Hybrids of CO2-Responsive Water-Redispersible Single-Walled Carbon Nanotubes by a Surfactant Based on Natural Rosin.

A kind of polymerizable dispersant based on natural rosin was used to disperse single-walled carbon nanotubes (SWNTs) in aqueous solution followed by in situ free-radical polymerization to achieve a controllable SWNTs dispersion, that not only can be controlled by CO2/N2, but can also be recycled and redispersed in CO2-saturated water after drying without sonication.

Introduction

Single-walled carbon nanotubes (SWNTs) have gained increasing interest because of their excellent thermal, electronic, and mechanical properties.[1] However, SWNTs tend to form aggregates because of their strong π–π stacking, resulting in their indispersibility in water and most conventional solvents,[2] and these aggregates are problematic for some practical applications where dispersed individual tubes are demanded.[3] To solve this problem, covalent and noncovalent approaches have been developed to modify SWNTs. Noncovalent functionalization is superior to covalent functionalization because the latter generates unwanted functional groups, changing the intrinsic electronic properties of the SWNTs.[4] In noncovalent functionalization, amphiphilic molecules have often been used to disperse SWNTs such as surfactants,[5−9] DNA,[10−13] ionic liquid,[14] lipids,[15] and polymers.[4,16] However, once these isolated SWNTs were freeze-dried, they were difficult to redisperse in water again without sonication and instead formed aggregates.[16]

Although Kim[16] reported a kind of water redispersible SWNTs fabricated by the in situ polymerization of micelles, SWNTs often need to be further modified with “intelligent” surfaces to satisfy the demands of responsiveness to the environment when they are used as sensors in medical or biological chemistry.[17] SWNTs modified noncovalently with a responsive surfactant have sparked great interest among researchers in recent years because SWNTs not only were debundled but were also functionalized “smartly”. Furthermore, the original structures and properties of the SWNTs were preserved.[18] Therefore, many kinds of responsive surfactants have been reported to disperse SWNTs including pH,[19−23] redox,[24] temperature,[21,25,26] and light[26] responsive surfactants. However, these systems required the addition of extra acids, bases, or oxidants, which result in the accumulation of byproducts and contamination of the system and utilization of UV–vis radiation or changes in temperature are unattainable in some situations.[27] Thus, as a green, biocompatible, cheap, and easily removed trigger, CO2 has also been drawn increasing attention.[18,28] However, homogenizing these precipitated hybrids after bubbling with N2 requires not only to be bubbled with CO2 but also to be sonicated. Furthermore, research studies on green chemistry have attracted widespread attention.[29,30] Rosin is a renewable biomass resource secreted by pine trees. It has excellent hydrophobicity because of its unique tricyclic rigid structure, which is similar to that of cholic acid, and may effectively interact with the hydrophobic surface of the SWNTs through interaction, most likely π-stacking.[22,31]

Here, we report a kind of CO2-responsive polymerizable dispersant, maleopimaric acid glycidyl methacrylate ester 3-dimethylaminopropylamine imide (MPAGN), based on natural rosin. It was used to disperse SWNTs in aqueous solution followed by in situ free-radical polymerization to achieve a controllable SWNT dispersion. This SWNT dispersion modified by a noncovalent approach that not only can be controlled by CO2/N2, but can also be redispersed in CO2-saturated water after drying without sonication just requiring only mild shaking.

Results and Discussion

The structures of the SWNT samples were characterized by IR, 1H NMR, as shown in Figures S1 and S2. The synthetic route and reversible transformation of P-MPAGN between its nonionic and cationic states are shown in Figure A and a schematic illustration of this dispersion is shown in Figure B.

Figure 1

(A) Synthetic route and reversible transformation of P-MPAGN between the cationic and nonionic states. (B) Schematic illustration of the SWNTs dispersed by MPAGNH+ that are locked in by free-radical polymerization in situ.

At first, neutral MPAGN was protonated to form cationic MPAGNH+ by bubbling with CO2 into solution to disperse SWNTs and then the cationic MPAGNH+ was polymerized on the surface of the SWNTs to obtain P-SWNTs dispersion; the zeta potential of the P-SWNTs dispersion was approximately +58 mV. The electrostatic repulsion among the charged nanotubes can prevent the bundling of debundled SWNTs. After N2 was bubbled through the dispersion at 60 °C, the cationic P-SWNT hybrids were deprotonated and the nonionic sample was formed leading to a decrease in the zeta potential of the dispersion to nearly 0 mV. The electrostatic repulsion disappeared among the nonionic P-SWNTs, and the P-SWNTs precipitated. The nonionic P-SWNTs were protonated and returned to cationic P-SWNTs by bubbling of the sample with CO2 leading to a zeta potential of the dispersion returning to approximately +58 mV. This process can be reversed several times without an obvious change in the zeta potential of the dispersion as shown in Figure A. The transmittance of the dispersion at 600 nm was also measured to illustrate the responsiveness of the P-SWNTs to CO2/N2. A black powder composed of P-SWNTs settled out of solution by bubbling N2 through the water solution at 60 °C with the transmittance at 600 nm of the supernatant increasing, and this powder can be redispersed into water easily by bubbling CO2 through the solution, resulting in the transmittance at 600 nm decreasing to nearly 0. This reversal also can be repeated many times without a marked change in the transmittance of the dispersion at 600 nm, as shown in Figure B. The inserted photos compared the appearance of the P-SWNTs after bubbling with N2 and bubbling with CO2. After precipitation by bubbling with N2, the precipitate could be recovered and reused after drying. It can be redispersed in a CO2-saturated water solution again with only mild shaking and no visible aggregates appeared over a long time. The zeta potential and transmittance of redispersed P-SWNTs dispersion after being dried with bubbling with CO2 and N2 alternately is shown in Figure S3. The redispersed P-SWNTs dispersion after being dried also can be responsive to CO2/N2 and the cycle zeta potential, and transmittance of P-SWNTs dispersion after being dried was almost no different from that of the freshly prepared ones. It can be concluded that the drying operation has little influence on the responsiveness of P-SWNTs.

Figure 2

(A) Cyclic changes in zeta potential of the P-SWNTs dispersion with bubbling with CO2 and N2 alternately; (B) cyclic changes in transmittance at 600 nm and appearance of the P-SWNTs dispersion with bubbling with CO2 and N2 alternately; (C) UV–vis–NIR spectra of the U-SWNTs, as-prepared P-SWNTs, redispersed P-SWNTs after drying and redispersed P-SWNTs after N2/CO2; (D) TGA data for pristine SWNTs, dried U-SWNTs, and dried P-SWNTs.

As shown in Figure C, the UV–vis–NIR spectra of the U-SWNTs and P-SWNTs dispersion showed van Hove transition peaks, which indicated the existence of individually isolated SWNTs in the U-SWNTs and P-SWNTs dispersion,[16] and the UV–vis–NIR spectra of P-SWNTs was almost the same as that of U-SWNTs. It can be concluded that the optical properties of SWNTs were not destroyed after in situ polymerization. The UV–vis–NIR spectra of the redispersed P-SWNTs dispersion after responsiveness with N2/CO2 and that of dispersion after drying were basically in line with that of the as-prepared sample, which indicated the excellent responsiveness of the P-SWNTs to CO2 and their excellent redispersibility after drying. The UV spectra of P-SWNTs prepared after 1 month was basically in line with that of the freshly prepared ones, which indicates the excellent durability of P-SWNTs dispersion, as shown in Figure S4.

Thermogravimetric analysis (TGA) was used to count the amount of MPAGN or P-MPAGN attached to the nanotubes. These dispersions were filtered through a membrane (0.22 μm) and dried before measurement. The pristine SWNTs had lost little weight, while the P-SWNTs had a 77.66% weigh loss at 800 °C, as shown in Figure D. This discrepancy confirmed that the surfactant had been combined with SWNTs. In other words, the ratio of P-MPAGN to SWNTs in P-SWNTs was 3.48:1 by weight. However, U-SWNTs had a 53.9% weight loss at 800 °C, such that the ratio of MPAGN to SWNTs in U-SWNTs was 1.17:1 by weight. It can be concluded that the MPAGN in the U-SWNT sample was partially washed away with the filter liquor in the process of filtration because the intermolecular forces between MPAGN and SWNT were weak. The amount of residue MPAGN attached to the SWNTs in the U-SWNTs was insufficient to redisperse SWNTs in water. In addition, the intermolecular forces between MPAGN and the SWNTs in the U-SWNTs may be destroyed by harsh processing procedures, such as freeze-drying. Therefore, although a black powder consisting of U-SWNTs also settled out of solution after bubbling with either N2 or air through the aqueous solution at 60 °C, it could not be redispersed in water after freeze-drying by bubbling CO2 through the aqueous solution without sonication, rather appearing to visibly aggregate. However, the MPAGN monolayer on the SWNT surface was solidly locked in after the in situ free-radical polymerization of P-SWNTs, therefore, the P-SWNTs can be redispersed easily. The photographs of redispersion of U-SWNTs and P-SWNTs after drying are shown in Figure S5.

The SWNTs exhibited bundled and networked microstructures in water, as shown in Figure S6A. However, some well-dispersed and individualized nanotubes of U-SWNTs with diameters of 10–12 nm were observed from the transmission electron microscopy (TEM) image, as shown in Figure A, which were larger than those of the original SWNTs (approximately 2 nm). This could indicate that a thick MPAGNH+ layer was attached to the surface of the nanotubes effectively. After polymerization, the diameter of the individual tubes of the P-SWNTs shown in Figure B was larger than that of the U-SWNTs, from which it can be concluded that the polymerization of MPAGNH+ has reacted on the surface of the nanotubes and the obtained polymer MPAGNH+ has been wrapped around the surface of the nanotubes and prevent the bundling of the SWNTs. The diameter of individualized nanotubes wrapped by MPAGN was much larger than that of the ones wrapped by surfactants based on long-chain alkane because of the steric hindrance of large rigid structures. After bubbling with N2, the P-SWNTs precipitated from the aqueous solution, and its TEM image is shown in Figure C. When CO2 was bubbled through the solution again, the P-SWNT hybrid suspension became homogeneous without the need for sonication, and the polymer was still wrapped around the surface of the SWNTs, based on their diameters, as shown in Figure D. The same was seen for the redispersion after drying, as shown in Figure S6B. Therefore, P-SWNT could be redispersed in CO2-saturated water without sonication, requiring only mild shaking.

Figure 3

TEM images of aqueous; (A) U-SWCNTs; (B) P-SWNTs; (C) precipitated P-SWNTs after bubbling N2; (D) P-SWNTs after responding to CO2.

To clarify the influence of the dispersion treatment on the intrinsic properties of SWNTs, Raman spectrometry was carried out because it was sensitive to the change in both the electronic and optical properties of SWNTs. If there was charge transfer or the optical property change of SWNTs, the shift and peak shape change of the G band and radial breathing mode (RBM) would be observed. The tangential (G) band at 1589.7 cm−1 and the represent active second disordered band (D) at 2678.3 cm−1 were revealed from the Raman spectrometry of SWNTs, respectively. RBM of the nanotubes was found at 100–200 cm–1. However, the Raman spectra of U-SWNTs and P-SWNTs were almost the same as that of raw SWNTs as shown in Figure . It can be concluded that there was no charge transfer between the surfactant and SWNTs and the intrinsic optical properties and electronic properties of SWNTs were still reserved.

Figure 4

Raman spectra of SWNTs, U-SWNTs, and P-SWNTs.

Conclusions

In summary, SWNTs were modified by a CO2-switchable polymerizable dispersant based on natural rosin in water followed by in situ free-radical polymerization, and the tubes were wrapped by the obtained polymer. The dispersion and precipitation in an aqueous solution can be controlled reversibly by bubbling with CO2 or N2. After precipitation, the polymer surfactant-fabricated SWNTs could be redispersed in water after drying without the need for sonication, requiring only mild shaking because the polymer was still solidly attached to the surface of the SWNTs after either bubbling with N2 or drying, while the MPAGN-fabricated SWNTs without polymerization could not.

Experimental Section

MPAGN was synthesized following our previous report.[32] A mixture of MPAGN (8 mg) and SWNTs (0.8 mg) in water (24 mL) bubbled with CO2 was treated by sonication for 20 min (750 W and 20 kHz, Sonics), followed by centrifugation at 8000 rpm for 10 min to obtain a homogeneous dispersion of SWNTs with an adsorbed layer of MPAGN (U-SWNTs). The MPAGN monolayer on the SWNT surface was locked in by in situ free-radical polymerization (P-SWNTs), and the initiator was 2,2′-azobis(2-methylpropionamide)dihydrochloride (V50). The dispersion was filtered through membranes (0.22 μm) to remove the unreacted monomer and poly-MPAGN that was not wrapped around the SWNTs (P-MPAGN).