sustainable energy & fuels - perovskite info · 2019. 1. 30. · perovskite solar cells. we...

SustainableEnergy & Fuels

PAPER

Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM

.

View Article OnlineView Journal

High open circui

aUniversity of Potsdam, Institut für Physik un

[email protected]̈t, Institut für Physik, BcHelmholtz-Zentrum Berlin, Institute for SilidHelmholtz-Zentrum Berlin, Young Investiga

Berlin, Germany. E-mail: steve.albrecht@heeHelmholtz-Zentrum Berlin für Materialien ufTechnical University Berlin, Faculty IV – Elec

10587 Berlin, Germany

† Electronic supplementary informa10.1039/c8se00509e

Cite this: DOI: 10.1039/c8se00509e

Received 19th October 2018Accepted 7th January 2019

DOI: 10.1039/c8se00509e

rsc.li/sustainable-energy

This journal is © The Royal Society of

t voltages in pin-type perovskitesolar cells through strontium addition†

Pietro Caprioglio, ad Fengshuo Zu,be Christian M. Wolff, a José A. MárquezPrieto,e Martin Stolterfoht, a Pascal Becker,e Norbert Koch,be Thomas Unold, e

Bernd Rech,c Steve Albrecht*df and Dieter Neher *a

The incorporation of even small amounts of strontium (Sr) into lead-base hybrid quadruple cation

perovskite solar cells results in a systematic increase of the open circuit voltage (Voc) in pin-type

perovskite solar cells. We demonstrate via absolute and transient photoluminescence (PL) experiments

how the incorporation of Sr significantly reduces the non-radiative recombination losses in the neat

perovskite layer. We show that Sr segregates at the perovskite surface, where it induces important

changes of morphology and energetics. Notably, the Sr-enriched surface exhibits a wider band gap and

a more n-type character, accompanied with significantly stronger surface band bending. As a result, we

observe a significant increase of the quasi-Fermi level splitting in the neat perovskite by reduced surface

recombination and more importantly, a strong reduction of losses attributed to non-radiative

recombination at the interface to the C60 electron-transporting layer. The resulting solar cells exhibited

a Voc of 1.18 V, which could be further improved to nearly 1.23 V through addition of a thin polymer

interlayer, reducing the non-radiative voltage loss to only 110 meV. Our work shows that simply adding

a small amount of Sr to the precursor solutions induces a beneficial surface modification in the

perovskite, without requiring any post treatment, resulting in high efficiency solar cells with power

conversion efficiency (PCE) up to 20.3%. Our results demonstrate very high Voc values and efficiencies in

Sr-containing quadruple cation perovskite pin-type solar cells and highlight the imperative importance of

addressing and minimizing the recombination losses at the interface between perovskite and charge

transporting layer.

Introduction

Organic–inorganic halide perovskites are considered one of themost promising materials for photovoltaic applications due totheir rather easy and low-cost fabrication, as well as outstandingoptoelectronic properties. Notably, these semiconductorscombine a high absorption coefficient with panchromaticabsorption of light1 and a long carrier diffusion length,2,3

allowing efficient photon absorption and charge extraction fora typical active layer thickness of only 500 nm. Another pecu-liarity of hybrid perovskites is that defects create mostly shallow

d Astronomie, Potsdam, Germany. E-mail:

erlin, Germany

con Photovoltaics, Berlin, Germany

tor Group Perovskite Tandem Solar Cells,

lmholtz-berlin.de

nd Energie GmbH, Berlin, Germany

trical Engineering and Computer Science,

tion (ESI) available. See DOI:

Chemistry 2019

energy levels, allowing high open circuit voltage (Voc) and longcarrier lifetime.4 Moreover, perovskite materials can be ob-tained from in-nature abundant precursors, which potentiallyreduce further the costs of future large scale production.Despite the fact that the rst full solid state perovskite solarcell was reported only in 2012, with a power conversion effi-ciency (PCE) of 9.7%,5 this technology has experienceda tremendous improvement,6–8 currently reaching a recordPCE of 22.7%.9 Regardless of the state of the art of perovskitesolar cells and their astonishing performances, the metrics llfactor (FF) and Voc are currently still limiting their PCE. Thus,in order to achieve higher efficiencies the limiting agents ofthese parameters need to be understood and minimized.10,11 Ithas been realized that non-radiative recombination at oracross the interface between the perovskite absorber and theadjacent charge-transporting layer (CTL) constitutes the majorrecombination loss.12,13 Recent studies have also shown howgenerally all common type of transport layer cause severeenergy losses and substantially dominates the non-radiativerecombination in the devices.14 To overcome this problematic,several approaches such as surface treatment or passivation,interlayers and compositional engineering have been applied

Sustainable Energy Fuels

http://crossmark.crossref.org/dialog/?doi=10.1039/c8se00509e&domain=pdf&date_stamp=2019-01-21http://orcid.org/0000-0002-3465-2475http://orcid.org/0000-0002-7210-1869http://orcid.org/0000-0002-4023-2178http://orcid.org/0000-0002-5750-0693http://orcid.org/0000-0001-6618-8403http://dx.doi.org/10.1039/c8se00509ehttps://pubs.rsc.org/en/journals/journal/SE

Sustainable Energy & Fuels Paper

Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM

. View Article Online

to boost the Voc of these devices.8,15–17 However, most of theseapproaches require a multistep deposition process, whereeither the existing perovskite layer is altered through a post-deposition treatment or an additional thin layer is depositedon top. Both approaches are not suitable when aiming at fastproduction schemes. Alternatively, attempts have beenmade to increase the performance of perovskite devices byadding suitable components to the perovskite precursorsolution.18–20 Such additives were shown to enrich at grain-boundaries or at the perovskite surface, paving the way forsuppressing unwanted non-radiative recombination whileavoiding a multistep preparation scheme [e.g. TOPO,17 andsome other small and large molecules]. One recent example ofthis strategy is the addition of SrI2 to the precursor solution ofhybrid and inorganic perovskites. It has been shown that Sr2+

partially substitute Pb2+ in the perovskite lattice,21,22 owing tothe nearly almost identical ionic radii of both ions (Sr2+ ¼ 132pm, Pb2+ ¼ 133 pm).22 Recent results suggested, however, thatSr segregates preferentially at the surface of solution-processedperovskites lms, going along with specic changes of thephotovoltaic parameters.

For example the addition of 1–2% SrI2 to the MAPbI3precursor solution increased the PCE of an pin-type device from12% to nearly 15%, mainly through an increase in Jsc and theFF, while the Voc was actually reduced.23 The overall improve-ment in device efficiency was attributed to an increased carrierlifetime in combination with surface passivation, resulting inan improved charge extraction. More recently Lau et al. reportedSr2+ insertion into a CsPbI3 perovskite.24 Here, addition of Sr ata concentration of 2% improved all photovoltaic parameters ofa nip-type solar cell, resulting in a stabilized PCE of nearly 11%.Based on an improved PL lifetime and predominant Sr surfaceaggregation, the authors concluded that Sr mainly acts asa surface passivating agent. Furthermore, also in Cl-containingperovskite the presence of Sr has shown positive effects on theperformance of actual devices.25,26 On the other hand, a morerecent paper demonstrated a signicant reduction of allphotovoltaic parameters of nip-type devices when adding Sr toa MAPI precursor solution.27

Here, we apply this approach to efficient pin-type devicescomprising a solution-processed quadruple cation perov-skite28 sandwiched between the hole-transporting polymerPTAA and a C60 electron-transporting layer. These pin-typecells attracted considerable attention as they require onlyultrathin undoped charge transport layers (e.g. 8 nmPTAA, 30 nm C60)11 and moderate annealing temperatures.However, the efficiency of such devices was shown to belargely limited by the perovskite/C60 interface, limiting theVoc to values of about 1.11 V for a perovskite with a band gapof around 1.62 eV bandgap, if no further treatments areapplied.29 We nd that addition of Sr leads to a large reduc-tion of non-radiative recombination loss in the device, withan increase of the Voc by 70 mV, up to a remarkable value of1.18 V. A combination of methods, including transient andabsolute photoluminescence (PL), second ion mass spec-troscopy (SIMS), scanning electron microscopy (SEM), andphotoelectron spectroscopy (PES) is applied to arrive at


a comprehensive picture of the morphology and surfaceenergetics of the neat perovskite layers, with and withoutSr. Our study conrms that Sr segregates mostly at theperovskite surface, where it widens the band gap and inducesa stronger n-type surface band bending. We demonstratethat these modications limit the accessibility of thesurface to photogenerated holes, thereby enhancing thequasi-Fermi level splitting and reducing interface-mediatednon-radiative recombination with C60 in the Sr-containingperovskite samples.

Device structure and materials

We employed solar cells with a simple “inverted” congurationwhere light enters through the hole-extracting contact (denotedhere as pin geometry),30 as shown in Fig. 1a. Acknowledgingthat, currently, best performing perovskite solar cells utilizea combination of inorganic and organic transport layers ina regular structure (nip), in this work we focus exclusively onundoped, fully organic transport layers. Inverted solar cells havereduced hysteresis, low processing temperatures and recentlygained strong interest in perovskite based multijunctiondevices.31 Here, a very thin layer (

Fig. 1 (a) Solar cell device structure representing the layer stackutilized here, (b) ABX3 perovskite cubic structure with the differentspecies implemented in this study. The figure indicates how the Sr2+

can take the place of Pb2+ into the perovskite lattice.

Paper Sustainable Energy & Fuels

Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


Results and discussionsOptoelectronic properties

Regardless the presence of Sr, our devices do not show anyevidence of hysteresis effect during J–V scans (Fig. S1A, ESI†).Importantly, already a small quantity of Sr (0.05%) affects thephotovoltaic performance, especially the Voc, which increasesfrom 1.10 V to 1.12 V. The rather low Voc of the Sr-free device ismost likely caused by the poor selectivity of C60 as an electron-extracting contact, as mentioned above. Further increasing theSr/Pb ratio to 2% in various steps, a consistent increase of Vocfrom 1.12 V to 1.18 V is found, see Fig. 2b, highlighting anexceptional Voc increase of 70 mV compared to the referencecell. A Voc of 1.18 V, for a perovskite absorber with a band gap of1.6 eV in a pin-type solar cell architecture is only enabled bya strong minimization of the energy losses. The resulting cellsshow a power conversion efficiency of 20.3% for the recorddevice, well exceeding the value of any other Sr containingperovskite reported so far.23–26 Higher Sr concentrations(Fig. S1B and S1D, ESI†) did not improve the Voc further, whilecausing a signicant drop of the FF and Jsc. We attribute this tothe signicant changes of the surface morphology whenincreasing the amount of Sr as discussed below. For this reason,

Fig. 2 (a) J–V characteristics in reverse scan (0.1 V s�1 with voltage stepcm�2 of the best devices for the two cases 0% Sr and 2% Sr. (b) Averagebased on 10 cells for each Sr concentration.

This journal is © The Royal Society of Chemistry 2019

we refer to a 2% Sr addition as the optimized amount and usethis concentration for further comparisons. Adding Sr until 2%does not compromise the Jsc, which varies in a non-systematicway around an average value of 22 � 0.5 mA cm�2 (Table 1),but we consistently observe a higher Jsc for the device with 2%Sr. A conclusive interpretation for this phenomenon is not clear,but we note that the higher Jsc of the 2% Sr device is also re-ected in the slightly larger external quantum efficiency (EQEPV)(Fig. S2A ESI†). On the other hand, increasing the Sr concen-tration causes a continuous decease of the FF. Notably, anopposite effect has been reported when Sr was added to MAPbI3in a solution-processed pin-type device, as described above.23

Here, the addition of Sr to led to a marked increase in Jsc and FF,assigned to improved charge extraction while the Voc decreasedby 50 meV.23 We cannot resolve this discrepancy but we notethat our devices differ from those employed in ref. 32substantially (the choice of the perovskite and all charge-transporting layers as well as a different way of perovskitelayer preparation). Following our previous studies,10 we tried tofurther enhance the Voc by adding a thin (less than 5 nm)insulating polystyrene (PS) layer at the interface betweenperovskite and C60. The resulting solar cell shows an extraor-dinarily high Voc of nearly 1.23 V due to the signicantlyreduced interface recombination. This result highlights the Vocpotential of well optimized electron selective contact. However,as reported in Fig. S1C, ESI,† the inclusion of the ultra-thininsulating PS interlayer lead to a signicant reduction of theFF by probably limiting the extraction of electrons from theperovskite layer. For a better comparison, a box chart repre-senting all parameters for the most important type of devicesanalysed in this work is presented in Fig. S1D, ESI.† As our workis mostly concerned with the suppression of interfacial recom-bination though mixed solution processing, we will notconsider this approach further in the work.

The position and shape of the onset of the EQEPV (Fig. S2C,ESI†) remain almost unaltered within sample reproducibility onSr addition in the range up to 2%. A widening of the band gap,which might be assumed as a cause of the Voc increase aercomposition modication,28,34,35 is not found here. This is also

of 0.02 V) under simulated AM 1.5G illumination calibrated to 100 mWd Voc values at 100 mW cm

�2 illumination as a function of Sr content


http://dx.doi.org/10.1039/c8se00509e

Table 1 Averaged J–V parameters for devices with different Sr concentrations including the standard errors based on a statistics of 10 cells foreach Sr concentration. Values are taken from J–V scans with scan rate of 0.1 V s�1 and voltage step of 0.02 V. The Jsc values calculated from theintegrated EQEPV spectrum (Fig. S2A) matchwithin a 5% deviation with the Jsc measured in J–V scans. As shown in Fig. S1 ESI forward and reversescan gives identical values showing the absence of hysteresis effect. The record parameters are reported in brackets

Sr [%] Voc [V] Jsc [mA cm�2] FF [%] PCE [%]

0 1.108 � 0.007 (1.121) 22.0 � 0.6 (22.9) 76.0 � 0.8 (77.6) 19.2 � 0.3 (19.4)0.05 1.127 � 0.005 (1.133) 21.4 � 0.5 (22.2) 73.0 � 2.0 (75.8) 18.4 � 0.6 (18.9)0.1 1.145 � 0.002 (1.150) 22.0 � 0.2 (22.2) 73.2 � 0.5 (74.0) 18.8 � 0.1 (19.1)0.3 1.150 � 0.003 (1.153) 21.4 � 0.3 (21.7) 71.0 � 1.0 (72.6) 18.2 � 0.4 (18.7)0.5 1.166 � 0.004 (1.169) 21.9 � 0.1 (22.0) 70.9 � 0.2 (71.1) 18.3 � 0.6 (18.7)1 1.170 � 0.003 (1.174) 21.4 � 0.4 (22.9) 71.0 � 1.0 (72.5) 18.8 � 0.4 (19.2)2 1.175 � 0.004 (1.180) 22.6 � 0.4 (23.2) 70.0 � 2.0 (74.0) 19.7 � 0.7 (20.3)5 1.168 � 0.006 (1.178) 15.0 � 0.5 (16.3) 64.8 � 2.0 (66.8) 12.3 � 0.6 (13.1)


Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


supported by absorption measurements (Fig. S2B, ESI†) whichshow no shi in the absorption onset when changing thecomposition from 0% Sr to 2% Sr. Overall, the data suggest thatthe addition of SrI2 to the perovskite precursor solution has noappreciable effect on the bulk energy gap, consistent with whatwas reported previously,19,23 meaning that the observed increasein Voc must be related to reduced recombination at grainboundaries or interfaces.

Time-resolved and steady-state photoluminescence

Time resolved photoluminescence lifetime (TRPL) measure-ments were applied to investigate the effect of Sr addition on thecharge carrier recombination in the bare perovskite material.The TRPL traces for neat perovskite lms on glass substratescontaining different amounts of Sr, shown in Fig. 3a, are inaccordance with a double exponential decay model:36

IPLðtÞ ¼ A e�� tsfast

�þ B e

�� tsslow

�(1)

In eqn (1), the parameters A and B are the relative amplitudesfor the fast and slow lifetimes, sfast and sslow, respectively.

Commonly, the initial fast decay (below 20 ns) is associatedto the capture of charges by trap states and recombination atgrain boundaries or at the surface.37 An increase of the fastphotoluminescence lifetime, indicated by the rst initial decayin Fig. 3a, point out that the addition of Sr either reduces theconcentration of such traps acting as a passivating agent, orlimits the accessibility of these traps to photogenerated freecarriers. At the same time, TRPL traces also show an improve-ment of the slow (longer than 50 ns) photoluminescence life-time, suggesting that non-radiative recombination, commonlyassociated with the slow decay and most likely due to processeshappening at the surface,38 is considerably attenuated. Aneffective (amplitude averaged) photoluminescence lifetime seffwas then calculated according to the following equation:39

seff ¼ Asfast þ BsslowAþ B (2)

The effective lifetimes are plotted in Fig. 3b for the differentcompositions. With the addition of 2% Sr an extraordinarily


long PL lifetime of almost 1 ms is measured. The positive effectof Sr addition on the PL lifetime found here is qualitatively inagreement with the enhancement of carrier lifetime aerinsertion of Sr in MAPbI3 as deduced from transient microwaveconductivity experiments23 or with the increase of the PL life-time in perovskites of different compositions.24,25 Given the factthat the addition of Sr increases both PL lifetime and Voc, but ithas basically no effect on the shape and amplitude of the EQE orabsorption spectrum, we propose that Sr mainly reduces thestrength of non-radiative recombination. In order to quantifythis reduction of energy losses in the bare absorber, wemeasured the quasi-Fermi level splitting (QFLS) by means of theabsolute PL intensity measurements. Here, we make use ofWürfel's generalized Planck's law40 describing the non-thermalemission of a semiconductor:

IPLðEÞ ¼ 2pE2aðEÞ

h3c21

exp�E � DEF

kT

�� 1

(3)

Here, h is Planck's constant, c is the speed of light, E is thephoton energy, DEF the QFLS, a(E) is the spectral absorptivity,and k is the Boltzmann constant. Approximating the Bose–Einstein distribution with a Boltzmann distribution (E � DEF[ kT) and assuming the absorptivity a(E) ¼ 1 for energiesabove the band gap, the PL intensity IPL can be expressed asfunction of E where the QFLS, DEF, can be extrapolated froma t of ln(IPL/E

2) above the band gap:39

ln

�h3c2IPLðEÞ

2pE2

�¼ DEF

kT� E

kT(4)

Fig. 4b shows the QFLS maps calculated from the PL inten-sity mapping using eqn (4) for two samples of neat perovskitewith 0% and 2% Sr. In Fig. 4a the histograms for the corre-sponding QFLS distributions are reported. In the case of 2% Sra high QFLS of DEF ¼ 1.194 � 0.001 eV (FWHM ¼ 2.3 meV) wasextrapolated from absolute PL mapping, whereas in the case of0% Sr the quasi Fermi levels splitting was only DEF ¼ 1.167 �0.002 eV (FWHM ¼ 5 meV), indicating that the losses correlatedto non-radiative recombination are considerably lower and themaximum achievable Voc higher when Sr is incorporated. Thisnding is consistent with the observed increase of the PL



Fig. 3 (a) Time resolved photoluminescence (TRPL) traces of neat perovskite layers for the different Sr concentrations measured in inertatmosphere with a fluence of �30 nJ cm�2 at a wavelength of 470 nm, normalized to the initial transient peak. (b) Calculated effective lifetimesfrom fitting of the TRPL traces according to eqn (1) and (2).


Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


lifetimes. Also, the QFLS distribution of the Sr-containingdevice is signicantly narrower, indicating that the Sr-addition also reduces the spatial inhomogeneity of recombi-nation pathways across the perovskite lm.

Electroluminescence efficiency

A simple method to investigate radiative and non-radiativerecombination in a full device is to measure the external elec-troluminescence efficiency (EQEEL) by applying a forward biasto the solar cell in the dark operating it as a light-emitting diode(LED). Fig. 5 displays EQEEL for the 0% and the 2% samples for

Fig. 4 (a) Histogram of the quasi-Fermi level splitting (QFLS) extrapolatedrepresented in (b). The maps represent the QFLS distribution of the bare pThe standard deviation of the histograms is considered as the error of th


a range of applied voltages around the Voc and the corre-sponding injected dark currents.

When Sr is added, the EL efficiency is increased bymore thanone order of magnitude, reaching a remarkably high EQEELvalue of 2.5 � 10�3 for injection currents approaching Jsc underillumination, denoting stronger emissive behaviour andreduction of non-radiative recombination also in the completedevice structure. A luminescence efficiency of 0.25% is amongthe best values reported in literature so far.10,41 EQEEL values aretaken for an injected current density equal to Jsc (being a goodapproximation of the generation current JG), meaning that theso-determined EQEEL is a good measure for the radiativerecombination efficiency at Voc under 1 sun illumination and,

from the PL maps (5 mm2) of a neat perovskite layer with 0% and 2% Srerovskite film of an area of 5� 5 mm2 in a range from 1.14 eV to 1.2 eV.e given absolute value.




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


therefore, relevant for the calculation of the radiative and non-radiative losses presented here. Following the approach fromRau42 and applied to perovskite materials by Tress et al.,43 wecalculated exceptionally low non-radiative losses of

DVoc;non-rad ¼ kBTe ln�

1EQEEL

�¼ 0:161 eV at T ¼ 300 K when Sr

is present. The radiative Voc limit for this cell is

Voc;rad ¼ kBTe ln

JGJrad;0

!¼ 1:337 eV, where the dark radiative

current Jrad,0 was calculated from the convolution of the EQEPVspectrum and blackbody radiation (eqn (S2), ESI†) at 300 K(Fig. S3, ESI†), and JG is set equal to Jsc. Combining these valuesresults in a predicted Voc ¼ Voc,rad � DVoc,non-rad ¼ 1.337� 0.161¼ 1.176 V, which is very close to the measured averaged Voc for2% Sr cells. The perovskite without Sr shows higher non-radiative voltage loss DVoc,non-rad ¼ 0.232 eV with an almostidentical radiative limit Voc,rad¼ 1.338 eV, leading to a predictedVoc ¼ 1.338 � 0.232 ¼ 1.106 V, again in very good agreementwith the measured averaged Voc for 0% Sr cells. The non-radiative voltage losses DVoc,non-rad of the two samples differby 70 mV, matching exactly the Voc enhancement measured byJ–V scans. The small difference of only 15 meV found betweenthe QFLS and the Voc for the Sr containing sample indicates thatthe energy losses due to the implementation of the chargetransporting layer are successfully minimized.

Morphology characterisation

As noted above, previous work showed that Sr segregates at theperovskite surface.23,24 Fig. 6 shows the elemental distributions ofseveral elements of our perovskite samples on ITO/glass asmeasured by secondary ion mass spectrometry (SIMS), utilizing Oas primary ion source. According to Fig. 6a the elementscomprising the perovskite, such as Cs, Pb, and Rb, are

Fig. 5 The external electroluminescence efficiency EQEEL (right y-axis)and the injected dark current (left y-axis) for complete solar cells without(0% Sr) and with 2% of Sr as a function of the applied voltages. Dashedlines and circles indicate the applied voltage, where the dark injectedcurrent is equal to the Jsc under simulated AM1.5G illumination.


homogeneously distributed when Sr is not present. On the otherhand, the SIMS in Fig. 6b clearly proves a signicant enrichmentof Sr at the surface and interface to ITO, with the Sr concentrationbeing considerably lower in the bulk. In both measurements thesignal of the negatively charged ions is low due to limitations ofthe sputtering yield; however those species have been detectedwith Cs ions as primary ion source and they have also shownhomogeneous distribution (Fig. S4d, ESI†). To exclude the possi-bility of lateral inhomogeneities, we analysed different spots of ca.250 � 250 mm2 on the surface and from samples of differentbatches. In all measurements (Fig. S4a–S4c, ESI†), Sr traces showthe same inhomogeneous distribution with a signicant enrich-ment at the surface/interface. Following this picture, when the Sris added to the precursor solution and the lm is formed aerspin coating, Sr clearly segregates the surface/interface, whereasthe bulk remains almost unaltered. Additional XPS measure-ments (Fig. S5A, ESI†) show Sr on the sample surface with the riseof the characteristic Sr 3p peaks, in accordance to the SIMSresults. A more detailed XPS surface composition analysis(Fig. S5B, ESI†) quantitatively shows that the relative atomic Srconcentration compared Pb is 10 times higher than what shouldbe predicted by stoichiometry, in good agreement with the nd-ings by Perez Del Rey et al.23 Detailed analysis of the core-levelspectra results in the Sr/Pb molar ratio of 0.21 and the Sr/Imolar ratio of 0.07, which are both much higher than the ex-pected stoichiometry. This nding is in good agreement withSIMS traces and it conrms a Sr enriched region at the surface.Moreover, we can exclude a coverage of the surface with non-reacted SrI2, since the I concentration found at the surface, andconsequently its ratio with Sr, should be consistently higher thanwhat found here. The I/Pb molar ratio, for the sample with Sr, ishigher than sample without Sr, which can be possibly ascribed tothe fact that Sr can partially replace Pb, leading to a decrease ofPb/I ratio at the surface in the perovskite lattice compared to whatpredicted by stoichiometry for a full Pb perovskite.

Scanning electron microscopy (SEM) conrms a notablechange of the surface when Sr is added to the perovskite, asrepresented in Fig. S6A, ESI.† Upon Sr addition the surfacestarts to be characterized by leaf-like bright areas. Thesebrighter islands appear to be characterized by different workfunction aer imaging them through energy sensitive SEM in-lens detector. We notice that these features can be resolvedexclusively through this in-lens detector and not with an Ever-hart–Thornley lateral detector, being more topographicalsensitive. This excludes the possibility of having a non-conductive material covering the surface, otherwise appearingevidently with area of different brightness due to strongcharging effect in both detectors. A comparison between the twodifferent imaging techniques on the same surface spot is pre-sented in Fig. S6B, ESI.† Features of similar size and shape havebeen reported previously on the surface of Sr-containingCsPbI2Br perovskite processed from solution.24 The crosssection (Fig. S6A(d), ESI†) shows that grains propagate from thetop to the bottom even in the presence of the additional featurespresent on top of the layer. The results indicate that the topsurface of the Sr-containing perovskite contains a material ofa different composition spread across the surface area



Fig. 6 Secondary ion mass spectrometry (SIMS) profiles for a specifiedgroup of elements as a function of depth for a neat perovskite layer onITO/glass substrates (a) with no Sr added and (b) with 2% Sr. Note thatin both figures a thickness equal to zero represent the top surface ofthe perovskite layer, while the rise of the indium signal (highlightedwith a black dotted line) indicates that the perovskite layer is almostcompletely sputtered off and that the ITO substrate becomes exposed.The initial indium signal for the Sr containing Sr might originate frompin holes or non-uniform coverage and therefore being detected atthe beginning of the scan. This agrees with the smeared profile of thesignals in the Sr containing sample as an indication of a roughersurface as confirmed in SEM images in Fig. S6Ad.†


Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


increasing the roughness. A series of images representing thesurface aer addition of Sr at different concentrations (Fig. S6Cand S6D, ESI†) show that the density of these islands on thesurface is directly correlated with the concentration of Sr added.In fact, the surface of 5% Sr perovskite is almost entirely coveredwith these features, while the paddlestones-type surfacetopography of the underlying perovskite is barely visible. Giventhe fact that solar cells made with the 5% perovskite absorberexhibit signicantly lower FF and Jsc we conclude that thesefeatures can potentially have larger bandgap/lower electronconductivity than the underlying Pb-containing perovskite. Onthe other hand, X-ray diffraction patterns (XRD) (Fig. S6E†) ofa series of sample containing 0%, 2% and 5% Sr show nosignicant shis of the reections assigned to the lead-halideperovskite phase, indicating no major changes in the unit cellvolume. However, a new reection at �11.6� appears in the Srcontaining samples suggesting the presence of a new phase,most probably containing Sr. Importantly, the intensity of theperovskite reections decrease with increasing Sr concentra-tion, potentially indicating that this new phase segregates at thesurface of the lm and attenuates the X-ray signals coming fromthe underlying perovskite phase. The broadening of the perov-skite peaks suggests a reduction in crystal domain size or anincrease in microstrain due to a decrease in compositionaluniformity upon Sr addition.

Electronic structure characterisation

The results above provide evidence that adding Sr to a quadruplecation perovskite has a signicant effect on the perovskite


surface, whereas the bulk properties seem to remain fairly unal-tered, asking for a detailed investigation of the electronic struc-ture of the perovskite surface with and without Sr. Photoemissionand inverse photoemission spectroscopy (PES and IPES) experi-ments were performed for solar cell related multilayer stackscomprising ITO, the hole-transporting PTAA, and the activeperovskite, to retrieve information on surface work function(WF), ionization energy (IE), electron affinity (EA), as well asposition of the valence band maximum (VBM) and the conduc-tion band minimum (CBM) with respect to the Fermi level (EF).

As shown in Fig. 7a and c, the valence band and VBM, whenSr is incorporated, are shied to higher binding energy whilethe conduction band and CBMmove closer to EF [Fig. 7a and d].VBM and CBM were evaluated by linear extrapolation of thevalence and the conduction band onsets towards the back-ground on a linear intensity scale, respectively, as shown inFig. 7a. For the 0% Sr-perovskite we extrapolated the VBM to beat 1.42 eV and the CBM at 0.31 eV relative to the Fermi level,giving a band gap of 1.73 eV. This value is comparable with theoptical band gap of 1.63 eV reported for this type of quadruplecation perovskite.28 We note that an overestimation of the bandgap (from UPS and IPES) compared to the optical gap can be dueto the linear extrapolation,44,45 and our results are consistentwith examples reported in literature.44–47 A lower band gapwould be expected if the extrapolation was based on logarithmicplots, but due to the uncertainty of cross-section effects andrather large experimental broadening in IPES we refrain fromsuch procedures here. Most importantly, we observe a strongeffect on the electronic structure when Sr is incorporated. Inparticular, the CBM for 2% Sr perovskite locates at the Fermilevel, indicating a strongly n-type surface. Concomitantly, theVBM is shied towards higher binding energy, i.e., 1.92 eV(relative to EF). From these measurements, we deduce that theSr-enriched perovskite surface features a ca. 190 meV widerband gap compared to the surface of Sr-free perovskite, which isconsistent with the smaller electronegativity of Sr compared toPb. An increased band gap was predicted by DFT calculationsfor Pb2+ being completely replaced by Sr2+ in MAPbI3.22 Sinceour samples have a different composition, a quantitativecomparison of our results and theory is precluded.

The observation that the surface of the two perovskitesappear strongly n-type in photoemission measurements doesnot readily imply that the bulk is also strongly n-doped. As thepresence of surface states at semiconductor surfaces, whichinduces surface band bending in electronic equilibrium, ismore the rule than the exception,48 we need to attend to thisissue. In fact, for the prototypical methylammonium lead iodideperovskite the presence of, seemingly ubiquitous, Pb0 surfacestates has been evidenced.49,50 These donor-type surface statesdonate electrons to the region close to the surface, leavingbehind an accumulation of immobile positive charges at thesurface. This gives rise to a surface space charge region witha downward band bending from the (intrinsic) bulk towards thesurface in electronic equilibrium (i.e., in dark), termed surfaceband bending. Even if the density of these Pb0-related surfacestates can oen not be directly measured by PES experiments(due to limited sensitivity of the method), their presence is yet



Fig. 7 (a) UPS and IPES spectra for 0% Sr and 2% Sr containing perovskites with magnified valence and conduction band regions near EF. Thebinding energy scale is referenced to EF, set to zero. (b) Energy levels with respect to EF for 0% and 2% Sr perovskites obtained fromUPS and IPESspectra. From this picture, we observe a strong n-type character of the perovskite surface, as well as a larger band gap for the case of 2% Sr. Widerange (c) valence and (d) conduction band regions.


Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


manifested by surface downward band bending.51 Since it islikely that Pb0-related surface states also occur at the surface ofour perovskites under study, in the following we provide

Fig. 8 (a) Schematic representation of the SPV effect of 0% and 2% Sr samvacuum level (VL), conduction band minimum (CBM) and valence bandmination conditions, the space charge region is completely screened bypositions of the relevant energy levels are depicted with respect to the EFof bulk-to-surface distance for different illumination conditions. The bulenergy levels are obtained from measurements of the work function (Wobtained from the corresponding measurements under illumination, assu(b) Corresponding WF and VBM values from UPS, measured without andBoth samples display very similar VL and VBM positions under illuminatiosurface band bending is significantly more pronounced in the Sr-contaarrows.


evidence that while the bulk of our samples does not exhibitpronounced n-type character, the surface does due to surfaceband bending.

ples. Dotted lines represent the effect of white light illumination on themaximum (VBM) in the surface-near region. At sufficiently high illu-photogenerated charges and flat band conditions are established. Theof the substrate (independent on illumination), and plotted as functionk and surface energetics in dark are represented by solid lines. SurfaceF) and the valance band onsets in the dark, while the bulk values areming that flat-band conditions are established as reported in Zu et al.50

with simultaneous white light illumination for 0% and 2% Sr samples.n, indicating nearly identical bulk energy levels for 0% and 2% Sr, whileining sample due to a much larger SPV effect as indicated with pink




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


The observation of a strongly n-type surface of the Sr-containing sample raises the question whether the addition ofSr introduces n-doping also in the perovskite bulk. Unfortu-nately, ultraviolet (UV) photoemission is not suited to addressthis issue due to the short inelastic electron escape depth. Onthe other hand, the pronounced concurrent increase in PLlifetime and absolute PL yield upon Sr addition as reportedabove excludes the presence of a high density of doping-induced background charges in the Sr-containing sample,similar to what has been reported by Bolink and coworkers onSr-doped MAPbI3.23 To support this claim, we simulated theTRPL decay and PL efficiency of the neat perovskite for varyingdoping concentration, using realistic parameters for therecombination coefficients (Fig. S7A, ESI†). According to thesesimulations, if the increase in PL efficiency is due to doping, wewould expect a concurrent reduction of the PL lifetime, con-rming previous studies.52,53 The pronounced increase of bothPL decay times and PL efficiencies, therefore rules out extensivebulk doping in the presence of Sr. An effective lifetime of nearly1 ms combined with a large QFLS suggests a doping density ofless than 1015 cm3, meaning that the perovskite bulk is nearlyintrinsic.

The low doping density in the bulk combined with thestrongly n-type surface suggests signicant surface bandbending to occur. To provide experimental evidence for thiseffect, surface photovoltage (SPV) measurements were per-formed on neat perovskite layers with and without Sr beingpresent. This technique determines the shi of the surfaceelectrostatic potential between measurements done in darkand under light illumination, e.g., through Kelvin Probe orUPS (indicated by parallel shis of work function and valenceband). Under illumination, photogenerated carriers arecreated due to band-to-band (or trap-to-band) transition.These carriers will redistribute under the inuence of thespace charge eld, thereby compensating the local excess ofpositive charges at the surface. As a result, upon increasingthe illumination intensity the level of surface band bandingwill gradually decrease until at band condition is reached,as schematically represented in Fig. 8a. SPV is a commonmethod to measure the degree of surface band bending ofsemiconductors,54–56 including halide perovskites.57–59 Forexample, surface band bending in perovskite has beenproven with the same methodology used in our study for bareMAPbIxCl3�x lms. There, the effects of an n-type surfacewith downward band bending was clearly demonstrated50

and attributed to the presence of donor levels at the perov-skite surface, likely consisting of reduced lead (Pb0).

The effect of white light illumination on the WF and VBM issummarized in Fig. 8b (see Fig. S8 ESI† for the correspondingUPS spectra). In the case of 0% Sr, the WF increases from4.17 eV in dark to 4.93 eV under illumination (light intensitysufficient to saturate the shi, i.e., the establishment of atband conditions), accompanied with the VBM shiing from1.51 eV to 1.00 eV binding energy. VBM and WF do not shiperfectly in parallel, most likely due to surface inhomogeneities,which affect SPV the VBM and WF in SPV experiments differ-ently, but cannot be differentiated in the area-averaging


photoemission experiment. A much stronger SPV effect wasobserved for 2% Sr samples where the WF shied from 4.00 eVin dark to 4.89 eV under illumination, and the VBM from1.87 eV to 1.03 eV, respectively. We note that the SPV wasreversible for multiple illumination/dark cycles, thus photo-chemical reactions and degradation can be excluded. Addi-tionally, in this measurements we notice a reduction in workfunction due to Sr addition, in agreement with the effectspreviously reported in literature.23 A change of sample workfunction can have manifold reasons, e.g., due to formation ofdipoles at the surface,44 a change in stoichiometry,60,61 or it canbe indeed associated with a more n-type surface.62 In conclu-sion, SPV reveals a much larger degree of surface band bendingof the Sr containing perovskite, resulting in a stronger SPVeffect compared to the 0% Sr sample, see magenta colouredarrows in Fig. 8a and b. At the same time, both perovskitesexhibit very similar values of the WF and VBM under illumi-nation, demonstrating that the bulk energy levels (notably theposition of EF in the gap) is almost independent of the Srcontent.

Discussion

Before proposing a model to explain the benecial effect ofadding Sr to a quadruple cation perovskite, we rst summarizethe key ndings from our studies (Table 2). From SIMS and XPS,we nd a strong enrichment in Sr at the surface denoted in a Sr/Pb ratio of 0.21 being much higher than the expected 0.02 froma homogeneous Sr addition assumption. In TRPL, we nd thatSr addition strongly suppresses the initial fast PL decay, indi-cating reduced trapping, but also prolongs the long-term decayattributed to non-radiative recombination. In accordance withthis, we observe a ca. 30 meV increase in the quasi-Fermi levelsplitting in the neat perovskite under 1 sun equivalent illumi-nation conditions if 2% Sr is added. In the complete device, Srsuppresses non-radiative recombination, strongly enhancingthe electroluminescence efficiency 25-fold and reducing thenon-radiative voltage loss from 230 to 160 meV. On the otherhand, very similar results of the absorption, EQEPV, and XRDmeasurements suggest an only minor effect of Sr addition onthe perovskite bulk properties. This is consistent with theoutcome of the SPV experiments where upon illuminationbands atten at the same position. Meanwhile, the morepronounced shi of the workfunction indicates stronger bandbending in the Sr-containing samples. In combination, theseproperties must be the cause of the signicant enhancement ofVoc by 70 mV.

In Fig. 9 we schematically represent two perovskitesamples, without and with Sr, combining ndings from allprevious studies as a summarizing gure. We propose that thisspecic energetic landscape is responsible for a considerablereduction of surface recombination as indicated in Fig. 9.Notably, a strong n-type character of the surface goes alongwith a nearly complete occupation of the electron traps alreadyin the dark, reducing the probability that photo-generatedelectrons become trapped by these states, fully consistentwith the almost complete absence of an initial fast PL decay in



Table 2 Summary of the most relevant photovoltaic and optoelectronic properties determined for 0% Sr and 2% Sr containing perovskite filmsand/or solar cells. The insertion of Sr results in a consistent improvement of all parameters

Sr [%]Sr/Pb surfaceratio

seff[ns] DEF [eV] EQEEL [%] DVoc,non-rad [eV]

WF shi underillum. [eV]

VBM shi underillum. [eV] Voc [V]

0 0 180 1.16 0.01 0.23 0.76 0.51 1.112 0.21 980 1.19 0.25 0.16 0.89 0.84 1.18


Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


the Sr-containing layer. More important, the stronger surfaceband bending, as measured in SPV when Sr is added, repelsholes from the surface, with the benet of limiting the prob-ability of a captured electron in a surface state to recombinewith a hole. The most well-known example of this phenomenais the so-called “back-surface eld”63 which is an establishedapproach for suppressing surface recombination in inorganicsemiconductors. In addition, the accessibility of the surfacefor electrons and/or holes may be reduced by the wider bandgap of the Sr-containing surface perovskite region, similar towhat has been proposed recently for a compositionally engi-neered perovskite/HTL interface.16 The Sr induced surfacemodication and the formation of the islands on the topsurface may also act as a passivating agent, reducing thenumber of interfacial states responsible for trap-assistedrecombination. Even though these three effects cannot bedisentangled based on the presented set of experiments, theyall may act in combination together to reduce non-radiative

Fig. 9 Schematic representation of the two perovskite samples with ana combination of morphological studies, PES and SPV experiments it is pinducing a wider band gap and a more pronounced n-type character. Onshows very similar characteristics for both samples. The energy levels in thare based on values from UPS and IPES (Fig. 7, 8, S7a and S7b†), taking intowith respect to the EF of the substrate. In the sample without Sr the recowith C60. The Sr enriched surface repels holes from the surface reducingthe selectivity of the contact with C60. The lack of trap states in the schemare permanently occupied.


losses in actual devices by suppressing interface recombina-tion with C60, as proposed in Fig. 9.

In order to prove the benecial effect of Sr in reducing non-radiative recombination at the perovskite/C60 surface, absolutePL efficiencies were measured for perovskite lms on glass,without and with 2% Sr, additionally covered with a 30 nm thicklayer of C60 (see Fig. S10, ESI† for the results). As expected, themeasurements reveal a substantial decrease of the PL efficiencyin the presence of C60 for both samples. However, when Sr ispresent the decrease in PL attributed to the presence of C60 isstrongly reduced. This conrms our proposal that Sr additionimproves the device performance by reducing the strength ofthe non-radiative recombination specically at the surface ofthe absorber and at the interface with C60. However, themodication of the interface morphology and energetics uponSr addition might induce an extra barrier for extraction ofelectrons or holes, as extensively discussed for Fig. S9, ESI,†which may be a main cause for the systematic decrease in FF.

d without Sr addition with respective recombination schemes. Fromossible to draw a picture where Sr segregates at the perovskite surfacethe other hand, optical measurements and SPV suggests that the bulke bulk and at the surface for the 0% Sr and the 2% Sr cases, respectively,account the surface band bending found from SPVmeasurements, all

mbination of charges can happen at the surface or across the interfacethe non-radiative recombination at or across the interface and improvee of the 2% Sr sample illustrates surface passivation or that these traps




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


Conclusions

In summary, we show that the addition of a small amount ofSrI2 to the precursor solution of a hybrid quadruple cationperovskite induces are remarkable increase of the Voc withoutthe need of any post-deposition treatment or interlayer depo-sition. The attained Voc of 1.18 V is among the highest of pin-type perovskite solar cells with a C60 electron-transporting layerfor the given bandgap of 1.62 eV. We note that during revision ofthis manuscript, a publication has appeared reporting a Voc ofup to 1.21 V for a similar pin-type perovskite architecture,utilizing a 2-step composition engineering to alter the ener-getics of the perovskite surface,64 not required here. Transientand absolute PL measurements in combination with PESsuggests that the addition of Sr causes a drastic reduction ofnon-radiative recombination at the interface between theperovskite and the C60 electron-transporting layer. We assignthis improvement to a preferential n-doping of the perovskitesurface and to the concurrent formation of a space-chargeregion which, combined with a locally larger band gap at thesurface, reduces the access of photogenerated holes to theperovskite surfaces, similarly to the well-known back-eld effectin inorganic semiconductor. Adding a thin layer of an insu-lating polymer on top of the Sr-enriched perovskite surfaceimproved the Voc further to nearly 1.23 V, corresponding toa non-radiative voltage loss of only 110 meV. With that, ourwork not only demonstrates very high Voc values and efficienciesin Sr-containing quadruple cation perovskite pin devices, butalso highlights the importance of addressing and minimizingthe losses located at the interface with the transport layers inworking solar cells.

Materials and methodsDevice preparation

Patterned indium-doped tin oxide (ITO, Lumtec, 15 ohm sq�1)was washed with acetone, Hellmanex III, DI-water and iso-propanol. Aer microwave plasma treatment (3 min at 200 W)poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA, Sigma-Aldrich, Mn ¼ 7000–10 000, PDI ¼ 2–2.2) in a concentration of1.5 mg mL�1 was spin-coated at 6000 rpm for 30 seconds andimmediately annealed for 10 minutes at 100 �C. The perovskitelayer was formed by spin coating a DMF : DMSO solution (4 : 1volume) at 4500 rpm for 35 seconds. Aer 10 seconds of spincoating, 500 mL of diethyl ether (antisolvent) was dripped ontop of the spinning substrate. Aer spin coating samples wereannealed at 100 �C for 1 h. Aerwards, samples were transferredto an evaporation chamber and C60 (30 nm), BCP (8 nm) andcopper (100 nm) were deposited under vacuum (p¼ 10�7 mbar).The active area was 6 mm2 dened as the overlap of ITO and thetop electrode.

Current density–voltage characteristics and EQEPV

J–V curves were measured under N2 on a Keithley 2400 system ina 2-wire conguration with a scan speed of 0.1 V s�1 and voltagestep of 0.02 V. One sun illumination at approximately 100 mW


cm�2 of AM1.5G irradiation was provided by a Oriel class ABAsun simulator. The real illumination intensity was monitoredduring the measurement using a Si photodiode and the exactillumination intensity was used for efficiency calculations. Thesun simulator was calibrated with a KG5 ltered silicon solarcell (certied by Fraunhofer ISE). The AM1.5G short-circuitcurrent of devices matched the integrated product of the EQEspectrum within 5–10% error. The latter was recorded usinga home build set-up utilizing a Philips Projection Lamp(Type7724 12 V 100 W) in front of a monochromator (OrielCornerstone 74100) and the light was mechanically chopped at70 Hz. The photo-generated current was measured using a lock-in-amplier (EG&G Princeton Applied Research Model 5302,integration times 300 ms) and evaluated aer calibrating thelamp spectrum with an UV-enhanced Si photodetector (cali-brated at Newport).

Photoluminescence and electroluminescence

Time resolved PL data was acquired with a TCSPC system(Berger & Lahr) aer excitation with a pulse-picked andfrequency-doubled output from a mode-locked Ti:sapphireoscillator (Coherent Chameleon) with nominal pulse durations�100 fs and uence of �30 nJ cm�2 at a wavelength of 470 nm.

Hyperspectral Absolute Photoluminescence Imaging wasperformed by excitation with two 450 nm LEDs equipped withdiffuser lenses. The intensity of the LEDs was adjusted to �1sun by illuminating a contacted perovskite solar cell (shortcircuit) and matching the current density to the short circuitcurrent measured in the JV sun simulator (the measured shortcircuit current density of the solar cell under this illuminationwas 22.2 mA cm�2). The photoluminescence image detectionwas performed with a CCD camera (Allied Vision) coupled witha liquid crystal tuneable lter. The system was calibrated toabsolute photon numbers in two steps in a similar way to theprocess described by Delamarre et al.65 For this purpose an IRlaser diode and a spectrally calibrated halogen lamp wascoupled to an integrating sphere. The pixel resolution of theimages corresponds to about 10 mm in diameter. Sets of imagesfrom 650 nm to 1100 nm with 5 nm step size were recorded. Allabsolute PL measurements were performed on lms with thesame thicknesses as used in the operational solar cells.

The EL spectra were acquired using an Andor SR393i-Bspectrometer equipped with a silicon detector DU420A-BR-DD(iDus). The response of this setup was measured with a cali-brated lamp (Oriel 63355). The cells were kept under forwardbias with a Keithley 2400 at an applied current Jinj z Jsc,stabilized for 20 seconds before recording a spectrum. AbsoluteEL was measured with a calibrated Si photodetector (Newport)attached to a Keithley 485 pAmeter. The photodetector wasplaced directly in front of the device pixel, then a forward biaswas applied with a Keithley 2400 source-meter and the resultingphoton ux was calculated considering the EL spectrum of thesolar cell and the spectral response of the Si photodiode.Injected current and photodetector response were monitoredwith a home written LabVIEW routine varying the voltage and




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


stabilising for 20 seconds aer every voltage step (dV typically0.02 V).

Morphological characterization

SIMS measurements were performed using a Cameca IMS4finstrument, using O2+ and Cs+ as primary sputtering ionswithin an energy range of 5–15 keV. The scanned area on thesample surface was 250 � 250 mm2.

SEM images were acquired with a Zeiss Ultra Plus SEM.Images has been acquired through the use of Secondary Elec-tron in-lens detector.

The structural characterization was performed with a PAN-alytical X'Pert MPD Pro X-ray diffractometer in Bragg–Brentanoconguration using a Cu K-a radiation source (l ¼ 0.15406 nm)with a step size of 0.013�. The measurements were performed inN2 atmosphere.

Photoemission and inverse photoemission spectroscopymeasurements

Photoemission experiments were performed at an UHV systemconsisting of sample preparation and analysis chambers (bothat base pressure of 1 � 10�10 mbar), as well as a load-lock (basepressure of 1 � 10�6 mbar). All the samples were transferred tothe UHV chamber using a transfer rod under rough vacuum (1� 10�3 mbar). UPS was performed using helium discharge lamp(21.22 eV) with a lter to reduce the photo ux and to blockvisible light from the source to hit the sample. XPS was per-formed using Al Ka radiation (1486.7 eV) generated from a twinanode X-ray source. All spectra were recorded at room temper-ature and normal emission using a hemispherical SPECSPhoibos 100 analyzer. All perovskite layers were deposited ona ITO/PTAA stack in order to be as much close as possible to realdevice conditions. The resolution and energy calibration of thePES and IPES were determined by measuring the Fermi edge ofa clean Au (111) single crystal. The overall energy resolution was140 meV and 1.2 eV for UPS and XPS, respectively. The IPESmeasurements were performed in the isochromat mode usinga low-energy electron gun with a BaO cathode and a band passlter of 9.5 eV (SrF2 + NaCl). All presented PES and IPES spectraare given in binding energy (BE) referenced to the Fermi level.The overall energy resolution for IPES was 0.74 eV. The UPSspectra of thin lms under consecutive on–off illuminationcircles were conducted using a white halogen lamp at 150 mWcm�2 (daylight rendering spectrum). The same experimentalconditions and setup presented in the work of Zu et al.50 hasbeen used for our set of measurements. C60 molecules werepurchased from Novaled, and were used as received and ther-mally evaporated from resistively heated quartz crucibles. Thenominal deposited thickness was monitored by a quartz crystalmicrobalance.

Conflicts of interest

The authors declare no competing nancial interests.


Acknowledgements

S. A. acknowledges funding from the German Federal Ministryof Education and Research (BMBF), within the project “Mate-rialforschung für die Energiewende” (grant no. 03SF0540), andthe German Federal Ministry for Economic Affairs and Energy(BMWi) through the “PersiST” project (grant no. 0324037C).Additional funding came from HyPerCells (a Joint GraduateSchool of the Potsdam University and the HZB) and the GermanResearch Foundation (DFG) within the collaborative researchcenter 951 “Hybrid Inorganic/Organic Systems for Opto-Electronics (HIOS)”. We thank RTG Mikroanalyse GmbH,Schwarzschildstraße 1, 12489 Berlin for performing the SIMSmeasurements. The authors thank Sebastián Caicedo-Dávilaand Dr Ulrich Hörmann for fruitful discussions.

Notes and references

1 G. Hodes, Science, 2013, 342, 317–318.2 Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao andJ. Huang, Science, 2015, 347, 967–970.

3 S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou,M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza andH. J. Snaith, Science, 2013, 342, 341–344.

4 W.-J. Yin, T. Shi and Y. Yan, Appl. Phys. Lett., 2014, 104,063903.

5 H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl,A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum,J. E. Moser, M. Grätzel and N.-G. Park, Sci. Rep., 2012, 2, 591.

6 J.-P. Correa-Baena, A. Abate, M. Saliba, W. Tress, T. JesperJacobsson, M. Grätzel and A. Hagfeldt, Energy Environ. Sci.,2017, 10, 710–727.

7 W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo andS. I. Seok, Science, 2015, 348, 1234–1237.

8 D. Bi, C. Yi, J. Luo, J.-D. Décoppet, F. Zhang,S. M. Zakeeruddin, X. Li, A. Hagfeldt and M. Grätzel, Nat.Energy, 2016, 1, 16142.

9 W. S. Yang, B.-W. Park, E. H. Jung, N. J. Jeon, Y. C. Kim,D. U. Lee, S. S. Shin, J. Seo, E. K. Kim, J. H. Noh and S. IlSeok, Science, 2017, 356, 1376–1379.

10 C. M. Wolff, F. Zu, A. Paulke, L. P. Toro, N. Koch andD. Neher, Adv. Mater., 2017, 29, 1700159.

11 M. Stolterfoht, C. M. Wolff, Y. Amir, A. Paulke, L. Perdigón-Toro, P. Caprioglio and D. Neher, Energy Environ. Sci., 2017,10, 1530–1539.

12 W. Tress, Adv. Energy Mater., 2017, 7, 1602358.13 M. Stolterfoht, C. M. Wolff, J. A. Márquez, S. Zhang,

C. J. Hages, D. Rothhardt, S. Albrecht, P. L. Burn,P. Meredith, T. Unold and D. Neher, Nat. Energy, 2018, 3,847.

14 M. Stolterfoht, P. Caprioglio, C. M. Wolff, J. A. Márquez,J. Nordmann, S. Zhang, D. Rothhart, U. Hörmann,A. Redinger, L. Kegelmann, S. Albrecht, T. Kirchartz,M. Saliba, T. Unold and D. Neher, 2018, arXiv:1810.01333.

15 G. Huang, C. Wang, H. Zhang, S. Xu, Q. Xu and Y. Cui, J.Mater. Chem. A, 2018, 6, 2449–2455.




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


16 K. T. Cho, S. Paek, G. Grancini, C. Roldán-Carmona, P. Gao,Y. Lee and M. K. Nazeeruddin, Energy Environ. Sci., 2017, 10,621–627.

17 I. L. Braly, D. W. Dequilettes, L. M. Pazos-Outón, S. Burke,M. E. Ziffer, D. S. Ginger and H. W. Hillhouse, Nat.Photonics, 2018, 12, 355–361.

18 M. Abdi-Jalebi, Z. Andaji-Garmaroudi, S. Cacovich,C. Stavrakas, B. Philippe, J. M. Richter, M. Alsari,E. P. Booker, E. M. Hutter, A. J. Pearson, S. Lilliu,T. J. Savenije, H. Rensmo, G. Divitini, C. Ducati,R. H. Friend and S. D. Stranks, Nature, 2018, 555, 497–501.

19 M. T. Klug, A. Osherov, A. A. Haghighirad, S. D. Stranks,P. R. Brown, S. Bai, J. T.-W. Wang, X. Dang, V. Bulovi,H. J. Snaith and A. M. Belcher, Energy Environ. Sci., 2017,236, 236–246.

20 A. D. Jodlowski, C. Roldán-Carmona, G. Grancini, M. Salado,M. Ralaiarisoa, S. Ahmad, N. Koch, L. Camacho, G. DeMiguel and M. K. Nazeeruddin, Nat. Energy, 2017, 2, 972–979.

21 J. Navas, A. Sánchez-Coronilla, J. Jesús Gallardo, N. CruzHernández, J. Carlos Piñero, R. Alcántara, C. Fernández-Lorenzo, D. M. De los Santos, T. Aguilar and J. Mart́ın-Calleja, Nanoscale, 2015, 7, 6216–6229.

22 T. J. Jacobsson, M. Pazoki, A. Hagfeldt and T. Edvinsson, J.Phys. Chem. C, 2015, 119, 25673–25683.

23 D. Perez-del-Rey, D. Forgacs, E. M. Hutter, T. J. Savenije,D. Nordlund, P. Schulz, J. J. Berry, M. Sessolo andH. J. Bolink, Adv. Mater., 2016, 28, 9839–9845.

24 C. F. J. Lau, M. Zhang, X. Deng, J. Zheng, J. Bing, Q. Ma,J. Kim, L. Hu, M. A. Green, S. Huang and A. Ho-Baillie,ACS Energy Lett., 2017, 2319–2325.

25 H. Zhang, H. Wang, S. T. Williams, D. Xiong, W. Zhang,C. C. Chueh, W. Chen and A. K. Y. Jen, Adv. Mater., 2017,29, 1606608.

26 X. Shai, L. Zuo, P. Sun, P. Liao, W. Huang, E.-P. Yao, H. Li,S. Liu, Y. Shen, Y. Yang and M. Wang, Nano Energy, 2017,36, 213–222.

27 M.-C. Wu, W.-C. Chen, S.-H. Chan and W.-F. Su, Appl. Surf.Sci., 2018, 429, 9–15.

28 M. Saliba, T. Matsui, K. Domanski, J.-Y. Seo,A. Ummadisingu, S. M. Zakeeruddin, J.-P. Correa-Baena,W. R. Tress, A. Abate, A. Hagfeldt and M. Grätzel, Science,2016, 354, 206–209.

29 J.-P. Correa-Baena, M. Saliba, T. Buonassisi, M. Grätzel,A. Abate, W. Tress and A. Hagfeldt, Science, 2017, 358, 739–744.

30 Q. Wang, Y. Shao, Q. Dong, Z. Xiao, Y. Yuan and J. Huang,Energy Environ. Sci., 2014, 7, 2359–2365.

31 M. Jošt, E. Köhnen, A. B. Morales-Vilches, B. Lipovšek,K. Jäger, B. Macco, A. Al-Ashouri, J. Krč, L. Korte, B. Rech,R. Schlatmann, M. Topič, B. Stannowski and S. Albrecht,Energy Environ. Sci., 2018, 11, 3511–3523.

32 Q. Wang, Q. Dong, T. Li, A. Gruverman and J. Huang, Adv.Mater., 2016, 28, 6734–6739.

33 L. Calió, S. Kazim, M. Grätzel and S. Ahmad, Angew. Chem.,Int. Ed., 2016, 55, 14522–14545.


34 M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress,A. Abate, A. Hagfeldt and M. Grätzel, Energy Environ. Sci.,2016, 9, 1989–1997.

35 D. P. Mcmeekin, G. Sadoughi, W. Rehman, G. E. Eperon,M. Saliba, M. T. Hörantner, A. Haghighirad, N. Sakai,L. Korte, B. Rech, M. B. Johnston, L. M. Herz andH. J. Snaith, Science, 2016, 351, 151–155.

36 H. Cho, S.-H. Jeong, M.-H. Park, Y.-H. Kim, C. Wolf,C.-L. Lee, J. Hyuck Heo, A. Sadhanala, N. Myoung, S. Yoo,S. Hyuk Im, R. H. Friend and T.-W. Lee, Science, 2015, 350,1222–1255.

37 D. Shi, V. Adinol, R. Comin, M. Yuan, E. Alarousu, A. Buin,Y. Chen, S. Hoogland, A. Rothenberger, K. Katsiev,Y. Losovyj, X. Zhang, P. A. Dowben, O. F. Mohammed,E. H. Sargent and O. M. Bakr, Science, 2015, 347, 519–522.

38 Y. Yang, M. Yang, D. T. Moore, Y. Yan, E. M. Miller, K. Zhuand M. C. Beard, Nat. Energy, 2017, 2, 16207.

39 T. Unold and L. Gütay, in Advanced Characterization Techniquesfor Thin Film Solar Cells, ed. D. Abou-Ras, T. Kirchartz andU. Rau, Wiley-VCH, 2011, pp. 151–175.

40 P. Würfel, J. Phys. C: Solid State Phys., 1982, 15, 3967–3985.41 D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier,

K. Schenk, A. Abate, F. Giordano, J.-P. Correa Baena,J.-D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin,M. Grätzel and A. Hagfeldt, Sci. Adv., 2016, 2, 1501170.

42 U. Rau, Phys. Rev. B, 2007, 76, 085303.43 W. Tress, N. Marinova, O. Inganäs, M. K. Nazeeruddin,

S. M. Zakeeruddin and M. Graetzel, Adv. Energy Mater.,2015, 5, 1–6.

44 S. Olthof, APL Mater., 2016, 4, 091502.45 P. Schulz, E. Edri, S. Kirmayer, G. Hodes, D. Cahen and

A. Kahn, Energy Environ. Sci., 2014, 7, 1377.46 E. M. Miller, Y. Zhao, C. C. Mercado, S. K. Saha, J. M. Luther,

K. Zhu, V. Stevanović, C. L. Perkins and J. van de Lagemaat,Phys. Chem. Chem. Phys., 2014, 16, 22122–22130.

47 M. F. Lo, Z. Q. Guan, T. W. Ng, C. Y. Chan and C. S. Lee, Adv.Funct. Mater., 2015, 25, 1213–1218.

48 Z. Zhang and J. T. Yates, Chem. Rev., 2012, 112, 5520–5551.49 B. Conings, L. Baeten, C. De Dobbelaere, J. D'Haen, J. Manca

and H. G. Boyen, Adv. Mater., 2014, 26, 2041–2046.50 F. S. Zu, P. Amsalem, I. Salzmann, R. Bin Wang,

M. Ralaiarisoa, S. Kowarik, S. Duhm and N. Koch, Adv.Opt. Mater., 2017, 5, 1700139.

51 F. Zu, P. Amsalem, M. Ralaiarisoa, T. Schultz, R. Schlesingerand N. Koch, ACS Appl. Mater. Interfaces, 2017, 9, 41546–41552.

52 J. F. L. Salas, S. J. Heise, M. Richter, V. Gerliz, M. S. Hammer,J. Ohland and I. Hammer-Riedel, Thin Solid Films, 2017, 633,40–44.

53 J. F. L. Salas, M. Richter, J. Parisi and S. J. Heise, J. Appl.Phys., 2017, 122, 203103.

54 Z. Zhang and J. T. Yates, Chem. Rev., 2012, 112, 5520–5551.55 J. Liqiang, S. Xiaojun, S. Jing, C. Weimin, X. Zili, D. Yaoguo

and F. Honggang, Sol. Energy Mater. Sol. Cells, 2003, 79, 133–151.

56 L. Kronik and Y. Shapira, Surf. Interface Anal., 2001, 31, 954–965.




Publ

ishe

d on

09

Janu

ary

2019

. Dow

nloa

ded

by H

elm

holtz

Cen

tre

Ber

lin o

n 1/

22/2

019

11:4

7:25

AM


57 J. R. Harwell, T. K. Baikie, I. D. Baikie, J. L. Payne, C. Ni,J. T. S. Irvine, G. A. Turnbull and I. D. W. Samuel, Phys.Chem. Chem. Phys., 2016, 18, 19738–19745.

58 L. Barnea-Nehoshtan, S. Kirmayer, E. Edri, G. Hodes andD. Cahen, J. Phys. Chem. Lett., 2014, 5, 2408–2413.

59 W. Zhang, S. Pathak, N. Sakai, T. Stergiopoulos, P. K. Nayak,N. K. Noel, A. A. Haghighirad, V. M. Burlakov,D. W. Dequilettes, A. Sadhanala, W. Li, L. Wang,D. S. Ginger, R. H. Friend and H. J. Snaith, Nat. Commun.,2015, 6, 10030.

60 T. G. Kim, S. W. Seo, H. Kwon, J. Hahn and J. W. Kim, Phys.Chem. Chem. Phys., 2015, 17, 24342–24348.


61 H. Xie, X. Liu, L. Lyu, D. Niu, Q. Wang, J. Huang and Y. Gao,J. Phys. Chem. C, 2016, 120, 215–220.

62 Q. Wang, Y. Shao, H. Xie, L. Lyu, X. Liu, Y. Gao and J. Huang,Appl. Phys. Lett., 2014, 105, 163508.

63 A. Cuevas and D. Yan, IEEE Journal of Photovoltaics, 2013, 3,916–923.

64 D. Luo, W. Yang, Z. Wang, A. Sadhanala, Q. Hu, R. Su,R. Shivanna, G. F. Trindade, J. F. Watts, Z. Xu, T. Liu,K. Chen, F. Ye, P. Wu, L. Zhao, J. Wu, Y. Tu, Y. Zhang,X. Yang, W. Zhang, R. H. Friend, Q. Gong, H. J. Snaith andR. Zhu, Science, 2018, 360, 1442–1446.

65 A. Delamarre, L. Lombez and J.-F. Guillemoles, Appl. Phys.Lett., 2012, 100, 131108.



High open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509e

High open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509e

High open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509eHigh open circuit voltages in pin-type perovskite solar cells through strontium additionElectronic supplementary information (ESI) available. See DOI: 10.1039/c8se00509e

sustainable energy & fuels - perovskite info · 2019. 1. 30. · perovskite solar cells. we...

Documents