Rigid Tightly Packed Amino Acid Crystals as Functional Supramolecular Materials

Abstract

The formation of ordered nanostructures by metabolites is gaining increased interest due to the simplicity of the building blocks and their natural occurrence. Specifically, aromatic amino acids possess the ability to form ordered supramolecular interactions due to their limited solubility in aqueous solution. Unexpectedly, l-Tyrosine (l-Tyr) is almost 2 orders of magnitude less soluble in water compared to l-phenylalanine (l-Phe). However, the underlying mechanism is not fully understood as l-Tyr is more polar. Here, we explore the utilization of insoluble tyrosine assemblies for technological applications and their molecular basis by manipulating the basic building blocks of tightly packed dimers. We show that the addition of an amyloid inhibition agent increases l-Tyr solubility due to the disruption of the dimer formation. The molecular organization grants the l-Tyr crystal higher thermal stability and mechanical properties between three amino acids. Additionally, l-Tyr crystals are shown to generate high and stable piezoelectric power outputs under mechanical pressure in a sandwich device. By incorporating the rigid l-Tyr crystals into a soft polymer, a mechano-responsive bending composite was fabricated. Furthermore, the l-Tyr crystalline needles exhibit an active photowaveguiding property, making them promising candidates for the generation of photonic biomaterial-based devices. The present work exemplifies a feasible strategy to explore physical properties of supramolecular self-assemblies comprises minimalistic naturally occurring building blocks and their applications in energy harvesting, photonic devices, stretchable electronics, and soft robotics.

Organic molecular self-assembly mediated by bottom-up organization is of high interest due to the potential fabrication of advanced smart materials.^1–13 These are produced through the spontaneous associations of small molecules such as polypeptides or nucleic acids into well-defined and relatively stable organizations under thermodynamic equilibrium conditions through noncovalent interactions.^14–24 Recently, very simple molecules (e.g., small metabolites and single amino acids) have been investigated for their self-assembly as potential functional biological or bioinspired materials, thereby harnessing their diverse advantages, such as low cost, ease of production, and biodegradability.^25–32 The combination of these attributes with the versatile nature of noncovalent supramolecular polymers proposes metabolite nanotechnology to be highly attractive for the field of materials science. Such metabolite supramolecular structures were already shown to exhibit exceptional optical properties, such as fluorescence at a broad range of the visible spectrum for cell imaging without any external dyes.³³ In addition, a recent study revealed the notably high piezoelectric properties of single amino acids crystals.³⁴ This is especially impressive due to the fact the single amino acids were thoroughly researched, yet such, and more superb, attributes are still to be discovered, and some fundamental questions regarding the inherent properties of single amino acids remain unanswered. Understanding the molecular basis and physical properties of amino acids and their supramolecular assemblies could be important for exciting discoveries in various technological and medicinal fields.

Herein, we have investigated the self-assembly and the resulting physicochemical properties of l-Phe, l-Tyr, and l-DOPA, three amino acids with very similar chemical structures. All three amino acids can form elongated crystal structures, which were analyzed by X-ray diffraction (XRD) in order to understand their molecular organization. In addition, the water solubility, self-assembly morphology, thermal stability, and mechanical properties were fully studied by optical microscopy, scanning electron microscopy (SEM), differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and atomic force microscopy (AFM). Surprisingly, l-Tyr is the least soluble in water compared to l-Phe and l-DOPA, and the thermal stability and mechanical properties of l-Tyr structures are the highest of the three. We explore the molecular basis for the poor solubility of l-Tyr by investigating the molecular organization comprises tightly packed dimers as the basic building block. We show that the addition of epigallocatechin-3-gallate (EGCG), known to act as an amyloid inhibition agent, can increase l-Tyr solubility due to the disruption of the dimer formation. Furthermore, the power generation of an l-Tyr crystal film-based nanogenerator device could achieve high values of current and voltage, namely, 35 nA and 0.5 V, respectively. Mechano-responsive bending and surface patterning composite is further fabricated, composed of l-Tyr crystal needles and polydimethylsiloxane (PDMS). Upon simply incorporating guest dyes, l-Tyr crystalline needles could exhibit the feature of optical waveguiding property. These results suggest potential future applications of the rigid and thermally stable l-Tyr-based molecular self-assembly in energy harvesting, photonic devices, stretchable electronic devices, and soft robotics.

Results and Discussion

The chemical structures of L-Phe, l-Tyr, and l-DOPA (Figure 1a) are highly similar, with the exception of the number of hydroxyl groups on the aromatic ring, namely, 0, 1, and 2, respectively. Based on this similarity, we set out to investigate the macroscopic properties of the respective molecular selfassembly. All three amino acids formed elongated crystals, as observed by optical microscopy and SEM (Figure 1b-d and Figure S1). The diameter of the crystals was 0.1-1, 1-10, and 20-200 μm for l-Phe, l-Tyr, and l-DOPA, respectively. The crystal structures of the three amino acids in this work were verified by comparing the patterns of experimental powder XRD with that of simulated ones from different polymorphic crystal structures of amino acids. As shown in Figure S2, the same peaks in XRD patterns were found to be in accordance with previous findings.^35–37 Hydrogen bonds are the main driving force for the packing of all the three amino acids. However, the face-to-face dimer building block of l-Tyr crystals is distinct in the hydrogen-bonded unit and in the crystal packing (Figure 1e-j).

Figure 1.

(a) Chemical structures of L-Phe, l-Tyr, and l-DOPA. (b-d) Optical microscopy images of amino acid crystals: (b) l-Phe (300 mM), (c) l-Tyr (10 mM), and (d) l-DOPA (100 mM). (e-j) Crystal structures of hydrogen-bonded unit and crystal packing of the three amino acids: (e,h) l-Phe, (f,i) l-Tyr, and (g,j) l-DOPA. The face-to-face dimer building block of l-Tyr crystals is distinct in the hydrogen-bonded unit and in the crystal packing.

Based on the amino acid polarity, we expected the water solubility of the three amino acids to follow the trend l-DOPA > l-Tyr > l-Phe. Upon experimentally investigating the water solubility of the three amino acids by the “heat-switch” method, where we dissolved the molecule by heating and allowed assembly to take place by cooling to room temperature, we were surprised to find out that this was not the case. Our results showed that l-Tyr has the lowest value of water solubility, approximately 0.5 mg/mL, whereas l-DOPA and l-Phe could be dissolved at ~5.9 and ~33 mg/mL, respectively (Figure 2a and Figure S3). It should be noted that the water solubility of l-Phe is over 60-fold higher than that of l-Tyr, even though they have very similar chemical structures and l-Tyr is more polar.

Figure 2.

(a) Water solubility of l-Phe, l-Tyr, and l-DOPA. (b) Thermal stability measurements of the three amino acid crystals using DSC. (c) Mass spectrum of l-Tyr in water showing both monomers and dimers. (d) Chemical structure of EGCG. (e) Inhibition of l-Tyr crystallization by adding EGCG after 3 h. l-Tyr concentration of 3 mg/mL (15 mM) for all of the experiments; molar ratio of EGCG/L-Tyr is as follows: left vial (0:1), middle vial (0.5:1), right vial (1:1). (f) Turbidity assay of various EGCG/L-Tyr molar ratios dissolved in water. (g,h) SEM images of assembled nanostructures from (g) l-Tyr and (h) l-Tyr/EGCG (1:1) at the concentration of 25 mM. (i) Powder XRD of only l-Tyr, only EGCG, and the mixture of l-Tyr/EGCG (1:1).

l-Phe and l-Tyr formed gels at a concentration of 300 and 50 mM, respectively (Figure S3a, b). l-DOPA is not a gelator and could not form a gel in water. In addition, DSC and TGA experiments were used to evaluate the thermal stability of the l-Phe, l-Tyr, and l-DOPA crystals (Figure 2b and Figure S4). DSC data allowed measurement of the melting temperatures and enthalpies of the material before and after deformation. A single endothermic peak in the DSC curve was observed for all three crystals at 260.2 °C (l-Phe), 282.6 °C (l-DOPA), and 312.0 °C (l-Tyr) with a melting enthalpy of 533.1, 509.8, and 765.5 J/g, respectively. TGA experiments showed that the crystals remained stable until the temperature reached 250.3, 277.4, and 308.3 °C for l-Phe, l-DOPA, and l-Tyr, respectively, in good agreement with the thermal stability results observed by DSC analysis. The results of DSC and TGA suggested the l-Tyr crystals to be the most thermally stable in the three amino acid crystals.

To better understand the cause for the poor water solubility and enhanced thermal stability of l-Tyr in comparison to l-Phe and l-DOPA, we examined the crystal structures of the assembled molecules. Strong face-to-face dimerization was found for l-Tyr (Figure 1f, i), which formed three-dimensional networks within the crystal as basic building blocks, in contrast to l-Phe and l-DOPA (Figure 1e,g,h,j and Figure S5). We suggest that the dimer formation by the hydroxyl group in l-Tyr inhibits the molecular binding with water molecules and therefore decreases the water solubility. To corroborate this hypothesis, we analyzed l-Tyr in water using mass spectrometry (MS), which indeed showed a strong dimer peak at m/z 383.4 (Figure 2c).

To further examine this explanation, a dimer disintegration experiment was conducted by using EGCG (Figure 2d), a principal polyphenol present in green tea that has been shown to be effective at preventing amyloid aggregation.³⁸ When EGCG was added into a solution of l-Tyr at molar ratios ranging from 0.1:1 to 2:1, the solubility dramatically increased, indicating that EGCG had an influence on the aggregation behavior of l-Tyr. When 15 mM concentration of l-Tyr was employed, a clear solution was observed when the molar ratio of EGCG/L-Tyr was higher than 1:1 (Figure 2e and Figure S6a), whereas an opaque solution with needle crystals could be found for only l-Tyr. By further increasing the concentration of l-Tyr to 25 mM, the kinetics of l-Tyr crystallization was also slowed down, but some precipitates were formed on the bottom of vials from 0.25:1 to 1:1 (Figure S6b). A negative control of EGCG solely formed a transparent aqueous solution. These clearly suggested that EGCG affected the local structure of l-Tyr by inhibiting the dimer formation through well-ordered intermolecular H-bonding between Tyr monomers as observed in crystal packing. It should be noted that the tyrosine crystals will not be disassembled by EGCG after stable dimer structures are formed. A turbidity kinetics assay was performed by measuring time-dependent UV-vis absorption at different EGCG/l-Tyr ratios (Figure 2f). The absorbance of l-Tyr solely in solution started to increase after 68 s and reached an optical density (OD) of 2.98 after 435 s due to the crystallization of l-Tyr. A delay in crystallization was observed for EGCG/l-Tyr (0.5:1) compared to l-Tyr solely, and a significant decrease in the OD was observed for EGCG/ l-Tyr (1:1), where the maxima of OD could reach only 0.7 after 1800 s. The result reveals that EGCG can inhibit the crystallization of l-Tyr in water to form an assembled opaque solution. Mass spectra of l-Tyr solution prepared in the presence of increasing amounts of EGCG showed a decrease of the dimer peak at m/z 383.4 in a dose-dependent manner, further confirming the inhibitory role of EGCG for l-Tyr dimer formation (Figure S7).

We further used SEM to study the morphological modulation of l-Tyr self-assemblies by the presence of EGCG as observed on the bottom of vials (Figure S6b). As shown in Figure 2g,h and Figure S8, l-Tyr self-assembled and formed short needles with the width of 1-5 μm without EGCG; however, significant morphological transition from needle-shaped crystals to thinner fibril structures was observed by increasing the amounts of EGCG from 0.25:1 to 2:1. The size of induced fibrils is in the range of 100-300 nm (Figure 2h). Powder XRD was employed to check the change of packing mode in the fibrous structures compared to only l-Tyr needles. As depicted in Figure 2i, the same pattern was observed for needles from l-Tyr and fibrils from l-Tyr/EGCG (1:1), indicating that EGCG can only disturb the self-assembly of l-Tyr by retarding the rate of dimer formation, but no specific co-assembly between l-Tyr and EGCG could occur. Taken together, the results support the hypothesis that l-Tyr dimers are much less soluble than the monomer, and that dimer formation could be the main underlying mechanism for the poor water solubility of the l-Tyr powder.

To examine the effect of molecular packing on mechanical properties of amino acid based crystals, AFM nanoindentation was applied to the surface of the crystalline needles on a flat glass coverslip (Figure S9). Both Young’s modulus and point stiffness were measured to check the mechanical rigidity (Figure 3a-f). The cantilever was brought to the surface of crystals and retracted at a constant speed of 2 μm s⁻¹, and the force-distance traces were obtained (Figure 3g). The Young’s modulus of the crystalline needles could be calculated at different positions after fitting the approaching traces using the hertz mode. As shown in Figure 3a-c, the values of Young’s modulus along the thickness direction were found to be 13.6 ± 6.07, 177.03 ± 54.39, and 55.98 ± 18.7 GPa for l-Phe, l-Tyr, and l-DOPA, respectively. Furthermore, the value of point stiffness could be further calculated according to the force–distance traces (Figure 3d,e). The statistical point stiffness values of the amino acid crystal needles were 84.21 ± 36.28, 632.77 ± 220.44, and 197.99 ± 69.64 N/m for l-Phe, l-Tyr, and l-DOPA, respectively. Thus, the rigidity sequence of the three crystals is l-Tyr > l-DOPA > l-Phe (Figure 3h), showing a negative correlation to the water solubility and a positive correlation to the thermal stability. These observations could be ascribed to the strong supramolecular packing of l-Tyr crystals formed by tightly packed dimers. Compared to the nanomechanical properties of natural biomaterials (Figure 3i), the Young’s modulus of l-Phe crystals is in the region of protein amyloids, whereas l-DOPA crystals reach a value between bone and glass.³⁹ Interestingly, a metallic-like rigidity was found for l-Tyr crystals in the nanoscale range. The physicochemical properties of l-Phe, l-Tyr, and l-DOPA are also summarized in Table S1 based on the results above.

Figure 3.

(a-c) Statistical Young’s moduli distributions of (a) L-Phe, (b) l-Tyr, and (c) l-DOPA. (d,e) Statistical point stiffness distributions of (d) l-Phe, (e) l-Tyr, and (f) l-DOPA. (g) Typical force-distance traces of l-Tyr crystals. The blue line corresponds to the fitting of the contact region in the “extend” trace using the hertz model. (h) Statistical Young’s modulus and point stiffness of the three crystals. (i) Comparison of the Young’s moduli of different biological and nonbiological materials, as well as l-Phe (red), l-Tyr (green), and l-DOPA (blue).

The noncentrosymmetric structure of l-Tyr with supra-molecular packing networks signifies internal polarization, implying potential piezoelectric properties. Moreover, due to its high thermal stability and mechanical rigidity, we further explored the piezoelectric properties of l-Tyr crystal powder. Computed piezoelectric strain constants of l-Tyr crystals, as obtained by density functional theory calculations, range from 5.0 to 9.7 pC/N.⁴⁰ The predicted piezoelectric coefficient is in the same order as those of traditional inorganic piezoelectric materials such as ZnO and CdS⁴¹ and biological materials such as bone, viruses, and fibrillar rat tail collagen.^42–44 The schematics of a fabricated l-Tyr-based nanogenerator and the energy conversion process are depicted in Figure 4a,b. The sandwich structure of the power generator was designed and fabricated by trapping a film of l-Tyr crystals with a 0.6 × 0.6 cm² filling area between two Ag-coated silicon substrates connected to an external low noise voltage preamplifier (Figure S10). As the linear actuator continues to apply the pressing force to the l-Tyr crystals, an electric dipole is generated in the piezoelectric l-Tyr crystal film, resulting in electrical current flowing to the top electrode. When the pressing force is released, the l-Tyr crystal film is no longer compressed and the current flows back to the bottom electrode.

Figure 4.

(a) Sandwich structure of an l-Tyr crystal-based nanogenerator. (b) Energy conversion process of the generator by pressing and releasing using l-Tyr crystals as active components. (c,d) Open-circuit voltage and (f,g) short-circuit current obtained from the generator in forward (c,f) and reverse (d,g) setups. (e) Linear fitting of the open-circuit voltage as a function of the applied force. Three replicates were measured and produced the standard deviation.

The output open-circuit voltage and short-circuit current of l-Tyr crystal film were measured during force application. By applying a pressing force of 31 N, the values of open-circuit voltage and short-circuit current reached up to 0.5 V and 35 nA, respectively (Figure 4c,f), exceeding the value achieved by other piezoelectric nanogenerators based on ZnO, monolayer MoS₂, polyvinylidene difluoride, and other biological materials such as bacteriophages.^45–47 Furthermore, the open-circuit voltage of the nanogenerator remained at a constant level at different pressing points, indicating the good dielectric property and negligible current leakage of the l-Tyr crystal film. To further confirm that the measured signal was truly from the piezoelectric l-Tyr crystals rather than a variation of contact resistance or parasitic capacitance, the testing was repeated with reversed connections, and the outputs were accordingly reversed for both open-circuit voltage and shortcircuit current (Figure 4d,g). The applied force-dependent open-circuit voltages were also tested from 0 to 31 N, showing a good linear fit with a slope of 10 mV/N and coefficient of determination (R²) of 98% (Figure 4e). These results suggest the potential application of the thermally stable and rigid l-Tyr crystals in energy generation biodevices.

Stimuli-responsive bending materials have gained great interest due to their potential applications in stretchable electronic devices, wearable biointegrated devices, and soft robotics.⁴⁸ In order to further explore the potential application of l-Tyr crystals, a composite of soft polymer PDMS coated with a thin film of rigid l-Tyr crystals on the surface was produced (Figure 5a). Surprisingly, as shown in Figure 5b, the composite showed mechano-responsive bending by tensile stress, whereas noncoated PDMS did not bend by the same mechanical treatment. We suggest that the mechano-responsive bending is induced by the differences in the mechanical properties of the l-Tyr crystals on the surface, compared to PDMS; for example, the Young’s modulus of l-Tyr crystals is much higher than that of PDMS (<1 GPa).⁴⁹ After being triggered by tensile stress, the elastic behavior of the PDMS is different from that of the l-Tyr crystal surface. As a control, PDMS alone and a composite of PDMS mixed with l-Tyr within the sample, rather than on the surface, showed very little bending under the same tension force (Figure 5b right and Figure S11). The recovery between bending and the flat shape was also confirmed by a mechanical force, as shown in Figure S12. By applying a compressing force of 20 mN, the displacement of the bending height was observed at the value of 2.75 mm. After the forces were released, the relatively flat shape was retrieved at the original bending height. This result reveals the l-Tyr crystals as potential rigid components for the design of soft-matter-based electronics as well as for biointegrated materials. Moreover, surface patterning was obtained by preparing and coating “TAU” shapes made of l-Tyr crystals on the PDMS surface (Figure 5c). The microscopic structures of l-Tyr and PDMS in the pattern were studied and are shown in the right of Figure 5c, with fibrous crystals visible in the region of l-Tyr crystals, in contrast to the few nanostructures visible in the PDMS region. The maximum excitation and emission wavelengths of l-Tyr crystals are 400 and 450 nm, respectively.33 Blue fluorescence was observed for the l-Tyr crystals under UV light excitation, whereas the PDMS area did not emit any fluorescence (Figure S13). Based on the differences between the rigidity, piezoelectricity, and fluorescence properties of l-Tyr and PDMS, the patterned surface composite holds great potential for future applications in soft robotics, stretchable electronics, and cell imaging.

Figure 5.

(a) Schematic depiction of the mechano-responsive bending composite composed of l-Tyr and PDMS. (b) Bending property of the composite (left) and PDMS alone (right). (c) l-Tyr surface patterning of “TAU” on PDMS and an optical microscopy image of the l-Tyr and PDMS regions. (d) Bright-field image of the needle crystal incorporating RhB dye. (e,f) Photoluminescence (PL) images of direct observation of waveguiding with local excitation at one end of the needle crystal. The red circle marks the excitation area and the green arrow denotes the out-coupling of PL emission at the other end.

Recently, bioinspired self-assembled nanostructures with defined spatial dimensions from peptides have been considered as promising candidates for fabrication of photonic and electric materials in biomedical fields.⁵⁰ However, to achieve this purpose by using biological amino acids as effective building blocks is still rarely reported. Thus, we further studied the optical waveguiding property of l-Tyr needle crystals with the incorporation of a guest dye rhodamine B (RhB) by using an experimental setup of local illumination microscopy (Figure S14). Figure 5d shows a bright-field image of the single isolated Tyr-RhB crystal. By focusing the excitation light on one end of the needle (red circles), the emitted photoluminescence (PL) was observed on the other end (green arrows), which confirms a typical feature of optical waveguiding (Figure 5e,f). If the excitation point is moved to the middle position, bright PL spots were observed at both ends of the crystal (Figure S15), whereas the PL emission should be only observed at the local area of the excitation position without guiding. The result indicates that the l-Tyr crystal upon simply incorporating guest dyes allows for the propagation of PL emission along the longitudinal axis and can be used as potential optical waveguide material.

Conclusion

In summary, the effect of small differences in the chemical structure of specific amino acids on the physicochemical properties of the resulting crystals is thoroughly investigated. Properties such as water solubility, self-assembly morphology, thermal stability, and mechanical properties were evaluated. l-Tyr crystals are highly insoluble due to previously unreported closely packed dimers, which serve as the building blocks for the crystallization process. Compared to L-Phe and L-DOPA, l-Tyr crystals also show a higher thermal stability and mechanical properties. The l-Tyr crystals generate relatively high and stable electric power outputs under pressure force in a sandwich piezoelectric device. By incorporating the rigid l-Tyr crystals in the soft polymer PDMS, a mechno-responsive bending composite was successfully developed. Furthermore, the l-Tyr crystalline needles with simple incorporation of dyes exhibit active optical waveguiding, making them promising candidates for the generation of photonic devices toward amino acid based biomaterials. This study exemplifies a feasible strategy to explore the utility of thermally stable and rigid molecular self-assemblies formed by minimalistic building blocks, such as amino acids, for power generation and smart advanced materials. Based on the advantages of ease of production, low cost, and environmentally friendly nature, these amino acid crystals hold great potential for future applications in energy harvesting, optical waveguiding, stretchable electronics, and soft robotics.

Methods

DSC and TGA

Melting and decomposition temperatures were recorded by differential scanning calorimetry (NETZSCH STA 449F5, Germany) at a heating rate of 5 K·min⁻¹ under nitrogen atmosphere with a flow rate of 50 mL min⁻¹ and at a temperature range between 50 and 500 °C.

Young’s Modulus Measurement

The AFM nanoindentation experiments were performed with a commercial AFM (JPK, Nanowizard II, Berlin, Germany). The force curves were recorded using the commercial JPK software and analyzed by a custom-written procedure based on Igor pro 6.12 (Wavemetrics Inc.). The RTESPA-525 cantilevers (Bruker Company, half-open angle of the pyramidal face of θ <10°, tip radius ~5 nm, frequency in air ~525 kHz) were used in the experiments. The spring constant of the cantilevers was approximately 200 N m⁻¹. The maximum loading force was set at 3 μN for the crystals. All AFM experiments were carried out at room temperature.

In a typical experiment, crystals were cast in the surface of a mica substrate, and the cantilever needed to be moved above the samples. For l-DOPA crystals, the locations of the samples were accomplished with the help of an optical microscope. For the measurement of F and Y crystals, the location of the crystals on the substrate was determined by AFM scanning due to the small size of the crystals. Then the cantilever was brought to the crystal samples at a constant speed of 2 μm s⁻¹ and held on the crystal surface at a constant force of 3 μN. After the extend-retract force curve was recorded, the cantilever retracted and moved to another spot to perform the next cycle. Young’s modulus of the crystals was obtained by fitting the extend curve to the hertz model (eq 1).

$$F(h)=\frac{2}{\pi }\tan \alpha \frac{E}{1-{\nu }^2}{h}^2$$ (1)

in which F corresponds to the cantilever, h corresponds to the depth of the crystal pressed by the cantilever tip, α corresponds to the half angle of the tip, E corresponds to the Young’s modulus of the crystals, and v corresponds to the Poisson ratio. We chose v = 0.3 in our calculations. Typically, five to eight such regions (2 × 2 μm, 400 pixels) were randomly selected on each sample to construct the elasticity histogram.

Point Stiffness Measurement

By considering the crystal and cantilever as two serial springs, the measured point stiffness (k_meas) comprises the stiffness constants of the cantilever (k_can) and the crystals (k_cry). To estimate the material property of the crystals, we assume that the mechanical behavior of the crystals can be described as linear elastic, which is extremely suitable for solids under small strains. Thus, the point stiffness of the crystal can be calculated using eq 2.

$$k_{\text{cry}}=\frac{k_{\text{can}}\cdot k_{\text{meas}}}{k_{\text{can}}-k_{\text{meas}}}$$ (2)

The point stiffness histograms were also constructed according to the calculation from extend-retract force curves in AFM nanoindentation experiments.

Fabrication of an l-Tyr-Based Nanogenerator

A silver film was fully coated on a silicon substrate by magnetron sputtering. l-Tyr crystal powders (4 mg) were filled into a PDMS (Sylgard184) protective layer with a 0.6 × 0.6 cm² filling area, which was sandwiched between two silver-coated silicon substrates serving as top and bottom electrodes. A Kapton tape and a double-sided adhesive tape were closely absorbed between the PMDS layer and the bottom and top electrode, respectively, to avoid the formation of gaps inside the device. A PDMS film coated on the top electrode of silicon served as a buffer layer that prevented breakage caused by a large impact force from a linear motor. A large glass was placed on the bottom electrode for easy testing. Two copper wires were fixed on the top and bottom silver films as electrodes to connect to the outside circuit. Finally, a Kapton tape layer was tightly pasted on the surface of the device as an encapsulation layer.

Power Generation

The nanogenerator based on l-Tyr crystals was firmly fixed onto a stainless-steel plate. The force periodically applied to the nanogenerator was produced from a linear motor (E1100-RS-HC type with force control, LinMot). A Stanford SR560 low noise voltage preamplifier was used to collect the output signal of the generator. A force sensor (DYZ-101, Bengbu Ocean Sensing System Engineering Co. Ltd.) was used to measure the driving force applied to the device. To avoid a Faraday signal from the environment, all measurements were conducted on a Faraday cage and an aluminum foil was fully coated on the nanogenerator and stainless-steel plate.

PDMS and l-Tyrosine Composite

l-Tyrosine was dissolved in double-distilled water to a concentration of 5 mg/mL. As l-Tyrosine is poorly dissolved in water, the solution was heated to boiling temperature and stirred until no precipitation was visible in the vial. The solution (3 mL) was then poured into a container (either a small Petri dish, or a plastic weighing boat) and allowed to cool to room temperature. Once the solution cooled, l-Tyrosine self-assembly-initiated and elongated needle-like structures were formed. The container was incubated at room temperature for at least 24 h, allowing the water to evaporate completely, forming a thin film of l-Tyrosine structures. Four milliliters of PDMS (Sylgard 184, 10:1 of silicone elastomer base/silicone elastomer curing agent) was poured at the desired amount on top of the l-Tyrosine thin film and incubated at 63 °C for 1 h for curing. After curing, the PDMS and l-Tyrosine composite was obtained by removing the container. As a control, PDMS was prepared without the l-Tyrosine film. The size of the composite and only PDMS in Figure 5b are 31.7 mm × 3.1 mm × 7.2 mm and 32.8 mm × 2.4 mm × 6.6 mm in length × width × height, respectively. To achieve the patterns of l-Tyrosine, Scotch tape was used, designing the desired pattern on the container (plastic weighing boat, in the case of the TAU patterning). The Scotch tape was removed before pouring the PDMS on top of the film, leaving just the desired pattern on the surface of the container. The compression and release of the recovery form bending to a flat shape was achieved using a tensile-compressive tester (Instron-5944 with a 2 kN sensor) in air at room temperature. The strain rate was maintained at 1.65 mm min⁻¹, and the force and the displacement were recorded.

Optical Waveguiding

To evaluate the potential of the l-Tyr crystals as active waveguide materials, rhodamine B (RhB) was used as a probe as light has a certain loss in the transport of tyrosine crystals. RhB at concentration of 1 mg/mL in hexafluoroisopropanol was incubated with l-Tyr crystals overnight, leading to the spontaneous accommodation of the dye within the needles. The crystals were then cleaned with double distilled water three times to remove residual RhB. The optical waveguide photographs were obtained using an Olympus FV1000 microscope with a CCD camera or a Leica DM IRBE microscope equipped with a digital camera. A laser with an excitation source of 561 nm was employed for the measurement. A scheme of the needle crystal waveguide characterization setup is shown in Figure S14.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsnano.9b08217.