R. Craveiroa,
M. Martinsbc,
G. B. Santosa,
N. Correiaad,
M. Dionísioa,
S. Barreirosa,
A. R. C. Duartebc,
R. L. Reisbc and
A. Paiva*a
aREQUIMTE/CQFB, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal. E-mail: alexandre.paiva@fct.unl.pt; Fax: + 351 212 948 385; Tel: +351 212 949 681
b3B's Research Group-Biomaterials, Biodegradable and Biomimetic, University of Minho, Headquarters of the European Institute of Excellence on Tissue Engineering and Regenerative Medicine, Avepark 4806-909 Taipas, Guimarães, Portugal
cICVS/3B's PT Government Associated Laboratory, Braga, Guimarães, Portugal
dUnité Matériaux et Transformation (UMET), UFR de Physique, UMR CNRS 8207, BAT P5, Université Lille 1, 59655 Villeneuve d'Ascq, France
First published on 27th March 2014
In this work, blends of starch and poly-ε-caprolactone (PCL) doped with different concentrations of 1-butyl-3-methylimidazolium acetate ([BMIM]Ac) or 1-butyl-3-methylimidazolium chloride ([BMIM]Cl) were studied. The blends were characterized by mechanical analysis, infra-red spectroscopy (FTIR), differential scanning calorimetry (DSC) and dielectric relaxation spectroscopy (DRS), evaluating the IL doping effect. The samples were subjected to supercritical carbon dioxide foaming and the morphology of the structures was assessed. DSC shows a single glass transition and melting endotherm for foamed and unfoamed samples, having no effect upon IL doping, and DRS shows increased molecular mobility for blends with higher IL concentrations, and some hindrance for lower ones. The conductivity for SPCL doped with 30% [BMIM]Cl, before and after foaming, is comparable to the conductivity of the IL but exhibits more stable conductivity values, opening doors for applications as self-supported conductive materials.
Supercritical fluid foaming (SCF foaming) constitutes an alternative to conventional processes to design and processing 3D structures with high porosity, interconnectivity and uniform distribution. This technology has many advantages, for example, avoiding the use of organic solvents and high temperature and decrease saturation time compared with conventional methods. Polymeric foams are used in a wide variety of applications from the automotive to the chemical industry or biotechnology.
It is clear that ILs constitute a generation of solvents which have an enormous potential as “green” recyclable substitutes to organic solvents. Furthermore, they can improve the properties of the materials by conferring them different characteristics, such as conductivity. It has been reported the use of ionic liquids coupled with biopolymers, such as gelatin, to form efficient solid electrolytes to be used in solid-state batteries.20 Therefore by doping IL in natural-based polymers it is possible to create self-supported conductive materials. Further treatment by SCF foaming, may increase the efficiency of the final material.
In this work, we will focus our attention in imidazolium-based ionic liquids (IBILs) composed by alkylimidazolium as a cation, paired with a counter-ion (chlorine and acetate). The effect of IL doping in the mechanical, thermal and conductivity characteristics of the natural-based polymer, starch/poly-ε-caprolactone (SPCL), was studied. The effect of SCF floaming in the properties of SPCL and doped SPCL was also efficiency.
Sample | IL | IL concentration (wt%) | Designation |
---|---|---|---|
SPCL | — | — | SPCL |
[BMIM]Ac | 10 | SPCL10Ac | |
30 | SPCL30Ac | ||
[BMIM]Cl | 10 | SPCL10Cl | |
30 | SPCL30Cl | ||
SPCL + SCF foaming | — | — | SPCLF |
[BMIM]Ac | 10 | SPCL10AcF | |
30 | SPCL30AcF | ||
[BMIM]Cl | 10 | SPCL10ClF | |
30 | SPCL30ClF |
Compression testing was carried out at a crosshead speed of 2 mm min−1, until a maximum reduction in sample weight of 60%. The compressive modulus is defined as the initial linear modulus on the stress/strain curves. The data presented is the result of the average of at least five measurements.
In tensile mode, the dimensions of the specimens used were 60 mm of length, 1 mm width and 3 mm of thickness. The load was placed midway between the supports with a span (L) of 30 mm. The crosshead speed was 1:
5 mm min−1. For each condition the specimens were loaded until core break.
Sample | Compression | Tensile | |
---|---|---|---|
E (MPa) | E (MPa) | Elongation at break (%) | |
SPCL | 6.2 ± 0.7 | 352 ± 44.6 | 5.7 ± 1.4 |
SPCL10Cl | 2.2 ± 0.3 | 114 ± 35.8 | 5.0 ± 1.1 |
SPCL30Cl | 0.8 ± 0.3 | 104 ± 26.4 | 4.7 ± 0.5 |
SPCL10Ac | 5.2 ± 0.5 | 223 ± 29.3 | 8.9 ± 1.9 |
SPCL30Ac | 2.6 ± 0.4 | 133 ± 53.3 | 8.0 ± 2.1 |
![]() | ||
Fig. 1 Effect of ionic liquid and ionic liquid concentration of the Young modulus of the matrices in (a) compressive and (b) tensile mode. |
The results presented demonstrate that the Young modulus is statistically different in the case of the raw material and in the mixture with IL. This was observed for both compressive and tensile tests, which indicates that the use of ionic liquid has indeed an effect on the polymeric chain mobility. The compressive tests, showed in SPCL with 30% of [BMIM]Cl the Young modulus is lower than the other samples (Table 2). Lowest concentration of [BMIM]Ac seems not to affect the mechanical properties of the matrix prepared. The tensile tests revealed higher elongation at break for SPCL with [BMIM]Ac in comparison with both unmodified samples and those containing [Bmim]Cl, which is another indication of stronger interactions between these IL and the polymer.
The compressive and tensile tests were carried out at a different crosshead speeds to assure that crystalline morphology does not depends on the stimulus to which the samples are being submitted to; no differences in the compression modulus calculated were observed.
The cross-section of the samples was analyzed by scanning electron microscopy (SEM) and the morphological parameters of the samples were determined by micro-computed tomography. Fig. 2 shows the cross-section images of the structures after supercritical fluid foaming. These images showed in presence of ionic liquid there is an improvement on porosity of the blends which is dependent of ionic liquid concentration used. Moreover, the blends doped with 30% of [Bmim]Ac, seemly present a pore size higher than the blends doped with 30% of [Bmim]Cl (Table 3).
Sample | ST | Morphological parameters |
---|---|---|
0 | Porosity | |
SPCL | 0.2 ± 0.3 | |
SPCL10Cl | 6.2 ± 3.3 | |
SPCL30Cl | 3.8 ± 1.4 | |
SPCL10Ac | 2.4 ± 0.7 | |
SPCL30Ac | 2.1 ± 0.9 | |
SPCLF | 1 | 14.3 ± 1.9 |
SPCL10ClF | 20.0 ± 3.4 | |
SPCL30ClF | 38.8 ± 1.0 | |
SPCL10AcF | 32.8 ± 4.5 | |
SPCL30AcF | 40.9 ± 4.8 |
The supercritical fluid foaming technique relies on the ability of CO2 to decrease the glass transition temperature of the polymer, plasticizing it.22,23 In the case of SPCL, without the presence of IL the foaming process produces matrices with a much lower porosity when compared with the ones doped with ionic liquid. On the other hand the presence of [BMIM]Ac and [BMIM]Cl greatly enhance the formation of a porous structure in order to 40%. A number of different works reported in the literature refer the high solubility of carbon dioxide in imidazolium-based ionic liquids.24 A previous work has demonstrated that the enhancement of the solubility of carbon dioxide in polymeric matrices doped with IL promotes the foaming of the structures.19 Therefore, we suggest that the foaming ability is in this case related with the higher or lower solubility of carbon dioxide in either [BMIM]Ac or [BMIM]Cl. In this work we evaluate the effect of the anion on the foaming ability as well as the effect of the concentration of the IL in the polymeric matrix.
Material | Wavenumber (cm−1) | Characteristic peak | Ref. |
---|---|---|---|
Starch | 1080; 1362–1338 | C–O stretching | 26 |
1413 | O–H bending | ||
1456 | CH2 bending | ||
925 | CH2 rocking | ||
Poly-ε-caprolactone | 961 | CH2 rocking | 26 |
1366 | C–O stretching | ||
1418 | O–H stretching | ||
1471 | C–H bending | ||
[Bmim]Ac | 1173 | C![]() |
27 |
1379 | C–O acetate carboxylate | ||
1578 | C![]() |
||
[Bmim]Cl | 1173 | C![]() |
The analysis of the spectra reveals a characteristic peak of the C–O–H groups of starch at 1080 cm−1 when IL is present, which is valid for both ILs studied, [BMIM]Cl and [BMIM]Ac.25,10 This peak confirms the occurrence of interactions between the polymer and IL. Furthermore, is known that ionic liquids are extremely hydroscopic and can easily catch water from air, nevertheless the water content of the samples was always less than 3%. Thus, another indication of these interactions is the decrease on the intensity of this peak, when ionic liquid is present at higher concentration which is more evident in case of [Bmim]Ac.
Hanke et al. reported that each ion pair of the ionic liquid forms complex interactions with the polymer.28 In the case of imidazolium-based ILs the ions will interact, more specifically with the C–O–H groups of starch. While the anion interacts with hydrogen, the cation will interact with the oxygen atom. These interactions will decrease the ability of the polymeric matrix to form hydrogen bonds and promoting hereafter a higher mobility of the polymer chains.
The thermogram collected upon the subsequent heating run (run 2) is dominated by an endothermic transition at ca. 60 °C for all samples. The SPCL blend without ionic liquid also exhibits an endothermic peak in the same temperature region, so this peak is attributed to SPCL melting. Moreover, its temperature location is in close agreement with the value reported for poly-ε-caprolactone,29 so it should be due to the melting of the PCL constituent. Upon further cooling (run 3), all the samples crystallize around 30 °C, as seen by the emergence of a sharp exothermic peak in the thermogram. Subsequent melting occurs again around below 60 °C (Tm of 57 °C estimated in the minimum of the melting peak for SPCL, Fig. 4(c)).
The temperature of the minimum of the melting is slightly shifted towards lower temperatures for the blends containing ionic liquid, which could be related with relatively lower spherulites size and/or spherulites thickness in these materials30 or even less dense crystalline regions; this can be due to the incorporation of the ionic liquid that interferes with the chain packing, as defects.
The effect of water content is observed in Fig. 4(d). With the removal of water the Tg increases about 15 °C. This is due to the plasticizing effect of water.
Furthermore, a close inspection in the low temperature region reveals a discontinuity in the heat flux typical of the Tg, evident in both cooling and heating runs. To allow a better comparison with the results reported for samples submitted to foam, all the samples were annealed at −70 °C. The glass transition that emerges in subsequent heating run becomes enhanced due to an accentuation in the heat flux jump; Fig. 4(b) is a scale-up of the Tg observed upon heating for the annealed samples (the curves were vertically displaced for better comparison).
The simultaneous exhibition of melting and glass transition, occurring on amorphous regions, indicates that all samples are semi-crystalline.
The Tg of the SPCL matrix itself, −64.05 °C determined from the midpoint, does not seem to undergo significant changes upon IL doping. The Tg value is in good accordance with the Tg provided for PCL in literature.29 Therefore the thermal behavior of the SPCL blends is mainly determined by the presence of PCL constituent, as seen by the agreement between both Tg and Tm values of PCL and SPCL.
No other heat flux step is observed in the SPCL + IL thermograms that could be assigned to the glass transition of ILs; the glass transition in the first heating scan is located under −90 °C for [BMIM]Cl and around −74 °C for [BMIM]Ac (Fig. 4(d)). This can be interpreted as no “rich-like” regions of IL exist, at least, until the limit of the DSC detection. Moreover, for the ionic liquids, upon water removal it is obvious a shift of the glass transition to higher temperatures as evidenced in the thermograms taken before and after heating up to 200 °C for [BMIM]Ac, where a deviation of ca. 10 °C occurs (Fig. 4(d)); the effect is much more pronounced for [BMIM]Cl as will be shown by DRS in the next section. The same effect is not observed for the glass transition detected at temperatures around −60 °C in the SPCL + IL samples, confirming that it is associated with the PCL constituent rather than to the ionic liquid.
Although FTIR-ATR analysis confirms IL-matrix interactions, those are not evident from DSC analysis. In fact, the infrared spectra reveals that the C–OH vibration is altered in the presence of IL showing that the IL anion interacts with the C–OH moiety of the starch component. Since DSC for the SPCL + IL composite is blind to starch, and the only transformations are originated by PCL, it is not surprising that no changes in the thermogram occur. Moreover, the invariance in the glass transition detected in both SPCL and SPCL + IL blends, can be interpreted as the IL presence not interfering with the cooperative mechanism that is in the origin of the Tg of SPCL. In spite of some decrease in melting temperature of the SPCL + IL, meaning that crystallization is slightly affected by the incorporation of the IL, after thermal treatment by heating at 200 °C, neither crystallization nor further melting appear to be influenced by the presence of IL.
DSC thermograms of samples submitted to SCF foaming (Fig. 5) were also collected under a similar thermal treatment, the only difference being the lowest temperature attained upon cooling (−90 °C for foamed samples instead of −150 °C).
As before, the thermograms are dominated by the melting and crystallization peaks. The endothermic broad band emerging at ca. 100–120 °C is related to the evaporation of residual water, being more evident for SPCL10AcF and SPCL30ClF.
A close inspection of the low temperature region seems to reveal the typical discontinuity of the heat flux associated with the glass transition, although ill-defined. To enhance this transition, samples that were submitted to SCF foaming were submitted to annealing at −70 °C, after which a temperature scan was carried out. The obtained thermograms are included in Fig. 5(b) confirming the presence of the Tg at ca. −63 °C. As observed previously for the samples not submitted to foaming, the glass transition region does not undergo significant changes between the different materials.
A detailed analysis of the melting region by analyzing Tm taken at the mid-point, reveals a slight shift of the minimum towards lower temperatures for SPCL10AcF and SPCL30ClF; the higher Tm occurs near 62.2 °C (SPCLF) while the lower at 58.2 °C SPCL30ClF. These alterations were already observed for the unfoamed materials. After the first heating scan (run 2) all the samples undergo crystallization near 30 °C (run 3). Both crystallization and further melting occur at temperatures very near to which was observed previously for the unfoamed materials. Therefore, this seems to be caused by the presence of the ionic liquid rather than by foaming.
The initial water content of the samples was measured and with values between 1.9% and 3.0%. Naturally that it is expected that the water present in the samples will have a slight influence in the conductivity measurements. Nevertheless the influence of water content in conductivity is expected to be relatively constant in all samples.
The property under measurement is the complex dielectric function,
ε*(ω) = ε′(ω) − iε′′(ω) | (1) |
σ*(ω) = iωε0ε*(ω) | (2) |
M*(ω) = 1/ε*(ω) | (3) |
Since the propagation of mobile charge carriers also contribute to the complex function,37 it can be advantageous to analyze the dielectric response through the complex conductivity function,
σ*(ω) = σ′(ω) + iσ′′(ω) | (4) |
To get an insight on dipolar polarization due to reorientational motions of permanent dipoles, the imaginary part of the complex electric modulus, M*(ω)
M*(ω) = M′(ω) + iM′′(ω) | (5) |
![]() | (6) |
Spectra of real conductivity, σ′(f) of SPCL and SPCL + IL were collected in order to evaluate the effect of the IL presence. Fig. 6 shows the conductivity isotherms at −80 °C and for 0 °C respectively.
As seen previously by DSC, SPCL is a semi-crystalline material and, at −80 °C, it is below its glass transition temperature. This means that at this temperature no appreciable cooperative motions of the matrix exist, which may explain why no significant conductivity (σ′) is noticed for SPCL with no exhibition of a plateau due to dc conductivity. The same is observed for SPCL + [BMIM]Ac samples since at this temperature, [BMIM]Ac is below its Tg. By other side, at −80 °C, [BMIM]Cl is above its Tg and the conductivity in SPCL10Cl and SPCL30Cl is higher, even showing a conductivity plateau meaning that dc conductivity takes place. Therefore, in the [BMIM]Cl doped samples, the conductivity is dominated by the IL response increasing with its concentration. This means that the charge transport mechanism, in which conductivity is based, is not correlated or even enabled by the segmental motion of the polymer matrix itself.
At 0 °C, both SPCL matrix and ILs are above their glass transitions (confirmed by DSC) which reflects in a conductivity enhancement also in samples doped with [BMIM]Ac, nevertheless, it is clear that [BMIM]Cl has a more significant effect in the material conductivity. According to the work of Ramesh et al.38 the doping of IL in polymer matrices weakens the interactions within the starch polymer chains via H-bonds; this was confirmed in the present work for SPCL + IL by FTIR-ATR analysis. The increase in conductivity is a consequence of the higher polymer flexibility and enhanced segmental mobility, which facilitates ionic transport, providing more conductive pathways.38
It is interesting to note that, at 0 °C, the conductivity of SPCL30Cl is very close to the pure ionic liquids with the advantage of exhibiting more stable conductivity values overall the tested frequency range; the measurements for the ionic liquids are repeatedly affected by electrical interferences. The SPCL matrix, although not contributing significantly to conductivity, acts as a solid support to the ionic liquid allowing obtaining a self-sustained conductive material. This opens doors to different applications of the produced materials.
An identical behavior is observed in the isothermal curves collected by DRS experiments, at both −80 °C and 0 °C, for the samples that were submitted to SCF foaming (Fig. 6(c) and (d)). This can be taken as a further evidence of the interconnected structure of the porous matrix, otherwise the conductivity pathways would be interrupted disabling the dc conductivity and decreasing the measured σ′.
An alternative way to compare the dielectric response of the different materials is through the modulus which temperature dependence of the imaginary part (M′′(T)) for all the tested samples is presented in Fig. 7(a) at 1 kHz; data taken from the dielectric spectra measured isothermally.
It is important to note that the dielectric response of a material to an alternating electrical field has contributions from dipolar relaxation, i.e., reorientational polarization of permanent dipoles and from conductivity, besides interfacial polarization. The study of the dielectric response of a material trough the dielectric modulus is advantageous since it suppresses the electrode polarization effect, due to the blocking of charges in the electrode–dielectric sample interface, and facilitates the identification of dipolar relaxation phenomenon.37 The dipolar contribution emerges at the lower temperatures while conductivity is felt at the higher temperatures (corresponding to the lowest frequencies in the isothermal dielectric spectra).
In Fig. 7(a) the peaks that are observed in the temperature region from −100 to −25 °C are due to dipolar relaxation from which it is possible to infer about the molecular mobility that originates the dynamical glass transition, going further on the analysis relative to DSC. In this plot, subtle differences arise which were not seen in the calorimetric measurements. First, the maximum of the M′′(T) peak obtained for [BMIM]Cl is located at the lowest temperatures. In Fig. 7(c) the modulus is plotted at a lower frequency, 100 Hz, in which the maximum emerges around −100 °C confirming its low Tg and the reason why it was not detected by calorimetry in runs starting at −90 °C; with the water removal (heating up to 120 °C) it significantly shifts to higher temperatures up to ca. −70 °C, undergoing a stronger effect upon drying relatively to [BMIM]Ac, included also in Fig. 7(c), and as seen previously by calorimetry.
Secondly, it is possible to observe that, in the first heating run, the blends with 10% of IL either [BMIM]Cl or [BMIM]Ac, are slightly shifted to higher temperatures relatively to the SPCL polymer blend and to the neat IL, indicating some mobility hindrance. This can be caused by the interaction of the ionic liquid with the starch moiety which slows down the dipoles reorientational motions. On the other hand, the composites containing 30% of IL are slightly shifted to lower temperatures denoting some molecular mobility enhancement. This can be attributed to water that interacts mainly with the IL; the plasticizing effect of water in the glass transition of the ionic liquids was calorimetrically demonstrated for [BMIM]Ac (Fig. 4(d)) and now for [BMIM]Cl by a clear shift of Tg to higher temperatures upon water removal. Therefore, some care must be taken on evaluating the effect of addition/blending ionic liquids with polymers in the case of high IL hydrophilicity.
At the highest temperatures, the dielectric response is mainly determined by the conductivity contribution. This is shown in Fig. 7(b) for SPCL and foamed SPCL where the dielectric response increases significantly on approaching melting, clear seen as a discontinuity in M′′(T) ca. 60 °C; the region of the glass transition emerges as a well-defined peak for both materials. It is possible to observe that, besides intensity, caused by thickness differences between the tested samples (the foamed ones being thicker), the overall dielectric behavior is essentially the same for both materials as the calorimetric measurements have previously shown as well.
FTIR-ATR spectroscopy has shown some molecular interactions between the ionic liquids used and the polymeric blend. Mechanical properties have also demonstrated the same effect.
The calorimetric analysis of the SPCL blends doped with ionic liquids revealed a glass transition and melting at temperatures, respectively Tg and Tm, close to the homologous transitions in poly-ε-caprolactone (PCL), with no phase transformation assigned to the starch constituent. In the SPCL + IL samples, melting occurs at slightly lower temperatures relative to SPCL which could indicate that the incorporated ionic liquid impairs a so closely packing in the crystalline PCL phase. The glass transition undergoes no effect upon IL addition. This goes against to what is described as a usual effect in literature, where the addition of low molecular weight compounds, including ILs, could lower the glass transition temperature due to a plasticizer effect. A possible reason for this behavior could be due to a greater influence of the IL over the starch constituent to which DSC is insensitive. The interaction of the IL mainly with starch was confirmed by FTIR analysis. Plasticization was instead proved for water over the ILs, with a stronger effect for [BMIM]Cl.
The dielectric measurements revealed some subtle differences in the mobility of the different materials. It confirmed the lower Tg for [BMIM]Cl compared with all tested materials, and a mobility enhancement relatively to SPCL blend for SPCL doped with the higher IL concentrations. Oppositely, for blends mixed with the lower IL contents, some hindrance in the mobility was observed probably due to the interaction of ILs dipolar moieties with starch.
In what concerns the conductivity, an appropriate comparison should take in account the glass transition of the IL additive and the water content. It was observed that conductivity of SPCL + IL depends mostly on the IL without being enabled by the segmental motion of the matrix. For SPCL doped with 30% [BMIM]Cl the conductivity was comparable to the one of the neat IL, without being affected by the IL electrical interferences upon measuring, which opens up novel potential applications as self-supported conductive system.
Scanning electron microscopy (SEM) and micro-computerized tomography (micro-CT) revealed some morphological changes, upon SCF foaming, concerning porosity. In presence of [BMIM]Cl, the morphological characteristics seem to improve and materials with a higher porosity were obtained. It is postulated, hereafter that, the increase in porosity in the presence of IL may be due to the high solubility of carbon dioxide in the IL which favors the foaming process.
The increase in porosity after the foaming process increases the interconnectivity of the porous matrix thus allowing for a continuous conductivity pathway.
This journal is © The Royal Society of Chemistry 2014 |