Cellulose Nanopaper for Solar Cells

As mentioned before, the cellulose nanopapers w ith both high transparency and haze have been applied in the field of solar cells. These nanopapers with the single function, i.e., light management, achieved a dramatic efficiency enhancement of solar cells by simple coating when compared w'ith the bare cells (Fang et al. 2014. Ha et al. 2014, Jia et al. 2017). However, the accumulation of dust on the surface of solar cells outside would decrease light flux into the active layer and thus working efficiency of solar cells. To address this problem, the superhydrophobic surface with water contact angles (CAs) of larger than 150° and sliding angles (SAs) of less than 10°can be introduced to the solar cells, and the dust accumulation can be avoided via self-cleaning process.

In 2018, our group fabricated a superhydrophobic, highly transparent, and hazy nanopaper composed of TEMPO-oxidized CNFs and 3D nanostructured polysilox- anes via a facile process: vacuum filtration and in situ siloxanes growth. This nanopaper can not only lead to a significant enhancement in short circuit density and conversion efficiency of a solar cell by light management but also recover most of the photovoltaic performance losses due to dust accumulation by a self-cleaning process.

Preparation of the Cellulose Nanopaper

As mentioned before, we obtained three types of nanocellulose, i.e., CNCs, CNFs, and 2,2,6,6-tetramethylpiperidine-l-loxy oxidized CNFs (TOCNFs), by chemical processes and/or mechanical treatment. Herein, the CNC, CNF, and TOCNF dispersions (0.1 wt %, 50 g) after stirring for 1 h at 800 rpm were vacuum filtrated with a glass suction filtration kit using a mixed cellulose ester membrane filter (pore size: 0.22 pm). Then, the wet sheets were carefully placed between filter papers and dried under pressure (0.01 MPa) at 105 °C for 10 min. Three types of cellulose nanopapers with ~40 pm thickness were fabricated and coded as CNC-P, CNF-P, and TOCNF-P, respectively.

To construct the superhydrophobic surface on the cellulose nanopaper, the nanopaper was further modified by the methyltrichlorosilane. In detail, the nanopapers were immersed into the 0.5 M methyltrichlorosilane solution of anhydrous toluene at room temperature for different time, as reported previously (Khoo and Tseng 2008, Gao and McCarthy 2006). Vessels were closed to the air but exposed to the chamber environment during the solution and sample introductions. After reaction, the samples were then removed and rinsed with various solvents in the follow'ing order: toluene, ethanol, ethanol/deionized water (1:1), and DI water. Finally, the samples were dried in an oven at 55°C for 4 h. As an example, the obtained hydrophobic TOCNF-Ps after different modification time (10, 20, 30, 40 min) were coded as HIO-TOCNF-P, H20-TOCNF-P, H30-TOCNF-P, and H40-TOCNF-P, respectively.

Characterization of the Cellulose Nanopaper

The surface morphology and structure of different nanopapers play a critical role in their properties. Herein, as can be seen in Figure 3.8, we compared the morphology of regular paper and three types of nanopapers that are made of different-sized cellulose nanofibres. The regular paper built up by the original cellulose fibres had a significantly rough surface and porous structure. These cavities would cause light scattering, thereby limiting the optical transparency of the regular cellulose paper. The various cellulose nanopapers illustrated different surface morphologies. The CNC-based nanopaper (CNC-P) had a significantly flattening and smooth surface; the CNF-P had a rough and uneven surface. For the TOCNF-P. its surface roughness was in between the CNC-P and CNF-P. While, all the cellulose nanopapers exhibited the dense structure rather than the porous structure in the regular paper. The surface morphology of the H40-TOCNF-P with different magnifications was also demonstrated. After methyltrichlorosilane (MTCS) modification for 40 min, the fibrous siloxanes network was constructed on the TOCNF-P surface. The pearl-necklace- like fibres were composed of interconnected spherical siloxanes particles, and the obtained siloxanes fibres w'ere irregularly bent and randomly cross-linked with each other. However, the siloxanes were homogeneously distributed on the H40-TOCNF-P surface, as illustrated by the elemental mapping of Si.

A possible mechanism of polysiloxanes formation was proposed based on the literature. The trace water in the toluene solution led to the hydrolysis of MTCS. then the obtained silanols would react with the isolated hydroxyl groups on the surface of

FIGURE 3.8 (a—d) The surface morphology of regular paper and nanopapers before modification under SEM. (e-h) The surface morphology of H40-TOCNF-P under SEM with different magnifications (left three) and elemental mapping of Si by energy dispersive X-ray spectroscopy (EDS) (rightmost).

TOCNF-P or other silanols. During this reaction, the Si-O-Si linkages were established on the substrates. Because the silanols with both hydrophilic groups (-OH) and hydrophobic groups (-CH,) have the propensity to self-assemble. the nanospheres or nanofibres of siloxanes continued to grow in three dimensions and react with excessive Si-OH groups. Therefore, the 3D fibrous siloxanes network that consisted of rough pearl-necklace-like fibres was formed on the nanopaper surface. Besides, the successful silanization of the nanopaper was further confirmed by the Fourier transform infrared (FTIR) spectra. Compared with all the unmodified nanopapers, two new peaks were observed in the H40-TOCNF-P sample: the absorption bands at 1273 and 781 cm^-1, which are assigned to the asymmetric stretching vibrations of Si-CH, and the characteristic vibrations of Si-O-Si. respectively (Duan et al. 2014). These groups and linkages were consistent with the predicted chemical structures of polysiloxanes.

The optical properties of nanopapers are critical for substrates towards widespread applications. Both the total transmittance and transmittance haze of nanopapers were measured by the UV-Vis spectrometer and are shown in Figure 3.9. The CNC-P presented the highest transmittance of 91.3% but the lowest haze of only 20.1% at 550 nm. This may be due to the rod-like CNCs with small length constructed densely, allowing more light to propagate through and suppressing light scatter behaviour. Conversely, CNF-P showed the lowest transmittance (69.7% at 550 nm) but the highest haze (61.4% at 550 nm), which may be due to their rough surface caused by the large dimensions of CNFs. Interestingly, TOCNF-P not only had a high optical transmittance (90.4% at a wavelength of 550 nm), which was close to that of CNC-P, but it also exhibited a higher transmission haze (49.3% at 550 nm) than that of CNC-P. A possible explanation for this could be that each individual TOCNF led to small forward scattering rather than significant back scattering due to its nanoscale diameter and appropriate length, and thus the obtained densely laminated TOCNF-P allowed most of light to propagate through and retained an appropriate light-scattering effect, as reported in the literature (Hu et al. 2013). Furthermore, the obtained H40-TOCNF-P after modification still retained a high transmittance (90.2% at 550 nm) and high haze (46.5% at 550 nm).

FIGURE 3.9 (a) Total transmittance and (b) transmittance haze of unmodified nanopapers and H40-TOCNF-P.

The mechanical properties like strength and toughness of nanopapers are critical for their wide range applications. Both the regular paper and CNC-P exhibited a very low strength of around 49 MPa. This may be attributed to that the native wood fibres with large dimensions in regular paper cannot bind tightly and densely, and the rod-like and rigid CNCs have limited contact area between fibres in the CNC-P. However, the CNF-P, TOCNF-P, and H40-TOCNF-P were much stronger (with the tensile strength of 92.8-103.7 MPa) and much tougher (with the toughness of 1.12-2.45 J M"³) than the regular paper and CNC-P. This improvement in mechanical properties may be related to the enhanced cohesion, such as dispersion force and hydrogen bonds, and increased contact area between CNFs orTOCNFs (Nishiyama 2018). Interestingly, H40-TOCNF-P exhibited the highest toughness of 2.45 J M~³, which was probably due to the formation of 3D polysiloxanes networks that increased the ability of H40-TOCNF-P to absorb energy or sustain deformation without breaking.

The wettability of regular paper and different cellulose nanopapers was characterized by the static CAs, which is illustrated in Figure 3.10. Due to outstanding hydro- philicity of original wood pulp and lots of cavities, the regular paper exhibited a completely hydrophilic property, as indicated by the zero CA and the wetted paper by water droplet. However, the TOCNF-P with CA of 93.5° presented hydrophobic- ity; this may be due to its dense structure and tight binding of TOCNFs inside. As expected, after modification by MTCS for 40 min, the obtained H40-TOCNF-P had

FIGURE 3.10 (a) Water contact angle of regular paper. TOCNF-P, and H40-TOCNF-P. (b) Photos of regular paper. TOCNF-P, and H40- TOCNF-P with the water droplet on the surface.

a significant water-repellency, as demonstrated by the sphere-like water droplet that settled on its surface. The H40-TOCNF-P has achieved the superhydrophilicity with the CA of 159.6°. Furthermore, the slide angle (SA) for H40-TOCNF-P is only 5.8°, indicating that the adhesion force between water droplet and nanopaper surface is ultralow. Therefore, the as-prepared superhydrophobic nanopaper was endowed with the self-cleaning function. The dust particles on the surface of H40-TOCNF-P could be easily collected and taken away by rolling water. This important characteristic reveals the potential applications of H40-TOCNF-P in many interdisciplinary technological fields.

Enhancement of Solar Cell Efficiency

The as-prepared superhydrophobic, transparent, and hazy cellulose nanopaper with both light-management and self-cleaning functions was used to improve the efficiency of solar cells. The photovoltaic performances of the bare cell and the cells coated with various nanopapers were evaluated by open circuit voltage (V_oc), short circuit density (J_sc), fill factor (FF). and the overall conversion efficiency. As shown in Table 3.2, both the TOCNF-P and H40-TOCNF-P significantly enhanced the conversion efficiency of the solar cell. This was mainly due to the optical properties, i.e., high transmittance and haze of nanopapers. High transmittance allow'ed most light to propagate through: high haze made the light diffusive, which increased the travelling path length and thus time of photos in the solar cell’s active layer.

However, after contaminated with dust, the solar cell coated with nanopaper exhibited obvious decrease in photovoltaic performances, which was mainly contributed to the dust particles which blocked the incident light significantly. After selfcleaning process, the conversion efficiency of solar cell was dramatically increased. Compared to that of solar cell w'ith native H40-TOCNF-P, the Jsc and the corresponding efficiency had recovered 94.80% and 96.76%, respectively. This photovoltaic performance was still slightly better than that of the bare solar cell.

TABLE 3.2

Photovoltaic Characteristics of the Bare Solar Cell and the Solar Cell with Different Nanopapers