Interfacial engineering of electron transport layer using Caesium Iodide for efficient and stable organic solar cells

Accepted Manuscript Title: Interfacial engineering of electron transport layer using Caesium Iodide for efficient and stable organic solar cells Authors: Mushfika Baishakhi Upama, Naveen Kumar Elumalai, Md Arafat Mahmud, Matthew Wright, Dian Wang, Cheng Xu, Faiazul Haque, Kah Howe Chan, Ashraf Uddin PII: DOI: Reference:

S0169-4332(17)31198-4 http://dx.doi.org/doi:10.1016/j.apsusc.2017.04.164 APSUSC 35845

To appear in:

APSUSC

Received date: Revised date: Accepted date:

15-2-2017 11-4-2017 20-4-2017

Please cite this article as: Mushfika Baishakhi Upama, Naveen Kumar Elumalai, Md Arafat Mahmud, Matthew Wright, Dian Wang, Cheng Xu, Faiazul Haque, Kah Howe Chan, Ashraf Uddin, Interfacial engineering of electron transport layer using Caesium Iodide for efficient and stable organic solar cells, Applied Surface Sciencehttp://dx.doi.org/10.1016/j.apsusc.2017.04.164 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Interfacial engineering of electron transport layer using Caesium Iodide for efficient and stable organic solar cells Mushfika Baishakhi Upama, Naveen Kumar Elumalai*, Md Arafat Mahmud, Matthew Wright, Dian Wang, Cheng Xu, Faiazul Haque, Kah Howe Chan And Ashraf Uddin* School of Photovoltaic and Renewable Energy Engineering, University of New South Wales, 2052, Sydney, Australia. * E-mail address: [email protected], [email protected]

Highlights: 

Effect of pure CsI ETL in PTB7:PC71BM polymer solar cell is studied



CsI/ZnO bi-layer ETL is incorporated instead of CsI or ZnO ETL



CsI layer modifies energy level alignment at ITO/ZnO interface



New ETL improves both device efficiency and UVO stability

ABSTRACT: Polymer solar cells (PSCs) have gained immense research interest in the recent years predominantly due to low-cost, solution process-ability, and facile device fabrication. However, achieving high stability without compromising the power conversion efficiency (PCE) serves to be an important trade-off for commercialization. In line with this, we demonstrate the significance of incorporating a CsI/ZnO bilayer as electron transport layer (ETL) in the bulk heterojunction PSCs employing low band gap polymer (PTB7) and PC71BM as the photo-active layer. The devices with CsI/ZnO interlayer exhibited substantial enhancement of 800% and 12% in PCE when compared to the devices with pristine CsI and pristine ZnO as ETL, respectively. Furthermore, the UV and UV-ozone induced degradation studies also revealed that the devices incorporating CsI/ZnO bilayer possess excellent decomposition stability (~23% higher) over the devices with pristine ZnO counterparts. The 1

incorporation of CsI between ITO and ZnO was found to favorably modify the energy-level alignment at the interface, contributing to the charge collection efficiency as well as protecting the adjacent light absorbing polymer layers from degradation. The mechanism behind the improvement in PCE and stability is analyzed using the electrochemical impedance spectroscopy and dark I-V characteristics. Keywords: organic solar cell, Cesium Iodide, PTB7, stability, work function

1. Introduction Polymer solar cells (PSCs) are low cost, flexible, light weight, simple to fabricate.1-3 The power conversion efficiency has exceeded 10% with the development of novel low bandgap polymers.4-7 However, the stability of such cells still needs improvement for creating a worldwide market. Inverted structure of PSCs has exhibited better stability compared to the normal/ traditional structure by elimination of hygroscopic material such as PEDOT:PSS8 and highly reactive low work function metal electrodes such as Ca, Al.9 Inverted device structure normally uses transition metal oxide such as Molybdenum (VI) oxide (MoO3),10 Vanadium (V) oxide (V2O5)11 as hole transport layer and n-type metal oxide semiconductors such as Titanium dioxide (TiO2),12 Zinc Oxide (ZnO)13 as electron transport layer. As an electron transport layer, ZnO has become a popular choice due to its low temperature processing,14 good optical transparency (bandgap ~3.5 eV),15 excellent electron mobility (10-3-10-2 cm2/V.s),15-16 easy solgel fabrication and convenient energy alignment with fullerene acceptors.17 Despite the advantages of ZnO, it has some major drawbacks. The degree of crystallinity of such metal oxides directly influences the trap state distribution and conductivity of the transport layer.1820

Low temperature processed (~160 °C) ZnO shows lower degree of crystallinity, which

increases the depth of trap distribution inside the electron transport layer and degrades overall device performance.18 Furthermore, the chemisorption of oxygen raises the density of trap 2

states in ZnO films.21 Consequently, the electrons from light absorbing organic layer hop into these trap states instead of being collected at the Indium Tin Oxide (ITO) cathode. The interfacial trap-assisted recombination phenomenon adversely affects the charge extraction near cathode and subsequent reductions in device performance (short-circuit current density (Jsc), open circuit voltage (Voc) and fill factor (FF)). Another comparatively less used ETL in organic solar cell is Cesium Iodide (CsI).22-24 It is a water-based, environment-friendly material and previous reports suggest that CsI can modify the ITO cathode work function and improve its (ITO) charge/electron collection functionality as a cathode in inverted PSCs. The problem with CsI is, it segregates and Cs diffuses to the organic layer.24 Moreover CsI only ETL can also produce fractal-like aggregates in the organic layer which eventually degrade device performance.22 Therefore, it is highly indispensable to develop an efficient electron transport layer which not only improves the performance but also increases the device stability significantly. In search of an efficient electron transport layer for PSCs, in this work, we demonstrate a hybrid sol-gel processed CsI/ZnO bilayer as electron transport layer (ETL) between the cathode and the low band gap active layer, PTB7:PC71BM (a high performance organic donor-acceptor system).25 The solution engineered CsI/ZnO bilayer thus combines the benefits of their individual functionality while simultaneously overcoming the drawbacks of their material systems when used individually, as explained earlier. Together, the CsI/ZnO bilayer can be an efficient electron transporting layer when compared to pristine ZnO and CsI based devices – which are also shown in this study for comparison. To the best of our knowledge, this is the first study to report CsI/ZnO bilayer ETL in any kind of organic solar cells. The devices with this bilayer ETL show an average PCE of 7.3% which is 12.3% higher than that of a structurally identical control device with ZnO only ETL; the ZnO only devices showed an average PCE of 6.5% which fall in the range of PCE reported in previous literature, having similar device structure.14, 26 In addition, the bilayer ETL based devices have ~8 times

3

higher PCE compared to devices with CsI only ETL. We have employed X-ray photoelectron spectroscopy (XPS) to comprehend the electronic modification in the ITO cathode due to the incorporation of CsI layer between ITO and ZnO. Detailed impedance spectroscopy (IS) analysis has been employed to probe into the superior performance of CsI/ZnO devices by extracting the active layer/ETL interface contact resistance, device recombination resistance, chemical capacitance and flat-band potential, using the fitted data from Nyquist plots and MottSchottky (MS) curves. In addition, the UV and UV-ozone stability of devices with CsI/ZnO and ZnO ETL was also compared. Study of UV induced degradation is critical for devices employing ZnO ETL since oxygen is chemisorbed on ZnO layer and can be released upon UV illumination.27-28 The chemisorption induces high energetic disorder at grain boundaries of ZnO layer.21 Such oxygen incorporation cannot be avoided in an industrial environment since the coating process is conducted in air. Besides, after UVO exposure, the ITO cathode releases oxygen over time which can degrade the underneath layers in the long term.29 UVO treatment was also carried out on the fabricated devices, in order to purposefully degrade the cells by introducing oxygen. Incorporation of oxygen from environment can potentially degrade the photo absorbing layer (PTB7),30 as well as the ZnO layer.21 The improved UV and UVO stability of the CsI/ZnO devices over ZnO only devices were analyzed and explained with the help of Impedance Spectroscopy and dark IV characteristics of the fabricated devices.

2. Experimental Section 2.1 Material and fabrication of solar cell Various inverted organic solar cells were fabricated with the structure: ITO/ ETL/PTB7:PC71BM/MoO3/Ag. The ETLs were plain ZnO or plain CsI or CsI/ZnO bilayer.

4

Patterned ITO glass substrates (12 mm × 12 mm, Lumtec) were cleaned via ultrasonication in soapy DI water, DI water, acetone and isopropanol, each for 10 min. ZnO precursor was formed by dissolving zinc acetate dihydrate (Zn(CH3COO)2•2H2O, Aldrich, 99.9%) and ethanolamine (MEA, NH2CH2CH2OH, Aldrich, 99.5%), in 2-methoxyethanol (CH3OCH2CH2OH, Aldrich, 99.8%). CsI precursor solutions of various concentrations were prepared by dissolving CsI (Sigma Aldrich) in 2-methoxyethanol (CH3OCH2CH2OH, Aldrich, 99.8%). Both CsI and ZnO precursors were spin-coated at 4000 rpm for 60 s. After spin coating ZnO, the samples were annealed on a hotplate in air at 170 °C for 30 min. For CsI layer deposition, the CsI coated samples were annealed in N2 environment at 150 °C for 5 min. PTB7 (1- Material, Inc.; 20.7 mg mL-1) and [6,6]-Phenyl C71 butyric acid methyl ester (PC71BM) (1- Material, Inc.; 31.0 mg mL-1) were dissolved in chlorobenzene (polymer: fullerene ratio: 1:1.5) with 3.0 vol% 1, 8diiodooctane in a N2-filled glovebox and stirred overnight at 60 °C. The solution was then cooled to room temperature and spin-cast on top of the ETL coated films at 900 rpm for 120 s. 10 nm MoO3 and 100 nm Ag layers were thermally evaporated on the active layer at a background pressure of 1 × 10−5 mBar. The device area used was 0.12 cm2, as defined by a shadow mask. At least 12 devices were made for each case. 2.2 Measurements and characterization The current density–voltage (J–V) measurements were performed using a solar cell I–V testing system from ABET technologies, Inc. (using a Keithley 2460 source meter) under illumination power of 100 mW/cm2 by an AM 1.5G solar simulator (Sun 3000 Solar Simulator; 100 mW/cm2). For optical characterization of the electrodes, a UV–VIS-NIR spectrometer (Perkin Elmer – Lambda 950) was used. X-ray diffraction (XRD) with Cu Kα radiation was done at an angle ranging from 20° to 60° by step-scanning with a step size of 0.02°. External quantum efficiency (EQE) measurements were performed using a QEX10 spectral response system from PV measurements, Inc. The film surface morphology was observed by atomic force 5

microscopy (AFM) with a Bruker Dimension ICON SPM. X-ray photoelectron spectroscopy (XPS) was carried out in ultrahigh vacuum environment (Thermo Scientific, UK; Model: ESCALAB 250Xi). Al Kα radiation source (hν= 1486.68 eV) and Thermo Scientific™ Avantage Software were used for the XPS analysis. The impedance measurements were performed with an Autolab PGSTAT-30 equipped with a frequency analyzer module in the frequency range of 106-1 Hz. AC oscillating amplitude was as low as 20 mV (rms) to maintain the linearity of the response. Bioforce’s UV-ozone Procleaner™ Plus system was used for UV and UV-ozone accelerated ageing test.

3. Results and Discussion 3.1 Effect of CsI ETL on PTB7:PC71BM solar cell First, the CsI only devices were made by spin-coating varied concentrations of CsI as electron transport layer (ETL), on top of ITO cathode, in PTB7:PC71BM cells. The concentrations are 0.1, 0.5, 1 and 2 mg/ml. These devices showed very low photovoltaic performance with highest average PCE of 0.92% only. The photovoltaic parameters of these cells are summarized in Table 1. Fractal-like aggregates are formed near the ITO cathode after active layer deposition. Such aggregates from CsI layer have also been reported in P3HT:PCBM solar cells by Xiao et al.22 The density of the aggregates is largely dependent on the CsI concentrations. Figure S1 displays the photographs of PTB7:PC71BM films deposited on top of the CsI coated ITO glass substrates. As can be seen, visible aggregates are formed and their density increases with CsI concentration. The PTB7:PC71BM films seem to decompose and lose their original dark color when CsI concentration exceeds 0.5 mg/ml. The film with 0.1 mg/ml CsI retains PTB7:PC71BM film color, but it still contains aggregates. The aggregates are responsible for the poor photovoltaic performance of PTB7:PC71BM solar cells with CsI only ETL. 6

The CsI only devices with low CsI concentrations (≤0.5 mg/ml) show very poor device performance (0.67% and 0.92%) (Table 1). The short circuit current density (Jsc) is moderate (around 10 mA/ cm2) which indicates the photoactive layer was less affected by aggregate formation. However, the fill factor (FF) and open-circuit voltage (Voc) are extremely low which can stem from defective interlayer between photoactive layer and ITO cathode. No significant photovoltaic behavior can be seen in devices with high CsI concentrations; there was not even any notable photocurrent in such devices. This observation supports the images presented in Figure S1 where the photoactive layer can be seen to decompose when CsI concentration exceeds 0.5 mg/ml. 3.2 Solar cell performances with CsI/ZnO ETL Next, PTB7:PC71BM devices were made with ZnO layer and CsI/ZnO bilayer as ETL. Since films with 0.1 mg/ml CsI concentration showed lowest decomposition and reasonable photovoltaic performance compared to other films (especially at high concentrations), this concentration has been used during the preparation of CsI/ZnO bilayer and reported as the CsI layer in rest of the paper. Figure 1(a) displays the device architecture of the fabricated cells. The J-V characteristics of the cells from both structures are presented in Figure 1(b). From Figure 1(b), it can be seen that the values of Jsc of CsI/ZnO devices are higher than that of the ZnO devices. Detail photovoltaic parameters of the cells are summarized in Table 2. The Jsc increases from 17.02 mA/ cm2 in ZnO devices to 17.64 mA/ cm2 in CsI/ZnO devices. The FF increases from 52.9% in ZnO devices to 56.6% in CsI/ZnO devices. The Voc also increases slightly from 722.2 mV in ZnO devices to 726.4 mV in CsI/ZnO devices. The efficiency increases by 12.3% for the CsI/ZnO devices. To rule out the possibility of the increment in device parameters originating from process related variations, the average data has been

7

collected from more than 12 devices for each case. The device parameters with error bars are displayed in detail in figure S2.

3.3 Surface and bulk properties: Transmission, XRD, EQE and AFM Figure 2 displays the transmission and X-ray diffraction (XRD) spectra of plain ZnO and CsI/ZnO bilayer thin film on top of the ITO coated glass substrate. In the transmission spectra (Figure 2(a)), there is a broad absorption dip near 500 nm which originated from the absorption of sputtered ITO electrode on top of the glass substrate.31 No significant reduction in transmission is observed due to the addition of CsI layer in between ITO and ZnO layer. A small decrease can be seen near 410 nm wavelength. However, this does not affect the absorption of the photovoltaic layer since PTB7 has peak absorption between 570 and 700 nm.32 Figure 2(b) shows the XRD spectra of both plain ZnO and CsI/ZnO bilayer thin film. The XRD measurement of plain ZnO confirms the characteristic peaks of the hexagonal wurtzite ZnO phase.13, 18 However, the (100) and (101) peak show relatively broader peak width compared to that of (002). Lin et al.13 showed that such broadening can happen in low temperature annealing of ZnO due to reduced crystallinity. Despite low crystallinity, such solgel ZnO buffer layer can contribute to high efficiency of organic solar cell since other significant factors, such as morphology, energy level alignment, substrate coverage and the contact quality (roughness) of ZnO layer with the active layer at the interface, play more vital role in determining the performance of OPV devices.14 In the CsI/ZnO bilayer, the (002) peak is diminished, however two new peaks appear at the (110) and (211) orientation. The new peaks belong to the preferred peaks for CsI crystal.33 Hence, although the CsI concentration is quite low (0.1 mg/ml), the XRD scan confirms the presence of CsI crystal in the bilayer ETL. 8

EQE profiles of ZnO and CsI/ZnO devices are shown in Figure S3. There is a slight reduction in the EQE of CsI/ZnO devices at short wavelengths (390-440 nm), which can be related to the reduction in light transmission through CsI/ZnO films (Figure 2(a)). The EQE of CsI/ZnO device starts to increase after 500 nm and remains higher than that of ZnO devices in the midwavelength regions (500-700 nm). Since both ZnO and CsI/ZnO device structures share similar active layer (PTB7:PC71BM) and the transmission through ZnO and CsI/ZnO films are also similar in this region, the increase in EQE of CsI/ZnO device is not coming from increased absorbance inside the photoactive layer. Rather, it is related to efficient charge extraction and transport mechanism, which will be discussed in details later in this paper. Due to increased EQE, the Jsc in CsI/ZnO device also increases, as displayed in Table 2. Figure 3 displays the AFM images of the ZnO only and CsI/ZnO films. The RMS surface roughness for the ZnO layer, extracted from the AFM analysis, is 3.64 nm, which matches the value of literature for similar concentration of ZnO pre-cursor solution.13 After applying a CsI layer in between ITO and ZnO layer, the surface roughness decreases to 2.93 nm. Since the topography of a thin so-gel processed ZnO is largely influenced by the surface roughness of ITO,13 it is possible that the CsI is helping to reduce the ITO surface roughness, which is intrinsically rough (SRRMS≈ 6.99 nm). It is well-known that reduced surface roughness can create a better contact between the ZnO buffer layer and the organic active layer (here, PTB7:PC71BM) and reduce the contact resistance.34-35 The reduced surface roughness of CsI/ZnO buffer layer can play a key role in increasing the fill-factor of devices with this buffer layer (FF= 56.6%) compared to that of devices with ZnO only ETL (FF= 52.9%). 3.4 Modification of energy level alignment via CsI layer To find any modification at the CsI/ITO interface that can enhance the device photovoltaic performance, X-ray photoelectron spectroscopy (XPS) was employed. The XPS spectra of

9

plain ITO and CsI-coated ITO, over a binding energy range of 0-1350 eV, are displayed in Figure 4 (a). The Cesium (Cs) peak is present in the survey scan of ITO/CsI film, along with other characteristics peaks of the ITO layer, which is another evidence of the presence of CsI on top of the ITO, although the used CsI concentration was as low as 0.1 mg/ml. Figure 4(b) shows Cs 3d and O 1s spectra and their characteristics peaks from the XPS study of CsI film on ITO. All the characteristics peaks have been detected and analyzed using Thermo Scientific™ Avantage Software. CsI film shows a lower binding energy of Cs (~724.5 eV) on ITO, even after annealing, compared to that of a Cs2CO3 film on ITO reported in literature.36 In case of Cs2CO3, the higher binding energy is attributed to the bonding of In and Sn to oxygen with Cs.36 Since CsI is devoid of oxygen, such bonding cannot be found at ITO/CsI interface. From the O 1s spectra in Figure 4(b), the O 1s spectra show three characteristic peaks at 530, 531.22, and 532.40 eV. The first peak is assigned to the ITO. The other two peaks are at nearly similar positions as observed by Liao et al.36 in Cs2CO3 on ITO. The observation of the later peaks suggests formation of Cs-O bond at the ITO/CsI interface. Xiao et al.22 also proposed such bond formation in solution-processed CsI interlayer and ITO, which can modify the ITO work function and improve cell performance. To divulge further into the device electronic properties contributing to augmented PCE of CsI/ZnO devices over ZnO devices, we have also conducted Mott-Schottky analysis. Figure 5(a) inset illustrates the Mott–Schottky plots of ZnO and CsI/ZnO devices at 10 kHz frequency under dark. The x-axis intercept of the extrapolated linear section of the Mott–Schottky curve provides the flat band potential. As can be seen from the zoomed Mott-Schottky curve in Figure 5(a), the flat band potential for CsI/ZnO devices (~710 mV) is slightly higher by 50 mV than that of ZnO devices (~660mV). Flat-band potential, Vfb, can be defined as the potential required to compensate the energetic differences between the quasi Fermi levels of the active layer and ETL/cathode electron selective contact (EP and EFn respectively):37 10

𝑞𝑉𝑓𝑏 = 𝐸𝐹𝑛 − 𝐸𝑃 ,

(3)

The active layer is similar in both device structures (PTB7:PC71BM), hence the variation in Vfb is related to the quasi Fermi level, EFn, of ETL/cathode; not EP. Since the Vfb in CsI/ZnO devices is higher than that in ZnO devices, the EFn position in CsI/ZnO devices has an upward shift. Such spatial variation in the quasi Fermi level indicates that, the work-function of ETL/Cathode work function is reduced when ITO is modified by a thin layer of CsI, which is a pre-requisite for more efficient electron extraction.38 This reduction in ITO work function can be attributed to the Cs-O bond at the ITO/CsI interface, as observed in the XPS spectra (Figure 4(b)). The high Vfb can also be attributed to the slight improvement observed in the Voc of CsI/ZnO devices (Table 2). The device Voc can be defined as the energetic offset between the quasi Fermi levels of ETL/cathode contact, EFn and the quasi Fermi level of HTL/anode contact, EFp:39-40 𝑞𝑉𝑜𝑐 = 𝐸𝐹𝑛 − 𝐸𝐹𝑝 ,

(4)

Since, the HTL (MoO3) remains constant for both device structures, the improvement in Voc in CsI/ZnO devices can be attributed to the upshift of quasi Fermi level of ETL/cathode contact due to the Cs-O bond formation. The mechanism of ITO work-function reduction is explained in Figure 6. According to the band-bending concept,41 when a metal and a semiconductor are in contact, there will be transfer of free electrons between the metal and semiconductor due to the work function difference. In this explanation, we have considered ITO as metal due to its n-type highly degenerate semiconductor behaviour with a wide band gap in the range between 3.5 and 4.3 eV.42-44 Here, the ITO work-function is higher than the ZnO work-function; hence the electrons will flow from ZnO to ITO until there fermi levels (EF,ZnO and EF,ITO) are aligned. Electron concentration near the ZnO surface is depleted compared with the bulk and creates a space charge region. In this region, the energy band edges in the ZnO are shifted continuously due to the electric field

11

between ZnO and ITO due to the charge transfer, also known as “band bending”.41 The energy bands bend upward toward the interface of ITO and ZnO (Figure 6(a, b)). The Schottky barrier formed at the metal−semiconductor interface, 𝜙𝐵1 can be defined as the difference between the work function of ITO (WFITO) and electron affinity of ZnO (𝜒𝑠 ): 𝜙𝐵1 = 𝑊𝐹𝐼𝑇𝑂 − 𝜒𝑠 ,

(1)

In order to decrease the Schottky barrier for better electron transport at the interface, the ITO work function (WFITO) needs to be reduced. When CsI layer is applied on ITO, WFITO is reduced by the formation of strong O-Cs dipole layer at the ITO/CsI interface (Figure 6(c, d)).45 ITO surface with O-Cs dipole layer has low effective work function owing to the highest dipole moment. As such, the Schottky barrier reduces from 𝜙𝐵1 to 𝜙𝐵2 , 𝜙𝐵2 <𝜙𝐵1 . Reduced ITO work function is denoted by WFITO” in Figure 6(c, d). 𝜙𝐵2 = 𝑊𝐹𝐼𝑇𝑂" − 𝜒𝑠 ,

(2)

The reduction in Schottky barrier can be correlated to the increase in Jsc and fill factor because of superior interfacial charge transfer rate at the cathode. Voc of inverted solar cells also improves due to the lower cathode work function.45 The improvement in Voc is also demonstrated by Mott-Schottky analysis in figure 5(a). 3.5 Impedance analysis of fresh CsI/ZnO devices To further understand the improved fill factor for devices with CsI/ZnO ETL, AC impedance measurements were carried out at high bias voltage (650 mV), close to the open-circuit voltage (Voc). At low bias (0.3 V or lower), the trap states present in the photo-active layer become quite influential and respond to the modulation signal, affected by trapping-detrapping process. On the other hand, at high bias, these traps are suppressed by strong charge carrier injection.46 Hence, the effect of traps can be avoided and useful information regarding the diffusivity and mobility of injected carriers can be achieved. Figure 5(b) displays the impedance spectra in 12

dark conditions from 1 Hz up to 1 MHz for the PTB7:PC71BM devices with ZnO only and CsI/ZnO ETL. Measured data resemble the typical semicircle shape that can be accurately modeled with a parallel R-C circuit.47-48 The cole-cole plots, in both cases, are composed of two arcs at low and high frequency regions, which require two sets of parallel R-C elements. Belén et al.49 proposed that the low frequency arc is attributed to the charge accumulation that cannot be extracted through the contacts. One limitation of Impedance Spectroscopy is, it cannot determine in which electrode (cathode or anode) the accumulation is taking place.49 There has been reports of charge accumulation at ZnO interface under both light and dark conditions.49 Exposure of anode to air and water in inverted devices can also result in charge accumulation phenomenon.50 In summary, the low-frequency arc in the Nyquist plot is responsible for device resistivity and charge accumulation near electrode. Figure 5(b) shows that the low-frequency arc of the ZnO devices is bigger compared to that of CsI/ZnO devices, which indicates slow diffusion of carriers and charge accumulation near the electrode. Since both device structures share the same anode, the reduction of low-frequency arc in CsI/ZnO devices is largely coming from the addition of a CsI layer before ZnO layer. The measured results were then fitted with an equivalent electrical circuit, shown in inset of Figure 5(b). The fitted data are displayed as straight lines in Figure 5(b). The equivalent circuit consists of two parallel R-CPE elements connected in series where R is resistance and CPE (Constant Phase Element) is a non-ideal capacitance that takes into account non-homogeneities such as porosities, roughness and surface states.47 In the circuit, Rs represents the series resistance, which includes the resistance of metallic wires, ohmic components, for example, ITO and Ag electrodes.51-52 Rt can be interpreted as the resistance associated with electron transport and Cg represents the dielectric component of the diode.48 Rrec accounts for the recombination resistance,48 in this case only non-geminate recombination since the measurement was carried out in dark and no photocarriers were generated. Cµ is the distributed chemical capacitance48

13

which suggests charge accumulation that cannot be effectively extracted at the contacts.49 In organic solar cells, the capacitance is dominated by the charge carriers injected from the contact. Chemical capacitance (Cµ) is also known as “diffusion capacitance” and it normally exceeds the space charge region capacitance at higher forward voltages.53 Cµ is related to the variation of the electron chemical potential in the electron transport layer “plate of the capacitor”,54 in our case, ZnO and CsI/ZnO layer. The fitted parameters of the equivalent circuit are summarized in Table 3.

From table 3, it can be observed that the series resistance Rs is slightly increased in the CsI/ZnO relative to ZnO devices, which could be attributed to the additional spatial thickness of the electron transport layer i.e. combined CsI/ZnO as opposed to ZnO only ETL. However, the slight increase in the series resistance is compensated by the significant reduction of transport resistance Rt which is about 30% lower in the case of CsI/ZnO device compared to that of ZnO only devices. The reduction in Rt suggests enhanced interfacial charge transfer rate at the cathode45 and contributes to the improved Jsc and FF for the CsI/ZnO devices (Table 2). The reduction in ZnO surface roughness (Figure 3) and the modification of ITO work-function45 (Figure 6) play a key-role in reducing the transport resistance. The electron diffusion or transit time, τd, is the product of electron transport resistance, Rt and chemical capacitance, Cµ. τd has been calculated for the fabricated devices (Table 3). The electron diffusion in CsI/ZnO is 1.4 times faster than that in ZnO devices, which means electron diffusivity is higher in CsI/ZnO devices. The electron diffusivity is calculated using the following formula:48 𝐷𝑛 =

𝐿2 𝜏𝑑

,

(5)

14

Here, L is the thickness of the diffusion layer, L = 200 nm. Electron diffusivity in CsI/ZnO device is one order higher than that in ZnO device. The electron mobility is calculated from the diffusivity, Dn, using the following relation:48 µ𝑛 =

𝑒×𝐷𝑛 𝑘𝑇

,

(6)

Here e is the charge on an electron and kT is thermal energy. The electron mobility is increased by 40% in CsI/ZnO devices than that in ZnO devices. Charge carrier mobility is a crucial parameter in determining the fill factor of an organic solar cell.55-56 The increase in mobility of CsI/ZnO devices correlates with the increase in device fill factor, displayed in Table 2. 3.6 UV and UVO stability enhancement Apart from the efficiency enhancement, the CsI/ZnO devices also show better UV and UVozone (UVO) stability compared to the ZnO devices. To manifest the stability enhancement, UV and UVO-assisted acceleration aging tests were carried out. During the UV-only ageing process, the devices were encapsulated so that moisture and oxygen cannot degrade the PSCs. The variations in photovoltaic parameters of fresh and degraded cells (UV and UVO-assisted) are summarized in Table 4. The J-V characteristics curves of the degraded cells (with both ZnO and CsI/ZnO ETLs) are plotted in Figure 7(a). From Table 4, it can be seen that, the Jsc in all devices (both ZnO and CsI/ZnO) decreases sharply under both UV and UVO exposure. In ZnO devices, all the device parameters (Jsc, Voc, FF, and PCE) go down. Jsc, Voc and FF drop by 14.8%, 3.3% and 3.1% under UV exposure and 17%, 5.9% and 8.5% under UVO exposure. The reduction in Jsc is higher in UVO-degraded cells since the ZnO and active layers are degraded by not only UV illumination but also the presence of oxygen.28, 57-58 As a result, the PCE falls from 6.5% to 5.19% (under UV) and 4.66% (under UVO). However, in CsI/ZnO devices, the drop in FF and Voc is negligible which suggests that UV and ozone induced degradation has smaller impact on 15

the cathode of CsI/ZnO devices, compared to that on the cathode of ZnO devices. The PCE largely drops due to the drop in Jsc. Even then, the percentage of PCE drop in CsI/ZnO devices is lower than that in ZnO devices (from PCE= 7.3% in fresh cells to PCE= 6.03% (under UV) and PCE= 5.8% (under UVO)). After UV and UVO induced degradation, the CsI/ZnO devices retain higher PCEs than ZnO devices. An interesting observation in the degradation pattern is that, the Jsc of CsI/ZnO devices, after both UV and UVO exposure, has similar values (15.17 and 15.39 mA/cm2 respectively). Under UV illumination, CsI/ZnO devices have higher photocurrents compared to that of ZnO devices (15.17 and 14.5 mA/cm2, respectively). The reason can be attributed to lower energetic disorder in ZnO ETL, coming from chemisorbed oxygen, since Cs-O bonds are formed at the CsI/ZnO contact. Hence, the degradation in the Jsc is coming only from the UV-induced oxidative degradation inside the photo active layer57-58 and not the disorder induced trap states in ZnO ETL. Under UVO exposure, the Jsc of ZnO devices drops to even lower value (14.1 mA/cm2). Surprisingly, the Jsc does not alter at all in CsI/ZnO devices. The pattern in Jsc is similar to the CsI/ZnO cells under UV-only illumination, although ozone is present in UVO exposure. It seems that, the CsI/ZnO devices, to some extent, are even immune to the externally introduced oxygen, as if the CsI layer is working as an oxygen-blocking layer. To find more evidence for less oxygen induced traps in ZnO in CsI/ZnO devices, dark J-V characteristics of the ZnO and CsI/ZnO devices were investigated. The dark J-V curves are depicted in Figure 7(b). A large variation can be seen in the dark currents between ZnO and CsI/ZnO devices, with ZnO devices (under both UV and UVO exposure) showing orders of magnitude higher leakage current in the low and reverse bias region, as compared to CsI/ZnO devices. This leakage current can be defined as an undesirable current, injected from the electrodes prior to the turn on voltage. Since its direction, under operating condition (0V to Voc), is opposite to the photocurrent, it reduces the photocurrent of the device.59 The only 16

change in the proposed device structures under comparison (Figure 1(a)) is the electron transport layer. The higher leakage current in ZnO devices is possibly coming from the deep localized trap states inside the ZnO crystalline structure, emerging from chemisorption of oxygen.21 These traps prevent efficient charge extraction from the photoactive layer due to electron hopping from active layers into the deep traps of ZnO. The leakage current is also inversely proportional to the device shunt resistance (Rsh). The relationship between Rsh and leakage current from electrode (Jsh) can be expressed in following equation:59 𝐽𝑠ℎ =

𝑉−𝐽𝑅𝑠 𝑅𝑠ℎ

,

(7)

Here, J is the net output current density of the solar cell, V is the applied bias and Rs is series resistance. The lower the magnitude of Rsh, the higher the leakage current that runs through it. High leakage current reduces the overall shunt resistance of the device. The shunt resistance values of the fabricated devices are displayed in Table 5.

Rsh values for ZnO devices, both fresh and aged, are lower than that of CsI/ZnO devices. After UV and UVO radiation, the Rsh of CsI/ZnO devices also decreases slightly. The decrease in Rsh is mainly contributed by the UV induced degradation of the active layer, as can be seen from the leakage current of CsI/ZnO devices, which is at least an order of magnitude lower than that of ZnO devices. Similar to the device Jsc trend, the leakage currents of CsI/ZnO devices, under both UV and UVO exposure, do not vary in magnitude and are essentially similar (Figure 7(b)). This trend again shows that, CsI can block the leakage current from ITO/ZnO electrode, arising from the oxygen-induced degradation of ZnO layer. Apart from the Jsc and fill factor, the Voc is also much stable in CsI/ZnO devices, compared to ZnO devices. From Table 4, the Voc of ZnO devices reduces by 24 mV and 43 mV after UV and UVO exposure, respectively, whereas the Voc of CsI/ZnO aged devices has only a minor 17

shift from the fresh devices, by 14 mV and 13 mV after UV and UVO exposure, respectively. To explain the stability in Voc, light impedance spectroscopy is carried out on all the aged devices at 650 mV bias. The Nyquist plots are presented in Figure 8. The plots are fitted to the same electrical circuit that was presented in Figure 5(b) inset. The fitted recombination resistance (Rrec) and chemical capacitance (Cµ) are displayed in Table 6 for comparison.

From the fitted model, the values of Rrec of the CsI/ZnO devices are higher than that of the ZnO devices, under both UV and UVO exposure. For UV degraded cells, the Rrec of ZnO devices is 235 Ω, whereas CsI/ZnO devices have an Rrec of 296 Ω. For UVO degraded cells, the Rrec of ZnO devices is 327 Ω, whereas CsI/ZnO devices have an Rrec of 448 Ω. The values indicate that the highest carrier recombination takes place in the devices with ZnO only ETL. Low Rrec of ZnO only devices can be attributed to the trap-assisted recombination in the degraded ZnO layer; it is also consistent with the low value of device shunt resistance, Rsh from the light J-V data (Table 5) and high leakage current under dark (figure 7). Trap-assisted recombination in ZnO layer and subsequent reduction in device recombination resistance plays vital role in reducing the Voc of ZnO devices significantly.21 Since CsI prevents the oxygen-induced degradation of ZnO layer, Rrec of CsI/ZnO devices is much higher than that of ZnO devices, which helps to retain device Voc and improve voltage stability even after UV and ozone induced degradation. In addition, the chemical capacitance, Cµ, of the CsI/ZnO devices is lower than that of ZnO devices. For UV degraded cells, the Cµ of ZnO devices is 27.1 nF, whereas CsI/ZnO devices have a Cµ of 25.5 nF. For UVO degraded cells, the Cµ of ZnO device is 39.8 nF; whereas CsI/ZnO devices have a Cµ of 22.2 nF. Cµ suggests charge accumulation at contacts that cannot be efficiently extracted.47 Low Cµ in CsI/ZnO means better charge extraction through ITO/CsI

18

cathode, which was also observed in the fresh CsI/ZnO devices, due to the reduction in ITO/CsI cathode work function (Figure 6) and ZnO surface roughness (Figure 3). The benefit of which is still present in UV and ozone degraded device. Aside from the short-term UV and UVO ageing test, the CsI/ZnO devices can supposedly prevent long term degradation of devices with ITO cathode. Earlier literature suggests that ITO cathode can release oxygen that can diffuse through the underneath ZnO and photo active polymer material and slowly degrade them over time.60-61 Since CsI can potentially block the oxygen diffused from ITO contact, it can reduce long-term degradation of devices. Although this analysis is beyond the scope of current study, it will be interesting to observe the long-term degradation mechanism of inverted PSCs incorporating CsI/ZnO as ETL in the future studies.

4. Conclusions In summary, PSCs with high stability and efficiency - incorporating CsI/ZnO as ETL in PTB7:PC71BM based light harvesting system has been demonstrated. The solution engineered CsI/ZnO bilayer was found to combine the benefits of their individual functionality (CsI and ZnO) while simultaneously overcoming the drawbacks of their material systems when used individually. It was also found that the CsI/ZnO bilayer can prevent both organic layer contaminations by Cs and build-up of aggregates near the cathode as observed in CsI only devices. In addition, the devices with bilayer ETL (average PCE= 7.3%) shows an improvement of 12.3% in the photovoltaic performance, compared to the devices with ZnO only ETL (average PCE= 6.5%), and ~8 times higher PCE compared to devices with CsI only ETL (average PCE= 0.92%). XPS analysis combined with the EIS studies showed that the incorporation of CsI between ITO and ZnO was found to favorably modify the energy-level alignment at the interface, which improved the charge collection efficiency by reducing the

19

charge transport resistance concomitantly lowering the charge recombination in the device. Furthermore, under UV and UVO exposure, the CsI/ZnO devices show better stability compared to ZnO devices. The ZnO devices suffer from 20.2% and 28.3% drop in PCE after UV and UVO radiation, respectively. On the other hand, CsI/ZnO device PCE drops by only 17.4% and 20.5% after UV and UVO radiation, respectively. We believe that the stability enhancement is coming from the formation of Cs-O bond at ITO/CsI and CsI/ZnO interface, where CsI works as an oxygen blocking layer and protects the underneath layers from generating additional trap states. Given the demonstrated results, future studies probing into the detailed mechanism behind the stability improvement via CsI incorporation between the ITO and ZnO would be of wide interest for the polymer solar cell research community.

Acknowledgements The authors would like to thank the Australian Centre for Advanced Photovoltaics, UNSW staff and technicians for their support. We acknowledge Future Solar Technologies for providing funding. We also thank the staffs of the Photovoltaic and Renewable Energy Engineering School (SPREE), the Electron Microscope Unit (EMU), and Dr Bill Gong and Outi Mustonen from the Solid State and Elemental Analysis Unit (SSEAU) of University of New South Wales (UNSW).

Appendix A. Supplementary material

References: 1. Espinosa, N.; Hosel, M.; Angmo, D.; Krebs, F. C., Solar Cells with One-Day Energy Payback for the Factories of the Future. Energy & Environmental Science 2012, 5, 5117-5132. 2. Wright, M.; Uddin, A., Organic—Inorganic Hybrid Solar Cells: A Comparative Review. Solar Energy Materials and Solar Cells 2012, 107, 87-111. 3. Krebs, F. C.; Espinosa, N.; Hösel, M.; Søndergaard, R. R.; Jørgensen, M., 25th Anniversary Article: Rise to Power – Opv-Based Solar Parks. Advanced Materials 2014, 26, 29-39.

20

4. Liu, Y.; Zhao, J.; Li, Z.; Mu, C.; Ma, W.; Hu, H.; Jiang, K.; Lin, H.; Ade, H.; Yan, H., Aggregation and Morphology Control Enables Multiple Cases of High-Efficiency Polymer Solar Cells. Nat Commun 2014, 5. 5. Ma, W., et al., Influence of Processing Parameters and Molecular Weight on the Morphology and Properties of High-Performance Pffbt4t-2od:Pc71bm Organic Solar Cells. Advanced Energy Materials 2015, 5. 6. Zhao, J.; Zhao, S.; Xu, Z.; Qiao, B.; Huang, D.; Zhao, L.; Li, Y.; Zhu, Y.; Wang, P., Revealing the Effect of Additives with Different Solubility on the Morphology and the Donor Crystalline Structures of Organic Solar Cells. ACS Applied Materials & Interfaces 2016, 8, 18231-18237. 7. Ouyang, X.; Peng, R.; Ai, L.; Zhang, X.; Ge, Z., Efficient Polymer Solar Cells Employing a NonConjugated Small-Molecule Electrolyte. Nat Photon 2015, 9, 520-524. 8. He, Z.; Zhong, C.; Su, S.; Xu, M.; Wu, H.; Cao, Y., Enhanced Power-Conversion Efficiency in Polymer Solar Cells Using an Inverted Device Structure. Nat Photon 2012, 6, 591-595. 9. Han, D.; Yoo, S., The Stability of Normal Vs. Inverted Organic Solar Cells under Highly Damp Conditions: Comparison with the Same Interfacial Layers. Solar Energy Materials and Solar Cells 2014, 128, 41-47. 10. Lin, R.; Wright, M.; Chan, K. H.; Puthen-Veettil, B.; Sheng, R.; Wen, X.; Uddin, A., Performance Improvement of Low Bandgap Polymer Bulk Heterojunction Solar Cells by Incorporating P3ht. Organic Electronics 2014, 15, 2837-2846. 11. Ongul, F., Solution-Processed Inverted Organic Solar Cell Using V2o5 Hole Transport Layer and Vacuum Free Egain Anode. Optical Materials 2015, 50, Part B, 244-249. 12. Kim, J. Y.; Kim, S. H.; Lee, H. H.; Lee, K.; Ma, W.; Gong, X.; Heeger, A. J., New Architecture for High-Efficiency Polymer Photovoltaic Cells Using Solution-Based Titanium Oxide as an Optical Spacer. Advanced Materials 2006, 18, 572-576. 13. Lin, R.; Miwa, M.; Wright, M.; Uddin, A., Optimisation of the Sol–Gel Derived Zno Buffer Layer for Inverted Structure Bulk Heterojunction Organic Solar Cells Using a Low Band Gap Polymer. Thin Solid Films 2014, 566, 99-107. 14. Jagadamma, L. K.; Abdelsamie, M.; El Labban, A.; Aresu, E.; Ngongang Ndjawa, G. O.; Anjum, D. H.; Cha, D.; Beaujuge, P. M.; Amassian, A., Efficient Inverted Bulk-Heterojunction Solar Cells from Low-Temperature Processing of Amorphous Zno Buffer Layers. Journal of Materials Chemistry A 2014, 2, 13321-13331. 15. Sun, Y.; Seo, J. H.; Takacs, C. J.; Seifter, J.; Heeger, A. J., Inverted Polymer Solar Cells Integrated with a Low-Temperature-Annealed Sol-Gel-Derived Zno Film as an Electron Transport Layer. Advanced Materials 2011, 23, 1679-1683. 16. Hau, S. K.; Yip, H.-L.; Baek, N. S.; Zou, J.; O’Malley, K.; Jen, A. K.-Y., Air-Stable Inverted Flexible Polymer Solar Cells Using Zinc Oxide Nanoparticles as an Electron Selective Layer. Applied Physics Letters 2008, 92, 253301. 17. White, M. S.; Olson, D. C.; Shaheen, S. E.; Kopidakis, N.; Ginley, D. S., Inverted BulkHeterojunction Organic Photovoltaic Device Using a Solution-Derived Zno Underlayer. Applied Physics Letters 2006, 89, 143517. 18. Elumalai, N. K.; Vijila, C.; Jose, R.; Zhi Ming, K.; Saha, A.; Ramakrishna, S., Simultaneous Improvements in Power Conversion Efficiency and Operational Stability of Polymer Solar Cells by Interfacial Engineering. Physical Chemistry Chemical Physics 2013, 15, 19057-19064. 19. Elumalai, N. K.; Saha, A.; Vijila, C.; Jose, R.; Jie, Z.; Ramakrishna, S., Enhancing the Stability of Polymer Solar Cells by Improving the Conductivity of the Nanostructured Moo3 Hole-Transport Layer. Physical Chemistry Chemical Physics 2013, 15, 6831-6841. 20. Elumalai, N. K.; Jin, T. M.; Chellappan, V.; Jose, R.; Palaniswamy, S. K.; Jayaraman, S.; Raut, H. K.; Ramakrishna, S., Electrospun Zno Nanowire Plantations in the Electron Transport Layer for HighEfficiency Inverted Organic Solar Cells. ACS Applied Materials & Interfaces 2013, 5, 9396-9404.

21

21. Mahmud, M. A.; Elumalai, N. K.; Upama, M. B.; Wang, D.; Chan, K. H.; Wright, M.; Xu, C.; Haque, F.; Uddin, A., Low Temperature Processed Zno Thin Film as Electron Transport Layer for Efficient Perovskite Solar Cells. Solar Energy Materials and Solar Cells 2017, 159, 251-264. 22. Xiao, T.; Cui, W.; Cai, M.; Leung, W.; Anderegg, J. W.; Shinar, J.; Shinar, R., Inverted Polymer Solar Cells with a Solution-Processed Cesium Halide Interlayer. Organic Electronics 2013, 14, 267272. 23. El Jouad, Z.; Louarn, G.; Praveen, T.; Predeep, P.; Cattin, L.; Bernède, J.-C.; Addou, M.; Morsli, M., Improved Electron Collection in Fullerene Via Caesium Iodide or Carbonate by Means of Annealing in Inverted Organic Solar Cells. EPJ Photovolt. 2014, 5, 50401. 24. Lindemann, W. R.; Xiao, T.; Wang, W.; Berry, J. E.; Anderson, N. A.; Houk, R. S.; Shinar, R.; Shinar, J.; Vaknin, D., An X-Ray Fluorescence Study on the Segregation of Cs and I in an Inverted Organic Solar Cell. Organic Electronics 2013, 14, 3190-3194. 25. Liang, Y.; Xu, Z.; Xia, J.; Tsai, S.-T.; Wu, Y.; Li, G.; Ray, C.; Yu, L., For the Bright Future—Bulk Heterojunction Polymer Solar Cells with Power Conversion Efficiency of 7.4%. Advanced Materials 2010, 22, E135-E138. 26. Loser, S.; Valle, B.; Luck, K. A.; Song, C. K.; Ogien, G.; Hersam, M. C.; Singer, K. D.; Marks, T. J., High-Efficiency Inverted Polymer Photovoltaics Via Spectrally Tuned Absorption Enhancement. Advanced Energy Materials 2014, 4, 1301938-n/a. 27. Madogni, V. I.; Kounouhéwa, B.; Akpo, A.; Agbomahéna, M.; Hounkpatin, S. A.; Awanou, C. N., Comparison of Degradation Mechanisms in Organic Photovoltaic Devices Upon Exposure to a Temperate and a Subequatorial Climate. Chemical Physics Letters 2015, 640, 201-214. 28. Prosa, M., et al., Enhanced Ultraviolet Stability of Air-Processed Polymer Solar Cells by Al Doping of the Zno Interlayer. ACS Applied Materials & Interfaces 2016, 8, 1635-1643. 29. Lo, M.-F.; Ng, T.-W.; Mo, H.-W.; Lee, C.-S., Direct Threat of a Uv-Ozone Treated Indium-TinOxide Substrate to the Stabilities of Common Organic Semiconductors. Advanced Functional Materials 2013, 23, 1718-1723. 30. Razzell-Hollis, J.; Wade, J.; Tsoi, W. C.; Soon, Y.; Durrant, J.; Kim, J.-S., Photochemical Stability of High Efficiency Ptb7:Pc70bm Solar Cell Blends. Journal of Materials Chemistry A 2014, 2, 2018920195. 31. Na, S.-I.; Kim, S.-S.; Jo, J.; Kim, D.-Y., Efficient and Flexible Ito-Free Organic Solar Cells Using Highly Conductive Polymer Anodes. Advanced Materials 2008, 20, 4061-4067. 32. Huang, D.; Li, Y.; Xu, Z.; Zhao, S.; Zhao, L.; Zhao, J., Enhanced Performance and Morphological Evolution of Ptb7:Pc71bm Polymer Solar Cells by Using Solvent Mixtures with Different Additives. Physical Chemistry Chemical Physics 2015, 17, 8053-8060. 33. Triloki; Garg, P.; Rai, R.; Singh, B. K., Structural Characterization of “as-Deposited” Cesium Iodide Films Studied by X-Ray Diffraction and Transmission Electron Microscopy Techniques. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 2014, 736, 128-134. 34. Liang, Z.; Zhang, Q.; Wiranwetchayan, O.; Xi, J.; Yang, Z.; Park, K.; Li, C.; Cao, G., Effects of the Morphology of a Zno Buffer Layer on the Photovoltaic Performance of Inverted Polymer Solar Cells. Advanced Functional Materials 2012, 22, 2194-2201. 35. Yip, H.-L.; Hau, S. K.; Baek, N. S.; Jen, A. K.-Y., Self-Assembled Monolayer Modified Zno/Metal Bilayer Cathodes for Polymer/Fullerene Bulk-Heterojunction Solar Cells. Applied Physics Letters 2008, 92, 193313. 36. Liao, H.-H.; Chen, L.-M.; Xu, Z.; Li, G.; Yang, Y., Highly Efficient Inverted Polymer Solar Cell by Low Temperature Annealing of Cs2co3 Interlayer. Applied Physics Letters 2008, 92, 173303. 37. Mahmud, M. A.; Elumalai, N. K.; Upama, M. B.; Wang, D.; Wright, M.; Sun, T.; Xu, C.; Haque, F.; Uddin, A., Simultaneous Enhancement in Stability and Efficiency of Low-Temperature Processed Perovskite Solar Cells. RSC Advances 2016, 6, 86108-86125. 38. Lai, T.-H.; Tsang, S.-W.; Manders, J. R.; Chen, S.; So, F., Properties of Interlayer for Organic Photovoltaics. Materials Today 2013, 16, 424-432. 22

39. Boix, P. P.; Ajuria, J.; Etxebarria, I.; Pacios, R.; Garcia-Belmonte, G.; Bisquert, J., Role of Zno Electron-Selective Layers in Regular and Inverted Bulk Heterojunction Solar Cells. The Journal of Physical Chemistry Letters 2011, 2, 407-411. 40. Elumalai, N. K.; Uddin, A., Open Circuit Voltage of Organic Solar Cells: An in-Depth Review. Energy & Environmental Science 2016, 9, 391-410. 41. Zhang, Z.; Yates, J. T., Band Bending in Semiconductors: Chemical and Physical Consequences at Surfaces and Interfaces. Chemical Reviews 2012, 112, 5520-5551. 42. Du, J.; Chen, X.-l.; Liu, C.-c.; Ni, J.; Hou, G.-f.; Zhao, Y.; Zhang, X.-d., Highly Transparent and Conductive Indium Tin Oxide Thin Films for Solar Cells Grown by Reactive Thermal Evaporation at Low Temperature. Applied Physics A 2014, 117, 815-822. 43. Alam, M. J.; Cameron, D. C., Optical and Electrical Properties of Transparent Conductive Ito Thin Films Deposited by Sol–Gel Process. Thin Solid Films 2000, 377–378, 455-459. 44. Han, Y.; Kim, D.; Cho, J.-S.; Koh, S.-K.; Song, Y. S., Tin-Doped Indium Oxide (Ito) Film Deposition by Ion Beam Sputtering. Solar Energy Materials and Solar Cells 2001, 65, 211-218. 45. Huang, J.; Li, G.; Yang, Y., A Semi-Transparent Plastic Solar Cell Fabricated by a Lamination Process. Advanced Materials 2008, 20, 415-419. 46. Burtone, L.; Ray, D.; Leo, K.; Riede, M., Impedance Model of Trap States for Characterization of Organic Semiconductor Devices. Journal of Applied Physics 2012, 111, 064503. 47. Arredondo, B.; Romero, B.; Del Pozo, G.; Sessler, M.; Veit, C.; Würfel, U., Impedance Spectroscopy Analysis of Small Molecule Solution Processed Organic Solar Cell. Solar Energy Materials and Solar Cells 2014, 128, 351-356. 48. Garcia-Belmonte, G.; Munar, A.; Barea, E. M.; Bisquert, J.; Ugarte, I.; Pacios, R., Charge Carrier Mobility and Lifetime of Organic Bulk Heterojunctions Analyzed by Impedance Spectroscopy. Organic Electronics 2008, 9, 847-851. 49. Arredondo, B.; Martín-López, M. B.; Romero, B.; Vergaz, R.; Romero-Gomez, P.; Martorell, J., Monitoring Degradation Mechanisms in Ptb7:Pc71bm Photovoltaic Cells by Means of Impedance Spectroscopy. Solar Energy Materials and Solar Cells 2016, 144, 422-428. 50. Norrman, K.; Madsen, M. V.; Gevorgyan, S. A.; Krebs, F. C., Degradation Patterns in Water and Oxygen of an Inverted Polymer Solar Cell. Journal of the American Chemical Society 2010, 132, 16883-16892. 51. Perrier, G.; de Bettignies, R.; Berson, S.; Lemaître, N.; Guillerez, S., Impedance Spectrometry of Optimized Standard and Inverted P3ht-Pcbm Organic Solar Cells. Solar Energy Materials and Solar Cells 2012, 101, 210-216. 52. Aprilia, A.; Wulandari, P.; Suendo, V.; Herman; Hidayat, R.; Fujii, A.; Ozaki, M., Influences of Dopant Concentration in Sol–Gel Derived Azo Layer on the Performance of P3ht:Pcbm Based Inverted Solar Cell. Solar Energy Materials and Solar Cells 2013, 111, 181-188. 53. Kirchartz, T.; Gong, W.; Hawks, S. A.; Agostinelli, T.; MacKenzie, R. C. I.; Yang, Y.; Nelson, J., Sensitivity of the Mott–Schottky Analysis in Organic Solar Cells. The Journal of Physical Chemistry C 2012, 116, 7672-7680. 54. Bisquert, J., Chemical Capacitance of Nanostructured Semiconductors: Its Origin and Significance for Nanocomposite Solar Cells. Physical Chemistry Chemical Physics 2003, 5, 5360-5364. 55. Bartelt, J. A.; Lam, D.; Burke, T. M.; Sweetnam, S. M.; McGehee, M. D., Charge-Carrier Mobility Requirements for Bulk Heterojunction Solar Cells with High Fill Factor and External Quantum Efficiency >90%. Advanced Energy Materials 2015, 5, 1500577. 56. Ebenhoch, B.; Thomson, S. A. J.; Genevičius, K.; Juška, G.; Samuel, I. D. W., Charge Carrier Mobility of the Organic Photovoltaic Materials Ptb7 and Pc71bm and Its Influence on Device Performance. Organic Electronics 2015, 22, 62-68. 57. Liu, H.; Wu, Z.; Hu, J.; Song, Q.; Wu, B.; Lam Tam, H.; Yang, Q.; Hong Choi, W.; Zhu, F., Efficient and Ultraviolet Durable Inverted Organic Solar Cells Based on an Aluminum-Doped Zinc Oxide Transparent Cathode. Applied Physics Letters 2013, 103, 043309.

23

58. Jeong, J.; Seo, J.; Nam, S.; Han, H.; Kim, H.; Anthopoulos, T. D.; Bradley, D. D. C.; Kim, Y., Significant Stability Enhancement in High-Efficiency Polymer:Fullerene Bulk Heterojunction Solar Cells by Blocking Ultraviolet Photons from Solar Light. Advanced Science 2016, 3, 1500269. 59. Proctor, C. M.; Nguyen, T.-Q., Effect of Leakage Current and Shunt Resistance on the Light Intensity Dependence of Organic Solar Cells. Applied Physics Letters 2015, 106, 083301. 60. Scott, J. C.; Kaufman, J. H.; Brock, P. J.; DiPietro, R.; Salem, J.; Goitia, J. A., Degradation and Failure of Meh‐Ppv Light‐Emitting Diodes. Journal of Applied Physics 1996, 79, 2745-2751. 61. de Jong, M. P.; van IJzendoorn, L. J.; de Voigt, M. J. A., Stability of the Interface between Indium-Tin-Oxide and Poly(3,4-Ethylenedioxythiophene)/Poly(Styrenesulfonate) in Polymer LightEmitting Diodes. Applied Physics Letters 2000, 77, 2255-2257.

LIST OF FIGURES

Figure 1 (a) Schematic diagram of device structure of the fabricated inverted PTB7:PC71BM organic solar cells with ZnO or CsI (0.1 mg/ml)/ZnO bi-layer ETL. (b) Current–voltage characteristics of the solar cells, with both ZnO and CsI (0.1 mg/ml)/ZnO ETL, at room temperature under AM1.5G illumination at 100 mW/ cm2.

24

Figure 2 (a) Optical transmission profiles of ITO/ZnO (0.1 g/ml) and ITO/CsI (0.1 mg/ml)/ZnO (0.1 g/ml) films. (b) X-ray diffraction (XRD) patterns of the ZnO (0.1 g/ml) and CsI (0.1 mg/ml)/ZnO (0.1 g/ml) thin films. (CsI peaks: labeled red, ZnO peaks: labeled black)

25

Figure 3 AFM images (5 µm x 5 µm) of (a) ZnO only and (b) CsI/ZnO bilayer thin film on top of an ITO/glass substrate.

26

Figure 4 (a) XPS binding energy spectrum of Cesium (Cs 3d) and Oxygen (O 1s) peaks for ITO/CsI film. An XPS spectrum of plain ITO film is also displayed. (b) X-ray photoelectron spectroscopies of Cs 3d (black line) and O 1s (red line) of CsI layer on ITO substrate. All the characteristics peaks have been detected and analyzed using Thermo Scientific™ Avantage Software.

27

Figure 5 (a) Mott–Schottky curve at 10 kHz frequency with ZnO (black square) and CsI/ZnO (red circle) devices under dark. Inset shows full Mott-Schottky spectra. (b) Nyquist plot of PTB7:PC71BM devices with ZnO and CsI/ZnO ETLs at a bias of 650 mV under dark. Inset displays the electrical equivalent circuit used for fitting the experimental data. The fitted parameters are summarized in Table 3.

28

Figure 6 Effect of dipole layer formation on ITO/CsI interface explained by energy band diagrams of ITO and ZnO contacts (a, b) and ITO/CsI and ZnO contacts (c, d). EVL, vacuum energy; EC, energy of conduction band minimum; EV, energy of valence band maximum; WFITO, ITO work function; WFITO”, modified ITO work function; 𝜒𝑠 , electron affinity of ZnO; 𝜙𝐵1 , 𝜙𝐵2 , Schottky barriers at ITO/ZnO and CsI modified ITO/ZnO interfaces.

29

Figure 7 Current–voltage characteristics of the solar cells, with both ZnO and CsI/ZnO ETL, exposed to UV and UVO treatment (duration: 20 min each) for accelerated ageing, (a) under illumination (b) in the dark, at room temperature under AM1.5G illumination at 100 mW/cm2

30

Figure 8 Nyquist plot of PTB7:PC71BM devices with ZnO and CsI/ZnO ETLs at a bias of 650 mV under illumination. The devices have been exposed to UV and UVO treatment (duration: 20 min each) for accelerated ageing. Figure 6(b) Inset displays the electrical equivalent circuit used for fitting the experimental data. The fitted parameters are summarized in Table 6.

31

Table 1: Average photovoltaic parameters for PTB7:PC71BM inverted devices with CsI electron transport layer prepared from various concentrations of CsI solutions.

CsI concentration Jsc (mA/cm2) (mg/ml)

Voc (mV)

FF (%)

PCE (%)

0.1

9.64

238.88

28.6

0.67

0.5

10.28

289.58

30.6

0.92

1

0.79

154.86

43.4

0.05

2

0.73

126.18

32.3

0.03

Table 2: Average photovoltaic parameters for PTB7:PC71BM devices with ZnO and CsI/ZnO ETL

ETL

Jsc (mA/cm2)

VOC (mV)

FF (%)

PCE (%)

ZnO

17.02

722.2

52.9

6.5

CsI/ZnO bi-layer

17.64

726.4

56.6

7.3

Table 3: Electrical parameters extracted from Nyquist plot for ZnO only and CsI/ZnO devices in figure 5(b). Parameters

ZnO Devices

CsI/ZnO Devices

Unit

Series resistance, Rs

20.5

23.8

Ω

Transport resistance, Rt

444

314

Ω

Recombination resistance, Rrec

2.23

2.24

kΩ

Bulk capacitance, Cg

4.63

5.82

nF

Chemical capacitance, Cµ

10.8

10.9

nF

Electron diffusion time, 𝜏𝑑 = 𝑅𝑡 × 𝐶µ

4.80

3.42

µs

Active layer thickness, L

2x10-05

2x10-05

cm

𝐿2

48

𝜏𝑑 𝑒×𝐷𝑛

8.34x10-05

1.17x10-04

cm2/s

48

3.22x10-03

4.51x10-03

cm2/V.s

Electron diffusivity, 𝐷𝑛 = Electron mobility, µ𝑛 =

𝑘𝑇

32

Table 4: Average photovoltaic parameters for devices with ZnO and CsI/ZnO ETL, fresh and exposed to UV and UVO treatment, for accelerated ageing. ETL type

ZnO devices

CsI/ZnO bilayer devices

Device Status

Jsc (mA/cm2)

VOC (mV)

FF (%)

PCE (%)

Normalized PCE

Fresh cell

17.02

722.2

52.9

6.5

1

After UV

14.5

698.71

51.24

5.19

0.79

After UVO

14.12

679.54

48.39

4.66

0.64

Fresh cell

17.64

726.4

56.6

7.3

1

After UV

15.17

712.32

55.79

6.03

0.83

After UVO

15.39

713.44

52.84

5.8

0.79

Table 5: List of average shunt resistance (Rsh) of fresh and aged PTB7:PC71BM devices with ZnO and CsI/ZnO ETL

ETL Type Fresh UV UVO

ZnO 735 701 708

Rsh (Ω) CsI/ZnO 956 786 742

Table 6: Fitted Recombination resistance (Rrec) and Chemical capacitance (Cµ) extracted from Nyquist plot for ZnO only and CsI/ZnO devices, exposed to UV and UVO treatment, for accelerated ageing.

ZnO [UV]

Recombination resistance, Rrec (Ω) 235

Chemical capacitance, Cµ (nF) 27.1

CsI/ZnO [UV]

296

25.5

ZnO [UV]

327

39.8

CsI/ZnO [UV]

448

22.2

Fitted Parameters

33