Towards improved efficiency of bulk-heterojunction solar cells using various spinel ferrite magnetic nanoparticles

Organic Electronics 39 (2016) 118e126

Contents lists available at ScienceDirect

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

Towards improved efficiency of bulk-heterojunction solar cells using various spinel ferrite magnetic nanoparticles Alexander Kovalenko a, *, Raghvendra Singh Yadav a, Jan Pospisil a, Oldrich Zmeskal a,
a, Martin Vala a, Jaromir Havlica a, Daniela Karashanova b, Patricie Heinrichova a Martin Weiter a b

ova 118, 612 00, Brno, Czech Republic Brno University of Technology, Faculty of Chemistry, Materials Research Centre, Purkyn Institute of Optical Materials and Technologies, Bulgarian Academy of Sciences, 109, Acad. G. Bontchev Str., P.O. Box 95, 1113, Sofia, Bulgaria

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 July 2016 Received in revised form 27 September 2016 Accepted 28 September 2016

A detailed study of organic solar cells (OSC) doped with various ferromagnetic and superparamagnetic (Fe3O4, ZnFe2O4 NiFe2O4) nanoparticles (MNPs) is presented. Additionally to previously used magnetite nanoparticles, various magnetic moment spinel ferrites were applied. By impedance spectroscopy (IS) analysis it is shown how the doping with various MNPs influences solar cells' performance by the charge carrier effective lifetime extension. In this regard, we introduced a convenient illustrative method to define time constants from the impedance measurements. It is also shown that, photovoltaic performance of the solar cells directly depends on the magnetic moment and alignment of the superparamagnetic single-domain MNPs. Alignment of the MNPs within the OSCs' active layer results in MNPs dipole-dipole interaction, thus further-improves photovoltaic performance due to efficient charge collection at the short-circuit condition. OSC doping with ferromagnetic MNPs showed negative influence on the device performance, however in dark conditions, devices doped with CoFe2O4 showed higher forward current presumably due to leakage current through the large MNP aggregation or electron-polaron hopping. © 2016 Elsevier B.V. All rights reserved.

Keywords: Organic solar cells Magnetic nanoparticles External magnetostatic field Spinel ferrite Impedance spectroscopy

1. Introduction Power conversion efficiencies (PCE) of organic solar cells [1e6] seemingly approaching their top empirical formulations to the efficiency limits of 10e12% [7], reporting top PCE values over 10% for single junction polymer [8e11] and over 9% [12,13] for smallmolecular solar cells. Localized nature of the electronic states [14] and low dielectric constant (ε~3) [15e17] of organic semiconductors result in short exciton length (~10 nm) [18] and large Coulombic barrier [19] to dissociate the photo-excited charge carriers. Thus, PCE of organic solar cell strongly depends on the phase morphology in so-called bulk heterojunction (BHJ) [4e6] where two different semiconductors with the offset energy levels forming nanometric domains in the bulk, which is however, a challenging task. The behavior of the organic semiconductors can be hardly predicted in terms of forming interpenetrating phase-separation in BHJ solar cells, and most of the recent records of PCEs directly

* Corresponding author. http://dx.doi.org/10.1016/j.orgel.2016.09.033 1566-1199/© 2016 Elsevier B.V. All rights reserved.

associated with BHJ morphology improvement. Another way to improve OSC performance is to enlarge effective lifetime [20] by doping of the photoactive layer with magnetic nanoparticles, thus attain efficient photogenerated charge carriers' collection [21e23]. It has been reported, that there are two possible ways, how doping with MNP improves photovoltaic characteristics: on one hand [23] is a coercive electric field, due to dipole interactions [24] from Fe3O4 magnetic nanoparticles (MNPs) dispersed in BHJ layer show potential to strengthen the performance of BHJ polymerfullerene OSCs. Moreover, alignment of the MNPs within the BHJ layer with an external magnetic field by placing the devices onto the magnet at elevated temperature has shown a further increment of the device performance [23], due to dipole reorientation of the MNPs, thus the coercive electric field of aligned MNPs was augmented. On the other hand it is spin-orbit coupling [20] which increases the efficiency of the exciton intersystem crossing process in the device, and thus extends efficient life-time. This correlates with the previous experiment where the reverse bias was applied to the OSC to create an internal 50e70 V/mm electric field to ensure efficient charge collection at the short-circuit condition [25,26].

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

However in conventional BH OSCs, charge collection electrode layers with offset in a work-function, provide a difference of less than 2 eV producing an internal electric field in the range of 20 V/ mm for a 100 nm OSC photoactive layer, which is below the requirements for an effective charge collection [26,27]. In the present paper, the influence of various magnetic nanoparticles, on the performance of organic bulk-heterojunction solar cells is described. Ferromagnetic (CoFe2O4) and three types of different superparamagnetic nanoparticles (Fe3O4, ZnFe2O4 e high magnetization and NiFe2O4 e low magnetization) were dispersed in the blend of DPP(TBFu)2:TC60BM with 1% DPP(TBFu)2:MNPs weight ratio. Magnetic nanoparticles of spinel ferrites were used in the present paper due to the possibility of fine-tuning the magnetic properties of nanoparticles by chemical manipulations. Generally spinel ferrites are of a great interest in fundamental science, especially for addressing the fundamental relationships between magnetic properties and their crystal chemistry and structure [30]. Synthesized by co-precipitation method spinel ferrite MNPs were stabilized with oleic acid. Here, oleic acid is used as surfactant which provides a steric stabilization of the nanoparticles against the van der Waals and magnetic attraction interactions and thereby prevents agglomeration [28]. This surfactant controls the growth of nanoparticles and prevents the Ostwald ripening process. Also, previous results [29] show that oleic acid capped nanoparticles, are able to prevent water adsorption, oxidation and are capable of being dispersed stably in organic solvents or mineral oils. 2. Experimental section 2.1. Materials Oleic acid stabilized magnetic iron oxide nanoparticles dispersion in toluene (6.5 nm ± 3.0 nm Fe-Oxide-Nucleus) (Aldrich) [6,6],Thienyl C61 butyric acid methyl ester (TC60BM) (Aldrich, 99%), chloroform (Aldrich, 99.9%), 3,6-bis[5-(benzofuran-2-yl)thiophen2-yl]-2,5-bis(2-ethylhexyl)pyrrolo [3,4-c]pyrrole-1,4-dione (DPP(TBFu)2) and Al (Aldrich, 99.99%) were used as received without further purification. All the materials were weighted in ambient atmosphere and consequently, as an additional drying step, prior dissolving, materials for the active layer, were kept in vacuum chamber at 60
C overnight, and then transferred to the nitrogen atmosphere. The solution for the photoactive layers was prepared by dissolving DPP(TBFu)2:TC60BM blend (1.5:1 by weight) in chloroform with a total concentration 20 mg/ml and then placed to the ultrasonic bath for 5 min until complete DPP dissolution. Then, the solution was filtered with 0.45 mm PVDF filter into the vial with MNPs (0.25, 0.5, 1% or 2% of DPP(TBFu)2 by weight). Subsequently vials were additionally sealed with parafilm and placed in the ultrasonic bath overnight to obtain stable dispersion. All further manipulations were carried out in a glovebox under a nitrogen atmosphere unless otherwise stated.

119

temperature of 120
C for 45 min and continuous stirring for 2 h in order to complete the growth process. The obtained precipitate was washed by distilled water and ethanol to remove any impurities present in the precipitate. After washing, the material was dried at 45
C for 4 h in ambient atmosphere. The obtained sample was designated as CoFe2O4/OA. Similar above synthesis process was followed for ZnFe2O4/OA, NiFe2O4/OA sample. These obtained samples CoFe2O4/oleic acid, ZnFe2O4/OA, NiFe2O4/OA were characterized and utilized in PV solar cells. 2.3. Device fabrication Pre-patterned glass/ITO substrates were preliminary cleaned in ultrasonic bath firstly in the 5% NaOH solution at room temperature, then rinsed in water and consequently washed twice in deionized water and then cleaned in acetone to remove residual water. Finally, ultrasonic treatment in isopropanol bath was applied for 10 min. Following step was the deposition of PEDOT:PSS layer onto glass/ITO substrates. Clevios P PEDOT:PSS was deposited by means of spincoating at 5500 rpm for 60 s with an immediate annealing for approximately 5e10 min at 150
C in the ambient conditions followed by transfer to the nitrogen atmosphere and 10 min annealing at 110
C to remove a residual moisture. DPP(TBFu)2:TC60BM heterojunction layer was deposited by dynamic spin coating dropping 25 ml of the prepared solution on pre-rotated at 2500 rpm substrates for 40 s. DPP(TBFu)2 (see Figure S1), which is one of the commonly used small-molecule diketopyrrolopyrrole (DPP) derivatives, was used as an electron donor due to its high reproducibility, high thermal and long-term stability and easy processing. Consequent to the active layer deposition, devices were annealed at 110
C. Finally, 200 nm Al electrodes were deposited. After that step, to operate the solar cells in the ambient environment, devices were encapsulated with glass slides and Ossila epoxy resin and treated by UV lamp for 30 min. To align magnetic nanoparticles in the active layer readymade devices (including reference undoped devices) were placed onto the neodymium magnet at 80
C for 10 min. 2.4. Characterization techniques Thickness of all the layers measured by the Decktak XT profilometer was in a range of 100 ± 10 nm. PCE measurement were performed under standard condition by AAA certified Abet Sun solar simulator with an air mass (AM) 1.5G filter. The simulated

2.2. Spinel ferrites nanoparticles synthesis Spinel ferrite AFe2O4 (A ¼ Co, Zn, Ni) magnetic nanoparticles were synthesized by co-precipitation method in presence of oleic acid (OA). The stoichiometric amount of Co(NO3)3. 6H2O (or Ni(NO3)3. 6H2O or Zn(NO3)3. 6H2O) and Fe(NO3)3. 6H2O were taken in 1:2 M ratio with distilled water in a beaker. This solution was heated to 50
C and mixed by the magnetic stirring. A specified amount of oleic acid was added to the abovementioned mixture as a capping agent. A 200 mL of NaOH with 1 M was heated to 100
C in another beaker. The mixture was added dropwise in NaOH solution at a rate of 4 mL/s with stirring. Within few minute, a precipitate was obtained. The precipitate was kept to a reaction

Fig. 1. X-ray Diffraction patterns of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles synthesized by co-precipitation method in presence of oleic acid.

120

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

Table 1 Crystal structural parameters for CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles synthesized by co-precipitation method in presence of oleic acid. Sample

Crystallite size (nm)

Lattice Parameter (Å)

rx (g/cm3)

dA(Å)

dB(Å)

CoFe2O4 ZnFe2O4 NiFe2O4

8.73 4.12 2.51

8.36 8.45 8.71

5.33 5.31 4.70

3.62 3.66 3.77

2.96 2.99 3.08

Where: rxeX-ray density, dAehopping length for tetrahedral site, dBehopping length for octahedral site.

light intensity was adjusted to 1000 Wm2 by using a NRELcalibrated Si solar cell. Statistical results are based on 24 electrodes (4 devices of 6 electrodes for each measurement) for each type of device. During the measurements a shadow mask (area 0.04675 cm2) was used. External quantum efficiency (EQE) was obtained using a combined apparatus consisting of: Keithley 6478 picoampermeter, LSH502 LOT Oriel xenon lamp, MSH101 LOT ORIEL monochromator, SiQE120 RaRe Solutions photometric head connected to Keithley 485 picoampermeter. Optical absorption spectra of the samples were measured with Varian Cary Probe50 UVevisenear IR spectrometer, Impedance spectroscopy (IS) analysis was performed on Solartron SI 1260 Impedance/Gain-Phase Analyzer with Solartron Dielectric Interface 1296 device in ambient conditions, with no light illumination applied. TEM study was performed by High Resolution Scanning Transmission electron microscope HR STEM JEOL JEM 2100, acceleration voltage 80e200 kV, maximum resolutione0.23 nm between two points, maximum magnification 1500000 in conventional and 2000000 in scanning mode, with 5 basic regimes e bright field and dark field microscopy, diffraction from selected and nano sized area and diffraction in focused beam, equipped with CCD camera GATAN Orius 832 SC1000 and GATAN Microscopy Suit Software.

patterns of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles synthesized by co-precipitation method in presence of oleic acid. The reflection peaks were noticed to (220), (311), (222), (400), (422), (511) and (440) planes of spinel ferrites with the cubic symmetry (JCPDS file no. 22e1086 for CoFe2O4, JCPDS file no. 22e1012 for ZnFe2O4, and JCPDS file no. 3e846 for NiFe2O4 [30]). The average crystallite size was calculated from the full width at half-maximum of the (311) peak by using the Scherrer formula [31].

3. Result and discussion

where (h,k,l) are the miller indices. The lattice constant was 8.3639 Å, 8.4480 Å, 8.7143 Å for CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles, respectively, as mentioned in Table 1. X-ray density (rX-Ray) of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles has been calculated by considering that a basic unit cell of the cubic spinel structure contained 8 ions:

3.1. Structural study of the spinel ferrite MNPs The crystal structure of synthesized ferrite nanoparticles was analyzed by X-ray Diffraction. Fig. 1 represents X-ray Diffraction

kl b Cosq where k is a shape factor with a typical value of 0.9 and b is the full width at the half of the maximum intensity. The calculated values of average crystallite size was 8.73 nm, 4.12 nm, 2.51 nm for CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles, respectively, as mentioned in Table 1. The lattice constant of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles was calculated by d-spacing relation:

dhkl

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi h2 þ k2 þ l2

Fig. 2. TEM images and SAED patterns of the Fe3O4 (A), ZnFe2O4 (B), NiFe2O4 (C) and CoFe2O4 (D) magnetic nanoparticles dispersed onto the amorphous carbon coated TEM grid.

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

121

cations in spinel ferrite and crystallite size of ferrite. The hopping length between the magnetic ions (the distance between the magnetic ions) in the tetrahedral (A) site and in the octahedral [B] site for CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles are calculated from the following relations [32],

pffiffiffi 0:25a 3 pffiffiffi 0:25a 2 where, dA is hopping length for tetrahedral site and dB is hopping length for octahedral site. The dA and dB values of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles are presented in Table 1. It can be observed that as the lattice parameter increases, the hopping lengths between the magnetic ions dA and dB also increases for CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles, respectively. As it is shown in Fig. 1 XRD data show that spinel-ferrite MNPs synthesized by a co-precipitation method, have a low degree of crystallinity, especially the NiFe2O4 sample (blue line) generally, crystallinity of the samples depends on their particle size. 3.2. TEM images

Fig. 3. Hysteresis curve of CoFe2O4 (a) ZnFe2O4 (b) and NiFe2O4 (c) nanoparticles synthesized by co-precipitation method in presence of oleic acid.

8M Na a3 where M is the molecular weight of sample, Na is Avogadro's number (6.0225  1023 particles/mole), and a is lattice parameter converted into cm units. The X-ray density (rx) values of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles are mentioned in Table 1. It was found that the densification process was influenced by type of

The microstructure of the obtained MNPs studied by TEM is presented on Fig. 2. Magnetic nanoparticles of all types were dispersed onto the amorphous carbon coated TEM grid to obtain a picture of individual nanoparticles. As it is seen from the TEM analysis NiFe2O4 despite of the smallest particle size is not completely dispersed and tends to create an agglomeration. Commercially available Fe3O4 was found to be the crystalline as it was measured by selected area electron diffraction (SAED) pattern e bright spots are visible, however synthesized by co-precipitation method were polynanocrystalline e spots and diffuse rings are observed. Similarly to the results obtained by the X-ray diffraction, SAED pattern indicated NiFe2O4 to be the most amorphous out of all the samples. 3.3. Magnetic properties of the MNPs The magnetic properties of CoFe2O4, ZnFe2O4 and NiFe2O4 nanoparticles were studied by vibrating sample magnetometer (VSM) at room temperature with applied magnetic field of 15 kOe.

Fig. 4. TEM images: Magnetic Fe3O4 nanoparticles aligned by an external magnetic field (A). Undoped DPP(TBFu)2:TC60BM active layer (B). Active layer doped with Fe3O4 (C), ZnFe2O4 (D) NiFe2O4 (E) and CoFe2O4 (F) MNPs. All the images are of aligned layers.

122

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

organic layer. As it is seen in Fig. 5 doping of the DPP(TBFu)2:TC60BM active layer do not significantly affects absorption spectra, due to the fact, that the concentration of the MNPs is relatively small (1% by weight). 3.5. Photovoltaic performance

Fig. 5. Absorption spectra of the variously doped (1% by weight) DPP(TBFu)2:TC60BM layers.

Magnetic hysteresis loop of CoFe2O4 nanoparticles is shown in Fig. 3a. The magnetic results showed that CoFe2O4 nanoparticles exhibit ferromagnetic behavior. Distinctly ZnFe2O4 and NiFe2O4 nanoparticles show superparamagnetic behavior (Fig. 3bec). The magnetization curve of ZnFe2O4 and NiFe2O4 MNPs demonstrates zero coercivity and remanence as the applied field is cycled between 15 and 15 kOe. The magnetization of NiFe2O4 MNPs increases with applied magnetic field and it does not attain saturation in the presence of strong magnetic field even 15 kOe. Further, the magnetization curve of NiFe2O4 nanoparticles exhibits no remanent magnetization and coercivity, implies that NiFe2O4 nanoparticles are superparamagnetic at room temperature. 3.4. Active (DPP(TBFu)2:TC60BM) layer doping with MNPs Active layer doped with MNPs was investigated using TEM microscopy to prove the presence of the nanoparticles in the BHJ layer and to study the quality of the MNPs dispersion. As it is depicted in Fig. 2, undoped organic active layer (Fig. 2b) is clearly distinct from the doped ones (Fig. 4c-4f), MNPs create relatively homogeneous dispersion (dark spots) in the active layer. However it has to be noted, that NiFe2O4 (Fig. 4e) and CoFe2O4 (Fig. 4f) doped layer possessed agglomeration notably larger than the particle size. In case of CoFe2O4 it can be associated with magnetic interaction between ferromagnetic particles, thus various sized agglomeration are observed. On the other side, in case of NiFe2O4 we observed 20e50 nm interspersing in the active layer, we assume that a low crystallinity material NiFe2O4 does not completely disperse in the

Firstly, to find the optimum concentration of the MNPs in the active layer, four different MNPs' concentration (0.25, 0.5, 1 and 2%) was tested (see table in SI), similarly to previously reported optimization using Fe3O4 nanoparticles optimum doping concentration, efficient for photovoltaic properties augmentation was about 1% for all types of MNPs. Higher concentration caused decrease in solar cells performance. Also, due to presence of high amount of nanoparticles the reproducibility was very low. Analyzing performance of the fabricated OSCs under light illumination, with and without MNPs (Table 2 and Fig. 6) it is clearly notable, that doping of the photoactive BHJ layer with superparamagnetic NPs result in short-circuit current augmentation. Non-aligned as-deposited doped with Fe3O4, NiFe2O4 or ZnFe2O4 MNPs organic solar cells possess 3.5%, 5.2% and 2.1% higher short-circuit current in average, than the reference devices respectively. It is worth to be noted, that in all the above mentioned cases open-circuit voltage was slightly lower than the reference value. This can be explained, that the coercive field unequally affects molecular orbitals of the donor and acceptor [33] (see Table 3). To align MNPs within the active layers, all devices (including reference undoped ones) were placed onto neodymium magnet at elevated temperature (80
C) for 10 min. This resulted in MNPs alignment in one direction, thus coercive electric field due to dipole interaction of MNPs was increased, which lead better dissociation of the photogenerated charge carriers at the short circuit, which resulted in augmentation of the Jsc. Notably, along with Jsc increment observed further drop of Voc, in the range of ~30 mV in comparison with the reference cell. Thus, coercive field changes molecular orbitals in a direct way. Totally, after MNPs alignment OSCs doped with Fe3O4, NiFe2O4 or ZnFe2O4 possessed Jsc increment of 16.6%, 6.2% and 13.4% respectively. Doping with CoFe2O4 possessed very poor photovoltaic behavior which was worsened after alignment. It has to be noted that the annealing time was empirically chosen (no further PCE changes was observed), also no notable changes in the undoped device was observed. It has to be emphasized, that Jsc enhancement was observed in the same devices with the alignment of nanoparticles. Comparing J-V curves in the dark (Fig. 7) we can observe a characteristic curve for solar cell in semilog scale for the undoped, as well as for Fe3O4, ZnFe2O4 and NiFe2O4 doped devices with a contact barrier near the Voc. Interestingly, analyzing the leakage dark current at in backward characteristics of the J-V curves, the

Table 2 Main photovoltaic parameters of the fabricated solar cells.

Reference cell Fe3O4 Fe3O4ealigned NiFe2O4 NiFe2O4ealigned ZnFe2O4 ZnFe2O4ealigned CoFe2O4 CoFe2O4ealigned

Jsc, mA/cm2

Voc, mV

FF,%

PCE,%

Rshunt, Ohm$cm2

Rseries, Ohm$cm2

8.87 9.18 10.34 9.33 9.42 9.06 10.06 6.69 6.71

906 902 864 891 873 898 877 320 286

50 49 49 50 50 52 52 38 39

4.02 4.05 4.37 4.16 4.11 4.23 4.58 0.81 0.75

10899 8863 8113 10109 9671 10462 10604 2762 2836

15.27 13.19 12.44 14.30 12.65 12.26 10.30 13.59 12.58

Where: Jsceshort circuit current; Voceopen circuit voltage; FFefill factor; PCEepower conversion efficiency, Rshunteshunt resistance, Rserieseseries resistance from the illuminated curves. Best performing devices are stated in the table. Deviation of the photovoltaic parameters based on 24 electrodes was in a range of ±5%.

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

123

Table 3 Calculated diffusion and effective lifetime constants. Material (aligned)

tn (ms)

td (ms)

Pure CoFe2O4 NiFe2O4 ZnFe2O4 Fe3O4

0.08 0.07 4.06 13.75 86.27

1.19 1.15 1.21 1.14 1.29

Where: tneeffective lifetime constant; tdediffusion time constant.

current values observed for doped layers is significantly lower than in the case of reference device. Also it can be noted, that the diode characteristics (difference between forward and backward current) in the dark improves with the magnetic moment of the doping material. Nevertheless CoFe2O4 possessed very low photovoltaic performance under 1 sun illumination (PCE less than 1%), J-V current in the dark shows much higher current in the forward characteristics, distinctly to the other samples, with higher shunt resistance (particularly lower contact barrier is observed) [34]. One of the possible explanations of this phenomenon can be the aggregation of the CoFe2O4 nanoparticles, due to the remanent magnetization of ferromagnetic NPs. Moreover, CoFe2O4 nanoparticles had the biggest crystalline size (~8 nm), all this could create a pathway for charge transport, which is observed by a dramatic decrease of the shunt resistance in Fig. 6. Also this occurrence can be associated with electron-polaron hopping in CoFe2O4 nanoparticles [35]. Due to localized electrons of the cobalt electrical conductivity increases owing to the electron hopping in the crystal lattice between the Fe2þ and Fe3þ ions and hole hopping between the Co3þ and Co2þ ions.

Fig. 7. J-V curves in the dark of the aligned devices.

accordance with measured current. As it is depicted in Fig. 8, the highest EQE was observed for the aligned Fe3O4 and ZnFe2O4 doped samples.

3.6. Impedance spectroscopy (IS)

Therefore we assume, that however doping of the BHJ layers with CoFe2O4 nanoparticles improves characteristics of the diode in the dark, hopping between the ions can act as a trap for the photogenerated carriers. It is also has to be noted that measured external quantum efficiencies of the abovementioned devices changed in the

IS, [36e40] was used to analyze the behavior of the abovementioned devices. The dependences of an impedance magnitude and phase shift on frequency defined by ffirelation pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z ¼ jZjexpðj4Þ ¼ Z0 þ jZ 00 , and whence jZj ¼ Z 02 þ Z 00 2 4 ¼ arctanðZ 00 =Z0Þ, where Z0 and Z 00 ¼ 1uC are resistance and reactance of a parallel circuit. From the phase angle-frequency plot (Fig. 9a) is seen that at low frequencies devices have a resistive character (4 ¼ 0
) further phase angle change to 90
indicates resistor-capacitor parallel circuit (4 ¼ arctg(-uRC)), phase angle shift from 90
to þ90
at high frequencies indicates capacitorresistor-inductor in series and 4 ¼ arctg(R1(uL1uC)), however it has to be noted that the inductance of the circuit is cause by the measurement equipment and will not be further considered. Two characteristic frequencies, where 4 ¼ 0, indicate recombination or parallel (rrec) and series (rs) resistances at the equivalent circuit of the solar cell.

Fig. 6. J-V curves under 1 sun illumination.

Fig. 8. EQE spectra of the devices under study.

⃡ Co3þ Co2þ þeþ (polaron hopping) ⃡ Fe2þ Fe3þ þ e (electron hopping) [35].

124

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

Fig. 9. Impedance phase angle vs. Frequency (A) and Impedance Magnitude vs. Frequency (B) plots of the devices under study. Measurements of aligned devices are plotted by solid lines.

Aligning of the MNPs within the active layer (with a neodymium magnet), showed interesting behavior. Spinel ferrites NiFe2O4, ZnFe2O4 and Fe3O4 possessed shift of the capacitance at lower frequencies depending on their magnetization i.e. larger shift was observed for Fe3O4 doped samples. The effect is associated with the alignment of the nanoparticles inside the organic layer, thus the sum of dipole designating the coercive field increases. However, as it depicted in Fig. 9b, ferromagnetic CoFe2O4 nanoparticle doped devices show an inverse behavior. Impedance magnitude (Fig. 9b) shows the augmentation at low frequency for the devices doped with superparamagnetic nanoparticles Fe3O4, NiFe2O4 and ZnFe2O4. In the presence of magnetic field one can observe further increment of the impedance at low frequencies, which, how it will be furtherexplained, is associated with an improved recombination resistance. Interestingly, impedance magnitude varies with an alignment of the MNPs depending on the material's magnetization. CoFe2O4 doped devices have shown an inverse shift. In case of non-aligned devices, all the results except CoFe2O4 doped samples has shown reversible results, with neodymium magnet placed onto the device during the measurement, when ferromagnetic CoFe2O4 NPs due to remanent magnetization possessed same behavior when an external magnetic field was removed. In case of the aligned samples, we did not observe any notable changes in the magnetic field, we assume that all the MNPs were aligned parallel to the external magnetic field and the Lorenz force equals to zero.

Fig. 10. (A) Nyquist plot measured in dark (logarithmic scale) and (B) zoomed Nyquist plot in linear scale [37].

As it has been reported [39,41] effective lifetime of the charge carriers depends on the devices chemical capacitance and recombination resistance by the equation: trec ¼ rrec cm , [41] recombination resistance can be calculated from the Nyquist plot and in simplified case equals to the arc size [37,39,42]. From the Nyquist plots depicted in Fig. 10 it is clearly visible, that doping of the BHJ layers with MNPs increases recombination resistance in the dark [41] by several orders of magnitude, and can be further-increased by applying of the magnetic field (MNPs alignment) in case of NiFe2O4 and ZnFe2O4 doped samples or decreased in case of CoFe2O4.

3.7. Diffusion and effective lifetime constants Diffusion-recombination model, initially described by Bisquert et al. [40] was applied to evaluate diffusion and lifetime constants in the devices. Diffusion of the charge carriers result in the appearance of the straight line at high frequencies in a Z0 , Z00 plot, which simply means series component of the equivalent circuit [40]. In Fig. 11 we show a simple description of the series component explaining diffusion-recombination model: if we assume that no diffusion is observed, and geometrical capacitance Cg ≪ cm only one arc is observed at Z0 , Z00 plot. We can also assume that a plot vertically intercepts Z0 axis, as far as transport resistance (rt) and chemical capacitance (cm) are in series and parallel Cg can be neglected (at the same time an arc from series circuit rt cm will be observed at Y0, Y00 plot). However as a result of diffusion charge

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

Fig. 11. Equivalent circuit of the BHJ OSC.

carriers flow in opposite direction (red arrow) resulting in the charge accumulation and, in this case Cg is approaching cm, thus we can assume, that diffusion time in the diffusion-recombination model depends on series parameters rt and chemical capacitance cm influenced by a parallel Cg, only when cm ~ Cg (modelling is shown in SI). Whence, td ~ rtcm. To estimate diffusion time constant here, we introduce a variable ts derived from series parameters:

125

ts ¼ Z'sC's, which frequency dependence is depicted in Fig. 12a, thus it is clearly seen, that the diffusion is presented in the measured devices, and its time constant can be found from the graph in the area where it nearly frequency independent (beginning of the flat area will be characteristic frequency ud). However it is visible, that doping of the solar cells do not significantly influence the diffusion time, which is also can be estimated from Fig. 10b (the length of straight line at high frequencies is relatively similar). In similar manner we introduced the tp value, which is the product of parallel Z'p by C'p, thus at low frequencies, where the graph plotted in Fig. 12b is nearly independent on frequency gives us the effective lifetime constant tn. Calculated diffusion and effective lifetime constants are stated in Table 2. As was to be proved, lifetime of the charge carriers in BHJ OSCs can be tremendously improved by MNPs doping, providing more efficient at short-circuit conditions (thus Jsc is increased). It is worth to be noted that the effect directly depends on the magnetic moment of the nanoparticles. Measured in the dark, effective lifetime changes from 80 ms for the reference material up to 86 ms for the best performing magnetite doped solar cell, when diffusion time is not significantly affected. Interestingly, IS spectroscopy under light irradiation (see SI) possessed a decrease of the resistance, which is in agreement with previously reported results [23], and shows that with MNP doping photoconductivity increases depending on the magnetic moment of MNPs. 4. Conclusion In conclusion, the present work presents how doping of the OSC with various MNPs affects photovoltaic performance. Spinel-ferrite oleic acid capped superparamagnetic (NiFe2O4 and ZnFe2O4) and ferromagnetic (CoFe2O4) MNPs compared with previously reported magnetite MNPs. Generally, it is fair to say, that superparamagnetic MNPs incorporated into the BHJ organic layer improve performance of the solar cells regarding to their magnetic moments. Moreover, alignment of the MNPs with an external magnetic field at elevated temperature further increases photovoltaics performance of the devices predominantly by the efficient charge collection at the short-circuit condition, resulting in higher Jsc. At the same tome coercive electric field due to superparamagnetic NPs dipole interactions affects energetic orbitals of the organic materials, thus Voc reduces with the MNPs' alignment, depending on the magnetic moment of the doping material. Ferromagnetic CoFe2O4 NPs doping reduces photovoltaic characteristics, however increase current in forward characteristics as a diode in the dark. This phenomenon can be associated with an external current pathway due to MNPs aggregation caused by the remanent magnetization, large MNPs size or an electron-polaron hopping within the MNP crystal lattice. The effect of OSC doping with superparamagnetic nanoparticles can be increased by MNP alignment (in the external magnetic field). MNP alignment has shown an increment of the impedance at low frequencies at the same time capacitive character is preserved, depending on the magnetic moment of MNPs. Additionally; it should be pointed out that doping of the OSCs with MNPs significantly increase lifetime of the charge carriers depending on the magnetic properties of the MNPs, which is assumed to be the result of spin-orbit coupling. In the present paper we introduced simple and illustrative method to observe diffusion and effective lifetime constant. Acknowledgements

Fig. 12. : Introduced values ts and tp dependences on frequency.

This work was supported by Grant Agency of the Czech Republic, Czechia via project No. 15-05095S, research infrastructures

126

A. Kovalenko et al. / Organic Electronics 39 (2016) 118e126

was supported by project Ministry of Education Youth and Sports,  Czechia (MSMT) No. LO1211. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.orgel.2016.09.033. References [1] S. Günes, H. Neugebauer, N.S. Sariciftci, Conjugated polymer-based organic solar cells, Chem. Rev. 107 (4) (2007) 1324e1338. [2] C.J. Brabec, N.S. Sariciftci, F. Padinger, T. Fromherz, 2.5% efficient organic plastic solar cells, Appl. Phys. Lett. 78 (6) (2001) 841e843. [3] H. Hoppe, N.S. Sariciftci, Organic solar cells: an overview, J. Mater. Res. 19 (07) (2004) 1924e1945. [4] H. Hoppe, N.S. Sariciftci, Morphology of polymer/fullerene bulk heterojunction solar cells, J. Mater. Chem. 16 (1) (2006) 45e61. [5] A. Guerrero, S. Loser, G. Garcia-Belmonte, C.J. Bruns, J. Smith, H. Miyauchi, S.I. Stupp, J. Bisquert, T.J. Marks, Solution-processed small molecule:fullerene bulk-heterojunction solar cells: impedance spectroscopy deduced bulk and interfacial limits to fill-factors, Phys. Chem. Chem. Phys. 15 (39) (08/2013). [6] P.P. Boix, A. Guerrero, L.F. Marchesi, G. Garcia-Belmonte, J. Bisquert, Currentvoltage characteristics of bulk heterojunction organic solar cells: connection between light and dark curves, Adv. Energy Mater. 1 (6) (11/2011).
A.J. Janssen, Jenny Nelson, Factors limiting device efficiency in organic [7] Rene photovoltaics, Adv. Mater. (2012), http://dx.doi.org/10.1002/adma.201202873. [8] Ikerne Etxebarria, Jon Ajuria, Roberto Pacios, Solution-processable polymeric solar cells: a review on materials, strategies and cell architectures to overcome 10%, Org. Electron. 19 (April 2015) 34e60. [9] Jing-De Chen, Chaohua Cui, Yan-Qing Li, Lei Zhou, Qing-Dong Ou, Chi Li, Yongfang Li, Jian-Xin Tang, Single-junction polymer solar cells exceeding 10% power conversion efficiency, Adv. Mater. 27 (2015) 1035e1041. [10] Zhicai He, Biao Xiao, Feng Liu, Hongbin Wu, Yali Yang, Steven Xiao, Cheng Wang, Thomas P. Russell, Yong Cao, Single-junction polymer solar cells with high efficiency and photovoltage, Nat. Photonics 9 (2015) 174e179, http://dx.doi.org/10.1038/nphoton.2015.6. [11] Chang Liu, Chao Yi, Kai Wang, Yali Yang, Ram S. Bhatta, Mesfin Tsige, Shuyong Xiao, Xiong Gong, Single-junction polymer solar cells with over 10% efficiency by a novel two-dimensional donoreacceptor conjugated copolymer, ACS Appl. Mater. Interfaces 7 (8) (2015) 4928e4935, http://dx.doi.org/ 10.1021/am509047g. [12] Y. Liu, Ch-Ch Chen, Z. Hong, J. Gao, Ya Yang, H. Zhou, L. Dou, G. Li & Y. Yang, Solution-processed small-molecule solar cells: breaking the 10% power conversion efficiency, Sci. Rep. 3, Article number: 3356. [13] Q. Zhang, Bi Kan, F. Liu, G. Long, X. Wan, X. Chen, Y. Zuo, W. Ni, H. Zhang, M. Li, Zh Hu, F. Huang, Y. Cao, Z. Liang, M. Zhang, T.P. Russell, Y. Chen, Smallmolecule solar cells with efficiency over 9%, Nat. Photonics 9 (2015) 35e41. €hler, D.A. dos Santos, D. Beljonne, Z. Shuai, J.-L. Bre
das, A.B. Holmes, [14] A. Ko A. Kraus, K. Müllen, R.H. Friend, Charge separation in localized and delocalized electronic states in polymeric semiconductors, Nature 392 (30 April 1998) 903e906, http://dx.doi.org/10.1038/31901. [15] Solmaz Torabi, Fatemeh Jahani, Ineke Van Severen, Catherine Kanimozhi, Satish Patil, Remco W.A. Havenith, Ryan C. Chiechi, Laurence Lutsen, Dirk J.M. Vanderzande, Thomas J. Cleij, Jan C. Hummelen, L. Jan Anton Koster, Strategy for enhancing the dielectric constant of organic semiconductors without sacrificing charge carrier mobility and solubility, Adv. Funct. Mater. 25 (2015) 150e157. [16] Xien Liu, Kwang S. Jeong, Bryan P. Williams, Kiarash Vakhshouri, Changhe Guo, Kuo Han, Enrique D. Gomez, Qing Wang, John B. Asbury, Tuning the dielectric properties of organic semiconductors via salt doping, J. Phys. Chem. B 117 (49) (2013) 15866e15874, http://dx.doi.org/10.1021/jp408537p. [17] V. Bobnar, A. Levstik, C. Huang, Q.M. Zhang, Intrinsic dielectric properties and charge transport in oligomers of organic semiconductor copper phthalocyanine, Phys. Rev. B 71 (2005), 041202(R) e Published 20 January. [18] Oleksandr V. Mikhnenko, Paul W.M. Blom, Thuc-Quyen Nguyen, Exciton diffusion in organic semiconductors, Energy Environ. Sci. 8 (2015) 1867e1888. [19] M. Muntwiler, Q. Yang, W.A. Tisdale, X.Y. Zhu, Coulomb barrier for charge separation at an organic semiconductor interface, Phys. Rev. Lett. 101 (19) (2008 Nov 7) 196403. Epub 2008 Nov 6.


lez, Volker Ko €rstgens, Yuan Yao, Lin Song, [20] Daniel Moseguí Gonza Gonzalo Santoro, Stephan V. Roth, Peter Müller-Buschbaum, Improved power conversion efficiency of P3HT: PCBM organic solar cells by strong spineorbit coupling-induced delayed fluorescence, Adv. Energy Mater. (2015) 1401770. [21] Wenfeng Zhang, Ying Xu, Haitao Wang, Chenhui Xu, Shangfeng Yang, Fe3O4 nanoparticles induced magnetic field effect on efficiency enhancement of P3HT: PCBM bulk heterojunction polymer solar cells, Sol. Energy Mater. Sol. Cells 95 (2011) 2880e2885. [22] Kai Wang, Chao Yi, Xiaowen Hu, Chang Liu, Yan Sun, Jianhui Hou, Yongfang Li, Jie Zheng, Steven Chuang, Alamgir Karim, Xiong Gong, Enhanced performance of polymer solar cells using PEDOT: PSS doped with Fe3O4 magnetic nanoparticles aligned by an external magnetostatic field as an anode buffer layer, ACS Appl. Mater. Interfaces 6 (2014) 13201e13208, http://dx.doi.org/10.1021/ am503041g. [23] Kai Wang, Chao Yi, Chang Liu, Xiaowen Hu, Steven Chuang & Xiong Gong; Effects of magnetic nanoparticles and external magnetostatic field on the bulk heterojunction polymer solar cells; Sci. Rep. 5 : 9265 DOI: 10.1038/ srep09265. [24] J.G. Hwang, et al., Effects of nanoparticle charging on streamer development in transformer oil-based nanofluids, J. Appl. Phys. 107 (2010) 014310e014326. [25] V.D. Mihailetchi, L.J.A. Koster, J.C. Hummelen, P.W.M. Blom, Photocurrent generation in polymer-fullerene bulk heterojunctions, 216601e1-4, Phys. Rev. Lett. 93 (2004). [26] C.G. Shuttle, et al., Bimolecular recombination losses in polythiophene: fullerene solar cells, 113201-1-4, Phys. Rev. B 78 (2008). [27] C.G. Shuttle, R. Hamilton, B.C. O'Regan, J. Nelson, J.R. Durrant, Chargedensitybased analysis of the current-voltage response of polythiophene/Fullerene photovoltaic devices, Proc. Natl. Acad. Sci. 107 (2010) 16448e16452. [28] Michael T. Sang-Yup Lee, Harris, Surface modification of magnetic nanoparticles capped by oleic acids: characterization and colloidal stability in polar solvents, J. Colloid Interface Sci. 293 (2) (15 January 2006) 401e408. [29] Shuang Chena, Weimin Liub, Oleic acid capped PbS nanoparticles: synthesis, characterization and tribological properties, Mater. Chem. Phys. 98 (1) (1 July 2006) 183e189. [30] Daliya S. Mathew, Ruey-Shin Juang, An overview of the structure and magnetism of spinel ferrite nanoparticles and their synthesis in microemulsions, Chem. Eng. J. 129 (1e3) (1 May 2007) 51e65. [31] A.L. Patterson, The scherrer formula for X-ray particle size determination, Published 15 November, Phys. Rev. 56 (1939) 978. [32] M. Satalkar, N. Ghodke, S.N. Kane, Influence of high temperature sintering on the structural and magnetic Properties of Mn1-xZnxFe2O4, J. Phys. Conf. Ser. 534 (2014) 012016. [33] D.G. Bounds, A. Hinchliffe, M. Barber, A molecular orbital study of the effect of an external electric field on methanol;, J. Mol. Struct. 37 (2) (April 1977) 283e288. [34] J.A.W. Zhao, X. Dai, M.A. Green, S.R. Wenham, Improvements in silicon solar cell performance, in: 22nd IEEE PV Specialists Conference, 1991, pp. 399e402. [35] C. Behera, R.N.P. Choudhary, Piyush R. Das, Size effect on electrical and magnetic properties of mechanically alloyed CoFe2O4 nanoferrite, J. Mater. Sci. Mater. Electron 26 (2015) 2343e2356, 2351. [36] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo, A. Hagfeldt, Impedance spectroscopy study of the influence of electrolyte conditions in parameters of transport and recombination in dye-sensitized solar cells, Sol. Energy Mater. Sol. Cells 87 (2005) 117e131. [37] G. Garcia-Belmonte, A. Guerrero, J. Bisquert, Elucidating operating modes of bulk-heterojunction solar cells from impedance spectroscopy analysis, J. Phys. Chem. Lett. 4 (2013) 877e886. [38] I. Mora-Sero, G. Garcia-Belmonte, P.P. Boix, M.A. Vazquez, J. Bisquert, Impedance characterisation of highly efficient silicon solar cell under different light illumination intensities, Energy Environ. Sci. 2 (2009) 678e686. [39] J. Bisquert, Theory of the impedance of electron diffusion and recombination in a thin layer, J. Phys. Chem. B 106 (2002) 325e333. [40] J. Bisquert, I. Mora-Sero, F. Fabregat-Santiago, Diffusion-recombination impedance model for solar cells with disorder and nonlinear recombination, Chem. Electro. Chem. 1 (2013) 289e296. [41] G. Garcia-Belmonte, et al., Charge carrier mobility and lifetime of organic bulk heterojunctions analyzed by impedance spectroscopy, Org. Electron. 9 (2008) 847e851, 37 J. Bisquert, J. Phys. Chem. B 106 (2002) 325e333. [42] Juan Bisquert, Iv
an Mora-Sero, Francisco Fabregat-Santiago, Diffusionerecombination impedance model for solar cells with disorder and nonlinear recombination, Chem. Electro. Chem. 1 (1) (January 3, 2014) 289e296, http://dx.doi.org/10.1002/celc.201300091.

́

́