Introducing Cu2O Thin Films as a Hole-Transport ... - ACS Publications

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article pubs.acs.org/JPCC

Introducing Cu2O Thin Films as a Hole-Transport Layer in Efficient Planar Perovskite Solar Cell Structures Soumyo Chatterjee and Amlan J. Pal* Department of Solid State Physics, Indian Association for the Cultivation of Science, Jadavpur, Kolkata, West Bengal 700032, India S Supporting Information *

ABSTRACT: In this work, we introduce Cu2O thin films as a hole-transport layer in planar perovskite solar cells. Here, a Cu2O layer was formed through successive ionic layer adsorption and reaction (SILAR) method. With methylammonium lead triiodide (MAPbI3) we form a direct structure (p− i−n), where the perovskite layer is sandwiched between a layer of p-type Cu2O and another layer of n-type PCBM (phenylC61-butyric acid methyl ester), which acted as hole- and electron-transport materials, respectively. We locate band edges of the materials with respect to their Fermi energy by recording scanning tunneling spectroscopy that has correspondence to their density of states (DOS). We observe that the energy levels of the materials form type II band alignments at each of the two interfaces (p−i and i−n) for charge separation and uninterrupted carrier transport upon illumination. Such a band alignment enabled charge transfer from MAPbI3 as evidenced from quenching of its photoluminescence emission when the perovskite was in contact with either the hole- or the electron-transport layer. With the direct p−i−n structure having appropriate energy levels for carrier separation, the planar perovskite solar cell (Cu2O/MAPbI3/PCBM) yielded an energy conversion efficiency (η) of 8.23% under 1 sun illumination.

1. INTRODUCTION

hole mobility apart from energy levels remaining compatible with the valence band edge of the perovskite. A direct structure (p−i−n) is always preferred in thin-film solar cells, since an expensive and high work-function metal (5.1 eV for gold) used as a top electrode in an inverted geometry often creates pinholes. In perovskite solar cells, a p− i−n structure is more advantageous, since with a lower workfunction metal as a hole-collecting electrode we may then have further choice in replacing spiro-MeOTAD. With indium tin oxide (ITO) as a hole-collecting electrode, the HOMO (or valence band) of the HTL can then be extended up to 4.7 eV without any barrier for holes to be transported through the ITO/HTL interface (4.7 eV is the well-known work-function of ITO). Here HOMO represents the highest occupied molecular orbitals of conjugated organics. In such direct structures, while PCBM remained the only ETL,12 nickel oxide (NiO),10,13 copper-doped NiO (Cu@NiO), 14 CuSCN, 10 and PEDOT:PSS12 have so far been used as HTL in ITO/HTL/ MAPbI3/PCBM/Al structures. Here, PCBM and PEDOT:PSS represent phenyl-C61-butyric acid methyl ester and poly(3,4ethylenedioxythiophene) polystyrenesulfonate, respectively. In our quest for a suitable HTL in such direct thin-film structures, we report introduction of Cu2O in perovskite solar cells. The HTL has been grown through a successive ionic layer adsorption and reaction (SILAR) method.15,16 In a very recent

In recent years, research on solar cells received a quantum jump with the introduction of organometal halides.1 This unique class of perovskites, namely, methylammonium lead triiodide (CH3NH3PbI3, MAPbI3) and its derivatives, is sandwiched between a hole-transport layer and an electron-transport layer in the form of n−i−p structures to form mesoscopic or planar (thin-film) perovskite solar cells having an inverted device geometry.1−5 After light absorption, charge generation as well as charge extraction occurs in the perovskite layer, followed by transport of holes and electrons through the hole- and electrontransport layers (HTL and ETL), respectively. Due to such a 2fold role played by the perovskite in solar cells, the devices have produced promisingly high power conversion efficiency. The high output is consistent with its ambipolar transport property and long-range electron−hole diffusion lengths in addition to high carrier mobility of the material.6,7 In consideration of the band edges of the perovskite material, either TiO2 or ZnO was used as an ETL.2−4,8,9 2,2′,7,7′-Tetrakis(N,N-di-4-methoxyphenylamino)-9,9′-spirobifluorene (commonly known as spiroMeOTAD) has so far been the only major choice as the HTL so that holes did not face any energy barrier at the MAPbI3/ HTL and HTL/gold interfaces.1,4,8 In efforts to replace spiroMeOTAD with a more stable and less expensive suitable material, such as CuI,2 CuSCN,3,10,11 carbazole,9 or copper phthalocyanine,5 the efficiency of solar cells had to be compromised. To obtain high photovoltaic performance, it is of critical importance that the HTL should also possess high © 2016 American Chemical Society

Received: November 26, 2015 Revised: December 24, 2015 Published: January 7, 2016 1428

DOI: 10.1021/acs.jpcc.5b11540 J. Phys. Chem. C 2016, 120, 1428−1437

Article

The Journal of Physical Chemistry C

Figure 1. (a) UV−visible absorption spectrum, (b) Tauc plot estimating the optical band gap, (c) XRD pattern, and (d) Raman spectrum of MAPbI3 thin film. Insets show (a) d[(αhν)2]/d(hν) versus energy plot of the material and (b) lattice structure of the perovskite.

material begins at that wavelength. The shouldered peak at 490 nm was similarly assigned to the transition from the lower valence band (VB2) to the CB in the two-valence-band model; an iodide-to-lead charge-transfer transition has also been assigned to the higher-energy peak. A better view of these two photoinduced peaks can be obtained from d[(αhν)2]/ d(hν) versus hν plots as shown in the inset of Figure 1a. The plot shows two distinct peaks at energy values 1.67 and 2.49 eV, which correspond to 742 and 498 nm, respectively. The positions match the peaks appearing in the absorption spectrum. In Figure 1b, we show a plot of (αhν)2 versus energy (Tauc plot) to obtain the optical gap of the perovskite. From extrapolation of the linear part of rgw Tauc plot (Kubelka− Munk theory26), the optical gap of MAPbI3 could be estimated as 1.54 eV, which is in close agreement with previous reports.27 Here, the optical gap inferred the low-energy transition between VB1 and CB. Since the crystalline nature of semiconductors is an important parameter in forming solar cells, we have recorded X-ray diffraction (XRD) patterns of the perovskite thin films. The patterns moreover infer formation of the material when they are compared with reported results in this direction.8,28 In Figure 1c, we show XRD patterns of MAPbI3 that yielded intense and sharp peaks at appropriate diffraction angles, ensuring formation of the corresponding material in a single phase. The material adopted a typical tetragonal structure for CH3NH3MX3 hybrid perovskites (M = Pb, Sn; X = halogen) formed by a three-dimensional anionic framework of PbI6 octahedron with methylammonium cations in the interstitial

report, which came to our attention during manuscript review, solution-processed Cu2O (and CuO) were used as holetransport materials in perovskite solar cells.17 Apart from its natural p-type conductivity, Cu2O also possesses a high carrier mobility of about 100 cm2/(V·s) and a long carrier diffusion length ranging up to several micrometers.18−20 Since light enters through the HTL in a direct (p−i−n) structure, a band gap of 2.16 eV21 moreover would allow Cu2O to absorb higherenergy photons while remaining transparent to the perovskite. In this work, we report fabrication and characterization of planar perovskite solar cells based on Cu2O and MAPbI3.

2. RESULTS AND DISCUSSION 2.1. Characterization of Perovskite Material. To ensure formation of the perovskite, the material was characterized by conventional techniques. In Figure 1, we present the optical absorption spectrum of a MAPbI3 thin film deposited on a quartz substrate. The spectrum that extends until the near-IR region has features characteristic of the perovskite material, indicating that MAPbI3 indeed formed during the reaction.22 The spectrum shows two dominant and distinct photoinduced peaks located at around 490 and 740 nm and a decrease in the absorbance in the 500−520 nm region, which are characteristics of MAPbI3 films deposited from N,N-dimethylformamide (DMF) solvent.23,24 The peaks have been assigned to transitions arising due to a dual valence band structure (VB1 and VB2) of methylammonium-based perovskites.25 The peak at 740 nm has been attributed to a direct gap transition from the first valence band maximum (VB1) to the conduction band (CB) minimum, representing that the photogeneration in the 1429


Article

The Journal of Physical Chemistry C space (inset of Figure 1b). A closer look of the patterns reveals that in the region of the (110) diffraction peak at 14.12° there is only a small signature of a peak at 12.65° (marked with an asterisk) that corresponds to the (001) diffraction peak for unreacted PbI2. In the spectrum of MAPbI3−xClx, the peak at 12.65° did not appear, implying absence of PbI2 in the material. (It may be recalled that PbI2 was not required to form MAPbI3−xClx.) The representative peak of MAPbCl3 that in general appears at 15.68° was also absent in the spectrum, implying formation of MAPbI3−xClx with a high level of phase purity. The optical absorption spectrum and XRD pattern of MAPbI3−xClx are presented in Figure S1 (Supporting Information). We have further characterized the perovskite by recording its Raman spectrum. The films appeared to be very sensitive to laser power. We therefore used a low laser power that was just sufficient to obtain adequate Raman signal; when the power was a little higher, the material was found to be affected by the laser beam, as evidenced by a change in its color. In the Raman spectrum, as presented in Figure 1d, a couple of modes in the low-frequency range were observed at 67.2 and 86.4 cm−1. The former could be assigned to the inorganic component of the perovskite, arising out of bending vibrations of the I−Pb−I bonds and a consequent libration of the cations due to deformation of the inorganic cage. The latter mode at 86.4 cm−1 also represents the inorganic component of the material and is known to be associated with both Pb−I stretching and libration modes of the cations. The sharp peak at 143.2 cm−1 appears due to the motion of the organic cations (methylammonium). Modes in the 200−400 cm−1 range could be attributed to torsional vibration frequencies of the organic cations. Quarti et al.29 have proposed these modes as possible markers of orientation order of the organic cations in the material, and thus of the whole crystal. The Raman spectrum hence substantiated the formation of CH3NH3PbI3 for use in perovskite solar cells. When we analyzed the elemental composition of the perovskites through energy-dispersive Xray (EDX) analysis (Figure S2 in Supporting Information), the spectra matched the reported results and were reproducible over the whole scan region, implying homogeneity of the synthesized material.30 We characterized the layers of NiO and Cu@NiO and of Cu2O also through optical absorption, Tauc plots, and XRD patterns. The characteristics matched the reported results of the respective materials.16,21,31,32 In Figures S3 and S4 (Supporting Information), we present such plots of the materials. 2.2. Estimation of Band Edges by Scanning Tunneling Spectroscopy. Since energy levels of a p−i−n heterojunction must form a type II band alignment at both interfaces for uninterrupted carrier transport, we aimed to locate the CB and VB of the individual materials with respect to their Fermi energy. In order to do so, we recorded tunneling current versus tip voltage characteristics of ultrathin films of the materials formed separately on n-type highly doped Si ⟨111⟩ substrates. From the obtained data, we have calculated the differential conductance (dI/dV) of the materials that has correspondence to their density of states (DOS). In the DOS spectra, which are shown in Figure 2a, since tip voltage was applied with respect to the substrate, the peaks in the positive and negative voltages denoted VB and CB of the semiconductors, respectively. For PCBM, the peaks represented HOMO and lowest unoccupied molecular orbital (LUMO) levels, respectively. The work-

Figure 2. (a) Differential conductance (dI/dV) plots of NiO, Cu@ NiO, Cu2O, MAPbI3, and PCBM thin films. VB and CB edges of the materials and HOMO and LUMO levels of PCBM with respect to the Fermi energy are marked with vertical continuous and broken straight bars in the positive and negative voltage regions, respectively. (b) Schematic energy level diagram of Cu2O/MAPbI3/PCBM heterojunction. The dashed line represents the Fermi energy after contact.

function of the substrate electrode is set to be aligned to the Fermi energy of the semiconductors. The DOS spectrum of MAPbI3 shows that while the CB edge was located at 0.69 eV above the Fermi energy, the first valence band maximum (VB1) could be located at 0.75 eV below. Location of the lower VB (VB2) could not be seen in the DOS spectrum, since tunneling to VB1 remained efficient also at higher positive voltages. The DOS spectrum moreover evidenced that the Fermi energy was mostly at the middle of CB and first VB edges, inferring the intrinsic nature of the semiconductor in pristine thin films. The DOS spectra of NiO, Cu@NiO, and Cu2O show that the Fermi energy was closer to the valence band edge, implying the p-type nature of the materials. Similarly, PCBM can be seen to act as an electron-acceptor material (n-type semiconductor). Transport gap of the semiconductors can be estimated by calculating the difference between CB and VB edges in the DOS spectra. The results show that the transport gap was lower than the optical gap of each of the materials. This is expected since trap states also contribute to tunneling current and hence appear in the DOS spectrum. From the VB and CB edges of the semiconductors, we have formulated the energy band diagram of the proposed device architectures that are shown in Figure 2b and Figure S5 (Supporting Information). The band diagram implied that the heterojunctions formed a p−i−n architecture, which is appropriate for carrier separation and to act as solar cells. The expected carrier separation at the HTL/MAPbI3 and MAPbI3/ETL interfaces can be monitored also by recording PL spectrum of the heterojunctions. It may be recalled that in case of TiO2/MAPbI3/Spiro-MeOTAD n−i−p structures, when PL of MAPbI3/Spiro-MeOTAD and TiO2/MAPbI3 heterojunctions were recorded, PL of the perovskite that appeared at around 763 nm was found to be quenched due to a holetransfer to Spiro-MeOTAD and an electron-transfer to TiO2, respectively.33 In our Cu2O/MAPbI3/PCBM p−i−n structures, we accordingly went on to record PL emission of the perovskite in the Cu2O/MAPbI3 and MAPbI3/PCBM heterojunctions. In Figure 3, we present photoluminescence (PL) spectra of such heterojunctions in addition to a spectrum of the perovskite film in its pristine form. PL emission of MAPbI3 appeared at 763 nm, representing transition from the VB1 edge. The emission of the perovskite quenched in the heterojunctions with HTL and ETL separately. Quenching of PL emission in the two 1430


Article


that the sandwiched structures may act as solar cells under illumination. For each type of devices, we characterized at least 20 cells; a typical I−V of each of the devices under 1 sun condition is shown in the corresponding panel. All the devices acted as solar cells, the parameters of which are listed in Table 1. While the VOC was high and ranged between 0.89 and 1.03 V, Table 1. Photovoltaic Parameters of HTL/MAPbI3/PCBM Heterojunctions with Different HTLsa device structure NiO/MAPbI3/PCBM Cu@NiO/MAPbI3/PCBM Cu2O/MAPbI3/PCBM Cu2O/MAPbI3−xClx /PCBM

Figure 3. PL emission spectra of MAPbI3 thin films and Cu2O/ MAPbI3 (p−i) and MAPbI3/PCBM (i−n) heterostructures. Excitation wavelength of the PL measurements was 600 nm.

ISC (mA/cm2)

VOC (V)

fill factor (%)

η (%)

9.75 14.90 16.52 15.60

0.91 1.03 0.89 0.92

43 51 56 58

3.81 7.80 8.23 8.30

a

Parameters of device based on MAPbI3−xClx are shown in the last row.

heterojunctions occurred due to hole transfer to Cu2O and electron transfer to PCBM from the perovskite, respectively. Scanning tunneling spectroscopy (STS) results also inferred that such transfers were energetically possible. Photoluminescence of MAPbI3 in Cu2O/MAPbI3 heterojunction was quenched the most. The quenching process in Cu2O/ MAPbI3 was efficient, presumably due to a higher hole mobility of the oxide that would result in efficient charge extraction. The other HTLs, namely NiO and Cu@NiO, also quenched PL emission of the perovskite. The results hence support the inference of carrier separation at the two interfaces of the p−i− n heterojunctions that is necessary for solar cell applications. 2.3. Photovoltaic Performance. We hence proceeded to characterize the p−i−n heterojunctions under dark and illumination conditions. In Figure 4a−c, we present I−V characteristics of the heterojunctions with a layer of (a) NiO, (b) Cu@NiO, and (c) Cu2O as separate HTL, respectively. The dark characteristics were rectifying in nature, indicating

the fill factor was moderate (up to 56%). The ISC in Cu2Obased devices exhibited a higher value. Although the fill factor of Cu2O-based devices was not high enough, the energy conversion efficiency (η) reached up to 8.23% in such devices. Such efficiency in a direct planar structure is indeed worth reporting in perovskite solar cell research. Histograms of VOC, ISC, η, and fill factor of the solar cells, as measured in 20 cells, are presented in Figure S6 in Supporting Information. Excelling in energy conversion efficiency with Cu2O as the HTL occurs primarily due to a near-perfect alignment of its VB edge with that of the perovskite. In addition, a high degree of crystallinity of Cu2O formed through a SILAR method, as evidenced from XRD patterns, led to an improved holeextraction efficiency of the material. When the performances of devices based on NiO and Cu@NiO as HTL are compared, V OC and correspondingly η could be found to show improvement upon inclusion of copper in the NiO layer. The

Figure 4. Current−voltage characteristics of (a) NiO/MAPbI3/PCBM, (b) Cu@NiO/MAPbI3/PCBM, and (c) Cu2O/MAPbI3/PCBM heterojunctions under dark and white light illumination conditions. 1431


Article


which reached up to 8.30%. Solar cell parameters of Cu2O/ MAPbI3−xClx/PCBM heterojunction have been appended in Table 1. Since the efficiencies of MAPbI3−xClx- and MAPbI3based devices were comparable, we continued further characterization of the Cu2O/MAPbI3/PCBM heterojunction. We may add here that we did not target to use CuO as a HTL that can be formed through high-temperature annealing of Cu2O layer, since such annealing reduces carrier mobility and hence also the conductivity.34 Moreover, as compared to Cu2O, cupric oxide has a smaller band gap (∼1.4 eV);21 CuO would hence form an opaque layer in front of the perovskite, reducing its light-absorbing capability drastically. In the case of a planar junction perovskite solar cell, it is of critical importance that the layers are uniform and smooth for better contact with the next material. For example, a smooth MAPbI3 film would lead to better contact between the perovskite and PCBM layers, enabling efficient electron extraction in the devices. To compare the morphology of HTL and MAPbI3 films deposited on different HTLs, we recorded their atomic force microscopic (AFM) images (Figure S7 in Supporting Information). When we compared surface morphology of NiO, Cu@NiO, and Cu2O layers, we observed that the root-mean-square (RMS) roughness (RRMS) of the Cu2O layer was the least. This implied that the Cu2O/MAPbI3 interface will facilitate hole extraction in the devices. To compare morphology of MAPbI3 films deposited on different HTLs, we recorded AFM images of HTL/MAPbI3 films. While RRMS of the perovskite film deposited on bare glass substrate was around 62 nm, the smoothness of the MAPbI3 film improved by a large extent when the perovskite layer was formed on a HTL, implying a more compact and denser nature of the film with less voids and traps. With a layer of NiO, Cu@

presence of copper, which is known to occupy vacant metal sites, led to a decrease in carrier concentration (as evidenced by a shift in Fermi energy away from the VB edge); in the band diagram of Cu@NiO/MAPbI3/PCBM (Figure S5b in Supporting Information), upon alignment of Fermi energy, a large energy difference hence existed between VB of Cu@NiO and CB of MAPbI3. Since the energy difference is known to contribute to VOC, this parameter excelled in Cu@NiO/ MAPbI3/PCBM heterojunction, exceeding 1 V. We introduced Cu2O also with MAPbI3−xClx, that is, as a Cu2O/MAPbI3−xClx/PCBM p−i−n heterojunction device. I−V characteristics under dark and white light are shown in Figure 5.

Figure 5. Current−voltage characteristics of Cu2O/MAPbI3−xClx/ PCBM heterojunctions under dark and white light illumination conditions.

Here also, the dark characteristics were rectifying in nature; under illumination, the device acted as a solar cell, the η of

Figure 6. (a) −dV/dI vs (ISC − I)−1, (b) ln (ISC − I) vs (V + IRseries), and (c) dI/dV vs V plots of the heterojunctions with a range of HTLs as stated in the legend. In the former two plots, broken lines represent linear fits to the experimental points. 1432


Article

The Journal of Physical Chemistry C Table 2. Diode Parameters of Different Device Structures device structure NiO/MAPbI3/PCBM Cu@NiO/MAPbI3/PCBM Cu2O/MAPbI3/PCBM

Rseries (Ω·cm2) 18.2 9.4 9.1

I0 (mA/cm2)

ideality factor (n)

−2

9.2 × 10 5.2 × 10−2 8.0 × 10−4

7.9 7.0 3.8

calcd VOC (V)

Rshunt (kΩ·cm2)

0.93 1.06 0.97

132 165 380

Figure 7. Photocurrent hysteresis behavior of Cu2O/MAPbI3/PCBM heterojunction with different (a) voltage sweep directions and (b) delay times. Symbols have the same meaning in the two plots. (c) Photovoltaic parameters of the heterojunction measured over a span of 15 days.

⎤ ⎛ V + IR series ⎞ ⎡ q (V + IR series)⎥ − ⎜ I = ISC − I0 exp⎢ ⎟ ⎦ ⎝ R shunt ⎠ ⎣ nKBT

NiO, or Cu2O beneath, the RRMS of MAPbI3 layer was 19.4, 15.0, and 5.3 nm, respectively. Such a smooth surface of Cu2O/ MAPbI3 heterojunction would allow formation of a uniform PCBM layer leading to better contact between these two layers and hence unhindered electron extraction in the devices. Improved η in Cu2O/MAPbI3/PCBM could therefore be due to efficient hole and electron extractions at the two interfaces. A homogeneous ETL would furthermore allow deposition of a smooth top Al electrode, ensuring reflection of unabsorbed photons. 2.4. Junction Properties. Apart from the solar cell parameters, such as VOC, ISC, fill factor, and η, we have analyzed the I−V characteristics further. It is imperative that an efficient solar cell structure would possess low series resistance and high shunt resistance. While low series resistance (Rseries) improves charge extraction possibilities, large shunt resistance (Rshunt) would reduce recombination losses. We therefore analyzed the I−V curves to evaluate the resistances (Rseries and Rshunt) of the solar cells. Planar-structured solar cells can in general be treated as a single-junction diode; I−V characteristics of such diodes can be expressed as

(1)

where I is current flowing through the external load, ISC is lightinduced current, I0 is dark saturation current density, V is applied voltage, n is the ideality factor, KB is the Boltzmann constant, T is temperature, and q is electron charge. Equation 1 reduces to ln(ISC − I ) = ln I0 +

q (V + IR series) nKBT

(2)

when the contribution of Rshunt is neglected. From the first derivative of eq 2, Rseries of the devices can be evaluated as −

nKBT dV 1 = R series + dI q (ISC − I )

(3)

Equation 3 allowed us to determine the series resistances of our devices, based on the three HTLs, when we plotted −dV/dI versus (ISC − I)−1 (Figure 6a). Equation 3 in addition yielded the ideality factor (n) from the slope of the linear fit. When Rseries of the device is available, it may be used in eq 2 to determine I0 as intercept, with the ordinate of the linear fit to ln 1433


Article

The Journal of Physical Chemistry C (ISC − I) versus (V + IRseries) plots in the higher ISC − I region (Figure 6b). Rshunt of devices could also be evaluated from eq 1 as −

q dI 1 = + I0eqV / nKBT dV R shunt nKBT

0.35 mA/cm2 and 0.01 V, which amounted to 2% and 1% with respect to the magnitude of the parameters, respectively. A low hysteresis effect in Cu2O/MAPbI3/PCBM could be achieved, presumably due to a morphologically superior Cu2O layer on which the perovskite film was deposited. The low hysteresis in these devices moreover implied that the parameters achieved in the solar cells appeared truly due to efficient charge separation at p−i and i−n interfaces of the direct p−i−n heterojunctions. 2.6. Stability. In perovskite solar cells, stability has remained an issue since their inception.39,40 The instability apparently occurs due to degradation of the perovskite (and also spiroMeOTED) upon exposure to oxygen, moisture, UV radiation, temperature, and the solvents and additives used during processing of the materials. In order to protect the materials and the devices from oxygen and moisture, we fabricated the devices and characterized them in a glovebox. We did not use any additives; neither did we use any encapsulation. No special effort was undertaken to protect the devices from UV radiation or from normal indoor illumination. Parameters of the solar cell recorded once a day over a fortnight are shown in Figure 7c. The devices were found to possess about 60% of their initial performance after a fortnight.

(4)

Here, the contribution of Rseries was neglected. From plots of dI/dV versus V (Figure 6c), Rshunt could be estimated from the reciprocal of the ordinate at V = 0. The other term, namely (q/ nKBT)I0, could be neglected due to a low value of I0. The values of Rseries, Rshunt, I0, and n, as obtained from the above analysis, are listed in Table 2. The values show that Rseries was the lowest and Rshunt was the highest in the Cu2O/MAPbI3/PCBM heterojunction as compared to the corresponding resistances of the devices based on the other two hole-transport materials. The improved fill factor strongly indicates a reduction in recombination loss as a result of Cu2O incorporation. However, the value is still lower than the reported results and thus may have limited the η of the devices. The value of I0 was at least 2 orders of magnitude lower in the Cu2O/MAPbI3/PCBM heterojunction. The ideality factor was also substantially lower in the Cu2O based devices. The junction properties are hence favorable in the Cu2O/MAPbI3/PCBM heterojunction, and the analyses are in concurrence with the superior solar cell parameters that we observed in this device. In a heterojunction solar cell, the value of the ideality factor represents junction quality and the carrier recombination mechanism. For such a solar cell, the ideality factor (n) is typically between 1 and 2. The value of n approaches unity when the diode current of a pn junction is dominated by carrier diffusion in the neutral region of the semiconductors. When carrier recombination in the depleted region dominates the diode current, the ideality factor approaches 2. In most cases, both phenomena take place simultaneously and hence n lies between 1 and 2. In our efforts, the ideality factor was higher than 2 in all the devices. Such a value meant that recombination of trapped charges was prevalent in those devices.35 From eq 1, VOC of the devices could be derived as VOC =

nKBT ⎛ ISC ⎞ ln⎜ ⎟ q ⎝ I0 ⎠

3. CONCLUSION In conclusion, we have introduced Cu2O as a hole-transport material in perovskite solar cells having a direct p−i−n structure. When we determined the band edges of Cu2O and the perovskite (MAPbI3) and the energy levels of PCBM from dI/dV spectra that have correspondence to the material’s DOS, we observed that p−i and i−n interfaces formed type II band alignment, enabling charge separation at both of them. Due to such an appropriate band diagram, the Cu2O/MAPbI3/PCBM heterojunction when sandwiched between ITO and aluminum electrodes yielded an efficiency of 8.23%. The value excelled when compared to heterojunctions based on other HTLs, such as NiO and copper-doped NiO. In the rest of the paper, we present our results on diode parameters. Such parameters showed that the higher efficiency in Cu2O/MAPbI3/PCBM devices was due to low series resistance, high shunt resistance, and low ideality factor, leading to a low recombination loss as a result of Cu2O incorporation in a direct-structured perovskite solar cell.

(5)

With the values of n and I0 being available, we went on to evaluate VOC of the devices, and these values are listed in Table 2. The values matched reasonably well with VOC values observed in the I−V characteristics under illumination, implying correctness of the device parameters, such as Rseries, Rshunt, I0, and n. 2.5. Hysteresis in Current−Voltage Characteristics. Perovskite solar cells having mesoscopic36 or planar junction37 structure have both been reported to show anomalous hysteresis in their I−V characteristics, making it difficult to accurately represent the device performance. The hysteresis behavior has been reported to depend on the voltage sweep direction, sweep speed or delay time, and light soaking of the devices.38 These effects led to either over- or underestimation of the device performance. To ensure a high degree of accuracy of our I−V measurements, we characterized the devices under two voltage sweep directions and at various sweep rates or delay times. As can be seen in Figure 7a,b, the Cu2O-based device exhibited a not-so-significant amount of hysteresis in the I−V characteristics under illumination conditions. The deviations in ISC and VOC with a range of delay times were

4. EXPERIMENTAL DETAILS 4.1. Materials. Methylamine (MA, CH3NH2), hydrogen iodide (HI), anhydrous lead iodide (PbI2) and lead chloride (PbCl2), N,N-dimethylformamide, anhydrous, 99.8% (DMF), copper(II) chloride dihydrate (CuCl2·2H2O), and chlorobenzene (99.5%) were purchased from Sigma−Aldrich Chemical Co. Phenyl-C61-butyric acid methyl ester (PCBM) (99%) was purchased from M/s SES Research, Houston, TX. 4.2. Synthesis of Perovskites. We synthesized CH3NH3PbI3 and CH3NH3PbI3−xClx following a reported route that contained a couple of steps. 23,27 At first, methylammonium iodide (CH3NH3I) was synthesized by reacting methylamine (CH3NH2) (13.5 mL, 40 wt % in aqueous solution) and hydroiodic acid (15.0 mL, 57 wt % in water) in an ice bath for 2 h. When the reaction that produced CH3NH3I in the form of white powder was complete, the solvent was evaporated by use of a rotary evaporator at 50 °C for 1 h. The precipitate was then collected, washed with diethyl ether (99.7% Merck) three times, dried at 60 °C in a nitrogen1434


Article


Information. Thickness of NiO and Cu@NiO thin films was around 40 nm each. Cu2O thin films were grown on ITO substrates by the SILAR technique, in which highly pure copper(II) chloride dihydrate (CuCl2·2H2O), ammonium hydroxide (NH4OH), and hydrogen peroxide (H2O2) were used as copper precursor, complexing agent, and oxidizing agent, respectively. Dehydrated methanol acted as a solvent during the film formation process. The cationic precursor was prepared by dissolving 0.1 M of the copper salt (∼341 mg) in 20 mL of methanol, followed by complexation with 25% NH4OH at pH 11.8. In doing so, the light green homogeneous solution of the copper salt turned deep blue. The film formation process was carried out at room temperature in an automated multivessel dipcoating unit (Apex Instruments Co. Pvt. Ltd.). The substrates were dipped in (i) copper precursor, (ii) methanol, (iii) H2O2 solution (1%, diluted in methanol) for 10 s to oxidize the adsorbed copper ions that could be inferred from formation of bubbles, and (iv) methanol, all kept in separate beakers, in sequence for the growth of Cu2O films. The steps were cycled to obtain a desired thickness of Cu2O thin films (40 nm), which were finally annealed at 170 °C for 20 min. Following a transfer of HTL-coated ITO substrates to the glovebox, thin films of the perovskites were spun on such substrates at 2500 rpm for 60 s. The films were left in the glovebox for 30 min for slow evaporation of the solvent, preceded by annealing at 100 °C for 30 min (45 min for CH3NH3PbI3−xClx) to promote crystallization of the perovskites. Thickness of CH3NH3PbI3 and CH3NH3PbI3−xClx layers was 270 nm for both. As the n-type ETL layer, PCBM was spun from a 17 mg/mL chlorobenzene solution followed by further annealing at 100 °C for 15 min. Thickness of the ETL was 15 nm, which was measured by forming its separate film on a substrate. The device fabrication process was completed by thermal evaporation of aluminum at a pressure below 5 × 10−6 Torr as the top electrode. Here, the evaporation chamber was attached to the glovebox, allowing easy transfer of the films from the glovebox to the chamber. Thermal evaporation through a shadow mask was carried out at 0.2 Å/s until a thickness of 100 nm was achieved. With the widths of ITO and Al strips being 1/8 in. each and orthogonal to each other, this completed the device fabrication process of four cells, each one having an area of 10 mm2, on a single glass substrate. 4.5. Device Characterization. Current−voltage (I−V) characteristics of the devices under dark and illumination conditions were recorded with a Keithley 2636 electrometer using LabTracer software. Electrodes of the devices, which were kept in the glovebox, were connected to the electrometer via pressure-loaded spring contacts. A 300 W solar simulator (Newport-Stratfort model 76500) attached with an AM1.5 filter placed outside the glovebox acted as a source for illumination. Intensity of the simulated solar light on the device was 100 mW/cm2. While I−V characteristics were recorded under illumination conditions, regions outside the cell area were masked to avoid any contribution from neighboring areas or cells. Bias was swept toward both directions and also in loops. To study stability of the devices, the cells, shelved in the glovebox, were characterized once per day for a fortnight. The cells were neither encapsulated nor protected from exposure to usual room light.

filled glovebox overnight, and finally stored in the glovebox for further use. To prepare thin films of CH3NH3PbI3 and CH3NH3PbI3−xClx, a one-step process was followed in both cases. A solution of CH3NH3PbI3 was prepared by mixing CH3NH3I and PbI2 at 1:1 equimolar ratio in DMF. The solution was allowed to stir overnight at 70 °C in the glovebox that had oxygen and moisture levels below 0.1 ppm each to obtain a yellow homogeneous and transparent solution of CH3NH3PbI3 perovskite. To form CH3NH3PbI3−xClx, the reactants, composed of 3:1 molar ratio of CH3NH3I and PbCl2 in DMF, underwent the same reaction protocol in the glovebox. PbI 2 + CH3NH3I → CH3NH3PbI3 PbCl 2 + 3CH3NH3I → CH3NH3PbI3 − xCl x + CH3NH3Cl

Upon heating, methylamine hydrochloride decomposed into hydrochloric acid and methylamine as CH3NH3Cl → HCl( ↑ ) + CH3NH 2( ↑ )

4.3. Characterization of Materials. The materials were characterized by optical absorption and photoluminescence (PL) spectroscopy, X-ray diffraction (XRD) patterns, and Raman spectroscopy. The measurements were carried out with a Shimadzu UV-2550 spectrophotometer and Spex Fluoromax 4P emission spectrophotometer, Bruker D8 advanced X-ray powder diffractometer (Cu Kα radiation, λ = 1.54 Å), and a Horiba Jobin-Yvon triple-grating Raman spectrometer system (model number T64000) using 514 nm excitation of an argonion laser source (Stabilite 2017, Spectra Physics), respectively. In addition, the materials were characterized with a scanning tunneling microscope (STM). We used a Nanosurf Easyscan2 STM in ambient conditions to record tunneling current through ultrathin films of Cu2O and other HTLs, perovskites, and PCBM. From the tunneling current, we determined differential conductance (dI/dV) that is known to have a correspondence to the density of states (DOS) of semiconductor materials.41 This enabled us to locate conduction and valence band edges (CB and VB, respectively) of the inorganic semiconductors with respective to their Fermi energy. For PCBM, the dI/dV spectrum located its highest occupied molecular orbitals (HOMO) and lowest unoccupied molecular orbitals (LUMO), respectively. For STM measurements, tip voltage was applied with respect to the substrate. At a positive voltage, the tip therefore withdrew electrons from the semiconductor; the peaks in the positive voltage of DOS spectrum hence denoted the VB (or HOMO) of the semiconductors. Similarly, the peaks at negative voltages, at which electrons could be injected from the tip to a semiconductor, provided the location of the CB (or LUMO). 4.4. Device Fabrication. All the planar perovskite solar cells were built on patterned ITO-coated glass substrates (sheet resistance = 15 Ω/square). Prior to deposition of a HTL, the substrates were cleaned following a usual protocol, followed by UV−ozone treatment for 20 min that ensured better contact with the layer. NiO and copper-doped NiO (5 at. %, Cu@NiO) layers were formed via spin-coating from their respective solutions. Both pristine and copper-doped NiO thin films were also formed by spin-coating their corresponding hydroxide solutions at 3000 rpm for 30 s, followed by annealing at 425 °C for 15 min. The reported method followed to form NiO and Cu@NiO layers31,32 is also briefly stated in Supporting 1435


Article


■

(10) Subbiah, A. S.; Halder, A.; Ghosh, S.; Mahuli, N.; Hodes, G.; Sarkar, S. K. Inorganic Hole Conducting Layers for Perovskite-Based Solar Cells. J. Phys. Chem. Lett. 2014, 5, 1748−1753. (11) Ye, S. Y.; Sun, W. H.; Li, Y. L.; Yan, W. B.; Peng, H. T.; Bian, Z. Q.; Liu, Z. W.; Huang, C. H. CuSCN-Based Inverted Planar Perovskite Solar Cell with an Average PCE of 15.6%. Nano Lett. 2015, 15, 3723− 3728. (12) Chiang, C. H.; Tseng, Z. L.; Wu, C. G. Planar Heterojunction Perovskite/PC71BM Solar Cells with Enhanced Open-Circuit Voltage via a (2/1)-Step Spin-Coating Process. J. Mater. Chem. A 2014, 2, 15897−15903. (13) Hu, L.; Peng, J.; Wang, W. W.; Xia, Z.; Yuan, J. Y.; Lu, J. L.; Huang, X. D.; Ma, W. L.; Song, H. B.; Chen, W.; et al. Sequential Deposition of CH3NH3PbI3 on Planar NiO Film for Efficient Planar Perovskite Solar Cells. ACS Photonics 2014, 1, 547−553. (14) Kim, J. H.; Liang, P. W.; Williams, S. T.; Cho, N.; Chueh, C. C.; Glaz, M. S.; Ginger, D. S.; Jen, A. K. Y. High-Performance and Environmentally Stable Planar Heterojunction Perovskite Solar Cells Based on a Solution-Processed Copper-Doped Nickel Oxide HoleTransporting Layer. Adv. Mater. 2015, 27, 695−701. (15) Pathan, H. M.; Lokhande, C. D. Deposition of Metal Chalcogenide Thin Films by Successive Ionic Layer Adsorption and Reaction (SILAR) Method. Bull. Mater. Sci. 2004, 27, 85−111. (16) Rafea, M. A.; Roushdy, N. Determination of the Optical Band Gap for Amorphous and Nanocrystalline Copper Oxide Thin Films Prepared by SILAR Technique. J. Phys. D: Appl. Phys. 2009, 42, No. 015413. (17) Zuo, C. T.; Ding, L. M. Solution-Processed Cu2O and CuO as Hole Transport Materials for Efficient Perovskite Solar Cells. Small 2015, 11, 5528−5532. (18) Malerba, C.; Biccari, F.; Ricardo, C. L. A.; D’Incau, M.; Scardi, P.; Mittiga, A. Absorption Coefficient of Bulk and Thin Film Cu2O. Sol. Energy Mater. Sol. Cells 2011, 95, 2848−2854. (19) Musa, A. O.; Akomolafe, T.; Carter, M. J. Production of Cuprous Oxide, A Solar Cell Material, by Thermal Oxidation and A Study of its Physical and Electrical Properties. Sol. Energy Mater. Sol. Cells 1998, 51, 305−316. (20) Olsen, L. C.; Addis, F. W.; Miller, W. Experimental and Theoretical-Studies of Cu2O Solar-Cells. Sol. Cells 1982, 7, 247−279. (21) Ito, T.; Yamaguchi, H.; Okabe, K.; Masumi, T. Single-Crystal Growth and Characterization of Cu2O and CuO. J. Mater. Sci. 1998, 33, 3555−3566. (22) Malinkiewicz, O.; Yella, A.; Lee, Y. H.; Espallargas, G. M.; Graetzel, M.; Nazeeruddin, M. K.; Bolink, H. J. Perovskite Solar Cells Employing Organic Charge-Transport Layers. Nat. Photonics 2014, 8, 128−132. (23) Kim, H. S.; Lee, C. R.; Im, J. H.; Lee, K. B.; Moehl, T.; Marchioro, A.; Moon, S. J.; Humphry-Baker, R.; Yum, J. H.; Moser, J. E.; et al. Lead Iodide Perovskite Sensitized All-Solid-State Submicron Thin Film Mesoscopic Solar Cell with Efficiency Exceeding 9%. Sci. Rep. 2012, 2, No. 591. (24) Xing, G. C.; Mathews, N.; Sun, S. Y.; Lim, S. S.; Lam, Y. M.; Gratzel, M.; Mhaisalkar, S.; Sum, T. C. Long-Range Balanced Electronand Hole-Transport Lengths in Organic-Inorganic CH3NH3PbI3. Science 2013, 342, 344−347. (25) Manser, J. S.; Kamat, P. V. Band Filling with Free Charge Carriers in Organonietal Halide Perovskites. Nat. Photonics 2014, 8, 737−743. (26) Tauc, J. Optical Properties and Electronic Structure of Amorphous. Mater. Res. Bull. 1968, 3, 37−46. (27) Zhao, Y. X.; Zhu, K. Charge Transport and Recombination in Perovskite (CH3NH3)PbI3 Sensitized TiO2 Solar Cells. J. Phys. Chem. Lett. 2013, 4, 2880−2884. (28) He, M. H.; Chen, Y.; Liu, H.; Wang, J. L.; Fang, X. S.; Liang, Z. Q. Chemical Decoration of CH3NH3PbI3 Perovskites with Graphene Oxides for Photodetector Applications. Chem. Commun. 2015, 51, 9659−9661. (29) Quarti, C.; Grancini, G.; Mosconi, E.; Bruno, P.; Ball, J. M.; Lee, M. M.; Snaith, H. J.; Petrozza, A.; De Angelis, F. The Raman Spectrum

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b11540. Additional text describing growth of NiO and Cu@NiO thin films and seven figures showing optical absorption and XRD of CH3NH3PbI3−xClx perovskites; EDX of perovskites; optical absorption, Tauc plots, and XRD of NiO and Cu@NiO thin films and of Cu2O thin film formed through SILAR technique; energy band diagrams of NiO/MAPbI 3 /PCBM and Cu@NiO/MAPbI 3 / PCBM; histogram of VOC, ISC, η, and fill factor of Cu2O/MAPbI3/PCBM solar cells; and surface morphologies of perovskite films (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Telephone +91-33-24734971; fax +91-33-24732805; e-mail [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS SC acknowledges DST INSPIRE Fellowship [IF 140158]. The authors acknowledge financial assistance from DeitY, SERIIUS, and Nano Mission (DST) projects.

■

REFERENCES

(1) Liu, M. Z.; Johnston, M. B.; Snaith, H. J. Efficient Planar Heterojunction Perovskite Solar Cells by Vapour Deposition. Nature 2013, 501, 395−398. (2) Christians, J. A.; Fung, R. C. M.; Kamat, P. V. An Inorganic Hole Conductor for Organo-Lead Halide Perovskite Solar Cells. Improved Hole Conductivity with Copper Iodide. J. Am. Chem. Soc. 2014, 136, 758−764. (3) Qin, P.; Tanaka, S.; Ito, S.; Tetreault, N.; Manabe, K.; Nishino, H.; Nazeeruddin, M. K.; Gratzel, M. Inorganic Hole Conductor-Based Lead Halide Perovskite Solar Cells with 12.4% Conversion Efficiency. Nat. Commun. 2014, 5, No. 3834. (4) Liu, D. Y.; Kelly, T. L. Perovskite Solar Cells with a Planar Heterojunction Structure Prepared using Room-Temperature Solution Processing Techniques. Nat. Photonics 2014, 8, 133−138. (5) Kumar, C. V.; Sfyri, G.; Raptis, D.; Stathatos, E.; Lianos, P. Perovskite Solar Cell with Low Cost Cu-Phthalocyanine as Hole Transporting Material. RSC Adv. 2015, 5, 3786−3791. (6) Giorgi, G.; Fujisawa, J. I.; Segawa, H.; Yamashita, K. Small Photocarrier Effective Masses Featuring Ambipolar Transport in Methylammonium Lead Iodide Perovskite: A Density Functional Analysis. J. Phys. Chem. Lett. 2013, 4, 4213−4216. (7) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.; Alcocer, M. J. P.; Leijtens, T.; Herz, L. M.; Petrozza, A.; Snaith, H. J. Electron-Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber. Science 2013, 342, 341− 344. (8) Qiu, J. H.; Qiu, Y. C.; Yan, K. Y.; Zhong, M.; Mu, C.; Yan, H.; Yang, S. H. All-Solid-State Hybrid Solar Cells based on a New Organometal Halide Perovskite Sensitizer and One-Dimensional TiO2 Nanowire Arrays. Nanoscale 2013, 5, 3245−3248. (9) Wang, H.; Sheikh, A. D.; Feng, Q. Y.; Li, F.; Chen, Y.; Yu, W. L.; Alarousu, E.; Ma, C.; Haque, M. A.; Shi, D.; et al. Facile Synthesis and High Performance of a New Carbazole-Based Hole-Transporting Material for Hybrid Perovskite Solar Cells. ACS Photonics 2015, 2, 849−855. 1436


Article

The Journal of Physical Chemistry C of the CH3NH3Pbl3 Hybrid Perovskite: Interplay of Theory and Experiment. J. Phys. Chem. Lett. 2014, 5, 279−284. (30) Dar, M. I.; Arora, N.; Gao, P.; Ahmad, S.; Gratzel, M.; Nazeeruddin, M. K. Investigation Regarding the Role of Chloride in Organic-Inorganic Halide Perovskites Obtained from Chloride Containing Precursors. Nano Lett. 2014, 14, 6991−6996. (31) Mashford, B. S.; Nguyen, T. L.; Wilson, G. J.; Mulvaney, P. AllInorganic Quantum-Dot Light-Emitting Devices Formed via LowCost, Wet-Chemical Processing. J. Mater. Chem. 2010, 20, 167−172. (32) Kim, K. H.; Takahashi, C.; Abe, Y.; Kawamura, M. Effects of Cu Doping on Nickel Oxide Thin Film Prepared by Sol-Gel Solution Process. Optik 2014, 125, 2899−2901. (33) Docampo, P.; Ball, J. M.; Darwich, M.; Eperon, G. E.; Snaith, H. J. Efficient Organometal Trihalide Perovskite Planar-Heterojunction Solar Cells on Flexible Polymer Substrates. Nat. Commun. 2013, 4, No. 2761. (34) Figueiredo, V.; Elangovan, E.; Goncalves, G.; Barquinha, P.; Pereira, L.; Franco, N.; Alves, E.; Martins, R.; Fortunato, E. Effect of Post-Annealing on the Properties of Copper Oxide Thin Films Obtained from the Oxidation of Evaporated Metallic Copper. Appl. Surf. Sci. 2008, 254, 3949−3954. (35) Hegedus, S. S.; Shafarman, W. N. Thin-Film Solar Cells: Device Measurements and Analysis. Prog. Photovoltaics 2004, 12, 155−176. (36) Kim, H. S.; Park, N. G. Parameters Affecting I-V Hysteresis of CH3NH3PbI3 Perovskite Solar Cells: Effects of Perovskite Crystal Size and Mesoporous TiO2 Layer. J. Phys. Chem. Lett. 2014, 5, 2927−2934. (37) Snaith, H. J.; Abate, A.; Ball, J. M.; Eperon, G. E.; Leijtens, T.; Noel, N. K.; Stranks, S. D.; Wang, J. T. W.; Wojciechowski, K.; Zhang, W. Anomalous Hysteresis in Perovskite Solar Cells. J. Phys. Chem. Lett. 2014, 5, 1511−1515. (38) Unger, E. L.; Hoke, E. T.; Bailie, C. D.; Nguyen, W. H.; Bowring, A. R.; Heumuller, T.; Christoforo, M. G.; McGehee, M. D. Hysteresis and Transient Behavior in Current-Voltage Measurements of Hybrid-Perovskite Absorber Solar Cells. Energy Environ. Sci. 2014, 7, 3690−3698. (39) Hwang, I.; Jeong, I.; Lee, J.; Ko, M. J.; Yong, K. Enhancing Stability of Perovskite Solar Cells to Moisture by the Facile Hydrophobic Passivation. ACS Appl. Mater. Interfaces 2015, 7, 17330−17336. (40) Jung, M. C.; Raga, S. R.; Ono, L. K.; Qi, Y. B. Substantial Improvement of Perovskite Solar Cells Stability by Pinhole-Free Hole Transport Layer with Doping Engineering. Sci. Rep. 2015, 5, No. 9863. (41) Mocatta, D.; Cohen, G.; Schattner, J.; Millo, O.; Rabani, E.; Banin, U. Heavily Doped Semiconductor Nanocrystal Quantum Dots. Science 2011, 332, 77−81.

1437
