Polymers and polymer composites have been present in human life for a long time. Every day, we use many objects made of polymers, e. g. disposable tableware made of polyethylene (PE) or polypropylene (PP) are with us during the family picnics and holiday trips. We drink water from plastic bottles, eat using plastic cutlery, and so on. On the other hand, we see polymers and polymer composites as essential parts in equipment at homes, public buildings, cars, aircrafts, and different machines. Moreover, the role of these materials has become significant for our organisms. Nowadays, polymeric biomaterials play important roles in medicine for example as resorbable substrates for tissue regeneration control, construction implants of tailored mechanical properties and controlled resorption time, construction of prostheses, dental materials, artificial coronary vessels, a drug carrier, biostable joining elements, surgical sutures, and dressing materials. The polymers most often used in medical applications are polyvinyl chloride (PVC), polyvinylidene chloride (PVDC), polystyrene (PS), PE, polytetrafluoroethylene (PTFE), PP, polymethyl methacrylate (PMMA), polycarbonate (PC), and many others.
Moreover, biomaterials are generally made of polymer composites. Additionally, they can include metals and metal alloys (e. g. Fe, Co, Cr-Ti, and amalgams), ceramic materials (e. g. bioglass and hydroxyapatite) or possibly cells and human tissues.
Generally, it is worth noting that organic–inorganic hybrid materials such as polymer-ceramics and polymer-carbon very often are applied as biomaterials or materials used in electronics and optoelectronics.
According to literature, essential and very often used components of polymer composites are different allotropic forms of carbon, e. g. graphene, graphene nanoribbons (GNRs), and single- or multiwalled carbon nanotubes (MWCNTs). Thus, owing to interesting electronic and mechanical properties of GNRs, they have become promising materials for preparation of conductive composites for applications as transparent electrodes, heat circuits, and supercapacitors . On the other hand, chemical modification of graphene by photopolymerization with styrene results in self-organized intercalative growth and delamination of graphene with few layers. Such composites are extremely promising for a wide range of mechanical, thermal, and electrical applications . An interesting report was that on electron field emission (FE) from reduced graphene oxide (rGO):poly(3-hexylthiophene) (P3HT) composite layers . Reduced graphene oxide can also have potential applications in vacuum microelectronic devices, such as microwave power amplifiers and FE-based electronic devices such as flat panel FE displays (FEDs) [4, 5, 6]. Graphene fibers provide new ways to study interfacial interactions between the polymer and graphene for producing high-performance graphene-enhanced polymer nanocomposites. As an example, transcrystallization of isotactic polypropylene (iPP) surrounding the graphene fibers can be shown. The results indicate that fabrication of effective graphene-enhanced polymer nanocomposites can be used for various emerging applications . Graphene-based polymer composites are also developed for potential use in biomedical applications. Soft biomedical polymers can be strengthened by incorporating graphene in the polymer matrix for use in hard tissue applications such as orthopedic repair and regeneration in the form of fracture fixation devices, tissue scaffolds, etc. .
Another form of carbon used as fillers in polymer composites or hybrid materials is carbon nanotube (CNT), generally due to its extremely high Young’s modulus, stiffness, and flexibility. Considering first single-walled carbon nanotubes (SWCNTs) as a candidate for application in producing composite materials, it should be noted that successful application requires well-dispersed nanotubes with good adhesion to the host matrix, which is not easy to realize. Moreover, weak nanotube–polymer interactions result in poor interfacial adhesion, which can cause nanotube aggregation within the matrix. A novel approach to in situ composite synthesis, by attachment of PS chains to full-length pristine SWCNTs without disrupting the original structure, has been proposed . For some applications, solid-phase deposition of CNTs at the site of action is realized, but for other ones solution-phase processing and manipulation are required to achieve appropriate assemblies, orientations, and homogeneous dispersion of CNTs within host materials [10, 11]. In the last decade, the interest has been focused on functionalization of CNTs and especially SWCNs with various organic, inorganic, and organometallic structures using both covalent and noncovalent approaches [12, 13]. Multifunctional nanotube–polymer composites (e. g. SWCNT/MWCNT-polycarbonate (PC)/polystyrene (PS)) have been developed mainly due to improvement in mechanical and thermal properties and an increase in electrical conductivity [14, 15, 16]. The interest in transparent and flexible conductors as components of various devices used, e. g. as paper displays and plastic solar cells, has resulted in novel CNT/polymer composites (e. g. PS, PMMA, P3HT) with highly aligned nanotubes inside. The composite films obtained show high optical transparency, robust flexibility, and excellent conductivity [17, 18]. Polymer colloids with an interfacial coating of purified SWCNs have been synthesized from length- and type-sorted SWCNTs. Such composite particles exhibit electrical conductivities comparable to or higher than those of bulk SWCNT–polymer composites at nanotube loadings lower by more than 1 order of magnitude, maintaining unique electronic and optical characteristics of the parent SWCNT solution with potential applications as microelectronic and microoptical components . In the last few years, hybrid nanocomposites containing SWCNTs and ordered polyaniline (PANI) have been prepared through in situ polymerization reaction. The SWCNT–PANI nanocomposites show both higher electrical conductivity and Seebeck coefficient than pure PANI, which could be attributed to the enhanced carrier mobility in the ordered chain structures of PANI. Such nanocomposites with thermoelectric effect (TE) have great potential for applications in both power generation in waste heat recovery systems and solid-state cooling or heating devices .
Hybrid materials fabricated on the basis of polymers and allotropic forms of carbon constitute a wide group of composite materials; however, other groups of hybrid materials synthesized from polymers and metals, metal oxides and metal alloys have been discussed in the past decade. Ordered mesostructured materials fabricated from polymerizable silica species and organic structure-directing agents, such as amphiphilic block copolymers or charged surfactants, are important examples of self-assembling hybrids. However, optical devices made from silica have a limited refractive index (n = 1.43) and must be supported by ultralow refractive index materials when efficient waveguiding is important . One of the innovations proposed for fabrication of ordered self-assembling optical hybrids required replacing the silica framework with a higher refractive index material, such as titania [22, 23]. Dye-doped hybrid waveguides with trifluoroacetate-modified titania frameworks and high effective index of refraction (n = 1.6–1.7) have been fabricated [23, 24]. Commodity polymers such as PC films and polyester (PET) filaments are selectively functionalized through UV oxidation and then used as templates to control the nucleation and growth of ZnO nanorods and microplates forming periodically ordered microarrays directly out of aqueous solution. Such a process allows preparation of materials for biomimetics and makes this strategy more technologically applicable . On the other hand, metal-insulator core-shell structure has been employed to fabricate composites with high dielectric constant and low energy losses. An example is Zn-ZnO/polyvinylidene fluoride (PVDF) composite with enhanced dielectric constant due to duplex interfacial polarizations induced by metal–semiconductor interface and semiconductor–insulator interface. Such polymer-based dielectric composites with conductive fillers have potential applications in electric power systems and electronic devices . Thermally conducting but electrically insulating polymer-based composites have been widely used in electronic devices. Such composites are mostly manufactured by introducing highly thermally conducting particles such as ceramics, metals, or metal oxides into a polymer matrix. Examples are hexagonal boron nitride (hBN) platelets widely used as the reinforcing fillers for enhancement of thermal conductivity of polymer-based composites .
The above-mentioned groups of different composite materials and their applications are far from exhausting all materials reported in literature and in the following, spectroscopic characterization of only chosen polymer composites is given.
1.1 Review of results
Among different spectroscopic techniques, Raman spectroscopy is a noninvasive and powerful tool for investigation of different materials. For quite a long time, Raman spectroscopy was mainly dedicated to fundamental research, but development of instrumentation (laser miniaturization, CCD detection, Rayleigh filters, high-resolution holographic gratings and data processing software) has rendered it a general characterization method. An important breakthrough in the development of Raman spectrometers was integration with optical microscope, ensuring spatial resolution of less than 1 μm, upon visible excitation. It is worth noting that infrared absorption spectroscopy is complementary to Raman spectroscopy; however, it is not as versatile, useful and informative as Raman spectroscopy. Moreover, standard equipment of today’s Raman microscopes is a motorized stage allowing Raman mapping and depth profiling of the sample.
A wide group of polymer composite materials contains allotropic forms of carbon, such as graphene flakes, carbon nanoribbons, or CNTs. Raman spectroscopy is known as a very useful technique to study compounds containing carbon allotropes. It is one of the few techniques sensitive to the full range of structural states present in this class of materials, from perfectly crystalline to amorphous. Raman spectroscopy allows distinction between diamond, graphite, graphene, or CNT through the measurement of a single spectrum. The common crystalline phases of carbon yield very simple spectra: diamond (sp3 hybridization) gives a strong single mode of T2g symmetry at 1,332 cm–1, whereas graphite (sp2 hybridization) gives doubly degenerated E2g modes at 42 and 1,582 cm–1. The latter is referred to as G band and corresponds to vibration in graphene planes, whereas the former corresponds to weak interplanar Van der Waals interactions . Two additional modes called D and D’ (the letter stands for “disorder”) are detected whenever flaws/defects appear in the structure. D’ mode results from the splitting of the G band and appears around 1,620 cm–1, at the value at which the dispersion curve of graphite is the flattest. The intensity ratio of G to D’ bands depends on the proportion of distorted graphene planes . D band results from the resonant enhancement of the modes from graphite dispersion curves, having the same wavevector k as the incident photons. This assignment provided an explanation for some peculiarities of the D band, such as its excitation dependence . A SWCNT can be considered as a cylinder of graphene of a few nanometers in diameter and a length ranging from tens of nanometers to millimeters. The properties of SWCNT are different and depend on the cylinder diameter and rolling direction. The number of Raman-active modes of SWCNT depends on the symmetry of the tube but is independent of its diameter. In fact, only four to five Raman bands are observed. Two of them recorded at around 1,600 cm–1 are assigned to G+ and G–; however, the mode at around 1,350 cm–1 is attributed to defect-induced vibration (D mode). Moreover, in the Raman spectrum of SWCNT, besides D, G (G+, G–), and D* (marked also in literature as 2D) modes, a low wavenumber radial breathing mode (RBM) is detected in the range 80–350 cm–1 .
1.2 Polymer composites with graphene
Owing to the unusual mechanical properties of graphene (Young’s modulus of an order of 1 TPa), production of reinforced high-performance composites based on polymer and graphene seems to be possible. Upon a composite deformation, internal stress is induced and in such materials stress is transferred from polymer matrix to the monolayer graphene. Raman spectroscopy is one of the best techniques for characterization of both graphene and its deformations. The positions of Raman bands in the graphene spectra shift with stress and the stress-induced Raman bands shifts can be used for determination of the stress in the material and in consequence for estimation of its effective Young’s modulus [32, 33]. It is important to evaluate and compare the levels of reinforcement in polymer nanocomposites by exfoliated graphene flakes made of different numbers of layers. Many authors have reported that the shift of Raman bands for the nanocomposites containing multilayer graphene flakes is smaller than that for the material with monolayer graphene. Moreover, the band shift rate for multilayer graphene without a top coat (i. e. polymer is deposited on only one surface of the flake) is very low [34, 35]. Gong et al. have discussed the influence of tensile strain on vibrational spectra of monolayer and bilayer graphene flakes before and after applying the SU-8 top coat (SU-8 is an epoxy resin). The changes in the peak position of 2D Raman band vs. strain for coated and uncoated monolayer and bilayer graphene on PMMA polymer beam are presented in Figure 1.
It is well known that the 2D band from the spectrum of bilayer graphene consists of four components and should be fitted using four lines. As seen in Figure 2(a), a red shift of four components of 2D band is observed. Comparing the slopes of 2D1A and 2D2A components recorded for bilayer graphene to those recorded monolayer graphene, one can notice very similar dependence of the 2D component position as a function of strain. To avoid problems with orientation of the graphene flake, the measurements were carried out for the same flake coated with mono-, bi-, and trilayer of graphene, which ensures the same orientation of the flake toward the laser light polarization. The shift of the adjacent monolayer region is shown for comparison in Figure 2(a) (open circles) . Figure 2(b) shows changes in the position of 2D band vs. strain for four different coated graphene structures, i. e. mono-, bi-, tri-, and other multilayers of graphene. The 2D band was fitted in all cases with a single Lorentzian line shape function, to compare the shift of the maximum of this band. The slope of 2D band maximum for mono- and bilayer materials is comparable but somewhat lower for the trilayer material. However, the slope for multilayer graphene is significantly lower.
A shift of 2D band can be also caused by a change in the measurement conditions, e. g. different wavelength, different relative orientation of graphene lattice to the straining direction, and direction of polarization of laser light. Gong at al. have performed a systematic study for more than 30 different graphene flakes on polymer beams in different orientations, with different numbers of layers, both uncoated and with a polymer coat, recording changes in the 2D band position upon deformation. For uncoated samples, a decrease in the band shift rate (dω2D/dε in cm–1/% strain) was noted when the number of layers increased from one to three, from –48.8 to –32.4, respectively. The value obtained for a multilayer uncoated flake was –37.4 but with a high error ±8.2. The shift rate obtained for uncoated graphite flake was very low and equaled –3.0. For coated specimens, the band shift rates observed were generally higher and reached from –57.7 to –46.6 cm–1/% strain for the samples with a single layer and that with three layers, respectively. Additionally, the differences between mono- and bilayer flakes were comparable within the limits of experimental error. The value obtained for a multilayer coated flake was –40.2 cm–1/% strain, but similarly as for uncoated specimen, it was charged with a high experimental error ±14.2 cm–1/% strain. For coated graphite, the band shift rate is 0 cm–1/% strain. Since the shift of the 2D band with strain (dω2D/dε in cm–1/% strain) is proportional to the effective Young’s modulus of graphene, it follows that, if the polymer–graphene interface remains undisturbed, the mentioned above band shift rates are an indication of the efficiency of internal stress transfer within the graphene layers. Concluding, it has been established that although there is good stress transfer between the polymer matrix and a monolayer graphene, monolayer graphene is not the optimum material to use for reinforcement of graphene-based polymer nanocomposites. Similarly, in bilayer material, good stress transfer between polymer matrix and graphene is observed, and there is no slippage between the layers for fully encapsulated material. Less efficient stress transfer has been found for trilayer and other multilayer graphene due to slippage between the internal layers, indicating that such materials will have a lower effective Young’s modulus than either monolayer or bilayer graphene in polymer-based nanocomposites .
Young et al. have demonstrated a possibility of using Raman mapping for investigation of strain distribution in graphene flake and stress transfer from the polymer matrix (PMMA beam) to a graphene monolayer for model composites sample . It was found that the shift rate with strain of 2D band was –61 ± 2 cm–1/% strain showing good stress transfer between underlying polymer and a graphene monolayer. Moreover, a significant broadening of 2D band is observed as a result of deformation. However, it should be mentioned that broadening or even splitting of 2D band strongly depends on the angle between the direction of laser light polarization and high-symmetry directions in graphene.
In order to avoid discrepancies in interpretation, the direction of the laser light polarization was parallel to the strain axis in the graphene monolayer for all measurements. Figure 3presents Raman maps of the strain in graphene monolayer at different strain levels applied in horizontal direction along the flake. Scale bar shows the relationship between the color of maps and graphene strain. The contour of the graphene flake around the maps is also shown. The black dots on the maps represent the points at which measurements were carried out. As seen from Figure 3, the first two maps were obtained for applied strains of 0 % and 0.4 % for graphene on PMMA beam without SU-8 top coating. These maps indicate that the strain in graphene is relatively uniform with some evidence of lower strain at the left-hand end at 0.4 % strain. Then, the specimen was unloaded, coated with SU-8, and reloaded to 0.4 % and 0.6 % strain, and Raman maps were obtained (see Figure 3). It is seen from Figure 3that at 0.4 % strain the graphene strain distribution is virtually identical in both the uncoated and coated samples. Increasing the applied strain to 0.6 % strain causes the strain in the graphene to increase to around 0.6 % strain over most of the monolayer, with a lower level of strain at the left-hand end. The distribution of strain over a linear region along the middle of the long axis of the monolayer is presented in Figure 4. The increase in strain at the right-hand tip of the tapering monolayer can be explained in the same way as for the fibers in composites with differently shaped ends [36, 37]. If the fiber has a square end in cylindrical fiber, there is a gradual decrease in fiber strain toward the end of the fiber (left-hand end of graphene flake), whereas if the fiber has a conical tip, then for high-modulus reinforcements the strain actually rises as the fiber tapers (right-hand end of graphene flake) and drops to zero only very close to the end of the tip .
Further significant information related to stress transfer to the graphene monolayer at higher strain can be gathered from the Raman maps of strain in graphene (see Figure 5).
As shown, the strain is relatively uniform around 0.15 % in the relaxed state, whereas if the applied strain is increased to 0.8 %, the strain distribution becomes very nonuniform in the graphene monolayer. Further analysis carried out for the specimen reloaded to 0.6 % applied strain (Figure 5) indicates that variation of strain along the middle of the long axis of the graphene flake changes quasi-periodically with the repetition within 10–20 μm. After repeated relaxation of the sample, the strain is less uniform than for the originally relaxed sample. After further reloading to 0.6 % applied strain, the strain distribution is very similar to that obtained for 0.8 % applied strain, but it is different than that obtained for the initially loaded specimen in which the distribution of strain was relatively uniform (see Figure 3). It probably indicates that the specimen was damaged by loading up to 0.8 % strain and the damage was retained on reloading to the lower strain . The results presented by Young et al. indicated that Raman mapping can be used for the analysis of strain distribution in specimens containing graphene and allows following the levels of reinforcement in such systems with high precision.
1.3 Polymer composites with carbon nanotubes
Owing to the application potential of CNTs in the production of molecular wires, the next generation of electronic devices, fibers with exceptionally high tensile length, novel thermoelectric materials, etc., new high-performance composite materials have been recently reported by many authors.
Yao et al.  have studied hybrid nanocomposites containing SWCNTs and PANI prepared in a simple one step by in situ polymerization reaction. Because of excellent electrical properties of SWCNTs much better than those of MWCNTs, resulting from fewer structure defects in the SWCNTs, both electrical conductivity and Seebeck coefficient of highly ordered chain structure of SWCNT/PANI nanocomposite are much enhanced with increasing SWCNT content. In this in situ polymerization process, the PANI could be considered to grow along the surface of SWCNTs because of the strong π–π interactions between the two components. Meanwhile, the chain packing of PANIs is also ordered due to the template effect of 1D nanostructured CNTs. Transmission electron Microscopy (TEM) and SEM images of SWCNT/PANI powder containing 25 wt % SWCNT are shown in Figure 6. The composites show nanocable structure in which a bundle of SWCNTs were coated and bounded by PANI.
Moreover, both X-ray diffraction and Raman spectra analyses confirmed the ordered chain packing of PANIs. Raman spectra analysis gives evidence for the ordering of PANI along SWCNT. As follows from Figure 7, the Raman spectrum of pure CNTs shows a strong peak at 1,588 cm–1 assigned to the G band (in-plane stretching E2g mode). For the pure PANI, C–H bending of the quinoid/benzenoid ring at 1,162 cm–1, weak C–N stretching at 1,218 cm–1, C=N stretching of the quinoid ring at 1,483 cm–1, and C–C stretching of the benzenoid rings at 1,590 cm–1 are observed. It is worth noting that the intensity of modes at 1,483 cm–1 and 1,164 cm–1 decreases with increasing SWCNT content with a simultaneous increase in the intensity of the mode at 1,590 cm–1. The decrease in the intensity of modes at 1,164 cm–1 and 1,483 cm–1 with increasing SWCNT content is attributed to the site-selective interactions between the quinoid rings and CNTs, which induce the chemical transformation of quinoid rings to benzenoid rings and cause the conformational changes of PANI from a coil-like structure to an extended one.
Increasing intensity of mode at 1,590 cm–1 suggests a decrease in the quinoid unit concentration and an increase in the benzenoid unit concentration. Moreover, it has been also found that all modes at 1,218, 1,483, and 1,590 cm–1 shift to the lower wavenumbers after addition of SWCNTs, which may result from the additional π–π conjugated interactions between the PANI and SWCNTs that induce the red shift of Raman modes. The ordered structures result in the increase in carrier mobility. In consequence, both electrical conductivity and Seebeck coefficient of the PANI-based polymer composites are dramatically improved as compared with those of pure PANI. However, the thermal conductivities of the composites, even with high SWCTN content, do not change much and still keep very low values, which is attributed to the phonon scattering effect of nanointerfaces produced by the SWCNT/PANI nanocable structure [20, 38].
Successful applications of composites reinforced by addition of SWCNTs require well-dispersed nanotubes with good adhesion to the host matrix, which, unfortunately, is not easily realized. The reason for this is related to poor solubility of SWCNTs and weak nanotube–polymer interactions, and the poor adhesion causes nanotube aggregation within the matrix. One approach to in situ composite synthesis is the attachment of PS chains to full-length pristine SWCNTs without disrupting the original structure. Anionic polymerization process requires no nanotube pretreatment and works well with as-produced SWCNTs. In such a process, well-defined composites with a homogeneous dispersion of nanotubes were obtained. Carbanions are introduced on the SWCNT surface by treatment with the anionic initiators that serve to exfoliate the bundles and provide initiating sites for the polymerization of styrene (see Figure 8). When styrene is added, both free sec-butyllithium and the nanotube carbanions initiate polymerization, resulting in an intimately mixed composite system. The polymerization was terminated using degassed n-butanol, and the composite was recovered by precipitation with methanol .
Evidence for formation of carbanions and subsequent attachment of PS chains was obtained using Raman spectroscopy. The Raman spectra of pristine SWCNTs, butylated SWCNTs, and PS-grafted SWCNTs are presented in Figure 9.
Apart from two characteristic Raman-active modes recorded for SWCNTs (E2g and RBM), the third mode marked as D band, indicative of disorder or sp3 character within the nanotube framework, is detected at about 1,320 cm–1. As shown in Figure 9, the intensity of D band with respect to E2g mode at about 1,580 cm–1 is indicative of the extent of covalent modification of the nanotube surface. In the Raman spectra of butylated and PS-grafted SWCNTs, the relative intensity of the D band increases when compared to that recorded for pristine SWCNTs. Analysis of the RBM wavenumbers was inconclusive because of the presence of characteristic C–X bending modes of PS in the same region, and no useful information could be obtained. Excessive chemical modification of the nanotube surface can lead to degradation of the nanotube mechanical strength and also result in the loss of electronic structure. It has been reported that nanotube electronic structure can be retained at low levels of functionalization [9, 39].
Li et al.  have reported the application of the Huisgen cycloaddition to functionalization of SWCNTs with PS. To achieve a high degree of functionalization, they chose to introduce alkyne groups on the nanotube surface using the Pschorr-type arylation, which was shown to modify a significant proportion of carbons within the nanotube sidewall. Subsequent introduction of PS was achieved by first installing an azide functionality at the polymer chain end. The Cu(I)-catalyzed formation of 1,2,3-triazoles by coupling azide-terminated polymer and alkyne-functionalized SWCNTs was found to occur in an efficient manner under a variety of favorable conditions. It resulted in relatively high nanotube graft densities, full control over polymer molecular weight, and good solubility in organic solvents .
Raman spectroscopy was not only used to verify the structural integrity of the modified SWCNT materials, but also to get the information on the degree of nanotube functionalization. Figure 10presents the Raman spectra of pristine SWCNTs (A), alkyne-functionalized SWCNTs (B), and PS-functionalized SWCNTs (C). Spectrum A exhibits the characteristic RBM at about 250 cm–1 and tangential G mode at about 1,590 cm–1. In addition, weak disorder band at about 1,290 cm–1 can be observed, indicating the presence of a small number of sp3 hybridized carbons within the nanotube framework. Intensity of disorder band D increases dramatically relative to both the radial breathing and tangential modes, indicating that a large number of sp2 hybridized carbons have been converted to sp3 hybridization. The intensity of tangential mode also increased with respect to that of the RBM. Upon reaction with azide-functionalized PS, the intensity of the disorder band relative to the RBM and G modes remained unchanged, as expected, because the click reaction does not alter the hybridization of carbon atoms within the nanotube framework .
IR spectroscopy provided the information about the structures appended to the surface of the SWCNTs, which is not available from Raman data. Figure 11presents the IR spectra of unmodified SWCNTs (spectrum A), the alkyne-functionalized SWCNTs (spectrum B), and the polymer-functionalized SWCNTs (spectrum C).
In spectrum A of unmodified SWCNTs, only one band around 3,500 cm–1 is recorded and assigned to trace the amount of water present in KBr used for preparation of pellet. However, spectrum B shows small, but clearly noticeable, signals at 2,120 and 3,280 cm–1, corresponding to the C–C and C–H stretching vibrations of appended terminal alkyne functionalities, respectively. Additionally, the signals corresponding to the C–C and C–H stretches of the aromatic ring that serves as a linker between the nanotubes and the alkyne functionality can be seen at 1,660 and 2,920 cm–1, respectively. Upon cycloaddition, the IR spectrum of the product (see Figure 11, spectrum C) bears the bands characteristic of PS, indicating that the polymer was grafted. Although the alkyne band at 2,120 cm–1 is weak, magnification of the three spectra in the spectral range 2,000–2,250 cm–1 indicates that it was not present prior to the reaction with p-aminophenyl propargyl ether or after the Huisgen cycloaddition. The disappearance of the alkyne stretching after the click coupling indicates that most of the alkynes must have been consumed during this reaction, although the low intensity of this IR absorption makes the quantitation of conversion difficult .
Yao et al.  have investigated the use of atom transfer radical polymerization (ATRP), which has been shown to be a highly versatile technique for the controlled radical polymerization of acrylate-based monomers from the surface of the CNTs. The nanotubes functionalized with poly(methyl methacryalte) were found to be insoluble, while those functionalized with poly(tert-butyl acrylate) were soluble in a variety of organic solvents. The resulting polymerized nanotubes were analyzed among others by IR and Raman spectroscopy.
Figure 12(A) presents the IR spectrum of a SWCNT specimen after 48 h of polymerization with MMA, clearly indicating the expected carbonyl stretching vibrational modes at about 1,730 cm–1 and C–H stretching at about 2,950 cm–1 attributed to nanotube-attached PMMA. The presence of CNTs in this residue was confirmed by Raman spectroscopy (Figure 12(B)), which revealed a tangential G band characteristic of them at about 1,590 cm–1 and RBMs at 184 and 205 cm–1, corresponding to 1.2 and 1.1 nm diameter tubes, respectively. The peak at 1,338 cm–1 corresponds to the presence of relatively small amount of sp3-hybridized carbon atoms, formed as a result of sidewall functionalization.
This combination of IR and Raman data indicates that both components are present in the sample and cannot be separated by washing with good polymer solvents. It should be noted that in control experiment, in which nanotubes were mixed with preformed PMMA, filtration and washing resulted in complete removal of the free polymer from nanotube residue, as indicated by the absence of IR stretches at about 1,730 and 2,900 cm–1 .
The IR and Raman spectra of poly(tert-butyl acrylate)-functionalized SWCNTs are presented in Figure 13(A) and (B). Characteristic bands at about 1,730 and 2,950 cm–1 attributed to stretching vibrations in the IR absorption spectrum indicate the presence of polymer, while the Raman bands detected at about 185 and 1,590 cm–1 correspond to the RBMs and tangential modes of SWCNT. It is not entirely clear why the RBM in this sample corresponds solely to nanotubes having a diameter of 1.2 nm, but this is likely due to heterogeneity of different batches of starting material. In addition, there is the band at 1,338 cm–1 corresponding to the presence of sp3-hybridized carbon atoms within the nanotubes formed as a result of the functionalization process. This disorder peak is proportional to the extent of nanotube functionalization.
1.4 Polymer composites with GNRs
Polymer-functionalized graphene nanoribbons (PF-GNRs) can be promising low-cost materials that could be useful for transparent electrodes and heat circuits, electroactive polymer/graphene supercapacitors, and conductive nanocomposites. The synthetic strategy for the one-pot synthesis of PF-GNRs is shown in Figure 14. MWCNTs were converted into edge-negatively charged polymerization macroinitiators via intercalation and splitting. On the basis of the proposed mechanism, it is assumed that the edges of the split tubes are lined by aryl anions and associated to them metal cations. Moreover, anionic polymerization of vinyl monomers starting at the negatively charged GNR edges results in PF-GNRs. The potassium naphthalenide liquid-phase intercalation is described below along with the relevant methodology. Briefly, MWCNTs, potassium metal, naphthalene, and THF (tetrahydrofuran) are placed in a Schlenk flask and subjected to three freeze–pump–thaw cycles to remove oxygen. The intercalation of solvent-stabilized potassium cations into MWCNTs may lead to expansion of the d-space between MWCNT layers, causing the MWCNTs to partially or fully split. The fissures in the sidewalls of the MWCNTs serve as the starting points for vinyl or epoxide monomers to anionically polymerize from the GNR edges. Because of polymerization probably proceeding between the GNR layers, only a small amount of olefin is needed to effect the exfoliation of the MWCNTs. The nonattached polymer was removed by extracting the raw product with boiling chloroform in a Soxhlet extractor .
Raman spectroscopy was used to characterize the graphitic structure of the PF-GNRs. An increase in the intensity of the D band over the G band from 0.15 for MWCNTs to 0.35 PF-GNRs was observed in Figure 15.
Upon splitting of MWCNTs, a prominent D band is an indication of disorder in the graphene structure due to the high edge content. The disordered structure also results in slight broadening of the G band and the 2D band, as well as a combination of D + G bands at 2,700 cm–1 in PF-GNRs. However, a splitting of the G band, corresponding to an intercalated graphitic structure, is not observed in the Raman spectrum, implying that little residual intercalants, if any, or solvents are left between the PF-GNRs .
To explore the flexibility of the proposed protocol, two other sources of MWCNTs, NanoTechLabs MWCNTs (NTL MWCNTs) and Bayer MWCNTs (Baytubes), were subjected to the reaction to compare the results to those obtained when the Mitsui MWCNTs were used in the former two experiments. Upon liquid-phase intercalation followed by polymerization, NTL MWCNTs were split but not further flattened to form GNRs (Figure 16(a)). With the Baytubes MWCNTs, although some partially flattened GNRs could be identified, most of the MWCNTs remained intact (Figure 16(b)).
Raman spectroscopy was used to differentiate the degree of disorder in the structure of the host materials by calculating the intensity ratio of D to G band. The ratio of intensities of disorder-induced D band to that of crystalline G band, ID/IG is 0.15 for Mitsui MWCNTs, 0.27 for NTL MWCNTs, and 0.92 for Baytubes, as shown in Figure 17.
Defect sites on graphite do not favor the formation of well-defined intercalation structure and thus, the complete exfoliation of highly defective Baytubes by intercalation is likely to be more difficult. This is corroborated by recent work on reductive alkylation of MWCNTs with potassium naphthalenide, in which the outer surface of highly defective MWCNTs (ID/IG > 1) was functionalized with decanoic acid and ribbon-like structure was observed in the SEM images. Although NTL MWCNTs have fewer defects, flattening of ultra-long split tubes may require further treatment. Most NTL MWCNTs remained split and stacked rather than completely flattened. It is difficult to precisely establish the structural threshold (i. e. the critical value for ID/IG) that should be used to predict if the MWCNTs can be split and exfoliated; however, it is noteworthy that the higher the degree of graphitization of the carbon host or the less defective the carbon host, the easier the exfoliation of the MWCNTs via intercalation. The PF-GNRs or split tubes could be used for reinforcing polymers, since the sword-in-sheath type failure of MWCNTs due to interlayer slip could be retarded owing to the entangled polymer chains anchored at the edges. Through the compatibilizing appended polymer chains, the load might be effectively transferred from the polymer matrix to the rigid PF-GNRs, thus making stronger composites .
1.5 Examples of other polymer composites
Hybrid materials, the composites that combine attractive qualities of dissimilar materials, have received great interest because of a wide range of mechanical, electronic, biological, and optical application. Boettcher et al.  have reported the synthesis and characterization of highly ordered, stable, titania-based hybrid optical materials, fabricated from self-assembling block copolymers (P123 or F127) and trifluoroacetic acid (TFAA)-modified titania precursors, from molecular to macroscopic length scale. To achieve a deep understanding of a hybrid material’s structure, it is necessary to investigate a hierarchy of size scales with a diverse array of analysis techniques. Analysis on molecular level considering bonding interactions can be carried out using IR and Raman spectroscopy. In the intermediate mesoscale size regime, SAXS (small-angle X-ray scattering) and TEM are used to understand the assembly and structure of ordered hexagonal and cubic mesostructures. The chemical bonding mode between the carboxylic acid groups and the titanium centers was investigated by IR and Raman spectroscopy, among the most well-developed analytical techniques for the characterization of organometallic complexes. The IR vibrations of the hybrid sample were assigned by comparison with the IR spectra of the individual components as well as literature sources . The observed IR/Raman spectra show TFA–Ti asymmetric (νas = 1,653 cm–1) and symmetric (νsym = 1,464 cm–1) carboxylate stretching vibrational modes (see Figure 18).
Low-intensity stretching vibrations of free TFAA at 1,781 cm–1 are apparent in the freshly dipped films but completely disappear within 20 min of drying at room temperature. Identical vibrational transitions are observed in the ethanolic precursor solution. The difference between carboxylate stretching modes wavenumbers Δ = νas – νsym is useful in identifying the bonding mode of the carboxylate ligand (see Figure 19).
Generally, monodentate complexes (Figure 19(A)) exhibit Δ values much higher than those of the corresponding ionic structure (i. e. sodium trifluoroacetate). Bidentate chelating complexes (Figure 19(B)) exhibit Δ values significantly smaller than the ionic values, and bridging complexes (Figure 19(C)) have the delta values less than, but close to, the ionic value. The observed Δ value for the hybrid material is 189 cm–1, much smaller than that for isolated hydrogen bonded dimers (345 cm–1) or the sodium salt (~252 cm–1). This establishes that TFA binds to the titanium center in a bidentate bridging or chelating fashion (Figure 19(C) or (B)), both in solution and in the final hybrid. Because chelation (Figure 19(B)) strains the ~120° free carboxylate OCO bond angle, it is supposed that the TFA ligands bridge adjacent oxo/hydroxybridged titanium atoms (Figure 19(C)) within the amorphous hybrid material. The three broad, full-width-at-half-maximum (fwhm) 50–100 cm–1, Raman modes positioned at about 600, 475, and 280 cm–1 (see Figure 18) can be attributed to the Ti–O– network  and correspond approximately to the three vibrational transitions of the rutile phase of TiO2 (612, 447, 232 cm–1) . The shoulder peaks in this region were identified as corresponding to structural deformations of the TFA ligand. The Raman data suggest that the nearest neighbor environment around the central Ti ions consists of an octahedral “rutile-like” oxygen coordination, consistent with previous studies on acetic acid-modified titanium alkoxides. The lack of long-range order between complexed titania species leads to an overall amorphous network .
Conducting polymers can be expected to serve as molecular wires connecting organic molecules and electrodes; therefore, many attempts have been made to elucidate the properties of single conducting polymer chains [44, 45]. Confinement of polymer chains in porous materials is a feasible method to isolate single polymer chains, permitting the study of their properties, such as conductivity and fluorescence of macromolecules in nanochannels of zeolites, nanoporous silicas, and organic crystalline hosts [46, 47, 48, 49]. Kitao et al.  have reported the incorporation of polysilane into the nanochannels of porous coordination polymers (PCPs) as individual chains. Polysilanes exhibit unique optical and electrical properties, such as backbone electronic transition with UV-vis absorption, high quantum yield of fluorescence, and high mobility of charge carriers [51, 52], which are attributed to delocalization of σ-electrons along the main chains. The conformation of polymer chain, which is sensitive to temperature and to solvent, strongly influences the σ-conjugation system. However, the σ-conjugation of polysilane becomes inefficient under UV exposure because of photodegradation. Therefore, improvement in photostability is highly desirable for light-emitting devices and solar cells .
Host–guest composites of 1⊃PMPrS were fabricated by inclusion of polymethylpropysilane (PMPrS) in the nanochannels of two PCPs with distinct channel size (Figure 20). The used PCPs can be described by the following chemical formula [Al(OH)(L)]n in which the ligand (L) corresponds to 2,6-naphthalenedicarboxylate (DUT-4, marked as 1a) and 4,4ʹ-biphenyldicarboxylene (DUT-5, marked as 1b), see Figure 20.
The molecular conformation of PMPrS encapsulated in PCP was investigated using Raman spectroscopy. Figure 21presents the room temperature Raman spectra of 1a, 1b, 1a⊃PMPrS, 1b⊃PMPrS and bulk PMPrS and Raman spectra of bulk PMPrS and 1a⊃PMPrS recorded at 80 °C. The intensity of the peak corresponding to the symmetric stretching vibrations ν(Si–C) at about 670 cm–1 strongly depends on the main chain conformation. The relative intensity of this peak increases with increasing amount of s-trans conformation. The intensity of the ν(Si–C) peak of 1a⊃PMPrS was higher than that of 1b⊃PMPrS, indicating that PMPrS chains preferred to form a linear structure with increasing s-trans conformation in the smaller channels of 1a. At room temperature, the intensity of the ν(Si–C) peak of bulk PMPrS is similar to that of 1a⊃PMPrS. This is because bulk PMPrS includes chains of s-trans conformation. On heating the bulk PMPrS from room temperature to 80 °C, the intensity of the peak underwent a large change, which showed the transformation from solid to liquid states accompanied by a large conformational change in the PMPrS chains. However, it was striking that the intensity of PMPrS in 1a was almost unchanged at the same temperature of 80 °C (Figure 21). A stretched conformation of PMPrS resulted from the confinement effect of 1a, in which the PMPrS chains were forced to form an extended linear structure in the narrow 1D channels .
Vibrational spectroscopy can be a very useful tool for characterization of different types of materials. Although infrared absorption spectroscopy is a method complementary to Raman spectroscopy, however, it is not as versatile, useful, and informative as Raman spectroscopy. For many composite materials, especially those containing carbon allotropes, Raman spectroscopy can be a basic and widely used characterization technique. Moreover, polymers and other compounds used in composite materials possess their own vibrational fingerprints, so the vibrational spectroscopy methods can appear as essential characterization techniques. However, it is obvious that in the case of some complex hybrid materials, vibrational spectroscopy methods are not the only characterization methods but they are often useful, informative, and quite easy to apply.
This article is also available in: Tylkowski, Polymer Engineering. De Gruyter (2017), isbn 978–3–11–046828–1.
Lu W, Ruan G, Genorio B, Zhu Y, Novosel B, Peng Z, et al. Functionalized graphene nanoribbons via anionic polymerization initiated by alkali metal-intercalated carbon nanotubes. ACS Nano. 2013;7:2669–2675. CrossrefPubMedGoogle Scholar
Jha A, Ghorai UK, Banerjee D, Mukherjee S, Chattopadhyay KK. Surface modification of amorphous carbon nanotubes with copper phthalocyanine leading to enhanced field emission. RSC Adv. 2013;3:1227–1234. CrossrefGoogle Scholar
Abdou JP, Braggin GA, Luo Y, Stevenson AR, Chun D, Zhang S. Graphene-induced oriented interfacial microstructures in single fiber polymer composites. ACS Appl Mater Interfaces. 2015;7:13620–13626. CrossrefPubMedGoogle Scholar
Kumar S, Raj S, Kolanthai E, Sood AK, Sampath S, Chatterjee K. Chemical functionalization of graphene to augment stem cell osteogenesis and inhibit biofilm formation on polymer composites for orthopedic applications. ACS Appl Mater Interfaces. 2015;7:3237–3252. PubMedCrossrefGoogle Scholar
Viswanathan G, Chakrapani N, Yang H, Wei B, Chung H, Cho K, et al. Single-step in situ synthesis of polymer–grafted single-wall nanotube composites. J Am Chem Soc. 2003;125:9258–9259. CrossrefPubMedGoogle Scholar
Yao Z, Braidy N, Botton GA, Adronov A. Polymerization from the surface of single-walled carbon nanotubes - preparation and characterization of nanocomposites. J Am Chem Soc. 2003;125:16015–16024. CrossrefPubMedGoogle Scholar
Ding W, Eitan A, Fisher FT, Chen X, Dikin DA, Andrews R, et al. Direct observation of polymer sheathing in carbon nanotube–polycarbonate composites. Nano Lett. 2003;3:1593–1597. CrossrefGoogle Scholar
Xu H, Wang X, Zhang Y, Liu S. Single-step in situ preparation of polymer–grafted multi-walled carbon nanotube composites under 60Co γ-ray irradiation. Chem Mater. 2006;18:2929–2934. CrossrefGoogle Scholar
Hwang J-Y, Nish A, Doig J, Douven S, Chen C-W, Chen L-C, et al. Polymer structure and solvent effects on the selective dispersion of single-walled carbon nanotubes. J Am Chem Soc. 2008;130:3543–3553. PubMedCrossrefGoogle Scholar
De S, Lyons PE, Sorel S, Doherty EM, King PJ, Blau WJ, et al. Transparent, flexible, and highly conductive thin films based on polymer–nanotube composites. ACS Nano. 2009;3:714–720. CrossrefPubMedGoogle Scholar
Yao Q, Chen L, Zhang W, Liufu S, Chen X. Enhanced theromelectric performance of single-walled carbon nanotubes/polyaniline hybrid nanocomposites. ACS Nano. 2010;4:2445–2451. CrossrefPubMedGoogle Scholar
Yang PD, Wirnsberger G, Huang HC, Cordero SR, McGehee MD, Scott B, et al. Mirrorles lasing from mesostructured waveguides patterned by soft lithography. Science. 2000;287:465–467. PubMedCrossrefGoogle Scholar
Alberius PCA, Frindell KL, Hayward RC, Kramer EJ, Stucky GD, Chmelka BF. General predictive syntheses of cubic, hexagonal, and lamellar silica and titania mesostructured thin films. Chem Mater. 2002;14:3284–3294. CrossrefGoogle Scholar
Zhang Y, Wang Y, Deng Y, Li M, Bai J. Enhanced dielectric properties of ferroelectric polymer composites induced by metal-semiconductor Zn-ZnO Core–shell structure. ACS Appl Mater Interfaces. 2012;4:65–68. CrossrefPubMedGoogle Scholar
Yuan C, Duan B, Li L, Xie B, Huang M, Luo X. Thermal conductivity of polymer-based composites with magnetic aligned hexagonal boron nitride platelets. ACS Appl Mater Interfaces. 2015;7:13000–13006. CrossrefPubMedGoogle Scholar
Dienig T, Hollricher O, Toporski J. Confocal Raman microscopy. Berlin Heidelberg: Springer-Verlag, 2010. Google Scholar
Frank O, Tsoukleri G, Riaz I, Papagelis K, Parthenios J, Ferrari AC, et al. Development of a universal stress sensor for graphene and carbon fibres. Nat Commun. 2011;2:1–6. Google Scholar
Goh KL, Meakin JR, Hukins DWL. Influence of fiber taper on the interfacial shear stress i fiber-reinforced composites materials during elastic stress transfer. Compos. Interfaces. 2010;17:75–81. Google Scholar
Holzinger M, Vostrowsky O, Hirsch A, Hennrich F, Kappes M, Weiss R, et al. Sidewall functionalization of carbon nanotubes. Angew Chem Int Ed. 2001;40:4002–4005. Google Scholar
Li H, Cheng F, Duft AM, Adronov A. Functionalization of single-walled carbon nanotubes with well-defined polystyrene by “click” coupling. J Am Chem Soc. 2005;127:14518–14524. PubMedCrossrefGoogle Scholar
Socrates G. Infrared characteristic group frequencies. New York: Wiley, 1980. Google Scholar
Grozema FC, Siebbeles LDA, Warman JM, Seki S, Tagawa S, Scherf U. Hole conduction along molecular wires: σ-bonded silicone versus π-bond-conjugated carbon. Adv Mater. 2002;14:228–231. CrossrefGoogle Scholar
Sozzani P, Comotti A, Bracco S, Simonutti R. A family of supramolecular frameworks of polyconjugated molecules hosted in aromatic nanochannels. Angew Chem Int Ed. 2004;43:2792–2797. CrossrefGoogle Scholar
Uemura T, Horike S, Kitagawa K, Mizuno M, Endo K, Bracco S, et al. Conformation and molecular dynamics of single polystyrene chain confined in coordination nanospace. J Am Chem Soc. 2008;130:6781–6788. CrossrefPubMedGoogle Scholar
Cucinotta F, Carniato F, Paul G, Bracco S, Bisio C, Caldarelli S, et al. Incorporation of a semiconductive polymer into mesoporous SBA-15 platelets: Toward new luminescent hybrid materials. Chem Mater. 2011;23:2803–2809. CrossrefGoogle Scholar
Kamata N, Terunuma D, Ishii R, Satoh H, Aihara S, Yaoita Y, et al. Efficient energy transfer from polysilane molecules and its application to electroluminescence. J Organomet Chem. 2003;685:235–242. CrossrefGoogle Scholar