Perovskite solar cells (PSCs) have attracted a great deal of attention due to their rapid development of photo-to-electric power conversion efficiency (PCE). The efficiency has increased from 3.8% in 2009^[1] to the latest 25.2%,^[2] which is appealing for the potential commercialization. However, it has been seriously hampered by two problems, price and stability. Usually these PSCs use precious metals (such as Au, Ag, and Cu^[3–5]), and expensive hole-transporting materials (HTMs, such as PTAA^[3] and spiro-OMeTAD^[6]). This not only increases the manufacturing costs, but also affects device stability. Replacing metal electrodes and HTMs by carbon-electrode could help to solve these problems, since the latter is cheap in price, inert to electrochemical corrosion, highly conductive and holding suitable work function (∼ 5.0 eV). These merits trigger the so-called carbon-electrode-based PSCs (CPSCs, and often “hole-conductor-free” is titled). Since the pioneered mesoscopic CPSCs proposed by Han’s group,^[7,8] until now several structures have been proposed, for example, embedment,^[9,10] quasi-planar,^[11,12] and planar CPSCs.^[13] These devices have shown excellent stability. However, the power conversion efficiencies (PCEs) of these CPSCs are right now relatively lower than those of devices based on metal-electrodes, thus it is essential to be upgraded.^[14] To improve the device performance, healing the interface between perovskite and carbon electrode was observed to be helpful due to the accelerated charge transfer.^[13,15,16] On the other hand, bulk engineering also works. In the studies of metal-electrode-based PSCs, improving crystallinity of the active layer or the lead halide perovskite (PVSK for short) is found to be the key to obtain higher efficiencies. Several strategies have been proposed. For example, in 2014, Huang and co-workers found that, solvent annealing could improve the crystallinity of MAPbI₃ (MA = CH₃NH₃) and thus the device performance.^[17] You et al. observed that moisture-assisted annealing was also useful to the crystallization of the MA-based perovskite.^[18] For FA [CH(NH₂)₂]-based perovskite, several studies showed that importing small ions such as MA⁺, Rb⁺, Cs⁺, K⁺, Cl⁻, Br⁻ could facilitate the nucleation and growth of the desirable α-FAPbI₃ (black phase).^[19–23] Impressively, Han et al. in 2014 reported that solvent additive of dimethyl sulfoxide (DMSO) could retard the crystallinity of the PbI₂, which then led to improved crystallization of MAPbI₃ in sequential formation routine.^[24] Chen et al. observed that intermediate phase could be formed between DMSO and PbI₂, which favored the formation of PVSK crystallites.^[25,26] On the other hand, porous structured PbI₂ films could improve the crystallinity of PVSK films,^[27–29] while adding polymers or nanomaterials into the perovskite precursor could also favor the crystallizing process of photo-active materials.^[30–35]

As an alternative, in this work, SiO₂ nanoparticles (SiO₂ NPs thereafter) were mixed with PbI₂ solution, aiming at favoring the crystallization process of PVSK (FA_0.75MA_0.25PbI_2.81Cl_0.19) and thus improve performance of carbon-electrode-based devices. As will be shown later, these nanoparticles could retard the crystallinity of PbI₂ films, through which crystallinity of resultant PVSK films can be upgraded. The improved crystallinity then upgrades the device efficiency of the low-temperature planar CPSCs from < 15%^[13] to 15.28% (reverse scan, AM1.5G, 100 mW/cm², corresponding to 14.46% for the maximum power point tracking).

Fabrication process of the low-temperature planar CPSCs is shown in Fig. 1(a), following the protocol described in the previous work.^[13] A sandwiched structure of “FTO/SnO₂/PVSK/C” is adopted (as shown by the typical cross-sectional SEM image in Fig. S1). To make it compatible with the low temperature procedures, treating temperature of the SnO₂ layer was reduced from the previously adopted 180 °C to the currently used 150 °C.^[13] To tune the crystallization of both PbI₂ and afterwards PVSK films, trace amount of SiO₂ NPs was imported to the system by mixing SiO₂ precursor (1 g/10 mL) into PbI₂ solution with varied volume ratios, or 0, 0.1%, 1%, 10%, etc. To help discussion, the volume ratio between SiO₂ NPs and PbI₂ is also estimated, for example, it is about 1/2570 when volume ratio of SiO₂ precursor is 0.1%. As a result, expansion effect of the SiO₂ NPs could be neglected. For comparison, x-ray diffraction study was performed on these SiO₂ NPs. As is expected, these nanoparticles are mainly amorphous when considering the low intensity. Transmission electron microscope (TEM) image in Fig. 1(c) shows that the SiO₂ NPs are well-dispersed, with particle sizes of about 10 nm.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) Schematic figures for the preparation of low temperature, planar CPSCs with addition of SiO2 NPs in PbI2 solution. (b) XRD pattern of the SiO2 NPs. (c) TEM image of the SiO2 NPs.

Effect of SiO₂ NPs on the crystallization behavior of PbI₂ and the afterwards-formed PVSK films were studied. XRD study showed that adding SiO₂ NPs could retard the crystallization of PbI₂. As seen in Fig. 2(c), with volume ratio of the SiO₂ precursor, the FWHM of the peak of PbI₂ (001) increases; while for the PVSK films, the FWHM decreases. To show it more clearly, average crystallite size is calculated by Scherrer’s formula using the FWHM peaked. As shown in Fig. 2(d), opposite trends are illustrated. With amount of SiO₂ NPs increases, PbI₂ crystallites tend to be smaller, while PVSK crystallites become larger. The improved crystallization of PVSK could also be reflected from morphology studies. As shown by the scanning electron microscopy (SEM) images, with increase of the volume ratio of SiO₂ NPs, larger PVSK grains could be harvested. For example, average grain size is around 500 nm before SiO₂ incorporation, but increases to be around 700 nm after adding 1% (volume ratio) SiO₂ precursor into PbI₂ solution. The SiO₂ NPs distributed uniformly in both of the PbI₂ and PVSK films, as being reflected by EDS mapping (Figs. S2 and S3). The above studies tell us that SiO₂ NPs have retarded the crystallization of PbI₂, but favored the crystallization of PVSK.

Fig. 2.

Figure Option ViewDownloadNew Window

Fig. 2. XRD patterns of (a) PbI2 and (b) PVSK films with addition of SiO2 NPs (volume ratios of the SiO2 precursor were marked). (c) Enlarged diffraction peaks of PbI2 (001) and PVSK (110) [dashed lines show the full width at half maximum (FWHM) of each peak]. (d) Effect of SiO2 NPs on the average crystallite size of PbI2 and PVSK. (e)–(h) Typical scanning electron microscopy images of PVSK films with addition of SiO2 NPs. Distribution histograms of grain sizes are shown in inset of the images.

The improved crystallinity of the PVSK could also be reflected by absorption behavior and the lifetime measurement. UV-vis absorption spectra in Fig. 3(a) show that, after incorporation of SiO₂ NPs, extinction of PVSK is improved, while time-resolved photoluminescence (TRPL) in Fig. 3(b) shows that lifetime of the photo-generated charge-carriers is increased after incorporation of SiO₂ NPs. The prolonged lifetime would be helpful for the power conversion of the perovskite solar cells.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (a) UV-vis absorption spectra, and (b) time-resolved photoluminance spectra (TRPL) of PVSK films with incorporation of SiO2 NPs.

From the above studies one could see that retarding the crystallization of PbI₂ is beneficial to obtain higher quality PVSK films. The regulated crystallite growth of PVSK could be well understood by considering the interaction between SiO₂ and PbI₂. X-ray photoemission spectroscopy test in Fig. 4(a) reflects that, after incorporation of SiO₂ NPs, binding energies of both Pb 4f and I 3d shifted, indicating possible interaction between them. For better illustration about the regulation brought by SiO₂ NPs, schematics are drawn in Fig. 4(b). Due to the disruption brought by SiO₂ NPs, crystallization process of PbI₂ is affected. Thus the PbI₂ film tends to be more amorphous, which is beneficial to the crystallization process of the afterwards-growing PVSK film. Since the amorphous state of PbI₂ makes it easier for organic ions (like FA⁺, MA⁺, I⁻, etc.) to diffuse and react with PbI₂.^[27,29,36] It is worth noting that the SiO₂ NPs had not obviously modified the film morphology of PbI₂. This is different from the result revealed by Yi et al., who observed that porous structure could be brought by evaporating DMSO in fresh PbI₂ films.^[28] Here it is shown that the amorphous SiO₂ NPs could retard the crystallization of PbI₂, which is similar to that brought by DMSO as proposed by Han et al.^[24] Thus one more strategy is provided to regulate the crystallization process of the PVSK films.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (a) X-ray photoemission spectra of PbI2 films with or without incorporation of SiO2 NPs. (b) Schematic of the effect of SiO2 NPs on regulated crystallization of PVSK film.

It was observed that adding small amount of SiO₂ NPs into PbI₂ solution (and then the PVSK films) could upgrade device efficiency of the low-temperature planar CPSCs. As reflected by the typical current density–voltage (J–V) curves shown in Fig. 5(a), the corresponding performance parameters including open-circuit voltage (V_OC), short-circuit current density (J_SC), fill factor (FF) and power conversion efficiency (PCE) are also listed in Table 1. It was observed that addition of SiO₂ NPs upgrades J_SC, V_OC and FF in all. For better illustration, these three performance parameters, as well as the {PCE}, are all collected and plotted against the volume ratio of SiO₂ precursor in Fig. 5(b). About 20 separated devices were used. Compared with the control devices (w/o SiO₂ NPs), all of the four parameters have been upgraded, especially for the case of moderate ratio of 1%. Concretely, FF increases from 52.92(±9.64)% to 61.17(±8.37)%, J_SC increases from 23.05(±0.55) mA/cm² to 23.51(±0.63) mA/cm², V_OC increases from 0.949(± 0.019) V to 0.988(±0.014) V, these help to upgrade {PCE} from 11.44(±1.83)% to 14.01(± 2.14)%, with the optimized at 15.28% (Fig. S4).

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (a) Typical current density–voltage (J–V) curves of as-fabricated devices recorded from reverse scan under simulated illumination (AM 1.5G, 100 mW/cm2, testing area is 0.0514 cm2). (b) Effect of volume ratio of SiO2 precursor on the performance parameters (reverse scan) and hysteresis index (HI).

To verify the improvement in photocurrent, external quantum efficiency (EQE) tests were performed. As shown in Fig. S5, the similar trend is observed in the integrated current density. Hysteresis behavior is also examined by collecting the PCEs from both reverse/forward scans. Hysteresis index (HI) factors are calculated using the method described before.^[6] As shown in bottom of Fig. 5(b), incorporation of SiO₂ NPs is also favorable to reach lower hysteresis. HI decreases slightly from 13.69(± 3.29)% to 12.72(± 2.24)% (ratio of 1%).

Table 1.

Table 1.

Table 1. Performance parameters picked from the J–V curves in Fig. 5(a). . SiO2 precursor/vol% VOC/V JSC/mA⋅cm−2 FF/% PCE/% 0 0.954 22.91 54.77 11.97 0.1 0.968 23.18 57.18 12.83 1 0.987 23.83 61.94 14.57 10 0.977 23.22 61.53 13.96

Table 1. Performance parameters picked from the J–V curves in Fig. 5(a). .

Charge transport and recombination dynamics of the low-temperature planar CPSCs were explored with respect to the addition of SiO₂ NPs, based on transient photocurrent/voltage (TPC/TPV) decay curves tested. Typical TPC/TPV curves are shown in Fig. S6. For comparison, extraction time (t_d) and lifetime (τ) of photo-generated carriers were picked from these transient decay curves, which are shown in Figs. 6(a) and 6(b), respectively. Similar to those performed in device performance parameters, statistics were also performed. It could be clearly seen that importing SiO₂ NPs could accelerate charge transport, while retard recombination processes of the photo-generated charge carriers. Take charge extraction time t_d as an example, it starts from 12.84(±1.71) μs of the control devices (without addition of SiO₂ NPs), and decreases to 11.55(±1.10) μs, 6.75(± 1.49) μs after adding SiO₂ NPs at volume ratios of 0.1% and 1%. As for life time (τ), it begins at 10.68(± 3.00) μs of the control devices, and prolongs to 15.13(± 4.55) μs, 56.24(± 11.9) μs for volume ratios of 0.1% and 1%, respectively. The case of 1% comes out with smallest extraction time (t_d) and longest life time (τ). Comparing Figs. 6(a) and 6(b) to the performance parameters in Fig. 6(c) could find that the evolution trend of t_d is opposite to J_SC, while the trend of lifetime τ is in parallel with V_OC. Such a phenomenon is quite similar to the previous studies.^[21] Thus accelerating charge extraction is beneficial to the output of photocurrent, while prolonged lifetime is favorable to output of V_OC.^[37] The prolonged lifetime is due to the reduced recombination. This could also be reflected by dark current test. As shown in Fig. 6(c), at given bias, devices with incorporation of SiO₂ NPs produce lower dark current. Additionally the case of volume ratio 1% comes out with lowest dark current.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. Effect of volume ratio of SiO2 precursor on (a) charge extraction time [td, extracted from transient photocurrent (TPC) decay curves], (b) lifetime [τ, extracted from transient photocurrent (TPV) decay curves], and (c) dark current density–voltage curves of the low-temperature planar CPSCs.

Storage stability and maximum-power-point tracking (MPPT) were tested at ambient air and in dark, with RH of about 45%, and no encapsulation was used.^[13] As shown in Fig. 7(a), besides the slight increase in the first 30 days, the device efficiency changed slowly in the afterwards 100 days, showing sound storage stability of the low-temperature planar CPSCs. As for the increase in the starting 30 days, it may be ascribed to the slow-strengthening contact between the PVSK film and the top carbon electrode.^[13] Careful examination in Fig. 7(b) could observe that devices assembled with SiO₂ NPs-regulated PVSK films come out with better stability. Photo-stability was also measured. Similar to the storage stability, the device assembled from the SiO₂ NP-regulated PVSK film comes out with better stability. The improved stability is ascribed to the upgraded crystallization of the PVSK films. As stated above, with incorporation of moderate amount of SiO₂ NPs, larger crystallites and also larger grains were obtained, which reduced the defects in the PVSK films, and thus favored the stability of the device system. As a result, higher crystallinity of the PVSK film is favorable to device stability.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. Stability test of the low-temperature, carbon-electrode-based PSCs. (a) Storage stability and (b) photo-stability or the test of so-called maximum-power-point tracking (MPPT). No encapsulation was used, and for storage-test, the devices were stored at dark with relative humidity of 45%(± 10%).

Finally, it comes to show that SiO₂ NPs could retard the crystallization process of PbI₂ films, making them more amorphous and also easier for the organic ions to diffuse in reaction with PbI₂. Such a strategy is verified to be effective in achieving better crystallinity of PVSK, and helpful in accelerating charge transport and retarding recombination process, and then upgrading the device performance of the low-temperature-processed carbon-electrode-based planar perovskite solar cells. This work provides an effective method in regulating the crystallization of the PVSK.