S. patent application (no. 62/643403) from the University of California, Irvine, which describes the design and working principle of the reported adaptive infrared-reflecting platforms. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the supplementary materials.
View Abstract

Subjects

Navigate This Article