Silsesquioxane-cored star amphiphilic polymer as an efficient dispersant for multi-walled carbon nanotubes

Abstract

One octopus-shaped amphiphilic polymer was used to functionalize MWNTs for the first time. Characterization results showed that the octopus-shaped amphiphilic polymer was much more efficient at modifying MWNTs than linear polymers in aqueous or acetone solution. The hydrophobic cubic-silsesquioxane core locally anchored to the surface of the nanotubes, while the arms of the RAFT polymer extended into the aqueous solution, thus imparting steric stabilization to the nanotube dispersion.

---

Since the first discovery of carbon nanotubes (CNTs) in 1991,¹ they have received increasing attention because of their unique physical, chemical, and structural properties, leading their use in all sorts of fields, such as polymeric composites,² electronic devices,³ field emission displays⁴ and hydrogen storage.⁵ It is crucial to obtain homogeneous and fine dispersions of CNTs in solution and recently, a few approaches have been successfully developed,^6–8 which mainly include covalent surface modification^9,10 and non-covalent surface modification. Non-covalent modification^11,12 of the surface has attracted considerable interest as a promising means for the modification of CNTs surface properties with no damage to the sp²-hybridized structure of the CNTs, and involves π–π interaction, hydrogen bonds, ionic interaction and CH–π interaction. Organic modifiers widely employed for this purpose are either small-molecule surfactants or amphiphilic block copolymers. As for the latter, surface wrapping and adsorption of polymer aggregates represent a mechanism that governs the dispersing efficiency. In previously published works on modification in either aqueous or organic media, polymeric structures have shown the possibility for a better dispersion of CNTs due to the great affinity of the hydrophobic domains of the polymers for the nanotubes.^13–15 Furthermore, the versatile chemistry of the polymer modifier allows for designing and tailoring nanotube/polymer composites with finely tuned properties.^16–20

Herein, we report a new method to disperse multi-walled carbon nanotubes (MWNTs) with an “octopus” copolymer in solution via noncovalent binding interactions. In our previous work, we had demonstrated that amino-functionalized linear amphiphilic quaternary copolymers could form a uniform and stable physical adsorption layer on the surface of the nanocarbon material, yielding better efficiencies toward dispersion and functionalization.^21,22 Silsesquioxane cages (CSQ)^23,24 are regarded as organic/inorganic hybrid materials at the molecular level, which can be introduced into polymers to readily tune the properties of materials for their special structural characteristics. Recently, considering that the densely packed morphology of polymer chains renders segment/CNT surface interaction more favorable,^25–27 we were motivated to extend MWNT modification to cubic silsesquioxane-cored star-like polymers. Firstly, multifunctional CSQ would be very conducive to forming star-shaped polymers.²⁸ Secondly, star-shaped polymers that contained multiple binding sites could form larger polymer aggregates than linear analogs.¹³ Finally, the lipophilic silsesquioxane shows a tendency to stabilize latexes,²⁹ which could generate hydrophobic–hydrophobic interactions between amphiphilic polymers and CNTs.¹⁴

Our synthetic strategy was based on a two-step procedure involving the condensation of carboxylic acid functionalized trithiocarbonate with octa(aminophenyl)silsesquioxane, followed by a reversible addition–fragmentation chain transfer (RAFT) polymerization of N,N-(dimethylamino)ethyl methacrylate to yield a grafted polymer (denoted as CSQ-PDMAEMA). In order to further explore the dispersing efficiency of the star-like modifiers, we also grafted N-vinyl carbazole onto a PDMAEMA segment (denoted as CSQ-PDMAEMA-CZ). FTIR, ¹H NMR and GPC analyses confirmed that we indeed obtained star-shaped polymers with narrow molecular weight distribution. For comparative purposes, their linear analogues, PDMAEMA and poly(N,N-dimethylaminoethyl)-b-poly(vinylcarbazole) (PDMAEMA-CZ) were also prepared by the RAFT method. Their chemical structures are illustrated in Scheme S1.†

In order to assess the efficiency of dispersion of MWNTs modified by star-like and linear modifiers in aqueous solution, we prepared four modified MWNT (m-MWNT) dispersions (M1, M2, M3, M4) and investigated their UV-vis absorption (Fig. 1a). Nearly no signal in the UV-vis region could be detected for the MWNT dispersions; there was no evident change in the absorption spectra with or without the modifiers, leading us to tentatively postulate that the solubilization process had not resulted in the destruction of the nanotubes or significant alteration of their structure. As expected, only carbazole-containing dispersions had ultraviolet absorption peaks near 340 nm.³⁰ We assumed that the amount of modifiers in the supernatants would be reduced as the modifiers attached onto the surface of MWNTs and therefore, the UV absorption intensity would decrease. From Fig. 1b, it was clearly indicated that the absorption intensity of the supernatants for both star-like (M2) and linear (M4) modifiers decreased significantly in comparison with that of the original modifier solutions diluted in the same ratio.

Fig. 1 UV-vis spectra of the modifiers and MWNT dispersions. (a) Absorption profiles for MWNT dispersions in an aqueous solution with modifiers. The dispersions were prepared by dispersing MWNTs in 0.5 g L⁻¹ polymer aqueous solution (initial mass ratio of 1 : 1 modifiers/MWNT). M1, M2, M3, M4 denote MWNT aqueous dispersions in which MWNTs were dispersed in CSQ-PDMAEMA, CSQ-PDMAEMA-CZ, PDMAEMA and PDMAEMA-CZ aqueous solution, respectively. (b) Absorption profiles for modifiers and supernatants of MWNTs dispersions. The supernatant was obtained by merging the centrifugal supernatant liquid and the washing liquid (the total volume of washing liquid was four times as much as the volume of dispersion) and all modifier solutions were diluted to 1/5 of their original concentration.

The sidewall surface of pristine MWNTs is highly hydrophobic; they could not be dispersed homogeneously in aqueous solution in the absence of a dispersant, even despite the vigorous ultrasonication, as confirmed in this sedimentation experiment (Fig. 2). We visually assessed the modification ability of linear and “octopus” polymers on MWNTs by comparing the stability of m-MWNTs dispersions. All modifiers enhanced the dispersing of MWNTs, and m-MWNTs in water/acetone solutions were more stable than these in aqueous solution. In particular, it seemed that the “octopus” polymers/MWNT dispersions were more stable, especially in water/acetone solutions. The stability of these “octopus” polymers/MWNT dispersions was observed in vials for 20 days without any precipitation or aggregation. We therefore inferred that the CSQ core played a key role in determining the nanotube dispersion. Firstly, CSQ cores and the graphite structure of the nanotube wall render both of them hydrophobic, thus allowing for an effective hydrophobic interaction between the modifying polymer and the sidewall surface, which primarily promotes nonpolar substances to minimize their contacts with water.^25,27 This speculation can be supported by the fact that the grafting of hydrophobic carbazole groups to linear modifiers resulted in an improved dispersibility. Then, thanks to the tentacles of “octopus” polymers, the solvation segments would have more possibilities to generate a great affinity for nanotubes. The “octopus” modifiers exhibited similar behaviors in weak polar organic solvent, while the presence of organic solvent was possibly beneficial to the stretching of polymer chains.

Fig. 2 Digital images of MWNT dispersions. M1–M5 denote the MWNTs dispersion; A1, A2, A3, A4 denote MWNT dispersions in which CSQ-PDMAEMA, CSQ-PDMAEMA-CZ, PDMAEMA and PDMAEMA-CZ, respectively are dissolved in 1 : 1 volume ratio of water/acetone solution as dispersant.

Direct evidence of modifying polymer-MWNT interactions were then obtained using FTIR (Fourier transform infrared spectroscopy), thermogravimetric analyses and transmission electron microscopy (TEM) observation after dispersion and centrifugation processes. As shown in Fig. S3,† FTIR spectra of pristine MWNT and all the modified MWNT show clear absorption bands at 3440 cm⁻¹, which correspond essentially to the –OH stretching vibration of surface hydroxyl groups on the nanotubes and the adsorbed water molecules. Modified MWNT exhibited an additional shoulder absorption at around 3567 cm⁻¹, which could be possibly attributed to carboxyl groups in the RAFT agent. An intense band centered at 1634 cm⁻¹ appeared in all samples and was assigned to the stretching vibration of carbonyl groups (C=O) present in carboxylic acids and conjugated C=C groups on the nanotubes.^21,22 Besides, in the spectra of modified MWNT, we observed an obvious C=O stretching vibration of ester groups of polymer arms (ca. 1719 cm⁻¹) and a C–O stretching in the vicinity of 1232 and 1149 cm⁻¹. The above observation suggested that the modifying polymers could be effectively adsorbed onto the surfaces of MWNT by non-covalent interactions, even after rinsing.

To confirm the strong and stable interfacial interactions between MWNTs and the “octopus” modifiers, we performed transmission electron microscopy (TEM) analysis. As observed from high-resolution TEM images (Fig. 3), all MWNTs retained their structural integrity; no damages to the nanotube walls were evident in the processes of the experiment. The pristine MWNTs aggregated and were entangled by way of large ropes, without any modification; however, m-MWNTs apparently tended to exist as individuals or in very small bundles, which was in sharp contrast to the case of the pristine nanotubes. Moreover, these star-like polymer-modified MWNTs had more individually dispersed nanotubes and less nanotube bundles, compared to linear polymer-modified MWNTs. Among the linear polymer-modified MWNTs, PDMAEMA-CZ-modified MWNTs had less entanglement (Fig. 3d). In addition, the tube surfaces of m-MWNTs were rougher, especially for star-like polymer-modified MWNTs, in contrast to the case of pristine nanotubes. Individual MWNTs showed thickness variations along their lengths, attributed commonly to regions of irregular polymer coating. These visual observations agreed well with UV-vis spectra and FTIR results, revealing the adherence of polymer to the nanotube surface.

Fig. 3 HR-TEM images of the modified CNTs and the interaction mechanism between MWNTs with modifiers. (a) M1 residue, (b) M2 residue, (c) M3 residue, (d) M4 residue, (e) M5 residue.

Thus, the possible mechanism for the interaction between MWNTs and modifiers in Fig. 3 was put forward.

Due to an effective hydrophobic interaction between the CSQ-cores and the sidewall, star polymers are readily adsorbed onto the nanotube surfaces. The tentacles on the side of the carbon tube would therefore wrap the tube by forming non-covalent forces and the other tentacles on the opposite side had to stretch into the solvent to form “buoys” for the steric hindrance of CSQ. In contrast, with the irregular distribution of linear modifier in solvent, the “octopus” modifiers were more inclined to adhere to the sidewall surface to form polymer aggregations; such aggregations sequentially reduced the hydrophobic area of the sidewall surface to minimize its contact with water or other tubes. Meanwhile, these hydrophilic “buoys” were able to greatly increase the solubilization of MWNT in water, thus leading to well-dispersed nanotubes in solution. For the carbazole-containing polymers, hydrophobic carbazole groups were able to form a π–π stacking system with the graphite structure of the nanotube to increase adhesion with tubes.

To gain a more quantitative picture of the extent of nanotube modification, TG analysis was performed on the residues (Fig. 4). The CSQ-DMAEMA, CSQ-DMAEMA-CZ and m-MWNTs residues showed a two-step thermal degradation that might be caused by the loss of ester bonds followed by backbone decomposition in the temperature range of 200–500 °C, while pristine MWNT (M5) had almost no mass loss in the same range. In the final stage of thermal degradation, the decomposition temperature of CSQ-DMAEMA was higher than that of PDMAEMA (Fig. S6†). This was probably attributed to the CSQ degradation.²² Meanwhile, the decomposition temperature of CSQ-DMAEMA-CZ was higher than CSQ-DMAEMA, probably due to the thermal enhancement of carbazole degradation.³¹ Considering that the modifiers would be completely decomposed at a temperature of 500 °C, as indicated in Fig. 4, the mass losses from the modified CNTs residues at 500 °C were used to estimate the amount of copolymers that were wrapping the MWNTs. The weight losses of the M1, M2, M4 and M4 residues deducting the mass loss of pristine MWNTs were respectively about 35%, 40%, 32% and 24% between 200 °C and 500 °C. Obviously, MWNT modified by star-like polymers had more mass losses than by the linear ones, thus providing a sound support to our previous suggestion that more “octopus” modifiers attached onto the MWNT surfaces. From TG curves, it also appeared that more carbazole-containing polymers adhered to the surface of MWNT.

Fig. 4 TG curves of CSQ-DMAEMA, CSQ-DMAEMA-CZ and the residues of M1, M2, M3, M4, M5.

Conclusions

In summary, we successfully designed and synthesized one octopus-shaped amphiphilic polymer (CSQ-PDMAEMA) via RAFT polymerization; it was firstly used to functionalize MWNTs. We also synthesized three other polymers as control samples (CSQ-PDMAEMA-CZ, PDMAEMA, PDMAEMA-CZ). The experimental results indicated that the “octopus” polymers/MWNT dispersions were more stable, especially in water/acetone solutions. Direct evidence of modifying polymer-MWNT interactions were obtained using UV-vis, FTIR, HR-TEM and TGA. Characterization results showed that the octopus-shaped amphiphilic polymer was much more efficient at modifying MWNTs than linear polymers in aqueous or acetone solution. The possible mechanism for the interaction between MWNTs and the octopus-shaped amphiphilic polymer is that the hydrophobic cubic-silsesquioxane core locally anchored to the surface of the nanotube, while the arms of the RAFT polymer extended into aqueous solution and amphiphilic polymer aggregations sequentially reduced the hydrophobic area of the sidewall surface to minimize its contact with water or other tubes.

Acknowledgements

The study was financially supported by the National Nature Science Foundation of China (No. 21174162) and the Integration of Industry, Education and Research of Guangdong Province Project (2011A091000007). Shuxi Gao and Zhiwei Yu contributed equally to this work.

Notes and references

1. S. Iijima, Nature, 1991, 354, 56–58.
2. X. Q. He, M. Rafiee, S. Mareishi and K. M. Liew, Compos. Struct., 2015, 131, 1111–1123.
3. S. Ahn, A. Choe, J. Park, H. Kim, J. G. Son, S.-S. Lee, M. Park and H. Ko, J. Mater. Chem. C, 2015, 3, 2319–2325.
4. S. Sridhar, L. Ge, C. S. Tiwary, A. C. Hart, S. Ozden, K. Kalaga, S. Lei, S. V. Sridhar, R. K. Sinha, H. Harsh, K. Kordas, P. M. Ajayan and R. Vajtai, ACS Appl. Mater. Interfaces, 2014, 6, 1986–1991.
5. T. Das, S. Banerjee, K. Dasgupta, J. B. Joshi and V. Sudarsan, RSC Adv., 2015, 5, 41468–41474.
6. A. Ghozatloo, A. M. Rashidi and M. Shariaty-Niasar, Int. Commun. Heat Mass Transfer, 2014, 54, 1–7.
7. N. Kemnade, C. J. Shearer, D. J. Dieterle, A. S. Cherevan, P. Gebhardt, G. Wilde and D. Eder, Nanoscale, 2015, 7, 3028–3034.
8. H. Karimi-Maleh, P. Biparva and M. Hatami, Biosens. Bioelectron., 2013, 48, 270–275.
9. J. L. Bahr and J. M. Tour, J. Mater. Chem., 2002, 12, 1952–1958.
10. R. Lv, E. Cruz-Silva and M. Terrones, ACS Nano, 2014, 8, 4061–4069.
11. B. Reuillard, A. Le Goff and S. Cosnier, Chem. Commun., 2014, 50, 11731–11734.
12. C. Wang, R. Yuan, Y. Chai, S. Chen, Y. Zhang, F. Hu and M. Zhang, Electrochim. Acta, 2012, 62, 109–115.
13. P. D. Petrov, G. L. Georgiev and A. H. E. Müller, Polymer, 2012, 53, 5502–5506.
14. Y. Sun, W. Fu, Z. Li and Z. Wang, Langmuir, 2014, 30, 8615–8620.
15. T. J. Simmons, J. Bult, D. P. Hashim, R. J. Linhardt and P. M. Ajayan, ACS Nano, 2009, 3, 865–870.
16. B. Yilmaz, J. Bjorgaard, C. L. Colbert, J. S. Siegel and M. E. Köse, ACS Appl. Mater. Interfaces, 2013, 5, 3500–3503.
17. B.-S. Kim, D. Kim, K.-W. Kim, T. Lee, S. Kim, K. Shin, S. Chun, J. H. Han, Y. S. Lee and H.-j. Paik, Carbon, 2014, 72, 57–65.
18. B. Rosner, D. M. Guldi, J. Chen, A. I. Minett and R. H. Fink, Nanoscale, 2014, 6, 3695–3703.
19. M. M. Hamedi, A. Hajian, A. B. Fall, K. Håkansson, M. Salajkova, F. Lundell, L. Wågberg and L. A. Berglund, ACS Nano, 2014, 8, 2467–2476.
20. N. Yeole, S. N. R. Kutcherlapati and T. Jana, J. Colloid Interface Sci., 2015, 443, 137–142.
21. J. Wu, H. Cai, K. Xu, Z. Fu, X. Liu, M. Chen and X. Zhang, Chem. Phys. Lett., 2012, 552, 78–83.
22. J. Wu, H. Cai, K. Xu, Z. Fu, X. Liu and M. Chen, J. Mater. Chem. C, 2013, 1, 646–656.
23. J. Peng, K. Xu, H. Cai, J. Wu, W. Lin, Z. Yu and M. Chen, RSC Adv., 2014, 4, 7124–7131.
24. K. Tanaka and Y. Chujo, J. Mater. Chem., 2012, 22, 1733–1746.
25. I. Choi, D. D. Kulkarni, W. Xu, C. Tsitsilianis and V. V. Tsukruk, Langmuir, 2013, 29, 9761–9769.
26. H. Durmaz, A. Dag, U. Tunca and G. Hizal, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 2406–2414.
27. Z. Iatridi and C. Tsitsilianis, Soft Matter, 2013, 9, 185–193.
28. D.-G. Kim, J. Shim, J. H. Lee, S.-J. Kwon, J.-H. Baik and J.-C. Lee, Polymer, 2013, 54, 5812–5820.
29. Y. Xing, J. Peng, K. Xu, W. Lin, S. Gao, Y. Ren, X. Gui, S. Liang and M. Chen, Chemistry, Weinheim an der Bergstrasse, Germany, 2016.
30. B. Zhang, Y. Chen, X. Zhuang, G. Liu, B. Yu, E.-T. Kang, J. Zhu and Y. Li, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 2642–2649.
31. M. Häussler, J. Liu, R. Zheng, J. W. Y. Lam, A. Qin and B. Z. Tang, Macromolecules, 2007, 40, 1914–1925.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00130k

---