Aqueous two-polymer phase extraction of single-wall carbon nanotubes using surfactants

Abstract

This review details the current state of the art in aqueous two-phase extraction (ATPE) based separations of surfactant dispersed single-wall carbon nanotubes by their chemical species, i.e., (n,m) structure, semiconducting or metallic nature, and enantiomeric handedness. Discussions of the factors affecting each separation, including workflow effects, variations of different surfactant and nanotube materials, and the underlying physical mechanism are presented. Lastly an outlook on the applications of ATPE at bench scale and implementation to larger scales is discussed, along with identification of research directions that could further support ATPE development.

Fig. 1. Phase diagram of a 6 kDa PEG and dextran 70 ATPE system in water determined by the cloud point method without (green diamond symbols) or in the presence of surfactants (red filled circles) at room temperature (≈21 °C). Compositions of PEG and DEX within the two-phase region, such as at the open blue circle, will at equilibrium form two phases at compositions located along the coexistence curve (dashed green eyeguide/dot-dash red eyeguide), and connected by a tie-line (black dotted eyeguide). In the PEG–dextran ATPE system the upper phase is PEG-rich phase, and the lower phase is dextran-rich. All initial compositions along a tie-line produce identical composition of the two phases, but at different volume ratios due to mass conservation (e.g. schematic volumes at the blue filled circles). The red circles report the coexistence curve, also called the binodal, of the polymers in the presence of 0.1% (mass basis) DOC and 0.75% (mass basis) SDS, which shifts the coexistence curve to slightly greater polymer mass fractions. Most conditions encountered in the use of surfactant-ATPE for SWCNTs will display similar coexistence curves.

Fig. 2. (A) Schematic of a multistage ATPE separation. (B) Schematic of the separation coefficient functionality for a semiconducting–metallic SWCNT separation at constant SC and oxidant (NaClO) concentration as a function of SDS addition. (C) Schematic of the separation coefficient functionality of different (n,m) SWCNTs for SDS/DOC competition separations as a function of SDS concentration for constant DOC. At the SDS concentration shown by the red (dashed) vertical line, approximately all of the (8,6), (9,4), (6,6) and most of the (7,5) would partition to the top phase, with the other shown species partitioning to the bottom phase. (D) Points showing the approximate SDS concentration required to partition the labelled species into the top phase of the PEG 6 kDa:DEX 70 ATPE system at ≈20 °C and a DOC concentration of 0.05%. The data quality was insufficient to discriminate enantiomers. Adapted from Fagan et al.,ACS Nano, 2015, 9, 5377–5390.

Fig. 3. (A) Schematic of SWCNT dispersion purification by ATPE as described by Subbaiyan et al., and with the variation of adding a second step to transfer the purified dispersion to a top phase for collection. For DOC dispersed SWCNT populations, the DOC concentration is reduced to a quantity that enables partition of many impurity types into the top phase, which can be discarded, while retaining the individualized SWCNTs in the bottom phase. In the variation including a second step, addition of SDS and reduction of the DOC concentration with addition of more top phase polymer, is specified to just at the condition needed to partition all SWCNT species to the top phase. Additional impurities will remain in the bottom phase if done correctly. (B) Figures from Subbaiyan et al. demonstrating the purification of SWCNT populations of several different average diameters that partition to the bottom phase relative to the original dispersion as shown by the reduction in absorbance that does not display the specific SWCNT optical transitions. Reprinted with permission from Springer Nature: Subbaiyan et al., Nano Res., 2015, 8(5), 1755–1769. Copyright 2014.

Fig. 4. (A) Photographs of water-filled electric arc synthesis SWCNTs being separated on the basis of semiconducting and metallic electrical properties from start to finish. The separation conditions for the semiconducting/metallic separation in the in the 3^rd, 4^th and 6^th photos were 0.7% SDS, 0.9% SC and (4 to 5) μL mL⁻¹ 1/100^th conc. NaClO. After separation, the red-orange-colored semiconducting SWCNTs primarily reside in the top phase, and the blue-green metallic SWCNTs primarily in the bottom phase. The photos are representative of the SWCNT concentration at which the separation is typically conducted in our labs. (B) A possible workflow for increasing semiconducting purity of the top fraction by repeatedly performing ATPE steps to remove contaminant metallic SWCNTs to fresh bottom phase. (C) A recommended replacement workflow in which a temperature change is used to shift all of the SWCNTs into a fresh bottom phase before application of a second, partitioning, ATPE step (back at room temperature). (D) Absorbance spectra of the semiconducting empty electric arc SWCNT synthesis SWCNTs separated as in (A) (red line) compared to the parent dispersion (black line). Absorbance features due to metallic SWCNT such as the peak features around ≈700 nm in the parent dispersion are near completely removed after the separation. The filled areas highlight the integrated regions of the absorbance spectra used for semiconducting purity assessment in literature methods. The ratio of A/(A + B) as measured over the shown wavenumber range (8110–15 575) cm⁻¹ is 0.426.

Fig. 5. (A) Schematic of additional workflow of ATPE separations for producing high purity metallic SWCNT populations starting at the 3M_B (i.e. M_BM_BM_B) phase of Fig. 4. After separation of semiconducting SWCNTs to the top phase and their removal via pipetting, DOC is added to a concentration of 0.05% in the bottom phase and mixed. After a waiting period of 30 min to 1 h, fresh top phase mimic containing DOC and SDS at a concentration just sufficient to partition all (n,m) species to the top phase is added, mixed and ATPE performed. The last step may need to be repeated on the resultant bottom fraction to collect all of the desirable SWCNTs. (B) Scaled absorbance spectra of metallic-enriched empty arc SWCNTs separated in the manner of workflow in (A) compared to the parent dispersion. The workflow of (A) results in a purer metallic population than achieved without use of DOC; a repetition of the separation further increases metallic purity. Also shown is the effect of re-semiconducting–metallic sorting the DOC-transferred top fraction, followed by a repetition of the workflow. This results in an even greater metallic purity with little recovery loss of the metallic population. Spectra are scaled to an absorbance of A = 5 arb. units at the ≈240 nm peak for all samples.

Fig. 6. (Top) Schematic of band gap fractionation of a SWCNT sample by sequential ATPE separations at successively reduced oxidant concentrations. (Bottom) Photographs of separated fractions isolated from a small diameter SWCNT population. Adapted with permission from Gui et al., Nano Lett., 2015, 15, 1642–1646. Copyright 2015 American Chemical Society.

Fig. 7. Simplified schematic of the hypothesized mechanism driving the selection of ATPE partition for each individual (n,m) species in the SDS–DOC competition ATPE system for a constant DOC concentration. (A) At low SDS concentration all SWCNT (n,m)s are covered primarily or entirely by DOC molecules, which would result in partition to the bottom phase (red arrows) in our PEG–DEX system. (B) and (C) At increasing SDS concentrations, SDS begins to comingle with, or outcompete, the DOC on the interface of some SWCNT (n,m)s; at sufficient SDS fraction these SWCNTs partition to the top phase (green arrows). The SDS concentration necessary for this for each (n,m) corresponds to the empirical critical point for that (n,m) shown in Fig. 2D. (D) At high SDS concentrations, (almost) all SWCNTs are coated with SDS, or a sufficient fraction of SDS to cause partitioning into the top phase. Note that the number of shown surfactant molecules in each box is not quantitative. However, from left to right the number of unbound DOC molecules remains constant while the number of SDS molecules increases dramatically to reflect equilibrium with the bulk volume beyond the schematic frame. It is possible that the amount and type(s) of polymer(s) affects the competition of the surfactants.

Fig. 8. An example of fractions resulting from the gradient process ATPE separation of empty metallic PT SWCNTs adapted from Fagan et al. (A) Photographs of the SDS gradient ATPE separated diameter fractions in order from lowest SDS for extraction (left) to greatest SDS for extraction (right). (B) Absorbance spectra of the fractions SWCNT fractions in order of their appearance in the photograph in panel (A) (left most = penultimate top), the topmost spectrum is of the metallic parent dispersion shown in (A). The changes in the SWCNT species distribution from systematic ATPE separation are demonstrated by the changes in absorbance peak wavelengths, with specifically enriched armchair (n,m) species labelled. Adapted from Fagan et al.,ACS Nano, 2015, 9, 5377–5390.

Fig. 9. Absorbance spectra of sequential chromatographically-separated SWCNT populations using a CCC instrument. The emergence and disappearance of absorbance features is due to the differential partitioning of the different SWCNT species along an SDS % gradient starting from a low concentration at the bottom of the figure to high concentration near the top. Reprinted with permission from Zhang et al., Anal. Chem., 2014, 86, 3980–3984. Copyright 2014 American Chemical Society.

Fig. 10. Schematic of the workflow to concentrate a dilute top phase fraction using a temperature swing. Cooling to <10 °C, and the addition of a small amount of bottom phase, results in partitioning of all SWCNTs to the bottom phase. Returning to room temperature, and adding a small volume of top phase, re-enables partitioning into the top phase. Large volumetric concentration can be achieved with minimal SWCNT mass loss via this method.