Effect of Various Aminated Single-Walled Carbon Nanotubes on the

Mar 29, 2011 - Atousa Aris , Akbar Shojaei , and Reza Bagheri. Industrial & Engineering Chemistry .... Gary P. Halada , Alexander Orlov. 2018,225-239 ...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/JPCC

Effect of Various Aminated Single-Walled Carbon Nanotubes on the Epoxy Cross-Linking Reactions J. M. Gonzalez-Domínguez,† M. Gonzalez,† A. Anson-Casaos,† A. M. Díez-Pascual,‡ M. A. Gomez,‡ and M. T. Martínez*,† †

Carbon Nanostructures and Nanotechnology Group, Instituto de Carboquímica (CSIC), Miguel Luesma Castan 4, 50018 Zaragoza, Spain ‡ Department of Polymer Physics and Engineering, Institute of Polymer Science and Technology (CSIC), Juan de la Cierva 3, 28006 Madrid, Spain

bS Supporting Information ABSTRACT: Arc discharge single-walled carbon nanotubes (SWNTs) have been functionalized with terminal primary amines through four experimental approaches that induce either tip/edge or sidewall functionalization. A full characterization of functionalized SWNTs has been carried out through several techniques. The aminated SWNTs have been solvent-free incorporated into a trifunctional high-performance epoxy system, and the mixture has been studied using differential scanning calorimetry (DSC) and isoconversional kinetic calculations. The thermodynamic parameters obtained from DSC and the isoconversional plots reveal that SWNTs aminated through sidewall addition reactions participate in cross-linking processes. Unfunctionalized SWNTs or those aminated on tips and edges do not show any evidence of a covalent anchoring to the matrix. The type and chemical nature of the amine moieties play an important role in the dispersibility and reactivity of SWNTs in the epoxy, but the functionalization degree and metal content do not affect the epoxy curing process. This study sets the way toward the choice of proper amine moieties for controlling the thermomechanical properties of epoxy/SWNT composites.

1. INTRODUCTION Because of their superior physical properties, carbon nanotubes (CNTs) offer unique possibilities for a wide variety of applications, including molecular electronics, biotechnology, sensors, and particularly as the polymer matrix filler for high performance composites. With respect to nanocomposite materials based on thermosetting epoxy matrices, the use of CNTs as the fillers is of a critical interest in lightweight aerospace and automotive applications; hence a huge research effort has been undertaken in this field.13 The incorporation of CNTs into an epoxy system is carried out before the curing reaction, and their presence affects the cross-linking reactions taking place during the curing stage of manufacturing. The in-depth study of these effects has been considered as highly necessary4,5 to identify the filler impact during the formation of the thermosetting architecture, whose glass transition temperature (Tg) may be altered. The thermomechanical stability and overall performance in the working conditions can be influenced or even tailored through the filler choice. Several studies have been performed with this aim. In different epoxy systems, an accelerating effect of CNTs at the early stages of the curing process has generally been found610 by using isothermal or nonisothermal differential scanning calorimetry (DSC) measurements. The most accepted explanation for this fact seems to be the high thermal conductivity of the filler6 or the presence of r 2011 American Chemical Society

surface oxygen groups,7,9 which may act as catalysts for epoxide rings opening. The incorporation of CNTs into epoxy matrices also seems to lower the overall cross-linking degree of the thermosetting network8,10 as inferred from curing enthalpies and Tg values. However, full understanding of this issue remains still unreached,4 especially regarding Tg changes, due to the disparity of some results and the wide variety of materials and experimental procedures employed. According to the literature, the use of solvents can lead to a decrease of Tg values, even after careful evaporation.4 On the other hand, one of the most interesting CNT assets is their ability to undergo a very large variety of chemical reactions, which can attach almost any desirable moiety. Over the past years, CNTs have been treated as nanoscale reagents in a wide variety of different chemical approaches, enabled by the reactivity of their graphene walls. This fact has been extensively analyzed in many review papers1114 and represents a remarkable upgrade toward specific applications that require chemical modifications and/or the purification of pristine tubes. Chemical covalent functionalization is believed to counteract the intertube stacking forces and to lead to Received: November 12, 2010 Revised: January 23, 2011 Published: March 29, 2011 7238

dx.doi.org/10.1021/jp110830y | J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

ARTICLE

Scheme 1. Reaction Routes for the Different Amino Derivatizations Applied to Arc SWNTs

debundling. This effect may be caused by the intercalation of the attached moieties.11 In the field of CNT composites, there is a great interest in the incorporation of covalently functionalized fillers.3 As a general trend, covalent modifications are pursued to enhance the solubility of CNTs in organic solvents, which is exploited to achieve homogeneous dispersions in the polymer matrix. In epoxy composites, the primary amine functional group is of special interest due to its high nucleophilicity and reactivity similar to that of hardener amines. In a few works, amine-functionalized CNTs were integrated into an epoxy matrix using a solvent-assisted integration method, and their effect on the curing process was evaluated by DSC.10,15 In these works, it was reported that the effect of aminated CNTs was a reduction in the curing enthalpy and an increase in the activation energy. However, there has been no clear attempt to evaluate the effect of covalent functionalities on the dispersion behavior of CNTs into epoxy matrices in the absence of solvents, which would be of particular interest for the evaluation of the chemical affinity between the functionalized fillers and the host matrix. There is also a lack of systematic studies with different functionalization routes and different amine moieties, coupled to a full characterization of the functionalized filler, to properly assess the impact of CNT amine groups in the epoxy matrix. In this Article, single-walled carbon nanotubes (SWNTs) have been covalently functionalized with different moieties ending in a primary amine, following diverse chemical routes. The moieties are intended to have different chemical nature but inevitably lead to different functionalization degrees. Aminated SWNTs have been directly integrated without using solvents into a trifunctional epoxy system for aerospace applications. The effect of the different functionalized SWNTs on the curing reaction of the epoxy matrix and the Tg values of the as-prepared nanocomposites has been studied using nonisothermal DSC scans and isoconversional kinetic calculations. A previous characterization of the functionalized filler has also been made to correlate the SWNTs effect with the observed phenomena.

2. EXPERIMENTAL SECTION 2.1. Materials and Reagents. SWNTs were produced in our research center by the arc-discharge technique, using graphite as

the carbon source, and Ni/Y as metal catalysts in 4/1 atomic ratio. Solvents were purchased from Sigma Aldrich in reagent grade and used as received. The other reagents (purity g97%) were also obtained from Sigma Aldrich and used as received. The epoxy system was an aerospace-grade resin based on a trifunctional precursor, triglycidyl-p-aminophenol (TGAP), and 4,40 diaminodiphenyl sulfone (DDS) was the curing agent. Both were supplied by Huntsman and used as received. 2.2. Functionalization of SWNTs. Within the vast field of CNT chemical functionalization, several routes stand out for their feasibility and versatile applications. In this Article, the most significant amination approaches have been selected and carried out. The most commonly employed amination route consists of the nucleophilic attack to the carboxylic groups present in CNTs after an oxidation treatment, to produce an ester or amide. This has been our first approach to obtain covalently aminated SWNTs (SWNT-oxa). After a proper activation of carboxylic groups and the reaction with an amine-terminated nucleophile, the amine functional group mainly sets on CNT tips and on the edges of shortened open-ended CNTs.2 Other “less-classical” approaches take advantage of the electronic density on CNT walls to make them react via nucleophilic, electrophilic, cyclo-, or radical additions. A good example of radical addition is represented by the thermal decomposition of diacyl peroxides on SWNTs surface.16 Another breakthrough in the chemistry of SWNTs consists of the SWNT fluorination with F2, which provides high sidewall reactivity, opening interesting alternatives in the subsequent derivatization.17 The reductive intercalation of alkaline ions in polar aprotic solvents, which implies the negative charging of SWNT sidewalls, has been used for further derivatizations with alkyl radicals. After the pioneer works by Billups et al.18 and Penicaud et al.,19 a more recent article20 proposed the functionalization of the reduced SWNTs with diacyl peroxides at room temperature as a facile means to derivatizate SWNTs. This approach has been carried out as the second amination route in the present Article (SWNT-nfp). In 2001,21 J. M. Tour and co-workers developed the derivatization of SWNTs with in situ generated aryl diazonium salts, a versatile and easy way to functionalize nanotubes with a large variety of chemical 7239

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C species containing an aromatic ring as spacer between CNT and the grafted moiety. We have recently employed this methodology to obtain amine-functionalized SWNT buckypapers.22 Prato and his research group developed the 1,3-dipolar cycloaddition of azomethine ylides onto CNT sidewalls as a way to achieve organic functionalization and to greatly increase the CNT solubility and processability for many applications.23 The fourth functionalization approach is an aminoacid functionalization through dipolar cycloaddition (SWNT-cda). A general overview of the performed functionalizations is displayed in Scheme 1. 2.2.1. Route 1: Functionalization via Acid Treatment þ Amide Formation (SWNT-oxa) Oxidative Acid Treatment. Around 0.5 g of as-grown SWNTs and 150 mL of a 1.5 M HNO3 solution were placed in a roundbottom flask. The system was heated to the boiling temperature and was refluxed for 2 h under constant magnetic stirring. Subsequently, the reaction medium was poured into several centrifugation vials and centrifuged at 3500 rpm for 15 min. The supernatant liquid was decanted off, and the vials were refilled with aqueous HCl (pH = 12), bath sonicated for 30 min, and centrifuged again under identical conditions. The supernatant was then decanted off, and the SWNTs were swept with deionized water, vacuum-filtered through a 1.2 μm polycarbonate membrane, and thoroughly rinsed with deionized water until the pH of the falling filtrate was neutral. As-produced oxidized SWNTs were dried at 60 °C in a vacuum oven for 24 h. These nanotubes have a high content in oxygen groups, especially carboxylic groups. Activation of Carboxylated SWNTs. 200 mg of nitric acidtreated SWNTs was bath sonicated in ∼10 mL of DMF inside a round-bottom flask, and then 40 mL of thionyl chloride was carefully added. The flask was coupled to a reflux, and the system was kept at 120 °C for 24 h under constant magnetic stirring. Afterward, the product was vacuum-filtered through a 0.1 μm pore size PTFE membrane and washed with anhydrous THF. Finally, it was dried under vacuum at room temperature for 2 h. Amidation of Activated SWNTs. 225 mg of freshly activated SWNTs was placed in a round-bottom Schlenk and bath sonicated in 100 mL of anhydrous DMF for 15 min. The suspended SWNTs were magnetically stirred under argon atmosphere at 90 °C for about 1 h. Next, ∼3 g of N-Boc-1,6diaminohexane was added by injection through a septum cap using a purged syringe. The system was allowed to react under unchanged conditions for 4 days. The product was vacuumfiltrated through a 0.1 μm pore size PTFE membrane and copiously washed with methanol. Drying was carried out at 60 °C in a vacuum oven. 2.2.2. Route 2: Functionalization via Alkaline Reduction þ Diacyl Peroxide (SWNT-nfp) Synthesis of the N-Fmoc-6-Aminohexanoyl Peroxide (NFP). The synthesis of this organic peroxide was achieved through reaction of hydrogen peroxide with the acyl chloride derivative of the N-Fmoc-6-aminohexanoic acid. Following the experimental protocol described by Martinez-Rubi et al.,20 a CH2Cl2 solution of N-Fmoc-6-aminohexanoic acid (1.76 g) was first acylated with SOCl2 (3.6 mL) and then precipitated in hexane. The acyl chloride (1 g) was subsequently dissolved at 0 °C in a chloroform/diethyl ether mixture, and a simultaneous dropwise addition of hydrogen peroxide (206 μL, 30% concentration) and pyridine (260 μL) was made. The NFP precipitate was dissolved in chloroform, purified by liquid-phase extraction with cold

ARTICLE

water, and then the pure product was precipitated from the organic phase by addition of diethyl ether. Reduction of SWCNTs with Sodium Naphthalide Radical Ion. 400 mg of arc as-grown SWNTs was bath-sonicated for 30 min in 30 mL of anhydrous tetrahydrofuran (THF) inside a round-bottom Schlenk flask, under an argon (Ar) atmosphere achieved by connection to a manifold with constant gas flow. After sonication, a fine suspension of the SWNTs was observed, which was kept under argon at a constant magnetic stirring. Parallel, sodium naphthalide radical ion was prepared by bathsonicating around 80 mg of small sodium pieces in a naphthaleneTHF solution (7 mg/mL, anhydrous THF) for 1520 min in Ar atmosphere. A dark green color appeared, indicating the successful radical ion synthesis. The sodiumnaphthalide ion was added to the SWNTTHF suspension with the aid of a glass syringe previously purged with Ar. Care must be taken to prevent the green solution contacting the outer air atmosphere; otherwise, the radical ion is deactivated, hence losing its color. About 70 mL of the radical ion solution was slowly injected into the SWNTTHF suspension, through a septum cap, until a persistent green color was observed in the medium. The mixture was bath-sonicated for 1 h and left overnight at room temperature under Ar and magnetic stirring. It was then placed in open air and poured into centrifuge vials. Centrifugation was applied at 8000 rpm for 30 min, and then the supernatants were decanted and discarded. The vials were filled with wet THF, sonicated in bath for 30 min, and centrifuged under identical conditions. Again, supernatants were decanted and discarded. The process was subsequently repeated according to the following sequence of solvents: methanol (two times), distilled water, HCl 5% v/v, and distilled water. SWNTs were collected and filtered through a 0.1 μm pore size polytetraflouroethylene (PTFE) membrane, washed with isopropanol, methanol, and acetone, and finally dried under vacuum at 60 °C. Reaction of Reduced SWNTs with NFP. 150 mg of the asprocessed SWNTs was taken and subjected to a reduction procedure identical to that described in the previous section, with an identical volume of the radical ion green solution. After the overnight reduction, about 1 g of NFP was added to the mixture, and a sudden loss of the green color was observed. After a few hours of magnetic stirring at room temperature, the reaction medium was removed from the Schlenk and subjected to the same centrifugation sequence, as described above. Deprotection of the Fmoc Group. In a typical deprotection, 100 mg of NFP-functionalized SWNTs was submerged in 10 mL of a DMF/piperidine mixture (80:20 v/v) and bath-sonicated for a few minutes. Next, magnetic stirring was applied for 1 h at room temperature. The medium was filtered through 0.1 μm pore size PTFE membrane and washed with N,N0 -dimethylformamide (DMF) until the filtrate fell colorless. Finally, it was rinsed with diethyl ether and dried under vacuum at room temperature. 2.2.3. Route 3: Functionalization via in Situ Diazonium Compounds (SWNT-dba). 100 mg of as-grown SWNTs were bath sonicated in 25 mL of DMF for 1 h and then tip sonicated for 30 min (60% amplitude 0.5 cycle time). Separately, 1.3 mL of 4-amino benzylamine (10 mmol) was dissolved into 25 mL of acetonitrile and bubbled with argon for a few minutes. Both liquids were blended in a glass vial and kept at 60 °C and constant magnetic stirring. A small slit was left in the vial seal to avoid overpressure. After the system temperature was stabilized, 1 mL of isoamyl nitrite was added, and it was left overnight at 60 °C with constant stirring. The product was vacuum-filtered through 7240

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C a PTFE 0.1 μm pore size membrane and washed with DMF until the filtrate fell colorless. The functionalized SWNTs were rinsed with diethyl ether and dried under vacuum at room temperature. 2.2.4. Route 4: Functionalization via 1,3-Dipolar Cycloaddition of Azomethine Ylides (SWNT-dca) Synthesis of the Amine-Terminated R-Aminoacid. The synthesis of the N-Boc-protected R-aminoacid, N-Boc-ethyleneglycol-bis(2-aminoethylether)-N0 -acetic acid, was carried out through the synthetic route proposed by Kordatos et al.24 in which a commercial symmetric diamine, ethylene glycol-bis(2aminoethylether), is monoprotected with Boc-anhydride, then alkylated with benzyl-2-bromoacetate, and finally reduced with hydrogen. Reaction between the R-Aminoacid, an Aldehyde, and the SWNTs. The 1,3-dipolar cycloaddition reaction was performed with the aforemementioned R-aminoacid and paraformaldehyde. In a typical experiment, 100 mg of pristine arc SWCNTs was mixed with 100 mL of DMF inside a 200 mL round-bottom flask and sonicated in an ultrasound bath for about 30 min. Subsequently, 150 mg of the R-aminoacid and 150 mg of paraformaldehyde were added to the resuting suspension. After 2 min of magnetic stirring at room temperature, the mixture was bath sonicated for 30 min. Next, the flask was set in an oil bath at 115 °C and refluxed under a constant magnetic stirring for 4 days. During the course of the reaction, three more additions of the same amounts of the R-aminoacid and paraformaldehyde were made in periods of approximately 24 h. These additions were carried out previously allowing the flask to cool to room temperature and were followed by 30 min of bath sonication. The reaction medium was vacuum-filtered (0.1 μm pore size PTFE membrane) after a period of 4 days.. The solid filtrated material was washed by successive submersions in 100 mL of solvents, bath sonication, and filtration in the same filter type. The solvent washing sequence was DMF (four times) and methanol (two times). After the last methanol filtration, the sample was rinsed with diethyl ether (2  25 mL) in the filtration equipment. The functionalized material was dried overnight under vacuum at room temperature. 2.2.5. Deprotection of the Boc Groups. Amine-SWNTs functionalized through both the acid þ acyl route and the 1,3-dipolar cycloaddition route were obtained as N-Boc-protected products and required the detachment of the protective group to have the free amine. Around 100 mg of amino-Boc SWNTs was bath sonicated for 15 min in 25 mL of an HCl/1,4-dioxane mixture (4% v/v HCl) and stirred at room temperature for 2 h. The reaction product was vacuum-filtered through a 0.1 μm pore size PTFE membrane, thoroughly washed with 1,4-dioxane and diethyl ether, and vacuum-dried at room temperature overnight. The deprotected SWNTs were then sonicated for 15 min in 100 mL of a diluted NaOH solution and stirred at room temperature for 24 h to neutralize the ammonium salt and yield the free primary amine groups. The aqueous suspension was filtered through a 1.2 μm pore size polycarbonate membrane, rinsed with deionized water (until the falling liquid pH was neutral), and finally dried at room temperature under vacuum overnight. 2.3. Nanocomposite Blend Preparation. The integration of amino-functionalized SWNTs into the trifunctional epoxy system was carried out using a previously optimized solvent-free procedure.25,26 Because no solvents are used, the chemical affinity of functionalized SWNTs is directly manifested, and thus the obtained DSC parameters can be exclusively attributed to the

ARTICLE

SWNTepoxy interaction. The integration was carried out by stirring SWNTs in 1 g of TGAP at 60 °C for 45 min and then tip sonicating (Hielscher DRH-UP400S, 50% amplitude, continuous application) the mixture for 15 min. Afterward, DDS incorporation (stoichiometric ratio, 0.67 g per TGAP gram) was stepwise made within a period no longer than 30 min with constant stirring at 60 °C. Samples containing TGAP and DDS had a SWNT loading of 0.5 wt % (8.35 mg of SWNTs into 1.67 g of TGAPþDDS). In the absence of DDS, the same amount of SWNTs was maintained, and the SWNT loading was recalculated. 2.4. Experimental Techniques. Raman spectroscopy was performed using a HORIBA Jobin Yvon Raman spectrometer model HR 800 UV, working with a 532 nm laser. The spectra were analyzed, baseline corrected, and normalized with the NGS LabSpec software. Thermogravimetric analyses (TGA) were performed under nitrogen atmosphere (60 mL/min flow) and registered with a SETARAM Setsys Evolution 16/18 device at a heating rate of 10 °C/min, in the range from room temperature to 900 °C. Fourier transformed infrared (FTIR) spectra were obtained with a Bruker Vertex 70 instrument. All of the samples were prepared as pellets using spectroscopic grade KBr. The amount of free primary amine groups was quantified using the Kaiser test,27 which was designed to monitor solid-state protein coupling. Later, it has successfully been applied to characterize amino-functionalized CNTs.28 The intramolecular redox reaction undergone by the ninhydrin molecules in the presence of primary amines (Kaiser reaction) was carried out using the standard procedure on ∼2 mg of amino-functionalized SWNTs. The reaction extent was measured by UVvisible spectroscopy (Shimadzu UV-2401 PC spectrophotometer) on the supernatant in the range of 450650 nm, and the maximum absorbance was used to quantitatively assess the number of free amines in functionalized SWNTs. The procedure was applied at least twice to ensure repeatability, and the reported are average values. Nickel and yttrium were quantitatively determined by the induced couple plasma-optical emission spectroscopy (ICPOES) method. Samples were first subjected to an alkaline fusion with sodium peroxide and then leached to an aqueous solution, which was analyzed in a Jobin-Ybon 2000 ICP instrument. Absorption spectroscopy in the near-infrared (NIR) region was performed using a Bruker VERTEX 70 spectrometer. An aqueous suspension of each SWNT sample (0.1 g/L) in sodium dodecyl benzene sulfonate (SDBS) surfactant (1 wt %) was made with the aid of an ultrasound bath. The suspended SWNTs were measured inside 2 mL quartz cells. The sample absorbance was adjusted within the range of 0.40.5, by dilution with the 1 wt % SDBS aqueous solution. Background correction was effected with the same surfactant solution. Differential scanning calorimetry (DSC) measurements were made in a Mettler DSC-823e equipment, calibrated using an indium standard (heat flow calibration) and an indiumlead-zinc standard (temperature calibration). Nonisothermal experiments were performed from room temperature to 320 °C (samples with curing agent) or 400 °C (samples without curing agent) at different heating rates (2, 5, 10, and 20 °C/min) on ∼10 mg (with curing agent) or ∼3 mg (without curing agent) of sample, exactly weighed, placed into standard 40 μL aluminum crucibles, under a 100 mL/min flow of N 2 . T g values were taken as the inflection point of the heating DSC curves after nonisothermal curing inside the equipment, in the subsequent cooling 7241

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

ARTICLE

Figure 1. TGA plots for SWNT-asg and functionalized SWNT samples.

Table 1. Thermogravimetric Data and Kaiser Test Results extent of functionalization TGA weight loss in % (from 100 to 700 °C, inert atmosphere)

a

X values (one functional

amine content

group per X carbon atoms)a

(μmol/gCNT)b

340

SWNT-asg SWNT-oxa

4.3 25.4

31

SWNT-nfp

17.1

34

64

SWNT-dba

9.0

88

198

SWNT-dca

11.0

115

20

Calculated from TGA data using eq 1. b Calculated from Kaiser test data as the difference between unfunctionalized (SWNT-asg) and functionalized samples.

cycle (at 20 °C/min). Enthalpy values (ΔH) were calculated from the area under the exothermic curing peak. The DSC data obtained at different heating rates were subjected to the modelfree calculations of the so-called “advanced isoconversional method” developed by Vyazovkin,29 which allows one to obtain activation energy (AE) plots versus the degree of conversion (R) and has extensively been applied to many epoxy systems and also to CNTepoxy composites.30,31

surface groups.32 The rest of the SWNT samples experience higher weight loss values, which follow the order: SWNT-oxa > SWNT-nfp > SWNT-dca > SWNT-dba. To calculate the extent of functionalization, weight loss values were employed together with the molecular weight of the different moieties, and the following equation was applied:

3. RESULTS AND DISCUSSION

where X stands for the number of carbon atoms in the SWNT sample per each covalent functional group, R (%) is the residual mass at 700 °C in the TGA plot, L (%) is the weight loss in the range of 100700 °C, and Mw is the molecular weight of the desorbed moieties. This calculation considers SWNT samples to be entirely constituted by carbon, which is indeed an approximation. Functional groups desorbed in the SWNT-asg sample were discounted in the calculations because they are not a direct consequence of the functionalization. In Table 1, X values for the functionalized samples are displayed, and the functionalization degree follows the order: SWNT-oxa > SWNT-nfp > SWNT-dba > SWNT-dca. Additionally, the amount of free amines was determined by the Kaiser test, whose results can be found in Table 1. The functionalization extent determined by the Kaiser test follows a trend identical to TGA data with the exception of

3.1. Characterization of Functionalized SWNTs. The degree of functionalization was determined from TGA experiments under inert atmosphere (Figure 1). Weight losses observed upon nonisothermal heating reveal the presence of covalently attached functional groups. Up to 100 °C a slight weight loss is detected, which corresponds to residual moisture within SWNT samples. This moisture content is especially visible in SWNT-oxa and SWNT-dca samples, consistent with their higher hydrophilicity as they bear large amounts of surface oxygen groups. The progressive desorption of the chemical groups starts at 100 °C. Weight losses between 100 and 700 °C for the functionalized SWNTs are listed in Table 1. As-grown SWNTs experience a small weight loss (4.3%), which comes from certain native

X ¼

7242

R ð%Þ 3 Mw ðg=molÞ L ð%Þ 3 12 g=mol

ð1Þ

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

ARTICLE

Figure 3. Raman spectra for the set of SWNT samples.

Table 2. NIR Purity Ratio and Metal Content for the Different SWNT Samples metal content NIR purity ratio

Figure 2. FTIR spectra for SWNT-asg and the functionalized SWNT samples.

SWNT-nfp, which appears to have less amine groups than expected. The classical oxidation-amidation approach generates the highest number of amines, while the 1,3-dipolar cycloaddition gives the lowest functionalization degree. The identification of the grafted functionalities was conducted through FTIR spectroscopy. Figure 2 shows solid-state FTIR spectra of all the SWNT samples. The as-grown SWNT sample (SWNT-asg) shows no relevant IR features except the band at 1559 cm1, which corresponds to the CdC phonon modes.33 A wide band at around 1190 cm1 (CO stretching vibration) is also visible, ascribed to the aforementioned native surface groups, and detected in TGA (Table 1). Oxidized SWNTs present a band at 1722 cm1 that can be attributed to the CdO stretching vibration of carboxylic groups. The presence of carboxylic groups can also be confirmed by the band at 1584 cm1, typical of the carboxylate anion.34,35 At 1217 cm1, the CO stretching vibrations of esters34 and phenolic groups can be observed.34,35 CNTs treated with nitric acid become a complex heterogeneous material with a wide variety of oxygen groups.36 All of the aminated SWNTs have a common band in the range of 16301640 cm1, which can be assigned to the NH deformation vibration of primary amine groups, and a double band at 2915/2848 cm1 (Csp3H vibrations). This doublet presents the lowest intensity in the SWNT-dca sample, which has the lowest functionalization degree. In the SWNT-oxa sample, the CdO stretching band downshifts to 1700 cm1, due to the formation of amide bonds, and the presence of unreacted groups is detected through the carboxylate band as well as the band at 1216 cm1. In SWNT-dca sample, a band at 1147 cm1 indicates the presence of aliphatic ether groups, and the band at 1206 cm1 corresponds to the CN stretching vibration in tertiary amines, both contained in the attached moiety (see Scheme 1). The SWNT-nfp sample has a band at 1447 cm1 consistent with the CH bending in methylene groups, and another at 748 cm1 that could be ascribed to the rocking vibrations of successive (CH2)n (n g 3).35 In the SWNT-dba sample, different bands

a

b

Ni (%)

Y (%)

SWNT-asg

0.027 /0.045

13.13

6.64

oxidized SWNTsc

0.047

3.94

0.65

SWNT-oxa

0.031

3.27

0.42

reduced SWNTd

0.063

SWNT-nfp

0.055

10.3

1.73

SWNT-dba SWNT-dca

0.052 0.053

6.14 12.1

14.5 2.03

a

NIR purity ratio of SWNT-asg employed for the nitric acid treatment and the subsequent amidation reaction. b NIR purity ratio of SWNT-asg employed for the rest of the functionalization procedures and for composite blends. c SWNT-asg oxidized with nitric acid (see Experimental Section). d Arc SWNTs after the first alkaline reduction cycle (see Experimental Section).

corresponding to the benzene ring of the attached moiety are observed: two located at 1515 and 1615 cm1 (CdC stretching) and another one at 814 cm1 (CH in-plane vibration of a 1:4 disubstituted benzene). As a consequence of covalent functionalization, SWNT Raman features are altered, particularly the D-band (13001400 cm1). This band is usually taken as the indicative of covalent functionalization,37 when comparing its relative intensity to the G-band (15001700 cm1). Generally, an increase in the D-band intensity is assumed as a successful covalent attachment onto SWNT surface. Thus, the G/D intensity ratio (after baseline correction) offers an idea of covalent modifications. As can be observed in Figure 3, the G/D ratios of functionalized SWNTs are significantly lower than that of the SWNT-asg sample. 3.2. SWNT Purity Control and Dispersibility. NIR purity data and the ICP metal content of the different SWNT samples are displayed in Table 2. The NIR purity ratio is a quantitative measurement of the carbonaceous purity of arc-grown SWNT samples.38 This method reliably assesses the purity of bulk SWNTs by only taking a representative portion of the sample, dispersing it in a liquid medium, and registering the absorption spectrum in the NIR region. The purity index is calculated as the ratio of the baseline subtracted peak area and the total peak area for the SWNT S22 7243

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

Figure 4. Example of DSC nonisothermal heatings for epoxy or epoxy/ SWNT blends in the presence () or absence (  ) of the curing agent.

transition between 7750 and 11 750 cm1. In our functionalized SWNTs, NIR purity ratios are visibly affected by the functionalization route. The nitric acid treatment causes a nearly 2-fold increase in the SWNT-asg purity, while the amidation step lowers this ratio to approximately the initial value. Sidewall addition reactions yield higher NIR ratios, particularly the alkaline reduction, which causes a ∼50% increase in the NIR ratio of SWNT-asg. The bundles exfoliation induced by negatively charging the SWNTs, combined with the centrifugation, could have caused the removal of carbonaceous impurities, hence increasing the NIR ratio. As-grown SWNTs possess a total metallic content of around 20 wt %, which corresponds to the catalysts remaining from the SWNT synthesis. As expected, the nitric acid treatment yields SWNTs with a noticeably reduced amount of metals (∼77% reduction). SWNTs aminated through carboxylic activation (SWNT-oxa) present metal content similar to that of nitric acid-treated SWNTs. The lower amount of metals present in the SWNT-oxa sample as compared to the rest of functionalized SWNTs is due to the nitric acid treatment that precedes the amination step. SWNTs functionalized through alkaline reduction (SWNT-nfp) and 1,3-dipolar cycloaddition (SWNT-dca) present a ∼3040% reduction in the metal content. While in SWNT-nfp, the removal of metals could arise from the negative charging and exfoliation of SWNTs (which may detach the impurities), in the SWNT-dca it has been reported as an effect of the organic functionalization.39 The dispersibility of SWNTs and the homogeneity of SWNT dispersion in the epoxy were evaluated through several ways. A primary macroscopic observation was made by direct inspection of the TGAP/SWNT mixtures. The black color intensity and the appearance (or not) of aggregates in the vial walls were the main features observed. After incorporation of DDS, the persistence of the black color in the mixture was checked macroscopically. Bad dispersed SWNTs presented a grayish color after adding DDS, instead of a fully black aspect. Besides, a microscopic observation

ARTICLE

was made with an optical microscope on the TGAP/SWNT mixtures before adding the DDS curing agent. The corresponding images are depicted in the Supporting Information. 3.3. Nonisothermal DSC Parameters. Nonisothermal DSC scans of the epoxy/SWNT mixtures led to substantially different profiles depending on the presence or absence of the DDS curing agent (Figure 4). In the presence of the curing agent, the system undergoes a typical cross-linking process upon heating. The crosslinking is due to the reaction of amines with oxirane rings by means of nucleophilic ring-opening, and it is composed of several stages (including primary amine addition, secondary amine addition, and etherification), which are often overlapped.40 In the absence of the curing agent, nonisothermal heating causes the autocuring of the TGAP monomer, through the formation of hydroxyl groups from oxirane rings, and the subsequent etherification with other oxirane rings. This mechanism is known to happen at higher temperatures than those involving the curing agent.41 The TGAP sample after a nonisothermal curing is a pulverulent black solid. Unlike in TGAP/ DDS systems, the TGAP autocuring overlaps with degradation at the exothermic peak temperatures, and thus Tg could not be measured in the absence of DDS. Thermodynamic parameters extracted from nonisothermal DSC scans at 5 °C/min are listed in Table 3. As nonfunctionalized SWNTs (SWNT-asg) are incorporated into the TGAP/ DDS epoxy, an increase in ΔH (17.3 J/g) and a very slight increase of Tg (1.6 °C) are observed. The latter is consistent with an increase in the rigidity of the system, and also with the fact that the nanofiller presence does not hinder the cross-linking process, unlike other authors’ findings in epoxy/SWNT composites with bifunctional epoxy monomers.8,10 The absence of organic solvents during our experimental procedure of preparation could be a suitable explanation for the disagreement. Besides, the as-made solvent-free integration of SWNT-asg results in apparently homogeneous mixtures with the epoxy probably due to favorable physical interactions (i.e., ππ stacking of aromatic species, polar interactions with SWNT native surface groups, etc.), which would enhance the intimate contact between matrix and filler. The increase in enthalpy could be due to the surface groups present on SWNT-asg, which undergo chemical changes during the curing process and thereby increase its reaction heat. As there is no practical change in the curing onset or maximum peak temperature (Tm), no catalytic effect is observed. This fact sheds light on the assumption that metal catalysts remaining in SWNT samples can affect the curing reaction,4 which does not seem to happen here. Because SWNT-asg have a high metal content (Table 2), the effects of functionalized SWNTs can be entirely ascribed to the functional groups and not to metals. Functionalized SWNT samples (SWNT-dba, SWNT-dca, and SWNT-nfp) exhibit a clear decrease of the curing enthalpy (of 18.8, 7.9, and 17.3 J/g, respectively), compatible with a decrease in the extent of curing conversion as the epoxide/amine ratio is modified.10 Additionally, the involvement of SWNT attached amines in the enthalpy changes can be confirmed by observing the SWNT-dca sample (which shows the lowest extent of functionalization), because it exhibits considerably lower ΔH decrease. On the other hand, Tg values show a noticeable increase in comparison with the neat matrix (of 7.3, 8.7, and 7.8 °C for composites containing SWNT-dba, SWNT-dca, and SWNT-nfp, respectively), indicating the successful cross-linking between the filler and the matrix, and hence an efficient covalent anchoring. Both onset and Tm values decrease, pointing to a catalytic effect of the amine groups in the epoxy medium. In contrast, the SWNT-oxa sample does 7244

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

ARTICLE

Table 3. DSC Parameters Obtained for Different TGAP/DDS/SWNT and TGAP/SWNT Mixtures Heated at a Rate of 5 °C/min with DDS

without DDS

fillera

onset (°C)

Tm (°C)

ΔH (J/g)

Tg (°C)b

onset (°C)

Tm (°C)

ΔH (J/g)

neat epoxy

167.5

199.7

203.9

209.9

301.3

312.9

497.7

SWNT-asg

167.5

199.0

220.6

211.5

298.0

309.2

439.5

SWNT-oxa

168.5

200.7

256.9

211.0

311.2

323.6

509.2

SWNT-nfp

162.3

197.6

186.6

217.7

298.7

309.7

364.0

SWNT-dba

166.3

198.1

185.1

217.2

297.7

309.0

440.0

SWNT-dca

164.6

196.4

196.0

218.6

295.8

306.6

351.8

SWNT content is 0.5 wt % (in the presence of DDS) or 0.835 wt % (in the absence of DDS). b Tg values are obtained during the cooling stage (20 °C/ min) after nonisothermal curing. a

not show catalytic behavior or decrease in enthalpy. Its DSC parameters are nearly those of SWNT-asg but with a higher ΔH. Because SWNT-oxa has the highest functionalization degree, this striking fact can only be explained by the extremely bad dispersibility of nitric acid-treated SWNTs into epoxy throughout solvent-free techniques.26 The SWNT-oxa amino groups are unable to develop their catalytic effect because they do not reach proper contact with the parent matrix, albeit the high ΔH suggests that oxygen groups (highly abundant in SWNT-oxa) experience thermal modifications during the curing process, raising the reaction heat. In the absence of the curing agent (Table 3), onset and Tm changes are small except for the TGAP/SWNT-oxa mixture, in which both parameters increase by about 10 °C, indicating no catalytic behavior. The rest of the samples show catalytic effects that can only be explained by the amino groups actuation because no DDS is present. For all of the TGAP mixtures with sidewall functionalized SWNTs, a large decrease in the ΔH (up to 133.7 J/g, for SWNT-nfp sample) is observed. This is perfectly consistent with an enhanced dispersion of functionalized SWNTs in the TGAP monomer by the solvent-free technique. Well-dispersed SWNTs can act as a physical hindrance to the monomer mobility and thus lower the reaction heat.8 This premise has been used for the DSC evaluation of the SWNT dispersion in epoxy matrices,42 considering the total heat of reaction being influenced mainly by well-dispersed SWNTs. In this sense, the incorporation of sidewall-functionalized SWNTs into the epoxy matrix leads to an improved state of dispersion (see the Supporting Information), as compared to SWNT-asg or SWNT-oxa, which is reflected in the DSC parameters. The enhanced dispersion would be directly caused by the attached moieties and could also be argued in the TGAP/DDS-based composites. 3.4. Isoconversional Kinetics Characterization. Epoxy/ SWNT mixtures were subjected to nonisothermal curing through DSC scans at different heating rates to apply the advanced isoconversional kinetic calculations. AE versus R plots for TGAP/ DDS/SWNT samples at a 0.5 wt % SWNT loading are depicted in Figure 5. In a general view, isoconversional plots show a fairly linear section, which takes most of the curing process (R ≈ 1090%). This section, which would stand for the propagation of the amine-oxirane ring-openings,40 has an average AE value of 6570 kJ/mol, consistent with literature data for TGAP/DDS curing studies using classical Arrhenius-based equations.43 In the extreme stages of the curing process, there is a significant dependence of AE with R. Regarding the initial steps of curing (up to R ≈ 10%), which can be considered mostly conducted by the primary amine reaction with

Figure 5. Activation energy plots for the neat epoxy resin and TGAP/ DDS/SWNT samples at a 0.5 wt % ratio.

oxirane rings, the neat TGAP/DDS sample shows no practical AE dependence. However, the incorporation of SWNTs leads to different behaviors depending on the functionalization route. The presence of SWNT-asg in the TGAP/DDS system causes a sudden increase of AE at the beginning of the curing process, reaching the linear value at R ≈ 10%. This could be considered as a decelerating effect. The lower initial values of AE could be attributed to an enhanced thermal diffusivity, which would promote the initiation of the curing reaction. The decelerating effect would arise upon viscosity increase during the initiation of curing, as it undergoes mainly by diffusional control.25 Furthermore, the oxygen groups on SWNT-asg surfaces could also create hydrogen-bridge bonds with TGAP upon oxirane opening, and this fact is well-known to affect the reaction course and the eventual cross-linking density.44 For the composite blend containing SWNT-oxa, an analogous discussion can be made, especially regarding the oxygen surface groups, whose larger amount would cause a more sharpened decelerating effect. In the SWNT-oxa blend, no influence of the amine groups is observed. For the rest of the functionalized samples, the opposite effect appears in the initial curing stages. A sharp drop of AE occurs, which can be perfectly compatible with an accelerating effect spawned by the 7245

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

Figure 6. Activation energy plots for the epoxy monomer and different TGAP/SWNT mixtures: (a) effect of different functional groups at a fixed loading, and (b) effect of loading for SWNT-dba sample.

surface amino groups on SWNTs. The higher AE initial values for these samples could be explained as an effect of oxirane rings preopening by the surface amino groups present in these SWNTs. During the blend manufacturing, SWNT amines could have partly reacted with epoxide groups, raising the viscosity of the system, as was experimentally observed. Because any of the amine moieties on SWNTs are more reactive than amines coming from DDS,45 the successful covalent interaction between TGAP and functionalized SWNTs is suggested, and this effect exceeds that of an enhanced thermal diffusivity. The strongest acceleration is caused by the SWNT-dca, which reaches an average AE value ∼5 kJ/mol lower than that of the neat matrix. According to the chemical nature of the covalent moieties, the one attached to SWNT-dca is structurally the most similar to the epoxy cross-linked network, due to the

ARTICLE

ethylenoxide (EO) groups. This could generate an additional affinity based on noncovalent interactions that would decrease the viscosity and promote the curing reaction by counteracting the diffusional control. We have shown how small amounts of EO accelerate initial stages of TGAP/DDS curing upon integration of SWNTs wrapped in a PEO-based block copolymer, due to an increase in the mobility of chemical species.25 With respect to the latter stages of curing, the neat matrix and SWNT-dca composite do not show a clear dependence of AE with R, but the rest of the composites exhibit a sharp increase of AE. Taking into account that the etherification reactions would predominate in the latter stages, this increase in AE could be ascribed to a hindrance of these reactions by surface groups in SWNTs (OH or amine). Indeed, it has been reported a delay on epoxy etherification reactions with OH-containing SWNTs, which raises the curing temperature at this stage.41 More interestingly, these effects do not depend on the functionalization degree, although they seem to be strongly dependent on the dispersibility of the filler, which is influenced by the chemical nature of the attached moiety and consequently by the functionalization route. The initial purity of the SWNT sample did not show to cause a determinant influence on the curing process, as was tested in a control experiment using a commercial high purity arc-discharge SWNT (see the Supporting Information). The isoconversional plots for TGAP/SWNT mixtures are displayed in Figure 6. In the absence of DDS, the same TGAP/ SWNT ratio for the 0.5 wt % composites end up in a 0.835 wt % when SWNT load is recalculated (Figure 6a). AE plots in these blends show the course of the self-curing and degradation of the TGAP monomer, which presents a dependence of AE with R (up to R ≈ 25%). The effect of all SWNT samples on the TGAP selfcuring is to hinder its initiation. The reaction between surface (either amine or oxygen) groups would compete with the oxirane self-opening. During most of the ongoing process, a linear section at ∼100 kJ/mol is observed, with no influence of SWNTs. The most noticeable AE dependences (in the form of a peak) arise from R ≈ 75% on, the moment in which the autocuring is much likely overlapping with degradation. In this stage, the presence of SWNTs greatly raises the AE (as determined from the SWNT-asg sample). The rest of the functionalized SWNTs exhibit an upshift of the R peak and lower AE values. This behavior correlates with the thermal stability observed for these SWNTs in TGA (Figure 1), suggesting that the degradation of the TGAP monomer is intimately related to the stability of the filler functional groups. To study the effect of the amount of functionalized SWNTs in the nonisothermal TGAP heating, a series of composite samples were recorded using SWNT-dba at different loadings (Figure 6b). The nanotube content plays an important role in this matter because the AE peak at the highest conversions progressively rises in intensity and shifts toward higher R values (i.e., temperatures). The successful integration by the solvent-free method of SWNTs not only catalyzes the TGAP/DDS curing but also prevents the monomer degradation.

4. CONCLUSIONS In the present work, four different experimental methodologies have been employed to obtain arc discharge SWNTs functionalized with covalently attached primary amines. The characterization by FTIR, Raman, Kaiser test, and TGA confirmed the successful functionalization and allowed the identification and quantification of covalent moieties. The integration of 7246

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C these SWNTs into a high performance epoxy system (TGAP/ DDS) has been carried out by a solvent-free protocol, and their effect on the curing reaction has been examined through nonisothermal DSC and isoconversional kinetic calculations. The incorporation of nonfunctionalized SWNTs, or those that undergo nitric acid oxidation prior to amination, raises the curing enthalpy with a slight change in Tg of the TGAP/DDS epoxy matrix and causes a sudden increase in the AE of the initial curing stages. Aminated SWNTs through different sidewall addition reactions exhibit a decrease in enthalpy and a big increase in Tg coupled to a sharp decrease in AE during the initial curing stages of TGAP/DDS, consistent with an effective covalent anchoring of SWNT amino groups to the target matrix. Further studies in the absence of the curing agent reveal the hindrance of autocuring and degradation of the TGAP monomer by the SWNTs, regardless of the functionalization degree albeit dependent on the SWNT ratio. The remaining metal catalysts do not seem to be the cause of all these facts; nevertheless, the dispersibility of SWNTs (influenced by the attached moiety and the functionalization procedure) plays a crucial role in the solvent-free integration. The type and chemical nature of the attached covalent moieties are directly responsible of these effects, not the functionalization degree. The results herein offer useful insights into the design of an epoxySWNT composite material for specific purposes, without negatively affecting the crosslinking reactions while promoting the covalent anchoring (fillermatrix adhesion). Manufacturing and full characterization of these composite materials will be reported in the near future.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional DSC study using a commercial arc-discharge SWNT sample of high purity is presented with comparative purposes (Table S1 and Figure S1). Optical microscopy images of TAGP/SWNT blends showing their dispersibility (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Fax: þ34976733977. E-mail: [email protected].

’ ACKNOWLEDGMENT The present work has been performed with financial support from the NRCCSIC collaboration project and fellowships from MICINN Spanish Ministry (FPU grant) and CAI-DGA program (research stays funds). J.M.G.-D. greatly thanks Dr. Benoit Simard and Dr. Yadienka Martinez for valuable training on the alkaline reductive functionalization and for their kindness and professionality. Special thanks go to Prof. Maurizio Prato for kindly allowing a short research stay in his group, and all of his team. Epoxy reagents were received as a Huntsman’s gift, which is greatly acknowledged. ’ REFERENCES (1) Spitalsky, Z.; Tasis, D.; Papagelis, K.; Galiotis, C. Prog. Polym. Sci. 2010, 35, 357. (2) Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chem. Rev. 2006, 106, 1105. (3) Bose, S.; Khare, R. A.; Moldenaers, P. Polymer 2010, 51, 975. (4) Allaoui, A.; El Bounia, N. eXPRESS Polym. Lett. 2009, 3, 588.

ARTICLE

(5) Gerson, A. L.; Bruck, H. A.; Hopkins, A. R.; Segal, K. N. Composites, Part A 2010, 41, 729. (6) Puglia, D.; Valentini, L.; Armentano, I.; Kenny, J. M. Diamond Relat. Mater. 2003, 12, 827. (7) Xie, H.; Liu, B.; Yuan, Z.; Shen, J.; Cheng, R. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 3701. (8) Tao, K.; Yang, S.; Grunlan, J.; Kim, Y. S.; Dang, B.; Deng, Y.; Thomas, R. L.; Wilson, B. L.; Wei, X. J. Appl. Polym. Sci. 2006, 102, 5248. (9) Zhou, T.; Wang, X.; Liu, X.; Xiong, D. Carbon 2009, 47, 1112. (10) Qiu, J.; Wang, S. Mater. Chem. Phys. 2010, 121, 295. (11) Dyke, C. A.; Tour, J. M. J. Phys. Chem. A 2004, 108, 1151. (12) Peng, X.; Wong, S. S. Adv. Mater. 2009, 21, 625. (13) Singh, P.; Campidelli, S.; Giordani, S.; Bonifazi, D.; Bianco, A.; Prato, M. Chem. Soc. Rev. 2009, 38, 2214. (14) Bahr, J. L.; Tour, J. M. J. Mater. Chem. 2002, 12, 1952. (15) Valentini, L.; Armentano, I.; Puglia, D.; Kenny, J. M. Carbon 2004, 42, 323. (16) Peng, H.; Alemany, L. B.; Margrave, J. L.; Khabashesku, V. N. J. Am. Chem. Soc. 2003, 125, 15174. (17) Khabashesku, V. N.; Billups, W. E.; Margrave, J. L. Acc. Chem. Res. 2002, 35, 1087. (18) Liang, F.; Sadana, A. K.; Peera, A.; Chattopadhyay, J.; Gu, Z.; Hauge, R. J.; Billups, W. E. Nano Lett. 2004, 4, 1257. (19) Penicaud, A.; Poulin, P.; Derre, A.; Anglaret, E.; Petit, P. J. Am. Chem. Soc. 2004, 127, 8. (20) Martinez-Rubi, Y.; Guan, J.; Lin, S.; Scriver, C.; Sturgeon, R. E.; Simard, B. Chem. Commun. 2007, 48, 5146. (21) Bahr, J. L.; Tour, J. M. Chem. Mater. 2001, 13, 3823. (22) Anson-Casaos, A.; Gonzalez-Domínguez, J. M.; Terrado, E.; Martinez, M. T. Carbon 2010, 48, 1480. (23) Tagmatarchis, N.; Prato, M. J. Mater. Chem. 2004, 14, 437. (24) Kordatos, K.; Da Ros, T.; Bosi, S.; Vazquez, E.; Bergamin, M.; Cusan, C.; Pellarini, F.; Tomberli, V.; Baiti, B.; Pantarotto, D.; Georgakilas, V.; Spalluto, G.; Prato, M. J. Org. Chem. 2001, 66, 4915. (25) Gonzalez-Domínguez, J. M.; Castell, P.; Anson, A.; Maser, W. K.; Benito, A. M.; Martinez, M. T. J. Nanosci. Nanotechnol. 2009, 9, 6104. (26) Gonzalez-Dominguez, J. M.; Anson-Casaos, A.; Diez-Pascual, A. M.; Ashrafi, B.; Naffakh, M.; Backman, D.; Stadler, H.; Johnston, A.; Gomez, M. A.; Martinez, M. T., ACS Appl. Mater. Interf. (in press). (27) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal. Biochem. 1970, 34, 595. (28) Quintana, M.; Prato, M. Chem. Commun. 2009, 40, 6005. (29) Vyazovkin, S. J. Therm. Anal. 1997, 49, 1493. (30) Vyazovkin, S. Anal. Chem. 2010, 82, 4936. (31) Gonzalez-Dominguez, J. M.; Anson-Casaos, A.; Castell, P.; Diez-Pascual, A. M.; Naffakh, M.; Ellis, G.; Gomez, M. A.; Martinez, M. T. Polym. Degrad. Stab. 2010, 95, 2065. (32) Martínez, M. T.; Callejas, M. A.; Benito, A. M.; Cochet, M.; Seeger, T.; Anson, A.; Schreiber, J.; Gordon, C.; Marhic, C.; Chauvet, O.; Fierro, J. L. G.; Maser, W. K. Carbon 2003, 41, 2247. (33) Uhlmann, U. K.; Jantoljak, H.; Pfander, N.; Bernier, P.; Journet, C.; Thomsen, C. Chem. Phys. Lett. 1998, 294, 237. (34) Cross, A. D.; Jones, R. A. An Introduction to Practical Infrared Spectroscopy; Butterworths: London, 1969. (35) Coates, J. Interpretation of Infrared Spectra, a practical approach. In Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; John Wiley & Sons Ltd.: Chichester, 2000; pp 1081510837. (36) Gorgulho, H. F.; Mesquita, J. P.; Gonc-alvez, F.; Pereira, M. F. R.; Figueiredo, J. L. Carbon 2008, 46, 1544. (37) Graupner, R. J. Raman Spectrosc. 2007, 38, 673. (38) Itkis, M. E.; Perea, D. E.; Niyogi, S.; Rickard, S. M.; Hamon, M. A.; Zhao, B.; Haddon, R. C. Nano Lett. 2003, 3, 309. (39) Georgakilas, V.; Voulgaris, D.; Vazquez, E.; Prato, M.; Guldi, D. M.; Kukovecz, A.; Kuzmany, H. J. Am. Chem. Soc. 2002, 124, 14318. (40) George, G. A.; Cash, G. A.; Rintoul, L. Polym. Int. 1996, 41, 169. (41) Choi, W. S.; Shanmugharaj, A. M.; Ryu, S. H. Thermochim. Acta 2010, 506, 77. 7247

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248

The Journal of Physical Chemistry C

ARTICLE

(42) Kim, S. H.; Lee, W. L.; Park, J. M. Carbon 2009, 47, 2699. (43) Patel, R. D.; Patel, R. G.; Patel, V. S. J. Therm. Anal. 1988, 34, 1283. (44) Bellenger, V.; Verdu, J.; Francillette, J.; Hoarau, P.; Morel, E. Polymer 1987, 28, 1079. (45) Nucleophilicity and basicity of an aromatic amine such as DDS are lower as compared to an aliphatic one because the direct linking to a benzene ring partly delocalizes the lone electron pair on the nitrogen atom. All of the SWNT functional groups contain aliphatic or benzylic amines, so their basicity and nucleophilicity are expected to be higher than DDS amines as well as their reactivity toward oxirane rings.

7248

dx.doi.org/10.1021/jp110830y |J. Phys. Chem. C 2011, 115, 7238–7248