The effect of charge extraction layers on the photo-stability of vacuum-deposited versus solution-coated organic solar cells

Organic Electronics 15 (2014) 47–56

Contents lists available at ScienceDirect

Organic Electronics journal homepage: www.elsevier.com/locate/orgel

The effect of charge extraction layers on the photo-stability of vacuum-deposited versus solution-coated organic solar cells Graeme Williams, Hany Aziz ⇑ Department of Electrical and Computer Engineering & Waterloo Institute for Nanotechnology, University of Waterloo, 200 University Avenue, Waterloo, ON N2L 3G1, Canada

a r t i c l e

i n f o

Article history: Received 31 July 2013 Received in revised form 2 October 2013 Accepted 16 October 2013 Available online 5 November 2013 Keywords: Organic photovoltaics Stability Degradation Extraction layer Buffer layer Interfacial

a b s t r a c t Organic solar cells (OSCs) are studied for their photo-stability in inert atmosphere. Polymer solar cells with a bulk heterojunction (BHJ) of poly(3-hexylthiophene) (P3HT) and phenylC61-butyric acid methyl ester (PCBM) are contrasted with small molecule solar cells with a BHJ of chloroindium phthalocyanine (ClInPc) and C60-fullerene. A series of charge extraction layers at the hole and electron collecting contacts are examined for their role in OSC performance and stability. The inter-compatibilities of these extraction layers in vacuum-deposited small molecule OSCs (SM-OSCs) versus solution-coated polymer OSCs (P-OSCs) are explored. Through photo-stability studies, we show that interfacial extraction layers are necessary to avoid contact photo-degradation, which otherwise leads to strong reductions in OSC efficiencies. We also highlight certain extraction layer combinations that result in strong inter-electrode degradation, and we discuss incompatibilities in extraction layers among SM-OSCs versus P-OSCs. Our results suggest that the presence of excitons at the organic-electrode interface likely plays a critical role in contact photo-degradation. By minimizing contact photo-degradation, which dominates the majority of short-term OSC degradation, a new avenue for studying OSC stability behavior and opportunities to focus on other losses in OSCs become possible. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Organic solar cells (OSCs) have achieved significant gains in efficiency in the past decade, now allowing for >10% power conversion efficiency (PCE) tandem structures [1–3]. However, in contrast, OSC stability research and the underlying degradation phenomena in OSCs are much less studied. The stability of OSCs is particularly difficult to study due to the large number of device parameters and experimental factors that may affect degradation, as well as the possibility for interplay among these factors. To this end, researchers have identified various potential avenues toward degradation, including photo-degradation [4–8], ambient (H2O/O2) degradation [9–13] and thermal degradation [14–16] of the photo-active organic layer. Photo⇑ Corresponding author. Tel.: +1 5198884567x36848. E-mail address: [email protected] (H. Aziz). 1566-1199/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.orgel.2013.10.017

and thermal-instability of the organic-electrode interfaces have also recently been highlighted as one of the most critical avenues toward OSC degradation [17–22]. We recently showed, by using X-ray photo-electron spectroscopy (XPS) measurements, that the organic-electrode interfaces commonly used in OSCs are inherently unstable and that they exhibit photo-induced chemical changes [20–22]. It behooves us to gain a clearer understanding of the basic role of the OSC extraction layer, and to elucidate the phenomena involved in the interfacial degradation mechanism. The umbrella term ‘OSC’ may be broken up into small molecule OSC (SM-OSC) and polymer OSC (P-OSC). While P-OSCs are always solution-coated, SM-OSCs can be either vacuum deposited or solution-coated. Despite the different fabrication methodologies, high efficiency devices can be achieved for all of these cases [1–3], usually in combination with suitable extraction layers. It is thus intriguing that the various approaches in device fabrication

48

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

(solution-coated versus vacuum-deposited) generally employ different extraction layers, while making use of the same electrode materials – indium tin oxide (ITO) for the anode and Al for the cathode. P-OSCs and solution-coated SM-OSCs usually employ a poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS) hole extraction layer (HEL) and a thin (1 nm) LiF electron extraction layer (EEL) [23]. In contrast, vacuum deposited SM-OSCs often use a plasma treatment-based HEL [24,25] and a bathocuproine (BCP) EEL [26,27]. The relative stabilities and efficacies of these and several other extraction layers (such as CF4 plasma treatment [25,28,29] and MoO3 [30,31]) when directly comparing the different fabrication methodologies are, as of yet, unstudied. It follows that a comprehensive study of the inter-compatibilities and stabilities of these interfacial layers for solution-coated versus vacuumdeposited OSCs would be of substantial value. In this work, we conduct a systematic study on the effect of commonly employed extraction layers on OSC photo-stability, as well as on the inter-compatibilities of these extraction layers in both solution-coated P-OSCs and vacuum-deposited SM-OSCs. By illuminating both SM-OSCs and P-OSCs in a controlled N2 environment, we elucidate the impact of the interfacial layers on device lifetime and photovoltaic parameters. Our results further demonstrate that, although both SM-OSCs and P-OSCs suffer from contact photo-degradation, the use of extraction layers generally improves their photo-stability. Optimal HEL/EEL combinations are identified for both vacuumdeposited SM-OSC and P-OSC materials systems. Incompatible HEL/EEL combinations, as observed by inter-electrode degradation, are also highlighted. We suggest that the presence of excitons at the organic-electrode interface may be fundamentally important in contact photo-degradation. By minimizing contact photo-degradation, which dominates the majority of short-term OSC degradation, a new avenue for studying OSC stability behavior and opportunities to focus on other losses in OSCs become possible.

2. Experimental In this work, both solution-coated and vacuum-deposited OSCs were fabricated on patterned ITO/glass slides. P-OSCs are comprised of a 1:1 mixture of the ubiquitous poly(3-hexylthiophene) (P3HT) donor and [6,6]-phenylC61-butyric acid methyl ester (PCBM) acceptor, while SM-OSCs are comprised of a 1:3 mixture of chloroindium phthalocyanine (ClInPc) donor and C60/fullerene acceptor. Sublimed ClInPc was provided by the Xerox Research Centre of Canada, while other materials were purchased commercially. Detailed information regarding materials, as well as further information regarding solution preparation and substrate cleaning can be found in our previous work [20,25]. All devices were made following the standard procedure that involves the successive application of HEL, active layer, EEL and top cathode. For solution-coated OSCs, a 70 nm active P3HT:PCBM layer was formed by spincoating 1:1 P3HT:PCBM in chlorobenzene (20 mg/mL solids) at a spin speed of 1100 RPM for 60 s, followed by annealing at 110 °C for 10 min. For vacuum-deposited OSCs, the

active layer comprised a 20 nm ClInPc:C60 (1:3) mixed layer, followed by a 30 nm pure C60 layer, deposited at a total rate of 2 Å/s. All materials were deposited by thermal evaporation under high vacuum at a base pressure below 5 microtorr. For HELs: 5 nm MoO3 films were deposited by thermal evaporation; 30 nm PEDOT:PSS films were spincoated from Clevios P VP Al4083 PEDOT:PSS at 2300 RPM, followed by annealing at 180 °C for 10 min; CF4 plasma treatment was accomplished by exposure to CF4:O2 (3:1) plasma using a reactive ion etch system equipped with an inductively coupled plasma source at 100 W and gas pressure of 20 Pa for 2 min. For EELs: 1 nm LiF and 8 nm BCP films were deposited by thermal evaporation. A 100 nm Al top cathode was deposited by thermal evaporation to complete the devices. All tests and photo-/thermalstability experiments were conducted in a dry N2 environment. Photo-stability tests were carried out with 100 mW/ cm2 white light, with the device temperatures maintained at 40 °C. Photovoltaic parameters were measured with 1sun AM1.5G radiation from an ABET Sun 3000 Class AAA Solar Simulator and a Keithley 2400 SourceMeter. Atomic Force Microscope (AFM) images were obtained in air using VEECO Dimension 3100 AFM in tapping mode.

3. Results and discussion 3.1. Photo-stability of organic-electrode interfaces with various interfacial layers An illustration of the relevant devices structures for the P-OSCs and SM-OSCs are provided in Fig. 1a and b respectively. The presented device structures and comprising materials were chosen for their high degree of OSC performance reproducibility, which is essential for stability studies, and also because their initial PCE values were roughly equivalent (2–2.2% PCE for optimized HEL/EEL), allowing for simpler comparisons across the two different fabrication methodologies. While the acceptor materials are fundamentally similar (C60 versus PCBM)), the donor materials for the two systems are different in both their structure and their energy levels, the latter illustrated in Fig. 1c [29,32–37]. As such, the present work considers two very different OSC systems – in fabrication methodology, constituent materials and device structure – to elucidate possible commonalities regarding their stability. As will be proven in the body of this work, in spite of these differences, both of the examined SM-OSCs and P-OSCs show strong similarities in their device stability that are largely associated with photo-induced organic-electrode interfacial degradation. Both P-OSC and SM-OSC samples were continuously illuminated over a period of 84 h in a dry N2 environment. The photovoltaic output parameters were measured at fixed time intervals during this period. In order to gauge any degradation simply due to the storage of the devices (i.e. device shelf life), identical samples were also fabricated and kept in the dark in a dry N2 environment and measured at the same time intervals. The initial photovoltaic parameters for representative SM-OSC and P-OSC samples (prior to illumination) are detailed in Tables 1 and 2. It

49

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

(a)

 PEDOT:PSS works well for P-OSCs, and it may work for SM-OSCs; however, it is incompatible with ClInPc:C60 SM-OSCs that rely on the Schottky junction architecture (discussed further in Section 3.3).  P-OSCs that use a PEDOT:PSS HEL and no EEL can show reasonable performance.

(b)

– EEL:  LiF works well for P-OSCs, but does not work well for SM-OSCs as it does not suitably protect the organic layers from damage during metal deposition, especially given the relatively high roughness of films including C60 (discussed further in Section 3.1.1).  BCP works well for both SM-OSCs and P-OSCs; however, it suffers from incompatibilities with PEDOT:PSS (discussed further in Section 3.1.2).

(c)

Fig. 1. Illustration of the OSC device structures used in this study for (a) vacuum-deposited small molecule organic solar cells and (b) solutioncoated polymer solar cells. (c) Energy level diagram for the constituent materials in (a) and (b).

is immediately obvious that, in nearly all cases, using either no HEL or no EEL results in very poor device performance, as is well established in the field [38]. The following general observations can be made regarding the initial performance of OSCs made by the two different fabrication methodologies (vacuum-deposited and solutioncoated OSCs): – HEL:  CF4 plasma treatment gives performance on par with MoO3, and both HELs work for SM-OSCs and P-OSCs.

3.1.1. Vacuum-deposited small molecule organic solar cells Since the majority of stability studies have focused primarily on P-OSCs [8,10,20], it is of immediate interest to observe the photo-stability of vacuum-deposited SM-OSCs. The normalized PCE values of the ClInPc:C60 SM-OSCs with varying HELs/EELs during the 84-h light illumination scheme are plotted in Fig. 2. In order to better isolate the effects of illumination on the samples, the data presented in Fig. 2 has been normalized relative to the PCE of identical devices kept in the dark at the given time. This normalization helps to remove any degradation effects or other variations in efficiency simply due to storage of the devices, and therefore it allows us to better observe degradation purely due to light exposure. The original nonnormalized data, including data for both the illuminated samples and the samples kept in dark, are provided in the Supplemental Information. It is worth noting that the devices that had suitable interfacial layers and that were kept in the dark generally degraded very little, immediately suggesting that the shelf life of these devices is much longer than the lifetime of illuminated devices. As it will become evident from the vast amount of raw data (as provided in the Supplemental Information), this normalization methodology assists in cleanly analyzing and identifying trends, while avoiding the obvious consequences of ‘big data.’

Table 1 Initial (t = 0) power conversion efficiency values (bolded) for ITO/HEL/ClInPc:C60/C60/EEL/Al solar cells with various hole extraction and electron extraction Layers. Jsc, Voc and FF are also shown in adjacent rows. HEL

EEL None

BCP

BCP + LiF

None

0.01%

0.19 mA/cm2 59 mV 25%

0.27%

3.18 mA/cm2 213 mV 34%

0.14%

2.70 mA/cm2 151 mV 29%

PEDOT:PSS

0.30%

2.54 mA/cm2 511 mV 23%

1.24%

4.51 mA/cm2 652 mV 42%

1.25%

4.58 mA/cm2 660 mV 41%

CF4

0.27%

2.52 mA/cm2 420 mV 25%

2.17%

4.57 mA/cm2 919 mV 52%

2.12%

4.44 mA/cm2 921 mV 52%

MoO3

0.42%

2.98 mA/cm2 516 mV 28%

2.15%

4.63 mA/cm2 982 mV 47%

2.17%

4.67 mA/cm2 985 mV 47%

50

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

Table 2 Initial (t = 0) power conversion efficiency values (bolded) for ITO/HEL/P3HT:PCBM/EEL/Al solar cells with various hole extraction and electron extraction Layers. Jsc, Voc and FF are also shown in adjacent rows. HEL

EEL None

LiF 2

BCP

BCP + LiF

None

0.02%

0.58 mA/cm 135 mV 22%

0.04%

1.26 mA/cm 201 mV 18%

0.04%

1.43 mA/cm 188 mV 17%

0.03%

1.05 mA/cm2 160 mV 18%

PEDOT:PSS

1.72%

6.18 mA/cm2 553 mV 50%

2.07%

6.22 mA/cm2 615 mV 54%

1.99%

6.29 mA/cm2 610 mV 52%

1.85%

6.12 mA/cm2 617 mV 49%

CF4

0.44%

5.27 mA/cm2 277 mV 29%

1.83%

6.60 mA/cm2 584 mV 47%

2.08%

6.89 mA/cm2 609 mV 50%

1.94%

6.75 mA/cm2 608 mV 47%

MoO3

0.94%

5.37 mA/cm2 382 mV 46%

1.87%

5.72 mA/cm2 593 mV 55%

2.11%

5.80 mA/cm2 630 mV 58%

2.06%

5.86 mA/cm2 632 mV 56%

From Fig. 2, it is clear that, without some form of HEL or EEL, the ClInPc:C60 SM-OSCs are very photo-unstable. Within 36 h, the PCE values for all devices without an HEL (Fig. 2a) decrease to 0% due to a strong deterioration in photocurrent density (Jsc), open circuit voltage (Voc) and fill factor (FF). Similarly, those devices with no EEL suffer very strong photo-deterioration, although not as severe as the devices with no HEL. To this end, the ‘No EEL’ devices with PEDOT:PSS and MoO3 HELs decrease to 40% and 65% of their original PCE within 84 h of continuous light exposure. In contrast, the use of either no HEL or a CF4 HEL with no EEL results in complete degradation of photovoltaic characteristics in the 84 h of illumination. It is also worth noting that the initial efficiencies of devices without HELs and EELs is quite poor, as shown in Table 1, which makes them very sensitive to further deteriorations in Jsc, Voc and FF. The results strongly demonstrate the photoinstability of these devices. We now turn our attention to those devices with both an HEL and an EEL, shown in Fig. 2b–d. It is worth noting that SM-OSCs generally employ an organic EEL – typically BCP – in contrast to the most commonly used inorganic EEL for P-OSCs: LiF. In the past, we have found that light harvesting devices with organic-metal interfaces are photo-unstable [20,22]; however, in those experiments, the organic material also served as a photo-active component in the device operation and thus contained excitons during illumination. The use of an organic extraction layer (i.e. BCP) that has no role in the generation of excitons can help show whether this interfacial degradation is merely due to the weak chemical nature of organic-metal interfaces in general, or due to exciton interactions at the organic–cathode interface. To better elucidate the effect of the EEL in the stability of SM-OSCs we have included three EEL variants in this study: BCP (8 nm), LiF (1 nm) and BCP(8 nm)/LiF(1 nm). Comparisons of BCP with BCP/LiF will immediately show whether an organic–inorganic– cathode interface offers stability enhancements over the simple organic-cathode interface. Upon examination of Fig. 2b–d, it is clear that the presence of a BCP or BCP/LiF EEL significantly improves the photo-stability of ClInPc:C60 OSCs with all three of the

2

2

examined HELs (PEDOT:PSS, CF4 plasma treatment and MoO3). PEDOT:PSS/. . ./BCP and PEDOT:PSS/. . ./BCP/LiF configurations lose 20% of their original PCE after 84 h of illumination, with a 5% loss to FF and more severe 10% losses to both Jsc and Voc. The use of a MoO3 HEL instead grants slightly better photo-stability, with only 15% reduced PCE after 84 h, due to only minor (few percent) losses to Voc, but the same 5% loss to FF and a 10% loss to Jsc. The most stable ClInPc:C60 device, however, employs a CF4 HEL and a BCP or BCP/LiF EEL, allowing for only a 10% reduction in PCE after 84 h of illumination. In this case the PCE reduction is due to few-percent losses to both Jsc and Voc, and the same 5% loss to FF. Plots of these major photovoltaic parameters (PCE, Jsc, Voc and FF) versus illumination time are available in the Supplemental Figures. Therefore, based on these results, the most stable SMOSC employs a CF4 plasma treatment as an HEL and a BCP EEL. Based on our previous work, which employed XPS to show that the organic-electrode interfaces are especially susceptible to photo-induced chemical changes [21,22], the instabilities of the organic-electrode interfaces observed here are suggested to be due to detrimental photo-chemical interactions between the photo-active organic species and the electrode. These interactions are found to occur due to exposure to light, as samples kept in dark showed either no change or substantially smaller changes in their photovoltaic parameters when compared to those devices exposed to light. In all configurations, the SM-OSCs lose 5% in FF, indicating the likelihood of an additional degradation mechanism beyond the photoinstability of the organic-electrode interface. Interestingly, the SM-OSCs with an LiF EEL showed degradation behavior very similar to that of a device with no EEL whatsoever. Note that the photo-stability data for SM-OSCs with LiF EELs are not provided for PEDOT:PSS and MoO3 HELs because the initial performances of these devices were generally too low to be useful. Furthermore, the combined BCP/LiF EEL showed stabilities similar to that of BCP alone – the curves of BCP and BCP/LiF EEL devices largely overlap. To understand the inefficacy of LiF here, it is useful to examine the function of the EEL in SM-OSCs as established from literature:

51

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

0.8

0.8

0.6

0.6

PCE

(b)1.0

PCE

(a) 1.0

0.4 0.2 0.0

0.4 0.2

0

20

40

60

80

0.0

100

0

20

Time (hrs)

0.8

0.8

0.6

0.6

PCE

(d)1.0

PCE

(c) 1.0

0.4 0.2 0.0

40

60

80

100

80

100

Time (hrs)

0.4 0.2

0

20

40

60

80

100

Time (hrs)

0.0

0

20

40

60

Time (hrs)

Fig. 2. Normalized power conversion efficiencies for ClInPc:C60 solar cells with varying HELs (a->d) and EELs over 84 h of Illumination. HELs include: (a) No HEL, (b) PEDOT:PSS, (c) CF4 plasma treatment and (d) MoO3.

– To aid in work function alignment of the cathode and/or aid with conduction of electrons to the cathode, same as with P-OSCs [39,40]. – To block excitons from reaching the adjacent electrode, unique to SM-OSCs (generally not considered for POSCs) [26,41]. – To protect the underlying organic layers from damage during deposition of the top metal cathode, also unique to SM-OSCs [26,41]. While LiF may help to satisfy the first role of work function alignment, a 1 nm LiF layer is insufficient to protect the underlying organic layers from metal damage during cathode deposition or to physically block excitons from reaching the adjacent electrode. This is particularly relevant when considering the RMS roughness values of the organic layer adjacent to the EEL, as determined by AFM. An AFM scan of the C60 layer (i.e. the layer adjacent to the EEL in the SM-OSC) is provided in Fig. 3a. For reference, an AFM image of the 1:1 P3HT:PCBM layer (i.e. the layer adjacent to the EEL in the P-OSC) is also provided in Fig. 3b. From this data, we found the roughness of the neat C60 layer is quite high (3.3 nm RMS roughness), especially when compared to that of the 1:1 P3HT:PCBM BHJ in the P-OSC (1.4 nm RMS roughness). It is worth noting that using a 1:1 BHJ of ClInPc:C60 does decrease the roughness slightly (2.5 nm RMS roughness), but not to the same level as with the P-OSC BHJ. To this end, LiF is similarly ineffective for a simple 1:1 ClInPc:C60 BHJ SM-OSC (i.e. without a neat C60 layer). The penetration of aluminum through 1 nm thick LiF might also have undesirable effects on device stability

and lifetime by the same photo-chemical degradation phenomenon that occurs at organic-electrode interfaces. Finally, in the case of the BCP/LiF EEL, the presence of LiF does not have any effect, whether positive or negative, on the device characteristics or photo-stability, indicating that the BCP EEL is already sufficient to carry out all of its functions denoted above. From these results, it is becoming clear that the cathode photo-degradation phenomenon is strongly linked to the presence of excitons at the organic-metal interface (and not merely due to the creation of a ‘weak’ organic-metal interface). BCP, with its large bandgap compared to either the donor or acceptor material, provides a necessary buffer to physically separate excitons from this interface and thus it can effectively limit photo-degradation. As such, there are two viable strategies for the selection of extraction layers for high stability OSCs: either form highly stable interfaces (e.g. through the use of inorganic layers, as is the case with LiF in P-OSCs), or prevent direct contact between the electrode and the photo-active species (e.g. with a wide bandgap organic extraction layer). 3.1.2. Solution-coated polymer solar cells The normalized PCE values of the P3HT:PCBM P-OSCs with varying HELs/EELs during the 84-h illumination scheme are plotted in Fig. 4. Following the same approach as in Section 3.1.1, the data has been normalized relative to the PCE of identical devices kept in the dark at the given time. The original non-normalized data, including data for both the illuminated samples and the samples kept in dark, are provided in the Supplemental Information. As

52

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

Fig. 3. Atomic force microscope image of (a) C60 and (b) 1:1 P3HT:PCBM films. RMS roughness values are 3.3 nm and 1.4 nm for (a) and (b) respectively.

0.8

0.8

0.6

0.6

PCE

(b) 1.0

PCE

(a) 1.0

0.4 0.2 0.0

0.4 0.2

0

20

40

60

80

0.0

100

0

20

Time (hrs)

0.8

0.8

0.6

0.6

PCE

(d) 1.0

PCE

(c) 1.0

0.4 0.2 0.0

40

60

80

100

80

100

Time (hrs)

0.4 0.2

0

20

40

60

80

100

Time (hrs)

0.0

0

20

40

60

Time (hrs)

Fig. 4. Normalized power conversion efficiencies for P3HT:PCBM solar cells with different HELs and EELs over 84 h of Illumination. HELs include: (a) No HEL, (b) PEDOT:PSS, (c) CF4 plasma treatment and (d) MoO3.

with the ClInPc:C60 OSCs, the storage of the devices with suitable HELs/EELs resulted in small variations in photovoltaic parameters (with only one exception, PEDOT:PSS/ P3HT:PCBM/BCP devices, which are discussed further below). In concurrence with our previous results [20], P3HT:PCBM P-OSCs with either no EEL or no HEL exhibit poor photo-stability. Devices with no HEL suffer a significant decrease in PCE to 0–20% of their initial PCE within 84 h of illumination regardless of the choice of EEL. P-OSCs with no EEL also photo-degrade, but not nearly as much as SM-OSCs, losing only 40%, 20% and 15% of their initial PCE for CF4 plasma treatment, PEDOT:PSS and MoO3 HELs respectively. P-OSCs that employ both an HEL and an EEL exhibit a significantly enhanced photo-stability. Following similar

behavior as the SM-OSCs, BCP and BCP/LiF EELs give roughly the same trends during illumination, so the latter are not discussed in depth here. As shown in Fig. 4b, PEDOT:PSS/. . ./LiF P-OSCs are relatively photo-stable, with only a 5–10% decrease in PCE after 84 h of light exposure. In this case, the reductions are due to slight losses in Voc and FF. From Fig. 4c, devices with either LiF or BCP in combination with a CF4 HEL are shown to exhibit reasonable stability. The CF4/. . ./BCP configuration allows for only a 20% reduction in PCE, while the CF4/. . ./LiF configuration shows only a 10% decrease in PCE after 84 h of illumination. From Fig. 4d, MoO3/. . ./LiF P-OSCs are also reasonably stable with only a 10% loss in PCE. The highest stability POSCs employ the MoO3/. . ./BCP configuration, allowing for only a 5% reduction in PCE over 84 h of illumination,

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

which appears to arise primarily from losses in FF. Therefore, based on these results, the most stable P-OSC employs a MoO3 HEL and a BCP EEL. From Fig. 4b, PEDOT:PSS/. . ./BCP devices are demonstrated to exhibit very poor stability, losing 80% of the original PCE within 84 h of light exposure – a trend in direct contrast to all other devices with a BCP EEL. Since PEDOT:PSS HEL devices show reasonable stabilities with an LiF EEL, and BCP EEL devices show reasonable stabilities with both CF4 and MoO3 HELs, it is clear that BCP and PEDOT:PSS cannot be employed simultaneously. PEDOT:PSS has been suggested to contain residual humidity that can adversely affect device performance [19], and BCP has been shown to suffer from moisture-induced degradation [42]. It is thus logical to study the thermal stability of these devices, on the basis that any thermally assisted out-diffusion of moisture from PEDOT:PSS would adversely affect the OSC lifetime by interacting with BCP. As such, PEDOT:PSS/. . ./BCP P-OSCs were heated in the dark in a N2 environment at the temperature that the illuminated samples reach (40 °C). Their output parameters were measured periodically up to a period of 36 h (i.e. the first data point in the concurrent photo-stability experiments). The normalized PCE values for the heated samples are plotted alongside illuminated data (for the same structure P-OSC) in Fig. 5a. The PEDOT:PSS/. . ./BCP samples are shown to degrade rapidly with heat, with a 90% loss in PCE in only 36 h. Furthermore, the 36-h heated sample data point matches reasonably well with the 36-h illuminated sample data point. This indicates that, for this particular combination of materials, heat-induced degradation likely dominates photo-degradation. The results suggest that this poor stability may be largely due to the heat-mediated release of moisture from PEDOT:PSS and its subsequent interaction with the BCP interlayer. While heat plays a major role in the OSC losses for P-OSCs with a PEDOT:PSS HEL and a BCP EEL, it is worth recalling that heat-induced degradation is not always the primary cause of OSC contact degradation and the associated PCE losses. In our previous work, we established that P-OSCs with no HEL and no EEL have similar PCE versus time behavior whether they are heated or simply kept in the dark in N2 [20]. These same devices, however, showed strong reductions in PCE when illuminated. We have further shown through XPS measurements that organic-electrode interfaces are especially susceptible to photo-induced chemical changes [21,22]. In these latter experiments, the light intensities were generally low enough that the devices did not experience any significant heating. Our results thus stress the importance of considering all possible avenues toward device degradation in OSC stability studies. Interestingly, for the ClInPc:C60 SM-OSCs, the PEDOT:PSS/. . ./BCP combination did not show as significant deterioration, with only a 20% loss in PCE after 84 h of illumination. Since it is unknown if this loss is similarly due to heat-induced effects, the same experiment was conducted for the SM-OSCs, and the PCE values for the heated and illuminated devices are both plotted in Fig. 5b. From this data, it is clear that 15% of the reduction in PCE in the first 36 h can be attributed exclusively due to thermal effects. Thus, it is supposed that this moisture release still

53

affects SM-OSCs, but not to the same degree as P-OSCs. Work completed by Voroshazi and coworkers suggests that C60 may act as a form of ‘getter’ material for residual moisture in OSCs [19]. Since the SM-OSCs employ both a high concentration of C60 and a neat C60 layer, this may explain the higher stability compared to the P-OSCs. If this caveat holds true, it implies that PCBM is not capable of acting in the same regard as a getter, at least when mixed with P3HT with a 1:1 mixing ratio. From this data, it is shown that the reduction in the PCE of OSCs under illumination is a widespread phenomenon, as it is observed in both SM-OSCs and P-OSCs and for all HEL/. . ./EEL configurations. It is also clear that the contacts play a critical role in the photo-stability behavior of these devices, with proper choices of HEL and EEL allowing for considerable stability improvements for illuminated samples. Finally, there is interplay between the HEL and the EEL that must be considered when studying the photo-stability behavior of OSCs, whether solution-coated or vacuum-deposited. It follows that the HEL and the EEL must be chosen carefully when considering the device’s ultimate stability. 3.2. Observations for the photo-stability of SM-OSCs versus P-OSCs It was established that the most stable configuration of SM-OSC employs a CF4 plasma treatment HEL and a BCP EEL. The most stable configuration of P-OSC, however, employs a MoO3 HEL and a BCP EEL. It is worthwhile to examine the different features of these two HELs that make them suitable for SM-OSCs and P-OSCs respectively, and thus make them suitable for the two different OSC fabrication methodologies (with either vacuum-deposited or solution-coated photo-active layers). For CF4 plasma treatments, the ITO surface is passivated with fluorine atoms; however, spincoating is a mechanically harsh process that may disrupt or alter the ITO surface chemistry compared to vacuum deposition. In contrast, MoO3 is much more suited as a foundation for subsequent spincoating processes of the polymer layers, and thus less likely to show regions of uncovered/untreated ITO. Such untreated regions in CF4 HEL-based P-OSCs would ultimately undergo photodegradation processes to lead to a strong reduction in photovoltaic output properties – essentially due to a loss in active device area. Given the similarity between the CF4 plasma surface treatment and self-assembled monolayer surface treatments, it is also feasible that the variations observed here are related to the selectivity of the contact [43], which may vary for the different materials systems and for the different fabrication methodologies. Regardless, given that inorganic oxides can be deposited as much thicker, denser films, it is suggested that they provide a more reliable platform for subsequent device fabrication and are therefore much more reliable extraction layers in general. In this work, we have established that a majority of the photo-degradation of output parameters, especially in the short-term, can be minimized through the choice of certain combinations of extraction layers. Once the optimum HEL/ EEL combinations have been chosen, the losses to Jsc and

54

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

(a) 1.0

Light Heat

0.6 0.4 0.2 0.0

Light Heat

0.8

PCE

PCE

0.8

(b)1.0 0.6 0.4 0.2

0

20

40

60

80

100

0.0

0

20

40

Time (hrs)

60

80

100

Time (hrs)

Fig. 5. Normalized power conversion efficiencies for (a) P3HT:PCBM and (b) ClInPc:C60 solar cells with a PEDOT:PSS HEL and a BCP EEL while Illuminated and heated.

V (V) 0

0.0

0.2

0.4

0.6

0.8

1.0

I (mA)

(a)

-1 ClInPc:C60, Pre-Illumination ClInPc:C60, Post-Illumination P3HT:PCBM, Pre-Illumination P3HT:PCBM, Post-Illumination

I (A)

(b) 4

I (mA)

Voc during illumination are strongly reduced. However, in spite of extraction layer optimization, all device configurations show a few percent decrease in Voc and a 5% decrease in FF. This trend is observed more clearly in Fig. 6a, which shows the pre-/post-illumination light IV characteristics of SM-OSCs and P-OSCs with optimal HELs and EELs. Further, from the dark IV characteristics of these devices, as plotted in Fig. 6b, post-illuminated devices have more leakage current in both forward and reverse bias. As the Voc is fundamentally related to the reverse saturation current [44], leakage current under reverse bias necessarily leads to a reduction in Voc. Since we observe an increase in current in both forward bias as well as reverse bias, it is suggested that there are additional shunts through the post-illuminated devices. This shunting and the associated decrease in FF merits further investigation. To this end, by minimizing this fast photo-induced contact degradation, it is now feasible to focus on additional degradation pathways that happen over a larger time scale, including photo-deterioration of the organic bulk materials or even photo-induced alterations at the organic–organic interfaces throughout the mixed donor–acceptor active layer.

2 -8

-3x10

-0.5

V (V)

0.0

3.3. Further general observations As established in Section 3.1, ClInPc:C60 SM-OSC samples that employ a PEDOT:PSS HEL exhibit a reduced initial PCE of 1.24%, which is much lower than the CF4 and MoO3 HEL samples. To understand this effect, it is necessary to examine each photovoltaic parameter in greater detail. SM-OSCs with a PEDOT:PSS HEL have Jsc values on par with those of the CF4 and MoO3 HEL devices; however, their FF is lower and their Voc is strongly reduced (650 mV for PEDOT:PSS versus 920 mV and 980 mV for CF4 and MoO3 respectively). The 650 mV Voc observed with the PEDOT:PSS HEL is near (slightly lower than) the theoretical Voc defined by the ClInPc HOMO/C60 LUMO offset in the bulk heterojunction (BHJ). The Voc enhancements observed for MoO3 and CF4 HELs are due to the Schottky junction effect that arises when a high work function electrode contacts a BHJ with high acceptor concentration, as has recently been established in literature [25,33,45]. It can thus be concluded that PEDOT:PSS may act as a suitable HEL for

0

-0.5

0.0 V (V)

0.5

1.0

Fig. 6. (a) Light IV and (b) dark IV curves for ClInPc:C60 and P3HT:PCBM solar cells before and after Illumination (Inset of B shows the dark current at negative bias at a magnified scale). Device structures are ITO/CF4/ ClInPc:C60/C60/BCP/Al and ITO/MoO3/P3HT:PCBM/BCP/Al.

these SM-OSCs, but it does not provide a sufficiently high work function to allow for the band bending and the subsequent Voc improvement essential for Schottky OSCs. In contrast, both MoO3 and CF4 plasma treatment HELs have sufficiently high work functions to allow for the Schottky OSC architecture. It is also worth noting that PEDOT:PSS has been used successfully in Schottky P-OSCs employing high concentrations of PCBM in P3HT:PCBM cells [46]. As such, the HELs that may allow for the formation of a

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56

Schottky OSC must also be considered in the context/ choice of the principal photo-active materials. Through this work, we have shown BCP to be a suitable EEL for both SM-OSCs and P-OSCs, granting good initial PCE values and high OSC lifetimes. This implies that BCP satisfies all of the EEL criteria established in Section 3.1.1 – exciton blocking, metal damage protection and work function alignment. However, the role of BCP in work function alignment is, as of yet, unclear. To this end, considering P3HT:PCBM P-OSCs, Reese et al. previously demonstrated that the high work function of an Al cathode causes a reduction in both Voc and FF; however, the use of an LiF/Al contact recoups these losses to grant good performance, near that of P-OSCs employing a Ca/Al or Ba/Al cathode [47]. The observed improvements with the use of an LiF EEL are attributed to the shallow work function of LiF/Al compared to Al alone, which is due to dipole formation at the LiF/Al interface [40]. Given that we observe a similar level of performance enhancement for P3HT:PCBM P-OSCs with a BCP EEL, especially with regard to the Voc (>600 mV) , it is suggested that the thin BCP EEL must also help in some regard with work function alignment. Such a possibility was recently shown by Xiao et al. who observed that solution-coated films of a similar wide bandgap electron transport material, Bathophenanthroline (Bphen), can be used as an EEL for inverted P3HT:PCBM P-OSCs [48]. For these inverted OSCs, the adjacent cathode is ITO (work function of 4.3–4.8 eV [49]), so the Bphen EEL must necessarily result in a low work function contact to grant good device performance and the observed Voc of 600 mV.

4. Conclusions We have conducted a systematic comparison of the photo-stability of vacuum-deposited SM-OSCs versus solution-coated P-OSCs, with a focus on the photo-stability of the organic–electrode interface. For both device fabrication methodologies, the organic–electrode interface is shown to be photo-unstable. The use of both an HEL and an EEL can drastically suppress contact photo-degradation. While commonly used HELs and EELs in SM-OSCs (CF4 plasma treatment and BCP) can be applied to P-OSCs, the opposite is not necessarily true – both PEDOT:PSS and LiF result in low efficiency SM-OSCs. Optimized HEL/EEL configurations allow for high efficiencies and highly photo-stable SMOSCs and P-OSCs, which exhibit only slight reductions to efficiency largely due to losses in FF. Our data suggest that it is not merely the existence of the organic–electrode interface in OSCs that leads to photo-unstable devices, but rather the direct contact between the photo-active layer and the electrode. It is thus hinted that the presence of excitons at the organic-electrode interface may be the root cause behind contact photo-degradation. The results establish that the contact photo-degradation mechanism is widespread, applying to both vacuum-deposited and solution-coated OSCs. Minimizing the short-term contact photo-degradation allows for the opportunity to address other degradation mechanisms that may occur in a larger timescale.

55

Acknowledgments We would like to acknowledge Dr. Matthew Heuft and Dr. Richard Klenkler from the Xerox Research Centre of Canada (XRCC) for helpful discussions. ClInPc used in this study was generously provided by XRCC. Financial support to this work from the Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged. GW acknowledges financial support through NSERC Alexander Graham Bell Canada Graduate Scholarship, Ontario Graduate Scholarship, and WIN Nanofellowship. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.orgel.2013.10.017. References [1] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Solar cell efficiency tables (version 41), Progr. Photovoltaics: Res. Appl. 21 (2013) 12–20. [2] Heliatek, Heliatek consolidates its technology leadership by establishing a new world record for organic solar technology with a cell efficiency of 12%. (January 16, 2013, accessed: June 16.06.2013). [3] J. You, L. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, K. Emery, C.-C. Chen, J. Gao, G. Li, Y. Yang, A polymer tandem solar cell with 10.6% power conversion efficiency, Nat. Commun. 4 (2013) 1446. [4] M. Manceau, E. Bundgaard, J.E. Carlé, O. Hagemann, M. Helgesen, R. Søndergaard, M. Jørgensen, F.C. Krebs, Photochemical stability of pconjugated polymers for polymer solar cells: a rule of thumb, J. Mater. Chem. 21 (2011) 4132–4141. [5] M. Manceau, S. Chambon, A. Rivaton, J.-L. Gardette, S. Guillerez, N. Lemaître, Effects of long-term UV–visible light irradiation in the absence of oxygen on P3HT and P3HT: PCBM blend, Sol. Energy Mater. Sol. Cells 94 (2010) 1572–1577. [6] E. Palacios-Lidon, E. Escasain, E. Lopez-Elvira, A.M. Baro, J. Colchero, Photobleaching of MEH-PPV thin films: correlation between optical properties and the nanoscale surface photovoltage, Sol. Energy Mater. Sol. Cells 117 (2013) 15–21. [7] A. Rivaton, S. Chambon, M. Manceau, J.-L. Gardette, N. Lemaître, S. Guillerez, Light-induced degradation of the active layer of polymerbased solar cells, Polym. Degrad. Stab. 95 (2010) 278–284. [8] N. Grossiord, J.M. Kroon, R. Andriessen, P.W.M. Blom, Degradation mechanisms in organic photovoltaic devices, Org. Electron. 13 (2011) 432–456. [9] M. Hermenau, M. Riede, K. Leo, S.A. Gevorgyan, F.C. Krebs, K. Norrman, Water and oxygen induced degradation of small molecule organic solar cells, Sol. Energy Mater. Sol. Cells 95 (2011) 1268– 1277. [10] M. Jorgensen, K. Norrman, F.C. Krebs, Stability/degradation of polymer solar cells, Sol. Energy Mater. Sol. Cells 92 (2008) 686–714. [11] B. Paci, A. Generosi, V.R. Albertini, P. Perfetti, R. de Bettignies, J. Leroy, M. Firon, C. Sentein, Controlling photoinduced degradation in plastic photovoltaic cells: a time-resolved energy dispersive X-ray reflectometry study, Appl. Phys. Lett. 89 (2006) 043507. [12] M. Seeland, R. Rosch, H. Hoppe, Luminescence imaging of polymer solar cells: visualization of progressing degradation, J. Appl. Phys. 109 (2011) 064513. [13] F.C. Krebs, K. Norrman, Analysis of the failure mechanism for a stable organic photovoltaic during 10,000 h of testing, Prog. Photovoltaics 15 (2007) 697–712. [14] S. Bertho, G. Janssen, T.J. Cleij, B. Conings, W. Moons, A. Gadisa, J. D’Haen, E. Goovaerts, L. Lutsen, J. Manca, Effect of temperature on the morphological and photovoltaic stability of bulk heterojunction polymer: fullerene solar cells, Sol. Energy Mater. Sol. Cells 92 (2008) 753–760. [15] S. Bertho, I. Haeldermans, A. Swinnen, W. Moons, T. Martens, L. Lutsen, D. Vanderzande, J. Manca, A. Senes, A. Bonfiglio, Influence of

56

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

G. Williams, H. Aziz / Organic Electronics 15 (2014) 47–56 thermal ageing on the stability of polymer bulk heterojunction solar cells, Sol. Energy Mater. Sol. Cells 91 (2007) 385–389. D.E. Motaung, G.F. Malgas, C.J. Arendse, Insights into the stability and thermal degradation of P3HT: C60 blended films for solar cell applications, J. Mater. Sci. 46 (2011) 4942–4952. M.O. Reese, A.J. Morfa, M.S. White, N. Kopidakis, S.E. Shaheen, G. Rumbles, D.S. Ginley, Pathways for the degradation of organic photovoltaic P3HT: PCBM based devices, Sol. Energy Mater. Sol. Cells 92 (2008) 746–752. S. Schäfer, A. Petersen, T.A. Wagner, R. Kniprath, D. Lingenfelser, A. Zen, T. Kirchartz, B. Zimmermann, U. Würfel, X. Feng, Influence of the indium tin oxide/organic interface on open-circuit voltage, recombination, and cell degradation in organic small-molecule solar cells, Phys. Rev. B 83 (2011) 165311. E. Voroshazi, B. Verreet, T. Aernouts, P. Heremans, Long-term operational lifetime and degradation analysis of P3HT: PCBM photovoltaic cells, Sol. Energy Mater. Sol. Cells 95 (2011) 1303– 1307. G. Williams, Q. Wang, H. Aziz, The photo-stability of polymer solar cells: contact photo-degradation and the benefits of interfacial layers, Adv. Funct. Mater. 23 (2012) 2239–2247. Q. Wang, G. Williams, H. Aziz, Photo-degradation of the indium tin oxide (ITO)/organic interface in organic optoelectronic devices and a new outlook on the role of ITO surface treatments and interfacial layers in improving device stability, Org. Electron. 13 (2012) 2075– 2082. Q. Wang, G. Williams, T. Tsui, H. Aziz, Photochemical deterioration of the organic/metal contacts in organic optoelectronic devices, J. Appl. Phys. 112 (2012) 064502. R. Po, C. Carbonera, A. Bernardi, N. Camaioni, The role of buffer layers in polymer solar cells, Energy Environ. Sci. 4 (2010) 285–310. Y. Hashimoto, M. Hamagaki, Effect of oxygen plasma treatment of indium tin oxide for organic solar cell, Electr. Eng. Japan 154 (2006) 1–7. S. Sutty, G. Williams, H. Aziz, Role of the donor material and the donor–acceptor mixing ratio in increasing the efficiency of Schottky junction organic solar cells, Org. Electron. 14 (2013) 2392–2400. H. Gommans, B. Verreet, B.P. Rand, R. Muller, J. Poortmans, P. Heremans, J. Genoe, On the role of bathocuproine in organic photovoltaic cells, Adv. Funct. Mater. 18 (2008) 3686–3691. M. Vogel, S. Doka, C. Breyer, M.C. Lux-Steiner, K. Fostiropoulos, On the function of a bathocuproine buffer layer in organic photovoltaic cells, Appl. Phys. Lett. 89 (2006) 163501. D.M. Han, J.H. Lee, K.H. Jeong, J.G. Lee, Effects of substrate bias power on the surface of ITO electrodes during O-2/CF4 plasma treatment and the resulting performance of organic light-emitting diodes, Met. Mater. Int. 16 (2010) 627–632. J. Kim, N.H. Kim, H. Kim, D. Jung, H. Chae, Simultaneous plasma and thermal treatments of ITO surfaces for organic solar cells, J. Nanosci. Nanotechnol. 11 (2011) 6490–6493. A.K.K. Kyaw, X.W. Sun, C.Y. Jiang, G.Q. Lo, D.W. Zhao, D.L. Kwong, An inverted organic solar cell employing a sol–gel derived ZnO electron selective layer and thermal evaporated MoO3 hole selective layer, Appl. Phys. Lett. 93 (2008) 221107. C. Tao, S.P. Ruan, X.D. Zhang, G.H. Xie, L. Shen, X.Z. Kong, W. Dong, C.X. Liu, W.Y. Chen, Performance improvement of inverted polymer solar cells with different top electrodes by introducing a MoO(3) buffer layer, Appl. Phys. Lett. 93 (2008) 193307. V. Shrotriya, G. Li, Y. Yao, C.-W. Chu, Y. Yang, Transition metal oxides as the buffer layer for polymer photovoltaic cells, Appl. Phys. Lett. 88 (2006). 073508–073508–073503.

[33] M. Zhang, H. Ding, Y. Gao, C. Tang, Organic Schottky barrier photovoltaic cells based on MoOx C60, Appl. Phys. Lett. 96 (2010) 183301. [34] S. Shaheen, G. Jabbour, M. Morrell, Y. Kawabe, B. Kippelen, N. Peyghambarian, M.-F. Nabor, R. Schlaf, E. Mash, N. Armstrong, Bright blue organic light-emitting diode with improved color purity using a LiF/Al cathode, J. Appl. Phys. 84 (1998) 2324–2327. [35] K.L. Mutolo, E.I. Mayo, B.P. Rand, S.R. Forrest, M.E. Thompson, Enhanced open-circuit voltage in subphthalocyanine/C60 organic photovoltaic cells, J. Am. Chem. Soc. 128 (2006) 8108–8109. [36] Z. Wu, H. Yang, Y. Duan, W. Xie, S. Liu, Y. Zhao, Improved efficiency of organic light-emitting devices employing bathocuproine doped in the electron-transporting layer, Semicond. Sci. Technol. 18 (2003) L49. [37] N.M. Bamsey, A.P. Yuen, A.M. Hor, R. Klenkler, J.S. Preston, R.O. Loutfy, Heteromorphic chloroindium phthalocyanine films for improved photovoltaic performance, Sol. Energy Mater. Sol. Cells 95 (2011) 2937–2940. [38] R. Steim, F.R. Kogler, C.J. Brabec, Interface materials for organic solar cells, J. Mater. Chem. 20 (2010) 2499–2512. [39] H. Ishii, K. Sugiyama, E. Ito, K. Seki, Energy level alignment and interfacial electronic structures at organic/metal and organic/ organic interfaces, Adv. Mater. 11 (1999) 605–625. [40] C.J. Brabec, S.E. Shaheen, C. Winder, N.S. Sariciftci, P. Denk, Effect of LiF/metal electrodes on the performance of plastic solar cells, Appl. Phys. Lett. 80 (2002) 1288–1290. [41] J. Huang, J. Yu, H. Lin, Y. Jiang, Detailed analysis of bathocuproine layer for organic solar cells based on copper phthalocyanine and C60, J. Appl. Phys. 105 (2009) 073105. [42] Q. Song, F. Li, H. Yang, H. Wu, X. Wang, W. Zhou, J. Zhao, X. Ding, C. Huang, X. Hou, Small-molecule organic solar cells with improved stability, Chem. Phys. Lett. 416 (2005) 42–46. [43] E.L. Ratcliff, A. Garcia, S.A. Paniagua, S.R. Cowan, A.J. Giordano, D.S. Ginley, S.R. Marder, J.J. Berry, D.C. Olson, Investigating the influence of interfacial contact properties on open circuit voltages in organic photovoltaic performance: work function versus selectivity, Adv. Energy Mater. 3 (2013) 647–656. [44] K. Vandewal, K. Tvingstedt, A. Gadisa, O. Inganäs, J.V. Manca, On the origin of the open-circuit voltage of polymer–fullerene solar cells, Nat. Mater. 8 (2009) 904–909. [45] M. Zhang, H. Wang, H. Tian, Y. Geng, C. Tang, Bulk heterojunction photovoltaic cells with low donor concentration, Adv. Mater. 23 (2011) 4960–4964. [46] B. Yang, F. Guo, Y. Yuan, Z. Xiao, Y. Lu, Q. Dong, J. Huang, Solutionprocessed fullerene-based organic Schottky junction devices for large-open-circuit-voltage organic solar cells, Adv. Mater. 25 (2012) 572–577. [47] M.O. Reese, M.S. White, G. Rumbles, D.S. Ginley, S.E. Shaheen, Optimal negative electrodes for poly(3-hexylthiophene): 6,6-phenyl C61-butyric acid methyl ester bulk heterojunction photovoltaic devices, Appl. Phys. Lett. 92 (2008) 053307. [48] T. Xiao, F. Fungura, M. Cai, J.W. Anderegg, J. Shinar, R. Shinar, Improved efficiency and stability of inverted polymer solar cells with a solution-processed BPhen interlayer and polystyrene beads, Org. Electron. 14 (2013) 2555–2563. [49] K. Sugiyama, H. Ishii, Y. Ouchi, K. Seki, Dependence of indium–tin– oxide work function on surface cleaning method as studied by ultraviolet and x-ray photoemission spectroscopies, J. Appl. Phys. 87 (2000) 295–298.