Fractioning HiPco and CoMoCAT SWCNTs viadensity gradient ultracentrifugation by the aid of a novel perylene bisimide derivative surfactant

Abstract

HiPco and CoMoCAT single-walled carbon nanotubes (SWCNT) were fractionated with the aid of a novel perylene bisimide surfactant by combined co-surfactant and replacement density gradient ultracentrifugation (DGU).

---

Even though carbon nanotubes promise to become a subject of technological progress in electronics, photonics and as sensors due to their remarkable physical and electronic properties,¹ their applications have so far been limited by the inhomogeneity of the as-synthesized material. Thus, separation of SWCNTs according to helicity and diameter is unambigous and has been addressed by a variety of techniques including dielectrophoresis,² chemical functionalization,³anion exchange chromatography⁴ and size-exclusion chromatography.⁵

Recently, the separation of SWCNTs by exploiting subtle differences in the buoyant density of surfactant encapsulated nanotubes in density gradient ultracentrifugation (DGU) has been reported.^6–9 In DGU, a density gradient is created by a gradient medium such as Iodixanol or nycodenz, as a response to the centripetal force. As soon as the gradient is formed, SWCNTs travel to their isopycnic points in the centrifuge vial, resulting in their separation according to their buoyant density. As has been outlined by Green and Hersam,¹⁰ the buoyant density of a dispersed CNT is both a function of the geometry of the nanotube (diameter) and the surfactant coating. On the one hand, a completely uniform surfactant coating, e.g.sodium cholate (SC) encapsulation of CoMoCAT SWCNTs,⁶ results in separation by diameter while on the other hand, the use of a surfactant (or a combination of surfactants) that encapsulates nanotubes in a manner dependant on the electronic structure, allows separation by electronic properties, as has been shown by co-surfactant DGU with 1 : 4 sodium dodecyl sulfate (SDS)/SC.⁶ Therefore, the investigation of different surfactant systems and various co-surfactant experiments is crucial to exploit the tunability of the separation process by DGU.

Herein, we report on the fractionation of both CoMoCAT and HiPco SWCNTs in DGU by the use of a novel perylene bisimide surfactant, 1, which has been shown to individualize SWCNTs in buffered aqueous media to a high degree.¹¹

The degree of individualization is specifically important in DGU, as bundled SWCNTs possess a higher density compared to individual nanotubes and so sediment to the bottom of the centrifuge vial without being separated. Furthermore, as mentioned above, the surfactant coating has a tremendous impact on the separation process, so that exploration of a variety of surfactants is vital.

Prior to the density differentiation, both HiPco and CoMoCAT SWCNTs have been dispersed in a buffered aqueous solution of 1 (0.1 wt%, pH = 7) by the aid of ultrasonication (30 min, 80 W, 45 kHz) to yield initial SWCNT concentrations of 0.5 g l⁻¹. The dispersions were centrifuged (15 000 rpm, 30 min) in order to remove coarse aggregates. For DGU, a procedure similar to that described by Arnold et al. has been carried out.⁶ To this end, a step gradient containing the surfactant in 60 wt%, 40 wt% and 20 wt% nycodenz, respectively, was prepared. The nanotubes dispersed in the aqueous surfactant solution containing 10 wt% nycodenz were layered on top of the gradient. The vials were then centrifuged in a Beckman Coulter ultracentrifuge equipped with a swinging bucket rotor (SW60Ti) for 18 h at 15 000 rpm. Fig. 1 shows the result for CoMoCAT (Fig. 1a) and HiPco SWCNTs (Fig. 1b), respectively.

Fig. 1 Centrifuge vials after DGU (18 h, 15 krpm) in a nycodenz gradient (a) CoMoCAT SWCNTs, (b) HiPco SWCNTs; from left to right: (I) SWCNTs dispersed in SDS in a gradient containing SDS ([SDS] = 2 wt%); (II) SWCNTs dispersed in 1 with [1] = 0.1 wt% throughout the centrifuge tube; III) SWCNTs dispersed in 1 with [1] = 0.1 wt% and [SDS] = 2 wt% throughout the centrifuge tube; IV) SWCNTs dispersed in 1 [1] = 0.1 wt% in a gradient containing SDS ([SDS] = 2 wt%).

When nanotubes dispersed in 1 ([1] = 0.1 wt%) are centrifuged in a nycodenz gradient also containing 1, no clear separation of the material in bands could be observed (trace II, Fig. 1). However, density differentiation occurred with the aid of SDS as a co-surfactant. For this purpose, SDS (2 wt%) was added in the step gradient (except for the top layer containing the nanotubes), in addition to 1. In the case of the CoMoCAT tubes, the formation of four bands was observed, while five bands could be identified for the HiPco material (trace III, Fig. 1). It is important to note that no band formation could be observed for SDS alone under the same centrifugation conditions (trace I, Fig. 1).

The clearest band formation for both HiPco and CoMoCAT SWCNTs (trace IV, Fig. 1) was obtained when nanotubes dispersed in a buffered aqueous solution of perylene 1 were added on top of a step gradient containing 2 wt% of SDS (without perylene in the step gradient). Presumably, this procedure not only resulted in nanotube band formation, but also in the replacement of the perylene units from the sidewall of the nanotube by SDS during the movement of the nanotubes through the gradient, as SDS is present in large molar excess compared to 1. It shall be mentioned that the diffusion of perylene throughout the centrifuge vial is very slow as indicated by the dark red color attributed to free bulk perylene in the top one third of the centrifuge tube which is not spread all over the vial (trace IV, Fig. 1).

In order to analyze the nanotube fractions, UV/Vis/nIR absorption spectroscopy was carried out after isolating the clearly separated bands of the right centrifugation vials in Fig. 1. It shall be mentioned that the detection range was limited to 600–1400 nm due to the strong absorption of the perylene core centered at 520 nm.

Fig. 2a reveals that, on the one hand, fractions 3–4 (F3–4) of the CoMoCAT material contain a large amount of SWCNT bundles, as the intensity of the resonant absorption peaks attributed to transitions between the mirror image spikes in the density of states of the SWCNTs (van Hove singularities—vHS) compared to the nonresonant background of the absorption remains low. On the other hand, F1 and F2 both show a significantly higher degree of individualized tubes as can be concluded from the higher resonant ratios¹² compared to F3–4.

Fig. 2 (a) As recorded UV/Vis/nIR absorption spectra of the CoMoCAT fractions 1–4 (F1–4) as indicated in Fig. 1 compared to pristine SWCNTs dispersed in a solution of SDS ([SDS] = 2 wt%), b) UV/Vis/nIR absorption spectrum normalized to the minimum at 795 nm of CoMoCAT-fractions 1–2 after addition of SDBS ([SDBS] ≈ 5 wt%) compared to the pristine material dispersed in SDBS ([SDBS] = 1 wt%)—vertical offset for clarity.

An evaluation of the enrichment and depletion of certain (n,m)-SWCNT species is challenging, however, as the nanotubes are dispersed in aqueous media containing different relative amounts of surfactants, e.g. F1 contains free bulk perylene, while free perylene 1 is absent in F2. Thus, the dielectric environment of the nanotube is different for F1 and F2 resulting in changes in the absorption pattern, e.g. adsorption of the perylene unit results in a red-shift of the vHS for F1.

To increase comparability, a solution of SDS was added to SWCNTs dispersed in a buffered aqueous solution of 1 ([1] = 0.1 wt%) in order to simulate the replacement of perylene by SDS during DGU. As depicted by Fig. S1, the peak position is then the same as for F1 mirrowing the same chemical environment for both systems. Fraction 2 exhibits similar optical absorption as pristine CoMoCAT SWCNTs dispersed in a solution of SDS (2 wt%), as shown by Fig. S2, further underlining the differing chemical surrouding of the nanotubes in F1 and F2, respectively.†

In order to evaluate differences in the fractionated nanotube material, furthermore, the anionic surfactant SDBS (sodium dodecylbenzenesulfonate) was added to the fractions in a high excess ([SDBS] = 5 wt%) compared to 1 and SDS, as it has previously been shown¹¹ that a solution of SDBS is capable of replacing 1 from the nanotube sidewall. Moreover, introducing SDBS on the nanotube sidewall is advantageous with regard to two aspects: (i) the formation of well resolved absorption and emission features of the nanotube is ensured, (ii) comparison to literature is possible as SDBS is commonly used.¹³ As shown by Fig. 2b, DGU of CoMoCAT SWCNTs with the aid of 1 leads to an enrichment of the (6,4) and (9,1)-SWCNTs in the region of the S₁₁-transitions in F1, while fraction 2 is depleted of the (6,5), (6,4) and (9,1) SWCNTs.

A similar behavior is observed for the HiPco material, although the differences in the absorption spectra are more striking (Fig. 3). In the case of the HiPco tubes, F3–5 appear rather bundled as can be concluded from the low resonant ratio as indicated by Fig. 3a. Again, F1 appears comparable in the structure of the absorption to the pristine material dispersed in 1 ([1] = 0.1 wt%) after addition of SDS as depicted by Fig. S3, while absorption of F2 resembles HiPco tubes dispersed in SDS (Fig. S4).†

Fig. 3 (a) As recorded UV/Vis/nIR absorption spectra of the HiPco fractions 1–5 (F1–5) as indicated in Fig. 1 compared to pristine SWCNTs dispersed in a solution of SDS ([SDS] = 2 wt%), (b) UV/Vis/nIR absorption spectrum normalized to the minimum at 1000 nm of HiPco-fractions 1–2 after addition of SDBS ([SDBS] ≈ 5 wt%) compared to the pristine material dispersed in SDBS ([SDBS] = 1 wt%)—vertical offset for clarity.

After addition of SDBS, the S₁₁-transitions of the smaller diameter nanotubes (from 975–1175) are more dominant compared to the pristine material, as represented by the green arrows pointing upwards in Fig. 3b. Additionally, the spectral features in the region from 1175 nm–1250 nm are significantly decreased indicating depletion of larger diameter tubes in F1. This is consistent with the observation that F2 is enriched in larger diameter tubes (blue arrows pointing upwards in Fig. 3b), while being depleted in smaller diameter tubes, as would be expected from DGU with a uniform surfactant coating where density differentiation is related only to the diameter as outlined by Hersam at al.⁹ However, presumably, density differentiation is more complex in our cases, as no clear band formation is observed when nanotubes dispersed in 1 are ultracentrifuged in a nycodenz gradient without the presence of SDS (trace II, Fig. 1).

To the best of our knowledge, this is one of the first reports on fractioning pristine HiPco SWCNTs by DGU, as DGU with HiPco SWCNTs has not shown to be successful under the conditions of separation by DGU with CoMoCAT material⁶ due to the wider diameter distribution of the HiPco material. Therefore, we believe that fractioning in our case is based on the selective replacement of the perylene moieties adsorbed on the sidewall of the SWCNT by SDS which would increase the possibilities of tailoring the separation process of SWCNTs by DGU. Further investigations concerning the use of 1 as surfactant in DGU and the expansion of the concept of a combined co-surfactant and replacement DGU are currently under way in our laboratories.

This study was supported by BASF SE. Part of this work was conducted within the Polymer Physics and Nanotechnology Innovation Team at BASF. We also thank the Deutsche Forschungsgemeinschaft (DFG) and the Interdisciplinary Center for Molecular Materials (ICMM) for financial support.

Notes and references

1. M. S. Dresselhaus, G. Dresselhaus and P. Avouris, Carbon Nanotubes: Synthesis, Structure, Properties, and Applications, Springer, Berlin, 2001.
2. R. Krupke, F. Hennrich, v. Lohneysen Hilbert and M. Kappes, Science, 2003, 301, 344–347.
3. M. S. Strano, C. A. Dyke, M. L. Usrey, P. W. Barone, M. J. Allen, H. Shan, C. Kittrell, R. H. Hauge, J. M. Tour and R. E. Smalley, Science, 2003, 301, 1519–1522.
4. M. Zheng, A. Jagota, M. S. Strano, A. P. Santos, P. Barone, S. G. Chou, B. A. Diner, M. S. Dresselhaus, R. S. McLean, G. B. Onoa, G. G. Samsonidze, E. D. Semke, M. Usrey and D. J. Walls, Science, 2003, 302, 1545–1548.
5. G. S. Duesberg, M. Burghard, J. Muster, G. Philipp and S. Roth, Chem. Commun., 1998, 435–436.
6. (a) M. S. Arnold, A. A. Green, J. F. Hulvat, S. I. Stupp and M. C. Hersam, Nat. Nanotechnol., 2006, 1, 60–65; (b) M. S. Arnold, S. I. Stupp and M. C. Hersam, Nano Lett., 2005, 5, 713–718.
7. J. Crochet, M. Clemens and T. Hertel, J. Am. Chem. Soc., 2007, 129, 8058–8059.
8. Y. Miyata, K. Yanagi, Y. Maniwa and H. Kataura, J. Phys. Chem. C, 2008, 112, 3591–3596.
9. M. C. Hersam, Nat. Nanotechnol., 2008, 3, 387–394.
10. A. A. Green and M. C. Hersam, Mater. Today, 2007, 10, 59–60.
11. C. Backes, C. Schmidt, F. Hauke, C. Böttcher and A. Hirsch, J. Am. Chem. Soc., 2009 DOI:10.1021/ja805660b.
12. Y. Tan and D. E. Resasco, J. Phys. Chem. B, 2005, 109, 14454–14460.
13. B. R. Priya and H. J. Byrne, J. Phys. Chem. C, 2008, 112, 332–337.

---

Footnote

† Electronic supplementary information (ESI) available: Absorption spectra of the nanotube fractions compared to the spectra of pristine nanotubes in co-surfactant solutions of 1 and SDS. See DOI: 10.1039/b818141a

---