8986 | Chem. Commun., 2015, 51, 8986--8989 This journal is©The Royal Society of Chemistry 2015

Cite this:Chem. Commun., 2015,

51, 8986

Modified deposition process of electron transportlayer for efficient inverted planar perovskite solarcells†

Hua Dong,‡a Zhaoxin Wu,‡*a Bin Xia,a Jun Xi,a Fang Yuan,a Shuya Ning,a

Lixin Xiao*b and Xun Houa

A highly-efficient inverted heterojunction perovskite solar cell was

prepared. A homogeneous and compact perovskite (CH3NH3PbI3)

layer was prepared via a two-step solution deposition method, and

subsequently a double-layer PCBM film was deposited by a sequential

spin-coating/vapor deposition process as the electron transport layer.

The optimised device could achieve a 12.2% (average 11.09%) efficiency.

Hybrid organometal halide perovskite solar cells have receivedmuch attention due to their superior intrinsic properties forsolar energy conversion. This type of solar cell was first reportedby Miyasaka et al. with a power conversion efficiency (PCE) of4%,1 and despite only several years elapsing, the efficiency soonevolved to over 19%.2–5 The hybrid perovskite materials exhibitappealing features such as high absorption coefficients, excellentambipolar charge mobility and appropriate band gaps.6–8 Moreover,the simple cell configuration and low cost fabrication process ofthis type of solar cell could further address the scalability andapplications in the photovoltaic field.9

Recently, typical perovskite solar cells have employed anelectron-transport layer (ETL)/perovskite material/hole-transportlayer (HTL) structure. Initial meso-superstructured-type solar cellsbased on a mesoporous TiO2 layer as the electron-transport layerrequired high-temperature processing (>450 1C), which hinders thecommercialization.10–12 Therefore, many approaches have beendeveloped to avoid the high-temperature barrier. One promisingapproach, a planar-inverted heterojunction device architecture, isparticularly interesting due to its simple cell configuration andpossible low-temperature fabrication.8,13 The use of phenyl-C61-butyric acid methyl ester (PCBM) as the electron-transport layerhas been investigated by several groups.14,15 However, perovskite

films with incomplete and non-homogeneous coverage via one-stepspin-coating has usually been observed in planar perovskite solarcells, and these subsequent disadvantages have been regarded asthe barrier which results in decreased device performance.16,17

Many efforts have been made to control the morphology ofperovskite thin films. Various modified one-step spin-coatingdeposition methods have been investigated, such as addingDIO and GBL as additives to achieve a high coverage perovskitelayer.17,18 However, the bareness and inhomogeneity problemsstill cannot be completely avoided. Bolink et al. proposed anevaporation/sublimation method, achieving a solar cell with >12%efficiency.4 Those studies demonstrated the great potential ofthe perovskite–PCBM architecture. However, the complicatedand rigorous fabrication process of the perovskite layer may placeconstraints on its application, and meanwhile the monitoring andcontrol of perovskite layer deposition by vacuum sublimation isdifficult.3,4,19

It is known that perovskite film based on a two-step solutionprocess could promise strong light-harvesting, good charge transportand complete coverage, which have been demonstrated in thehigh-temperature mesoporous devices.18,20 However, such atechnology has rarely been applied in planar perovskite–PCBMsolar cells, and reported perovskite–PCBM devices based on atwo-step solution process only show a less than 8% efficiency.21,22

This is mainly because the two-step solution-processed perovskitelayer is quite rough, and the subsequent ETL hardly fills in theundulations and smooths the interface between the ETL and themetal electrode. As a relatively easy-to-fabricate planar perovskite–PCBM solar cell with low-temperature fabricability, it is quitedesirable to increase its efficiency.

In this communication, we report a simple and effective methodto enhance the performance of perovskite–PCBM solar cells basedon a two-step solution-processed perovskite layer. The deviceconfiguration is shown in Fig. 1(a). The structure is designed asfollows: glass, ITO substrate, HTL PEDOT:PSS, CH3NH3PbI3

(MAPbI3), ETL PCBM (layer-by-layer) and a Ag electrode.As one type of fullerene derivative, PCBM has been the most

pervasive ETL applied in planar heterojunction perovskite solar cells.

a Key Laboratory of Photonics Technology for Information, Key Jiaotong University,

Xi’an 710049, P. R. China. E-mail: zhaoxinwu@mail.xjtu.edu.cn;

Tel: +86-29-82664867b State Key Laboratory for Mesoscopic Physics and Department of Physics,

Peking University, Beijing 100871, P. R. China. E-mail: lxxiao@pku.edu.cn

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cc01483b‡ These authors contributed equally to this work.

Received 21st February 2015,Accepted 9th April 2015

DOI: 10.1039/c5cc01483b

www.rsc.org/chemcomm

ChemComm

COMMUNICATION

Publ

ishe

d on

09

Apr

il 20

15. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

28/0

5/20

15 1

2:39

:56.

View Article OnlineView Journal | View Issue

This journal is©The Royal Society of Chemistry 2015 Chem. Commun., 2015, 51, 8986--8989 | 8987

Considering the membranous nature and solubility of the material,the reported thicknesses of spin-coated PCBM layers are in the rangeof dozens of nanometers. This range could satisfy the coveringrequirement of solar cells based on one-step solution or vapordeposition technology, to obtain considerable efficiency.

However, the spin-coating process alone could hardly completelyfill in and cover the rough perovskite layer by a two-step solutionprocess, which would lead to charge recombination losses andinfluence the transport at the interface. Here, we have designed asequential solution–vapor deposition process for preparing thePCBM layer, the highlight in our study, to solve this problem.The initial spin-coating process could fill in the depressions ofthe MAPbI3 layer preliminarily, then the sequential vapor deposi-tion process could further smooth the interface and separate theMAPbI3 layer and the Ag electrode. Through optimizing thefabrication process, a final PCBM layer with moderate thickness,good film properties and a planar surface can be obtained. Solarcells based on the sequential PCBM layer achieved high lightharvesting efficiencies as well as good charge transport. Thebest device showed a 12.2% efficiency with Voc = 0.99 V, Jsc =18.11 mA cm�2 and FF = 0.68; this efficiency is superior to thoseof the reported perovskite–PCBM devices based on a two-stepMAPbI3 layer.21,22 We hope that as an easy-to-fabricate andhighly-efficient structure, this modified solution-processed-based perovskite solar cell could provide a new method forpromoting the development of this hot solar cell.

In Fig. 1(b), we first compare the X-ray diffraction patterns offilms of CH3NH3PbI3 formed by either a one-step or a two-stepdeposition method. Strong peaks at 14.031 and 27.521, corre-sponding to the (110) and (220) planes, confirm the formationof a tetragonal perovskite structure.3 There is only a smallsignature of a peak at 12.651 (the (001) diffraction peak forPbI2). This indicates that the two MAPbI3 films both presenteda high level of phase purity, which is also a precondition forachieving highly-efficient perovskite solar cells.

The top-view SEM images in Fig. 2a and b show thedifferences between the MAPbI3 film morphologies producedby the two deposition processes. It can be seen that the MAPbI3

film based on a one-step process has a low coverage and largevoids. This shapeless perovskite morphology has been identified asbeing very detrimental to the performance because it not onlycauses electrical shorting, but also leads to the charge dissociation/transport/recombination.23 However, the cube-like MAPbI3 filmwith B200 nm domain sizes based on a two-step process presentshomogeneous and complete coverage on the substrate. It is known

that a highly roughened interface strengthens internal lightscattering, and that larger crystallites and a larger interface areaare also beneficial for charge transport.24 These properties havebeen demonstrated in mesoporous TiO2-based perovskite solarcells, and we hope to develop these mentioned advantages inplanar PCBM–perovskite solar cells.

As for the planar perovskite solar cell, a high PCE requiresboth strong light harvesting and good charge transport at theinterfaces. So how to effectively fill in and smooth the roughperovskite layer is important; only then can good interfacecontact for charge transport be achieved, as well as the advantagesof the two-step perovskite layer be utilized. Modification of thefabrication process of the PCBM layer has been studied in order toachieve this win–win situation in our study. Here, we designeddevices using three PCBM deposition methods in order to developthe most appropriate deposition technique: spin-coating only;vaporing only and sequential spin-vapor deposition. A device basedon a one-step MAPbI3 layer was also prepared as a reference. Fig. S1(ESI†) shows the schema of the four types of devices. Each typeemployed a PCBM layer with gradient thickness to find the optimalperformance.

Before preparing the devices, all the thicknesses of thePCBM mentioned in the following text were calibrated by anellipsometer on the same standard planar glass substrates.

To prove the repeatability of the devices, 30 cells of eachstructure were prepared. The performances of the devices weremeasured and are shown in Table S1 (ESI†). For structure type A,based on a one-step MAPbI3 layer, the device with 60 nm PCBMexhibited an average performance of PCE = 7.37% (Table S2,ESI†). The absence of material from some regions in the one-stepMAPbI3 film (pinholes) will result in direct contact of the ETLPCBM and the HTL PSS:PEDOT, leading to a shunting path thatis probably partially responsible for the lower FF and open-circuit voltages in planar heterojunction devices.25

The use of a two-step-based MAPbI3 layer could avoid directcontact between the HTL and the ETL. Before preparing thestructures of types B, C and D, the roughness of the two-stepMAPbI3 layer was observed, to aid in the design of the devices.Fig. 3a and e show the tilt-angle SEM and 3-dimensional AFMimages of the MAPbI3 layer; the root mean square (RMS) roughnessis 25.5 nm. This rough surface requires a certain thickness ofPCBM to completely fill in and smooth the MAPbI3 layer.

We first tried the spin-coating only method for PCBM,shown as structure B, and a series of devices with variousthicknesses (60–120 nm) were prepared. It can be seen that

Fig. 1 (a) Device structure and energy band diagram of the perovskite–PCBM solar cell. (b) X-ray diffraction profiles of one-step and two-stepsolution-processed perovskite (CH3NH3PbI3) films.

Fig. 2 SEM top views of the one-step (a) and two-step (b) solution-processed deposited perovskite films.

Communication ChemComm

Publ

ishe

d on

09

Apr

il 20

15. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

28/0

5/20

15 1

2:39

:56.

View Article Online

8988 | Chem. Commun., 2015, 51, 8986--8989 This journal is©The Royal Society of Chemistry 2015

devices with 60–80 nm PCBM thickness exhibit a considerable8.1% efficiency. This efficiency is comparable with the reportedrecord. Fig. 3b and f show the tilt-angle SEM and 3-D AFMimages of structure B (PCBM thickness = 60 nm) without Ag.We can see that the rough MAPbI3 layer was relatively coveredby the PCBM layer, and the RMS roughness obviously decreased(RMS = 12.4 nm). It is noted that the top of the MAPbI3 crystalscan still be observed in some regions, which led to directcontact between the MAPbI3 layer and the Ag electrode whenthe real cell operated. This problem may have limited thepossibility of a higher performance. When the thickness wasover 100 nm, the PCE obviously decreased, which was mainlydue to the low film-forming quality for excessively thick films.Fig. S2 (ESI†) shows the tilt-angle SEM image of structure B(PCBM thickness = 100 nm) without Ag. It was found that thesurface changed to become irregular and rough, and theconsequent poor interface would have influenced the charge-transport. The study of structure B clarified that the spin-coating process alone could not provide a suitable PCBM filmfor achieving a highly-efficient device.

Vapor deposition is another practical method to obtain acontinuous film with tunable thickness, and devices based on avapor deposition method only for the ETL are shown asstructure C. The results in Table S1 (ESI†) show that with a100 nm-thickness of PCBM, an average of 7.61% efficiency forthe devices was obtained. As we know, films formed via a vapordeposition process just inherit the morphology of the under-layer materials, thus this process could hardly buffer themorphology of the architecture. Fig. 3c and g show the tilt-angle SEM and 3-D AFM images of structure C (PCBM thickness =100 nm) without Ag. It was found that although the MAPbI3

layer was almost covered by the PCBM, the surface was stillrough with a 21.7 nm RMS roughness. Such an asperous andinhomogeneous ETL would be adverse to fluent charge-transportat the interface, and the relatively low FF of the devices directlyreflected this. The study of structure C indicated that a PCBMlayer with sufficient thickness can be obtained by vapor deposi-tion, but that the roughness of the whole device was hardlydecreased.

Based on structures B and C, we designed structure D inorder to explore the best performance. 60 nm-thick PCBM wasfirst deposited by spin-coating for initial filling, then a tunablePCBM layer was prepared by vapor deposition for furthersmoothing the interface and isolating the under-MAPbI3 layerand the metal electrode. Considering that excessively thickPCBM could decrease the conductivity of the film, the rangeof the vaporing layer is limited to 10–80 nm. A surprisingefficiency with an average of 11.09% appeared for the typicalstructure (ETL thickness: 60 nm + 30 nm). Fig. 3d and h showthe tilt-angle SEM and 3-D AFM images of structure D (ETLthickness: 60 nm + 30 nm) without Ag. We can see that thisdouble-layer deposition process could efficiently cover theMAPbI3 layer and obviously decrease the roughness. The rathersmall RMS roughness (7.8 nm) shows that the surface of thestructure was nearly flat, which is the basis of the planarinverted PCBM–MAPbI3 solar cell. The good performance ofthe Voc, FF and Jsc all clarified that strong light-harvesting aswell as good charge-transport were achieved. As the thicknessof the PCBM film by vaporing increased, the performance of thedevice decreased, which was mainly attributed to the lowconductivity of the excessively thick PCBM layer.

As a response to the charge transport/recombination tosome extent, dark-current was studied to further investigatethe performances of different devices. Fig. 4a shows the averageJ–V characteristics of the devices based on structures B (ETLthickness = 60 nm), C (ETL thickness = 100 nm) and D (ETLthickness = 60 + 30 nm) under illumination and in the dark.The results show that device D has the lowest dark-current,which demonstrated the good performance on reducing thecharge recombination loss and benefiting the charge transport.The larger dark-current values of devices B and C was explainedas follows: as for the device B, the bare protuberance of theMAPbI3 layer (shown in Fig. S1 (ESI†)) in some regions would bein direct contact with the Ag electrode, which led to interfacerecombination and was reflected in the low FF and open-voltage values. Coming to device C, the film-forming propertiesof the PCBM layer were poor and the surface was rough, whichreduced the fluency and uniformity of the charge-transport at

Fig. 3 Tilt-angle SEM and 3-D AFM images of the MAPbI3 layer (a and e), structure B: ITO/PEDOT/MAPbI3(two-step)/PCBM(spin-coating) (b and f),structure C: ITO/PEDOT/MAPbI3(two-step)/PCBM(vaporing) (c and g) and structure D: ITO/PEDOT/MAPbI3(two-step)/PCBM(spin-coating)/PCBM(vaporing) (d and h).

ChemComm Communication

Publ

ishe

d on

09

Apr

il 20

15. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

28/0

5/20

15 1

2:39

:56.

View Article Online

This journal is©The Royal Society of Chemistry 2015 Chem. Commun., 2015, 51, 8986--8989 | 8989

the MAPbI3/ETL/Ag interface. The low FF of the device waspartially a consequence of these disadvantages.

The highest performance of device D is shown in Fig. 4b,exhibiting a performance of Voc = 0.99 V, Jsc = 18.11 mA cm�2,FF = 0.68 and PCE = 12.2% under reverse scanning, and Voc =0.97 V, Jsc = 18.09 mA cm�2, FF = 0.63 and PCE = 11.04% underforward scanning. The hysteresis between the forward and thereverse J–V scans is the manifestation of a slow response time ofthe cell to a change in load.26 The relative standard deviationsof the parameters was less than 10% (shown in Table S1 (ESI†)and Fig. S3, ESI†), which indicated that the performance of thedevice was highly reproducible with low variation. The inte-grated current density derived from the EQE spectra in Fig. S4(ESI†) was in close agreement with the value measured undersimulated sunlight shown in Fig. 4b. Fig. S5 (ESI†) shows across-section SEM image of the inverse device D. The cross-section SEM image also depicts a uniform deposition along thelength of the device. Finally, the evolution of the PCE for thebest device, D, was observed and is shown in Fig. S6 (ESI†); thisindicated a considerable stability.

In summary, we have developed a sequential layer-by-layerPCBM deposition method to fabricate planar perovskite solarcells. The solution–vapor process to deposit the ETL couldeffectively fill in and smooth the rough MAPbI3 layer. Thecube-like and homogeneous MAPbI3 layer promised stronglight-harvesting and benefited the charge transport. Mean-while, planar and continuous PCBM film could separate theMAPbI3 layer and the Ag electrode, for reducing recombinationloss. With a rather low variation, the best PCE presents a 12.2%efficiency, much higher than the reported PCBM–MAPbI3 solarcells with the same structure.

This work was financially supported by the Basic ResearchProgram of China (2013CB328705), National Natural ScienceFoundation of China (Grant No. 61275034) and PhD Programs

Foundation of Ministry of Education of China (Grant No.20130201110065).

Notes and references1 A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc.,

2009, 131, 6050.2 J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao,

M. K. Nazeeruddin and M. Gratzel, Nature, 2013, 499, 316.3 M. Z. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501.4 O. Malinkiewicz, A. Yella, Y. H. Lee, G. M. Espallargas, M. Graetzel,

M. K. Nazeeruddin and H. J. Bolink, Nat. Photonics, 2014, 8, 128.5 H. P. Zhou, Q. Chen, G. Li, S. Luo, T. B. Song, H. S. Duan, Z. R. Hong,

J. B. You, Y. S. Liu and Y. Yang, Science, 2014, 345, 542.6 S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer,

T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013,342, 341.

7 G. C. Xing, N. Mathews, S. Y. Sun, S. S. Lim, Y. M. Lam, M. Gratzel,S. Mhaisalkar and T. C. Sum, Science, 2013, 342, 344.

8 S. Y. Sun, T. Salim, N. Mathews, M. Duchamp, C. Boothroyd,G. C. Xing, T. C. Sum and Y. M. Lam, Energy Environ. Sci., 2014, 7, 399.

9 D. Bryant, P. Greenwood, J. Troughton, M. Wijdekop, M. Carnie,M. Davies, K. Wojciechowski, H. J. Snaith, T. Watson andD. Worsley, Adv. Mater., 2014, 26, 7499.

10 M. D. McGehee, Nature, 2013, 501, 323.11 S. Kazim, M. K. Nazeeruddin, M. Gratzel and S. Ahmad, Angew.

Chem., Int. Ed., 2014, 53, 2812.12 Y. Z. Ma, L. L. Zheng, Y. H. Chung, S. S. Chu, L. X. Xiao, Z. J. Chen,

S. F. Wang, B. Qu, Q. H. Gong, Z. X. Wu and X. Hou, Chem.Commun., 2014, 50, 12458.

13 J. Y. Jeng, Y. F. Chiang, M. H. Lee, S. R. Peng, T. F. Guo, P. Chen andT. C. Wen, Adv. Mater., 2013, 25, 3727.

14 G. E. Eperon, V. M. Burlakov, P. Docampo, A. Goriely andH. J. Snaith, Adv. Funct. Mater., 2014, 24, 151.

15 J. B. You, Z. R. Hong, Y. Yang, Q. Chen, M. Cai, T. B. Song,C. C. Chen, S. R. Lu, Y. S. Liu, H. P. Zhou and Y. Yang, ACS Nano,2014, 8, 1674.

16 B. Conings, L. Baeten, C. De Dobbelaere, J. D’Haen, J. Manca andH. G. Boyen, Adv. Mater., 2014, 26, 2041.

17 P. W. Liang, C. Y. Liao, C. C. Chueh, F. Zuo, S. T. Williams, X. K. Xin,J. J. Lin and A. K. Y. Jen, Adv. Mater., 2014, 26, 3748.

18 H. B. Kim, H. Choi, J. Jeong, S. Kim, B. Walker, S. Song and J. Y. Kim,Nanoscale, 2014, 6, 6679.

19 C. W. Chen, H. W. Kang, S. Y. Hsiao, P. F. Yang, K. M. Chiang andH. W. Lin, Adv. Mater., 2014, 26, 6647.

20 L. L. Zheng, Y. Z. Ma, S. S. Chu, S. F. Wang, B. Qu, L. X. Xiao,Z. J. Chen, Q. H. Gong, Z. X. Wu and X. Hou, Nanoscale, 2014, 6, 8171.

21 L. Hu, J. Peng, W. W. Wang, Z. Xia, J. Y. Yuan, J. L. Lu, X. D. Huang,W. L. Ma, H. B. Song, W. Chen, Y. B. Cheng and J. Tang, ACSPhotonics, 2014, 1, 547.

22 Z. L. Zhu, Y. Bai, T. Zhang, Z. K. Liu, X. Long, Z. H. Wei, Z. L. Wang,L. X. Zhang, J. N. Wang, F. Yan and S. H. Yang, Angew. Chem., Int.Ed., 2014, 53, 12571.

23 P. Docampo, J. M. Ball, M. Darwich, G. E. Eperon and H. J. Snaith,Nat. Commun., 2013, 4, 2761.

24 D. Y. Liu and T. L. Kelly, Nat. Photonics, 2014, 8, 133.25 Q. Wang, Y. C. Shao, Q. F. Dong, Z. G. Xiao, Y. B. Yuan and

J. S. Huang, Energy Environ. Sci., 2014, 7, 2359.26 H. J. Snaith, A. Abate and J. M. Ball, J. Phys. Chem. Lett., 2014, 5, 1511.

Fig. 4 (a) J–V characteristics of the devices B–D under illumination and inthe dark. (b) J–V characteristics of device D with the highest efficiency.

Communication ChemComm

Publ

ishe

d on

09

Apr

il 20

15. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

28/0

5/20

15 1

2:39

:56.

View Article Online