Article pubs.acs.org/Macromolecules
Synthesis and Rheological Behavior of Supramolecular Ionic Networks Based on Citric Acid and Aliphatic Diamines M. Ali Aboudzadeh,† M. Eugenia Muñoz,‡ Antxon Santamaría,‡ M. José Fernández-Berridi,‡ Lourdes Irusta,‡ and David Mecerreyes*,†,§ †
POLYMAT, University of the Basque Country UPV/EHU, Joxe Mari Korta Center, Avda. Tolosa 72, 20018 Donostia-San Sebastian, Spain ‡ POLYMAT and Polymer Science and Technology Department, Faculty of Chemistry, University of the Basque Country UPV-EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia-San Sebastián, Spain § Ikerbasque, Basque Foundation for Science, E-48011 Bilbao, Spain ABSTRACT: Novel supramolecular ionic networks were obtained by reacting citric acid and aliphatic diamines. A proton transfer reaction takes place between the carboxylic acid of citric acid and the amine group leading to the corresponding ionic carboxylate and quaternary ammonium groups. By this method, a series of supramolecular ionic networks were obtained due to the multiple ionic interactions between the corresponding citrate and diammonium molecules as observed by FTIR spectroscopy. Rheological analysis of the ionic networks was carried out considering frequency and temperature sweeps in small-amplitude oscillatory flow and viscous measurements in continuous flow. At low temperatures and/or high frequencies the ionic interactions brought about an elastic network or gel which vanished at high temperatures and/or low frequencies. The viscoelastic behavior was governed by a single relaxation time and a very high plateau modulus, Gp = 5 × 106 Pa. The relaxation time showed an Arrhenius-like dependency with temperature, leading to draw diagrams of the physical states for each sample. The obtained supramolecular ionic networks based on different aliphatic diamine molecules did not show differences in their respective solid and liquid states. However, the frequency-dependent network−liquid transition temperature, Tnl, varied with the chemical nature of the diamines. The higher Tnl (45 °C) was found for the system that contains 1,3-diaminopropane which is attributed to stronger ionic bonds involving primary amines, with respect to ionic bonds with tertiary amines (between −1 and 32 °C). Comparing ionic networks obtained from different tertiary diamines, such as tetramethyl-1,3-propanediamine and tetraethyl-1,3-propanediamine, the lower Tnl was observed in the latter, ascribed to a higher mobility of the aliphatic pendant groups.
■
INTRODUCTION The introduction of new cations and anions coming from ionic liquids chemistry is extending the properties and classical applications of polyelectrolytes1 to other fields such as energy,2 analytical chemistry, gas membranes,3 or smart surfaces.4 These new polyelectrolytes are being named polymeric ionic liquids or poly(ionic liquids) in analogy to their monomeric constituents. This emerging ionic chemistry has also been translated to the development of supramolecular polymer assemblies such as anisotropic ion conducting films,5 block copolymers and interpenetrating polymer networks,6 supramolecular ion gels,7 or self-organized polythiophenes.8 Recently, Grinstaff et al. designed supramolecular ionic networks using multicationic and multianionic compounds.9−12 Interestingly in these works, supramolecular networks were demonstrated using ionic bonds as an alternative to hydrogen bonds as the reversible noncovalent links in supramolecular networks composed of weakly bonded small molecules.13 Whereas, these pioneering reports made use of complex ionic liquid molecules such as alkyl phosphonium dications, very recently we reported the facile synthesis of supramolecular ionic networks by using commercially available di- or trifunctional amines and carboxylic acids.14 Among the investigated molecules, citric © 2012 American Chemical Society
acid indicated a great potential leading to supramolecular ionic networks with unique self-healing and ionic conductivity properties. The viscoelastic properties of thermoreversibe supramolecular polymer gels based on hydrogen bonding have been extensively investigated in recent years.15−20 However, little effort has been devoted to investigate the viscoelastic response of a network formed through electrostatic interactions. Therefore, although certain similitudes between the viscoelastic response of supramolecular reversible networks based on hydrogen bonding and those based on ionic interactions can be expected, the fact is that experimental work on the latter is an urgent demand. To our knowledge this paper is the first work that deepens on the viscoelastic properties of a supramolecular ionic network based only on ionically bonded small molecules. Thus, the first goal of this paper is to investigate the effect of the chemical nature of the diamine and the strength of ionic interactions on the rheological properties of the supramolecular ionic networks. A second goal in this Received: May 14, 2012 Revised: August 29, 2012 Published: September 11, 2012 7599
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
Scheme 1. Proposed Reaction Pathway for the Synthesis of Supramolecular Ionic Polymers Based on Citric Acid and Diamines
atmosphere. The proton transfer reaction was carried out for 30 min. Next, the mixture was placed under vacuum at 80 °C to remove the solvent. A transparent solid was obtained. 1H NMR (D2O): δ 2.27 (s, 3H, CH3), 2.83 (t, 4H, CH2), 2.88 (s, 12H, CH3), 3.31 (t, 4H, CH2), 2.65 (d, 2H, CH2), 2.73 (d, 2H, CH2). 13C NMR (D2O): δ 180.56 (COOH), 177.24 (COOH), 74.96 (C−OH), 54.56 (CH2−N), 51.89 (CH2−N), 45.06 (CH2−COOH), 43.33 (N−CH3), and 39.97 (N− CH3). FTIR: 2974, 1714, 1594, 1469, 1390, 1179 cm−1. Synthesis of Ionic Network Based on Citrate and 1,3Propanediammonium. Citric acid (19.212 g, 0.1 mol) and 1,3propanediamine (0.7412 g, 0.1 mol) were simultaneously added to a round-bottom flask under stirring with methanol in it as a solvent without a nitrogen atmosphere. The proton transfer reaction was carried out for 30 min. Next, the mixture was placed under vacuum at 80 °C to remove the solvent. A transparent solid was obtained. 1H NMR (D2O): δ 2.10 (m, 2H, CH3), 3.14 (t, 4H, CH2), 2.69 (d, 2H, CH2), 2.79 (d, 2H, CH2). 13C NMR (D2O): δ 180.58 (COOH), 177.33 (COOH), 74.92 (C−OH), 45.16 (CH2−COOH), 36.97 (CH2−N), and 25.20 (R2−CH2). FTIR: 3432, 3063, 1715, 1577, 1388 cm−1. Control Experiment, Using a Triol Instead of Citric Acid. Glycerol (9.209 g, 0.1 mol) and N,N,N′,N′-tetramethyl-1,3-propanediamine (13.023 g, 0.1 mol) were simultaneously added to a roundbottom flask under stirring in the absence of a solvent and nitrogen atmosphere. The reaction was left to stir for about 2 h. General Procedure for Rheological Measurements. The rheological measurements were performed on a Thermo-Haake Rheostress I viscoelastometer, using oscillatory and continuous flow tests. Tests were repeated twice observing a very good repeatability of the results. Considering that the systems could be eventually sensitive to water, comparative experiments were carried out with and without nitrogen atmosphere, observing that the results were exactly the same. The testing protocol consisted of the follwoing: (1) Angular frequency sweeps (from 0.0628 to 62.8 rad/s) at constant strain amplitude in isothermal experiments at the temperatures reflected in the text. The strain value γ = 0.005 su was low enough to ensure that both moduli G′ and G″ were obtained in the linear viscoelastic regime. (2) Temperature sweeps to determine G′ and G″ on heating and cooling cycles at an angular frequency of 6.28 rad/s and temperature intervals defined for each sample. The temperature ramp rate was 1 °C/min on heating and cooling cycles. For citrate−pentamethyldiethylenetriammonium three isochrone tests were performed at respective angular frequencies of 0.628, 6.28, and 62.8 rad/s. (3) Viscosity measurements in continuous steady state flow at different temperatures, covering a maximum shear rate interval of 0.01−1000 1/ s. Cycles of increasing shear rate followed by decreasing shear rate were applied.
paper is to report a synthetic method to allow citric acid to be used in the design of supramolecular networks. Historically, citric acid was isolated from citrus fruits, but it is currently industrially produced by fermentation. Citric acid is a global product that is manufactured in over 20 countries. It is mainly used for its flavoring and buffering properties by the food industry, metal-chelating ability in the petroleum, industrial cleaning, and cosmetics industries. Interestingly, and related to the increasing use of natural products in polymer industry, significant efforts have been carried out in the past years to incorporate citric acid into conventional polymers such as polyesters, polyurethanes, or polyacrylates.21−23 Here we will determine the properties and design rules of new supramolecular ionic polymer networks based on citric acid.
■
EXPERIMENTAL SECTION
General. All chemicals were purchased from Aldrich as highest purity grade and used without further purification. NMR spectra were recorded on a Bruker AC-500 spectrometer (for 1H and 13C, at 500 and 125 MHz, respectively). Fourier transform infrared spectra (FTIR) were measured in a Nicolet 6700 spectrometer. The rheological measurements were performed on a Thermo-Haake Rheostress I viscoelastometer, using oscillatory and continuous flow tests. Synthesis of Ionic Network Based on Citrate and N,N,N′,N′Tetramethyl-1,3-propanediammonium. Citric acid (19.212 g, 0.1 mol) and N,N,N′,N′-tetramethyl-1,3-propanediamine (13.023 g, 0.1 mol) were simultaneously added to a round-bottom flask under stirring with methanol in it as a solvent without a nitrogen atmosphere. The proton transfer reaction was carried out for 30 min. Next, the mixture was placed under vacuum at 80 °C to remove the solvent. A transparent solid was obtained. 1H NMR (D2O): δ 2.9 (s, 12H, CH3), 2.21 (m, 2H, CH2), 3.2 (t, 4H, CH2), 2.69 (d, 2H, CH2), 2.76 (d, 2H, CH2). 13C NMR (D2O): δ 180.02 (COOH), 176.17 (COOH), 74.80 (C−OH), 54.45 (CH2−N), 44.83 (CH2−COOH), 43.18 (CH3−N), and 20.13 (R2−CH2). FTIR: 2938, 2678, 2479, 1711, 1597, 1472, 1368, 1223 cm−1. In order to investigate the effect of mole ratio on the final properties, this ionic network was also prepared in two other mole ratios of citric acid to N,N,N′,N′-tetramethyl-1,3-propanediamine, which were 1:1.25 and 1:1.5. Synthesis of Ionic Network Based on Citrate and N,N,N′,N′Tetraethyl-1,3-propanediammonium. Citric acid (19.212 g, 0.1 mol) and N,N,N′,N′-tetraethyl-1,3-propanediamine (18.634 g, 0.1 mol) were simultaneously added to a round-bottom flask under stirring with methanol in it as a solvent without a nitrogen atmosphere. The proton transfer reaction was carried out for 30 min. Next, the mixture was placed under vacuum at 80 °C to remove the solvent. A transparent viscous liquid was obtained. 1H NMR (D2O): δ 1.41 (t, 12H, CH3), 2.21 (m, 2H, CH2), 3.27 (t, 4H, CH2−N), 3.27 (q, 8H, CH2−N) 2.70 (d, 2H, CH2), 2.86 (d, 2H, CH2). 13C NMR (D2O): δ 180.02 (COOH), 177 (COOH), 74.69 (C−OH), 48.75 (CH2−N), 47.76 (CH2−N), 44.93 (CH2−COOH), 19.14 (R2−CH2), and 8.51 (R−CH3). FTIR: 2925, 2654, 2479, 1711, 1593, 1480, 1388, 1051 cm−1. Synthesis of Ionic Network Based on Citrate and N,N,N′,N″,N″-Pentamethyldiethylenetriammonium. Citric acid (19.212 g, 0.1 mol) and N,N,N′,N″,N″-pentamethyldiethylenetriamine (17.730 g, 0.1 mol) were simultaneously added to a round-bottom flask under stirring with methanol in it as a solvent without a nitrogen
■
RESULTS Synthesis and Characterization of Supramolecular Ionic Networks Based on Citric Acid and Diamines. The basic chemical reaction used for the synthesis of our ionic supramolecular polymers is the proton transfer reaction used in the synthesis of protic ionic liquids also known as acid−base complex formation. Protic ionic liquids are easily synthesized by the neutralization and subsequent proton transfer between a Brønsted acid such as citric acid and a base such as a diamine. In our case, we evaluated citric acid in the role of Brønsted acid and different multifunctional amines such as N-tetramethyl-1,37600
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
propanediamine, N-tetraethyl-1,3-propanediamine, N-penthamethyldiethylenetriamine, and 1,3-propanediamine in the role of Brønsted bases. In all the cases, the reactants (citric acid and diamine) were mixed at room temperature by simultaneous addition in an equimolar amount in the presence of a solvent such as methanol to eliminate the effects of the exothermic neutralization reactions. The proton transfer reactions were fast and by simple evaporation of the solvent a series of products were readily synthesized according to Scheme 1. It is worth to remark that this synthetic method has the advantage that does not need any previous ionic monomer synthetic step but just the reacting/mixing of commercially available citric acid and amine molecules. The possibility of undergoing hydrogen bonding in these systems was considered, and a control experiment with the equivalent triol (glycerol), instead of citric acid, with 1,3-propanediamine in mole ratio 1:1, was carried out. As a result, a very low viscosity liquid was obtained, indicating the small influence of the hydrogen bonding as compared to the ionic interactions. Whereas citric acid is a white crystalline solid and the diamines were liquid at room temperature, all the reaction products were soft amorphous solids at room temperature. The products were analyzed by FTIR spectroscopy. It is worth remarking that the infrared spectral features of these compounds differ from those of the pure citric acid, which responds to its dimeric form in the solid state (Figure 1).
the central carboxylic group behaves in a different way to those terminal groups.24 In the case of the product obtained by reacting the citric acid with the corresponding diamine, this region displays mainly two bands at 1705 and 1597 cm−1, which can be attributed to hydrogen-bonded carbonyl groups and carboxylate anions, respectively. The coexistence of these two bands can be explained by the following. First, the excess of carboxylic groups in the citric acid molecule with respect to the diamine can generate different hydrogen-bonding structures because of their interactions with both hydroxyl and acid OH groups. Second, proton transfer processes between terminal carboxylic groups and nitrogen from the amines leading to carboxylate moieties. In addition, the presence of a new band at about 1400 cm−1, attributable to the carboxylate symmetric stretching vibration, and a complex band in the region between 2800 and 2600 cm−1, assigned to quaternary ammonium groups, confirms the proton transfer reactions, as proposed in Scheme 1. Rheological Characterization. Rheological analysis of the ionic networks were carried out considering frequency and temperature sweeps in small-amplitude oscillatory flow and viscous measurements in continuous flow. At low temperatures and/or high frequencies the ionic interactions brought about an elastic network or gel which vanished at high temperatures and/or low frequencies. As a representative example, Figure 2
Figure 2. Dynamic moduli G′ (filled symbols) and G″ (empty symbols) as a function of frequency, (left) below and (right) above the network−liquid transition for supramolecular ionic polymer based on citrate and 1,3-diammoniumpropane. Figure 1. FTIR spectrum of (top) citric acid and (bottom) supramolecular ionic network based on citrate N,N,N′,N′-tetramethyl-1,3-propanediammonium.
shows the dynamic viscoelastic functions G′ (storage moduli) and G″ (loss viscous moduli) at different temperatures for an ionic network composed by citrate and 1,3-diammoniumpropane. A similar rheological behavior was observed for the rest of the products. At low temperatures (Figure 2, left) the existence of an elastic network or gel, characterized by no effect of frequency on both moduli and by G′ > G″, was denoted, whereas at high temperatures (Figure 2, right) the typical response of viscoelastic liquids was observed. Given this viscoelastic behavior, we assume the existence of a transient network originated by the ionic interactions explained above. The transition temperature from the network to the liquid state, Tnl (network−liquid transition), is obtained from the variation of the dynamic moduli with increasing temperature, taken at a frequency of 6.28 rad/s (1 Hz). The transition temperature from the network to the liquid state, Tnl, is defined
Therefore, the different OH stretching modes observed can be attributed to central hydroxyl groups, (3498 cm−1) central carboxylic OH (3445 cm−1), and terminal hydrogen-bonded carboxylic OH moieties (3290 cm−1). However, all these bands disappear when the supramolecular ionic molecules are formed and a broad OH stretching vibration band (3400 cm−1) is observed instead, indicating the generation of new hydrogen bonds, whose average strength is weaker than those of the dimer citric acid. Moreover, the carbonyl stretching region of citric acid is characterized by two well-resolved bands that can be assigned to “free (1750 cm−1)” and “associated (1700 cm−1)” carboxylic carbonyl groups as a consequence of the dimer nature where 7601
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
by the maximum in G″, which coincides with the temperature at the crossing point G′ = G″ as shown in Figure 3.
Table 1. Chemical Compositions and Obtained Viscoelastic Resultsa
a Network−liquid transition temperature (Tnl) (at ω = 6.28 rad/s), activation energy (Ea), relaxation time (τ) (at Tnl + 20 °C), and plateau modulus Gp.
temperatures of the different supramolecular ionic networks Interestingly enough, in contrast to the terminal viscoelastic behavior of ionic liquids, which implies multiple relaxation times,25 the dynamic viscoelastic functions of our supramolecular ionic polymers at high temperatures are very well fitted to the most simple model of the linear viscoelasticity: the Maxwell model with a single relaxation time.26 The following equations are deduced from this model:
Figure 3. Evolution of the dynamic moduli G′ (filled symbols) and G″ (empty symbols) with increasing temperature at a constant frequency of 6.28 rad/s for three different supramolecular ionic networks.
Figure 3 shows the evolution of the dynamic moduli G′ and G″ with increasing temperature at a constant frequency of 6.28 rad/s for three different supramolecular ionic polymers. The adequate temperature range to detect the most significant viscoelastic change was selected for each sample. In all the cases, at low temperatures the ionic interactions brought about a soft solid characterized by a high G′ values (>106 Pa) which vanished at high temperatures to a viscoelastic liquid with G″ ≫ G′ values. As can be seen, the network−liquid transition varies from −1 °C for the case of the network formed by citrate with ethyl-substituted tertiary diamine such as tetraethyl-1,3propanediamonium to 32 °C for the one formed by citrate and methyl-substituted diamine analogue. The highest transition temperature was observed for the ionic network based on citrate and primary 1,3-diammoniumpropane. It is worth to remark that changing the citrate/diamonium mole ratio from 1:1 to 1:1.25 or to 1:1.5 did not change significantly the values and shape of the rheology curves and Tnl = 34 °C and Tnl = 33 °C obtained for the mole ratios 1:1.25 and 1:1.5, respectively. DSC analysis of the thermal transitions have not been considered so far in the literature for systems similar to ours. Significantly, among extensively investigated thermoreversibe supramolecular polymer gels based on hydrogen bonding (see refs 15−20), only in the paper of Van Gemert et al.19 an allusion is made to DSC results, just to report that the glass transition temperature being Tg =28 °C the system had to be plasticized to obtain soft materials. Our DSC results obtained for the network based on citrate and N,N,N′,N′-tetramethyl1,3- propanediammonium indicated the presence of a transition at T = 2 °C, clearly below the network−liquid transition of this systems which was 32 °C. Notwithstanding an analysis of the thermal properties of our systems is out of the scope of this paper, we may hypothesize that the transition we detected by DSC is the glass transition of the ionic network. Table 1 shows the numeric values obtained from the analysis of the rheological curves for the network−liquid transition
G′(ω) =
G″(ω) =
Gp(ωτ )2 1 + (ωτ )2
(1)
Gpωτ 1 + (ωτ )2
(2)
Examples of the obtained fittings are shown in Figure 4. The relaxation time τ is found from the inverse of the frequency at which G′ equals G″ because G′/G″ = 1 = ωτ . This coincides with the maximum of G″ because the analytical calculation on eq 2 gives that the maximum should take place at ωτ = 1.
Figure 4. Dynamic moduli G′ (filled symbols) and G″ (empty symbols) as a function of frequency fitted to Maxwell model: (left) supramolecular ionic polymer formed by an stoichiometric mixture of citric acid and N,N,N′,N′-tetramethyl-1,3-propanediamine at 40 °C and (right) supramolecular ionic polymer formed by an stoichiometric mixture of citric acid and 1,3-diaminopropane at 50 °C. Dotted lines are Maxwell fitting. 7602
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
The effect of temperature on the relaxation time of the Maxwell model for all the samples is shown in Figure 5. To
Figure 6. Variation of the dynamic moduli with temperature at the indicate frequencies for citrate−N-pentamethyldiethylenetriammonium supramolecular ionic network.
⎛ E ⎞ ω = ω0 exp⎜ − a ⎟ ⎝ RTnl ⎠
Figure 5. Effect of temperature on the relaxation time for three samples. Lines correspond to the fit of the relaxation times to an Arrhenius-like equation, with Ea values shown in Table 1: (a) citrate N,N,N′,N′-tetramethyl-1,3-propanediammonium; (b) citrate and 1,3propanediammonium; (c) citrate N,N,N′,N′-tetraethyl-1,3-propanediammonium.
where the activation energy is Ea = 155 000 J/mol. This value is close to the activation energy obtained in frequency scans, Ea = 189 500 J/mol (Table 1). These results constitute an exponent of the validity of the time−temperature equivalence in the investigated system and reveal a single-exponential dependence using two different rheological procedures, such as frequency scans and temperatures scans. A significant aspect of these ionic networks is the observed rapid thermal reversibility, since the same G′ and G″ values are obtained on continuous heating and cooling. This is compatible with a basic feature of supramolecular polymers, also called reversible or equilibrium polymers,28 which can reversibly break and re-form rapidly by noncovalent interactions. Figure 7 shows G′ and G″ data of both cooling and heating cycles, showing the thermal reversibility for an ionic network based on citrate and N,N,N,N′-tetramethyl-1,3-propanediammonium.
obtain the lines of this figure, data were adjusted to Vogel− Fulcher−Tamman−Hesse equation,26 observing that the best fits were obtained with a characterisitic temperature T0 = 0 K, which actually leads to an Arrhenius-like dependency (τ = A exp Ea/RT). The corresponding activation energies are presented in Table 1, together with other characteristic parameters of the viscoelastic response at high temperatures. The plots of Figure 5 constitute actually maps that delimitate the network and liquid states. At experimental times larger than the relaxation time (at the right of the curves), viscous flow sets in because at these time scales molecular associations are able to break and recover, allowing diffusion and flow. But at experimental times shorter than the lifetime of molecular associations, the system behaves like an elastic network or gel. Therefore, the transition from the network gel state to the liquid state depends on both temperature and time scale. Oscillatory flow experiments at low frequencies imply large experimental times, which, as can be deduced from Figure 5, bring about a liquid behavior at lower temperatures than when the oscillatory experiments are carried out at high frequencies. This shift of the network−liquid transition to high temperatures as the frequency increases, which responds to the time−temperature equivalence observed in homogeneous polymers,26 is verified in Figure 6 for the pentamethyldiethylenetriammonium sample. In this figure the variation of the dynamic moduli with temperature, taken at respective constant frequencies of 62.8, 6.28, and 0.628 rad/s, is presented. As in Figure 3, in Figure 6 the network−liquid transition temperature, Tnl, is defined by the maximum in the loss modulus that coincides with the temperature at the crossing point G′ = G″. The corresponding Tnl are 27 °C for a frequency of 62.8 rad/s, 19 °C for 6.28 rad/s, and 10 °C for 0.628 rad/s. Similarly to the dependence of the glass transition on frequency which has been shown to follow an Arrhenius-like equation,27 the effect of frequency on our Tnl is given by
Figure 7. Thermal reversibility. Temperature sweep of G′ (filled symbols) and G″ (empty symbols): 20 to 60 °C (heating cycle) and 60 to 20 °C (cooling cycle +) for ionic network based on citrate and N,N,N′,N′-tetramethyl-1,3-propanediammonium, measured at ω = 6.28 rad/s and a strain of 0.5%. 7603
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
Consequently, the interpretation of our rheological results lies on the association−dissociation processes of these ionic interactions. The dynamic viscoelastic data, in particular the reversibility and repetitiveness in temperature scan tests, show that reversible networks, susceptible to temperature and frequency/time effects, are formed. In Scheme 2, a graphical
This experiment was sistematically carried out for other combinations, and the same behavior was obtained. The remarked results, together with the existence of a single relaxation time in the Maxwell model, associated with a singleexponential stress, suggest that the dynamics of our supramolecular polymers is governed by the formation and breakage of the ionic interactions detected by FTIR spectroscopy. The existence of another mechanism of deformation and flow would imply a deviation from the simple Maxwell model and more than one plateau in the storage modulus G′(ω), each corresponding frequency being associated with the time constants of one of the temporary cross-links.18,29,30 The continuous flow experiments shown in Figure 8 reveal a shear independent viscosity (called Newtonian or linear
Scheme 2. Proposed Representation of the Temperatureand Time-Dependent States of a Supramolecular Ionic Network Based on Citrate and Citric Acid and NPentamethyldiethylenetriaminea
Figure 8. Continuous flow experiments for three samples at 60 °C. Cycles of increasing shear rate followed by decreasing shear rate showed a reversible (nonhysteresis) behavior. The choice of the shear rate ranges was decided according to best measuring conditions, which in turn were determined by the viscosity of each sample.
a
At low temperatures or short times (high frequencies) the ionic interaction associations hold on consitituting a network, whereas at high temperatures or long experimental times (low frequencies) the associations are broken.
representation of the reversible cross-linking process is presented. At low temperatures or for short time tests (high frequencies) the ionic interactions bring about a network which vanishes at high temperatures or long experimental times (low frequencies). The transition from the network to liquid state is determined by the variation of the relaxation time with temperature, as is shown in Figure 5. Interestingly, our rheological results indicate a simpler mechanism than the generally accepted dynamics of supramolecular polymer networks as governed by two different characteristic time scales: (i) the time scale of formation and breakage of supramolecular associations and (ii) the time scale for the relaxation of chains by reptation due to entanglements.18 Actually, we do not have any result that would justify the existence of two relaxation mechanisms. The presence of two relaxation times should lead to two plateaus in the elastic modulus: the classical plateau of the entanglement network and a second plateau due to the additional reversible ionic interaction. A double plateau has been reported, for instance, by Craig et al.29 for metallopincer-cross-linked poly(4-vinylpyridine). In our case, however, the viscoelastic data are well fitted to a Maxwell model with a single relaxation time and a well-defined plateau at high frequencies, which leads us to discard the existence of more than one relaxation process. This absence of entanglements is also expected from the molecular nature of our supramolecular network building blocks compared to the polymeric ones used in those previous works.
viscosity), except at very high shear rates the citrate−Npentamethyldiethylenetriammonium supramolecular ionic network. No hysteresis effect (neither tixotropy nor rheopexy) was observed, as the viscosity curves were identical in shear rate increasing and decreasing processes. This linear rheological behavior suggests that the density of ionic interactions, which dissociate and reassociate when an external macroscopic deformation is applied, remains constant during the flow test. The process is similar to entanglement dynamics in conventional polymer melts31 where Newtonian viscosity is observed because the entanglement density remains the same as the applied shear rates correspond to experimental times larger than the relaxation time. Within this context, the shear thinning behavior observed in our supramolecular polymers at very high shear rates (Figure 8) should be a consequence of a reduction in the density of ionic interactions. Increasing shear rate implies experimental times shorter than the relaxation time, so the observed shear thinning behavior at high shear rates responds to broken associations which are not able to recover due to the very short experimental times.
■
DISCUSSION The analysis of the supramolecular ionic networks by NMR and FTIR spectroscopies reveals that carboxylate and ammonium molecules are formed which anounced the existence of weak ionic interactions between the anionic and cationic groups. 7604
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
ammounium bands appears at lower wavelengths, which indicates that the ionic bonds involving primary amines are stronger than ionic bonds with tertiary amines, which may cause the observed higher network−liquid transition temperature. As can be seen in Table 1, besides the same plateau modulus (Gp) values found for the four combinations, the relaxation times in the liquid state, at the same temperature (T = Tnl + 20) above the network−liquid transition (see Table 1), are also similar. Therefore, the investigated samples do not show differences in their respective solid and liquid states. Consequently, the specificity of each diamine is only reflected in the transition from the network to the liquid state. Seeing this result, the comparison with the glass transition of polymers seems pertinent, since the molecular characteristics of the monomer affect to a great extent the glass transition temperature, but much less to the values of the elastic modulus in the glassy state.
A distinguishing feature of the dynamic viscoelastic results of our supramolecular polymers is the high values of the plateau modulus, Gp = 5 × 106 Pa (Table 1), considerably higher than any reported elastic modulus associated with entanglements29 or reversible interactions (H-bonding), as far as we know. Assuming that the elasticity is dominated by the entropy associated with the conformational distribution between ionic interactions, we can apply the equation of the rubber elasticity Gp = ρRT/Mi, which correlates the plateau modulus and the molar weight, Mi, between elastically effective interactions, to our Gp data; the values of Mi for all the samples are around 550 g/mol. This value is lower than the smallest value of the entanglement molecular weight Mi found for polymers.31 This result leads to assume that the observed plateau modulus at high frequencies is not due to entangled physical interactions. Besides the described general viscoelastic behavior of these new supramolecular polymers, some preliminary conclusions can be extracted on the relation between the chemical nature of the molecular constituents and the physical properties. On the one hand, as remarked above, all the systems give rise to the same plateau modulus (5 × 106 Pa) at high frequencies, which indicates that the density of effective interactions or temporary cross-links is the same and, therefore, independent of the amine used to prepare the supramolecular polymers. However, the results of the relaxation time as a function of temperature establish a significant difference between the macromolecules obtained using tetraethyl-1,3-propanediamine and the rest of the samples. As seen in Figure 5, the combination which contains tetraethyl-1,3-propanediamine brings about a shift of the relaxation times to lower temperatures. This is compatible with the lower temperature for the transition from network to liquid taken at a frequency of 6.28 rad/s (Figure 3 and Table 1), as compared with the other samples. The dependence of the relaxation time with temperature is also smaller for this system, as can be seen comparing the activation energies shown in Table 1. From general results on reversible polymer gel and networks, we know that the elastic modulus, Gp, depends basically on the number of temporary cross-links, although the energy necessary to break them, triggering gel−liquid transition, depends on their strength. This would lead to suggest that weaker links are involved in the case of the samples prepared with tetraethyl-1,3-propanediamine. But ionic interactions are the same for tertiary amines (tetramethyl-1,3propanediamine, tetraethyl-1,3-propanediamine, and pentamethyldiethylenetriamine), whereas only tetraethyl-1,3-propanediamine presents a significantly lower Tnl temperature. A possible explanation arises from the comparison with liquid crystals with side-chain mesogenic units.32,33 In the latter systems increasing the length of the flexible group which links the mesogenic unit to the main chain produces a decrease in the anisotropic−isotropic transition temperature because mobility is higher. Similarly, in the case of tetraethyl-1,3propanediamine the length of the unit linked to nitrogen is larger than for the other samples (ethyl units face to methyl units). This can cause more strain upon the temporary crosslinks, favoring its breakage and reducing the relaxation time and the network−liquid transition temperature, as observed in Figure 5 and Table 1. In a more specific analysis of the results, i.e., comparing each combination with the others, we also remark the higher Tnl found for the system that contains 1,3-diaminopropane. In this case the explanation may lay on the difference in the strength of the ionic interaction. As observed by FTIR, the carboxylate and
■
CONCLUSION By a simple acid−base reaction between citric acid and commercially avalaible diamines, novel supramolecular ionic polymers were obtained. As shown by FTIR analysis, a proton transfer reaction between the carboxylic acid of citric acid and the amine takes place, leading to the corresponding ionic carboxylate and quaternary ammonium groups. The dynamic viscoelastic results denote the existence of a reversible elastic network at low temperatures, which turns into a viscoelastic liquid at high temperatures. It was found that a simple Maxwell model (with one relaxation time) can be used to fit the viscoelastic data of the studied samples at different temperatures, which suggests that the dynamics of our supramolecular polymers is solely governed by the ionic interactions. This hypothesis that leads to discard any effect of entanglements is supported by the presence of only one elastic plateau. The value of this plateau modulus, Gp = 5 × 106 Pa, is the same for all the samples and considerably higher than any reported elastic modulus associated with entanglements. On the other hand, in the liquid state, at the same equivalent temperature with respect to network−liquid transition T = Tnl + 20, all the samples give a similar relaxation time value τ = 0.02 s. Thus, it can be stated that all the investigated supramolecular networks showed a similar behavior in their respective solid and liquid states.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The financial support of the Spanish Ministerio de Ciencia e Innovación through projects MAT2010-16171 and MAT201127993, the University of the Basque Country UPV/EHU (UFI11/56), and the Basque Government through SUPRAPOL Saiotek project is acknowledged.
■
REFERENCES
(1) (a) Mecerreyes, D. Prog. Polym. Sci. 2011, 36, 1629−1648. (b) Yian, J.; Antonietti, M. Polymer 2011, 52 (7), 1469−1482. (c) Lu, J.; Yan, F.; Texter, J. Prog. Polym. Sci. 2009, 34, 431−448. 7605
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606
Macromolecules
Article
(32) Finkelmann, H. In Polymer Liquid Crystals; Ciferri, A., Krigbaum, W. R., Meyer, R. B., Eds.; Academic Press: New York, 1982. (33) Lipatov, Yu. S.; Tsukruk, V. V.; Shilov, V. V. J. Macromol. Sci., Part C: Polym. Rev. 1984, 24, 173−23.
(2) (a) Shaplov, A. S.; Lozinskaya, E. I.; Vygodskii, Y. S. In Electrochemical Properties and Applications of Ionic Liquids; Torriero, A. A. J., Shiddiky, M. J. A., Eds.; Nova Science Publishers Inc.: New York, 2011; Chapter 9. (b) Appetechi, G. B.; Kim, G. T.; Montanina, M.; Carewska, M.; Marcilla, R.; Mecerreyes, D.; De Meatza, I. Power Sources 2010, 195, 3668−3675. (3) (a) Bara, J. E.; Camper, D. E.; Gin, D. L.; Noble, R. Acc. Chem. Res. 2010, 43, 152−159. (b) Ho, T. D.; Canestraro, A. J.; Anderson, J. L. Anal. Chim. Acta 2011, 695 (1), 18−43. (4) Döbbelin, M.; Arias, G.; Loinaz, I.; Llarena, I.; Mecerreyes, D.; Moya, S. Macromol. Rapid Commun. 2008, 29, 871−875. (5) Yoshio, M.; Mukai, T.; Ohno, Y.; Kato, T. J. Am. Chem. Soc. 2004, 126, 994−995. (6) (a) Vijayakrishna, K.; Jewrajka, S. K.; Ruiz, A.; Marcilla, R.; Pomposo, J. A.; Mecerreyes, D.; Taton, D.; Gnanou, Y. Macromolecules 2008, 41, 6299−6308. (b) Noro, A.; Ishihara, K.; Matsushita, Y. Macromolecules 2011, 44, 6241−6244. (c) Becht, G. A.; Sofos, M.; Seifert, S.; Firestone, M. A. Macromolecules 2011, 44, 1421−1428. (7) He, Y.; Li, Z.; Simone, P.; Lodge, T. P. J. Am. Chem. Soc. 2006, 128 (8), 2745−2750. (8) (a) Dobbelin, M.; Tena-Zaera, R.; Marcilla, R.; Iturri, J.; Moya, S.; Pomposo, J. A.; Mecerreyes, D. Adv. Funct. Mater. 2009, 19, 3326− 3333. (b) Lee, S.; Becht, G. A.; Lee, B.; Burns, C. T.; Firestone, M. A. Adv. Funct. Mater. 2010, 20, 2063−2070. (9) Godeau, G.; Navailles, L.; Nallet, F.; Lin, X.; McIntosh, T. J.; Grinstaff, M. W. Macromolecules 2012, 45, 2509−2513. (10) Wathier, M.; Grinstaff, M. W. Macromolecules 2010, 43, 9529− 9533. (11) Wathier, M.; Grinstaff, M. W. J. Am. Chem. Soc. 2008, 130, 9648−9649. (12) Craig, S. Angew. Chem., Int. Ed. 2009, 48, 2645−2647. (13) Bosman, A. W.; Sijbesma, R. P.; Meijer, E. W. Mater. Today 2004, 4, 34−39. (14) Aboudzadeh, M. A.; Muñoz, M. E.; Santamaría, A.; Marcilla, R.; Mecerreyes, D. Macromol. Rapid Commun. 2012, 33, 314−318. (15) Noro, A.; Matsushita, Y.; Lodge, T. P. Macromolecules 2008, 41, 5839−5844. (16) Noro, A.; Matsushita, Y.; Lodge, T. P. Macromolecules 2009, 42, 5802−5810. (17) Nair, K. P.; Breedveld, V.; Weck, M. Macromolecules 2008, 41, 3429−3438. (18) Seiffert, S.; Sprakel, J. Chem. Soc. Rev. 2012, 41, 909−930. (19) Van Gemert, G. M. L.; Peeters, J. W.; Söntjens, S. H. M.; Janssen, H. M.; Bosman, A. W. Macromol. Chem. Phys. 2012, 213, 234−242. (20) Lei, Y.; Lodge, T. P. Soft Matter 2012, 8, 2110−2120. (21) Doll, K. M.; Shogren, R. L.; Willett, J. L.; Swift, G. J. Polym. Sci., Part A: Polym. Chem. 2006, 14, 4259−4267. (22) Noordover, B. A. J.; Duchateau, R.; Van Benthem, R. A. T. M.; Ming, W.; Koning, C. E. Biomacromolecules 2007, 8, 3860−3870. (23) Bednarz, S.; Lukasiewicz, M.; Mazela, W.; Pajda, M.; Kasprzyk, W. J. Appl. Polym. Sci. 2011, 119, 3511−3520. (24) Bichara, L. C.; Lanus, H. E.; Ferrer, E. G.; Gramajo, M. B.; Brandán, S. A. Adv. Phys. Chem. 2011, 347072 DOI: 10.1155/2011/ 347072. (25) Pogodina, N. V.; Nowak, M.; Läuger, J.; Klein, C. O.; Wilhelm, M.; Friedrich, C. J. Rheol. 2011, 55, 241−256. (26) Ferry, J. D. Viscoelastic Properties of Polymers, 3rd ed.; Wiley: New York, 1980. (27) Murayama, T. Dynamic Mechanical Analysis of Polymeric Material; Elsevier Scientific Publishing Co.: Amsterdam, 1978. (28) Vermonden, T.; Van Steenbergen, M. J.; Besseling, N. A. M.; Marcelis, A. T. M.; Hennink, W. E.; Sudholter, E. J. R.; Cohen Stuart, M. A. J. Am. Chem. Soc. 2004, 126, 15802−15808. (29) Serpe, M. J.; Craig, S. L. Langmuir 2007, 23, 1626−1634. (30) Jongschaap, R. J. J.; Wientjes, R. H. W.; Duits, M. H. G.; Mellema, J. Macromolecules 2001, 34, 1031−1038. (31) Graessley, W. W. Polymeric Liquids & Networks: Dynamics and Rheology; Taylor & Francis, Inc.: London, 2004. 7606
dx.doi.org/10.1021/ma300966m | Macromolecules 2012, 45, 7599−7606