**4.1. Device fabrication**

The MoO3/Ag or Ag electrodes were thermally evaporated on glass or PET substrates at a base pressure of 5.0 × 10−4 Pa, with an evaporation rate of 0.02 nm/s for MoO3 and 0.1 nm/s for Ag, respectively. The thicknesses and evaporation rates of MoO3, Ag and Al were estimated in situ with a calibrated quartz crystal monitor. Since the 2 nm ultrathin MoO3 is not smooth and closed, the given thickness had to be a nominal value obtained by the monitor, representing the amount of MoO3 on the sample. The sheet resistances of these MoO3/Ag or Ag electrodes were measured by using a four point probe setup system. The transmission spectra were recorded by using a spectrophotometer (Lambda950, PerkinElmer). The fabricated devices has a structure of glass (or PET)/(MoO3)/Ag/MoO3/P3HT:PCBM/Al and an active area of 12.5 mm2 . The detail fabrication process is the same as that in previous sections.

**Figure 11.** Measured (symbol) and calculated (solid line) J‐V characteristics for OSCs fabricated on the (a) Ag or (b) MoO3/Ag anodes. Inset: A photograph of MoO3 (2 and 10 nm)/Ag (9 and 11 nm) electrodes. Adapted with permission [11]. Copyright 2014, Elsevier.

#### **4.2. Results and discussion**

In details, the OSCs based on Ag thin‐film electrode are first fabricated with a structure of glass/Ag/MoO3/P3HT:PCBM/Al, and then, a MoO3 interlayer is introduced between glass and Ag electrode to further improve the device performance. As shown in **Figure 11a** and **Table 3**, the device with 11 nm Ag shows a higher PCE of 2.57% and lower *R*s of 2.0 Ω cm2 . Further, by inserting a MoO3 interlayer, a senior PCE of 2.71% is obtained by employing a MoO3 (2 nm)/Ag (9 nm) electrode as well as a relatively low *R*s of 3.0 Ω cm2 , which is compa‐ rable to that of reference OSCs based on ITO electrode. Meanwhile, an increase in Ag or MoO3 layer thickness only degrades the device performance. To understand the underlying mechanism, the optical and electrical properties as well as the surface morphology are studied as follows.

by metal clips during the J‐V test may also be a factor of device degradation. As a result, considering the comparable PCEs and better device air‐stability, the AZO cathode is a prom‐

In this section, ITO‐free OSCs based on (MoO3/)Ag thin‐film electrodes on glass or PET substrates are fabricated, and the best performance of OSCs is obtained by optimizing the thicknesses of Ag film and MoO3 interlayer. And the underlying mechanism, especially the Ag

The MoO3/Ag or Ag electrodes were thermally evaporated on glass or PET substrates at a base pressure of 5.0 × 10−4 Pa, with an evaporation rate of 0.02 nm/s for MoO3 and 0.1 nm/s for Ag, respectively. The thicknesses and evaporation rates of MoO3, Ag and Al were estimated in situ with a calibrated quartz crystal monitor. Since the 2 nm ultrathin MoO3 is not smooth and closed, the given thickness had to be a nominal value obtained by the monitor, representing the amount of MoO3 on the sample. The sheet resistances of these MoO3/Ag or Ag electrodes were measured by using a four point probe setup system. The transmission spectra were recorded by using a spectrophotometer (Lambda950, PerkinElmer). The fabricated devices has a structure of glass (or PET)/(MoO3)/Ag/MoO3/P3HT:PCBM/Al and an active area of 12.5

**Figure 11.** Measured (symbol) and calculated (solid line) J‐V characteristics for OSCs fabricated on the (a) Ag or (b) MoO3/Ag anodes. Inset: A photograph of MoO3 (2 and 10 nm)/Ag (9 and 11 nm) electrodes. Adapted with permission

In details, the OSCs based on Ag thin‐film electrode are first fabricated with a structure of glass/Ag/MoO3/P3HT:PCBM/Al, and then, a MoO3 interlayer is introduced between glass and

. The detail fabrication process is the same as that in previous sections.

ising alternative of ITO to fabricate the long‐lifetime IOSCs.

**4. ITO-free OSCs based on Ag thin-film electrodes**

thin‐film growth and film properties are also investigated.

**4.1. Device fabrication**

174 Nanostructured Solar Cells

[11]. Copyright 2014, Elsevier.

**4.2. Results and discussion**

mm2


**Table 3.** Photovoltaic performance parameters for OSCs fabricated on different anodes/glass substrates. Reproduced with permission [11]. Copyright 2014, Elsevier.

The thermally evaporated Ag film prefers 3D island growth, namely Volmer‐Weber growth, which starts from disconnected nuclei [21]. Thus, for the deposition of the first few nanometers of Ag, separate nuclei are formed. According to the SEM images in **Figure 12**, optical trans‐ mittance and sheet resistance in **Figure 13**, it can be seen that the percolation threshold thickness of Ag thin film in this study is about 11 nm. At this thickness, the Ag islands are closed and a continuous Ag layer is formed, while the relatively high transmittance and low sheet resistance (6.29 Ω/square) are obtained. This is in good line with the corresponding device performance. By introducing a MoO3 interlayer, as shown in **Figure 13a**, MoO3 (2 nm)/Ag (9 nm) anode not only shows similar spectral shape of the transmission curve as Ag (11 nm) electrode but also presents a higher transparency with a maximum of 74% at 361 nm. Partic‐ ularly, in the visible spectral range, the transparency of the electrode between 56 and 70% is achieved, showing the potential of this electrode. With the introduction of 2 nm thick MoO3 interlayer between the Ag layer and glass substrate, the sheet resistance of the electrode is decreased to 9.32 Ω/square. The excellent properties of MoO3 (2 nm)/Ag (9 nm) electrode in transparency and conductivity lead to the best device performance among all the ITO‐free OSCs and verify the fact that the percolation threshold of Ag has been reduced to 9 nm by introducing a 2 nm MoO3 interlayer. As we know, the thickness of Ag film is strongly related to its transmittance and conductivity, and the percolation threshold thickness determines the smallest thickness for a metal film electrode, which is the most important parameter in the 3D growth of Ag film during the thermal evaporation process. Thus, the decrease in percolation threshold thickness not only could maintain the high conductivity, but also could enhance the optical transmission of Ag film and lower the fabrication cost as well.

**Figure 12.** SEM images of (a) Ag (9 nm), (b) Ag (11 nm), (c) MoO3 (2 nm)/Ag (9 nm) and (d) MoO3 (10 nm)/Ag (9 nm) electrodes deposited on glass substrates. The white scale bar represents 100 nm. Reproduced with permission [11]. Copyright 2014, Elsevier.

**Figure 13.** (a) Transmittance spectra with corresponding sheet resistances and (b) J‐V characteristics for ITO‐free OSCs fabricated on PET or glass substrates. (a) Adapted with permission [11]. Copyright 2014, Elsevier. (b)Reproduced with permission [26]. Copyright 2015, IEEE.

In our opinion, the introduction of the MoO3 interlayer can effectively improve the wetting of Ag on the substrate and reduce the percolation threshold of Ag. However, the mechanism of the smoothening effect of the MoO3 layer still remains to be determined. Here, MoO3 works as a surfactant to modify the surface of Ag film. When the thickness of MoO3 is 2 nm, the unclosed layer may create preferred nucleation sites on the glass substrate to enhance the lateral growth of Ag film. Similar results are also found in the recent report [23]. However, with an increase in MoO3 thickness to 10 nm, things become different. Since the surface energy of MoO3 (*γ* = 0.06 J m−2) is much less than that of Ag (*γ* = 1.25 J m−2) [24, 25], the Ag‐Ag interactions are stronger than the Ag‐substrate interactions, which weakens the surface‐modifying effect of the MoO3 layer. Thus, the effect of a thick MoO3 interlayer (here 10 nm) on improving the wetting of Ag on the substrate is inferior to that of a thin MoO3 interlayer (2 nm).

**Figure 14.** Normalized photovoltaic performance parameters of flexible ITO‐free OSCs as a function of the number of (a) outer or (b) inner bending cycles. Adapted with permission [26]. Copyright 2015, IEEE.

Furthermore, flexible devices using our optimized MoO3 (2 nm)/Ag (9 nm) anode are fabricated on PET substrates with P3HT:PCBM films as the active layer. A PCE of 2.50% is achieved for such flexible ITO‐free device (**Table 3**), which is comparable to the PCE (2.71%) of the glass/MoO3/Ag‐based devices and the PCE (2.85%) of glass/ITO‐based OSCs. Simultaneously, the corresponding flexible ITO‐free OSCs based on MoO3 (2 nm)/Ag (9 nm) anode show good mechanical flexibility. As shown in **Figure 14**, about 10% degradation in PCE is observed after 500 inner bending cycles with a bending radius of 1.5 cm, whereas a 5% decrease in PCE is observed after 500 outer bending cycles. It shows the huge potential of our flexible electrodes, and it may be instructive for further research on flexible electrodes and roll‐to‐roll mass production of OSCs.
