Smart Silk Origami as Eco-sensors for Environmental Pollution

Origami folding is an easy, cost-effective, and scalable fabrication method for changing a flat material into a complex 3D functional shape. Here, we created semicrystalline silk films doped with iron oxide particles by mold casting and annealing. The flat silk films could be loaded with natural dyes and folded into 3D geometries using origami principles following plasticization. They performed locomotion under a magnetic field, were reusable, and displayed colorimetric stability. The critical parameters for the design of the semi-autonomous silk film, including ease of folding, shape preservation, and locomotion in the presence of a magnetic field, were characterized, and pH detection was achieved by eye and by digital image colorimetry with a response time below 1 min. We demonstrate a practical application—a battery-free origami silk boat—as a colorimetric sensor for waterborne pollutants, which was reusable at least five times. This work introduces silk eco-sensors and merges responsive actuation and origami techniques.

979.2 mg L -1 MgCl2; 999.2 mg L -1 MgCO3; 600.0 mg L -1 MgOH2; or real, stagnant river water randomly grab sampled in Glasgow, Scotland. The images were processed as for loading. Films were washed three times in ultrapure water for 0.25 h, dabbed dry with a tissue and dried at room temperature for 0.17 h before submersion in a new solution. The relative color change (S) was calculated by subtracting the mean intensities in the red (Ro), green (Go), and blue (Bo) channels of the loaded film from the red (R), green (G) and blue (B) intensities in the films following exposure to each medium, according to equation 1. 1 The color change was also calculated based on the Lab model, according to equation 2. 1 S4 = √( − ) 2 + ( − ) 2 + ( − ) 2 (2) In the Lab color space published by the International Commission on Illumination (CIE), the L*, a*, and b* channels constitute image lightness, red to green intensities, and blue intensity, respectively. Each condition was repeated with three films.

Colorimetric stability of native silk and azosilk films
Native silk and azosilk films doped with 0.1% (w/w) iron oxide and loaded with curcumin or anthocyanin were stored for 31 days in vacuum sealed boxes at 4 °C or 20 °C in the dark. Silk films were then imaged before and after being submerged in 0.2 M potassium phosphate buffer (pH 9.15) for 2 min on an iPhone SE (Apple, Cupertino, CA, U.S.A.) reverse camera at a focal length of 9.7 cm. Photographs were standardised as described for loading, the RGB values measured for 2 boxes on the grid (42840 pixels) and the averages were calculated. Each condition was repeated with three films.
Digital image colorimetry was also achieved using the ColorAssist Lite© (FTLapps, Inc., Broadlands, VA, U.S.A.) smartphone application with a spatial sample aperture of 50 × 50 pixels, temporal sample aperture of 30 frames (1 second) and mean output intensities measured in the RGB color space. Each film was sampled in three regions and the average calculated. Each condition was repeated with three films.

Scanning electron microscopy
For surface imaging, samples were added to aluminium stubs with sticky carbon tabs, with the surface uppermost. For section imaging, a titling SEM stub was used to rotate the samples to 90° with the cross-section uppermost. Samples were sputter coated (ACE200, Leica Microsystems, S5 Wetzlar, Germany) with a 20 nm gold layer to minimise charging in the SEM. Samples were viewed with a TM4000Plus SEM (Hitachi Ltd., Tokyo, Japan) operated at beam voltage 10000eV, probe current setting 2, standard vacuum level (M) and with data collected in backscattered electron mode at magnifications of 100×, 1000× and 10000×.

Contact angle measurement
The films were placed on a glass slide. The contact angle was measured using a DSA30 drop shape analyser (KrussGmbH, Hamburg, Germany) equipped with a manual syringe and needle (diameter 0.8 mm). Droplet size was controlled manually. Results were analysed in Advance software (KrussGmbH, Hamburg, Germany) with manual droplet shape fitting.

Fourier Transform Infrared Spectroscopy analysis
Positive silk II controls were provided by autoclaved silk films and silk films treated with 70% v/v ethanol/ultrapure H2O, while air-dried silk films and freeze-dried silk were used as positive controls for silk I structure. Secondary structures of silk films, freeze-dried powders and freezedried particles were analysed by Fourier transform infrared spectroscopy (FTIR) on an ATRequipped TENSOR II FTIR spectrometer (Bruker Optik GmbH, Ettlingen, Germany). Each FTIR measurement was recorded in absorption mode over the wavenumber range of 400 to 4000 cm −1 at 4 cm -1 resolution for 128 scans and then corrected for atmospheric absorption using Opus (Bruker Optik GmbH, Ettlingen, Germany). The second derivative of the background-corrected FTIR absorption spectra was analysed in OriginLab 19b® (Northampton, MA, U.S.A.) by adapting a literature protocol. 2 The second derivative was smoothed twice using a seven-point Savitzky-Golay function with a polynomial order of 2. The amide I region was analysed by interpolation of a non-zero linear baseline between 2-3 of the highest values in the 1600-1700 cm -S6 1 range. Peak positions were identified by applying the second derivative, and the peaks were fitted in the amide I region using non-linear least squares with a series of Gaussian curves. The position, width and height of each peak were allowed to vary, while peak area could take any value less than or equal to 0. Deconvoluted spectra were then area-normalised, and the relative area of each band was used to calculate the secondary structure content according to literature band assignments. 3,4 Thermal analysis First-cycle differential scanning calorimetry and thermogravimetric analysis were carried out on the dried samples (3-5 mg) in aluminium pans from 20-350 °C at a scanning rate of 10 °C min -1 and under a nitrogen flow of 50 mL min -1 using an STA Jupiter 449 (Netzsch, Gerätebau GmbH, Selb, Germany). Thermograms were analysed using Proteus® (Netzsch, Gerätebau GmbH, Selb, Germany).

Swelling analysis
The swelling of silk fibroin films was monitored over 30 min. Each film was split into 3 pieces and placed in ultrapure water (20 mL), and the weight was measured at defined intervals. The films were removed, and any excess water on the film was dabbed dry with paper towels. This was repeated three times for each film.

Electromagnetic field strength
The strength of a magnetic field was measured by the ability to pull a floating rectangular silk film (0.1 g, 25 μm thickness) along the water-air interphase using a N52 round cylinder magnet (25 × 20 mm Rare Earth Neodymium; (BH)max 52 MGOe). The magnet used in the study was moved across water-glass interphase (2 mm thickness) parallel to the water-air interphase (where silk was S7 floating) at fixed distances defined by the volume of water in a 5000 mL glass beaker. Every added 50 mL was equivalent to a 1 mm increase in the distance between the silk film and the magnet.
The distance from the surface of the magnet at which locomotion was first observed was converted to magnetic flux density (B) and magnetic field strength (H).

Contraction and relaxation on water
The time of contraction from the original shape of the water-annealed dried silk films (ca. 7 × 5 cm) was measured in seconds as the films were placed in water (floating at the water-air interface).
The relaxation and return to the original shape were followed by measuring time in seconds.   Table S2. First-cycle thermal analysis data of azosilk films with different thickness, fiber contents and iron oxide particle contents. ± SD, n = 3.  Asterisks denote statistical significance determined using the post-hoc tests as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.   S19 Figure S7. The change in mean pixel channel intensities in the RGB color space of curcuminloaded and anthocyanin-loaded azosilk and native silk medium thickness films containing 0.1% (w/w) iron oxide particles in response to pH. ± SD, n = 3. Sample pairs were analyzed using the independent t-test. Asterisks denote statistical significance determined using the t-test as follows:

S23
S24 Figure S10. The change in color and mean pixel channel intensities of curcumin-loaded and anthocyanin-loaded azosilk and native silk medium thickness films containing 0.1% (w/w) iron oxide particles after exposure to (a) surfactants and heavy metal salts serving as model pollutants, and (b) after folding into 3D origami silk canoes and exposure to media at the indicated pH. Scale bars = 0.5 cm. ±SD, n = 3. Sample pairs were analyzed using the independent t-test. Asterisks denote statistical significance determined using the t-test as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
S25 Figure S11. The colorimetric stability of curcumin and anthocyanin-loaded native silk and azosilk medium thickness films containing 0.1% (w/w) iron oxide particles. Following 31 days of storage in vacuo at 4 or 20 °C, the color changes after exposure to 0.2 M potassium phosphate buffer (pH 9.15) were monitored in (a) the RGB color space using the mean pixel channel intensities, and (b) in the Lab color space to calculate the color change (ΔE) of silk films. Digital image colorimetry was undertaken using image pre-processing and analysis in ImageJ ® or directly in ColorAssist Lite © . ± SD, n = 3. Sample pairs were analyzed using the independent ttest. Asterisks denote statistical significance determined using the t-test as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ±SD, n = 3.
S26 Figure S12. Origami folding and the resulting silk film origami structures.