Multisource energy conversion in plants with soft epicuticular coatings

Turning common plants into devices harvesting electricity from wind and radio frequency radiation endows a surprising prospect for energy-autonomous sensors.

. Circuit model used to estimate the behavior of two leaf generators on the same or separate plants. Figure S2. Effect of the leaf coating on self-healing of wounded leaves. Figure S3. Mechanical energy conversion by c-u leaf pairs. Figure S4. Scanning electron microscopy images of the coated H. helix leaves and effect of coating thickness on the voltage output. Figure S5. Mechanical-to-electrical energy conversion efficiency of a c-u leaf pair. Figure S6. Influence of climber H. helix' support material on the mechanical energy conversion. Figure S7. Model for the estimation of the behavior of signal generation and energy harvesting with multiple leaves on the same plant. Figure S8. Influence of leaf-wetting on c-u leaf pair energy conversion. Figure S9. Electric field characteristics and field strength-dependent energy harvesting using H. helix. Figure S10. Effect of ion conduction obstruction by tissue drying on energy harvesting and impedance analysis. Figure S11. Application scenario: Signals received during plant-based RF energy harvesting in an urban outdoor environment. Table S1. Overview of plant-hybrid energy harvesting and sensing techniques.  helix leaves on the process of wilting after the leaves (oL: old leaves, yL: young leaves) were removed from the plant and subsequently observed for 65 days. A clear color change and wrinkling of both, coated and uncoated leaves can be seen which was not observed when the leaves were kept alive on the plant. c) Effect of coating on recovery from wounding in a Ficus benjamina for which self-repair mechanism by coagulation of plant sap latex is well known. 41 The experiments show that both, coated and uncoated leaves self-heal and remain alive. d) Microscopy of the latex coagulation of a coated leaf within 30 min after wounding.
All experiments clearly confirmed that coated leaves overcome a wounding event similar to uncoated leaves and that thus the coating does not reduce the leaves' survival. Instead, it may even have a protective effect against wounding by, e.g., herbivores. Self-healing was tested on 5 leaves of the same plant per condition (coated/uncoated) treated in the same manner, the results give representatives images. Alternating voltage signals generated by a c-u pair of F. microcarpa @ 5 Hz, 1 N stimulus showing significant enhancement of output voltage after epicuticular coating. c) Zoom-in of the voltage signal generated by the u-u pair with an amplitude of ~45 mV. The smoothing spline reveals the 5 Hz stimulus signal. The amplitude of the c-u pair is thus ~450 times higher as achieved by the u-u pair. d) Short circuit current generated by a c-u pair of F. microcarpa @ 5 Hz, 1 N stimulus. stimulation (1 N, 5 Hz, 1 cm² contact area) between the coated and uncoated leaf was applied by an actuator in a Faraday cage and the voltage was measured between electrodes inserted in the petiole of the two leaves. The x and y error bars represent average and standard deviation of six leaves subject to 150 contacts and six different coating thicknesses (a total of 900 tests). The results suggest that there is an optimum of the coating thickness (here 135 µm) that, under the conditions of the experiments, leads to highest voltage generation (up to 120 V). The increase could be related to the dielectric properties but also to an optimal softness and adaptation to the contact surface. Controlling coating thickness could thus be used further tune the output of the c-u leaf pairs. Figure S5. Mechanical-to-electrical energy conversion efficiency of a c-u leaf pair. A certain mechanical energy E mech,in was applied to a c-u leaf pair of H. helix by a pendulum as described in the methods section and the resulting electrical energy produced was determined from the current measured over a 10 or 100 MΩ resistor, respectively. The results show that the leaves' mechanical-to-electrical energy conversion efficiency n M, overall varies within a magnitude dependent on the kinetic energy introduced. It was found higher at lower input energy reaching values of up to 0.14 %. A reason for this behavior could be that energy conversion efficiency is a function of the impact force which increases at higher E mech,in and has an maximum at a given force as suggested earlier. 36 Each datapoint represents mean and standard deviation of 3 individual c-u pairs that were tested by 6 current measurement per E mech,in (total of 360 measurements). given distance from a support material. The graph below shows the generated voltage signals (10 Hz stimulus) using PMMA, wood, PDMS, PTFE, and another leaf as support material when analyzing c-u and u-u pairs. PDMS and PTFE enhance the signal which can be explained by an additional contact charging due to contact of the leaf with this surface.
PMMA, wood, and leaf as a substrate does not significantly change the signal. Using the u-u pair, the signal is expectedly lower and only slightly increased by the substrate and the main contribution for generated voltage is the c-u and u-u leaf pair.  Figure 4j). The cumulative energy could be even higher than when each signal could be harvested individually as positive peaks occurring simultaneously sum up as well.
Moreover, the estimation shows that E in both cases increases the more leaves are used.
This suggests that upscaling using multiple leaves that produce signals on the same plant could expectedly be possible. This suggests that upscaling using multiple leaves that produce signals on the same plant expectedly is possible. Our experimental results with eight leaves (Figure 4j and 5f) show a positive power balance and signals did not cancel out. Yet, the model should be confirmed by further experimental results with larger amounts of leaves.
Limitations of the model that require experimental observations are for example the fact that leaves on the same branch often underly similar mechanical constraints that lead to similar and potentially to motion which is to some extent synchronized. Hence signals from multiple leaves would not appear randomly but moderately synchronized which could further improve the power balance by reducing the event that a positive spike hits a negative spike of another leaf leading to canceling out of the signals. Furthermore, the circuit model needs to be extended by including factors like capacitances of the tissue or different internal resistances for the individual leaf generators. Yet, plant-based harvesters using a few leaves presented here can already power commercial electronics like wireless sensor nodes. Further upscaling and optimizing the model as well as the energy outputs is part of our future investigation. Figure S8. Influence of leaf-wetting on c-u leaf pair energy conversion. A H. helix c-u leaf pair was exposed to a 10 Hz mechanical stimulus and voltage amplitude is recorded before and after wetting the leaves by spraying water. Wetting expectedly strongly reduces contact electrification and the obtained voltage amplitudes. However, the signal recovers when the leaves dry again.