Nanostructured fluids from pluronic® mixtures
Keywords: Pluronic® Micellization Calorimetry Rheology Nanostructure
Micellization and gelation of binary mixtures of EO99PO69EO99 (pluronic® F127) and EO80PO27EO80 (pluronic® F68) in aqueous solutions were investigated by means of micro-differential scanning calorime- try and rheology and for a total copolymer concentration fixed at 20 wt%. The aim of this investigation is to determine the interplay between micellization and macroscopic gelation of the mixed solutions. Micro-DSC reveals the formation of two distinct populations in F127/F68 mixtures during heating and subsequent cooling of the solutions. The enthalpies of micellization of each copolymer and the respective onset temperatures remained constant after mixing indicating the predominance of two independent processes of micellization in the mixtures. The F127 exhibits a crystallization transition, at a distinct temperature which persists, but increases in the mixtures with concentrations higher than 10 wt%. Rheological measurements were performed during heating ramps or after maturation periods versus frequency. They showed two types of gelation transitions: either a steep increase of the storage and the loss moduli, which corresponds to the crystallization temperature of the F127 micelles or a progressive jamming transition when no crystal can form. Maturation process has a major effect on the rheologi- cal properties of the mixed gels, possibly related to local rearrangements of the two micellar phases. This investigation highlights the unique features of the binary pluronic® mixtures, compared to dilution effects of single component aqueous solutions.
1. Introduction
Amphiphilic triblock copolymers with block composi- tions EOn–POm–EOn (where EO refers to ethylene oxide and PO to propylene oxide) are extensively studied. It is well agreed that during heating the rupture of hydrogen bonds of poly(propyleneoxide) (PPO) leads to an increase of hydrophobicity while the poly(ethyleneoxide) (PEO) chains remain hydrophilic (Alexandridis et al., 1994). When the ratio n/m is greater than 0.5 and above a critical micelle concentration (CMC) spherical micelles are formed with a hydrophobic core and a hydrophilic corona (Wanka et al., 1994). These block copolymers are currently pro- duced with various lengths of EO or PO sequences and are known under their trade names of pluronic® or poloxamer. They are useful as biocompatible materials for biomedical and personal care applications mainly because of their gelation properties. Indeed, at high concentrations of copolymers the aqueous solutions exhibit a dramatic change of the viscoelastic moduli and become soft solids or gels. The ability to tune gel formation temperatures for drug release delivery systems has attracted considerable attention.
Use of mixtures of pluronic® block copolymers is considered as a strategy to modulate structural and rheological properties without the need of synthesizing new molecules. Interestingly, we demonstrated in previous work that gelling temperature (Tgel) can be modulated by varying the concentra- tion of EO99PO69EO99 (pluronic® F127) but also by varying the proportion between two pluronics F127 and F68 (EO80PO27EO80) (Aka-Any-Grah et al., 2010). The decrease of Tgel with increasing F127 polymer concentration for pluronics is an expected result. Indeed, when F127 is alone in aqueous solution, its gelation is explained by the organization of the pluronic® micelles to form a highly viscous and elastic gel (Pham Trong et al., 2008). However, we also observed that Tgel was decreased when the proportion of F68 was increased in the formulation composed of F127/F68 (Aka- Any-Grah et al., 2010).
Basically, two opposing views have been supported when mixtures of two pluronics were studied: either cooperative micellization of the two copolymers (mixed micelle formation) through favourable interactions between the unlike components or coexistence of two distinct populations of micelles assembling each component independently.
In the first case, cooperative micellization was reported for mix- tures of two copolymers containing similar corona chain lengths, but core blocks with different chain lengths. This is the case for Newby et al. (2009) who investigated binary mixtures of pluronic® copolymers P85 (EO26PO41EO26) and P123 (EO20PO72EO20). Micelle dimensions were probed by dynamic light scattering. The authors found that the hydrodynamic radius of micelles for the 50/50 mix- ture is larger than that for either P85 or P123 alone, due to the formation of mixed micelles with a higher association number. In another vein, Gaisford et al. (1997) studied binary pluronic® mix- tures selected from a series containing the same ratio of PEO:PPO (70% PEO, 30% PPO) and varying polymeric molecular weights. The pluronic® samples for their study were F77, F87, and F127. Because of differences in molecular weights, each copolymer had a different number of PPO units. Data demonstrated that binary mix- tures 50%/50% of F77/F87 showed cooperative aggregation over the experimental concentration range. These two polymers have sim- ilar core block composition and a PPO:PPO ratio of 0.96. However, binary mixtures 1:1 of F77/F127 and F87/F127 having PPO:PPO ratios of 0.54 and 0.56 respectively showed non-cooperative aggre- gation because the composition of the polymers was markedly different. Therefore, the difference in micellization appears to be PPO dependent: polymers having similar PPO moieties showed cooperative aggregation and those having different PPO moieties non-cooperative binding.
Although the micellization and gelation of individual pluronic® copolymers have been extensively investigated, far fewer studies are known on mixtures of pluronic® F127 and F68 solutions. The aim of this investigation is to determine the interplay between micellization and macroscopic gelation of the mixed solutions of F127 and F68. The solutions have a fixed total concentration of 20 wt% and contain various proportions of these two pluronics. They are fully transparent within the range of temperatures and proportions investigated. The experiments are mainly based on microcalorimetry and rheology. Several aspects which have raised controversies in other systems are examined:
– First of all, whether this system generates mixed or individual micelles when heated.
– How to characterize the elastic gel properties with a rigorous approach?
– What is the gel structure in mixed systems?
Generally speaking, it is important to identify the specific fea- tures associated to the nanostructured fluids, with high volume fractions of micelles, compared to solutions of small molecules or to concentrated suspensions of solid spheres whose behaviour was more systematically reported in the literature.
2. Materials and methods
2.1. Materials
Pluronics F127 (poloxamer P407, EO98PO69EO98) and F68
• For the preparation of hydrogels containing F127 or F68 alone, the pluronic® powder was gradually added under magnetic agitation to cold-water phase contained in an ice bath (0–5 ◦C). The final concentration of pluronic® in the hydrogel was varied as 10, 16, 17 and 20 wt%.
• For the preparation of hydrogels composed of F127 and F68 mix- tures, the final concentration of pluronic® was kept constant (20 wt%) while the proportion of F127/F68 in the preparation was varied as 0/20, 3/17, 5/15, 10/10, 17/3, 18/2, 19/1 and 20/0 wt%. The different preparations were denominated by two numbers indicating the weight percentage of pluronic® F127 and pluronic® F68 respectively.
2.3. Methods of characterization
2.3.1. Micro-DSC
Micro-DSC experiments were carried out with a calorimeter Micro-DSC III Setaram (Caluire, France) (Aka-Any-Grah et al., 2010; Pham Trong et al., 2008). The cells used were batch type (1 mL). Two empty cells with the caps were weighed after complete dry- ing and the joints were chosen to obtain the same mass (±0.2 mg). The reference (distilled water) and the sample were introduced to the cells at room temperature and weighed for the identical mass (±0.3 mg). After their introduction into the oven at room temper- ature, the temperature of the oven was lowered until 5 ◦C with —1 ◦C/min, and then was kept at this constant temperature to stabilize the heat flow. After that, a temperature scan was performed at +0.1 ◦C/min until 70 ◦C. Finally, the samples were kept 1 h at 70 ◦C to balance the heat flow again before the beginning of melting with —0.1 ◦C/min to 5 ◦C.
2.3.2. Rheology
Temperature ramp: all rheological measurements were per- formed on a controlled stress AR-G2 rheometer (TA instruments) using a cone-plate geometry (diameter 6 cm, angle 2◦, gap 54 µm). A solvent trap maintained a solvent-saturated atmosphere around the cell, and evaporation was not significant for the temperatures and the time scales investigated.
2.2. Sample preparation
The solutions of the formulations were prepared using the pre- viously described “cold method” but by using a magnetic stirrer instead of mechanical one (Koffi et al., 2006; Aka-Any-Grah et al., 2010; Bouchemal et al., 2009).
2.3.3. X-ray scattering experiments
X-ray scattering experiments were performed at SOLEIL syn- chrotron (St Aubin, France) on the SWING beamline operated at 11 keV. The scattered intensity was reported as a function of the scattering vector q = 4π sin θ/λ where 2θ is the scattering angle and λ the wavelength of the incident beam. Data were collected by a two-dimensional CCD detector.
Calibration of the q-range was carried out with silver behenate. Intensity values were normalized to account for beam inten- sity, acquisition time and sample transmission. Each powder-like diffraction pattern, displaying a series of concentric rings, was then integrated circularly to yield the intensity as a function of q. The samples were loaded into thin quartz capillaries of 1.5 mm diame- ter (GLAS W. Muller, Berlin, Germany). Samples were thermostated in a microcalorimeter, MICROCALIX, designed to allow simulta- neous DSC and X-ray diffraction measurements, which was placed in the beam pathway.
A secondary weak peak appears at the end of micellization. It corresponds to the crystallization temperature (Tc) (Fig. 1B) where the micelles arrange in a crystalline structure (Pham Trong et al., 2008; Jain et al., 1997). At 60 ◦C, well above Tc = 23 ◦C, X-ray experiments showed that the F127 (20 wt%) micelles arranged in a two-dimensional hexagonal close-packed structure (2D hcp) with a parameter of ~21 nm (see Supporting material).
Few reports are found for the crystallization peak of pluronic® aqueous solutions. Most of authors reported no enthalpic changes after Tendset for pluronic® F127 using classical DSC or a low con- centration of copolymer (Wanka et al., 1994; Jain et al., 1997). Pham Trong et al. (2008) studied the micellization of F127 single of short range order. The mean correlation distance between the micelles was ~13.4 nm. SAXS experiments therefore clearly estab- lish that the single solutions of these pluronics form micelles at concentrations similar to the ones used in Micro-DSC experiments; temperature was above the micellization temperatures determined by the enthalpic traces.
In Fig. 2B two distinct micellization peaks appeared for the mixtures F127/F68 10/10 and 5/15 wt% corresponding to micelle formation of the two copolymers in distinct temperature ranges. The first peak on heating is attributed to F127 and the second to F68 providing direct evidence of separate micelle formation of the two copolymers. Both copolymers can form micelles because they have large enough concentrations in the mixture, whereas only one peak is seen in the mixtures of 19/1, 18/2 and 17/3 wt% corre- sponding to the unique F127 micelle formation or either to a small enthalpic effect (no detectable) for F68. For concentrations 19/1 and 17/3 wt%, increase in F68 concentration shifted the crystalliza- tion temperature of F127 to higher values. SAXS characterization of F127/F68 17/3 wt% showed that at 25 ◦C (Tendset) the correlation distance between the micelles was d ~ 14.3 nm. At 60 ◦C a 2D-hexagonal compact structure was observed. This structure was similar to the one of F127 20 wt% at the same temperature, but the lattice parameter was higher for F127/F68 17/3 wt% ~22 nm.
The crystallization peak of F127 does not appear for F127/F68 mixtures 10/10 and 5/15 wt% because it was too small (the crys- talline fraction decreases with the F127 concentration) or either because the intricate structure containing micelles of two different sizes does not enable crystallization. Neither any ordered structure between micelles was seen in SAXS experiments. The high concen- trations of micelles lead to close packing of micelles rather than to crystal formation. The rheological measurements should be able to point out the difference between these two states (see below).
In Fig. 3 and Table 1, the variation of the characteristic temper- atures obtained from Micro-DSC experiments (Tonset, Tpeak, Tendset,
solutions and showed that a small secondary endothermic peak appeared at the end of micellization for F127 concentration higher than 15 wt%. Lau et al. (2004) reported the micellization and the gelation for aqueous solutions of pluronic® F108 (EO133PO50EO133) with Micro-DSC. They found the secondary small peak for con- centrations higher than 20 wt% and attributed it to the gelation of F108.
In our study, several proportions of F127/F68 mixtures with a fixed total pluronic® concentration 20 wt% were analysed follow- ing the same protocol as for F127 and F68 single solutions (Fig. 2). We can observe from Fig. 2A that the mixtures 19/1, 18/2, and 17/3 wt% had the same enthalpogram that the solution containing only 20 wt% F127 except that the crystallization peak was shifted to higher temperatures and became smaller (Fig. 3A). Only one peak of micellization was observed for these systems indicating that only F127 was able to form micelles because, in these conditions, F68 had concentrations lower than the CMC or that the enthalpic con- tribution on micellization was too small to be detected (see below). In Fig. 2B, F68 20 wt% single solution shows a clear endothermic peak due to micelle formation, but no crystallization peak. The micellization peak is smaller than the F127 at the same concen- tration and broader in temperature.
The results of SAXS experiments show that F127 single solution, 10 wt% at 60 ◦C, forms spherical micelles with radius of ~11 nm. At 60 ◦C, F68 10 wt% solutions are composed of non-interacting spher- ical micelles with radius of ~6.4 nm (see Supporting material). At 60 ◦C, F68 20 wt% pattern displayed a correlation peak indicative and Tc) of F127 micelle formation is reported versus the concentration of F127 in the solution. In binary pluronic® mixtures data shows that the transition temperatures and the enthalpic traces are fully reversible as previously seen for single component solutions, suggesting that the micellizations are for the mixtures investigated in thermodynamic equilibrium versus temperature. For the best of our knowledge, this observation has not been clearly reported so far in the literature.
The major effects due to copolymer mixing are evidenced in Fig. 3. In a single component solution of F127, the micellization temperatures (Tonset, Tpeak and Tendset) steadily increased when the concentration decreased. For instance, between 20 wt% and 10 wt%, the Tonset increased from 12.5 to 18.8 ◦C. In the binary mixtures F127/F68 these temperatures remained constant when “dilution” of F127 occurs through increasing the proportion of F68 (Fig. 3A).
The Tpeak is only slightly shifted upon mixing with F68. This type of “dilution” effect is unexpected. We want to emphasize that these solutions are completely transparent and phase separation normally associated to cloud point is not suspected to occur. The enthalpy measurements are fully reversible, as stated before. The endset temperature in the binary pluronic® mixtures (Fig. 3B) is increased by a few degrees, from 21.3 to 26.4 ◦C, upon mixing with F68. Therefore, mixing of F127 with F68 broadens the range of micellization temperatures of F127 compared to single solutions, but leaves unchanged the onset temperatures.
In solutions of F127, the position of Tc due to micelle crystalliza- tion increases upon addition of F68, from 22.4 to 26.6 ◦C (Fig. 3C) when concentrations of F127 decrease from 20 to 17 wt%. The change for Tc is larger than for Tendset, as shown in Fig. 3B. Upon dilution of F127 single component solutions, the gap between end- set and crystallization temperatures is increased from +1.1 (20 wt%) to +5.1 ◦C (16 wt%). In binary mixtures F127/F68, this effect is much more amplified because Tc changes from 22.4 ◦C to 36.2 ◦C (by +13.8 ◦C) when F127 concentration decreased from 20 to 17 wt%, whereas the Tendset are constant in this range of concentrations (Fig. 3D).
The Tonset of F68 micelle behaves in a similar way to F127; how- ever, because of the overlapping of the micellization peaks of F127 and F68, the onset temperatures are more difficult to locate pre- cisely. The two pluronics in mixed solutions behave as if they keep their own environment when micellization occurs as if concentrations were locally identical to those where the copolymer was dissolved. This suggests that during (or just before) micelle formation seg- regation takes place in the mixtures because each copolymer is surrounded by its own family. The onset of micellization of F127 starts at a temperature close to the dissolution temperature of the batch solution, meaning also that the composition is close to the solubility limit.
3.1.2. Enthalpy of micellization
From the heat flow traces, the area of large peaks can be ascribed to the enthalpy of micellization ∆H for each component indepen- dently, which is in agreement with the suggestion of two separate micellizations. Because of the presence of overlapping peaks in thermal runs, the contribution of each copolymer, in particular for concentrations 10/10 and 5/15 wt% is more difficult to evaluate. We chose to define the baseline by a procedure of interpolation of the measurements far away from peak positions, with five intermedi- ate points leading to a baseline with a monotonic slope.
The enthalpy of micellization of F127 by the total mass in solu- tion as single component or in mixtures is shown in Fig. 4. The enthalpy of micellization normalized by the mass of PPO derived either from the enthalpy peaks of F127 or F68 is shown in Table 2.From these data, the enthalpy of micellization by the mass of F127 in solution was found approximately proportional to the copolymer concentration. From the slope of the linear regression fit of the curve it is deduced that the enthalpy of micellization of F127 in single or mixed solutions is equal to ∆HF127 = 26 ± 1 J/g with a small error bar. Besides, the enthalpy of micellization of F68 was also derived ∆HF68 = 13 ± 2 J/g of F68. The error bar here is larger. The literature reports on several values for micellization enthalpy of F68. Tsui et al. (2008) collected such data for samples from various manufacturers, concentrations and temperatures. Concentrations are either 5 g/L (0.5 w/v%) or 1–5 wt% in their own investigation. Our measurement of enthalpy of micellization per gram of F68 cor- responds to any value between 76 kJ/mol and 162 kJ/mol because there is a large uncertainty on the molecular weight of F68 sam- ple according to the supplier. Furthermore, Borbely and Pedersen (2000) found at 40 ◦C, that a considerable amount of polymer is still in monomer form and at 60 ◦C as aggregation progresses further, most of the chains are in aggregated state, but still there is a small amount of monomers in the solution. The authors also showed that the mean aggregation number increased substantially both with concentration and temperature (from 4 to 50 ◦C), while the poly- dispersity of micelles decreased with the progress of aggregation. Consequently, the F68 sample has a less cooperative micellization transition than F127, therefore inducing an additional spread in enthalpy associated to micellization or aggregation number. In our investigation, the total enthalpy of micellization of F68 appears to be independent on the concentration thus it supposes that within the extended range of the temperature ramp, up to 70 ◦C, almost all copolymers are incorporated into the micellar aggregates and that the enthalpy per unit mass of F68 is constant.
The normalized enthalpies by the mass of PPO can be derived from these measurements ∆H = 96 ± 8 J/g PPO. In the paper by Beezer et al. (1994) it is reported that the enthalpy of micelliza- tion depends both on the PPO molecular weight and of the PEO content. We used in our calculations the molecular weights given by the manufacturer. Our results show, within the experimental precision, a constant value for the enthalpy per gram of PPO.
3.2. Rheological experiments on mixtures
In rheological terms, a gel is defined as a state where the stor- age modulus Gr and the loss modulus Grr are frequency independent and the phase angle δ (tan δ = Grr/Gr) is low at all frequencies (Wei et al., 2002). The inverse tube method is also used as a crude method to discriminate between fluid and solid-like behaviour (Chaibundit et al., 2007). The thermosensitive properties of hydrogels are gen- erally evaluated by sol/gel transition temperature. The oscillatory shear measurements are used mainly in order to determine the sol–gel transition temperature, by identifying the temperature for which Gr undergoes a critical variation and to characterize the gel shear modulus beyond the gel point. Frequency dependence is used to establish the relative contributions of Gr and Grr to the final shear modulus of the gel. The higher the Gr value, the more pronounced the elastic character and conversely, the higher Grr, the more pro- nounced the viscous properties. The elastic modulus is a measure the mixture 10/10 wt% than in 17/3 wt% solution. The rheological “transition” observed in the mixture 10/10 wt% around 50 ◦C cannot be ascribed to the crystal formation of F127 micelles. The sol state is a fluid containing large volume fractions of micelles of the two sizes. Around 50 ◦C, a sudden increase of the moduli is observed, which may be comparable to a jamming transition or formation of a glassy state. The volume occupied by the micelles reaches a criti- cal value, where the flow is strongly affected, similar to the critical volume fraction of particles in hard spheres suspensions (random close packing), when the viscosity increases asymptotically.
All the results of rheology experiments are summarized in Table 1 together with Micro-DSC experiments. In F127 single solu- tions and mixtures with concentrations 18/2, 17/3 and 16/4 wt% the gelation temperature coincides with the crystallization peak of F127 and increases from 22.4 ◦C to 29 ◦C upon mixing with F68. No gelation was observed for 5/15 wt% and for pure F68 20 wt% solutions in the range of temperatures up to 70 ◦C.
Usually, the gelation temperatures of pharmaceutical formula- tions are considered to be suitable if they are lower than human body temperature (37 ◦C). If gelation temperature is higher than 37 ◦C, the gels will lose their elastic properties resulting in a leak- age of the formulations from the site of administration. According to this part of the work, it appears that 17/3 wt% (Tgel = 35.4 ◦C) is a suit- able mixture for pharmaceutical applications because the Tgel was lower than 37 ◦C and higher than room temperature, which allows administration in liquid form. The results presented in Table 1, at F127 concentrations decreasing from 20 to 16 wt% show that the storage modulus Gr of the hydrogel measured just after gelation decreased from 21.6 kPa to 5.7 kPa. In addition, the phase angle (δ) increased from 2.63◦ to 13.6◦. For the same total concentration of pluronic®, addition of F68 to hydrogels resulted in a decrease of the storage modulus and a gain in the loss modulus. The vis- cous contribution is likely to be related to the non-crystalline phase of the energy stored and recovered per cycle of deformation and reflects the solid-like component or elastic behaviour.
3.2.1. Heating at constant rate
The shear moduli and the enthalpy measurements are shown in combined plots in Fig. 5A and B for two pluronic® mixtures, 17/3 wt% and 10/10 wt%, versus temperature with identical ther- mal protocols. The rheological measurements are performed at a fixed frequency 1 Hz. In both cases, Gr was very low in sol state despite micelle formation and increased drastically together with Grr when “gelation” temperature is reached (Koffi et al., 2006; Aka- Any-Grah et al., 2010). In single F127 solutions, the changes of Grr during micelle formation can be related to the volume fraction of the micelles, providing some simple assumptions (Pham Trong et al., 2008). The “gelation” temperature is usually defined as the intersection of the two curves Gr and Grr. In Fig. 5A the small crys- tallization peak obtained from Micro-DSC measurements perfectly matched with the sol–gel transition observed by rheological mea- surements. Therefore, in 17/3 wt% solutions (Fig. 5A) the sol–gel transition temperature Tg coincided with the micelle crystallization temperature Tc of F127 component.
There is no crystallization peak on the heat flow trace for the mixture 10/10 wt% (Fig. 5B); two successive micellization processes for F127 and F68 induced only a slight increase of the loss modu- lus in the thermal range from 5 to 45 ◦C, but the storage modulus is small (<10—1 Pa) and remains constant. At the end of the sec- ond micellization peak, one can see that both Gr and Grr increased sharply and the shear modulus Gr overtook the loss modulus Grr. However, the amplitude of the change of the shear modulus at the end of the “transition” is less important and more progressive in mainly composed by the F68 unimers or micelles. This phase prob- ably occupies a small volume (a few percent of the total volume of the solution). It is also expected that the gain in the loss modu- lus will make the spreading of the hydrogel easier after its in vivo administration, which can be a favourable contribution. 3.2.2. Time resolved experiments and dynamical spectra Rheology is a complementary approach to microcalorimetry for understanding the complex phase diagram of the binary pluronic® mixtures. It is interesting to establish from rheological experiments quantitative or at least qualitative differences between gels with a dominant crystalline phase (such as 17/3 wt% mixture) and those with a bimodal distribution of spheres, like the 10/10 wt% mixture. It is suggested from the discussion of our results, that the binary mixtures near the gel point contain a complex nanostruc- ture. The enthalpy is fully reversible with temperature scans, which is normally characteristic of phase transformations following an equilibrium path. It is important to confirm if rheology also shows time independent states at fixed temperatures, in agreement with microcalorimetry. With this aim in view, we designed rheological experiments which include time resolved experiments and sam- ple characterization steps at fixed temperatures. The samples were first heated under a constant ramp (+0.1 ◦C/min) as before, up to a certain temperature and were kept at this temperature for several hours in the rheometer. Time resolved experiments were performed at a frequency of 1 Hz during periods of 12 h. Two mix- tures were studied with this protocol: 18/2 wt% and 10/10 wt% mixtures with two final temperatures, one below the rheological transition T < Tgel and a second above the transition T > Tgel. The gela- tion temperature for the 18/2% mixture is found towards 29.3 ◦C. Fig. 6A displays the time dependence of the moduli Gr and Grr at 27 ◦C for the mixture 18/2 wt%. It is evidenced that approaching “gelation” temperature, the solutions containing large amounts of the predominantly crystalline phase shows a stable state, with- out any change versus time. The storage modulus is high, ~13 kPa, compared to the storage modulus after 12 h at 27 ◦C (280 Pa). Therefore, the crystalline structure is responsible for both a high modulus at T > Tgel and a stable value during 12 h, compared to the measurements in the concentrated micellar suspension, before crystallization.
Another important feature is provided by the dynamic spectrum of the systems before and after “gelation”. The frequency sweep between 0.002 Hz and 10 Hz can be performed after 12 h of matu- ration at the fixed temperature (Fig. 6B).
Fig. 6B shows the spectra at the two temperatures at 27 ◦C and 37 ◦C after 12 h. At 27 ◦C the spectrum corresponds to a vis- coelastic fluid with a characteristic frequency fc = 0.016 Hz where Gr = Grr = 140 Pa. Any measurement at a frequency above this limit f > fc shows a “solid like” behaviour, as Gr > Grr. In particular, this is the case for f = 1 Hz, which is a frequency often preferred in many experiments. At lower frequencies, below fc = 0.016 Hz, the mate- rial appears “liquid like”, because Gr < Grr. It is expected that both Gr and Grr decrease with the frequency. The characteristic frequency fc is the inverse of a characteristic time: at long times, longer than 50 s = 1/fc (inverse of frequency fc) the system is liquid-like, it can flow, by rearrangements of the micellar structures. On the other hand, as shown in Fig. 6B at 37 ◦C the frequency dependence of the moduli in the crystalline gel state is much less pronounced: the characteristic frequency where Gr = Grr is not appearing in this frequency range. It is probably located one or two decades below (characteristic time between 500 and 5000 s), where the storage and loss moduli seem to converge, meaning that the “crystal” might also flow or creep, but much more slowly than the suspension of micelles randomly distributed. Fig. 6C summarizes the frequency dependence of the tan δ for the 18/2 wt% mixture after 12 h. The loss angle δ in the jammed micellar structure at 27 ◦C is more strongly frequency dependent; large values are seen at low frequencies δ = 53◦ (tan δ = 1.32) show- ing the liquid-like tendency. When δ = 45◦, tan δ = 1 corresponds to Gr = Grr. Finally, in Fig. 7C the tan δ is plotted versus the frequency. Again, the values are very close to each other for the two temperatures; in all cases (frequency range from 0.004 Hz to 10 Hz) tan δ < 1 mean- ing a solid like behaviour. However, tan δ increases rapidly with decreasing frequencies and tends towards limit where tan δ could be >1 at both temperatures which eventually is a liquid like state, with predominantly high loss moduli at very low frequencies (or and temperature dependent transitions are known in glass forming systems. In the pluronic® mixture, the rheological states below and above 51 ◦C are viscoelastic fluids in both cases. Initially during the ramps, it was not possible to measure the whole frequency spec- trum in the “sol” state because the system is not stable. According to well-known behaviour of polymer melts at the glass transition, we may suppose that the characteristic frequency fc is initially high and that it decreases with time, during structural rearrangements. The characteristic frequencies after long maturation periods are close to fc ≈ 0.003 Hz according to Fig. 7B at both temperatures with a char- acteristic time ≈330 s. This characteristic time is however much shorter that the predominantly crystalline state discussed before.
3.3. Micellization and gelation scenarios for F127/F68 mixed solutions
Combining Micro-DSC and rheology experiments allows the micellization and the gelation scenarios for F127/F68 mixed solu- tions to be explained. As presented in Fig. 8, three regions were identified. The following scenario is suggested: at low temper- atures, pluronics F127 and F68 exist in the form of unimers in aqueous solution (region (i)). Staring at low temperatures around 5 ◦C for dissolution of the pluronics and upon heating, first the dehydration of the F127 unimers takes place and they begin to associate together to form micelles (Fig. 8A, region (ii)). The onset micellization temperature of single component F127 solutions, at a concentration of 20 wt% is Tonset = 12.5 ◦C. Micelles are composed of a hydrophobic core formed by the PPO blocks surrounded by the hydrophilic PEO blocks. Their hydrodynamic radius is about 11 nm measured by dynamic light scattering in dilute solutions (Prud’homme et al., 1996) and with our own SAXS measurements at 10 wt% at 60 ◦C.
In single solutions, when the temperature is increased above Tc = 22.4 ◦C, F127 micelles arrange themselves in a crystalline struc- ture (Fig. 8A (iii)). The same scenario is described for F127 solution at concentrations 16 and 17 wt% from the presence of the small secondary endothermic peak.
The micellization of F68 starts at a lower temperature compared to F127, Tonset ≈ 28 ◦C for a concentration of 20 wt% as explained. F68 micelles at 60 ◦C, 10 wt%, have a radius of 6.4 nm measured by SAXS, but according to Borbely (1998), the radius of gyration measured by small angle neutron scattering (SANS) in 3.1 nm. It is not expected that the two radii should be identical as SANS detects mostly the core, whereas SAXS can see the PEO/water interface. The hydrodynamic radius can be compared to the radius determined by SAXS.
F68 is considered as the more hydrophilic component in the mixture, its micellization temperature is higher (Kozlov et al., 2000).
Micro-DSC experiments unambiguously demonstrate that mixed micelles of F127 and F68 are not formed. Initially the pluron- ics were individually dissolved at low temperature (5 ◦C) and upon long times). The structures are still evolving in time for both tem- peratures after several hours.
Rheological measurements performed during a continuous heating ramp do not allow the mixed solutions to reach a time independent state. When we stopped the ramp, we observed that the moduli increased, probably through slow local rearrangements of the structure. This suggests that the “transition” temperature in Fig. 5B depends on the rate of heating of the solutions. Such time mixing of the batch solutions homogeneous, clear, fluid phase form. However, nucleation of the individual micelles starts at the same temperature as in the initial batch solutions. It is well observed that Tonset for F127 remains constant upon mixing of two batch solutions of 20 wt% concentrations, with variable volume fractions. This behaviour is completely different from what is observed by simple dilution of F127 in water. The two populations coexist at high enough temperatures. In dilute solutions of pluronics, it was known that the difference in length of hydrophobic PPO blocks gen- erally prevents mixed micelles formation. Our experiments prove that this is true in a large concentration range.
When equal concentrations of F127 and F68 are dissolved, and the batch solutions with equal volumes are mixed, Fig. 8C, (10/10 wt%) no crystal formation is detected during the tempera- ture scan. The rheological experiments in Fig. 5B exhibit a transition around 51 ◦C which cannot be ascribed to crystal formation and corresponds to the endset of micellization of F68. This complex solution contains equal volumes of nanostructured phases, where almost all the copolymers are in the micellar forms. Rheologi- cal experiments performed below the “transition temperature” show a slow structural relaxation versus time. Above the transition temperature, the time increase of the moduli is limited. Eventu- ally, a similar viscoelastic spectrum of the shear moduli is observed at the two temperatures. The rheology experiments suggest that the gelation “transition” is similar to a glass forming system. The storage and loss moduli of crystalline and non-crystalline gels (Figs. 6B and 7B respectively) differ significantly. The mechanisms underlying the slow structural relaxation before or after gelation may correspond to a slow nucleation of micellar crystals of F127. Extensive development of the crystalline phase should imply a structural arrest, because of the very long characteristic relaxation times of the colloidal crystals. We did not try to detect these changes by Micro-DSC experiments, after long maturation times at fixed temperatures. The sensitivity of the Micro-DSC might not be suffi- ciently high. Small angle X ray diffraction could be an alternative solution to evidence the development of the crystalline nuclei, from random packing of micelles.
In the literature, the glass transition of model suspensions of colloidal spheres was analysed in the early work of Pusey and van Megen (1987) who reported experimental evidence for the structural arrest. Later, glass transition and phase diagrams of strongly interacting binary colloidal mixtures were investigated by diffusion wave spectroscopy by Meller and Stavans (1992). More recently, Williams and van Megen (2001) performed dynamic light- scattering experiments on binary mixtures of hard-sphere colloidal suspensions with a particle size ratio of 0.6. They observed the motions in binary mixtures with total volume fractions between 0.51 and 0.58 and showed that on the introduction of the smaller spheres, the structural relaxation is released. They called this effect the “glass melting”. In their investigation, suspensions of the smaller particles do not crystallize due to their larger poly- dispersity, whereas the larger particles show the crystallization transition expected for hard spheres. The authors reported that “dilution by smaller particles instead of solvent, shifts the posi- tion of the primary maximum to higher scattering vectors” (smaller distances).
In their work, other differences appear in the structure factor compared to a simple “dilution”. In the shift of the main structure factor peak and the enhancement, in excess of that obtained by sim- ple dilution, of the mean-squared amplitude of long-wavelength fluctuations, they see the effects of the “attractive depletion poten- tial of mean force that emerges between the larger particles upon introduction of the smaller particles”. They argue that the size ratio between spheres (0.6) in their case is too large for depletion effects to incur a fluid–fluid phase separation. Biben and Hansen estab- lished theoretically that dense binary mixtures of hard spheres phase separate when the size ratio of their diameters is less than 0.2 and the partial packing fractions of the two species are com- parable. In our micellar mixture, this ratio is close to 0.5, which is similar to the case reported by Williams and van Megen (2001) and Meller and Stavans (1992). In addition, we observed in F127/F68 mixtures, that the presence of the smaller micelles at 60 ◦C (17/3) changes the compact hexagonal phase parameter of F127 crys- talline phase from 21 nm to 22 nm which is a slight increase. The presence of a few small micelles can distort the lattice of the large ones.
4. Summary and conclusions
This investigation suggests that aqueous solutions which con- tain two pluronics F127 and F68 at a total high concentration of 20 wt% form, upon heating, nanostructured phases display- ing a bimodal distribution of micelles. Solutions and gels are totally transparent. The micellization temperatures and the heat traces are fully reversible under heating and subsequent cooling at constant rates of ±0.1 ◦C/min, which is usual for systems in thermodynamic equilibrium. However, the rheological behaviour is much more complex and shows that internal relaxations and slow reorganizations take place in the thermal range preceding or fol- lowing the “gelation” temperatures. In F127 rich mixtures the micelles are able to arrange in crystalline domains whereas in F127/F68 10/10 the gel is similar to a glass forming system which can undergo a continuous reorganization.
In mixtures, unlike single pluronic® solutions, the two micellar phases which coexist at high temperature undergo a continuous reorganization: in the F127 rich phase, the micelles should be able to slowly arrange themselves in crystalline domains; the major- ity phase should form a connected path throughout the solution, while the minority phase forms inclusions of variable sizes. The high elasticity modulus of the crystalline phase (F127 rich) pre- vents the process to arrive to completion, and the nanostructure is arrested in a way similar to a glassy state.Pluronic F-68 Small angle X-ray diffraction studies would be very helpful to reveal the local mechanisms in future studies.