Perovskite Solar Cells Yielding Reproducible Photovoltage of 1.20 V

High photovoltages and power conversion efficiencies of perovskite solar cells (PSCs) can be realized by controlling the undesired nonradiative charge carrier recombination. Here, we introduce a judicious amount of guanidinium iodide into mixed-cation and mixed-halide perovskite films to suppress the parasitic charge carrier recombination, which enabled the fabrication of >20% efficient and operationally stable PSCs yielding reproducible photovoltage as high as 1.20 V. By introducing guanidinium iodide into the perovskite precursor solution, the bandgap of the resulting absorber material changed minimally; however, the nonradiative recombination diminished considerably as revealed by time-resolved photoluminescence and electroluminescence studies. Furthermore, using capacitance-frequency measurements, we were able to correlate the hysteresis features exhibited by the PSCs with interfacial charge accumulation. This study opens up a path to realize new record efficiencies for PSCs based on guanidinium iodide doped perovskite films.


Introduction
Over the past few years, perovskite solar cells (PSCs) have attracted great attention in the applied and theoretical research fields [1]. The appealing photophysical properties, such as small exciton binding energy, tunable bandgap, high absorption coefficient, relatively high carrier mobility and diffusivity, and tolerance to defects, broaden the application horizon of perovskites from photovoltaics to light emitting diodes and lasing [2][3][4][5][6]. Since PSCs were first introduced by the Miyasaka group in 2009, the power conversion efficiency (PCE) has been improved from 3.8% to over 23% using solution-based deposition methods [2,7]. However, these methods can generate films exhibiting pinholes and defects, which are detrimental to the device performance due to the occurrence of parasitic charge carrier recombination under operational conditions [8,9]. Therefore, to develop high-efficiency PSCs using solution-based approaches, one of the challenges is to achieve a rational control over the quality of the absorber layer [10,11]. Recent reports on PSCs have shown that nonradiative charge carrier recombination or charge injection at the interfaces to charge transport layers limits the open circuit voltage ( OC ) [12][13][14][15]. The role of grain boundaries acting as recombination centers is under debate as such processes depend on the processing conditions and composition of the perovskite layer [16][17][18][19][20].
As it has become important to suppress the nonradiative recombination losses occurring at the interfaces or throughout the bulk of the perovskite film [21], many efforts have been devoted towards passivating the defects, which consequently retard the trap-assisted nonradiative charge carrier recombination. For example, a series of dialkylammonium and phenyltrimethylammonium halides were successfully explored to passivate the interfaces and grain boundaries resulting in a much-improved PCE [22][23][24]. In a similar manner, the guanidinium (Gua) cation (CH 6 N 3 + ) was employed for the passivation of nonradiative recombination centers at CH 3 NH 3 PbI 3 grain boundaries enhancing the OC up to 1.07 V [25][26][27]. The Gua cation may also act as a crosslinker between the neighboring perovskite grains in the film [28]. It is worth emphasizing that the Gua cation was mainly investigated as an additive to CH 3 NH 3 PbI 3 based PSCs displaying moderate power conversion efficiencies. In addition to forming mixed-cation phases exhibiting relatively large bandgap, Gua cation can mix with MA and/or FA cations in the 3D perovskite lattice [29,30] and consequently influence the excitonic and structural properties of the host phase remarkably. However, its influence on the bandgap, film quality, and performance of multication-based PSCs has not been explored so far.
In this article, we illustrate the effect of guanidinium iodide (GuaI) on the photophysical properties of mixedcation and mixed-halide perovskite films. Thorough analyses based on scanning electron microscopy and emission studies were carried out to investigate the effect of GuaI addition on the morphology and emission of the perovskite material. Furthermore, electroluminescence and time-resolved photoluminescence studies were performed, respectively, to investigate the fully assembled devices and the charge carrier recombination occurring within the perovskite films. The operational stability carried out at a maximum power point under continuous full-sun illumination for 200 h showed that the PSCs containing Gua cations are as stable as control PSCs. In addition to providing new physical insights, we demonstrate for the first time robust stability for Gua based PSCs which are >20% efficient and yield photovoltage as high as 1.20 V.

Steady-State and Time-Resolved Photoluminescence.
To unravel the impact of GuaI on the emission properties of the absorber layer, steady-state photoluminescence (PL) was measured for Cs 5 Pb.5GuaI films (Figure 1(a)). Upon addition of 5% GuaI, a slight red shift from 757 to 766 nm in the peak position indicates that incorporation of GuaI in the perovskite lattice decreases the band gap slightly. In addition, the emission intensity enhanced dramatically with the incorporation of GuaI in Cs 5 Pb precursor solution indicating a decrease in the nonradiative recombination.
Variation in the steady-state PL intensity apparently demonstrates that the level of defects promoting nonradiative carrier recombination in the perovskite film decreases upon introducing GuaI into the precursor solution. We employed time-resolved photoluminescence (TRPL) to further characterize the perovskite films quantitatively (Figure 1(b)). While exciting the films from the perovskite side using 408 nm photons, we studied the charge carrier dynamics occurring within the reference and Cs 5 Pb.xGuaI perovskite films. Figure 1(b) shows that the incorporation of 5% GuaI dramatically increases the charge carrier lifetime. Charge carrier lifetime of 215 ns and 627 ns was estimated, respectively, for the Cs 5 Pb, and Cs 5 Pb.0.05GuaI films. The initial fast decay is strongly reduced in the Cs 5 Pb.0.05GuaI film, consistent with the high PL intensity [32].

Morphological and Photovoltaic
Characterization. Furthermore, the surface morphology of perovskite films was characterized via scanning electron microscopy (SEM) (Figure 2). Ostensibly, top view (Figures 2(a) and 2(b)) and crosssectional (Figures 2(c) and 2(d)) SEM micrographs acquired from Cs 5 Pb, Cs 5 Pb.0.05GuaI perovskite films and the devices based on them confirm that the morphology of absorber layers is not affected by the addition of GuaI. The effect of GuaI incorporation on the photovoltaic performance is further examined in a device configuration of FTO/c-TiO 2 /m-TiO 2 /perovskite/spiro-OMeTAD/Au (for more details see Methods). Figure 3(a) depicts the J-V curves of the devices with 5% GuaI content in comparison to the reference Cs 5 Pb device, and the corresponding photovoltaic parameters are summarized in Table 1. For the Cs 5 Pb device, the highest PCE obtained is 20.03% with an open circuit voltage ( OC ) of 1.13 V, a short-circuit current density (J SC ) of 23 mA cm −2 , and a fill factor (FF) of 77%, which is in agreement with the values reported for the cesium-containing triple cation PSCs previously [31].
The PCE of Cs 5 Pb.0.05GuaI based device is comparable to that of the reference device (Table 1). With respect to the Cs 5 Pb based device, the Cs 5 Pb.0.05GuaI devices showed higher J SC values of 23.6 mA/cm 2 . Intriguingly, the addition of 5% GuaI to Cs 5 Pb significantly improved the OC from 1.13 V to 1.20 V. To the best of our knowledge, a OC of 1.20 V is among the highest values reported in the literature for a similarly configured PSC. Consistent with the PL data, Figure 3(b) shows the incident photon-to-current efficiency (IPCE) and integrated current density as a function of wavelength. From the IPCE spectra, we found that upon addition of 5% GuaI (Figure 3(b)), the same number of photons appears to generate more current compared to the reference Cs 5 Pb device.
Based on the photovoltaic characterization, we found that the incorporation of 5% GuaI into mixed-cation and mixedhalide perovskite films improved the overall performance of PSCs. The reproducibility of the device performance was further ascertained by measuring 22 devices each from Cs 5 Pb and Cs 5 Pb.0.05GuaI as shown in Figure 3(c). We noted that a OC between 1.19-1.20 V is consistently achievable. By contrast, the fill factor of the Cs 5 Pb.0.05GuaI devices is lower than that of the Cs 5 Pb reference. Generally, the FF of a solar cell is affected by the series resistance (R S ) and ideality factor of the device [33]. By taking the slope of light J-V characteristics at the OC , R S was calculated and summarized in Table S1. The Cs 5 Pb.0.05GuaI device exhibits a higher value of R S as compared to the Cs 5 Pb device. Therefore, by

Insight about the Hysteresis
Behaviour. The Cs 5 Pb and Cs 5 Pb.0.05GuaI based devices were also examined for hysteresis ( Figure S2) and despite favorable properties of Gua ions, such as zero dipole moment, the hysteresis effects were more prominent in GuaI based devices [34]. To understand the cause of the dominant hysteresis feature [35], capacitancefrequency measurements at zero bias in the dark were carried out (Figure 4(a)), where features in the low and high frequency of C-f response are clearly distinguishable. A constant capacitance element at a frequency >10 3 Hz for Cs 5 Pb and >10 4 Hz for Cs 5 Pb.0.05GuaI based devices is associated with the dielectric response of the absorber material. In turn, the capacitance in the low-frequency spectra could be associated with the ionic characteristics [35,36]. Apparently, excess Gua ions, which are not confined within the perovskite crystal lattice and have weak bonding capability, could pile up near the contact interface and screen the local electric field leading to a higher value of capacitance in the low-frequency region [37]. By considering the electrostatic interactions at room temperature, space charge densities of 6.04×10 17 cm −3 and 1.4×10 18 cm −3 were obtained, respectively, for the reference and Cs 5 Pb.0.05GuaI devices. The accumulation of relatively higher ions causes an excess capacitance in the low frequency, which consequently leads to the hysteresis. Furthermore, the motion of these ions under the applied bias can screen the internal electric field at the interface (TiO 2 /perovskite), thus amplifying the hysteresis effect.

Electroluminescence Study.
Electroluminescence (EL) measurements were carried out to compare the radiative emission properties of Cs 5 Pb and Cs 5 Pb.0.05GuaI PSCs. The EL was measured during a voltage sweep from 0 to 2 V and back. Figure S3 shows the J-V curve (solid lines) and emitted photon flux (dashed lines) obtained from Cs 5 Pb and Cs 5 Pb.0.05GuaI devices by a voltage loop starting from 0 V with a scan rate of 20 mV/s. The current onset is shifted towards higher voltages upon addition of GuaI (solid lines), consistent with the increased OC . This comes along with a roughly four-times higher radiative emission yield when compared at the same driving current (Figure 4(b)). Therefore, the increased PL signal and lifetime of charge carriers observed in perovskite films could be maintained.
Following the approach in [38], OC can be derived from the emission yield (  (Figure 4(c)) [39]. It is evident that Cs 5 Pb and Cs 5 Pb.0.05GuaI remain relatively stable for almost 200 h, which indicates that the incorporation of Gua cations, an organic species, does not introduce any additional source of degradation [40]. Thus far, such a promising operational stability has not been demonstrated for >20% efficient guanidinium-based PSCs yielding photovoltage as high as 1.20 V, which makes this investigation highly important.

Discussion
We investigated the effect of guanidinium iodide incorporation into mixed-cation and mixed-halides perovskite films. Apparently, a slight red shift from 757 to 766 nm in the PL peak position indicates the incorporation of GuaI in the perovskite lattice. A gradual increase in the content of GuaI shows a significant enhancement of the open circuit voltage from 1.13 V to 1.20 V leading to the realization of >20.3% PCE. The increased photoluminescence and lifetime of charge carriers observed in GuaI containing perovskite films were maintained in the fully assembled device, justifying the trends in the OC . PSCs containing a definite amount of GuaI showed four-times higher radiative emission yield than Cs 5 Pb devices. The accumulation of higher ions in GuaI based PSCs causes an excess capacitance in the lowfrequency response of capacitance-frequency measurements, which eventually increased the hysteresis.

Experimental Design.
We aimed to fabricate highly reproducible and highly efficient perovskite solar cell yielding remarkable photovoltages by judiciously tailoring the photophysical properties of perovskite structures. In this direction, the following steps and detailed characterization techniques, including spectroscopy, scanning electron microscopy, Xray diffraction, and other device characterization techniques, were employed.

Materials.
All materials were purchased from Sigma-Aldrich and were used as received, unless stated otherwise. The precursor solution was spin coated onto the mesoporous TiO 2 films in a two-step program at 1000 and 6000 r.p.m. for 10 and 30 s, respectively. During the second step, 130 l of chlorobenzene was dropped on the spinning substrate 10 s prior the end of the program. This was followed by annealing the films at 100 ∘ C for 40 min. The device fabrication was carried out under controlled atmospheric conditions with humidity <2%. For completing the fabrication of devices, 2,2' ,7,7'-tetrakis(N,Ndi-p-methoxyphenylamine)-9,9-spirobifluorene (spiro-OMeTAD, 60 mM in chlorobenzene), the HTM was doped with bis(trifluoromethylsulfonyl)imide lithium salt, tris(2-(1H-pyrazol-1-yl)-4-tert-butylpyridine)-cobalt(III) tris(bis(trifluoromethylsulfonyl) imide) (FK 209, from Dyenamo), and 4-tert-Butylpyridine in a molar ratios of 0.5, 0.05, and 3.3, respectively. Finally, ∼75 nm gold (Au) layer was thermally evaporated. photovoltaic (J-V) measurements. The spectral output of the lamp was filtered using a Schott K113 Tempax sunlight filter (Präzisions Glas & Optik GmbH, Germany) to reduce the mismatch between the simulated and actual solar spectrum to less than 2%. The photo-active area of 0.16 cm 2 was defined using a dark-coloured metal mask. The IPCE measurements were performed using a LED light source (Ariadne EQE from Cicci Research).

Morphological Characterization.
Scanning electron microscopy (SEM) was performed on a ZEISS Merlin HR-SEM using an In-lens detector.

Structural
Characterization. X-ray diffraction data were collected on a Bruker Advance D8 X-ray diffractometer with a graphite monochromator, using Cu K radiation. laser head NanoLED-405LH (Horiba) emitting <200 ps duration pulses at 408 nm with a repetition rate of 1 MHz was used as an excitation source.

Electroluminescence
Measurements. The emitted photon flux was detected with a large-area (1 cm 2 ) Si-photodiode (Hamamatsu S1227-1010BQ) positioned close to the sample. The voltage scan was performed using a Bio-Logic SP300 potentiostat, which was also used to simultaneously measure the short-circuit current of the photodiode connected to a second channel.  for financial support. Neha Arora gratefully acknowledges financial support from Greatcell Solar. Figure S1. Photoluminescence spectra recorded from Cs 5 Pb. xGuaI films deposited onto the mesoporous Al 2 O 3 substrate. Table S1. Summarized J-V characteristic of the devices with 5% GuaI content in comparison to the reference Cs 5 Pb device. Figure S2. Forward and reverse current-voltage scans for (a) Cs 5 Pb and (b) Cs 5 Pb.0.05GuaI devices. Figure S3. J-V curve (solid lines) and emitted photon flux (dashed lines) obtained from Cs 5 Pb and Cs 5 Pb.0.05GuaI devices by a voltage loop starting from 0 V with a scan rate of 20 mV/s. Differences in current at low voltages are due to the hysteresis phenomenon. Figure S4. X-ray diffraction patterns corresponding to pure and GuaI containing perovskite films. (Supplementary Materials)