Assessment of the Potential of Polymer/HDI-GO Nanocomposites for Use in Organic Solar-Cells †

: Conducting films comprising conducting polymers and carbon nanomaterials have gained a lot of interest for applications in several fields, including transparent electrodes, superca-pacitors, light-emitting diodes (LEDs), polymer solar cells (PSCs), etc. One of the main motivations is the replacement of costly oxides and degradable materials, like indium tin oxide (ITO). On the other hand, graphene oxide (GO) has emerged as an ideal filler to reinforce polymeric matrices owing to its large specific surface area, transparency, flexibility, and very high mechanical strength. Nonetheless, functionalization is required to improve its solubility in common solvents and expand its practical uses. In this work, the potential of polymer nanocomposites based on hexamethylene diisocyanate (HDI)-functionalized GO (HDI-GO)/poly(3,4-ethylenedioxythiophene):poly(styrenesul-fonate)(PEDOT:PSS) for use as active layers (ALs) or interfacial layers (IFLs) in PSCs has been assessed. Conventional deposition techniques applied to thin films were tested for the developed nanocomposites. Deposition methods included drop and spin casting, where different type of substrates, as clean glass and glass/ITO were tested. The results of deposition essays were analyzed by atomic force microscopy (AFM) and UV-Vis spectroscopy. In addition, thermal evaporation was tried with the aim of obtaining homogeneous layers. The layers obtained by drop casting showed poor film quality, with large aggregates. On the other hand, spin coating lead to layers not fully wetting the substrate. New synthesis procedures for the nanocomposites and/or alternative treatments of the substrate surface will be investigated in the future to optimize their composition and properties (i.e., transparency) and improve their suitability for use in PSCs.


Introduction
Conducting polymers are materials with a wide range of applications, such as lightemitting diodes (LEDs) [1], supercapacitors [2], transparent electrodes [3], polymer solar cells [4][5][6][7][8][9][10], etc.The combination of conducting polymers with graphene-based materials has also proved to be suitable for specific applications in some fields [11].In polymer or organic solar cells, the typical structure involves costly oxides and degradable materials, like indium tin oxide (ITO) as a transparent electrode, but the recent tendency is to avoid the use of ITO due to indium scarcity [12].Graphene oxide (GO) has become an ideal filler to enhance the properties of polymeric matrices in different layers and different kind of cells [13,14], but its poor solubility in common solvents requires improvements using functionalization methods.Previous works presented different approaches for GO functionalization and its possible applications [15][16][17][18][19][20][21][22].In this work, the potential of nanocomposites based on the mixtures of hexamethylene diisocyanate (HDI)-functionalized GO (HDI-GO)/poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate)(PEDOT:PSS) for use in organic solar cells, as an alternative to ITO or as an interfacial layer to assist on charge transport at the interface with ITO (as a secondary hole transport layer, HTL), has been assessed.The chemical structure of HDI-GO is shown in Scheme 1.The most typical deposition techniques employed in the preparation of organic solar cells, the characterization of the samples using UV-Visible spectroscopy and atomic force microscopy (AFM) measurements, as well as the evaporation of samples to obtain homogeneous thin layers are described in this paper.

Synthesis of GO and HDI-GO
The preparation of GO was carried out using Hummers' method from graphite powder as reported in previous works [19].Firstly, graphite powder was heated with P2O5 and K2S2O8, and later deionized water was added and the product was filtered and then mixed with KMnO4, H2O2, and H2SO4.The final product was purified by centrifugation, followed by several cycles of purification with H2O2/H2SO4 washing, bath ultrasonication, and finally washed with deionized water followed by vacuum-drying.
The synthesis of HDI-GO was achieved by applying the procedure described in previous works [19], which can be summarized as: (1) GO functionalization step using HDI as reagent and TEA as catalyst, added dropwise at 60 °C under Ar atmosphere, followed by stirring overnight (12 h). ( 2) preparation of HDI-GO samples with different functionalization degree (FD) to study their effect on the deposition process: (a) 1:1:1 (FD of 3.12%, named as HDI-GO4), and (b) 1:1:1 with a previous ultrasonic tip treatment (5 min) (FD of 17.20%, named as HDI-GO5).

Deposition of HDI-GO and HDI-GO/PEDOT:PSS Films
HDI-GO samples were dispersed in different solvents (N-Methyl-2-pyrrolidone, NMP, 2-propanol and dimethyl sulfoxide, DMSO) at a concentration of 5 wt.% and the dispersions were stirred to achieve a good dispersion; this fact is very important since photovoltaic device preparation requires very homogeneous dispersions to avoid imperfections in the layer surface.Then, the dispersions were deposited by two different methods: drop casting and spin coating.The samples were deposited onto glass substrates, except some spin coating essays that employed glass/ITO substrates.
The drop casting method consists in the dropwise addition onto the substrate surface until the whole surface is covered.In our samples, a thermal annealing step was applied to dry the samples.
Spin coating deposition involves the use of a speed-regulated rotative support where the substrate to be covered is placed.The substrate is hold using a vacuum pump, and subsequently, the solution is added and the rotation of the substrate is started.

Instrumentation
The thermal annealing of drop casted and spin coated layers was carried out with a conventional heating plate.The spin coating was performed using a spin coater.UV-Visible spectra were recorded using a Cecil 2700 spectrophotometer.The AFM measurements were obtained using a Nano-Observer-Model 5100 microscope.

Layer Deposition
The dispersions obtained presented the tendency to get aggregated after a small time of no use, so a previous sonication cycle is recommended before the employment of solution if they are not prepared at the moment of the essay.
The HDI-GO/PEDOT:PSS samples obtained by drop casting onto glass resulted in little transparent layers to be employed either as HTL before the ITO layer or in between ITO and the photoactive layer [23,24], considering DMSO and NMP solvents.Furthermore, a clear phase separation between HDI-GO and PEDOT:PSS was observed, which suggests a very heterogeneous layer deposition.These two facts, opacity and heterogeneity, are not compatible with the requirements for HTL layers, i.e., transparency in the visible spectral region and homogeneous deposition.Regarding HDI-GO4 and HDI-GO5, the samples showed higher transparency in the visible than the samples with PEDOT:PSS, although some agglomerates were present.Further, the substrates' surfaces were not covered in all their extension.The results of drop casting deposition (Figure 1) suggest that the high hydrophobic nature of the nanomaterials, and the strong Van der Wall forces between flakes, hinders a homogeneous deposition and forms agglomerates.
The second approach was the spin coating deposition of HDI-GO5 in DMSO onto glass.The deposition of each sample follows different preparation conditions.The values of each deposition variable considered are collected in Table 1.In comparison with the drop casting reference, the spin coating method produces thinner and more transparent layers than drop casting.However, the samples do not show a clear layer deposition on to the surface, and small aggregates can be observed in all samples.The size of these aggregates seems to be reduced with increasing coating speed, which can be explained considering that it provides less time to the samples to get attached to the surface.Furthermore, it was found that surface plasma treatment had no effect on the morphology of samples deposited with the same speed.

UV-Visible Spectra Analysis
The UV-Visible spectra were recorded for the spin coated samples and the drop casting reference.The other drop casted layers were not considered due to their opacity.The spectra of the samples studied are shown in Figure 2.
There is a clear difference between the transmittance of the drop casting sample (GOdro) and the spin coating samples.This fact is consistent with the visual observations.The spin coated samples do not show significant changes in their transmittance values, neither comparing the treatment with plasma (cp vs. sp) or the coating speed (1000 rpm vs. 1800 rpm).These results point out that the coating speed or plasma treatment have no direct correlation with the transmittance values, and the differences found among them could be due to scattering or experimental errors in the measurements.The effect of different solvents was also studied.Three samples using DMSO without plasma treatment, 2-propanol with plasma treatment, and 2-propanol without plasma treatment were prepared.The substrate, in this case, was ITO onto glass.The results reveal no noticeable changes between DMSO and 2-propanol transmittance, and neither for both of the 2-propanol samples, hence it was concluded that a possible solvent effect was negligible with these nanocomposites.
The ultrasonication effect on HDI-GO deposition was also investigated by UV-Vis spectra.A set of samples were prepared according to two different sonication cycles: (a) 1 h of sonication, followed by a rest of 12 h and another 2 h of sonication; and (b) 2 h of sonication.Both sonication cycles were applied for HDI-GO4 and HDI-GO5 nanocomposites, with and without plasma treatment.In the case of HDI-GO4, no changes between samples with different sonication cycles or plasma treatments were found.However, in the case of HDI-GO5, the sample without plasma treatment presents a slightly better transmittance in the range 300-570 nm, but also shows less scattering when the sonication cycle was a).This result can be related with the higher FD value of the sample comparing with HDI-GO4: the higher FD, the higher the HDI-GO remnant at the hydrophobic surface.

AFM Results
The AFM technique was applied to study HDI-GO5 samples, since UV-Visible spectra results showed a slightly better deposition than HDI-GO4.DMSO and 2-propanol were selected as solvents, and typical results are shown in Figures 3 and 4.  In both figures, the phase images do not show distinct regions and the topography images resemble those of bare ITO surface, which means that HDI-GO5 was poorly or not even deposited onto the ITO surface.No conclusive results could be obtained from the AFM images, given that the layers were too heterogeneous to be properly characterized.

Thermal Evaporation
The thermal evaporation was applied to obtain a homogeneous thin film of HDI-GO5, with the aim of avoiding the problem of processing it as a dispersion.The evaporation was carried out under high vacuum-atmosphere (2 × 10 −6 mbar).However, the sample did not evaporate under the tested conditions.This can be due to the high thermal stability of HDI-GO5 (higher than 600 °C under air [25]).

Conclusions
In this work, the potential of HDI-GO based nanomaterials to form transparent and homogeneous layers on top of glass and glass/ITO substrates was studied considering the most typical deposition techniques employed for organic photovoltaic solar-cells.Samples prepared via spin coating technique also showed heterogeneities, but only small irregular depositions were detected, however, the loading of HDI-GO on the surface was very low.These visual observations were corroborated by UV-Visible spectra.A plausible solution for these issues could be to modify the synthesis procedure of the nanocomposites in order to achieve transparent and homogeneous layers.The effect of solvents, surface plasma treatment and ultrasonication cycle were also studied, and the results were almost the same considering the different conditions investigated.Overall, it can be concluded that the synthesis procedure of these promising nanomaterials should be tailored and optimized, which will be the task of future works.

Table 1 .
Experimental conditions for spin coating essays.

Figure 2 .
Figure 2. UV-Visible spectra of samples: the abbreviation "sc" means spin coating sample; the abbreviation "drop" denotes drop casting sample; the numbers indicate the coating speed; the abbreviation "cp" denotes surface plasma treatment, and "sc" no surface plasma treatment.