Solution-Processed Nanocomposites Containing Molybdenum Oxide and Gold Nanoparticles as Anode Buffer Layers in Plasmonic-Enhanced Organic Photovoltaic Devices

Solution-Processed Nanocomposites Containing Molybdenum Oxide

and Gold Nanoparticles as Anode Bu

ffer Layers in

Plasmonic-Enhanced Organic Photovoltaic Devices

Kim-Shih Tan,

Ming-Kai Chuang,

Fang-Chung Chen,*

and Chain-Shu Hsu

*

ABSTRACT: Solution-processed nanocomposites containing molybdenum oxide (MoO₃) and gold nanoparticles (Au NPs) have been used as anode buffer layers in organic photovoltaic devices (OPVs). The resulting devices exhibit a remarkable enhancement in power conversion efficiency after Au NPs were incorporated into the device. Such enhancements can be attributed to the localized surface plasmon resonance induced by the metallic nanostructures. We have also found that the rate of exciton generation and the probability of exciton dissociation were increased. Furthermore, the devices made of the MoO₃buffer layer containing Au NPs exhibited superior stability. This work opens up the possibility of fabricating OPVs with both high efficiency and a prolonged lifetime.

KEYWORDS: molybdenum oxide, gold nanoparticles, plasmonic, polymer, photovoltaic devices, stability

1. INTRODUCTION

Organic photovoltaic devices (OPVs) have been extensively studied because they have many advantageous properties, including low-cost fabrication, light weight, mechanical flexibility, and an amazing short period of energy pay-back time, compared to photovoltaic cells based on inorganic semiconductors.1−4Over the past few years, the efficiencies of OPVs have been improved substantially to over 10%.4 Nevertheless, power conversion efficiency (PCE) of these devices still needs to be improved to compete with inorganic photovoltaic devices. Additionally, inferior operating stability is also a problem requiring emphatic concern.

Organic semiconductors usually exhibit low charge-carrier mobility and short exciton diffusion length. As a result, there exists a trade-off between light absorption and charge-collection efficiency in OPVs.5,6The thickness of the photoactive layer is limited largely to∼100 nm for effective charge collection and reduced charge-recombination loss. To increase the absorption efficiency using thin films, many light-trapping approaches have been proposed.5 For example, the introduction of a surface plasmon effect induced by metallic nanostructures7 and the

incorporation of optical spacers for redistribution of the electrical field in the device8,9 have been demonstrated successfully. Between these approaches, light trapping on the basis of localized surface plasmon resonance (LSPR) effects has received much attention as a means of improving the absorption efficiency.7,10−17 Specifically, metal nanoparticles (NPs) are often introduced into the devices for triggering LSPRs. Their unique optical properties, including strong light scattering and localfield enhancement, improve the absorption efficiency of OPVs substantially.

One simple method for incorporating the metal NPs into the devices is through doping soluble NPs in the anode buffer layers, such as poly(3,4-ethylenedioxythiophene):poly(styrene-sulfonate) (PEDOT:PSS).12,13 The use of PEDOT:PSS, however, usually suffers from the problem of long-term stability because of its acidic and hygroscopic nature.18 Alternatively, transition-metal oxides, such as molybdenum oxide (MoO3)

Received: August 13, 2013

Accepted: November 15, 2013

Published: November 15, 2013

3 ffer

layers of OPVs. The embedded Au NPs induced LSPR, which enhanced the local electromagnetic field of the active layer, thereby improving the device performance. The AuNP/MoO3

nanocomposite layer is one of the most promising candidates for increasing the light-harvesting ability without significantly sacrificing the charge transport and/or collection efficiencies in OPVs. More importantly, the devices made of the MoO3buffer

layer containing Au NPs exhibited superior stability. This work opens up the possibility of fabricating OPVs with both high efficiency and prolonged lifetime.

2. MATERIALS AND METHODS

2.1. Preparation of MoO3and AuNP/MoO3 Thin Films. To

prepare the MoO3 solutions, ammonium heptamolybdate was

dissolved in deionized water to form a 0.2 wt % solution.22The pH of the solution was adjusted to between 1 and 1.5. Then, the precursor

solution was heated at 80 °C for 2 h in air. Ammonium

heptamolybdate was decomposed to three components: MoO3,

NH3, and H2O.22 Gaseous NH3 was easily evaporated under the

experimental conditions, and MoO3was left as the only solute in the

solution. The Au NPs were synthesized following a procedure developed in a previous work.23 Figure 1b displays the scanning

electron microscopy (SEM) image of the octahedral Au NPs. We observed that the average particle size of the Au NPs was approximately 40 nm.17For the preparation of the composite buffer layers, the Au NPs were mixed with the as-prepared MoO3solution

(Figure 2a). The anode buffer layers were prepared by spin coating the aqueous mixtures containing both Au NPs and MoO3precursors onto

the UV−O3-pretreated indium tin oxide (ITO)-coated glass substrates.

The samples were annealed at 150°C for 10 min in air. Figure 2b

reveals the transmittance spectra of the MoO3solution incorporating

Au NPs with different concentrations. We can see that the neat MoO3

film exhibited high transmittance in the visible spectral region. Because the plasmon band of Au NPs was around 530 nm, the transmittance decreased upon the addition of Au NPs.

2.2. Fabrication of Photovoltaic Devices. The devices were fabricated on ITO-coated glass substrates. The device structure in this work is illustrated in Figure 1a. The ITO substrates were treated with UV-ozone prior to use. The Au NP/MoO3films were prepared using

spin coating as described above. The photoactive layer comprising regioregular poly(3-hexylthiophene) (P3HT) and 1-(3-methoxycarbonyl)propyl-1-pheny[6,6]methanofullerene (PCBM) at a weight ratio of 1:1 was spin-coated from a solution of 1,2-dicholorobenzene on the anode buffer layers. After solvent annealing,24,25the driedfilm was further annealed at 110 °C for 10 min. The resulting thickness of the active layer was∼180 nm. Finally, the metals Ca (30 nm) and Al (100 nm) were thermally evaporated sequentially to form the cathodes. The device active area was 0.1 cm2_.

The thickness of the MoO3buffer layers was ca. 10 nm.

2.3. Characterization. For the device characteristics, a Keithley 2400 source measure unit was used for recording the current density− voltage (J−V) curves. The electrical properties of the devices under illumination conditions were measured using a 150 W Thermal Oriel solar simulator equipped with standard AM 1.5G filters. The light intensity of the simulator was further calibrated using a Si photodetector containing a KG5 filter.26 The incident photon-to-electron conversion efficiency (IPCE) curves were obtained from a measurement system (Enli Technology) composed of a quartz-tungsten-halogen lamp, a monochromator, an optical chopper, a lock-in amplifier, and a calibrated silicon-based diode.

3. RESULTS AND DISCUSSION

Figure 3 displays the J−V characteristics of the devices fabricated with various amounts of Au NPs; the measurements were performed under standard illumination conditions (AM 1.5G, 100 mW cm−2). The controlled device, in which a layer of pristine MoO₃ was used as the anode buffer, exhibited an open-circuit voltage (Voc) of 0.60 V, a short-circuit current (Jsc)

of 10.2 mA cm−2, and afill factor (FF) of 0.60. The calculated PCE was 3.68%. The results were consistent with those

Figure 1.(a) Device structure of the OPVs in this study. (b) SEM image of the Au NPs. The average size was 40± 5 nm. (c) UV−vis absorption spectrum of the Au NPs suspended in water. The inset shows a photograph of the Au NP solution.

Figure 2.(a) Photograph of the MoO3 solution containing various

volume ratios of Au NPs for preparing the anode buffer layers. (b) Transmittance spectra of the MoO3thinfilms deposited on ITO glass

substrates. Different concentrations of Au NPs were embedded in the thinfilms.

reported previously.22After the incorporation of Au NPs to the MoO3 buffer layer, the photocurrent and FF were improved.

For the optimized device, the Jscand FF values were improved

to 10.9 mA cm−2 and 0.65, respectively. The values of Voc

remained unchanged (0.60 V), indicating that the anode interface was not significantly affected by the Au NPs. As a result, the PCE was increased to 4.20%. We also extracted the device series resistance (Rs) from the J−V curves measured in

the dark (Figure S1). The values of Rsincreased from 1.7 to 1.9

Ω cm2_{after the addition of 10% Au NPs, suggesting that the}

device improvement was not due to the reduction in device resistance. Rather, the advantageous LSPR effect induced by the Au NPs overwhelmed the inferior electrical properties of the devices containing Au NPs.17 We also note that a further increase in the amount of the Au NPs resulted in a decrease in the value of Jsc. We suspected that the excess Au NPs might

lead to enhanced backward scattering. Furthermore, the high concentration of NPs might also increase the resistivity of the buffer layers. From the J−V curves measured in the dark, we indeed obtained an increased Rsto 2.4Ω cm2after the addition

of 30% Au NPs. Table 1 summarizes the device performance in this work.

Figure 4a shows the IPCE spectra of the solar cells prepared with different anode buffers. We can see that the trends in the IPCE were consistent with those for the values of Jsc. The

devices containing Au NPs exhibited higher efficiencies. Furthermore, we also noted that the spectral shape changed after the use of Au NP solutions. The efficiencies increased substantially within the spectral range from 450 to 600 nm. The peak efficiency was improved to 66% at 550 nm for the device containing 10% Au NPs. Figure 4b displays the corresponding curve of the increase in IPCE (ΔIPCE) for the solar cell containing 10% Au NPs; the extinction spectrum of the Au NPs is also included in Figure 4b for comparison. The spectral regime in which the IPCE values were increased was matched with the plasmonic band of the Au NPs. The spectral

consistence suggested that the LSPR contributed to the device enhancement.

The morphology changes of the MoO3:Au NP layer were

investigated using atomic force microscopy (AFM), and the topographic images are displayed in Figure 5. The root-mean-square (rms) roughness of the MoO3film was 8.8 nm (Figure

5a). Upon the addition of Au NPs, we observed obvious changes in the surface morphology of the MoO3films. The rms

roughness was increased to 10.4, 11.9, and 12.7 nm after the incorporation of 10, 20, and 30% of Au NP solutions, respectively. The rough surface might induce trap states at the interfaces and somehow affect the charge-transport/ collection efficiencies of the OPVs, resulting in a lower photocurrent. The increased Rsvalue of the devices containing

20 and 30% Au NPs further confirmed this assumption. For the device containing 10% Au NPs, however, the drawback of the poor interfaces was overwhelmed by the higher degree of absorption upon the addition of Au NPs, thereby improving the overall device PCE.

The transmission electron microscopy (TEM) image of the MoO3containing 10% Au NP solution was obtained to study

further the morphology of the thin films (Figure S2). We observed large particles, which might be due to the formation of MoO3. However, the dark dots should be Au NPs in the

Figure 3.J−V characteristics obtained under AM 1.5G illumination

(100 mW cm−2) of the OPVs prepared with MoO3 layers

incorporating various amounts of Au NPs.

Table 1. Electrical Characteristics of Devices Incorporating Various Amounts of Au NPs

concentration of Au NPs (%) V_oc(V) J_sc(mA cm−2) FF PCE (%) R_s(Ω cm2₎a

0 0.60± 0.01 10.2± 0.17 0.60± 0.02 3.68± 0.08 1.7± 0.1 10 0.60± 0.01 10.9± 0.20 0.64± 0.02 4.20± 0.11 1.9± 0.1 20 0.60± 0.01 10.7± 0.25 0.62± 0.03 3.97± 0.12 2.1± 0.2 30 0.60± 0.01 10.6± 0.21 0.61± 0.02 3.86± 0.10 2.4± 0.1

a_{Device series resistance (R}

s) was obtained from the inverse slope of the dark J−V curve.

Figure 4.(a) IPCE spectra for the photovoltaic devices prepared with various concentrations of Au NP solutions. (b) Increase in IPCE (ΔIPCE) after the use 10% Au NPs in the anode buffer layer and the absorption spectrum of the Au NPs. The absorption spectrum was obtained using a neat MoO3thinfilm as the baseline.

nanocopposites, which did not aggregate in the thinfilms. The results suggest that the rougher surface, as revealed in the AFM images, could be due to the aggregation of MoO₃particles.

The photocurrent behavior of the plasmonic-enhanced device was further investigated to study the LSPR effects on the device performance. Through the calculation of the maximum generation rate (Gmax), we were able to evaluate

the overall light harvesting. The devices were biased swept from +1 to −10 V under 100 mW cm−2illumination (AM 1.5G). Figure 6a displays the dependence of the photocurrent (Jph) on

the effective applied voltage (Veff). The photocurrent accounts

for the dark current (JD) using the equation

27,28

such a photocurrent at high biases to be the saturation photocurrent density (Jsat), which should be theoretically

independent of the applied electricalfield and temperature.27,28 Under the saturation conditions, all of the excitons generated in the device could be dissociated and contributed to the current owing to the high electric potential, which could overcome the exciton binding energy. Therefore, we express Jsatas

=

J_sat qG_maxL ₍₃₎

where q is the electronic charge and L is the thickness of the active layer.27,28 Therefore, the maximum exciton generation rate (Gmax) can be calculated from the measured results of Jsat.

Because the value of Gmaxis relative to the maximum number of

photons absorbed in the devices, it can be considered as one key parameter for evaluating the absorption ability of typical OPVs. From Figure 6a, the values of Gmaxwere 3.82× 1027and

4.13× 1027 _m−3_s−1_{for the device prepared without and with}

Au NPs, respectively. The noticeable increase in Gmax for the

device containing the Au NPs suggests an enhanced light-harvesting ability of the plasmonic device.

Furthermore, the exciton dissociation probabilities [P(E,T)] were also affected by the LSPR.13

For OPVs, only a portion of the exciton can be dissociated into free carriers under normal operation conditions. Therefore, Jph can be expressed as the

following equation27,28 =

J_ph qG_max( ,E T L) ₍₄₎

Assuming the dissociation probability is 100% under saturation conditions, the value of P(E,T) under various bias conditions can be calculated from the plot of the normalized photocurrent density (J_ph/J_sat). Figure 6b displays the normalized photo-current density. Under the short-circuit conditions (Va= 0 V),

the result indicates that the probability of P(E,T) was increased from 78.7 to 84.9% for the device incorporating 10% AuNPs. We infer that excitation of the LSPR could also facilitate exciton dissociation, increasing the photocurrent of the plasmonic devices.

From the device results, we also noted that FF values were increased after the use of Au NPs. In fact, a higher exciton dissociation probability decreases the recombination rate.13 Furthermore, the device series resistance was increased slightly, suggesting that the improvement in the FF values was not due to the reduction of resistance. Therefore, we suggest that the LSPR effects increased the exciton dissociation probability, thereby increasing the FF values.13

The results of steady-state photoluminescence (PL) measurements of the P3HT:PCBM thin films prepared on various buffer layers also helped to understand the LSPR effect on the exciton generation behavior; Figure 7 shows the PL results. The peak intensity of the thinfilm containing Au NPs apparently became higher by approximately 60%. As was noted previously, because of the good overlap between the resonance

Figure 5. AFM images (10 × 10 μm2_{) of the MoO}

3 buffer layers

prepared with different concentrations of Au NP solutions: (a) 0, (b) 10, (c) 20, and (d) 30% Au NP solution doping.

Figure 6. (a) Plots of photocurrent density as a function of the effective bias for devices prepared with neat MoO3 and MoO3

containing 10% Au NP solution. (b) Exciton dissociation probability [P(E,T)] plotted with respect to Vefffor the device prepared with and

peak of the Au NPs and the absorption spectrum of P3HT, we attributed the increased PL intensity to the increased level of photon absorption. The enhanced PL intensity probably arises from the coupling between the plasmonicfield of the Au NPs and the excited state of P3HT.

The long-term stability of devices was investigated by testing the shelf lifetime. The solar cells were stored under ambient conditions in which the relative humidity during the period of measurement was controlled at ca. 60%. The test protocols could be qualified as an ISOS-D-1 experiment.29Figure 8 shows

the normalized PCEs of the devices as a function of time. It can be seen that the device made of a PEDOT:PSS buffer layer degraded much more rapidly. The PCE decreased to 50% of the initial efficiency within the first 3 days. No apparent electrical response could be measured after 5 days. In contrast, the devices fabricated with MoO3buffer layers exhibited higher

stability. The addition of Au NPs slightly affects the device stability, which is probably due to the degradation of the interface as the surface of the buffer layer becomes rougher (Figure 5).30For the device containing 10% Au NPs, however, the PCE still remained at ca. 80% of its initial value after storing the device for 7 days. The replacement of the hygroscopic PEDOT:PSS with MoO3 might reduce the diffusion rate of

humidity within the devices, thus reducing the degree of oxidation of the cathode. Furthermore, unlike the acidic PEDOT:PSS, the MoO3buffer could also avoid the corrosion

of the ITO anode.

4. CONCLUSIONS

We have developed a simple, solution-processed method for preparing oxide buffer layer containing Au NPs for OPVs. The plasmonic effects induced by the Au NPs enhanced the photocurrent and PCE of the devices. More importantly, the devices fabricated with these Au NP/MoO₃ buffer layer exhibited a longer device lifetime. We foresee that the metal/ oxide nanocomposites reported herein might be of further use in other material systems, such as low-band-gap polymers, to achieve even higher efficiencies and prolonged stability.

■

*

J−V characteristics of the devices measured in the dark, and TEM image of the nanocomposites. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competingfinancial interest.

■

We thank the National Science Council of Taiwan (grant nos. NSC 102-2221-E-009-130-MY3, NSC 102-3113-P-009-004, and NSC 101-2628-E-009-008-MY3) and the Ministry of Education of Taiwan (through the ATU program) forfinancial support. We also thank Prof. Michael H. Huang and Mrs. Ching-Wen Liao from National Tsing Hua University for providing the Au nanoparticles.

■

(1) Yip, H. L.; Jen, A. K.-Y. Energy Environ. Sci. 2012, 5, 5994−6011. (2) Servaites, J. D.; Ratner, M. A.; Marks, T. J. Energy Environ. Sci. 2011, 4, 4410−4422.

(3) Roes, A. L.; Alsema, E. A.; Blok, K.; Patel, M. K. Prog. Photovoltaics 2009, 17, 372−393.

(4) Li, G.; Zhu, R.; Yang, Y. Nat. Photonics 2012, 6, 153−161. (5) Ko, D. H.; Tumbleston, J. R.; Gadisa, A.; Aryal, M.; Liu, Y.; Lopez, R.; Samulski, E. T. J. Mater. Chem. 2011, 21, 16293−16303.

(6) Park, S. H.; Roy, A.; Beaupre, S.; Cho, S.; Coates, N.; Moon, J. S.; Moses, D.; Leclerc, M.; Lee, K.; Heeger, A. J. Nat. Photonics 2009, 3, 297−302.

(7) Gan, Q.; Bartoli, F. J.; Kafafi, Z. H. Adv. Mater. 2013, 25, 2385− 2396.

(8) Kim, J. Y.; Kim, S. H.; Lee, H.-H.; Lee, K.; Ma, W.; Gong, X.; Heeger, A. J. Adv. Mater. 2006, 18, 572−576.

(9) Chen, F. C.; Wu, J.- L.; Hung, Y. Appl. Phys. Lett. 2010, 96, 193304-1−193304-3.

(10) Wang, D. H.; Park, K. H.; Seo, J. H.; Seifter, J.; Jeon, J. H.; Kim, J. K.; Park, J. H.; Park, O. O.; Heeger, A. J. Adv. Energy Mater. 2011, 1, 766−770.

(11) Li, X.; Choy, W. C. H.; Lu, H.; Sha, W. E. I.; Ho, A. H. P. Adv. Funct. Mater. 2013, 21, 2728−2735.

(12) Chen, F. C.; Wu, J.-L.; Lee, C. L.; Hong, Y.; Kuo, C. H.; Huang, M. H. Appl. Phys. Lett. 2009, 95, 013305-1−013305-3.

(13) Wu, J. L.; Chen, F. C; Hsiao, Y. S.; Chien, F. C.; Chen, P.; Kuo, C. H.; Huang, M. H.; Hsu, C. S. ACS Nano 2011, 5, 959−967.

(14) Yang, X.; Chueh, C. C.; Li, C. Z.; Yip, H. L.; Yin, P.; Chen, H.; Chen, W. C.; Jen, A. K. Y. Adv. Energy Mater. 2013, 3, 666−673.

(15) Baek, S. W.; Noh, J.; Lee, C. H.; Kim, B.; Seo, M. K.; Lee, J. Y. Sci. Rep. 2013, 3, 1726.

(16) Hsiao, Y. S.; Charan, S.; Wu, F. Y.; Chien, F. C; Chu, C. W.; Chen, P.; Chen, F. C J. Phys. Chem. C 2012, 116, 20731−20737. Figure 7.Photoluminescent spectra of the P3HT:PCBM thinfilms

deposited on the MoO3layers prepared with and without Au NPs. A

470 nm laser was used as the excitation source.

Figure 8.Normalized PCEs of devices fabricated with PEDOT:PSS or MoO3anode buffer layers plotted as a function of time. Au NPs were

incorporated into the buffer layers. The devices, which were not encapsulated, were stored under ambient conditions.

(23) Chang, C. C.; Wu, H. L.; Kuo, C. H.; Huang, M. H. Chem. Mater. 2008, 20, 7570−7574.

(24) Li, G.; Yao, Y.; Yang, H.; Shrotriya, V.; Yang, G.; Yang, Y. Adv. Funct. Mater. 2007, 17, 1636−1644.

(25) Chen, F. C.; Ko, C. J.; Wu, J. L.; Chen, W. C. Sol. Energy Mater. Sol. Cells 2010, 94, 2426−2430.

(26) Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y. Adv. Funct. Mater. 2006, 16, 2016−2023.

(27) Mihailetchi, V. D.; Koster, L. J. A.; Hummelen, J. C.; Blom, P. W. M. Phys. Rev. Lett. 2004, 93, 216601-1−216601-4.

(28) Mihailetchi, V. D.; Xie, H. X.; de Boer, B.; Koster, L. J. A.; Blom, P. W. M. Adv. Funct. Mater. 2006, 16, 699−708.

(29) Reese, M. O.; Gevorgyan, S. A.; Jørgensen, M.; Bundgaard, E.; Kurtz, S. R.; Ginley, D. S.; Olson, D. C.; Lloyd, M. T.; Morvillo, P.; Katz, E. A.; Elschner, A.; Haillant, O.; Currier, T. R.; Shrotriya, V.; Hermenau, M.; Riede, M.; Kirov, K. R.; Trimmel, G.; Rath, T.; Inganas, O.; Zhang, F.; Andersson, M.; Tvingstedt, K.; Lira-Cantu, M.; Laird, D.; McGuiness, C.; Gowrisanker, S. (J.); Pannone, M.; Xiao, M.; Hauch, J.; Steim, R.; DeLongchamp, D. M.; R€osch, R.; Hoppe, H.; Espinosa, N.; Urbina, A.; Yaman-Uzunoglu, G.; Bonekamp, J.-B.; van Breemen, A. J. J. M.; Girotto, C.; Voroshazi, E.; Krebs, F. C. Sol. Energy Mater. Sol. Cells 2011, 95, 1253−1267.