Observation of ferromagnetic ordering in conjugated polymers exhibiting OMAR effect

Organic Electronics 21 (2015) 66–72

Contents lists available at ScienceDirect

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

Observation of ferromagnetic ordering in conjugated polymers exhibiting OMAR effect Sayani Majumdar a,b,⇑, Jan-Olof Lill c, Johan Rajander c, Himadri Majumdar d a

Nanospin, Department of Applied Physics, Aalto University School of Science, FI-00076 Aalto, Finland Wihuri Physical Laboratory, Department of Physics and Astronomy, University of Turku, 20014, Finland c Accelerator Laboratory, Turku PET Center, Åbo Akademi University, 20500, Finland d VTT Technical Research Center of Finland, FI-02150 Espoo, Finland b

a r t i c l e

i n f o

Article history: Received 24 November 2014 Received in revised form 9 February 2015 Accepted 26 February 2015 Available online 2 March 2015 Keywords: Organic magnetoresistance Magnetic semiconductor PIXE

a b s t r a c t We report observation of ferromagnetic (FM) ordering in p-conjugated polymeric semiconductors, namely regio-regular poly (3-hexyl thiophene) (RRP3HT) and 1-(3-methoxycarbonyl)propyl-1-phenyl[6,6]-methanofullerene (PCBM), in the temperature range of 5–300 K. Diodes made from these polymers exhibit sizable magnetoresistance (known commonly as OMAR). However, upon blending these two materials, the FM ordering is suppressed by a huge paramagnetic (PM) signal. In the diodes with RRP3HT:PCBM blend showing PM response, OMAR decreases substantially. Particle induced X-ray emission spectroscopy indicate presence of dilute magnetic impurities as residues from the synthesis process in the individual polymers. However the impurity signal is unable to explain the temperature dependence of magnetization in these materials and the observed paramagnetism of the RRP3HT:PCBM blend. We propose a hypothesis for the origin of the FM nature of these polymers based on ours and previously reported experimental observations and pointed out that the reported organic magnetoresistance might have a close correlation with the ferromagnetic interaction in these polymers. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Over the past two decades, the growth of organic electronics has been phenomenal [1–3]. However, until recently, all research has been predominantly focused towards the understanding of charge transport in organic semiconductors (OS) small molecules and pi-conjugated polymers (PCP) and its implications in devices. Study of spin in organics is a relatively unexplored territory. It provides a whole new scope of physics along with immense potential for application. Spintronics (spin-based electronics) refers to the study of the role played by electron spin in solid state physics, and possible devices that specifically exploit spin properties instead of or in addition to charge degrees of freedom [4]. In recent years semiconductors are studied widely as promising spin transport materials [5] for spintronic applications. Although great research effort is dedicated to explore the inorganic magnetic semiconductors for spintronic applications very little attention has been paid to the use OS molecules and PCPs as spintronic materials. Long spin-correlation length is expected in OS compared to their ⇑ Corresponding author at: Wihuri Physical Laboratory, Department of Physics and Astronomy, University of Turku, 20014, Finland. E-mail address: sayani.majumdar@utu.fi (S. Majumdar). http://dx.doi.org/10.1016/j.orgel.2015.02.025 1566-1199/Ó 2015 Elsevier B.V. All rights reserved.

inorganic counterparts as they are composed of mainly light molecules and possess low spin–orbit interaction and hyperfine interaction [6]. PCPs are especially the better choice as their conjugation length is much higher than the small molecules and oligomers, leading to better transport properties and simpler fabrication process [7]. The ability to manipulate electron spin in organic molecules offers an alternative route to conventional spintronics, both from fundamental and technological points of view and with the emergence of different experimental results, it is fast becoming a subject of huge research interest [8–12]. Increased ferromagnetism at the ferromagnet–OS interface [12 and references therein] especially raised high interest for highly spin polarized interface (termed as ‘‘spinterface’’) in the organic spintronic devices. Magnetic field effects on OS were demonstrated already in the early nineties [13]. However, demonstration of successful spin injection and transport in lateral [14] and vertical [15] spin-valve devices and at the same time large magnetoresistance (OMAR) and magneto-electroluminescence (MEL) response in OS diode devices with non-magnetic electrode at room temperature under a small applied magnetic field (10–100 mT) [16–18] caused resurgence of interest. Several experiments have so far successfully demonstrated efficient spin injection and transport in organic spin-valve structures and their successful operations, even at room

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

temperature [19,20], paving the way for future commercial applications. However, the reason for large OMAR effect in organic diodes has not been explained very successfully yet. Different models have been proposed to explain the observed effects [21–25] but differences between theory and experimental results still exist. Ferromagnetism in conjugated polymers has been demonstrated before. Ferromagnetic (FM) ordering was shown in doped poly(3-methyl thiophene) (P3MT) using Superconducting quantum interface designed (SQUID) magnetometer and electron spin resonance (ESR) measurements and it was found that FM ordering in P3MT is independent of magnetic impurities [26]. Hopping transport in conducting polymers has also been shown to be significantly affected in presence of such FM ordering [27]. More recently ferromagnetism in polythiophene was observed using different techniques like Faraday rotation [28] and magnetic force microscopy (MFM) measurements [29]. In our previous works we have observed that there are signatures of magnetic domain formation in OS [30] from magneto-transport measurements. In organic spin-valves fabricated with regio-regular poly (3-hexyl thiophene) (RRP3HT) [17] as the spacer we observed enhanced switching field than the expected values from using La0.7Sr0.3MnO3 and Co electrodes. Furthermore, ITO/RRP3HT/Al diodes showed hysteretic nature of device resistance as a function of magnetic fields [30]. These observations indicate signatures of trapped spins in the bulk of the PCPs forming magnetic domains. The magnetic properties and their physical origin in the PCPs is therefore an essential characteristic that needs to be investigated in order to fully understand different organic spintronic devices and for designing devices with new functionalities. In an earlier report [31] we chose the RRP3HT:PCBM blend as the model system to study the effect of electron–hole (e–h) recombination on the OMAR response of the device. We found that with decreasing probability of Coulombically bound e–h pair formation, the OMAR response decreased significantly. In our present experiment also, we have chosen RRP3HT, 1-(3-methoxycarbonyl)propyl-1-phenyl-[6,6]-methanofullerene (PCBM) and the RRP3HT:PCBM blend (1:1) as the test systems to investigate the magnetic properties. In the present paper, we report the magnetic characterization together with their OMAR characteristics which points towards a very intriguing correlation between magnetic and magnetotransport properties of these p-conjugated polymers, so far unnoticed.

67

polarization in dark and under light, the thin-film sample was illuminated using an Ar-ion laser working at k = 514.5 nm (2.42 eV). To re-confirm the magnetic data, ac-susceptibility was measured from the RRP3HT powder using the vibrating sample magnetometer of the Quantum Design Physical Property Measurement System (PPMS) at two different frequencies of 1 and 5 kHz between 10 and 300 K. For elemental analysis by the particle induced X-ray emission (PIXE) technique, about 15 mg of the polymer sample material was pressed to a pellet (diameter 13 mm). Pure graphite was used as backing material to minimize the amount of sample material [32]. The samples were irradiated with a 3 MeV proton beam from the Åbo Akademi MGC-20 cyclotron. The acquisition time was about 500 s with a beam current of 10 nA. All irradiations were performed in air to avoid heating and charge build-up. A strong ion luminescence from the irradiated spot on the polymer samples was observed in the beginning of the proton irradiation but faded away within a few seconds. This phenomenon indicates some changes in the molecular structure but does not affect the elemental concentration measurements. The radiation emitted from the sample during the irradiation was measured with an IGP X-ray detector for PIXE analysis. The integrated charge on the target needed for quantification was determined from measurements of light induced in air by the proton beam [33]. The obtained PIXE spectra were analyzed using the GUPIX software package [34]. The calibration was checked using the USGS granite CRM G-2. The procedure with the quality assurance has been described earlier [35]. 3. Results and discussion Fig. 1 shows magnetoresistance curves for an ITO/RRP3HT/Al diode (device schematics as shown in figure) showing a typical positive magnetoresistance (closed circle) and a hysteretic (open circle) curve between 300 mT and +300 mT. In the typical scan,

2. Experimental The device structure used for the magneto-transport experiments is described earlier [31]. For the magnetic measurements, RRP3HT, either pure powder (as received from Sigma–Aldrich) or drop-casted films from chloroform or dichlorobenzene (DCB) solution on ITO or SiO2 substrates were used. The PCBM powder was measured as received. For the blend, 1:1 weight ratio of RRP3HT:PCBM was dissolved in DCB and films were made by drop-coating method on SiO2 substrates to minimize any magnetic signal from the substrates. For all the films, a solution of density 5 mg/ml was used and the approximate thicknesses of the films were 1 lm. The sample preparation was done in a nitrogen-filled glove-box and using anhydrous solutions. After fabrication, the films are transferred in a nitrogen atmosphere to the SQUID magnetometer. Temperature dependence of magnetization, M, was measured both after cooling the sample under zero field and then measuring the sample while heating under an applied field (ZFC) and during cooling the sample under the measuring field (FC) at temperatures between 5 and 300 K. Magnetic hysteresis curves were recorded in the field of B = ±150 mT. All measurements in this case were made in dark. The external field B was always applied along the plane of the films. For the measurement of spontaneous

Fig. 1. Device structure and magnetoresistance response as a function of magnetic field (B) at 300 K for an ITO/RRP3HT/Al diode measured with 1 lA current (Black square) when B was scanned from 0 to +300 mT and 0 to 300 mT with a scan speed of 1 mT/s and showing resistance hysteresis (Red circle) behavior when B was scanned from 0 to +300 mT, +300 mT to 300 mT and back to +300 mT at a scan speed of 20 lT/s. Arrows indicate the direction of the sweeping field. (Inset) Current–voltage characteristics of one such diode at 300 K. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

(a)

Magnetization/unit mass -3 2 (10 Am /Kg)

4

B = 150 mT

3

RRP3HT ZFC RRP3HT FC PCBM ZFC PCBM FC

FC

ZFC FC

2

ZFC

0

50

100

150

200

250

300

T (K) 7.2

(b)

2

Magnetization (10 Am )

B = 150 mT

-8

RRP3HT:PCBM ZFC RRP3HT:PCBM FC

FC

7.0

6.8 ZFC

6.6 0

50

100

150

200

250

300

T (K) 4 2

(c)

1 k Hz 10 K

' (10-4 Am2)

the diode was scanned from 0 to +300 mT and then scanned from 0 to 300 mT after a sufficient time interval. In the hysteretic scan the magnetoresistance was measured continuously from 0 to +300 mT, +300 mT to 300 mT and back to +300 mT. The scan speed for the hysteresis measurement was 20 lT/s. In the hysteresis measurement the %MR plot differed completely from a typical MR plot. The measurement was done by sending a constant current and measuring the voltage in presence of a certain magnetic field. For the measurement with 1 mT/s, the sample was subjected to the external magnetic field (between 0 and 300 mT) for only 5 min while in the 20 lT/s. case, the sample was subjected to the magnetic field for almost 1 day and the difference between them clearly indicate that the magnetoresistance behavior of the organic diodes are time dependent. It is important to mention here that no evidence of such hysteretic behavior in current–voltage characteristics (inset Fig. 1), measured without a magnetic field, was observed and the device was stable for the whole measurement period and afterwards. Therefore, such hysteretic nature of %MR could possibly indicate a long relaxation time of polarized spins at a particular site leading to formation of magnetically aligned domains within the device. In case of PCPs, a localized carrier in a deep trap state could become magnetically polarized under an applied magnetic field and can form bound magnetic polarons that can lead to long range magnetic interactions and modify the magnetotransport properties of the material. To verify this hypothesis, magnetization measurements of the individual polymer components is essential and we performed a thorough DC magnetization measurement using SQUID magnetometer and also some confirmation experiments using AC susceptibility measurements. For magnetization measurements, first RRP3HT and PCBM powders were separately wrapped in clean TeflonÓ [36] tape and measured separately in the SQUID magnetometer in the temperature range of 5–300 K both in ZFC and FC directions with an applied field of 150 mT (Fig. 2(a)). Magnetization value increases fast below 100 K for both RRP3HT and PCBM. The ZFC and FC curves start to separate from each other below 150 K, which is usually observed for frustrated or low-dimensional magnetic systems like spin glasses and indicate domain freezing behavior. The measured saturation magnetic moment (MS) for both materials are given in Table 1. It was additionally found that for films made from RRP3HT dissolved in chloroform and in DCB the MS values were widely different (not shown here). While the RRP3HT-from-DCB film showed clear FM signal till room temperature, for the RRP3HT-from-chloroform film, it was almost impossible to measure the signal above 50 K. This difference in MS values suggests a dependence of the magnetic response of organic films on its morphology. It has been proven extensively in literature that uses of either chloroform or DCB lead to complete different crystallization of the thiophene polymer. One of the very early work by the group in Cambridge [37] clearly demonstrate that change in solvent changes microstructure of the thiophene. This leads to formation or nonformation of defect states that works as traps and change carrier concentration. For the films made from the RRP3HT:PCBM blend, no difference in FC and ZFC were observed and the M was found to be temperature independent (Fig. 2(b)). To confirm the magnetic data, ac-susceptibility of RRP3HT powder and film was measured using a completely different setup (vibrating sample magnetome-

' (10-4 Am2)

68

RRP3HT

1.0 0.5 0.0 0

0

100 200 T (K)

300

-2 -4 -300 -200 -100

0

100

200

300

B (mT) Fig. 2. Zero-field cooled (ZFC) (open) and field cooled (FC) (closed) magnetization of (a) RRP3HT and PCBM powders under an applied field of 150 mT. The arrows indicate the domain freezing temperatures. (b) ZFC (open) and FC (closed) magnetization of RRP3HT:PCBM (1:1) blend film on SiO2 under 100 mT magnetic field. (c) Real part of ac-susceptibility as a function of bias field at 10 K for RRP3HT powder measured with a driving frequency of 1 kHz.

ter in the PPMS) where the powder was wrapped in a tissue and placed inside a plastic straw. Fig. 2 (c) shows real part of ac-susceptibility as a function of magnetic field at 10 K. The time-dependent magnetic moment induced in the sample due to an applied field of 1 kHz frequency shows a non-zero value dependent on the applied magnetic field confirming ferromagnetic interaction. Also temperature dependent AC susceptibility data confirms similar trend

Table 1 Magnetic moment (MS) and coercive field (HC) values of different polymers at 5 K and 300 K. Polymer ; RRP3HT powder RRP3HT film from dichlorobenzene solution RRP3HT film from chloroform solution PCBM

MS @ 5 K

MS @ 300 K 20

1

5.57  10 lB kg 1.08  1020 lB cc 1 5.9  1018 lB cc 1 2.5  1020 lB kg 1

20

3.98  10 lB kg 8.85  1019 lB cc Not measurable 1.99  1020 lB kg

1 1

1

HC @ 5 K (mT)

MS from impurities

11.2 4.5 16.5 6.7

3.5–4.5  1020 lB kg 1 – – 9 ± 1.8  1020 lB Kg 1

69

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

of a PCBM diode device made similarly as the RRP3HT diode device. Although the OMAR response is smaller in this case compared to the RRP3HT diode, clear OMAR signal was obtained from each device. Next we studied the magnetization behavior of the RRP3HT:PCBM blend and found that the 1:1 weight ratio of RRP3HT and PCBM is paramagnetic (PM) at all temperatures between 5 and 300 K. Fig. 3(e) shows the M–B plot of a typical 1:1 blend of RRP3HT and PCBM at room temperature. Upon closer inspection, it is observed that FM from the individual components exists but it is suppressed by the large PM signal originating from the blend. Suppression of FM interaction is concluded from the significant decrease of the remanent magnetization and the HC values of the blend compared to the individual components. PM response of the similar blends was reported earlier using light induced ESR spectroscopy and the result was explained as separation of charge

0.008 0.004

RRP3HT:PCBM 300 K

0.04

RRP3HT:PCBM 300K

0.03

0.000

% MR

Magnetization/ unit area (A)

in magnetic moment as shown and discussed before for DC magnetization (inset of Fig. 2(c)). Further measurements were done to trace the origin of the ferromagnetism in RRP3HT and PCBM. The RRP3HT powder shows open magnetization vs. magnetic field (M–B) curves (Fig. 3a) at all measured temperatures. This also indicates FM ordering even up to 300 K. The magnetic hysteresis shows the saturation field to be 100 mT. Fig. 3(b) shows the OMAR response of the RRP3HT diode as a function of magnetic field at room temperature. The OMAR was measured using 100 lA current in a standard diode configuration with ITO and Al electrode and the measurement technique has been described previously in details [31]. Similar measurements were done on PCBM powder. The magnetic hysteresis shows the saturation field to be 100 mT (Fig. 3(c)). The saturation magnetic moment (MS) and coercive field (HC) values are listed in Table 1. Fig. 3(d) shows the OMAR response

-0.004

0.02 0.01

(e)

-0.008 -200

-100

0

B (mT)

100

200

(f)

0.00 -300

-200

-100

0

100

200

300

B (mT)

Fig. 3. Magnetization as a function of magnetic field (B) at 300 K for (a) RRP3HT powder, (c) PCBM powder and (e) RRP3HT:PCBM blend film on SiO2 substrate. % Magnetoresistance as a function of magnetic field at 300 K for diode devices with (b) RRP3HT, (d) PCBM and (f) RRP3HT:PCBM solar cells with ITO as the hole and Al as the electron injector.

70

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

carriers into different paths on the polymer and fullerene blends [38]. In a RRP3HT:PCBM blend, PCBM is embedded in the conjugated polymer and acts as a strong electron acceptor whose lowest unoccupied molecular orbital (LUMO) lies below the excitonic state. It has been shown earlier that upon irradiation, electron is transferred from a polymer chain to a fullerene molecule within femtosecond time scale [39], whereas the electron back transfer is much slower and this electron transfer results in a charge-separated state. Krinichnyi [38] showed that upon irradiation, this charge separated state can give rise to a huge PM signal and the effective PM susceptibility of the generated polarons and fullerene anion-radicals, is inversely proportional to the probability of their recombination, in excellent agreement with the transport data in Ref. [25]. However, we observe a strong PM signal from the 1:1 blend of RRP3HT and PCBM already in the dark, implying that a ground state charge transfer complex is formed in the RRP3HT:PCBM blend at the interface between the polymer and fullerene molecule, leading to separation of free carriers. Existence of such charge-transfer complexes have been reported before [40,41]. The PM signal arising from these free carriers suppress the individual FM ordering in RRP3HT and PCBM and we observe only a large PM signal. Fig. 3(f) shows the OMAR response of a diode made from the blend – orders of magnitude smaller than that of a RRP3HT or PCBM diode. Recently Yang et al. [29] has shown ferromagnetic ordering using MFM imaging from P3HT:PCBM blend which was annealed for more than 12 h. They reported magnetic domains of sizes from tens to hundreds of nanometers, distributed uniformly over the entire film, in the crystalline phase of the blend that was achieved using the slow drying process. However, this fabrication technique is completely different from ours and hence should not be directly compared to our results. While preparing our samples for magnetic measurements, we have kept the preparation condition identical to the diode conditions. Therefore it can be concluded that depending on film morphology and carrier concentration, magnetic properties of organic semiconductors can change significantly. In order to investigate the origin of the FM ordering in the polymer RRP3HT and fullerene PCBM, impurities in these materials were checked by PIXE analysis. Fig. 4 shows a typical PIXE spectrum, with clear peaks from S, Ni and Br. Sulfur can be found in the RRP3HT backbone and bromine in the ends of the polymer chain, the other elements are considered to be impurities. Elemental concentrations and statistical uncertainties for the samples are shown in Table 2. From the PIXE data we find that both RRP3HT and PCBM samples contain small amounts of magnetic impurities like Fe, Ni, Co, Mn etc. Traces of Ni were found in both the polymers while Fe was the main magnetic impurity in PCBM. Earlier [42–44] very weak FM was observed in C60 molecule and the FM ordering was attributed to the presence of Fe impurities. In RRP3HT the Ni impurity amount is 12 lg/g and in PCBM the Fe impurity amount is

S

RRP3HT Br(Kα)

4

Intensity [a.u.]

10

Br(Kβ) esc

3

10

Ni(Kα)

102

101

2

4

6

8

10

12

14

X-ray energy [keV] Fig. 4. A typical energy spectrum of regio-regular poly(3-hexyl thiophene) (RRP3HT) showing presence of Ni among different impurities.

Table 2 Elemental concentrations in the materials given in lg/g of dry weight. Typical statistical errors in % and limit of detection (LOD) in lg/g are in the columns to the right. Elemental content ;

RRP3HT

PCBM

Error (%)

LOD

S Br Mn Fe Ni

180,360 3975 bdl bdl 12

bdl bdl bdl 15 1.7

0.2 0.4 – 20 14

190 1.4 8.7 4.9 1.2

bdl – below detection limit.

15 lg/g (Table 2). Now, spin magnetic moment of Ni varies between 2.9 and 3.3 lB/atom in octahedral complexes and between 3.7 and 4.0 lB/atom for tetrahedral complexes [45]. For Fe, the spin moments vary between 5.1 and 5.7 lB/atom in octahedral complexes and between 5.3 and 5.5 lB/atom in tetrahedral complexes. So, the magnetic moment from Ni impurity (in RRP3HT) can be calculated as 3.5–4.5  1020 lB/kg and that for Fe impurity (in PCBM) can be 9  1020 lB/kg. Both the values are similar in magnitude and shows quite good matching with the experimentally obtained results of RRP3HT at room temperature while that for PCBM is a bit lower suggesting that the measured magnetic moment in the pure polymer powder can arise from the magnetic impurities present in the samples. However, we observed from Fig. 2(a) that the magnetization of both RRP3HT and PCBM is temperature dependent, in contrary to the temperature independent magnetization expected from Ni and Fe in the 5–300 K range. The impurity concentration is also too dilute to lead to long-range magnetic interaction between the impurities. Hence, it is also a possibility that the FM interaction in RRP3HT and PCBM might be mediated via an intermediary that gives rise to longrange FM ordering, as previously shown in EuO where the O2 ligands mediates the super-exchange interaction [46]. Comparison between C60 and PCBM shows that presence of hydrogen atoms, i.e. hyperfine interaction, in PCBM increases FM ordering (results not shown here). This indicates that the hyperfine interaction might play the role of the intermediary in the long-range FM ordering in OS. From our measurements, the magnetization value suggests that the strength of magnetic interaction inside the polymer is of the order of few milliteslas (mT) which is quite comparable to the hyperfine strength of the organic semiconductors. [47]. However, detailed nuclear magnetic resonance (NMR) study of hydrogen atoms is needed to further elaborate the role of hyperfine interaction in determining FM ordering in OS. Another important point is the morphology dependence of the magnetic saturation moment in RRP3HT. Although the amount of magnetic impurity is same, the saturation moment varies by almost 2 orders of magnitude depending on the film morphology (Table 1). So, it becomes evident that depending on the crystallinity and the domain size FM interaction can vary significantly. Earlier Nascimento et al. [48] have shown similar FM ordering in doped P3MT using SQUID magnetometer and ESR spectroscopy data and they found that FM ordering in P3MT is independent of magnetic impurities. They attributed the magnetic property to the interaction of spin ½ polarons, which can be either FM or anti-ferromagnetic (AFM) depending on morphology. It is, therefore, possible that a trapped carrier mediated FM interaction, as is usually observed in metal-oxide systems [40], could also play an important role here. Recently, Gulacsi et al. [49] showed that pentagon-chain polymers with electron densities above half filling can become ferromagnetic or half metallic due to an unexpected mechanism in multi-orbital polygon chains with different size-dependent Coulomb interaction strengths. This interaction can turn the dispersive band into an effectively flat band in a pentagon chain polymer system leading to long range FM ordering. The carrier concentration of

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

Magnetization (A/m)

B swiched off 103

300 K RRP3HT

B = 150 mT Sample under optical excitation (B = 0)

Sample in dark (B = 0)

102 102

103

104

time (sec) Fig. 5. Magnetization as a function of time in RRP3HT thin films under optical excitation and in dark after switching the external magnetic field (B) off showing significantly increased spontaneous magnetic moment under light.

the pure P3HT (without doping) is 1016/cc. According to the calculations of Gulacsi et al. [49] this carrier density can produce enough site dependent Coulomb interactions and is sufficient to turn the dispersive band into flat band leading to ferromagnetism and half metallicity. de Paula et al. [50] have shown that depending on number of polarons, the ferro and paramagnetic properties can be modified and they attributed this to the distance dependent interactions between spin ½ polarons in different polymer chains. We have also observed that carrier concentration plays a significant role in determining the magnetic interaction strength in polymers. Fig. 5 shows that relaxation in spontaneous magnetization in RRP3HT thin films in darkness and under light. After the external field is switched off, it can be seen clearly that spontaneous magnetization decays much faster in absence of light (i.e. when carrier concentration is low) and almost 1 order of magnitude lower than that under light. This again indicates that number of carriers and Coulomb interaction strength can play an important role in inducing FM interaction. Finally, we have observed that when a sizable FM ordering in a material exists, we always find a large OMAR response in the diodes with the respective material as shown in Fig. 3. However, when we observe a large PM signal, the OMAR response is significantly reduced. In order to explain this observation we consider the probable effect of spin dynamics on the hopping transport. Depending on whether the carrier spins are aligned in a certain direction or not, the available numbers of hopping sites change significantly. Earlier Zuppiroli et al. [27] showed that hopping transport in conducting polymers can be significantly affected in presence of FM ordering. When a carrier with a particular spin orientation hops to its neighboring site, the hopping probability is not the same if the spin of the carrier is parallel or anti-parallel to the neighboring spin. According to Pauli exclusion principle, two carriers with same spin configuration cannot occupy the same site and therefore polarized spins at some hopping sites eventually ‘‘block’’ some transport channels for the carriers which lead to magnetic field dependent resistance in organic based devices, as also pointed out by Harmon and Flatte [51]. Also another possibility is that like in many other frustrated or weak magnetic systems, inside the polymers there could be regions where spins are aligned in the direction of magnetic field (FM domains) together with regions of random spin orientations (PM domains). Even magnetizations in the grains and the grain boundaries are different leading to the formation of domains and domain walls. Now there could be also tunneling of carriers between magnetically aligned domains separated by thin domain walls and could lead to grain boundary tunneling kind of magnetoresistance response. Therefore we conclude that whether the carriers are aligned or not can significantly modify the hopping of carriers through the system. Together with our present observation, it can be inferred that with sizable number of polarized spins inside a material, magneto-transport properties inside the material would

71

be modified and therefore FM ordering in OS might have a correlation with the OMAR effect. However, further experiments like electric-field induced magnetization study are necessary to clarify the true origin of this effect. In conclusion, we have observed FM ordering in RRP3HT and PCBM but in 1:1 blends of these materials a large PM signal is observed. PIXE analysis of the individual materials suggests the presence of magnetic impurities which can almost account for the observed magnetic moment in the powders. However, these dilute magnetic impurities are unable to explain the morphology and temperature dependence of magnetization in these polymer thin films. In samples with FM ordering, we observe significant OMAR effect but OMAR disappears in RRP3HT:PCBM blend films having PM response. This is in accordance with a charge transfer complex formed in the blend giving rise to free carriers. Therefore it can be concluded that transport mechanism in conjugated polymers can be modified due to presence of polarized spins and this knowledge can be very useful for future printed magnetic sensors using these polymers. However, further work is needed to elucidate whether the coexistence of OMAR and FM ordering in conducting polymers is coincidental or strongly correlated. Acknowledgements The authors gratefully acknowledge the Kone and Jenny and Antti Wihuri Foundation and Prof. Ronald Österbacka for financial support. References [1] T.W. Tang, S.A. Van Slyke, Appl. Phys. Lett. 51 (1987) 913. [2] A. Tsumura, H. Koezuka, T. Ando, Appl. Phys. Lett. 49 (1986) 1210. [3] H.E.A. Huitema, G.H. Gelinck, J.B.P.H. van der Putten, K.E. Kuijk, C.M. Hart, E. Cantatore, P.T. Herwig, A.J.J.M. van Breemen, D.M. de Leeuw, Nature 414 (2001) 599. [4] G.A. Prinz, Science 282 (1998) 5394. [5] I. Zutic, J. Fabian, S. Das Sarma, Rev. Mod. Phys. 76 (2004) 323. [6] P.P. Ruden, D.L. Smith, J. Appl. Phys. 95 (2004) 4898. [7] S. Majumdar, H.S. Majumdar, R. Laiho, R. Österbacka, J Alloys Compd. 423 (2006) 169. [8] D. Sun, E. Ehrenfreund, Z.V. Vardeny, Chem. Commun. 50 (2014) 1781. [9] S. Majumdar, H.S. Majumdar, Org. Electron. 13 (2012) 2653. [10] F. Djeghloul et al., Sci. Rep. 3 (2013) 1272. [11] A.J. Drew et al., Nat. Mater. 8 (2009) 109. [12] S. Majumdar, S. Dey, H. Huhtinen, J. Dahl, M. Tuominen, P. Laukkanen, S. van Dijken, H.S. Majumdar, Spin 4 (2014) 1440009. [13] E.L. Frankevich, A.A. Lymarev, I. Sokolik, F.E. Karasz, S. Blumstengel, R. Baughman, H.H. Hörhold, Phys. Rev. B 46 (1992) 9320. [14] V. Dediu, M. Murgia, F.C. Matacotta, C. Taliani, S. Barbanera, Solid State Commun. 122 (2002) 181. [15] Z.H. Xiong, D. Wu, Z.V. Vardeny, J. Shi, Nature 427 (2004) 821. [16] J. Kalinowski, M. Cocchi, D. Virgili, V. Fattori, P. Di Marco, Phys. Rev. B 70 (2004) 205303. [17] T.L. Francis, Ö. Mermer, G. Veeraraghavan, M. Wohlgenannt, New J. Phys. 6 (2004) 185. [18] Ö. Mermer, G. Veeraraghavan, T.L. Francis, Y. Sheng, D.T. Nyugen, M. Wohlgenannt, A. Köhler, M.K. Al-Suti, M.S. Khan, Phys. Rev. B 72 (2005) 205202. [19] M. Cinchetti, K. Heimer, J.-P. Wüstenberg, O. Andreyev, M. Bauer, S. Lach, C. Ziegler, Y. Gao, M. Aeschlimann, Nat. Mater. 8 (2009) 115. [20] S. Majumdar, H.S. Majumdar, P. Laukkanen, J. Värynen, R. Laiho, R. Österbacka, Appl. Phys. Lett. 89 (2006) 122114. [21] V.N. Prigodin, J.D. Bergeson, D.M. Lincoln, A.J. Epstein, Synth. Met. 156 (2006) 757. [22] P. Desai, P. Shakya, T. Kreouzis, W.P. Gillin, N.A. Morley, M.R.J. Gibbs, Phys. Rev. B 75 (2007) 094423. [23] P.A. Bobbert, T.D. Nyugen, F.W.A. van Oost, B. Koopmans, M. Wohlgenannt, Phys. Rev. Lett. 99 (2007) 216801. [24] B. Hu, Y. Wu, Nat. Mater. 6 (2007) 985. [25] F.J. Wang, H. Bässler, Z.V. Vardeny, Phys. Rev. Lett. 101 (2008) 236805. [26] O.R. Nascimento, A.J.A. de Oliveira, E.C. Pereira, A.A. Correa, L. Walmsley, J. Phys. Condens. Matter. 20 (2008) 035214. [27] L. Zuppiroli, M.N. Bussac, S. Paschen, O. Chauvet, L. Forro, Phys. Rev. B 50 (1994) 5196. [28] P. Gangopadhyay, G. Koeckelberghs, A. Persoons, Chem. Mater. 23 (2011) 516. [29] B. Yang, Z. Xiao, Y. Yuan, T.V. Jayaraman, J.E. Shield, R. Skomski, J. Huang, Polymer 54 (2013) 490. [30] S. Majumdar, H.S. Majumdar, H. Aarnio, R. Laiho, R. Österbacka, Phys. Status Solidi RRL 3 (2009) 242.

72

S. Majumdar et al. / Organic Electronics 21 (2015) 66–72

[31] S. Majumdar, H.S. Majumdar, H. Aarnio, D. Vanderzande, R. Laiho, R. Österbacka, Phys. Rev. B 79 (2009) 201202. [32] K.-E. Saarela, J.-O. Lill, F.J. Hernberg, L. Harju, A. Lindroos, S.-J. Heselius, Nucl. Instr. Meth. B 1003 (1995) 466. [33] J.-O. Lill, Nucl. Instr. Meth. B 150 (1999) 114. [34] J.A. Maxwell, J.L. Campbell, W. Teesdale, Nucl. Instr. Meth. B 43 (1989) 218. [35] J.-O. Lill, Quantitative elemental analysis of thick solid samples using the PIXE technique and a millimetre-sized proton beam (Dissertation), Åbo Akademi University, Turku, 2008. [36] Teflon tape has negligible magnetic moment in the magnetic field range <300 mT. Above 1 T, it shows diamagnetic behaviour. [37] J.-F. Chang, B. Sun, D.W. Breiby, M.M. Nielsen, T.I. Sölling, M. Giles, I. McCulloch, H. Sirringhaus, Chem. Mater. 16 (2004) 4772. [38] V.I. Krinichnyi, Sol. Energy Mater. Sol. Cells 92 (2008) 942. [39] C.H. Lee, G. Yu, D. Moses, K. Pakbaz, C. Zhang, N.S. Sariciftci, A.J. Heeger, F. Wudl, Phys. Rev. B 48 (1993) 15425.

[40] Z. Xu, H. Zang, B. Hu, J. Miner. Met. Mater. Soc. 60 (9) (2008) 49. [41] P. Parkinson, J. Lloyd-Hughes, M.B. Johnston, L.M. Herz, Phys. Rev. B 78 (2008) 115321. [42] R. Höhne, P. Esquinazi, Adv. Mater. 14 (2002) 753. [43] D. Spemann, Nucl. Imstrum. Methods B 210 (2003) 531. [44] T. Makarova, F. Palacio (Eds.), Carbon-Based Magnetism, Elsevier, Amsterdam, 2006. [45] A.B.P. Lever, Inorg. Chem. 4 (1965) 763. [46] B.T. Matthias, R.M. Bozorth, J.H. Van Vleck, Phys. Rev. Lett. 7 (1961) 160; A. Mauger et al., J. Phys. CS 41 (1980) 263. [47] P. Bobbert, Nat. Mater. 9 (2010) 288. [48] A. Kaminski, S. Das Sarma, Phys. Rev. Lett. 88 (2002) 247202. [49] Z. Gulacsi, A. Kampf, D. Volhardt, Phys. Rev. Lett. 105 (2010) 266403. [50] F.R. de Paula, D. Schiavo, E.C. Pereira, A.J.A. de Oliveira, J. Magn. Magn. Mater. 370 (2014) 110. [51] N.J. Harmon, M.E. Flatte, Phys. Rev. Lett. 108 (2012) 186602.