Arch ī um Ateneo Arch um Ateneo Polyaniline In Situ Grafted to Graphene Sheets Polyaniline In Situ Grafted to Graphene Sheets

Graphene is one of the most promising allotropes of carbon with wide applications in nanotechnology. Modification of graphene by chemical means can further expand its uses. Here, liquid-exfoliated graphene is functionalized with p -aminophenyl moiety using p nitrophenyl diazonium salts which were diazotized in situ then reduced by tin (II) chloride. The aminophenyl-graphene thus produced is further modified to form polyaniline-graphene (PANI-GNH) by in situ oxidative graft polymerization of aniline using ammonium peroxydisulfate as oxidizing agent. The properties of the PANI-GNH were compared with polyaniline/graphene blends by Raman, infrared and UV-Visible spectroscopy, and cyclic voltammetry. Indeed, PANI-GNH registered different spectrochemical and electrochemical properties compared with the physically blended PANI and GNH, a manifestation of the effect of chemical grafting on the overall property of the modified graphene.


INTRODUCTION
Graphene is a two-dimensional form of sp 2hybridized carbon with atoms arranged in a honeycomb lattice. It is the basic building block for graphitic materials of all other dimensionalities (Wei et al., 2009;Wehling et al., 2008). It has high surface area, tensile strength, and thermal and electrical conductivity. It is flexible, transparent and has low coefficient of thermal expansion. It is a zero-band gap semiconductor and is chemically inert. It is considered as one of the most promising materials for post-silicon electronics because it combines high electron mobility with atomic thickness (Wei et al., 2009;Wehling et al., 2008;Lu et al., 2009;Liu et al., Graphene, 2008;Li et al., Synthesis, 2012;Kumar et al., Polyaniline, 2012;Zhao et al., Effect, 2012;Li and Shi).
Polyaniline is an intrinsically conducting polymer (ICP) known for its facile synthesis, stability and reversible acid/base-doping/ dedoping (Li et al., Synthesis, 2012;Wu et al., Supercapacitors, 2010;Bo et al., 2011;Reddy et al., 2009;Bhadra et al., 2009;Lai et al., 2012). It has a fairly good processability and conductivity, and from an economic point of view, it is significantly superior to other ICPs because its monomer, aniline, is less expensive than other monomers used for other ICPs. Its synthesis is very simple and its properties can be tuned easily (Kumar et al., Polyaniline, 2012;Bhadra et al., 2009).
Covalently-linked graphene/ polyaniline composites have been synthesized recently through in situ graft polymerization of aniline that has been immobilized to graphene oxide through its functional groups (Kumar, et al., Polyaniline, 2012;Sahoo et al., 2012). However, the use of graphene oxide diminishes the potential of the composite, particularly its electronic properties vis-à-vis composites with pristine graphene.
In this study, graphene is functionalized with polyaniline by in situ graft polymerization of aniline in the presence of p-aminophenylfunctionalized graphene using ammonium peroxydisulfate as oxidizing agent. Covalentlylinked graphene/polyaniline was characterized and compared with graphene, polyaniline, and their physical blend using Raman, infrared and UV-Visible spectroscopy, and cyclic voltammetry.
Commercially available glass slides were used as substrates for Raman analysis.

Apparatus. Centrifugation was done using
Hettich Zentrifugen Mikro 220 and Sorvall Legend RT Centrifuge. Sonication was done using GE 130 Ultrasonic Processor (Thomas Scientific) and Cole-Parmer 8891 Ultrasonic bath.

Synthesis of Graphene.
Graphene was synthesized from graphite powder using microabrasive cleavage from a modified procedure of Hamilton et al. (2009). A 100-mg (~8.3 mmol) graphite powder was sonicated for 1 hour in 100 mL N,N-dimethylformamide (DMF), centrifuged at 4,600 rpm for 30 min, and the resulting supernatant was collected. A 240-mL graphene dispersion in DMF was filtered over a 0.45 μm nylon membrane filter. The membrane filter was briefly sonicated in 30 mL DMF to dislodge graphene from it.
Functionalization of p-Aminophenyl (Aniline) Moiety in Graphene. Diazonium functionalization was conducted by filtering 240 mL DMF graphene and then redispersing it in 30 mL DMF as a pre-concentration step to reduce the amount of solvent used for the reaction. The 240 mL DMF was used to exfoliate more graphene; this amount of DMF would be wasted if the reaction was conducted without pre-concentration. In addition, the functionalization of p-aminophenyl (aniline) moiety was conducted in two steps, namely: diazonium functionalization of p-nitrophenyl moiety from p-nitroaniline and then reduction of aryl nitro to aryl amino groups because p-phenylene-diamine (PPD), the reduced form of p-nitroaniline, cannot be used as both amino groups would have been available for diazotization.
The summary of steps is shown in a scheme in Figure 1. The residue from filtering 240 mL DMF graphene was re-sonicated in 30 mL DMF for 10 min at 40% amplitude. 0.26 g (1.92 mmol) p-nitroaniline and 0.160 mL (1.92 mmol) HCl was added and the mixture was purged with nitrogen gas for 10 min and pre-heated to 70 °C. 0.13 g (1.92 mmol) sodium nitrite was added and the temperature was maintained for 1 h with constant agitation while protected from light. Water was added and the residue was collected by centrifugation at 4,600 rpm for 20 min. The residue was washed with water and acetone until the supernatant was clear.

Figure 1. Functionalization of p-Aminophenyl (Aniline) Moiety in Graphene (represented by the dark line).
Reduction was done using a procedure adopted from Bellamy et al. (1984). p-Nitrophenyl-functionalized graphene (GNO) was sonicated in 30 mL ethanol for 10 min at 40% amplitude. 2.16 g (9.60 mmol) tin(II) chloride dihydrate was added and the mixture was heated to 70 °C. The temperature was maintained for 3 h with constant agitation. The residue was collected by centrifugation at 4,600 rpm for 20 min and it was washed three times with ethanol and also with water and finally acetone.

In Situ Oxidative Graft Polymerization of Aniline.
Here, p-Aminophenyl-functionalized graphene (GNH) was reacted in situ by oxidative graft polymerization of aniline using ammonium peroxydisulfate (APS) as oxidizing agent as summarized in Figure 2.
p-Aminophenyl-functionalized graphene (GNH) was sonicated in 15 mL water for 10 min at 40% amplitude. 0.75 mL (8 mmol) hydrochloric acid was added to 0.75 g (8 mmol) aniline and dried under vacuum with slight heating. The resulting precipitate was added to the GNH dispersion and bubbled with nitrogen gas for at least 10 min in a septum-capped vessel. 2.28 g (10 mmol) ammonium peroxydisulfate was dissolved in 15 mL water and slowly introduced to the GNH/aniline mixture using a syringe while continuously bubbled with nitrogen. The resulting mixture was agitated at 120 rpm for 6 h at 25 °C under nitrogen. The precipitate was collected by centrifugation at 4,600 rpm for 20 min and washed thrice with water and acetone. For comparison, a physical blend of polyaniline and graphene was prepared by substituting GNH with pristine graphene (Section 2.3). Polyaniline was prepared in the same way without graphene.

Spectroscopic
Characterization of Polyaniline/Graphene Composites. For UV-Vis, the samples were dispersed by ultrasonication in ethanol for 10 min and centrifuged at 300 rpm for 10 min. The supernatants were collected and their corresponding absorption spectra were acquired in a quartz cuvette with a 0.5 nm scanning resolution using a Shimadzu UV-2401PC UV-Vis Recording Spectrophotometer.
For Raman studies, the exfoliated graphene flakes or modified exfoliated graphene flakes were dispersed in ethanol. The dispersion was then drop-cast onto 1" × 0.5" glass slides and the solvent evaporated at room temperature. The Raman spectra were then analyzed using a home-built dispersive Raman confocal microspectroscope (550 mm focal length, 1200 grooves/mm grating, excitation source of 514.5 nm, and with a resolution better than 2 cm -1 .) of the National Institute of Physics, UP-Diliman. Spectral analysis, which include baseline correction and normalization of the peaks relative to the most prominent carbon peak (G peak), were performed using Origin Pro 8 Peak Fit.
For FTIR, the samples were pelletized with KBr and run in transmittance mode at 4 cm -1 resolution using a Shimadzu IRAffinity-1; parameters used were: Happ-Ganzel apodization, at least 25 scans averaged, and baseline correction and single smoothing were done.

Electrochemical
Characterization of Chemically Modified Graphene. The product was concentrated in the native solvent and drop-casted onto pre-treated pencil-lead graphite electrodes (PLGEs). Pretreatment consisted of a potential sweep from -0.2 V to 1.0 V and back for 50 cycles with a sweep rate of 50 mV/s in a 1 M HCl solution using a three electrode cell. The prepared electrodes were then left overnight to dry. The electrodes were then attached to a Metrohm Autolab µAutolabIII/FRA2 in a three-electrode configuration with a silver/silver chloride reference electrode and a platinum counter electrode. Two different electrolytes were used, 0.1 M KCl with 1 mM K3Fe(CN)6 and 1 M H2SO4. Cyclic voltammetry (CV) was performed within a potential range of -0.2 to 1.0 V with varying sweep rates for 10 cycles each.
The triethylamine, THF and diethyl ether used in the synthesis of the salicylaldehyde precursor were freshly distilled prior to use. All other starting materials and solvents were purchased commercially and used as received. Proton NMR spectra and infrared (IR) spectra were measured using a 500 MHz JNM-LA500 and a Shimadzu FTIR-8300, respectively. UV-Vis spectra were recorded on a Shimadzu MultiSpec-1500 at room temperature with CH2Cl2 as solvent. For low temperature UV-Vis measurements, an Alpha Engineering Ltd. CS-96-2 cryostat and Model TC 22-A thermo controller were used. High resolution and electron-ionization mass spectrometric measurements were carried out on a JEOL JMS-SX102A. For the complex, fast atomic bombardment (FAB) with p-nitro-benzyl alcohol (NBA) as matrix was utilized. Elemental analyses were obtained with the use of a Perkin-Elmer 2400-II and magnetic susceptibility measurements were done by way

RESULTS AND DISCUSSION
Functionalization of p-Aminophenyl (Aniline) Moiety in Graphene. GNO and GNH ( Figure 1) were characterized vis-à-vis pristine graphene (PG) using FTIR and Raman spectroscopy to confirm successful functionalization. The spectrum of exfoliated graphene (PG) hardly have any features indicating the absence of functional groups visible in the IR spectrum ( Figure 3). This contrasts with the FTIR spectrum of the starting material used by Kumar et al. (2012) in which graphite oxide (GO) was used. The spectra of GNO and GNH in this region have peaks corresponding to the functional groups (bending vibrations for N-H, N-O and C-N) introduced; e.g. for the GNH sample, a more prominent peak at 1620 cm -1 typically attributed to the -NH2 scissoring vibration is present. These results are indicative of successful modification of graphene. However, peaks at 2840 cm -1 and 2900 cm -1 , which correspond to aliphatic C-H stretch, cannot be accounted for in the structures of GNO and GNH while the peak at ~3400 cm -1 , which correspond to O-H or N-H stretch, cannot be accounted for that of GNO.
Furthermore, the unmodified and modified graphenes have different dispersibilities in ethanol. GNO in ethanol remains dark after centrifugation whereas GNH easily settles down. On the other hand, in water, GNH is easier to disperse as fine particles whereas PG hardly breaks up and easily agglomerates when allowed to stand. This difference in dispersibility lies in the nitro and amino groups present: nitro and amino groups are polar, so they favorably interact with ethanol; however, amino groups are easily ionizable and the presence of these charges make them less dispersible in ethanol which is only weakly polar compared with water. In summary, their respective dispersibilities in ethanol are in the order of GNO >> PG > GNH.
Though the FTIR data confirmed the presence of the nitro and amino functional groups, it could not differentiate between covalent linkage and non-covalent interaction between graphene and the moieties. Washing  Wavenumber (cm -1 ) alone may not have removed residual aminocontaining reagents. Thus, Raman spectroscopy was used to confirm covalent functionalization as this is able to detect a change in the hybridization of the carbon in graphene, from sp 2 to sp 3 , including when it forms a covalent bond with a functional moiety. At the excitation of 514.5 nm, diamond-like carbons produce a Raman shift at ~1350 cm -1 , denoted as the D peak, while graphitic carbons have a Raman shift at ~1580 cm -1 , the G peak. These two shifts, respectively, are correlated to the sp 3 and sp 2 carbons present in carbonaceous material, with the former resulting from symmetrybreaking effects (Jorio, 2011). The G peak (~1580 cm -1 ), which corresponds to sp 2 carbon-carbon bond vibration, remains largely unaffected by chemical reactions in the absence of resonance enhancement or formal charging (Sharma R et al., 2010). The D peak at ~1350 cm -1 corresponds to the sp 3 carbon vibrational modes and is indicative of defects in the graphene structure. The D/G peak ratio is usually used to monitor the extent of hybridization, and therefore extent of reaction or functionalization, where the G peak can serve as reference peak (Hsiao et al., 2010).
The Raman spectra for the samples are compared with the exfoliated graphene in Figure 4. Here, we note that the D peak is present in exfoliated graphene as these indicate defect areas particularly at the edges of the graphene flakes. We also note a major difference in these samples compared with the work by Kumar et al. (2012)-the Raman peaks in our samples are sharper where the D and G peaks are non-overlapping, with G still more prominent than the D peak, compared with the broader, overlapping bands in Kumar's work. This is indicative of less disruption of the graphene structure in our scheme of direct functionalization of graphene which did not pass through the graphene oxide step of Kumar et al.
The D/G peak ratios are summarized in Table  1. The values for the modified graphene are much higher compared to PG: 2.5 times for GNO and 3.4 times for GNH.  This implies structural modification of the carbon network of graphene, which arises from the change in hybridization of the affected carbon atoms, from sp 2 to sp 3 . This conforms with the FTIR results and further confirms successful covalent functionalization of presumably p-nitrophenyl and p-aminophenyl moieties on graphene. It is expected that the D/G ratios for the GNO and GNH should be the same because the reduction of the nitro groups in GNO does not affect the graphene-phenyl bond. However, the D/G ratios are somewhat different. This may possibly be due to the difference in the effect of scattering of the functional groups but may also be caused by the polydispersity of the sizes of the exfoliated flake samples. Polydispersity arises from solution phase techniques used to produce graphene dispersions (Green and Hersam, 2010). Single graphene layers seem to be more reactive than bi-or multilayers under Raman spectroscopy because the process samples the inner carbon atoms which are shielded from chemical treatment; the G peak intensity increases for a given D peak caused by the functionalization of the outer layers for multi-layered graphene. Also, as previously reported, there is preferential functionalization of the edges of graphene over the bulk. An altered electronic structure is expected at the edges compared to the bulk region due to symmetry breaking of the honeycomb lattice at the edges. In addition, the dangling bonds at the edges can also be a major contribution to the enhanced edge reactivity (Sharma R et al., 2010).
The efficiency/yields of these two processes cannot be reliably quantified. FTIR was used to confirm the presence of nitro and amino groups in the sample; however, the peaks cannot be used to quantify the functional groups present due to the lack of an internal standard. In addition, peaks that differentiate the two samples overlap, so it is impossible to distinguish the amino groups from nitro groups in GNH if it had undergone incomplete reduction. On the other hand, Raman spectroscopy was used to confirm the covalent functionalization of graphene with p-substituted phenyl moieties but the polydispersity of the exfoliated graphene flakes makes it difficult to quantify the exact yield. Further studies are needed to quantify the yield of these two processes.
In Situ Oxidative Graft Polymerization of Aniline. FTIR spectroscopy confirmed the successful synthesis of polyaniline and its presence in either the physical blend of PANI with graphene (PANI-G) and in the covalently functionalized PANI-grafted-G (PANI-GNH) composites showing the typical bands for PANI ( Figure 5). The broad peak around 3430 cm -1 appears due to the N-H stretch in PANI. The peaks around 1560 cm -1 and 1480 cm -1 are due to the C=C stretching deformations of the quinonoid and benzenoid structures of PANI, respectively. The characteristic peak around 1120 cm -1 is due to the N-Q-N-Q stretch of the quinonoid ring.
Nonetheless, the general properties of the two polyaniline-graphene composites differ from each other. Right after polymerization, PANI/G precipitates were large and loose, did not break up into smaller particles in water when agitated, and there were green precipitates of polyaniline. On the other hand PANI-GNH precipitates clumped together but were easily broken down into smaller particles in water when agitated, and are uniformly black to dark green. Upon drying, optical microscope images of PANI/G precipitates retained the appearance of graphene and polyaniline with the presence of loose graphene sheets (light gray). In comparison, PANI-GNH precipitates clumped together ( Figure 6). This difference in behavior in water is indicative of a different type of linkage (physical versus chemical) for these composites. However, gross morphology is not a direct proof of structure.
UV-Visible spectroscopy further reveals the difference between the two composites, and compared with PG or PANI alone (Figure 7). The peak at 270 nm in graphene, PANI/G and PANI-GNH confirms the presence of graphene (Shan et al., 2010) while the absence of a peak at 230 nm confirms the absence of graphene oxide. The absorption band around  380 nm for PANI appears to have a bathochromic shift in the two composites which may be due to the interaction between polyaniline and graphene. This interaction, mainly due to - stacking, stabilizes the  → * transition at 320-380 nm (Saini and Basu, 2012) in polyaniline, which resulted in a corresponding bathochromic shift (Sharma and Kumar, 2010). It is notable that the graphene and PANI optical properties are retained in the case of the grafted composite (PANI-GNH)-dominated mostly by the graphene band, with small PANI contribution.
The relatively low PANI absorbance in PANI-GNH indicates low density of functionalization per flake. Further studies are needed to quantify the extent of functionalization of PANI relative to GNH.
The characterization methods used to analyze the two graphene/polyaniline composites cannot be used to quantify the extent of polymerization of aniline relative to graphene. This is due to the lack of an internal standard in the FTIR. On the other hand, the differences in the dispersibility of graphene, polyaniline and graphene/polyaniline composites in a solvent transparent to UV-Vis and the tendency of the samples to scatter light makes quantification using Beer-Lambert's Law inconclusive. Further studies are needed to quantify the extent of functionalization of PANI relative to GNH.
Electrochemical Characterization. The electrochemical properties of the graphene composites were compared alongside with PG and PANI alone, using each of them as working electrode (coated on graphite lead) and using potassium hexacyanidoferrate(III) as internal reference redox couple. The cyclic voltammograms are shown in Figure 8. The redox peaks of the potassium hexacyanidoferrate(III) is seen for the PG electrode with an additional irreversible reduction peak appearing at around +0.82 V (vs Ag|AgCl) (Bo et al., 2011), which may be attributed to the reduction of the native oxidized carbon at the edges of the graphene flakes. These CV peaks were swamped by the PANI redox bands in the case of PANI/G or PANI-GNH (compare with PANI for example). There is slight shifting of the peak positions but overall they are similar for the two composites. This indicates that PANI remains electroactive and dominates the property even when grafted to graphene.

CONCLUSION
Polyaniline in situ grafted to graphene was successfully synthesized. The process did not go through the usual graphite oxide intermediate to produce a PANI-GNH covalent composite. This composite was compared vis-à-vis polyaniline/graphene blend showing similar FTIR spectra, different UV-Vis absorption spectra indicating possibly low grafting density, and similar electrochemical property. Compared with the physical blend of PANI and graphene, the PANI-GNH is predicted to be more robust because of the covalent link. Additional characterization and studies are needed to evaluate the extent of grafting and subtle changes in the PANI and/or graphene properties.

ACKNOWLEDGEMENT
This work is done in part with funding support from the Philippine Council for Industry, Energy, and Emerging Technology Research and Development of the Department of Science and Technology (PCIEERD-DOST). EPE acknowledges the Chevron Philippines, Inc. Endowed Professorial Chair in the Ateneo. The Commission on Higher Education (CHED) is also acknowledged for the KIMIKA Writing Workshop held last 10-12 April 2013.