Natural cellulose shows a large number of hydroxyl groups, being demonstrated by its chemical distinction in the FTIR spectrum (Fig. S1), which leads to a high Zeta potential value (~-25 mV) (Fig. S2) (Li et al. 2019). Such distinction enables ion-selective transport, such as cations in the nanochannels of cellulose membrane, thus being capable of harvesting osmotic energy. Conventionally, the cellulose membrane shows a dense structure due to the cellulose-molecule interwinding behavior, which can hardly support the effective transport of sufficient ions in such few nanochannels or pores of the cellulose membrane. Via PVP nanoparticle inserting-removing engineering, enough nano-scaled pores are reasonably formed in the PCM. As shown in Fig. 2a, 2b, and S3, the PCM demonstrates a rough surface and porous structure, more specifically, a large amount of nanopores is observed in the PCM system, which is absent for the OC. We further investigated the pore size distribution of cellulose membrane, in which PCM shows an increased amount and larger volume of nanopores than OC (Fig. 2c and 2d). Note that the nanopores of PCM with a diameter less than 10 nm feature a desirable volume of 0.0125*10− 3 cm3/g, much higher than that of OC, just 0.00853*10− 3 cm3/g (Fig. S4).
As a result of porous structure, our PCM shows excellent porosity of 89%, a water retention value of 141.1%, and a specific surface area of 12.86 m2/g, while the OC is just 66.2%, 113.4%, and 7.99 m2/g, respectively (Fig. 2g, S5, S6, and S7). Such porous structure also endows the high hydrophilicity (low water contact angle) to PCM (59.8°vs. 44.5° for OC and PCM in Fig. 2e and 2f), which supports the water and ions transport in its nanopores (Sun et al. 2020). The nanopore structure of PCM strongly relies on the PVP dosage, where exceeding PVP forms large-sized pores with a diameter > 10 nm, which on the contrary decreases its porosity and specific surface area.
The PCM was subsequently secured within a dual-chamber electrolytic cell to assess its ion transport performance. Ag/AgCl electrodes were utilized to measure the current-voltage (I-V) characteristic of cellulose membrane in varying concentrations of NaCl electrolyte (Fig. 3a). As depicted in Fig. 3b, the I-V curve of PCM demonstrates linear ohmic behavior of cellulose membrane across NaCl concentrations from 0.01 to 1 M, which suggests the symmetric configuration of the PCM. Such a result is in agreement with the conclusion of the I-V curve of the PCM in the forward and backward models, where it shows the same absolute value in voltage and current as the two models (Fig. S8).
Additionally, we evaluated the ionic conductivity of the PCM (Fig. S9). In regions of high ionic concentration, the ion conductivity of cellulose membrane exhibits a linear relationship with ionic concentration. However, when the concentration falls below 10− 3 M, the conductivity stabilizes. Such a result demonstrates the surface charge of nanopores controls ion transport in PCM structure. Moreover, the PCM has higher ionic conductivity than the OC, enhancing its suitability for osmotic energy harvesting.
Figure 3c and 3d display the I-V curves of cellulose membranes under varying NaCl concentrations. It is observed that the greater the concentration gradient, the greater the Voc and Isc for both OC and PCM. More importantly, both the values of Isc and Voc for the PCM exceed these of the OC membrane under all tested conditions, indicating superior selectivity and efficiency of ion transport. In addition, the cationic transfer number (t+) of the PCM, as shown in Fig. S10, is higher than 0.88, further evidencing its exceptional ion selectivity. The current-time curve results (Fig. 3e) demonstrate that our PCM maintains stable ion transport, with both adverse and positive currents showing minimal fluctuation over extended periods exceeding 3600 seconds under alternating external bias voltages ranging from − 1V to 1V.
The osmotic energy captured by the PCM was quantified by linking it to an external circuit equipped with a load resistance (RL). The current density (CD) decreases as the load resistance increases, yet the power output peaks at a resistance of 4k Ω (Fig. 4a). With a 50-fold NaCl concentration gradient, the PCM delivers an output PD of 1.33 W/m2, doubling that of the OC membrane. When the salinity gradient is raised to 500 folds, the PCM achieves a PD of 4.16 W/m2. As the concentration gradient intensifies, the resulting CD correspondingly rises (Fig. S11).
Furthermore, we examined the impact of PVP content on the performance of PCM. Shown in Fig. 4b and 4c are the PD under the 50-time concentration difference of NaCl. It is observed that the PD of the PCM firstly increases and then decreases as the PVP content rises, achieving the highest value at a PVP dosage of 0.02 g. PVP effectively adjusts the microstructure of cellulose membrane, more specifically, endowing the PCM with higher porosity (89%) and higher surface charge (12.86 m2/g), correspondingly supporting the high-performance selective transport of ions in PCM structure, thus superior capability of harvesting salinity energy.
As we know, the electrolyte type remarkably affects the efficiency of energy conversion (Duan & Majumdar, 2010). We assessed the PD of the PCM with various electrolytes, including KCl, NaCl, and LiCl, as illustrated in Fig. S12. Notably, the KCl setup provided the highest PD to the PCM, peaking at 1.6 W/m2 for a 50-fold KCl concentration gradient. This enhancement is linked to the high ionic diffusion coefficient of KCl. (Siria et al. 2017). As depicted in Fig. 4d, the PCM sustains a high PD of 1.33 W/m2 over 8000 seconds without needing electrolyte replacement, illustrating its robustness in osmotic energy production. Furthermore, the PCM exhibits remarkable operational stability, delivering a power output of ~ 1.3 W/m2 for a continuous period of 14 days (Fig. S13), which validates the desirable applicability of our PCM toward practical osmotic energy harvesting.