Influence of the Hole Transporting Layer on the Thermal Stability of Inverted Organic Photovoltaics Using Accelerated-Heat Lifetime Protocols.

High power conversion efficiency (PCE) inverted organic photovoltaics (OPVs) usually use thermally evaporated MoO3 as a hole transporting layer (HTL). Despite the high PCE values reported, stability investigations are still limited and the exact degradation mechanisms of inverted OPVs using thermally evaporated MoO3 HTL remain unclear under different environmental stress factors. In this study, we monitor the accelerated lifetime performance under the ISOS-D-2 protocol (heat conditions 65 °C) of nonencapsulated inverted OPVs based on the thiophene-based active layer materials poly(3-hexylthiophene) (P3HT), poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl]] (PTB7), and thieno[3,2-b]thiophene-diketopyrrolopyrrole (DPPTTT) blended with [6,6]-phenyl C71-butyric acid methyl ester (PC[70]BM). The presented investigation of degradation mechanisms focus on optimized P3HT:PC[70]BM-based inverted OPVs. Specifically, we present a systematic study on the thermal stability of inverted P3HT:PC[70]BM OPVs using solution-processed poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) and evaporated MoO3 HTL. Using a series of measurements and reverse engineering methods, we report that the P3HT:PC[70]BM/MoO3 interface is the main origin of failure of the P3HT:PC[70]BM-based inverted OPVs under intense heat conditions, a trend that is also observed for the other two thiophene-based polymers used in this study.

Introduction

Organic photovoltaics (OPVs) have attracted great scientific interest during the past decade because of their ease of manufacture with printable techniques and their potential to become flexible, lightweight, and low-cost energy sources.[1,2] High power conversion efficiencies (PCEs) and prolonged lifetimes are essential for OPV commercialization. Exciting PCEs of 10% have been demonstrated,[3] but long stabilities of OPVs are the next barrier that needs to be overcome. Understanding the degradation mechanisms that influence the stability of different device configurations under various environmental stress factors is still a challenging task.

In the inverted structure,[4,5] the use of an air-stable metal such as silver results in enhanced lifetime compared with normal structured OPVs, which are limited in stability mainly because of the oxidation of the metals such as calcium or aluminum.[6] Poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) is the most common material used as the hole-transporting layer (HTL) in both architectures. It is known that because of its hygroscopic nature, ingress of moisture and oxygen from the edges into the device can cause degradation.[7,8]

Furthermore, heat is one of the environmental factors found to significantly affect the long-term stability of OPVs. Heat stability studies at the operating temperature performed by Sachs-Quintana et al. on normal structured OPVs show that a thin polymer layer forms at the back (metal) contact, which creates an electron barrier between the active layer and the cathode.[9] This is related to the interface between the back contact and the bulk heterojunction and is shown to be the first step in thermal degradation in organic solar cells with normal architecture and especially while using low glass-transition temperature-conjugated polymers. However, in inverted OPVs, this barrier formation is favorable as it helps hole selectivity. This feature has been claimed to be an additional advantage of inverted OPVs.[9] Thus, the inverted structure is preferable concerning product development targets as it can allow more design flexibility and prolonged lifetime.

However, despite its prolonged lifetime, the top electrode (the anode) of the inverted OPVs has been identified as one of the most vulnerable parts of the device under various environmental stress factors. Several studies were performed, which reveal the crucial influence of the interface between the active layer and HTL on the lifetime performance of the inverted OPVs.[10−13] Inverted OPVs with PEDOT:PSS treated with different wetting agents present different lifetime behaviors under accelerated humidity conditions, thereby showing the crucial influence of processing and HTL/active layer interface formation.[14] Furthermore, Norrman et al. have suggested that PEDOT:PSS is the main factor of degradation of inverted OPVs under ambient conditions. They have demonstrated that the phase separation of PEDOT:PSS into PEDOT-rich and PSS-rich regions and the active layer/PEDOT:PSS interface are the main sources of failure of the long-term lifetime of inverted OPVs.[15]

Therefore, substitution of the moisture-sensitive PEDOT:PSS in inverted OPVs with metal oxide HTLs such as MoO3, V2O5, and WO3 could be beneficial in some cases concerning the stability of inverted OPVs. However, the use of solution-processed metal oxides as HTLs in inverted OPVs is still challenging because major processing issues have to be addressed. Recent work has shown that solution-processed MoO3 can be used to produce normal structure and inverted OPVs that are efficient and stable when used as the sole HTL[16,17] as well as when inserted in addition to PEDOT:PSS.[18] Mixing solution-based MoO3 with PEDOT:PSS has also yielded efficient inverted solar cells.[19] However, evaporated metal oxides are still more widely used as HTLs in inverted OPVs, leading to high PCEs with optimum hole selectivity.

Evaporated MoO3 is a promising material as an HTL in inverted OPVs in terms of efficiency because of the favorable energy-level alignment between its work function measured at 6.9 eV[20] and the highest occupied molecular orbital of typical active layers in the range of −5.0 eV for polythiophene-based materials.[21] There are promising developments regarding the stability of inverted OPVs comparing MoO3 and PEDOT:PSS HTLs in air;[17] however, lifetime studies under humidity comparing MoO3 and PEDOT:PSS HTLs show that evaporated MoO3 is also very sensitive to moisture and leads to inverted OPVs with limited lifetime.[22] The reason is that oxygen and humidity ingress can alter the conductivity and work function of MoO3.[23,24] On the other hand, the hygroscopic PEDOT:PSS in this architecture can act as a getter and protect the active layer from water interactions.[25] Furthermore, MoO3 may undergo a change in the oxidation state of the Mo(VI) metal center under exposure to heat or UV light, leading to changes in its work function and its optical response. Under heat conditions, MoO3 releases oxygen, leading to lower Mo oxidation states and a shift in the work function of the oxide.[24,26]

In addition, Voroshazi et al. have shown that the reduction of MoO3 can also occur under heat in dark conditions in inverted OPVs with MoO3/Ag/Al top contact because of the chemical interactions between MoO3 and Al atoms.[27] Another aspect of the degradation mechanism of inverted OPVs with the aforementioned top electrode was pointed out by Rösch et al. through different imaging techniques, whereby they attributed the failure to the migration of silver and diffusion into the MoO3 layer, changing its work function.[28] Greenbank et al. also discuss diffusion into the active layer, and its resulting weakening is correlated to the decline (specifically Jsc reduction) in the device performance when used in conjunction with MoO3 as the HTL.[29] However, a full understanding of the stability of MoO3 HTL in inverted OPVs is still lacking because of the complexity of this material and the varying environmental stress factors that can influence the degradation cause of inverted OPVs with the MoO3 HTL.

In this study, therefore, we aim to analyze the degradation mechanisms of nonencapsulated inverted OPVs using MoO3 as the HTL, focusing on accelerated-heat lifetime conditions using the ISOS-D-2 protocol [dark, low relative humidity (RH), 65 °C]. As a first step, inverted OPVs using various polymer/fullerene active layers are investigated. Photocurrent mapping measurements provide degradation images at 65 °C at various points in time. Second, different electrode configurations are investigated for our reference poly(3-hexylthiophene):[6,6]-phenyl C71-butyric acid methyl ester (P3HT:PC[70]BM) polymer/fullerene active layer. Using combinations of two interlayers—MoO3 and PEDOT:PSS—at the top electrode of P3HT:PC[70]BM-based devices, the device degradation is attributed primarily to the interface between the active layer and MoO3 and secondarily to that between MoO3 and Ag. Third, alternative reverse engineering methods are used to enhance our assumption that the MoO3/Ag interface significantly contributes to the degradation of devices under heat.

Experimental Section

Prepatterned glass–indium tin oxide (ITO) substrates (sheet resistance of 4 Ω/sq) were purchased from Psiotec Ltd. Zinc acetate dehydrate, 2-methoxyethanol, and ethanolamine were purchased from Sigma-Aldrich, P3HT was purchased from Rieke Metals, poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl]] (PTB7) was purchased from 1-Material, PC[70]BM was purchased from Solenne BV, PEDOT:PSS PH was purchased from H.C. Stark, MoO3 powder was purchased from Sigma-Aldrich, and silver pellets were purchased from Kurt J. Lesker.

For the fabrication of inverted solar cells, ITO substrates were sonicated in acetone and subsequently in isopropanol for 10 min. The ZnO electron-transporting layer was prepared using a sol–gel process as described in detail elsewhere.[30] The ZnO precursor was doctor-bladed on the top of ITO substrates and annealed for 20 min at 140 °C under ambient conditions. After the annealing step, a 40 nm ZnO layer is formed as measured using a Veeco Dektak 150 profilometer. The photoactive layers, which are blends of 36 mg/mL in chlorobenzene P3HT:PC[70]BM (1:0.8 by weight), 15 mg/mL in o-DCB DPPTTT:PC[70]BM (1:2 by weight), and 25 mg/mL PTB7:PC[70]BM (1:1.5 by weight), were doctor-bladed on the top of ZnO, resulting in a layer thickness of ∼180, 100, and 90 nm, respectively, as measured with a profilometer.

All PEDOT:PSS-based devices were treated with two wetting agents (Zonyl and Dynol) as described in detail previously[14,31] (the resultant is referred to as PEDOT:PSS:ZD hereafter). The P3HT:PC[70]BM-based inverted OPVs were annealed inside of a glovebox at 140 °C for 20 min before thermal evaporation of a silver layer with a thickness of 100 nm, resulting in four solar cells, each with an active area of 9 mm2. For the devices with the MoO3 HTL, 10 nm of MoO3 was evaporated before silver evaporation. For devices using MoO3 and PEDOT:PSS:ZD as HTLs, MoO3 was evaporated onto the active layer and PEDOT:PSS:ZD was doctor-bladed onto the MoO3 layer followed by silver evaporation. These devices were annealed for another 2 min at 140 °C after the PEDOT:PSS:ZD deposition. For devices with a PEDOT:PSS:ZD/MoO3/Ag top electrode, the procedure was the same as for PEDOT:PSS:ZD-based devices, except for an extra 10 nm MoO3 layer evaporated on the top of the PEDOT:PSS:ZD layer.

The current density–voltage (J/V) characteristics were measured using a Keithley source measurement unit (SMU 2420). For illumination, a calibrated Newport Solar Simulator equipped with a Xe lamp was used, providing an AM 1.5G spectrum at 100 mW/cm2 as measured using an Oriel 91150 V calibration cell equipped with a KG5 filter. Photocurrent measurements were recorded under a 405 nm laser excitation using a Botest PCT photocurrent system. Thermal aging of the devices was performed according to the ISOS-D-2 protocol;[32] namely, inverted OPVs under examination were placed in a dark chamber (Binder) at 65 °C with controlled ambient humidity of ∼45% and aged for several hours. Their J/V characteristics were measured at different stages of degradation. The devices were nonencapsulated, and the lifetime study was started 1 day after device fabrication.

Results and Discussion

From our initial heat lifetime experiments comparing inverted OPVs comprised of ITO/ZnO/P3HT:PC[70]BM with a MoO3/Ag versus PEDOT:PSS:ZD/Ag top electrode, we observed a dramatic drop in all photovoltaic parameters after just a few hours of aging of the inverted OPVs using MoO3 as the HTL (data not shown).

Investigating the Degradation Mechanisms of Inverted OPVs under the ISOS-D-2 Protocol Using Different Polymer/Fullerene Blends and Top Electrode Configurations

To obtain a general picture about the heat degradation origins of inverted OPVs, three different polymer/fullerene blends were tested. The inverted OPVs with the following general configuration of ITO/ZnO/polymer:PC[70]BM/MoO3/Ag or PEDOT:PSS:ZD/Ag were exposed to heat conditions based on the ISOS-D-2 protocol (T = 65 °C, dark, and constant ambient RH of ∼40%). P3HT:PC[70]BM (1:0.8),[30,33] PTB7:PC[70]BM (1:1.5),[34,35] and DPPTTT:PC[70]BM (1:4)[36,37] were used as the active layers. The polymer/fullerene ratios used were chosen from the best performing devices reported in the literature. The average absolute values of the photovoltaic parameters of the inverted OPVs under study were obtained just before starting the heat-aging and are summarized in Table .

Table 1

Average Absolute Photovoltaic Parameter Values and Standard Deviation out of Eight Inverted OPVs in Each Case, Obtained before Initiating the Heat-Aging

inverted OPVs	Voc (V)	Jsc (mA/cm2)	FF (%)	PCE (%)
ActiveLayer/MoO3/Ag
P3HT:PC[70]BM	0.560 ± 0.009	11.17 ± 0.33	56.41 ± 1.69	3.54 ± 0.15
PTB7:PC[70]BM	0.678 ± 0.023	12.49 ± 0.55	46.75 ± 6.29	3.97 ± 0.67
DPPTTT:PC[70]BM	0.553 ± 0.010	11.22 ± 0.57	52.95 ± 3.73	3.32 ± 0.41
ActiveLayer/PEDOT:PSS:ZD/Ag
P3HT:PC[70]BM	0.582 ± 0.007	9.26 ± 0.38	58.26 ± 2.56	3.16 ± 0.15
PTB7:PC[70]BM	0.691 ± 0.010	8.26 ± 0.68	50.28 ± 1.70	2.86 ± 0.31
DPPTTT:PC[70]BM	0.544 ± 0.008	8.07 ± 1.11	54.32 ± 1.42	2.53 ± 0.11

It should be noted that although some of the device performance values presented in Table may be below some literature values for the device structure, the purpose of this paper is not to produce best-performing “hero” devices. Thus, the relative performance values can still be used to obtain a general picture of the degradation behavior. Figure shows the normalized photovoltaic parameters over time of exposure under heat conditions of all inverted OPVs under study.

Figure 1

Degradation trends of the OPV parameters at 65 °C over time for nonencapsulated inverted OPVs with different active layers (red, P3HT:PC[70]BM; green, DPPTTT:PC[70]BM; and blue, PTB7:PC[70]BM) using either MoO3 (dashed lines) or PEDOT:PSS (solid lines) as the HTLs, plotted as a function of (a) normalized Voc, (b) normalized Jsc, (c) normalized fill factor (FF), and (d) normalized PCE.

As shown in Figure , photovoltaic parameters of all inverted OPVs with different polymer/fullerene blends using MoO3/Ag top electrodes drop dramatically within the first few hours of exposure under accelerated-heat conditions. By contrast, the lifetime performance of the corresponding inverted OPVs using the PEDOT:PSS HTL is significantly better than that observed with MoO3 for all different polymer/fullerene blend combinations studied.

It is worth noting that all polymers used in this study are thiophene-based (P3HT, PTB7, and DPPTTT). For PTB7:PC[70]BM-based inverted OPVs, a fast degradation pattern under heating at 85 °C in the dark has previously been observed and has been primarily attributed to morphological changes within the active layer.[38] Furthermore, PTB7 has been reported to be sensitive to several environmental factors,[36,39] whereas DPPTTT has shown a greater stability under light and oxygen[37] but not under thermal conditions.[40] On the other hand, P3HT:PC[70]BM-based inverted OPVs are known to be resistive to several environmental stress factors, and impressive lifetime performances under accelerated and outdoor conditions are demonstrated elsewhere.[41−43]

All above observations indicate that the incorporation of MoO3 as the HTL at the top electrode of inverted OPVs has a dominant influence on the lifetime performance when subjected to the ISOS D-2 protocol despite the complicated degradation patterns that might also arise because of the variety of active layers. In fact, our study focuses specifically on the electrode configurations in the inverted device structure and the effects of this on the degradation behavior of OPVs. Detailed investigations of the active layer would be beyond the scope of this work. However, the interested reader is referred to the work by Wantz et al.,[44] Rumer and McCulloch,[45] Wang et al.,[43] and Schaffer et al.[46] Our further analyses in the remainder of this study focus on the understanding of the degradation mechanisms of inverted OPVs using our reference and well-optimized P3HT:PC[70]BM as the active layer.

Investigating the Degradation Mechanisms of P3HT:PC[70]BM/MoO3-Based Inverted OPVs under the ISOS-D-2 Protocol Using Buffer Layer Engineering

To visualize this fast degradation of P3HT:PC[70]BM-based solar cells, photocurrent images that map the active area of the device and its degradation are shown in Figure . This is a powerful technique to understand the spatially resolved photocurrent distribution of complete devices showing a photoelectric effect.[47] A laser is moved across the device through the transparent electrode, giving an indication of defects and therefore possibly diminished device performance. Inverted OPVs using P3HT:PC[70]BM as the active layer and MoO3 as the HTL were aged, and photocurrent mapping images were extracted at different stages of degradation. The degradation of inverted OPVs with MoO3 as the HTL occurs very fast, even after 5 h of degradation under heat, and the photocurrent drops dramatically (see Figures and 2).

Figure 2

Normalized photocurrent mapping images of nonencapsulated inverted OPVs using MoO3 as the HTL in P3HT:PC[70]BM-based solar cells, showing degradation at 65 °C over time of exposure.

The photocurrent mapping measurements indicate that for nonencapsulated inverted P3HT:PC[70]BM solar cells incorporating evaporated MoO3 HTL, the photocurrent generation drops dramatically in just 5 h of exposure. The photocurrent intensity observed is significantly less within the whole device area. Despite the obvious drop near the edges of the devices, which could be attributed to the ingress of air and moisture, the center of the device also shows significantly less photocurrent generation compared with the fresh devices. This is in accordance with Figure b, which shows a fast Jsc degradation of P3HT:PC[70]BM/MoO3 within the first few hours of exposure.

To investigate the interfacial interaction of MoO3 with the P3HT:PC[70]BM active layer, we use buffer layer engineering to isolate the interfaces under study. Inverted OPVs based on P3HT:PC[70]BM with four different configurations of the top electrode—MoO3/Ag, PEDOT:PSS:ZD/Ag, PEDOT:PSS:ZD/MoO3/Ag, and MoO3/PEDOT:PSS:ZD/Ag—were tested under 65 °C. The photovoltaic parameters of these devices are shown in Table .

Table 2

Initial Photovoltaic Parameters and Standard Deviation of Inverted OPVs with Different Hole-Transporting Layers

inverted P3HT:PC[70]BM OPVs using different top electrode configurations	Voc (V)	Jsc (mA/cm2)	FF (%)	PCE (%)
PEDOT:PSS:ZD/Ag	0.582 ± 0.007	9.26 ± 0.38	58.26 ± 2.56	3.16 ± 0.15
PEDOT:PSS:ZD/MoO3/Ag	0.568 ± 0.014	10.03 ± 0.47	53.65 ± 2.21	3.08 ± 0.17
MoO3/Ag	0.560 ± 0.009	11.17 ± 0.33	56.41 ± 1.69	3.54 ± 0.15
MoO3/PEDOT:PSS:ZD/Ag	0.546 ± 0.009	8.52 ± 0.73	57.20 ± 0.81	2.72 ± 0.26

We observe that the solar cells using the MoO3/Ag top electrode show the highest initial efficiency of 3.54%. This is followed by very similar device efficiencies of 3.16 and 3.08% obtained using PEDOT:PSS:ZD/Ag and PEDOT:PSS:ZD/MoO3/Ag, respectively. Finally, a 2.72% PCE is obtained for inverted OPVs with an MoO3/PEDOT:PSS:ZD/Ag top electrode. Interestingly, these values are comparable to those reported by Wang et al. using a solution-based hybrid electrode consisting of MoO3 and PEDOT:PSS,[19] indicating an efficient bilayer in our system.

As shown in Figure and in line with the previous observations from Figure , inverted OPVs using the MoO3/Ag top electrode degrade very fast under heat conditions from the initial efficiency of 3.54% compared with inverted OPVs with PEDOT:PSS:ZD as the HTL. As also reported elsewhere,[48] all J/V parameters exponentially drop without allowing precise identification of the failure mechanism. However, by incorporating a PEDOT:PSS:ZD layer between the active layer and MoO3 as well as between MoO3 and Ag, the degradation behavior can be influenced.

Figure 3

Degradation trends of inverted OPV parameters at 65 °C over time for ITO/ZnO/P3HT:PC[70]BM with different top electrode configurations as a function of (a) normalized Voc, (b) normalized Jsc, (c) normalized FF, and (d) normalized PCE.

The inverted OPVs with MoO3/PEDOT:PSS:ZD/Ag as a top contact present an intermediate PCE decay between devices with MoO3 and PEDOT:PSS:ZD as HTLs. The majority of the PCE drop in those devices is attributed to FF (Figure ). Thus, we assume that the influence on FF in devices with MoO3/Ag top contact is attributed to the interface of MoO3 and active layer. By inserting the layer of PEDOT:PSS:ZD between MoO3 and Ag, the suppressed decay of those devices indicates that a direct contact between the MoO3 and Ag influences device degradation. Furthermore, the decline in the Voc of devices with a direct interface between P3HT:PC[70]BM and MoO3 (organic/metal oxide interface) is more obvious than that of the devices with P3HT:PC[70]BM and PEDOT:PSS:ZD (organic/organic interface). The latter can again link the failure of the device to the interface of the P3HT:PC[70]BM active layer with MoO3, as has been reported previously.[48]

On the other hand, inverted OPVs with a PEDOT:PSS:ZD/MoO3 HTL present lifetime behavior similar to that of inverted OPVs with only PEDOT:PSS:ZD as the HTL. In that case, no obvious degradation is observed when MoO3 and Ag are in contact. Two possible phenomena could explain this.

The first is that carrier selectivity occurs through PEDOT:PSS:ZD and carrier (i.e., hole) collection occurs through MoO3/Ag, suggesting that electrons are responsible for the degradation at the MoO3/Ag interface and that degradation is prevented when the MoO3/Ag is protected from electrons. This effect was observed by Zhu et al. on inverted hybrid quantum dot/polymer solar cells,[49] in which use of a dual HTL (PEDOT:PSS/MoO3) led to enhanced electron blocking. This could explain the reduced degradation characteristics observed in our OPVs in which PEDOT:PSS interfaces with the active layer.

The second might be that silver migration and diffusion in the active layer through defects of the MoO3 layer is prevented by inserting the PEDOT:PSS:ZD layer, as assumed elsewhere.[28,29,48] This effect of atoms diffusing into the active layer has been investigated in several works: The Wantz group has observed a diffusion of silver atoms in MoO3 and organic layers,[29,48] but interestingly, they did not observe oxygen migration.[50] Additionally, the Österbacka group saw that diffusion of molecules into the active layer material causes doping effects that are detrimental to the device performance.[51] Our findings for the MoO3/Ag top contact seem to reflect these behaviors. However, as the detrimental effects have slowed down drastically upon insertion of the PEDOT:PSS:ZD layer in our inverted OPVs, we believe that this is a strong indication that indeed the silver top contact influences these degradation effects also.

Therefore, in our experimental approach, our observations reveal that the interfaces between the active layer and MoO3 as well as between MoO3 and Ag majorly contribute to the degradation of the P3HT:PC[70]BM-based inverted OPVs under heat-aging conditions and that electron blocking as well as diffusion of both MoO3 and Ag species may play a role in this.

To better understand the above observations on the degradation mechanism, the dark J/V characteristics of the different types of OPVs with different buffer layers at the initial and at different stages of degradation were obtained and are shown in Figure .

Figure 4

Current density vs voltage characteristics (J/V) in dark over time of exposure under heat conditions for inverted OPVs using different HTLs: (a) PEDOT:PSS:ZD, (b) PEDOT:PSS:ZD/MoO3, (c) MoO3, and (d) MoO3/PEDOT:PSS:ZD.

The inverted OPVs containing an MoO3 HTL as produced and measured in the dark (see Figure ) have a low leakage current at reverse bias, implying a high shunt resistance (Rp). In addition, at high forward bias, they have high current density and thus low series resistance (Rs) values and therefore their diodelike behavior is better than that of PEDOT:PSS:ZD-containing devices and accordingly they present higher PCE values, as shown in Table .

The dark J/V curves of devices with PEDOT:PSS:ZD as the HTL show more stable Rs and Rp values, which are slightly improved over time (Figure a). This shows that the organic–organic interface improves over time under heat conditions and that hole transportation is favored at the first stages of degradation. In addition, the performance may be assisted by the formation of a hole-selective thin P3HT layer on the top of the active layer under heating.[9]

Figure b shows the dark J/V for devices with PEDOT:PSS:ZD/MoO3 as the HTL. The Rp of these devices is slightly improved over time, whereas Rs is slightly decreased over time. We assume that the Rp is improved because of the changes in the organic/organic interface. However, we attribute the slight reduction in Rs to the interactions in the double interlayers affecting the carrier transportation in the vertical direction.

Inverted OPVs using MoO3 as the HTL show slightly reduced Rs over time of exposure. On the other hand, a significant leakage current and ideality factor increase over time, implying poor diode properties and MoO3/active layer interactions (Figure c). The increase in Rp of those devices is more intense and more obvious than that of the other inverted OPVs under study, and it occurs even after some hours of degradation. In addition, the Rp and ideality factor issues over time of exposure are also observed when inverted OPVs using MoO3/PEDOT:PSS:ZD as the HTL are tested (Figure d).

The latter reveals the detrimental influence of the P3HT:PC[70]BM/MoO3 interface in carrier selectivity.[52] This interfacial interaction over time of exposure promotes high leakage current, resulting in poor Rp and ideality factors, which are reflected in FF, as also shown in Figure .

Investigating the Degradation Mechanisms of P3HT:PC[70]BM/MoO3-Based Inverted OPVs under the ISOS-D-2 Protocol Using Reverse Engineering

Having seen that the inverted OPVs in which MoO3 directly interfaces with the active layer degrade the fastest, we wanted to investigate whether this interfacial interaction was reversible. We, therefore, fabricated incomplete device structures with ITO/ZnO/P3HT:PC[70]BM and ITO/ZnO/P3HT:PC[70]BM/MoO3 stacks. These were aged at 65 °C before the deposition of the subsequent fresh MoO3/Ag and Ag layers, respectively.

Figure illustrates the illuminated and dark J/V curves for devices with the MoO3 HTL as-produced and after aging for 20 h at 65 °C. Also, shown are the J/V characteristics for devices with degraded stacks after fresh evaporation of the top contact, enhancing the aspect that the MoO3/Ag interface also plays a role in this degradation. After reverse engineering,[11] illuminated J/Vs show that the Voc value is recovered when a new MoO3/Ag is deposited and also only when a new Ag layer is deposited on the ITO/ZnO/P3HT:PC[70]BM/MoO3 aged stack. This seems to suggest that the MoO3/Ag interface is responsible for the decay in Voc in devices using MoO3/Ag top contact.

Figure 5

(a) Illuminated and (b) dark J/V characteristics of complete devices with an MoO3 HTL as produced (black rectangles) and aged for 20 h (blue circles). The incomplete stacks were aged for 20 h at 65 °C and then coated with the required fresh electrode. The stacks were ITO/ZnO/P3HT:PC[70]BM, which was coated with fresh MoO3/Ag (red triangles), and ITO/ZnO/P3HT:PC[70]BM/MoO3, which was coated with fresh Ag (green diamonds).

In addition, the rectification of devices with an active layer/MoO3 interface changes over time and is getting more symmetrical (Figures c,d and 5b). However, in complete devices based on aged ITO/ZnO/P3HT:PC70BM/MoO3 upon evaporation of a fresh silver electrode or aged ITO/ZnO/P3HT:PC70BM with fresh MoO3/Ag, the ideality factor of devices with ITO/ZnO/P3HT:PC70BM/MoO3 and ITO/ZnO/P3HT:PC70BM aged stacks is recovered (see Figure ) and is similar in shape to the as-produced devices. This is an indicator that in these devices, aging of the MoO3/Ag interface is responsible for the change in the ideality factor.[48]

However, the fresh evaporation of Ag by itself does not lead to a recovery of Jsc. This happens only upon a fresh evaporation of both MoO3 and Ag on the aged stack of ITO/ZnO/P3HT:PC[70]BM. This could be another indication that the P3HT:PCBM/MoO3 interface is responsible for the degraded Jsc of the device over time of exposure under heat conditions. Overall, with a freshly evaporated MoO3/Ag electrode, we recover 90% of the efficiency of the as-produced devices, which is a clear indication that the device degradation is not affected by ITO/ZnO/P3HT:PCBM interfaces but almost exclusively by the interfaces of the top electrode P3HT:PCBM/MoO3/Ag, as shown in Figure . It should be noted that in principle, we expected to reach the same efficiency with the as-produced devices. However, we assume that the exposure to moisture in combination with the heat causes faster degradation and interface modification than the uncapped devices, thereby preventing full recovery by reverse engineering the top contact.

Conclusions

We monitored the lifetime performance of P3HT:PC[70]BM, PTB7:PC[70]BM, and DPPTTT:PC[70]BM inverted OPVs using thermally evaporated MoO3 as the HTL under heat conditions according to the ISOS-D-2 protocol (T = 65 °C, dark, and constant ambient RH of ∼40%). Inverted P3HT:PC[70]BM OPVs using thermally evaporated MoO3 as the HTL presented the most dramatic decay in all OPV parameters under 65 °C even after a few hours of exposure. A series of experiments were performed using photocurrent mapping and different device architectures to isolate the interfaces and reveal the failure mechanisms of P3HT:PC[70]BM devices under accelerated-heat conditions.

Upon incorporation of the two compared HTLs (evaporated MoO3 and solution-processed PEDOT:PSS:ZD) in different configurations in inverted OPVs with the P3HT:PC[70]BM active layer, we observed that the P3HT:PC[70]BM/MoO3 interface is the most sensitive part of this degradation triggered at 65 °C. This is primarily shown by the reduction in FF and slower degradation of inverted OPVs with the MoO3/PEDOT:PSS:ZD double interlayer, which isolates the MoO3 from the Ag. This was in agreement with the dark J/Vs of inverted OPVs, which show that the ideality factor of OPVs with the P3HT:PC[70]BM/MoO3 interface is affected, thereby further revealing the detrimental influence of this interface. After an alternative reverse engineering method, we showed that the Voc and the ideality factor of such devices are recovered when fresh Ag is deposited on the top of the aged ITO/ZnO/P3HT:PC[70]BM/MoO3 stacks. This showed that the Ag electrode influences the stability of inverted OPVs with MoO3 as the HTL under heat conditions and was confirmed by the suppressed decay of inverted OPVs using MoO3/PEDOT:PSS:ZD as the HTL.

To conclude, in the case of inverted P3HT:PC[70]BM OPVs with MoO3/Ag top contact, which were extensively studied in this paper using a series of measurements and device/reverse engineering methods, the results presented indicate that the heat degradation of the inverted OPVs is not affected by the ITO/ZnO/P3HT:PCBM interfaces but almost exclusively by the interfaces of the top electrode P3HT:PCBM/MoO3/Ag. Although the interface between MoO3 and Ag contributes to degradation, the interface between the P3HT:PC[70]BM active layer and MoO3 is the main origin of failure of inverted OPVs under intense heat conditions, a trend that was observed for the three thiophene-based materials used in this study of inverted OPVs. The PEDOT:PSS:ZD HTL resulted in a more stable inverted OPV performance compared with MoO3 HTL under the ISOS-D-2 heat protocol, a fact which should be taken into account when reporting high-efficiency devices using MoO3 as the HTL.