Cesium carbonate as a surface modification material for organic–inorganic hybrid perovskite solar cells with enhanced performance

Abstract

Cs₂CO₃ has been employed as a new surface modification material for inorganic–organic hybrid perovskite solar cells. With the optimized modifying process, 14.2% power conversion efficiency (PCE) was obtained, enhanced by nearly 20% compared with the control devices. Further studies showed that the PCE improvement mainly came from the retarded back recombination.

---

Though they have been known for several decades, organic–inorganic hybrid perovskite materials CH₃NH₃PbX₃ (X = Cl Br I) have attracted great attention recently due to their huge success in photovoltage devices.^1–3 Since Miyasaka et al. first reported 3.81% dye-sensitized solar cells with CH₃NH₃PbI₃ as light absorber in 2009, the conversion efficiency of solar cells based on hybrid perovskites has been surging.⁴ The latest certified record published by NREL was 17.9%, got by Sang II Seok from KRICT. With the conversion efficiency improving fast, perovskite solar cells' structure has also been evolved. Previous studies were all based on the structure of dye-sensitized solar cells, with a more or less 10 μm thick mesoporous TiO₂ layer and I⁻/I₃⁻ liquid electrolyte.⁵ The instability of the devices due to the dissolution of the perovskites in the liquid electrolyte was a big challenge. By now, this problem was solved through replacing the liquid electrolyte with the solid hole transport materials (such as spiro-OMeTAD), and the thickness of mesoporous layer was also reduced to hundreds of nanometers, or even without the mesoporous layer.^6–10

Similar to dye-sensitized solar cells and organic solar cells, trap states at each layer's surface were easy to form charge-recombination centers.¹¹ Considering that perovskite layer itself had little deep traps, the situation of mesoporous TiO₂ surface became a key point for high conversion efficiency.¹² Surface modification with other materials was a useful method to passivate the TiO₂ surface traps and reduce the current loss resulting from the charge recombination.¹³ Surface modification materials such as TiO₂, Al₂O₃, Y₂O₃, C₆₀, graphene quantum dots and Sb₂S₃ had been reported.^14–19 Michael Gratzel et al. studied using sub-nanometer conformal TiO₂ layer by atomic layer deposition as the blocking layer and 11.5% PCE was got.¹⁴ Shuzi Hayase et al. used Al₂O₃ or Y₂O₃ to passivate the mesoporous TiO₂ layer and after passivation with Y₂O₃, the efficiency increased from 6.59% to 7.5%.¹⁵

Here we reported Cs₂CO₃ as a new surface modification material for perovskite solar cells. Cs₂CO₃ was widely used in organic light-emitting devices (OLEDs) to improve electron-injection from the metal cathod.²⁰ Our results showed that it could also act well in perovskite solar cells. Cs₂CO₃ had a work function of 2.2 eV, much lower than TiO₂.²¹ A thin Cs₂CO₃ layer between the mesoporous-TiO₂/CH₃NH₃PbI₃ interface would not affect the crystal of perovskite layer, but could reduce the charge recombination significantly. As a result, both the short-circuit current and fill factor of the photovoltaic devices improved, and the power conversion efficiency increased from 11.9% to 14.2%, giving a nearly 20% enhancement.

Perovskite solar cells were fabricated in the configuration mesoporous-TiO₂/perovskite/spiro-OMeTAD/Au. Detailed fabrication process could be found elsewhere.²² For Cs₂CO₃ surface modification, 100 μL Cs₂CO₃/DI water solution with certain concentration was dropped onto the sintered mesoporous-TiO₂ film (2.5 cm × 2.5 cm), soaking for 30 s, and then spinning at 2500 rpm for 60 s. The films were then sintered at 300 °C for 30 min before spinning the PbI₂ layer. Control devices were made under the same condition expect without the spinning of Cs₂CO₃ solution. We had tried to characterize the Cs₂CO₃ layer with TEM, but it seemed that it was too thin to be found. So we moved to XPS. As shown in Fig. 1, XPS results clearly showed the existence of Cs. Considering that the decomposition temperature of Cs₂CO₃ was higher than 600 °C, we thought this could indicate that the Cs₂CO₃ layer was introduced into the mesoporous TiO₂. The atomic ratio of films without and with Cs₂CO₃ surface modification tested by XPS could be found in Table S1 in the ESI.† To identify weather Cs₂CO₃ treatment had intermixing or doping effect to mp-TiO₂, we washed the modified mp-TiO₂ film softly with DI water for three times, and did XPS test again. The result could be found in Fig. S1.† Peaks from Cs_3d disappeared totally, showing that Cs₂CO₃ was washed out. That indicated that Cs₂CO₃ left in the film formed only physical contact with mp-TiO₂ and should not affect the physical properties of mp-TiO₂. UV-Vis absorption curves of the TiO₂ films without and with Cs₂CO₃ modified were shown in Fig. 2. As expected, the absorption onset didn't change, indicating the bandgap of the mp-TiO₂ hadn't changed after the Cs₂CO₃ modification.

Fig. 1 XPS results of films without and with Cs₂CO₃ surface modification.

Fig. 2 UV-Vis absorption of TiO₂ films without and with Cs₂CO₃ modified.

In order to get the highest conversion efficiency, the thickness of the modification layer should be optimized. It could be easily controlled by adjusting the concentration of the solution used. Photovoltaic performance of devices modified with different Cs₂CO₃ concentration was shown in Table 1. Fig. S2† illustrated the J–V curves of devices without and with Cs₂CO₃ modified using the optimized concentration (30 mmol L⁻¹) under AM 1.5, 100 mW cm⁻² solar irradiance. The results were got under reverse scan with a 50 mV s⁻¹ scanning speed, from 1.05 V to 0 V. J–V curves tested under different scanning directions were shown in Fig. S3.† There was a little difference for the forward and reverse scan, as noticed by others.²³ The reason for that still was not clear. We could see that, even with quite low concentration, Cs₂CO₃ modification gave a significant enhancement on the power conversion efficiency (PCE). J_sc, V_oc and FF were three key parameters that decided the final PCE. It was quite clear that J_sc increased regularly with increasing the Cs₂CO₃ concentration and started to decrease after reaching the optimized concentration (30 mmol L⁻¹). This was not hard to understand because Cs₂CO₃ was a kind of insulating material; too thick Cs₂CO₃ layer would block the electrons' transport from perovskite into the mp-TiO₂. But when it was thin enough, electrons could pass through that with the tunnelling effect. This would be explained more detailed in the later discussion. V_oc also increased by 60 mV from 0.97 V to 1.03 V after Cs₂CO₃ modification. This improvement of V_oc was always found in the surface modification work of dye-sensitized solar cells and it was always attributed to the better passivated TiO₂ surface traps.²⁴ Different batches of devices were fabricated to make the results more convincing. Table S2(a) and (b)† gave the parameters of one batch of devices.

Table 1 Photovoltaic parameters of perovskite solar cells modified with Cs₂CO₃ solution of different concentration

CCs2CO3/mmol L^−1	Voc/V	Jsc/mA cm^−2	FF	PCE/%
0	0.97	19.8	0.62	11.9
5	0.95	20.3	0.68	13.2
15	0.96	20.7	0.68	13.5
30	1.03	21.3	0.65	14.2
45	1.02	19.1	0.68	13.3

---

To fully investigate possible reasons for the improved cell performance, we firstly checked weather the Cs₂CO₃ modification would affect the crystallization process of perovskite layer. Fig. S4(a) and (b)† showed the SEM results of perovskite layers without and with surface modification. In both cases, a compact perovskite capping layer was formed on the mesoporous TiO₂. No clear difference on the shape of the crystals was found, indicating that Cs₂CO₃ could not act with or get into the perovskite layer to affect its morphology. XRD was also done and the results were shown in Fig. 3(a). In both cases, sharp diffraction peaks could be found at 14.1°, 24.5°, 28.4° and 31.8°, which were assigned as the (100), (111), (200), (210) lattice planes of the perovskite crystal, respectively. There were some changes on the relative peak intensities, showing that the crystallization of the perovskite changed after the Cs₂CO₃ surface modification. This might also affect the final performance of the devices and needed to be further researched. Peaks belonging to PbI₂ were also found in both films, indicating the incomplete conversion to perovskite. This was also noticed by others.²⁵ Considering the high PCE we got, the residual PbI₂ seemed to be not a trouble. Yang Yang et al. recently even reported the formation of PbI₂ in the perovskite boundaries and at the relevant interface as the hole blocking layer for improved cells performance.²⁶ As expected, no peaks belonging to Cs₂CO₃ were found, consistent with the fact that Cs₂CO₃ layer for the surface modification was very thin. Fig. 3(b) was the UV-vis absorption curves of the films without and with surface modification. After Cs₂CO₃ modification, the absorption of the perovskite film decreased slightly. This was because that after Cs₂CO₃ modification, Cs₂CO₃ filled into the pores in the mp-TiO₂, reducing the space left in the mp-TiO₂ film. As a result the total amount of perovskite was also reduced, leading to the decrease of absorption. But we could also notice that the absorption of the film was quite strong. Considering that we still had Au back contact that could act as a mirror to reflect the incident light, the perovskite film should be able to absorb nearly all the incident light. So this little decrease on the absorption would not affect largely on the final J_sc. Ultraviolet photoelectron spectroscopies (UPS) of the TiO₂/perovskite and TiO₂/Cs₂CO₃/perovskite were shown in the Fig. S5.† No peak position change was noticed, indicating that Cs₂CO₃ modification also hadn't led to the shift of the Fermi level and valence band position of the perovskite.

Fig. 3 (a) XRD results and (b) UV-vis absorption results of the films without and with Cs₂CO₃ modification.

Now that the surface modification had not affect a lot on the morphology, crystallization and absorption of the perovskite layer, the performance improvement should be contributed to the internal physical changes. Fig. 4(a) showed the J–V curves of devices without and with Cs₂CO₃ modification measured in dark. Compared with the control device, the onset of the dark current of surface modified device moved nearly 50 mV to higher voltage. This indicated that the back flow of electrons from TiO₂ to perovskite layer or the hole conductor was significantly suppressed. Considering that the electron recombination mainly happened due to the surface traps of the mesoporous TiO₂, the retarded dark current also showed that these surface traps were well passivated, which was in good agreement with the higher V_oc for surface modified devices.

Fig. 4 (a) J–V curves of devices without and with Cs₂CO₃ surface modification tested in dark; (b) EIS cycles of devices tested in dark.

Electrochemical impedance spectroscopy (EIS) was a simple but useful method to analyze the electron transport and interfacial properties inside the solar cells.²⁷ Fig. 4(b) showed the EIS results of devices measured in dark. Open-circuit voltage of the device was used as the bias for the testing. Due to the complexity arising from the mesoporous TiO₂ layer, only the middle frequency semicircles were found in the Nyquist plots, which presented the charge transfer resistance (R_ct) related to the recombination of electrons in TiO₂ with perovskite or hole transport layer. The larger the R_ct was, the less the electrons recombined at the surface. Fig. 4(b) clearly showed that, after surface modification, the R_ct increased largely, which meant that the charge recombination was retarded, consistent with the conclusion from the dark J–V curves. This also explained the enhancement of the V_oc. Similar to the conventional solid-state DSCs, Gratzel et al. also noticed that in perovskite solar cells, the recombination rate increased when the quasi Fermi level of electrons in TiO₂ moved toward its conduction band.¹⁴ Inversely, the retarded charge recombination meant the quasi Fermi level of n-type TiO₂ shifted up, leading to the increase of the device's V_oc. Fig. 5 showed the incident photon to current efficiency (IPCE) curves for devices without and with Cs₂CO₃ surface modification. After the surface modification, IPCE increased in a large range, being consistent with the result the J_sc increased after the surface modification. In the range of 650–800 nm, IPCE was nearly the same for both cases. We attributed this to the composite of the increased electron transport efficiency but decreased light harvesting efficiency.

Fig. 5 IPCE curves of devices without and with Cs₂CO₃ surface modification.

Combining with the above discussion, we gave a simple schematic diagram of the working principle of Cs₂CO₃ surface modification in Fig. 6. After surface modification, a ultrathin insulating layer was introduced into the mp-TiO₂/CH₃NH₃PbI₃ interface. This ultrathin layer would not affect the forward electron transport, but could prevent the electrons in mp-TiO₂ from recombining with holes left in perovskite layer or in the hole transport layer (HTL). So that little current would loss during the electron transport process, giving an enhanced J_sc.

Fig. 6 Schematic diagram of the working principle of Cs₂CO₃ surface modification.

In summary, we employed Cs₂CO₃ as the surface modification material for perovskite solar cells. Both the J_sc and V_oc enhanced after the surface modification. Further research showed that, the surface modification would not affect a lot on the crystal of the perovskite layer, but mainly contribute to the improved electron transport and retarded back recombination in the solar cells. With the optimized concentration, 14.2% conversion efficiency was finally got, increasing nearly 20% compared with the control devices.

Acknowledgements

This work was supported by the national natural science foundation of China under grant No. 51273104.

Notes and references

1. R. F. Service, Science, 2014, 344, 458.
2. S. Kazim, M. K. Nazeeruddin, M. Gratzel and S. Ahmad, Angew. Chem., Int. Ed., 2014, 53, 2812.
3. T. C. Sum and N. Mathews, Energy Environ. Sci., 2014, 7, 2518.
4. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050.
5. J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park and N.-G. Park, Nanoscale, 2011, 3, 4088.
6. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Gratzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
7. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643.
8. J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gratzel, Nature, 2013, 499, 316.
9. M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395.
10. Q. Chen, H. Zhou, Z. Hong, S. Luo, H.-S. Duan, H.-H. Wang, Y. Liu, G. Li and Y. Yang, J. Am. Chem. Soc., 2014, 136, 622.
11. T. Leijtens, B. Lauber, G. E. Eperon, S. D. Stranks and H. J. Snaith, J. Phys. Chem. Lett., 2014, 5, 1096.
12. W.-J. Yin, T. Shi and Y. Yan, Appl. Phys. Lett., 2014, 104, 063903.
13. A. K. Chandiran, M. K. Nazeeruddin and M. Gratzel, Adv. Funct. Mater., 2014, 24, 1615.
14. A. K. Chandiran, A. Yella, M. T. Mayer, P. Gao, M. K. Nazeeruddin and M. Gratzel, Adv. Mater., 2014, 26, 4309.
15. Y. Ogomi, K. Kukihara, S. Qing, T. Toyoda, K. Yoshino, S. Pandey, H. Momose and S. Hayase, ChemPhysChem, 2014, 15, 1062.
16. W. Li, J. Li, L. Wang, G. Niu, R. Gao and Y. Qiu, J. Mater. Chem. A, 2013, 1, 11735.
17. A. Abrusci, S. D. Stranks, P. Docampo, H.-L. Yip, A. K. Jen and H. J. Snaith, Nano Lett., 2013, 13, 3124.
18. Z. Zhu, J. Ma, Z. Wang, C. Mu, Z. Fan, L. Du, Y. Bai, L. Fan, H. Yan, D. L. Phillips and S. Yang, J. Am. Chem. Soc., 2014, 136, 3760.
19. S. Ito, S. Tanaka, K. Manabe and H. Nishino, J. Phys. Chem. C, 2014, 18, 16995.
20. J. Huang, Z. Xu and Y. Yang, Adv. Funct. Mater., 2007, 17, 1966.
21. G. Li, C. W. Chu, V. Shrotriya, J. Huang and Y. Yang, Appl. Phys. Lett., 2006, 88, 253503.
22. W. Li, H. Dong, L. Wang, N. Li, X. Guo, J. Li and Y. Qiu, J. Mater. Chem. A, 2014, 2, 13587.
23. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nat. Mater., 2014, 13, 897.
24. E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz and J. R. Durrant, J. Am. Chem. Soc., 2003, 125, 475.
25. V. Somsongkul, F. Lang, A. R. Jeong, M. Rusu, M. Arunchaiya and T. Dittrich, Phys. Status Solidi RRL, 2014, 8, 763.
26. Q. Chen, H. Zhou, T.-B. Song, S. Luo, Z. Hong, H.-S. Duan, L. Dou, Y. Liu and Y. Yang, Nano Lett., 2014, 14, 4158.
27. A. Dualeh, T. Moehl, N. Tetreault, J. Teuscher, P. Gao, M. K. Nazeeruddin and M. Gratzel, ACS Nano, 2014, 8, 362.

---

Footnote

† Electronic supplementary information (ESI) available: Elemental ration result from XPS, J–V curves and SEM results. See DOI: 10.1039/c4ra08565e

---