Photophysical and photovoltaic properties of a polymer–fullerene system containing CdSe nanoparticles

Synthetic Metals 164 (2013) 69–77

Contents lists available at SciVerse ScienceDirect

Synthetic Metals journal homepage: www.elsevier.com/locate/synmet

Photophysical and photovoltaic properties of a polymer–fullerene system containing CdSe nanoparticles João Paulo de Carvalho Alves, Jilian Nei de Freitas, Teresa Dib Zambon Atvars, Ana Flávia Nogueira ∗ Institute of Chemistry, University of Campinas – UNICAMP, P.O. Box 6154, 13083-970 Campinas, SP, Brazil

a r t i c l e

i n f o

Article history: Received 2 July 2012 Received in revised form 5 November 2012 Accepted 15 December 2012 Available online 31 January 2013 Keywords: Conducting polymers CdSe nanoparticles Hybrid solar cells Photophysical studies

a b s t r a c t In this report, the photophysical studies and the photovoltaic performance of the P3HT:PCBM system containing CdSe nanoparticles were investigated. The addition of CdSe into P3HT:PCBM system promoted a decrease in the photocurrent and in the efficiency of the solar cells. The reduction in photocurrent and efficiency observed after addition of even small portions of CdSe can be associated to a strong interaction between polymer and CdSe nanoparticles, as evidenced by absorption and emission measurements. To investigate the possible contribution of a morphological effect induced by CdSe in P3HT:PCBM film, atomic force microscopy images were also obtained. To get a better understanding of how this ternary system works, comparison was made with another system: PCBM:CdSe and poly(9,9-n-dihexyl2,7-fluorenylenevinylene-alt-2,5-thienylenevinylene (PFT), as reported by de Freitas et al. [Journal of Material Chemistry 20 (2010) 4845]. Our results indicate that the chemical structure of the chosen polymer is an important issue in ternary systems and, and that P3HT, a polymer whose chain contains a large concentration of thiophene units, favors polymer–nanoparticle aggregation. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Organic solar cells are a promising alternative for renewable energy resources [1–3]. To our best knowledge, the highest efficiency achieved for an organic solar cells is 10.1% by Mitsubishi Chemical [4,5]. Several of these devices are assembled by mixing conducting polymers and fullerene derivatives. The conducting polymer acts as light absorber, electron donor and hole transporter while the fullerene acts as an electron acceptor and transporter [6,7]. The mechanical flexibility, large area and low-cost production of devices based on conducting polymers can enable large-scale implementation [8,9]. However, organic solar cells have lower charge mobility and narrower spectral range absorption of solar energy compared to inorganic-based photovoltaic devices [10]. In this scenario, hybrid solar cells based on the combination of conducting polymers and inorganic nanoparticles have been studied as promising alternatives. Among the conjugated polymers, poly(3-hexylthiophene) (P3HT) is one of the most studied materials and the incorporation of several types of inorganic particles has been tested. For example, Sun et al. [11] reported devices based on mixtures of P3HT with CdSe tetrapods with 2.8% efficiency. Zhou et al. [12] removed the capping ligand treating the HDA–CdSe

∗ Corresponding author. Tel.: +55 19 3521 3029; fax: +55 19 3521 3023. E-mail address: anafl[email protected] (A.F. Nogueira). 0379-6779/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.synthmet.2012.12.014

QD with hexanoic acid. This treatment was beneficial resulting in devices with power conversion efficiencies of up to 2%. More recently, Dowland et al. [13] reported a successful alternative to eliminate the passivation layer by in situ formation of CdS network in P3HT films. The authors have obtained devices with an efficiency of 2.17%. The best results were reported by Dayal et al. in a hybrid devices made of a bulk heterojunction containing CdSe tetrapods and poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta[2,1-b;3,4b ]dithiophene)-alt-4,7-(2,1,3-benzothiadiazole)] (PCPDTBT) with an efficiency of 3.2% [14]. Despite the partially successful application of inorganic nanoparticles in polymer solar cells, there are only a few papers investigating the incorporation of inorganic nanoparticles into polymer–fullerene systems. Kim and Carrol [15] observed an increase of 50–70% in efficiency in poly(3-octylthiophene) (P3OT):buckminsterfullerene C60 bulk heterojunction photovoltaic devices after addition of Ag and Au nanoparticles. Topp et al. [16] reported that addition of Au nanoparticles into the active layer of P3HT:[6,6]-phenyl-C61 butyric acid methyl ester (PCBM) solar cells promoted a decrease in the efficiency of these devices assigned to poor hole mobility because of the disordered polymer phase. Wang et al. [17] reported an increase in the photocurrent (Jsc ) and incident photon-to-current efficiency (IPCE) when large Au nanoparticles (70 nm) are introduced in P3HT:PCBM heterojunction. The improvement was attributed to enhanced light absorption due to light scattering of the large Au nanoparticles in the active layer. The introduction of metallic nanoparticles in ternary system,

70

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

Fig. 1. Structures of P3HT, PFT and PCBM and HRTEM images of CdSe nanoparticles (4 nm of preferential diameter) used in this work.

especially Au nanoparticles, is followed by contradictory results. Wang’s report raised the important observation that nanoparticle size is a crucial parameter in such complex systems. Fu et al. investigated hybrid solar cells based on P3HT:PCBM:CdSe system in an inverted cell configuration. The incorporation of 10 wt.% pyridinecapped CdSe nanodots (ratio to P3HT or PCBM) into P3HT:PCBM system increased the power conversion efficiency from 2.06 to 3.05% in devices with high stability [18]. The work reported here is an extension of that carried out by de Freitas et al. [19] which investigated hybrid solar cells based on a ternary system composed of poly(9,9-n-dihexyl-2,7fluorenylenevinylene-alt-2,5-thienylenevinylene) (PFT), PCBM and CdSe nanoparticles. They observed that the best results of photocurrent and efficiency were obtained for the system in which the amount of polymer was kept at 20 wt.% with 1:1 wt.% PCBM:CdSe ratio. The improved device performance was associated with increased absorption of light after the addition of nanoparticles and with the formation of a more suitable nanomorphology of the active layer, as demonstrated by AFM images. In the previous work no comparison with other studies could be made because of the particular properties of the conjugated polymer PFT. In this work, the effects of the addition of CdSe nanoparticles into the well-known and well-established P3HT:PCBM system are more systematically investigated. Fig. 1 shows the structure of the materials detailed in this manuscript: P3HT, PFT, PCBM and CdSe nanoparticles. The results are discussed and compared to those obtained for the previously reported PFT:PCBM:CdSe system [19].

2. Experimental 2.1. Materials Regioregular P3HT and PCBM were purchased from Rieke Metals and Nano-C Inc., respectively. These materials were used without any further purification. The synthesis and characterization of CdSe

nanoparticles [19] with 4 nm of preferential diameter capped with trioctylphosphine oxide (TOPO) and the characterization of PFT [6] were published elsewhere. 2.2. Sample preparation For the photophysical studies, solutions of P3HT and PFT (0.025 mg mL−1 ) were prepared by dissolving these polymers in chlorobenzene under stirring for 24 h. PCBM was dissolved in chlorobenzene to form 1 mg mL−1 solution and CdSe was dispersed in chlorobenzene in the concentration of 1 mg mL−1 . For morphological studies, nanocomposites of P3HT:PCBM, P3HT:CdSe and P3HT:PCBM:CdSe were obtained by preparing chlorobenzene solutions containing a fixed amount of P3HT (8 mg mL−1 ) and 50 wt.% of PCBM and/or CdSe, where the PCBM:CdSe ratio was varied. Films of these nanocomposites were deposited by spin-coating (1500 rpm, 60 s) onto indium tin oxide (ITO) coated glass substrate. 2.3. Fabrication of photovoltaic devices Bulk-heterojunction solar cells were assembled in the following configuration: ITO/PEDOT:PSS/Active Layer/Ca/Al. Indium tin oxide (ITO) coated glass substrate was first cleaned with water and detergent, ultrasonicated in acetone and isopropanol for 15 min each and dried under a nitrogen flow. Poly(ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS, in aqueous solution, Aldrich) layer (40 nm) was spin-coated onto the ITOglass and dried at 120 ◦ C for 20 min. In a glovebox, the active layer (250 nm), consisting of P3HT:PCBM mixture (2.2% in dichlorobenzene, Aldrich) before or after the addition of CdSe nanoparticles (10, 20 and 30 wt.%). The percentage relates to the amount of CdSe, considering as the total amount all the three components (P3HT + PCBM + CdSe). Finally, it was then deposited on top by spincoating (700 rpm, 60 s) and subjected to annealing of 140 ◦ C for 10 min, followed by deposition of the Ca (20 nm) and Al electrodes

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

71

Vacuum level (0 eV) 0

Current / mA cm

-2

-3.0 eV

-3.1 eV

LUMO -3.6 eV

HOMO

-5.2 eV

P3HT

-10

-5.4 eV PFT

-5.6 eV CdSe

-4.2 eV

-6.1 eV PCBM

(4.0 nm) 0.0

0.5

Fig. 3. Energy level diagram for P3HT, PFT, PCBM and CdSe nanoparticles.

Voltage / V Fig. 2. J–V characteristics for solar cells assembled with P3HT/PCBM and different concentrations of CdSe nanoparticles (active area ∼0.1 cm2 ; irradiation ∼100 mW cm−2 ). CdSe (wt.%): (䊉) 0, () 10, (*) 20, () 30.

(100 nm) by thermal evaporation under high vacuum conditions. The active area of devices was 0.1 cm2 . 2.4. Characterization UV–vis absorption spectra of solutions and films were recorded with a HP 8452A diode array spectrophotometer. Photoluminescence (PL) spectra were obtained with an ISS photon counting spectrofluorometer. UV–vis absorption and PL spectra of P3HT or PFT solutions were obtained before and after the addition of PCBM and CdSe. atomic force microscopy (AFM) images of those spincoated films were obtained with a Agilent Pico Scan 5500 in the tapping mode. The current density versus voltage (J–V) characteristics of the solar cells were determined by a Source-Meter (Keithley, model 2400) under a Xe lamp illumination with 100 mW cm−2 irradiation intensity. 3. Results and discussions 3.1. Photovoltaic performance Fig. 2 shows the J–V characteristics for solar cells assembled with P3HT:PCBM and with different amounts of CdSe nanoparticles. Table 1 shows the photovoltaic parameters obtained for the devices. The addition of CdSe to the P3HT:PCBM system promoted a decrease in efficiency () and photocurrent (Jsc ) values. The higher the concentration of CdSe, the larger the loss of performance, the opposite of what one should expect due to the contribution of CdSe bandgap absorption (see Fig. 4a). Also, there is a decrease in the open circuit potential values (Voc ) compared with the standard device which seems to be independent of the CdSe concentration. For PFT:PCBM system, de Freitas et al. [19] observed an increase in both the Jsc and Voc after incorporation of certain amounts of CdSe. The best results of photocurrent and efficiency were obtained for the system in which the amount of polymer was kept at 20 wt.% Table 1 Photovoltaic parameters obtained for the solar cells based on P3HT/PCBM and CdSe (assembled in glove box, illuminated under ∼ 100 mW cm−2 ). CdSe (wt.%)

Jsc (mA cm−2 )

Voc (V)

FF

 (%)

0 10 20 30

10.79 5.31 2.24 1.28

0.61 0.57 0.58 0.57

0.50 0.56 0.52 0.52

3.27 1.70 0.68 0.38

with 1:1 wt.% PCBM:CdSe ratio, in other words, both PCBM and CdSe were equally important to device’s performance. The improved performance of the devices was associated with increased lightharvesting after the addition of nanoparticles and with a more suitable morphology of the active layer. Comparing the photovoltaic performance of the systems containing P3HT:PCBM:CdSe and PFT:PCBM:CdSe, it can be clearly seen that their electrical properties are remarkably different. More recently, and of particular relevance to this paper, Peterson et al. [20] also investigated the ternary system based on P3HT:PCBM:CdSe. They observed that the addition of CdSe nanoparticles in the P3HT:PCBM system promoted an increase in the photocurrent in a region corresponding to the nanoparticles absorption (560–600 nm). However, for a low ratio of CdSe to PCBM, they observed that the photocurrent was accompanied by a space charge build up that limited the performance of the device. The authors suggest that the space charge build up could be due to deep trapping states either on the nanoparticle surface or at an interface of the nanoparticle with another component of the device (P3HT, LiF, fullerene or PEDOT). In next sections, we combined photophysical and morphological data to explain why the hybrid solar cells with P3HT respond differently to those based on PFT polymer. There are several possible hypotheses for these remarkable differences. Among them we must consider changes in the energy levels, light absorption, the presence of some specific interaction between the polymer and the particles, such as the formation of a P3HT–CdSe charge-transfer complex and so on. Some of these hypotheses were analyzed here. Fig. 3 shows the energy levels for PFT, P3HT, PCBM and CdSe determined by cyclic voltammetry. The similarity between the HOMO and LUMO energies of both polymers cannot explain the observed differences of the photovoltaic performances. 3.2. Electronic absorption in solutions and films The electronic absorption spectra of P3HT and PFT in chlorobenzene solutions before and after the addition of several aliquots of PCBM and CdSe are displayed in Fig. 4. P3HT shows a well-defined absorption band with maximum around 450 nm, corresponding to ␲ → ␲* transition, while for PFT this band is observed at 460 nm. Separate additions of PCBM or CdSe nanoparticles to both polymer solutions contribute to the increase of the absorption band in the visible spectral region. The increase of the relative amount of CdSe to the P3HT and PFT solutions promotes the appearance of the absorption band around 600 nm (region of the excitonic absorption peak of CdSe nanoparticles). For P3HT, the addition of CdSe to the solution promotes a blueshift (∼5 nm) in the maximum absorption band of the polymer possibly due to a solvatochromism effect. Khan

72

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

(b)

(a) CdSe

Absorption / a.u.

Absorption / a.u.

PCBM

300

400

500

600

700

300

400

Wavelength / nm

500

600

Wavelength / nm

(c)

(d)

Absorption a.u.

Absorption / a.u.

CdSe

300

400

500

700

600

700

Wavelength / nm

300

PCBM

400

500

600

700

Wavelength / nm

Fig. 4. Electronic absorption spectra of P3HT with increasing addition of (a) CdSe and (b) PCBM; and PFT with increasing additions of (c) CdSe and (d) PCBM. P3HT concentration of 0.025 mg mL−1 ; PFT concentration of 0.025 mg mL−1 .

et al. [21] also observed a blueshift in the absorption band of P3HT and a photoluminescence quenching after CdS addition, which they attributed to a charge transfer complex formation involving P3HT and nanoparticles. It is well known that an interaction between sulfur atoms of thiophene rings in the polymer chains and the nanoparticles can occur [22]. The driving force for this complexation is the strong dipole–dipole interaction between the Cd2+ ions, on the surface of nanoparticles, and sulfur atoms of the polymer [23]. P3HT has a higher amount of sulfur atoms (from the thiophene rings) in the polymer chain available to interact more effectively with the CdSe nanoparticles than PFT, thus, complexation becomes more effective (Fig. 5).

Fig. 6 shows electronic UV–vis absorption spectra of films consisting of fixed amount of P3HT and different PCBM:CdSe ratios. These spectra are red-shifted compared to the spectra in solutions (Fig. 4), indicating a higher planar orientation of P3HT chains in the solid film [22]. The spectra of P3HT and PCBM films have bands with absorption maxima at approximately 340 nm, corresponding to ␲ → ␲* transitions of the PCBM molecule, and 500 nm for the ␲ → ␲* transitions of the P3HT chains [16,24]. The absorption band of the CdSe nanoparticles is not clearly observed due to the overlap with the absorption of P3HT in the solid state. Furthermore, two shoulders are observed at approximately 550 and 600 nm. These shoulders are attributed to the vibronic coupling in the polymer chain resulting in a higher ordering of the P3HT

Fig. 5. Cartoon describing the possibilities of interaction between the polymer and the CdSe nanoparticles, resulting from the strong dipole–dipole interaction between the Cd2+ ions and sulfur atoms.

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

Absorption

Absorption / a.u.

Table 2 Maximum of the electronic absorption band of P3HT for films with different PCBM:CdSe ratio.

PFT + PCBM

PFT + CdSe

300

73

max absorption (nm)

PCBM:CdSe

400

500

1:0 0.75:0.25 1:1 0.25:0.75 0:1

600

Wavelength / nm

488 508 512 516 520

results in a closer packing of the polymer chain and increases polarizability [21,23,26,27].

300

400

500

600

700

800

3.3. Fluorescence quenching in solutions

Wavelength / nm Fig. 6. Electronic UV–vis absorption spectra of hybrid films containing 50 wt.% of ) 1:0, () 0.75:0.25, (*) 1:1, () 0.25:0.75 P3HT and different PCBM:CdSe ratios: ( and (䊉) 0:1. The inset shows the absorption spectra of the PFT/PCBM and PFT/CdSe films.

chains, through their interchain interactions from ␲–␲ stacking in solid state [16,25]. Interestingly, no redshift was observed in the UV–vis absorption spectra of the PFT:PCBM and PFT:CdSe films ratios (inset, Fig. 6). No photoluminescence spectra were obtained in solid state since the emission is completely quenched under our experimental conditions.Table 2 shows the absorption maxima of P3HT upon addition of different PCBM:CdSe ratios. There is a redshift in the P3HT absorption with increasing CdSe concentration, indicating that the polymer chains become more planar. This planarization

Fluorescence quenching studies were carried out in solution as an attempt to get some insights about the competition of the charge transfer processes from the donor (P3HT) and the acceptors (PCBM and CdSe nanoparticles). We are assuming that the processes occurring in solution are much more efficient in the solid state because of the shorter polymer–particles distances. In solid state the fluorescence is completely quenched and experiments cannot be carried out. Moreover, it is assumed that in the solid state the quenching mechanism may have an additional non-diffusional component contributing to polymer quenching. Energy transfer or energy transport process is related to the specific interaction between two components in solution or in solid state and there are several reports describing the data treatment [28]. There is no attempt to transfer the data from solution to the solid state, but the information is crucial for a better understanding of the quenching process.

(b) PL intensity / a.u.

PL intensity / a.u.

(a)

CdSe

500

550

600

650

700

750

PCBM

500

550

Wavelength / nm

600

650

700

CdSe

550

600

Wavelength / nm

(d) PL intensity / a.u.

PL intensity / a.u.

(c)

500

750

Wavelength / nm

650

700

PCBM

500

550

600

650

700

Wavelength / nm

Fig. 7. Photoluminescence spectra for solutions of P3HT (exc = 450 nm) (a, b) and PFT (exc = 460 nm) (c, d) with increasing addition of CdSe (a, c) or PCBM (b, d). P3HT and PFT concentrations of 0.025 mg mL−1 .

74

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

Table 3 Stern–Volmer constants (Ksv ) for all systems investigated in solution.

quenchers, Ai is the integrated emission band in the presence of quenchers and [S] is quencher concentration [30].

Solution

Quencher

Ksv (L mol−1 )

P3HT P3HT PFT PFT P3HT + CdSe P3HT + PCBM

PCBM CdSe PCBM CdSe PCBM CdSe

2579 1445 2711 665 1010 1355

A0 = 1 + Ksv [S] Ai

A decrease in the fluorescence intensity of P3HT with addition of CdSe nanoparticles has been reported by Tang et al. [29], and it was ascribed to the nanocrystals acting as electron acceptor species. As the charge transfer between the P3HT and PCBM or CdSe nanoparticles is a well-known process, the Förster energy process was not considered here. Fluorescence quenching of P3HT in the presence of PCBM was reported before by Hu et al. [26] and was attributed to efficient charge transfer from P3HT to PCBM, providing a fast non-radiative decay of the excited state. The quenching of polymer photoluminescence in the presence of CdSe and/or PCBM particles was observed for both polymers, according to Fig. 7. To avoid any other additional effects, these fluorescence quenching studies were carried out initially in diluted systems and after in solid state (films). Obviously in the films all of the processes are more complicated but it is assumed that all those processes occurring in solution may also occur in films in addition to others. These data are rationalized by comparing the results with both P3HT and PFT polymers. In order to determine quantitatively the fluorescence quenching of these two polymers by separate additions of PCBM and CdSe, the Stern–Volmer constants (Ksv ) (Table 3) were calculated using Eq. (1), where A0 is the integrated emission band in the absence of

The Stern–Volmer plots A0 /Ai versus quencher concentration (CdSe or PCBM) are shown in Fig. 8 and the Ksv values found from the slopes of these curves are 2579 and 1445 Lmol−1 for PCBM and CdSe to P3HT solution, respectively. The value of Ksv for P3HT/CdSe system is higher when compared to literature [31]. These values show that PBCM quenches more efficiently than CdSe the electronically excited P3HT. For comparison, in the PFT system, the Ksv values were found to be 2711 and 665 L mol−1 , for the addition of PCBM and CdSe, respectively. Interestingly, similar values were found for the quenching efficiency of P3HT and PFT by PCBM, while the values found for polymer/CdSe systems differ considerably. Thus, individually PCBM is a better quencher for both polymers in solution than CdSe particles. The quenching efficiency of the electronic excited state of P3HT by CdSe (1445 L mol−1 ) is much higher than that of PFT (665 L mol−1 ). This difference may be explained by the nature of the nanoparticle/polymer interaction ascribed to the nanoparticlethiophene complex, as proposed in Fig. 5. The Stern–Volmer constants were also determined for solutions containing ternary mixtures of P3HT:PCBM:CdSe in order to analyze the competitive quencher behavior. Initially different amounts of CdSe were added to fixed amounts of P3HT and PCBM and subsequently PCBM was added to solutions with fixed amounts of P3HT and CdSe. The first experiment gave additional quenching by CdSe in a solution of P3HT:PCBM and the latter the additional quenching produced by PCBM in a solution of P3HT:CdSe. The value of the Ksv found by the addition of PCBM to the solution of P3HT:CdSe was 1010 L mol−1 . This value differs considerably from that found by the addition of PCBM to the solution

1,8

1,8

(b)

1,6

1,6

1,4

1,4

A0 / Ai

A0 / Ai

(a)

1,2

1,0

1,2

1,0

0,8

0,8 0

2

4

6

8 -5

10

12

0

10

-1

20

30

40

-5

[PCBM] / 10 mol L

50

-1

[CdSe] / 10 mol L 1,8

1,8

(d)

(c) 1,6

1,6

1,4

A0 / Ai

1,4

A0 / Ai

(1)

1,2

1,2

1,0

1,0

0,8

0,8 0

2

4

6

8 -5

10 -1

[PCBM] / 10 mol L

12

0

10

20

30 -5

40

50

-1

[CdSe] / 10 mol L

Fig. 8. Photoluminescence quenching for solutions of P3HT (exc = 450 nm) (a, b) and PFT (exc = 460 nm) (c, d) with increasing addition of CdSe (a, c) and PCBM (b, d).

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

75

Fig. 9. AFM images obtained in the tapping mode for films of P3HT with PCBM and/or CdSe deposited onto ITO/PEDOT:PSS substrates.

of pure P3HT (2579 L mol−1 ). By successive additions of CdSe to the P3HT:PCBM solution, the Ksv value was 1355 L mol−1 , which is similar to that obtained by the addition of CdSe to pure P3HT (1445 L mol−1 ). These results indicate that there is competition between PCBM and CdSe for the quenching of the polymer emission and that the quenching efficiency in the ternary system is lower than in the binary systems. Furthermore, the PCBM quenching efficiency is more strongly affected in the ternary system, while CdSe nanoparticles show a similar quenching efficiency in ternary or binary systems. This implies that the CdSe component of the ternary P3HT:PCBM:CdSe system is protecting the polymer excited state from the quenching by PCBM, probably due to the strong interaction with P3HT (Fig. 5). 3.4. Solid state properties and morphology To investigate the possible contribution of a morphological effect induced by the CdSe nanoparticles, atomic force microscopy (AFM) images were obtained for films of P3HT:PCBM (1:1 wt.%), P3HT:CdSe (1:1 wt.%) and P3HT:PCBM:CdSe (1:0.5:0.5 wt.%) (Fig. 9). These films were deposited by spin-coating of the solution with the components on ITO/PEDOT:PSS substrates.

AFM images show that the morphology of the P3HT:PCBM mixture has low surface roughness (1.0 nm) with a homogeneous dispersion of the PCBM domains and a sharp distribution of the domains in terms of size and shape. The morphology of the P3HT:CdSe composite changes notably when compare to the standard P3HT:PCBM, achieving the highest surface roughness (7.02 nm). As expected, the ternary system, P3HT:PCBM:CdSe, has an intermediate surface roughness of 3.67 nm. The addition of CdSe nanoparticles significantly changes the morphology in the pristine polymer or in the P3HT:PCBM mixture. The films became less homogeneous, more aggregated and we observed an increase in the size of the domains. It is well known that grain sizes and grain distributions strongly influences the performance of photovoltaic devices [32]. There are several reasons for this influence but two of them are very important. Grain sizes larger than the exciton diffusion pathway inhibit charge separation and transport, fundamental requirements for good device performance [33–35]. Under this condition the nonradiative path for exciton deactivation will be faster than the charge transfer processes and thus the absorbed energy is not converted into current. On the other hand, heterogeneous distribution of widely spaced grains inhibits the formation of percolation channels

76

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77

required for charge transport and the localized exciton recombination becomes more effective than charge transport. The study of the quenching by charge transfer process in solution showed that CdSe is a more efficient quencher than PCBM in ternary system. Therefore the excitons formed in the polymer segments around the CdSe grains will be efficiently quenched by electron transfer to the CdSe and, because of the lack of a percolation pathway for the charge transport, recombination will be the preferential process. The excitons formed in the vicinity of the PCBM grains may dissociate and be transported if they are in close connection with a percolation channel between the two electrodes. The specific interaction between the CdSe nanoparticles and P3HT as described in photophysical measurements is expected to favor the formation of aggregates. This is in agreement with the redshift observed in the absorption spectra of films (Fig. 6). The planarization effect induces the formation of polymer crystalline domains around the nanoparticles, which prevents both the charge transfer interaction and the dispersion of PCBM in the polymer matrix. This is supported by the increased surface roughness of the films. Thus, although there is a specific interaction between the CdSe nanoparticles and the polymer, they are not giving additional efficiency to the device due to the poor dispersion in the film and poor charge carrier mobility of CdSe nanoparticles, due to stabilizing shell of organic ligands (TOPO) [36–38]. In summary, our explanation for why in the ternary system with PFT, the addition of CdSe was positive and the best efficiency was achieved with the mixture containing equal amounts of PCBM and CdSe, can be a consequence of the fact that PFT is a “less coordinating” polymer which allows some PCBM molecules to be intermixed with the polymer phase. In the case of P3HT, the situation is changed to a scenario where the CdSe nanoparticles are so complexed with P3HT chains that the PCBM molecules are “isolated”, not taking part in the electronic transport. Such strong interaction between P3HT and CdSe can be reason for the space charge build up that limited the device’s performance in the work by Peterson et al. [20]. The apparent discrepancy between our results and those reported by Fu et al. [18] can be supported by the fact that the authors employed CdSe nanodots covered with pyridine instead of CdSe–TOPO. The authors only use one CdSe concentration. We suspect that morphology plays an important role here. This raises an important observation that when ternary systems are used in solar cells, direct comparison can only be made between systems with the same preparation conditions and assembly/testing protocols. Because of this argument, in the present work we compared two different polymers (P3HT and PFT) containing PCBM and CdSe–TOPO, with the same compositions and prepared using exactly the same protocol. 4. Conclusions In this work, hybrid systems based on polymer–fullerene mixtures and CdSe nanoparticles were investigated and applied in solar cells. The addition of CdSe to the P3HT:PCBM system promoted a decrease in efficiency and photocurrent of the devices. These results are opposite to those reported for PFT:PCBM systems. The difference observed for the devices with polymers P3HT and PFT was associated with different polymer–nanoparticle interactions, as evidenced by absorption and emission spectroscopy measurements and morphological characterization. The higher concentration of thiophene in the P3HT polymer contributes significantly to the strong interaction existing between P3HT and CdSe (which quenches efficiently the electronic excited state of the polymer) and to the deactivation of electron transfer process between the polymer and PCBM, as demonstrated by Stern–Volmer calculations. These results are a possible explanation for the accumulate

charges on the surface of the nanoparticle or interfaces between the nanoparticle and another constituent of the device, observed by Petterson et al. [20]. Electrons can be transferred from P3HT to the CdSe in the excited state and, since the TOPO-capped CdSe nanoparticles are not good charge carriers, electrons can be trapped on their surface, not being collected by the electrodes. From this perspective, only the P3HT segments not complexed with CdSe would be free to perform an effective electron transfer to PCBM, as the CdSe acts as a trap for most of electrons from P3HT. This is a possible explanation for the difference observed when using PFT. Besides, as PFT is less coordinating, we could reach an “ideal” morphology where both CdSe and PCBM, not aggregated, are contributing to light absorption and transport. Ternary systems applied to polymer solar cells have already been proved to be an interesting way to increase device performance. Care must be taken when analyzing such complex systems. The size of the nanoparticle was recently demonstrated to be responsible for contradictory results in the literature for P3HT:PCBM/Au nanoparticle systems. The results described here show how the choice of the conducting polymer in terms of structure is also an important issue that cannot be neglected. Acknowledgements The authors thank Fapesp (fellowships 2011/51593-6, 2009/ 15428-0, 2008/53059-4), CNPq and INEO (National Institute of Organic Electronic/CNPq/FAPESP) for financial support and fellowships; LME/LNNano/CNPEM for the HR-TEM image of CdSe nanoparticles; Luiz Carlos P. Ameida and Monica A. Cotta for AFM images; Prof. Leni C. Akcelrud for the PFT polymer; and Prof. Carol Collins for English revision. References [1] G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270 (1995) 1789. [2] W. Ma, C. Yang, X. Gong, K. Lee, A.J. Heeger, Advanced Functional Materials 15 (2005) 1617. [3] G. Dennler, M.C. Scharber, C.J. Brabec, Advanced Materials 21 (2009) 1323. [4] T. Song, S.T. Lee, B. Sun, Journal of Materials Chemistry 22 (2012) 4216. [5] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Progress in Photovoltaics: Research and Applications 20 (2012) 12. [6] J.N. de Freitas, A. Pivrikas, B.F. Nowacki, L.C. Akcelrud, N.S. Sariciftci, A.F. Nogueira, Synthetic Metals 160 (2010) 1654. [7] C.J. Brabec, N.S. Sariciftci, J.C. Hummelen, Advanced Functional Materials 11 (2001) 15. [8] S. Sun, Z. Fan, Y. Wang, J. Haliburton, Journal of Materials Science 40 (2005) 1429. [9] K.M. Coakley, M.D. McGrehee, Chemistry of Materials 16 (2004) 4533. [10] D. Majundar, S.K. Saha, Chemical Physics Letters 489 (2010) 219. [11] B. Sun, H.J. Snaith, A.S. Dhoot, S. Westenhoff, N.C. Greenhan, Journal of Applied Physics 97 (2005) 014914. [12] Y. Zhou, F.S. Riehle, Y. Yuan, H.F. Schleiermacher, M. Niggemann, G.A. Urban, M. Krüger, Applied Physics Letters 96 (2010) 013304. [13] S. Dowland, T. Lutz, A. Ward, S.P. King, A. Sudlow, M.S. Hill, K.C. Molloy, S.A. Haque, Advanced Materials 23 (2011) 2739. [14] S. Dayal, N. Kopidakis, D.C. Olson, D.S. Ginley, G. Rumbles, Nano Letters 10 (2010) 239. [15] K. Kim, D.L. Carrol, Applied Physics Letters 87 (2005) 203113. [16] K. Topp, H. Borchert, F. Johnrn, A.V. Tunc, M. Knipper, E. von Hauff, J. Parisi, K. Al-Shamery, Journal of Physical Chemistry A 114 (2010) 3981. [17] D.H. Wang, D.Y. Kim, K.W. Choi, J.H. Seo, S.H. Im, J.H. Park, O.O. Park, A.J. Heeger, Angewandte Chemie International Edition 50 (2011) 5519. [18] H. Fu, M. Choi, W. Luan, Y.S. Kim, S.T. Tu, Solid-State Electronics 69 (2012) 50. [19] J.N. de Freitas, I.R. Grova, L.C. Akcelrud, E. Arici, N.S. Sariciftci, A.F. Nogueira, Journal of Materials Chemistry 20 (2010) 4845. [20] E.D. Peterson, G.M. Smith, M. Fu, R.D. Adams, R.C. Coffin, D.L. Carroll, Applied Physics Letters 99 (2011) 073304. [21] M.T. Khan, R. Bhargav, A. Kaur, S.K. Dhawan, S. Chand, Thin Solid Fims 519 (2010) 1007. [22] D. Verma, V. Dutta, Journal of Renewable and Sustainable Energy 1 (2009) 023107. [23] M.T. Khan, A. Kaur, S.K. Dhawan, S. Chand, Journal of Applied Physics 110 (2011) 044509. [24] D. Chirvase, J. Parisi, J.C. Hummelen, V. Dyakonov, Nanotechnology 15 (2004) 1317.

J.P.d.C. Alves et al. / Synthetic Metals 164 (2013) 69–77 [25] I.J. Wang, S.C. Shiu, M.Y. Lin, J.S. Huang, Y.H. Lin, C.F. Lin, Solar Energy Materials and Solar Cells 94 (2010) 1681. [26] Z. Hu, D. Tenery, M.S. Bonner, A.J. Gesquiere, Journal of Luminescence 130 (2010) 771. [27] S. Verma, A.R. Rao, V. Dutta, Solar Energy Materials and Solar Cells 93 (2009) 1482. [28] A. Javier, C.S. Yun, J. Sorena, G.F. Strouse, Journal of Physical Chemistry B 107 (2003) 435. [29] A. Tang, F. Teng, H. Jin, Y. Gao, Y. Hou, C. Liang, Y. Wang, Materials Letters 61 (2007) 2178. [30] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd ed., Kluwer Academics & Plenum Publishers, New York, 1999.

77

[31] K. Kumari, U. Kumar, S.N. Sharma, S. Chand, R. Hakkar, V.D. Vankar, V. Kumar, Journal of Physics D: Applied Physics 41 (2008) 235409. [32] D.E. Motaung, G.F. Malgas, C.J. Arendse, S.E. Mavundla, C.J. Oliphant, D. Knoesen, Journal of Materials Science 44 (2009) 3192. [33] S. Wu, Jinhua Li, Q. Tai, F. Yan, Journal of Physical Chemistry C 114 (2010) 21873. [34] H. Hoppe, N.S. Sariciftci, Journal of Materials Chemistry 16 (2006) 45. [35] J.E. Slota, X. He, W.T.S. Huck, Nano Today 5 (2010) 231. [36] N.C. Greenham, X. Peng, A.P. Alivasatos, Physical Review B 54 (1996) 17628. [37] D. Yu, B.L. Wehrenberg, P. Jha, J. Ma, P.G. Sionnesi, Journal of Applied Physics 99 (2006) 104315. [38] D.V. Talapin, S.K. Poznyak, N.P. Gaponik, A.L. Rogach, A. Eychmüller, Physica E 14 (2002) 237.