WO2008142375A2

WO2008142375A2 - Method of carbon nanotube selection

Info

Publication number: WO2008142375A2
Application number: PCT/GB2008/001676
Authority: WO
Inventors: Robin John Nicholas; Adrian Nish
Original assignee: Oxford University Innovation Ltd
Current assignee: Oxford University Innovation Ltd
Priority date: 2007-05-18
Filing date: 2008-05-15
Publication date: 2008-11-27
Anticipated expiration: 2009-11-18
Also published as: GB0709593D0; WO2008142375A3

Abstract

A method of purifying a sample of single walled carbon nanotubes to isolate nanotubes having particular properties or groups of properties, comprising dispersing the sample of nanotubes in a polymer solution containing a polymer of formula (I): wherein A, A', A', X, m and n are as herein defined, such that at least a portion of the nanotubes forms polymer-nanotube complexes with the polymer, and separating the polymer-nanotube complexes from the solution.

Description

METHOD OF CARBON NANOTUBE SELECTION

The present invention relates to a method of selecting carbon nanotubes, for example to purify synthesized carbon nanotubes, to separate carbon nanotubes from a mixture, or to select particular species from a mixture of carbon nanotubes.

Single-walled carbon nanotubes (SWNTs) can be grown or synthesized by the action of metal catalyst on a carbon vapour at high temperature. Two particular known growth processes are known as HiPCO and CoMoCAT. The multitude of possible ways of wrapping of a graphene sheet into a tube leads to many distinct possible SWNTs structures, each characterized by its diameter and the angle of its graphene lattice to the nanotube axis. All known growth methods produce ensemble samples of SWNTs with a distribution of chiral indexes n and m centred on a mean diameter. The physical properties of SWNTs are strongly affected by their structure, defined by the chiral indexes. For example metallic nanotubes are generated when the graphene wrapping condition n—m=3p+q corresponds to g=0, and semiconducting nanotubes are found for q=±\ . There exists the problem of finding purification methods which can separate a given SWNT ensemble into its distinct species. Another problem is that nanotubes typically form in aggregates known as bundles. The individual nanotubes from these bundles need to be dispersed so that any selection or purification process can be performed. A further problem is that the product of carbon nanotube synthesis typically contains a significant amount of contaminants such as soot, graphite and metallic impurities.

Some known techniques for solubilising and isolating SWNTs include using surfactants, such as sodium dodecyl sulfate or sodium dodecylbenzene sulfonate (SDBS), to disperse SWNTs. Coupled with strong sonication and ultracentrifugation these methods can yield solutions with high proportions of isolated individual tubes. Further developments include using multiple, density-gradient ultracentrifugation to introduce some tube selectivity, but these methods require significant experimental complexity and constitute a very tedious and elaborate technique. Another route which has been used to disperse SWNTs is by using DNA.

This differs from the surfactant method in that DNA is a long helical molecule and is thought to enclose the nanotube by wrapping its helix around the cylindrical structure of the tube. In contrast, surfactants are thought to coat all the material, including impurities, in a given raw nanotube sample. The use of DNA has enabled the diameter distribution of a SWNT ensemble to be narrowed, and DNA has also indicated preferential interaction with certain nanotube species. However, this technique still has significant limitations and disadvantages.

Accordingly, the present invention provides a method of purifying a sample comprising single-walled carbon nanotubes which comprises:

(a) dispersing the sample in a polymer solution in order to allow at least a portion of the nanotubes to associate with the polymer and form polymer-nanotube complexes, wherein the polymer is of formula (I):

-[A-A" -{X -A')_m[ - (I)

wherein: - A is selected from phenyl, biphenyl, triphenyl, naphthyl, anthracenyl, phenanthryl and fluorenyl;

A " is a bond or a phenyl group;

X is a bond, a vinylene group or a group -NR-, wherein R is phenyl, naphthyl or a 5- to 10-membered heteroaryl group; - A' is phenyl, naphthyl or a 5- to 10-membered heteroaryl group and is unfused or fused to a further phenyl or 5- to 10-membered heteroaryl group; m is zero, 1 or 2; n is an integer of from 5 to 10,000; and - A, R and the aryl and heteroaryl moieties in A^' and A" are unsubstituted or substituted by 1 , 2 or 3 substituents which are the same or different and are selected from halogen atoms and C)_-I0 alkyl,

C_2-)_O alkenyl, Ci_-I0 alkoxy, C_2-I0 alkenyloxy, Ci_-I0 haloalkyl, C_2-I0 haloalkenyl, Ci_-I0 haloalkoxy, C_2-I0 haloalkenyloxy, hydroxyl, cyano, nitro, Ci-io hydroxyalkyl or -NR 'R" where R' and R" are the same or different and represent hydrogen or Ci_-4 alkyl; and (b) separating the polymer-nanotube complexes from the polymer solution.

In one embodiment the method allows separation of single-walled carbon nanotubes having a specific chiral angle or range of chiral angles from a sample comprising single-walled carbon nanotubes. In another embodiment the method allows separation of single-walled carbon nanotubes having a specific diameter or range of diameters from a sample comprising single- walled carbon nanotubes. More preferably the method is used to select single-walled carbon nanotubes having both a specific chiral angle or range of chiral angles and a specific diameter or range of diameters.

The invention also provides the use of a polymer of formula (I) as defined above for purifying single-walled carbon nanotubes. In one embodiment the invention provides the use of a polymer as described above for selecting carbon nanotubes having a specific chiral angle or range of chiral angles. In another embodiment the invention provides the use of a polymer described above for selecting carbon nanotubes having a specific diameter or range of diameters. In a preferred embodiment the invention provides the use of a polymer as described above for selecting carbon nanotubes having both a specific chiral angle or range of chiral angles and a specific diameter or range of diameters. The invention further provides carbon nanotubes obtainable by the method described above.

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings in which:

Fig. 1 shows PLE maps of SWNT solutions produced according to the method of an embodiment of the invention and according to the method of a comparative example;

Fig. 2 shows absorbance spectra for SWNTs grown by two different processes and then treated according to a method embodying the present invention and a comparative method; and

Fig. 3 shows maps of PL intensities for comparing the SWNTs selectivity of a method embodying the present invention and two comparative methods.

Fig. 4 shows PLE maps of SWNT solutions produced using four different polymers, wherein the nanotubes were grown by the HiPCO process. Fig. 5 shows a map of PL intensities to demonstrate the selectivity of a particular embodiment of the invention and, inset, a histogram illustrating the distribution of diameter of SWNT species solublilised in an embodiment of the invention and a comparative example. Fig. 6 and Fig. 7 show maps of PL intensities illustrating the SWNTs which are solubilised in three embodiments of the invention.

DETAILED DESCRIPTION OF THE INVENTION

As used herein, a C_MO alkyl group or moiety is a linear or branched alkyl group or moiety containing from 1 to 10 carbon atoms, for example a C_M alkyl group or moiety containing from 1 to 4 carbon atoms. Examples Of C]_-4 alkyl groups include methyl, ethyl, n-propyl, i-propyl, n-butyl, i-butyl and t-butyl. Examples of higher alkyl groups include all isomers of pentyl, hexyl, heptyl, octyl, nonyl and decyl. Preferred such alkyl groups include hexyl and octyl groups, for example straight-chain -C₆Hi₃ and -CgHi₇. For the avoidance of doubt, where two alkyl moieties are present in a group, the alkyl moieties may be the same or different.

As used herein, a C_2-I₀ alkenyl group or moiety is a linear or branched alkenyl group or moiety containing from 2 to 10 carbon atoms, for example a C_2-6 or C_2-4 alkenyl group or moiety containing from 2 to 6 or 2 to 4 carbon atoms respectively, such as -CH=CH₂ or -CH₂-CH=CH₂. For the avoidance of doubt, where two alkenyl moieties are present in a group, they may be the same or different.

As used herein, a halogen is typically chlorine, fluorine, bromine or iodine and is preferably chlorine, bromine or fluorine.

As used herein, a C_MO alkoxy group or C_2-I0 alkenyloxy group is typically a said C_MO alkyl group or a C_2-I0 alkenyl group respectively which is attached to an oxygen atom.

A haloalkyl, haloalkenyl, haloalkoxy or haloalkenyloxy group is typically a said alkyl, alkenyl, alkoxy or alkenyloxy group respectively which is substituted by one or more said halogen atoms. Typically, it is substituted by 1, 2 or 3 said halogen atoms. For example, preferred haloalkoxy groups include methoxy groups substituted by 1 or 2 said halogen atoms, such as -OCH₂F and -OCHF₂. Other preferred haloalkyl and haloalkoxy groups include perhaloalkyl and perhaloalkoxy groups such as -CX₃ and -OCX3 wherein X is a said halogen atom, for example chlorine and fluorine. Perhaloalkyl groups include groups such as -CF₃ and -CCl₃. As used herein, a C_M 0 hydroxyalkyl group is a C_MO alkyl group substituted by one or more hydroxy groups. Typically, it is substituted by one, two or three hydroxy groups. Preferably, it is substituted by a single hydroxy group.

When a phenyl or naphthyl group or moiety is fused to a further phenyl or 5- to 10-membered heteroaryl group or moiety it is preferably a phenyl group fused to a 5- to 10-membered heteroaryl group or a naphthyl group fused to a further phenyl group. Suitable examples of a phenyl group fused to a 5- to 10-membered heteroaryl group include a phenyl group fused to a 5- to 6-membered heteroaryl group, more preferably a phenyl group fused to an oxazolyl, oxadiazolyl, thiazolyl or thiadiazolyl group, most preferably a phenyl group fused to a thiadiazolyl group (i.e. a benzothiadiazolyl group).

As used herein, a 5- to 10- membered heteroaryl group or moiety is a monocyclic 5- to 10- membered aromatic ring, such as a 5- or 6- membered ring, containing at least one heteroatom, for example 1 , 2, 3 or 4 heteroatoms, selected from O, S and N. When the ring contains 4 heteroatoms these are preferably all nitrogen atoms. Examples include pyridyl, pyrazinyl, pyrimidinyl, pyridazinyl, furanyl, thienyl, pyrazolidinyl, pyrrolyl, oxadiazolyl, oxazolyl, isoxazolyl, thiazolyl, thiadiazolyl, imidazolyl, pyrazolyl, triazolyl and tetrazolyl groups. More preferred heteroaryl groups or moieties include oxazolyl, oxadiazolyl, thiazolyl and thiadiazolyl, more preferably thiadiazolyl.

As used herein, a 5- to 10- membered heteroaryl group or moiety fused to a further phenyl or 5- to 10- membered heteroaryl group or moiety is preferably a 5- to 10-membered heteroaryl group or moiety fused to a further phenyl group or moiety. In this case, preferably the 5- to 10-membered heteroaryl group or moiety is an oxazolyl, oxadiazolyl, thiazolyl or thiadiazolyl group or moiety fused to a further phenyl.

The aryl and heteroaryl groups are unsubstiruted or substituted by 1 , 2 or 3 substituents, which are the same or different when 2 or 3 are present. Suitable substituents include halogen atoms and C_MO alkyl, C_2-I₀ alkenyl, C_1-I0 alkoxy, C_2-I0 alkenyloxy, Ci_-10 haloalkyl, C_2-10 haloalkenyl, Ci_-I0 haloalkoxy, C_2-io haloalkenyloxy, hydroxyl, cyano, nitro, Ci-I₀ hydroxyalkyl or -NR'R" where R' and R" are the same or different and represent hydrogen or Ci_-4 alkyl. Preferred substituents include Ci_-I0 alkyl and Ci-I₀ alkoxy. Particularly preferred Ci_-I0 alkyl substituents include methyl, octyl (all isomers thereof), methoxy and 2-ethylhexyloxy. The polymer for use in the method of the invention is of general formula (I) as defined above. Preferably A is selected from phenyl, naphthyl, anthracenyl, phenanthryl and fluorenyl. More preferably A is selected from phenyl or fluorenyl. Most preferably A is fluorenyl.

Preferably A is unsubstituted or substituted by 1 or 2 substituents. More preferably A is unsubstituted or substituted by 2 substituents. When A is substituted it is most preferably substituted by two groups which are the same or different and are Ci_-I0 alkyl or Ci_-I0 alkoxy groups, preferably Ci_-I0 alkyl groups, more preferably hexyl or octyl groups, in particular octyl groups. When A is fluorenyl, most preferably it is substituted by two Ci_-I0 alkyl groups, and these groups are preferably both present at the 9 position. More preferably these alkyl groups are C_4-I0 alkyl, more preferably C_6-I0 alkyl, most preferably C₆ or C₈, especially C₈ groups. m is zero, 1 or 2. When m is zero, the group (X-A') is absent, and the polymer is effectively a homopolymer of A groups. When m is one or two, the group (X-A') is present, and the polymer is effectively a copolymer of A groups and (X-A') groups.

X represents a bond, a vinylene group (-CH=CH-) or a group -NR-. When X is a vinylene group, preferably A' is a C_6-H aryl group. When X is -NR-, R is preferably a phenyl group which is unsubstituted or substituted with one, two or three, for example one, substituent. The substituent(s) are typically selected from Ci-io alkyl groups and Ci_-I0 alkoxy groups, preferably Ci-io alkyl groups, more preferably Ci_-6 alkyl groups, for example butyl. When X is -NR-, A' is preferably phenyl, more preferably unsubstituted phenyl.

When X is a bond, A' can be a C_6-I4 aryl group or a 5- to 10-membered heteroaryl group which is unfused or fused to a further phenyl or 5- to 10-membered heteroaryl group. When X is a bond, preferably A' is a 5- to 10-membered heteroaryl group which is unfused or fused to a further phenyl or 5- to 10-membered heteroaryl group. More preferably X represents a bond. Preferably A^' is a phenyl or naphthyl group which is unfused or fused to a further phenyl or 5- to 10-membered heteroaryl group. More preferably A' is an unfused phenyl, a phenyl group which is fused to a 5- to 10-membered heteroaryl group, or a naphthyl group which is fused to phenyl group. More preferably A' is an unfused phenyl, a phenyl group which is fused to a 5- to 6-membered heteroaryl group, or a naphthyl group which is fused to phenyl group (i.e. an anthracenyl group). More preferably A' is an unfused phenyl, a benzothiadiazolyl or an anthracenyl group. More preferably still A^' is a benzothiadiazolyl group.

Preferably A' is unsubstituted or substituted by 1 or 2 substituents. More preferably A' is unsubstituted or substituted by 2 substituents. When A' is substituted it is most preferably substituted by two groups which are the same or different and are Ci_-I0 alkyl or Ci_-I0 alkoxy groups, preferably Ci_-I0 alkyl groups. Preferred substituents include Ci_-4 alkyl groups (such as methyl and ethyl, more preferably methyl), C_4-I0 alkyl groups (more preferably C_6-I0 alkyl groups such as C₆ and C₈ alkyl groups), Ci_-4 alkoxy groups (such as methoxy and ethoxy, more preferably methoxy) and C_4-I0 alkoxy groups (more preferably C_6-J₀ alkoxy groups such as C₈ alkoxy). Most preferred substituents include methyl, hexyl, octyl, methoxy and octyloxy and isomers thereof. For example, a suitable octyloxy substituent is 2-ethylhexyloxy. A" is either a phenyl group, which may be unsubstituted or substituted, or is a bond. Where A" is a substituted phenyl group, one two or three substituents are typically present, the substituents being the same or different and typically being selected from Ci_-I0 alkyl and Ci_-I0 alkoxy groups. Preferably, A^" either represents a bond or an unsubstituted phenyl group. The number of monomers defined in formula (I) which are present in the polymer is represented by integer n. This is generally an integer from 5 to 10,000, more preferably from 10 to 2,000, more preferably from 10 to 1,000, more preferably from 10 to 750. As will be apparent to a person skilled in the art, samples of polymers are generally polydisperse, so the value of n will vary within a range. The polymers defined by formula (I) can terminate in a number of different groups. These terminal groups are usually described as "end-capping" the polymer. A huge variety of terminal groups are suitable for use in the invention. In particular, as the integer n generally represents a large number and the main body of the polymer is hence large relative to the size of the terminal groups, these terminal groups generally have little bearing on the properties of the polymer as a whole or are chosen to ensure that they have little bearing on these properties. Suitable terminal groups include single aryl or heteroaryl groups, or small oligomers or dendrimers of such aryl and/or heteroaryl groups optionally linked via linking groups. For example, suitable terminal groups may comprise phenyl or 5- to 6- membered heteroaryl groups or oligomers of the same optionally linked via linking groups such as vinylene, acetylene or nitrogen-containing groups. Exemplary terminal groups include phenyl groups which are unsubstituted or substituted with 1, 2 or 3 small substituents such as halogen atoms or Ci_-4 alkyl groups. As an example there can be mentioned dimethylphenyl (DMP). Other terminal groups include oligomers comprising 2 to 6 phenyl and/or 5- to 6-membered heteroaryl groups. As an example there can be mentioned 2,5-diphenyl-l,2,4-oxadiazole. Further terminal groups include dendritic groups comprising 3 to 10 phenyl and/or 5- to 6-membered heteroaryl groups optionally linked or branched via linking or branching groups such as vinylene, acetylene or nitrogen. As an example there can be mentioned N,N- bis(4-methylphenyl)-4-aniline. Particularly preferred terminal groups in accordance with the invention are dimethylphenyl groups. Other suitable terminal groups will be apparent to the skilled person.

In accordance with the method of the invention, the first step is to disperse a sample comprising carbon nanotubes in a polymer solution. The sample may comprise a mixture of nanotubes (e.g. where the method is to separate a certain species of nanotube from a mixture of nanotubes), or may comprise nanotubes and other substances such as impurities (e.g. soot, graphite and metallic impurities). The polymer solution comprises the polymer of formula (I) and a suitable solvent. The mixture of polymer solution and sample may be agitated to assist in dispersing the nanotubes. In many instances, nanotubes are present as a bundle and some means of mixing or agitation helps to break up these bundles. Sonication is an appropriate means of agitation.

The nature of the solvent to be used may depend on the polymer of formula (I) and the nanotubes which are to be purified. A wide range of solvents can be used, although it is preferable to choose a solvent in which the polymer is readily soluble. Exemplary solvents include water and a wide range of common organic solvents such as toluene, chlorobenzene, tetrahydrofuran (THF), chloroform, dimethylformamide (DMF), alcohols (e.g. methanol, ethanol and higher alcohols), acetonitrile, dichloromethane, dichloroethane, dioxane, N,N-dimethylacetamide, methyl ethyl ketone, acetone, heptane, cyclohexane, and combinations thereof. Preferred solvents include water, toluene, tetrahydrofuran, dimethylformamide, chlorobenzene, chloroform and alcohol. More preferred solvents include toluene and tetrahydrofuran, most preferably toluene. Without wishing to be bound by theory, it is believed that once the nano tubes have been dispersed in the solution, the polymer of formula (I) selectively wraps around nanotubes having a certain chiral angle or a certain range of chiral angles. Other substances (e.g. impurities such as soot or graphite) are not wrapped by the polymer. Furthermore other nanotubes, having different chiral angles, are not wrapped by the polymer. This method therefore selects nanotubes of a specific chiral angle or range of chiral angles. As described herein, the nanotubes wrapped with the polymer of formula (I) are called polymer-nanotube complexes.

In a second step of the method the polymer-nanotube complexes are separated from the rest of the solution. The rest of the solution will contain the solvent, as well as any polymer which has not become associated into a polymer- nanotube complex. The solution will also contain nanotubes which have not been wrapped by the polymer, as well as other impurities (e.g. soot and graphite). The polymer-nanotube complexes can be separated by any suitable method which will be known to the skilled person. For example, this separation can conveniently be achieved by centrifugation.

In some embodiments it may be sufficient simply to isolate the polymer- nanotube complexes from the rest of the solution, without then dissociating the complex into the constituent nanotubes and polymers. However, in an alternative embodiment the method further comprises the step of isolating the nanotubes from the polymer-nanotube complexes by treating the product of step (b) to dissociate the nanotubes from the polymer. This can be achieved by a variety of methods which will be familiar to the skilled person in the art. For example, addition of an alternative solvent to the polymer-nanotube complex, addition of a surfactant or by dielectrophoresis, heating, burning off or exposure to light.

In a first preferred embodiment of the invention m is zero, A" is a bond and A is fluorenyl. In this embodiment preferred polymers are poly(9,9-dioctylfluorenyl- 2,7-diyl) and poly(9,9-dihexylfluorenyl-2,7-diyl. Poly(9,9-dioctylfluorenyl-2,7-diyl) is particularly preferred. In this embodiment, the most preferred solvent is toluene.

In a second preferred embodiment of the invention m is one, X represents a bond, A is fluorenyl, A' is benzothiadiazolyl and A" is a bond. In this embodiment a most preferred polymer is poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l,4-benzo- {2,1 ',3}-thiadiazole)]. In the second preferred embodiment, the preferred solvent is toluene.

In a third preferred embodiment, m is one, X represents a bond, A is fluorenyl, A' is phenyl and A" is a bond. In this embodiment a most preferred polymer is poly(9,9-dioctylfluorenyl-2,7-diyl)-co-(l,4-phenylene). In this embodiment, the most preferred solvent is THF.

In a fourth preferred embodiment, m is two, X represents -NR-, wherein R is phenyl, A is fluorenyl, A' is phenyl and A" is phenyl. In this embodiment a most preferred polymer is poly(9,9-dioctylfluorenyl-2,7-diyl)-co-(N,N'-diphenyl)- N,N'di(p-butyl-oxy-phenyl)-l,4-diaminobenzene). In this embodiment, the most preferred solvent is THF. The first and second embodiments are particularly preferred, with the most preferred polymers being poly(9,9-dioctylfluorenyl-2,7-diyl) and poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l ,4-benzo- {2, 1 ',3 } -thiadiazole)].

These embodiments will be described in more detail in the Examples section below.

Examples

Example 1

Poly(9,9-dioctylfluorenyl-2,7-diyl), a common semiconducting organic polymer, referred to as PFO, was purchased from American Dye Source Inc. Samples of single- walled carbon nanotubes grown by both the 'HiPCO' and 'CoMoCAT' processes were purchased from Carbon Nanotech. Inc. and Southwest Nanotech. Inc. respectively. All other chemicals used were purchased from Sigma Aldrich. The nanotube powders were dispersed in PFO/toluene solutions in the ratio 5 mg SWNT: 6 mg polymer: 10 ml solvent. The solutions were then homogenised in a sonic bath for 60 minutes followed by vigorous sonication using an ultrasonic disintegrator for 15 mins. This was then promptly followed by centrifugation at 9000 g for 3 minutes. Also used as a comparative example in place of PFO was another common light emitting polymer, poly[2-methoxy-5-(2-ethylhexyloxy)-l,4- phenylene-vinylene], MEHPPV. Samples were prepared using this polymer with the same procedure as for PFO. The preparation procedure for dispersions in sodium dodecylbenzene sulfonate (SDBS), used here along with MEHPPV for comparison with PFO, consisted of 5 mg SWNT: 350 mg surfactant: 35 ml D₂O. Sonication was done for 30 minutes using an ultrasonic disintegrator and centrifugation was using an ultracentrifuge at 85,000 g for 4 hours. It is worth noting that a much lower centrifugal force and time duration was found to be sufficient to remove the non- dispersed nanorubes in our polymer samples compared to the aqueous surfactants. Absorbance measurements were taken using a Perkin-Elmer Lambda 9 Photospectrometer. Photoluminescence Excitation (PLE) mapping was done using an automated custom built system consisting of a 75 Watt Xenon lamp focused into a monochromator which then illuminated the sample in a quartz fluorescence cell. Normalisation for the lamp's spectral response was done using a silicon photodiode. Luminescence from the sample was collected at 90 degrees to the excitation beam and focused into a spectrograph fitted with a liquid nitrogen cooled InGaAs photodiode array.

Fig. 1 shows photoluminescence-excitation (PLE) maps of SWNTs produced by both the HiPCO and CoMoCAT processes dispersed in PFO using the solvent toluene. The peaks correspond to the resonantly enhanced emission from the SWNTs' primary En electronic transitions when the excitation matches their secondary E₂₂ electronic levels. Comparison with solutions dispersed with conventional aqueous surfactant SDBS on the same starting material are also shown and demonstrate that there has been a dramatic reduction in the number of peaks and hence tube species. The points indicate theoretical positions for the energy gaps of the corresponding n,m indexed tube species. For the PFO-SWNT maps these points have been redshifted by 1.3 % from those given in previously published work for the case of SWNTs dispersed using aqueous surfactant. This shift is attributed to a difference in the local dielectric environment where the increased dielectric screening causes a reduction of the electron-electron and excitonic Coulomb interactions which produce a net reduction in the energy gap. Six peaks corresponding to fundamental nanotube resonances are visible in the HiPCO map and five may be seen in the CoMoCAT map for the samples processed according to this embodiment of the invention. Other resonant features which are visible at higher and lower energy excitation relative to E₂₂ also result in emission from the same En energy gaps and can be attributed to previously established phonon and weakly allowed absorption effects.

Fig. 2 shows the absorbance spectra of HiPCO- and CoMoCAT-produced SWNTs dispersed in aqueous SDBS and in PFO/toluene solutions. Three significant differences are observed between the two solutions. Firstly the total absorption is substantially smaller in the case of the polymer/toluene solutions. Secondly the absorption features are significantly narrowed and much more clearly resolved than in the case of the aqueous solution, and thirdly the underlying background is very significantly reduced. The difference is due to a very substantial reduction in the number of different nanotube species present in the polymer/toluene solutions, combined with an enhanced purification of the solutions. As a result the peaks observed in absorbance can easily be resolved and assigned to those seen in the PLE maps, in contrast to the absorption spectra from the aqueous solutions where the presence of multiple species causes convolution of the absorption peaks making detailed analysis very difficult. The increased resolution now means that it is possible to deduce the relative molecular fractions of the different tubes directly from both optical absorbance and peak intensity in the PLE maps as shown in Table I. In practice, the relative values deduced by the two methods may differ if the individual n,m species have different PL efficiencies. In particular, it has been suggested that the q=+\ nanotubes have a significantly lower PL yield than the q=—\ tube types. Further chiral angle and diameter dependencies are also thought to exist. Limited evidence for this may be seen by comparing the relative molecular fractions deduced for the (7,6), (8,7) and (9,8) tubes from absorption and PL. The apparent proportion of these tubes is significantly smaller when deduced from the PL intensities for either the HiPCO or CoMoCAT samples, consistent with these tubes having a much lower quantum efficiency than the other species present. A further notable observation from the absorption and PLE maps is that the (6,5) SWNT species which is particularly prominent for the CoMoCAT sample has been almost totally removed from the sample processed in PFO/toluene. This can be seen from the absence of its primary transition peak at around 1000 nm and also of its secondary peak at just below 600 ran.

TABLE I

The spectra shown in Fig. 2 are all raw spectra on which no background subtraction has been made. However, the almost total absence of the background is particularly striking for the CoMoCAT in PFO sample, though some background is still present in the HiPCO sample. The origin of this background is not certain. It is thought to come from the π-plasmon which may originate in residual sample impurities, such as C₆₀ type fullerenes or other amorphous and graphitic carbon. It is also possible that some of this background could come from the presence of metallic SWNTs. The absorption data of Fig. 2 show clear peaks at 505 and 455 nm, corresponding to the metallic branches beginning with the (7,7) and (6,6) tubes, for the aqueous CoMoCAT solution, but no detectable trace of these peaks remains for the polymer/toluene sample. Removal of the metallic SWNTs from the CoMoCAT material is also evidence in Raman spectra which show no detectable signal from the metallic species which were present in the aqueous solutions. , In addition to the improvements in resolution and purity, the polymer/toluene solutions also show a significant increase in quantum efficiency. The luminescence 5 quantum yield has been found by comparing the measured luminescence intensity and absorbance with that of a solution of known quantum yield. Rhodamine 6G, a common laser dye with a quantum efficiency of 95%, was used as a standard solution. For the SDBS suspended CoMoCAT nanotubes a quantum efficiency of 0.1 % was measured. However the quantum efficiency for the same CoMoCAT

10 nanotubes suspended in PFO was found to increase to ~1.5%. The main cause of this difference is thought to be the absence of the large non-resonant background which absorbs a significant fraction of the excitation light but does not contribute to the photoluminescence. This suggests that previous reports of low quantum yield in nanotubes are in fact primarily a result of the parasitic absorption and not an intrinsic

15 property of isolated nanotubes. Measurements on isolated single nanotubes have found quantum efficiencies of 7%.

The ability of PFO to selectively wrap a small number of SWNT species is best illustrated using schematic representations called graphene sheet maps. These show hexagons corresponding to the carbon nanotube species that give measurable

20 PL shaded in a greyscale from white to black, proportional to their relative intensity. Examples of this are shown in Fig. 3 for the HiPCO starting material, which has the wider initial spread (at least 22 observed) of tube species. Maps are shown for both the SDBS and PFO suspensions, as well as a third suspension using the light emitting polymer MEHPPV in toluene. The MEHPPV polymer introduces some diameter

25 selectivity, with a significant reduction in the number of visible species to -14.

Much more dramatic selectivity is demonstrated, however, by the PFO suspension, where only 6 species can be detected, with a very strong chiral bias since all of the observed species are very close to the armchair configuration. Absorbance spectra can give a more direct measure of the relative species populations, however previous

30 measurements have been hampered due to difficulties from the background and in de-convoluting peaks with matching energy gaps. The results presented in Table I show however that for the case of the PFO wrapped tubes, where an accurate measurement of the absorbance is possible, the relative fractions of tubes present are similar when deduced by the two spectroscopic techniques. Both methods show a similar chiral selectivity for the CoMoCAT PFO/toluene solutions, and in particular that around 60% of all the suspended tubes are from the (7,5) species. From the results it is clear that PFO has a very strong preference for SWNTs which have a large chiral angle. In other words the polymer only wraps and suspends individual nanotubes in a very narrow range of species. The total removal of the formerly dominant (6,5) species in the CoMoCAT sample also shows that there is not only a chiral angle selective process but also a very high degree of diameter selection. Such strong chirality selection is unprecedented and indicates that the PFO polymer has a very intimate interaction with the nanotube surface. This is unlike previous assumptions that polymers wrap the surface of the tube with some diameter preference but are chirality insensitive. In addition, the disappearance of absorption and Raman peaks associated with metallic nanotubes also suggests that the wrapping process is inhibited for metallic species, due possibly to charge transfer changing the conformation of the polymer. This effect provides a route to the bulk separation of metallic from semiconducting nanotube species.

Example 2 A further embodiment of the invention involves the use of a the polymer

Poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l ,4-benzo-{2,l ',3}-thiadiazole)] (PFO-BT), available from American Dye Source, Inc. under the tradename ADS 133YE. This polymer is used in place of PFO in the above Example 1, but all other features of the method are the same so the description will not be repeated. Fig. 4D shows photoluminescence-excitation (PLE) maps of SWNTs^' produced by the HiPCO process dispersed in PFO-BT using the solvent toluene. The number of peaks which appear in the PLE map is significantly lower than the number of peaks in the corresponding map for the aqueous dispersion depicted in Figure 1 , demonstrating the selectivity of PFO-BT for a small number of SWNT species. A graphene sheet map showing the selectivity of PFO-BT to HiPCO- produced SWNTs in toluene is shown in Figure 5. PFO-BT can be seen to most strongly select the (10,5) SWNT species, with a total of only nine species ((8,6), (8,7), (9,4), (9,7), (10,5), (10,6), (11,4), (12,2), (13,2)) being detected. The inset histogram of Figure 5 compares the normalised PL intensity for SDBS (white bars) and PFO-BT (black bars) solutions using HiPCO produced SWNTs. A strong selectivity for nanotubes with a diameter around 1.05nm can be seen for the PFO-BT experiment.

Example 3

The methods set out in Example 1 were repeated using the polymer Poly (9,9-dihexylfluorenyl-2,7-diyl) (PFH), available from American Dye Source, Inc.. Fig. 4B shows photoluminescence-excitation (PLE) maps of SWNTs produced by the HiPCO process dispersed in PFH using the solvent toluene. The graphene sheet map of Figure 6 clearly depicts the selected HiPCO-produced SWNTs using toluene as the solvent. PFH can be seen to most strongly select the (8,7) and (9,7) SWNT species, with a total of only ten species ((7,5), (7,6), (8,6), (8,7), (9,4), (9,5), (9,7), (10,3), (11,3), (12,1)) being detected.

Example 4

The methods set out in Example 1 were repeated using the polymer Poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l,4-phenylene)] (PFO-P), available from American Dye Source, Inc.. The methods were repeated a second time with PFO-P, but replacing the toluene solvent with THF. Fig. 4C shows photoluminescence- excitation (PLE) maps of SWNTs produced by the HiPCO process dispersed in PFH using the solvent toluene. The graphene sheet map of Figure 7 depicts the selected HiPCO-produced SWNTs using THF as the solvent. PFO-P can be seen to most strongly select the (9,4) SWNT species in THF.

Example 5

The methods set out in Example 1 were repeated using the polymer Poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(N,N'-diphenyl)-N,N'di(p-butyl-oxy- phenyl)- 1 ,4-diaminobenzene)] (PFO-PBAB), available from American Dye Source, Inc. and using THF as the solvent. The graphene sheet map of Figure 7 depicts the selected HiPCO-produced SWNTs. PFO-PBAB can be seen to most strongly select the (7,6) SWNT species in THF.

Claims

1. A method of separating single-walled carbon nanotubes having a specific chiral angle or range of chiral angles from a sample comprising single- walled carbon nanotubes which comprises: (a) dispersing the sample in a polymer solution in order to allow at least a portion of the nanotubes to associate with the polymer and form polymer-nanotube complexes, wherein the polymer is of formula (I):

wherein:

A is selected from phenyl, biphenyl, triphenyl, naphthyl, anthracenyl, phenanthryl and fluorenyl;

A " is a bond or a phenyl group; - X is a bond, a vinyl ene group or a group -NR-, wherein R is phenyl, naphthyl or a 5- to 10-membered heteroaryl group; A' is phenyl, naphthyl or a 5- to 10-membered heteroaryl group and is unfused or fused to a further phenyl or 5- to 10-membered heteroaryl group; - m is zero, 1 or 2; n is an integer of from 5 to 10,000; and A, R and the aryl and heteroaryl moieties in A' and A" are unsubstituted or substituted by 1 , 2 or 3 substituents which are the same or different and are selected from halogen atoms and CMO alkyl, C_2-]_O alkenyl, C_1-Io alkoxy, C_2-I0 alkenyloxy, CMO haloalkyl, C_2-!o haloalkenyl, CMO haloalkoxy, C_2-I0 haloalkenyloxy, hydroxyl, cyano, nitro, CM_O hydroxyalkyl or -NR R" where R' and R" are the same or different and represent hydrogen or Ci ^ alkyl; and (b) separating the polymer-nanotube complexes from the polymer solution.

2. A method of separating single-walled carbon nano tubes having a specific diameter or range of diameters from a sample comprising single-walled carbon nanotubes which comprises:

(a) dispersing the sample in a polymer solution in order to allow at least a portion of the nanotubes to associate with the polymer and form polymer-nanotube complexes, wherein the polymer is of formula (I) as defined in claim 1 ; and

(b) separating the polymer-nanotube complexes from the polymer solution.

3. A method as claimed in claim 1 or claim 2 wherein A, R and the aryl and heteroaryl moieties in A' and A" are unsubstituted or substituted by 1 or 2 substituents which are the same or different and are selected from C_1-I0 alkyl and Ci- io alkoxy.

4. A method as claimed in any one of the preceding claims wherein A is fluorenyl.

5. A method as claimed in claim 4 wherein A is unsubstituted or substituted by 1 or 2 groups which are the same or different and are selected from C_MO alkyl and

Ci_-Io alkoxy, more preferably C_MO alkyl.

6. A method as claimed in claim 5 wherein A is substituted by 1 or 2 groups selected from C_6-io alkyl.

7. A method as claimed in any one of the preceding claims wherein m is zero.

8. A method as claimed in any one of the preceding claims wherein the polymer is poly(9,9-dioctylfluorenyl-2,7-diyl), poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l ,4- benzo-{2,l ',3}-thiadiazole)], poly(9,9-dihexylfluorenyl-2,7-diyl), poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(l ,4-phenylene)] or poly[(9,9- dioctylfluorenyl-2,7-diyl)-co-(N,N'-diphenyl)-N,N'di(p-butyl-oxy-phenyl)-l,4- diaminobenzene)] .

9. A method as claimed in any one of the preceding claims wherein the polymer is poly(9,9-dioctylfluorenyl-2,7-diyl).

10. A method as claimed in any one of the preceding claims wherein the solvent is selected from water, toluene, tetrahydrofuran, dimethylformamide, chlorobenzene, chloroform and alcohol.

11. A method as claimed in claim 10 wherein the solvent is toluene.

12. A method as claimed in any one of the preceding claims which further comprises the step of isolating the nanotubes from the polymer-nanotube complexes by treating the product of step (b) to dissociate the nanotubes from the polymer.

13. Use of a polymer of formula (I) as defined in any one of claims 1 to 9 for selecting carbon nanotubes having a specific chiral angle or range of chiral angles.

14. Use of a polymer of formula (I) as defined in any one of claims 1 to 9 for selecting carbon nanotubes having a specific diameter or range of diameters.

15. Carbon nanotubes obtainable by the method as claimed in any one of claims 1 to 12.