Polymeric films based on blends of 6FDA–6FpDA polyimide plus several copolyfluorenes for CO₂ separation

Abstract

Three emitting copolyfluorenes, based on 2,7-(9,9-dihexyl)fluorene and different aryl groups (1,4-bencene, PFH-B; 1,4-bencen-1,2,5-thiadiazole PFH-BT; 1,4-naphthalen-1,2,5-thiadiazole, PFH-NT), showing diverse acceptor character, in different proportions were blended with a polyimide 6FDA–6FpDA to make a series of films. These copolyfluorene–polyimide blends were prepared and characterized in the solid state, using several techniques. The fluorescence of conjugated polymers can be used as a tool to understand the formation of the membrane and also to increase permeability and selectivity in comparison to films which do not fluoresce. The relationship between the intrinsic fluorescence of conjugated polyfluorenes and their gas separation properties has been explored in order to establish the influence of the composition and the nature of the aryl group in the conjugated polymer, on the gas separation performance. In all cases, a low proportion of copolyfluorene (<1% weight) gives better CO₂/CH₄ permselectivity properties than the original pure polyimide matrix. The best results were found for the samples that contain PFH-NT. These samples gave over 23% increase in the CO₂ permeability with a 15% increase in CO₂/CH₄ selectivity. Finally, the loss of efficiency in conjugation mechanisms of absorption and emission of the samples could be explained on the basis of the π-stacking of the polymer chains, produced when a certain low percentage of conjugated polymers in the blend is surpassed. When this π-stacking starts, the gas permeation properties start to decline too.

---

1. Introduction

The growing world-wide energy demand causes an increasing interest in exploiting contaminated EOR (enhanced oil recovery) wells, for example, by using CO₂ injections in miscible-flood secondary recovery procedures to rejuvenate mature wells. Moreover global perceptions on climate change are demanding a reduction of carbon emissions. Therefore the gas separation processes for applications in chemical and petrochemical industries are gaining increasing importance.^1–3 The removal of acid gases (such as CO₂) from natural gas (typical pipeline specifications require 2–5% CO₂) is an important process since it increases the heating value of natural gas; prevents corrosion of the pipeline during gas transport; decreases the volume of gas to be transported in pipelines and reduces atmospheric pollution.^4–6 Therefore, natural gas sweetening (CO₂/CH₄ separation) is increasingly important.

Membrane processes are commercially demonstrated technologies for natural gas processing applications due to their low energy consumption, easy operation and maintenance, environmental friendliness and small footprint compared to amine absorption, pressure swing adsorption, rectification, and cryogenic distillation.⁷ In order to have a good gas separation performance, membranes must have high permeability and permselectivity, excellent chemical resistance (for example against corrosive materials such as H₂S), great thermal stability (for high temperature applications), good mechanical properties (for high pressure applications), and good plasticization resistance.⁸

Polyimides are well known for their excellent thermal oxidative stability, good organic solvent resistance and exceptional mechanical properties, along with an extraordinary ability of separating complex mixtures of gases in diverse applications.^9–11 Specifically, polyimides with –CF₃ groups in their structure produce a restriction of torsional motion of the neighboring phenyl rings that enhances permselectivity. For this reason, one of the most studied polyimides for gas separation is based on 2,2-bis(3,4-dicarboxyphenyl) hexafluoropropane dianhydride (6FDA), and the diamine 4,4′-(hexafluoroisopropylidene) dianiline (6FpDA).^12–14 Fully aromatic polyimides have interesting photophysical properties due to both intra- and inter-chain charge transfer interactions.¹⁵ Nevertheless, the four trifluoromethyl groups in 6FDA–6FpDA have a significant electronegativity, thus withdrawing electrons and avoiding significant π-bonds, or causing π-stacking and the subsequent quenching of optical activity of the polyimide.^16,17

Among various approaches, the use of polymer blends has been recognized as one of the most cost- and time-effective routes to get good separation properties as they combine the advantages of different materials into a new compound with unique and synergetic properties that are difficult to obtain by other synthesis means.¹⁸ In this way, we explore the possible compatibility and miscibility of blends of polyimides and conjugated polymers. Conjugated polymers are defined as systems connected by π-orbitals with delocalized electrons. One of the most typical conjugated polymers are polyfluorenes and their copolymers, which are a family of blue to red-emitting polymers with large extinction coefficients and great photoluminescence quantum yields.^19,20 Moreover, conjugated polymers are widely used for applications far away from gas separation.^21–24 For example, they have been used in electronic and photonic devices, showing a great ability and capacity to transfer charge and energy.

Referring to gas separation, copolyfluorenes and 6FDA–6FpDA show accumulation of charges that would give them a clear polar nature, leading to dipolar moments that should interact with the intrinsic dipole of the CO₂ molecule. As the CH₄ molecule does not have a relevant polarity, the CO₂/CH₄ selectivity should improve.

This paper presents the results obtained for some blends developed from the polyimide 6FDA–6FpDA and several copolyfluorenes (PFH–B, PFH-BT and PFH-NT) in several proportions. Our aim was to correlate the optimal concentration of copolyfluorenes in these blends for gas separation with the properties of fluorescence that, in turn, give us some clues on their structure.

2. Experimental

2.1. Reagents

All solvents were HPLC grade, used as received without further purification, and were purchased from Aldrich Corporation. N,N-Dimethylacetamide (DMAc), chloro(trimethyl)silane (CTMS), pyridine (Py), 9-dihexylfluorene-2,7-diboronic acid, 1,4-dibromobenzene, 1,4-dibromobenzo{2,3}-1,2,5-thiadiazole and tetrakis(triphenylphosphane) palladium(0) were also purchased from Aldrich. Specific reagents: 2,3-diaminonaphthalene was supplied by TCI; 4,4′-(hexafluoroisopropylidene) diphthalic anhydride (6FDA) and 4,4′-(hexafluoroisopro-pylidene)dianiline (6FpDA) were acquired from ABCR and sublimed just before use.

2.2. Synthesis of copolyfluorenes

The synthesis of the monomers and the corresponding copolyfluorenes: copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene] (PFH-B); copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene-1,2,5-thiadiazole] (PFH-BT) and copoly-[9,9-bis(n-hexyl)fluorene-alt-1,4-naphthalen-(2,3)-1,2,5-thiadiazole] (PFH-NT) are shown in Fig. 1. They were synthesized via the Suzuki coupling reaction, using the same methods described in previous studies.^25–29

Fig. 1 Synthetic approach of PFH-B, PFH-BT and PFH-NT: (i) Br₂/AcOH, RT, 3 h (87%); (ii) SOCl₂, Pyr, CHCl₃; 80 °C, overnight (64%); (iii) K₂CO₃, [Pd(PPh₃)₄], THF/H₂O, 85 °C, 48 h.

The required non-commercially available monomers were synthesized as follows:

• 1,4-Dibromo-2,3-diaminonaphthalene: 1,4-dibromo-2,3-diaminonaphthalene was prepared from 2,3-diaminonaphthalene (12.5 mmol) as described in Fig. 1 (step i). Here, the solid obtained was purified using a Versatflash silica column with a dichloromethane/hexane mixture (2 : 1). After evaporation of the fractions and drying under vacuum at 40 °C overnight, a white powder was obtained (2.08 g, 88%).

• 4,9-Dibromonaphto{2,3}-1,2,5-thiadiazole: 1,4-dibromo-2,3-diaminonaphthalene (0.81 g, 2.56 mmol) was dissolved in chloroform (25 ml) as described in Fig. 1 (step ii) and was cooled in an ice bath. Subsequently, a mixture of thionyl chloride (1.27 ml, 17.5 mmol), pyridine (3 ml) and chloroform (7.6 ml) was slowly dropped into the solution over 10 minutes. After complete addition the mixture was stirred overnight at 80 °C. Evaporation of the solvent and posterior purification using silica with a dichloromethane/hexane mixture (3 : 2) yielded an orange solid (564 mg, 64%).

In Table 1, some characteristics of the new monomers and copolymers synthesized are shown.

Table 1 Some characteristics of the newly synthesized chemicals^a

a ^1H and ^13C NMR spectra were recorded on a Bruker AVANCE 500 spectrometer, with tetramethylsilane as an internal reference. FTIR spectra were obtained using a Nicolet model 520P spectrometer with samples prepared as KBr pellets. Coupled size exclusion chromatography (SEC-GPC) was carried out using an HP-1090 liquid chromatograph with autosampler, autoinjector and light scattering detector; ELSD 2200 Alltec. The size exclusion chromatogram was carried out in THF as eluent and data were calibrated using Polymer Laboratories EasiCal PS standards using a two column PLGel 5 μm MIXED-C at constant room temperature.
Monomers
1,4-Dibromo-2,3-diaminonaphthalene	mp: 159–161 °C
	^1H NMR (500 MHz, CDCl3) δ (ppm): 8.00 (“q”, 2H), 7.40 (“q”, 2H) and 4.27 (bs, 4H)
	^13C NMR (125 MHz, CDCl3) δ (ppm): 134.5, 127.8, 125.7, 125.0 and 106.5
	FT-IR (pellet BrK, cm^−1): 3423, 3380, 3347, 3267 (ν –NH2), 1646 (m), 1446 (s), 1252, 752 (s, ν-napht), 703, 638, 574 and 462 (νAr-Br)
4,9-Dibromonaphto{2,3}-1,2,5-thiadiazole	mp: 237–238 °C
	^1H NMR (500 MHz, CDCl3) δ (ppm): 8.40 (“q”, 2H), 7.59 (“q”, 2H)
	^13C NMR (125 MHz, CDCl3) δ (ppm): 150.8, 128.9, 128.0, 123.7 and 114.3
	FT-IR (pellet BrK, cm^−1): 1406, 1264 (s), 1165, 892 (s), 755 (s, ν-napht), 707, 615, 514 (s) and 448
Copolyfluorenes
Copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene] (PFH-B)	^1H NMR (500 MHz, CDCl3) δ (ppm): 7.85–7.82 (m, 4H), 7.71–7.50 (m, 6H), 2.08 (4H), 1.7–1.10 (12H) 0.78 and 0.77 (2 × s, 10H)
	^13C NMR (125 MHz, CDCl3) δ (ppm): 151.8, 141.0, 140.7, 140.1, 128.2, 126.8, 121.2, 120.8, 55.7, 40.6, 31.5, 29.7, 24.0, 22.6, 14.0
	GPC (polystyrene standard): Mn = 6400; Mw = 10 900; PDI: 1.7
Copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene-1,2,5-thiadiazole] (PFH-BT)	^1H NMR (500 MHz, CDCl3) δ (ppm): 8.09–7.95 (m, 6H), 7.79–7.73 (m, 2H), 2.15 (4H), 1.17 (12H), 0.95 (4H) and 0.80 (6H)
	^13C NMR (125 MHz, CDCl3) δ (ppm): 154.7, 151.9, 141.0, 136.6, 133.7, 133.2, 128.5, 124.1, 120.2, 55.6, 40.4, 31.6, 29.9, 23.8, 22.7, 14.2
	GPC (polystyrene standard): Mn = 2050; Mw = 4500; PDI: 2.2
Copoly-[9,9-bis(n-hexyl)fluorene-alt-1,4-naphthalen-(2,3)-1,2,5-thiadiazole] (PFH-NT)	^1H NMR (500 MHz, CDCl3) δ (ppm): 8.06–7.90 (m, 6H), 7.78–7.72 (m, 4H), 2.19 (4H), 1.7–1.3 (12H), 1.13 (4H) and 0.83 (6H)
	^13C NMR (125 MHz, CDCl3) δ (ppm): 154.5, 151.6, 141.2, 140.7, 138.2, 133.8, 128.5, 127.8, 124.3, 123.7, 120.6, 55.6, 40.2, 30.5, 29.5, 24.1, 22.4, 13.4
	GPC (polystyrene standard): Mn = 4500; Mw = 8600; PDI: 1.9

---

2.3. Synthesis of the polyimide

Polyimide 6FDA–6FpDA was synthesized by us following the classical in situ silylation two-step method from the 6FDA dianhydride and the diamine 6FpDA, improved by Muñoz et al.³⁰

2.4. Blends and film preparation

A sketch of the polymers blended is shown in Fig. 2a. The copolyfluorenes were added in different percentages 0.1, 1, 5 and 10% w/w by simple mixing with 6FDA–6FpDA. The membranes were manufactured by using tetrahydrofuran (THF) as a solvent. 10% w/v solutions of the blends in THF were prepared and filtered with a 1 mm (Titan, PTFE) membrane. The films were obtained by the method of casting on a leveled glass plate kept at 30 °C. Afterwards they were kept at 60 °C during 12 h in order to evaporate most of the solvent. The so-obtained films were separated from the glass by immersion in distilled water and introduced in a vacuum oven at 180 °C where they were kept during 6 h. The films (some of them are shown in Fig. 2b) had a thickness between 35–45 μm.

Fig. 2 Chemical structure of the polymers used here (a) and the films obtained (b).

2.5. Instrumentation

A Thermal Analysis Q500 instrument was used for thermogravimetric analysis (TGA). Disc samples, from 5 to 15 mg, were cut from films and tested. When running dynamic scans, it was done in Hi-Resolution mode, where the heating rate is automatically adjusted in response to changes in the rate of weight loss, which results in improved resolution, with an initial heating rate of 10 °C min⁻¹ under a flux of nitrogen.

Differential scanning calorimetry (DSC) analyses were carried out in a Mettler Toledo (DSC 822e) calorimeter equipped with a liquid nitrogen accessory. Disc samples, also weighting 5–15 mg, were sealed in aluminium pans. Samples were afterwards heated at 10 °C min⁻¹ to each target temperature.

UV absorption spectra were measured with a Shimadzu UV-265FS spectrophotometer. Steady-state emission and excitation spectra were acquired at room temperature using a PTI QuantaMaster Model QM-62003SE spectrofluorometer for the different films, and devices using a solid support were aligned to keep light beams at normal incidence (it is well known that results may vary depending on the angle between the excitation source and detector). Obviously, the signal intensity obtained depends on the optical path, this is why, although the thickness of the samples is very similar, the intensity has been corrected by taking into account the actual thickness of each sample.

2.6. Gas permeation and selectivity

The permeability, P, for several gases (namely: O₂, N₂, CH₄ and CO₂) was determined by using a permeator with constant volume and variable pressure which uses the “time-lag” operation method. Measurements were carried out at 3 bar and 30 °C. The equations used and the mode of operation of the equipment, as well as an outline of it, has been described previously.^31,32

3. Results and discussion

3.1. Characterization of the blends

3.1.1. Thermal stability. TGA results for the synthesized films containing 10% emitting copolyfluorenes and pure 6FDA–6FpDA after being treated at 180 °C over 6 hours are shown in Fig. 3 (similar results were obtained for other percentages).

Fig. 3 TGA curves in dynamic conditions for the blends with the highest percentage of copolyfluorene and pure 6FDA–6FpDA without any copolyfluorene.

The initial loss from ambient temperature to 100 °C, can be attributed to water adsorbed on the sample. A second step, approximately from 100 to 250 °C, is due to water adsorbed into the film along with some residual solvent, with a weight loss from 1 to 1.7%. In the third step the main differences from sample to sample were found. It is worth mentioning that the lowest thermal stability is found for the blend containing PFH-NT which is the most voluminous. In all cases, a comparison with the polymer without the presence of copolyfluorene (pure 6FDA–6FpDA) shows that this step cannot be due to an induced degradation of a part of the polyimide structure because, in this range, the degradation for the blends containing copolyfluorene happens at higher temperatures than for the pure polyimide. The final stage of weight loss is due to the thermal decomposition of the remaining aromatic segments. For all copolymers, the residual carbon content at 800 °C was around 52%, with the exception of the blend with a PFH-NT residue which was about 46%.

3.1.2. Glass transition temperature. The samples were heated in a DSC instrument with a cyclic method in order to monitor the changes in the thermal properties of the films. First, T_g values for the copolyfluorenes were obtained between 120 and 150 °C (116 °C for PFH-B, 148 °C for PFH-BT, and 149 °C for PFH-NT), while for 6FDA–6FpDA the T_g was 287 °C. Then, the analysis was conducted with the blends with the highest content of the copolyfluorenes (10%). The results confirmed the existence of a single T_g. For the blend with PFH-B the T_g was 279 °C; in the case of the polymer with PFH-BT the T_g was 276 °C and finally for the polymer with PFH-NT the value obtained was 278 °C. Obviously all values of the blends are below the value of the polyimide 6FDA–6FpDA.

The existence of a single T_g for different blends confirms the complete miscibility of the polymers.³³ One of the major drawbacks when dealing with polymer blends is a lack of miscibility.

3.1.3. Optical properties. For all the blends with different copolyfluorene concentration, tests of absorption and emission have been performed both in solution and the solid state, with insignificant differences. Spectra of solutions in tetrahydrofuran (THF) show maximal absorption wavelengths at 230, 390, 445 and 492 nm for the 6FDA–6FpDA matrix, PFH-B, PFH-BT and PFH-NT, respectively. These values were used to excite the samples and to register the subsequent solid state emission spectra.

Here, a wide range of emitting wavelengths is covered (PFH-B has a blue emission; PFH-BT has a green emission and PFH-NT has a red emission), and furthermore the 6FDA–6FpDA matrix gives a transparent film with negligible signals of fluorescence, at the excitation wavelength used. The series of membranes with blends in different ratios have been measured under the same conditions.

A. Copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene] (PFH-B) blends. The emission spectra obtained for the polymer blends derived from the 6FDA–6FpDA matrix and PFH-B are shown in Fig. 4. The resulting steady-state emission spectra have a shape which is typical of polyfluorenes. It is composed of a 0–0 band centered at 417–421 nm with a vibronic shoulder at 440 nm, assigned to the mode 0–1.

Fig. 4 Emission spectra for 6FDA–6FpDA/PFH-B films at different percentages (λ_exc = 390 nm). Inset: normalized emission intensity as a function of the percentage of PFH-B in the blend.

The blend with a content of 0.1% PFH-B, gives a low fluorescence signal or intensity, which rises for the 1% blend and remains almost constant for the 5% blend, followed by a decrease at 10% content. This quenching effect of the emission is normally found for the films of conjugated polymer blends. PFH-B provides π-type interactions between chains, decreasing the fluorescence intensity, because not all the delocalized electrons in the structure decay to produce fluorescence. Some of these delocalized electrons are being used in interactions with other chain-sites of the same polymer, or with the trifluoromethyl groups in 6FDA–6FpDA, with a decrease of the conjugation capacity.

In order to determine the presence and influence of the film on the fluorescence activity of the conjugated polymer (PFH-B in this case), it was extracted by simple solution: one milligram of membrane was dissolved in 25 ml of THF. Thus, the maximum wavelength (0–0 band/transition) obtained for pure PFH-B in solution was at 407 nm while the 0–1 transition was at 428 nm (data not shown). The maximal intensities of fluorescence for the solution test and solid state were compared (inset in Fig. 4) where intensities were normalized to the first band with the lowest concentration of polyfluorene (0.1%).

For samples extracted in solution, as expected, the tendency is to increase the intensity when the percentage of polyfluorene is higher. The lack of proportionality is much stronger in the solid state samples, which could be explained as a consequence of concentration and optical pathway effects. In effect, the reduction of the intensity of fluorescence for polyfluorene contents over 0.1% could be caused by low thickness or self-absorption of high-energy photons.³⁴

B. Copoly-[9,9-bis(n-hexyl)fluorene-alt-1,4-bencene-1,2,5-thiadiazole] (PFH-BT) blends. The insertion of a benzothiadiazole (BT) group into the copolyfluorene structure provides an acceptor of electron charge that acts as a powerful excitation trap. In fluorene copolymers, the addition of BT in the main chain (PFH-BT) produces a quenching of the blue emission of the fluorene units, as compared to that seen in the films of PFH-B blends, due to the efficient excitation energy transfer from fluorene to the BT ring.³⁵ Moreover, in analogous polyfluorene derivatives containing BT, their photoluminescence promotes strong dipole–dipole coupling, and nearly complete quenching of the blue luminescence.^36,37

Fig. 5 shows the emission spectra obtained for the blends containing PFH-BT copolyfluorenes with different weight proportions. The incorporation of the BT group enhances the fluorescence and causes the appearance of a broad fluorescence band in the range of 535–550 nm for the sample with a very low content of copolyfluorene (0.1%). Furthermore, an increase of emission intensity was registered from 0.1% to 1% weight of PFH-BT blends while for 5 and 10% blends the emission decreased. The maximum intensity obtained was found for the blends with 1% of PFH-BT copolymer. Possibly, with 5% of conjugated polymer the effect of auto-quenching due to interaction between chains appeared. For the sample with 10% the effect was even more pronounced. Here, a high concentration of conjugated polymer in the bulk matrix produces a stabilization of the intra- or inter-molecular charge transition state because torsion is not possible anymore. This has been previously described in an oligomer of fluorene–dibenzothiophene-S,S-dioxide (FSF) dispersed in a Zeonex matrix.³⁸

Fig. 5 Emission spectra for 6FDA–6FpDA/PFH-BT films at different percentages (λ_exc = 445 nm). Inset: normalized emission intensity as a function of the percentage of PFH-BT in the blend.

The maximum wavelength emission for solutions of extracted conjugated polymer in THF solution was at 538 nm. Like for the PFH-B blends, it was observed that the samples in solution (1 mg of blend per 25 ml THF) have much higher relative intensities than the film samples. In the solid state, the blend 6FDA–6FpDA/PFH-BT 1% showed the greatest electron delocalization.

C. Copoly-[9,9-bis(n-hexyl)fluorene-alt-1,4-naphthalen-(2,3)-1,2,5-thiadiazole] (PFH-NT) blends. The naphthodithiadiazole (NT) group increases red emission due to its extended conjugation length, with any strong donor group in the copolymer structure leading to a possible suppression of intermolecular dipole–dipole interactions.²⁸ This group produces an increase in emission with respect to the BT group due to its intrinsic planar character and different π-stacking interaction in the films. Fig. 6 shows the emission spectra obtained for the 6FDA–6FpDA/PFH-NT blends.

Fig. 6 Emission spectra for 6FDA–6FpDA/PFH-NT films at different percentages (l_exc = 492 nm). Inset: normalized emission intensity as a function of the percentage of PFH-NT in the blend.

In this case, the maximum intensity appears for the blends with 5% of the PFH-NT polymer. For the polymer with 10% polyfluorene, the quenching effect between chains is significant, producing a strong decrease of the peak intensity, probably due to the same stabilization of the inter-molecular charge transition commented on above, but appearing now for larger concentrations.

It was observed for the solution samples (1 mg per 25 ml THF) that the maximum emission wavelength of the pure conjugated polymer appears at 618 nm, corresponding in this case, with the value for the sample with 5% weight of PFH-NT. The relation between the fluorescence intensity for the films and the solutions can indicate that there isn’t any influence of the interactions between the chains for blends until a proportion of 5% of copolyfluorene (see inset in Fig. 6). Again, the variation observed at higher percentages may be produced by hindering the chain torsion and stabilization of the inter-molecular charge transition state. This explanation agrees with the red-shift wavelength (5 nm) observed for the 10% blend with respect to the 5% blend of around 10 nm, when the content of PFH-NT increased from 0.1 to 5% (Fig. 7). This phenomenon is normally produced when the surroundings produce a collapse or π-stacking of the chains and/or could be due to the appearance of a new preferred orientation produced by the planar effect of the naphthalene group in films.³⁹ These results for PFH-NT blends are promising and suggest that they should have influence on the permselectivity of the membrane.

Fig. 7 Red-shift of the emission spectrum versus polyfluorene content.

It is worth noting that there is a clear red-shift of the maximum wavelength for increasing percentages of all polyfluorenes that can be observed in Fig. 4–6 and is shown in Fig. 7. This phenomenon can be attributed to low wavelength quenching as mentioned previously.

3.2. Permeation properties

3.2.1. Permeability and selectivity. The evolution of the permselective properties of the films are certainly influenced by the amount of copolyfluorene in the blend. Fig. 8a shows the CO₂ permeability obtained for the samples.

Fig. 8 CO₂ permeability (a) and CO₂/CH₄ selectivity (b) as a function of the copolyfluorene in the blend. Lines are only for guidance and have been extrapolated asymptotically to 0% of polyfluorene.

Pure polyimide films show lower permeability than the blend membranes with a low content of conjugated polymers PFH-BT and PFH-NT, which is probably due to the voluminous character of the corresponding groups. This is not the case for the blends containing PFH-B that for equally low percentages of copolyfluorene have permeabilities quite similar to those of the pure polyimide. Actually all of the copolymers have similar permeabilities compared to those of pure 6FDA–6FpDA for percentages over 0.1% to 3%. Over this percentage all of the blends give lower permeabilities.

The decrease of permeability for high copolyfluorene content could be attributed to two effects. This type of conjugated polymer containing charge transfer complexes (CTC), produces quenching that can be attributed to the interaction of π electrons of the aromatic rings in the structures of the polymers.⁴⁰ This causes a minimization of charge delocalization along the polymer chain, thus decreasing chain-to-chain repulsion and decreasing both polarity (solubility of polar species) and free volume (diffusivity). Then both solubility and diffusivity would decrease for increasing amounts of polyfluorene. The initial increase of permeability for low polyfluorene content could be attributed to initial repulsion of delocalized charges, with initial increases in free volume and polarity.

Fig. 8b shows the results of CO₂/CH₄ selectivity as a function of the percentage of polymers in the blend. As in the previous case, it is observed that for a content over 2%, the selectivity results for this couple of gases are below those of pure polyimide for the blends containing PFH-BT and PFH-B while selectivities higher than those of the pure polyimide are obtained for up to 5% of copolyfluorene in the case of PFH-NT. The blends containing PFH-BT show selectivities quite similar to those of pure 6FDA–6FpDA for very low percentages of the copolyfluorene. In any case, low percentages of polyfluorene show the best results. The improvement is substantial for example in the case of the blend with 1% of PFH-NT copolymer. These improvements of selectivity, show that there should be some kind of favorable interaction with CO₂, which does not affect CH₄, probably due to their different polarities, hence the selectivity increases.

In summary, the results show that both the permeability and the selectivity improve with respect to the pure polyimide for percentages of polyfluorene around 1% with ulterior deterioration of both of these properties for percentages over 3%. Is also seems clear that the variation of these properties with the percentage of polyfluorene is strongly influenced by the type of polymer conjugate used for the formation of the blend. The highest relative improvement was found for PFH-NT. In this case, permeability improves by more than 23% over that of the initial polymer, while selectivity increases 15% over the pure polyimide results. It is clear that optimal polyfluorene content is around 1%, when the blends show the best balance between permeability and selectivity enhancements. In addition, the percentage of rise depends on the type of polyfluorene.

A visual way of representing the permselective properties obtained for different gases are the Robeson’s representations.^41,42 In these representations the distance to the so-called upper bond gives us an idea of the permeation properties of the samples. In Fig. 9 the Robeson’s plot for the CO₂/CH₄ pair is shown.

Fig. 9 Robeson’s plot for CO₂/CH₄.

3.2.2. Diffusivity and solubility. Fig. 10 shows the results of solubility and diffusivity for the PFH-NT polymer blends. It is seen that diffusivity changes only slightly, with a maximum value for blends containing 0.1% of polyfluorene and decreasing slowly for increasing percentages. In turn, solubility is maximal for 1% polyfluorene in the blend and decreases strongly for increasing contents. This seems to indicate that solubility should play the key role in the decrease of the permeability with increasing polyfluorene content.

Fig. 10 CO₂ solubility and diffusivity as a function of the percentage of PFH-NT in the blends.

3.2.3. Correlation of permeation and fluorescence. In Fig. 11 we present the improvement in permeability and selectivity over those of the pure polyimide as a function of the relative intensity of fluorescence in the solid state, I_S, as compared to that in solution, I_L, for the blends containing PFH-NT (similar results are obtained for the other copolyfluorenes). It seems clear that when fluorescence is not affected by quenching (i.e. when the chain–chain structures induced by delocalized electrons are preserved) the fluorescence intensity in the solid state is more similar to that in solution leading to higher permeability and selectivity. In these conditions the CO₂ interactions with the π electrons of the aromatic rings and the free volume effect of the voluminous polyfluorenes have maximum effects.

Fig. 11 Percentage of increment of CO₂ permeability (a) and CO₂/CH₄ selectivity (b) as a function of fluorescence intensity in the solid state compared to that in liquid state (solid state quenching).

4. Conclusions

A series of new blends have been prepared. These polymer blends have been synthesized by the simple mixture between a polyimide (6FDA–6FpDA) and different conjugated polymers (copolymers copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene] (PFH-B), copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-bencene-1,2,5-thiadiazole] (PFH-BT), and copoly-[9,9-bis(n-hexyl) fluorene-alt-1,4-naphthalen-(2,3)-1,2,5-thiadiazole] (PFH-NT)) in several proportions.

TGA provided information about the stability of the mixtures of the polyimide and the chromophore polymers at the treatment temperature. DSC showed that there was a single T_g for each of the blends, confirming the full miscibility of the blends. It was determined that the T_g values were between the values of the pure polyimide and the polyfluorenes.

The comparison between the intensities of the samples in solid state and polymers in solution gave us information about the state of the polyfluorenes in the films. It was observed that above 5% polyfluorene content the polymers showed a quenching effect, losing effectiveness in fluorescence. The analysis of the maximum showed a red-shift when the percentage of fluorophore polymer in the blend increases, again due to the effect of the interactions with the medium.

The best permselectivity was found for the blends with a low fluorene content. The samples with a content lower than 5% of conjugated polymer showed permeabilities and selectivities better than the pure polyimide. CO₂ permeability reached an improvement of above 23% and CO₂/CH₄ selectivity reached an almost 15% enhancement. Especially good results were obtained for the PFH-NT blends.

The solubility of the samples with a 0.1 and 1% content increased despite the diffusivity remaining almost constant. This could be attributed to the polarity of the copolyfluorene molecules that is high enough to increase the CO₂ solubility but not too high; the copolyfluorene content should be kept low enough so as to allow significant chain-to-chain construction and loss of fluorescence due to quenching.

To sum up, it seems clear that this type of modification by the formation of polymer blends with conjugated polymers is suitable for separations where CO₂ is involved, and especially for CO₂/CH₄ separation. This separation has great importance in the processing of natural gas. The simplicity of blending makes these materials attractive alternatives to the current polymeric materials for these applications.

Acknowledgements

We are indebted to the Spanish Junta de Castilla y León for financing this work through the GR-18 Excellence Group Action and to the Ministry of Science and Innovation in Spain for their economic support of this work (MAT2011-25513, MAT2011-23007). We also acknowledge financial support from the programme Consolider Ingenio 2010 (project CSD-0050-MULTICAT) and from the Junta de Castilla y León (project VA-248U13). The help provided by Sara Rodriguez in measuring gas permeability and selectivity is greatly appreciated.

References

1. F. Ahmad, K. K. Lau, A. M. Shariff and G. Murshid, Comput. Chem. Eng., 2012, 36, 119–128.
2. A. Hart and N. Gnanendran, Energy Procedia, 2009, 1, 697–706.
3. F. G. Kerry, Industrial Gas Handbook: Gas Separation And Purification, CRC Press, 2007.
4. R. W. Baker, Ind. Eng. Chem. Res., 2002, 41, 1393–1411.
5. B. D. Bhide, A. Voskericyan and S. A. Stern, J. Membr. Sci., 1998, 140, 27–49.
6. Y. Xiao, B. T. Low, S. S. Hosseini, T. S. Chung and D. R. Paul, Prog. Polym. Sci., 2009, 34, 561–580.
7. P. Bernardo, E. Drioli and G. Golemme, Ind. Eng. Chem. Res., 2009, 48, 4638–4663.
8. C. A. Scholes, S. E. Kentish and G. W. Stevens, Recent Pat. Chem. Eng., 2008, 52–66.
9. M. I. Bessonov, M. M. Koton, V. V. Kudryavtsev and L. A. Laius, Polyimides: Thermally Stable Polymers, Plenum, New York, USA, 1987.
10. Polyimides: Fundamentals and Applications, ed. M. K. Ghosh and K. L. Mittal, Marcel Dekker, New York, USA, 1996.
11. Polyimides, ed. D. S. Wilson, H. D. Stenzenberger and P. M. Hergenrother, Blackie, Glasgow, UK, 1990.
12. M. Das and W. J. Koros, J. Membr. Sci., 2010, 365, 399–408.
13. A. Tena, L. Fernández, M. Sánchez, L. Palacio, A. E. Lozano, A. Hernández and P. Prádanos, Chem. Eng. Sci., 2010, 65, 2227–2235.
14. R. Wang, C. Cao and T.-S. Chung, J. Membr. Sci., 2002, 198, 259–271.
15. M. Hasegawa and K. Horie, Prog. Polym. Sci., 2001, 26, 259–335.
16. C. A. Hunter and J. K. M. Sanders, The nature of π–π interactions, J. Am. Chem. Soc., 1990, 112(14), 5525–5534.
17. S. E. Wheeler and J. W. G. Bloom, Toward a More Complete Understanding of Noncovalent Interactions Involving Aromatic Rings, J. Phys. Chem. A, 2014, 118, 6133–6147.
18. L. M. Robeson, Polymer Blends: A Comprehensive Review, Hanser Gardener Publications, Cincinnati, OH, 2007.
19. M. Leclerc, J. Polym. Sci., Part A: Polym. Chem., 2001, 39, 2867–2873.
20. U. Scherf and E. J. W. List, Adv. Mater., 2002, 14, 477–487.
21. C. L. Chochos and S. A. Choulis, Prog. Polym. Sci., 2011, 36, 1326–1414.
22. A. C. Grimsdale, K. L. Chan, R. E. Martin, P. G. Jokisz and A. B. Holmes, Chem. Rev., 2009, 109, 897–1091.
23. M. J. Tapia, M. Montserín, A. J. M. Valente, H. D. Burrows and R. Mallavia, Adv. Colloid Interface Sci., 2010, 158, 94–107.
24. C. Zhu, L. Liu, Q. Yang, F. Lv and S. Wang, Chem. Rev., 2012, 112, 4687–4735.
25. B. Liu and G. C. Bazan, Nat. Protoc., 2006, 1, 1698–1702.
26. R. Mallavia, F. Montilla, I. Pastor, P. Velásquez, B. Arredondo, A. L. Álvarez and C. R. Mateo, Macromolecules, 2005, 38, 3185–3192.
27. R. Molina, S. Gómez-Ruiz, F. Montilla, A. Salinas-Castillo, S. Fernández-Arroyo, M. Del Mar Ramos, V. Micol and R. Mallavia, Macromolecules, 2009, 42, 5471–5477.
28. P. Wei, L. Duan, D. Zhang, J. Qiao, L. Wang, R. Wang, G. Dong and Y. Qiu, J. Mater. Chem., 2008, 18, 806–818.
29. J. Yang, C. Jiang, Y. Zhang, R. Yang, W. Yang, Q. Hou and Y. Cao, Macromolecules, 2004, 37, 1211–1218.
30. D. M. Muñoz, J. G. de la Campa, J. de Abajo and A. E. Lozano, Macromolecules, 2007, 40, 8225–8232.
31. A. Tena, A. Marcos-Fernández, A. E. Lozano, J. G. de la Campa, J. de Abajo, L. Palacio, P. Prádanos and A. Hernández, J. Membr. Sci., 2012, 387–388, 54–65.
32. A. Tena, A. Marcos-Fernandez, L. Palacio, P. Cuadrado, P. Pradanos, J. De Abajo, A. E. Lozano and A. Hernandez, Ind. Eng. Chem. Res., 2012, 51, 3766–3775.
33. D. R. Paul and S. Newman, Polymer blends, Academic Press, USA, 1978.
34. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, 2006.
35. Z. Ma, S. Lu, Q.-L. Fan, C.-Y. Qing, Y.-Y. Wang, P. Wang and W. Huang, Polymer, 2006, 47, 7382–7390.
36. J. Chappell, D. G. Lidzey, P. C. Jukes, A. M. Higgins, R. L. Thompson, S. O’Connor, I. Grizzi, R. Fletcher, J. O’Brien, M. Geoghegan and R. A. L. Jones, Nat. Mater., 2003, 2, 616–621.
37. E. B. Namdas, T. D. Anthopoulos, I. D. W. Samuel, M. J. Frampton, S. C. Lo and P. L. Burn, Appl. Phys. Lett., 2005, 86, 1–3.
38. F. B. Dias, S. King, A. P. Monkman, I. I. Perepichka, M. A. Kryuchkov, I. F. Perepichka and M. R. Bryce, J. Phys. Chem. B, 2008, 112, 6557–6566.
39. R. Traiphol, N. Charoenthai, T. Srikhirin, T. Kerdcharoen, T. Osotchan and T. Maturos, Polymer, 2007, 48, 813–826.
40. R. A. Dine-Hart and W. W. Wright, Macromol. Chem., 1972, 153, 237.
41. L. M. Robeson, J. Membr. Sci., 1991, 62, 165–185.
42. L. M. Robeson, J. Membr. Sci., 2008, 320, 390–400.

---

Footnote

† Present address: Helmholtz-Zentrum Geesthacht, Institute of Polymer Research, Polymer Synthesis Dept, Max-Planck-Straße, Geesthacht, Germany.

---