Degradable conjugated polymers

Abstract

A polymer comprising at least one unit of the formula (1) wherein T.sup.1 is a carbon atom or a nitrogen atom, T.sup.2 is a carbon atom if T.sup.1 is a nitrogen atom, or is a nitrogen atom if T.sup.1 is a carbon atom, r is 1, 2, 3 or 4, s is 1, 2, 3, or 4, M.sup.1 is preferably selected from the group consisting of M.sup.2 is preferably The polymers are prepared by reacting monomers (1a) with monomers (2a) H.sub.2N-[-M.sup.1-]r-NH.sub.2 (1a) OHC-[-M.sup.2-]s-CHO (2a) or the step of reacting monomers (1b) with monomers (2b) OHC-[-M.sup.1-]r-CHO (1b) H.sub.2N-[-M.sup.2-]s-NH.sub.2 (2b). ##STR00001##

Claims

1. A process for separating semiconducting single-walled carbon nanotubes from a mixture of semiconducting and metallic single-walled carbon nanotubes, the process comprising: (i) providing a mixture A of semiconducting and metallic single-walled carbon nanotubes, (ii) dispersing the mixture A in a solvent using a degradable polymer comprising at least one unit of formula (1): ##STR00057## wherein T.sup.1 is a carbon atom or a nitrogen atom, T.sup.2 is a carbon atom if T.sup.1 is a nitrogen atom, or is a nitrogen atom if T.sup.1 is a carbon atom, r is 1, 2, 3 or 4, s is 1, 2, 3, or 4, wherein M.sup.1 is selected from the group consisting of ##STR00058## and M.sup.2 is selected from the group consisting of ##STR00059## ##STR00060## wherein R.sup.1 is C.sub.1-36-alkyl, R.sup.102 and R.sup.103 are independently selected from the group consisting of H and C.sub.1-20-alkyl, which is optionally prepared through a direct condensation of diamines comprising r units of M.sup.1 and dialdehydes comprising s units of M.sup.2 or of dialdehydes comprising r units of M.sup.1 and diamines comprising s units of M.sup.2, as a dispersing agent to obtain a dispersion B comprising the semiconducting single-walled carbon nanotubes in a solvent, the dispersion B further comprising metallic single-walled carbon nanotubes, (iii) separating the metallic single-walled carbon nanotubes from the dispersion B, to obtain an enriched dispersion C of the semiconducting single-walled carbon nanotubes in a solvent, (iv) degrading by hydrolysis the degradable polymer in the enriched dispersion C, and (v) separating the semiconducting single-walled carbon nanotubes from a solution D comprising a solvent and the degradable polymer.

2. The process according to claim 1, wherein the mixture A is prepared by arc discharge or a plasma torch process.

3. The process according to claim 1, where the dispersing is carried out by means of ultrasonication.

4. The process according to claim 1, wherein the hydrolysis of the degradable polymer in the degrading is aided by ultrasonication.

5. The process according to claim wherein the solvent in the providing is toluene.

6. The process according to claim 1, further comprising: (vi) isolating pure monomers (1a) and (2a):
H.sub.2N-[-M.sup.1-].sub.r-NH.sub.2 (1a)
OHC-[-M.sup.2-].sub.s-CHO (2a) or pure monomers (1b) and (2b)
OHC-[-M.sup.1-].sub.r-CHO (1b)
H.sub.2N-[-M.sup.2].sub.s-NH.sub.2 (2b) from the solution D for re-synthesis of the degradable polymer.

7. The process according to claim 6, wherein the pure monomers are isolated by flash column chromatography.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

(1) 1. Polymer Synthesis and Characterization

(2) Materials and General Methods

(3) All reagents and starting materials were purchased from commercial sources and used without further purification. Thermal gravimetric analyses (TGA) were performed using a Mettler Toledo TGA/SDTA 851e at a heating rate of 10 C./min under a nitrogen flow (20 mL/min). Gel permeation chromatography (GPC) was performed on Tosoh High-temperature EcoSEC (RI detector) at high-temperature of 180 C. using 1,2,4-tricholorobenzene (TCB) as eluent. Compounds 1 was prepared according to the literature procedure. 2 was purchased from Sigma-Aldrich.

(4) ##STR00056##

(5) Compared with traditional conjugated polymers that synthesized with cross-coupling reactions (e.g. Suzuki or Stille coupling), the polymers of the invention can be synthesized without any noble metal catalysts and toxic phosphorous ligands. Thus the degradable polymers of the invention are cheaper and more environmentally friendly. The imine bonds are environmentally stable at ambient conditions, and no polymer degradation was observed for over six-month storage. Both polymers show excellent thermal stability with decomposition temperature over 400 C. (FIG. 1).

(6) FIG. 1 shows Thermogravimetric analyses (TGA) of PDPP-PD (5% loss, 404 C.) and PFPD (5% loss, 400 C.).

Example 1

(7) 9,9-didodecyl-9H-fluorene-2,7-dicarbaldehyde (F-CHO): To a 100 mL round bottom flask, compound 1 (3 g, 4.54 mmol) and ethyl ether (40 mL) was added. n-BuLi (2.5 M, 4.54 mL, 11.4 mmol) was added at 78 C. After stirring at 78 C. for 30 min, dry DMF (1.16 g, 15.9 mmol) was added dropwise at 78 C. The mixture was allowed to warm up to room temperature and stirred for 1 h. Then the mixture was quenched with HCl (50 mL, 1 M aqueous solution). The aqueous layer was extracted with dichloromethane (350 mL). The combined extracts were washed with distilled water and dried over anhydrous Na.sub.2SO.sub.4. After removal of the solvents under reduced pressure, the residue was purified via chromatography with silica (eluent: hexane/dichloromethane=4/1 to 1/1) to afford F-CHO as an off-white solid. Yield: 2.23 g (88%). .sup.1H NMR (CDCl.sub.3, 300 MHz, ppm): 10.10 (s, 2H), 8.00-7.84 (m, 6H), 2.13-2.00 (m, 4H), 1.38-0.93 (m, 36H), 0.85 (t, J=6.8 Hz, 6H), 0.62-0.44 (m, 4H).

Example 2

(8) 5,5-(2,5-bis(2-octyldodecyl)-3,6-dioxo-2,3,5,6-tetrahydropyrrolo[3,4-c]pyrrole-1,4-diyl)bis(thiophene-2-carbaldehyde) (DPP-CHO): To a 100 mL round bottom flask, diisopropylamine (0.98 mL, 6.96 mmol) and THF (50 mL) was added. n-BuLi (1.6 M, 2.9 mL, 4.64 mmol) was added and stirred at 0 C. for 30 min to prepare fresh lithium diisopropylamide (LDA). Compound 2 (1.0 g, 1.16 mmol) was then added dropwise into the flask at 78 C. After stirring at 78 C. for 30 min, dry DMF (0.46 mL, 6.96 mmol) was added dropwise at 78 C. The mixture was allowed to warm up to room temperature and stir for 1 h. Then the mixture was quenched with 20 mL of water. The aqueous layer was extracted with dichloromethane (350 mL). The combined extracts were washed with distilled water and dried over anhydrous Na.sub.2SO.sub.4. After removal of the solvents under reduced pressure, the residue was purified via chromatography with silica (eluent: hexane/ethyl acetate=20/1 to 10/1) to afford DPP-CHO as a dark red solid. Yield: 0.82 g (77%). .sup.1H NMR (CDCl.sub.3, 300 MHz, ppm): 10.03 (s, 2H), 9.04-9.02 (d, J=4.2 Hz, 2H), 7.88-7.86 (d, J=4.2 Hz, 2H), 4.05-4.03 (d, J=7.7 Hz, 4H), 1.89-1.86 (m, 2H), 1.45-1.18 (m, 64H), 0.94-0.81 (m, 12H).

Example 3

(9) Polymerization for PF-PD: To a Schlenk tube (100 mL), F-CHO (800 mg, 1.43 mmol), p-phenylenediamine (154.8 mg, 1.43 mmol), p-toluenesulfonic acid (PTSA) (13.6 mg, 0.0715 mmol, 5 mol %), anhydrous CaCl.sub.2 (200 mg), and anhydrous toluene (50 mL) were added under nitrogen atomsphere. The tube was then sealed under nitrogen atomsphere. The mixture was stirred for 48 h at 110 C. After completion, dry K.sub.2CO.sub.3 (20 mg) was added and the mixture was stirred at 110 C. for 30 mins. Then the mixture was filtered through a nylon filter to remove the drying agent and any insoluble salts. After removing the solvent in the filtrate, the polymers were collected to afford a yellow solid (890 mg, yield 99%). .sup.1H NMR (C.sub.2D.sub.2Cl.sub.4, 400 MHz, ppm): 8.64-8.57 (m, 2H), 8.00-7.83 (m, 6H), 7.40-7.23 (m, 2H), 6.77-6.74 (m, 2H), 2.12-2.05 (m, 4H), 1.28-1.06 (m, 36H), 0.88-0.85 (t, J=6.8 Hz, 6H), 0.69-0.55 (m, 4H). Molecular weight from HT-GPC, M.sub.w: 15,580 Da, PDI: 2.19.

Example 4

(10) Polymerization for PDPP-PD: To a Schlenk tube (100 mL), DPP-CHO (250 mg, 0.273 mmol), p-phenylenediamine (29.5 mg, 0.273 mmol), p-toluenesulfonic acid (PTSA) (2.6 mg, 0.014 mmol, 5 mol %), anhydrous CaCl.sub.2 (100 mg, drying agent), and anhydrous toluene (30 mL) were added under nitrogen atomsphere. The tube was then sealed under nitrogen atomsphere. The mixture was stirred for 48 h at 110 C. After completion, dry K.sub.2CO.sub.3 (10 mg) was added and the mixture was stirred at 110 C. for 30 mins. Then the mixture was filtered through a nylon filter to remove the drying agent and any insoluble salts. After removing the solvent in the filtrate, the polymers were purified via Soxhlet extraction for 2 h with dry acetone, 2 h with hexane, and finally was collected with chloroform. The chloroform fraction was then evaporated to remove the solvent and afford a dark green solid (213 mg, yield 79%). .sup.1H NMR (C.sub.2D.sub.2Cl.sub.4, 400 MHz, 393 K, ppm): 10.07-10.05 (m, 2H), 8.95-8.67 (m, 3H), 7.89-7.67 (m, 2H), 7.40 (m, 1H), 7.26-7.24 (m, 1H), 6.76-6.74 (m, 1H), 4.15-4.07 (m, 4H), 2.06-2.19 (m, 2H), 1.51-1.32 (m, 64H), 0.96-0.93 (m, 12H). Molecular weight from HT-GPC, M.sub.w: 39,574 Da, PDI: 2.64.

(11) 2. Polymer Degradation Under Acidic Conditions

Example 5

(12) To explore the polymer degradation process, we prepared polymer solutions in THF and monitored their degradation process by UV-Vis-NIR spectroscopy (Cary 6000i spectrophotometer, Varian) after adding a small amount of trifluoroacetic acid (TFA) and a drop of DI water. The degradation can also be performed in toluene, which requires sonication (bath sonicator) to accelerate the degradation reaction, because water has low solubility in toluene.

(13) FIG. 2a, b show polymer degradation under the catalysis of acid. Absorption spectrum changes of (a) PF-PD and (b) PDPP-PD after adding a small amount of trifluoroacetic acid (TFA) and a drop of water.

(14) FIG. 2a, b display the degradation process of polymer PF-PD and PDPP-PD, respectively, under acidic conditions at room temperature. Both polymers can be degraded into monomers under a catalytic amount of acid. After complete degradation, their absorption spectra were almost identical as their monomers. PF-PD degrades faster, usually within several minutes, whereas PDPP-PD degrades much slower. It was found that sonicating or heating the polymer solutions could accelerate the degradation process. For example, the degradation of PDPP-PD can be completed within 30 mins using a bath sonicator at 50 C.

(15) 3. Selective Dispersion and SWNT Characterization

(16) General Sorting Procedure for Polymer PF-PD Using Plasma Grown SWNTs:

Example 6

(17) 5 mg of PF-PD and 5 mg of plasma grown SWNTs (RN-020, as obtained from Raymor Nanotech, contains 30% SWNTs) were mixed in 25 mL of toluene and ultrasonicated for 30 min at an amplitude level of 50% (Cole Parmer ultrasonicator 750 W). The solution was then centrifuged at 17 000 rpm (22 000 g) for 30 min at 16 C. The supernatants were collected and provided the final sorted s-SWNT solutions.

(18) The absorption spectrum of PF-PD do not overlap with the absorption of metallic SWNTs. Therefore, both the dispersion concentration and the selectivity can be evaluated by UV-Vis-NIR measurement.

(19) General Sorting Procedure for PDPP-PD Using Arc-Discharged SWNTs:

Example 7

(20) 5 mg of PDPP-PD and 10 mg of arc-discharged SWNTs (P2-SWNTs, purchased from Carbon Solutions, Inc., contains 65% SWNTs) were mixed in 25 mL of toluene and ultrasonicated for 30 min at an amplitude level of 70% (Cole Parmer ultrasonicator 750 W). The mixture was then centrifuged at 17 000 rpm for 30 min at 16 C. The supernatant (80%) was collected to afford the final sorted SWNT solution.

(21) The dispersion concentration was evaluated by UV-Vis-NIR measurements using a 1 cm path-length quartz cells with toluene as a background. To determine the selectivity, SWNT thin films on glass substrate were prepared by drop-casting the supernatant on the substrate and annealing the film at 500 C. under Ar for 1 h to remove the wrapping polymers.

(22) General Procedure for Polymer Removal and Recycling:

Example 8

(23) To degrade the wrapping polymers, a small amount of TFA (10 uL) and several drops of water was added to the sorted SWNT solution. The solution was bath sonicated for 0.51 h to complete the hydrolysis reaction. SWNT precipitates were formed after polymer degradation. Then the solution was centrifuged at 17 000 rpm for 5 min to sediment the SWNTs. After collecting the supernatant, 20 mL acetone was added. The mixture was sonicated for 10 min in a bath sonicator and centrifuge again to wash the SWNT sediments. Repeating the washing step twice. Then the s-SWNT sediments were collect by filtration.

(24) The sediments produced in the polymer sorting, which contains amorphous carbon, met-SWNTs, and undispersed s-SWNTs, absorbed a significant amount of polymers. To recycle these polymers, 20 mL THF was added to the sediments. The mixture was sonicated for 10 min in a bath sonicator and centrifuge at 17 000 rpm for 15 min to wash the SWNT sediments. Repeating the washing step twice. Then the sediments were removed by filtration.

(25) All the supernatants and filtrates from the above steps were combined for polymer recycling. After removal of the solvents under reduced pressure, the monomers were purified via flash chromatography and provided pure monomers. The purified and cycled monomers can be used for polymerization another time.

(26) Determination of the Sorting Yield and Polymer Cycling Efficiency:

Example 9

(27) 50 mg of PF-PD and 50 mg of plasma grown SWNTs (RN-020) were mixed in 25 mL of toluene and ultrasonicated for 30 min at an amplitude level of 50% (Cole Parmer ultrasonicator 750 W) while externally cooled with a dry ice bath. The solution was then centrifuged at 17 000 rpm (22 000 g) for 30 min at 16 C. 80% of the supernatants (20 mL) was collected. Absorption spectroscopy revealed a peak absorption of A=2.486 at =939 nm for a d=1 cm cuvette (FIG. 4a).

(28) Following the above polymer removal steps for the supernatants, the sorted SWNTs were finally filtered through a 0.2 um pore-size PTFE membrane, washed with toluene 3 times (30 mL), and dried in vacuum at 60 C. A SWNT film was formed on the membrane. Weighting the SWNT film gives a total mass of 1.41 mg. Thus the SWNT concentration of the sorted solution (20 mL) was calculated to be 0.0705 mg/mL. The optical density at 939 nm was used to calculate the absorbance coefficient (, mL mg.sup.1 cm.sup.1) via Beer-Lambert law: A=OD.sub.939=/c, where / is the path length (cm) of the cuvette and c is the SWNT concentration (mg/mL). The absorbance coefficient () was determined to be 35.3 mL mg.sup.1 cm.sup.1, consistent with a recently reported value of 34.9 mL mg.sup.1 cm.sup.1 for similar type of s-SWNTs. With this absorbance coefficient (), the sorting yields for different conditions can be calculated using their absorption peaks at 939 nm. The yield calculated for PF-PD sorted SWNTs was based on the semiconducting SWNT amounts of the raw SWNTs, which is 20% of the total mass of the raw SWNTs (RN-020).

(29) Raman and PLE Characterization of Sorted SWNTs:

(30) Raman spectroscopy was carried out at 2.33 eV (532 nm), 1.93 eV (638 nm) and 1.58 eV (785 nm) excitation at 100 magnification and 1-m spot size. The peak positions were calibrated with the Si line at 521 cm.sup.1. The Raman peaks can be assinged to metallic or semiconducting SWNTs according to the Raman Kataura plot. The PLE spectra of various SWNT samples in toluene were taken according to our previous reported method.

(31) FIG. 3 shows a comparison for using different purity SWNTs under the same conditions (polymer/SWNT ratio=1:1).

(32) As-produced 30% SWNTs (RN-020, 10 USD/g) and Semi-purified SWCNT (RN-220, 6070% purity, 45 USD/g) were purchased from Raymor Industries Inc. Semi-purified SWNTs display obviously higher SWNT contents than as-produced raw SWNTs, as demonstrated by their NMP dispersed solutions. After PF-PD sorting, 30% SWNTs showed a value of 0.407 and a yield of 23.7%, while 70% SWNTs showed a value of 0.408 and a yield of 19.8%. The yield calculation was based on the s-SWNT amounts of the total mass of raw SWNTs (for RN-020, the s-SWNTs amount is 20%; for RN-220, the s-SWNT amount is 46%). Therefore, compared with 70% semi-purified SWNTs, 30% raw SWNTs showed very similar selectivity and even higher yield.

(33) FIG. 4 shows (a) Absorption spectra of raw SWNTs (30% purity) dispersed by PF-PD using different polymer concentrations (polymer/SWNT ratio is 1:1). Spectra were taken in a 1 cm path length cuvette. (b) Plot of yields and values for different polymer/SWNT concentrations. Higher concentrations gives lower yield but provides higher throughput for large-scale production.

(34) FIG. 5 shows (a) Absorption spectra of SWNTs dispersed by PDPP-PD with various polymer/SWNT ratios in toluene. (b) Normalized absorption spectra of the PDPP-PD sorted SWNT thin films for various polymer/SWNT ratios. The SWNT thin films were prepared by drop-casting the supernatants on glass substrates and then annealing at 500 C. under Ar for 1 h to remove the wrapping polymers.

(35) As the SWNT amount increases while keeping a constant polymer amount, the dispersion concentration first increased and then decreased. The decrease in the dispersion concentration might due to the absorption of the polymers by the excessive carbonaceous sediment, which then decreased the available polymer concentration for sorting in solution. The sorting selectivity is also dependent on the polymer/AD-SWNT ratio, 5 mg/10 mg ratio gives the best selectivity.

(36) FIG. 6 shows purity comparison of the s-SWNTs dispersed by PF-PD and PDPP-PD with commercial available 99.9% pure s-SWNTs from Nanointegris. PDD-PD sorted s-SWNTs with larger tube diameters showed similar selectivity, while PF-PD sorted s-SWNTs with similar tube diameters showed obviously better selectivity.

(37) FIG. 7 shows Raman spectra of pristine SWNTs and SWNTs sorted by polymer PF-PD, excited using (a) 532 nm and (b) 638 nm lasers. Under 532 nm and 638 nm excitation, several RBM peaks of s-SWNTs were observed in the range of 150-200 cm.sup.1 in the pristine SWNTs. After sorting, some peaks remained and the relative peak intensity also changed, indicating that the polymer prefers to disperse certain chiralities of s-SWNTs. To compare peak intensities of pristine and sorted SWNTs, their G.sup.+ peaks were normalized to the same intensity.

(38) FIG. 8 shows Raman spectra of pristine SWNTs and SWNTs sorted by polymer PDPP-PD, excited using (a) 785 nm, (b) 532 nm and (c) 638 nm lasers. To compare peak intensities of pristine and sorted SWNTs, their G.sup.+ peaks were normalized to the same intensity.

(39) FIG. 9 shows absorption spectra of the solutions using monomers for dispersion. No SWNT absorption was detected, indicating that both monomers cannot disperse any SWNTs.

(40) First the ability of PF-PD to disperse a low-cost raw SWNTs with diameters of 0.9-1.5 nm (RN-020, $10/g from Raymor Industries Inc., 30% SWNTs purity) was tested. FIG. 10a demonstrates that PF-PD can disperse high concentrations of large-diameter s-SWNTs with barely detectable metallic peak in the M.sub.11 region. The optical density (OD) at 939 nm for PF-PD sorted SWNTs is up to 2.498 in a 1 cm path length cuvette, corresponding to a SWNT concentration of 0.0705 mg/mL. The OD value is significantly higher than other reported conjugated polymers used for large diameter SWNTs, indicating the strong dispersion ability of PF-PD. Compared with previous sorting methods using more expensive purified SWNTs (5070% purity, prices are usually 48 times higher), the results indicate that as-produced raw SWNTs (30% purity) can be directly used for polymer wrapping without affecting the sorting yield and selectivity (FIG. 3). Polymer/SWNT ratios and concentrations are factors that affect the yield and selectivity of SWNT dispersion. The sorting yield was determined by absorption intensity of the S.sub.22 peak and calculated relative to s-SWNT amounts of the total mass of SWNTs. The selectivity was evaluated based on value, which is defined by the ratio of peak and background area of both S.sub.22 and M.sub.11 absorptions. Higher values indicate higher s-SWNT purities and values>0.40 were correlated to a purity>99%.

(41) FIG. 10b plots the relationship between the yield and selectivity for different polymer/SWNT ratios. Compared with literature reports that reaching high values>0.40 usually resulted in low yields (<5%) for large-diameter SWNTs, polymer PF-PD demonstrated both high value of 0.407 and high yield of 23.7% simultaneously (polymer/SWNT ratio=1 in FIG. 10b), suggesting the strong dispersion ability and high selectivity of this polymer. Reducing the polymer/SWNT ratios lead to even higher selectivity with a value up to 0.445 but relatively lower yield. Polymer/SWNT concentrations were also varied from 0.2 mg/mL to 2 mg/mL (FIG. 4a). Under the same polymer/SWNT ratio (1:1), higher concentration resulted in higher selectivity but lower yield. This inverse relationship between yield and selectivity was also observed in many other polymer systems.

(42) Compared with traditional conjugated polymers, DPP based polymers tend to disperse larger diameter SWNTs. Arc-discharged SWNT were used to investigate the ability of PDPP-PD for dispersing SWNTs (FIG. 5). PDPP-PD can disperse a significant amount of arc-discharged SWNTs with an estimated yield of 2.3%. Because the absorption spectra of DPP based polymers overlap with the M.sub.11 region of the sorted SWNTs, the polymers were removed by thermally annealing polymer-SWNT films at 500 C. under Ar before estimating the selectivity for different sorting conditions. Since the value cannot be directly used to estimate the selectivity for DPP sorted SWNTs due to interference from the polymers, the sorted absorption spectra were compared with that of commercially available 99.9% s-SWNTs (FIG. 6). Both PDPP-PD sorted s-SWNTs and PF-PD sorted s-SWNTs showed similar or even better selectivity compared with the 99.9% s-SWNTs, indicating the high selectivity of the degradable polymers.

(43) To further validate the enrichment of s-SWCNTs, Raman spectra of the pristine SWNTs and the polymer sorted SWNTs were measured (FIG. 10c, FIG. 7, and FIG. 8). Under the 785 nm excitation, strong radial-breathing-mode (RBM) peaks of metallic tubes in the range of 140190 cm.sup.1 were observed for the pristine SWNTs. After sorting, this peak almost disappeared. In the G peak region (1500-1600 cm.sup.-1), the intensity of G.sup. peaks mainly from m-SWNTs at 1550 cm.sup.1 significantly decreased, whereas the G.sup.+ peaks mainly from s-SWNTs at 1600 cm.sup.1 remained unchanged. Other laser excitations (532 nm and 633 nm) also confirmed the removal m-SWNTs and the dispersion of certain charilites of s-SWNTs (FIG. 7).

(44) Both dialdehyde monomer of PF-PD and PDPP-PD are not able to disperse SWNTs (FIG. 5), suggesting their weak interactions with SWNTs. To release the s-SWNTs, a catalytic amount of TFA and some DI water were added to the sorted solution. The solution was sonicated to accelerate the hydrolysis reaction. SWNT precipitates were formed after polymer degradation. The SWNT precipitates were centrifuged and collected by filtration though a 0.2 m PTFE membrane. The absorption spectra of the filtrate only showed the monomer absorption, indicating all the enriched s-SWNTs were quantitatively collected (FIG. 10d). The filtered s-SWNTs can be redispersed in NMP and aqueous solutions using surfactants with almost identical absorption spectra in both M.sub.11 and S.sub.22 region (FIG. 10d), indicating that the sorted SWNTs can also be used for NMP or aqueous based deposition methods in electronic applications. No polymers or monomers can be detected by absorption spectroscopy for the redispersed SWNTs in NMP, demonstrating the quantitative removal of the polymers. This result was further confirmed by the by X-ray photoelectron spectroscopy (XPS) measurement of the SWNT solids, where the N 1 s peaks attributed to the N atoms in the polymers disappeared in the released SWNT samples (FIG. 10e). It was found that to of the conjugated polymers were absorbed by the sediments containing undispersed s-SWNTs, m-SWNTs, and amorphous carbons. Thus, to recycle the polymers, polymers in both supernatants and sediments were degraded and filtered. The filtrates can be purified by flash column chromatography to obtain pure monomers for polymer re-synthesis. Almost all of the monomers can be recycled with recycling yield up to 93% after purification, indicating that degradable polymer approach of the invention will effectively lower the polymer costs for SWNT separation.

(45) 4. Thin-Film Transistor Characterization of Sorted SWNTs

Example 10

(46) Thin-film transistors (TFTs) were fabricated by drop-casting the PF-PD sorted SWNT solution on patterned Au (source-drain)/SiO.sub.2 (300 nm)/doped Si substrates. Then the substrates were rinsed with toluene and annealed at 200 C. for 20 min under ambient conditions. Toluene rinsing was used to remove most of the polymers, which can lead to better tube-tube junctions and improved charge carrier mobility. FIG. 11a displays the atomic force microscopy (AFM) image showing a SWNT network with a tube density ca. 50 tubes/m.sup.2. Most of the sorted tubes exhibited tube lengths in the range of 0.5 to 2 m. The TFT devices exhibited high hole mobilities in the range of 20-49 cm.sup.2 V.sup.1 s.sup.1 with on/off ratios>10.sup.5 (FIG. 11b). Compared to the unsorted raw SWNTs with on/off ratios lower than 10, the high on/off ratio of our sorted s-SWNTs further confirms the high purity of the sorted s-SWNTs.

(47) In summary, imine-based conjugated polymers based on different types of building blocks have been tested for selective dispersion of s-SWNTs. They exhibited strong dispersion ability for large-diameter s-SWNTs with high yield (up to 23.7%) and high selectivity (>99%). The inventive approach offers significant advantages (e.g. low-cost, high-selectivity, recyclable, and less-damage to SWNTs) over other methods but still employs regular building blocks without specific structure design. More importantly, the degradable polymer design can be readily used for other -conjugated structures, thus demonstrating a general approach for low-cost separation of clean s-SWNTs.

(48) FIG. 10a-e show selective dispersion and release of s-SWNTs using PF-PD. (a) Absorption spectra of raw SWNTs (30% purity) dispersed by PF-PD (PF-PD/SWNT ratio=1:1, concentration 2 mg/mL) and NMP. Spectra were taken in a 1 cm path length cuvette. (b) Plot of yields and values for different polymer/SWNT ratios (polymer concentration 0.2 mg/mL). (c) Raman spectra of the raw SWNTs and PF-PD sorted SWNTs (excited by 785 nm laser). Polymers were degraded and removed for the sorted SWNTs. (d) Absorption spectra of the as-sorted SWNTs, sorted SWNTs redispersed in NMP and surfactant SC (1% aqueous) polymer PF-PD, and monomer (F-CHO) after polymer degradation. (e) XPS results for the PF-PD wrapped SWNTs and released SWNTs.

(49) FIG. 11b shows transfer characteristics of a typical TFT device (V.sub.DS=40V).