Enhanced photovoltage for inverted planar heterojunction perovskite solar cells

See allHide authors and affiliations

Science  29 Jun 2018:
Vol. 360, Issue 6396, pp. 1442-1446
DOI: 10.1126/science.aap9282

Perovskite layers make the grade

Inverted planar perovskite solar cells offer opportunities for a simplified device structure compared with conventional mesoporous titanium oxide interlayers. However, their lower open-circuit voltages result in lower power conversion efficiencies. Using mixed-cation lead mixed-halide perovskite and a solution-processed secondary growth method, Luo et al. created a surface region in the perovskite film that inhibited nonradiative charge-carrier recombination. This kind of solar cell had comparable performance to that of conventional cells.

Science, this issue p. 1442


The highest power conversion efficiencies (PCEs) reported for perovskite solar cells (PSCs) with inverted planar structures are still inferior to those of PSCs with regular structures, mainly because of lower open-circuit voltages (Voc). Here we report a strategy to reduce nonradiative recombination for the inverted devices, based on a simple solution-processed secondary growth technique. This approach produces a wider bandgap top layer and a more n-type perovskite film, which mitigates nonradiative recombination, leading to an increase in Voc by up to 100 millivolts. We achieved a high Voc of 1.21 volts without sacrificing photocurrent, corresponding to a voltage deficit of 0.41 volts at a bandgap of 1.62 electron volts. This improvement led to a stabilized power output approaching 21% at the maximum power point.

Recently, perovskite solar cells (PSCs) with inverted planar heterojunction structures, wherein a polycrystalline perovskite film is sandwiched between a hole- and an electron-extraction layer, have gained attention because they offer the promise of easy fabrication, compatibility with flexible substrates, versatility of energy-band engineering, and the possibility of fabricating multijunction cells (14); moreover, they have achieved power conversion efficiencies (PCEs) exceeding 20%. Further enhancement of their PCEs is now mainly hampered by the relatively low open-circuit voltages (Voc), typically <1.10 V, in comparison with >1.20 V reported for regular PSCs using similar bandgap perovskites (2, 5). The less-than-ideal Voc of inverted PSCs is attributed to the nonradiative recombination losses both inside the perovskite bulk material and at the interfacial contacts (68), owing to the presence of a considerable density of defects or recombination centers (915). Critically, although “isolated” perovskite films can exhibit very high radiative efficiencies, the radiative efficiency drops considerably when perovskite films contact the charge-extraction layers (16, 17). Several approaches to reducing the nonradiative recombination—including increasing the grain size (18, 19), surface passivation (3), ion compensation (11), and heterojunction engineering (20, 21)—have been proposed, pushing the Voc up to 1.15 V in inverted PSCs. In addition, the precise origin of the higher Voc for regular PSCs, which often include a thin mesoporous scaffold layer, is not clear. Evidence suggests that the perovskite film crystallized within this “scaffold region” may be more n-type in nature, leading to a favorable n-type contact at the perovskite n-type charge-extraction region (2224). However, the perovskite film crystallized on a p-type substrate has been shown to be more p-type (or less n-type) in nature at this contact interface (25). These results suggest that the nature of the perovskite film, and specifically its doping state (n, p, or intrinsic) near the charge-extraction layers, are strongly influenced by the polarity of the substrates on which the films are crystallized. Consequently, controlling the nature of the perovskite film at the contact interface may be an effective way to achieve high Voc in inverted PSCs. Here we introduce a means to deliver a wider bandgap region near the top surface of the film and a more n-type perovskite film by using a solution-processed secondary growth (SSG) technique, leading to a substantial increase in Voc.

The SSG procedure includes two steps: (i) the preparation of perovskite films by solution processing and (ii) the secondary growth with the assistance of guanidinium bromide (which we abbreviate SSG-G). We started with the preparation of a mixed-cation lead mixed-halide perovskite layer using a nonstoichiometric recipe of (FA0.95PbI2.95)0.85(MAPbBr3)0.15 (14, 26), where MA and FA denote methylammonium and formamidinium, respectively. In Fig. 1A, we show the scanning electron microscopy (SEM) image of the control perovskite film. The bright crystals on the surface indicate regions of higher electron density, or regions that accumulated charges during the measurement. From energy-dispersive x-ray (EDX) analysis (fig. S1), we determined that the lead halide complex mainly dominated the composition of the bright crystals in the control sample, absent of carbon, so we assigned it as PbI1.50Br0.50. We further confirmed this through x-ray diffraction (XRD) analysis of the control sample (Fig. 1B), in which we assigned the diffraction peak (2θ = 13.1°, where θ is the work angle between the x-ray beam and the plane of the sample surface) to PbX2 (X is a mixture of I and Br at a ~1.5:0.5 ratio), using the standard diffraction peaks of PbI2 at 2θ = 12.67° and PbBr2 at 2θ = 14.37° as the references. Previously, the presence of the PbX2 diffraction peak has been interpreted as indicating the presence of a PbX2 “shell” surrounding the perovskite grains (26, 27). In contrast, our XRD analysis shows that the PbX2 constitutes entirely separate grains. From the Sheerer broadening of the PbX2 peak, we estimated an average crystal grain size of 42 nm, which was consistent with the size of the bright grains observed in the SEM images. Through SSG-G, the excess PbX2 crystals were “digested,” and the perovskite film displayed distinct surface morphologies (Fig. 1A and fig. S2). In the XRD patterns (Fig. 1B), we observed considerably reduced PbX2 diffraction peak intensity.

Fig. 1 Improved morphologies and crystal structures.

(A) Top-view SEM images of the control and SSG-G films. Scale bar, 2 μm. (B) XRD patterns (α, δ, and the black square denote the identified diffraction peaks corresponding to the black perovskite phase, the nonperovskite phase, and PbI1.50Br0.50, respectively). a.u., arbitrary units.

We evaluated the photovoltaic performances of the perovskite films with and without the SSG-G process by fabricating inverted planar heterojunction PSCs {device structure: indium tin oxide (ITO)/poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA)/perovskite/[6,6]-phenyl-C61-butyric acid methyl ester (PC61BM)/buckminsterfullerene (C60)/bathocuproine (BCP)/copper (Cu)} and measuring the current density-voltage (J-V) curves under simulated AM (air mass) 1.5G (global) illumination at 100 mW cm−2. We systematically optimized the processing parameters for the SSG technique, variations of which we show in tables S1 to S5. All the devices exhibited negligible hysteresis (table S6). In Fig. 2A, we present J-V curves of the PSCs obtained from reverse scans. We found that the SSG-G devices delivered increased Voc by up to 100 mV compared with the control devices, without compromising short-circuit current density (Jsc) and fill factor (FF). The champion SSG-G device showed a PCE of 21.51% and a stabilized power output (SPO) of 20.91% (Fig. 2B). We integrated the external quantum efficiency (EQE) spectra over the AM 1.5G solar spectrum, and the resulting Jsc values were in close agreement with the values determined from the J-V scans (fig. S3). We fabricated 200 devices from different batches, and we present the histograms of average Voc values in Fig. 2C (we show histograms of the corresponding PCEs in fig. S4). The average Voc of the control devices was about 1.10 V, whereas it was about 1.20 V for the SSG-G devices, with a record value of 1.21 V (Fig. 2C). The “stabilized photovoltage” values that we measured under continuous 1-Sun illumination were consistent with the Voc values obtained from the scanned J-V curves (Fig. 2D). To confirm our in-house device efficiency measurements, we sent one of our non-encapsulated devices to the National Institute of Metrology, China, for external certification and obtained a PCE of 20.90% (Voc = 1.175 V, Jsc = 21.86 mA cm−2, and FF = 81.37%) (fig. S5). A Voc of 1.21 V rivals the values reported for regular PSCs with similar bandgaps, corresponding to a voltage deficit of 0.41 V. We compare our results with other published results for inverted PSCs in fig. S6 and table S7. The SSG-G process is also applicable to devices with varied electron-extraction layers, buffer layers, and metal electrodes (fig. S7 and table S8).

Fig. 2 Photovoltaic performances.

(A) The J-V curves of the inverted planar heterojunction PSCs, obtained from reverse scans under simulated AM 1.5G illumination at 100 mW cm−2. The inset shows a schematic diagram of the solar cell used in this work. (B) The stabilized power output (SPO) for the champion device and the control device. The SPO at the maximum power point is indicated. (C) Histograms of the Voc for 200 control and 200 SSG-G devices. (D) Stabilized photovoltage as a function of illumination time at the open-circuit conditions for the champion device and the control device.

To assess the impact of the SSG-G process on the long-term stability of the PSCs, we first tested the thermal stability of non-encapsulated devices in a nitrogen atmosphere. The SSG-G device showed a slight decay (~5%) in PCE after aging for 500 hours at 85°C (fig. S8). In contrast, the control device degraded more substantially to a PCE of <60% of its original value, indicating that the SSG-G process improved the thermal stability of the device. We further examined the operational stability of non-encapsulated devices in a nitrogen atmosphere by comparing the PCE decay of the control and the SSG-G devices aged under a xenon lamp–based simulator with an ultraviolet component (100 mW cm−2) at room temperature (Fig. 2B). The SPO of the control device showed a fast decay initially and then a subsequent decay at a relatively slow rate. In contrast, we did not see an obvious decay in the SPO of the SSG-G device over the measurement time. These results indicate that the positive impact of the SSG-G process is stable for an operational device and that it improves the overall stability of the device.

We investigated the changes in the bandgap, microstrain, and electronic disorder induced by the SSG-G process (figs. S9 and S10) and observed some improvements, including a reduction in the Urbach energy (described in the supplementary materials). However, by calculating the Voc in the radiative limit (28), we confirmed that the reduced electronic disorder within the SSG-G film only contributed a few millivolts to the enhancement in Voc (fig. S11).

To uncover the origin of the increase in Voc, we performed surface photovoltage (SPV), ultraviolet photoelectron spectroscopy (UPS), and photoluminescence quantum yield (PLQY) measurements on a series of samples. We performed SPV measurements with a Kelvin probe on perovskite films coated on ITO (structure 1) or ITO/PTAA (structure 2) and on close-to-complete cells consisting of ITO/PTAA/perovskite/PC61BM (structure 3) (Fig. 3A). A >100-meV shift in surface potential occurred for all films treated with the SSG-G process (compared with the control films), regardless of structure or illumination intensity. This shift was in the right direction to result in a ~100-mV increase in Voc in a cell.

Fig. 3 Surface photovoltage and ultraviolet photoelectron spectroscopy.

(A) Surface photovoltage measurements of the perovskite films coated on ITO (structure 1) or ITO/PTAA (structure 2) and of close-to-complete cells consisting of ITO/PTAA/perovskite/PC61BM (structure 3). (B) Helium Iα (hν = 21.22 eV) spectra of secondary electron cutoff and valence band of control and SSG-G perovskite thin films. The blue-green vertical lines indicate the valence band maximum (Ev) with respect to the Fermi level. (C) Schematic energy-level diagrams of control and SSG-G films.

We further investigated the electronic structures of the perovskite films by using UPS; we show the data for the control and the SSG-G perovskite films in Fig. 3B. From the secondary electron cutoff spectrum, we observed a change in work function from 5.02 to 4.42 eV between the control and the SSG-G film. This is consistent with the direction of change, although slightly smaller in magnitude, that we determined from the SPV measurements. The valence band maximum (Ev) with respect to the Fermi level (EF) in the SSG-G film was also shifted relative to the control film (the absolute value of Ev can be calculated from the relationship shown in fig. S12). From these measurements, we built energy-level diagrams for the two films (Fig. 3C). From the control to the SSG-G film, the Ev shifts by about 200 meV toward vacuum, and the EF shifts by a further 400 meV toward the conduction band minimum (Ec). The EF shift indicates a more n-type nature for the SSG-G film, resulting from a surface and/or a bulk effect. The absolute shift in energy levels toward vacuum in the SSG-G film is likely to originate from a change in the ratio of lead halide– to organic halide–terminated surfaces. Recently, lead iodide termination has been shown to lead to deeper energy levels than ammonium iodide termination (29). Because we have added additional guanidinium bromide as a secondary process, we would expect a larger fraction of the surface to be terminated with organic cation halide.

Therefore, combining the SPV and UPS analyses, we obtained consistent results indicating a more n-type perovskite film produced by the SSG-G process, with the energy levels shifted toward vacuum. We characterized the uniformity of the SSG-G film by Kelvin probe force microscopy (KPFM) mapping (fig. S13) and found variation in surface potential similar to that of other reported conventional perovskite films (30).

To investigate the influence of the SSG-G process on the radiative recombination at the perovskite charge-extraction layer heterojunctions, we measured the PLQY for the films with various structures (Table 1). Isolated SSG-G films on quartz exhibited a higher PLQY of ~2.8%, compared with the control films at ~0.17%. However, on ITO substrates, which are the semitransparent electrodes used in the devices, the SSG-G films exhibited PLQY values similar to those of the control films. After the introduction of PTAA, the PLQYs of the SSG-G films were 10 times those of the control films, reaching an average value of 8.9%. The PLQY values were still >2.5% when the PTAA and PC61BM layers simultaneously contacted the SSG-G films, whereas the PLQY for the control film was quenched to an undetectably low level (<0.1%). The radiative efficiencies of the perovskite films are influenced by the nature of the underlying substrate on which they are crystallized (31). Therefore, the factors that influence the PLQY are the “quality” of the as-crystallized perovskite films, the extent to which the substrate or subsequent layer extracts charges, and the degree to which new nonradiative recombination pathways are introduced at the perovskite contacting–layer heterojunctions. Evidently, ITO/PTAA is a good substrate on which to crystallize perovskite films; few nonradiative pathways appear to be introduced by the SSG-G films contacting PTAA, and with the subsequent coating with PC61BM, few additional nonradiative pathways are introduced. As is evident from the high PCEs of the solar cells, these interfaces allow charge extraction.

Table 1 Photoluminescence quantum yield (PLQY).

Summary of PLQY results for the control and SSG-G perovskite films on quartz, ITO, and ITO/PTAA and in close-to-complete cells consisting of ITO/PTAA/perovskite/PC61BM.

View this table:

To corroborate the PLQY results with the improved Voc in the solar cells, we measured the electroluminescence from the PSCs. We observed an electroluminescence spectrum at 778 nm for the SSG-G device (fig. S14A), with an external radiative efficiency (ERE) of 1.2% under a forward bias current density of 23 mA cm−2, but we could not detect electroluminescence from the control device. The ERE value is lower than the PLQY of the perovskite films contacting both PTAA and PC61BM. This is likely due to the isolated film emitting in both directions in the PLQY measurements, whereas the PSCs can only emit in the forward direction, implying more reabsorption of emitted light for the latter. Furthermore, there were likely imperfections in the deposition of the films, resulting in some shunting paths between the top metal electrode and the perovskite film, which would increase dark current and reduce the ERE of the PSCs. We estimated a Voc of 1.21 V on the basis of the measured radiative efficiency (fig. S14B), which is in direct agreement with our device measurements under simulated 1-Sun illumination.

To examine whether the electronic changes to the film are homogeneous throughout the thickness of the perovskite layer or solely present in a surface region, we measured time-resolved photoluminescence (PL) spectra of perovskite films [coated on quartz substrates: quartz glass/perovskite/poly(methyl methacrylate) (PMMA)] and excited the film from the glass and PMMA sides, using a 400-nm laser with a ~90-fs pulse width. We observed a single band emission of the control film when exciting the film from both the PMMA and the glass sides (Fig. 4A). In contrast, for the SSG-G film, we observed a dual peak at an early time (0 ns) when exciting the film from the PMMA (top) side, and this peak shifted from 735 to 767 nm over a few nanoseconds (Fig. 4C). However, when we excited the film from the glass side, we only observed the redshifted band (~767 nm). We further performed transient absorption (TA) measurements by exciting the film with 400-nm (90-fs) pump pulses and probing with broadband visible pulses. For the control film, we did not observe a considerable shift in the ground-state bleach upon exciting from either side (Fig. 4B). In contrast, again, for the SSG-G film, we observed a redshift of the ground-state bleach maxima from 730 to 750 nm over the first few nanoseconds when exciting and probing from the PMMA side (Fig. 4D); we did not observe such a redshift when exciting and probing from the glass side.

Fig. 4 Time-resolved photoluminescence and transient absorption spectra.

(A and C) Photoluminescence spectra of films consisting of quartz glass/perovskite/PMMA at different time scales after excitation, recorded from both the PMMA and glass sides in the reflection geometry, for (A) Control and (C) SSG-G samples. (B and D) Transient absorption spectra of films consisting of quartz glass/perovskite/PMMA in the transmission geometry, obtained by using 400-nm (90-fs) excitation pump pulses from different sides, for (B) control and (D) SSG-G samples.

These time-resolved PL and TA results suggest the presence of a wider bandgap (~80 meV wider) in the SSG-G film, close to the surface. The resulting redshift in the time-resolved PL and TA spectra over time indicates that the charge population migrates from the wider bandgap region to the narrower bandgap bulk. Because we pumped the perovskite films with high-energy photons (400 nm), we expect that the light was absorbed strongly in the top region (1/e absorption depth, ~62 nm), and the initial photoexcited charge population was hot. The time scale that we determined for the transfer of the charge population from the top region to the narrower bandgap bulk region is on the order of a few nanoseconds—much longer than the charge population cooling time, which is expected to occur within 100 fs (32). Therefore, we expect that an equilibrium population of free carriers is created in the topmost region of the film, before charge transfer toward the bulk, most likely driven by diffusion. These results are consistent with establishing a wider bandgap close to the film surface.

In EDX bromide elemental mapping (fig. S15) and time-of-flight secondary ion mass spectroscopy (ToF-SIMS) depth profiles (fig. S16), we observed a bromide-rich region close to the top surface of the SSG-G film. The Br enrichment near the top surface is consistent with a wider bandgap existing in this region of the film. Our studies of the spectroscopy and the stability of the PSCs suggest that the presence of the wider bandgap top layer is stable. Unexpectedly, we observed that the guanidinium (GA+) is well distributed across the perovskite layer. Because GA+ is a large cation, and because in XRD measurements, we observed negligible shifts in characteristic perovskite reflection peaks, we would not expect a considerable fraction of GA+ to be incorporated into the perovskite crystal lattice. Therefore, its location is most likely at the surface and grain boundaries. A substantive contribution to the reduction in nonradiative recombination may therefore be through a passivation role (33), given that GA+ may be present at both top and bottom surfaces and at the grain boundaries.

Multiple factors are thus responsible for the reduced nonradiative recombination. First, the predominant trap species leading to trap-assisted recombination in perovskite films are electron traps (34). The more n-type perovskite film could result in a larger fraction of occupied versus vacant traps and a reduction in the rate of trap-assisted recombination. Secondly, a wider bandgap near the top surface of the SSG-G film might reduce the total electron-hole recombination rate in this region of the film where surface defects are likely to exist, by making it energetically favorable for one or both charge carriers to reside predominantly within the bulk of the film, away from the surface. Thirdly, the presence of guanidinium halide throughout the films and at the surfaces may inhibit unwanted trap-assisted recombination at the heterojunctions by trap passivation. Hence, the SSG-G process that we have presented to improve the perovskite film quality results in considerable improvements to the electronic nature at the heterojunctions between the perovskite film and charge-extraction layers. Our findings should be broadly applicable to PSCs and perovskite light-emitting diodes.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S16

Tables S1 to S8

References (3548)

References And Notes

Acknowledgments: We thank S. Hinder from the Faculty of Engineering and Physical Sciences, University of Surrey (UK), for his kind assistance with ToF-SIMS measurements and helpful discussion; B. Wenger from the University of Oxford (UK) for his assistance with PLQY measurements; S. Mahesh from the University of Oxford (UK) for his help with radiative loss estimation; and P. Li and Z. Lu from the Department of Materials Science and Engineering, University of Toronto (Canada), for helpful suggestions on UPS measurements and analysis. Funding: This work was partly funded by the 973 Program of China (2015CB932203), the National Natural Science Foundation of China (61722501, 91733301, 91433203, and 61377025), EPSRC (UK), the European Union Seventh Framework Programme under grant agreement 604032 of the MESO project, and AFOSR through project FA9550-15-1-0115. W.Z. acknowledges financial support from a Royal Society Research Grant (2017; RG160742), the Royal Society International Exchanges Scheme (2016; IE160511), and the University of Surrey Sustainability Conference Support Programme. A.S. and R.H.F. acknowledge support from EPSRC, Indo-UK APEX, and UKIERI projects. R.Sh. acknowledges a Newton-Bhabha international fellowship. Author contributions: D.L. and R.Z. conceived of the work. D.L. and W.Y. fabricated and characterized solar cells. D.L. and W.Y. conducted UPS, KPFM, electroluminescence, ERE, and EQE measurements. Z.W. conducted SPV measurements and estimated microstrain and the voltage radiative limit. Z.W. and R.Sh. conducted PLQY measurements. A.S. and R.H.F. contributed to the photothermal deflection spectroscopy data. A.S., R.Sh., and R.H.F contributed to time-resolved PL and TA data. Q.G. analyzed and discussed time-resolved PL and TA spectra. Q.H. conducted SEM experiments on the perovskite films, performed SEM analysis, and conducted the XRD measurements. L.Z. and F.Y. analyzed the XRD data. R.Su., T.L., K.C., and D.L. contributed to the certification of solar cells. R.Su. and P.W. conducted thermal stability tests on the devices. Y.T. and J.W. conducted EDX measurements, and L.Z. contributed to the EDX analysis. Y.Z. prepared metal oxide buffer layers. Y.Z. and X.Y. measured ultraviolet–visible light absorption spectra. G.F.T., J.F.W., R.Su., D.L., and W.Z. contributed to the ToF-SIMS measurement and data analysis. Z.X. conducted statistical analysis of device efficiencies. D.L. and Z.W. estimated the voltage deficit from bandgap to voltage. W.Z., H.J.S., and R.Z. directed and supervised the project. D.L. and W.Y. wrote the first draft of the paper. Z.W., W.Z., H.J.S., and R.Z. revised the paper. All authors analyzed their data and reviewed and commented on the paper. Competing interests: H.J.S. is chief scientific officer of Oxford PV, a company commercializing perovskite solar cells. Data and materials availability: All data needed to evaluate the conclusions of the paper are present in the paper or the supplementary materials.

Stay Connected to Science

Navigate This Article