Abstract

Keratin, obtained from chicken feathers, was grafted on the surface of commercially available carbon nanotubes. The original procedure developed allows a covalent interaction between some specific chemical groups characteristic of the keratin, with some functional groups introduced on purpose on the surface of the nanotubes, as revealed by infrared and Raman spectroscopies, which also allowed to determine structural changes introduced during the process, such as crystallinity, which lead to changes in other properties, as well.

1. Introduction

Carbon nanotubes (CNTs) have become one of the main focuses of contemporary materials science and engineering thanks to their extraordinary electrical, thermal, and mechanical properties [1] and the rolling up of the graphene sheet at a nanometer scale [2], high stiffness, and axial strength, as a result of the carbon-carbon sp2 bonding [3]. Also, thanks to the inherent graphene structure of the nanotube shell walls, the thermal conductivity of these materials approaches that of on-axis graphite [4]. The precise magnitude of these properties depends on the diameter and chirality of the nanotubes and whether they are single-walled (SWNTs), double-walled (DWNTs), or multiwalled nanotubes (MWNTs) in form [5].

Despite their low degree of crystallinity, CNTs still maintain high mechanical properties. This characteristic makes them ideal as reinforcing agents for high performance polymer nanocomposites. However, the high specific surface area (200  and 400 m2/g) [6] results in a strong tendency to agglomeration. Nanoropes (agglomerates) due to Van der Waals forces are extremely difficult to disperse and align in a polymer matrix [5, 7]. Thus, the way to improve the dispersion of the carbon nanotubes in a polymer matrix is the functionalization of CNTs [5, 810], in particular, covalent functionalization, for which one has first to generate defects on the surface of the carbon nanotubes by hydrolysis, thus forming the covalent bond.

On the other hand, keratin is a protein complex (biopolymer) formed by long chains with a molecular mass of approximately 10 kDa [11]. Relevant to our purposes is to mention the numerous reactive side chains, which can form a variety of surface patterns as result of bending of the backbone, feature that makes it ideal to functionalize, in situ, the surface of carbon nanotubes. Keratin is a useful subject for the study of orientation because it occurs naturally in a α-helix configuration in mammalian fibers such as hair and wool but predominantly adopts a β-sheet configuration in avian fibers such as feathers. These fibers, as well as keratin in skin, have been the target of a number of recent studies according to Rintoul et al. [12].

In this present work, multiwalled carbon nanotubes (MWNTs) were oxidized by hydrolysis to generate defects and different functional groups such as –OH and –COOH in its walls (MWOH’s) [13, 14] aiming to be able to produce a novel hybrid organic-inorganic nanocomposite, through the grafting of keratin onto the carbon nanotubes.

2. Experimental

Multiwalled carbon nanotubes (MWNTs) were obtained from a commercial source, which produced them by chemical vapor deposition, with a diameter from 10 to 30 nm and 1 to 10 μm long, with a purity >90% and surface area of 90 to 350 m2/g. KMnO4, urea 98%, EDTA 90%, 2-mercaptoethanol 98%, tris-hydroxymethylamine 97% at a pH of 9.0, NHO3 (95%), and H2SO4 (98%) were purchased from a commercial supplier. The Bradford reactive was purchased also from a commercial supplier.

Based on previous reports [9, 13, 1518], the purification/oxidation of the MWNTs was achieved in liquid phase with a 3 : 1 mix of NHO3 (95%) and H2SO4 (98%) at 85°C for 3 h in a reflow process. The resulting material was vacuum washed to a neutral pH with deionized water.

Keratin was extracted from chicken feathers, as previously described [11, 19, 20] as follows: 3 g of biofibre were solubilized using 75 mL of distilled water to form 8 M urea 98%, 3 mM EDTA 90%, 2-mercaptoethanol 98%, and 200 mM tris-hydroxymethylamine 97% at a pH of 9.0. The mixture was left at room temperature under stirring for 24 h. The keratin extracted in saline solution passes through a process of dialysis, due to the presence of salts such as urea. Dialyzed solution was preserved at 4°C.

Infrared (FTIR) spectra were recorded in a Vector 33 Bruker spectrophotometer at 32 scans, with a resolution of 4 cm−1. Solid samples were imbedded in KBr disks. Raman spectra were recorded in a Micro-Raman Dilor with a resolution of 515 cm−1, with a 514.5 nm laser with 15 s of integration time. The microplate reader was a Bio-Rad Model 680 XR to 100 V, 50 Hz, and 550 nm. Elemental analysis by combustion has been performed under the following experimental conditions: the sample was weighed on a microbalance Mettler Toledo model XP6 and analyzed at a combustion temperature of 1150°C and a reduction temperature of 850°C in an elemental analyzer from elemental model vario-MICRO-cube. Oxygen 99.996% was the combustion gas, and helium 99.997% was gas carrier. Each sample was analyzed by duplicate. The analyses X-ray Photoelectron Spectrometer (XPS) were performed with a JPS-9200 Photoelectron Spectrometer (ESCA).

Covalent functionalization of the previously oxidized multiwalled carbon nanotubes (MWOHs) with keratin was done [19, 20] by using 0.25% (wt/v) KMnO4, 0.33% (wt/v) malic acid, and 5% (v/v) H2SO4 (98 wt%). These concentrations are constant at 55°, 65°, and 75°C for 2 h with MWOHs/keratin 1 : 1, 1 : 2, and 1 : 3 (wt/wt). In the REDOX reaction, the malic acid/KMnO4 implicates the generation of macroradicals formed by the attack of the species to primary radicals of the peptide bonds. The keratin has hanging functional groups as –NH2, –COOH, –SH, and –OH, which can form covalent bonds with functional groups COOH and COO– present in the walls of the MWOHs. Samples prepared are described in Table 1.

Bradford analysis [21] was employed to quantify the degree of keratin grafting onto the oxidized multiwalled carbon nanotubes, thanks to the color changes, from red to blue, due to the union of the coloring agent to the protein. From an aliquot (60 μL) of keratin-g-MWOHs solution (80 μg/mL) 4.8 g of solid sample is obtained. A calibration curve, by using known solutions of bovine serum albumin protein (BSA 50 μg/mL) in a range of 0 to 4 μg of protein by final volume, was produced. With the linear fit of the resulting curves, the -Square is obtained [22].

3. Results and Discussion

The purification procedure developed here removed most of the amorphous carbon and catalyst residues on the MWNTs, as was confirmed by spectroscopy, thus obtaining oxidized multiwalled carbon nanotubes, key for the covalent interaction (grafting) with the keratin. Figure 1 shows the calibrated absorbance curve, prepared using bovine serum albumin, which allowed quantify the keratin on the surface of the MWOHs.

Table 2 summarizes the quantification of keratin grafted in the carbon nanotubes according to absorbance experiments with Bradford assay, where the sample CK7511 (1:1 wt/wt MWOHs/Keratin ratio to 75°C conditions) contain the highest percentage of keratin grafted in contrast to the 65°C case, with 1 : 1 wt/wt MWOHs/Keratin ratio (CK6511) which contains the least quantity. On average, the reaction condition at 55°C with 1 : 3 wt/wt MWOHs/keratin ratio (CK5513) is the most representative.

Raman spectra can be regarded as a sort of fingerprint of nanostructured carbon [23, 24]. The Raman spectra of the MWNTs pristine, oxidized and all grafted samples with keratin show three important bands at ~1570 cm−1 (G band), ~1340 cm−1 (D band), and ~2684 cm−1 (D′ band) which have been reported by several groups [23, 2533], as can be observed in Figure 2.

The G band in carbon nanotubes involves an optical phonon mode between the two dissimilar carbon atoms in the unit cell. It is known to originate from sp2 carbon sites present in the sample [31]. This band essentially represents the (C=C) stretching vibration of all pairs of sp2 atoms in the ring. According to group theoretical analysis, the G band corresponds to the modes and has been assigned to the movement in opposite directions of two neighboring carbon atoms in a graphite sheet [23]. The second signal (D band) is due to the disorder present in the MWNTs [23]. D band had been assumed to be associated with both the presence of disorder and finite size effect of carbons; this did not always mean that it could originate only because of the disorder present in the material [31]. The D band as characteristic for disordered graphite shows quite significant dependency of its intensity on the excitation wavelength. It is to be noted that this graphitic lattice vibration mode has been assigned to symmetry [31]. Also, in a review recently [33], the D band is attributable to the presence of disordered amorphous carbon, double resonance effects in sp2 carbon. The frequency of the D band peak increases with increasing laser energy [33]. Note that this peak results from amorphous carbon, not defects, in the tube walls [33]. Finally show one broadband assigned to the G′ band to ~2684 cm−1. The G′ band observed in the Raman spectra of all the carbonaceous materials, which is basically the second harmonic of the D band [31, 34], appears at ~2ωD wavenumber position, when sp2 bonded carbon atoms are present [31]. This peak exhibits a strong dispersive behavior as a function of excitation energy [31]. DiLeo et al. show decreasing the intensity of G′ band directly with mass fraction decreases of the MWNTs is also evident [35].

Since the G′ band results from a two-phonon process, it is plausible that its intensity should decrease as the sample becomes less ordered (i.e., more impurities present) not allowing for the coupling effect which is necessary for the two-phonon process [35]. The relative changes in the various peak intensities clearly correlate with the amount of carbonaceous impurities [35]. In our samples, the Raman intensities are summarized in Table 3. A change in the D/G area ratio of MWNTs in a covalent functionalization is possible by modifying the external walls of carbon nanotubes [23] and, as a function of MWNts purity, there are changes in the Raman peak ratios (i.e., D/G, G’/G, and G’/D). Further evidence of this is the G′/D ratio, since the G’ band is a specific indicator of MWNTs purity and the D band is due to nano-carbons (presence of disordered carbons phases), therefore, the higher the value, the purer.

The oxidation of multiwalled carbon nanotubes (MWOHs) was also confirmed by FTIR. The vibrational normal modes of MWNTs in infrared spectroscopy are and , near to 868 and 1575 cm−1 [36]; however, Kuhlmann et al. [37] refer to 850 and 1590 cm−1. In our samples, these modes are present, with a slight shoulder at ~1580 cm−1 for and ~800 cm−1 for . In Figure 3 the spectra confirmed the oxidation of multiwalled carbon nanotubes [38]. The 1726 cm−1 band is due to (C=O) present in the carboxylic groups (COOH). At a wavenumber of 1466 cm−1 appears the vibration δ(O–H) present in the carboxyl groups; however, in this zone the vibrations (C–O) are present too. Also, according to Stobinski et al. [39], the spectra indicate intense bands at 3443 cm−1 (stretching vibrations of isolated surface –OH moieties and/or –OH in carboxyl groups and in sorbed water). The shifts in characteristic wavenumbers toward lower wavenumbers indicate the presence of strong hydrogen bonds between –OH groups. The bands in the 1750–1550 cm−1 range can be assigned to C=O groups in different environments (carboxylic acid, ketone/quinone) and to C=C in aromatic rings, whereas the bands in the range of 1300–950 cm−1 show the presence of C–O bonds in various chemical surroundings [33]. A band near 1466 cm−1 is due to sorbed water (OH in-plane deformation), and overlapping bands in region characteristic for C–O moiety, for example, C–O–C (structure of oxides, oxygen bridges, etc.) are present in the FTIR transmission spectra of multiwalled carbon nanotubes pristine. The relative increase and partial separation of bands in the 1250–950 cm−1 wave region point to increase in the amounts of hydrated surface oxides (O–H deformation and C–O stretching combination in surface phenols, hydroquinones, and aromatic carboxylic acids).

Figure 4 shows the characteristic vibrations of dialyzed keratin according to the amino acids present in the structure of the keratin [4043]. The protein IR absorption is visible as an amino acid side-chain absorption and normal modes of the amide group. Amino acid side-chain absorption provides valuable information when the mechanism of protein reactions is investigated. This is because side chains are often at the heart of the molecular reaction mechanism. The vibrational normal modes of the amide groups are due to the contribution of the internal coordinates, when the amide groups are incorporated into a polypeptide [41, 43]. In Figure 4 of the keratin dialyzed spectrum, there is no evidence of cysteine v(SH) in the 2550–2600 cm−1 region of the infrared spectrum but there are carbonyl groups of protonated carboxyl groups (1710–1790 cm−1) because, in the dialyzed keratin, the cysteine is oxidized. Intensity values for the vibrations correspond to cysteic acid monoxide (1077 cm−1), cysteic acid dioxide (1125 cm−1) and (S–O) (1170 cm−1).

The aliphatic moieties of amino acid side chains give rise to several absorbance bands of medium to weak intensity. While the δas(CH3), the δ(CH2), and the δs(CH3) vibrations between 1445 and 1480, 1425 and 1475, and near to 1385 cm−1 are more intense when next to a C–O group [41], the δ(CH) is representative between 1315 and 1350 cm−1 [41]. The frequency of the δ(CH2) vibration of aspartic acid and glutamic acid residues is sensitive to the protonation state and the δs(CH3) vibration to the branching of the hydrocarbon chain [44]. Additionally, the characteristic infrared bands of these amino acid side chains are to 1720 of C=O stretching and 1560 of CO2– asymmetric stretching [45].

Of the different structures of aliphatic amino acids, the proline (Pro) is remarkable for not forming the usual amide group with the amino acid that precedes in the sequence, but rather an N,N-disubstituted amide group due to the additional linkage of the side chain to the amide N-atom. The Pro amino acid is present with (CN) band near 1430 cm−1 [44].

Of the aromatic side chains, the tyrosine is a relatively strong infrared absorber due to its polar character, and its bands originated from the (CC), the (C–O), and the δ(COH) mode near 1517, 1238, and 1209 cm−1. The phenylalanine band is covered by a peak of greater intensity, which is not observable at the intensity of 1498 cm−1; however, at 1334 and 1455 cm−1, the tryptophan is seen with a little intensity, due to the small amount in grams of this amino acid in keratin [20, 44].

The bond peptide of the protein is characteristic of the amide groups. The amide II stretching occurs by bending in the plane δ(NH) group and (CN), with contributions in the planes from (CC), δ(CO), and (NC). The amide III stretching depends on the chain structure with the combination of δ(NH) in the plane, (CN), (CC), and δ(CO). In the above, at 3293 cm−1 is clear the stretching (NH) of amide A, and 3069 cm−1 belongs to (NH) of amide B [20]; therefore, this is related to the β or extended polypeptide chain, class of fibrous proteins [46]. It is exclusively localized on the NH group and is therefore in proteins insensitive to the conformation of the polypeptide backbone [41].

Near to 2930 and 2850 cm−1 are present characteristic infrared bands of aliphatic hydrocarbons of methylene asymmetric C–H stretching and symmetric C–H stretching, respectively [45]. However, for keratin protein, the methyl (CH3) asymmetric and symmetric modes are observed at 2955 and 2933 cm−1, respectively, and the methylene (CH2) asymmetric and symmetric modes at 2875 and 2855 cm−1 [47]. Therefore, similar to the above, the band at 2929 cm−1 is assigned to (CH3), and the band at 2858 cm−1 corresponds to (CH2); both signals originated from keratin and can be assigned to the aliphatic chain, that is, increased with grafting in carbon nanotubes, similar to chemical modification of keratin biofibers by graft polymerization of methyl methacrylate using redox initiation [20]. The infrared spectrum has a slightly parallel amide I component at 1660 cm−1. This band has been reasonably assigned to polypeptide chains; in an unordered state corresponding slightly perpendicular amide II bands are found at 1544 cm−1  [42, 48, 49]; these peaks were assigned to a α-helical component [41, 42].

The 1800–1500 cm−1 region (Figure 5) contains the important amide I and II modes, (CONH) and δ(NH2), respectively, of the keratin. A strong band was reported in the 1652 cm−1 and was assigned to the (C=O) stretching amide I vibration band, showing that the α-helix exists predominantly conformation [50]. Conversely, a strong band at 1666 cm−1 corresponds to the (C=O) stretching amide I vibration of a β-sheet conformation. Therefore, it is possible to observe a weaker shoulder on this strong band for the amide I β-sheet conformation, indicating that the keratin contained both α-helix and β-sheet conformations [50]. In summary, amide I is in the range of 3360–3340 for asymmetric NH2 stretching, 3190–3170 for symmetric NH2 stretching, 1680–1660 for C=O stretching, and 1650–1620 for NH2 bending [45].

A strong amide II band is observed at 1546 cm−1 and a weaker shoulder at 1521 cm−1. Peptides and proteins with an antiparallel β-sheet structure have strong amide II bands at 1534 cm−1; a parallel β-sheet structure is found at 1646 cm−1. Inspection of both amide I and II generally provides little help in distinguishing between turn and sheet conformation. In favorable situations, however, the amide II band may assist in assigning random structure and allow a more accurate estimate of the helix and random components [51].

Spectral features in the 1000–450 cm−1 regions (Figure 5) are the C–S stretching modes in the spectra of alkyl thiols appearing in the 730–620 cm−1 region. For methionine, there is a correlation between the frequency of the C–S stretching vibration and the conformation of the –S–CH3 group relative to the CH2–CH2– groups. In addition to the bands of amide I and III, two bands at 960–930 cm−1 region have been reported as being characteristic of α-helical keratin conformation.

In Figure 6 we compare the spectra of keratin, multiwalled carbon nanotubes, multiwalled carbon nanotubes oxidized, and three samples of experiments at 75°C with MWOHs/keratin 1 : 1 wt/wt (CK7511), 65°C with MWOHs/keratin 1 : 1 wt/wt (CK6511), and 55°C with MWOHs/keratin 1 : 3 wt/wt (CK5513). Based on the characteristic peaks of keratin, MWNTs pristine, and MWOHs spectra, we can observe the covalent bond between the hydroxyl from carbonyl and carboxyl of MWOHs with different functional groups of keratin dialyzed. In samples CK7511, CK6511 and CK5513, it is possible see (SO3) at 1220 and 1035 cm−1 due to the stretching of sulphur-oxygen [52, 53]. On the other hand, at 3443 cm−1 (associated with O–H another at 1466 cm−1) in the samples grafting, we can see a widening in the spectra in this wavenumber, where it is clear the presence of (NH) at 3293 cm−1, that is, invisible in MWOHs and MWNTs.

To support the presence of keratin in the walls of carbon nanotubes, the studies of elemental analysis by combustion and X-ray Photoelectron Spectrometer are the key to that. With a quality control, the sulfanilamide was used as standard. The experimental results were N = 16.24%, C = 41.79%, H = 4.64%, and S = 18.64%, with the theoretical values of N = 16.25%, C = 41.81%, H = 4.65%, and S = 18.62%, and the recovery percent is 99.94%, 99.95%, 99.78%, and 100.11% for N, C, H, and S, respectively. Therefore, for the samples of MWNTs, MWOHs, and the most important assay (CK7511, CK5513, and CK6511), the results are shown in Table 4.

The presence of nitrogen in the samples is due to peptide bend [44], on the other hand, the presence of sulfur is due to cysteine containing amino acid responsible for the sulfur-sulfur bonding in the keratin [54]. The percentages of this characterization are lower to Bradford assay because the studies of elemental analysis by combustion are only the elemental percentage of nitrogen, hydrogen, carbon, and sulfur; however, Bradford assay determines the quantity of protein on the samples. With that analysis, the reduced percentage of nitrogen and sulfur elemental with relation to the amount of protein with Bradford assay is clear.

X-ray photoelectron spectroscopy (XPS) technique can give information about the chemical structure of carbon nanotubes [36]. But the most widely used data refers to the structure modification of the CNT walls due to the chemical interaction with organic compounds [36].

Compared to keratin-g-MWOHs samples with MWNTs and MWOHs in wide range of 250 to 540 eV (Figure 7), the nitrogen into the nanotube structure is appreciable near to 400 eV [36, 55], characteristic of peptide bend of keratin [46, 47, 56], and viewed very clearly in narrow range of 406–396 eV. On the other hand, in the narrow range of 294–281 eV showed in Figure 7, the C1s peak exhibits a width according to oxidized and grafted the keratin on carbon nanotubes due to different hybridization sp2 and sp3 of the carbon with different elements of protein [57]. In a like manner, in Figure 7, the O1s peak exhibits the same behavior in the range of 537–528 eV.

4. Conclusions

The procedure developed in this paper was successful in grafting, in a controlled fashion, of keratin, obtained directly from chicken feathers, onto standard carbon nanotubes, provided they are properly functionalized. The spectroscopy analysis, both FTIR and Raman, in addition to confirming the covalent bonding involved, allowed to deduce the structural changes, both chemically and crystallographically speaking, associated with the grafting technology, which lead to variation on the properties of the resulting hybrid (i.e., organic-inorganic). Finally, since our results showed the details of the interactions, the method developed could be employed for grafting other natural proteins onto carbonaceous nanostructures.

Acknowledgments

The authors are indebted to Dr. Genoveva Hernández-Padrón, Mrs. Carmen Vazquez, Maria Lourdes Palma Tirado, and Mrs. Alejandra Núñez Pineda of Universidad Nacional Autonoma de Mexico and Dr. Gustavo López Téllez of Universidad Autonoma del Estado de Mexico for their technical support with the synthesis, FTIR, Raman, elemental analysis by combustion, and XPS studies. They thank the National Council of Science and Technology of Mexico (CONACyT) and FOMIX-CONACYT (ZAC-2010-C04-149908) for their financial support.