3.1 Influence of delignification on water transport and distribution in wood
Herein, basswood as a light, easy-to-work-with wood was used and it could be easily processed into a variety of shapes as required (Fig. 2a). To acquire the wood matrix with hydrophilic channel alignment for efficient water transport and thus favor the generation of ion gradients, the removal of hydrophobic lignin from wood feedstock was first performed. As shown in Fig. 2b–g, original wood possesses a hierarchical cellular structure with well-oriented hollow channels along the longitudinal direction. After delignification, wood became light-colored but its external dimension and honeycomb-like structure were well preserved. High-resolution SEM images reveal that numerous nanoscale pores were generated and densely aligned cellulose nanofibrils were exposed on the surface of the cell walls due to subtractive chemical modification (Fig. 2d, g and Fig. S1, S2). From the results of the contact angle measurement, it can be found that delignification makes the wood more hydrophilic, allowing passive capillary force to move water quickly. As shown in Fig. 2h, DW allows the water droplet to be absorbed in less than one second while the absorption of the water droplet on NW cannot be completed in ten seconds. The chemical composition of wood specimens before and after delignification was further identified using FT-IR spectroscopy and chemical composition testing (Fig. 2i, j). The absence of the characteristic lignin bands at 1594 cm–1, 1505 cm–1, and 1457 cm–1, and presence of the hemicellulose-related bands at 1735 cm–1 and 1238 cm–1 were observed after chemical treatment, suggesting the selective removal of lignin [34]. The quantitative chemical composition analysis also reveals that most of the lignin and a portion of hemicellulose were removed after delignification. Additionally, it could be seen from the XPS spectra that oxygen-containing functional groups including surface-exposed –OH and –COOH groups in DW were significantly increased after delignification [5] (Fig. 2k and S3).
Considering that the transport of water in wood would significantly affect the hydroelectric power generation, different migration behaviors of water in NW and DW were clarified in detail. As shown in Fig. 3a, c and d, no matter the transport direction, DW had a faster water transport rate than NW due to the increased hydrophilicity within pores and channels of various sizes. With a continuous infusion of water, the progression of the flow front varied as DW// > NW// > DW⊥ > NW⊥ (//: parallel growth direction; ⊥: vertical growth direction). Obviously, there was a faster transport rate at parallel growth direction than at vertical growth direction. It could be ascribed to pump-free water transport via capillary forces and anisotropic water transport owing to the pronounced directionality of channels [31]. A similar result was obtained via screenshots of sessile drop tests tracking the flow of water through wood before and after delignification (Fig. 3b, Fig. S4, and Video S1). In addition, DW can absorb more water than NW if in contact with it; the longer the time of wood delignification, the stronger is its water absorbing capacity (Fig. 3e). It is because that decreasing lignin level over time results in an increase in hydrophilicity (Table S1). As shown in Fig. 3f, when placed vertically along the parallel growth direction, DW exhibits an obvious water content-height correlation compared with NW. More hydrophilic wood contains more water in it, and the decrease of the height results in gain in gravimetric water content because of the force of gravity. It would be crucial to in-situ generation of the concentration gradient of ions, with consequent effects on the preparation and performance of WEGs.
3.2 Water-induced electricity generation via ion concentration gradients
It is well-known that untreated industrial wastewater as a complex matrix is usually too acidic or too alkaline. Hence, HCl and NaOH as typical strong acid and strong base, respectively, were first utilized to investigate their effect on WEG performance. Generally, a typical WEG is composed of a pair of Pt mesh electrodes and a piece of DW in between (Fig. 4a). As mentioned above, numerous nanochannels between the aligned cellulose nanofibrils and additional nanopores were generated upon delignification. These water-conducting nanoscale pathways can easily have charged surfaces because of the presence of abundant oxygen-containing functional groups. When in contact with DI water, the WEG mainly relies on a water-evaporation-induced flow of H+ ions dissociated from negatively charged functional groups (–OH, –COOH) on the wood surface to generate electricity (Fig. 4b). However, the resulting streaming potential and current was small due to limited capabilities to dissociate. As shown in Fig. 4d, and the voltage of the WEG is only 0.25 V. To obtain a pronounced macroscopic surface charge gradient over the wood, unequal distribution of HCl/NaOH solution was created by discontinuous addition of water (Fig. 4c). Unless otherwise specified, DI water addition at a rate of 3 mL/h was performed via an electronic peristaltic pump. When adding water to the top of the DW immersed in a HCl/NaOH reservoir, capillary action helps bring the HCl/NaOH solution upward and gravity pulls water downward. A combination of these two effects leads to a bottom-to-top asymmetric distribution of the HCl/NaOH solution and thus constructs an ionization gradient over the wood as anions remain immobile on it. Under such a gradient, the surface-charge-governed ion selectivity and transport along the fibre direction could be easily regulated.
As shown in Fig. 4e, when raising the concentration of sodium hydroxide solution from 0.01 to 4 M, the output voltage varies from 0.26 to 1.1 V. Obviously, the higher the concentration, the stronger the base. It could bring a large difference of ionic strength and surface charge between both ends of DW, resulting in significant high voltage output. Of note, the WEG immersed in a reservoir of NaOH had a wider variation range of the output voltage, as compared to the counterpart in a HCl solution (Fig. S5). In the case of utilizing a reservoir of NaOH, hydrogen bonding between polar functionalities (i.e. hydroxyls, carboxyls) on the cellulose nanofiber surface could significantly weaken with the increase of the OH− ion concentration. The functional groups of cellulose molecular chains were gradually deprotonated, and the surface negative-charge density was continuously increased as evidenced by the change of zeta potential of cellulose surface under different conditions (Fig. 4f, and Fig. S6a). When using a NaOH reservoir with a concentration above 0.1 M, electric energy harvesting from water evaporation accounts for a low proportion; that is, the main source of electricity production is the ion concentration gradient over the wood. The large deprotonation variations that occur due to a large concentration difference of OH− ions between both ends of the wood result in an enhanced selective ion diffusion between the positive and negative ions. In turn, the facilitated upward diffusion of Na+ ions and exclusion of OH− ions via ion selectivity of the electric double layer (EDL) effect [30] brings about the increased potential difference and the current. When the reservoir falls within the acidic range, there is a decrease in voltage output (Fig. S5, and Fig. S6b). The generation of a notable potential difference is impeded due to the protonation-induced decrease in negative charge density and ion selectivity of nanochannels (Fig. S6b) [35]. Significantly, the use of strong alkaline electrolytes as the reservoir enhances the negative charge density and intermittent water addition ensures the continuing presence of the ion concentration gradient over the wood for highly conductive and selective ionic transport.
To further clarify the key role that water addition-induced concentration difference along the fiber alignment direction in DW plays in enhancement of electrical performance, the output voltages of the WEG using a reservoir of NaOH (2 M) with and without addition of water on top were measured. As shown in Fig. 4g, a big variation in the longitudinal distribution of negative surface charges in aligned nanochannels could be created upon adding water, thus bringing about an ion concentration gradient from bottom to top over the wood. Consequently, the output voltage was significantly increased to ~ 1 V. By contrast, the generation of an ionization gradient over the wood could not be maintained without adding water, resulting in an output voltage of less than 0.2 V. As the evaporation enthalpy of the NaOH solution is greater than that of pure water under the same conditions, reduced water evaporation results in a smaller voltage than that of the WEG immersed in DI water. Besides, DI water was added to the DW and NW, respectively, which served as a supplement to illustrate the significance of an ion concentration gradient over the wood for the electrical outputs. As shown in Fig. 4h, DW generated an output voltage of ~ 1 V while NW had an output voltage of ~ 0.5 V. It can be ascribed to the decreased content of surface functional groups in NW, and the inconvenience of DI water infiltration into NW with poor hydrophilicity (Fig. S7, S8, and Video S1). These observations once again indicate that the existence of nanochannels combined with dissociated functional groups and NaOH concentration difference are crucial factors in enhancing voltage output. As shown in the Fig. 4i, when the voltage of DW reached about 1 V, the continuing addition of water only increased the voltage by a small margin (from 1 to 1.1 V) and the water addition rate could be lowered to 200 µL/h to maintain a stable voltage output. Because the NaOH concentration difference between both ends of DW gradually tended to be stable, no obvious changes in the output voltage were observed. In addition, cyclic voltammetry (CV) was performed on the whole WEG device in DI water and NaOH solution, respectively, and no distinct redox peaks were observed (Fig. S9).
3.3 Evaluation of electrical performance of WEGs
To further investigate the output performance of the WEG, the ionic conductivity of DW was characterized using electrochemical impedance spectroscopy (EIS) at room temperature. The results show that the conductivity of DW was better than that of NW, and the conductivity increased with NaOH concentration owing to high ionic strength and highly charged surfaces inside the nanospace (Fig. S10 and S11). When the concentration of NaOH was not less than 2 M, the WEG could generate a remarkably high output voltage approaching 1.1 V, and a current with orders of magnitude higher than that in DI water (Fig. 5a, b, and Fig. S12). Compared with previously reported all-biomass-based power generators using ion concentration gradients, our WEG has the highest voltage and current outputs (Table S2). The comparison focuses on, in our case, the performance of green WEGs with inert electrodes since prior studies have confirmed that non-inert electrodes could initiate a redox reaction at the material interface to affect the electrical performance [5, 36–39]. Besides, as the height was increased from 0.5 to 2 cm, the voltage of the WEG reached 1.1 V, up from 0.5 V. Further increase in height would exert a negligible influence on voltage output performance, ensuring a stabilized output voltage (Fig. 5c and Fig. S13). Obviously, insufficient height is not conducive to the generation of large ion concentration gradients. When it comes to the cross-sectional area of the WEG, the voltage remains unchanged and the current increases linearly with an increase in the cross-sectional area (Fig. 5d). Moreover, common wood species including balsa wood, basswood, and beech wood have similar voltage output performance, illustrating the wide applicability of the proposed design (Fig. 5e). Considering that the ionization degree of bases could affect the concentration of OH− ions in the solution, we further compared the voltage output performance of the WEG in different alkaline solutions. The results show that the voltage output in the strong alkaline solution was higher than that in the weak solution (Fig. 5f). As shown in Fig. S14, weak bases with the same molar concentration exhibited a relatively low pH (i.e. a low concentration of OH− ions) due to incomplete ionization. Considering that the output power acts as an important indicator of its operating performance in practical applications, we tested the power outputs under different load resistance. The curve of load resistance versus voltage and current outputs were shown in the Fig. 5g. The power output of WEG was about 27 µW when the load resistance was about 3000 Ω (Fig. 5g, h), which was 3 orders of magnitude higher than that in DI water. A 3-unit WEG could charge a 47 µF capacitor to about 3 V in less than 60 s (Fig. 5i, and S15). The above results indicate that it is an effective way to improve the output performance of the WEG by utilizing a strong alkaline reservoir and then introducing the concentration gradient.
3.4 Preparation and application of wastewater-based WEGs
As we all know, alkaline wastewater widely exists in typical industries, such as mineral processing, pulp manufacturing, and packaging and printing. High levels of alkalinity as a bane can lead to serious environmental issues. But here, it may prove a boon to energy production since we have successfully demonstrated the feasibility of such WEGs using alkaline electrolyte reservoirs. Ideally, the WEG utilizes untreated alkaline wastewater to produce electricity, and promotes energy sustainability (Fig. 6a). To check for this possibility, a series of common alkaline wastewater, including aluminium smelting wastewater, printing wastewater, pulping black liquor, and silicate wastewater, was selected and utilized (Fig. 6b). As shown in Fig. 6c, all the four WEGs could provide a voltage not less than 0.46 V, and a voltage of up to 0.98 V was obtained for the WEG using pulping black liquor (Video S2). It is notable, that in spite of having similar pH values, the other three types of wastewater except printing wastewater exhibited varying output behaviors. Obviously, industrial wastewater has a wide variety and complex chemical composition, and these organic and inorganic substances have an impact on the electrical performance of the WEG. Besides, the WEG in the pulping black liquor could deliver stable voltage output (~ 0.98 V) with almost no attenuation for more than 11 h, indicating its long-term stability of the electrical output (Fig. 6d).
Obviously, our WEG with superior electrical output performances could fit into various application scenarios. Also, our design is economically acceptable and environmentally responsible. As mentioned above, very little water is required to maintain a stable electrical output performance of the WEG. The easy access to water harvesting provides self-sufficiency to water use, such as rainwater collecting and dew collecting. As shown in Fig. 7a–c, the two-unit WEG could drive the electronic calculator to work without any rectifiers, and the four-unit WEG could directly light up the LED (Video S3, S4). Also, the WEG stored the generated electrical energy in the supercapacitor, and three supercapacitors in series could light up a LED. In addition, by capturing water vapor in the exhaled air, the WEG could well monitor the breathing rhythm especially for people with respiratory system disorders (Fig. 7d–f). It is because that breathing causes asymmetric distribution of ions in DW to generate electrical signals. In another example, the WEG could employ the skin sweat to produce electricity, making it possible to be used in wearables (Fig. 7g–i and S16). The above applications fully illustrate that the WEG with excellent electrical performance and high output power could meet the application requirements of various scenarios, and promote the application of wood in sustainable energy and green electronics.