Functionalization of multiwalled carbon nanotubes ...

3 downloads 0 Views 1MB Size Report
The chemistry of carbon nanotube is attractive for a variety of purposes. .... diameter PVDF filter membrane with 0.45 μm pore size which was tied with nylon.
Functionalization of multiwalled carbon nanotubes with urethane segments and their interaction with solvents and a polyurethane elastomer

Borja Fernández-d’Arlas1, Arantxa Eceiza1,* 1

Grupo“Materiales+Tecnologías” (GMT), Departamento de Ingeniería Química y del Medio Ambiente, Escuela Politécnica.Universidad del País Vasco/Euskal Herriko Unibertsitatea (UPV/EHU), Pza. Europa 1, 20018 Donostia-San Sebastián, Spain

ABSTRACT Manipulation of nano-particles is easier when they are appropriately functionalized with adequate chemical groups. In this work we functionalized multiwalled carbon nanotubes with different chemical groups and studied the influence of functionalization on their stability in different solutions and its impact on elastomeric polyurethane composites. We observed that the functionalization yielded in all cases nanotubes with approximately the same dimensions (diameter: ∼8–40 nm and length: ∼1–25 m) and 24-29 wt% of organic tallow per mass of nanotube. The chemical behavior of nanotubes in solutions was highly dependent on functionality. The impact on mechanical properties was function of the chemical nature of nanotube functional groups and moderately followed the expected tendency according to their solubility into solvents.

Keywords: Carbon nanotubes, functionalization, urethane, nanocomposites

2

Introduction The chemistry of carbon nanotube is attractive for a variety of purposes. Multiwalled carbon nanotubes (MWCNT) are attractive for different applications due to their unique combination of interesting properties such as stiffness and strength (Falvo el al. 1997; Wong et al. 1997) and high electrical (Ebbesen et al. 1996) and thermal conductivity. (Berber et al. 2000; Che et al. 2000; Saito et al. 1999) These properties have inspired interest in using CNT as filler in polymer composite systems (Ajayan et al. 1994) to obtain light structural materials with enhanced mechanical, electrical, thermal and optical characteristics or, conversely, the possibility to greatly enhance the mechanical properties of polymers with little amounts of nanofiller. (Breuer et al. 2004) Carbon nanotubes chemistry is not only important for improving polymers performance but also is considered to be interesting to manipulate nanotubes for purposes involving drug-carrying systems (Zhag et al. 2011) or biomedical patches (Weber et al. 2005), catalysis supports, hydrogen storing cells (Dillon et al. 1997), chemical sensors (Zhao et al. 2002; Snow et al. 2005; Yáñez-Sedeño et al. 2010), bendable transistors (Cao et al. 2006), printable solar cells (Gruner et al. 2007), or even filtrating and purification devices. For most of these applications is desired a proper chemical functionalization that minimizes nanotubes damage and therefore preserves their properties but also an appropriated nanotubes purity (Andrade et al. 2013). For the introduction of carbon nanotubes in multiphase systems such as copolymers it appears reasonable to functionalize nanotubes in order to target a particular phase and to enhance its specific properties. In this paper we present results on functionalization of MWCNT with different functional groups with the aim of studying the functional group nature impact on MWCNT solubility in different solutions and on mechanical properties of polyurethane 3

composites made with them. Functionalization started with a simple oxidation treatment (Goyanes et al. 2007) to introduce carboxylic groups and proceeded by reacting them with 1,6-hexamethylene diisocyanate to incorporate either polyurethane hard segments or soft segments by continuing the reaction with either 1,4-butanediol or poly(caprolactone-b-hexamethylene

carbonate-b--caprolactone)diol,

respectively.

We

compare the chemical behaviour of these nanotubes with nanotubes functionalized by the classic octadecylamine route functionalization (Hamon et al. 1999; Chen et al. 2001).

Experimental Materials Thin multiwalled carbon nanotubes (diameter: ∼8–40 nm and length: ∼1–25 m) were purchased from Nanocyl (Nanocyl, Belgium). The trade name of the product is Nanocyl 3100 with a purity over 95 wt% of MWCNT, with less than 5 wt% of amorphous carbon and catalysis residues. Nevertheless the goal was to study the effect of pure MWCNT over polyurethane matrices so the nanotubes were always purified before use. Nitric and sulphuric acids were purchased from Panreac with 65 wt% and 96 wt% of purity,

respectively.

1,6-Hexamethylene

diisocyanate,

poly(-caprolactone-b-

hexamethylene carbonate-b--caprolactone)diol and 1,4-butanediol used for nanotube functionalization used for polyurethanes synthesis. Octadecylamine (ODA) was purchased from Aldrich. Polyurethane with 10 wt% hard segment content was synthesized by the common two steps method as described previously (Fernándezd’Arlas et al. 2008). N, N´-dimethylformamide was from Panreac, with 99.8 wt% purity and tetrahydrofuran was from Lab-Scan, Analytical Sciences.

4

Acid treatment of MWCNT To purify nanotubes from catalysts residues and functionalize nanotubes with carboxylic groups the nanotubes were sonicated in a sonic bath for 2 h in a mixture of H2SO4:HNO3 (3:1) with a ratio of 100 mg nanotubes/130 mL of mixture (Scheme 1b). While sonication the temperature varied in the range (~25-45 ºC). After sonication the mixture was poured into distilled water in order to stop the reaction. Then this suspension was vacuum filtered through PTFE membranes with 0.20 m average pore diameter and washed with water until the pristine utilised water pH was obtained in the filtrate. Then, the acid treated nanotubes were washed with acetone and left to dry in a vacuum oven for at least 12 h. The number of carboxylic groups was determined by means of a chemical backward titration (Hu et al. 2001) giving an average value of 1.4 ± 0.5 x 10-3 mol g-1 nanotube (Fernández-d’Arlas et al. 2009). For the chemical titration typically three solutions of 8 mg of acid treated nanotubes were incubated into 2.2 mL 0.02 N solution of NaOH and over 15 mL extra H2O added to easier manipulate the solutions. Solutions were left reacting overnight at room temperature. Then the solutions were vacuum filtrated through 0.45 m pore size polyamide filtrate membranes and the filtrates were immediately potentiometrically titrated adding 0.2 mL of HCl 0.01 N at each step before data acquisition. The addition was finished when the pristine used water pH was reached. Acid functionalized nanotubes were named as MWCNT-COOH. Functionalization of nanotube with polyurethane soft segments Nanotube functionalization with soft segments (Scheme 1c) was performed as follows: 100 mg acid treated MWCNT (~ 1.4 x 10-4 mol of COOH) were suspended in 300 mL of dried DMF and dispersed by means of a sonic tip (3 min), followed by a 30 min mild

5

power bath sonication and finally again with 1 min sonic tip. DMF was chosen as solvent for its demonstrated capability to effectively disperse carbon nanotubes (Furtado et al. 2004; Bergin et al. 2008). The dispersion was charged into a three necked reactor and nitrogen was left to flow for about 1 min. Then 200 L of HDI (1.19 x 10-3 mol; ~ 8 time excess respect to the COOH) were charged into the reactor provided with a magnetic stirrer and the mixture was left react under reflux at 95 ºC for 5 h. After this time 4.4 g of PCL-b-PHMC-b-PCL (2.18 x 10-3 mol; ~ 2 fold excess respect to the HDI so that not severe branching occurred) dissolved into 20 mL of DMF, were charged and left react for 1 h at 100 ºC and overnight at 95 ºC. After reaction, the mixture was filtered through a poly(vinilydene fluoride) (PVDF) 0.45 m pore size membrane and the nanotubes were washed with acetone. Then the nanotubes were wrapped into a 9 cm diameter PVDF filter membrane with 0.45 m pore size which was tied with nylon cords and put into Soxhlet extractor, with THF as extractive solvent, for 48 h to remove the unreacted PCL-b-PHMC-b-PCL. After this extraction the black powders were introduced in a vacuum oven at 60 ºC and full vacuum for at least 12 h. The obtained black powders were stored into vials inside a desiccator. These nanotubes were named as MWCNT-g-SS. Functionalization of nanotubes with polyurethane hard segments The protocol regarding temperature and reaction times to functionalize nanotubes with hard segments was the same as for the case of soft segments, with the difference that instead polyol, 0.4 g of BD (4.44 x 10-3 mol; ~ 4 time excess respect to the HDI) dissolved into 20 mL DMF were added. Drying and washing protocol was the same as for the previous functionalization. Nanotubes functionalized with HDI-BD hard segments were obtained (Scheme 1d). These nanotubes were named as MWCNT-g-HS. Functionalization of nanotubes with ODA 6

This functionalization was performed following the method proposed by Hammond et al. (1999) which has been extensively applied for dispersion of single walled carbon nanotubes (Chen et al. 2001). The amount of 0.1 g of acid treated nanotubes were mixed with 1.1 g ODA in a 25 mL round bottom reactor and taken to 110 ºC and when the ODA was melt the mixing was continued using the tip of a spatula. The reactor was covered with a top and the mixture was left reacting at 120 ºC for 72 h. When this time had lapsed the mixture was left to cool down. Afterwards the black powders were dispersed in 160 mL ethanol using an ultrasonic tip (20% amplitude) for 3 min. This dispersion was filtered and the black powders were collected from the filter and redispersed in 200 mL THF using the ultrasonic tip for 2 min. This dispersion was filtered and the black powders were left to dry into a vacuum oven at 60º C at full vacuum for at least 12 h. The obtained black powders were stored into vials inside desiccators (Scheme 1e). These nanotubes were named as MWCNT-COO(-)(+)ODA. Presence of carboxylated carbonaceous fragments (CCFs) is discarded due to the consecutive filtering and Soxhlet extraction processes which enable the CCFs to be separated from the nanotubes (Stéfani et al. 2011; Andrade et al. 2013). Hazards Nanotube functionalizations were carried out into a fume hood, in order to avoid toxicity from acid fumes or dimethyl formamide and 1,6-hexamethylene diisocyanate vapours. Composites preparation The required amount of nanotubes for a concentration

3wt% in 300 mg of total

composite mass was suspended in 6 mL of a mixture of DMF/THF (1:1) and sonicated under a high energy sonication tip (750 W, Bioblock Scientific) operated with an

7

amplitude of 20% for 10 min. Afterwards the right amount of polymer was pipeted (around 6 mL) from a polymer solution in the same solvent mixture, with a concentration of 50 mg mL-1. Then the nanotube/polymer/solvent mixture was sonicated again under the same conditions as mentioned above, after what the mixture underwent a mild bath sonication for 2 h. Solutions were then drop cast into PTFE moulds of 4 cm x 4 cm x 1 cm and the solvent was evaporated under controlled temperature and vacuum conditions. Evaporation cycle was: 24 h at room temperature and 700 ± 40 mbar, 24 h at 60 ºC and 700 ± 40 mbar and finally 24 h at 60 ºC and 200 ± 20 mbar. Solvents were recovered and disposed accordingly. Scanning electron microscopy (SEM) SEM images were taken in a Zeiss instrument operated at voltages in the range of 2-5 kV and with focal distances of 5-8 mm. Thermogravimetric analysis Thermogravimetric analysis (TGA) was performed in a Mettler Toledo TGA/SDTA 851e equipment under air atmosphere using N2 as purge gas. Typically around 1-2 mg of nanotubes were analysed heating from 25 ºC to 1000 ºC at a heating rate of 10 ºC min -1. Scans were performed two times to confirm reproducibility. Raman spectroscopy A InVia Raman (Renishaw) equipment was used with laser excitation wavelength was 514 nm. The analysed range was 150-3200 cm-1, and 5 accumulations every 10 s were taken for each sample. For data acquisition sample powders were just fitted under the equipment probe over a slide. Fourier transformed infrared spectroscopy

8

Samples for FT-IR analysis were prepared by mixing the different nanotubes into KBr with a concentration of around 2.5 mg nanotube g-1 KBr. That is, around 0.5 mg of nanotube were pestled into a mortar with 200 mg KBr. A 200 mg KBr blank was also prepared. Scans were performed in a Nicolet-Nexus-FTIR in the range of 400-4000 cm-1 and performed at 120 ºC to avoid water noise. A number of 500 scans with a resolution of 1 cm-1 were averaged for each sample. Ultraviolet-visible spectroscopy UV-Vis spectra were recorded using a Jasco V-630 spectrophotometer in the range of 200-1000 nm using quartz cuvettes as samples holders. Employed solvents and solvent mixtures were used as blanks. The stability of differently functionalized nanotubes in THF, THF/DMF and DMF solvents was studied by means of sedimentation analysis measuring ultraviolet (UV) absorbance at 600 nm before and after sample centrifugation. Initial nanotubes concentrations were of 0.015 mg mL-1, and they were dispersed by means of an ultrasonic tip energetic treatment for 10 min (750W, 20% amplitude) followed by 2 h mild bath sonication. Centrifugation was carried out at 5500 rpm for 90 min for all samples. Considering the Lambert-Beer relation applies for both initial and centrifuged dispersions (Bergin et al. 2008) the nanotubes concentration remaining after centrifugation can be estimates as follows:

C c  Ci 

Ac Ai

(1)

where Cc and Ci, are the final supernatant and initial nanotube concentration, respectively, and Ac and Ai is the absorbance at 600 nm after and before centrifugation, respectively. The nanotubes % remaining in solution was calculated as Acx100/Ai. Tensile tests. Tensile tests were carried out in a MTS-Insight10 equipment with

9

specimens of about 2.5 mm and ~0.1 mm thickness. Initial testing length was of ~8.5 mm and the cross-head speed of 100 mm min-1.

Results and discussion Structural characterization As can be observed in SEM images of Fig. 1 the general structure of nanotubes was preserved along different functionalization steps, from the pristine to organically decorated nanotubes. In the images, it could be observed that oxidized nanotubes appear thicker than oxidized ones. This is in agreement with previous data obtained statistically by measuring pristine and oxidized nanotubes by atomic force microscopy (Fernándezd’Arlas et al. 2012) were we found that the average value of pristine nanotube diameter was 14 ± 6 nm, while that of acid treated nanotubes was reduced to 12 ± 5 nm. This might be related to the full destruction of outer nanotube walls. The average length of pristine nanotubes was 1639 ± 1331 nm and that of the oxidized nanotubes was 740 ± 519 nm, thus indicating an important reduction in average length. This is related with a preferential oxidation of nanotubes at the tips of the nanotubes, where more defects are prone to be found (Goyanes et al. 2007). Further functionalization did not produce any appreciable morphological variations of nanotubes as can be observed in Fig. 1c for MWCNT-g-HS. Raman spectroscopy combined with Fourier transform infrared spectroscopy are useful tools for analysing both structure and functional groups present in carbonaceous materials like carbon nanotubes (Dresselhaus et al. 2002). In Fig. 2, Raman and FT-IR spectra of pristine and acid treated nanotubes are presented. In FT-IR, the polarised functional groups such as carbonyl (~1735 cm-1), ethoxy (~1050-1200 cm-1) or hydroxyl (~3400 cm-1) are easily detected. It is seen that acid treatment increases the intensity of

10

carbonyl and hydroxyl group bands. In Raman spectra the polarisable bonds such as CC, C-H or C=C are more detectable. The main features of Raman spectra are the disorder-induced D band at 1350 cm-1 and the G band corresponding to crystalline graphitic structures at 1585 cm-1, with a new peak emerging at 1610 cm-1 with acid treatment, the so-called D´ band (Bacsa el at. 1994; Rosca et al. 2005). By analyzing the D to G band intensity ratio, it is possible to estimate the density of defects (Andrade et al. 2013) which increase parallel to the introduction of carboxylic groups. Introduction of oxygen based functional groups was also detected by elemental analysis in which carbon, hydrogen and oxygen ratios {C,H,O} were measured giving {99.3 %, 0.5 %, ~0 %} for the pristine nanotubes, and {88.1 %, 1.1 %, 9.6 %} for the acid treated nanotubes. As seen an increase on hydrogen was also perceived as a consequence of the introduction of oxygenated species containing hydrogen, such as carboxylic or hydroxyl groups, which can give more polar and acid-base interactions with solvents such as water as well as electrostatic repulsion with other nanotubes. Functionalization of nanotubes with the different organic tallows was analysed by FTIR spectra, as gathered in Fig. 3. The band corresponding to the nanotube C=C backbone appearing around 1570 cm-1 is common to all of them indicating that MWCNT structural backbone was preserved. Acid treatment increased hydroxyl and carboxyl groups, as mentioned above, but also introduces some aliphatic C-H groups, as can be seen in the region of 2950 cm-1. Functionalization with hard segments introduces hydroxyl groups but also urethane groups, which may contribute to the increased overall intensity of the wide band at around 3360 cm-1. The introduction of aliphatic CH2 units increases C-H stretching band at 2930 cm-1. The ratio between these two bands reduces when increasing the number of aliphatic C-H in the pendant chain, decreasing in the order MWCNT-g-HS > MWCNT-g-SS > MWCNT-COO(-)(+)ODA, as also seen in Fig.

11

3. The reduction in the carbonyl band when functionalizing the acid treated nanotubes with either HS or SS could be due to a decrease in the dipole moment of the parental functional group, changing from a carboxylic to an amide group, -NH-CO-, resulting from the reaction between isocyanate groups, -NCO, and carboxylic groups, -COOH. The reduction of the band intensity in respect to MWCNT-COOH could also be due to the presence of urethanes and esters and carbonates, in the case of SS-functionalized nanotubes. While some bands changed in intensity, some other spectra features, as baseline and some peaks relative intensity, did not vary notoriously suggesting some nanotube intrinsic absorption properties were still detected. Thermogravimetric analysis Thermogravimetric analysis was performed over the functionalized nanotubes in order to measure the amount of organic tallows grafted after the functionalization and washing treatment. Fig. 4a and b show the derivative (dW/dT) and weight loss curve of differently functionalized MWCNT along with pristine MWCNT. It is seen that the maximum weight loss peak of the organic tallows starts at 250ºC, with maximum degradation peaks between 300-350 ºC, while the peak corresponding to the degradation of the nanotubes backbone occurs in the range ~550-610 ºC. The peaks corresponding to the degradation of MWCNT-g-HS and MWCNT-g-SS appear in the region typical of degradation of polyurethane (Corcuera et al. 2010). The area of the peak corresponding to the degradation of the organic tallows was integrated in all cases and related to the area of the whole thermogram to obtain the amount of organic material remaining after the treatments. Table 1 gathers the functionalization ratios

12

obtained in each case. As seen, the weight losses were all below 30 wt%, with a maximum value for the hard segment functionalized nanotubes while soft segment and ODA functionalized nanotubes had a similar value. Despite the similarities in weight losses, the grafted branches or tallows must be different as deduced from the tallows molecular weights. Although it should be regarded with caution (due to the large standard deviation of the chemical COOH titration) the high value of 1.57 x 10-3 mol of -(HDI-BD)n- g-1 nanotube can be due to parallel reactions of isocyanate groups with hydroxyl groups present onto the nanotubes, which were not accounted in the chemical titration employed to measure COOH groups. The bigger size of soft segment molecules and its consequent lower reactivity must affect the final reactivity and the lower amount of tallow molecules attached. The amount of ODA tallows grafted to nanotubes fits well to the number of –COOH groups. This must be due to the acid-base type of reaction that undergoes this functionalization. This reaction has actually been used to determinate –COOH groups present in the nanotubes (Marshall et al. 2006). Dispersability of functionalized nanotubes Solubility of the differently functionalized nanotubes can be understood with the aid of Table 2, where the employed solvent solubility parameters are listed along those of the different nanotube functionalities. Fig. 5 gathers stability tests carried out by UV after sedimentation experiments. As seen in Fig. 5a, acid treated nanotubes were not as stable in THF as in DMF. This fact can be related to the high value of solubility parameters as a consequence of the presence of polar groups such as carboxylic onto MWCNT surface as well as the known capability of amidic solvents to solubilize nanotubes attributed to

13

the equality between the surface tension of this class of solvents and that of the nanotubes (Bergin et al. 2008). The solubility/stability in THF increased as solubility parameter of functional group decreased. Nanotubes functionalized with hard segments were more soluble in DMF than in THF or in the solvents mixture. This can also be related to the high value of hard segment solubility parameter owed to large hydrogen bonding contribution of the urethane units (see Table 2). Solubility in DMF decreased as nanotube functionalities solubility parameter decreased. This solubility analysis served to confirm that the functionalization onto the nanotubes with the desired chemical groups was attained. Impact of functionalization onto MWCNT/PU composites One of the purposes of this work was to study the effect of functionalization on preferential hierarchical assembling arising from a favoured dispersion of nanotubes in the different polyurethane phases and its impact on mechanical properties. In Fig. 6 the black curve represents a neat HDI-10 tensile curve. Addition of 3 wt% acid treated MWCNT-COOH (red curve) reduced max but increased c and Ec. When preparing composites with hard segment functionalized nanotubes, MWCNT-gHS, the elastic modulus improved in the same manner as with MWCNT-COOH, but the ductility was higher. This fact can be related to a better targeting of nanofiller into hard domains (Liff et al. 2007) therefore favouring the soft segment provided extensibility. For composites with MWCNT-g-SS the tensile curve laid on stresses below those of nanocomposites with MWCNT-COOH and MWCNT-g-HS, probably due to the presence of more amounts of a low Tg material which might act as plasticizer. At the same time, as can be observed in the inset of Fig. 6 the modulus of composites with MWCNT-g-SS (14.7 ± 2.3 MPa) was higher than those of composites with MWCNT14

COOH (10.8 ± 0.4 MPa) and MWCNT-g-HS (12.7 ± 2.3 MPa). This could be attributed to a better inclusion of MWCNT-g-SS nanotubes into the PU soft phase, acting as new crosslinking points stiffening the matrix. For nanocomposites with MWCNT-COO()(+)

ODA this decrease on the stress was even more notable, probably due to the

plasticizing effect of the aliphatic methylenes present on the ODA tallows. It is also noted that yield and plateau stresses increased as functional tallow chain mobility decreased. The organic ODA tallows might have the highest mobility as they are formed by hydrocarbon chains of low molecular weight, which generally have very low glass transitions, followed by the SS (Tg ≈ -80 ºC) and by HS (Tg ≈ 50 ºC). This fact might affect the global nanotubes free volume, decreasing matrix interchain attraction and the cross-linking effect provoked by nanotubes. This phenomenon should be more important as nanotubes tallows mobility increases, therefore influencing on final nanocomposites elastic modulus, plateau stress and ductility. The higher stresses presented at the viscous-plateau region by nanocomposites containing nanotubes with higher solubility parameter, MWCNT-COOH and MWCNTg-HS, is believed to be due to a higher amount of nanotubes in the sample powders in the first case and for a higher amount of rigid hard segments in the second, which may induce a better interaction between nanotubes and polyurethane HS, but also for the ability of these nanotubes to form cross-linked hard segment structures in comparison to nanotubes functionalized with lower solubility parameter and Tg tallows such as MWCNT-g-SS and MWCNT-COO(-)(+)ODA, as discussed above. This fact is also assessed when comparing MWCNT-g-HS reinforced HDI-10 polyurethane with that reinforced with MWCNT-g-SS. In both cases the amount of organic tallows on the nanotubes was of the same order (~ 25 wt%) but in the case of nanotubes functionalized with soft segments their functional groups may favour the 15

dispersion of nanotubes into soft domains rather than into hard domains. As in this work Khan et al (2010) have shown that the presence of nanotubes into PU soft domains has a hindrance effect on the final composite ductility, and this phenomenon has as also demonstrated for other type of soft segment functionalized nanofillers such as Laponite (Liff et al. 2007).

Conclusions Acid treatment resulted into a reduction of nanotube aspect ratio from ~117 down to ~62. The treatment introduced more defects, as seen in the increase of aliphatic C-H group stretching vibration, along carboxylic groups and hydroxyl groups, but preserved the general structure of MWCNT. Nanotubes were functionalized with polyurethane hard and soft segments and were characterized by different methods. The diisocyanate intermediate step can be regarding as an interesting methodology to anchor different types of nucleophilic molecules onto carbon nanotubes walls. Organic functionalization of nanotubes yielded in all cases around 30 wt% of organic tallows over nanotubes. Considering the molecular weights of soft and hard segments, the functionalization degree, expressed as mol g-1 nanotubes, decreased with increase in molecular weight of the reactive chain. The solubility of the differently functionalized nanotubes in THF, DMF and DFM/THF mixtures varied accordingly with the solubility parameters calculated for the organic tallows. MWCNT functionalization with hard segments improved ductility with respect to acid treated MWCNT, what can be matched with a preferential targeting of MWCNT into polyurethane hard domains.

Acknowledgments. B. Fernández-d’Arlas acknowledges the University of the Basque Country (UPV/EHU) post-doc grant “Ayuda a la Especialización de Doctores”, for its financial support and all

16

authors acknowledge the General Research Services of the University of the Basque Country (SGIker), and especially Macroconducta-Mesoestructura-Nanotecnología unit for their technical support and to Dr. Luis Bartolomé for the elemental analysis. Authors also acknowledge funding from Basque Government through SAIOTEK 11-S-PE11UN132 program and in the frame of Grupos Consolidados (IT-776-13). Authors also acknowledge “Low dimensional nanostructures group” from the School of Physics of Trinity College Dublin for the SEM images.

17

Figures and tables captions

Table 1. Organic quatification by TGA. Weight loss obtained from TGA along with molecular weights of the tallows and as-calculated functional group moles obtained.

Table 2. Solubility parameters. Solubility parameters of the used solvents along those calculated for the different organic tallows.

Scheme 1. Functionalization routes. From (a) pristine nanotubes to (b) MWCNTCOOH, (c) MWCNT-NCO, (d) MWCNT-COO(-)(+)ODA, (e) MWCNT-g-SS, and (f) MWCNT-g-HS.

Fig. 1. SEM images of (a) pristine MWCNT, (b) MWCNT-COOH and (c) MWCNT-gHS. Scale bar corresponds to 100 nm in all cases.

Fig. 2. Raman and FT-IR spectra of (a) pristine MWCNT and (b) MWCNT-COOH.

Fig. 3. FT-IR spectra of the pristine and functionalized MWCNT.

Fig. 4. Thermogramiteric curves. (a) DTGA and (b) weight loss curves.

Fig. 5. Remaining nanotubes in dispersion for different solvent systems ordered from low to high solubility parameter of the organic tallow. (a) MWCNT-COO(-)(+)ODA, 18

(b) MWCNT-g-SS, (c) MWCNT-g-HS and (d) MWCNT-COOH.

Fig. 6. Representative stress-strain curves of the polyurethane matrix and the composites prepared with 3 wt% of differently functionalized MWCNT.

19

Table 1. Organic quatification by TGA. Weight loss obtained from TGA along with molecular weights of the tallows and as-calculated functional group moles obtained. Weight loss (wt%)

Tallow Mw (with n = 1) (g mol-1)

103 x mol functional groups g-1 nanotube

-

45

1.4 ± 0.5*

- (HDI-BD)n-

28.8

258

1.57

- (HDI- PCL-b-PHMC-b-PCL)n-

24.0

2192

0.14

H3N+-(CH2)17-CH3

24.4

270

1.2

Functionality

-COOH

*Determined by backward titration

20

Table 2. Solubility parameters. Solubility parameters of the used solvents along those calculated for the different organic tallows.

 Material

(J1/2cm-3/2)

THF

18.6§

DMF

24.7§

-COOH

≥ 25Ø

- (HDI-BD)-

23*

PCL-b-PHMC-b-PCL

20.9*

H2N-(CH2)17-CH3

19.1*

§

Taken from literature (Krevelen V, 1990). Assumed to be higher than DMF (which is a good pristine nanotubes solvent) for the introduction of polar groups upon acid treatment. * Calculated using Feedors molar volumes and Hoftyzer-Van Krevelen method (Krevelen V, 1990).. Ø

21

Scheme 1

22

Fig. 1

23

Fig. 2

24

Fig. 3

25

Fig. 4

26

Fig. 5

27

Fig. 6

28

References Ajayan PM, Stephan O, Colliex C, Trauth D (1994) Aligned carbon nanotube arrays formed by cutting a polymer resin-nanotube composite. Science 265: 1212-1214 Andrade NF, Stéfani D, Martinez T, Paula AJ, Silveira JV, Alves OL, Filho AGS. Temperature effects on the nitric acid oxidation of industrial grade multiwalled carbon nanotubes. J. Nanopart. Res. 15: 1761-1768 Bacsa WS, Ugarte D, Chatelain A, de Heer WA (1994) High-resolution electron microscopy and inelastic light scattering of purified multishelled carbon nanotubes. Phys Rev. B, 50: 473-476 Berber S, Kwon YK, Tománek D (2000) Unusually high thermal conductivity of carbon nanotubes. Phys. Rev. Lett. 84: 4613-4616 Bergin SD, V. Nicolosi V, Streich PV, Giordani S, Sun Z, Windle AH, Ryan P, Niraj NPP, Wang ZT, Carpenter L, Blau WJ, Boland JJ, Hamilton JP, Coleman JN (2008) Towards solutions of single-walled carbon nanotubes in common solvents. Adv. Mater. 20: 1876-1881. Breuer O, Sundararaj U (2004) Big returns from small fibre: A review of polymer/carbon nanotube composites. Polym. Compos. 5: 630-645 Cao Q, Hur SH, Zhu ZT, Sun Y, Wang C, Meitl MA, Shim M, Rogers JA (2006). Highly bendable, transparent thin-film transistors that use carbon-nanotube-based conductors and semiconductors with elastomeric dielectrics. Adv. Mater. 18: 304309 Che J, Cagin T, Goddard WA (2000) Thermal conductivity of carbon nanotubes. Nanotechnology 11: 65-69.

29

Chen J, Rao AM, Lyuksyutov S, Itkis ME, Hamon MA, Hu H, Cohn RW, Eklund PC, Colbert DT, Smalley RE, Haddon RC (2001) Dissolution of full-length singlewalled carbon nanotubes. J. Phys. Chem. B 105: 2525-2528 Corcuera MA, Rueda L, Fernández-d’Arlas B, Arbelaiz A, Marieta C, Mondragon I, Eceiza A (2010) Microstructure and properties of polyurethanes derived from castor oil. Polym. Degr. and Stab. 95: 2175-2184 Dillon AC, Jones KM, Bekkedahl TA, Kiang CH, Bethune DS, Heben MJ (1997) Storage of hydrogen in single-walled carbon nanotubes. Nature 386: 377-379 Dresselhaus MS, Dresselhaus G, Jorio A, Filho AGS, Saito R (2002) Raman spectroscopy on isolated single wall carbon nanotubes. Carbon, 40: 2043-2061 Ebbesen TW, Lezec HJ, Hiura H, Bennett JW, Ghaemi HF, Thio T (1996). Electrical conductivity of individual carbon nanotubes. Nature 382: 54-56 Falvo MR, Clary GJ, Taylor RM, Chi V, Brooks FP, Washburn S, Superfine R (1997) Bending and buckling of carbon nanotubes under large strain. Nature 389: 582584 Fernandez-d´Arlas B, Goyanes S, Rubiolo GH, Mondragon I, Corcuera MA, Eceiza A (2009) Surface modification of multiwalled carbon nanotubes via esterification using a biodegradable polyol. J. Nanosci. Nanotechnol. 9: 6064-6071 Fernández-d’Arlas B, Khan U, Rueda L, Martin L, Ramos JA, Coleman JN, Gonzalez ML, Valea A, Mondragon I, Corcuera MA, Eceiza A (2012) Study of the mechanical, electrical nand morphological properties of PU/MWCNT composites obtained by two different processing routes. Comp. Sci. Techn. 72: 235-242 Fernandez-d´Arlas B, Rueda L, de la Caba K, Mondragon I, Eceiza A (2008) Microdomain 30

composition

and

properties

differences

of

biodegradable

polyurethanes based on MDI and HDI. Polym. Eng. Sci. 48: 519-529 Furtado CA, Kim UJ, Gutierrez HR, Pan L, Dickey EC, Eklund PC (2004). Debundling and dissolution of single-walled carbon nanotubes in amide solvents. J. Am. Chem. Soc. 126: 6095-6105 Goyanes S, Rubiolo GR, Salazar A, Jimeno A, Corcuera MA, Mondragon I (2007) Carboxylation treatment of multiwalled carbon nanotubes monitored by infrared and ultraviolet spectroscopies and scanning probe microscopy. Diamond. Rel. Mater. 16: 412-417 Gruner G (2007) Carbon nanonets spark new electronics. Sci. Amer. 296: 76-83 Hamon MA, Chen J, Hu H, Chen Y, Itkis ME, Rao AM, Eklund PC, Haddon RC (1999) Dissolution of single-walled carbon nanotubes. Adv. Mater. 11: 834-840 Hu H, Bhowmik P, Zhao B, Hamon MA, Itkis ME, Haddon RC (2001) Determination of acid sites of purified single-walled carbon nanotubes by acid-base titration. Chem. Phys. Lett. 345: 25-28 Khan U, Blighe FM, Coleman JN (2010) Selective Mechanical Reinforcement of Thermoplastic Polyurethane by Targeted Insertion of Functionalized SWCNTs. J. Phys. Chem. C 114: 11401–11408 Krevelen V (1990) Properties of polymers. Elsevier Science, Amsterdam, pp. 189 Liff

SM,

Kumar

M,

McKinley GH

(2007)

High-performance

elastomeric

nanocomposites via solvent-exchange processing. Nature Mat. 6: 76-83 Marshall MW, Popa-Nita S, Shapter JG (2006) Measurement of functionalized carbon nan otube carboxylic acid groups using a simple chemical process. Carbon, 44: 1137-1141 Rosca ID, Watari F, Uo M, Akasaka T (2005) Oxidation of multiwalled carbon 31

nanotubes by nitric acid. Carbon, 43: 3124-3131 Saito R, Dresselhaus G, Dresselhaus MS (1999) Physical properties of carbon nanotubes. Imperial College Press, London Snow ES, Perkins FK (2005) Capacitance and conductance of single-walled carbon nanotubes in the presence of chemical vapors. Nano Lett. 5: 2414-2417 Stéfani D, Paula AJ, Vaz BG, Silva RA, Andrade NF, Justo GZ, Ferreira CV, Filho AGS, Eberlin MN, Alves OL (2011) Structural and proactive safety aspects of oxidation debris from multiwalled carbon nanotubes. J. Hazard. Mat. 189: 391396. Weber J, Holman T, Eidenshink T, Chen JJ (2005) Using buckypapers as a therapeutic aid in medical application. U. S Patent 2005/0074479 A1 Wong EW, Sheehan PE, Lieber CM (1997) Nanobeam mechanics: elasticity, strength, and toughness of nanorods and nanotubes. Science 277: 1971-1975 Yáñez-Sedeño P, Pingarrón JM, Riu J, Rius X (2010) Electrochemical sensing based on carbon nanotubes. TrAC Trends. Anal. Chem. 29: 939-953 Zhang W, Zhang Z, Zhang Y (2011) The application of carbon nanotubes in target drug delivery systems for cancer therapies. Nanosca. Res. Lett. 6: 555-577 Zhao Q, Gan Z, Zhuang Q (2002) Electrochemical sensors based on carbon nanotubes. Electroanalysis 12, 1609-1613

32