polyaniline composite for transparent hole transporting layer

polyaniline composite for transparent hole transporting layer

Journal of Industrial and Engineering Chemistry 65 (2018) 309–317 Contents lists available at ScienceDirect Journal of Industrial and Engineering Ch...

2MB Sizes 0 Downloads 22 Views

Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec

Graphene nanosheet/polyaniline composite for transparent hole transporting layer Olga D. Iakobsona , Oxana L. Gribkovaa , Alexey R. Tameeva,b,* , Alexander A. Nekrasova , Danila S. Saraninc , Aldo Di Carloc,d a A.N. Frumkin Institute of Physical Chemistry and Electrochemistry of the Russian Academy of Sciences, Leninsky Prosp. 31, bld. 4, Moscow 119071, Russian Federation b National Research University Higher School of Economics, Myasnitskaya Str. 20, Moscow 101000, Russian Federation c National University of Science and Technology “MISiS”, Leninsky prosp. 4, Moscow 119049, Russian Federation d C.H.O.S.E. (Centre for Hybrid and Organic Solar Energy), Department of Electronic Engineering, University of Rome “Tor Vergata”, Via del Politecnico 1, Rome 00133, Italy

A R T I C L E I N F O

Article history: Received 1 March 2018 Received in revised form 21 April 2018 Accepted 27 April 2018 Available online 18 May 2018 Keywords: Graphene Polyaniline Surface roughness Conductivity Photovoltaics

A B S T R A C T

Composites based on graphene and water-dispersible polyaniline-poly(2-acrylamido-2-methyl-1propanesulfonic acid) complex were shown as materials promising for the development of hole transporting layers (HTLs). By using multimodal atomic force microscopy, transmission electron microscopy, cyclic voltammetry and conductivity measurements, the relationship between the oxidation degree of graphene and the morphology, electrical conductivity and electron energy structure of HTLs was revealed. Utilizing the HTLs in organic solar cells it was shown that graphene nanostacks, rather than oxidized graphene nanostacks, enhanced the performance of the cells by increasing the roughness of the HTL surface, which caused the improvement of the short circuit current. © 2018 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

Introduction Graphene, a well-known 2D carbon nanomaterial with unique electrical, thermal, mechanical and surface properties, attracts the attention of many researchers especially as a component of various composites. Combined with organic (polymer) [1,2] and inorganic [3–5] materials it can be used in many fields, such as automotive and aerospace industries, electronics, packaging, water purification, biomedicine. A lot of efforts is devoted to the development of graphene-based polymer composites as charge transporting layers [1,2]. The development of hole transporting layer (HTL) used as an anode buffer layer for organic photovoltaic (OPV) cells is of special interest [6,7]. In particular, the feasibility of OPV devices fabricated using graphene of different oxidation degree as HTL was demonstrated in

* Corresponding author at: A.N. Frumkin Institute of Physical Chemistry and Electrochemistry of the Russian Academy of Sciences, Leninsky Prosp. 31, bld. 4, Moscow 119071, Russian Federation. E-mail addresses: [email protected] (O.D. Iakobson), [email protected] (O.L. Gribkova), [email protected], [email protected] (A.R. Tameev), [email protected] (A.A. Nekrasov), [email protected] (D.S. Saranin), [email protected] (A. Di Carlo).

[8,9]. Although HTL is believed to improve the performance of OPV cell through adjustment of the anode work function [10,11], there are several works implying the absence of obvious correlation between the work function change and performance of OPV devices [12–14]. On the contrary, increased surface roughness [15–18] and high electrical conductivity [10,15,19] of HTL are shown to be beneficial for the device performance. A correlation between the electrical conductivity and surface morphology is revealed to be of great importance [12,15]. Conducting polymers stabilized by water-soluble polyacids are considered as promising materials due to possibility to form their layers from environment-friendly solution. Recently we have reported the procedure for the preparation of conducting polyaniline (PANI) material that does not precipitate from its aqueous dispersion for up to 2 years [20]. The polymer material was created by performing PANI synthesis in the presence of poly(2acrylamido-2-methyl-1-propanesulfonic acid) resulting in the formation of PANI–PAMPSA complexes. The complexes possess good film-forming ability, along with temporal stability of optical and electrical properties of the layers. The introduction of graphene or graphene oxide into such PANI–polyacid complexes is of growing interest [21–24]. The use of aqueous dispersion of PANI–polyacid complex opens new ways for

https://doi.org/10.1016/j.jiec.2018.04.042 1226-086X/© 2018 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

310

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

technological easy-to-scale procedure of nanocomposite preparation by simple mixing graphene (graphene oxide) and PANI dispersions. Such an approach is impossible for common PANI because of its poor solubility in common solvents. However, the only result of the application of graphene oxide/PANI–polyacid complex composite as a counter electrode (anode) in dyesensitized solar cell is presented in Ref. [24]. Yet, the correlation between the nanocomposite structure and characteristics of a related device has not been studied elsewhere. The novelty of our approach is a simple design of nanocomposites based on graphene of a different oxidation degree and a PANI–PAMPSA complex. The variation of graphene oxidation degree enables us to obtain materials with tunable electronic structure and to find a proper combination between insoluble and infusible features of graphene and insulating nature of graphene oxide. The successful application of these water-dispersible graphene/PANI–PAMPSA nanocomposites as HTL in OPV cells based on a conventional blend of regioregular poly(3-hexylthiophene-2,5-diyl) (P3HT) and fullerene derivative PC70BM is demonstrated. The relationship between the oxidation degree of graphene and electrical, morphological properties of the nanocomposite HTL as well as parameters of OPV cells is revealed for the first time. Experimental Initial materials, nanocomposite preparation and characterization The chemical oxidative aniline polymerization was carried out in the presence of PAMPSA (Mw  2,000,000, 15 wt.% in H2O, Aldrich) with the formation of stable aqueous dispersion of PANI– PAMPSA (1:2 mol/mol) complex. The detailed description of the synthesis was published earlier [25]. Graphene, G, (dispersion in ethanol), graphene oxide with a low oxidation degree, GOx, (aqueous dispersion) or highly oxidized graphene, GO, (aqueous dispersion) were used for preparation of the nanocomposites with PANI–PAMPSA complexes. The detailed procedure of graphene preparation and characterization was given in Refs. [26,27]. Graphene/PANI nanocomposites were obtained by mixing graphene dispersion with dialyzed PANI–PAMPSA dispersion followed by 10 min of sonication. In all studied nanocomposites, the concentration of graphene (G, GO or GOx) in PANI–PAMPSA complex was 1 wt.%. Nanocomposite layers were prepared by drop-casting of graphene/PANI–PAMPSA dispersions on pre-cleaned substrates. The following substrates were used: glass (for conductivity and atomic force microscopy (AFM) measurements, optical spectra recording), FTO glass with the sheet resistance of ca. 7 V/square (for electrochemical investigations); ITO glass with the sheet resistance of ca. 15 V/square (for photovoltaic devices, conductive AFM measurements); highly oriented pyrolytic graphite (HOPG) plate (for Kelvin-probe studies). The HOPG substrates were purchased from NT-MDT. In the case of HR TEM the images were taken in a cross-section of free-standing films. The thickness of the nanocomposite layers was 50  60 nm (KLA-Tencor D-100 Profiler). Electron spectra of the layers were recorded using a Shimadzu UV-3101PC spectrophotometer. Electrochemical measurements The nanocomposite layers were treated in a 10% aqueous solution of CaCl2 for 5 min to prevent dissolution in aqueous electrolyte and ensure better adhesion to the substrate. Cyclic voltammetry (CVA) was performed in a 1 M aqueous HCl in the range of potentials of the first PANI oxidation/reduction stage

(leucoemeraldine–emeraldine transition) at different potential scan rates (10, 20, 50, 100 mV/s). The electrochemical parameters were controlled and measured by a computer-driven combination of HA-501G potentiostat/galvanostat (Hokuto Denko Ltd.) and a digital storage oscilloscope Nicolet 2090 (Nicolet Inc.). A platinum foil was applied as a counter electrode, saturated Ag/AgCl electrode – as a reference electrode. All values of the electrochemical potentials in this work are presented relatively to this electrode. Conductivity The DC-conductivity in longitudinal direction of a layer was measured by the four-probe technique as described earlier [28]. The conductivity measurement error did not exceed 5%. Photovoltaic characteristics OPV cells of the architecture ITO/HTL/P3HT:PC70BM/LiF/Al in which P3HT:PC70BM form a bulk heterojunction (BHJ) were prepared. A HTL nanocomposite layer (thickness of 40 nm) was spin-coated on an ITO-glass substrate (15 V/sq., Kintec) and annealed 10 min at 70  E. For preparing the cells, onto the surface of the HTL a photoactive layer of the thickness of 200 nm was spin-coated at 900 rpm from a solution in chlorobenzene of a blend (1:0.8 wt.) of regioregular P3HT (4002-EE; Rieke Metals) and PC70BM (Sigma–Aldrich). The purchased materials were used as received without further purification. The samples were then transferred into a glovebox with argon atmosphere (MBraun MB200MOD) and annealed at 90  C for 5 min and at 140  C for another 5 min. A LiF cathode buffer layer (0.9 nm) and Al cathode (60 nm) were deposited sequentially on the P3HT:PC70BM layer by low vacuum (10 6 mbar) thermal evaporation in Edwards Evaporator Auto500. For each composition of HTL 20 OPV cells were tested inside the glovebox. J–V characteristics were recorded by a SMU Keithley 2400 at AM1.5 conditions provided by a solar simulator (Xe lamp 150 W Oriel Solar Simulator, Newport Corp.). Multimodal probe microscopy The surface topography of the nanocomposite layers was recorded using Enviroscope AFM with a Nanoscope V controller (all by Bruker) in a tapping mode. For the evaluation of nanoscale fluctuations of the work function (WF) along the surface of graphene/PANI nanocomposites Kelvinprobe studies were carried out. For this purpose, the same Enviroscope AFM with the Nanoscope V controller in an interleave mode was used. Titanium nitride coated tips NSG01/TiN (NT-MDT) with the curvature radius of 35 nm, the force constant of ca. 5.1 N/m and the resonant frequency of ca. 150 kHz were used. The samples were prepared as follows. One-half of the surface of a freshly cleaved HOPG was isolated with an adhesive tape. Graphene/PANI–PAMPSA film was applied on the second half of the HOPG surface by solution casting method. The film was dried for 24 h at a room temperature. The first stage of Kelvin-probe experiment was carried out on the film-coated part of the HOPG substrate and appropriate topographical and surface potential (F) distribution images were registered. Then the adhesive tape was removed from the sample accompanied by exfoliation of the upper sheet of HOPG producing a freshly cleaved HOPG surface, which (according to Ref. [29]) has a WF of 4.6 eV. This clean HOPG surface was used in the second stage of Kelvin-probe experiment. The averaged F value calculated from a cross-section on the potential distribution image for HOPG was then used as a reference to establish a correlation between F (in V) and energy (in eV) scales. From the fluctuations of F along a cross-section of Kelvinprobe image for the freshly cleaved HOPG surface one can estimate

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

the accuracy of WF determination, which in our case was equal to 0.05 eV. Conductive AFM (c-AFM) measurements were performed by Multimode V AFM with a Nanoscope IV controller (all by Bruker) in contact + interleave mode. Cantilevers HA_C/Pt (TipsNano) had two tips with curvature radius <10 nm, force constant of 0.65 and 0.26 N/m, respectively, and resonant frequencies of 37 kHz and 19 kHz, respectively. The AFM image sizes were 5  5 mm and 256  256 points resolution. The topography and current images were recorded simultaneously with 1 V bias and at a scanning rate of 1 Hz. The current sensitivity was 10 nA/V. The load force was maintained below 20 nN to avoid damages of the tip and samples. All of the c-AFM experiments were carried out at ambient conditions. High-resolution transmission electron microscopy The distribution of graphene in graphene/PANI–PAMPSA layer was analyzed by high-resolution transmission electron microscopy (HR TEM) using a Cs-corrected JEM2100 F transmission electron microscope operating at 200 kV. Results and discussion Electrical conductivity and correlated features The introduction of graphene or graphene oxide in the nanocomposites can regulate the nanocomposite properties. It is the graphene oxidation degree that defines stacking ability, hydrophilic–hydrophobic properties of this 2D nanostructure and, therefore, its interactions with other components of the nanocomposite [30]. In Table 1, it is clearly seen that electrical properties of the graphene/PANI–PAMPSA nanocomposites depend on the graphene oxidation degree. Graphene with low oxidation degree (GOx) is found to be the most effective for the conductivity improvement (up to 20 times compared to initial PANI–PAMPSA complex). Optical spectra of the nanocomposite layers demonstrate that their optical transparency in the visible-NIR spectral range is above 80% (Fig. 1). The influence of graphene oxidation degree on the optical properties of the graphene/PANI–PAMPSA complex nanocomposites is also observed clearly. Moreover, the correlation between electrical conductivity increase and the red shift of the localized polaron absorption band (which maximum is positioned at around 750 nm [31]) of the nanocomposite layers is detected (Table 1 and Insert in Fig. 1). Compared to the pristine PANI–PAMPSA complex, the red shift of the localized polaron absorption band is pronounced in all the nanocomposite layers. The largest red shift of 15 nm is noticed for PANI–PAMPSA composite with GOx. We attribute this red shift to the increase of electron delocalization. Indeed, the results of the CVA investigation of the graphene/ PANI–PAMPSA nanocomposite layers (Table 2, SI, Figs. 1S and 2S) support the idea. First, comparing the averaged half-wave potentials, E1/2, (Table 2) for 20 and 10 mV/s (in this range they are practically

Table 1 Longitudinal electrical conductivity of graphene/PANI–PAMPSA (1 wt. %) nanocomposite layers. Additive to PANI–PAMPSA

Conductivity, S/cm

– G GOx GO

1.5  10 3.0  10 2.5  10 2.0  10

2 2 1 2

311

independent on the scan rate (w)) one can see that these values for the nanocomposites are more cathodic compared to the initial PANI–PAMPSA complex. Taking into account that Fermi level approximately corresponds to the half-wave potential position on the energy scale (eV), one can conclude that energy levels slightly depend on the oxidation degree of graphene in the nanocomposites. Since PANI is a p-type semiconductor and its Fermi level lays close to the energy level of the highest occupied molecular orbital (HOMO), we consider that Fermi level for PANI is equal to the HOMO level. Using correlation between electrochemical and physical (energy) scales [32] the calculated HOMO values are presented in Table 2. These values are close to those determined by ultra-violet photoelectron spectroscopy [33]. Second, anodic and cathodic peaks are more pronounced for nanocomposite layers than that for pristine PANI–PAMPSA layer (SI, Fig. 1S). It is an evidence of more rapid layer relaxation at a potential change and decreased portion of capacitive component of the current. Linear dependence of anodic peak currents on the square root of scan rate (SI, Fig. 2Sa) in combination with the scanrate-dependent anodic-to-cathodic peak separation (DE) indicates a diffusion-controlled quasi-reversible electrochemical behavior of all the layers [34]. Moreover, one can see the reduction of DE for the nanocomposite layers compared to initial PANI–PAMPSA complex that is especially noticeable in the case of using G and GOx additives (Table 2). Interestingly, the scan rate (w) dependence of DE for GOx/PANI–PAMPSA is best fitted by a square root function (SI, Fig. 2Sb) as distinct from PANI–PAMPSA complex and other graphene/PANI–PAMPSA nanocomposites, which demonstrate linear dependences. Extrapolation of the square root fitting of the DE-w dependence for GOx/PANI–PAMPSA passes through the origin of coordinates while none of the linear DE-w dependences for PANI–PAMPSA, G/PANI–PAMPSA and GO/PANI–PAMPSA nanocomposites does. Thus, at low scan rates GOx/PANI–PAMPSA layer demonstrates more ideal electrochemical behavior [34]. This may be connected with the lesser dependence of DE on the dramatic change in PANI bulk conductivity (being a common feature of the redox processes in electroactive conducting polymers [35]), which in our case occurs during leucoemeraldine–emeraldine transition. Thus, both the spectroscopic and electrochemical data confirm that the introduction of graphene in PANI–PAMPSA complex increases the electron exchange between conductive PANI domains, therefore this result correlates well with the enhanced conductivity of the G/PANI–PAMPSA and GOx/PANI–PAMPSA nanocomposites. Atomic-force, probe and electron microscopy During the preparation of OPV cells of the conventional architecture, a photoactive layer is coated onto an anodic buffer layer (i.e. HTL), so the surface topography of the latter is of importance for the quality of the interface. As shown earlier [36], specific interactions between the nanocomposite components seem to be the reason of the influence of graphene oxidation degree on the electron delocalization, electron exchange and charge transport processes within the layers of nanocomposites based on graphene and PANI–polyacid complexes. They define the spatial distribution of graphene nanoobjects within the polymer matrix, the tendency of graphene to aggregate and the size of graphene stacks. This results in the dependence of the nanocomposite layer topography on the graphene oxidation degree, as demonstrated by the AFM images presented in Fig. 2. Naturally, according to the particle analysis (Bruker Nanoscope Analysis Software), the pristine PANI–PAMPSA layer has a regular surface with insignificant height fluctuations (mean height = 2.5

312

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

Fig. 1. Transmittance spectra of the graphene/PANI–PAMPSA nanocomposite layers.

Table 2 Anodic–cathodic peak separation (DE), half-wave potentials (E1/2) and corresponding energy level of the highest occupied molecular orbital (HOMO) for the nanocomposites of graphene with a different oxidation degree and PANI–PAMPSA. Sample

Scan rate, mV/s

DE, V

E1/2, V

PANI–PAMPSA

100 50 20 10 100 50 20 10 100 50 20 10 100 50 20 10

0.394 0.278 0.188 0.157 0.288 0.206 0.139 0.128 0.364 0.259 0.168 0.115 0.334 0.235 0.149 0.121

0.240 0.216 0.190 0.191 0.212 0.195 0.186 0.183 0.207 0.195 0.178 0.180 0.203 0.194 0.184 0.178

G/PANI–PAMPSA

GOx/PANI–PAMPSA

GO/PANI–PAMPSA

a

HOMOa , eV 4.87

4.86

4.86

4.86

The standard deviationis 0.0 eV.

nm with 1.1 nm dispersion (s)) (Fig. 2a). After introducing GOx or GO into the PANI–PAMPSA complex the surface remains comparatively regular, however, the mean height increase up to about 6.9 (s = 2.6) and 8.3 (s = 4.9) nm, correspondingly (Fig. 2b and c). The picture is completely different in the case of G/PANI–PAMPSA nanocomposite (Fig. 2d), which has the most irregular surface among the studied samples (the mean height exceeds 44 (s = 30.4) nm). The G/PANI–PAMPSA surface looks like PANI– PAMPSA surface with the segregated islands of nanostacked graphene, these sharp islands being distributed more or less uniformly in the PANI–PAMPSA matrix. Therefore, the relief of the nanocomposite layers is defined by the graphene used: the lower is the graphene oxidation degree the higher are the height fluctuations of the surface. The conclusions about nanocomposite

surface irregularity were also proved by us when calculating the exponents for surface profile gradient [28]. The arrangement of graphene within the graphene/PANI– PAMPSA layers and the tendency of graphene to aggregation were analyzed by HR TEM (Fig. 3). As seen from the HR TEM images, both G and GOx aggregate within the polymer layer (pointed by arrows on Fig. 3). The functional groups on the surface of GOx prevent formation of large stacks while in the absence of surface functionalization the G sheets tend to aggregate due to p–p-interaction between them. This conclusion is in good correlation with the above AFM results. The charge transport through HTL in OPV cells occurs in the transvers direction to the layer surface, so the c-AFM technique is suitable to probe spatial distribution of the charge transport pathways across HTLs on the nano- or molecular scale [37]. Moreover, using a c-AFM probe (the current flowing between the cantilever probe to the conductive substrate) one can establish a correlation between the local vertical electrical conductivity of the sample and the topography [38,39]. 2D and 3D views of typical current distributions in PANI–PAMPSA and G/PANI–PAMPSA layers are shown in Fig. 4. One can see relatively inhomogeneous current distribution through the PANI–PAMPSA layer (the highest current difference is 5 nA) (Fig. 4a and b). Such inhomogeneity is characteristic of PANI films and the size of conductive regions depends on PANI doping level [38–40], film thickness [39], method of film preparation and doping anion [38]. The distribution of current on the surface of G/PANI–PAMPSA layer is much more inhomogeneous compared to that on PANI–PAMPSA (Fig. 4c and d). C-AFM current image for the majority of G/PANI– PAMPSA surface demonstrates current distribution similar to that for PANI–PAMPSA, yet the localized regions of high current (15  90 nA) are clearly seen also. Generally, we observe the correspondence between topography (Fig. 2d) and current (Fig. 4c) images. As for GOx/ PANI–PAMPSA and GO/PANI–PAMPSA layers, no high current areas were observed (c-AFM images are not shown). Thus, in the G/PANI– PAMPSA nanocomposite layer, graphene nanostacks form vertical

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

313

Fig. 2. Topography AFM-images of PANI complex or nanocomposite layer surface: PANI–PAMPSA (a), GO/PANI–PAMPSA (b), GOx/PANI–PAMPSA (c), G/PANI–PAMPSA (d).

Fig. 3. HR TEM-images taken from G/PANI–PAMPSA (a) and GOx/PANI–PAMPSA layers (b).

pathways for the charge carrier transport, thereby improving the overall charge transport through the layer. Another method to estimate positions of the energy levels as well as submicron fluctuations of the work function along the layer surface is Kelvin-probe microscopy (KPM) [41]. In Fig. 5 one can see a KPM image of G/PANI–PAMPSA sample (a) and its cross-section in the area of graphene inclusions (b, curve 1). The cross-section of potential (Fig. 5b, curve 2) obtained from KPM image of freshly

cleaved HOPG (the KPM image is not shown) was used as a reference to establish a correlation between F and WF scales. One can see distinct peaks of potential in the areas of G inclusions with the amplitude of 0.67 V relative to the HOPG level (2.1 V). In terms of energy this is 4.6 eV + 0.67 eV = 5.27 eV. Potential in the areas of PANI–PAMPSA surface without graphene inclusions exceeds the HOPG level by 0.27 V which gives the WF value 4.6 eV + 0.27 eV = 4.87 eV. This result correlates well with the position of HOMO level

314

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

Fig. 4. C-AFM current images of layers: (a) PANI–PAMPSA; (c) G/PANI–PAMPSA and corresponding 3D images; (b) PANI–PAMPSA; (d) G/PANI–PAMPSA.

for PANI–PAMPSA layer determined from CV experiment (Table 2). So, G inclusions in G/PANI–PAMPSA layer have 0.4 eV higher WF (and, therefore, deeper laying energy level) than the surface of PANI–PAMPSA matrix, which may contribute to more effective hole injection from the composite layer. OPV cells and conductivity of nanocomposites As shown above, the developed graphene/PANI–PAMPSA nanocomposites completely meet the requirements for application as an HTL in OPV cells due to both the enhanced electrical conductivity and the vis–NIR optical transparency.

The performance data of OPV cells with the HTL made from the developed nanocomposites and PANI–PAMPSA (as a reference) are presented in Fig. 3S and Table 3. We observed different influence of the oxidation state of graphene on the short circuit current, Jsc, and, consequently, on the power conversion efficiency (PCE) (Table 3 and Fig. 3S). In general, there is a tendency of decreasing the short circuit current with increasing the oxidation degree of graphene (Table 3). Decreasing the short circuit current correlates with decreasing the intrinsic conductivity of the used graphene nanostructures in the order: G > GOx > GO, which caused by disrupting sp2 bonding networks with graphene oxidation [30]. This trend does not

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

315

Fig. 5. KPM image of G/PANI–PAMPSA layer (a) and the cross-sections of G/PANI–PAMPSA layer along the white line (1) (b) and freshly cleaved HOPG (2, the image is not shown). F-to-WF correlation was established by admitting HOPG WF to be 4.6 eV [29].

Table 3 Photovoltaic parameters of ITO/HTL/P3HT:PC70BM/LiF/Al cells under illumination with an intensity of 100 mW cm 2. Jsc, Voc, FF and PCE are short-circuit current density, open circuit voltage, fill factor and power conversion efficiency, respectively. HTL

Jsc, mA/cm2

Voc, V

FF

PCE, %

PANI–PAMPSA G/PANI–PAMPSA GOx/PANI–PAMPSA GO/PANI–PAMPSA

8.1  0.2 11.1  0.6 7.9  0.3 5.9  0.2

0.58  0.01 0.60  0.01 0.60  0.01 0.63  0.01

0.45  0.02 0.44  0.02 0.46  0.02 0.44  0.02

2.12  0.10 2.92  0.36 2.19  0.25 1.62  0.08

correlate with that of the influence of the oxidation degree of graphene on the conductivity of graphene/PANI–PAMPSA nanocomposites (Table 1). The high longitudinal conductivity of the GOx/ PANI–PAMPSA layer (Table 1) turns out to be insufficient to increase the short circuit current that flows across the HTL (Table 3). Only OPV cells with HTL of G/PANI–PAMPSA demonstrated the increased short circuit current and PCE when compared with those of graphene-free PANI–PAMPSA HTL. As the concentration of graphene (oxidized graphene) nanostacks is far below the percolation threshold, the longitudinal conductivity is defined generally by the charge transport over PANI macromolecules. On the contrary, as the size of the G nanostacks is comparable with the thickness of the HTLs studied (Figs. Figure 2d and Figure 5), the transverse conductivity can be increased due to the additional contribution of the charge transport over the graphene nanostacks. The latter is not the case of the GO nanostacks in the GO/PANI–PAMPSA layers. Furthermore, photochemical reduction of oxidized graphene can occur under the action of light with a photon energy of 3.2 eV [15,46]. The photoreduction of GO involves the detachment of the OH, CO, CO2 and O2 in electron excited states according to dissociation mechanism considered in Ref. [43] and Refs therein. The detached moieties and defects on the GO nanosheet [30] eventually act as traps for charge carriers in the HTL and thereby limit the short circuit current. Thus, we suggest the following two reasons of increasing the parameters of the OPV cells with G nanostacks in the HTL. One reason is associated with an increased surface roughness of the HTL. G nanostacks above PANI layer can work as series of nanopillars filled with a photoactive blend of P3HT:PC70BM. A surface of a homogeneous conductor is an equipotential surface of the electrostatic field so the

electric fields are strongest at locations along the surface where the conductor is most curved. In particular, on wedge-shaped edge and apex edge (where one of the radii of curvature tends to zero), the surface charge density becomes infinite, so that the field strength near such edges also tends to infinity as the surface is approached [44]. Sharp edges of the G nanostacks penetrated into the photoactive layer create the electric fields as well. In the photoactive layer, an electronhole pair (known as an exciton) generated upon light absorption is bound due to Coulombic attraction. The electric fields from the sharp edges of the G nanostacks facilitate the dissociation of excitons to electrons and holes, so the concentration of the free charge carriers in the photoactive layer increases. Moreover, the enhanced surface roughness (i) enlarges the interface area between the HTL and photoactive layer [16–18,45], and (ii) elongates optical path in photoactive layer by increasing light scattering [17,18,45,46]. Additional reason is associated with the intrinsic conductivity of the used graphene nanostructures. GO/PANI–PAMPSA demonstrated the worst OPV characteristics (Table 3) despite this HTL displayed the high surface roughness (Fig. 2c) and the reasonable longitudinal conductivity (Table 1). Thus, the influence of charge carrier traps on reducing the electrical conductivity along GO nanostacks prevails over the influence of the surface roughness on increasing the free charge carriers concentration. As for graphene G, in addition to the high conductivity of G nanostacks, the work function of G (5.27 eV) lays in the intermediate position (Fig. 5) between that of PANI (4.87 eV) and P3HT (5.4 eV) so the hole transfer from the P3HT:PC70BM layer to anode is facilitated. The only publication describing the use of graphene oxide composite with poly(styrenesulfonic acid)-graft-polyaniline (PSSAg-PANI) in BHJ OPV cell is Ref. [10]. When PSSA-g-PANI with 2.5 wt.% GO was used as HTL, the P3HT:PC61BM-based solar cell device exhibited a PCE by 10% higher than the device with pristine PANI-gPSSA. We achieved enhancing of the OPV device performance by 38% using HTL layer containing only 1 wt.% of graphene. It should be especially emphasized that the polyaniline grafting procedure is far more complicated compared to our procedure. Conclusions We developed a facile route for the formation of water-dispersible polymer nanocomposites based on graphene nanosheets and

316

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317

PANI–PAMPSA complex. Using graphene of different oxidation degree enabled us to obtain a set of materials with such tunable properties as the electron exchange rate between PANI macromolecules, electrical conductivity, and surface roughness of the layer. For the first time, the relationship between the oxidation degree of graphene and the morphology, electrochemical properties and electrical conductivity of the nanocomposite thin layers was revealed. We also developed the conductive thin layers of the nanocomposites with the transparency of 80% which are promising for organic electronics. Namely, G/PANI–PAMPSA nanocomposite used as an HTL in OPV cells increased the short circuit current by 40% and 80% in comparison with that for OPV cells with HTLs based on graphene free PANI–PAMPSA complex and GO/PANI–PAMPSA nanocomposite, respectively. The improvement of the short circuit current was caused by the G nanostacks penetrated into the BHJ photoactive layer. The sharp edges of the G nanostacks serve as sources of the electric fields which can facilitate the exciton dissociation to free electrons and holes at the interface between the G nanostacks and the photoactive layer. On the contrary, the GO nanostacks are located within the PANI– PAMPSA bulk and serve as deep traps for holes decreasing the short circuit current dramatically. In parallel with the increase of the charge carrier concentration, the G nanostacks provide additional pathways for the charge transport across the nanocomposite HTL. Moreover, the G nanostacks possess the larger work function with respect to that of PANI–PAMPSA and consequently facilitate the hole transfer from the photoactive layer to anode. Thus, tuning the surface roughness of the G nanocomposite HTL along and matching electron energy levels at the interface between the functional layers rather than increasing the electrical conductivity of G nanocomposite HTLs are important for the improvement of the OPV cell operation. Acknowledgements The work was supported by the Russian Foundation for Basic Research, the grant No. 16-29-06423 (the study of structure and physico-chemical properties of 2D nanocomposites). A.R. Tameev is indebted to the Basic Research Program of the National Research University Higher School of Economics (Moscow) for the support. D.S. Saranin and A. Di Carlo acknowledge the financial support of the Ministry of Education and Science of the Russian Federation in the framework of Increase Competitiveness Program of NUST «MISiS» (No. K2-2017-25), implemented by a governmental decree dated 16th of March 2013, N 211. The authors thank Dr. O.Yu. Posudievsky (Pisarzhevskii Institute of Physical Chemistry of the National Academy of Sciences, Ukraine) for providing themwith graphene of different oxidation degree, Dr. A. V. Egorov (M.V. Lomonosov Moscow State University, Russia) for performing and discussion HR TEM, Dr. V.I. Zolotarevsky (Frumkin Institute of Physical Chemistry and Electrochemistry, Russia) for the help in AFM-, Kelvin- and conductive AFM-measurements. The scanning probe microscopy and Vis–NIR spectroscopy measurements were performed using the equipment of CKP FMI at the Frumkin Institute of Physical Chemistry and Electrochemistry. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.jiec.2018.04.042. References [1] H. Kim, A.A. Abdala, C.W. Macosko, Macromolecules 43 (2010) 6515, doi:http:// dx.doi.org/10.1021/ma100572e. [2] G. Mittal, V. Dhand, K.Y. Rhee, S.-J. Park, W.R. Lee, J. Ind. Eng. Chem. 21 (2015) 11, doi:http://dx.doi.org/10.1016/j.jiec.2014.03.022.

[3] X. Zang, W. Wang, C. Liu, X. Tang, Sens. Actuators B: Chem. 252 (2017) 1179, doi: http://dx.doi.org/10.1016/j.snb.2017.07.144. [4] J. Wei, Z. Zang, Y. Zhang, M. Wang, J. Du, X. Tang, Opt. Lett. 42 (2017) 911, doi: http://dx.doi.org/10.1364/OL.42.000911. [5] H. Huang, J. Zhang, L. Jiang, Z. Zang, J. Alloys Compd. 718 (2017) 112, doi:http:// dx.doi.org/10.1016/j.jallcom.2017.05.132. [6] R. Po, C. Carbonera, A. Bernardi, N. Camaioni, Energy Environ. Sci. 4 (2011) 285, doi:http://dx.doi.org/10.1039/C0EE00273A. [7] C. Huang, H. Wu, Y. Cao, alcohol-soluble conjugated polymer-based interlayers for, Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2014, pp. 301, doi:http://dx.doi.org/10.1002/9783527656912.ch09. [8] R. Wu, Y. Wang, L. Chen, L. Huang, Y. Chen, RSC Adv. 5 (2015) 49182, doi:http:// dx.doi.org/10.1039/C5RA02099A. [9] J. Liu, M. Durstock, L. Dai, Energy Environ. Sci. 7 (2014) 1297, doi:http://dx.doi. org/10.1039/C3EE42963F. [10] S. Bae, J.U. Lee, H. Park, E.H. Jung, J.W. Jung, W.H. Jo, Sol. Energy Mater. Sol. Cells 130 (2014) 599, doi:http://dx.doi.org/10.1016/j.solmat.2014.08.006. [11] A. Zampetti, A.H. Fallahpour, M. Dianetti, L. Salamandra, F. Santoni, A. Gagliardi, M. Auf Der Maur, F. Brunetti, A. Reale, T.M. Brown, A. Di Carlo, J. Polym. Sci. B: Polym. Phys. 53 (2015) 690, doi:http://dx.doi.org/10.1002/ polb.23685. [12] J. Huang, P.F. Miller, J.S. Wilson, A.J. De Mello, J.C. De Mello, D.D.C. Bradley, Adv. Funct. Mater. 15 (2005) 290, doi:http://dx.doi.org/10.1002/adfm.200400073. [13] Y. Kim, A.M. Ballantyne, J. Nelson, D.D.C. Bradley, Org. Electron. Phys. Mater. Appl. 10 (2009) 205, doi:http://dx.doi.org/10.1016/j.orgel.2008.10.003. [14] F.L. Zhang, A. Gadisa, O. Inganäs, M. Svensson, M.R. Andersson, Appl. Phys. Lett. 84 (2004) 3906, doi:http://dx.doi.org/10.1063/1.1739279. [15] B. Kadem, W. Cranton, A. Hassan, Org. Electron. 24 (2015) 73, doi:http://dx.doi. org/10.1016/j.orgel.2015.05.019. [16] S. Woo, J. Jeong, H. Lyu, Y. Han, Y. Kim, Nanoscale Res. Lett. 9 (2014) 506, doi: http://dx.doi.org/10.1186/1556-276X-9-506. [17] X. Wang, S. Hu, Q. Li, F. Li, K. Yao, M. Shi, RSC Adv. 5 (2015) 52874, doi:http://dx. doi.org/10.1039/C5RA07362F. [18] T. Xiao, W. Cui, J. Anderegg, J. Shinar, R. Shinar, Org. Electron. Phys. Mater. Appl. 12 (2011) 257, doi:http://dx.doi.org/10.1016/j.orgel.2010.11.008. [19] Z. Hu, J. Zhang, Z. Hao, Y. Zhao, Sol. Energy Mater. Sol. Cells 95 (2011) 2763, doi: http://dx.doi.org/10.1016/j.solmat.2011.04.040. [20] O.D. Iakobson, O.L. Gribkova, A.A. Nekrasov, V.A. Tverskoi, V.F. Ivanov, P.V. Mel’nikov, E.A. Polenov, A.V. Vannikov, Prot. Met. Phys. Chem. Surf. 52 (2016) 1005, doi:http://dx.doi.org/10.1134/S207020511606023X. [21] C. Basavaraja, G.T. Noh, D.S. Huh, Colloid Polym. Sci. 291 (2013) 2755, doi: http://dx.doi.org/10.1007/s00396-013-3016-8. [22] J. Luo, S. Jiang, Y. Wu, M. Chen, X. Liu, J. Polym. Sci. A: Polym. Chem. 50 (2012) 4888, doi:http://dx.doi.org/10.1002/pola.26316. [23] J. Luo, S. Jiang, R. Liu, Y. Zhang, X. Liu, Electrochim. Acta 96 (2013) 103, doi: http://dx.doi.org/10.1016/j.electacta.2013.02.072. [24] L. Wan, B. Wang, S. Wang, X. Wang, Z. Guo, H. Xiong, B. Dong, L. Zhao, H. Lu, Z. Xu, X. Zhang, T. Li, W. Zhou, React. Funct. Polym. 79 (2014) 47, doi:http://dx.doi. org/10.1016/j.reactfunctpolym.2014.03.012. [25] O.L. Gribkova, A.A. Nekrasov, M. Trchova, V.F. Ivanov, V.I. Sazikov, A.B. Razova, V.A. Tverskoy, A.V. Vannikov, Polymer (Guildf) 52 (2011) 2474, doi:http://dx. doi.org/10.1016/j.polymer.2011.04.003. [26] O.Y. Posudievsky, O.A. Khazieieva, V.G. Koshechko, V.D. Pokhodenko, J. Mater. Chem. 22 (2012) 12465, doi:http://dx.doi.org/10.1039/c2jm16073k. [27] O.Y. Posudievsky, O.A. Khazieieva, V.V. Cherepanov, V.G. Koshechko, V.D. Pokhodenko, J. Nanoparticle Res. 15 (1–9) (2013) 2046, doi:http://dx.doi.org/ 10.1007/s11051-013-2046-y. [28] O.D. Omelchenko, O.L. Gribkova, A.R. Tameev, S.V. Novikov, A.V. Vannikov, Prot. Met. Phys. Chem. Surf. 50 (2014) 613, doi:http://dx.doi.org/10.1134/ S207020511405013X. [29] M.M. Beerbom, B. Lägel, A.J. Cascio, B.V. Doran, R. Schlaf, J. Electron Spectros. Relat. Phenomena 152 (2006) 12, doi:http://dx.doi.org/10.1016/j.elspec.2006.02.001. [30] D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, Chem. Soc. Rev. 39 (2010) 228, doi:http://dx.doi.org/10.1039/B917103G. [31] A.G. MacDiarmid, A.J. Epstein, Synth. Met. 69 (1995) 85, doi:http://dx.doi.org/ 10.1016/0379-6779(94)02374-8. [32] S. Trasatti, Pure Appl. Chem. 58 (1986) 955, doi:http://dx.doi.org/10.1351/ pac198658070955. [33] C. Tengstedt, A. Crispin, C. Hsu, C. Zhang, I. Parker, W. Salaneck, M. Fahlman, Org. Electron. 6 (2005) 21, doi:http://dx.doi.org/10.1016/j.orgel.2005.02.001. [34] G. Inzelt, Conducting Polymers: A New Era of Electrochemistry, second ed., Springer Berlin Heidelberg, Berlin, Heidelberg, 2012, doi:http://dx.doi.org/ 10.1007/978-3-642-27621-7. [35] M.A. Vorotyntsev, J. Heinze, Electrochim. Acta 46 (2001) 3309, doi:http://dx. doi.org/10.1016/S0013-4686(01)00623-5. [36] O.L. Iakobson, A.R. Tameev, V.V. Kravchenko, A.V. Egorov, A.V. Vannikov, Synth. Met. 211 (2016) 89, doi:http://dx.doi.org/10.1016/j.synthmet.2015.11.018. [37] J.M. Mativetsky, Y.-L. Loo, P. Samorì, J. Mater. Chem. C 2 (2014) 3118, doi:http:// dx.doi.org/10.1039/C3TC32050B. [38] D.-H. Han, S.-M. Park, J. Phys. Chem. B 108 (2004) 13921, doi:http://dx.doi.org/ 10.1021/jp0494279. [39] C.-G. Wu, S.-S. Chang, J. Phys. Chem. B 109 (2005) 825, doi:http://dx.doi.org/ 10.1021/jp046259b. [40] G.B. Meshkov, V.F. Ivanov, I.V. Yaminsky, Polym. Sci. Ser. B 47 (2005) 327. [41] W. Melitz, J. Shen, S. Lee, J.S. Lee, A.C. Kummel, R. Droopad, E.T. Yu, J. Appl. Phys. 108 (2010) 23711, doi:http://dx.doi.org/10.1063/1.3462440.

O.D. Iakobson et al. / Journal of Industrial and Engineering Chemistry 65 (2018) 309–317 [43] V.A. Smirnov, N.N. Denisov, V.G. Plotnikov, M.V. Alfimov, High Energy Chem. 50 (2016) 51, doi:http://dx.doi.org/10.1134/S0018143916010070. [44] L.D. Landau, E.M. Lifshitz, Electrostatics of conductors, Electrodynamics of Continuous Media, Volume 8 in Course of Theoretical Physics, Elsevier, 1984, pp. 1, doi:http://dx.doi.org/10.1016/B978-0-08-030275-1.50007-2.

317

[45] L. Müller-Meskamp, Y.H. Kim, T. Roch, S. Hofmann, R. Scholz, S. Eckardt, K. Leo, A.F. Lasagni, Adv. Mater. 24 (2012) 906, doi:http://dx.doi.org/10.1002/adma. 201104331. [46] R. Geethu, C. Sudha Kartha, K.P. Vijayakumar, Sol. Energy 120 (2015) 65, doi: http://dx.doi.org/10.1016/j.solener.2015.06.037.