Open-Circuit Voltages Exceeding 1.26 V in Planar Methylammonium

Dec 6, 2018 - Zhifa Liu†‡ , Lisa Krückemeier*† , Benedikt Krogmeier† , Benjamin ... Wang, Huang, Zhao, Sun, Huang, Xue, Lee, Zhu, Huang, Li, ...
0 downloads 0 Views 2MB Size
Download PDF
http://pubs.acs.org/journal/aelccp

Open-Circuit Voltages Exceeding 1.26 V in Planar Methylammonium Lead Iodide Perovskite Solar Cells ACS Energy Lett. Downloaded from pubs.acs.org by AUSTRALIAN NATL UNIV on 12/06/18. For personal use only.

Zhifa Liu,†,‡,# Lisa Krückemeier,*,†,# Benedikt Krogmeier,† Benjamin Klingebiel,† José A. Márquez,§ Sergiu Levcenko,§ Senol Ö z,∥ Sanjay Mathur,∥ Uwe Rau,† Thomas Unold,§ and Thomas Kirchartz*,†,⊥ †

IEK5-Photovoltaics, Forschungszentrum Jülich, 52425 Jülich, Germany State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002, People’s Republic of China § Department of Structure and Dynamics of Energy Materials, Helmholtz-Zentrum-Berlin, Hahn-Meitner-Platz 1, 14109 Berlin, Germany ∥ Institute of Inorganic Chemistry, University of Cologne, Greinstrasse 6, 50939 Cologne, Germany ⊥ Faculty of Engineering and CENIDE, University of Duisburg−Essen, Carl-Benz-Strasse 199, 47057 Duisburg, Germany ‡

S Supporting Information *

ABSTRACT: We demonstrate open-circuit voltages exceeding 1.26 V for inverted planar CH3NH3PbI3 solar cells fabricated using a combination of lead acetate and PbCl2 precursors leading to smooth films and large grain sizes. Surface recombination is suppressed by careful optimization of the PTAA hole transport and PCBM electron transport layers. Suppression of bulk and surface recombination is verified by absolute photoluminescence measurements with external quantum efficiencies of ∼5% in complete cells. In addition, we find exceptionally long photoluminescence lifetimes in full cells and in layer stacks involving one or two contact layers. Numerical simulations reveal that these long photoluminescence lifetimes are only possible with extremely low recombination velocities at the interfaces between absorber and contact materials.

F

that a quasi-Fermi level splitting corresponding to an internal open-circuit voltage of 1.28 V is possible in MAPI films on glass, proving that the bulk material has the potential to come within ∼40 mV of its thermodynamic limit.9 However, the presence of contacts has so far limited the open-circuit voltage Voc achievable in a completed device. The lowest values of Voc loss relative to the radiative limit correspond to the highest values of external luminescence quantum efficiency (Qe,lum) at injection conditions corresponding to 1 sun.10,11 Saliba et al.12 showed record values for Qe,lum ≈ 1% corresponding to voltage losses of about kT/q × ln(100) = 120 mV in Rb-containing perovskites with a band gap slightly above 1.6 eV. Any progress beyond this value toward lower nonradiative recombination

or mature photovoltaic technologies, the biggest challenge when improving their power conversion efficiencies is approaching the radiative limit of the open-circuit voltage by minimizing nonradiative recombination processes in the bulk, at surfaces, and at contact layers.1,2 This statement also holds for the rapidly developing class of lead halide perovskite solar cells. From the beginning of their development, perovskite solar cells have demonstrated astonishingly high open-circuit voltages in relation to their band gap energy. Already the first paper on CH3NH3PbI3 (MAPI) based solar cells with >10% efficiency3 showed an open-circuit voltage of Voc ≈ 1.1 V. Nowadays, photovoltaic power conversion efficiencies of lead halide perovskite solar cells have risen above 23%,4 and open-circuit voltages of perovskites with band gaps of Eg ≈ 1.6 eV are now approaching or sometimes exceeding 1.2 V (with the radiative limit for the Voc in MAPI being ∼1.32 V5−8). It has recently been shown © XXXX American Chemical Society

Received: October 6, 2018 Accepted: November 30, 2018

110

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

Cite This: ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters

and as far as we are aware, a mixture of these two has not been used with the combination of PTAA- and PCBM-based contact materials. This interlayer combination has been shown to have very low voltage losses also with multication-based perovskites.21 However, the combination with our MAPI absorber seems to currently have an even higher potential to achieve high open-circuit voltages. Because high open-circuit voltages require well-passivated surfaces, optimization also had to include the right choice of processing conditions for PCBM. There are various degrees of freedom in processing the PCBM layer that we can use to optimize the MAPI/PCBM interface for the lowest possible recombination rates. These degrees of freedom include, besides the spin-coating parameters, the choice of solvent, postdeposition annealing treatment, and the atmosphere during spin-coating and subsequent annealing. It was reported17 that higher boiling point solvents can lead to more ordered layers because the solvent has more time to evaporate when compared to lower boiling point solvents. In addition, by annealing, preheating of the solution, or solventassisted annealing, crystallization of the absorber layer can be accelerated or slowed down. As part of device optimization, we used various solvents for PCBM, namely, the lower boiling point solvents toluene (T) (110.6 °C) and chlorobenzene (CB) (130 °C) and the higher boiling point (180.5 °C) solvent 1,2-dichlorobenzene (1,2-DCB). Shao et al.22 have previously reported an improvement in Voc in a very similar device stack when using 1,2-DCB as a solvent for PCBM combined with solvent vapor annealing in 1,2-DCB vapor. In contrast, we found here that the PCBM films deposited using the lower boiling point solvent CB or a mixture of CB/T on top of our PTAA/MAPI stack led to substantially higher opencircuit voltages. Note that PCBM is not completely soluble in pure toluene in the desired concentration. During our studies, we also noted that the ideal processing conditions for PCBM did vary depending on the choice of hole transport material and depending on the MAPI processing conditions. If we use 1,2-DCB as the solvent for PCBM, substantial losses in Voc occur and we only achieve open-circuit voltages less than 1.2 V (SI Figure S5b). The opposite trend of Voc with PCBM processing conditions was observed when another HTL layer (PEDOT:PSS) was chosen on which the perovskite crystallizes differently (Figure S2). We therefore assume that the ideal process parameters must always be considered holistically and have to be individually adjusted to each other in order to achieve the best possible results. The process parameters leading to the highest open-circuit voltages on PTAA are described in detail in the experimental section in the SI. Figure 2a shows the current density−voltage (JV) curve of a typical device optimized for high open-circuit voltages (blue line, device A) and a device further optimized for efficiency (red line, device B). All measurements were recorded at a scan speed of 100 mV/s with a class AAA solar simulator under forward (dashed line) and reverse (solid line) scan conditions. The efficiencies are around 18% (device A) and 20% (device B), but the most remarkable feature is in both cases the high open-circuit voltage at Voc ≈ 1.26 V. Figure 2b shows the efficiency and Figure 2c the open-circuit voltage measured as a function of time under white light provided by a LED adjusted to 1 sun conditions, confirming that the open-circuit voltage stabilizes above 1.26 V at an efficiency of ∼18% and ∼20.2%, respectively. Furthermore, as illustrated by the histogram in Figure 2d, this high Voc value is reproducible for a series of

losses requires careful control of all nonradiative recombination pathways, in the bulk and at the interfaces toward the electron and hole extracting layers. Here, we prepare and characterize MAPI solar cells fabricated from a mixture of lead acetate (Pb(CH3COO)2 or Pb(Ac)2) (90 mol %) and lead chloride (PbCl2) (10 mol %) that incorporate no other cations or anions in the crystal lattice. We use the organic charge extraction layers poly(triaryl amine) (PTAA) for hole extraction and [6,6]-phenyl-C61butyric acid methyl ester (PCBM) for electron extraction and show how small processing improvements allow one to substantially suppress recombination at interfaces and in the bulk to achieve open-circuit voltages around 1.26 V. This result outperforms the open-circuit voltages previously reported for other halide perovskites with similar band gaps and approaches the levels previously shown to be possible only in films on glass9 that were lacking charge extracting contacts. We use steady-state photoluminescence (PL) spectroscopy to relate the properties of full solar cells to the ones of films with different contacts, as introduced by Sarritzu et al. 13 Furthermore, we combine steady-state PL with transient photoluminescence (tr-PL) data of different layer stacks to correlate the kinetics of the decay with the absolute intensity of the luminescence. Figure 1 shows a cross-sectional schematic of the full solar cell stack as used in this study, featuring ITO (150 nm) as the transparent conductive oxide, PTAA (∼12−30 nm) as the hole conductor, MAPI (∼280−510 nm) as the absorber, PCBM (∼45 nm) as the electron transport material, as well as bathocuproine (BCP) (∼8 nm) and Ag (80 nm) finishing up the cathode. In this study, we aimed at optimizing simple MAPI layers grown on PTAA for the highest possible open-circuit voltage. The process to deposit the MAPI layers by spin-coating involved a solution of Pb(CH3COO)2, PbCl2, and MAI in DMF or DMF/DMSO combined with a very short annealing time of 2 min at 75 °C. Lead acetate-based precursors have previously been shown to yield smooth,15 high-quality pinholefree layers16 and were therefore chosen to achieve the highest open-circuit voltages. Addition of PbCl2 has been shown to lead to substantially increased PL lifetimes,17,18 increased grain sizes (see Supporting Information (SI) Figure S2) and improved crystallinity due to methylammonium chloride (MACl) in the perovskite precursor solution.19,20 In this study, we find continuous improvement in open-circuit voltages when the concentration of PbCl2 is increased from 0−8%, while the optimal concentration of PbCl2 in terms of efficiency is in the range of 8−10% (see Figure S4). We note that while PbAc2- and PbCl2-based precursors have both been shown to be beneficial for suppressing recombination,15 there is a comparably small amount of literature on these precursors,

Figure 1. Schematic drawing of the layer stack used in this study combined with a SEM cross section of a solar cell.14 111

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters

Figure 2. (a) Current density J as a function of voltage V taken at a 100 mV/s scan speed measured under a class AAA solar simulator for two representative MAPI solar cell in this study. Device A (blue) was optimized for a high open-circuit voltage, and device B (red) was also optimized for efficiency by increasing the thickness of the absorber layer and making the PTAA layer thinner. (b) Open-circuit voltage Voc and (c) efficiency η tracking for 300 s measured with a white LED resulting in a stabilized values of above 1.26 V and around 18% (device A) and 20.2% (device B). (d) Histogram of the open-circuit voltage for the number of devices using the recipe of device A in this study. (e) Quantum efficiency Qe,dir as a function of photon energy E with the accumulated short-circuit current density Jsc of the cells.

Figure 3. (a) Electroluminescence spectrum ϕEL,dir and quantum efficiency Qe,dir from EQE and FTPS of the same cell measured to determine the radiative open-circuit voltage Voc,rad by using the optoelectronic reciprocity relation. The calculation gives a thermodynamic limit for the open-circuit voltage of 1.324 V for the respective device stack. (b) Overview of the open-circuit voltage of lead halide perovskite solar cells as a function of band gap energy Eg, comparing the current state in the literature to the Voc in this work. This comparison demonstrates the outstanding record value that we have achieved by device optimization. The surface plot illustrates the radiative limit for different external quantum efficiencies (1−10−6) calculated with a step-function-like absorptance.

improvements in performance, but we note that, while we observe improvements in Voc similar to the ones presented by Tsai et al.,25 we do not observe any changes in band gap (based on PL data shown in Figure S9) that were suggested to be a result of photoinduced lattice expansion.25 In contrast, the observations of defect passivation and reported fullerene doping23,24,26 are fully in agreement with our results of improved fill factor and Voc after light soaking. Figure 2e shows the photovoltaic quantum efficiency Qe,dir of the same solar cells as a function of photon energy E, providing further evidence for the measured short-circuit current density

different devices fabricated using the recipe of device A. In order to reach these high open-circuit voltages, all devices have to be photoactivated either by measuring several current voltage curves or by being kept under open-circuit conditions for ∼10 min by the white light LED (Figures S7 and S8). There have been various reasons proposed to explain the improvement of electronic properties by light soaking such as defect passivation by redistribution of halide ions,23,24 lattice expansion reducing local strain in the crystal,25 and doping of the fullerene-based electron contact layers.26 Here, we do not aim at further elucidating the mechanisms for light-induced 112

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters

Figure 4. (a) Steady-state PL measurements on different interlayer stacks as a function of laser excitation, converted into the quasi-Fermi level splitting ΔEf and (b) the external PLQE Qe,lum. (c) PL image of the perovskite layer on PTAA and (d) PL image of the complete cell. The data points in (a,b) were measured in Jülich, and the images in (c,d) were measured in Berlin to provide independent confirmation.

and illustrating the key differences between the two device preparation processes. The absorber thickness of device A is around 280 nm, implying that the quantum efficiency at the band gap is not perfectly sharp due to the kink in the absorption coefficient of MAPI (Figure S18). In addition, we observe substantial losses in the blue spectral region that can be partly associated with the ∼30 nm thick PTAA layer. For process B, we therefore reduced the PTAA layer thickness to ∼12 nm to reduce resistive losses and parasitic absorption. The reduction in parasitic absorption is clearly visible in the increased quantum efficiency at energies of ∼3.4 eV, shown in Figure 2e. An increase in the sharpness of the absorption onset of process B vs process A is achieved by a slight increase in perovskite layer thickness of the perovskite (∼330 nm). This is enabled by using a more highly concentrated solution (0.67 M) combined with the addition of DMSO to the precursor solution to maintain a good quality of the film. DMSO in the MAI−PbI2−DMSO phase retards the rapid reaction between PbI2 and MAI during evaporation of the solvent in the spincoating process27 and slows down the crystallization speed. If only a more highly concentrated precursor solution is used, the layer thickness increases, but also the crystal growth changes and very rough layers are formed. To counteract the formation of pinholes in the perovskite layer, which results from the poorer wetting of the PTAA surface for a DMSO-doped perovskite solution, and to further improve grain sizes, the solution is heated to 75 °C before spin-coating.28 On the basis of the quantum efficiency shown in Figure 2e, we calculated the thermodynamic limit for the open-circuit voltage. Figure 3a shows the electroluminescence spectrum ϕEL,dir of device A as well as the photovoltaic quantum efficiency Qe,dir on a logarithmic scale. Using the optoelectronic reciprocity theorem,11 we calculated the solar cell quantum efficiency for the low-energy range from the electroluminescence spectrum ϕEL,dir and combined it with Qe,dir to an overall Qe,PV, which was then used to determine the radiative saturation-current density

∫0

J0,rad = q

ϕbb(E) =

(2)

is the blackbody spectrum at temperature T (of the solar cell), h is Planck’s constant, and c is the speed of light in vacuum. From J0,rad, we may calculate the open-circuit voltage Voc,rad in the limit that there is only radiative recombination via Voc,rad =

yz kT ijjj Jsc lnjj + 1zzzz z q j J0,rad k {

(3)

Using the values Jsc = 18.18 mA/cm and J0,rad = 9.498 × 10−22 mA/cm2 as calculated from eq 1, we obtain a radiative opencircuit voltage Voc,rad = 1.324 V, which is similar to the values that have been obtained for MAPI in refs 5, 6, and 8. If we compare this radiative open-circuit voltage with the actual open-circuit voltages obtained in this study, we estimate the luminescence quantum efficiency Qe,lum as a measure of the nonradiative loss using the theory developed originally by Ross10 and refined and used by many others.11,29−31 We obtain from 2

qVoc,rad − qVoc = qΔVoc,nr = − kT ln(Q e,lum)

(4)

a value of 64 mV for the voltage loss ΔVoc,nr due to nonradiative recombination. This voltage loss corresponds to an external luminescence quantum efficiency of Qe,lum = 8.4%. This external luminescence quantum efficiency, while not being measured directly, is substantially higher than the highest one reported so far, which was ∼1% at 1 sun conditions.12 Figure 3b shows a comparison between measured values of the open-circuit voltage of lead halide perovskite solar cells as a function of band gap Eg and the Shockley−Queisser (SQ) limit for Voc. The SQ and radiative limit for different external quantum efficiencies is calculated for simplicity with a stepfunction-like absorptance as opposed to the measured photovoltaic quantum efficiency used for Figure 3a. While the data points used for Figure 3b are certainly a nonexhaustive representation of the literature, we took care to include the results that reported particularly high open-circuit voltages relative to their band gaps (additional information in Figure S21 and Table S3). Figure 3b shows that, so far, only one data point in the literature has reached a Voc that corresponds to

∞

Q e,PV(E)ϕbb(E) dE

2πE2 1 2πE2 i −E yz zz ≈ expjjj 3 2 3 2 h c [exp(E /kT ) − 1] hc k kT {

(1)

where q is the elementary charge 113

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters ∼1% external luminescence quantum efficiency. The present work shows that this level can be substantially overcome in working devices of pure MAPI (with PbCl2 being used during fabrication) without the inclusion of molecules and elements such as K,32 Rb,12 or Cs33 that have previously been used to achieve the highest open-circuit voltages. As it has recently been shown by Stolterfoht et al.21,34 in a similar device stack, interface recombination between the perovskite absorber layer and the electron and hole transport layers PCBM and PTAA is a critical factor in achieving the highest possible open-circuit voltages. In order to study interfacial recombination, we use steady-state and tr-PL on completed cells and on different layer stacks with only one contact layer (PTAA) or with both contact layers (PTAA and PCBM) but without metal back contact. We additionally show the PL of a MAPI layer grown on PTAA passivated with ntrioctylphosphine oxide (TOPO), allowing us to evaluate the quality of the perovskite itself as well as its interface to PTAA and the uncovered surface. Figure 4a shows the steady-state PL as a function of laser fluence used for sample excitation for the four investigated stacks (passivated perovskite, hole contact only, both contacts, full cell). We show the data for samples that were exposed to light soaking before the measurement. While the left y-axis shows the PL counts in arbitrary units, the right y-axis shows the quasi-Fermi level splitting that corresponds to the PL counts on the y-axis using the luminescence of the solar cell at 1 sun to calibrate the quasiFermi level axis. The quasi-Fermi level splitting ΔEf of the solar cell at 1 sun corresponds to qVoc of the cell, and all other quasiFermi level splittings are calculated from ϕPL ji zyz zz ΔEf = qVoc(ϕsun) + kT lnjjjj jϕ (ϕsun) zz PL,cell k {

the addition of contacts always led to reduced luminescence and quasi-Fermi level splitting. This is no longer the case in the MAPI samples presented here at least if the detectable luminescence emitted through the front surface is compared. The reason for this observation may be that photon recycling has to be a strong factor in devices being that close to the radiative limit. Thus, the presence of a back reflector that increases reabsorption of light relative to outcoupling toward the back may increase the concentration of charge carriers in the device as well as the luminescence that can be detected. We obtained similar results for the PL images of the layer systems measured on a hyperspectral absolute PL imaging setup, which are shown in Figure 4c,d. Because this setup is calibrated to provide absolute quantum yields, the data gave us additional evidence and independent validation of the high quasi-Fermi level splitting observed in our samples. Here we obtain an external quantum efficiency of 5% for the cell, which is smaller than the results depicted in Figure 4b, but it still represents the highest value reported so far for a full lead halide perovskite solar cell device. We note however that higher external luminescence quantum efficiencies of up to ∼15% have been recently reported for perovskite light-emitting diodes.37−41 The image section in Figure 4d shows one solar cell of the four existing on the substrate. In the case of the layer system without an ETL and silver, a contrast between the areas without and with ITO is visible. Contact layers typically quench the luminescence and therefore should also be clearly visible in tr-PL experiments with the expectation being that the transient becomes substantially faster in the presence of a contact. In the past, the longer time decays of luminescence of samples with at least one electron or hole accepting layer have been analyzed and interpreted in terms of surface recombination velocities.34,42,43 However, in the present case, the contacts did not lead to any substantial quenching of the steady-state luminescence. Thus, it is not surprising to see that the transients of the three types of samples shown in Figure 5a decay quite slowly after a quick initial decay. Figure 5b shows the differential lifetimes that follow from

(5)

assuming flat quasi-Fermi levels where ϕPL,cell is the PL intensity of the solar cell and ϕPL is the PL intensity of any of the four types of samples. Thus, any difference in PL intensity of a factor of 10 relative to the PL intensity of the solar cell at an excitation corresponding to 1 sun would correspond to kT ln(10) ≈ 60 meV change in the quasi-Fermi level splitting. We find that the TOPO-passivated sample (gray) has the highest quantum efficiency and a quasi-Fermi energy splitting of ∼1.27 eV at 1 sun, which is comparable to the 1.28 eV one previously reported by Braly et al.9 However, as an electrically insulating interface minimizing recombination, TOPO does not meet the key requirement for a functional solar cell with reasonably low resistive losses. Nevertheless, using TOPO as a passivation experiment,35 we demonstrate that our bulk material is prepared in such a way that nonradiative recombination is already very low. When comparing the passivated sample to the other samples, it stands out that the quasi-Fermi level splitting in the solar cell (red) and in the other layer stacks is only about 10−20 meV smaller. This very small difference demonstrates how well the interlayers in our samples are adapted and suggests that there are only minor nonradiative losses due to surface recombination. We observed that the solar cell actually has a slightly higher quasi-Fermi level splitting than their two respective layer stacks without metal. This apparent superiority of the device relative to the other two samples may at first seem surprising, because the presence of metal contacts should lead to additional surface recombination relative to the typically well-passivated bare surfaces of MAPI.36 Thus, in previous publications,13,21,32,34

ij d ln(ϕPL) yz zz τ = −jjj zz j dt k {

−1

(6)

The lifetimes increase for longer times and saturate for long times in the low microsecond range for all three scenarios. The quantitative interpretation of PL transients of layer stacks involving charge extracting contacts is complicated by the multitude of competing mechanisms that should be taken into account. Here, we therefore performed numerical simulations with the software TCAD Sentaurus42 to obtain a more quantitative understanding of the surface recombination velocity at the perovskite/PTAA interface. On the basis of the parameters given in the SI (Table S1), we performed simulations for different values of the surface recombination velocity that show that these long transients are most compatible with values of S ≈ 1 cm/s or smaller. The shape of the transients is over a large range of times already wellapproximated by assuming radiative recombination only. At early times, Auger recombination contributes, and at longer times, the transient may be affected by nonradiative recombination in the bulk or at the MAPI/PTAA interface. However, the contribution of these monomolecular recombination terms is consistent with high internal and external 114

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters

photon recycling 44−46 in a similar way as efficiency optimization in GaAs solar cells has been achieved in the past.47

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsenergylett.8b01906.

■

Method section on materials, device fabrication and optimization; experimental details and additional data on device characterization using JV, EQE, FTPS, PL, EL, tr-PL, XPS, UPS, and PDS measurements and hyperspectral absolute PL imaging; and parameters for the transient simulations (TCAD Sentaurus) (PDF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (T.K.). *E-mail: [email protected] (L.K.). ORCID

José A. Márquez: 0000-0002-8173-2566 Sanjay Mathur: 0000-0003-2765-2693 Thomas Kirchartz: 0000-0002-6954-8213 Author Contributions #

Z.L. and L.K. contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge support from the Impuls- und Vernetzungsfonds der Helmholtz Gemeinschaft via the project PEROSEED. T.K. and L.K. thank the Bavarian Ministry of Economic Affairs and Media, Energy and Technology for the joint projects in the framework of the Helmholtz Institute Erlangen-Nürnberg. S.O. and S.M. acknowledge support from Merck KGaA Darmstadt and the University of Cologne (Excellence Program “Quantum Matter and Materials”). J.A.M. acknowledges support by the German Federal Ministry of Economics and Technology (BMWi) through the EFFCIS project (Grant No. 0324076D). T.K. would like to thank Henry Snaith (Oxford) and Eva Unger (Lund, Berlin) for fruitful discussions on the origin of the high open-circuit voltages and the effect of ion movement in the presence of organic contact layers.

Figure 5. (a) tr-PL measurements on different layer stacks, namely, the full solar cell (red) and MAPI grown on PTAA without (green) and with PCBM/BCP (blue) as an ETL. (b) Differential lifetime τ obtained by taking the derivative of the curves in (a) for every time t. To obtain a smoother derivative, the data in (a) was first fitted (thin lines shown in (a)), and then, the derivate was computed from the fitted data according to eq 6. (c) tr-PL simulations to model the ITO/PTAA/perovskite stack for different surface velocity S, demonstrating that the experimentally measured decay (green) only fits the simulations for very small S.

■

luminescence quantum efficiencies also in the presence of selective contacts (cf. Figure S11). In conclusion, we have shown that by using a combination of PbAc2- and PbCl2-based precursors and process optimization of the contact layers PTAA and PCBM the open-circuit voltage of simple MAPI solar cells can be tuned up to 1.26 V without any addition of passivating alkali metals such as Cs, K, or Rb. In addition, extremely low surface recombination velocities have been measured that are required in order to rationalize the high open-circuit voltage. While the current study focused on eliminating and characterizing losses due to bulk and interface recombination, future improvements in open-circuit voltage will require optimization of the layer stack for minimum parasitic absorption and maximum efficiency of

REFERENCES

(1) Green, M. A. Radiative Efficiency of State-Of-The-Art Photovoltaic Cells. Prog. Photovoltaics 2012, 20, 472−476. (2) Rau, U.; Blank, B.; Müller, T. C. M.; Kirchartz, T. Efficiency Potential of Photovoltaic Materials and Devices Unveiled by DetailedBalance Analysis. Phys. Rev. Appl. 2017, 7, 044016. (3) Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith, H. J. Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites. Science 2012, 338, 643−647. (4) Jeon, N. J.; et al. A Fluorene-Terminated Hole-Transporting Material for Highly Efficient and Stable Perovskite Solar Cells. Nat. Energy 2018, 3, 682. (5) Tvingstedt, K.; Malinkiewicz, O.; Baumann, A.; Deibel, C.; Snaith, H. J.; Dyakonov, V.; Bolink, H. J. Radiative Efficiency of Lead Iodide Based Perovskite Solar Cells. Sci. Rep. 2015, 4, 6071. 115

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters (6) Tress, W.; Marinova, N.; Inganas, O.; Nazeeruddin, M. K.; Zakeeruddin, S. M.; Graetzel, M. Predicting the Open-Circuit Voltage of CH3NH3PbI3 Perovskite Solar Cells Using Electroluminescence and Photovoltaic Quantum Efficiency Spectra: The Role of Radiative and Non-Radiative Recombination. Adv. Energy Mater. 2015, 5, 1400812. (7) Correa-Baena, J. P.; Saliba, M.; Buonassisi, T.; Grätzel, M.; Abate, A.; Tress, W.; Hagfeldt, A. Promises and Challenges of Perovskite Solar Cells. Science 2017, 358, 739. (8) Yao, J. Z.; et al. Quantifying Losses in Open-Circuit Voltage in Solution-Processable Solar Cells. Phys. Rev. Appl. 2015, 4, 014020. (9) Braly, I. L.; deQuilettes, D. W.; Pazos-Outon, L. M.; Burke, S.; Ziffer, M. E.; Ginger, D. S.; Hillhouse, H. W. Hybrid Perovskite Films Approaching the Radiative Limit with Over 90% Photoluminescence Quantum Efficiency. Nat. Photonics 2018, 12, 355−361. (10) Ross, R. T. Some Thermodynamics of Photochemical Systems. J. Chem. Phys. 1967, 46, 4590−4593. (11) Rau, U. Reciprocity Relation between Photovoltaic Quantum Efficiency and Electroluminescent Emission of Solar Cells. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 76, 085303. (12) Saliba, M.; et al. Incorporation of Rubidium Cations into Perovskite Solar Cells Improves Photovoltaic Performance. Science 2016, 354, 206. (13) Sarritzu, V.; et al. Optical Determination of Shockley-Read-Hall and Interface Recombination Currents in Hybrid Perovskites. Sci. Rep. 2017, 7, 44629. (14) Albrecht, W.; Moers, J.; Hermanns, B. HNF-Helmholtz Nano Facility. JLSRF 2017, 3, 112. (15) Zhang, W.; et al. Ultrasmooth Organic−Inorganic Perovskite Thin-Film Formation and Crystallization for Efficient Planar Heterojunction Solar Cells. Nat. Commun. 2015, 6, 6142. (16) Qiu, W.; et al. Pinhole-Free Perovskite Films for Efficient Solar Modules. Energy Environ. Sci. 2016, 9, 484−489. (17) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.; Alcocer, M. J.; 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. (18) Hutter, E. M.; Eperon, G. E.; Stranks, S. D.; Savenije, T. J. Charge Carriers in Planar and Meso-Structured Organic−Inorganic Perovskites: Mobilities, Lifetimes, and Concentrations of Trap States. J. Phys. Chem. Lett. 2015, 6, 3082−3090. (19) Yang, M.; et al. Perovskite Ink with Wide Processing Window for Scalable High-Efficiency Solar Cells. Nat. Energy 2017, 2, 17038. (20) Zhao, Y.; Zhu, K. CH3NH3Cl-Assisted One-Step Solution Growth of CH3NH3PbI3: Structure, Charge-Carrier Dynamics, and Photovoltaic Properties of Perovskite Solar Cells. J. Phys. Chem. C 2014, 118, 9412−9418. (21) Stolterfoht, M.; et al. The Perovskite/Transport Layer Interfaces Dominate Non-Radiative Recombination in Efficient Perovskite Solar Cells. http://arxiv.org/abs/1810.01333 (2018). (22) Shao, Y.; Yuan, Y.; Huang, J. Correlation of Energy Disorder and Open-Circuit Voltage in Hybrid Perovskite Solar Cells. Nat. Energy 2016, 1, 15001. (23) deQuilettes, D. W.; et al. Photo-Induced Halide Redistribution in Organic-Inorganic Perovskite Films. Nat. Commun. 2016, 7, 11683. (24) Mosconi, E.; Meggiolaro, D.; Snaith, H. J.; Stranks, S. D.; De Angelis, F. Light-Induced Annihilation of Frenkel Defects in OrganoLead Halide Perovskites. Energy Environ. Sci. 2016, 9, 3180−3187. (25) Tsai, H.; et al. Light-Induced Lattice Expansion Leads to HighEfficiency Perovskite Solar Cells. Science 2018, 360, 67. (26) De Bastiani, M.; et al. Ion Migration and the Role of Preconditioning Cycles in the Stabilization of the J−V Characteristics of Inverted Hybrid Perovskite Solar Cells. Adv. Energy Mater. 2016, 6, 1501453. (27) Jeon, N. J.; Noh, J. H.; Kim, Y. C.; Yang, W. S.; Ryu, S.; Seok, S. I. Solvent Engineering for High-Performance Inorganic−Organic Hybrid Perovskite Solar Cells. Nat. Mater. 2014, 13, 897.

(28) Chiang, C.-H.; Lin, J.-W.; Wu, C.-G. One-Step Fabrication of a Mixed-Halide Perovskite Film for a High-Efficiency Inverted Solar Cell and Module. J. Mater. Chem. A 2016, 4, 13525−13533. (29) Smestad, G.; Ries, H. Luminescence and Current Voltage Characteristics of Solar-Cells and Optoelectronic Devices. Sol. Energy Mater. Sol. Cells 1992, 25, 51−71. (30) Kirchartz, T.; Nelson, J.; Rau, U. Reciprocity between Charge Injection and Extraction and Its Influence on the Interpretation of Electroluminescence Spectra in Organic Solar Cells. Phys. Rev. Appl. 2016, 5, 054003. (31) Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J. V. On the Origin of the Open-Circuit Voltage of Polymer-Fullerene Solar Cells. Nat. Mater. 2009, 8, 904−909. (32) Abdi-Jalebi, M.; et al. Maximizing and Stabilizing Luminescence from Halide Perovskites with Potassium Passivation. Nature 2018, 555, 497. (33) Saliba, M.; et al. Cesium-Containing Triple Cation Perovskite Solar Cells: Improved Stability, Reproducibility and High Efficiency. Energy Environ. Sci. 2016, 9, 1989−1997. (34) Stolterfoht, M.; et al. Visualization and Suppression of Interfacial Recombination for High-Efficiency Large-Area pin Perovskite Solar Cells. Nat. Energy 2018, 10, 847−854. (35) deQuilettes, D. W.; Koch, S.; Burke, S.; Paranji, R. K.; Shropshire, A. J.; Ziffer, M. E.; Ginger, D. S. Photoluminescence Lifetimes Exceeding 8 μs and Quantum Yields Exceeding 30% in Hybrid Perovskite Thin Films by Ligand Passivation. ACS Energy Lett. 2016, 1, 438−444. (36) Staub, F.; Hempel, H.; Hebig, J. C.; Mock, J.; Paetzold, U. W.; Rau, U.; Unold, T.; Kirchartz, T. Beyond Bulk Lifetimes: Insights into Lead Halide Perovskite Films from Time-Resolved Photoluminescence. Phys. Rev. Appl. 2016, 6, 044017. (37) Ban, M.; et al. Solution-Processed Perovskite Light Emitting Diodes with Efficiency Exceeding 15% Through Additive-Controlled Nanostructure Tailoring. Nat. Commun. 2018, 9, 3892. (38) Zou, W.; et al. Minimising Efficiency Roll-Off in HighBrightness Perovskite Light-Emitting Diodes. Nat. Commun. 2018, 9, 608. (39) Yang, X.; et al. Efficient Green Light-Emitting Diodes based on Quasi-Two-Dimensional Composition and Phase Engineered Perovskite with Surface Passivation. Nat. Commun. 2018, 9, 570. (40) Xiao, Z.; Kerner, R. A.; Zhao, L.; Tran, N. L.; Lee, K. M.; Koh, T.-W.; Scholes, G. D.; Rand, B. P. Efficient Perovskite Light-Emitting Diodes Featuring Nanometre-Sized Crystallites. Nat. Photonics 2017, 11, 108. (41) Cho, H.; et al. Overcoming the Electroluminescence Efficiency Limitations of Perovskite Light-Emitting Diodes. Science 2015, 350, 1222−1225. (42) Krogmeier, B.; Staub, F.; Grabowski, D.; Rau, U.; Kirchartz, T. Quantitative Analysis of the Transient Photoluminescence of CH3NH3PbI3/PC61BM Heterojunctions by Numerical Simulations. Sust. Energy & Fuels 2018, 2, 1027−1034. (43) Hutter, E. M.; Hofman, J. J.; Petrus, M. L.; Moes, M.; Abellón, R. D.; Docampo, P.; Savenije, T. J. Charge Transfer from Methylammonium Lead Iodide Perovskite to Organic Transport Materials: Efficiencies, Transfer Rates, and Interfacial Recombination. Adv. Energy Mater. 2017, 7, 1602349. (44) Abebe, M. G.; Abass, A.; Gomard, G.; Zschiedrich, L.; Lemmer, U.; Richards, B. S.; Rockstuhl, C.; Paetzold, U. W. Rigorous WaveOptical Treatment of Photon Recycling in Thermodynamics of Photovoltaics: Perovskite Thin-Film Solar Cells. Phys. Rev. B: Condens. Matter Mater. Phys. 2018, 98, 075141. (45) Kirchartz, T.; Staub, F.; Rau, U. Impact of Photon Recycling on the Open-Circuit Voltage of Metal Halide Perovskite Solar Cells. ACS Energy Lett. 2016, 1, 731−739. (46) Pazos-Outon, L. M.; Xiao, T. P.; Yablonovitch, E. Fundamental Efficiency Limit of Lead Iodide Perovskite Solar Cells. J. Phys. Chem. Lett. 2018, 9, 1703−1711. 116

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117

Letter

ACS Energy Letters (47) Miller, O. D.; Yablonovitch, E.; Kurtz, S. R. Strong Internal and External Luminescence as Solar Cells Approach the Shockley− Queisser Limit. IEEE J. Photovolt. 2012, 2, 303−311.

117

DOI: 10.1021/acsenergylett.8b01906 ACS Energy Lett. 2019, 4, 110−117